SCRIP: Scholarly Research In Progress 2021 by GeisingerCollege

Volume 5 • November 2021

Scholarly Research In Progress

Table of contents 2

Geographic Variability in Antibiotic Prescribing Rates in Medicaid

Alexia G. Aguilar, Priscilla C. Canals, Kimberly A. Miller, and Brian J. Piper

Declining National Codeine Distribution in United States Hospitals and Pharmacies

Preventability Associated with Elevated Maternal Mortalities Among Black Women Colleen G. Jordan, Sophia A. Klevan, Kendra C. Benn-Francis, Ofonime E. Emah, and Amy L. Kennalley

Recent Trends in Gabapentin Usage Among Medicaid Patients Christopher Logan SanCraint and Joshua P. Mills

Assessment and Improvement of Sepsis Bundle Compliance at Geisinger Community Medical Center Sean B. Watson, Michael S. Pheasan, Steven A. Picozzo, John R. Wroblewski, Jeffrey D. Perluke, and Igor Georgievsky

Conflicts of Interest Differ Among Male and Female Pediatric Journal Authors Rebecca L. Petlansky, Amadea D. Bekoe-Tabir, Vanessa N. Bueno, AnnMarie N. Onwuka, Michael R. Gionfriddo, and Brian J. Piper

Ethics and Current Climates Surrounding HPV Vaccination Yezhong Lu

Virtual Newborn Nursery Rounds: An Alternative Experience During the COVID-19 Pandemic Tara E. Avery and Ashley L. Shamansky

Battling Trainee Biases and Reconstructing Perceptions in Global Neurology Maia X. Giombetti and Kelly J. Baldwin

Primary Ectopic Breast Carcinoma of the Vulva: A Case Report Youngeun C. Armbuster, Paula Ronjon, Cletus Baidoo, and Waqarun N. Rashid

Trends and Regional Differences in the Diversion of Stimulants in the United States, 2015–2019 Holly E. Funk, Susannah E. Pitt, Alison T. Varano, and Brian J. Piper

Hydroxychloroquine, Azithromycin, and Chloroquine Prescribing Patterns in Medicaid Mansi S. Khurana, Uzoamaka V. Eziri, Taylor S. Mewhiney, Cathie-Allegra Z. Nkabyo, Jennifer Szpernoga, and Daniela I. Velasquez

Changes in Morphine Distribution in the United States Megan E. Dowd, E. Jessica Tang, Kurlya T. Yan, Kenneth L. McCall, and Brian J. Piper

Chronic Traumatic Encephalopathy: A Literature Review Yvette M. Johnson, Chloé E. Mballa, Taylor S. Mewhiney, Cathie-Allegra Z. Nkabyo, and Grace L. Tieko

A Review of the Literature: How Intestinal Microbiota Shape the Immune System and the Related Impact on Human Health and Autoimmune Disease Shane P. Bross

Measuring the Change in Use of Generic Oxcarbazepine (OXC) Versus Brand Names for Medicaid Enrollees Throughout the United States in 2018-2019 Isra Amin, Rizelyn Benito, Daniela Velasquez, Megan Yeung, and Alyssa Trajan

Rat-Bite Fever in a 14-Year-Old Male Wyatt L. Keck, Michael S. Pheasant, Desiree N. Wagner, and Lindsay M. Dittman

Conflict of Interest Disclosure Accuracy Among Physician Authors of Cancer Research Journals Shuyi Chen, Alivia L. Roberts, Kevin Zhao, Abigail C. Burke, Jesse E. Ritter, Katherine M. Musto, and Brian J. Piper

Amy L. Kennalley, Youcef A. Boureghda, Jay G. Ganesh, Adam M. Watkins, Kenneth L. McCall, and Brian J. Piper

Exploring the Impact of Medicaid Expansion on Colorectal Cancer (CRC) with a Focus on Individuals Below the Standard Screening Age in the United States Sandybell J. Anorga, Mukta C. Bhatnagar, Chantel V. Golding, Brian M. Grodecki, Eric M. Watiri, Brian J. Piper, and Elizabeth C. Kuschinski

Pelvic Examinations Under Anesthesia (EUA) Informed Consent Policy Albena Gesheva, Caitlin Tillson, and Genevieve Conway

Hunger vs. Heart Failure: Can Food Insecurity Screening Prevent CHF Exacerbations? Sarah Eidbo, Johanna Dungca, Amanda Goetz, Nicholas Fiala, Andrew Denisenko, Julie Sturzen, and John Pamula

103 Effects of Medical Cannabis on Patients with Charcot-Marie-Tooth Disease Priscilla C. Canals, Alexia G. Aguilar, Myriam Moise, Jasmine Bernard, Gregory T. Carter, Allison Moore, Robert Moore, Joy Aldrich, Margaret D’Elia, Andrew Westerkamp, Miyabe Shields, and Brian J. Piper

110 Short- and Long-Term Outcomes of Breastfeeding on Children’s Mental and Physical Health Taylor S. Mewhiney

116 An Examination of the Impact of COVID-19 on Black, Latino and Asian New York City Residents and the Factors Associated with the Social Determinants of Health Stephanie Ren, Naomi Francois, and Vicki T. Sapp

123 The Natural History of Genu Valgum in the Pediatric Obese Patient Mark Mandel, Brandi Woo, Benjamin Wheatley, Amanda Young, Peter Fabricant, and Mark Seeley

128 Review of Selected Contemporary Treatments for Ischemic Stroke Gwendolene K. Conteh

133 The Role of Socioeconomic Factors in Influencing Tuberculosis Rates: A Comparison of New York State and New York City 2007-2016 Raskirth P. Singh, Rachael O. Oluwasanmi, Smita S. Bajgain, Matthew M. Bradley, Michael I. Abdool, and Kylar J. Harvey

141 Environmental Influences on Childhood Asthma Prevalence in Philadelphia Raeva N. Mulloth, Alexander J Blackley, Peter J. Koszuta, Kaitlyn M. Nemes, Maddison M. Vail, and Monglin L. Zhang

146 Declines and Pronounced Regional Disparities in Prescription Opioids in the United States Joshua D. Madera, Amanda E. Ruffino, Adriana Feliz, Kenneth L. McCall, and Brian J. Piper

154 The Association Between Early Menarche and Gestational Diabetes: A Secondary Analysis Annabel S. Alfonseca, Uzoamaka V. Eziri, Anmol K. Kaur, Taylor S. Mewhiney, and Grace L.Tieko

160 Review of Ketamine as a Rapid Antidepressant for TreatmentResistant Depression Sankung X. Darboe, Peter J. Koszuta, Paul W. Lee, and Mannaa I. Mannaa

168 The Relationship between Food Deserts and the Prevalence of Type 2 Diabetes in Communities in Southeastern Pennsylvania Stephen O. Adeniyi, Adewale O. Awoyemi, Daniel O. Ayodele, Cheryl A. Frazier, Yvette M. Johnson, Cathie-Allegra Z. Nkabyo, and Theodore J. Tucker

174 Characterizing the Behavioral and Cellular Effects of the R904S Variant of OPA1 as a Tourette Disorder Probable Risk Gene Kinza Abbas, Cara Nasello

183 Why is There a Ten-Fold Variation between States in Clozapine Usage among Medicaid Enrollees in the United States? Rizelyn A. Benito, Michael H. Gatusky, Mariah W. Panoussi, Anisa S. Suparmanian, and Brian J. Piper

189 The Changing Use of Opioids in the U.S. from 2017 to Early 2020 Rachel Gifeisman, Aneesha Morris, Bianca Sanchez, Kenneth L McCall, and Brian J Piper

194 Examining Health Literacy and Health Outcomes Among United States Immigrants and Non-Immigrants Jason L. McLeod, Muna M. Ahmed, Darin M. Chhing, Sami R. Hasan, Teresa N. James, and Yashoda T. Khatiwoda

201 Electroconvulsive Therapy Uses and Its Ability to Induce Neurogenesis: A Literature Review Kylar J. Harvey, Gwyneth J. Harris, Alexander I. Greenstone, Catherine L. Falzone, and Sami R. Hasan

209 Improving the Future of the Opioid Epidemic: Methocinnamox Colleen G. Jordan, Amy L. Kennalley, Tenzing Dolma, Kaitlyn M. Nemes, and Alivia L. Roberts

216 Treatment of Pediatric Lisfranc Injuries: A Systematic Review and Introduction of a Novel Treatment Algorithm Samuel Paek and Grant D. Hogue

223 The Relationship Between Treatment Center Services and Number of Opioidrelated Deaths in the United States Before and After a Declaration of a National Opioid Crisis Brittany N. Davis, Courtney L. Hatton, Mahamed A. Jama, and Nidha S. Samdani

233 Racial Differences in Insurance Type between Diabetes Mellitus Type 2 Patients in the United States Mannaa I. Mannaa

237 Pronounced Declines in Licit Fentanyl Utilization and Changes in Prescribing and Reimbursement Practices in the United States, 2010-2019 Raymond A. Stemrich, Jordan V. Weber, Kenneth L. McCall, and Brian J. Piper

246 A Novel Approach to Chest Wall Reconstruction Following Intrathoracic Scapular Dislocation Michael S. Pheasant and Shazad Shaikh

248 Differences in Multiple and Single-Drug Arrests by the Maine Diversion Alert Program (DAP) Maaz Siddiqui, John Piserchio, Misha Patel, Jino Park, Michelle Foster, Clare E. Desrosiers, John Herbert, Stephanie D. Nichols, Kenneth McCall, and Brian J. Piper

253 Investigating Potential Conflicts of Interest Among UpToDate and DynaMed Content Contributors SooYoung H. VanDeMark, Mia R. Woloszyn, Laura A. Christman, Michael Gatusky, Warren S. Lam, Stephanie S. Tilberry, and Brian J. Piper

258 Summer Research Immersion Program 259 Finding your way 260 Medical Research Honors Program 261 Cover art submissions

A message from the editor-in-chief As the Journal of Scholarly Research in Progress (SCRIP) enters its fifth year of publication, I would like to offer thanks to our readers, our contributors, our faculty reviewers, and our student editors for their continued support of the journal and its mission: to promote and disseminate student scholarly activity at Geisinger Commonwealth School of Medicine. In the last five years, SCRIP has seen continual growth in the form of submissions and accepted published works. This year was no different, as we received well over 50 quality submissions from our students which included literature reviews, case reports, and original research manuscripts on topics ranging from virtual newborn nursery rounds to investigating conflict of interest disclosures among physician authors. Throughout these five years, the number of manuscript submissions from students in our MBS program and from female first authors has doubled and the number of submissions from students participating in the Biomedical Research Club has tripled. Clearly, our students recognize the importance of research and view publication as being important for their career progression. In large part, this pursuit is due to the commitment of faculty throughout Geisinger who serve as mentors to our students and who recognize the importance of research and scholarship for their education. Brian Piper, PhD, assistant professor of neuroscience, believes “publishing in SCRIP is an invaluable experience for students! The scientific publication process may seem foreign and unattainable, but the vast majority of students learn that this is well within their capabilities. The feedback they receive from reviewers often strengthens their work for when it is later submitted to another outside journal.” Indeed, many SCRIP authors have gone on to publish their work in prominent external research journals. I am heartened by the level of student interest in disseminating their research and scholarly work. Those submitting authors who have had their work accepted should be proud of their achievement. As always, I take the responsibility of sustaining and building upon the SCRIP’s quality and success very seriously. Suggestions from our contributors and readers to further develop or improve the journal are more than welcome; if you would like to share your thoughts, please email me at slobo@som.geisinger.edu. Lastly, I would like to take this opportunity to invite students who have an interest in being involved in the editorial work of the journal. The role of the student editor is to assist in the writing, peer review, publication, and editing of content for the journal. It is a great way to build your academic scholarship portfolio, and it helps to ensure the journal’s growth and sustainability. Send your updated CV (including all relevant research and/or creative scholarship experience and all relevant writing, editing, or peer critique experience) to scrip@som.geisinger.edu with the subject "Application for Student Editor." Sincerely,

Sonia Lobo, PhD Editor-in-Chief

Student editors Jaclyn Podd, MD Class of 2024 Julia Schroer, MD Class of 2023 Niraj Vyas, MD Class of 2024 Acknowledgments The SCRIP would not be possible without the contributions of faculty and student volunteers committed to the review and assessment of submitted articles. Their feedback provides student authors with an opportunity to strengthen their writing and to respond to critiques. We gratefully acknowledge the following faculty members for their support in providing peer review. John A. Arnott, PhD David Averill, PhD Carmine Cerra, MD Thomas M. Churilla, MD Anthony Gillott, MD Michael Gionfriddo, PharmD, PhD Elizabeth Kuchinski, MPH William McLaughlin, PhD Kimberly Miller, PharmD Jacob Parrick, MD Brian Piper, PhD Jaclyn Podd, MD Class of 2024 Cyamatare Felix Rwabukwisi, MD, MPH Vaibhav Sharma, MD Class of 2022 Angela Slampak-Cindric, PharmD, BCPS, BCCCP Youssef Soliman, MD, PhD Michael Sulzinski, PhD Niraj Vyas, MD Class of 2024 Jennifer K. Wagner, JD, PhD Gabi Waite, PhD Mark White, MD Cathy Wilcox, PhD Eric Wright, PharmD, MPH

Office of Research & Scholarship MSB, Suite 2024, Second Floor West 570-504-9662 Sonia Lobo, PhD, RYT Associate Dean for Research & Scholarship Professor of Biochemistry Michele Lemoncelli Administrative Assistant to the Associate Dean for Research & Scholarship Laura E. Mayeski MT(ASCP), MHA Manager, Research Compliance Adam Blannard, MS Manager, Research Education Resources Tracey Pratt, MPH Grants Specialist

Volume 5 • November 2021

Scholarly Research In Progress

On the cover: The cover image is a chalk pastel drawing of a snake plant by Elana Stains, MD Class of 2025. Snake plants are excellent air filters— unlike most plants, they convert CO2 to O2 at night. Snake plants are also known for their ability to absorb benzene, formaldehyde, and xylene.

Scholarly Research In Progress • Vol. 5, November 2021

Pelvic Examinations Under Anesthesia (EUA) Informed Consent Policy Albena Gesheva1†, Caitlin Tillson1†, and Genevieve Conway1† ¹Geisinger Commonwealth School of Medicine, Scranton, PA 18509 † Doctor of Medicine Program Correspondence: agesheva@som.geisinger.edu

Abstract It is a common practice for medical students to perform pelvic exams under anesthesia (EUAs) on women undergoing procedures without the patient explicitly consenting to this exam, despite various governing bodies advising against this practice for over a decade. The Association of American Medical Colleges (AAMC) in 2003 and American College of Obstetricians (ACOG) in 2011 released statements against pelvic exams under anesthesia (EUA) without explicit consent. Additionally, 15 states have outlawed non-consensual pelvic exams, with 2 states in the process of passing legislation, including Pennsylvania. At Geisinger Commonwealth School of Medicine, we established a policy to protect patients and students from participating in sensitive genitourinary and breast examinations under anesthesia without explicit patient consent. Our policy is in keeping with the guidelines set by the governing bodies and expands on them by including breast and prostate exams, to ensure high-quality patient care and maintain ethical medical training.

Introduction Pelvic examination under anesthesia (EUA) is an internal exam during which the vagina, cervix, uterus, bladder, and/or rectum are visualized by scopes or examined by a digital exam for abnormalities. Risks of this exam are minimal but can include infection, bleeding, or traumatic damage. It is indicated when a patient cannot be adequately examined without sedation or general anesthesia for reasons of physical or psychological discomfort, or to provide information that will help guide a subsequent surgical procedure. It is common practice during medical school training to allow students to perform pelvic exams on patients under anesthesia, involving inserting fingers of a gloved hand in the patient’s vagina. The Association of Professors of Gynecology and Obstetrics (APGO) supports this practice and considers it essential to student education, clarifying that student involvement is only indicated when the procedure has been “Explicitly consented to; Related to the planned procedure; Performed by a student who is recognized by the patient as a part of their care team; And done under direct supervision by the educator” (1). In their publication of “Professional Responsibilities in Obstetric–Gynecologic Medical Education and Training,” APGO acknowledges the inherent disparity in power and authority that students face in their learning environment and how essential medical student education is to maintaining standards of medical competence (2). They offer recommendations to student education, including respecting patient autonomy by allowing the patient to choose when not to be cared for by learners and EUAs to be performed only when specific informed consent is obtained prior to surgery (2). 2

Additionally, learners should not be placed in situations where they are required to provide care or perform procedures for which they were not consented and not adequately supervised (2). Other educational bodies have spoken on the subject, including the AMA Council on Ethical and Judicial Affairs and the Association of American Medical Colleges (AAMC), echoing the importance of obtaining explicit consent when students are anticipated to be involved in EUAs (3). In 2018, Bioethics published “Educational pelvic exams on anesthetized women: Why consent matters” (4). The article addresses the ethical justification for informed consent in maintaining the patient's autonomy, trust, and basic rights and how foregoing this process is a violation, regardless of whether she becomes aware of it. It acknowledges the objections based in utilitarianism that EUAs offer benefit to the student’s education and thus are justified. The practice of standardized patients to teach pelvic exams is widely accepted by teaching institutions and has been argued to be more valuable with the added benefit of guidance by the patient. Despite the guidance from several medical governing bodies, it has been common practice for medical students to practice pelvic exams on patients under anesthesia without first obtaining explicit consent. In 2019, ELLE conducted a survey of 101 medical students from seven major American medical schools (5). Ninety-two percent reported performing a pelvic exam on an anesthetized female patient. Of that group, 61% reported performing this procedure without explicit patient consent. Nearly one-third of the respondents felt unable to opt out of performing these exams. Since supervising residents and attending physicians write evaluations, students feared jeopardizing their grades and future careers. This elucidates that a common medical student experience is not as benign as many believe. A student is quoted after performing a prostate exam on an anesthetized elderly man, stating “I feel like I just sexually assaulted a patient…That I had to violate a patient’s bodily autonomy in order to check off a requirement for a pass/ fail one-week rotation is absurd.” It is evident that educational institutions need to change their standards to protect both patients and students.

Methods In order to develop a proposal for the need to establish a policy, as well as to write the policy once we had approval, we researched state laws on EUAs passed at the time our initiative began in August 2019, including those of California, Hawaii, Illinois, Iowa, Oregon, Utah, Virginia, Michigan, and New York. Additionally, we referenced established policies from the AAMC and the American College of Obstetricians (ACOG) for

Pelvic Examinations Under Anesthesia (EUA) Informed Consent Policy

formatting and terminology. Beginning in October 2019, we presented our proposal to both school and Geisinger leadership to convince all parties of the need to establish a policy. Once approved to create a policy, we drafted it using the resources noted above and then worked with two administrative leaders who are practicing OB/GYN physicians to finalize the language. It was presented to and approved by the Medical Curriculum Committee in June 2020.

Discussion This policy was implemented to protect the patients’ autonomy and bodily rights as well as to protect students from participating in sensitive genitourinary and breast examinations under anesthesia without explicit consent obtained from the patient (Figure 1). A study reported that several of their respondents said they would feel “physically assaulted” if not explicitly consented (6). As many as 72% to 100% of women said that they would want to be consented before an educational pelvic EUA was performed on them (7). These statistics are especially important when put in context that, according to the Centers for Disease Control and Prevention, 1 in 3 women in the United States have experienced sexual violence (8).

consent only if the student was female, 18% are not sure, and only 14% would refuse (10). A study in Ireland tracked the number of women who agreed to having a pelvic EUA performed by a medical student and found that 74% consented (11). These studies support that obtaining explicit consent would not interfere with educational opportunities. Performing an EUA is only indicated when a normal pelvic examination cannot be adequately performed due to physical or psychological pain or when proper staging of vaginal or cervical cancer is needed for surgery. When a pelvic EUA is clinically warranted, informed consent should be performed including discussion from the primary physician with the patient about why the pelvic EUA is needed and who would be involved in the process. After the context of the pelvic EUA is explicitly given, it can be documented, and a surgical consent form can be signed stating that which was consented.

This school policy is crucial to establish ethical practice in students before they become practicing physicians. Students learn medical practice norms from mentors, and it is important to instill trust and respect supported by policy during this formative time. A 2003 study of 401 Philadelphia medical students found that trainees who had completed an OB/GYN rotation viewed consent as significantly less important than those who had not yet completed an OB/GYN rotation (51% compared to 70%) (9). These results suggest that the environment in which you learn to practice significantly influences attitudes about ethics, and the current environment is not teaching future physicians to value or respect patient autonomy. A common misconception revealed during our research suggested that obtaining explicit consent would hinder teaching. However, studies have shown that a majority of women would consent if explicitly asked; a survey in Canada found that the majority of women (62%) report that they would agree to have a pelvic EUA performed on them by a medical student, while 5% say they would Figure 1. Geisinger Commonwealth School of Medicine Informed Consent Policy 3

Pelvic Examinations Under Anesthesia (EUA) Informed Consent Policy

Students have been asked to perform EUAs during procedures when it is clinically indicated and when it is not; they sometimes even occur during non-OB/GYN procedures. However, a student performing an EUA is never indicated and is always for educational purposes. Even if an EUA is clinically warranted, it must be repeated by a licensed provider, evidencing the student’s exam as superfluous and possibly injurious. Therefore, it is important that patients be given the right to decide whether they want any person, particularly an unlicensed one, performing a sensitive exam that is not required for their care. A patient being unconscious does not negate the autonomy and respect due to them; the vulnerability of sedation requires us to be even more sensitive to upholding a patient’s rights when they are unable to for themselves. Revising policies on this practice at the school level will ensure that students will develop ethical practices regardless of where they subsequently practice, and ideally instill these ethics into institutions nationwide. While the policy will increase awareness and compliance with best practices, we advise that the baseline consent forms for all procedures requiring anesthesia should include a subsection requiring signature to explicitly consent to a medical student performing a pelvic, breast, or prostate EUA, with the option to not consent to this; this guarantees the patient is aware of the possibility for an exam and gives consent, eliminating the potential for human error in failing to obtain consent. Additionally, a nurse or physician should orally and explicitly discuss this with the patient prior to the exam.

Acknowledgments We would like to thank Thomas Samuelsen, MD, William Jeffries, PhD, Michael Ferraro, MD, and George Valenta, MD, for their guidance and efforts to implement this policy.

Disclosures The authors have nothing to disclose.

References

APGO Statement on Teaching Pelvic Exams to Medical Students [Internet]. APGO. 2019 [cited 6 March 2021]. Available from: https://apgo.org/page/ teachingpelvicexamstomedstudents

Committee Opinion No. 500: Professional Responsibilities in Obstetric–Gynecologic Medical Education and Training. Obstetrics & Gynecology. 2011;118(2):400-404.

Adashi E. Teaching Pelvic Examination Under Anesthesia Without Patient Consent [Internet]. JAMA Health Forum. 2019 [cited 6 March 2021]. Available from: https://jamanetwork.com/channels/health-forum/ fullarticle/2759681

Friesen P. Educational pelvic exams on anesthetized women: Why consent matters. Bioethics. 2018;32(5): 298-307.

Tsai J. Medical Students Regularly Practice Pelvic Exams on Unconscious Patients. Should They? [Internet]. ELLE. 2019 [cited 6 March 2021]. Available from: https://www.elle. com/life-love/a28125604/nonconsensual-pelvic-examsteaching-hospitals/

Hsieh P. Pelvic Exams on Anesthetized Women Without Consent: A Troubling And Outdated Practice [Internet]. Forbes. 2018 [cited 6 March 2021]. Available from: https://www.forbes.com/sites/paulhsieh/2018/05/14/ pelvic-exams-on-anesthetized-women-without-consent-atroubling-and-outdated-practice/?sh=42d59fee7846

Bibby J, Boyd N, Redman C, Luesley D. Consent for vaginal examination by students on an anesthetized patient. The Lancet. 1988;332(8620):1150.

Basile K, Smith S, Breiding M, Black M, Mahendra, R. Sexual violence surveillance. Atlanta, Georgia: Centers for Disease Control and Prevention, National Center for Injury Prevention and Control; 2014.

Ubel P, Jepson C, Silver-Isenstadt A. Don't ask, don't tell: A change in medical student attitudes after obstetrics/ gynecology clerkships toward seeking consent for pelvic examinations on an anesthetized patient. American Journal of Obstetrics and Gynecology. 2003;188(2):575-579.

10. Wainberg S, Wrigley H, Fair J, Ross S. Teaching Pelvic Examinations Under Anesthesia: What Do Women Think? Journal of Obstetrics and Gynecology Canada. 2010;32(1):49-53. 11. Martyn F, O'Connor R. Written consent for intimate examinations undertaken by medical students in the operating theatre-time for national guidelines? Irish Medical Journal. 2009;102(10):336–337.

Scholarly Research In Progress • Vol. 5, November 2021

Geographic Variability in Antibiotic Prescribing Rates in Medicaid Alexia G. Aguilar1†, Priscilla C. Canals1†, Kimberly A. Miller1, and Brian J. Piper1,2 Geisinger Commonwealth School of Medicine, Scranton, PA 18509 ²Center for Pharmacy Innovation and Outcomes, Forty Fort, PA 18704 † Doctor of Medicine Program Correspondence: aaguilar@som.geisinger.edu 1

Abstract Background: Antibiotic resistance is a persistent and growing concern. There is a lack of data identifying the current state of antibiotic prescription patterns in the Medicaid program. We analyzed temporal and regional trends in antibiotic prescribing data across the United States (U.S.) to identify regional disparities. Methods: We analyzed prescribing rates over the past 2 years in Medicaid Part D for 8 antibiotics. Four were broad spectrum: azithromycin, ciprofloxacin, levofloxacin, and moxifloxacin; and 4 were narrow spectrum: amoxicillin, cephalexin, doxycycline, and trimethoprim/sulfamethoxazole. We identified the geographical distribution of these antibiotics across the U.S. Furthermore, we evaluated total antibiotic prescriptions per state per quarter during 2018 and 2019 collected from the Medicaid State Drug Utilization database. Prescription rates were reported per 1,000 Medicaid enrollees. The states were divided into specific geographic regions according to the U.S. Census to determine which regions have the highest and lowest prescription rates. We analyzed the data and constructed figures using International Business Machine Corporation’s Statistical Package for the Social Sciences (IBM SPSS), Statistical Analysis System’s John’s Macintosh Project (SAS JMP), and GraphPad Prism. Results: Antibiotic prescriptions decreased 9.6% from 2018 to 2019. Amoxicillin was the predominant antibiotic, followed by azithromycin, cephalexin, trimethoprim/sulfamethoxazole, doxycycline, ciprofloxacin, levofloxacin, and moxifloxacin. Substantial geographic and quarterly variation in antibiotic prescribing existed. The South prescribed 52.2% more antibiotics (580/1,000) in 2019 than the West (381/1,000). We identified a significant correlation between the 2018 and 2019 prescription rates (r =0.95, p < 0.001). Conclusions: This study identified the geographical prescribing rates of 8 antibiotics during 2018 and 2019. The south had the highest prescribing rates among all the regions. Areas of high antibiotic prescribing rates may benefit from programs to reduce unnecessary prescribing. Further analysis on state level Medicaid or prescribing policies may be done to identify reasons for such high prescribing rates.

Introduction Due to the increasing use of antibiotics over the past century, antibiotic resistance has become one of the most pressing threats to public health today. According to the World Health Organization, increasing antibiotic resistance is one of the biggest threats to global health, food security, and development of human defenses today (1). Studies have discovered overuse

and inappropriate prescribing of antibiotics in a variety of health care settings (2–5). According to a 2019 report released by the Centers for Disease Control and Prevention, more than 2.8 million antibiotic-resistant infections occur in the United States (U.S.) each year, and more than 35,000 people die as a result. (6). Most of these antibiotic prescriptions are for respiratory infections commonly caused by viruses, which do not respond to antibiotics. (7). Antibiotics are often prescribed unnecessarily (8, 9), and 30–50% of prescriptions for them are not associated with an indication (10–12). Sulfonamides and urinary antiinfective agents are the classes most likely to be prescribed without documentation (11). Up to 25% of antibiotics prescribed in the outpatient setting to Medicaid beneficiaries were not associated with a provider visit and therefore are not screened by existing antimicrobial stewardship systems (13). Further, among 298 million prescriptions filled by 53 million Medicaid patients between 2004 and 2013, 45% of prescriptions for antibiotics were made without any clear rationale. Twenty-eight percent of antibiotics were prescribed without evidence of seeing the provider, and 17% were given without documentation for infection-related diagnosis. Inappropriate antibiotic prescribing leads to antibiotic resistance on individual as well as community levels, particularly against public health threats such as carbapenem-resistant Enterobacteriaceae and methicillin-resistant Staphylococcus aureus (14). Furthermore, overuse of antibiotics increases the risk of adverse effects such as rash, GI upset, and renal dysfunction (6). The risk of infection with Clostridium difficile increases with duration of exposure to antibiotics (15, 16). Recent data supports the relationship between early antibiotic exposure before age 2 and subsequent negative outcomes such as obesity, autoimmune disorders, and asthma (17). It has been found that geographic areas with high antibiotic consumption are associated with increased antibiotic resistance, and that broad-spectrum antibiotics are more likely to be associated with antibiotic resistance (18). The South census region of the United States has the highest prescribing rate among the rest of the country (19, 20). Examination of antibiotic prescription variation in the U.S. in 2011 found that the most common antibiotic categories were penicillins followed by macrolides. Among individual antibiotics, the most prescribed antibiotic agents were azithromycin followed by amoxicillin. When examining geographic trends in prescriptions, prescribing rates were highest in the South. Specifically, Kentucky had the highest prescriptions rates (1,281 prescriptions per 1,000 persons) which was about four-fold higher than Alaska with the lowest (348 per 1,000 persons) (19). A report of California Medicaid beneficiaries also found penicillins to be the most prescribed antibiotic class, followed

Geographic Variability in Antibiotic Prescribing Rates in Medicaid

by macrolides, cephalosporins, sulfonamides, quinolones, and tetracyclines (18). Amoxicillin, azithromycin, cephalexin, doxycycline, and sulfamethoxazole/trimethoprim were the most frequently used agents in their respective class (18). Due to the public health risks that arise from antibiotic resistance, it is imperative to be continuously vigilant to identify and avoid unnecessary prescriptions. This study examined the temporal and geographical patterns of antibiotic prescribing rates among U.S. Medicaid program beneficiaries.

Methods Participants Medicaid is a joint federal and state program that provides coverage for 75 million people or 21% of the U.S. population (21). It is one of the largest payers for health care in the U.S. All states provide coverage for outpatient prescription drugs (21). Procedures State antibiotic prescription totals and population data during 2018–2019 were collected from the Medicaid State Drug Utilization database. Total number of prescriptions and Medicaid enrollee data were used to calculate individual antibiotic prescribing rates per state. Antibiotics were identified using the generic and trade names (22). Antibiotics were further categorized as broad or narrow spectrum based on their potential for influencing antibacterial resistance as well as their spectrum of activity according to the National Committee For Quality Assurance's “antibiotics of concern.” Azithromycin, levofloxacin, moxifloxacin, and ciprofloxacin are broad-spectrum antibiotics, while amoxicillin, cephalexin, doxycycline, and trimethoprim sulfamethoxazole are narrow-spectrum antibiotics (23, 24).

Results In 2019, a total of 33,011,946 of the antibiotics included for analysis were prescribed with an average prescribing rate of 464 prescriptions/1,000 enrollees. The total amount prescribed was 9.61% lower than that observed in 2018 (36,519,951 total antibiotics, 494 prescriptions/1,000 enrollees). There was considerable variation in antibiotic prescribing rates among different regions (Figure 1). The South census region had the highest prescribing rate (580 prescriptions/1,000 enrollees in 2019, 601 prescriptions/1,000 enrollees in 2018) compared to the West census region, which had the lowest (381 prescriptions/1,000 enrollees in 2019, 420 prescriptions/1,000 enrollees in 2018). Kentucky and Louisiana were the top two states with the highest prescribing rates in 2018 and 2019 (Figure 2 and 3). Alaska and Oregon had the lowest prescribing rates in 2018, while Oregon and Washington state had the lowest in 2019. Figure 4 shows that from 2018 to 2020, amoxicillin was the most prescribed antibiotic (48.1%), followed by azithromycin (18.3%). Additionally, we found a similar trend in antibiotic prescriptions during different quarters (quarter 1: January to March, quarter 2: April to June, quarter 3: July to September, quarter 4: October to December). Prescriptions fell from April to the end of September, suggesting that there is a higher rate of prescription during cooler months.

The states were divided into geographic regions according to the U.S. Census. These regions were defined as: the South (Alabama, Arkansas, Delaware, District of Columbia, Florida, Georgia, Kentucky, Louisiana, Maryland, Mississippi, North Carolina, Oklahoma, South Carolina, Tennessee, Texas, Virginia, and West Virginia), the West (Alaska, Arizona, California, Colorado, Hawaii, Idaho, Montana, Nevada, New Mexico, Oregon, Utah, Washington, and Wyoming), the Midwest (Illinois, Indiana, Iowa, Kansas, Michigan, Minnesota, Missouri, Nebraska, North Dakota, Ohio, South Dakota, and Wisconsin), and the Northeast (Connecticut, Maine, Massachusetts, New Hampshire, New Jersey, New York, Pennsylvania, Rhode Island, and Vermont). Data analysis After the total number of prescriptions was collected and prescription rate per 1,000 enrollees was calculated, we input the data into IMB SPSS. We conducted t-tests and linear regressions to find associations between variables. Pearson correlations were completed between antibiotics for each year. We created figures and heatmaps through Statistical Analysis System’s John’s Macintosh Project (SAS JMP) and GraphPad Prism to show variation across regions in the United States. Variability was reported as the standard error of the mean (SEM). A p-value of < 0.05 was considered statistically significant.

Figure 1. Antibiotic prescription rates in Medicaid Part D by Census regions. There was considerable variation in antibiotic prescribing rates among different regions. The South census region had the highest prescribing rate (580 prescriptions/1,000 enrollees in 2019, 601 prescriptions/1,000 enrollees in 2018) compared to the West census region which had the lowest (381 prescriptions/1,000 enrollees in 2019, 420 prescriptions/1,000 enrollees in 2018).

Geographic Variability in Antibiotic Prescribing Rates in Medicaid

Figure 2. Total prescribing rate per state in 2018.

Figure 3. Total prescribing rate per state in 2019.

The correlation between population-corrected antibiotic prescriptions in 2018 is shown in Table 1. The correlation between the grand total was generally significantly associated with individual agents except for moxifloxacin. This pattern was replicated in 2019 (Table 2).

Discussion Antibiotic misuse, overuse, and subsequent resistance are major public health issues (1, 6). Patients prescribed antibiotics have been shown to develop resistance that is most impactful during the month after therapy but continues for up to a year (25). In one study at a single center, investigators found that 12.3% of antibiotics prescribed in the outpatient setting were not consistent with current guideline recommendations (26). The most common antibiotic prescribed in this study was azithromycin, followed by amoxicillin-clavulanate and amoxicillin. In a study that evaluated prescribing of antibiotics in children, azithromycin and amoxicillin were found to be the most frequently prescribed (27). This is consistent with our data, which showed amoxicillin to be the most prescribed antibiotic, followed by azithromycin. Infections of the upper respiratory and urinary tracts are the most common indications for antibiotic use in adults and children in the outpatient setting. Beta lactam and macrolide antibiotics are used for the treatment of upper respiratory tract infections in both adults and children. Doxycycline is used for upper respiratory tract infections in adults, while trimethoprim/ sulfamethoxazole (TMP/SMX) is prescribed for urinary tract infections and skin and skin structure infections. TMP/SMX is also often used in the treatment of upper respiratory tract infections despite recommendations against its use for this indication due to its spectrum of action (28). Our results showed a total of 36,519,951 antibiotics prescribed in 2018 and 33,011,946 prescribed in 2019. The prescribing rate was found to be 464 prescriptions/1,000 enrollees. Our findings show a higher rate of antibiotic prescribing compared to other studies (24). Further, we determined that the South census region had the highest prescribing rate, while the West census region had the lowest. This is consistent with findings from other studies (19, 29). Trends in antibiotic prescribing varied based on the quarter in the year. The highest antibiotic prescribing rates occurred found during October to March,

Figure 4. Quarterly antibiotic prescribing rates.

while prescribing rates fell from April to the end of September. This is consistent with trends in upper respiratory tract infections. These are often more prevalent during fall or winter months and decrease during warmer months. The correlations in population corrected antibiotic use per state showed that regional patterns were generally homogenous for all agents in both 2018 and 2019 except moxifloxacin. This suggests that the use pattern for this broad-spectrum fluoroquinolone was distinct from other agents. While we were fortunate to be using a database that collects data from across the entire United States, Medicaid only accounts for 20% of the population. Other limitations include limited data reporting from some states for certain antibiotics, for example amoxicillin/clavulanic acid was removed from the study due to limited data reported on the antibiotic.

Geographic Variability in Antibiotic Prescribing Rates in Medicaid

Table 1. Correlation matrix showing associations between antibiotics in 2018. Correlation coefficients were derived from pooled data from the Medicaid State Drug and Utilization database. * p < 0.001

Table 2. Correlation matrix showing associations between antibiotics in 2019. Correlation coefficients were derived from pooled data from the Medicaid State Drug and Utilization database. * p < 0.001

Conclusion

Disclosures

This study provides an overview of antibacterial prescribing practices in the Medicaid system. Findings include discrepancies in antibacterial prescribing rates among different regions across the U.S. as well as during different times of the year. Future research plans include identifying state-level antibiotic prescribing policies as well as how COVID-19 has impacted antibiotic prescription rates.

This work was supported by Health Resources and Services Association (HRSA) Center of Excellence (grant number D34HP310250).

Acknowledgments We thank Geisinger Commonwealth School of Medicine’s Center of Excellence for their work in organizing and coordinating this project.

References 1.

World Health Organization. Antibiotic Resistance 2020 [Available from: https://www.who.int/news-room/factsheets/detail/antibiotic-resistance.

Fridkin S, Baggs J, Fagan R, Magill S, Pollack LA, Malpiedi P, et al. Vital signs: improving antibiotic use among hospitalized patients. MMWR Morb Mortal Wkly Rep. 2014;63(9):194-200.

Geographic Variability in Antibiotic Prescribing Rates in Medicaid

Huttner B, Samore M. Outpatient antibiotic use in the United States: time to "get smarter." Clin Infect Dis. 2011;53(7):640-3.

Septimus EJ, Owens RC, Jr. Need and potential of antimicrobial stewardship in community hospitals. Clin Infect Dis. 2011;53 Suppl 1:S8-S14.

White AT, Clark CM, Sellick JA, Mergenhagen KA. Antibiotic stewardship targets in the outpatient setting. Am J Infect Control. 2019;47(8):858-63.

Centers for Disease Control and Prevention. Antibiotic Resistance Threats in the United States, 2019. 2019.

Centers for Disease Control and Prevention. Measuring Outpatient Antibiotic Prescribing 2020 [Available from: https://www.cdc.gov/antibiotic-use/data/outpatientprescribing/index.html.

Barnett ML, Linder JA. Antibiotic Prescribing for Adults With Acute Bronchitis in the United States, 1996-2010. JAMA. 2014;311(19):2020-2. Chua K-P, Fischer MA, Linder JA. Appropriateness of outpatient antibiotic prescribing among privately insured US patients: ICD-10-CM based cross sectional study. BMJ. 2019;364:k5092.

10. Fleming-Dutra KE, Hersh AL, Shapiro DJ, Bartoces M, Enns EA, File TM Jr., et al. Prevalence of Inappropriate Antibiotic Prescriptions Among US Ambulatory Care Visits, 20102011. JAMA. 2016;315(17):1864-73. 11. Ray MJ, Tallman GB, Bearden DT, Elman MR, McGregor JC. Antibiotic prescribing without documented indication in ambulatory care clinics: national cross sectional study. BMJ. 2019;367:l6461. 12. Havers FP, Hicks LA, Chung JR, Gaglani M, Murthy K, Zimmerman RK, et al. Outpatient Antibiotic Prescribing for Acute Respiratory Infections During Influenza Seasons. JAMA Netw Open. 2018;1(2):e180243. 13. Non-Infection-Related And Non-Visit-Based Antibiotic Prescribing Is Common Among Medicaid Patients. Health Affairs. 2020;39(2):280-8. 14. Ventola CL. The antibiotic resistance crisis: part 1: causes and threats. P t. 2015;40(4):277-83. 15. Brown KA, Fisman DN, Moineddin R, Daneman N. The magnitude and duration of Clostridium difficile infection risk associated with antibiotic therapy: a hospital cohort study. PLoS One. 2014;9(8):e105454. 16. Hensgens MP, Goorhuis A, Dekkers OM, Kuijper EJ. Time interval of increased risk for Clostridium difficile infection after exposure to antibiotics. J Antimicrob Chemother. 2012;67(3):742-8.

18. Gahbauer AM, Gonzales ML, Guglielmo BJ. Patterns of Antibacterial Use and Impact of Age, Race/Ethnicity, and Geographic Region on Antibacterial Use in an Outpatient Medicaid Cohort. Pharmacotherapy: The Journal of Human Pharmacology and Drug Therapy. 2014;34(7):677-85. 19. Hicks LA, Bartoces MG, Roberts RM, Suda KJ, Hunkler RJ, Taylor TH, Jr., et al. US outpatient antibiotic prescribing variation according to geography, patient population, and provider specialty in 2011. Clin Infect Dis. 2015;60(9):1308-16. 20. Kakpovbia E, Feng H, Feng PW, Cohen JM. Antibiotic prescribing trends among US dermatologists in Medicare from 2013 to 2016. Journal of Dermatological Treatment. 2021;32(1):70-2. 21. Medicaid. [Available from: https://www.medicaid.gov/ medicaid/index.html. 22. Trade Names [Internet]. 23. National Committee for Quality Assurance. HEIDIS 2017 Final NDC Lists 2017 [Available from: https://www.ncqa. org/hedis/measures/hedis-2017-national-drug-code-ndclicense/hedis-2017-final-ndc-lists/. 24. Lee GC, Reveles KR, Attridge RT, Lawson KA, Mansi IA, Lewis JS, et al. Outpatient antibiotic prescribing in the United States: 2000 to 2010. BMC Medicine. 2014;12(1):96. 25. Costelloe C, Metcalfe C, Lovering A, Mant D, Hay AD. Effect of antibiotic prescribing in primary care on antimicrobial resistance in individual patients: systematic review and meta-analysis. BMJ. 2010;340:c2096. 26. Shively NR, Buehrle DJ, Clancy CJ, Decker BK. Prevalence of Inappropriate Antibiotic Prescribing in Primary Care Clinics within a Veterans Affairs Health Care System. Antimicrob Agents Chemother. 2018;62(8). 27. Fleming-Dutra KE, Demirjian A, Bartoces M, Roberts RM, Taylor TH, Jr., Hicks LA. Variations in Antibiotic and Azithromycin Prescribing for Children by Geography and Specialty-United States, 2013. Pediatr Infect Dis J. 2018;37(1):52-8. 28. Wong DM, Blumberg DA, Lowe LG. Guidelines for the use of antibiotics in acute upper respiratory tract infections. Am Fam Physician. 2006;74(6):956-66. 29. Kabbani S, Palms D, Bartoces M, Stone N, Hicks LA. Outpatient Antibiotic Prescribing for Older Adults in the United States: 2011 to 2014. J Am Geriatr Soc. 2018;66(10):1998-2002.

17. Ni J, Friedman H, Boyd BC, McGurn A, Babinski P, Markossian T, et al. Early antibiotic exposure and development of asthma and allergic rhinitis in childhood. BMC Pediatrics. 2019;19(1):225.

Scholarly Research In Progress • Vol. 5, November 2021

Declining National Codeine Distribution in United States Hospitals and Pharmacies Amy L. Kennalley1*, Youcef A. Boureghda1†‡, Jay G. Ganesh1†‡, Adam M. Watkins1†‡, Kenneth L. McCall2, and Brian J. Piper1 ¹Geisinger Commonwealth School of Medicine, Scranton, PA 18509 ²University of New England, Portland, ME 04005 † Doctor of Medicine Program *Master of Biomedical Sciences Program ‡ Authors contributed equally Correspondence: akennalley@som.geisinger.edu

Abstract

Introduction

Background: Past research has identified pronounced regional disparities in use of different opioids but less is known for codeine. The primary objective of this study was to analyze the trends of distribution of prescriptions containing codeine in the United States (U.S.) from 2010 to 2019 based on business type (i.e., hospitals, pharmacies, health care practitioners, teaching institutions). This investigation also analyzed the distribution of prescriptions containing codeine per Medicaid enrollees. In addition, this study aimed to identify regional disparities in prescribed milligrams (mg) of codeine per person across states and identify any extreme values during 2019.

Since 1979, there have been over 600,000 deaths related to drug overdoses in the United States (U.S.) (1). Of those deaths, most have been attributed to opioids, varying from illegal narcotics such as heroin to prescriptions (2). When considering that the "first wave" of the opioid epidemic began in the late 1980s, this number offers an even bleaker picture (3). What began as a good-faith effort to be more proactive in managing pain has resulted in over 2 million Americans becoming addicted to opioids in some form (4). Recognizing this growing issue, various entities at the local, state, and federal levels implemented initiatives directed to combat the problem. One such initiative that began in 2016 was the U.S. surgeon general's "Turn the Tide Rx'' campaign, which focused on providing educational resources as well as a voluntary pledge for clinicians to uphold (5). Despite numerous efforts to "turn the tide," the number of people addicted and dying from opioids continues to increase, so perhaps weaker opioids may provide an alternative.

Methods: The distribution of codeine via pharmacies, hospitals, and practitioners in kilograms was obtained from the Drug Enforcement Administration’s Automated Reports and Consolidated Ordering System (ARCOS) from 2010 to 2019. In addition, the number of prescriptions of codeine per 1,000 Medicaid enrollees was obtained from the State Drug Utilization Database (SDUD) provided by Medicaid. Results: There was a steady decrease (-25.0%) in total grams of codeine through all distributors from 2010 through 2019. There was an increase in total grams of codeine distributed in the U.S. from 2014 to 2015 of +28.9%; this was the largest increase between 2 consecutive years. The largest decrease from 2010 to 2019 for a given distributor type was hospitals, with an -89.6% decrease. Within the same time frame, there was a -20.8% decrease in grams of codeine distributed through pharmacies. Total grams of codeine distributed in Texas was significantly elevated relative to the national average. Conclusion: The peak of prescription codeine in 2011 was consistent with the overall peak in prescription opioids, with a progressive decrease over the following decade. This could be explained by relatively recent recommendations regarding the therapeutic use of codeine and how other antitussive agents may be of better use. The precipitous rise of codeine in Texas that we observed has been recognized in prior studies. The national average prescription for codeine is 50 mg per person, while in Texas it is about 147 mg per person. Additionally, the peak in SDUD data in 2016 might be due to corresponding peak in Medicaid/CHIP enrollees for that year. In conclusion, more research investigating the variance of codeine prescription in Texas is needed.

Codeine is a prodrug converted to morphine by the cytochrome P450 enzyme CYP2D6 (6, 7). Morphine, the metabolite that provides analgesic effects, is further metabolized by glucuronidase enzymes to form the inactive metabolite morphine-3-glucoronide and the active metabolite morphine6-glucoronide (6). Genetic polymorphisms of CYP2D6 leads to varied efficacies of codeine based on the phenotypic classifications of ultrarapid, normal, reduced, or poor metabolizer (6, 7). Despite this variation, codeine is the most prescribed opioid in pediatrics (8). According to the 2016 Cochrane Review which examined the efficacy of codeine as a cough suppressant in children, there is a lack of sufficient evidence to support or oppose the use of cough and cold medications containing codeine (9). Various studies have recommended clinicians discover alternatives to codeine prescriptions in reaction to the continuous prescribing of codeine for children. This recommendation is due to the variety of effects and the danger of respiratory depression and mortality in children (10, 11, 12, 13). These suggestions are in accordance with the U.S. Food and Drug Administration's (FDA) 2012 "black box warning" and contraindications for codeine use in children (14, 15, 16).

Declining National Codeine Distribution in United States Hospitals and Pharmacies

Depending on the formulation, codeine is classified as a Schedule II, III or V drug in the U.S. (17, 18, 19). The Drug Enforcement Administration's (DEA) Automation of Reports and Consolidated Orders System (ARCOS) monitors the manufacturing and distribution of all Schedule I and II drugs as well as Schedule III narcotic and gamma-hydroxybutyric acid (GHB) substances (20). After rescheduling codeine products in Australia to be prescription-only in 2018, researchers examined the drug sales trends and noted a modest increase in sales (21). Interestingly, another study compared international postoperative opioid prescriptions and found codeine and tramadol accounted for approximately 58% and 45% in Canada and Sweden, and only 7% in the U.S. However, the U.S. prescribed higher doses of opioids overall following surgery more often than in Canada or Sweden (22). Though the U.S. prescribed comparatively lower postoperative codeine, the number of weaker opioid prescriptions at discharge from the emergency room in the U.S. grew from 2012 to 2017, according to the CDC's National Health Statistics Report. The formulation acetaminophen-codeine was 12.5% of all opioids prescribed in 2016–2017 (23). To date, there has been no comprehensive study analyzing national codeine prescription trends. This study's principal objective was to analyze the distribution trends of codeine-containing prescriptions within the U.S. from 2010 to 2019 based on business type (i.e., hospitals, pharmacies, and health care practitioners). Additionally, an analysis of the distribution of prescriptions containing codeine per Medicaid enrollees from 2014 to 2019 was conducted to further elucidate any trends. Finally, this study aimed to identify regional disparities in prescribed codeine across states and identify any extreme values within the 10-year time frame.

Methods The ARCOS database, expressly the Retail Drug Summary Reports, was the source of data regarding the national distribution of codeine within the U.S. from 2010 to 2019. The requirement for reporting certain controlled substances transactions from the point of manufacture to distributors — hospitals, retail pharmacies, practitioners, mid-level practitioners, and teaching institutions — to the Attorney General was made possible due to the Controlled Substances Act of 1970. However, since only certain controlled substances are monitored and recorded, it was not possible to consider all the formulations of codeine available nationally. Notably, the Schedule IV and V codeine formulations were not included in the analysis. Report 1 of the ARCOS Retail Drug Summary Reports provided data for grams of codeine distributed by zip code, while the total amount of codeine distributed nationally and by state per quarter (in grams) was collected from Report 4. Additionally, Report 5 was the source for grams of codeine purchased by healthcare business types (i.e., hospitals, pharmacies, and health care practitioners). While Report 5 did contain data on other healthcare business types (e.g., mid-level practitioners and narcotic treatment programs), the amount used by those businesses was deemed to be negligible. Data on the populations of interest were collected from American Community Survey and U.S. Census Bureau to normalize other findings for a given state or year.

The number of codeine prescriptions covered by state Medicaid agencies reported through Medicaid's State Drug Utilization Data (SDUD) database from 2014 to 2019 for all 50 states and Washington D.C. was also collected and corrected for the number of Medicaid enrollees. In addition, the number of prescriptions ordered nationally from 2014 to 2019 was collected. Finally, data analysis, figures, and heat maps were completed using GraphPad Prism, Version 9.1.0., JMP, and Microsoft Excel. This study was deemed exempt from review by the Geisinger Institutional Review Board.

Results From 2010 to 2019, the total amount of codeine distributed across all the major business types of distributors in the U.S. declined by 25% (Figure 1). Further analysis determined that hospitals experienced the largest decrease in codeine distribution with a decline of 89.6%, while pharmacies had a 20.8% decrease in codeine distribution. Practitioners experienced the smallest decrease in codeine distribution of the three business types, with a decrease of 0.15%. Though all business types experienced a decline in codeine distribution in the given period, there was an increase in codeine administration between 2014–2015. More specifically, there was an increase of 28.9% in total codeine distribution across the U.S. (Figure 1). The peak year for codeine distribution was in 2011, with an average of 53.7 mg per state. Comparing the peak year to the lowest and most recent data year in 2019, most states demonstrated a decline in distribution with an average 40.1% decrease nationwide (Figure 2). Figure 2 gives a quantitative illustration of Washington D.C. and the 46 states which experienced a reduction in codeine distribution over the 10-year focus. In contrast, four states showed an increase in codeine administration: Texas (240%), Oklahoma (28.7%), West Virginia (35.4%), and Arkansas (28.7%), which were all considered statistically significant with an alpha value of 0.05

Figure 1. Codeine (kg) by business activity as reported by the United States Drug Enforcement Administration’s Automated Reports and Consolidated Orders System (ARCOS) from 2010 to 2019.

Declining National Codeine Distribution in United States Hospitals and Pharmacies

(Figures 2 and 3). The heat map also illustrates that Nevada is the state with the most notable decline in codeine distribution between 2011 and 2019 with a decrease of 68%, which was considered statistically significant via paired t-test with an alpha value of 0.05 (Figure 3). Lastly, Figure 4 displays the distribution of codeine (mg/person) in the most recent year, 2019, per state. Medicaid prescriptions per 1,000 enrollees were examined between the years of 2014 to 2019. Figure 5 shows the prescriptions per 1,000 enrollees increased to a peak in the

third quarter of 2016 with a subsequent drop in 2017 and stabilization to the most recent year of 2019. The prescriptions per 1,000 enrollees increased from 21,517 prescriptions in the first quarter of 2014 to 391,214 prescriptions in the third quarter of 2016, reflecting a 1,718.16% increase. Progressing from this peak in the third quarter of 2016 (391,214 prescriptions) to the first quarter in 2017 (166,848 prescriptions) resulted in a 57.35% decrease in prescriptions per 1,000 enrollees.

Discussion

Figure 2. Percent change in distribution of codeine (mg/person) from 2011 to 2019 as reported by the United States Drug Enforcement Administration’s Automated Reports and Consolidated Orders System (ARCOS), where blue indicates a decrease in distribution and red indicates an increase.

This study investigated trends in codeine distribution and use in the U.S. between 2011 and 2019, with 2011 corresponding to the peak year for codeine administration. Our data suggest declining distribution of codeine by 25% across all major distributors in the U.S. (hospitals, pharmacies, practitioners) from 2011 to 2019. In addition, our data was indicative of regional disparities in codeine utilization when comparing 2011 to 2019. Forty-six out of 50 states experienced a decline in codeine distribution, while four states increased: Texas, Oklahoma, West Virginia, and Arkansas. When comparing the amount of codeine distributed in Texas in 2011 compared to 2019, there was a 240% increase in the amount of codeine utilized. Furthermore, an analysis of Medicaid prescriptions per 1,000 enrollees was examined from 2014 to 2019. The data indicated a peak in Medicaid prescriptions during the third quarter of 2016 (391,214 prescriptions per 1,000 enrollees) followed by a subsequent drop in prescriptions during the first quarter of 2017 (166,848 prescriptions per 1,000 enrollees), reflecting a 57.35% decline before stabilizing in 2019.

Figure 3. Percent change in codeine distribution (mg/person) between 2011 and 2019 as reported by the United States Drug Enforcement Administration’s Automated Reports and Consolidated Orders System (ARCOS) for 49 states and Washington D.C. (excluding Texas for clarity), (paired t-test, *p < 0.05). 12

Declining National Codeine Distribution in United States Hospitals and Pharmacies

This pattern of change in the declining distribution of codeine from the peak year of 2011 can be explained with a further examination into other prescription opioids. The peak codeine prescription rate in 2011 was consistent with the overall peak in prescription opioid rates during the same year (20). Comparing the codeine prescription rate between the peak year of 2011 and 2019, the average percent decrease of milligrams of codeine per person was 40.1%. This decrease could be explained by policies that were implemented across the U.S. As of 2015, 50 states and the District of Columbia put in place an Electronic Prescribing of Controlled Substances (EPCS) system; moreover, as of April 2021, 24 states have mandatory EPCS protocols in place (24). Given the ongoing trend towards more instituted policies directed towards prescribed substances, such as codeine, this may play an integral role in accounting for opioid prescription reduction. The system's ubiquitous nature and goal for monitoring the opioid crisis implemented impactful guidelines to control opioid mismanagement. The U.S. Food and Drug Administration (FDA) guidelines and policies may also be a decisive contributing factor to the progressive decline in codeine distribution. During the period of interest between 2011 and 2019, the FDA instituted policies requiring labeling changes for prescription opioid cough and

cold medicine to limit their use to adults 18 years and older (15). Included in these labeling changes is the addition of safety information about the risks of misuse, abuse, addiction, overdose, death, and slowed or difficult breathing in prescription cough and cold medicines containing codeine (16). These mandated FDA policies are likely another contributing factor to the declining distribution of codeine among different healthcare distributors. According to our regional analysis, 46 of 50 states demonstrated a precipitous decline in codeine utilization when comparing the peak year of 2011 to 2019. However, the four states with an increase in codeine administration were Texas, Oklahoma, West Virginia, and Arkansas. This indicates the presence of a regional disparity in codeine utilizations primarily centered around the southern region of the U.S. in close proximity to Texas, excluding West Virginia. Texas presented with the most notable increase in codeine administration — a 240% increase when comparing 2011 to 2019. According to ARCOS, the spike in codeine distribution for Texas began increasing largely in 2015 and continued to climb with each successive year. The precipitous rise in Texas has been recognized in another study as well (25). A possible explanation for the pronounced regional disparity in Texas, Oklahoma, and Arkansas could be a result

Figure 5. Number of codeine prescriptions per 1,000 Medicaid enrollees as reported by the State Drug Utilization Database (SDUD) from 2014 to 2019.

Figure 4. Distribution of codeine (mg/person) in 2019 as reported by the United States Drug Enforcement Administration’s Automated Reports and Consolidated Orders System (ARCOS). The light blue bars include data within one standard deviation of the mean with “a” denoting point outside one standard deviation and “b” denoting points within one standard deviation but outside the 95% confidence interval. 13

Declining National Codeine Distribution in United States Hospitals and Pharmacies

from the cultural roots of codeine consumption. On the other hand, the regional disparity could be a result of the complex interplay between variations in health status, attitudes and cultural responses to health care, and access to health care in Texas and its neighboring states. Future research should focus on sociocultural/economic factors that may mediate the pronounced regional disparity. Looking to the future, it would be beneficial to further investigate the decreasing trend in codeine in the principal distributors, such as hospitals and pharmacies. Perhaps this decrease resulted from a negative connotation regarding codeine, fear of the drug being misused, or the presence of safer alternatives less likely to cause addiction or abuse such as NSAIDS. Alternatively, it could have been caused by the implementation of stricter guidelines monitoring the control, spread, and prescription of codeine nationally or the adoption of systems designed to distribute codeine more efficiently with the concern about people's health, safety, and well-being. Finally, examining the Medicaid codeine prescriptions per 1,000 enrollees, it is clear that further research is necessary to investigate the 1,718.16% increase in prescriptions between the first quarter of 2014 and the third quarter of 2016. According to Medicaid data, the number of national Medicaid enrollees peaked around 2016 and mirrors the rise and fall of prescriptions per 1,000 enrollees between 2014 and 2019. This could be a contributing reason for the increase of codeine prescriptions observed in 2016, but further research should be conducted to discover other possible factors. Overall, gaining insight to combat the opioid epidemic will be aided by further analysis of codeine prescription trends in the U.S.

Acknowledgments The authors extend their gratitude to Colleen Jordan for her assistance with citations and Iris Johnston for providing access to articles.

Disclosures BJP is part of an osteoarthritis research team supported by Pfizer and Eli Lilly. The other authors have no disclosures.

References

Jalal H, Buchanich JM, Sinclair DR, Roberts MS, Burke DS. Age and generational patterns of overdose death risk from opioids and other drugs. Nat Med 2020; 26(5): 699-704.

Jalal H, Buchanich JM, Roberts MS, Balmert LC, Zhang K, Burke DS. Changing dynamics of the drug overdose epidemic in the United States from 1979 through 2016. Science 2018; 361(6408).

Brown R, Morgan A. The opioid epidemic in North America: Implications for Australia. Trends Issues Crime Crim. Justice 2019; (578): 1-15.

Kuehn, B. NIH strategy to combat opioid crisis. Jama 2017; 318(24): 2418-2418.

Murthy, VH. Ending the opioid epidemic—a call to action. N Engl J Med 2016; 375(25): 2413-2415.

Linares OA, Fudin J, Schiesser WE, Daly Linares AL, Boston RC. CYP2D6 phenotype-specific codeine population pharmacokinetics. J Pain Palliat Care Pharmacother 2015; 29(1): 4-15.

Fulton CR, Zang Y, Desta Z, Rosenman MB, Holmes AM, Decker BS, et al. Drug-gene and drug-drug interactions associated with tramadol and codeine therapy in the ingenious trial. Pharmacogenomics 2019; 20(06): 397-408.

Cartabuke RS, Tobias JD, Taghon T, Rice J. Current practices regarding codeine administration among pediatricians and pediatric subspecialists. Clin Pediatr 2014; 53(1): 26-30.

Gardiner SJ, Chang AB, Marchant JM, Petsky HL. Codeine versus placebo for chronic cough in children. Cochrane Database Syst Rev. 2016; 7(7): CD011914.

10. Rieder MJ, Jong GT. The use of oral opioids to control children's pain in the post-codeine era. Paediatr Child Health 2021; 26(2): 120-123. 11. Cartabuke RS, Tobias JD, Taghon T, Rice J. Current practices regarding codeine administration among pediatricians and pediatric subspecialists. Clin Pediatr 2014; 53(1): 26-30. 12. Chidambaran V, Senthilkumar Sadhasivam MM. Codeine and opioid metabolism–implications and alternatives for pediatric pain management. Curr Opin Anaesthesiol 2017; 30(3): 349. 13. Kohler JE, Cartmill RS, Kalbfell E, Schumacher J. Continued prescribing of periprocedural codeine and tramadol to children after a black box warning. J Surg Res. 2020; 256: 131-135. 14. FDA. Safety review update of codeine use in children; new Boxed Warning and Contraindication on use after tonsillectomy and/or adenoidectomy [Internet]. United States: USFDA; 2013 [cited 2021 May 20]. 4 p. Available from: https://www.fda.gov/media/85072/download 15. FDA. FDA Drug Safety Communication: FDA restricts use of prescription codeine pain and cough medicines and tramadol pain medicines in children; recommends against use in breastfeeding women [Internet]. United States: USFDA; 2017 [cited 2021 May 20]. Available from: https:// www.fda.gov/drugs/drug-safety-and-availability/fda-drugsafety-communication-fda-restricts-use-prescriptioncodeine-pain-and-cough-medicines-and 16. FDA. FDA Drug Safety Communication: FDA requires labeling changes for prescription opioid cough and cold medicines to limit their use to adults 18 years and older [Internet]. United States: USFDA; 2018 [cited 2021 May 20]. Available from: https://www.fda.gov/drugs/drugsafety-and-availability/fda-drug-safety-communicationfda-requires-labeling-changes-prescription-opioid-coughand-cold 17. US Department of Justice. PART 1308 - Section 1308.12 Schedule II [Internet]. United States: USDEA; 1974 [cited 2021 May 20]. Available from: https://www.deadiversion. usdoj.gov/21cfr/cfr/1308/1308_12.htm

Declining National Codeine Distribution in United States Hospitals and Pharmacies

18. US Department of Justice. PART 1308 - Section 1308.13 Schedule III [Internet]. United States: USDEA; 1974 [cited 2021 May 20]. Available from: https://www.deadiversion. usdoj.gov/21cfr/cfr/1308/1308_13.htm 19. US Department of Justice. PART 1308 - Section 1308.15 Schedule V [Internet]. United States: USDEA; 1974 [cited 2021 May 20]. Available from: https://www.deadiversion. usdoj.gov/21cfr/cfr/1308/1308_15.htm 20. US Department of Justice. Automation of Reports and Consolidated Orders System (ARCOS) [Internet]. United States: USDEA; 2021 [cited 2021 May 20]. Available from: https://www.deadiversion.usdoj.gov/arcos/index.html 21. Schaffer AL, Cairns R, Brown JA, Gisev N, Buckley NA, Pearson S. Changes in sales of analgesics to pharmacies after codeine was rescheduled as a prescription only medicine. Med J Aust. 2020; 212(7): 321–327. 22. Ladha KS, Neuman MD, Broms G, Bathell J, Bateman BT, Wijeysundera DN et al. Opioid prescribing after surgery in the United States, Canada, and Sweden. JAMA Netw Open. 2019; 2(9): e1910734. 23. Rui P, Santo L, Ashman JJ. Trends in opioids prescribed at discharge from emergency departments among adults: United States, 2006-2017. Natl Health Stat Report. 2020; (135): 1–12. 24. Jameson H. Don’t Panic Over Mandated E-Prescribing of Controlled Substances Laws. Academy of General Dentistry [Internet]. 2020 [cited 2021 May 21]; Newsroom: [about 1p.]. Available from: https://www. agd.org/publications-and-news/newsroom/newsroomlist/2020/08/10/don-t-panic-over-mandated-eprescribing-of-controlled-substances-laws 25. Ighodaro EO, McCall KL, Chung DY, Nichols SD, Piper BJ. Dynamic changes in prescription opioids from 2006 to 2017 in Texas. PeerJ. 2019; 7: e8108.

Scholarly Research In Progress • Vol. 5, November 2021

Rat-Bite Fever in a 14-Year-Old Male Wyatt L. Keck1†‡, Michael S. Pheasant1†‡, Desiree N. Wagner1, and Lindsay M. Dittman1,2 Geisinger Commonwealth School of Medicine, Scranton, PA 18509 Geisinger Wyoming Valley Medical Center, Wilkes-Barre, PA 18711 † Doctor of Medicine Program ‡ Authors contributed equally Correspondence: wkeck@som.geisinger.edu 1 2

Abstract Rat-bite fever (RBF) is a rare, zoonotic infection with an increasing incidence among the pediatric population as rats become more popular household pets. Here, we present a case of RBF in a previously healthy 14-year-old boy. The patient presented with nonspecific clinical findings and a social history significant for close contact with his pet rat, creating a scenario highly suspicious for RBF. Following clinical diagnosis and appropriate treatment, the patient experienced a rapid recovery and complete resolution of symptoms. In light of this case, we support and encourage the long-standing practice of conducting a thorough social history and inquiring about any pets or animal exposure in the home or place of work.

Introduction Rat-bite fever (RBF) is a rare, systemic, zoonotic infection that is most commonly caused by Streptobacillus moniliformis (North America) or Spirillum minus (Asia), and to a much lesser extent, Streptobacillus notomytis (1, 2). As the name suggests, sustaining a bite from an infected rat is the typical mode of transmission to humans, and each bite carries an approximate 10% risk of infection (1, 2). However, not all infections require a bite. According to current literature, as many as 30% of cases report no history of a bite, suggesting that close contact is another mode through which the disease can be transmitted (1). In fact, it has been reported that 10% to 100% of domestic rats and 50% to 100% of wild rats carry S. moniliformis, which resides in the upper respiratory tract, including the nasopharynx, larynx, upper trachea, and middle ear (1). Additionally, several case reports have documented the transmission of RBF via oral contact in children who kiss their pet rats or following accidental ingestion of food and water contaminated with infected rat feces (3, 4). Historically, RBF was a disease commonly seen among lab technicians and those living in poverty. However, with rats becoming increasingly popular as household pets, the demographics have now changed to include children, representing over half of all cases, laboratory personnel, and pet store employees (1). Here, we present a case of RBF with a typical clinical presentation and resolution following appropriate treatment.

Case Presentation A previously healthy 14-year-old male presented to the Emergency Department with his mother for evaluation of a 3-day history of rash and arthralgia. His symptoms began 3 days prior following a minor ankle injury during gym class. The ankle pain resolved, but new pain had developed in the right elbow.

The day prior to presentation, a rash appeared on his arms and legs and spread to his palms and soles over the course of the day. The trunk and genital area were largely spared. The rash on his soles was painful when walking. Associated symptoms included fever with chills and drenching sweats two days prior to presentation, which had resolved. Upon thorough review of systems, the patient reported a headache, but denied any additional symptoms. He was seen the previous day at another facility and was treated with tramadol and diphenhydramine. Further history revealed that the family spends significant time in the woods but denied known tick bites. The patient and family had no recent travel history. Household pets include two dogs, a rabbit, and a rat. The patient denied any sick human contacts; however, he reported close contact with his pet rat, which he allowed to kiss his nose and clean his teeth. The rat had reportedly been ill recently with sneezing and loud breathing. The patient had an unremarkable birth history, was up to date on his vaccinations and took no medications. Past surgical history was significant for a nasal endoscopy at age 4 for a foreign body removal and a tonsillectomy at age 6. On physical examination, vital signs were as follows: body mass index (BMI) 42.13 kg/m2, temperature 36.72oC, blood pressure 121/72 mmHg, pulse 93 beats per minute, respirations 20 breaths per minute, and oxygen saturation of 98% on room air. Skin exam revealed erythematous macules, papules, and pustules in an acral distribution on all four extremities, particularly on the extensor surfaces and extending to the palms and soles where the lesions were tender to palpation. There was also tenderness to passive extension of the right elbow, with active extension limited to approximately 110 degrees, but no significant effusion or erythema. Cardiovascular, pulmonary, abdominal, and neurological examinations revealed no abnormal findings. Pertinent laboratory results revealed an elevated C-reactive protein and erythrocyte sedimentation rate. The patient’s complete blood count with differential revealed a neutrophilic predominance, while his comprehensive metabolic and respiratory viral pathogen panels were negative for any abnormalities. A tick-borne pathogen panel was also obtained. A differential diagnosis of RBF versus anaplasmosis was considered. After blood cultures were obtained, the patient was started on intravenous ceftriaxone (1 g every 12 hours) and oral doxycycline (100 mg every 12 hours). The rash improved rapidly following initiation of treatment. The right elbow arthralgia was slower to respond but did show improvement prior to discharge. The final results of blood cultures were negative, including those for Babesia microti, Anaplasma phagocytophilum, and Lyme disease. He was discharged home on hospital day 6 to continue his remaining antibiotic course of 500 mg penicillin V potassium every 6 hours for a total of 21 days as an outpatient.

Rat-Bite Fever in a 14-Year-Old Male

Discussion

Conclusion

In the United States, RBF is predominantly caused by infection with S. moniliformis. The organism is a highly pleomorphic, extremely fastidious, gram-negative rod that is both filamentous and nonmotile (1). The typical incubation period ranges from 3 to 20 days, but most patients become symptomatic within 7 days or less (1, 2). In general, symptoms are characterized by fever, chills, migratory arthralgias, and a characteristic rash (1, 2). Abrupt, recurring fevers ranging from 38.0° C to 41.0° C with intense rigors are frequently the earliest manifestations, which resolve within 3 to 5 days. Other symptoms reminiscent of viral syndrome may be observed during the initial phase of infection and include nausea, vomiting, headache, and sore throat. As the disease progresses, half of patients will experience migratory arthralgias with swelling and erythema affecting both large and small joints of the extremities, and most commonly the knee or ankle, which can last up to several years in some patients (1, 2). In nearly 75% of cases, patients will also develop a maculopapular and petechial rash on the extensor surfaces of the extremities and can include exquisitely tender hemorrhagic vesicles on the palms and soles (1, 2).

RBF deserves greater awareness among emergency and pediatric physicians, given its nonspecific presentation, high potential for misdiagnosis or serious complications without treatment, and rising incidence among the pediatric population as rats become increasingly popular household pets (1, 4). A high clinical suspicion for RBF is appropriate for any patient with a known exposure to rats presenting with fever, arthralgias, and a rash, and antibiotics should be initiated early despite negative blood cultures (4). This further highlights the importance of taking a thorough social history and inquiring about any pets in the home. Doing so may additionally provide an opportunity to educate patients on hygienic precautions such as wearing gloves, washing hands, and properly cleaning any bites or scratches to minimize the risk of infection (3).

While the prognosis of adequately treated RBF is quite good, mortality rates have been reported as high as 13% if untreated, with an average mortality rate of 10% (1,2). Although it is possible for the infection to resolve spontaneously without treatment or intervention, it could take up to a year, with increased risk for complications (2). Of the reported complications and causes of death for untreated disease, endocarditis carries the highest risk of mortality (1). Other documented complications of untreated RBF include fulminant sepsis, pneumonitis, meningitis, focal abscesses, septic joint, and adrenal failure (1, 2). With proper antibiotic treatment, however, patients typically experience rapid resolution of symptoms. In some cases, the rash can be slow to resolve, with migratory arthralgias and fatigue persisting for months (1).

References 1.

Elliott SP. Rat bite fever and Streptobacillus moniliformis. Clin Microbiol Rev. 2007 Jan;20(1):13–22.

Gupta M, Bhansali RK, Nagalli S, Oliver TI. Rat-bite Fever. 2021 Apr 7. In: StatPearls [Internet]. Treasure Island (FL): StatPearls Publishing; 2021 Jan–April.

Miraflor AP, Davallow Ghajar L, Subramaniam S, de Abreu FB, Castanedo-Tardan MP, Samie FH, Mann JA, Holmes AV, Tsongalis GJ, Yan S. Rat-bite fever: An uncommon cause of fever and rash in a 9-year-old patient. JAAD Case Rep. 2015 Oct 24;1(6):371–4.

Vetter NM, Feder HM Jr, Ratzan RM. Rat bite fever caused by a kiss. Am J Emerg Med. 2016 Jun;34(6):1190.e3–4.

Regarding appropriate treatment, penicillin is the antibiotic of choice against S. moniliformis. However, for reasons including the nonspecific presentation of RBF, broad list of differential diagnoses, and fastidiousness of S. moniliformis in culture, most patients experience delays in treatment (1, 2). Therefore, even in the absence of positive blood cultures, antibiotics should be initiated if RBF is suspected (4). It is recommended that children receive 12–30 mg/kg/day of IV penicillin G for 5 to 7 days, followed by an additional 7 to 14 days of oral penicillin V at 25–50 mg/kg/day divided into four doses per day. Adults with RBF should be given 240–360 mg of IV penicillin G for 7 days, and if no improvement is observed within 2 days, should be increased to 720 mg per day (1). Alternatively, IV ceftriaxone at 1 gram per day may be given, and for penicillin-allergic patients, oral or IV doxycycline at 100 mg twice daily is appropriate (2). Lastly, general recommendations for avoidance of contact with rat respiratory secretions, handwashing after contact, and proper cleaning of any bites or scratches with antiseptics should be followed, especially for those who frequently work with or keep rodents as pets (2).

Scholarly Research In Progress • Vol. 5, November 2021

Preventability Associated with Elevated Maternal Mortalities Among Black Women Colleen G. Jordan1*, Sophia A. Klevan1*, Kendra C. Benn-Francis1*‡, Ofonime E. Emah1*‡, and Amy L. Kennalley1*‡ ¹Geisinger Commonwealth School of Medicine *Master of Biomedical Sciences Program ‡ Authors contributed equally Correspondence: cjordan02@som.geisinger.edu

Abstract Background: Maternal mortality is defined as the deaths of women during pregnancy while giving birth or soon after birth. The United States is the only developed country having a continuous rise in maternal mortality, with Black women being at the highest risk compared to other racial and ethnic groups. Methods: Data from the CDC WONDER database was used to analyze preventability associated with maternal mortality among Black women from ages 15 to 55 between the years 2010 to 2019. Demographic statistics such as age, year, and geographic location were utilized to complete this analysis. Only preventability associated with antenatal and perinatal maternal deaths among Black women in the United States was examined. Results: From 2010 to 2019, Black women in the United States experienced statistically significant increasing rates of preventable maternal mortality (p < 0.05); statistically significant association was indicated with age group (AMA and non-AMA) (p < 0.005) and geographic location (urban and rural) (p < 0.005). Conclusion: Further research is needed to analyze the racial disparities associated with Black women experiencing a higher maternal mortality rate compared to other racial and ethnic groups. In addition, health care policies and social determinants in other countries need to be investigated to develop new interventions aimed at improving maternal care of Black women in the United States.

Introduction The United States is the only developed country experiencing an increase in maternal mortality, which is a public health crisis in need of critical examination (1–3). Maternal mortality is defined as the deaths of women during pregnancy while giving birth or soon after birth (4). Maternal age is commonly divided into two classifications: non-advanced maternal age (non-AMA) which includes ages 35 and under and advanced maternal age (AMA) which includes ages above 35 during pregnancy (1). The World Health Organization (WHO) divides maternity care into three parts: antenatal, perinatal, and postpartum care (5). Antenatal care focuses on promoting health and wellness through education and referrals (6). Perinatal care, from 22 weeks of gestation through 1 week following birth, includes palliative and preventive measures for mother and child, although maternal mortality rates are highest during this time (7–9). The final phase is postpartum care, and this refers specifically to the mother’s care (10). For this study, we considered the antenatal and perinatal phases of pregnancy-related maternal deaths.

Racial and ethnic disparities in obstetric care and delivery outcomes have shown that Black women experience higher rates of pregnancy-related mortality and morbidity compared to other racial and ethnic groups (11). The condition of the hospital, in addition to the care provided, contributes to the patient outcome. A recent investigation found that racial and ethnic minority women deliver in lower-quality hospitals than white women and receive lower-quality care; this is believed to be correlated with more severe morbidity in Black women (12). Studies indicate that Black women are 2.4 to 3.3 times more likely to die from pregnancy complications than their counterparts in other racial and ethnic groups (12, 13). Additionally, Black women endure substantially higher cesarean delivery rates, which leads to further postpartum complications in addition to the widely known risks involved in major abdominal surgery (12, 13). In Black women, 46% of pregnancyrelated deaths could have been prevented compared to 33% of white women (14). Black women experience elevated rates of pregnancy-induced conditions such as chronic hypertension, asthma, placental disorders, gestational and preexisting diabetes, and blood disorders (15). However, no direct correlation has linked these pregnancy-induced conditions to the disparities surrounding maternal mortality rates of Black women in the United States (12). There is no explanation for why there has been a failure of health care providers to acknowledge these known disparities. Preventive measures are not readily available and accessible to all pregnant women; most notably in regard to hospitals in which Black women predominantly give birth (12). Based on a literature review, this study’s definition of prevention includes continual health checks, logistical access to health resources and nutrition, ability to obtain and take medications, and managing chronic illness and stress (9). There is an undeniable need to improve prenatal and delivery care in the United States, specifically in low-performing hospitals serving a disproportionate ratio of Black women (12). Structural and environmental racism consistently underlie the higher maternal mortality rates for Black women. An approach that generates quality care throughout pregnancy from preconception to postpartum care will be most effective in reducing maternal mortality (12, 16). This study examines associations of common population level factors, such as age and geographic location, to antenatal and perinatal maternal preventable and nonpreventable deaths among Black women in the United States.

Preventability Associated with Elevated Maternal Mortalities Among Black Women

Methods Participants Black females between the ages of 15 to 55 with a cause of death related to pregnancy, childbirth, and puerperium were selected for analysis. The most recent publicly accessible version of the NCHS Urban-Rural Classification Scheme from 2013 was used to categorize the geography of deaths in the database, and no classifications were excluded (17). Inclusion criteria consisted of all weekdays, autopsy values, and all death locations within the United States in order to have comprehensive data that reflects all maternal mortalities within the research parameters. Women of Hispanic, non-Hispanic, or unspecified origin were also included for this investigation. Death rates occurring from certain conditions (see Table 1) originating after the puerperium period (over 42 days following birth) were excluded due to the results being outside the range of the study. Based on these criteria a sample of approximately 2,700 women were analyzed out of the over 21,000 maternal mortalities in the U.S.

Procedures This study utilized secondary data analysis with data obtained from the Centers for Disease Control and Prevention’s Wide-ranging ONline Data for Epidemiologic Research (CDC WONDER) database (18). Starting in 1999, all death certificates in the United States are sent to the CDC and are compiled into the WONDER database. The secondary data analysis focused on maternity-related deaths of Black women from 2010 through 2019 and was collected from the WONDER database on January 29, 2021. Table 2 shows detailed description of variables and criteria entered into database search to obtain the appropriate sample. Demographic statistics such as age, year, geographic location, and preventability were assessed. In order to complete the analysis, causes of death were operationalized into nominal categories: Preventable, Non-preventable, and Undetermined. Causes of death were connected to public and community health resources based on published literature pertinent to the time frame of 2010 through 2019 as they fit into this study’s operationalized definition of preventable (19). Data analysis

Table 1. ICD-10 Code Classification

Table 2. CDC WONDER Request Form Entry Criteria, Retrieved: January 30, 2021

Age at the time of death was examined on a nominal scale including non-AMA (age ≤35) and AMA (age >35) pregnancy groups (18). Geography was used as a variable in nominal categories defined as rural or urban location of death as defined by the 2013 NCHS Urban-Rural Classification Scheme (17). The primary variable examined, which was compared to all other variables, was whether a cause of death is considered preventable. Preventability for each cause was defined by an extensive literature review in combination with this study’s definition. Preventable causes were defined as causes that could have been prevented if a mother had continual health checks, logistical access to health resources and nutrition, the ability to obtain and take medications, and the ability to manage chronic illness and stress (9). Each ICD-10 code was compared with this definition. If a cause met these criteria, it was designated as preventable. If a cause did not, it was non-preventable. If a cause was situationally dependent such as complications from anesthesia or clinical mistakes, it was designated as non-defined to exclude inconclusive data. Deaths considered preventable and not preventable were the primary focus of data analysis. This research used descriptive analyses with the quantitative data. Statistical analyses consisted of chi-squared tests. A chi-square analysis was used to evaluate the preventability of death in Black women versus age during pregnancy.

Preventability Associated with Elevated Maternal Mortalities Among Black Women

Several chi-squared tests comparing the preventability over time were run to capture the difference in preventability of deaths per year to examine the trends. A chi-squared test was run to find differences between preventability and urbanization. Lastly, a chi-square analysis was used to compare the top five causes of death and their preventability. The data was exported directly from the CDC WONDER database. These data were analyzed on GraphPad online software for statistical analyses, and figures were prepared with GraphPad Prism, version 9.0.2 (20, 21).

Results Preventability associated with maternal mortalities among Black women from ages 15 to 55 were examined from 2010 to 2019. The preventable versus non-preventable proportion of maternal mortalities among Black women is presented in Figure 1. Figure 1 presents the death with the highest rates of preventable deaths with 53.5%, followed by non-preventable deaths with 25.0%, followed by non-defined deaths with 21.5% among Black women from 2010-2019. Preventability associated with maternal mortalities among Black women comparing non-AMA and AMA groups are presented in Figure 2. Figure 2 displays a higher number of deaths for both preventable and non-preventable deaths among AMA Black women compared to non-AMA Black women from 2010 to 2019, presenting significant values (X2 (1) = 29.642, p ≤ 0.0001). The odds of experiencing a preventable death during AMA pregnancies is 1.71 times the odds of experiencing a preventable death during non-AMA pregnancies.

Figure 1. Maternal mortality of Black women from 2010 to 2019 categorized as preventable (53.5%), non-preventable (25%), or nondefined (21.5%) as reported by the CDC WONDER database.

Figure 2. Preventable maternal deaths of Black women from 2010 to 2019 as reported by the CDC WONDER database in AMA (age >35) versus non-AMA (age ≤35) pregnancies (** = p < 0.005).

Figure 3 shows a steady trend of non-preventable deaths from 2010 to 2019 (X2 (1) = 10.106, p = 0.0015), 2016 (X2 (1) = 11.348, p = 0.0008), 2017 (X2 (1) = 21.17, p ≤ 0.0001), and 2019 (X2 (1) = 9.047, p = 0.0026). In 2013, there were the highest number of preventable deaths compared to nonpreventable deaths. In 2018, there was a decline in preventable deaths with a subsequent increase in 2019. There was a significant increase for preventable deaths in 2012 (X2 (1) = 5.029, p = 0.0249), 2014 (X2 (1) = 6.956, p = 0.0084) and 2015 (X2 (1) = 12.267, p = 0.0005). Preventability associated with maternal mortalities among Black women comparing rural and urban geographic locations are presented in Figure 4. Figure 4 exhibits significantly higher numbers for both preventable and non-preventable deaths among Black women living in an urban environment compared to a rural environment from 2010 to 2019 (X2 (1) = 14.704, p ≤ 0.0001). The odds of experiencing a preventable death in a rural area is 2.29 times the odds of experiencing a preventable death in an urban area. Preventability associated with maternal mortalities among Black women comparing the top five causes of death to other causes of death are presented in Figure 5. There is a significantly higher number of preventable deaths caused by the top five causes of deaths (pre-existing hypertensive heart disease complicating pregnancy, other specified pregnancy-related conditions, obstetric blood-clot embolism, cardiomyopathy in the puerperium, and diseases of the circulatory system complicating pregnancy; ICD codes 010.1, 026.8, 088.2, 090.3, and 099.4) among Black women, as well as a significantly higher number of non-preventable deaths due to other causes among Black women from 2010 to 2019 (X² (1) = 368.044, p ≤ 0.0001). Figure 5 shows the summaries of the age, geographic, and top five causes explorations.

Discussion Figure 3. Preventable and non-preventable maternal deaths of Black women from 2011 to 2019 compared to preventable and non-preventable deaths of 2010 as a baseline as reported by the CDC WONDER database (* = p < 0.05, ** = p < 0.005).

This study supports the notion that Black women have alarmingly high rates of preventable deaths over deaths of preventable causes, and there are

Preventability Associated with Elevated Maternal Mortalities Among Black Women

surrounding racial and ethnic disparities in obstetric care and delivery outcomes have shown that Black women experience higher rates of pregnancy-related mortality and morbidity compared to all other racial and ethnic groups (11, 12).

Figure 4. Preventable maternal deaths of Black women from 2010 to 2019 as reported by the CDC WONDER database in rural versus urban locations (** = p < 0.005).

There are many complex reasons why Black women are more susceptible to preventable maternal deaths, including factors such as the lack of access to quality medical care, community resources, and low-quality obstetric care. Further explanation of these results could be due to societal factors such as lack of access to quality care and systemic, personal, and environmental racism leading to active discrimination contribute to these disparities. The increase over the 10-year period could be explained by increasing prevalence of effects from various forms of racism. Discrimination caused by racism of health care workers and the health care system itself leads to significant disparities in clinical care and implementation of antiracism intervention is paramount (22). Anti-Black hatred is part of American culture and Black populations have been imposed to live in underserved and physically unhealthy locations (23). The effects of environmental racism continually intensify due to global climate change causing increased health risks and poor outcomes (23). For the AMA and non-AMA comparison, the association could be explained by the increased risk of complications associated with AMA pregnancies in general, as AMA is a strong independent risk factor for morbidity (24). Knowing this association, more frequent and diligent testing and care may reduce the increased odds of preventable deaths in the AMA population. For the urban versus rural exploration, access to community health resources is limited in rural areas, and the increased odds of preventable maternal death in rural areas is likely due to this and other disparities (25, 26).

Figure 5. Preventable maternal deaths of Black women from 2010 to 2019 as reported by the CDC WONDER database in AMA versus nonAMA, rural versus urban, and top five causes of deaths versus all other causes of deaths (** = p < 0.005).

associations with age and geography. In both age groups, AMA and non-AMA, the data consistently displays more preventable maternal deaths than non-preventable maternal deaths for Black women. This proportional trend generally continues from 2010 to 2019 (Figure 2). Looking at 2010 preventable and non-preventable deaths as a baseline, maternal Black women are dying significantly more from preventable deaths than from non-preventable deaths. Specifically, the years 2012–2017 and 2019 show a significantly higher number of preventable deaths than the baseline in 2010. In respect to geographical distribution of rural versus urban areas (Figure 4), there are significantly more preventable maternal deaths than nonpreventable maternal deaths for Black women. In contrast to all other causes of death, the top five maternal mortality causes for Black women are significantly preventable (Figure 5). These data further uncovered the reality of maternal care for Black women in the United States, exposing the need for new public health interventions to improve this area of medicine. The findings from this study indicate that maternal mortality for Black women is primarily from preventable factors and illustrates the need for interventions to avoid Black maternal preventable deaths. In accordance with these findings, studies

The extent and depth of this research was limited due to data suppression when less than 10 deaths occurred sub-nationally (27). Additionally, data is defined as unreliable when the death count is below 20 (27). As this database solely collected data about cause of death, no exact conclusions can be drawn about specific contributing factors on a population or individual level. The data used is comprehensive, as no data was suppressed or hidden at the national level of this analysis. Due to the large sample size, over 2,700, results are reliable and have high statistical power. Further research is needed to compare preventable and non-preventable maternal deaths based on racial and ethnic groups in the United States. Furthermore, investigation on proposed actions to limit preventable deaths should be considered to determine their effectiveness. For a more comprehensive analysis, future research should compare the data on maternal outcomes to other countries. An investigation into best practices from countries with low maternal mortality rates among Black women can help provide a blueprint to solutions for United States health care systems.

Conclusion There is an alarming rate of maternal mortality among Black women in the United States. The findings from this study reveal the incline in mortality rate for Black women from 2010 to 2019 stem primarily from preventable causes. When compared to their counterparts, Black women have been reported to possess elevated risk factors from preexisting conditions and

Preventability Associated with Elevated Maternal Mortalities Among Black Women

pregnancy-induced conditions. Approaches that highlight the need to better understand how social determinants such as socioeconomic status, community, behaviors, beliefs, and barriers to quality care can be addressed to promote optimal outcomes for Black women (15). The United States healthcare system and government need to address racial and ethnic disparities within the maternal mortality crisis by grounding in equitable quality health improvement measures by utilizing comprehensive analyses of race, poverty, and access to health care, with individual narratives as a framework from a human-rights approach (28). Elimination of the racial disparities surrounding maternal mortality should be of utmost importance for the United States. Further research is needed to examine maternal mortality related to maternity care of Black women during antenatal and perinatal phases, as well as during postpartum care, which was not examined in this study. Health policies and social determinants within countries that have low Black maternal mortality also need to be investigated to understand the intricacies of health disparities and interventions that can improve the maternal care of Black women in the United States.

Acknowledgments The authors extend appreciation to Catherine Freeland, MPH, Elizabeth Kuchinski, MPH, and Brian Piper, PhD, for their guidance throughout the writing and investigative processes. Appreciation is also extended to Shantia Horsford for her support.

Disclosures The authors disclose no conflicts of interest.

References 1.

CDC. Maternal Mortality [Internet]. United States: CDC; 2020 [cited 2021 March 9]. Available from: https://www. cdc.gov/reproductivehealth/maternal-mortality/index.html

Martin N, Montagne R. U.S. has the worst rate of maternal deaths in the developed world. NPR [Internet]. 2017 May 12 [cited 2021 March 9]; Lost Mothers:[about 1 p.]. Available from: https://www.npr. org/2017/05/12/528098789/u-s-has-the-worst-rate-ofmaternal-deaths-in-the-developed-world

Gingrey JP. Maternal mortality: A US public health crisis. Am J Public Health. 2020; 110:462–464.

CDC. Pregnancy Mortality Surveillance System [Internet]. United States: CDC; 2020 [cited 2021 March 9]. Available from: https://www.cdc.gov/reproductivehealth/maternalmortality/pregnancy-mortality-surveillance-system.htm

WHO. Maternal and perinatal health [Internet]. Europe: WHO; 2020 [cited 2021 March 9]. Available from: https:// www.who.int/maternal_child_adolescent/topics/maternal/ maternal_perinatal/en/ Chalmers B, Mangiaterra V, Porter R. WHO principles of perinatal care: The essential antenatal, perinatal, and postpartum care course. Birth. 2001; 28:202–207.

WHO. Maternal mortality [Internet]. Europe: WHO; 2019 [cited 2021 March 9]. Available from: https://www.who.int/ news-room/fact-sheets/detail/maternal-mortality

Miller RS, Cummings JJ, Maccauley R, Ralston SJ. Perinatal palliative care: ACOG committee opinion, Number 786. Obstet Gynecol. 2019; 134:e84–e89.

CDC. Prevent pregnancy-related deaths [Internet]. United States: CDC; 2019 [cited 2021 March 9]. Available from: https://www.cdc.gov/vitalsigns/maternal-deaths/index.html

10. Stuebe A, Auguste T, Gulati M. Optimizing Postpartum Care: ACOG Committee Opinion No. 736. Obstet Gynecol. 2018; 131:e140–e150. 11. Tangel V, White RS, Nachamie AS, et al. Racial and ethnic disparities in maternal outcomes and the disadvantage of peripartum Black women: A multistate analysis, 20072014. Am J Perinatol. 2019; 36:835–848. 12. Howell EA. Reducing disparities in severe maternal morbidity and mortality. Clin Obstet Gynecol. 2018; 61:387–399. 13. Eichelberger KY, Doll K, Ekpo GE, Zerden ML. Black Lives Matter: Claiming a space for evidence-based outrage in obstetrics and gynecology. Am J Public Health. 2016; 106:1771–1772. 14. Berg CJ, Harper MA, Atkinson SM, Bell EA, Brown HL, Hage ML, et al. Preventability of pregnancy-related deaths: Results of a state-wide review. Obstet Gynecol. 2005; 106:1228–1234. 15. Warnecke RB, Oh A, Breen N, Gehlert S, Paskett E, Tucker KL, et al. Approaching health disparities from a population perspective: The National Institutes of Health Centers for Population Health and Health Disparities. Am J Public Health. 2008; 98:1608–1615. 16. Howell EA, Egorova NN, Balbierz A, Zeitlin J, Herbert PL. Site of delivery contribution to black-white severe maternal morbidity disparity. Am J Obstet Gynecol. 2016; 215:143– 152. 17. Ingram D, Franco S. 2013 NCHS Urban-Rural Classification Scheme for counties. Vital Health State 2. 2014; 2:1–73. 18. Sauer MV. Reproduction at an advanced maternal age and maternal health. Fertil Steril. 2015; 103:1136–1143. 19. CDC. Underlying Cause of Death 1999-2019 [Internet]. United States: CDC; 2021 [cited 2021 March 9]. Available from: https://wonder.cdc.gov/ucd-icd10.html 20. GraphPad. Analyze a 2x2 contingency table [Internet]. United States: GraphPad; 2021 [cited 2021 March 9]. Available from: https://www.graphpad.com/quickcalcs/ contingency1/ 21. GraphPad. Prism - GraphPad [Internet]. United States: GraphPad; 2021 [cited 2021 March 9]. Available from: https://www.graphpad.com/scientific-software/prism/ 22. Hassen N, Lofters A, Michael S, Mall A, Pinto AD, Rackal J. Implementing anti-racism interventions in healthcare settings: A scoping review. Int J Environ Res Public Health. 2021; 18:2993.

Preventability Associated with Elevated Maternal Mortalities Among Black Women

23. Zimring CA. Clean and white: A history of environmental racism in the United States. New York: NYU Press; 2017. 24. Pinheiro RL, Areia AL, Mota Pinto A, Donato H. Advanced maternal age: Adverse outcomes of pregnancy, a metaanalysis. Acta Med Port. 2019; 32:219–226. 25. National Rural Health Association. About Rural Health Care [Internet]. United states: NRHA; 2021 [cited 2021 May 2]. Available from: https://www.ruralhealthweb.org/ about-nrha/about-rural-health-care 26. Warshaw R. Health Disparities Affect Millions in Rural U.S. Communities [Internet]. United States: AAMC; 2017 [cited 2021 May 2]. Available from: https://www.aamc.org/ news-insights/health-disparities-affect-millions-rural-uscommunities 27. CDC. Multiple Cause of Death 1999-2019 [Internet]. United States: CDC; 2021 [cited 2021 March 9]. Available from: https://wonder.cdc.gov/wonder/help/mcd.html 28. Lu MC, Highsmith K, de la Cruz D, Atrash HK. Putting the ‘M’ back in the Maternal and Child Health Bureau: Reducing maternal mortality and morbidity. Matern Child Health J. 2015; 19:1435–1439.

Scholarly Research In Progress • Vol. 5, November 2021

Recent Trends in Gabapentin Usage Among Medicaid Patients Christopher Logan SanCraint1†‡ and Joshua P. Mills1†‡ ¹Geisinger Commonwealth School of Medicine, Scranton, PA 18509 † Doctor of Medicine Program ‡ Authors contributed equally Correspondence: csancraint@som.geisinger.edu

Abstract Background: Gabapentin is a drug that was originally designed as an anti-epileptic medication in the early 1990s. It was hailed as a sort of “miracle drug” due to its seemingly absent abuse potential and rapidly expanding list of uses including prescriptions for neuropathic pain and migraine treatment. These factors led to a prescription boom, propelling the drug to being one of the most prescribed in the country. However, recent evidence indicates a possibility of gabapentin misuse, particularly in high-risk populations, and subsequent control measures including drug scheduling have been enacted by several states. This study sought to examine current trends in gabapentin prescriptions among Medicaid patients. Methods: We utilized Medicaid state drug utilization data for years ranging from 2014 to 2020. These data sets encompassed units prescribed, reimbursed and amount reimbursed as a total from all 50 states. We isolated data only for gabapentin from the data set and subsequently quantified the trends as percentages of national totals. Results: The analysis of gabapentin prescription trends provided valuable insight to the national sense around gabapentin use. It was shown that gabapentin prescriptions increased each year from 2014 to 2017. From 2017 through 2020, however, there has been a plateau in the number of gabapentin prescriptions and units reimbursed. Lastly, the data analysis showed a decline in reimbursement for gabapentin through Medicaid every year since 2015 continuing through 2020. Conclusion: The results of this analysis indicate that while gabapentin was an extremely popular drug and being prescribed at a high rate, recent control measures and awareness may have led to a slowing or buffering of this trend. The results show that in recent years gabapentin prescription rate has plateaued, pointing to a hesitancy of providers to prescribe the drug. Further investigation is necessary to explain the reduced reimbursement of gabapentin as it could be the result of a combination of forces such as changing Medicaid policy, reduced prescriptions, and other factors. Gabapentin continues to be a popular medication and future examination and review are necessary to assess its abuse potential and usage.

Introduction Gabapentin was originally approved by the FDA in 1993 to treat epilepsy. Its approval for use was expanded to include post-herpetic neuralgia in 2004 (1). Since that point, prescribing and use has exploded, with steadily increasing prescriptions beginning in the 1990s and continuing through the early 2010s (2, 3) This large increase in use can be attributed to its tendency to be prescribed for off-label use. In fact, some studies estimate up to 95% of gabapentin prescriptions are for off-label uses (1). 24

There are several possible reasons for this usage trend. One explanation points to predatory, and illegal, marketing of the drug under the trade name Neurontin by Warner-Lambert (2). Warner-Lambert, who was later acquired by Pfizer Inc., was found to use marketing strategies including continuing medical education presentations and peer-to-peer selling by physicians. These marketing practices led to a boom in gabapentin use, particularly since it was marketed under the guise of research activities and education. These practices were so dominant and widespread that it eventually resulted in litigation, with the ultimate outcome of Warner-Lambert, being found guilty and requiring the payment of a of $430 million fine (2). However, illegal marketing does not solely inform physician prescribing decisions or explain the widespread use of gabapentin. To this day, the specific mechanism of action of gabapentin remains unknown (3). Due to the mystery surrounding its mechanism of action, and the fact that the drug is widely considered to have no misuse potential, off-label prescriptions continued to increase. Other common uses include insomnia, other types of neuropathic pain, alcohol addiction and withdrawal and migraines (3). However, in recent years, the excessive prescribing of gabapentin has become cause for concern amongst some health care providers. This is compounded with increasing evidence that gabapentin does indeed have abuse potential, especially in certain populations, and the drug can have dire side effects (4). Gabapentin was the 10th most prescribed drug in the United States in 2016, a figure which suggests its increased usage is not due to a rise in patients with conditions approved for treatment with gabapentin (4). One possible explanation for this rise in gabapentin prescriptions for a host of issues is the patient expectation to have their pain treated pharmacologically. The pressure on providers from patients to send them home with a prescription may lead them to increase off-label use due to its presumed safety. In addition, writing a prescription tends to be faster and easier than continued follow-ups and patient education (4). Recent literature has raised concerns about indiscriminate prescribing practices (4, 5, 6). First, many of the off label uses of gabapentin have not been rigorously tested in controlled experiments. As a result, the efficacy of gabapentin, and particularly the long-term effectiveness in chronic conditions, is unknown at best (4). Furthermore, gabapentin use can come with serious side effects. One reason for this is the nature of the disorders for which it is being prescribed. Since gabapentin can be effective in treating neurologic disorders, it may be used in conjunction with other drugs which have effects on the central nervous system (4). Patients can also experience withdrawal symptoms when gabapentin therapy is stopped abruptly (4). The widespread use of gabapentin for a variety of disorders has prompted reviews of the original claims that it has little to

Recent Trends in Gabapentin Usage Among Medicaid Patients

no abuse potential. Current evidence suggests that patients with a history of substance abuse disorders or psychiatric disorders are at a higher risk to misuse gabapentin (5). There exists emerging evidence that gabapentin can produce euphoric effects and that continued use can lead to dependence (3, 5). These trends have been described in countries around the world (5, 6). In the United States, recent evidence shows that prescriptions are being filled at three times the recommended dosage (1). In addition, there have been increasing reports of intoxications, suicides and accidents associated with gabapentinoid use (5). Some studies suggest gabapentin abuse and misuse to be as high as 65% among individuals with a prescription (5) — a cause for concern, giving the ever-rising numbers of prescriptions. The potentially dangerous effects of rampant gabapentin use have forced some states to act. In 2017, the Kentucky Board of Pharmacy resolved to make gabapentin a Schedule V drug (7). As of 2020, Kentucky, West Virginia, Tennessee, Michigan, North Dakota, and Virginia classified gabapentin as a Schedule V drug. Several other states now require gabapentin monitoring statewide (8). Although once described as a sort of miracle, cure-all drug, emerging evidence suggests danger associated with indiscriminate gabapentin use. As a result, we seek to determine if these trends have been acted upon by prescribers in the Medicaid program and postulate that gabapentin usage has declined as a result.

Methods Procedures The data used for this analysis was obtained from the Medicaid State Drug Utilization Data (SDUD) set maintained by the Centers for Medicaid and Medicare Services (9). This data represents drug prescriptions and reimbursement in all states from 2014 through 2020. To isolate gabapentin data from the data set specifically, several of its aliases were identified, as follows: Gabapentin, Neurontin, and Gralise. Upon isolation of these specific drugs, quarterly prescriptions, reimbursement, and units prescribed were quantified for each year examined. Subsequently, each of these parameters were normalized against national Medicaid enrollment, total yearly prescriptions, and total reimbursement, respectively. The Geisinger IRB approved this study.

Figure 1. (A) Gabapentin units prescribed to Medicaid patients as a percent of all medication units prescribed that year. (B) Gabapentin prescriptions as a percentage of all prescriptions to Medicaid patients. (C) Comparison of units prescribed with prescriptions as a percentage of total units prescribed and total prescriptions each year, respectively.

Data analysis The data are presented as gabapentin units prescribed as a percent of total units per year, gabapentin prescriptions per 100,000 Medicaid members nationally per year, gabapentin reimbursement amounts as a percent of total prescription reimbursement per year, and gabapentin prescriptions as a percent of total prescriptions per year. To visualize the data, we utilized the GraphPad Prism 9 data analysis tool.

Results After analyzing the number of gabapentin units, which includes pills and patches, as a percentage of the total units of all drugs prescribed to Medicaid members each year, an increase over time has been seen until approximately 2018, when the usage begins to plateau (Figure 1). This trend is concurrent with data shown in Figure 2, which displays gabapentin prescriptions as a function of total prescriptions per year. These data are displayed

Figure 2. Gabapentin prescriptions per 100,000 Medicaid enrollees

Recent Trends in Gabapentin Usage Among Medicaid Patients

may have contributed to the decrease in gabapentin usage, and subsequently, a substantial decrease in the reimbursement for this drug by Medicaid.

Figure 3. Gabapentin reimbursement as a percentage of total Medicaid drug reimbursement.

together in Figure 3. To better emphasize the data, gabapentin prescriptions were analyzed per 100,000 Medicaid members, ensuring that the data was normalized for changes in membership year over year. This data provides the same trend of a plateau in use, as seen below in Figure 4. As a part of the analysis, reimbursement values for gabapentin were calculated as a percentage of total reimbursement of all drugs. Accordingly, it was observed that reimbursement decreased over time. From its peak during the examined period in 2015 at 0.36% of all reimbursement to its most recent value of 0.23% in 2020, gabapentin reimbursement has declined by 63.8% (Figure 5).

Discussion The objective of this study was to better understand trends in gabapentin usage among Medicaid patients during the 2010s, its potential for abuse, and explanations for those trends. While usage was on the rise for many years, whether due to on-label or off-label prescription, the usage has since begun to plateau, beginning approximately in 2018. There are several potential explanations for a plateau in usage: increases in regulation on the prescription of gabapentin, increased awareness of off-label inefficacy, and increased awareness for potential for misuse. As mentioned in the introduction, beginning in 2017, several states which make up a substantial population of Medicaid enrollees have restricted gabapentin as a Schedule V drug. The criteria for reaching this scheduling under the Controlled Substances Act, most importantly, is that the drug in question has a potential for abuse and physical dependence with long term usage (10). Moving this drug to a scheduled classification will require that the prescribers of gabapentin attend training on opiates and controlled substances, with the intent of improving integration of treatments and exploring alternatives with drugs which have a lower abuse potential (11). This push for regulation in the wake of the opioid crisis

The boom in gabapentin prescriptions in the early 2000s can be attributed in part to providers’ willingness to prescribe gabapentin for a myriad of conditions, particularly those outside the scope of its FDA approved use. More recently however, the leveling off in prescriptions can be attributed to a reluctance to prescribe gabapentin for off-label uses, especially considering the possibility for misuse. Furthermore, the emphasis on following evidence-based care does not fall in line with this indiscriminate prescribing. The literature is lacking in clinical trials examining the efficacy of gabapentin for these off-label uses (4). Absence of scientific evidence makes it difficult for providers to assess the benefit of prescribing gabapentin versus the risk of abuse or adverse side effects. The data reflects a decreased tendency to use gabapentin off label as evidence by the stalled rise in prescriptions. Gabapentin’s potential for abuse is well documented, as evidenced by its scheduling in several states and studies indicating its ability to form physical dependence and produce a psychotic effect (5, 7). This is in stark contrast to the original belief that gabapentin had little to no abuse potential and was marketed as such. Our results indicate an increased awareness of the possible negative outcomes that can be associated with indiscriminate gabapentin prescribing. The plateau in gabapentin usage could point to physician’s increased education and perception around the possible results of general overuse of gabapentin.

Conclusion In conclusion, Medicaid State Drug Utilization Data shows that gabapentin was and remains an extremely popular prescription medication. However, we found that the prescription rate and amount of gabapentin units as functions of the totals for all Medicaid patients has remained stagnant since 2017. We also found that the total dollar amount reimbursed for gabapentin in comparison to the total amount reimbursed for prescription drugs through all of Medicaid has steadily declined since 2015. These findings indicate a rising awareness that gabapentin may not have zero misuse potential and may not be a cure-all, especially in treating various types of pain. Further investigation into the trends regarding gabapentin prescriptions, with a special focus on off-label use, is necessary to determine the prevalence and possible risks associated with these practices. Furthermore, increasing control by governmental agencies, namely at the state level will provide opportunity for further investigation and study of gabapentin.

Acknowledgments We would like to thank Brian Piper, PhD, for his continuous invaluable help and guidance during this project. We would additionally like to acknowledge Emily Pocius (MD Class of 2023) and Dhwani Patel (MBS Class of 2021) for their efforts in the analysis during the initial stages of this project.

Recent Trends in Gabapentin Usage Among Medicaid Patients

Disclosures We have no financial conflicts of interest to disclose for this research.

References 1.

Smith RV, Havens JR, Walsh SL. Gabapentin misuse, abuse and diversion: a systematic review. Addiction. 2016 Jul;111(7):1160-74. doi: 10.1111/add.13324. Epub 2016 Mar 18. PMID: 27265421; PMCID: PMC5573873.

Michael A. Steinman, Lisa A. Bero, Mary-Margaret Chren, et al. Narrative Review: The Promotion of Gabapentin: An Analysis of Internal Industry Documents. Ann Intern Med.2006;145:284-293. [Epub ahead of print 15 August 2006]. doi:10.7326/0003-4819-145-4-20060815000008

Pauly NJ, Delcher C, Slavova S, Lindahl E, Talbert J, Freeman PR. Trends in Gabapentin Prescribing in a Commercially Insured U.S. Adult Population, 2009-2016. J Manag Care Spec Pharm. 2020 Mar;26(3):246-252. doi: 10.18553/jmcp.2020.26.3.246. PMID: 32105169; PMCID: PMC7155217.

Goodman CW, Brett AS. Gabapentin and Pregabalin for Pain - Is Increased Prescribing a Cause for Concern? N Engl J Med. 2017 Aug 3;377(5):411-414. doi: 10.1056/ NEJMp1704633. PMID: 28767350.

Hägg S, Jönsson AK, Ahlner J. Current Evidence on Abuse and Misuse of Gabapentinoids. Drug Saf. 2020 Dec;43(12):1235-1254. doi: 10.1007/s40264-02000985-6. PMID: 32857333; PMCID: PMC7686181.

Campbell LS, Coomer TN, Jacob GK, Lenz RJ. Gabapentin controlled substance status. J Am Pharm Assoc (2003). 2021 Mar 2:S1544-3191(21)00027-3. doi: 10.1016/j. japh.2021.01.025. Epub ahead of print. PMID: 33674205.

"Important Notice: Gabapentin Becomes a Schedule 5 Controlled Substance in Kentucky" (PDF). Kentucky State Board of Pharmacy. March 2017. Retrieved 18 June 2018.

Dufrene H. Gabapentin to Become a Controlled Substance in Virginia [Internet]. Carlisle Medical. 2019 [cited 2021Apr27]. Available from: https://www.carlislemedical. com/2019/06/gabapentin-to-become-a-controlledsubstance-in-virginia/

State Drug Utilization Data [Internet]. Medicaid. Centers for Medicare & Medicaid Services; 2021 [cited 2021Mar1]. Available from: https://www.medicaid.gov/medicaid/ prescription-drugs/state-drug-utilization-data/index.html

10. CSA Schedules [Internet]. Drugs.com. [cited 2021May4]. Available from: http://www.drugs.com/csa-schedule.html. 11. Gabapentin Scheduled as Controlled Substance to help with State's Opioid Epidemic [Internet]. LARA - Gabapentin Scheduled as Controlled Substance to help with State's Opioid Epidemic. 2019 [cited 2021May4]. Available from: https://www.michigan.gov/lara/0,4601,7-154-11472487050--,00.html

Scholarly Research In Progress • Vol. 5, November 2021

Assessment and Improvement of Sepsis Bundle Compliance at Geisinger Community Medical Center Sean B. Watson1†‡, Michael S. Pheasant1†‡, Steven A. Picozzo1†‡, John R. Wroblewski1†‡, Jeffrey D. Perluke1†‡, and Igor Georgievsky1,2,3,4 ¹Geisinger Commonwealth School of Medicine, Scranton, PA 18509 ²Geisinger Community Medical Center, Scranton, PA 18510 ³Geisinger Northeast Internal Medicine Residency Program, Scranton, PA 18510 4 The Wright Center Internal Medicine Residency Program, Scranton, PA 18510 † Doctor of Medicine Program ‡ Authors contributed equally. Correspondence: swatson@som.geisinger.edu

Abstract Sepsis is a life-threatening emergency resulting from the body’s overwhelming response to infection. Annually, 1.7 million American adults develop sepsis, and nearly 270,000 will die as a result (1, 2). In an effort to improve patient outcomes, Geisinger Community Medical Center (GCMC) has adopted the recommendations of The Society of Critical Care Medicine and the European Society of Intensive Care Medicine’s Surviving Sepsis Campaign. The campaign consists of a definition of sepsis, severe sepsis, and septic shock as well as prescribed 3-hour and 6-hour treatment bundles (3). We have conducted a retrospective chart review of cases in which criteria for sepsis, severe sepsis, or septic shock was met and in which patients had a poor outcome defined as mortality or hospital readmission within 30 days. Our results demonstrated that in cases of sepsis, the 3-hour bundle compliance was 40.91% with the greatest deficit in proper fluid bolus administration. In cases of severe sepsis and septic shock, 3-hour bundle compliance was 53.13% and 79.17% respectively with deficits in appropriate fluid bolus administration. Six-hour bundle compliance in severe sepsis and septic shock was 31.25% and 33.33% respectively, with deficits in appropriate vasopressor use. Our team has identified multiple avenues for improvement. We suggest the development and implementation of an interdisciplinary sepsis team. Use of interdisciplinary teams is already widely established in the hospital, notably in patients experiencing cardiac arrest or trauma. By involving nonphysician providers in direct patient care, teams may be more effective in mobilizing and delivering hospital resources for patient care. We contend that it is both possible and reasonable to consider the development of an interdisciplinary sepsis team at Geisinger Community Medical Center to increase successful compliance with SCC guidelines.

Introduction Sepsis is the result of the body’s overwhelming response to infection. It is a life-threatening medical emergency that can result in local or diffuse tissue damage, organ failure, and death. These complications can develop quickly and thus physicians must intervene quickly and efficiently when sepsis is suspected. Annually, 1.7 million American adults develop sepsis, and nearly 270,000 will die as a result. Approximately 1 in 3 patients who die in a hospital will have sepsis (1, 2). The Society of Critical Care Medicine and the European Society of Intensive Care Medicine developed the Surviving Sepsis

Campaign (SSC) (3). These guidelines offer a succinct and specific definition of sepsis consisting of vital sign parameters and infection source. Additionally, their efforts resulted in the creation of sepsis bundles: designated orders consisting of laboratory specimens, intravenous fluids, and medication administration designed to maximally improve survivability in sepsis (3). A significant number of observational studies have been conducted regarding compliance with the SCC guidelines and sepsis bundles. These studies demonstrate the compliance is a critical challenge to the success of the SCC in reducing morbidity and mortality (4). One study conducted as a survey of critical care physicians, emergency medicine physicians, and internal medicine physicians demonstrated that compliance with individual bundle components could vary with personal preference. For example, only 62% of respondents reported always giving appropriate antibiotics in compliance with the 3-hour bundle (5). Systematic review and meta-analysis of implemented quality improvement programs demonstrate that these improvement programs are associated with both increased adherence to the sepsis bundles and reduced morbidity and mortality in patients with sepsis, severe sepsis, and septic shock (6). Geisinger Community Medical Center (GCMC) has adopted the Surviving Sepsis Campaign bundles in their effort to minimize mortality and readmission following a diagnosis of sepsis, severe sepsis, or septic shock. Per the SSC, sepsis is defined as: meeting two severe inflammatory response syndrome (SIRS) criteria in the setting of an identified source of infection. Criteria must be met within a 6-hour timeframe to qualify as sepsis (Figure 1). SIRS criteria consist of the following: fever greater than 38º C or hypothermia below 36º C; heart rate greater than 90 beats per minute; respiratory rate less than 20 breaths per minute; serum white cell abnormality defined as either greater than 12,000 wbc/mm3, less than 4,000 wbc/mm3, or greater than 10% band cells (3, 7). If sepsis is identified per the stated parameters, physicians are to initiate the 3-hour sepsis bundle consisting of four steps. 1. Measure blood lactate level. 2. Obtain blood cultures prior to the administration of antibiotics. 3. Administer a broad-spectrum antibiotic. 4. Administer 30 ml/kg crystalloid intravenous fluid bolus.

Assessment and Improvement of Sepsis Bundle Compliance at Geisinger Community Medical Center

Methods Preliminary discussion with staff at GCMC aligned with published studies across multiple Geisinger institutions, which directed investigation toward barriers to bundle compliance. Barriers consisted primarily of uncertainty as to appropriate sepsis alert usage, information overload, and burdensome order sets. We set out to find where bundle compliance may be lacking in our sample population.

Figure 1. Criteria for diagnosis of sepsis, severe sepsis, and septic shock.

Bundle compliance is achieved only if all four steps are implemented correctly within three hours of the patient meeting sepsis criteria (3, 8). One caveat to the 3-hour bundle is in the administration of the 30 ml/kg bolus. Physicians may decline to administer the required fluid amount based on clinical judgment if they feel the volume of fluid specified in the bundle would cause harm to the patient, for example in the case of a patient with pulmonary edema and respiratory compromise. One additional consideration is that physicians may have defaulted to use ideal body weight for bolus volume determination in the event of obesity, which was considered appropriate (3, 6). Severe sepsis and septic shock require further management due to the increased severity of disease (Figure 1). Severe sepsis is defined as meeting criteria for sepsis plus evidence of end organ dysfunction (7). Examples of end organ dysfunction include acute kidney injury or lactate greater than four. Septic shock is sepsis in a patient with hypotension defined as a mean arterial pressure less than 65 mmHg (8). In the event of severe sepsis or septic shock, physicians are to complete the steps of the 6-hour bundle: 1. Compliance with the 3-hour bundle as listed above. 2. Repeat measure of blood lactate. 3. Application of a vasopressor in the event of hypotension unresponsive to the initial fluid bolus. 4. Reassessment of volume status and tissue perfusion by a provider. Compliance is only achieved if the steps of the 3-hour bundle are achieved as described above, and if the additional steps of the 6-hour bundle are performed within 6 hours of when the patient met criteria for severe sepsis or septic shock (3, 8).

Geisinger monitors patients with diagnosed sepsis that results in poor outcome defined as readmission or mortality within 30 days of hospital discharge. Sample size was drawn from Geisinger’s established database of sepsis patients from January 2019 to February 2020. Exclusion criteria included patients who were included in the data set but failed to meet criteria for sepsis, patients admitted to hospice following discharge, and non-unique cases previously reviewed for sepsis bundle compliance by a Geisinger employee. Meeting criteria for sepsis was defined based on the SCC guidelines as well as cases where SCC criteria were not met but the physician diagnosed sepsis. The rationale behind the choice to include diagnosed sepsis that does not meet criteria is to ensure that clinical judgment is appreciated and to analyze all situations in which the bundle should be completed. Seventy-eight cases were found to meet inclusion criteria of both having sepsis/ severe sepsis/septic shock and being unique. Of note, our definitions of severe sepsis & septic shock differ from the official SCC guidelines to align with established Geisinger protocol. Thus, the definitions were expanded to include severe sepsis as having a lactate between 2 to 4 and septic shock also having lactate greater than 4. Using a standardized review method, patient charts were evaluated for sepsis-related admissions in the period of January 2019 to February 2020. Cases were evaluated for the following: 1. Presence or absence of sepsis alert in electronic health record. 2. Hospital fulfillment of overall 3-hour bundle, and its individual bundle components. 3. Hospital fulfillment of overall 6-hour bundle applied exclusively to severe sepsis/septic shock cases where appropriate, and its individual bundle components.

Results The initial data set containing the list of patients with poor outcome following sepsis diagnosis contained 132 patients. Eighty-four patient charts remained after removing charts that had previously been analyzed for bundle compliance. Of these, three patients were sent to hospice and were not included in analysis. Three additional patients were found after systematic chart review to not qualify as meeting criteria for sepsis or carry a sepsis diagnosis and were excluded. The total number of patients meeting inclusion criteria was 78. Of these, 22 met criteria for sepsis, 32 for severe sepsis, and 24 for septic shock (Figure 2).

We retrospectively analyzed prior sepsis cases at GCMC to assess compliance with sepsis bundle items to identify and address areas of improvement in bundle compliance with the goal of enhancing morbidity and mortality of sepsis patients. 29

Assessment and Improvement of Sepsis Bundle Compliance at Geisinger Community Medical Center

Figure 2 outlines results of sepsis case analysis (Figure 3). Sepsis alert was called in 18.18% of cases. Blood cultures were drawn in 77.27%. Appropriate broad-spectrum antibiotic was delivered after cultures were drawn in 81.82% of cases. And serum lactate was measured in 77.27% of cases. The appropriate 30 ml/kg bolus was delivered in 40.91% of patients. Three-hour bundle compliance was obtained in 40.91% of cases. Figures 3 and 4 outline results of severe sepsis and septic shock case analysis respectively (Figures 4 and 5). Sepsis alert was called in 34.38% and 58.33% of cases, respectively. Blood cultures were drawn in 71.88% and 87.50%, respectively. Appropriate broad-spectrum antibiotic was delivered after cultures were drawn in 75.00% and 91.67% of cases, respectively. Initial serum lactate was measured in 87.50% and 91.67% of cases. The appropriate 30 ml/kg bolus was delivered to patients in 37.50% and 58.33%, respectively. Three-hour bundle compliance was 53.13% and 79.17% for

Figure 2. Breakdown of patients assessed.

Figure 3. Bundle compliance by required intervention for sepsis patients.

severe sepsis and septic shock, respectively. Additionally, these cases were analyzed for 6-hour bundle compliance in which lactate was remeasured, vasopressor given, and the patient was reexamined. In severe sepsis cases the percentages were 53.13%, 0.00%, and 68.75%, which combined with 3-hour compliance resulted in an overall 6-hour bundle compliance of 31.25%. In septic shock cases the percentages were 79.17%, 8.33%, and 83.33%, which combined with 3-hour bundle compliance resulted in an overall 6-hour bundle compliance of 33.33%.

Discussion Mortality rates for sepsis range from 10–50% (9). Given the high mortality rate, an algorithmic approach may help streamline treatment, thus improving patient outcomes (6, 8). Physicians at GCMC have a high rate of success with several components of the bundle. Providers are largely compliant in collecting blood cultures, administering appropriate antibiotics, and measuring blood lactate levels. The data additionally suggests several potential areas of improvement in bundle compliance. The sepsis alert tool available in Epic is underutilized, being called in only 18.18% of sepsis cases. The sepsis alert tool has been experimentally demonstrated to enhance a provider’s ability to acknowledge sepsis criteria and rapidly place appropriate orders (10, 11). To date, Geisinger already employs several of the techniques described in papers detailing quality improvement measures at other hospitals, namely the use of a built in EHR sepsis alert, and preformed order sets (11). Tools are not effective however, if they are not utilized appropriately. Our current data is insufficient to establish a cause for the limited use of the sepsis alert. However, possibilities include burdensome interface, redundancy in order sets, and the ability to place necessary orders outside of the sepsis tool. The administration of the 30 ml/kg bolus is another potential for improvement in overall bundle compliance. The administration of this specified fluid bolus had the lowest rate of compliance of all necessary 3-hour bundle components, even when considering provider discretion for comorbid conditions. The research team noted that in many cases of diagnosed sepsis, severe sepsis, and septic shock, there was a fluid bolus delivered to the patient. In many of these cases however, the bolus fell short of 30ml/kg even when considering the patient’s ideal body weight. It is possible that physicians are under-dosing bolus fluid resuscitation in favor of using readily available or recognizable IV fluid volumes. The frequent use of 1-liter or 2-liter bolus volumes reinforces this possibility. Considering the finding, the research team has proposed the creation and addition of a built-in fluid bolus calculator that will automatically calculate a bundle complaint fluid volume. This

Assessment and Improvement of Sepsis Bundle Compliance at Geisinger Community Medical Center

pressors is dependent on individual patient status and provider comfort, improving compliance with these issues requires a multi-faceted approach.

Figure 4. Bundle compliance by required intervention for severe sepsis.

Given the complexity of sepsis management as well as the high demand on provider time and attention in the emergency department, where most sepsis cases are first recognized, the implementation of an interdisciplinary team model may increase bundle compliance. Other institutions have made successful use of an interdisciplinary team approach to managing sepsis (12). One study demonstrated, using 3 months’ data pre- and post-intervention, that not only was compliance improved, but the time to completion of individual bundle components was also improved (13). These interdisciplinary teams are generally composed of emergency or internal medicine physicians, nurses, pharmacists, and may include imaging technicians or infectious disease specialists. These teams are available 24/7 and typically operate in a dedicated role, allowing them to focus time and energy on the sepsis cases that present. (12, 13, 14, 15, 16). The use of interdisciplinary teams is already widely established in the hospital, notably in patients experiencing cardiac arrest or trauma. By involving nonphysician providers in direct patient care, teams may be more effective in mobilizing and delivering hospital resources for patient care (16). We believe that it is both possible and reasonable to consider the development of an interdisciplinary sepsis team at GCMC to increase successful compliance with SCC guidelines. Limitations to this study include a limited sample size. Cases were excluded from the data set if they had been previously analyzed by a Geisinger employee. This was done to expand the overall understanding of 3- and 6-hour bundle compliance at GCMC. It is important that moving forward, our data is joined with the additionally analyzed data to produce a more robust data pool from which to draw conclusions. Additionally, our initial intent was to develop, implement, and measure the outcome of a specific intervention to improve compliance. Due to the scope of the project and time required for accurate case review, we were unable to implement any quality improvement measure. Thus, our data serves as a foundation for future quality improvement measures.

Figure 5. Bundle compliance by required intervention for septic shock patients.

Acknowledgments

novel function would serve as a reminder to use the required volumes as well as decrease the effort involved in calculation.

Cheryl Fritzen developed the standardized chart review method utilized and instructed researchers in proper use of review methodology.

Regarding the 6-hour bundle in the case of severe sepsis and septic shock, vasopressors appear critically ignored. Severe sepsis & septic shock can present with dangerously low blood pressures for which fluid boluses are insufficient. We believe clinicians who use decline vasopressors rely on multiple boluses for quick, ephemeral boosts in blood pressure rather than transitioning to long-term pressors. Because the use of

Disclosures Igor Georgievsky, MD, is an employee of and receives compensation through Geisinger Community Medical Center. The authors have no other disclosures.

Assessment and Improvement of Sepsis Bundle Compliance at Geisinger Community Medical Center

References 1.

Centers for Disease Control and Prevention. “What is Sepsis [Internet]. Washington, D.C.: Department of Health and Human Services. [Updated 2021 Jan 27; cited 2021 May 21]. Available from: www.cdc.gov/sepsis/what-issepsis.html Liu V, Escobar GJ, Greene JD, et al. Hospital deaths in patients with sepsis from 2 independent cohorts. JAMA [Internet]. 2014 [Cited 2021 May 21];312(1):9092. Available From: https://pubmed.ncbi.nlm.nih. gov/24838355/

Michael D, Howell, MD. Management of Sepsis and Septic Shock. JAMA Network [Internet]. 28 Feb. 2017 [Cited 2021 May 21]. Available from: jamanetwork.com/journals/jama/ fullarticle/2598892

McVeigh S. Sepsis Management in the Emergency Department. Nursing Clinics of North America [Internet]. 2020 [Cited 2021 July 15]; 55(1): 71-79. Available from: https://www.sciencedirect.com/science/article/abs/pii/ S0029646519300829?via%3Dihub

Djurkovic S, Baracaldo J, Guerra J, et al. A survey of clinicians addressing the approach to the management of severe sepsis and septic shock in the United States. Journal of Critical Care [Internet]. 2010 [Cited 2021 July 15]; 25(4): 658.e1-658.e6. Available at https://www.sciencedirect. com/science/article/pii/S0883944110000882?casa_ token=k_EIGopbTnQAAAAA:LdP2jW7oZIYL MERCgki6SH3MSppU-u-gWaBPCgkwzsegZRrcYoVOf1H1fIYrvCHBJrcZvFELA#tbl1

Damiani E, Donati A, Serafini G, et al. Effect of performance improvement programs on compliance with sepsis bundles and mortality: a systematic review and meta-analysis of observational studies. PLoS One [Internet]. 2015 [Cited 2021 July 15]; 6;10(5):e0125827. Available from: https:// pubmed.ncbi.nlm.nih.gov/25946168/

Singer M, Deutschman CS, Seymour CW, et al. The Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3). JAMA [Internet]. 2016 [Cited 2021 May 21] ;315(8):801-810. Available from: https://pubmed.ncbi. nlm.nih.gov/26903338/

Rhodes A, Evans L, Alhazzani W, et al. Surviving Sepsis Campaign: international guidelines for management of sepsis and septic shock. Intensive Care Medicine [Internet]. 2017 [Cited 2021 May 21]; 34:304-377. Available from: https://link.springer.com/article/10.1007/ s00134-017-4683-6?source=content_type:react|first_ level_url:article|section:main_content|button:body_link DOI:10.1097/CCM.0000000000002255

Neviere R. Sepsis syndromes in adults: epidemiology, definitions, clinical presentation, diagnosis, and prognosis. Up to Date [Internet}. 2020 Feb 03 [Cited 2021 May 21]. Available from: https://www.uptodate.com/contents/ sepsis-syndromes-in-adults-epidemiology-definitionsclinical-presentation-diagnosis-and-prognosis

10. Hayden G, Tuuri R, Scott R, et al. Triage sepsis alert and sepsis protocol lower times to fluids and antibiotics in the ED. The American Journal of Emergency Medicine [Internet]. 2015 [Cited 2021 July 16]; 34(1): 1-9. Available from: https://pubmed.ncbi.nlm.nih.gov/26386734/ 11. Gatewood M, Wemple M, Greco S, et al. A quality improvement project to improve early sepsis care in the emergency department. BMJ Quality and Safety [Internet]. 2015 [Cited 2021 July 16]; 24(12):787-95. Available from: https://pubmed.ncbi.nlm.nih.gov/26251506/ 12. Delawder J, Hulton L. An Interdisciplinary Code Sepsis Team to Improve Sepsis-Bundle Compliance: A Quality Improvement Project. Journal of Emergency Nursing [Internet]. 2020 [Cited 2021 July 16]; 46(1):9198. Available from: https://pubmed.ncbi.nlm.nih. gov/31563282/ 13. Grek A, Booth S, Festic E, et al. Sepsis and Shock Response Team: Impact of a Multidisciplinary Approach to Implementing Surviving Sepsis Campaign Guidelines and Surviving the Process. American Journal of Medical Quality [Internet]. 2017 [Cited 2021 July 16]; 32(5): 500-507. Available from: https://journals.sagepub.com/ doi/10.1177/1062860616676887 14. Ju T, Al-Mashat M, Rivas L, et al. Sepsis Rapid Response Teams. Critical Care Clinics [Internet]. 2018 [Cited 2021 July 16]; 34(2): 253-258. Available from: https://www. criticalcare.theclinics.com/article/S0749-0704(17)300982/fulltext 15. Sebat F, Johnson D, Musthafa A, et al. A multidisciplinary community hospital program for early and rapid resuscitation of shock in nontrauma patients. Chest [Internet]. 2005 [Cited 2021 July 16]; 127(5): 1729-43. Available from: https://linkinghub.elsevier.com/retrieve/pii/ S0012-3692(15)34744-9

Scholarly Research In Progress • Vol. 5, November 2021

Conflicts of Interest Differ Among Male and Female Pediatric Journal Authors Rebecca L. Petlansky1†, Amadea D. Bekoe-Tabiri2‡, Vanessa N. Bueno3‡, AnnMarie N. Onwuka3‡, Michael R. Gionfriddo4,5, and Brian J. Piper1,4 ¹Geisinger Commonwealth School of Medicine, Scranton, PA 18510 ²Bryn Mawr College, Bryn Mawr, PA 19010 ³The University of Scranton, Scranton, PA 18510 4 Center for Pharmacy Innovation and Outcomes, Geisinger, Forty Fort, PA 18704 5 School of Pharmacy, Duquesne University, Pittsburgh, PA 15282 † Doctor of Medicine Program ‡ Authors contributed equally Correspondence: rpetlansky@som.gesinger.edu

Abstract The Physician Payments Sunshine Act requires disclosure of payments made by drug and medical device manufacturers to physicians or teaching hospitals. Academic literature extensively documents gender disparities in the medical profession regarding salary, promotion, and government-funded research. This investigation sought to quantify potential conflicts of interest (CoIs) in pediatric medical journals, specifically examining sex differences. To identify potential CoIs, we examined manuscripts published prior to 2019 in six pediatric journals (JAMA Pediatrics, Pediatrics, The Journal of Pediatrics, Pediatric Blood and Cancer, Pediatric Critical Care Medicine, and The Pediatric Infectious Disease Journal). We collected physician demographics and specialty from the National Plan and Provider Enumeration System National Provider Identifier Registry and compensation data from both ProPublica’s Dollars for Docs (PDD) and Centers for Medicare and Medicaid Services Open Payments (CMSOP). Data was collected on 2,747 authors from 929 manuscripts. Of the 1,088 authors based in the United States with medical degrees (40.5% female), 510 (46.9%) had entries in PDD and CMSOP. Overall, 11,791 payments to these physician-authors totaled $9,586,089.97. Males were 1.20 times more likely to receive payments (RR = 1.20, 95% CI [1.05, 1.37], p = 0.008). The mean amount received by males was $23,250.71 while that received by females was $10,970.78 (mean difference = 12,279.92, 95% CI = (2036.31, 22,523.54), p = 0.019). A comprehensive understanding of these CoIs can inform the disclosure policies of journals to promote transparency of authors.

Introduction A conflict of interest (CoI) exists when judgment regarding an individual’s primary interest, such as patient welfare or research integrity, may be unduly influenced by secondary interests, such as financial interests (1). Medical journals require authors to disclose their relationships with pharmaceutical and medical device companies because such relationships can serve as potential CoIs. However, these potential CoIs are not always reported, and disclosures are rarely verified (2, 3). Verification of CoIs, was facilitated by the passage of the Physician Payments Sunshine Act, which requires public reporting of payments made to physicians and teaching hospitals from pharmaceutical and medical device companies (4). This payment data can be explored using databases such

as the Centers for Medicare and Medicaid Services Open Payments (CMSOP) program and ProPublica’s Dollars for Docs (PDD) (5, 6). CMSOP is a systematic nationwide effort to report these payments to the public. PDD database also contains payment information made to physicians and teaching hospitals and serves as a tool to search these payments. Several studies have characterized CoIs including studies examining guidelines (7), point of care databases (8), textbooks (9, 10), and journal articles (11). Additionally, several reports have documented a disparity in payments between males and females (12, 13, 14). One area where CoIs have not been characterized is pediatrics. The proportion of women in pediatrics has grown over the past several decades such that women now represent most pediatricians (15,16). Our objective was to quantify potential CoIs of authors publishing in pediatric medical journals and determine any differences in payments made to male or female physicians.

Methods Procedures A total of six pediatric medical journals were selected based on journal rankings on Scimago Journal and Country Rank (17), a publicly available portal that provides bibliometric indicators of journals. We attempted to select high-impact journals with a large proportion of authors who are physicians practicing in the United States. We chose three general pediatric journals (JAMA Pediatrics 2018 Scimago Journal Rank [SJR] = 5, Pediatrics SJR = 3, and The Journal of Pediatrics SJR = 1.2) and three pediatric subspecialty journals (The Pediatric Infectious Disease Journal SJR = 1.3, Pediatric Blood and Cancer SJR = 1.3, and Pediatric Critical Care Medicine SJR = 1.3). Within each journal, we began examining articles published in December 2018 and continued in reverse chronological order with the intention to retrieve at least 100 authors with entries in CMSOP (3) and PDD (4). For each article, the first author’s name was recorded. If the author had a medical degree and was located in the United States, his or her name was entered into both CMSOP and PDD to determine if financial compensation was received from pharmaceutical or medical device companies. CMSOP and PDD contain data from payment reports released by the Centers for Medicare and Medicaid Services. Physicians

Conflicts of Interest Differ Among Male and Female Pediatric Journal Authors

are included in the database if payments total $10 or more. CMSOP categorizes payments as general payments, research payments, associated research funding, and ownership and investment interest. General payments include fees related to consulting, services other than consulting, travel and lodging, food and beverage, honoraria, education, royalties or licenses, speaking at an accredited/certified education program, and speaking at an unaccredited/non-certified education program. A total payment amount and number of payments is provided. Payment amounts and number of payments provided by specific companies or within specific categories of payment can also be viewed. PDD also provides a total payment amount and number of payments in the category of general payments, which includes promotional speaking, consulting, meals, travel, and royalties. Financial compensation that was received by the authors between 2015 and 2018 (a timeframe that included 36 months prior to publication, the standard for reporting potential CoIs) and that was classified as “general payments” on CMSOP was recorded. Payments classified as research payment, associated research funding, or ownership and investment interest were not included. Total payment amount and number of payments on PDD was also recorded for each of the years. The author’s name was also entered into the National Plan and Provider Enumeration System National Provider Identifier (NPPES NPI) Registry and degree (DO or MD), sex, specialty, and state of practice were also recorded. This process was repeated for the last author of the article as well as a middle author. Middle author was chosen based on their possession of a medical degree. If multiple middle authors had medical degrees, the individual with the highest compensation according to PDD and CMSOP was reported for inclusiveness.

Data analysis We summarized general payments to physicians using descriptive statistics including the proportion of physicians receiving payments and amounts received. We also conducted analyses based on journal type (general vs. specialty), sex (male vs. female), and payment database (CMSOP and PDD). Payment database was examined as prior work identified minor discrepancies across databases (9). To compare across groups, we utilized t-tests and Mann Whitney tests for continuous data and chi-square tests for categorical data. All statistical analysis was performed in SYSTAT (Version 13, Systat Software, Inc, San Jose, California). GraphPad Prism (Version 8, GraphPad Software Inc., San Diego, California) was used to generate figures.

Results We examined 2,747 authors from 929 manuscripts. Of these authors, 1,659 were excluded because they did not hold a medical degree or were located outside of the U.S. (Figure 1). Of 1,088 authors that were U.S.-based physicians (40.5% female), 510 (36.3% female) had received general payments within 2015 to 2018. Overall, 11,791 payments to these 510 physicians totaled $9,586,089.97. There was a strong correlation between data collected from PDD and CMSOP (r = 1.00, p < 0.0001), which ensured accuracy of recording of payments. Forty-one percent of authors who had published in general (JAMA Pediatrics, Pediatrics, or The Journal of Pediatrics) journals had received payments, while 54.3% of those who had published in subspecialty (Pediatric Critical Care Medicine, Pediatric Infectious Disease Journal, or Pediatric Blood and Cancer) journals had received payments. Authors who had published in subspecialty journals were 1.32 times more likely to have

Figure 1. Flow diagram showing the pediatric journals, manuscripts, and authors receiving industry payments as reported by ProPublica’s Dollars for Docs. 34

Conflicts of Interest Differ Among Male and Female Pediatric Journal Authors

Figure 2. Top 10 compensated authors as reported by ProPublica’s Dollars for Docs (2015-2018) for general (top) and subspecialty (bottom) pediatrics journals. Male authors (49) are shown in blue and females (11) in pink.

received payments than those who had published in the general pediatrics journals (RR = 1.32, 95% CI 1.16-1.50, p < 0.0001). Although the mean amount received was higher for authors of articles in the general pediatrics journals (mean difference = 5,640.83, 95% CI = (-6,377.37, 17,659.03) and p-value = 0.357), the difference was not statistically significant. Of 438 female authors, 42.2% received payments. Overall, males were 1.20 times more likely to have received payments than females (RR = 1.20, 95% CI 1.05-1.37, p = 0.008). The mean amount received by males was $23,250.71, while that received by females was $10,970.78 (mean difference = 12,279.924, 95% CI = (2036.31, 22,523.54), p = 0.019). The majority of top ten earners from each journal were males (Figure 2). Of 643 male authors who met inclusion criteria, 50.5% received payments.

Discussion Over 45% of U.S.-based physicians who authored articles published in six high-impact pediatric journals had received payments of at least $10 or more within 36 months prior to their publication. Physicians who had published in a subspecialty journal were more likely to have received payments. Disparities between payments made to male and females were also observed. Of the physicians who received payments, males were both significantly more likely to have received payments and, on average, had received higher amounts than females. These results align with previous findings that have documented a disparity between industry payments to male and female physicians of various specialties, including male and female otolaryngologists and radiation oncologists (12, 13). One report that examined payments to practicing physicians of various specialties found that women received approximately $3,600

per year fewer total dollars from industry in comparison to men (14). Because relationships between industry and physicians have the potential to increase understanding of and access to new technology and drugs, the disparity in these relationships between males and females may ultimately have a negative outcome on a female physician’s advancement of her career (12). Disparities between male and female physicians, beyond industry payments, are well documented. For example, previous analyses have found that on average female physicians are paid less and are less likely to publish articles, speak at conferences, or obtain grant funding (18, 19). These disparities persist despite increasing numbers of females entering medicine, including pediatrics (15). This may be due to lack of access to mentoring, childcare responsibilities, and other non-careeroriented responsibilities (13). Further research is needed to accurately elucidate the cause of these differences in payments to pediatricians. We found a perfect correspondence between CMSOP and PDD for amount received and the number of payments. A prior investigation noted that, on rare occasions, there were occasionally discrepancies between these databases (8) which presumably was due to contested payments which were corrected in CMSOP before PDD. There are some important limitations to note for this study. First, we only considered physicians practicing in the U.S. and did not account for the potential CoIs of authors who hold other degrees or who live outside the U.S., which is a group that accounted for 60.4% of authors. Similar tools have been developed in Australia (19) and the Netherlands (20), but no studies have examined cross country comparisons. Future studies should examine CoIs among non-U.S. authors and the extent to which these disparities persist across countries.

Conflicts of Interest Differ Among Male and Female Pediatric Journal Authors

Second, choosing the middle author based on receiving payments may have artificially elevated the values we obtained. However, when we looked at articles rather than authors, we found that 68% of articles had at least one U.S. physician who had received payments. Overall, we determined that two-fifths of pediatric authors in high-impact journals had significant CoIs totaling $9.6 million. We also observed sex differences between the payments received, and thus potential CoIs, of male and female pediatric journal authors. Future studies should assess the accuracy of disclosure as well as monitor the trends in the CoIs of journal authors and sex disparities among authorship and industry sponsorship. Because CoIs will continue to have the potential to influence patient welfare and research integrity, future studies will be important in continuing to quantify and describe potential CoIs as new data becomes available annually.

Disclosures BJP is part of an osteoarthritis research team supported by Pfizer and Eli Lilly. The other authors have no disclosures.

References

10. Piper BJ, Alinea AA, Wroblewski JR, et al. A quantitative and narrative evaluation of Goodman and Gilman's Pharmacological Basis of Therapeutics. Pharmacy (Basel) 2019; 8(1):1. doi: 10.3390/pharmacy8010001. 11. Riechelmann RP, Wang L, O'Carroll A, Krzyzanowska MK. Disclosure of conflicts of interest by authors of clinical trials and editorials in oncology. Journal of Clinical Oncology. 2007 Oct 10;25(29):4642-7. 12. Eloy JA, Bobian M, Svider PF, Culver A, Siegel B, Gray ST, Baredes S, Chandrasekhar SS, Folbe AJ. Association of gender with financial relationships between industry and academic otolaryngologists. JAMA Otolaryngol Head Neck Surg. 2017;143(8):796-802. 13. Weng JK, Valle LF, Nam GE, Chu FI, Steinberg ML, Raldow AC. Evaluation of sex distribution of industry payments among radiation oncologists. JAMA Network Open. 2019;2(1):e187377. 14. Rose SL, Sanghani RM, Schmidt C, Karafa MT, Kodish E, Chisolm GM. Gender differences in physicians' financial ties to industry: A study of national disclosure data. PLoS One. 2015;10(6):e0129197. doi: 10.1371/journal. pone.0129197.

Boyd EA, Bero LA. Improving the use of research evidence in guideline development: 4. Managing conflicts of interests. Health Res Policy Syst. 2006;4(1):1-6.

15. Age and Gender Distributions of Pediatricians. Accessed 5/16/2021 at: https://downloads.aap.org/AAP/Images/ pyramid2019.png

Cooper RJ, Gupta M, Wilkes MS, Hoffman JR. Conflict of interest disclosure policies and practices in peerreviewed biomedical journals. J Gen Internal Med. 2006;21(12):1248-52.

16. Pediatricians’ Gender Distribution. Accessed 5/16/2021 at: https://downloads.aap.org/AAP/Images/gender.png

Taheri C, Kirubarajan A, Li X, Lam AC, Taheri S, Olivieri NF. Discrepancies in self-reported financial conflicts of interest disclosures by physicians: a systematic review. BMJ Open. 2021 Apr 1;11(4):e045306.

Physician Payments Sunshine Act of 2009, S. 301

OpenPaymentsData.CMS.gov Accessed 5/11/2021 at: https://openpaymentsdata.cms.gov/search/physicians/byname-and-location

ProPublica’s Dollars for Docs, Accessed 5/11/2021 at: https://projects.propublica.org /docdollars

Neuman J, Korenstein D, Ross JS, Keyhani S. Prevalence of financial conflicts of interest among panel members producing clinical practice guidelines in Canada and United States: cross sectional study. BMJ 2011; 343: d5621 doi:10.1136/bmj.d5621.

Chopra AC, Tilberry SS, Sternat KE, Chung DY, Nichols SD, Piper BJ. Quantification of conflicts of interest in an online point-of-care clinical support website. Sci Eng Ethics. 2020; 26(2):921-930. doi: 10.1007/s11948-019-00153-9.

Piper BJ, Lambert DA, Keefe RC, Smukler PU, Selemon NA, Duperry ZR. Undisclosed conflicts of interest among biomedical textbook authors. AJOB Empir Bioeth. 2018; 9(2):59-68. doi: 10.1080/23294515.2018.1436095.

17. Scimag Journal and Country Rank. Accessed 5/11/2021 at: https://www.scimagojr.com/journalrank.php 18. Spector ND, Overholser B. Examining gender disparity in medicine and setting a course forward. JAMA Network Open. 2019 Jun 5;2(6):e196484-. 19. Berg EJ, Ashurst J. Patterns of recent National Institutes of Health (NIH) funding in general surgery: analysis using the NIH RePORTER system. Cureus. 2019 Jun;11(6). 20. Medicines Australia Transparency Reporting, Accessed 5/16/2021 at: https://www.disclosureaustralia.com.au/ search/ 21. Stichting Transparantieregister Zorg, Accessed 5/16/2021 at: transparantieregister.nl

Scholarly Research In Progress • Vol. 5, November 2021

Ethics and Current Climates Surrounding HPV Vaccination Yezhong Lu1† ¹Geisinger Commonwealth School of Medicine, Scranton, PA 18509 † Doctor of Medicine Program Correspondence: ylu@som.geisinger.edu

Abstract Human papillomavirus (HPV) is the most prevalent sexually transmitted infection (STI) worldwide, affecting both males and females. Roughly 150 HPV types have been identified, 35 to 40 of which can infect the anogenital tract. Infection with certain oncogenic strains is associated with cancers of the cervix, anus, vulva, vagina, and oropharynx. Although vaccination against HPV has been shown to be safe and effective at preventing HPV infection, controversies surrounding the widespread uptake of the available HPV vaccines persist. We examine the notable ethical debates surrounding HPV vaccination programs, including issues of cost efficacy, adolescent autonomy and consent, parental involvement in vaccine decision-making, male inclusion in vaccination programs, and perceived barriers to vaccination. A common theme that emerged within many of the topics we present involved the dynamics of adolescent-parent relationships, which we consider at length to inform future interventional strategies. Despite the safety and efficacy of the HPV vaccine, the potential of becoming infected by HPV remains high, likely due to overall low vaccination uptake. In this review, shortcomings in previous vaccination programs are addressed, while interventions that may enable the success of future strategies are highlighted.

Introduction Human papillomavirus (HPV) is the most prevalent sexually transmitted infection (STI) worldwide (1, 2). When it was discovered in the 1980s that HPV had a causal relationship to cervical cancer (1), vaccination research and investigation into the strategic implementation of HPV vaccination programs came to the forefront of international public health. While the HPV vaccination has been shown to be largely effective in combating cervical cancer and other HPV-associated cancers, the advent of national HPV vaccination programs has not been without ethical controversy. In this review we introduce and examine notable ethical debates surrounding the adoption of HPV vaccination programs. We explore issues including cost efficacy, adolescent decisionmaking capacity, consent, the role of sexual behavior, male inclusion in vaccination programs, educational strategies, and potential conflicts of interest. As these topics are presented, we will discuss how they relate to ethical considerations in the context of patient autonomy, beneficence, and social justice. The synthesis of these considerations is pertinent to the development and application of HPV-related public health policy.

Methods The electronic database PubMed was the primary source for article identification. The database was searched for articles from the years 2000 to 2021. Appropriate free text and MeSH

terms were used to identify all studies. Articles identified using search term “vaccine,” “epidemiology,” “etiology,” “prevention,” “gender,” “consent,” “parents,” “program,” “ethics,” and “education” were paired with “HPV” or “human papillomavirus.” Articles were independently appraised and assessed for quality, and reference lists were scanned for additional studies of potential relevance.

Discussion Background HPV is the most common sexually transmitted disease in the United States (2). To date, more than 100 HPV serotypes have been identified. Although the mechanisms by which these serotypes manifest are highly variable, ranging from innocuous lesions to various forms of cancer, transmission occurs primarily through direct skin-to-skin contact (3). In 1996, the World Health Organization, along with the International Agency for Research on Cancer, recognized the link between HPV and cervical cancer (1). Since then, the association between oncogenic strains of HPV and cervical cancer has been well established, with HPV types 16 and 18 associated with 70% of cervical cancers and cervical intraepithelial neoplasia (CIN) (4). Furthermore, oncogenic strains of HPV have been linked to cancers of the anus, vulva, vagina, and oropharynx (2). Indeed, while HPV has been implicated as a common cause of cervical cancer among women, it is notable that, among males, the incidence of HPV associated penile, anal, and oropharyngeal cancers are on the rise (5). Worldwide, cervical cancer is attributable to more than 250,000 deaths per year, having a disproportionately higher incidence rate in developing countries where resources for preventive methods and secondary treatment are scarce (2). Epidemiological studies have shown that, following the introduction of the Papanicolaou (Pap) cytology test, and the implementation of well-organized screening programs, rates of cervical cancer mortality have significantly declined (6–9). Despite these efforts, cervical cancer remains the second most common cancer among women worldwide (2).

Discussion HPV prevention Engaging in sexual activity at an early age, or having multiple partners, increases the risk of HPV infection. Likewise, “safesex” practices and condom usage as a method of contraception cannot guarantee protection from HPV infection, as transmission can occur via skin-to-skin contact (2). Thus, the primary means by which HPV infection is prevented is via vaccination.

Ethics and Current Climates Surrounding HPV Vaccination

HPV vaccine In 2006, the World Health Organization recognized the high efficacy of the tetravalent HPV vaccine in preventing genital warts and CIN associated with HPV 16 and 18 (2). Currently, the vaccine is recommended for adolescents 9 to 14 years of age and is most effective when administered prior to the onset of sexual relations. The HPV vaccine is safe and effective at preventing HPV infection, and reported side effects are rare, generally mild, and no vaccine-related deaths have been reported (2). There is also a bivalent HPV vaccine that is recognized to prevent CIN associated with HPV 16 and 18, but it does not protect individuals against HPV serotypes associated with genital warts (10). The most recent FDAapproved HPV vaccine is a nonavalent serotype that protects against 5 additional strains of HPV that have been linked to 20% of cervical cancers not previously protected by the tetravalent HPV vaccine (11). Male vaccination, herd immunity, cost efficacy Advocates for HPV vaccination of males emphasize the legitimate health risks HPV infection and associated sequelae confer upon the male sex. According to CDC data gathered between 2011 and 2015, while 60% of HPV-related cancers occur in women, 40% of occur in men. While cervical cancer is commonly attributed as the largest contributor to HPVassociated morbidity and mortality, HPV has also been linked to a substantial number of non-cervical cancers, including vaginal and vulvar cancer in females, and anal, oropharyngeal, and penile cancers in men (2). In fact, one study published in the Journal of Clinical Oncology projected that by 2020 the annual number of HPV-positive oropharyngeal cancers will surpass the annual number of cervical cancers, with the majority occurring in men (5). These considerations raise an issue as to whether female-selective vaccination programs unjustly exclude males. In addition to the direct health benefits that male vaccination provides against male HPV-related infections and sequelae, proponents of male HPV vaccination programs also argue that male inclusion will contribute to more comprehensive protection for females against HPV infections and cervical cancer. Because males comprise half of the epidemiological chain of HPV transmission, some argue that gender-neutral vaccination may be a more effective strategy to combat HPV transmission. By immunizing both sexes, it is argued that a collective, or herd, immunity effect can be created, which would help further mitigate the risk of HPV exposure to females and unvaccinated individuals of both sexes (2). One dynamic transmission model examined reduction rates in HPV-16 prevalence in unvaccinated females in response to a female-only vaccination program versus a male and female program and found that, assuming 80% vaccine coverage in both programs, the gender-neutral program resulted in a 86– 96% reduction in prevalence compared to a 7–31% reduction in the female-only program (12). While the data remain disputed, if it can be definitively demonstrated that the vaccination of males contributes significantly to herd immunity and thus, by extension, confers a population-wide benefit on the reduction of HPV transmission, then “public-good” or social justice arguments for male HPV vaccination may be valid.

Health economists proposing subsidies for female-only vaccination programs base their arguments on cost-effective means to optimally reduce disease burden. According to a study evaluating vaccination programs via disease transmission modeling, “vaccination against HPV-16 and HPV-18 can be costeffective, although including male participants in a vaccination program is generally not cost-effective, compared to femaleonly vaccination” (13). While the relative benefits of male vaccination programs remain largely debated, it appears that the contribution of said programs to HPV eradication is more significant under certain conditions. The cost efficacy of male vaccination programs appears to depend on a multitude of factors, including vaccine efficacy, female vaccination coverage, and the breadth of HPVrelated outcomes considered. According to the CDC, “adding male vaccination to female-only vaccination becomes more cost-effective when all HPV-associated health outcomes are included in the model and vaccine coverage of females is low” (14). This suggests that the relative benefit of implementing male vaccination programs may be more pronounced in countries with less extensive vaccine uptake and could thus be a more effective strategy for combating HPV transmission in underdeveloped and low-income nations. Conflicting interest: parents and children Due to the nature of the HPV vaccine as a method of STI prevention — with a target population consisting of adolescent males and females — a controversy arises regarding the conflicting interests of two parties: parent and child. Currently, minor consent laws in all states allow adolescents to provide informed consent for the confidential treatment of STIs; however, with respect to vaccination, these laws do not reflect issues of decision-making (15). Furthermore, it has been observed that adolescents seeking confidential medical treatment may be among those most likely to benefit from vaccines that prevent STIs, such as the HPV vaccine (15). Because these state consent laws do not specifically address the capacity of an adolescent to consent to vaccination, a consensus among some health care providers has been reached by interpreting these laws to include vaccination as part of routine medical care and STI treatment, circumventing the need to obtain additional parental written or verbal consent (15). Nonetheless, a minor’s right to consent to vaccination is not established in United States law (16); herein lies the foundation of the controversy at hand — when the views of the parent conflict with those of the adolescent, at what point, and to what extent, does the adolescent’s autonomy become restricted? Although conflicting interests between parent and child can present a potential barrier to vaccination, studies suggest that adolescents play an important role in the process of vaccine decision-making, especially when the parents’ perception of their adolescent’s sexual activity is considered (17–19). In this study, more than half of responding parents who perceived their adolescent to be sexually active were likely to report that their adolescent was involved in the decision-making process (17). While these findings suggest increased parent-adolescent communication may increase vaccine uptake, this may not be indicative of increased favorability with respect to parental perceptions on adolescent decision-making capacity. In a study comparing parent and adolescent views on decision-making, it

Ethics and Current Climates Surrounding HPV Vaccination

was reported that more than half of all responding adolescents considered adolescents to be capable of reasonable decisionmaking at an earlier age; on the other hand, more than half of all responding parents thought otherwise (20). Indeed, there is continued debate surrounding discussions of when an adolescent attains the capacity to make informed decisions. While some argue that adolescents should have the full authority to make their own personal health-related decisions, others question the capacity of adolescents to make informed decisions (15). Parent and adolescent consent Patient autonomy is one of the most important components of consent. As the sexual health of adolescents has become a major healthcare topic globally, countries have begun to support the term “adolescent autonomy” in an effort to counter the controversies surrounding adolescent ethical decision-making. The notion of adolescent autonomy is illustrated well in an example from the United Kingdom. In England, more than 30 years ago, contraception was made available for adolescent females under the age of 16, following proper education from a health care provider, but without parental consent (21). It is worth noting that, in this case, adolescent consent was emphasized, rather than parental rights or authority under the law. Although the term “adolescent consent” is not widely used in the United States, a crucial role of health care professionals is informing and educating patients, both adults and adolescents alike (21). However, United States law authorizes parents/ guardians with the right to refuse to consent to adolescent vaccinations. Therefore, when considering adolescent consent, the role of parental decision-making cannot be overlooked. Parental decision-making processes are deeply affected by methods of communication, disease prevalence prior to vaccination, and depth of knowledge regarding vaccine efficacy and safety (20). Indeed, it has been shown that parents who have access to high-quality information, perceive vaccines to be safe, or have a child presenting with a mild illness, will be more likely to consider and consent to the vaccination of their child (20). In a qualitative study conducted in England, parents unopposed to vaccination acknowledged that adolescent decision-making was an indication of maturity, citing that the will of an adolescent to request a vaccination was indicative of their capacity to make well-informed decisions (21). In this same study, the largest group of responding parents agreed that adolescents who were well informed and understood the implications of the HPV vaccine should be allowed to consent to vaccination, without parental involvement. However, this view is in direct contrast with that of the second largest group of responding parents, which reported that parental involvement in adolescent decision-making is necessary, and should be given precedence (21). Despite these conflicting views among parents, it has been shown that providing both parents and adolescents with high-quality education is an important factor in facilitating informed consent and increasing vaccine uptake (22). Because HPV is a sexually transmitted disease, and patient autonomy is a central component of informed consent, there is a logical argument for making the HPV vaccine readily available to adolescents who opt to be vaccinated, even if their parent/ guardian has refused to consent. If the notion of adolescent autonomy becomes more widely discussed in the United

States, additional research will be necessary in order to better understand how and why vaccine preferences differ between parent and adolescent. Likewise, future research should aim to provide insight into the nuanced conversations and discussions between parents and adolescents with respect to individual autonomy, ethical consent, adolescent decision-making, and the role of policy in healthcare delivery and access. Sexual abstinence and HPV vaccination Adolescents who intend to remain sexually abstinent, whether due to education, religious beliefs, or personal choice, would not technically be considered at risk for contracting or transmitting HPV, and thus would not be a necessary subgroup to vaccinate. However, even if an attitude of sexual abstinence is maintained throughout an individual’s entire premarital life, it is not likely that this practice will be continued after marriage, warranting the need for vaccine administration prior to marriage and the onset of sexual relations. While the HPV vaccine has been approved for use in adults younger than 26 years of age, current recommendations indicate routine vaccination of adolescents age 11 to 12 years old (23). When considering HPV vaccine uptake outside of this recommended age range, as would be the case of abstinent adolescents who opt to delay vaccine uptake prior to marriage, it is necessary to compare vaccine efficacy among both adults and adolescents in order to ensure accurate and informed decisions can be made. In a study comparing data on HPV vaccine immunogenicity between adolescent males and females 10 to 15 years of age and males and females 16 to 23 years of age, the immunogenic response was greater (1.7 – 2.7-fold) in the younger cohort than the older. However, no statistically significant difference was reported, suggesting that current data on vaccine efficacy is comparable between younger and older adolescents (24). Considering the high likelihood of becoming infected by some form of HPV (25), and the absence of studies assessing rates of HPV vaccine uptake, or rates of HPV infection, among adults who opted to delay vaccination prior to marriage, there is a strong case to be made for abstinent individuals to also opt for HPV vaccination during adolescence. Administration of the vaccine would not undermine their autonomy and choice but would rather serve to safely and effectively decrease the risk of future HPV infection (26). Sexual promiscuity and HPV vaccination One perceived barrier to HPV vaccination is the parental concern that vaccination may implicitly promote earlier and riskier sexual behaviors among adolescents. The prevalence of this perception is not well known, and while one regional study reported less than 1 in 5 parents (19%) considered this a possibility, there is no factual basis to support this concern, as vaccination against other infectious agents have not been associated with increases in sexual or otherwise risky behaviors (27). The perception that vaccination against STIs may facilitate or enable riskier sexual behaviors among adolescents is unique when considering the history of vaccine requirements for school attendance. The hepatitis B vaccine, which protects against a virus largely transmissible through sexual contact, is currently required for school attendance in 47 states and the District of Columbia (28).

Ethics and Current Climates Surrounding HPV Vaccination

The role education has on promoting parental acceptance of the HPV vaccine is unclear. After parents of adolescent girls and boys who originally opposed vaccination were provided an educational pamphlet on the prevalence and seriousness of HPV infection, 20% opted to have their child vaccinated; however, following the educational intervention, some parents incorrectly concluded that the vaccine would increase sexual promiscuity (22). While education is only one aspect of a multicomponent approach to promoting public health, high-quality physician recommendations have been positively associated with HPV vaccine uptake (29). An interesting line of question for future research may be to better understand if parental education, coupled with physician recommendations, is able to curb the misperception of adolescent promiscuity and thus increase vaccine uptake. To our knowledge, there is currently no literature available representative of this perceived barrier on a state or national level. With respect to HPV vaccination, efforts should be made through future research to better understand the implications of this incorrect perception, as parental acceptance of vaccination is an essential component for the immunization of adolescents (30). Ethical concerns of HPV education programs and strategies to avoid them When implementing programs that aim to increase public awareness and understanding of the HPV virus and vaccine, ethical issues that emerge tend to be rooted in the strategies used to increase vaccine uptake. The ethical question that each strategy must answer is “Do the ends justify the means?” A study conducted in Bangladesh demonstrates this very ethical dilemma. The study analyzed a vaccine education program which targeted adolescent females 10 to 12 years of age and was considered successful as far as achieving a vaccine coverage rate of 94% (10). However, the methods employed to achieve the observed rate of vaccine coverage may not have been justified. While this program used educational sessions to increase knowledge and awareness of HPV and cervical cancer risk, it deliberately rebranded the vaccine as a “cancer vaccine.” Details regarding the mechanisms of HPV transmission were also omitted in order to avoid confronting regional taboos regarding adolescent sexual behaviors, which may have otherwise negatively influenced vaccine uptake. It is likely that the combination of these two strategies contributed significantly to the high vaccine coverage achieved (10). Furthermore, informed consent was not technically attained in this case, but rather the absence of a parent at the time of vaccination was interpreted as "implied consent" (10). Vaccine programs may utilize several strategies in order to avoid ethical issues and mitigate factors that may negatively impact vaccine uptake. A meta-analysis and review of strategies found that behavioral interventions, such as reminder/recall systems, and informational interventions such as factsheets, brochures, and expert or peer-narrated videos, were the most effective in vaccine series initiation (31). However, with respect to vaccine series completion, behavioral interventions were the only strategy shown to produce significantly higher rates. Although informational interventions were shown to successfully improve beliefs and attitudes toward vaccination, they were shown to have no effect on vaccine completion (31). Additionally, 40

the effects of environmental interventions, which include programs implemented in schools, clinics, and postpartum units, were highly variable, affecting different target groups, or only certain aspects of vaccine behavior. For example, studies investigating the effect of a school-based vaccination program found that increased accessibility to vaccines led to increased vaccine uptake (31). Additionally, college females who viewed a combined peer-expert narrative video were significantly more likely to later vaccinate than females who viewed a control informational video that lacked a narrative which contained separate peer and expert narratives (31). As new vaccine programs and strategies are implemented, effective interventions should adhere to the highest ethical standards, maintain transparency about possible long-term consequences of HPV infection, while also incorporating behavioral and informational components in an effort to create an optimal vaccine program. Ethics and HPV screening Another issue that effective vaccine programs aim to resolve is increased awareness of and access to preventive cervical cancer screening technology, but sometimes programs (e.g., the Bangladeshi program) fail to mention the importance of screening, which may have established a false sense of security in vaccinated females who received the divalent serotype, which only prevents 70% of cervical cancer (10). Screening is used to detect any precancerous lesions and address them before they can cause harm. There are two screening methods currently accepted: the Pap cytology test and the hrHPV test. When compared to cytology alone, hrHPV testing alone, as well as co-testing (cytology plus hrHPV testing), have been shown to be twice as effective at detecting cervical dysplasia (32). However, Zhao et al. (2015) reported that roughly 30% of women participating in their study were diagnosed with cervical cancer 3 to 5 years after having previously received only one of the aforementioned screening methods, which demonstrates flaws in the precision of these screening techniques. Current CDC-recommended screening intervals for cytology alone are every 3 years for women 21 to 65 years of age; screening intervals for hrHPV testing alone, as well as cotesting, are recommended every 5 years for women 30 to 65 years of age (34). A concern among some physicians when consulting on screening preferences is that, for patients who opt to be screened on an interval of every 5 years, there is an increased likelihood that a patient may be lost to follow-up. In contrast, some physicians may be more comfortable with, and even promote, less frequent screening for patients when considering each patient’s socioeconomic status or potential risk. Regardless of the screening method a physician and patient agree on, it is imperative that vaccination programs promote regular cervical cancer screenings to the same degree that they promote vaccination.

Conclusion Current ethical dilemmas surrounding HPV vaccination range from vaccine program strategy to conflicts of interest between parents and adolescents, the capacity for adolescents to make informed decisions, and methods of implementing new state

Ethics and Current Climates Surrounding HPV Vaccination

and national legislation. While current vaccination strategies focus on female-only coverage, recent findings suggest there are additional benefits when males are included in vaccination programs, such as robust population-wide protection against HPV transmission and decreased disease burden. Issues regarding the cost efficiency of implementing and funding programs are controversial but warrant further examination and discussion. Vaccination may be declined in some cases because of personal beliefs or in other cases because of misinformation that parents have perceived to be true. Several strategies involve the facilitation of communication between parents/guardians and adolescents, while others aim to promote public awareness, but regardless of strategy, the focus remains increasing vaccine delivery. Moving forward, it will be imperative to establish a more accepted consensus regarding adolescent decisionmaking capacity, both with respect to vaccine uptake, and as it relates to health care delivery more broadly. HPV vaccination provides a means to reduce HPV infection and transmission, and extensive vaccine uptake is particularly important for conferring population-wide immunity and reducing HPV prevalence. To that end, there are logical arguments that could be made in support for expansion of HPV vaccination inclusion and coverage. In conclusion, we support the continued use, implementation, and expansion of HPV vaccination programs, in addition to the utilization of supplemental preventive methods, such as routine screening and proper patient education.

Disclosures The author has no financial relationship with a commercial entity producing health care related products and/or service.

References 1.

Burd EM. Human papillomavirus and cervical cancer [Internet]. Vol. 16, Clinical Microbiology Reviews. American Society for Microbiology (ASM); 2003 [cited 2021 Jun 2]. p. 1–17. Available from: /pmc/articles/PMC145302/

Navarro-Illana P, Aznar J, Díez-Domingo J. Ethical considerations of universal vaccination against human papilloma virus [Internet]. Vol. 15, BMC Medical Ethics. BioMed Central Ltd.; 2014 [cited 2021 May 7]. p. 1–7. Available from: http://www.biomedcentral.com/14726939/15/29

Bosch FX, Manos MM, Muñoz N, Sherman M, Jansen AM, Peto J, et al. Prevalence of human papillomavirus in cervical cancer: A worldwide perspective. J Natl Cancer Inst [Internet]. 1995 Jun 7 [cited 2021 May 7];87(11):796– 802. Available from: https://jhu.pure.elsevier.com/en/ publications/prevalence-of-human-papillomavirus-incervical-cancer-a-worldwide-3 Muñoz N, Bosch FX, de Sanjosé S, Herrero R, Castellsagué X, Shah K V., et al. Epidemiologic Classification of Human Papillomavirus Types Associated with Cervical Cancer. N Engl J Med [Internet]. 2003 Feb 6 [cited 2021 May 7];348(6):518–27. Available from: https://pubmed.ncbi. nlm.nih.gov/12571259/

Chaturvedi AK, Engels EA, Pfeiffer RM, Hernandez BY, Xiao W, Kim E, et al. Human papillomavirus and rising oropharyngeal cancer incidence in the United States. J Clin Oncol [Internet]. 2011 Nov 10 [cited 2021 May 7];29(32):4294–301. Available from: https://pubmed.ncbi. nlm.nih.gov/21969503/

Singh GK, Miller BA, Hankey BF, Edwards BK. Persistent area socioeconomic disparities in U.S. incidence of cervical cancer, mortality, stage, and survival, 19752000. Cancer [Internet]. 2004 Sep 1 [cited 2021 May 7];101(5):1051–7. Available from: https://pubmed.ncbi. nlm.nih.gov/15329915/

Kitchener HC, Castle PE, Cox JT. Chapter 7: Achievements and limitations of cervical cytology screening. Vaccine [Internet]. 2006 Aug 21 [cited 2021 Jun 2];24(SUPPL. 3). Available from: https://pubmed.ncbi.nlm.nih. gov/16950019/

Dobbs SP, Asmussen T, Nunns D, Hollingworth J, Brown LJR, Ireland D. Does histological incomplete excision of cervical intraepithelial neoplasia following large loop excision of transformation zone increase recurrence rates? A six year cytological follow up. Br J Obstet Gynaecol [Internet]. 2000 [cited 2021 Jun 2];107(10):1298– 301. Available from: https://pubmed.ncbi.nlm.nih. gov/11028584/

Safaeian M, Solomon D, Castle PE. Cervical Cancer Prevention-Cervical Screening: Science in Evolution [Internet]. Vol. 34, Obstetrics and Gynecology Clinics of North America. NIH Public Access; 2007 [cited 2021 Jun 2]. p. 739–60. Available from: /pmc/articles/PMC2762353/

10. Salwa M, Abdullah Al-Munim T. Ethical issues related to human papillomavirus vaccination programs: An example from Bangladesh [Internet]. Vol. 19, BMC Medical Ethics. BioMed Central Ltd.; 2018 [cited 2021 May 7]. p. 81–7. Available from: https://doi.org/10.1186/s12910-0180287-0 11. Kirby T. FDA approves new upgraded Gardasil 9 [Internet]. Vol. 16, The Lancet. Oncology. Lancet Oncol; 2015 [cited 2021 May 7]. p. e56. Available from: https://pubmed.ncbi. nlm.nih.gov/25532625/ 12. Regan DG, Philp DJ, Hocking JS, Law MG. Modelling the population-level impact of vaccination on the transmission of human papillomavirus type 16 in Australia. Sex Health [Internet]. 2007 [cited 2021 May 7];4(3):147–63. Available from: https://pubmed.ncbi.nlm.nih.gov/17931528/ 13. Taira A V., Neukermans CP, Sanders GD. Evaluating human papillomavirus vaccination programs. Emerg Infect Dis [Internet]. 2004 [cited 2021 May 7];10(11):1915– 23. Available from: https://pubmed.ncbi.nlm.nih. gov/15550200/ 14. Dunne EF, Markowitz LE, Harrell C, Robinette C, Mona S, Gee J, et al. Recommendations on the Use of Quadrivalent Human Papillomavirus Vaccine in Males — Advisory Committee on Immunization Practices (ACIP), 2011 [Internet]. Centers for Disease Control and Prevention Morbidity and Mortality Weekly Report. 2011 [cited 2021 May 7]. Available from: https://www.cdc.gov/mmwr/ preview/mmwrhtml/mm6050a3.htm 41

Ethics and Current Climates Surrounding HPV Vaccination

15. Ford CA, English A, Davenport AF, Stinnett AJ. Increasing Adolescent Vaccination: Barriers and Strategies in the Context of Policy, Legal, and Financial Issues. J Adolesc Heal [Internet]. 2009 Jun [cited 2021 May 7];44(6):568– 74. Available from: https://pubmed.ncbi.nlm.nih. gov/19465321/ 16. Vaccination Requirements and Laws | CDC [Internet]. [cited 2021 Jun 2]. Available from: https://www.cdc.gov/ vaccines/imz-managers/laws/index.html 17. McRee AL, Reiter PL, Brewer NT. Vaccinating adolescent girls against human papillomavirus-Who decides? [Internet]. Vol. 50, Preventive Medicine. NIH Public Access; 2010 [cited 2021 May 7]. p. 213–4. Available from: /pmc/ articles/PMC4046840/ 18. Hughes J, Cates JR, Liddon N, Smith JS, Gottlieb SL, Brewer NT. Disparities in how parents are learning about the human papillomavirus vaccine. Cancer Epidemiol Biomarkers Prev [Internet]. 2009 Feb [cited 2021 Jun 2];18(2):363–72. Available from: https://pubmed.ncbi.nlm. nih.gov/19190161/ 19. Ziarnowski KL, Brewer NT, Weber B. Present choices, future outcomes: Anticipated regret and HPV vaccination. Prev Med (Baltim) [Internet]. 2009 May [cited 2021 Jun 2];48(5):411–4. Available from: https://pubmed.ncbi.nlm. nih.gov/18996144/ 20. Rand CM, Humiston SG, Schaffer SJ, Albertin CS, Shone LP, Blumkin AK, et al. Parent and adolescent perspectives about adolescent vaccine delivery: Practical considerations for vaccine communication. Vaccine [Internet]. 2011 Oct 13 [cited 2021 May 7];29(44):7651–8. Available from: https:// pubmed.ncbi.nlm.nih.gov/21839793/ 21. Brabin L, Roberts SA, Kitchener HC. A semi-qualitative study of attitudes to vaccinating adolescents against human papillomavirus without parental consent. BMC Public Health [Internet]. 2007 Feb 9 [cited 2021 May 7];7(1):1–7. Available from: http://www.biomedcentral. com/1471-2458/7/20

May 7];118(5):2135–45. Available from: https://pubmed. ncbi.nlm.nih.gov/17079588/ 25. Broomall EM, Reynolds SM, Jacobson RM. Epidemiology, clinical manifestations, and recent advances in vaccination against human papillomavirus. Postgrad Med [Internet]. 2010 Mar [cited 2021 May 7];122(2):121–9. Available from: https://pubmed.ncbi.nlm.nih.gov/20203463/ 26. Ruiz-Sternberg ÁM, Moreira ED, Restrepo JA, LazcanoPonce E, Cabello R, Silva A, et al. Efficacy, immunogenicity, and safety of a 9-valent human papillomavirus vaccine in Latin American girls, boys, and young women. Papillomavirus Res [Internet]. 2018 Jun 1 [cited 2021 May 7];5:63–74. Available from: /pmc/articles/PMC5887018/ 27. Brewer NT, Cuite CL, Herrington JE, Weinstein ND. Risk compensation and vaccination: Can getting vaccinated cause people to engage in risky behaviors? Ann Behav Med [Internet]. 2007 [cited 2021 May 7];34(1):95–9. Available from: https://pubmed.ncbi.nlm.nih.gov/17688401/ 28. Califano S, Calo WA, Weinberger M, Gilkey MB, Brewer NT. Physician support of HPV vaccination school-entry requirements. Hum Vaccines Immunother [Internet]. 2016 Jun 2 [cited 2021 May 7];12(6):1626–32. Available from: / pmc/articles/PMC4964636/ 29. Gilkey MB, Calo WA, Moss JL, Shah PD, Marciniak MW, Brewer NT. Provider communication and HPV vaccination: The impact of recommendation quality. Vaccine [Internet]. 2016 Feb 24 [cited 2021 May 7];34(9):1187–92. Available from: https://pubmed.ncbi.nlm.nih.gov/26812078/ 30. Zimet GD, Mays RM, Sturm LA, Ravert AA, Perkins SM, Juliar BE. Parental attitudes about sexually transmitted infection vaccination for their adolescent children. Arch Pediatr Adolesc Med. 2005;159(2):132–7. 31. Rodriguez AM, Do TQN, Goodman M, Schmeler KM, Kaul S, Kuo YF. Human Papillomavirus Vaccine Interventions in the U.S.: A Systematic Review and Meta-analysis. Vol. 56, American Journal of Preventive Medicine. Elsevier Inc.; 2019. p. 591–602.

22. Davis K, Dickman ED, Ferris D, Dias JK. Human papillomavirus vaccine acceptability among parents of 10to 15-year-old adolescents. J Low Genit Tract Dis [Internet]. 2004 Jul 1 [cited 2021 May 7];8(3):188–94. Available from: https://augusta.pure.elsevier.com/en/publications/ human-papillomavirus-vaccine-acceptability-amongparents-of-10-to

32. Katki HA, Kinney WK, Fetterman B, Lorey T, Poitras NE, Cheung L, et al. Cervical cancer risk for women undergoing concurrent testing for human papillomavirus and cervical cytology: A population-based study in routine clinical practice. Lancet Oncol [Internet]. 2011 Jul [cited 2021 May 7];12(7):663–72. Available from: https://pubmed.ncbi.nlm. nih.gov/21684207/

23. Petrosky E, Bocchini JA, Hariri S, Chesson H, Curtis CR, Saraiya M, et al. Use of 9-valent human papillomavirus (HPV) vaccine: updated HPV vaccination recommendations of the advisory committee on immunization practices. MMWR Morb Mortal Wkly Rep [Internet]. 2015 Mar 27 [cited 2021 May 7];64(11):300–4. Available from: http:// www.ncbi.nlm.nih.gov/pubmed/25811679

33. Zhao C, Li Z, Nayar R, Levi AW, Winkler BA, Moriarty AT, et al. Prior high-risk human papillomavirus testing and papanicolaou test results of 70 invasive cervical carcinomas diagnosed in 2012: Results of a retrospective multicenter study. Arch Pathol Lab Med [Internet]. 2015 Feb 1 [cited 2021 May 7];139(2):184–8. Available from: https://pubmed.ncbi.nlm.nih.gov/24694342/

24. Block SL, Nolan T, Sattler C, Barr E, Giacoletti KED, Marchant CD, et al. Comparison of the immunogenicity and reactogenicity of a prophylactic quadrivalent human papillomavirus (types 6, 11, 16, and 18) L1 virus-like particle vaccine in male and female adolescents and young adult women. Pediatrics [Internet]. 2006 Nov [cited 2021

34. Kim JJ, Burger EA, Regan C, Sy S. Screening for cervical cancer in primary care a decision analysis for the us preventive services task force. J Am Med Assoc [Internet]. 2018 Aug 21 [cited 2021 May 7];320(7):706–14. Available from: https://jamanetwork.com/

Scholarly Research In Progress • Vol. 5, November 2021

Virtual Newborn Nursery Rounds: An Alternative Experience During the COVID-19 Pandemic Tara E. Avery1† and Ashley L. Shamansky1 ¹Geisinger Commonwealth School of Medicine, Scranton, PA 18509 † Doctor of Medicine Program Correspondence: tavery@som.geisinger.edu or teavery@geisinger.edu

Abstract Background: The novel coronavirus (COVID-19) has presented educators with new, unique challenges in delivering education that have underscored the need for innovative ways to prepare learners for the clinical setting without being physically present in the clinic. This virtual workshop was developed to simulate newborn nursery rounds for students removed from the clinical setting at the onset of the COVID-19 pandemic. Methods: Students were each assigned one case to present, with the facilitator guiding the discussion as if on clinical newborn nursery rounds. Student educational outcomes were measured using newborn nursery Entrustable Professional Activities (EPAs) assessment forms on a 5-point Likert scale. Performance on the EPAs was compared between students completing virtual rounds versus those who completed inperson rounds. Results: Of the 7 measured EPAs, only 2 showed a statistically significant differences (EPA 3, p-value 0.01 and EPA 9, p-value 0.003). Conclusion: These results demonstrated that virtual newborn nursery rounds can act as a substitution for clinical newborn nursery rounds with statistically significant differences in only 2 of the 7 measured EPAs. Virtual, simulated rounds are an alternative way to introduce medical students to the newborn nursery and review concepts that will help them to succeed in the clinical setting.

Introduction The novel coronavirus (COVID-19) has presented educators with new, unique challenges in delivering education to medical students. Although lectures during the didactic years of medical school are frequently replaced with podcasts and virtual resources, replacing hands-on clinical experiences poses a significant challenge. Prior research has demonstrated that implementing preclinical pediatric rounds can be successful in delivering educational content and improving students’ clinical skills (1). Clinical rotations are a key component of medical education and are often the first time that medical students interact with patients and are involved in decision-making as part of the health care team. Patient contact, a variety of patient presentations, and being part of a clinical team have been identified as core features of clinical learning (2). In the face of uncertainty and new guidelines for social distancing, it has become imperative to find new ways to prepare learners for the clinical setting without being physically present in the clinic. One of the most valuable skills that students develop during their clinical year is the art of the oral presentation during patient rounds. The oral presentation is crucial as it facilitates

patient care, directs both individual and group learning, and provides an opportunity for students to be evaluated and receive constructive feedback (3). When performed properly, the oral presentation communicates all important aspects of patient care to the rest of the health care team and makes for efficient and effective patient rounds. Developing proficiency in this area is one of the greatest challenges for students, as it requires integration of materials rather than simple rote memorization and regurgitation. Perhaps the most challenging aspect for medical students when giving oral presentations is identifying which data is clinically relevant due to student inexperience (4). The oral presentation can also be a source of anxiety for students. Prior studies have demonstrated that students required to deliver oral presentations had significant anxiety with increased cortisol levels on the day of their presentations compared to before and after giving the presentation (5). The art of the oral presentation is further complicated by the lack of a standardized framework for both teaching and evaluating students and the influence of each attending physician’s subjective preferences on presentation style. Despite this lack of framework, several standards have been identified that are commonly accepted among clerkship directors with the most important elements being a complete and accurate history of the present illness, pertinent physical exam findings, laboratory studies, and a prioritized assessment and plan addressing the problem at hand (3). However, students best learn to give oral presentations through trial and error rather than a standardized framework or specific model, thus underscoring the importance of continued practice (4). In the clinical setting, most of the time on rounds is spent presenting with less time spent on providing constructive feedback due to time constraints, reducing opportunities for medical student education. E-learning, or online learning, has become a popular feature of medical education as it is durable and reusable, allows for standardized content delivery, and permits asynchronous learning with fewer time constraints (6, 7). Furthermore, implementing technology into medical education can be considered a primer for interprofessional cooperation and future telemedicine encounters (8). Prior curricula developed at Vanderbilt University School of Medicine demonstrated that implementing preclinical pediatric rounds was successful in delivering educational content and improving students’ clinical skills such as eliciting a chief complaint, obtaining the history of present illness (HPI), and reviewing past medical history with patients. Additionally, student-reported comfort levels pertaining to interacting with the pediatric patient population increased (1). Another study demonstrated that pediatric interns who underwent simulationbased training in family-centered rounding demonstrated 43

Virtual Newborn Nursery Rounds: An Alternative Experience During the COVID-19 Pandemic

statistically significant improvement in their presentation skills and higher performance ratings overall (9). To prepare medical students removed from the clinical setting at the onset of COVID-19 for future clinical rotations, we developed five newborn nursery cases to be used during virtual rounds to simulate the experience of the newborn nursery. The Council on Medical Student Education in Pediatrics (COMSEP) identified newborn care as a core component of the MS3 General Pediatric Clerkship Curriculum, as noted in the 2019 COMSEP Curriculum Action Toolkit (10). Using these cases, we expect students to gather information from maternal and patient data and practice oral presentations to prepare for real-time rounding in the clinical setting. These cases present the opportunity for students to receive real-time feedback on their performance, and to address common topics covered on newborn nursery rounds, thus providing a foundation in newborn care and knowledge. Our primary goal is to identify if virtual, simulated newborn nursery rounds can act as a substitution for clinical newborn nursery rounds.

Methods We developed this workshop for third-year medical students, but the materials can be applied broadly to any learner preparing for pediatric clinical rotations or clerkships and are readily adaptable to any institution. We presented these cases in a workshop to 27 students during a virtual curriculum lasting 2 weeks via Zoom Video Communications. One day prior to the workshop, we sent the students workshop instructions and assigned each student a case with the following materials: maternal delivery summary, infant delivery summary, maternal history and physical, newborn information, as well as maternal medications and additional laboratory values depending on the specific case. We expected students to have a rudimentary understanding of how to properly construct an oral patient presentation which they would deliver during virtual rounding in groups of five. In preparation for their presentations, students were required to fully review all documents and complete a history and physical note for the newborn that they would be presenting. We also provided workshop instructions to the pediatric clinical facilitators, as well as case summaries and a summary of all teaching points. We conducted the workshop with learning groups composed of one facilitator and five learners, with each learner assigned to one of the five cases. Facilitators for the workshop were newborn nursery physicians that volunteered to guide the discussion. During the workshop, students presented their patient cases sequentially to the group, as if on rounds, with the facilitator guiding the discussion using the provided case summaries and teaching points. We repeated this process for all five cases and spent approximately 2.5 hours completing the simulated rounds, with 30 minutes spent on each individual case. This study was approved by the Geisinger Institutional Review Board as not-human research. We measured educational objectives and outcomes for the workshop using the GCSOM Clinical Assessment Tool, based on the Entrustable Professional Activities (EPA) assessment form on a 5-point Likert scale with anchors. Following the workshop, facilitators completed an assessment of student performance on each EPA. Each 44

assessment was returned to the student within one week of the workshop. We then compared student performance on virtual nursery rounds to student performance on clinical nursery rounds from the same year to identify if virtual rounds were an adequate substitution for clinical rounds. In order to do so, we looked at differences in performance on the EPAs between 36 students who completed clinical, in-person newborn nursery rounds and the 27 students who completed virtual newborn nursery rounds. There were 7 EPAs in total that evaluated a variety of skills pertaining to clinical competencies. These EPAs included gathering a pertinent history, prioritizing differential diagnoses and selecting a working diagnosis, recommending and interpreting common tests, discussing management and treatment plans, delivering an oral presentation, and demonstrating an eagerness to learn as well as professionalism (Table 1). EPA 5, documenting a clinical encounter in appropriate written format, was omitted from the EPA assessment. Although students completed HPIs for their cases, they were graded by other faculty and not the physicians facilitating the workshop and were therefore not included. EPA 8, performing a patient handover, was also not assessed in the virtual setting and was therefore excluded from the evaluation. We conducted statistical analysis of the data using a two-tailed t-test.

Results In total, we delivered this virtual curriculum to 27 third-year medical students. The students who completed the workshop were students removed from the clinical setting at the onset of the COVID-19 pandemic. These students had a rudimentary understanding of how to develop and write a HPI but had not had any clinical exposure to pediatrics or delivering oral presentations in the newborn nursery. Facilitators for this workshop were volunteer core pediatric faculty and newborn nursery physicians and therefore had pre-existing background knowledge on the topics covered during the simulated newborn nursery rounds. Facilitators evaluated students following virtual rounds using a newborn nursery EPA assessment form on a 5-point Likert scale with anchors (Table 1). These EPAs aligned with our educational objectives which we developed using Kirkpatrick’s pyramid. Our ultimate goal was to evaluate if simulated newborn nursery rounds are an adequate substitution for clinical newborn nursery rounds. To answer this question, we compared the EPA assessments of the third-year medical students completing virtual, simulated newborn nursery rounds to EPA assessments of students on clinical newborn nursery rounds from the same year. We found statistically significant differences in performance on two of the EPAs (Table 2), with improved performance among the in-person group as compared to virtual learners. EPA 3, recommending and interpreting common tests, differed between the two groups (p-value = 0.01) and EPA 9, demonstrating professionalism and teamwork (p-value = 0.003).

Discussion Virtual, simulated newborn nursery rounds are a novel method of preparing learners for pediatric rotations outside of the clinical setting. Implemented as a substitution for clinical newborn nursery rounds in the wake of the COVID-19

Virtual Newborn Nursery Rounds: An Alternative Experience During the COVID-19 Pandemic

pandemic, we found that the virtual newborn nursery rounds are successful in preparing student learners. Ultimately, students can gain a better understanding of how patient rounds function, practice and refine their oral presentation skills, and discuss various topics relevant to caring for a newborn as a pediatrician. Of the 7 measured EPAs, only 2 showed a statistically significant difference. EPA 3, interpreting and recommending common tests, was lower among the virtual group (p-value = 0.01). This could be because the materials provided to students already included documented clinical testing, reducing student recommendations for future laboratory tests. This could also be accounted for by a lack of newborn medicine experience at the time of the virtual rotation. These simulated newborn rounds were the first exposure to the newborn nursery for many students, requiring them to Table 1. Entrustable Professional Activities (EPAs) assessed. Students were assessed using the above EPAs. Each EPA was graded using a 5-point Likert scale rapidly acquire, assess, and incorporate information ranging from 1 to 5, with 1 being the lowest score possible and 5 being the highest. prior to presenting whereas students in the clinical EPA 5, documenting a clinical encounter in appropriate written format, and EPA 8, newborn nursery group were evaluated toward performing a patient handover, were not assessed in the virtual setting and therefore the end of their rotation (1 full week of newborn excluded from students’ evaluations. nursery), which provided them with more time to receive feedback and hone their skills prior to formal evaluation. The virtual formatting also allowed for more candid, rapid feedback which may be more representative of student performance. EPA 9, professionalism, was also lower among the virtual rotators versus the clinical rotators. We hypothesize that this difference may be due to the fact that students were delivering oral presentations from the comfort of their home environment. Additionally, when assessing professionalism in the virtual environment, there are not as many interactions with providers, support staff, and families that can be assessed. Interestingly, the standard deviation for professionalism is the largest among the data set, suggesting a large difference between virtual versus in-person rotators. This raises the question of whether students can have improved professionalism ratings by simply being present in person. Although these cases strive to replicate clinical scenarios, it is difficult to simulate the professional clinical environment in its entirety. Despite these differences, most EPAs showed no statistical difference.

Table 2. In-person newborn nursery vs. virtual curriculum nursery performance. Of the 7 measured EPAs, only 2 showed a statistically significant difference. EPA 3, interpreting and recommending common tests, and EPA 9, professionalism, were both lower among the virtual rotators versus the clinical rotators.

Potential limitations of this activity include reliance on a skilled facilitator with knowledge of the newborn nursery. Facilitators for the workshop were all pediatricians with background knowledge of the subject matter; however, any facilitator with knowledge of the newborn nursery could guide the discussion using the provided materials. Similar to in-person clinical rounds, there may be time constraints. Facilitators should be mindful of the time throughout the workshop to guarantee enough time to cover the teaching points and provide ample constructive feedback to each student. One area for improvement is the newborn physical exam, which could not be approximated with virtual rounds. Students were instructed to view the Loyola University Stritch School of Medicine Newborn Physical Exam Video and provided with newborn physical exam findings that they could present (11). This workshop helps to develop a framework for and discuss pertinent physical exam findings but does not give students the hands-on experience

Virtual Newborn Nursery Rounds: An Alternative Experience During the COVID-19 Pandemic

that they would receive in the actual newborn nursery. A useful adjunct to this workshop would include a session allowing students to simply handle newborn infants and become comfortable examining them with a more in-depth discussion of pertinent physical exam findings.

Heiman HL, Uchida T, Adams C, et al. E-learning and deliberate practice for oral case presentation skills: A randomized trial. Medical Teacher. 2012;34(12):e820-e826. http://www.tandfonline. com/doi/abs/10.3109/0142159X.2012.714879. doi: 10.3109/0142159X.2012.714879.

Conclusion

Ferrel MN, Ryan JJ. The impact of COVID-19 on medical education. Cureus (Palo Alto, CA). 2020;12(3):e7492. https://www.ncbi.nlm.nih.gov/ pubmed/32368424. doi: 10.7759/cureus.7492.

Rao P, Hill E, Palka C, et al. Improving pediatric resident communication during family-centered rounds using a novel simulation-based curriculum. MedEdPORTAL. 2018;14(1):10733. https://search.datacite.org/ works/10.15766/mep_2374-8265.10733. doi: 10.15766/ mep_2374-8265.10733.

In summary, simulated rounds are an adequate substitution for clinical rounds with differences in only 2 of the 7 measured EPAs. Although intended for use by third-year medical students, these cases can be applied broadly to any learner preparing for pediatric clinical rotations or clerkships and are adaptable to any institution. These cases can be delivered synchronously on a pediatric rotation, asynchronously as a “primer” prior to starting a newborn nursery rotation, or as clinical makeup time or remediation. Simulated, virtual rounds will never fully replicate the professional working environment of the hospital or working within an integrated health care team, but they are an alternative way to open the door for medical students and provide education on pertinent topics and presentation skills that will help them to succeed in the clinical setting.

Disclosures The authors have nothing to disclose.

References 1.

Apple R, Fleming A, Israel S. Pediatric clinical rounds teaching guide: A preclinical curriculum to improve medical student comfort with pediatric patients. MedEdPORTAL. 2012;8(1). doi: 10.15766/mep_2374-8265.9293.

Kandiah DA. Perception of educational value in clinical rotations by medical students. Advances in medical education and practice. 2017;8:149-162. https://www.ncbi.nlm.nih. gov/pubmed/28223855. doi: 10.2147/AMEP.S129183.

Green E, Green E, Durning S, et al. Expectations for oral case presentations for clinical clerks: Opinions of internal medicine clerkship directors. J Gen Intern Med. 2009;24(3):370-373. https://www.ncbi.nlm.nih.gov/ pubmed/19139965. doi: 10.1007/s11606-008-0900-x.

Haber R, Lingard L. Learning oral presentation skills. J Gen Intern Med. 2001;16(5):308-314. https://www.ncbi. nlm.nih.gov/pubmed/11359549. doi: 10.1046/j.15251497.2001.00233.x.

Merz CJ, Wolf OT. Examination of cortisol and state anxiety at an academic setting with and without oral presentation. Stress (Amsterdam, Netherlands). 2015;18(1):138-142. http://www.tandfonline.com/ doi/abs/10.3109/10253890.2014.989206. doi: 10.3109/10253890.2014.989206.

Ruiz JG, Mintzer MJ, Leipzig RM. The impact of E-learning in medical education. Academic Medicine. 2006;81(3):207212. https://www.ncbi.nlm.nih.gov/pubmed/16501260. doi: 10.1097/00001888-200603000-00002.

10. Narayan AM, Held M, Noelck A, Narayan M, Jamie. 2019 COMSEP CURRICULUM REVISION. COMSEP curriculum revision action team toolkit. 2019 COMSEP curriculum revision. 11. Normal newborn screening physical exam. [https://vault. lumc.edu/videos/videojump.cfm?hd=1&ID=4DE8CA15DAF6-42FC-AA95-40EA218E2942]. Chicago, IL: Loyola University Health System; 2001.

Scholarly Research In Progress • Vol. 5, November 2021

Battling Trainee Biases and Reconstructing Perceptions in Global Neurology Maia X. Giombetti1†, and Kelly J. Baldwin2 ¹Geisinger Commonwealth School of Medicine, Scranton, PA 18509 ²Maine Medical Center, Portland, ME 04102 † Doctor of Medicine Program Correspondence: mgiombetti@som.geisinger.edu

Abstract This article addresses gaps in medical education and training when it comes to tackling biases during the preclinical years through residency and beyond. An elective neurology rotation in Lusaka Zambia sheds light on ingrained biases and how they inhibit quality care. The concept of “single stories” looks to identify the source of preconceived perceptions and how to tackle the gap in health care created by biases.

Perspective Combating bias has been a long-term goal within the health care field. From expert to gender to racial (both individual and institutional), biases can severely hinder the care provided to patients. Innovative research and training programs have been implemented by health care systems and medical education to reduce biases among future and currently practicing physicians. Although small strides are being made through these efforts, there are still major blind spots in health care that affect patients. As the rate of globalization increases as well as the fight for social justice and equality, it is increasingly important to address biases preemptively. A prevalent theme in medical education and training emphasizes self-awareness, self-reflection, and cultural identity to encourage recognition of implicit biases (and otherwise). Following self-reflection, the hope is then to combat bias through practice. This approach alone seeks to improve the gap in delivering appropriate and accurate care to patients based upon inequalities and other social determinants of health. However, with only conceptual training before clinicals, how is bias reduction being ensured among future practicing physicians before affecting patient care? Integrating real experiences that force confrontation of biases plus self-reflection through a global health curriculum may hold the answer. To address the issue at hand, it must first be identified. A neurology elective rotation in Lusaka Zambia helped shed light on how ingrained biases could be tackled. A common theme of biases among participants surrounded poverty and lack of education. Some self-reflections on the experience by the participants included shock, expecting small, unclean, and understaffed infrastructures with limited to no specialty services and resources. Anything contradictory to these preconceived ideas brought on feelings of dissonance and inhibits quality care. In a post-reflection in June 2020 by Ashka Patel, DO, on her experience in Zambia, she stated, “I think we forget that when it comes to any country, there are rich areas and there are impoverished areas. There are private hospitals that have access to an array of testing and medications, compared to the public hospitals that struggle with the volume

of patients and lack of resources. We cannot forget this divide. There are underserved areas and people with no health care insurance who deserve access to good health care as a basic human right.” Along with assumptions of subpar training curriculums, these biased themes about developing nations are not uncommon from residents of a developed nation. These assumptive ideas are not just about developing nations as a whole. Common occurring biases and presumptions about minority groups exist in both developing and developed nations. The experience in Lusaka necessitated confrontations of biased perceptions within an individual about the nation, system, facility, and people. With individual biases and perceptions identified, the source must further be identified to effectively combat them. The concept of “single stories” aims to explain how ingrained biases in an individual come to be. Conceived by author Chimamanda Ngozi Adichie, “single stories” describes distinct accounts of individuals, groups, or nations that promote false perceptions. The danger of “single stories,” however, is not whether the accounts hold truth or not, but that they are the only story (1). A single narrative leaves incomplete understandings and fosters misconstrued ideas. If the only story of Africa is one that is poor and unsophisticated, then it will remain. The same is to be said about stories being told about minority groups. Persisting stereotypes and assumptions about groups persevere until more stories are told and retold. Unfortunately, many perceptions of developing nations and minority groups are reduced to “single stories” until more authentic experiences are encountered for individuals to rewrite for themselves. For health care to advance, “single stories” need to evolve. “Single stories” illustrate the source of ingrained biases and how they arise. Determining approaches on how to advance “single stories” past their sole plot could reveal best practices in reducing and eliminating biases among current and future practicing physicians. This points to curriculum and training past conceptual self-awareness. The rotation in Lusaka uncovered implicit biases about the nation of Zambia, the health system, and the people. Appropriate training experiences abroad combined with self-reflection led to identification of preconceived ideas and the resources to combat them in practice, whereas isolated self-awareness cannot definitively produce the same results. It is the extended encounters with the people, health care system, and communities that provide teachings and experiences that combat ingrained biases and remove feelings of shock and dissonance over time. These genuine encounters advance “single stories” to personal narratives and relationships that only add to overall care. What the genuine encounter does is enable empathy over sympathy. 47

Battling Trainee Biases and Reconstructing Perceptions in Global Neurology

Empathy is the cornerstone of patient-centered medicine and is hinged upon authentic connections. The ability to empathize involves the capacity to adopt different perspectives, nonjudgmentally. It only occurs in the absence of biases. Authentic experiences, without assumptions or perceptions, foster connections and allow profound relationships to form. Empathy grants holistic care due to its understanding of an individual’s personal narrative and all that it encompasses. On the opposite end of the spectrum is sympathy with its underlying foundation of preconceived perceptions. Sympathy drives disconnection via an uneven dynamic between parties. Sympathy does not demand perspective taking or vulnerability; it only elicits pity. Sympathizing merely offers brief condolences as opposed to meaningful connections to provide greater care. Sympathy feeds disengagement and is derived from disjointed awareness such as with “single stories.” Once again, with “single stories” as the only source of knowledge to fuel thought processes and decisions, blind spots within the health care system are created and gaps in care widen. It is a perpetual cycle. “Single stories” create biases. Biases reinforce sympathy. Sympathy feeds disengagement which then maintains the major blind spots in health care. Therefore, the cycle continues on and on. It is apparent that this cycle needs to be broken.

References Adichie CN. (2009, July). The danger of a single story [Video]. TED Conferences. https://www.ted.com/talks/chimamanda_ ngozi_adichie_the_danger_of_a_single_story

Scholarly Research In Progress • Vol. 5, November 2021

Primary Ectopic Breast Carcinoma of the Vulva: A Case Report Youngeun C. Armbuster1†, Paula Ronjon2, Cletus Baidoo3, and Waqarun N. Rashid4 ¹Geisinger Commonwealth School of Medicine, Scranton, PA 18509 ²Hematology and Oncology, Geisinger, Danville, PA 17822 ³Pathology, Geisinger, Danville, PA 17822 4 Obstetrics and Gynecology, Geisinger, Danville, PA 17822 † Doctor of Medicine Program Correspondence: yarmbuster@som.geisinger.edu

Abstract

Case Presentation

A 64-year-old woman presented to the Emergency Department with acute vomiting and moderate, sharp, diffused abdominal pain, and weakness. She reported chronic Bartholin’s cyst for 1 year, which prompted gynecology consult. Investigations had revealed a primary breast carcinoma of the vulva. Surgical excision was performed, and pathology of the mass demonstrated estrogen receptor weakly positive (20%), progesterone receptor negative (<1%), and HER2 oncoprotein positive (3+). PET/CT showed metastatic disease involving retroperitoneal, bilateral iliac chain and pelvic lymph nodes and the left T12 lamina. Mammogram showed no evidence of disease, and all her prior mammograms were negative. Due to the lack of standard treatment guidelines, the patient was managed utilizing the established breast cancer treatment guidelines. The purpose of this case report is to highlight the rarity of the diagnosis of primary breast carcinoma of the vulva and the importance of including this diagnosis in differential when evaluating patients with mass lesions of the vulva.

A 64-year-old obese, postmenopausal woman, gravida 1 para 1-0-0-1, presented to the Emergency Department with acute vomiting and moderate, sharp, diffused abdominal pain, and weakness. On physical exam, patient was febrile and tachycardic with diffuse abdominal tenderness. Her complete blood count with differential revealed neutrophilic leukocytosis. Blood culture grew group B Streptococcus. Computed tomography of abdomen and pelvis revealed several enlarged lymph nodes within the inguinal regions and iliac chains within the pelvis bilaterally (Figure 1). The patient was admitted for evaluation and management of her presenting symptoms.

Introduction Ectopic breast tissue is a common congenital condition found in 2% to 6% of women and 1% to 3% of males, which may develop along the embryologic mammary lines extending bilaterally from the axilla through the breast to the mons pubis (1, 2). The term ectopic breast tissue is used for both supernumerary and aberrant breast tissue. Supernumerary breasts have nipples, areolae or both with varied composition of glandular tissue, whereas an aberrant or accessory breast tissue has no organized secretory system and does not communicate with the overlying skin. The most common location for the accessory breast tissue is the axilla while other uncommon sites are infraclavicular, subscapular, epigastric and vulva (3). An accessory breast tissue is hormonally sensitive and may enlarge in response to pregnancy or exogenous hormones, and these tissues may also develop breast pathologies, including fibroadenoma, phyllodes tumor, Paget disease, and invasive adenocarcinoma (1, 2). Ectopic breast carcinoma is often not detected, or diagnosis is delayed until significant clinical symptoms due to lack of screening. We report herein a case of primary ectopic breast carcinoma of the vulva with distant metastasis to bones and lymph nodes in a postmenopausal woman.

Upon further investigation of unclear etiology of inguinal lymphadenopathy and bacteremia, patient reported chronic Bartholin’s cyst for 1 year, which prompted gynecology consult. External examination of genitalia showed 3 cm x 4 cm firm left vulvar mass with irregular border superiorly with erythema. There was no fluctuance or drainage. On her hospital stay day 3, she was taken to the operating room for simple excision of vulvar mass for management of sepsis as suspected source of infection. Firm, non-mobile, non-necrotic, 3.5 cm x 4.5 cm vulvar mass on left labia majora was excised for biopsy. Histopathologic evaluation reveals a 2.7 cm poorly differentiated infiltrative mass invading subcutaneous tissue, epidermis, and dermis with skin ulceration. Deep surgical

Figure 1. Computed tomography shows several enlarged lymph nodes within the inguinal regions and iliac chains within the pelvis bilaterally.

Primary Ectopic Breast Carcinoma of the Vulva: A Case Report

Figure 2. Hematoxylin and eosin stain of ulcerated skin with invasive tumor x20 magnification (left) and x200 magnification (right).

Figure 3. CK7 immunostain (top left), estrogen receptor immunostain (top right), GATA-3 immunostain (bottom left), and GCDFP-15 immunostain (bottom right) of invasive tumor x40 magnification.

Figure 4. PET/CT scan shows multiple FDG avid bilateral iliac chain and pelvic lymph nodes (left) and left T12 lamina (right).

Primary Ectopic Breast Carcinoma of the Vulva: A Case Report

the tumor cells. Therefore, the results from specimens that are negative must be evaluated accordingly. HER2 oncoprotein expression is positive with 3+ average membranous intensity.

Figure 5. Mammography from 2019 (top) and 2020 (bottom).

margins were involved, and the mass comprises infiltrative sheets and clusters of malignant ductal epithelial cells with comedo-type necrosis. The tumor cells show markedly enlarged pleomorphic nuclei with vesicular chromatin and a distinct-toprominent nucleoli (Figure 2). Myoepithelial cell layer is absent. The tumor shows the following immunophenotypic profile: CK7 and GATA-3 diffuse positivity; GCDFP-15 and estrogen receptor (ER) patchy positivity; BNC5 immunostain confirms a clonal ductal proliferation with loss of the myoepithelial cell layer; CK20, p40, CK5/6, uroplakin-II, mammaglobin, CD56, synaptophysin, and S100 immunostains are all negative; and p16 immunostain shows equivocal patchy staining. Therefore, usual markers of breast origin (CK7, GATA-3, GCDFP-15, and ER) are positive, while usual markers of melanocytic, neuroendocrine, and urothelial primaries are negative (Figure 3). Hence, the diagnosis of a primary ectopic breast carcinoma of the vulva, histologic grade 3 (poorly differentiated). The pathologic staging for this case is a pT1b pNX (for lesions more than 2 cm or any size with stromal invasion more than 1.0 mm, confined to the vulva and/or perineum; and regional lymph nodes cannot be assessed). Prognostic markers revealed: ER is weakly positive (20%), progesterone receptor (PR) is negative (<1%), and HER2 oncoprotein is positive (3+). The tumor cells show the usual profile of an invasive ductal carcinoma of breast origin. Evaluation of receptor protein expression is performed by visual analysis of formalin-fixed paraffin-embedded immunostaining of the invasive tumor using FDA-cleared antibodies and protocols with estrogen receptor protein (Ventana SP1 antibody), progesterone receptor protein (1E2 antibody) and FDA approved HER2 oncoprotein (Ventana 4B5 antibody). ER protein expression is weakly positive with 20% nuclear positivity and 2+ average intensity score (range 0 to 3+). PR protein expression is negative with <1% nuclear positivity. Assay external control immunoreactivity is appropriate. No internal control was present in the analyzed tissue. False negative results may occur when no internal control ducts are present in tissue with negative reactivity in

Positron emission tomography/computed tomography (PET/CT) scan was performed which showed multiple enlarged and fluorodeoxyglucose (FDG) avid retroperitoneal, bilateral iliac chain and pelvic lymph nodes most consistent with metastatic disease. Several indeterminate subcentimeter, mildly FDG avid bilateral subpectoral lymph nodes were visualized as well. Hypermetabolic osseous lesion involving the left T12 lamina is consistent with metastatic disease (Figure 4). There is an indeterminate 8 mm left upper lobe nodule with no abnormal FDG activity seen on CT, likely too small to be seen on PET. Mammogram showed no evidence of malignancy (BIRADS Category 1), and past mammography from 2019, 2018, 2016 and 2014 were all negative (Figure 5). The patient has a history of hypertension, diabetes mellitus Type 2, dyslipidemia, iron deficiency anemia, osteoarthritis, rheumatoid arthritis, sleep apnea, chronic diarrhea, and nausea. Past surgical history includes Cesarean section, dilation and curettage, fluorescein angioscopy, bevacizumab injection, and retina treatment (photocoagulation). There is no significant family history of cancer. The patient is being managed by a hematology oncology provider. A biopsy of lymph nodes or bone was requested but was not feasible. Initial treatment options were followed: chemotherapy with paclitaxel, trastuzumab and pertuzumab. A cycle is every 21 days with close monitoring of heart function. Echocardiogram was obtained prior to initiating the treatment to assess the baseline cardiac function which was within normal limits. Denosumab is given to prevent skeletal events with the plan for restaging every three to four cycles. Following the initiation of the treatment, Taxol was changed to Abraxane due to allergic reaction, even with oral high dose dexamethasone. A cycle is 21 days: 2 weeks on and 1 week off. The patient reported severe depression and anxiety, and she was referred to palliative care and support group. The patient also reported a new onset headache, which prompted CT scan of brain which was within normal limits. The patient declined MRI due to claustrophobia.

Discussion At the fifth or sixth week of fetal development, an ectodermic thickening starts to form the mammary ridges, which extends bilaterally from the axilla to the groin along the milk lines. These ridges are not prominent in the human embryo and disappear over the following months, except for small portions that may persist in the pectoral region (5, 6). Ectopic breast tissue is persistent epidermal thickenings along milk line from axilla to perineum or vulva due to clusters of primordial breast cells that fail to involute. Ectopic breast tissue may be combinations of breast glandular tissue and nipple, and it occurs in 2% to 6%

Primary Ectopic Breast Carcinoma of the Vulva: A Case Report

of females and 1% to 3% of males. Almost any type of known breast pathology can occur in such ectopic breast tissue, and primary breast carcinoma arising from accessory breast tissue has been reported in 60% to 70% of all forms of ectopic breast tumor (4).

Disclosures

Primary carcinoma of the ectopic breast is relatively common at the axilla, while primary carcinoma arising from ectopic breast tissue in the vulva is extremely rare with an incidence of 4%. The predominant pathology is that of invasive ductal carcinoma, however, ductal carcinoma in situ, lobular carcinoma, mucinous adenocarcinoma, phyllodes tumors, and fibroadenomas have also been reported in ectopic breast tissue (7). Multiple cases of this rare malignancy have been reported in the English-language clinical literature; however, ectopic breast carcinoma is difficult to diagnose due to the late expression of pathologic symptoms.

References

In the absence of concurrent breast carcinoma, the lesion of primary vulvar origin can be categorized by the following: a morphologic pattern consistent with breast carcinoma, the presence of estrogen and progesterone receptors, and/or positivity for common breast cancer markers such as epithelial membrane antigen, carcinoembryonic antigen, and glandular keratins (8). For a diagnosis of this disease, a thorough metastatic workup is necessary including, but not limited to history, physical examination, and radiologic examination of the breasts, to confirm that the vulvar lesion is the primary site as opposed to a metastasis from a primary breast cancer. Although primary breast cancer of the vulva tends to metastasize early and to have a poor prognosis, definitive treatment guidelines have been unavailable. Currently, this type of cancer is stage and treated according to current tumor, node, metastasis (TNM)-based classification applicable to primary breast cancer. The treatment should consist of individualized combination of surgery, chemotherapy, monoclonal antibody therapy, radiation, and adjuvant endocrine therapy, as appropriate.

Youngeun C. Armbuster, Paula Ronjon, Cletus Baidoo, and Waqarun N. Rashid declare that they have no conflict of interest.

Hoffman BL, Schorge JO, Halvorson LM, Hamid CA, Corton MM, Schaffer JI. Benign Disorders of the Lower Reproductive Tract. Williams Gynecology, 4e. McGrawHill; [cited 2020 Oct 24]. Available from: https:// accessmedicine-mhmedical-com.gcsom.idm.oclc.org/ content.aspx?bookid=2658&sectionid=219458833

Patel PP, Ibrahim AM, Zhang J, Nguyen JT, Lin SJ, Lee BT. Accessory breast tissue. Eplasty. 2012;12:ic5.

Husain M, Khan S, Bhat A, Hajini F. Accessory breast tissue mimicking pedunculated lipoma. BMJ Case Rep. 2014;2014:bcr2014204990. Published 2014 Jul 8. doi:10.1136/bcr-2014-204990

Lee J, Jung JH, Kim WW, et al. Ductal carcinoma arising from ectopic breast tissue following microcalcification observed on screening mammography: a case report and review of the literature. J Breast Cancer. 2014;17(4):393396. doi:10.4048/jbc.2014.17.4.393

C. Parker C, Damodaran S, Bland KI, Hunt KK. The Breast. In: Brunicardi F, Andersen DK, Billiar TR, Dunn DL, Kao LS, Hunter JG, Matthews JB, Pollock RE. eds. Schwartz's Principles of Surgery, 11e. McGraw-Hill; [cited 2020 Nov 1]. Available from: https://accessmedicinemhmedical-com.gcsom.idm.oclc.org/content. aspx?bookid=2576&sectionid=216206605

Al-Mansouri L, Poursoltan P, Simons M, Muljono A, Boyages J. Primary breast cancer of the vulva: A case report and literature review. J Obstet Gynaecol Res. 2018 Dec;44(12):2190-2194. doi: 10.1111/jog.13778. Epub 2018 Aug 20.

Lopes G, DeCesare T, Ghurani G, Vincek V, Jorda M, Glück S, Silva O. Primary ectopic breast cancer presenting as a vulvar mass. Clin Breast Cancer. 2006 Aug;7(3):278-9. doi: 10.3816/CBC.2006.n.041.

Irvin WP, Cathro HP, Grosh WW, Rice LW, Andersen WA. Primary breast carcinoma of the vulva: a case report and literature review. Gynecol Oncol. 1999 Apr;73(1):155-9. doi: 10.1006/gyno.1998.5269.

Conclusion Primary breast carcinoma, arising from embryonic mammary ridge remnants, is an extremely rare histologic subtype of vulvar cancer; however, this should be included in differential diagnosis when evaluating patients with mass lesions of the vulva. Obtaining adequate tissue biopsy is essential in establishing a morphologic diagnosis, since diagnosis rests on the pathologic findings, with recognition of the characteristic histologic features and the presence of estrogen, progesterone and/or HER2 receptors and the extent of disease. Therapy should consist of an individualized combination of surgery, radiotherapy, chemotherapy, antiestrogen therapy, and monoclonal antibody therapy, like cancer of the orthotopic breast of similar stage. Owing to the rarity of this lesion, clinical trials to determine optimum treatment are not available, and management guidelines will rely on small series or case studies.

Acknowledgments We thank the patient for allowing us to share her details and thank John S. Farrell, MD, Department of Radiology, Geisinger, for radiologic image acquisition.

Scholarly Research In Progress • Vol. 5, November 2021

Trends and Regional Differences in the Diversion of Stimulants in the United States, 2015–2019 Holly E. Funk1†‡, Susannah E. Pitt1†‡, Alison T. Varano1†‡, and Brian J. Piper1 ¹Geisinger Commonwealth School of Medicine, Scranton, PA 18509 † Doctor of Medicine Program ‡ Authors contributed equally Correspondence: spitt@som.geisinger.edu

Abstract

Introduction

Background: Stimulant drugs are commonly prescribed for the treatment of attention deficit hyperactivity disorder (ADHD) in the United States (U.S.). Sympathomimetics such as amphetamine, lisdexamfetamine, and methylphenidate are classified as Schedule II substances with a high potential for misuse. Drug diversion is defined as the distribution of legal prescription medication in an unlawful way. The purpose of this study was to examine and analyze epidemiological trends in stimulant drug diversion across the U.S. for the years 2015 to 2019.

Prescription drug misuse in the United States (U.S.) is challenging to monitor and quantify. There is limited data available and illicit use often goes unreported to medical providers, health officials and law enforcement. According to the drug scheduling classification, Schedule I drugs have the highest potential for abuse, dependence, and adverse effects, while Schedule V drugs have the least potential for abuse. Stimulants, such as amphetamine, lisdexamfetamine, and methylphenidate are labeled as Schedule II drugs with high potential for abuse (1). Prescription stimulants are commonly used to treat attention deficit hyperactivity disorder (ADHD), with other indications including binge-eating disorder and narcolepsy. ADHD is the most common childhood neurodevelopmental disorder (2). Symptomology of the disorder includes inattentiveness, hyperactivity, and impulsivity in multiple settings, such as the home and in school (3). The National Institute of Drug Abuse classifies stimulant misuse as taking an inappropriate dose of prescription medication and/or in a manner other than what was prescribed to you by a physician or misusing another individual’s prescription medication (4). While ADHD is commonly known as a childhood disorder, the Diagnostic and Statistical Manual (DSM) 5 criteria now addresses diagnostic issues for adult patients with ADHD, such as broadened onset age requirements and lower symptom number thresholds (5). According to the National Survey on Drug Use and Health (N=102,000), it was estimated that 6.6 % of U.S. adults used prescription stimulant drugs, 4.5% without any misuse and 2.1% with misuse (6). Abusing stimulants can lead to negative medical outcomes affecting nearly every organ system. Acute intoxication effects may include insomnia, anxiety, panic attacks, hallucinations, hypertension, tachycardia, and arrhythmias, while chronic misuse can lead to neurological, cardiovascular, pulmonary, and gastrointestinal complications. There is potential for sensitization, drug-craving and other addictive behaviors as well (7).

Methods: Drug data weight for the years 2015 to 2019 was obtained using StreetRx.com through a Data Use Agreement in all 50 states for the following drugs: amphetamine, lisdexamfetamine, and methylphenidate. StreetRx is a crowdsourcing public website which allows users to make submissions on the quantity, price, and location of illicit drug sales. The data set was divided into four regions, Northeast (NE), South (S), West (W), and Midwest (MW), for regional analysis. Mass of stimulants in grams was corrected for population estimates from the U.S. Census Bureau, Population Division. Results: The total mass of diverted stimulants per population (mg/population) in the United States decreased from 2015 to 2019 for amphetamine (-31.5%), lisdexamfetamine (-51.0%), and methylphenidate (-57.0%). The percent changes varied regionally, with trends in the NE and S closely mirroring the total percent changes for most years. However, the changes in the MW and W regions were much more modest. The W stands out as having a significantly higher average value of diverted stimulants per population (mg/population) for all included years. Conclusion: Reported diversion by StreetRx.com of stimulants on the illicit market has declined in recent years. Diversion in the W U.S. is higher than other regions. These trends do not mirror the rise in prescription distribution of amphetamine and lisdexamfetamine, nor the rise in diagnoses of ADHD over the past 10 years. Potential explanations include drug characteristics, misuser characteristics, and regional health care characteristics. This research provides health care teams with insight into the distribution of frequently prescribed stimulants in the illicit market and potentially in their communities. Further research is needed to shed light on the impact of COVID-19 on the already appreciable misuse of prescription stimulants.

Drug diversion is defined as the illegal acquirement and distribution of controlled pharmaceuticals (8). Diversion of prescription drugs may happen at any point in the supply chain course, from production to final distribution, but often occurs after the medical professional and patient interaction. The ability to effectively monitor and track the diversion of prescription drugs does not only reduce the negative health implications of prescription drug abuse but can also inform efforts to reform and improve strategies to mitigate diversion

Trends and Regional Differences in the Diversion of Stimulants in the United States, 2015–2019

(9). Additionally, education and training programs on the risk of misuse and diversion of prescription stimulants for healthcare providers may be necessary in order to confidently prescribe medications to patients. A national questionnaire administered to physicians (N= 826) reported that only 48% of physicians received official training on prescription drug diversion (10). It is evident that patients would benefit from programs that educated physicians and other health care providers on the prevalence, outcomes, and mitigation of stimulant drug misuse. The Researched Abuse, Diversion and Addiction-Related Surveillance (RADARS®) System is an anonymous, self-reported online platform named StreetRx.com, which allows individuals to track the price and distribution of prescription and illicit drugs across the U.S. (11). Users can make submissions that include information such as the generic name of the substance purchased, the price, raw dosage, and the geographical location of the transaction. An analysis of this platform provides valid estimates of distribution and pricing on the illicit market (12). Researchers have used StreetRx to analyze crowdsourced reporting of illicit drug use for other substances. General opioid misuse, as well as focused oxycodone and oxymorphone cross-sectional studies have been conducted (12, 13). One study concluded that street prices were often influenced by factors such as potency and crush-resistant formulations (13). In a study focused on buprenorphine diversion, researchers found that geographic distribution and socioeconomic factors influenced the pricing and diversion through the black market (14). StreetRx data is cited in several other studies and our aim was to use similar methods to analyze the diversion of stimulants (15). The purpose of this study was to examine trends in prescription stimulant drug diversion across the U.S. for 2015 to 2019. Previous studies have reported increases in stimulant medication use over recent years (16, 17, 18). We aimed to specifically analyze illicit use and determine if trends mirrored the increase in prescription distribution and ADHD diagnoses in recent years (16, 17, 18). Uncovering patterns of stimulant drug diversion may provide health care providers with insight regarding illicit distribution of commonly prescribed medications in their communities.

Methods Procedures A data use agreement between Geisinger Commonwealth School of Medicine and Rocky Mountain Poison and Drug Safety department of the Denver Health and Hospital Authority allowed access to StreetRx.com data reports. Information in the data use agreement included: generic name, drug dosage, and the date and location of submission. Drug data weight was obtained from Street Rx.com for the years 2015 to 2019 in all 50 states for the following drugs: amphetamine, lisdexamfetamine, and methylphenidate. Procedures were approved by the Geisinger Institutional Review Board. Data analysis The data set obtained from StreetRx.com was divided into four regions for regional analysis of distribution: Northeast (NE), South (S), West (W), and Midwest (MW). The raw dose in milligrams was totaled for each of the three drugs according to year, state, and region to determine the total quantity in milligrams of amphetamine, lisdexamfetamine and methylphenidate during the years 2015 to 2019. The mass was corrected from population estimates from the U.S. Census Bureau, Population Division for each state and region. Percent changes between each year, 2016 to 2019, were compared to 2015 for each drug. Figures were created using GraphPad Prism to plot the mean diversion distribution with standard error of mean (SEM) for each drug during the years noted. Heat maps were created using JMP Graph Builder to map the percent changes in total distribution of amphetamine, lisdexamfetamine and methylphenidate from 2015-2019. For each drug, a paired sample t-test was completed comparing the total diversion amount for each year, 2016 to 2019, relative to 2015 values. Note that there was no methylphenidate data available for Louisiana, Montana, and North Dakota for the year 2019. The corresponding t-test paired the remaining states for the analysis, while these three were excluded. Paired sample t-tests were conducted for analysis of statistical significance of annual changes in total mass drug diversion for each region of the U.S. GraphPad Prism was utilized to perform statistical analysis. Variability was depicted as the SEM.

Figure 1. Total diversion in milligrams per person of amphetamine (A), lisdexamfetamine (B), and methylphenidate (C) for all 50 states from 2015 to 2019 as reported by StreetRx and corrected for population.

Trends and Regional Differences in the Diversion of Stimulants in the United States, 2015–2019

Results Total mass diversion distribution Diverted amphetamine distribution decreased by an average of -31.54% across all 50 states from 2015 to 2019 (Figure 1A). Diverted lisdexamfetamine distribution decreased by an average of -50.97% (Figure 1B). Diverted methylphenidate distribution decreased by an average of -57.03% (Figure 1C). The percent change trends in the NE and S were relatively similar to the total percent changes for lisdexamfetamine and methylphenidate, while the percent changes in NE for amphetamine and the MW and W for all three stimulants were relatively modest. Specific percent change values for each drug according to region are summarized in Table 1. Paired t-tests comparing the total mass of distribution for illicit stimulants showed varying significance through the years 2015 to 2019. For amphetamine, the percent changes showed statistical significance between the years 2015 and 2016, 2017, and 2019, while 2018 did not show a statistically significant change relative to 2015 (P <0.05). For lisdexamfetamine, the percent changes showed statistical significance between all years relative to 2015 (P <0.05). For methylphenidate, the percent changes showed statistical significance between the years 2015 and 2017, 2018, and 2019 with only 2016 showing an insignificant change relative to 2015 (P <0.05). Statistical analysis for total distribution for the three stimulants is summarized in Table 2.

Table 1. Total and regional percent change values relative to 2015

Regional differences in drug diversion Average mass per population for each year was plotted to show regional differences in nonmedical prescription stimulant reporting (Figure 2). Heat maps of 2015 to 2019 percent changes showed varying degrees of both positive and negative percent changes by state (Figures 3-5). Paired t-tests comparing the regional data showed no statistical significance among intraregional averages for any year relative to 2015 (P <0.05).

Table 2. T-test results comparing StreetRx total diversion between years

Figure 2. Average values in milligrams per person of amphetamine (A), lisdexamfetamine (B), and methylphenidate (C) by region from 2015 to 2019 as reported by StreetRx and corrected for population. Green represents W, blue represents MW, red represents S, and yellow represents NE.

Trends and Regional Differences in the Diversion of Stimulants in the United States, 2015–2019

Discussion

Figure 3. 2015 to 2019 percent changes in amphetamine distribution for 2015 to 2019. The total grams of amphetamine diversion as reported by StreetRx has been corrected for population.

Figure 4. 2015 to 2019 percent changes in lisdexamfetamine distribution for 2015 to 2019. The total grams of lisdexamfetamine diversion as reported by StreetRx has been corrected for population.

Across all 50 states, reported diversion by StreetRx.com of all three stimulants decreased. Amphetamine reporting by raw number comprised greater than half of total submissions. The average mass per population for each year was significantly higher in the W and MW than in the NE and S regions, with the W showing the greatest value. The differences within each region by year were not statistically significant and are very modest compared to the interregional differences. While the overall percent changes for each drug from 2015 to 2019 were negative, some states showed positive percent changes from 2015 to 2019 for amphetamine and lisdexamfetamine, while all states showed negative percent changes from 2015 to 2019 for methylphenidate. The West is thus an interesting area of study in terms of stimulant drug distribution, both prescription and illicit (17). Hypothesized etiologies of these patterns can be summarized in three broad categories: drug characteristics, misuser characteristics, and regional health care characteristics. First, drug characteristics may include differences in efficacy, abuse potential, and availability of certain stimulants. For example, lisdexamfetamine may carry a lower abuse potential than other stimulants due to certain pharmacodynamic and route of administration parameters (19, 20). This could lead to decreasing misuse of this particular stimulant across the U.S. Second, misuser characteristics may include varying popularity of stimulants, pricing preferences, and demographics of misusers. For example, research shows that white college-aged students that self-report symptoms of ADHD are more likely to abuse stimulants, often to improve concentration or alertness (21). Additionally, misusers may prefer amphetamine over other stimulants due to perceived higher efficacy, greater availability, and higher rates of prescribing, leading to overall higher reporting of misuse (22). Third, regional healthcare factors may include differences in ADHD diagnoses, prescriber preferences, and number of prescribers per population. For instance, the median number of psychiatrists in the U.S. has continued to decline in recent years (23). There are pronounced differences in the availability of psychiatrists and mental health resources in certain areas of the U.S, particularly poverty-stricken rural areas (24). This could lead to differences in absolute distribution, as well as perceived need to obtain illicit substances if patients cannot access proper health care. As research moves forward, delineating the data into specific states and cities of interest may further elucidate potential etiologies. Focusing on Pennsylvania as an example, approximately 65% of the StreetRx.com entries included the city of purchase. From these, approximately 32.7% of the self-reported purchases were made in Philadelphia, 16.1% in Pittsburgh, and 3.4% in the Scranton/Wilkes-Barre area. It is noteworthy that Philadelphia makes up roughly 12.4% of Pennsylvania's population, yet almost one-third of data points collected were from Philadelphia. Moving forward, this type of analysis could identify particular cities and counties that could benefit from community level interventions.

Figure 5. 2015 to 2019 percent changes in methylphenidate distribution for 2015 to 2019. The total grams of methylphenidate diversion as reported by StreetRx has been corrected for population.

Understanding illicit drug use patterns is important to community health, specifically during times of global crisis. While it is difficult to fully capture trends in illicit drugs as compared to legal use, analysis is still necessary. Some of the

Trends and Regional Differences in the Diversion of Stimulants in the United States, 2015–2019

trends found in illicit stimulant use in this study can potentially be explained by the trends in prescription stimulant use. Elucidating socioeconomic factors that contribute to regional variations in stimulant misuse will require further investigation. Limitations An important limitation to note in our study is the data obtained from StreetRx.com is dependent on user submissions, therefore our study is not able to represent a complete understanding of all illicit drug use and distribution. Since submissions were self-reported, there is the potential for individuals to input information inaccurately and inconsistently. There is also the possibility that individuals did not want to self-report illicit drug use due to fear of self-incrimination or negative societal views. Furthermore, socioeconomic factors such as a lack of access to the internet could lessen use of the platform in some communities. Lack of awareness of the platform’s existence may also influence data, potentially influencing regional differences in reporting. This limitation can be diminished by correlating trends found in this study to other databases and resources. This research team and others analyzed pharmacoepidemiological trends in both prescription and illicit stimulant distribution of these drugs in the U.S. over similar time periods. The prescription data can be obtained from the Automation of Reports and Consolidated Orders System (ARCOS) retail drug summary reports (16). Future efforts to analyze potential correlations or differences among studies and databases may potentially solidify regional and nationwide patterns.

Conclusion Over the past 5 years, reported diversion by StreetRx.com of stimulants on the illicit market has declined. Diversion in the W of the U.S. is higher than other regions. We hypothesized that trends in stimulant misuse would mirror trends in stimulant prescription distribution, yet this relationship was not found for every drug, year, and region. Potential contributing factors to explain these trends include differences in drug, user, and regional health care characteristics. While states have some commonalities in regulations regarding controlled substances such as stimulants, the specific guidelines of each state may vary. Additionally, existing individual state and community-level programs may influence the use and distribution of stimulants in different regions. It is important to analyze data surrounding the efficacy of intervention and treatment programs in decreasing illicit stimulant use and distribution in order help future policymakers address the issue and allocate appropriate funding. Future research will focus on these state and community-level programs to further characterize regional trends. Future research exploring the trends of drug diversion through StreetRx.com may unveil additional patterns. Although our study looked at stimulant drug distribution including amphetamine, lisdexamfetamine and methylphenidate, broadening the Date Use Agreement to receive data on all drugs reported to StreetRx.com would allow for comparison of stimulant drug diversion to other drug classes, such as opioids and benzodiazepines.

Acknowledgments We would like to thank the Biomedical Research Club at Geisinger Commonwealth School of Medicine for their guidance and assistance throughout the development of this project.

Disclosures BJP is part of the osteoarthritis research team funded by Pfizer and Eli Lily. The other authors have no disclosures.

References 1.

United States Drug Enforcement Administration. Drug Scheduling [Internet]. U.S. Department of Justice; 2020 [updated 2020 April 20; cited 2021 Apr 24]. Available from: https://www.dea.gov/drug-scheduling

Perou, R., Bitsko, R. H., Blumberg, S. J., Pastor, P., Ghandour, R. M., Gfroerer, J. C., ... Huang, L. N. (2013). Mental health surveillance among children–United States, 2005–2011. MMWR Suppl. 2013 May 17;62(2):1-35.

Edition F. Diagnostic and statistical manual of mental disorders. Am Psychiatric Assoc. 2013;21.

NIDA. Prescription Stimulants DrugFacts [Internet]. National Institute on Drug Abuse; 2018 [cited 2021 Apr 24]. Available from: https://www.drugabuse.gov/ publications/drugfacts/prescription-stimulants

Epstein JN, Loren RE. Changes in the definition of ADHD in DSM-5: subtle but important. Neuropsychiatry. 2013 Oct 1;3(5):455.

Compton WM, Han B, Blanco C, Johnson K, Jones CM. Prevalence and correlates of prescription stimulant use, misuse, use disorders, and motivations for misuse among adults in the United States. Am J of Psychiatry. 2018 Aug 1;175(8):741-55.

Ciccarone D. Stimulant abuse: pharmacology, cocaine, methamphetamine, treatment, attempts at pharmacotherapy. Prim Care. 2011 Mar 1;38(1):41-58.

Inciardi JA, Surratt HL, Kurtz SP, Burke JJ. The diversion of prescription drugs by health care workers in Cincinnati, Ohio. Subst Use Misuse. 2006 Jan 1;41(2):255-64.

Wood D. Drug diversion. Aust Prescr. 2015 Oct;38(5):164.

10. Colaneri N, Keim SA, Adesman A. Physician training and qualification to educate patients on attention-deficit/ hyperactivity disorder stimulant diversion and misuse. J Child Adolesc Psychopharmacol. 2018 Oct 1;28(8):554-61. 11. Rocky Mountain Poison and Drug Safety. RADARS® System. Denver, CO: Denver Health; 2020. Available from: www.rmpds.org/consulting/radars%C2%AE-system.html. 12. Dasgupta N, Freifeld C, Brownstein JS, Menone CM, Surratt HL, Poppish L, Green JL, Lavonas EJ, Dart RC. Crowdsourcing black market prices for prescription opioids. J Medical Internet Res. 2013;15(8):e178. 13. Lebin JA, Murphy DL, Severtson SG, Bau GE, Dasgupta N, Dart RC. Scoring the best deal: Quantity discounts and street price variation of diverted oxycodone and oxymorphone. Pharmacoepidemiol Drug Saf. 2019 Jan;28(1):25-30. 57

Trends and Regional Differences in the Diversion of Stimulants in the United States, 2015–2019

14. Hswen Y, Zhang A, Brownstein JS. Leveraging black-market street buprenorphine pricing to increase capacity to treat opioid addiction, 2010–2018. Prev Med. 2020 Aug 1;137:106105. 15. Piper BJ, Ogden CL, Simoyan OM, Chung DY, Caggiano JF, Nichols SD, et al. Trends in use of prescription stimulants in the United States and Territories, 2006 to 2016. PloS One. 2018;13(11). Available from: doi: 10.1371/journal. pone.0206100 16. Vaddadi SM, Czelatka NJ, Guitierrez BD, Maddineni BC, Torres-Teran CD, Tron DN, McCall KL, Piper BJ. Rise and regional variations in schedule II stimulant use in the United States. medRxiv. 2020 Jan 1. 17. Danielson ML, Bitsko RH, Ghandour RM, Holbrook JR, Kogan MD, Blumberg SJ. Prevalence of parent-reported ADHD diagnosis and associated treatment among US children and adolescents, 2016. J Clin Child Adolesc Psychol. 2018 Mar 4;47(2):199-212. 18. Goodman DW. Lisdexamfetamine dimesylate (Vyvanse), a prodrug stimulant for attention-deficit/hyperactivity disorder. P T. 2010;35(5):273-287. 19. Jasinski DR, Krishnan S. Abuse liability and safety of oral lisdexamfetamine dimesylate in individuals with a history of stimulant abuse. J Psychopharmacology. 2009;23(4):41927. Available from: https://doi-org.gcsom.idm.oclc. org/10.1177/0269881109103113. 20. Wilens TE, Adler LA, Adams J, Sgambati S, Rotrosen J, Sawtelle R, Utzinger L, Fusillo S. Misuse and diversion of stimulants prescribed for ADHD: a systematic review of the literature. J Am Acad Child Adolesc Psychiatry. 2008 Jan 1;47(1):21-31. 21. Faraone SV, Buitelaar J. Comparing the efficacy of stimulants for ADHD in children and adolescents using meta-analysis. Eur Child Adolesc Psychiatry. 2010;19(4):35364. Available from: DOI 10.1007/s00787-009-0054-3 22. Satiani A, Niedermier J, Satiani B, Svendsen DP. Projected workforce of psychiatrists in the United States: a population analysis. Psychiatr Serv. 2018 Jun 1;69(6):710-3. 23. Butryn T, Bryant L, Marchionni C, Sholevar F. The shortage of psychiatrists and other mental health providers: causes, current state, and potential solutions. Int J Acad Med. 2017 Jan 1;3(1):5.

Scholarly Research In Progress • Vol. 5, November 2021

Hydroxychloroquine, Azithromycin, and Chloroquine Prescribing Patterns in Medicaid Mansi S. Khurana1*‡, Uzoamaka V. Eziri2*‡, Taylor S. Mewhiney2*‡, Cathie-Allegra Z. Nkabyo2*‡, Jennifer Szpernoga2*‡, and Daniela I. Velasquez2*‡ ¹Geisinger Commonwealth School of Medicine, Scranton, PA 18509 *Master of Biomedical Sciences Program ‡ Authors contributed equally Correspondence: mkhurana@som.geisinger.edu

Abstract

Methods

Over the last year, the COVID-19 pandemic has claimed the lives of many people throughout the world. As the virus spreads, affecting millions of patients, there has been a massive movement to discover readily available and effective treatment options (1). Inconsistent information regarding the benefits of hydroxychloroquine/chloroquine and azithromycin in COVID-19 treatment has been an obstacle in the delivery of clinical care during the pandemic (2). Limited data regarding the evolution of these therapies has created a knowledge gap that we aim to address by analyzing the experimental treatment options of COVID-19 using drug prescription patterns. This study used data from the Medicaid State Drug Utilization database and the Micromedex database to gather information on prescribed hydroxychloroquine, chloroquine, and azithromycin in Medicaid from 2016 to 2020. Our results show a decrease in azithromycin (-45.63%) and chloroquine prescription (-18.9%) from 2016 to 2020, and an increase in hydroxychloroquine prescription (+19.8%). Additionally, our results show a decrease in the average cost for hydroxychloroquine (-74.2%) and azithromycin (-20.4) and an increase in the average cost of chloroquine (138.4%). The increase in the number of prescriptions for hydroxychloroquine from quarter 1 of 2020 to quarter 3 of 2020 can be secondary to the COVID-19 pandemic in states, whereas the decrease in azithromycin prescriptions from 2016 to 2020 can be linked to emergence of new antibiotics with stronger function.

Procedures

Introduction Hydroxychloroquine and chloroquine are antimalarial drugs used to treat autoimmune diseases and prevent malaria. Use of these drugs have been rising in the United States (U.S.) as an experimental treatment option for COVID-19 (3). Azithromycin is an antimicrobial drug used to treat various bacterial infections such as chlamydia, mycoplasma, and mycobacterium (4). Hydroxychloroquine, with or without azithromycin, has been considered as a possible therapeutic agent for patients with COVID-19 (5). We explored the prescribing patterns and changes in average cost of hydroxychloroquine, azithromycin, and chloroquine in Medicaid throughout the U.S. in 2016 through 2020. We hypothesized that there was an increase in prescribing patterns of hydroxychloroquine, azithromycin, and chloroquine secondary to the COVID-19 pandemic.

We examined the prescribing patterns of hydroxychloroquine, azithromycin, and chloroquine prescribing patterns in Medicaid. The Micromedex database was used to find drug trade names. The Medicaid State Drug Utilization database served as a reference to extract the total number of prescriptions per state per quarter during 2016 to 2020. It was also used to gather information about enrollees and the total amount reimbursed. Statistics: For data analysis, a combination of Excel, Google sheets, and GraphPad Prism were used to create bar graphs, line graphs and to perform calculations.

Results The total number of prescriptions for azithromycin was 8,715,064; 8,206,847; 7,167,437; 6,836,130 and 4,459,327 for 2016, 2017, 2018, 2019, and 2020, respectively (Figure 1A and B). From 2016 to 2020, there was a decrease of 45.63% (Figure 2) in azithromycin for the total number of prescriptions per 100,000 Medicaid enrollees. The total number of prescriptions for hydroxychloroquine was 634,914; 680,191; 673,354; 629,653 and 716,091 for 2016, 2017, 2018, 2019, and 2020, respectively (Figure 1A and B). There was an increase in the total number of prescriptions per 100,000 Medicaid enrollees for hydroxychloroquine by +19.8% (Figure 2). The number of prescriptions for chloroquine was 2,264; 2,676; 2,850; 2,435 and 1,728 for 2016, 2017, 2018, 2019, and 2020, respectively (Figure 1). There was an overall decrease of -18.9% in the total number of prescriptions per 100,000 Medicaid enrollees for chloroquine (Figure 2). The average cost for azithromycin was $12.95 in 2016 and $10.30 in 2020 with a percentage decrease of -20.4% from 2016 to 2020 (Figure 3). The average cost of hydroxychloroquine was $103.5 in 2016 and $26.67 in 2020 with a percentage decrease of -74.2% from 2016 to 2020 (Figure 3). Our results indicated an increase in the average cost of primaquine by +139.42% from 2016 to 2020 (Figure 3). The prescribing pattern in 2020 for quarter 1 and quarter 2 is shown in Figure 4. There was an overall increase in the number of prescriptions of hydroxychloroquine from quarter 1 of 2020 to quarter 3 of 2020. There was a decrease in number of prescriptions of azithromycin from quarter 1 (1,721,804) to quarter 2 (791,207) but an increase in the number of prescriptions from quarter 2 of 2020 to quarter 3 (919,385) and quarter 4 (1,026,931) of 2020. The prescribing pattern for chloroquine stayed relatively constant through 2020.

Hydroxychloroquine, Azithromycin, and Chloroquine Prescribing Patterns in Medicaid

Figure 3. Percentage change in the average cost of each drug from 2016 to 2020 in Medicaid

Figure 1. (A) Total number of prescriptions per quarter from 2016 to 2020 from Medicaid and (B) total number of prescriptions per 100,000 enrollees in Medicaid.

Figure 4. Number of prescriptions for first three quarters in 2020

Figure 2. Percentage change in the total number of prescriptions per 100,000 enrollees from 2016 to 2020 60

Hydroxychloroquine, Azithromycin, and Chloroquine Prescribing Patterns in Medicaid

Discussion The decrease in azithromycin prescriptions from 2016 to 2020 can be linked to emergence of new antibiotics with stronger function and rapid emergence of resistant bacteria as well as increased antibiotic resistance crisis in the states (6). Studies have reported, but not confirmed, additional benefits of using these drugs with azithromycin (6). Hydroxychloroquine was briefly touted as a treatment, or prophylactic, for COVID-19 (7). Hydroxychloroquine was widely prescribed to patients who were hospitalized in the beginning of COVID-19 pandemic (1). The increase in the number of prescriptions for hydroxychloroquine from quarter 1 of 2020 to quarter 3 of 2020 can be secondary to the COVID-19 pandemic in states. The number of outpatients with prescription dispensed from retail pharmacies decreased substantially, likely related to transient interest in azithromycin in COVID-19 management in hospital settings (8). Although a later study showed that there were no significant differences in in-hospital mortality between patients who received hydroxychloroquine with or without azithromycin and patients who received neither drug (9).

Shehab N, Lovegrove M, Budnitz DS. US hydroxychloroquine, chloroquine, and azithromycin outpatient prescription trends, October 2019 through March 2020. JAMA Intern Med. 2020;180(10):1384–6.

Sandman Z, Iqbal OA. Azithromycin. In: StatPearls. Treasure Island (FL): StatPearls Publishing; 2021.

Ventola CL. The antibiotic resistance crisis: part 1: causes and threats. P T. 2015;40(4):277–83.

Vaduganathan M, van Meijgaard J, Mehra MR, Joseph J, O’Donnell CJ, Warraich HJ. Prescription fill patterns for commonly used drugs during the COVID-19 pandemic in the United States. JAMA. 2020;323(24):2524–6.

King LM, Lovegrove MC, Shehab N, Tsay S, Budnitz DS, Geller AI, et al. Trends in U.S. outpatient antibiotic prescriptions during the COVID-19 pandemic. Clin Infect Dis [Internet]. 2020 [cited 2021 Jun 24]; Available from: https://academic.oup.com/cid/advance-article/ doi/10.1093/cid/ciaa1896/6054971

Gouin KA, Creasy S, Beckerson M, Wdowicki M, Hicks LA, Lind JN, et al. Trends in prescribing of antibiotics and drugs investigated for COVID-19 treatment in U.S. nursing home residents during the COVID-19 pandemic. Clin Infect Dis. 2021;

Rosenberg ES, Dufort EM, Udo T, Wilberschied LA, Kumar J, Tesoriero J, et al. Association of treatment with hydroxychloroquine or azithromycin with in-hospital mortality in patients with COVID-19 in New York state. JAMA. 2020;323(24):2493–502.

Conclusion During the early period of COVID-19 pandemic, there was an increase in prescription of hydroxychloroquine and decrease in the number of prescriptions for azithromycin and chloroquine. There are some limitations that need to be considered in interpreting the results of our study. First, the intrinsic disadvantages of web-based, online survey research include uncertainty over the validity of the data and sampling issues, as well as concerns surrounding the design, implementation, and evaluation of an online survey. Secondly, the Medicaid State Drug Utilization database updates frequently. The data from 2020 is still not up to date and hence limits our interpretation. Third, Medicaid only accounts for 20% of the population in states (10). We hope to further our research on the prescribing patterns of hydroxychloroquine, azithromycin, and chloroquine on a state-by-state basis to further assess the impacts of the COVID pandemic.

10. Mikulic M. Medicaid- Statistics & Facts [Internet]. Statista. com. 2020 [cited 2021 Jun 25]. Available from: https:// www.statista.com/topics/1091/medicaid/

Acknowledgments We would like to thank Brian Piper, PhD, Sonia Lobo, PhD, and Kimberly Miller, PharmD, for the professional guidance and feedback on this topic.

Disclosures There is no financial relationship between this paper’s authors and any institution mentioned herein.

References 1.

Geleris J, Sun Y, Platt J, Zucker J, Baldwin M, Hripcsak G, et al. Observational study of hydroxychloroquine in hospitalized patients with covid-19. N Engl J Med. 2020;382(25):2411–8.

Golamari R, Kapoor N, Devaraj T, Sahu N, Kramer C, Jain R. Experimental therapies under investigation for COVID-19. J Community Hosp Intern Med Perspect. 2021;11(2):187–93.

Scholarly Research In Progress • Vol. 5, November 2021

Changes in Morphine Distribution in the United States Megan E. Dowd1†‡, E. Jessica Tang1†‡, Kurlya T. Yan1†‡, Kenneth L. McCall2, and Brian J. Piper1 ¹Geisinger Commonwealth School of Medicine, Scranton, PA 18509 2 University of New England, Portland, ME 04103 † Doctor of Medicine Program ‡ Authors contributed equally Correspondence: kyan@som.geisinger.edu

Abstract

Introduction

Background: Morphine is one of the most commonly prescribed opioids by hospitals and pharmacies in the United States (U.S.). Morphine’s potent analgesic properties have also been associated with the increase in addiction, misuse, and opioidrelated deaths in the U.S. since the 1990s. This has led to heavier regulation of opioid usage and prescription within the past decade. Despite federal regulations, population-adjusted morphine distribution varies markedly between states, in part due to varying state laws. The objective of this study was to describe trends in morphine distribution amounts nationwide and between states from 2012 to 2019.

Since the federal deregulatory policies of the late 1980s that gave pharmaceutical companies greater direct-to-consumer marketing access, the United States (U.S.) has observed an increase in demand and prescription rates of opioids. This is partially attributable to aggressive advertising of opioids that were oftentimes misleading and labeled opioids as non-addictive (1). The ongoing opioid crisis due to the subsequent rise in opioid addiction and misuse was declared a national public health emergency in 2017.

Methods: Drug weight and population data were obtained from Report 5 of the U.S. Drug Enforcement Administration’s Automation of Reports and Consolidated Orders System (ARCOS) to describe trends in the distribution of morphine across the U.S. between 2012 and 2019. Morphine distribution amounts were separated by state and business type and adjusted by total grams distributed to each state business by the population of the state. The percent change in grams of morphine distributed per state population from 2012 to 2019 was then calculated. Results: In 2012, U.S. pharmacies and hospitals dispensed 24,600 kilograms of morphine. In 2019, that number had decreased to 11,900 kilograms, a 51.7% decrease. Notably, Oregon experienced the largest decrease, 68.2%. States with the highest morphine usage in 2012 also observed the highest declines in morphine distribution amounts over the 7 years. Conclusion: Distribution of morphine has substantially decreased in the last decade. This is observed nationally and unanimously statewide. The decline in the distribution of morphine in the U.S. may be attributable to increased prioritization of the opioid crisis as a public concern, resulting in increased funding of opioid misuse and treatment programs and decreased production quotas for opioids, including morphine. This decline also coincides with the national shortage of parenteral opioids resulting in greater prescriptions of alternative opioids such as nalbuphine and buprenorphine. Guidelines may also be effective in changing prescribing practices and can be considered when comparing differences in decline of morphine distribution between states. The Oregon Health Authority (OHA) implemented an Opioid Initiative in 2015 that increased access to nonopioid pain treatment, decreased opioid prescribing, and used data to inform policies and interventions. This coincides with Oregon having the greatest decrease in morphine prescriptions compared to all other states.

Morphine, a potent analgesic commonly prescribed for aroundthe-clock treatment of moderate to severe pain, is one of the most commonly abused opioids in the U.S. Adverse effects of morphine range from increased risk of infections to neonatal abstinence syndrome to death. Treating opioid abuse is estimated to cost $72 billion each year, comparable to medical costs for treating asthma and diabetes (2). Morphine is most frequently prescribed for chronic, postoperative, and acute outpatient pain. Prescriptions increased by 64% between 2004 and 2011 — the greatest increase observed across all opioid analgesic prescriptions — but after peaking in 2012 has since declined (3). This coincides with the U.S. Department of Health and Human Services’ $1 billion fund allocated to states to be used for addiction prevention, treatment, and recovery services, data collection, pain management, overdose reversing drugs such as naloxone, and research. Usage of these funds has largely lied within individual state legislation (4). Despite an overall national decrease in morphine prescriptions, there is marked variation between states concerning the percent change in population-adjusted opioid distribution since the peak of the opioid crisis in 2012. Therefore, the objective of this report is to 1) describe trends in morphine prescription from 2012 to 2019 and 2) compare these trends on a state-tostate basis.

Methods Procedures All data was collected through the Automation of Reports and Consolidated Orders System (ARCOS). ARCOS is an automated, comprehensive drug reporting system that allows the U.S. Drug Enforcement Administration to monitor the flow of controlled substances from the point of manufacture through commercial distribution channels to point of sale or distribution at the dispensing/retail level — hospitals, retail pharmacies, practitioners, mid-level practitioners, and teaching institutions (5). ARCOS tracks controlled substances transactions and

Changes in Morphine Distribution in the United States

monitors the distribution of controlled substances by weight (grams). The analysis for this study was for morphine distributed to pharmacies and hospitals from 2010 to 2019. The total grams per year per state and dispenser/retailer was accessed through the publicly available ARCOS Report 5. Analysis The programs GraphPad Prism, Microsoft Excel, and JMP were used to graph and analyze the data. The weight (g) contributed by practitioners, mid-level practitioners, and teaching institutions was negligible in comparison to hospital and pharmacy weight (g). We added pharmacy and hospital weights of morphine for each year per state and divided them by the population of the state for a population adjusted calculation for 2012, which was the peak year of total morphine prescribed, and 2019. Population information for Figure 1. Amount of morphine in kilograms distributed to different business types from 2012 2012 and 2019 was obtained from the to 2019, as reported by the Drug Enforcement Administration’s Automated Reports and U.S. Census Bureau. Then, the percent Consolidated Ordering System (ARCOS). The percent change relative to the peak year is shown in change between 2012 and 2019 was parentheses. calculated for each state using their respective population adjusted weights. prescription, Illinois, dispensed 3,994.72 grams/100K in 2012 Statistical significance of the percent and 2,634.47 grams/100K in 2019 — a statistically significant change for each state was determined using a 95% confidence decrease of 34.1%. interval, which was calculated as 1.96 times the standard deviation from the national average in the U.S., calculated by taking the average of all percent changes. This study was Discussion approved by the Institutional Review Boards of Geisinger and This study identified two key findings about the use of the University of New England. morphine in the U.S. in the past decade. First, total morphine prescriptions from hospitals and pharmacies across the U.S. Results substantially decreased by 51.7% since peaking in 2012. However, the size of this reduction varied across the 50 states, The peak year for total morphine distributed between 2010 with an approximately two-fold difference between the largest and 2019 was 2012, during which hospitals and pharmacies percent decrease in morphine prescriptions (-68.19%) seen in dispensed 24,166070.11 grams, compared to 11,946,706.94 Oregon from 2012 to 2019, and the smallest percent decrease grams in 2019 (Figure 1). (-34.05%) seen in Illinois within the same time period. Hospital and pharmacy opioid distribution were broken down The overall reduction in morphine prescriptions nationally can by state, by the morphine weight distributed per person by be attributed to more aggressive and comprehensive policies each state’s population (Figure 2). When comparing 2012 and and initiatives prioritizing the opioid crisis as a healthcare issue. 2019, several changes were notable. Firstly, between the peak The US Drug Enforcement Administration (DEA) decreased year of 2012 and 2019, hospital and pharmacy distribution of quotas after the passage of the SUPPORT Act, which called on morphine declined by 51.7% nationally. Secondly, in 2012, the the DEA to quantify diversion of prescription opioids and “make states prescribing the most morphine, Tennessee (18,016.36 appropriate quota reductions” (6). In 2012, the DEA quota for grams/100K), Oregon (15,480.77 grams/100K), and Arizona morphine production (for sale) was 48,200,000 grams; this was (15,330.38 grams/100K) saw some of the largest decreases decreased by 39% to 29,353,655 in 2019 (7, 8). Interestingly, in prescription between 2012 and 2019; Tennessee had a this reduction in the overall production quotas of opioids is 61.6% reduction, Oregon had a 68.2% decline, and Arizona concurrent with the rise in the production of marijuana (6). had a 65.6% reduction. Thirdly, in 2012, the state of Oregon Indeed, another factor contributing to the decline in morphine was prescribing 15,480.78 grams/100K of morphine; in 2019, prescriptions may be the nationwide shortage of parenteral that number had dropped to 4,924.88 grams/100K — a 68% opioids (notably morphine, hydromorphone, and fentanyl) decrease, which is statistically significant and the largest that has resulted in diversion to alternative opioids such as decrease in the U.S. for morphine prescription between 2012 nalbuphine and buprenorphine (9). and 2019. The state with the smallest reduction in morphine

Changes in Morphine Distribution in the United States

Figure 2. (A) Hospital- and pharmacy-distributed morphine in grams per person per state in 2012 and 2019 as reported by the Drug Enforcement Administration’s Automated Reports and Consolidated Ordering System (ARCOS). (B and C) Percent decrease from 2012 to 2019 in A. Average decrease was -51.66%. *State values that were outside a 95% confidence interval, calculated as mean ± (1.96 x standard deviation), namely those of Oregon and Illinois, were considered statistically significant.

The variability among different states with regard to their success in reducing morphine prescriptions may be attributable to differences in state policies in handling the opioid crisis. In 2019, pharmacies and hospitals in Texas distributed 2116.52 grams/100K of morphine, the smallest population-adjusted amount out of all 50 states excluding Washington D.C. The decline in Texas’ prescription rates of morphine and all other opioids in general have been attributed to the implementation of several state policies, including the “pill mill” law passed in 2010 that led to significant decreases in monthly morphine prescription volume (10). Oregon is another example of a state in which morphine prescriptions have been reduced through policy efforts. The Oregon Health Authority launched The Opioid Initiative in 2015, which works to increase access to nonopioid pain treatment, support medication-assisted

treatment and naloxone access for those using opioids, decrease opioid prescribing, and use data to inform policies and interventions. From 2015 to 2017, the number of Oregonians on 90 or more morphine equivalent doses (MEDs) decreased by 37%, from 11.1 per 1,000 residents quarterly to 7.0 per 1,000 residents quarterly. Prescription opioid overdose deaths decreased 20% from 4.5 per 100,000 in 2015 to 3.6 per 100,000 in 2016 (11). Policies limiting opioid prescriptions (regulations for prescribing to “high-risk” Medicaid payers, required urine drugs tests and documentation of justification for high-dose opioid prescriptions) may account for the success of Oregon’s Opioid Initiative, especially as 4.1% of all prescribers were responsible for 60% of controlled substance prescriptions in Oregon (2).

Changes in Morphine Distribution in the United States

Drug Enforcement Administration, Department of Justice. (2012). Controlled substances: final adjusted aggregate production quotas for 2012. Retrieved 2021 from: 2012 - Final Adjusted Aggregate Production Quotas for 2012 (usdoj.gov)

Future directions for this analysis can include further investigation of the differences that may result in state-bystate variation, i.e., availability of morphine (including brand and generic manufacturers for both oral and IV morphine) and reliance on opioid adjuncts in each state. Characterization of the primary patient populations that are prescribed morphine in each state, as rural residents and Medicaid payers are associated with higher rates of morphine prescriptions and adverse effects of morphine usage (12).

Drug Enforcement Administration, Department of Justice. (2019). Proposed aggregate production quotas for schedule I and II controlled substances and assessment of annual needs for the list I chemicals Ephedrine, Pseudoephedrine, and Phenylpropanolamine for 2020. Retrieved 2021 from: https://deadiversion.usdoj.gov/ fed_regs/quotas/2019/fr0912.htm

In conclusion, this study was conducted to observe the national and state-by-state changes in morphine prescriptions by pharmacies and hospitals in the U.S. from 2012, the peak of the opioid crisis, to 2019, the most recent year with completed ARCOS data. Since 2012, morphine prescriptions have decreased in every state, although states with higher morphine distribution in 2012 were observed to have more dramatic decreases in morphine distribution over the 7-year period.

Davis, M. P., McPherson, M. L., Mehta, Z., Behm, B., & Fernandez, C. (2018). What parenteral opioids to use in face of shortages of Morphine, Hydromorphone, and Fentanyl. The American Journal of Hospice & Palliative Care, 35(8), 1118–1122. https://doi. org/10.1177/1049909118771374

10. Lyapustina, T., Rutkow, L., Chang, H. Y., Daubresse, M., Ramji, A. F., Faul, M., Stuart, E. A., & Alexander, G. C. (2016). Effect of a "pill mill" law on opioid prescribing and utilization: the case of Texas. Drug and Alcohol Dependence, 159, 190–197. https://doi.org/10.1016/j. drugalcdep.2015.12.025

A strength of this study was that it consolidated both hospital and pharmacy morphine distributions. This controls for potential inconsistencies in defining “hospital” vs “pharmacy” prescribed opioids, as postoperative morphine prescriptions that are written in a hospital but filled in an outside pharmacy are considered “pharmacy-distributed” by ARCOS. A potential limitation of this study is that the drug distribution amounts are listed in weight (grams) rather than number of prescriptions.

References 1.

Donohue J. A history of drug advertising: the evolving roles of consumers and consumer protection. Milbank Q. 2006;84(4):659-699. doi:10.1111/j.14680009.2006.00464.x

U.S. Department of Health and Human Services. Addressing prescription drug abuse in the United States: current activities and future opportunities. (2013, September). Retrieved 2021 from https://www.cdc. gov/drugoverdose/pdf/hhs_prescription_drug_abuse_ report_09.2013.pdf

Piper, B. J., Shah, D. T., Simoyan, O. M., McCall, K. L., & Nichols, S. D. (2018). Trends in medical use of opioids in the U.S., 2006-2016. American Journal of Preventive Medicine, 54(5), 652–660. https://doi.org/10.1016/j. amepre.2018.01.034

Assistant Secretary of Public Affairs (ASPA). (n.d.). What is the U.S. opioid epidemic? HHS.gov. Retrieved 2021 from https://www.hhs.gov/opioids/about-theepidemic/index.html

U.S. Department of Justice Drug Enforcement Administration Diversion Control Division. (n.d.). Automation of Reports and Consolidated Orders System (ARCOS). Retrieved 2021 from https://www.deadiversion. usdoj.gov/arcos/index.html

DEA proposes to reduce the amount of five opioids manufactured in 2020, marijuana quota for research increases by almost a third. (2019, September 11). Retrieved 2021 from https://www.dea.gov/pressreleases/2019/09/11/dea-proposes-reduce-amount-fiveopioids-manufactured-2020-marijuana-quota

11. Hedberg K, Bui LT, Livingston C, Shields LM, Van Otterloo J. Integrating public health and health care strategies to address the opioid epidemic: The Oregon Health Authority's Opioid Initiative. J Public Health Manag Pract. 2019 May/Jun;25(3):214-220. doi: 10.1097/ PHH.0000000000000849. PMID: 30048336. 12. Haegerich, T. M., Paulozzi, L. J., Manns, B. J., & Jones, C. M. (2014). What we know, and don't know, about the impact of state policy and systems-level interventions on prescription drug overdose. Drug and Alcohol Dependence, 145, 34–47. https://doi.org/10.1016/j.drugalcdep.2014.10.001

Scholarly Research In Progress • Vol. 5, November 2021

Chronic Traumatic Encephalopathy: A Literature Review Yvette M. Johnson1*, Chloé E. Mballa1*, Taylor S. Mewhiney1*, Cathie-Allegra Z. Nkabyo1*, and Grace L. Tieko1* ¹Geisinger Commonwealth School of Medicine, Scranton, PA 18509 *Master of Biomedical Sciences Program Correspondence: tmewhiney@som.geisinger.edu

Abstract Chronic traumatic encephalopathy (CTE) is a neurodegenerative disease that exhibits abnormally high amounts of tau, a protein found in brain cells. The cause of CTE is still relatively unknown. One’s prolonged experience of head trauma leads the medical community to believe that CTE is the main culprit. The phenomenon, occurring primarily with American football players, was first referred to as “punch drunk syndrome” in 1928. Currently, CTE cannot be diagnosed premortem. Examining the brain tissue post-mortem is the only way to officially diagnose CTE. Unfortunately, there are no known cures. However, pharmacologic and nonpharmacologic interventions can be used to lessen the symptoms. The purpose of this study was to critically analyze current methods of diagnosing CTE and identify limitations within the diagnostic methods by reviewing peer reviewed articles from various credible databases (PubMed, JSTOR, ClinicalKey) and journals. As more studies are conducted and technology advances to allow brain tissue diagnosis premortem, a cure for chronic traumatic encephalopathy may arise.

Introduction Chronic traumatic encephalopathy (CTE) is a neurodegenerative disease that typically comes from head impact related injuries (1). It can cause memory loss, changes in behavior, and cognitive impairment (2). In those under 50 with continual head trauma, it has been seen that traumatic brain injuries can also cause severe disabilities such as seizures, psychiatric disorders, and personality changes (3). In general, CTE is affiliated with serious morbidity and mortality rates (4). It is often found in people who partake in contact sports (1). However, in some cases, CTE has been found in people with no known history of neurotrauma (5). Typically, brain injuries have two phases, the initial injury and then the cascade or progressive aftermath to follow, which is common in repetitive injuries like CTE which often progress after multiple blows to the head (6). Currently there are no diagnostic criteria, which makes it hard to differentiate CTE from other neurodegenerative diseases (7). The only identification marker for risk factor is repetitive head trauma (8). Along with this, the only time a diagnosis is possible is postmortem, which makes CTE hard to detect and find (9). During the last decade, there are a lot of unanswered questions about the disease, which are slowly being discovered. Until recently, CTE was not as widely researched, and much of the research performed had sans primary data collection, potentially leading to biases and inaccuracies (10). The progressing recognition recently has caused an increase in the awareness and there has been an uptick in conferences and activity centered around the gaps in CTE (11, 12). Recently, many different sports leagues have been looking into means of prevention and enforcing rules and regulations to avoid

head trauma that may lead to CTE (13). The first reported case of CTE in a football player was in 2005, which ushered in a newfound interest in the topic (8). For athletes, this can cause long-term effects that lead to mental health issues such as anger, depression, and suicidal ideation that may result in death by suicide (14). There have been numerous studies that look at the psychological consequences, and the correlation between those with a diagnosis of CTE (postmortem), and suicide (2, 15). In some cases, CTE and post-traumatic stress disorder have been shown to overlap (16). The correlation between CTE and Alzheimer’s is also a frequently researched topic (17). At this point in time, pathological findings have shown that the brain has an increase in a protein called tau in the superior colliculus (18, 19). There is an extreme need for a new diagnosis process, as the suspicion in many causes added distress and anxiety (17). In lieu of this, there has been an increased use of magnetic resonance imaging (MRI) and magnetic resonance spectroscopy (MRS) on live subjects to diagnose prior to death. This has the potential to pave the way to track patterns that can be diagnosed in vivo (20). While there is research in the literature on CTE pathology and diagnosis, disadvantages in the developments in the literature can still be found.

Methods A critical examination of CTE pathology and diagnosis was conducted and several credible databases and scholarly articles on CTE were utilized. We conducted this study by gathering peer-reviewed literature from our databases. References were obtained through ClinicalKey, JSTOR PubMed, Elsevier, Springer, and NCBI. The journals completed a peer-review process and were deemed reliable and acceptable. Keywords searched for included: chronic traumatic encephalopathy, neuropathology, brain trauma, neurodegenerative, traumatic brain injury, Alzheimer’s, and tau phosphorylation.

Discussion Clinical symptoms of CTE Repetitive brain trauma can eventually lead to the deterioration of the brain. It is often seen in individuals that partake in aggressive contact and collision sports such as football, boxing, soccer, ice hockey or even rugby (21). Boxers were the first to be officially diagnosed with CTE, and the first bull rider to be diagnosed with CTE was in 2018 (22, 23). Fourteen retired soccer players who were considered headers participated in a longitudinal study from 1980 to 2010 monitoring their cognitive diseases that developed around 60 years old (24). In addition to the development of concussion through these sports, CTE can also manifest through minor repeated blows to the head via physical abuse from another individual or self-harm (25). Through brain tissue analysis, the gross neuropathological ﬁndings included cerebral, thalamus, and hypothalamus atrophy

Chronic Traumatic Encephalopathy: A Literature Review

(reduced gray and white matter), enlarged ventricles, cavum septum pellucidum and depigmentation of the substantia nigra and locus coeruleus, with microscopic evidence of neuroﬁbrillarytangles (NFT) in the cortex and brainstem, sometimes associated with senile plaques or neuronal loss (26, 27). This may lead to a variety of behavioral, physical, and psychological symptoms. CTE is clinically associated with symptoms of irritability, aggression, depression, short-term memory loss, and heightened suicidality that usually begins 8 to 10 years after experiencing repetitive mild traumatic brain injury (25, 28). There are four stages of physical symptoms in CTE. Frequent headaches, loss of attention, mild aggressive symptoms, and lack of concentration are seen in stage one (29). As the symptoms progress to stage two, individuals suffer from depression, loss of short-term memory, and sudden outbursts of anger, in addition to the symptoms from stage one. Some other less-common symptoms seen at this stage are executive dysfunction, language difficulties, and impulsivity (30). Stage three continues with a compilation of the previously mentioned symptoms with the addition of suicidal thoughts. Patients typically display more cognitive deficits, ranging from memory loss to executive and visuospatial functioning deficits and apathy (29). A study demonstrated that cognitive impairment symptoms were significantly higher in participants self-reporting CTE (31). The most severe stage is stage four, in which the development of memory loss with dementia, aggressive tendencies, paranoia, and issues with gait occur (30). A study explains that the most common causes of death for those with CTE are respiratory failure, cardiac disease, overdose, and symptoms associated with end-stage dementia and malignancy (30). A study showed that adverse events in childhood and ongoing life stressors can result in depression based on symptoms from neuropathology characteristic of CTE (32). All these clinical symptoms are linked to certain pathological abnormalities in the brain. Although these clinical symptoms can be identified pre-mortem, many of these are not diagnosed as CTE, but rather as a presenting symptom of other diseases. Due to this, diagnoses are often misinterpreted as comorbidities. Hence, the importance of evaluating CTE pathology is critical to identify unique differences in its physical presentations.

IV, there is further advanced cerebral, medial temporal lobe, hypothalamic, thalamic, and mammillary body atrophy, septal abnormalities, ventricular dilatation, and pallor of substantia nigra and locus coeruleus; phospho-tau in widespread regions including white matter, with prominent neuronal loss, gliosis of cortex, and hippocampal sclerosis. These stages build on one another, carrying on the symptoms and pathologies from the previous stage (34). Traumatic brain injury (TBI) can result in intracranial hemorrhage, which may lead to brain herniation across dural or skull-defined compartments. Because increased intracranial pressure can often lead to deterioration and death, herniation of the brain will be considered before other types of injuries are addressed (36). Based on the Glasgow Coma Scale (GCS), which measures the level of consciousness in a person, TBI severity ranges from a scale of zero (most severe) to 13 (minor) (37, 38). In the absence of hypoxic-ischemic injury, children with traumatic brain injury and Glasgow Coma Scale scores of 3 to 5 can recover independent function (39). In some circumstances, some individuals with CTE neuropathology show no symptoms or clinical indicators that correlate with TBI (40). Pediatric TBI occurs mostly in adolescents and young adults, followed by a secondary peak in infancy. Motor vehicle accidents, assault, and concussions in lieu of contact during sports are the most frequent causes for head injuries in children. Children are most susceptible to falls at age 5 and under, and infants are vulnerable to abusive head trauma (AHT) which causes severe TBI. TBI is two times more frequently found in boys than in girls, with a period of distinction between the childhood and adolescent years (41). One of the main contact sports played by children that may later result in the development of CTE is rugby. It has been demonstrated that children who participate in rugby have an increased risk of concussion, which may consequently result in serious impairments later in adulthood (42). The benefits of physical activity do not outweigh the costs when it comes to childhood

CTE pathology Along with the symptoms, there are also pathological defects to the brain. Gross anatomical abnormalities found during brain autopsies are consistent with the current understanding of CTE. These abnormalities, which may result from underlying neurodegenerative processes, include an overall reduction in brain weight (33). In Figure 1, the four stages of CTE symptoms are depicted. In stage I, there are perivascular phospho-tau neurofibrillary tangles in focal epicenters at the depths of the sulci which are measured in the dorso-lateral prefrontal cortex (34, 35). In stage II, there is a progression of neurofibrillary tangles in the superficial cortical layers adjacent to the focal epicenters and in the nucleus basalis of Meynert and locus coeruleus (34). In stage III, there is a mild cerebral atrophy, septal abnormalities, third ventricular dilatation, depigmentation of locus coeruleus and substantia nigra, dense phospho-tau pathology in the cortex, medial temporal lobe, diencephalon, brainstem, and spinal cord (34, 21). Lastly in stage

Figure 1. This figure shows the progression and development of symptoms of CTE.

Chronic Traumatic Encephalopathy: A Literature Review

brain trauma and injuries. The long-term effects are not worth the risks when it comes to the health of young children and adolescents (43). Based on the ongoing CTE research, current results demonstrate that there are no significant differences in the neuro pathological and clinical outcomes in all contact sports. However, does one have an increased chance of developing CTE depending on the position played in collision sports? Although unknown for all positions in collision sports, there needs to be more systematic research about all the possible contact sports in the world in order to further evaluate and understand this hypothesis. It is also unknown if the duration of time in recovery after a collision has an effect in the development of CTE. A larger longitudinal study should be conducted amongst athletes that have encountered repetitive head injuries throughout their lifetime for future research. These findings will be informative to understand the specific collision injury variables that may influence CTE. Clinical presentation: cognitive, motor, behavioral, psychiatric, neuroendocrine Current data suggests that there are three possible domains that CTE symptoms fall under for clinical presentation: motor functioning, cognition, and behavior/mood (44). In stages I through IV, symptoms such as mood swings, suicidality, paranoia, dementia, memory loss, and cognitive impairment are observed (45). By stage IV, about 10% of patients develop a motor neuron disease that closely resembles amyotrophic lateral sclerosis (ALS) (46). Currently, there is no distinct clinical diagnostic criteria for CTE, as symptoms are varied and not yet agreed upon (47, 48). The current symptoms of CTE that have been supported by literature consist of anxiety, anger, depression, headaches, suicidality, anger management problems, dysarthric speech, mild cognitive impairment, ataxia, Parkinsonism, and dementia (49). Furthermore, some people that do not exhibit any clinical features of CTE have still been diagnosed with the disease based solely on neuropathology (49, 50). There are reportedly four stages of CTE (45). Typically, the onset and advancement of the disease is slow and usually occurs 8 to 10 years after the first concussions (51). There is still much uncertainty regarding the causality of CTE (52, 53). Multiple studies on CTE research utilize recall of clinical symptoms from the deceased patient’s close family and

friends. Moreover, the disease can only be positively diagnosed postmortem (46). For example, the high-profile case of retired NFL player, Junior Seau, whose family donated his brain to the National Institutes of Health (NIH) after his suicide, was affirmatively diagnosed with CTE postmortem (54, 55). Observed in multiple cases of individuals suffering from CTE is a cavum septum pellucidum (56). Additionally, reported findings associated with this condition include volume loss in the cerebellum and hippocampus (56). Furthermore, amygdala atrophy is compatible with suicidal ideation (56). One study of a 39-year-old man and retired NFL player observed white matter damage and loss of brain volume (as shown in Figure 2) in deep gray matter structures utilizing diffusion tensor imaging (57). Severe CTE has been linked to impulsivity, anxiety, depressive symptoms, and explosivity, which are associated with prefrontal cortex, amygdala, and locus coeruleus regions of the brain (58). CTE often clinically presents indistinctly but is often distinguished by two apparent phenotypes: affective changes and cognitive impairment (59). Genetics have not been linked to any clear risk factors (59). TBI impairs cognitive control, which is the capability to manage actions and achieve goals (60). TBI can impact individuals in early childhood and can be evaluated using the Physical and Neurological Examination of Subtle Signs (PANESS), which examine subtle motor signs, utilizing MRI to measure total cerebral/motor/premotor volume (61). In addition to the clinical symptoms of CTE, the phosphorylation of tau and Aβ pathologies is also identified in Alzheimer’s disease (62, 63). Our understanding of CTE is still developing and many important questions remain unanswered. The causality of CTE is still not fully understood; is it a result of repetitive blows to the head or concussions or a combination? Diagnosing CTE postmortem does not allow for true intervention or prevention of the disease. Future studies should explore the effects of genetics, comorbidities, and lifestyle factors. Thus, there is room for further studies to be conducted. CTE and association with dementia/Alzheimer’s disease CTE is often correlated with Alzheimer’s disease due to similarities in their pathologies and cellular mechanisms, and PET imaging (64, 65, 66). Although it is unclear whether CTE is progressive or solely linked to repetitive head trauma, scientists can agree that a key player in both CTE and Alzheimer’s disease is tau (67). Tau is a protein in neurons that is encoded by a single gene on chromosome 17. As a result of alternative splicing, Tau can have six different isoforms that aid in stabilizing microtubules (64, 68). Tau proteins can precipitate into aggregates which can result in diverse neurofibrillary tangles (69). Upon aggregation in the central nervous system, Tau prions can cause tauopathies such as Alzheimer’s disease and CTE (71, 72). This happens in mice with repeated head trauma and causes severe frontal brain injury to humans (73, 74). This aggregation is caused by the level of phosphorylation of tau. Hyperphosphorylation of tau catalyzes microtubule detachment, the inhibition of axonal transport between the soma and synapse which subsequently results in neurodegeneration (71, 72).

Figure 2. This figure shows the difference between a normal brain and an advanced CTE brain. A decrease in brain volume is demonstrated.

Various analyses have demonstrated this neurodegeneration may be linked to abnormal phosphorylation in the sulci and peri-vascular

Chronic Traumatic Encephalopathy: A Literature Review

regions, microgliosis, and astrocytosis (75, 76). Other studies specifically looked at the neuropathological changes in the cortical degeneration of CTE and Alzheimer’s disease patients, and results indicated that tau pathologies in CTE and Alzheimer’s were more prevalent in the upper cortex in both disorders (70, 79). However, the neurodegeneration mechanism remains undiscovered (80). Additionally, the tau proteins were found to be more widely distributed in Alzheimer’s disease, which suggested that in the later stages of CTE, patients may develop Alzheimer’s disease (76). Although there is agreement on where the tau protein aggregates, there is still some debate about whether this may be surrounding the blood vessel or containing a blood vessel (66). Another investigation focused on identifying whether patients with CTE and dementia have distinct clinical features when compared to patients with Alzheimer’s disease (81). In this study, the researchers performed neurological examinations, neuropsychological testing, and emotional/behavioral testing. After testing, it was found that CTE’s clinical presentation was not significantly different from Alzheimer’s disease. Additionally, neurocognitive impairments and neurobehavioral symptom differences were not significant (81). These findings implied that the neuropathologic processes between Alzheimer’s disease and CTE may be associated (81). An additional study confirmed a correlation between these diseases' neuropathologic processes through postmortem analyses, which demonstrated white matter with oligodendrogenesis in both CTE and Alzheimer's disease patients (82). However, this study indicated that even with the similarities between the two diseases, it is still unclear how the severity of these neuropathologic processes play a role in triggering unfavorable behavioral outcomes (82). Some studies have suggested that neuroinflammation may drive the onset of tau phosphorylation, which consequently may lead to neuronal loss and synaptic dysfunction (83, 84). Because there is a long latency period between a history of TBI and the later onset of dementia symptoms, there may be a possibility of intervention in the early stages of CTE and Alzheimer’s disease (83, 84). One approach is the use of non-steroidal anti-inflammatory drugs (NSAIDs) and anti-inflammatory minocycline. The latter has been shown to reduce tau phosphorylation in serine residues in mice, reduce astrocytosis, and pro-inflammatory cytokines (83). Additionally, researchers believe NSAIDs may be able to modulate γ-secretase, which may decrease Aβ production, which may help in Alzheimer’s disease and traumatic brain injuries (83). Some investigators have analyzed targeted gene profiling in which downregulation of microtubule associated proteins may help with CTE pathology (80). All treatments are promising, but further research is needed into these new developments (83). As indicated by these studies, the pathological processes of CTE are still very unclear. Although some treatments have been suggested through mice studies, there is currently no definitive way of preventing CTE. Additionally, there is controversy with regards to the similarities between Alzheimer’s and CTE’s clinical and pathological features which demands further research into this topic. Further research may be limited due to the inability to diagnose pre-mortem; therefore, it may be of importance to focus research on distinct clinical symptoms and pre-mortem diagnosis. Possible imaging technologies may be able to help with analysis of neural changes while patients may still be alive.

Examining CTE genetics in contact sports through imaging technology CTE has various potential genetic distinctions. A variant in the TMEM106B gene is associated as a potential risk for developing CTE (84). Quantitative proteomics is a process used to identify certain proteins, such as high levels of tau, in CTE brain tissues (85). The expression of another gene, claudin-5, observed in areas of high levels of phosphorylated tau suggests that CTE may also cause blood-brain barrier dysfunction (86). A study that utilized brain tissue of military veterans from the Department of Veterans Affairs Biorepository Brain Bank (VABBB) found that the regions of the brain most affected by hyperphosphorylation of tau were the middle frontal gyrus, superior temporal gyrus, inferior parietal lobule, and hippocampus (87). Another study compared the positron emission tomography (PET) with 2-(1-{6-[(2-[F-18] fluoroethyl) (methyl)amino]-2naphthyl} ethylidene)malononitrile (FDDNP) that detects brain patterns of tau neuropathology distribution via β-pleated sheet conformation (88, 89). Results showed that military veterans have similar brain-binding compared to retired football players. Identifying CTE patterns in living brains with methods such as FDDNP-PET is giving researchers hope for a cure in the future (90). Ultimately, there are not many CTE studies that have female participants. This could continue to limit our understanding of CTE, thus prolonging the discovery of a cure. It must be addressed that one possibility for less female participation in such studies could be due to less female enrollment in contact sports compared to males. However, comparisons of CTE diagnosis between female and male participants of activities with repetitive head injuries should also be explored due to gonadal hormone levels affecting female and male brain structures and function differently, especially in response to trauma (91).

Conclusion The purpose of this paper was to explore the history, current science, and limitations behind CTE, its association with Alzheimer’s, dementia, and other diseases, and future CTE diagnosis developments. There are currently no standard guidelines for diagnosing CTE. Identifying CTE symptoms in patients premortem rely heavily on self-reporting which runs the risk of misdiagnosing patients with CTE (92). Factors such as standard aging, retirement lifestyle, substance abuse, sleep patterns, demographics, and surgical procedures should be considered before prematurely diagnosing patients with CTE (93). Male patients with a history of depression can exhibit clinical symptoms of CTE (94). Additionally, male patients with anger control problems could also be misdiagnosed with CTE if they have a history of repetitive head trauma (95). As previously mentioned, there is currently no cure for CTE, however, pharmacologic and nonpharmacologic interventions are available to treat symptoms of CTE. Medications like selective serotonin reuptake inhibitors or SSRIs, can help improve cognitive function (96). Cholinesterase inhibitors and antipsychotics can also be used to treat CTE symptoms. Nonpharmacologic treatments include living a healthy lifestyle by getting proper sleep, having a healthy diet, and getting

Chronic Traumatic Encephalopathy: A Literature Review

physical activity (96). Research has shown that the younger the person is exposed to contact sports, the earlier neurobehavioral symptoms are reported (97). In other words, the longer the participation in such activities, the greater the severity of CTE (98). Studies have evaluated the changes in white brain matter in ages 8- to 13-year-old tackle football players, and results suggest that children who play tackle football begin to see a decrease in neurodevelopment (99). This reduced density of white brain matter has been associated with dementia (100). Further research is needed to confirm that younger football players are at higher risk for developing CTE later in life. As more studies are conducted and technology advances to allow brain tissue diagnosis premortem, a cure for chronic traumatic encephalopathy may arise.

Acknowledgments This group of researchers would like to acknowledge the help received from Brian J. Piper, PhD, Sonia Lobo, PhD, and Semhal Mebrahtu, BA.

References

Banks SJ. Chronic Traumatic Encephalopathy (CTE). In: Neuro-Geriatrics. Cham: Springer International Publishing; 2017. p. 183–94.

Russell ER, McCabe T, Mackay DF, Stewart K, MacLean JA, Pell JP, et al. Mental health and suicide in former professional soccer players. J Neurol Neurosurg Psychiatry. 2020;91(12):1256–60.

Norins LC. The Beehive Theory: Role of microorganisms in late sequelae of traumatic brain injury and chronic traumatic encephalopathy. Medical Hypothesis. 2019;128:1–5.

Breen PW, Krishnan V. Recent preclinical insights into the treatment of chronic traumatic encephalopathy. Front Neurosci. 2020;14:616.

Iverson GL, Luoto TM, Karhunen PJ, Castellani RJ. Mild chronic traumatic encephalopathy neuropathology in people with no known participation in contact sports or history of repetitive neurotrauma. J Neuropathol Exp Neurol. 2019;78(7):615–25.

10. Wolfson DI, Kuhn AW, Kerr ZY, Brett BL, Yengo-Kahn AM, Solomon GS, et al. Chronic traumatic encephalopathy research viewed in the public domain: What makes headlines? Brain Inj. 2020;34(4):528–34. 11. Hay J, Johnson VE, Smith DH, Stewart W. Chronic traumatic encephalopathy: The neuropathological legacy of traumatic brain injury. Annu Rev Pathol. 2016;11:21–45. 12. McKee AC, Cairns NJ, Dickson DW, Folkerth RD, Keene CD, Litvan I, et al. The first NINDS/NIBIB consensus meeting to define neuropathological criteria for the diagnosis of chronic traumatic encephalopathy. Acta Neuropathol. 2016;131(1):75–86. 13. Nitrini R. Soccer (Football Association) and chronic traumatic encephalopathy: A short review and recommendation. Dement Neuropsychol. 2017;11(3): 218–20. 14. Iverson GL, Terry DP, Luz M, Zafonte R, McCrory P, Solomon GS, et al. Anger and depression in middle-aged men: Implications for a clinical diagnosis of chronic traumatic encephalopathy. J Neuropsychiatry Clin Neurosci. 2019;31(4):328–36. 15. U.S. Sports Academy. Chronic traumatic encephalopathy (CTE) and former national football league player suicides [Internet]. Thesportjournal.org. 2016 [cited 2021 May 29]. Available from: https://thesportjournal.org/article/chronictraumatic-encephalopathy-cte-and-former-nationalfootball-league-player-suicides/ 16. Nemani SK, Notari S, Cali I, Alvarez VE, Kofskey D, Cohen M, et al. Co-occurrence of chronic traumatic encephalopathy and prion disease. Acta Neuropathol Commun. 2018;6(1):140. 17. Graham NS, Sharp DJ. Understanding neurodegeneration after traumatic brain injury: from mechanisms to clinical trials in dementia. J Neurol Neurosurg Psychiatry. 2019;90(11):1221–33. 18. Armstrong RA, Mckee CCN. Pathology of the Superior Colliculus in Chronic Traumatic Encephalopathy. Journal of the Am Acad of Optometry. 2017 19. Lee BG, Leavitt MJ, Bernick CB, Leger GC, Rabinovici G, Banks SJ. A systematic review of positron emission tomography of tau, amyloid beta, and neuroinflammation in chronic traumatic encephalopathy: The evidence to date. J Neurotrauma. 2018;35(17):2015–24.

Morley WA. Environmental subconcussive injury, axonal injury, and chronic traumatic encephalopathy. Front Neurol. 2018;9:166.

Asken BM, Sullan MJ, DeKosky ST, Jaffee MS, Bauer RM. Research gaps and controversies in chronic traumatic encephalopathy: A review: A review. JAMA Neurol. 2017;74(10):1255–62.

20. Ruprecht R, Scheurer E, Lenz C. Systematic review on the characterization of chronic traumatic encephalopathy by MRI and MRS: Characterization of CTE by MRI and MRS. J Magn Reson Imaging. 2019;49(1):212–28.

Pearce AJ, Sy J, Lee M, Harding A, Mobbs R, Batchelor J, et al. Chronic traumatic encephalopathy in a former Australian rules football player diagnosed with Alzheimer’s disease. Acta Neuropathol Commun. 2020;8(1):23.

Pierre K, Dyson K, Dagra A, Williams E, Porche K, LuckeWold B. Chronic traumatic encephalopathy: Update on current clinical diagnosis and management. Biomedicines. 2021;9(4):415.

21. Alosco ML, Stern RA. Youth Exposure to Repetitive Head Impacts From Tackle Football and Long-term Neurologic Outcomes: A Review of the Literature, Knowledge Gaps and Future Directions, and Societal and Clinical Implications. Semin Pediatr Neurol. 2019;30:107–16. 22. McKee AC. The neuropathology of chronic traumatic encephalopathy: The status of the literature. Semin Neurol. 2020;40(4):359–69.

Chronic Traumatic Encephalopathy: A Literature Review

23. Keene CD, Latimer CS, Steele LM, Mac Donald CL. First confirmed case of chronic traumatic encephalopathy in a professional bull rider. Acta Neuropathol. 2018;135(2): 303–5.

39. Lieh-Lai MW, Theodorou AA, Sarnaik AP, Meert KL, Moylan PM, Canady AI. Limitations of the Glasgow Coma Scale in predicting outcome in children with traumatic brain injury. J Pediatr. 1992;120(2 Pt 1):195–9.

24. Ling H, Morris HR, Neal JW, Lees AJ, Hardy J, Holton JL, et al. Mixed pathologies including chronic traumatic encephalopathy account for dementia in retired association football (soccer) players. Acta Neuropathol. 2017;133(3):337–52.

40. Bailes JE, Origenes AK, Alleva JT. Chronic traumatic encephalopathy. Disease-a-Month. 2019;65(10):342–3.

25. Chen L. What triggers tauopathy in chronic traumatic encephalopathy? Neural Regen Res. 2018;13(6):985. 26. McKee AC, Alosco ML, Huber BR. Repetitive head impacts and chronic traumatic encephalopathy. Neurosurg Clin N Am. 2016;27(4):529–35. 27. McKee AC, Stein TD, Kiernan PT, Alvarez VE. The neuropathology of chronic traumatic encephalopathy: CTE neuropathology. Brain Pathol. 2015;25(3):350–64. 28. Randolph C. Is Chronic Traumatic Encephalopathy a Real Disease? Current Sports Medicine Reports. 2014;13(1):33–7. 29. Fesharaki-Zadeh A. Clinical characteristics of chronic traumatic encephalopathy. Psychiatric Times. 2019 Aug 29; 30. Ortiz A. Learn the symptoms in the four stages of CTE. The Boston globe [Internet]. 2017 Sep 21 [cited 2021 Jun 4]; Available from: https://www.bostonglobe.com/ metro/2017/09/21/symptoms-watch-for-four-stages-cte/ Q1wniQOnQXH1bU8OibU3WJ/story.html 31. Grashow R, Weisskopf MG, Baggish A, Speizer FE, Whittington AJ, Nadler L, et al. Premortem chronic traumatic encephalopathy diagnoses in professional football: Pre-mortem CTE diagnoses in football. Ann Neurol. 2020;88(1):106–12. 32. Iverson GL, Keene CD, Perry G, Castellani RJ. The need to separate chronic traumatic encephalopathy neuropathology from clinical features. J Alzheimers Dis. 2018;61(1):17–28. 33. Aldag M, Armstrong RC, Bandak F, Bellgowan PSF, Bentley T, Biggerstaff S, et al. The biological basis of chronic traumatic encephalopathy following blast injury: A literature review. J Neurotrauma. 2017;34(S1):S26–43. 34. McLauchlan D, Robertson NP. Chronic traumatic encephalopathy: quantifying risk and clarifying pathogenesis. J Neurol. 2020;267(3):870–2. 35. Goodman JC. Neuropathology of Traumatic Brain Injury. Youmans and Winn Neurological Surgery. 2017;338: 2765–77. 36. Mckee AC, Daneshvar DH. The neuropathology of traumatic brain injury. Grafman J, Salazar AM, editors. Handb Clin Neurol. 2015;127:45–66. 37. Budson AE, Solomon PR. Prevalence, Definition, Pathology, and Pathophysiology. Memory Loss, Alzheimer’s Disease and Dementia. 2021;15:172–9. 38. Petridou ET, Antonopoulos CN. Glasgow Coma Scale. In: Injury Epidemiology. 2017. p. 258–74.

41. Lerner JT, Giza CC. Traumatic Brain Injury in Children. In: Swaiman’s Pediatric Neurology. 2017. p. 781–93. 42. Kirkwood G, Parekh N, Ofori-Asenso R, Pollock AM. Concussion in youth rugby union and rugby league: a systematic review. Br J Sports Med. 2015;49(8):506–10. 43. Quarrie KL, Brooks JHM, Burger N, Hume PA, Jackson S. Facts and values: on the acceptability of risks in children’s sport using the example of rugby - a narrative review. Br J Sports Med. 2017;51(15):1134–9. 44. R. Martínez-Pérez, I. Paredes, P.M. Munarriz, B. Paredes, J.F. Alén. Chronic traumatic encephalopathy: The unknown disease. Neurology. 2014;32(3):185–91. 45. VanItallie TB. Traumatic brain injury (TBI) in collision sports: Possible mechanisms of transformation into chronic traumatic encephalopathy (CTE). Metabolism. 2019;100. 46. Vile AR AL. Chronic Traumatic Encephalopathy: The cellular sequela to repetitive brain injury. J Clin Neurosci. 2017;41:24–9. 47. Kenneth Perrine, Jacqueline Helcer, Apostolos John Tsiouris, David J. Pisapia, Philip Stieg. The Current Status of Research on Chronic Traumatic Encephalopathy. World Neurosurgery. 2017;102:533–44. 48. Smith DH, Johnson VE, Trojanowski JQ, Stewart W. Chronic traumatic encephalopathy - confusion and controversies. Nat Rev Neurol. 2019;15(3):179–83. 49. Burley C. Suicide as a clinical feature of chronic traumatic encephalopathy: What is the evidence? Aggression and Violent Behavior. 2020;54. 50. Gaetz M. The multi-factorial origins of Chronic Traumatic Encephalopathy (CTE) symptomology in post-career athletes: The athlete post-career adjustment (AP-CA) model. Medical Hypotheses. 2017;102:130–43. 51. Safinia C, Bershad EM, Clark HB, SantaCruz K, Alakbarova N, Suarez JI, et al. Chronic traumatic encephalopathy in athletes involved with high-impact sports. J Vasc Interv Neurol. 2016;9(2):34–48. 52. Finkel AM, Brand KP, Caplan AL, Evans JS, Wolpe PR. First report the findings: genuine balance when reporting CTE. Lancet Neurology. 2019;18(6):521–2. 53. Iacono D, Shively SB, Edlow BL, Perl DP. Chronic traumatic encephalopathy: Known causes, unknown effects. Phys Med Rehabil Clin N Am. 2017;28(2):301–21. 54. Azad TD, Li A, Pendharkar AV, Veeravagu A, Grant GA. Junior Seau: An Illustrative Case of Chronic Traumatic Encephalopathy and Update on Chronic Sports-Related Head Injury. World Neurosurgery. 2016;86:515.d11-515. e16.

Chronic Traumatic Encephalopathy: A Literature Review

55. Wilson L, Stewart W, Dams-O’Connor K, Diaz-Arrastia R, Horton L, Menon DK, Polinder S. The chronic and evolving neurological consequences of traumatic brain injury. Lancet Neurology. 2017;16(10):813–25. 56. Raji CA, Merrill DA, Barrio JR, Omalu B, Small GW. Progressive Focal Gray Matter Volume Loss in a Former High School Football Player: A Possible Magnetic Resonance Imaging Volumetric Signature for Chronic Traumatic Encephalopathy. American Journal of Geriatric Psychiatry. 2016;24(10):784–90. 57. Dickstein DL, Pullman MY, Fernandez C, Short JA, Kostakoglu L, Knesaurek K, et al. Cerebral [18 F]T807/ AV1451 retention pattern in clinically probable CTE resembles pathognomonic distribution of CTE tauopathy. Transl Psychiatry. 2016;6(9):e900. 58. Mez J, Daneshvar DH, Kiernan PT, Abdolmohammadi B, Alvarez VE, Huber BR, et al. Clinicopathological evaluation of chronic traumatic encephalopathy in players of American football. JAMA. 2017;318(4):360–70. 59. Fesharaki-Zadeh A. Chronic Traumatic Encephalopathy: A brief overview. Front Neurol. 2019;10:713. 60. Han K, Chapman SB, Krawczyk DC. Neuroplasticity of cognitive control networks following cognitive training for chronic traumatic brain injury. NeuroImage Clin. 2018;18:262–78. 61. Crasta JE, Sibel J, Slomine BS, Mahone EM, Mostofsky SH, Suskauer SJ. Subtle motor signs in children with chronic traumatic brain injury. Am J Phys Med Rehabil. 2019;98(9):737–44. 62. Puvenna V, Engeler M, Banjara M, Brennan C, Schreiber P, Dadas A, Bahrami A, Solanki J, Bandyopadhyay A, Morris JK, Bernick C, Ghosh C, Rapp E, Bazarian JJ, Janigro D. Is phosphorylated tau unique to chronic traumatic encephalopathy? Phosphorylated tau in epileptic brain and chronic traumatic encephalopathy. Brain Research. 2015;1630:225–40.

67. Iverson GL, Gardner AJ, Shultz SR, Solomon GS, McCrory P, Zafonte R, et al. Chronic traumatic encephalopathy neuropathology might not be inexorably progressive or unique to repetitive neurotrauma. Oxford Academic. 2019;142(12):3672–93. 68. Woerman AL, Aoyagi A, Patel S, Kazmi SA, Lobach I, Grinberg LT, McKee AC, Seeley WW, Olson SH, Prusiner SB. Tau prions from Alzheimer’s disease and chronic traumatic encephalopathy patients propagate in cultured cells. Proc Natl Acad Sci USA. 2021;113(50):E8187–96. 69. Ban VS, Madden CJ, Bailes JE, Hunt Batjer H, Lonser RR. The science and questions surrounding chronic traumatic encephalopathy. Neurosurg Focus. 2016;40(4):E15. 70. Cieplak AS. Tau inclusions in Alzheimer’s, chronic traumatic encephalopathy and Pick’s disease. A speculation on how differences in backbone polarization underlie divergent pathways of tau aggregation. Front Neurosci. 2019;13:488. 71. Roukerd MR, Stump A, Hawryluk GWJ. Understanding the Cause of Chronic Traumatic Encephalopathy and Related Dementias. World Neurosurgery. 2020;144:192–3. 72. Turner RC, Lucke-Wold BP, Robson MJ, Lee JM, Bailes JE. Alzheimer’s disease and chronic traumatic encephalopathy: Distinct but possibly overlapping disease entities. Brain Inj. 2016;30(11):1279–92. 73. Cheng H, Deaton LM, Qiu M, Ha S, Pacoma R, Lao, et al. Tau overexpression exacerbates neuropathology after repeated mild head impacts in male mice. Neurobiology of Disease. 2019;134. 74. Okamura Y, Kawakami I, Watanabe K, Oshima K, Niizato K, Ikeda K, et al. Tau progression in single severe frontal traumatic brain injury in human brains. Journal of the Neurological Sciences. 2019;407. 75. Hsu ET, Gangolli M, Su S, Holleran L, Stein TD, Alvarez VE, McKee AC, Schmidt RE, Brody DL. Astrocytic degeneration in chronic traumatic encephalopathy. Acta Neuropathol. 2021;136(6):955–72.

63. Lee EB, Kinch K, Johnson VE, Trojanowski JQ, Smith DH, Stewart W. Chronic traumatic encephalopathy is a common co-morbidity, but less frequent primary dementia in former soccer and rugby players. Acta Neuropathol. 2019;138(3):389–99.

76. Zhang W, Tarutani A, Newell KL, Murzin AG, Matsubara T, Falcon B, Vidal R, Garringer HJ, Shi Y, Ikeuchi T, Murayama S, Ghetti B, Hasegawa M, Goedert M, Scheres SHW. Novel tau filament fold in corticobasal degeneration. Nature. 2021;580(7802):283–7.

64. Arena JD, Smith DH, Lee EB, Gibbons GS, Irwin DJ, Robinson JL, Lee VM, Trojanowski JQ, Stewart W, Johnson VE. Tau immunophenotypes in chronic traumatic encephalopathy recapitulate those of ageing and Alzheimer’s disease. Brain. 2021;143(5):1572–87.

77. Armstrong RA, McKee AC, Stein TD, Alvarez VE, Cairns NJ. A quantitative study of tau pathology in 11 cases of chronic traumatic encephalopathy. Neuropathol Appl Neurobiol. 2017;43(2):154–66.

65. Lesman-Segev OH, La Joie R, Stephens ML, Sonni I, Tsai R, Bourakova V, et al. Tau PET and multimodal brain imaging in patients at risk for chronic traumatic encephalopathy. NeuroImage: Clinical. 2019;24. 66. Iverson GL, Luoto TM, Castellani RJ. Authors’ Reply: Age-Related Tau Aggregates Resemble Chronic Traumatic Encephalopathy Neuropathologic Change. Journal of Neuropathology. 2020;79(8):924–8.

78. Armstrong RA, McKee AC, Stein TD, Alvarez VE, Cairns NJ. Cortical degeneration in chronic traumatic encephalopathy and Alzheimer’s disease neuropathologic change. Neurol Sci. 2019;40(3):529–33. 79. Mufson EJ, He B, Ginsberg SD, Carper BA, Bieler GS, Crawford F, Alvarez VE, Huber BR, Stein TD, McKee AC, Perez SE. Gene Profiling of Nucleus Basalis Tau Containing Neurons in Chronic Traumatic Encephalopathy: A Chronic Effects of Neurotrauma Consortium Study. J Neurotrauma. 2021;35(11):1260–71.

Chronic Traumatic Encephalopathy: A Literature Review

80. LoBue C, Schaffert J, Cullum CM, Peters ME, Didehbani N, Hart J, White CL. Clinical and neuropsychological profile of patients with dementia and chronic traumatic encephalopathy. J Neurol Neurosurg Psychiatry. 2021;91(6):586–92.

94. Juengst SB, Nabasny A, Terhorst L. Neurobehavioral symptoms in community-dwelling adults with and without chronic Traumatic Brain Injury: Differences by age, gender, education, and health condition. Front Neurol. 2019;10:1210.

81. Briggs ID, Angoa-Perez M, Kuhn DM. Prolonged Repetitive Head Trauma Induces a Singular Chronic Traumatic Encephalopathy–Like Pathology in White Matter Despite Transient Behavioral Abnormalities. American Journal of Pathology. 2016;186(11).

95. Iverson GL, Gardner AJ. Risk for misdiagnosing chronic traumatic encephalopathy in men with anger control problems. Front Neurol. 2020;11:739.

82. Collins-Praino LE, Corrigan F. Does neuroinflammation drive the relationship between tau hyperphosphorylation and dementia development following traumatic brain injury? Brain, Behavior, and Immunity. 2016;60. 83. Kriegel J, Papadopoulos Z, McKee AC. Chronic traumatic encephalopathy: Is latency in symptom onset explained by tau propagation? Cold Spring Harb Perspect Med. 2018;8(2):a024059.

96. Cantu R, Budson A. Management of chronic traumatic encephalopathy. Expert Rev Neurother. 2019;19(10):1015– 23. 97. Alosco ML, Mez J, Tripodis Y, Kiernan PT, Abdolmohammadi B, Murphy L, et al. Age of first exposure to tackle football and chronic traumatic encephalopathy: Age of First Exposure to Football and CTE. Ann Neurol. 2018;83(5):886–901.

84. Cherry JD, Mez J, Crary JF, Tripodis Y, Alvarez VE, Mahar I, et al. Variation in TMEM106B in chronic traumatic encephalopathy. Acta Neuropathol Commun. 2018;6(1):115.

98. Standring OJ, Friedberg J, Tripodis Y, Chua AS, Cherry JD, Alvarez VE, et al. Contact sport participation and chronic traumatic encephalopathy are associated with altered severity and distribution of cerebral amyloid angiopathy. Acta Neuropathol. 2019;138(3):401–13.

85. Cherry JD, Zeineddin A, Dammer EB, Webster JA, Duong D, Seyfried NT, et al. Characterization of detergent insoluble proteome in chronic traumatic encephalopathy. J Neuropathol Exp Neurol. 2018;77(1):40–9.

99. Mez J, Daneshvar DH, Abdolmohammadi B, Chua AS, Alosco ML, Kiernan PT, et al. Duration of American football play and chronic traumatic encephalopathy. Ann Neurol. 2020;87(1):116–31.

86. Doherty CP, O’Keefe E, Wallace E, Loftus T, Keaney J, Kealy J, et al. Blood-brain barrier dysfunction as a hallmark pathology in chronic traumatic encephalopathy. J Neuropathol Exp Neurol. 2016;75(7):656–62.

100. Alosco ML, Stein TD, Tripodis Y, Chua AS, Kowall NW, Huber BR, et al. Association of white matter rarefaction, arteriolosclerosis, and tau with dementia in chronic Traumatic Encephalopathy. JAMA Neurol. 2019;76(11):1298–308.

87. Walt GS, Burris HM, Brady CB, Spencer KR, Alvarez VE, Huber BR, et al. Chronic traumatic encephalopathy within an amyotrophic lateral sclerosis brain bank cohort. J Neuropathol Exp Neurol. 2018;77(12):1091–100. 88. Chen ST, Siddarth P, Merrill DA, Martinez J, Emerson ND, Liu J, et al. FDDNP-PET tau brain protein binding patterns in military personnel with suspected chronic traumatic Encephalopathy1. J Alzheimers Dis. 2018;65(1):79–88. 89. Ercoli LM, Siddarth P, Kepe V, Miller KJ, Huang S-C, Cole GM, et al. Differential FDDNP PET patterns in nondemented middle-aged and older adults. Am J Geriatr Psychiatry. 2009;17(5):397–406. 90. Omalu B, Small GW, Bailes J, Ercoli LM, Merrill DA, Wong K-P, et al. Postmortem autopsy-confirmation of antemortem [F-18]FDDNP-PET scans in a football player with chronic traumatic encephalopathy. Neurosurgery. 2018;82(2):237–46. 91. Biegon A. Considering biological sex in traumatic brain injury. Front Neurol. 2021;12:576366. 92. Ojo JO, Mouzon BC, Crawford F. Repetitive head trauma, chronic traumatic encephalopathy and tau: Challenges in translating from mice to men. Exp Neurol. 2016;275 Pt 3:389–404. 93. Asken BM, Sullan MJ, Snyder AR, Houck ZM, Bryant VE, Hizel LP, et al. Factors influencing clinical correlates of chronic traumatic encephalopathy (CTE): A review. Neuropsychol Rev. 2016;26(4):340–63.

Scholarly Research In Progress • Vol. 5, November 2021

A Review of the Literature: How Intestinal Microbiota Shape the Immune System and the Related Impact on Human Health and Autoimmune Disease Shane P. Bross1† Geisinger Commonwealth School of Medicine, Scranton, PA 18509 Doctor of Medicine Program Correspondence: sbross@som.geisinger.edu

1 †

Abstract The symbiotic relationship between gut microbiota and surrounding tissue significantly influences the functioning of our immune system. This review examines the current understanding of how gut microbiota contribute to shaping the immune system by influencing immune system related cells in the surrounding environment. Additionally, it examines how gut microbiota can impact human health by contributing to both intestinal and non-intestinal autoimmune diseases. The microbiota-host interaction is explored by emphasizing the evolutionary and ecological mechanisms which established their relationship; further, how gut microbiota shape the immune system and contribute to autoimmune disease generation via dysbiosis and generation of metabolic byproducts is discussed. To summarize the current understanding of the research topic we searched the PubMed database using combinations of relevant key words (e.g., “microbiome”, “immune system”, “autoimmune”), time limiters of 5 and 10 years, and ancestry technique. A comprehensive understanding of gut microbiota and their impact on the immune system elucidates disease process, mechanism, and associated symptoms. In conclusion, this understanding translates to practicing improved investigative technique, which limits waste and maximizes utility of resources. Judicious experimental design and execution fosters production of safer, more efficacious medical therapies and treatments for relevant autoimmune diseases and symptom management.

Introduction There are roughly 7.5 billion people on Earth, and each one of us is unique in one way or another. When one thinks of the word unique as it applies to people, one often thinks of a person’s hair color, eye color, personality, DNA, fingerprints, and the like. However, every person is also unique in the way that they house a different population of nearly 100 trillion microorganisms in their gut that make up their individual gut microbiomes (1). These microbes occupy niches within the human host gut, presenting potential pathogens with a survival challenge: They make it difficult for the pathogen to survive and thrive in the already occupied environment (1). As such, microbial community structure is considered a factor that can influence disease presentation, occurrence, and severity depending on the host. Perhaps the most focused area of study between the human host, its microscopic occupants, and disease is how individual microbiota populations play a role in shaping the immune system.

The effects that intestinal microbiota have on the immune system paves the path that leads to discussion of how these bacteria have the potential to elicit autoimmune diseases. Inflammatory bowel disease (IBD) is a result of autoimmunogenic mechanism. Crohn’s disease and ulcerative colitis are the two primary forms of IBD; both are relapsing disorders characterized by chronic inflammation in regions of the intestinal tract (2). IBD is found globally, with the highest reported prevalence in Europe (Norway and Germany) and North America (USA and Canada). Additionally, since 1990, the incidence rates for Crohn’s disease and ulcerative colitis have been rising in newly industrialized countries in Africa, Asia, and South America (3). Furthermore, gut microbiota can influence devastating non-intestinal autoimmune diseases. One of the most profound examples is multiple sclerosis. Multiple sclerosis is the most frequently seen demyelinating disease in which the immune system attacks the protective coverings of nerve cells. It has a prevalence rate of nearly 100 per 100,000 people in North America and Europe (4). In addition to addressing the connection between gut microbiota and intestinal related autoimmune diseases (IRADs), this review will explore the less-investigated connection between gut microbiota and their potential impact on nonIRADs. Exploring these relationships allows for better understanding of the potential mechanisms that contribute to disease presentation and severity. Utilizing this knowledge, the following review will also discuss potential therapies and areas of study that could be further investigated to aid in improved control and management of multiple autoimmune diseases. By first taking an ecological and evolutionary look into the relationship between human and microbe interaction, one can explore how diversity of these bacterial populations came to be by interacting with the host. Building off this foundational framework, this review investigates how this interaction shapes the immune system of the host, contributing to autoimmune disease manifestation. The ultimate purpose of this literature review is to help gain insight and understanding on how gut bacterial populations shape the immune system and contribute to disease by elucidating connections between human host and microbe relationships.

Methods A comprehensive literature review was conducted. This extensive literature search was carried out by utilizing the PubMed database, allowing for the review and collection of relevant subject content. The content analyzed includes

How Intestinal Microbiota Shape the Immune System

foundational knowledge of the ecological and evolutionary forces that shape microbial diversity and host interaction, epidemiological data and background information concerning IRADs and non-IRADs, interactions between gut microbiome and host immunity, and experimental findings that support how gut microbiota influence IRADs and non-IRADs. The specific search techniques used to locate each information source are listed below, in respective order. A literature search for information pertaining to microbial diversity in humans was conducted using the keywords ecological, evolution, microbial, and diversity in the PubMed database. This initial search produced 733 articles. To narrow the search, the search limiter “humans” was used as a species subtopic selection. A second search was conducted with the same limiter; the key words “human” and “disease” were added to the search, producing 42 articles. After examining titles and abstracts of the articles produced from these searches, two articles were selected to be used in this literature review based on their relevance to the topic. A search for epidemiological literature discussing IRADs was conducted using PubMed database and entering combinations of the keywords “worldwide,” “incidence,” “prevalence,” “inflammatory,” “bowel,” and “disease”. Publication date of 5 years was used as a limiter to increase accuracy and relevance. This search produced 83 articles; the abstracts of multiple articles were examined based off title, with two articles being selected for use in this literature review. To examine epidemiology of non IRADs, the keywords “epidemiology,” “multiple,” and “sclerosis” were used as search terms in the PubMed database, again using publication date of 5 years as a limiter. This search produced 2,367 articles, ultimately leading to the selection of one article being used for this literature review based off applicability via article abstract examination. A literature search focused on targeting the influential effects that the gut microbiome has on shaping immune response was conducted using multiple combinations of the keywords “gut,” “microbiome,” “bacteria,” “shape,” “immune,” and “system” in the PubMed database. Two articles were selected for use in this literature review after confirming their relevancy to the topic by examining their abstracts. An additional search was carried out, yielding 554 articles. A search limiter of 10 years was selected for publication date to narrow the search based on recent findings, which produced 537 articles. One article was selected from this search after examining the abstract and confirming its relevance to the review topic. An extensive literature search was conducted to highlight experimental results that may serve as additional evidence to support the contention that gut microbiota play a role in influencing the development of IRADs and non-IRADs via their impact on host immune system modification. Searches of the PubMed database were conducted using a combination of terms and keywords. The terms used included “intestinal microbiota,” “gut homeostasis,” “human disease,” “autoimmune,” and “immune system.” Key words such as “metabolites” and “imbalance” were added to the searches to determine whether the scopes of certain studies were applicable. The search term “intestinal microbiota” produced 15,462 articles; the addition of “autoimmune diseases” narrowed the results down to 500 articles. Using a search limiter of 5 years further narrowed

down the results to 426 articles, of which one was chosen based off title and abstract. Searching the term “intestinal microbiota metabolites” with 5 years as a search limiter yielded 1,111 articles; adding the term “gut homeostasis” narrowed results to 121 articles, leading to the selection of one article to be used in the literature review. Searching the keywords “intestinal,” “microbiota,” “immune,” and “system” produced 2,883 articles. Investigating titles and abstracts for relevancy resulted in the selection of two articles for use in the literature review. A final search of the key terms “microbial,” “community,” and “imbalances” in the PubMed database produced 32 articles; after examining the title of each article for relevancy and investigating the abstracts, one article was chosen to be used in this literature review. Lastly, to better understand the topic and provide additional support, the ancestry technique was used and proved to be very beneficial. Using the ancestry technique, 17 articles were examined based on relevance and were eventually chosen to be included in this literature review.

Discussion Evolutionary and ecological perspective: the microbehost relationship Since the beginning of time, the coevolution of man and microbe has resulted in a relationship between the two that is as intimate as it is complex. The coevolution between microbial communities and their hosts, enforced by natural selection operating at multiple levels, resulted in diverse populations of microbes that are specific to an individual. Dethlefsen et al. provide ample support for these statements in their discussion about the evolution of mutualism and the potential impact a changing microbial environment might have on the occupying symbionts. It is understood that natural selection favors traits that contribute to the fitness of a species. However, in more recent times, there has been an increase in the number of studies that analyzed the evolution of traits that benefit others in addition to the trait bearer. No-cost mutualism is the result of an instance when an organism’s traits, which contribute to the fitness of that organism, coincidentally also prove to be beneficial to the members of another species. Such a case describes one’s relationship with their gut microbiota: these intestinal symbionts, as Dethlefsen et al. describe, are selected upon due to their effectiveness as consumers through direct effects on their fitness. However, this selection also benefits the host in the form of protection against pathogens. The niche occupied by these selected microbiota serves as competition to potential pathogens, effectively creating a barrier that prevents pathogens from prospering and colonizing (5). If these mutualistic interactions are possible, it would stand to reason that those traits which improve — or at a minimum, stabilize — this mutualistic relationship might evolve in one or both symbionts. Additionally, a species might benefit in increasing its own fitness by subsequently increasing the fitness of its mutualistic associate. In either case, the selection to increase mutualistic benefits between partners can be strengthened with the interaction of alike lineages of partners that spans multiple generations (5). Thus, the evolution of

How Intestinal Microbiota Shape the Immune System

mutualism can support why humans have such a prominent population of gut microbiota. It also supports how and why our gut microbiota have stayed with us as humans as a species evolved.

Figure 1. Diversity and distribution of bacterial phylotypes in female human. The distribution of bacterial phylotypes is broad and extensive, with the colon housing the greatest number of distinctive bacteria. These data are not all encompassing and provide a rough estimation of phylotype diversity. Data used for figure construction was obtained from reference (5).

Evolution is one way in which to look at the relationship between host and microbe. Evolution requires selection of traits, which are passed on via genes. As it is understood that genes can be affected by environment, so too can a host’s gut microbiota. This environmental impact introduces the utility of an ecological perspective to examine the relationship between host microbiota populations and their respective environments inside the human body. When a human infant is born, they acquire their microbiota from the environment. As the first weeks of life turn into months, multiple species of fauna and flora colonize, diversify, and begin to occupy their specific microbial habitats within the infant body. With time, the ingestion of solid foods facilitates the creation of an individualized microbiome like that of an adult (5). Thus, the bacteria that compose an individualized microbiome are diverse and broadly distributed throughout the human body (Figure 1).

Figure 2. Gut microbial metabolites and effect on host immune responses. Two major avenues eliciting host immune response via microbial metabolites involve short chain fatty acids (SCFAs) and quorum sensing (QS) signaling molecules. This image has been reproduced using original source image from reference (7). CSF: Competence and sporulation factor; G– : Gram-negative bacteria; G+ : Gram-positive bacteria; IECs: Intestinal epithelial cells; IkK: Kinase complex; MAPK: Mitogen-activated protein kinase; NF-kB: Nuclear factor-kappa beta complex.

How Intestinal Microbiota Shape the Immune System

Microbiota can undergo very drastic changes in response to a change in their environment. For example, a change in the environment of a non-pathogenic microbe can result in host tissue invasion. An immune response is then mounted to eliminate the infection, resulting in the mass death of numerous microbial occupants. These immune defenses introduce selective pressure. To combat death by the immune system in response to a change in environment, the microbe needs to adapt in such a way as to avoid detection from the host immune system. In this case, the microorganism begins adapting toward selection that favors its conversion to a pathogen to conserve fitness (5). These changes highlight the extreme sensitivity and response between the host microbiome and their environment with the human body. Appreciating the importance of evolutionary mechanisms and human ecology regarding the host-microbiota relationship is influential in understand why a disruption in either of these areas could contribute to ailments. An ecological or genetic change that disconnects the human-microbe relationship has the potential to result in serious diseased states. A closer look: gut microbiota metabolites and dysbiosis As introduced, the community structure of gut bacteria can be intimately related with proper immune system function. Important components of modulation and shaping of the immune system by gut microbiota involve microbial byproducts and dysbiosis, or an unnatural shift in the composition of microbiota. The metabolic byproducts of certain bacterial species can serve to prevent inflammatory autoimmune disease such as IBD, specifically Crohn’s disease and ulcerative colitis. These conditions are classically characterized by increases in pro-inflammatory cytokines, particularly tumor necrosis factor-alpha (TNF-α) and interferon-gamma. More recently, a population of inflammatory T helper lymphocytes has been noted as a factor that could contribute to pathogenesis of human and experimental colitis. These T helper lymphocytes are characterized by the expression of pro inflammatory cytokine interleukin-17 (IL-17) (6). The immunomodulatory effects of products from intestinal bacteria metabolism increases colonic IL-17. Additionally, other types of bacterial-released metabolites, including intermediates or end products of commensal gut bacteria dietary metabolism, are proposed to function in shaping immunity. Lin et al. thoroughly discusses two such metabolites: short chain fatty acid (SCFA) and quorum sensing (QS) signal molecules. There are multiple proposed mechanisms by which these metabolites exert their immunologic effects (Figure 2). The SCFA metabolites include propionate and butyrate. The latter is well known for its anti-inflammatory activities, in which accumulating evidence has shown that butyrate can diminish bacterial translocation across epithelia under stress, as well as improve the gut barrier by reinforcing intestinal epithelial cells. Both SCFAs have demonstrated binding capability to specific G-protein coupled receptors that mediate signal transduction involved in certain inflammatory pathways. Thus, these SCFAs serve to repress inflammation by binding these receptors (7). Quorum sensing—reliant on cell density—is a regulatory mechanism that bacteria use to promote synchronized behavior between and among the bacterial populations. The QS signaling

molecules allow for smooth operation of this process, which is utilized by pathogens to facilitate colonization and invasion into a host (7). Thus, it is proposed that QS signals may act as an important weaponry fallback to combat pathogenic encroachment via multiple immunomodulation pathways controlling anti-inflammatory processes, chemotaxis, and inflammatory gene expression. These findings suggest that a dysbiotic event leading to a shift in the proportions of certain bacteria would result in a change in the number and/or frequency of metabolites produced. Theoretically, this change in metabolite production due to dysbiosis could then lead to an immunocompromised state. Alternatively, discussion of how these metabolites hold potential to effect intestinal epithelia cells warrants deeper investigation into their effects on the gut mucosa and mucosal immune system. The degree of development of the mucosal immune system, and the lymphoid structures it contains, differ depending on the absence or presence of commensal bacteria in the gut. Within the mucosa of the small intestine lie Peyer’s patches. These lymphoid tissues are sites in which B and T lymphocytes are activated (8). They are integral members of the mucosal immune system. Histological comparisons of mouse model tissues show evidence that the mucosal immune system is underdeveloped in germ-free (GF) mice: the formation of Peyer’s patches is greatly affected (Figure 3). Additionally, comparisons show a greatly reduced number of immunoglobulin A (IgA)- producing plasma cells and a reduction in CD4+ T lymphocytes in the GF mice. This is a significant observation; IgA functions in protecting mucosal surfaces against pathogenic and non-pathogenic microorganisms (9). However, these immunologic abnormalities are not solely limited to the mucosal immune system. Splenic tissue in association with the systemic immune system also shows considerable changes upon comparison between the two mouse models. The tissue sample from the GF mouse shows poorly formed T and B lymphocytes zones. These observations provide a more concrete understanding of how dysbiosis in commensal microbiota can shape the immune system. Genetic blemishes and regulatory anomalies Furthering the discussion about bacterial impact on the mucosal immune system is the relationship between bacteria, genetics, and mucosal function. Activation of immune and inflammatory responses can be initiated by stimulating the luminal flora and/ or their products. It has also been suggested that genetically determined variations in key mucosal functions, including cell activation by bacterial pattern molecules, leads to differing immune function and susceptibility to IBDs (10). Pattern recognition receptors (PRRs) play a pivotal role in mediating host immune processes by initiating inflammatory responses upon recognition of a microbial molecule during infection. However, an issue exists in this detection-altering system: PRRs are not discriminatory between ligands produced from pathogenic bacteria and ligands produced from a resident microbiota upon colonization (11). Accordingly, and perhaps helping to clarify the histological abnormalities discussed above, recent studies reveal that these gut bacterial ligands from resident microbiota signal through PRRs to promote host tissue and immune development, thereby facilitating protection from disease (11). There are additional findings

How Intestinal Microbiota Shape the Immune System

an environment). Moreover, the transfer of the gut microbiota from the affected mice into wild-type recipients induced the disease, providing additional support for this assertion. These results spurred investigation into how the microbiota have effects on the functioning and/or production of other regulatory mechanisms associated with immune response and function. Using a mouse model, Ishikawa et al. provide results that help clarify thoughts of how intestinal microbiota effect the production of cells that function to contribute to immune response, thus shaping the immune system. Oral tolerance is a local and systemic immune unresponsiveness to orally ingested antigens (14). It plays a critical role in immune defense, specifically in the Figure 3. Impact of intestinal bacterial colonization on lymphoid structures in intestinal and gastrointestinal mucosa. Failure of oral systemic tissues from mouse model. Histological comparison of splenic and intestinal tissue sections taken from germ-free wild-type mouse and a mouse of the same strain colonized tolerance induction leads to susceptibility with intestinal bacteria; a: Germ-free mice have poorly formed T cell (pink) and B cell zones in to indigenous bacterial antigens in the splenic tissue; b: Low number of CD4+ cells (brown) from lamina propria are seen in the germgastrointestinal tract, potentially leading free mouse tissue; c: Intestinal tissue from the germ-free mouse demonstrates a conspicuous to IBD. It is reported that oral tolerance decrease in the number of IgA producing cells (brown) compared to colonized mouse. This can be maintained by regulatory T (Treg) image was reproduced using its original source from reference (8). IgA: Immunoglobulin A. lymphocytes by facilitating the characteristic hypo-responsiveness to the orally ingested antigen. Specifically, CD25+ CD4+ Treg that affirm the relevance of how gut microbiota shape the lymphocytes have been associated with immune system by exploiting host genetic defects. One such this phenomenon due to their production of the inflammatory theory contends that Crohn’s disease and ulcerative colitis suppressive cytokine, transforming growth factor-beta (15). are the result of continuous microbial antigenic stimulation Comparing specific pathogen free (SPF) mice to GF mice, of pathogenic immune responses because of host genetic Ishikawa et. al. observed that experimental GF mice seemed to mutations in mucosal barrier function, innate bacterial killing, be resistant to oral tolerance. It was concluded that both the or immunoregulation (12). Furthermore, polymorphisms in frequency and total number of CD25+ CD4+ Treg lymphocytes the host genome are said to likely interact with functional in the GF mice were significantly lower than those in the SPF bacterial changes — perhaps utilizing PRR signaling — to mice. Furthermore, cells from the mesenteric lymph node — the stimulate aggressive immune responses leading to chronic reported site where CD25+ CD4+ Treg are generated — in SPF tissue injury (12). Thus, even when gut microbiota are mice suppressed proliferation of certain T lymphocytes that themselves functioning as they should in an environment devoid produced pro-inflammatory cytokines in comparison of those of dysbiosis, they can still have devastating impacts on host from GF mice (15). Thus, the study provides evidence that the immune function by interacting with the genetic shortcomings microbiota play an important role in generating fully functional that affect pattern recognition, mucosal barrier, killing of innate CD25+ CD4+ Treg lymphocytes. The absence or incomplete bacteria, and perhaps least surprisingly, genetic imperfections functioning of these regulatory cells, due in part to the absence associated with immunoregulation. of a microbial population, has potential to lead to increased susceptibility and altered immune response against indigenous Having an established genetic program in the immune system intestinal bacterial antigens, contributing to inflammatory unquestionably aids in proper immune function. To reiterate autoimmune diseases such as IBD. how an immune system piloted by an ill-functioning genetic program can produce disease driven by gut microbiota, Garrett et. al. examined how communicable ulcerative colitis was induced by an immune cell-specific T-box transcription factor (Tbet) deficiency. Tbet orchestrates inflammatory genetic programs in both adaptive and innate immunity by regulating TNF-α, which drives tissue injury in experimental mouse model colitis. Tbet also fosters harmony between eukaryotes and prokaryotes by affecting the behavior of the colonic epithelium (13). Their study demonstrated that mice deficient of Tbet within the innate immune system developed spontaneous ulcerative colitis, inferring the possibility that episodes of spontaneous ulcerative colitis can be significantly influenced by a compromised immune system because of a malfunctioning colonic environment (and the bacteria that are affected by such 78

Resident gut-microbiota and non-intestinal related autoimmune disease Much of this review thus far has revolved around IRADs (Crohn’s disease and ulcerative colitis), which can be facilitated by an altered immune system due to gut microbiota. However, to more fully appreciate the far-reaching impact that gut commensals have on the immune system, it is necessary to investigate how they can affect systemic immunity. Information previously referenced demonstrates how gut microbiota hold potential to contribute to non-intestinal autoimmune diseases, specifically Hashimoto’s thyroiditis, Type 1 diabetes mellitus (T1D), and multiple sclerosis (16).

How Intestinal Microbiota Shape the Immune System

Hashimoto’s thyroiditis is characteristically identified by the infiltration of mononuclear cells in the thyroid along with production of autoantibodies against thyroglobulin and thyroid peroxidase (17). It has been documented that the transfer of microbiota from conventional rats into SPF rats increased the susceptibility of the SPF rats to experimental autoimmune thyroiditis (18). This finding provides evidence that supports the influence of the microbiota on the pathogenesis of Hashimoto’s thyroiditis. Furthermore, to reiterate the effects of dysbiosis and epithelial tissue barriers in the gut, it has been demonstrated that dysbiosis within the microbial environment can manifest as a condition known as “leaky gut” (19). From a histological perspective, leaky gut is characterized by changes in the epithelial cells of the gut, as well as lymphocyte infiltration. Notably, similar observations have been made upon examination of tissues from patients with Hashimoto’s disease (20). T1D is one of the best known and most researched non IRADs. It is characterized by CD4+ and CD8+ T lymphocyteguided destruction, targeting the insulin-storing and releasing beta cells of the pancreas (21). While there is a hereditary component associated with the pathogenesis of T1D, environmental factors — such as cesarean or vaginal birth — can also contribute to the induction of the disease (22). As mentioned, the resident microbiota population is established by the host environment after birth. Results from Endesfelder et al. provided evidence that the induction of T1D can be attributed by intestinal microbiota by their priming of the immune system at an early postnatal period (23). Additionally, a study of Italian children affected by T1D showed an increase in their intestinal permeability that was correlated with modifications in their resident gut microbiota population. Moreover, the same study found three very highly represented microbial biomarkers in the affected children when compared to healthy control children (24). As a final point, and perhaps most striking, is the observation that the entire Bacteroidaceae family of bacteria, as well as two other prominent species, were significantly disproportionately represented in children with T1D (Table 1) (25). Lastly, multiple sclerosis is the most frequently seen demyelinating disease in which the immune system attacks the protective coverings of nerve cells. Like the findings in children with T1D, it was found that multiple sclerosis patients could have a specific type of microbiome due in part to dysbiosis. Studies found abnormally low proportions of some gut specific species while other species were present in abnormally high amounts (26, 27). Importantly, samples of microbiome from patients affected by multiple sclerosis impaired the ability of certain T lymphocytes to properly differentiate into Treg cells (28). As evidenced, this is considerably impactful to the immune system since the proper functioning of these cells are critical in moderating effective and controlled immune responses. In relation to Tregs and the prior discussion of the effect of certain microbial metabolites on immune function, important evidence was derived using an autoimmune encephalomyelitis (AE) model. This model, mediated by T lymphocytes, is an animal model used to study multiple sclerosis due in part to its ability to reproduce most of the symptoms observed in people who have multiple sclerosis (29).

Interestingly, orally administered propionate (a SCFA which was one of the bacterial metabolic byproducts discussed earlier) decreased AE clinical scores and increased Tregs at lymph nodes (30). These findings are particularly promising because they demonstrate once more how metabolic products from gut bacteria metabolism could potentially contribute to therapies for multiple sclerosis and other non-intestinal autoimmune diseases.

Conclusion There have been major strides in research that have focused on identifying specific species of gut microbiota and how these species have evolved to reside in their respective environments within the host. However, clinical practice might benefit from research that focuses more heavily on examining the habitats of the human body and how these environments change over time and vary depending on an individual’s diet and genetic predispositions. This could improve care by allowing health care providers to better anticipate the consequences that a changing

Table 1. Changes observed in microbiota for associated autoimmune diseases. Significant changes in bacterial colonization were observed in multiple non-IRADs. This table was reproduced using material in original source from reference (16). IRADs; Intestinal related autoimmune diseases.

How Intestinal Microbiota Shape the Immune System

environment might have on our microbiota, and the associated impact on host immunity and severity of disease pathogenesis. By investigating these symbiotic “landscapes” with research focused on examining the ecology of human health, one gains better insight into the relationship between humans and their acquired microorganism populations and the diseases that present when this relationship goes awry. There have also been major advances in related research from a more macroscopic level that focuses on microbial communities and the relationship they have with their host environment, as well as their impact on mucosal and systemic immunity. As discussed, these communities can exploit host genetic defects to alter immune function and contribute to disease. By conducting research that focuses on identifying the polymorphisms that lead to host-microbial alterations, researchers and care providers can work in conjunction to use selective targeted interventions that correct the identified polymorphisms underlying these abnormalities. It is hopeful that these methods would provide improved therapies for treatment of autoimmune diseases via precision medicine. This domain of research requires disciplined and precise methods of examining and obtaining samples of gut microbiota and surrounding tissue. Thus, it is important to consider the technical and ethical limitations that exist when obtaining samples from humans to be used in research. The use and relevance of human studies are clear. However, obtaining samples to be used can be laborious, time consuming, and costly due to the limitations listed above. Focused experimental model systems could be used to obtain preliminary results and findings, guiding researchers towards practicality of methods in humans. Experimental model systems, such as the mice and rats mentioned above (which have experimentally induced and specific autoimmune diseases or malfunction), can fill this role by conserving evolutionary features that are likely to be crucial for function. Model systems would also allow for improved understanding of multiple microbial mechanisms that lead to the same immunologic outcome without compromising precious human samples. Additionally, it could give insight into the evolutionary relationship of host-microbe interaction by stratifying experimentation based on species complexity. Abiding by the ethical regulations set forth to protect animals and their utility in biomedical research is critically essential and must be vigilantly monitored. Nonetheless, results could provide a clearer understanding of microbial interactions with the immune system that could then be explored using human tissue in a more focused, precise, and efficacious manner.

References 1.

Ley RE, Peterson DA, Gordon JI. Ecological and evolutionary forces shaping microbial diversity in the human intestine. Cell. 2006;124(4):837-48.

Frank DN, St Amand AL, Feldman RA, Boedeker EC, Harpaz N, Pace NR. Molecular-phylogenetic characterization of microbial community imbalances in human inflammatory bowel diseases. Proc Natl Acad Sci USA. 2007;104(34):13780-5.

Ng SC, Shi HY, Hamidi N, Underwood FE, Tang W, Benchimol EI, et al. Worldwide incidence and prevalence of inflammatory bowel disease in the 21st century: a systemic review of population-based studies. Lancet. 2018;390(10114):2769-78.

Leray E, Moreau T, Fromont A, Edan G. Epidemiology of multiple sclerosis. Revue neurologique. 2016;172(1):3-13.

Dethlefsen L, McFall-Ngai M, Relman DA. An ecological and evolutionary perspective on human-microbe mutualism and disease. Nature. 2007;449(7164):811-8.

Round JL, Mazmanian SK. The gut microbiome shapes intestinal immune responses during health and disease. Nature reviews Immunology. 2009;9(5):313-23.

Lin L, Zhang J. Role of intestinal microbiota and metabolites on gut homeostasis and human diseases. BMC immunology. 2017;18(1):2.

Macpherson AJ, Harris NL. Interactions between commensal intestinal bacteria and the immune system. Nature reviews Immunology. 2004;4(6):478-85.

Macpherson AJ, Hunziker L, McCoy K, Lamarre A. IgA responses in the intestinal mucosa against pathogenic and non-pathogenic microorganisms. Microbes and infection. 2001;3(12):1021-35.

10. Podolsky DK. The current future understanding of inflammatory bowel disease. Best practice & research clinical gastroenterology. 2002;16(6):933-43. 11. Chu H, Mazmanian SK. Innate immune recognition of the microbiota promotes host-microbial symbiosis. Nature immunology. 2013;14(7):668-75. 12. Sartor RB. Microbial influences in inflammatory bowel diseases. Gastroenterology. 2008;134(2):577-94.

Acknowledgments

13. Garrett WS, Lord GM, Punit S, Lugo-Villarino G, Mazmanian SK, Ito S, et al. Communicable ulcerative colitis induced by T-bet deficiency in the innate immune system. Cell. 2007;131(1):33-45.

I would like to thank Jennifer Boardman, PhD, for her advice and guidance provided throughout the writing process.

14. Commins SP. Mechanisms of oral tolerance. Pediatric clinics of North America. 2015;62(6):1523-9.

Disclosures

15. Ishikawa H, Tanaka K, Maeda Y, Aiba Y, Hata A, Tsuji NM, et al. Effect of intestinal microbiota on the induction of regulatory CD25+ CD4+ T cells. Clinical experimental immunology. 2008;153(1):127-35.

The author has no financial affiliations to disclose from sponsors or organizations/associations.

16. Opazo MC, Ortega-Rocha EM, Coronado-Arrazola I, Bonifaz LC, Boudin H, Neunlist M, et al. Intestinal microbiota influences non-intestinal related autoimmune diseases. Frontiers in microbiology. 2018;9:432.

How Intestinal Microbiota Shape the Immune System

17. Antonelli A, Ferrari SM, Corrado A, Di Domenicantonio A, Fallahi P. Autoimmune thyroid disorders. Autoimmune reviews. 2015;14(2):174-80. 18. Penhale WJ, Young PR. The influence of the normal microbial flora on the susceptibility of rats to experimental autoimmune thyroiditis. Clinical and experimental immunology. 1988;72(2):288-92. 19. Vaarala O, Atkinson MA, Neu J. The “perfect storm” for type 1 diabetes: the complex interplay between intestinal microbiota, gut permeability, and mucosal immunity. Diabetes. 2008;57(10):2555-62. 20. Sasso FC, Carbonara O, Torella R, Mezzogiorno A, Esposito V, deMagistris L, et al. Ultrastructural changes in enterocytes in subjects with Hashimoto’s thyroiditis. Gut. 2004;53(12):1878-80. 21. Atkinson MA, Eisenbarth GS, Michels AW. Type 1 diabetes. Lancet. 2014;383(9911):69-82. 22. Rewers M, Ludvigsson J. Environmental risk factors for type 1 diabetes. Lancet. 2016;387(10035):2340-8. 23. Endesfelder D, Engel M, Zu Castell W. Gut immunity and type 1 diabetes: a mélange of microbes, diet, and host interactions? Current diabetes reports. 2016;16(7):60. 24. Maffeis C, Martina A, Corradi M, Quarella S, Nori N, Torriani S, et al. Association between intestinal permeability and faecal microbiota composition in Italian children with beta cell autoimmunity at risk for type 1 diabetes. Diabetes/ metabolism research and reviews. 2016;32(7):700-9. 25. De Goffau MC, Luopajarvi K, Knip M, Ilonen J, Ruohtula T, Harkonen T, et al. Fecal microbiota composition differs between children with beta-cell autoimmunity and those without. Diabetes. 2013;62(4):1238-44. 26. Jangi S, Gandhi R, Cox LM, Li N, von Glehn F, Yan R, et al. Alterations of the human gut microbiome in multiple sclerosis. Nature communications. 2016;7:12015. 27. Freedman SN, Shahi SK, Mangalam AK. The “gut feeling”: breaking down the role of gut microbiome in multiple sclerosis. Neurotherapeutics. The journal of the American Society for Experimental Neurotherapeutics. 2018;15(1):109-25. 28. Cekanaviciute E, Yoo BB, Runia TF, Debelius JW, Singh S, Nelson CA, et al. Gut bacteria from multiple sclerosis patients modulate human T cells and exacerbate symptoms in mouse models. Proc Natl Acad Sci USA. 2017;114(40):10713-8. 29. Stromnes IM, Goverman JM. Active induction of experimental allergic encephalomyelitis. Nature protocols. 2006;1(4):1810-9. 30. Mizuno M, Noto D, Kaga N, Chiba A, Miyake S. The dual role of short acid chains in the pathogenesis of autoimmune disease models. PLoS One. 2017;12(2):e0173032.

Scholarly Research In Progress • Vol. 5, November 2021

Exploring the Impact of Medicaid Expansion on Colorectal Cancer (CRC) with a Focus on Individuals Below the Standard Screening Age in the United States Sandybell J. Anorga1†‡, Mukta C. Bhatnagar1†‡, Chantel V. Golding1†‡, Brian M. Grodecki1†‡, Eric M. Watiri1†‡, Brian J. Piper¹, and Elizabeth C. Kuschinski¹ ¹Geisinger Commonwealth School of Medicine, Scranton, PA 18509 † Master of Biomedical Sciences program ‡ Authors contributed equally Correspondence: mbhat398@outlook.com

Abstract In the United States, the second most common cause of cancer death is colorectal cancer (CRC). While it is imperative that there are viable insurance options to get preventive tests, many states have opted not to participate in the Medicaid expansion, thus adding to health care discrepancies and disparities among CRC patients. The typical U.S. CRC screening age is 50 years old, but due to a rising incidence rate of CRC in younger individuals this standard is now being challenged. We used the Surveillance, Epidemiology, and End Results (SEER) program research data (1975-2017) from the National Cancer Institute (NCI) to conduct a secondary data analysis on CRC participants from a cancer registry based on geography, health insurance access, age, race, and stage of diagnosis. The resulting data analysis for 13 U.S. states (12 Medicaid expansion and one non-expansion) did not yield any association between race and CRC incidence across the provided age groupings. There was, however, a statistically significant association between age group and stage of diagnosis. The highest CRC mortality rates were also found in the southeastern U.S., where the largest proportion of non-expanded states are located. These findings demonstrate a need to lower the CRC screening age to cater to the increase in younger individuals developing this disease, as well as expand the benefits stemming from Medicaid expansion, both of which would ultimately reduce potential CRC disparities and poor health outcomes.

Introduction Colorectal cancer (CRC) remains the second leading cause of cancer-related death in the United States due to rising cases in adults under the age 50 (1). The negative correlation between CRC screening and mortality rates in individuals over the age of 50 years has increased; however, the same cannot be said for those below the screening age (1). Some individuals may fall within a coverage gap if they earn above their state’s eligibility for Medicaid, but below the minimum income required to afford private insurance. This jeopardizes their well-being because they are left without a viable option for obtaining necessary preventive health care. Medicaid expansion may provide a safety net for these lowincome individuals, but only in the states that have opted for its implementation. In order to better understand the impact of Medicaid expansion on CRC, this study examined the incidence and mortality rates with a special focus on individuals below the standard screening age.

Despite positive outcomes from screening and diagnostic tests, the U.S. continues to face socioeconomic disparities with CRC incidences. From an economic standpoint, financial barriers have continued to hinder individuals from receiving the necessary CRC screenings (2). It is hypothesized that the unfavorable diagnoses and prognoses of the CRC patients facing these current disparities will be further exacerbated in non-expanded Medicaid states compared to expanded states. Most CRC patients tend to suffer from the disease for quite some time before they are diagnosed due to the “delayed” requirements for CRC screenings starting at ages 50 years and older (3). Individuals below 50 were not originally thought to be at substantial risk for CRC; however, there has been a rise in CRC cases among younger adults once thought to be less prone to the disease (3). Although the incidence of CRC is decreasing for the standard age group, the incidence in those younger than 50 has been increasing by 2.0% annually for the last 9 years (4). These findings make it imperative that younger adults below the screening age of 45 to 50 have a viable insurance option in order to receive these preventive measures, especially if they are deemed to be at a higher risk (5). Moreover, for those young adults with insurance and access to primary care physicians, there are still issues with the recognition of this disease. Of the 1,195 CRC patients and survivors ages 20 to 49 surveyed by researchers in 2018 (mostly in the U.S.), 57% were diagnosed between ages 40 and 49, 33% were diagnosed between 30 and 39, and 10% were diagnosed before age 30 (6). Thus, not only is this younger population more likely to be uninsured and less likely to be screened, but their symptoms are more frequently overlooked by their physicians and themselves. It is no longer safe to assume that only those with a family history of cancer are at risk, as a new group with early onset CRC (EOCRCs) and no family history of the disease has recently emerged (3). These are all patients who were diagnosed under the current recommended screening age of 50 (3). The current diagnostic guidelines were established based on familial cancer and are typically insufficient for early onset diagnosis (3). The importance of this CRC-focused research is reflected in its novelty of investigating the impact of Medicaid expansion on the CRC incidences among the younger individuals as opposed to the typically studied older individuals. If individuals develop CRC at a younger age but lack viable options for diagnosis and treatment, then this will result in increased incidence and mortality rates over time (7). This study aimed to provide

Exploring the Impact of Medicaid Expansion on Colorectal Cancer

vital information about incidences and diagnoses, as well as background analysis on at-risk individuals by examining the collected CRC data from Medicaid expanded and nonexpanded states.

Methods Procedures Surveillance, Epidemiology, and End Results (SEER) is a cancer registry database including information acquired by the North American Association of Central Cancer Registries’ (NAACCR) Data Standards on patient demographics (age, sex, and race), cancer characteristics, stage of disease, treatment, and outcomes (8,9). This study was a secondary data analysis using the SEER cancer registry data — evaluating states from each region throughout the country — and addressed any CRCrelated discrepancies in the region based on several variables. The categorical variables included within this analysis were health insurance access, race, and sex, and the continuous variable was age of diagnosis. These variables were analyzed in order to describe any associations between them, as well as to isolate certain descriptive characteristics (e.g., age range). All the data in this paper was obtained from electronic records (SEER) and stored on password-protected electronic devices. The medium for data storage was a Microsoft Excel file with exclusive access to current members of the research group. No physical data with sensitive or identifiable information was used for this research. This systematic investigation utilized a broad range of measured demographic, functional, and health variables. It sought to interpret and inform generalizable knowledge, with protection of vulnerable groups, minimized risk to all participants, and it did not include any sensitive information. This study was reviewed by Geisinger’s Institutional Review Board and determined not to be human subjects research under the federal Common Rule, 45 CFR Part 46.102(d). The SEER program is used by the National Cancer Institute (NCI), which is a source of cancer-related statistical information (9). Specifically, the information relates to cancer incidence rates and survival data/mortality rates based upon cancer registries accounting for approximately 35% of the U.S. population (10). These reported cases are then reviewed to determine if the provided information should be stored in the cancer registry respective to each state (8). If so, the cancer registries then obtained the necessary cancer-related information from the patient’s medical records (8). A cancer registry can be either population-based or hospitalbased; the SEER program utilizes the population-based registry model for collecting the necessary cancer statistics (10). The overall design of this type of registry is meant to gather information to monitor the distribution of cancer among various demographic factors which provide a basis that can be used for future efforts in research and determining effective use of health resources currently available in cancer control efforts (10). The requested information from the SEER 1975–2017 Research Data was utilized when conducting the secondary data analysis (11).

Participants The SEER program data for this study contains information collected from cancer registries that have analyzed participants’ clinical data, demographic, and mortality rates associated with CRC, as well as their health insurance status (12). This primary data source provides a succinct population-based resource which was utilized in this study to analyze the impact of Medicaid on CRC screening and diagnosis. The SEER program contains CRC incidences of a population categorized by age, race, sex, and geographic location. This secondary research data considered the following variables: age of diagnosis, access to health insurance, race, and sex in various states. This sample size included participants from all races and sexes, within the specified age range of 20 to 64 years old. Specifically, the SEER program has age groups listed in 5-year increments starting at age 0 to 85+; however, the age range for this study started at 20 years and ended inclusively at 64 years of age. The selected lower limit of 20 corresponded to the earliest adult age group reflected in the primary source data. An upper age limit of 64 was selected because individuals age 65 and older are covered under Medicare and thus fell outside the scope of this study. Despite these exclusions, the sample size was large enough to allow for effective statistical analysis. All 13 of the states included in the SEER database for CRC incidence rates were included in this analysis. The SEER Registry makes it voluntary for individual states to include their cancer information in the database, hence only 13 states have data available for public use; all 13 states were included for this analysis to maximize the available data (13). Together, the chosen states represent the different regions of the U.S. (i.e., the Northeast, Mid-Atlantic, Southeast, Midwest, Southwest, Northwest, and West). Hawaii and Alaska were also included in this analysis, although their data in the SEER database was very limited. Statistical analysis We performed descriptive analyses including age range, count (e.g., sample size), and percent (e.g., CRC incidence sex and race breakdown within each state). We also conducted an inferential analysis, calculating a p-value from each chi-square test, and weighted the samples for all analyses used via SEER*Stat Software (14). The samples were weighted in order to reduce survey bias and adjust for unequal selection probabilities (15). Multiple chi-square tests were implemented on the data: the first included age of CRC diagnosis versus access to health care, to determine if access to health care had any relation to age of diagnosis. More specifically, it sought to ascertain if living in a Medicaid expansion state was associated with a difference in diagnosis age. The second chi-square test compared age of diagnosis versus race, to ascertain if there was any relationship between race demographic and age of CRC diagnosis. A third chi-square test evaluated the stage at which the CRC was diagnosed (localized or regional) versus the specified age ranges. The 5-year age ranges provided by SEER were collapsed into 15-year increments, and 2x2 chi-square tests were utilized to investigate any associations. The chi-square tests compared the 20–34 age group versus the 50–64 age group, as well

Exploring the Impact of Medicaid Expansion on Colorectal Cancer

as the 35–49 age group versus the 50–64 age group. No association was pursued between the 20–34 and 35–49 age groups, since these both fell below the common threshold for CRC screening in most states. This data analysis used the chisquare test in particular to evaluate statistical independence or association between two categorical variables within each state, whereupon it compared the findings from each state. The “age” variable was actually an age range, which is considered a categorical variable and appropriate for this type of test. The results of the chi-square tests are presented as a figure, in the form of a distribution curve. For the p-values obtained, values above 0.05 show that the data is statistically non-significant (although the indicated association may still be noteworthy) (16). Heat maps were used to delineate CRC adjusted mortality data by state for ages 20 to 64 from both 2010–2013 and 2015–2018. The heat map software used was provided by Microsoft Excel. These analyses could provide evidence supporting lowering the initial screening age for CRC in the U.S.

Results Mortality rates Due to states expanding at different times, four years pre- and post-initial Medicaid expansion were observed to provide a balanced range to account for varying mortality rates among the states. In 2010, the mortality rates ranged from 10 to 12 deaths per 100,000 in the population in the Northeast, Mid-Atlantic and Southeast region. The western regions’ mortality rates were predominantly less than 10 with a few exceptions, including Colorado and Nevada. In the Southeast region, Mississippi had the highest mortality rate, with 13.63 deaths per 100,000 in the population (Figure 1). The mortality rates in 2011 closely paralleled those in the previous year, with Mississippi remaining the state with the highest mortality rate of 14.01 deaths per 100,000 (Figure 2). In 2012, Arkansas and Mississippi had the highest mortality rates in the country at 14.73 and 14.71 per 100,000 in the population, respectively. A majority of the country had mortality rates below 10 or nearly 10. States with mortality rates equal to or greater than 10 were on the eastern side of the country (Figure 3). By contrast, in 2013 nearly all of the states in the country had a mortality rate of 6 to 12 per 100,000 in the population. Northeast, Mid-Atlantic and Southeast states had rates that were equal to or greater than 10 per 100,000. On the opposite side of the country, mortality rates in the western U.S. states were predominantly less than 10, with the exception of Nevada having a mortality of 10.54 per 100,000. Mississippi had the highest mortality rate, with 13.74 per 100,000 in the population (Figure 4). With Medicaid expansion taking place in 2014, there was an expected year or two delay before any notable trends could be exhibited throughout the country. In 2015, the mortality rates mirrored similarly to those in 2013, with the eastern half of the country having higher mortality rates than the western half. Nevada had the highest rate in the West, with 10.04 deaths per 100,000 in the population, and Mississippi had the highest mortality rate in the country, with 13.54 deaths per 100,000 in the population (Figure 5). Two years following the expansion, higher mortality rates were concentrated in the Southeast region, with Mississippi having the highest at 14.35 deaths

per 100,000 (Figure 6). A similar trend was found in 2017, with the Southeast having the greatest mortality rates of the entire country. Mississippi, Arkansas and Louisiana rates were 13.38, 13.53, and 13.39 deaths per 100,000 in the population, respectively. Nevada had the highest mortality rate in the West, with 10.43 deaths per 100,000 (Figure 7). In 2018, the mortality rates increased within the Southeast region and the Mid-Atlantic region. Mississippi had the highest, with 13.54 deaths per 100,000, and West Virginia had 13.48 deaths per 100,000 in the population (Figure 8). Chi-square tests In our comparison between race and the youngest age group versus the oldest age group, California, Iowa and Louisiana demonstrated a relationship between the two extreme age cohorts and race in terms of CRC incidence rates in the year 2014 (Table 1). During 2015, the combined incidence rates of Alaska and Hawaii exhibited a relationship between race and age cohorts. For our comparisons between race and the middle-age group versus the oldest age group, there was no relationship between them in terms of CRC incidence rates. Washington in 2015 was the only state that demonstrated a significant relationship between race and age influencing CRC incidence rates (Table 2). When the race and insurance of CRC-diagnosed individuals was examined, there was no relationship between the race of a patient and their insurance status for expanded states. However, in Georgia, the single non-expanded state, there was a relationship between race and insurance status for CRC diagnosis of patients (Table 3). In our comparison between age and the stage upon which CRC was detected, there was a suggested relationship between the youngest age group (20–34) experiencing statistically significant instances and advanced stages of CRC that had originally been exclusive to the older age group (Table 4). Additionally, we identified a significant relationship between the middle age group and more advanced stages of CRC at the time of diagnosis compared to the older age group (Table 4).

Discussion From observations made of the heat maps 4 years prior and post-Medicaid expansion, the Southeast region had the highest mortality rates before and after the expansion of the 7 regions. It must be noted that the majority of non-expanded states are located in this region, including Mississippi, Alabama, Florida, and Georgia. By contrast, the West, Mid-Atlantic, and Northeast regions of the country with predominantly expanded states have post-expansion mortality rates that are steady or slightly decreased from pre-expansion rates. This suggests that the adoption of Medicaid expansion may have a beneficial effect on the diagnosis and treatment of CRC. Compared to expanded states, the mortality rates of non-expanded states rose steadily (by 1 to 2 points) during the post-expansion years. However, there were some expanded states with mortality rates that continued to rise by a few points during the post-expansion years (2015–2018). This rise in mortality rates in expanded states called for acknowledging additional factors unique to each state having an impact on the overall CRC mortality rate of patients 20 to 64. Some of the possible reasons for this

Exploring the Impact of Medicaid Expansion on Colorectal Cancer

Figure 1. Age-adjusted per 100,000 CRC mortality rate for adults 20–64 in 2010. Darker colors reflect increased mortality.

Figure 2. Age-adjusted per 100,000 CRC mortality rate for adults 20–64 in 2011. Darker colors reflect increased mortality.

Figure 3. Age-adjusted per 100,000 CRC mortality rate for adults 20–64 in 2012. Darker colors reflect increased mortality.

Figure 4. Age-adjusted per 100,000 CRC mortality rate for adults 20–64 in 2013. Darker colors reflect increased mortality.

Figure 5. Age-adjusted per 100,000 CRC mortality rate for adults 20–64 in 2015. Darker colors reflect increased mortality.

Figure 6. Age-adjusted per 100,000 CRC mortality rate for adults 20–64 in 2016. Darker colors reflect increased mortality.

Figure 7. Age-adjusted per 100,000 CRC mortality rate for adults 20–64 in 2017. Darker colors reflect increased mortality.

Figure 8. Age-adjusted per 100,000 CRC mortality rate for adults 20–64 in 2018. Darker colors reflect increased mortality.

Exploring the Impact of Medicaid Expansion on Colorectal Cancer

Table 1. Calculated p-values from 2x2 chi-square tests between race (white and black) versus age ranges (20-34 and 50-64). Highlighted findings significant at the p<0.05 level. N/A: not applicable.

Table 2. Calculated p-values from 2x2 chi-square tests between race (white and black) versus age ranges (35-49 and 50-64). Highlighted findings significant at the p<0.05 level. N/A: not applicable.

with the combined values of Alaska and Hawaii (0.026). These instances conveyed a tentative association between race and CRC incidence for the younger versus older age bracket. For the middle versus old age group, the non-expansion state of Georgia consistently had p-values over 0.7 similar to the values of the aforementioned group. There were frequent values of 0.9 or higher throughout the data for the remaining Medicaid expanded states except for Washington’s p-value of 0.006 in 2015. The single instance that conveyed the association between race and incidence rate of the middle age group compared to the four instances of the younger age group were isolated exceptions of the calculations. Therefore, at best, the provisional characterization that can be made is that for most of the states in the used SEER database, race played little role in the CRC incidence rates.

The findings from analyses between race (white and black) and insurance status (insured versus uninsured) depicted that there was generally no conclusive relationship between these two factors in the Medicaid expansion states. California Table 3. Calculated p-values from 2x2 chi-square tests between race (white and black) and was the only expanded state with at least insurance status (insured versus uninsured). Highlighted findings significant at the p<0.05 two instances (0.031 and 0.039) where race level. N/A: not applicable. and insurance influenced the CRC incidence rate in 2012 and 2014, respectively. These instances possibly arose due to the largely diverse demographic of California residents alongside socioeconomic transitions between these two years. The single nonexpanded state of the cohort, Georgia, had three consecutive years (2012–2014) where race and insurance status did have an influence on CRC incidence rates for that state (0.001, 0.049, 0.006, respectively). Table 4. Calculated p-values from 2x2 chi-square tests between age cohorts (20-34, 35-49, The majority of expanded states not 50-64) and stage of CRC during screening (localized, regional, distant). Highlighted findings significant at the p<0.05 level. Any E-values with an integer greater than 10 were considered experiencing this relationship suggested not valid; therefore, not statistically significant. that Medicaid may reduce the possibility of race influencing CRC incidence based on insurance status. This coincided with age shift and increased mortality rates could be secondary to the true intention of the expansion to remove the possibility of environmental and lifestyle factors such as diet, exercise, and insurance status hindering or negatively impacting the health the rising prevalence of obesity among younger individuals (3). care individuals receive. However, states without the wideIn addition, the socioeconomic disparities already present in ranging public insurance option did not necessarily protect these locations further exacerbate the rising CRC mortality and prevent the disparities faced by those of different racial rates. groups and having access to health insurance. This in turn The results from the race versus age impact on CRC incidence would be anticipated to impact the possibility and availability trials were less informative, by virtue of inconsistent p-values for individuals without insurance to receive proper preventive and a concomitant lack of statistical significance from the health care services, thus increasing CRC incidence rates chi-square tests. For the younger versus older age group, most despite Medicaid expansion. of the expanded states and the single non-expanded state, The observed findings regarding the stage of CRC Georgia, had p-values that were statistically non-significant upon diagnosis by age range across the states generally over the 4-year span from 2012 to 2015. The exception to demonstrated that the middle age grouping (35–49) was at a this included three Medicaid expanded states in 2014 — higher risk for being diagnosed at a later stage of CRC. California (0.012), Iowa (0.01), and Louisiana (0.041) — along 86

Exploring the Impact of Medicaid Expansion on Colorectal Cancer

The years examined (2012–2017) showed a strong correlation between the middle age group and the older age group (50–64) and the stage of diagnosis. It is worth noting that a longer interval between confirmed CRC diagnosis to start of treatment has been associated with a significant increase in mortality rate across all cancer stages. Risk of death was 1.64 times more likely for those treated over 151 days after confirmed diagnosis compared to those treated within 30 days of diagnosis. Those under the age of 44 had a higher mortality rate (41.59%) than those in the 45–54 age cohort (36.79%) and the 55–64 age cohort (34.77%). Higher mortality rates in young patients are thought to be due in part to later diagnoses and longer treatment delays, both of which may be minimized with more age-appropriate diagnostics and a shift away from the perspective that CRC only affects older individuals (17). The main limitation for the study is that there was only one Medicaid non-expansion state available for CRC incidence in the SEER database (out of 21 total non-expansion states as of April 2015) (18). Some of the states included as part of this study (e.g., Alaska and Hawaii) had very minimal samples but were retained to refrain from discarding any of the limited data.

Conclusion Our findings suggest that many younger individuals have their CRC symptoms misdiagnosed or undiagnosed until the later stages of the disease, which are consequently more lethal. The direness of the situation could be further exacerbated for certain minority racial groups in having higher uninsured statuses, placing these individuals at higher risk for not getting proper preventive care and timely diagnosis for CRC. With this trend seen in Georgia, other non-expanded states located in the Southeast region may have CRC patients facing similar disparities. Further research is warranted to fully substantiate this data across the broader array of the U.S., but the results suggest that leveraging openings produced by Medicaid expansion to increase early screening and detection of CRC throughout these areas may reduce the recent uptick in mortality rates for younger individuals (birth–49) (4). Thus, there is an essential need for a more comprehensive analysis on the effects of Medicaid expansion and additional evidence affirming the benefits of lowering the CRC screening age in order to contemplate the need for vital, life-sustaining legislature.

Acknowledgments We would like to thank Brian Piper, PhD, and Elizabeth Kuschinski for their tireless guidance and feedback throughout this research process.

Disclosures The authors do not report any conflict of interest.

References 1.

Siegel RL, Miller KD, Goding Sauer A, Fedwea SA, Butterly LF, Anderson JC, et al. Colorectal cancer statistics, 2017. CA Cancer J. Clin. 2017 May 6 [cited 2021 Mar 1]; 67(3): 177-193. Available from: https://doi.org/10.3322/ caac.21395

Lee C, Eldridge HE, Reed EM, Lee KJ, Kushi HL, Spiegelman D. The impact of Affordable Care Act on colorectal cancer incidence and mortality: The case of Kaiser Permanente of northern California. MedRxiv. 2020 Oct 27 [cited 2021 Mar 1]; 1-12. Available from: https:// doi.org/10.1101/2020.10.26.20218503

Loomans-Kropp HA, Umar A. Increasing incidence of Colorectal Cancer in young adults. J. Cancer Epidemiol. 2019 Nov 11 [cited 2021 Mar 1]; 2019: 1-9. Available from: https://doi.org/10.1155/2019/9841295

Siegel LR, Miller KD, Goding Sauer A, Fedewa SA, Butterfly LF, Anderson CJ, et al. Colorectal cancer statistics, 2020. CA Cancer J. Clin. 2020 Mar 5 [cited 2021 Mar 1]; 70(3): 145-164. Available from: https://doi.org/10.3322/ caac.21601

McMorrow S, Kenney GM, Long SK, Anderson N. Uninsurance among young adults continue to decline, particularly in Medicaid expansion states. Health Affairs. 2015 Apr [cited 2021 Mar 1]; 34(4): 616-620. Available from: https://doi.org/10.1377/hlthaff.2015.0044

Yarden RI, Newcomer NI. Abstract 3347: Young onset colorectal cancer patients are diagnosed with advanced disease after multiple misdiagnoses. Clin. Cancer Res. 2019 Jul [cited 2021 Mar 1]; 79(13): 3347. Available from: https://doi.org/10.1158/1538-7445.AM2019-3347

Sung H, Siegel LR, Rosenberg SP, Jemal A. Emerging cancer trends among young adults in the USA: Analysis of a population-based cancer registry. Lancet Glob. Health. 2019 Feb 3 [cited 2021 Mar 1]; 4(3): E137-E147. Available from: https://doi.org/10.1016/S2468-2667(18)30267-6

What is a Cancer Registry? [Internet]. National Cancer Institute: NIH; 2020. Data Collection, Storage, & Management; 2020 [cited 2020 Nov 10]. Available from: https://seer.cancer.gov/registries/cancer_registry/data_ collection.html

Surveillance, Epidemiology, and End Results Program [Internet]. National Cancer Institute: NIH; 2020 [cited 2020 Nov 10]. Available from: https://seer.cancer.gov/

10. What is a Cancer Registry? [Internet]. National Cancer Institute: NIH; 2020. Cancer Registry; 2020 [cited 2020 Nov 10]. Available from: https://seer.cancer.gov/registries/ cancer_registry/cancer_registry.html 11. SEER Incidence Data, 1975-2017 [Internet]. National Cancer Institute: NIH; 2020 [cited 2020 Nov 10]. Available from: https://seer.cancer.gov/data/SEER 12. Medicaid Linked Data Resource [Internet]. National Cancer Institute: NIH; 2020 [cited 2020 Nov 10]. Available from: https://healthcaredelivery.cancer.gov/seer-medicaid/ 13. List of SEER Registries [Internet]. National Cancer Institute: NIH; 2020 [cited 2021 Mar 1]. Available from: https://seer.cancer.gov/registries/list.html 14. SEER*Stat Software [Internet]. National Cancer Institute: NIH; 2020 Sept 2 [cited 2020 Nov 10]. Available from: https://seer.cancer.gov/seerstat/

Exploring the Impact of Medicaid Expansion on Colorectal Cancer

15. Step 1: Calculating Age-adjusted Rates [Internet]. National Cancer Institute: NIH; 2020 [cited 2020 Nov 10]. Available from: https://seer.cancer.gov/seerstat/tutorials/aarates/ step1.html 16. Kuttappa S. A learning guide to IBM SPSS Statistics: Get the most out of your statistical analysis [Internet]. IBM; 2020 Jun 17 [cited 2021 Mar1]. Available from: https:// www.ibm.com/blogs/journey-to-ai/2020/06/a-learningguide-to-ibm-spss-statistics-get-the-most-out-of-yourstatistical-analysis/ 17. Lee YH, Kung PT, Wang YH, Kuo WY, Kao SL, Tsai WC. Effect of length of time from diagnosis to treatment on colorectal cancer survival: a population-based study. PLoS One. Jan 2019 [cited 2021 May 4]; 14(1): e0210465. Available from: https://doi.org/10.1371/journal. pone.0210465 18. Buettgens M, Holahan J, Recht H. Medicaid Expansion, Health Coverage, and Spending: An Update for the 21 States That Have Not Expanded Eligibility [Internet]. KFF; 2015 Apr 29 [cited 2021 Mar 1]. Available from: https:// www.kff.org/medicaid/issue-brief/medicaid-expansionhealth-coverage-and-spending-an-update-for-the-21states-that-have-not-expanded-eligibility/

Scholarly Research In Progress • Vol. 5, November 2021

Conflict of Interest Disclosure Accuracy Among Physician Authors of Cancer Research Journals Shuyi Chen1†, Alivia L. Roberts1*, Kevin Zhao1†, Abigail C. Burke1†, Jesse E. Ritter1†, Katherine M. Musto1†, and Brian J. Piper1 ¹Geisinger Commonwealth School of Medicine, Scranton, PA 18509 † Doctor of Medicine Program *Master of Biomedical Sciences Correspondence: schen01@som.geisinger.edu

Abstract Background: Clinical research is becoming increasingly reliant on industry sponsorship, with oncology physicians in particular demonstrating dependence on industrial payments. The introduction of a federally required reporting system, the Centers for Medicare & Medicaid Service’s (CMS) Open Payments Program, allows public access to financial exchanges between industry and clinical researchers. In this study, we investigated if a discrepancy between self-reported and CMSreported conflicts of interest (COI) existed among physician authors of cancer research journals in the United States. Methods: We explored three cancer research journals published from January 2017 to December 2018: Journal of Clinical Oncology, Cancer Research, and Nature Reviews Cancer. We collected a total of 525 published articles and stratified this data by selecting only MD or DO authors who practice in the U.S. to be further investigated. We inspected the COI statement of each article and then cross referenced the author to the CMS Open Payment Program “General Payments” section and determined the accuracy of COI disclosures. We also characterized the extent of this discrepancy among male and female physician authors with a chi-square analysis. Results: Sixty-seven authors met our inclusion criteria. Fortythree out of the 67 authors have received general payments; 65.2% of which correctly disclosed COIs while 34.8% authors failed to do so. The percentage of accurate disclosure for Journal of Clinical Oncology, Cancer Research, and Natural Reviews Cancer was 77.8%, 63.6% and 50.0%, respectively. In addition, female physician authors had a lower tendency to properly disclose COIs compared to their male counterparts (41.6% vs 74.2%, p = 0.0447) Conclusion: This study found that one-third of published cancer researchers did not accurately self-report their COIs. Thus, journals may need to reconsider their policies or include a link from CMS to improve transparency and maintain the public’s trust in oncology research.

Introduction Therapeutic discovery is an expensive process, and as clinical research becomes increasingly reliant on industry sponsorship (1), oncology in particular, emerges as a specialty with heavy ties to private sources of research funding, especially when compared to other specialties (2). Physicians in oncology specialties in 2014 were 1.09 to 1.75 times more likely than those in non-oncology specialties to receive industryrelated payments that could be considered potential financial conflicts of interest (COIs). However, it should be noted that

the presence of financial COIs is not indicative of biased or dishonest research. Recent publications have demonstrated the lack of significant correlations between the presence of financial COIs and research outcomes (3, 4). Yet, the body of literature is prevalent which raises concern to the contrary (5, 6), and should not be ignored in the characterization of COIs in the current landscape of cancer research. With the introduction of the Center for Medicare & Medicaid Service’s (CMS) Open Payments program in 2013, federally mandated reporting of private industry involvement in research made the subject of COIs in research both more public and more uniform than it was when COI disclosure was dictated by individual journal policy. Disclosure policy has been shown to be in need of significant standardization (7), making any current self-disclosure information obtained solely through articles themselves both inconsistent and incomplete. Open Payments, on the other hand, offers a free-access, centralized database of payment information, sortable for type, source, and nature of payment. The service is undoubtedly an invaluable tool in maintaining researcher accountability but suffers from serious issues of its own (8), making the need for an accurate self-disclosure protocol even more important to ensure neither source of information prevails with incomplete or inaccurate data. In addition to the benefit of having two comparable sources of disclosure information to maintain the accuracy of both, self-disclosed financial relationships in literature are still arguably the more accessible COI information available to readers, who may or may not choose, or know how to, look beyond the scope of the article to find potential COIs. In this study, we investigated whether discrepancies between self-reported and industry-reported conflicts of interest (COI) existed among physician authors (MD or DO) of cancer research journals. We also sought to characterize the extent of this discrepancy between male and female physician authors in order to better understand the current landscape of available disclosure transparency.

Methods Original research articles published in The Journal of Clinical Oncology, Cancer Research, and Nature Reviews Cancer from January 1, 2017, to December 31, 2018, were identified (n = 525). These general oncology journals were selected based on having high impact factors (33.0, 9.1, and 53.0 respectively) and having an appreciable portion of their authors being practicing physicians in the United States (U.S.). The last author for each article was evaluated and only articles with authors who are U.S.-based physicians (MD or DO) were selected for investigation. All articles written by authors who are non89

Conflict of Interest Disclosure Accuracy Among Physician Authors of Cancer Research Journals

U.S.-based physicians or who did not have a record in Open Payments were excluded. A total of 67 eligible authors were identified. Data collection for the 67 authors included sex, degree (MD or DO), and general payment information in CMS Open Payments. General payments include fees related to consulting, travel and lodging, food and beverage, services other than consulting, education, honoraria, royalties or licenses, speaking at an accredited/certified education program, and speaking at an unaccredited/non-certified education program. Payments classified as research payment, associated research funding, or ownership and investment were excluded. Selfreported financial disclosures included in the COI portion of the research articles were compared to data reported to the CMS Open Payment Program to determine the accuracy of COI disclosures among authors. A match between the two criteria was documented as proper COI disclosure whereas a mismatch between the two criteria was documented as failure to properly disclose COI. Discrepancies among male and female authors was investigated using chi-square analysis.

Results Based on the exclusion criteria, 67 last authors were eligible for the study. Among them, 43 authors (64.2%) received general payment according to CMS Open Payments. Together, these authors received $566,937.61. Of the 43 authors, 28 (65.2%) correctly disclosed the COI in the COI statement of their published articles, whereas 15 authors (34.8%) failed to disclose (Figure 1). In addition, we also examined the percentage of authors who correctly disclose the COI among Journal of Clinical Oncology, Cancer Research, and Nature Reviews Cancer. We found that the COI disclosure rate for the three journals respectively were 77.8%, 63.6% and 50.0%. The difference among the three journals was statistically non-significant (Figure 2).

Figure 2. COI disclosure rate of journals. In the 3 journals explored in this study: Journal of Clinical Oncology (JCO), Cancer Research (CR), Nature Reviews Cancer (NRC), with rates of disclosure of 77.8%, 63.6%, and 50.0% respectively (JCO vs. CR, P=.4084; JCO vs. NRC, P=.1008; CR vs. NRC, P=.4954).

Finally, we also investigated if there was a gender disparity in terms of accurately disclosing COI. Among the 43 eligible authors that received COI, 12 (27.9%) were female and 31 (62.1%) were male. 41.6% of the female authors versus 74.2% of the male authors correctly disclosed their COI (p < 0.05, Figure 3).

Figure 1. COI disclosure rate of last author. Among the 3 cancer research journals, 67 authors were eligible for the study as per our criteria, where 43 authors (64.2%) received a general payment from CMS Open Payments. Although 65.2% of these authors disclosed this information, 34.8% failed to do so. 90

Figure 3. COI disclosure rate between genders. 41.6% females and 74.2% males who had received payments were found to disclose COI correctly. The * indicates a statistically significant difference between male and female authors in disclosing COI via Chi-square analysis.

Conflict of Interest Disclosure Accuracy Among Physician Authors of Cancer Research Journals

Discussion

References

We found that more than a third of the authors failed to disclose COI information accurately. We did not find any significant differences in COI disclosure rates between the Journal of Clinical Oncology, Cancer Research and Nature Reviews Cancer, but did find that female physicians were less likely to accurately disclose potential financial COIs than their male counterparts. While it is not exactly clear why a disparity was observed between self-reported vs CMS-reported COI disclosures, variable journal disclosure policy (7) and non-standardized COI definitions (8) may certainly lend to the difficulties experienced by both author and journal in determining what financial relationships and payment types are relevant for self-disclosure. Furthermore, general payments not directly related to the studies in question may never be explicitly stated.

Buchkowsky SS, Jewesson PJ. Industry sponsorship and authorship of clinical trials over 20 years. Ann Pharmacother. 2004;38(4):579–85.

Marshall DC, Moy B, Jackson ME, Mackey TK, HattangadiGluth JA. Distribution and patterns of industry-related payments to oncologists in 2014. J Natl Cancer Inst [Internet]. 2016 [cited 2021 May 16];108(12). Available from: https://academic.oup.com/jnci/article/108/12/ djw163/2706926

Bariani GM, de Celis Ferrari ACR, Hoff PM, Krzyzanowska MK, Riechelmann RP. Self-reported conflicts of interest of authors, trial sponsorship, and the interpretation of editorials and related phase III trials in oncology. J Clin Oncol. 2013;31(18):2289–95.

Universal disclosure policy and the establishment of standardized COI language offers a solution to these, and other issues plaguing disclosure policy currently. It would alleviate the difficulties of deciding which financial information is relevant for physician authors, increase the accuracy and wealth of accessible reader information, and improve the security of all COI information available by providing a practical way to compare information between self- and industry-disclosed information for accuracy and completeness.

Miranda MC, Lera AT, Ueda A, Briones B, Lerner T, Del Giglio A, et al. The influence of self-reported conflicts of interest on the conclusions of editorial authors of phase III cancer trials. J Clin Oncol. 2011;29(15_suppl):6039–6039.

Friedberg M, Saffran B, Stinson TJ, Nelson W, Bennett CL. Evaluation of conflict of interest in economic analyses of new drugs used in oncology. JAMA. 1999;282(15):1453–7.

Jang S, Chae YK, Majhail NS. Financial conflicts of interest in economic analyses in oncology. Am J Clin Oncol. 2011;34(5):524–8.

Kesselheim AS, Lee JL, Avorn J, Servi A, Shrank WH, Choudhry NK. Conflict of interest in oncology publications: A survey of disclosure policies and statements. Cancer. 2012;118(1):188–95.

Hudis CA, Carlson RW. Real transparency in medicine: Time to act. Cancer. 2019;125(22):3924–6.

Conclusion We believe complete disclosure has the potential to improve the credibility of all works by authors with financial ties to industry, allowing readers to make more informed opinions about the information being shared in scientific literature. Further research with a larger sample of authors or journals is recommended to explore disparities in disclosure and outcomes of research to better understand how, if at all, the lack of self-disclosure in the real presence of potential financial COIs impacts the results of research. Additionally, to begin formulating universal policy and standard COI terminology, research aiming to characterize the profundity of variation within current disclosure policy and COI language may be of great benefit to the process as well.

Acknowledgements We would like to thank Darina L. Lazarova, PhD, for guidance and support throughout our research and Iris Johnston for her help in obtaining interlibrary loans for literature used in this project.

Disclosures BJP is part of an osteoarthritis research team supported by Pfizer and Eli Lilly. The rest of the authors have no financial disclosures.

Scholarly Research In Progress • Vol. 5, November 2021

Measuring the Change in Use of Generic Oxcarbazepine (OXC) Versus Brand Names for Medicaid Enrollees Throughout the United States in 2018-2019 Isra Amin1*‡, Rizelyn Benito1*‡, Daniela Velasquez1*‡, Megan Yeung1*‡, and Alyssa Trajan1* ¹Geisinger Commonwealth School of Medicine, Scranton, PA 18509 *Master of Biomedical Sciences Program ‡ Authors contributed equally Correspondence: myeung@som.geisinger.edu

Abstract Background: There is a contentious debate on the efficacy and cost between the use of generic versus brand name antiepileptic drugs (AEDs). Oxcarbazepine (OXC) is a common antiepileptic drug available as generic oxcarbazepine and as brand names Trileptal and Oxtellar XR, which Medicaid covers for qualifying applicants through a system called Commonwealth of Pennsylvania Access to Social Services (COMPASS). This study analyzed the nationwide 2018 and 2019 Medicaid drug utilization datasets to compare the usage of generic oxcarbazepine and its brand names Trileptal and Oxtellar XR for persons throughout the United States. Medicaid has restrictive administrative policies such as prior authorization (PA), preferred drug lists (PDLs), and step therapy (ST) which create socioeconomic disparities in public care by restricting Medicaid patients’ access to brand name alternatives. The hypothesis proposed that generic oxcarbazepine will be utilized significantly and constantly more in both years and therefore the change in generic and brand name use from 2018 to 2019 will not be significant. Methods: A heat map for each drug and year was created per 100,000 Medicaid enrollees. A pie chart for each year was made to compare the prescribed amount of generic medication, oxcarbazepine, to the brand-name medications, Trileptal and Oxtellar XR. A bar graph was created to show which states had the most oxcarbazepine use, and a two tailed paired t-test was used to establish the significance of our data. Results: The t-test revealed that there was not a significant change in use for generic oxcarbazepine (p = 0.513) and Oxtellar XR (p = 0.953). However, Trileptal (p = 0.025) had a significant change in use from 2018 to 2019. Conclusion: Our hypothesis was confirmed. The change in use of generic oxcarbazepine and Oxtellar XR between the years 2018 to 2019 was insignificant, but the change in use of Trileptal was significant. Prescriptions for years 2018 and 2019 were largely generic oxcarbazepine with 94% and above. Future plans include exploring socioeconomic implications between generic oxcarbazepine and brand name drugs on individuals with Medicaid as well as the relationship between the prescription prices of generic oxcarbazepine, Oxtellar XR, and Trileptal and how it varies per state based on preference.

Introduction There are approximately 3 million epileptic adults in the United States (U.S.) who use anti-epileptic drugs (AEDs). However, AED medications are also prescribed for diagnoses such as

bipolar I disorder, mania and others. As these drugs have biologically sensitive effects, a larger contentious discussion on the efficacy and access of generic and brand AEDs has ensued. This controversy affects patients who have Medicaid as some brand-name medications are not covered by Medicaid plans, leading patients to pay more for brand-name medications (1). Oxcarbazepine is a common anti-epileptic drug often prescribed to treat epilepsy and personality disorders and is available as generic OXC and as the brand-name alternatives Oxtellar XR and Trileptal. Both generic and brand-name AEDs vary slightly in chemical compounds due to different fillers that can lead to rare side effects in some patients (2). Table 1 denotes biochemical differences between generic OXC and its brand names. Although the drugs are meant to have similar clinical outcomes, patients and providers prefer to prescribe brandname medication compared to generic medication, as they believe it treats illnesses and diseases more effectively (3). The determination of generic drug efficacy of AEDs is important to establish, as it is central to maintaining ethical patient care and establishing pharmacological consistency in its effects. This requires AEDs, such as oxcarbazepine, which is often used in combination with other medicines, to be biologically effective, sensitive, and consistent with a patient’s history. Since patients react differently to various pharmacological agents, it is critical that generic and brand names be readily available. However, restrictive Medicaid policies like prior authorization that require providers of each state’s Medicaid agency in order to control and reduce Medicaid costs before prescribing a drug are enacted to hinder access for brand name drugs due to governmental costs (4). Additionally, Medicaid programs operate a preferred drug list (PDL) that lists generic drug brands and requires healthcare providers to prescribe generic brands before brand-name alternatives. This process is called step therapy (ST) where in order to obtain brand-name drugs, Medicaid patients would need to perform poorly on the generic options — patients have to either experience worsening conditions or have no effect with the drug prescribed (5). However, ST varies across states through a mechanism called “step edits” which are “fail-first” policies that insurance companies enact to ensure generic medications are tried first to curb costs. For example, according to the Epilepsy Foundation, Pennsylvania is the only state that is a quadruple step edit state for AED access. The continued implementation of step edits prompts contentious debate, as controlling Medicaid costs through ST results in long-term ineffective treatment and harmful effects. In particular, there exists a disparity between usage of brand name of generic OXC,

Measuring the Change in Use of Generic Oxcarbazepine (OXC) Versus Brand Names for Medicaid Enrollees

causing Medicaid patients who require two or more AEDs to incur higher health care costs in states with restricted AED access.

federal and state health coverage program that assists lowincome persons, disabled individuals, children, and the elderly with medical costs (6).

This study examines the novel topic of measuring change in use of generic OXC compared to brand name drugs for persons enrolled in Medicaid throughout the U.S. in 2018 to 2019. Due to the administrative restrictions of Medicaid coverage, we hypothesize generic OXC to be prescribed more frequently over the brand names and that the change in use for the analyzed drugs is not significant from 2018 to 2019.

Participants

Methods Data source We used the United States 2018 and 2019 Medicaid drug utilization datasets for generic oxcarbazepine and its brand names, Trileptal and Oxtellar XR, for patients. Medicaid is a joint

The proposed sample size consisted of patients that were enrolled in the United States Medicaid database in 2018 to 2019. The population type included patients from low-income families and the elderly. Patients must meet the required income eligibility for annual household income per household size. To meet eligibility criteria for our study sample, patients had to report use of either generic OXC, Trileptal, or Oxtellar XR for the Medicaid enrollment period of 2018 and 2019. Data collection Patients who require medical assistance must satisfy federal and state requirements regarding residency, immigration status, and documentation of citizenship. Eligibility also includes financial need requirements and considers disability-based circumstances. Applications may be submitted electronically through COMPASS, via telephone call, or in paper form to any of the insurance contractors. According to COMPASS, applications require patient information such as household income from jobs, housing and utility expenses, birth dates, social security, and proof of citizenship and property. Applications may also be sent to a contractor as referrals from the CAO (County Assistant Office) or from the FFM (Federally Facilitated Marketplace). A family will be given a maximum 15-day period to supply any missing information. However, missing information will not delay the application submission beyond the 15-day window.

Table 1. Generic OXC and Brand Name Drugs Fact Sheet. The table shows half-life (in hours), bioavailability, mechanism, and epidemiology for both generic OXC and brand-name drugs (23, 24, 25).

Medicaid obtains their primary source of statistical data from Medicaid Statistical Information System (MSIS), Medicaid Analytic extract (MAX) files, and the CMS64 reports. MSIS serves as a source of data which factors in payments, utilization, and characteristics based on individuals who have submitted statewide eligibility within Medicaid. The Statistical Enrollment Data System (SEDS) collects enrollment data by states via the forms CMS-64ES, CMS-64.21E, and CMS-21E. Previously referred to as State Medicaid Research Files (SMRFs), the Medicaid Analytic eXtract (MAX) data are person-level data files taken from MSIS data regarding eligibility for Medicaid, service utilization, and payments. Figure 1 notes this information flow. These data records are developed to support research and policy analysis initiatives for populations of low-income and for Medicaid.

Measuring the Change in Use of Generic Oxcarbazepine (OXC) Versus Brand Names for Medicaid Enrollees

was shown in Alabama (AL) with a value of 235.53. For data analysis purposes, the Washington, D.C., dataset was not used due to lack of data. Additionally, Vermont (VT) was removed for Oxtellar XR t-testing due to unavailable 2019 information.

Results

Figure 1. Conceptual model of data collection plan. Conceptual model depicts the relationships between integral administrative entities to produce Medicaid statistical databases. It includes patient application requirements and steps used to collect application information such as household income and other factors through FFM and CAO referrals. This model also details MSIS data collection factors such as payments and utilization. These factors get filtered into CMS and MAX which summarizes the Drug Utilization Data.

Data from the nationwide Medicaid "Drug Utilization 2018–2019” spreadsheet was examined to compare the use of generic and brand name OXC from Data.Medicaid.gov for low-income persons, disabled individuals, children, and the elderly. This study will utilize secondary data analysis, which will be completed on Excel. Data analysis The data from the Medicaid database for the years of 2018 and 2019 was utilized to compare the use of generic OXC and brand names in Medicaid patients nationwide. The data was downloaded from the Medicaid database and put into an Excel spreadsheet for further analysis, focusing on three specific drugs, generic OXC, Trileptal, and Oxtellar XR based on the year and number of prescriptions per 100,000 Medicaid enrollees. Heat maps were generated through summation of the number of prescriptions and dividing by the total number of state Medicaid enrollees per 100,000 for each state. Pie charts were formulated through summation of the number of prescriptions for each drug and divided by total value of the three drugs together. Bar graphs used the average value of the number of prescriptions for each drug per state. To analyze the trends of drug use nationwide, pie charts, bar graphs, and heat maps were formulated through Prism. The statistical test used to analyze the significance of change for each drug was a two tailed paired T-Test via Excel. The test used a significance level of α = 0.05. The categorical values for 2018 and 2019 that were inputted were the sums of the number of prescriptions for each state. In the Oxtellar XR 2018 dataset, North Dakota (ND) had one value of 3,573 prescriptions which is an extremely high outlier. This led to the 2018 Oxtellar XR heatmap not being representative of the data. To simplify data depiction, ND data was substituted with the next highest value of Oxtellar XR which

Figure 2A shows usage of OXC at 94.76%, Oxtellar XR with 2.70%, and Trileptal with 2.55% during the year of 2018. The total amount of prescriptions is 1.7435E+06. Figure 2B shows usage of OXC with 95.34%, Oxtellar XR with 2.87%, and Trileptal with 1.79% during the year of 2019. The total number of prescriptions is 1.76589E+06. Our findings showed that generic OXC was used at a significantly greater percentage than its brand names Oxtellar XR and Trileptal. This confirmed our hypothesis that patients would be using more generic OXC through Medicaid compared to brand names due to Medicaid’s usage of ST that required patients to use and fail generic drugs before prescribing brand-name drugs.

Figure 3 shows the average usage of generic oxcarbazepine and brand-name drugs per 100,000 Medicaid enrollees in each state from years 2018 to 2019. Texas prescribed more AEDs as a whole compared to other states in both years. Also, the usage of generic OXC in all states exceeds the amounts of brand name Trileptal and Oxtellar XR drugs. This confirmed our hypothesis that generic OXC would be more commonly used for patients under Medicaid. Figure 4 depicts heat maps for the usage of generic OXC and brand-name drugs per 100,000 Medicaid enrollees per state. Figure 4A shows that for the year 2018, oxcarbazepine depicted the highest number of prescriptions coming from Louisiana with 4,610.36, and West Virginia with 4,341.49. The lowest number of prescriptions is from Vermont, with 743.85. Figure 4B shows that the highest number of prescriptions for Oxtellar XR in 2018 comes from Texas with 224.08 and Alabama with 235.53. There are no prescriptions for Oxtellar XR in Arkansas, Minnesota, Alaska, New Hampshire, South Dakota, Vermont, and Wyoming for the year of 2018. Figure 4C shows that the highest number of prescriptions of Trileptal in 2018 comes from Wyoming with 357.89, North Dakota with 228.34, and South Dakota with 178.86. The lowest number of prescriptions for Trileptal in 2018 are in California, with 1.74 and New York with 3.19. Figure 4D shows the highest numbers of prescriptions of oxcarbazepine for the year 2019 come from Louisiana, with 5,234.34, West Virginia with 4,720.94, and Montana with 4,435.40. The lowest number of oxcarbazepine prescriptions in 2019 is from Vermont, with 735.80. Figure 4E shows that for the year 2019, the highest numbers of Oxtellar XR prescriptions come from Oklahoma with 444.45 and Texas with 220.39. There are no prescriptions in states Arkansas, New Hampshire, North Dakota, South Dakota, Vermont, and Wyoming for the year of 2019. Lastly, Figure 4F shows the highest numbers of prescriptions for Trileptal in 2019 come from North Carolina with 158.6, Texas with 130.83, and Maine with 111.85.

Measuring the Change in Use of Generic Oxcarbazepine (OXC) Versus Brand Names for Medicaid Enrollees

Figure 2. (A) Total number of prescriptions of OXC, Oxtellar XR, and Trileptal in 2018. (B) Total number of prescriptions of OXC, Oxtellar XR, and Trileptal in 2019.

Figure 3. (A) Average generic OXC and brand-name usage per 100,000 Medicaid enrollees in 2018. (B) Average generic OXC and brand-name usage per 100,000 Medicaid enrollees in 2019.

Measuring the Change in Use of Generic Oxcarbazepine (OXC) Versus Brand Names for Medicaid Enrollees

Figure 4. (A) Heat map of generic OXC usage per state among Medicaid patients in 2018 (dark blue: highest; light blue: lowest). (B) Heat map of Oxtellar usage per state among Medicaid patients in 2018 (dark blue: highest; light blue: lowest). (C) Heat map of Trileptal usage per state among Medicaid patients in 2018 (dark blue: highest; light blue: lowest). (D) Heat map of generic OXC usage per state among Medicaid patients in 2019 (dark blue: highest; light blue: lowest). (E) Heat map of Oxtellar usage per state among Medicaid patients in 2019 (dark blue: highest; light blue: lowest). (F) Heat map of Trileptal usage per state among Medicaid patients in 2019 (dark blue: highest; light blue: lowest).

Measuring the Change in Use of Generic Oxcarbazepine (OXC) Versus Brand Names for Medicaid Enrollees

There are no prescriptions in 2019 for Trileptal in the following states: Arkansas, Arizona, Kansas, New Hampshire, North Dakota, South Dakota, Vermont, and Wyoming. A two tail paired samples t-test was conducted to compare the sums of prescribed oxcarbazepine, Oxtellar XR, and Trileptal per each state to determine the significance of change in use of each drug from 2018 to 2019 nationwide. There was not a significant difference in the use of generic oxcarbazepine for 2018 (M = 33,041.2, SD = 33,151.4) and 2019 (M = 33,379.6, SD = 33,490.9); t(48) = 2.01, p = 0.513. There was not a significant difference in the use of Oxtellar XR for 2018 (M = 959.5, SD = 1,579.9) and 2019 (M = 955.3, SD = 1,534.2); t(47) = 2.01, p = 0.953. There was a significant difference in the use of Trileptal for 2018 (M = 888.5, SD = 1,393.6) and 2019 (M = 633, SD = 1,053.9); t(48) =2.01, p = 0.025. For Oxtellar XR statistical analysis, Vermont was excluded due to unavailable 2019 data.

Discussion The main objective of our study was to analyze the change in use of generic OXC versus its brand names from 2018 to 2019. Although OXC is a multipurpose drug with numerous benefits, the accessibility for it varies per state according to their respective Medicaid policies. The generic drug is prescribed significantly more than its brand name alternatives (Figure 2). In particular, generic OXC use increased from 2018 to 2019, while Oxtellar XR use remained stagnant and Trileptal use significantly decreased. Consistent with past literature, state Medicaid programs curb prescription costs via restrictive formularies such as PA, ST and PDLs to encourage generic drug use. To depict the national trends among states, geographical distribution of the data via heat mapping was used to determine the density of generic vs brand name OXC. As mentioned previously, Pennsylvania is the only state that is a quadruple state edit for AEDs. Implementing restrictive Medicaid policies is suggestive of mainly utilizing and/or having access to generic OXC before its brand names. Figures 3A and 3B were consistent in demonstrating that generic OXC was more likely to be used compared to its brand names secondary to the quadruple step edit process. The relationship between the use of Trileptal (Figure 4C and 4F) and Oxtellar XR (Figure 4B and E) versus the use of generic OXC (Figure 4A and 4D) also identified greater use in generic OXC than its brand names. Similarly, there were certain states such as Missouri, Kansas, and Louisiana that had very dense regions for the use of generic OXC. However, these states differed in their Medicaid AED access. Consequently, Kansas included AEDs on their PDLs whereas Louisiana excluded AEDs from their PDLs. Additionally, the heat maps (Figure 4B and 4E) for Alabama and Texas showed similar findings of significant use of Oxtellar XR in 2018 followed by a significant decrease in the use of Oxtellar XR the following year. While Texas and Alabama utilize PDLs without AEDs, it is important to note that Alabama limits brand name usage up to 4 times per month. Specifically in Texas, it remains unknown whether it covers classes of drugs excluded from its PDL. In contrast, Oklahoma was the only state that significantly increased Oxtellar XR use from 2018 to 2019. While its Medicaid includes AEDs on PDLs, it requires prior authorization (PA) for some brand name AEDS (26).

Based on the results, Trileptal was the only AED noted to have a significant change in use from 2018 to 2019. Both Wyoming and North Carolina were among the denser regions for Trileptal use in 2018 and 2019, respectively. While Wyoming Medicaid requires PA for brand names, it covers AEDS, despite their exclusion from its PDL. Additionally, North Carolina Medicaid does not have a PDL. Consequently, these states allowed for facilitated administrative access to Trileptal. While our reference data states that it is unknown whether Texas Medicaid covers drugs outside of PDLs, Texas was consistently depicted to be a denser region compared to other states for both Oxtellar XR and Trileptal use (26). Limitations were found while using the 2018 and 2019 dataset from Medicaid. The datasets were incomplete at the time we obtained the data, which was January 26, 2021. Therefore, many states either had no data entered or only had one value. The research that has been done mainly focuses on the difference of generic and brand name OXC; however, further studies could be implemented to explore this drug and its socioeconomic implications on individuals with Medicaid. The analysis of open access in 2018 and 2019 in Figure 2A and 2B showed that the use of Oxtellar XR is more than Trileptal. A potential future question to examine is whether there are differences in preference that exist between the two brandname medications by analyzing the accessibility of state Medicaid programs per state.

Acknowledgments We would like to thank Catherine Freeland, Elizabeth Kuchinski, and Brian J. Piper, PhD, for their support and feedback on this subject.

Disclosures There is no financial relationship between these paper’s authors and any institution mentioned herein.

References 1.

Al-Jazairi AS, Bhareth S, Eqtefan IS, Al-Suwayeh SA. (2008). Brand and generic medications: are they interchangeable?. Annals of Saudi medicine, 28(1), 33–41. https://doi. org/10.5144/0256-4947.2008.33

McCormack J, Chmelicek JT. (2014). Generic versus brand name: the other drug war. Canadian family physician Medecin de famille canadien, 60(10), 911.

Desai RJ, Sarpatwari A, Dejene S, Khan NF, Lii J, Rogers JR, Dutcher SK, Raofi S, Bohn J, Connolly JG, Fischer MA, Kesselheim AS, Gagne JJ. (2019). Comparative effectiveness of generic and brand-name medication use: A database study of US health insurance claims. PLoS medicine, 16(3), e1002763. https://doi.org/10.1371/ journal.pmed.1002763

Fishman J, Shrestha S, Wang L, Yuce H, Zorzoli J, Faught E, Farley J. (2018, April 10). Impact of Medicaid antiepileptic drug formulary restrictions on health outcomes and costs (P6.279). Retrieved December 09, 2020, from https://n. neurology.org/content/90/15_Supplement/P6.279

Measuring the Change in Use of Generic Oxcarbazepine (OXC) Versus Brand Names for Medicaid Enrollees

State Medicaid Formularies. (n.d.). Retrieved November 15, 2020, from https://advocacy.epilepsy.com/ StateFormularies

18. Wellington K, Goa KL. (2001). Oxcarbazepine. Mol Diag Ther 15, 137–163. https://doi.org/10.2165/00023210200115020-00005

Pennsylvania Medicaid Program. (n.d.). Retrieved December 09, 2020, from https://www.benefits.gov/ benefit/1148

Gifford K, Winter A, Wiant L, Dolan R, Tiana M, Garfield R. (2020, April 29). How State Medicaid Programs are Managing Prescription Drug Costs: Results from a State Medicaid Pharmacy Survey for State Fiscal Years 2019 and 2020 - Cost Containment and Utilization Control Strategies. Retrieved January 27, 2021, from https://www. kff.org/report-section/how-state-medicaid-programs-aremanaging prescription-drug-costs-cost-containment-andutilization-control-strategies/

19. By W, Team E. (2020, June 23). Pennsylvania (PA) Medicaid Program Eligibility. Retrieved December 05, 2020, from https://eligibility.com/medicaid/pennsylvania-pa-medicaid benefits

How to get prescription drug coverage. (n.d.). Retrieved November 16, 2020, from https://www.medicare.gov/drugcoverage-part-d/how-to-get-prescription-drug-coverage

Centers for Medicare & Medicaid Services. State Drug Utilization Data. 2018-2019. [Internet]. Maryland: Centers for Medicare & Medicaid Services. Medicaid Database. [cited 2021 January 26]. Available from: https://data. medicaid.gov/

10. McCormack J, Chmelicek JT. (2014). Generic versus brand name: the other drug war. Canadian family physician Medecin de famille canadien, 60(10), 911. 11. Medicaid & CHIP Enrollment Data. (n.d.). Retrieved November 30, 2020, from https://www.medicaid.gov/ medicaid/national-medicaid-chip-program-information/ medicaid-chip-enrollment-data/index.html 12. Medicaid Data Sources - General Information. (n.d.). Retrieved November 30, 2020, from https://www.cms. gov/Research-Statistics-Data-and-Systems/ComputerData-and Systems/MedicaidDataSourcesGenInfo/ index?redirect=%2FMedicaidDataSourcesGenInf o%2FMSIS%2Flist.asp 13. Musume and Priya Chidambaram Published: Jun 12, M. (2019, June 12). How Do Medicaid/CHIP Children with Special Health Care Needs Differ from Those with Private Insurance? Retrieved November 16, 2020, from https:// www.kff.org/0bbd44a/ 14. Pennsylvania Medicaid Program. (n.d.). Retrieved December 09, 2020, from https://www.benefits.gov/ benefit/1148 15. State Medicaid Formularies. (n.d.). Retrieved November 15, 2020, from https://advocacy.epilepsy.com/ StateFormularies 16. Step Therapy: Instead of “Fail First,” Put Patients First. (2019). Retrieved December 05, 2020, from https://www. rheumatology.org/ 17. The Centers for Medicare & Medicaid Services (CMS) Medicaid Integrity Group (MIG). (2015). Anticonvulsant Medications: Use in Adults [Fact sheet]. https://www.cms. gov/Medicare-Medicaid-Coordination/Fraud-Prevention/ Medicaid-Integrity-Education/Pharmacy-EducationMaterials/Downloads/ac-adult-factsheet11-14.pdf

20. Welcome to COMPASS. (n.d.). Retrieved December 11, 2020, from https://www.compass.state.pa.us/compass. web/Public/CMPHome 21. Zack MM, Kobau R. (2017). National and State Estimates of the Numbers of Adults and Children with Active Epilepsy United States, 2015. MMWR. Morbidity and mortality weekly report, 66(31), 821–825. https://doi.org/10.15585/mmwr. mm6631a1 22. September 2020 Medicaid & CHIP Enrollment Data Highlights. (n.d.). Retrieved January 26, 2021, from https:// www.medicaid.gov/medicaid/program-information/ medicaid-and chip-enrollment-data/report-highlights/ index.html 23. Food and Drug Administration. Clinical Pharmacology Review. 2012. [Internet]. Maryland: Food and Drug Administration. FDA [cited 2021 March 26]. Available from: https://www.fda.gov/files/drugs/published/Oxcarbazepine202810-Clinical-PREA.pdf 24. TRILEPTAL (oxcarbazepine). Drugs at FDA. 2017. [Internet]. Maryland: Food and Drug Administration. FDA [cited 2021 March 26]. Available from: https://www.accessdata.fda.gov/drugsatfda_docs/ label/2017/021014s036lbl.pdf 25. OXTELLAR XR (oxcarbazepine) extended-release tablets, for oral use. Drugs at FDA. 2018. [Internet]. Maryland: Food and Drug Administration. FDA [cited 2021 March 26]. Available from: https://www.accessdata.fda.gov/ drugsatfda_docs/label/2018/202810s010lbl.pdf 26. Schachter SC. (n.d.). Medicaid AED rules for Generic substitutions by state. Retrieved February 10, 2021, from https://www.epilepsy.com/learn/professionals/diagnosistreatment/therapeutic-and-generic-substitutions/ medicaid-aed-rules

Scholarly Research In Progress • Vol. 5, November 2021

Hunger vs. Heart Failure: Can Food Insecurity Screening Prevent CHF Exacerbations? Sarah Eidbo1†‡, Johanna Dungca1†‡, Amanda Goetz1†‡, Nicholas Fiala1†‡, Andrew Denisenko1†‡, Julie Sturzen2, and John Pamula2 ¹Geisinger Commonwealth School of Medicine, Scranton, PA 18509 ²Robert Packer Hospital, Sayre, PA 18840 † Doctor of Medicine Program ‡ Authors contributed equally Correspondence: seidbo@som.geisinger.edu

Abstract Food insecurity, defined by the United States Department of Agriculture as “access by all people at all times to enough food for an active, healthy life,” is experienced by 14.5% of American households. Food insecurity can depend on multiple factors, including transportation, financial stability, and geographic location — some areas termed “food deserts” have limited access to affordable and nutritious food. These factors can lead to higher cardiovascular health risks in addition to creating health disparities within many populations with chronic illness. With respect to patients with congestive heart failure (CHF), food security is increasingly relevant, as low-salt diets are one of the most important recommendations for management of heart failure. The purpose of this study was to examine the relationship between food insecurity with management outcomes, including readmissions and mortalities, in congestive heart failure patients within the Guthrie Clinic in Sayre, Pennsylvania. This quality improvement prospective study used retrospective review of records. Subjects were identified by an Epic report with baseline data from 2019. Identified adult patients with CHF were screened using the Household Food Insecurity Access Scale (HFIAS), and if identified as food insecure, were provided appropriate nutritional counseling. This study increased food insecurity screening rates from 86.6% to 94% over a 6-month span. Within the population of admitted CHF patients, readmission rates for a CHF exacerbation dropped from 27.6% to 21.2% of CHF patients after food insecurity screening and education was implemented. Mortality rates dropped from 24.1% to 19.8% after food insecurity screening and education was implemented. Although the differences between readmission and mortality rates before and after implementing food insecurity screenings in patients with CHF were not statistically significant, these differences are still important to note.

Introduction Food insecurity is a growing problem within the United States. With 14.9% of households experiencing food insecurity overall, and rates approaching 25% in black and Hispanic households, this is an issue that will affect communities with a variety of chronic illnesses that must be considered. The United States Department of Agriculture defines food insecurity as “access by all people at all times to enough food for an active, healthy life.” A variety of factors influence and are influenced by food insecurity and have been termed “social determinants of health” (1).

Last year, Healthy People 2020 created a “place-based” organizing framework to display how five key areas of social determinants of health (SDOH) can interact to influence a person’s well-being (1). These five key areas are education, social and community context, health and health care, neighborhood and built environment, and economic stability. Each of these five determinants interacts with the other four, giving a better idea of the social factors that determine a person’s well-being in addition to their psychological and physical health. Several important factors are covered in each of the five categories. A person’s education takes into account their early childhood education and development, enrollment in higher education, if they graduated high school, and their language and literacy level. A person’s social and community context considers their civic participation, any discrimination they experience, any experiences they have had being incarcerated, and their level of social cohesion. A person’s health and health care give information about their access to health care and primary care as well as their own health literacy. A person’s neighborhood and built environment gives us information about their quality of housing, the conditions of the environment that they inhabit, what kind of crime and violence occurs in their environment, and their access to foods that will help support their healthy eating patterns. A person’s economic stability includes their employment status, if they are living in poverty or are considered below the poverty line, any housing instability they experience, and if they are experiencing food insecurity. Thus, it is important to recognize how food insecurity fits into a larger picture of interweaving social determinants that can greatly influence a person’s health (1). Food insecurity can be influenced by multiple factors as well, including financial stability, transportation, and geographic location. For example, a patient living within a food desert may not have reliable access to nutritious food. A food desert is defined as an area with limited access to affordable and nutritious food — these areas can include urban areas as well as rural areas (2). In an area like Sayre, Pennsylvania, with a population of 5,500 as of 2019, the rural nature of the location is a large contributing factor. People living in a rural area without reliable transportation can have issues obtaining fresh produce if they live far from supermarkets or grocery stores. Unfortunately, food insecurity affects a person’s physical health. A study conducted by Morris et al. found that congestive heart failure-specific patients living in food deserts had higher rates of hospitalization when compared to those who did not live in food deserts (3). This could be partly due to the low-salt diet recommendation for managing heart failure.

Hunger vs. Heart Failure: Can Food Insecurity Screening Prevent CHF Exacerbations?

Table 2. Overview of sample characteristics.

Study interventions The quality improvement intervention was to identify CHF patients using the electronic health records, followed by screening of these patients for food insecurity Table 1. Questions asked by the Household Food Insecurity Access Scale, HFIAS. with the HFIAS, then providing those who screened positive with appropriate nutritional counseling. Data collected Diet is crucial to the management of this chronic condition. from the electronic medical record included gender, race, age, Currently at the Guthrie Clinic, food insecurity is assessed comorbidities, patients’ answers to the HFIAS, and whether using the Household Food Insecurity Access Scale (HFIAS), they were considered food insecure. which has been embedded within the Epic electronic health record system (Table 1). This was developed by Jean Miner, Assessment of food insecurity MD, and asks questions including if they have had difficulty paying for or obtaining food within the past 12 months (Table 1). Participants completed the HFIAS, with sample questions Given the high aging American population and high prevalence included in Table 1. Sample questions included rating how often of congestive heart failure (CHF) in the United States, this they felt that they had enough money to cover food costs, and project sought to utilize the HFIAS to identify CHF patients how often they felt unsure where their next meal was coming with food insecurity. This identification could help health from (Table 1). Answer options provided a range of frequencies care professionals facilitate appropriate enrollment into food for the participant to choose from. Participants who answered assistance programs or to disseminate information about local the questionnaire indicating any frequency of food insecurity food banks and other resources. This project aimed to examine within the last 12 months were considered food insecure and the relationship between food insecurity and management given nutritional counseling on food resources available to them. outcomes, including CHF-specific readmissions and mortalities, Assessment of management outcomes in congestive heart failure patients within the Guthrie Clinic.

Methods Participants and procedures Participants were 243 adult CHF patients admitted to Robert Packer Hospital of Sayre, Pennsylvania, between June 2020 and March 2021. Patients were identified via Epic report and included baseline data from 2019. Identified patients were screened using the HFIAS, and if identified as food insecure, were provided with appropriate nutritional counseling. Inclusion criteria included being 18 years or older, having a prior diagnosis of CHF upon admission to Robert Packer Hospital documented in the electronic medical record, and proficiency in English. Subjects under 18 years of age and pregnant patients were excluded from this study. Patients who were willing and eligible provided verbal consent and completed the HFIAS. There was no payment or cost to subjects to participate. Participants were placed at minimal risk in this study, as there is always a small chance of data becoming unsecured. To minimize this possibility, all protected health information was de-identified before results were shared. Any identifiable data was accessed using only Guthrie-approved applications. This project was IRB approved.

100

Electronic medical records were analyzed to determine management outcomes for participants. Readmissions at any time between the participant’s food insecurity counseling and the end date of the study were recorded. If the readmission was documented in the medical record as being due to an exacerbation of the patient’s congestive heart failure, it was counted as a readmission in this study. Readmissions for causes other than CHF-exacerbations were not included as a management outcome in this study. Mortalities of the participants were also recorded as a management outcome for this study. These management outcomes were analyzed among two groups of participants. One group consisted of 31 patients admitted to and discharged from Robert Packer Hospital for CHF exacerbations in June to September of 2020, before this project began screening patients for food insecurity. These patients were considered the baseline group to compare the intervention outcomes to. The second group included 212 patients admitted to and discharged from Robert Packer Hospital from September 2020 to March 2020, for which the HFIAS was administered, and management outcomes could be considered after the screening and food insecurity counseling.

Hunger vs. Heart Failure: Can Food Insecurity Screening Prevent CHF Exacerbations?

These groups were compared by the percent change for CHF-specific readmissions, mortalities, and number of patients screened for food insecurity. Microsoft Excel and GraphPad Prism were used for data analysis.

Results Sample characteristics demonstrated that among patients with congestive heart failure at Robert Packer Hospital, the majority are males with a mean age of 75.4 years (Table 2). Age was distributed across several decades, ranging from 40–49 to 80–89 (Figure 1). This study did not encounter any admitted CHF patients that were younger than 40 years or over 90 years. Among the CHF patients admitted, 7% were found to be food insecure (Table 2). Analysis of electronic medical records from both the control and screened groups revealed that when the HFIAS was used for

food insecurity screening, food insecurity screening rates rose from 86.6% to 94% among patients with CHF, an 8.5% increase in screening (Figure 2A). Between the two groups, CHF readmissions were found to drop by 6.4% from 27.6% to 21.2% after the HFIAS was utilized (Figure 2B). CHF mortality rates also had a drop of 4.3%, from 24.1% to 19.8% after the HFIAS was utilized (Figure 2C). A paired t-test was performed to determine that each of these changes was statistically nonsignificant.

Discussion Utilizing the HFIAS tool when screening CHF patients for food insecurity resulted in modest, nonsignificant readmission and mortality rate reductions within the Robert Packer Hospital. Although the differences between readmission rates and mortality rates before and after implementing food insecurity screenings were not statistically significant, these differences are still important to note. Noting that there was a slight increase in readmission rates and mortalities among CHF patients with food insecurity could help health care providers keep food insecurity in mind as an issue to address to help keep their patients healthier. Patients with CHF that screened positively for food insecurity fell evenly across a large age range. This could be due in part to food insecurity affecting people of all ages. It is one of many social determinants of health that influences other aspects of psychological and physical health. Since employment and socioeconomic status can vary among all age ranges, it is no surprise that food insecurity would follow.

Figure 1. Age distribution among CHF patients with identified food insecurity.

Although this study did reveal nonsignificant readmission and mortality decreases when the HFIAS was implemented, there are several limitations to note. This study took place during the COVID-19 pandemic, and as such, could have been influenced by the pandemic as well as by people’s perceptions of the

Figure 2. (A) Food insecurity screening rate in CHF patient groups observed prior to and after HFIAS implementation. (B) CHF readmission rates among CHF patient groups observed prior to and after HFIAS implementation. (C) CHF mortality rates among CHF patient groups observed prior to and after HFIAS implementation. 101

Hunger vs. Heart Failure: Can Food Insecurity Screening Prevent CHF Exacerbations?

pandemic. Surges in Sayre, Pennsylvania, could have impacted the patients’ willingness to leave their homes to seek treatment for CHF exacerbations, thus influencing data collection. These results could potentially be skewed toward people with more severe complications or exacerbations of CHF, thus leading them to become hospitalized regardless of the pandemic status. It could also affect a patient’s transportation to the hospital. Another determinant of health — transportation — could have been affected by COVID-19, especially in the rural area of Sayre, Pennsylvania. If a patient was unable to reach the hospital due to both the pandemic and being too far from the hospital to arrive here on their own, this could also have skewed our data toward encountering only CHF patients who lived within a reasonable distance of Robert Packer Hospital. Other confounding variables potentially include the weather conditions in Sayre, Pennsylvania. Due to poor road conditions for many months over the time period that the study took place, CHF patients too afraid to drive to the hospital or feeling unsure about driving too far in winter conditions may not have been included in this study. This could have skewed the results again toward patients with more severe complications or exacerbations of CHF, such as any exacerbation requiring an ambulance to take them to the hospital, thus negating the patient’s own fear of transportation during the winter months limiting their arrival. In addition to the limitations imposed by difficulties with transporting patients to the Robert Packer Hospital, there is a more insidious limitation to consider — the delicacy of the topic that this study addresses. Food insecurity is a sensitive topic for many people. Patients are wary to admit that they are having difficulty making ends meet in any context; these topics are rife with shame and guilt, making patients less likely to admit to them, let alone seek help or educational counseling. The patients that were able to admit to experiencing food insecurity in this study may not realize that they are far from alone in this struggle, but due to the stigma surrounding it, it is hard to know for certain whether the data presented here is fully accurate. Some CHF patients may have denied experiencing food insecurity when the HFIAS was administered. It is clear that in many ways, this study further emphasizes the interrelated nature of the five social determinants of health organized by Healthy People 2020. Food insecurity is one factor among many, but potentially a more significant factor among members of the community with chronic health conditions such as congestive heart failure.

102

Disclosures We have nothing to disclose.

References 1.

Decker D, Flynn M. Food Insecurity and Chronic Disease: Addressing Food Access as a Healthcare Issue. Rhode Island Medical Journal; 2018, 101(4), 28–30.

Schroeder K, Smaldone A. Food Insecurity: A Concept Analysis. Nursing Forum; 2015, 50(4), 274–284. https://doi. org/10.1111/nuf.12118

Morris AA, McAllister P, Grant A, Geng S, Kelli HM, Kalogeropoulos A, Quyyumi A, Butler J. Relation of Living in a “Food Desert” to Recurrent Hospitalizations in Patients With Heart Failure. The American Journal of Cardiology; 2019, 123(2), 291–296. https://doi. org/10.1016/j.amjcard.2018.10.004

Scholarly Research In Progress • Vol. 5, November 2021

Effects of Medical Cannabis on Patients with Charcot-Marie-Tooth Disease Priscilla C. Canals1†, Alexia G. Aguilar1†, Myriam Moise2, Jasmine Bernard3, Gregory T. Carter4, Allison Moore5, Robert Moore5, Joy Aldrich5, Margaret D’Elia6, Andrew Westerkamp7, Miyabe Shields7, and Brian J. Piper1 ¹Geisinger Commonwealth School of Medicine, Scranton, PA 18509 ²The University of Scranton, Scranton, PA 18510 ³Lincoln University, Chester County, PA 19352 4 Saint Luke’s Rehabilitation Institute, Spokane, WA 99202 5 Hereditary Neuropathy Foundation, New York, NY 10016 6 Champlain Valley Dispensary, Burlington, VT 05401 7 Real Isolates, Woburn, MA 01801 † Doctor of Medicine Program Correspondence: pcanals@som.geisinger.edu

Abstract Background: Charcot-Marie-Tooth (CMT) disease, also known as hereditary motor and sensory neuropathy, is a rare disease that affects 1 in 2,500 people. CMT can damage the nerves themselves or the myelin sheath surrounding them. CMT can cause symptoms such as neuropathic pain, weak ankles, clawlike hands, and muscle wasting in the extremities. Although treatments can help patients manage their symptoms, no known cure exists. In this study, we sought to determine the efficacy of using medical cannabis as a method of symptom relief in CMT patients. Methods: We collected patient-reported surveys through the Hereditary Neuropathy Foundation’s patient registry: Global Registry for Inherited Neuropathies. The online survey contained 52 multiple-choice questions. We analyzed the data through IBM SPSS and created figures through GraphPad Prism. The sample participants (N=56) consisted of 71.4% females. Ages ranged from 18 to 87. The majority (48.5%) were CMT1, 18.2% were CMT2, and 6.1% were CMT4. Results: Women were more likely to report experiencing pain than men (p<.05). Participants who perceived support from their providers were more likely to inform them of their cannabis use (p<.05). When asked about how much relief they experience from using cannabis as a method of symptom relief, respondents reported an average of 69.6% (SEM + 2.6) relief. Conclusion: Our results indicate that patients reported receiving substantial symptom relief from medical cannabis. Further prospective or controlled research is necessary to extend upon these findings.

Introduction Charcot-Marie-Tooth (CMT) disease is one of the most common inherited neuromuscular disorders, affecting about 1 in 2,500 people. It can be caused by mutations in over 90 genes that affect peripheral nerves or by diabetes or chemotherapy (1). There are five CMT types, with type 1 being the most prevalent. Mutations can either affect nerves themselves or the myelin sheath. In both cases, weaker messages travel between the brain and the extremities. Mutations can be inherited through autosomal dominant, autosomal recessive, or X-linked patterns (2). CMT1 is an autosomal dominant (3), demyelinating neuropathy (4).

CMT2 is an autosomal dominant axonal peripheral neuropathy (4). CMT4 is a rare autosomal recessive demyelinating neuropathy (5). Hereditary neuropathy with liability to pressure palsies (HNPP) is an autosomal dominant peripheral neuropathy, which leads to demyelination (6). Demyelinating forms lead to degradation of the myelin sheath. Axonal forms lead to axon deterioration in the peripheral nerves (7). Although CMT has a broad genetic heterogeneity, patients display similar phenotypes (8). Patients with CMT experience various motor deficit symptoms, such as clumsiness, ankle twisting, muscle atrophy of the extremities, hammertoes, and claw hands. They also experience sensory deficit symptoms, such as pain, pins and needles, and burning (9). Symptoms usually worsen over time, but the severity of symptoms vary from person to person. There are several ways to diagnose CMT disease. It can be diagnosed through genetic testing or through electrophysiological examination, which measures motor conduction velocity (VCM). Those with CMT will display a decreased VCM if they have demyelinating neuropathy (9). Genetic tests can be conducted to receive a CMT diagnosis. Although there is no known cure for CMT, there are treatments available such as physical therapy, occupational therapy, pain medication, and orthotics to manage symptoms. In the past, there has been taboo surrounding the use of cannabis. Researchers conducted a qualitative descriptive study on the stigma surrounding cannabis use. People who use cannabis for therapeutic purposes (CTP) feel they are perceived as “potheads” by their families and healthcare providers. They also state that when they explain the medical benefits of cannabis, their providers and families respond with mistrust. Due to negative reactions, many participants choose to keep their use undercover (10). Withholding information from providers can have negative consequences due to possible drug interactions. CBD has potential inhibitory effects on CYP3A4, which is a vital liver enzyme responsible for removing drugs from the body (11). CYP3A4 substrates include benzodiazepines, opioids, and antidepressants (12). Another cytochrome that is inhibited by THC and more so, CBD, is CYP2D6 (13). This enzyme is primarily active in the liver and its substrates include medications such as antipsychotics and antidepressants (14). Inhibition of enzymes leads to increased blood levels of substrate. As a result, there is a greater possibility of adverse effects, such as overdosing 103

Effects of Medical Cannabis on Patients with Charcot-Marie-Tooth Disease

on medications (12). Researchers conducted a survey study where they gathered oncologists’ thoughts regarding medical cannabis. Seventy percent of physicians said they did not feel as though they had enough knowledge on medical cannabis to recommend it to their patients (15). In another survey study, investigators found that 38.7% of primary care providers feel medical cannabis should be recommended for managing symptoms associated with medical illnesses (16). Currently, 36 states have legalized medical cannabis, furthering research, and encouraging clinicians to consider cannabis as a therapeutic alternative for patients. Cannabis has cannabidiol (CBD) and tetrahydrocannabinol (THC) components. Both components have been found to have anti-inflammatory and analgesic effects, while THC has an added psychoactive component (17). There are three types of strains: indica, sativa, and hybrids. Each produces different effects and combats separate issues. Indica is often recommended for medical conditions such as chronic pain, anxiety, insomnia and muscle spasms (18). As stated in the MMP directory website, benefits of sativa use include reduced fatigue, pain, headaches, nausea, and increased focus. Hybrid strains provide the effects of both sativa and indica strains. Dispensary staff commonly recommends hybrid strains for amyotrophic lateral sclerosis (ALS) (18). Humans have an endocannabinoid system (ECS) with cannabinoid receptors (CB1 and CB2) and endogenous lipid ligands (19). Endocannabinoid receptors recognize and bind exogenous plant-derived cannabinoids (20). Cannabinoids target nociceptors, which are receptors associated with painful stimuli. They exert their effects on nociceptors in the brain, spinal cord, and peripheral nerves, which conversely are the locations of CB1 receptors. When chronic or neuropathic pain is rooted in inflammation, cannabinoids also target CB2 receptors (20). CBD has potential therapeutic effects, such as anti-inflammatory, anti-anxiety, and anti-nausea. According to animal studies, cannabis is effective for treating chronic pain, specifically neuropathic and inflammatory pain. In a case study, a 67-year-old female with multiple sclerosis (MS) reported neuropathic pain and rated the pain a 9 out of 10. She was prescribed 1 g of a 2.5% THC and 5% CBD strain per day through vaporizer. She reported that this strain was unsuccessful at treating her symptoms, so they changed the prescription to a 9% THC and 13% CBD strain. During her 60 day follow-up, she rated her pain a 5 out of 10 and reported that she planned on lowering the dose of other medications (21). Thus, the effectiveness of medical cannabis may be dose dependent. Additionally, the substitution effect can occur when medical cannabis replaces recreational or pharmaceutical drug use (22). A cross-sectional survey study shows that medical cannabis use was correlated with a 64% decrease in opiate use (23). Another survey study found that cannabis served as a replacement for opiates (30%), alcohol (25%), benzodiazepines (16%), and antidepressants (12%) (24). Cannabis has been ruled a Schedule I drug under federal law. A Schedule I drug is said to have no medical purpose and to have high abuse potential. This creates a limitation in the research that can be conducted to determine its efficacy and safety. In this study, we seek to determine the efficacy of using medical cannabis as a source for symptom relief as reported by CMT patients. 104

Methods Participants Participants (N=56, 71.4% female, Age= 48.9, Min= 18, Max=87, Figure 1). Further demographic information was not collected to maintain anonymity. Survey The survey used for this study was modified from (22). Most of the items were kept the same, but some were tailored to address the sample we are studying. Surveys are posted on the Hereditary Neuropathy Foundation (HNF) website. A link was also created and sent to all the members of the foundation. Those affected by CMT were free to take the survey. It contained 52 questions, which were all multiple choice. The survey questions can be accessed in the Appendix section. The surveys were anonymous to maintain privacy and to minimize bias. Generally, it took participants about 10 minutes to complete. The survey was designed to gather information on the effectiveness of medical cannabis as a treatment for symptom relief for those affected by CMT. The purpose was to gather this information from a patient perspective. Since this study does not cause any harm to participants, the study was IRB approved as exempt by Advarra. Data analysis After the surveys were collected, we input the data into IMB SPSS, version 26. Through this software program, we conducted Chi-square and t-tests to find associations between variables. We created figures and tables through GraphPad Prism, version 8. A p<0.05 was considered statistically significant. The variability was reported as the Standard Error of the Mean (SEM). For the results listed in Table 1, the percentages do not add up to 100 because respondents were allowed to choose more than one symptom/mobility device. For the symptom category in Table 1, we combined the results for two variables, symptoms, and characteristics.

Results A total of 56 people answered the survey. When asked if they experienced pain associated with CMT, 90.9% of participants answered yes. The majority (48.5%) of respondents had CMT type 1A, which is the most prevalent CMT type; 27.3% were HNPP; 18.2% were CMT2; and 6.1% were CMT4 (Figure 1). In this sample population, about three-quarters of respondents had weak ankles, about one-third had muscle atrophy and/or tremors, and about a quarter had temperature sensitivity and/or poor or absent reflexes (Table 1). Other symptoms alleviated by medical cannabis according to participants were pain, substance cravings, and anxiety (Table 2). Furthermore, many respondents needed the aid of mobility devices. About one-fifth needed canes/walking sticks and/or braces, one-sixth needed custom foot orthotics/inserts, and about one-tenth needed walkers (Table 1). Of the sample population, 33.9% possessed a medical cannabis certificate. Most (90.9%) respondents said they experienced pain. When prompted with “How effective is medical cannabis in alleviating your symptoms with CMT?” respondents provided the percent of relief they experienced from medical cannabis.

Effects of Medical Cannabis on Patients with Charcot-Marie-Tooth Disease

Table 2. Other symptoms alleviated by medical cannabis.

Figure 1. Participants’ CMT types (bottom) and age of participants (top). HNPP: Hereditary neuropathy with liability to pressure palsies.

Table 1. Symptoms experienced and mobility devices used by respondents.

105

Effects of Medical Cannabis on Patients with Charcot-Marie-Tooth Disease

Figure 2. Percent CMT symptom relief from using medical cannabis (mean= 69.6, SEM= 2.6).

Figure 5. How likely participants are to inform providers of cannabis use based on perceived support.

Figure 6. Difference in plans of stopping usage of medical cannabis in participants who have and have not experienced negative side effects.

Most respondents said that cannabis provided 80% relief (Figure 2). Participants were asked if they noticed a change in their use of other medications after they began using medical cannabis. In Figure 3, we show how many respondents reported a decrease in medication use. Four-fifths (80%) reported using less opiates, 68.8% responded using less sleep medication, 50% reported using less anxiety medication, and 47.8% reported using less antidepressants (Figure 3). We wanted to see if there was an association between gender and pain. All (100%) of female respondents and 72.7% of male respondents reported feeling pain associated with CMT (Figure 4). Additionally, we ran a 2x2 chi-squared test on gender and pain variables. There was a significant relationship between gender and pain. Figure 5 shows physicians’ attitudes regarding patient medical cannabis use. Based on those attitudes, we measured how likely participants were to inform the provider of their cannabis use. As evident below, the more positive the provider’s response, the more likely they are to inform them. For those who reported receiving a supportive response from providers, 63.6% said that they would inform them of their medical cannabis use. Figure 3. Percent reducing medication use after using medical cannabis.

Figure 4. Gender differences in pain reporting.

106

For those who reported unsupportive responses from providers for medical cannabis use, only 18.8% said they would inform providers of cannabis use. We conducted a 2x2 chi-squared test to analyze whether provider attitudes regarding medical cannabis affect patients' decisions to inform them of cannabis use. When providers are perceived by patients as being supportive, patients are significantly more likely to inform them of cannabis use than if providers are unsupportive toward cannabis use (p < 0.05). Participants reported whether they experienced negative side effects due to cannabis use and if they planned on halting their usage. The preponderance (91.7%) of participants who experienced negative side effects and 87.2% of respondents who did not have negative side effects said they did not have plans to stop consuming medical cannabis (Figure 6).

Effects of Medical Cannabis on Patients with Charcot-Marie-Tooth Disease

Figure 7. Preferred method of consumption among participants.

Figure 8. Commonly used strains among participants.

Preferred methods of consumption among participants are smoking (34%), vaporize (28.3%), edibles (26.4%), sublingual tinctures (7.6%), and topical (3.8%) (Figure 7). Preferred strains were indica dominant strains (22.2%), 50/50 THC/CBD strains (20.4%), high CBD strains (16.7%), high THC strains (9.3%), sativa dominant strains (5.6%), and hybrid strains (3.7) (Figure 8).

neuropathic pain in patients with MS or with traumatic nerve injury (20). Per the literature (19, 20, 28,32), medical cannabis provides relief for symptoms such as spasticity and pain. Cannabis appears to be a promising therapy option for CMTrelated symptoms.

Discussion This novel study shows that medical cannabis provides symptom relief for CMT patients. Studies have examined the efficacy of medical cannabis for MS-related symptoms, such as spasticity and tremors. For MS patients with spasticity, oral cannabis extract (OCE) and THC are effective at reducing patientreported pain scores. However, for MS patients with tremors, OCE and THC are likely ineffective for treating this symptom (19). In a randomized controlled trial conducted among MS patients, those who smoked medical cannabis scored about 2.7 points lower on the Ashworth Scale than those in the control group (25). The Ashworth Scale measures spasticity in MS patients, with 0 being no resistance and 4 being rigidity. The U.S. FDA has approved a mucosal spray (nabiximols) that consists of CBD and THC in a 1:1 ratio for clinical trials. This drug may aid in decreasing spasticity in MS patients (26). In patients with amyotrophic lateral sclerosis (ALS), cannabis possibly aids in muscle relaxation, appetite stimulation, and pain reduction (27). In a survey examining the effects of cannabis on ALS patients, researchers found that cannabis is moderately effective at relieving symptoms such as spasticity, pain, and depression (28). Additionally, in open-label observational study, Parkinson’s disease (PD) patients showed an improvement in tremors, rigidity, and pain 30 minutes after smoking cannabis (29). Symptoms experienced in CMT overlap with those experienced in MS, ALS, and PD. Almost all participants said they experienced pain, one-third had tremors and one-fifth had spasticity and/or muscle rigidity. In general, patients with unmanaged neuropathic pain have poorer health outcomes and have an increased probability of developing anxiety and depression. The ECS plays an important role in regulating neuropathic pain. A series of studies have shown that compared to placebo groups, those exposed to short-term use of medical cannabis experienced significant alleviation of neuropathic pain (30, 31, 32). Additionally, in a randomized, double-blind control study, THC reduced

There was a significant association (p<0.05) between gender and pain. It is unknown whether there were biological differences or/and sociocultural expectations that lead to the differences in pain sensation. A literature review suggests that females are more sensitive to experimentally induced pain and have a higher prevalence of common forms of pain (33). Social and cultural factors also contribute to the differences in expression of pain between genders. By societal standards, it is more acceptable for females to display pain expression. In turn, this can lead to respondent bias, where men downplay their perception of pain (34). Additionally, studies show that neuropathic pain is more prevalent in females than in males (33). In an online survey study, researchers found that twice as many consumers preferred indica strains to sativa strains (22). In this study, we discovered that about a quarter of respondents preferred indica dominant strains, which are used to provide relaxing and sedating effects. One-fifth of respondents said they use 50/50 THC/CBD strains. One-sixth said they used high CBD strains to provide more relaxing effects and are used to relieve symptoms. One-tenth of participants said they use high THC strains that are more potent and provide the effect of being “high.” Only 5.6% of participants said they used sativa dominant strains, which provide an uplifting effect. Participants used hybrid strains, which provide a combination of indica and sativa effects, the least. Perceived provider attitudes affected patients' likelihood to inform providers of their medical cannabis use. Some physicians have conflicting views on medical cannabis. While physicians may be hesitant to endorse medical cannabis, with increasing evidence-based guidelines this hesitance is likely to decrease. Patients are also aware of the social stigma that surrounds cannabis use, so they are less likely to disclose their usage to friends and family (21). In a cross-sectional survey study among medical cannabis users in Canada, 79.3% of respondents said they withheld information regarding their cannabis use, mainly to avoid being judged. Additionally, 10.1% stated that their physicians were unsupportive of their medical cannabis use, while 38% perceived their physicians to be supportive. Furthermore, 32.6% of respondents said that their physician

107

Effects of Medical Cannabis on Patients with Charcot-Marie-Tooth Disease

had refused to provide them with a medical certificate, mostly because the physician was not sure of the therapeutic benefits or because they were afraid of negative consequences on behalf of the medical association (35). This supports the data seen in Figure 7, which shows that when people perceive positive feelings and decreased judgment from their providers, they were more likely to inform them that they are using cannabis to experience symptom relief. Medical cannabis has been legalized in 36 states, but many providers are hesitant to provide medical cannabis certificates. As seen in our study population, only 33.9% of participants had medical cannabis certificates. In a survey study done in Ohio, where medical cannabis has been legalized, they found that there are geographic limitations to where patients can find physicians that provide medical cannabis certificates. Providers who do provide certificates were usually located in cities, whereas there was a lack of certifying physicians in rural areas (36). Another study found that over time, both adolescents and adults perceive cannabis to be of decreased risk. In physicians, they did not find the same trends. Due to inconvenient reporting requirements, providers do not recommend and often dissuade patients from using medical cannabis (37). This leads to people obtaining cannabis illegally or to patients withholding information regarding cannabis use from their providers. There are several notable limitations to this study. Inherent to survey studies is respondent bias. In this study, all participants (N=56) used cannabis. Given the branched design of the survey, there were multiple questions where fewer than 20 respondents provided data. Therefore, the data is biased toward a positive cannabis experience, since those who are more inclined to answer the survey generally have more positive experiences with medical cannabis. Although the sample size is relatively small, CMT patients are not easy to come across since this is a relatively rare disease, making these findings exciting. Within the sample, not all the forms of CMT were accounted for, such as CMT3. Additionally, because of the small sample size, there may be a lack of generalizability. Cannabis not being used in a controlled setting (e.g., dosage, composition, methods of delivery) was also a limitation. Cannabis can be absorbed in many ways; whether it’s inhaled through vapor or ingested in food or drinks, each method has a different rate of absorption. Though some of the participants had reported a decrease in their pain levels, there is no indication as to how much cannabis they consumed for them to see that decrease. It is possible that those who reported decreased levels of pain consumed an increased dose of cannabis and/or had a different method that was faster at relieving pain than that of those participants who reported no change in their pain levels. Physical and psychological adverse events were not considered. Finally, they did not report if they had psychedelic effects.

innovative and shows that medical cannabis is a possible avenue to take when assessing treatments for neuropathic pain. For further research, it would be beneficial to continue receiving surveys to get more robust results. We could also build on the results that we have and hopefully provide a new and effective modality for symptom relief for CMT patients. Future studies will benefit with a wider range of participants to account for the missing CMT types. If possible, accounting for medical cannabis effects in a controlled environment and having the possibility of comparing placebo to treatment groups would provide valid data. Since this was the first study exploring CMT and medical cannabis, we believe this is exciting and opens a new door to providing more effective treatments for CMT patients.

Acknowledgments Thanks to Iris Johnston from the Geisinger Commonwealth School of Medicine Library for granting us access to the literature that was imperative in conducting our research.

Disclosures PCC, AGA, MM, JB, MM, and BJP were supported by a grant from the HRSA foundation. BJP is part of an osteoarthritis research team supported by Pfizer and Eli Lilly. A member of his immediate family was employed by a CBD company. MD is employed by a dispensary. The other authors report no disclosures.

References 1.

Zhao HT, Damle S, Ikeda-Lee K, et al. PMP22 antisense oligonucleotides reverse Charcot-Marie-Tooth disease type 1A features in rodent models. J Clin Invest. 2018 Jan 2;128(1):359–68.

Berciano J, García A, Gallardo E, et al. Intermediate Charcot–Marie–Tooth disease: an electrophysiological reappraisal and systematic review. J Neurol. 2017 Aug;264(8):1655–77.

Timmerman V, Raeymaekers P, De Jonghe P, et al. Assignment of the Charcot-Marie-Tooth neuropathy type 1 (CMT 1a) gene to 17p11.2-p12. Am J Hum Genet. 1990 Oct;47(4):680-85.

Barreto LCLS, Oliveira FS, Nunes PS, et al. Epidemiologic study of Charcot-Marie-Tooth disease: a systematic review. Neuroepidemiology. 2016;46(3):157–65.

Hu B, McCollum M, Ravi V, et al. Myelin abnormality in Charcot-Marie-Tooth type 4J recapitulates features of acquired demyelination. Ann Neurol. 2018 Apr;83(4):756– 70.

Chen B, Niu S, Wang X, et al. Clinical, electrophysiological, genetic, and imaging features of six Chinese Han patients with hereditary neuropathy with liability to pressure palsies (HNPP). J Clin Neurosci. 2018 Feb;48:133–7.

Morelli KH, Seburn KL, Schroeder DG, et al. Severity of demyelinating and axonal neuropathy mouse models is modified by genes affecting structure and function of peripheral nodes. Cell Rep. 2017 Mar 28;18(13):3178–91.

Conclusion Although studies have shown inconclusive results regarding the effectiveness of medical cannabis in pain relief, the findings from the surveys indicate that cannabis provides substantial symptom relief in CMT patients. Since CMT is a rare and relatively new disease, we believe that this sample (N=56) provides valuable insight into the possibility of using cannabis as a method of symptom relief. Additionally, we believe this data is

108

Effects of Medical Cannabis on Patients with Charcot-Marie-Tooth Disease

Lupo V, García-García F, Sancho P, et al. Assessment of targeted next-generation sequencing as a tool for the diagnosis of Charcot-Marie-Tooth disease and hereditary motor neuropathy. J Mol Diagn. 2016 Mar;18(2):225–34.

23. Boehnke KF, Litinas E, Clauw DJ. Medical cannabis use is associated with decreased opiate medication use in a retrospective cross-sectional survey of patients with chronic pain. J Pain. 2016 Jun;17(6):739–44.

Corrado B, Ciardi G, Bargigli C. Rehabilitation management of the Charcot–Marie–Tooth syndrome: a systematic review of the literature. Medicine (Baltimore). 2016 Apr;95(17):e3278.

24. Lucas P, Walsh Z. Medical cannabis access, use, and substitution for prescription opioids and other substances: a survey of authorized medical cannabis patients. Int J Drug Policy. 2017 Apr;42:30–5.

10. Bottorff JL, Bissell LJ, Balneaves LG, et al. Perceptions of cannabis as a stigmatized medicine: a qualitative descriptive study. Harm Reduct J. 2013 Feb 16;10(1):2.

25. Corey-Bloom J, Wolfson T, Gamst A, et al. Smoked cannabis for spasticity in multiple sclerosis: a randomized, placebocontrolled trial. CMAJ. 2012 Jul 10;184(10):1143–50.

11. Qian Y, Gurley BJ, Markowitz JS. The potential for pharmacokinetic interactions between cannabis products and conventional medications. J Clin Psychopharmacol. 2019;39(5):462–71.

26. Suryadevara U, Bruijnzeel DM, Nuthi M, et al. Pros and cons of medical cannabis use by people with chronic brain disorders. Curr Neuropharmacol. 2017;15(6):800-14.

12. Brown JD. Potential adverse drug events with tetrahydrocannabinol (THC) due to drug–drug interactions. J Clin Med. 2020 Mar 27;9(4):919.

27. Carter GT, Abood ME, Aggarwal SK, et al. Cannabis and amyotrophic lateral sclerosis: hypothetical and practical applications, and a call for clinical trials. Am J Hosp Palliat Care. 2010 Aug;27(5):347–56.

13. Yamaori S, Okamoto Y, Yamamoto I, et al. Cannabidiol, a major phytocannabinoid, as a potent atypical inhibitor for CYP2D6. Drug Metab Dispos. 2011 Nov;39(11):2049–56.

28. Amtmann D, Weydt P, Johnson KL, et al. Survey of cannabis use in patients with amyotrophic lateral sclerosis. Am J Hosp Palliat Care. 2004;21(2):95–104.

14. Ford KA, Ryslik G, Sodhi J, et al. Computational predictions of the site of metabolism of cytochrome P450 2D6 substrates: comparative analysis, molecular docking, bioactivation and toxicological implications. Drug Metab Rev. 2015 Aug;47(3):291–319.

29. Lotan I, Treves TA, Roditi Y, et al. Cannabis (medical marijuana) treatment for motor and non–motor symptoms of Parkinson disease. Clin Neuropharmacol. 2014;37(2):41– 4.

15. Braun IM, Wright A, Peteet J, et al. Medical oncologists’ beliefs, practices, and knowledge regarding marijuana used therapeutically: a nationally representative survey study. J Clin Oncol. 2018 Jul 1;36(19):1957–62. 16. Philpot LM, Ebbert JO, Hurt RT. A survey of the attitudes, beliefs and knowledge about medical cannabis among primary care providers. BMC Fam Pract. 2019 Jan 22;20(1):17. 17. Chiurchiù V, Stelt MVD, Centonze D, et al. The endocannabinoid system and its therapeutic exploitation in multiple sclerosis: clues for other neuroinflammatory diseases. Prog Neurobiol. 2018 Jan;160:82–100. 18. Haug NA, Kieschnick D, Sottile JE, et al. Training and practices of cannabis dispensary staff. Cannabis Cannabinoid Res. 2016 Dec 1;1(1):244–51. 19. Koppel BS, Brust JCM, Fife T, et al. Systematic review: efficacy and safety of medical marijuana in selected neurologic disorders: report of the Guideline Development Subcommittee of the American Academy of Neurology. Neurology. 2014 Apr 29;82(17):1556–63. 20. Pacher P, Bátkai S, Kunos G. The endocannabinoid system as an emerging target of pharmacotherapy. Pharmacol Rev. 2006 Sep;58(3):389–462. 21. Ko GD, Bober SL, Mindra S, et al. Medical cannabis — the Canadian perspective. J Pain Res. 2016 Sep 30;9:735-44.

30. Wilsey B, Marcotte T, Tsodikov A, et al. A randomized, placebo-controlled, crossover trial of cannabis cigarettes in neuropathic pain. J Pain. 2008 Jun;9(6):506–21. 31. Wallace MS, Marcotte TD, Umlauf A, et al. Efficacy of inhaled cannabis on painful diabetic neuropathy. J Pain. 2015 Jul;16(7):616–27. 32. Lee G, Grovey B, Furnish T, et al. Medical cannabis for neuropathic pain. Curr Pain Headache Rep. 2018 Feb 1;22(1):8. 33. Fillingim RB, King CD, Ribeiro-Dasilva MC, et al. Sex, gender, and pain: a review of recent clinical and experimental findings. J Pain. 2009 May;10(5):447–85. 34. Bartley EJ, Fillingim RB. Sex differences in pain: a brief review of clinical and experimental findings. Br J Anaesth. 2013 Jul;111(1):52-8. 35. Leos-Toro C, Shiplo S, Hammond D. Perceived support for medical cannabis use among approved medical cannabis users in Canada. Drug Alcohol Rev. 2018 Jul;37(5):627–36. 36. Lombardi E, Gunter J, Tanner E. Ohio physician attitudes toward medical cannabis and Ohio’s medical marijuana program. J Cannabis Res. 2020 Apr 21;2(1):16. 37. Carliner H, Brown QL, Sarvet AL, et al. Cannabis use, attitudes, and legal status in the U.S.: A review. Prev Med. 2017 Nov;104:13–23.

22. Piper BJ, Dekeuster RM, Beals ML, et al. Substitution of medical cannabis for pharmaceutical agents for pain, anxiety, and sleep. J Psychopharmacol. 2017 May;31(5):569–75.

109

Scholarly Research In Progress • Vol. 5, November 2021

Short- and Long-Term Outcomes of Breastfeeding on Children’s Mental and Physical Health Taylor S. Mewhiney1* ¹Geisinger Commonwealth School of Medicine, Scranton, PA 18509 *Master of Biomedical Sciences Program Correspondence: tmewhiney@som.geisinger.edu

Abstract Breast milk is a perfectly evolved substance that fits the dietary needs of the average infant to ensure it thrives during growth and development. Breastfeeding significantly impacts children’s lives as it works to maintain their health from the minute they are born until the mother chooses to cease lactation. Research has shown that after being fed breast milk, children had stronger immunity and their fatality rates were decreased compared to babies who were not breastfed. Breastfed children have higher IQs, healthier BMIs, decreased obesity and blood pressure, and reduced risk of rhinitis allergies, asthma, and skin conditions. A breastfeeding mother is also provided with health benefits, as her risk for ovarian and breast cancer, diabetes, and postpartum depression is reduced. When given proper instruction to feed her infant, a mother can better herself and her baby well into their future. For the first 6 months of a child’s life, breastfeeding is strongly encouraged to give the child its greatest source of nutrition. The purpose of this review is to enhance perspectives on breastfeeding and to increase awareness on the long-term and short-term outcomes of breastfeeding.

Introduction Breast milk is produced through the process of lactation in the mammary glands. During pregnancy, the hormones in a woman’s body begin to significantly fluctuate. Halfway through the pregnancy, lactation begins. There are two phases of lactation: lactogenesis I and II. Lactogenesis begins while the baby is in utero, and the mammary glands begin to produce colostrum (1). Colostrum is a discharge-like fluid produced in small quantities, as progesterone prohibits the mother from full lactation until lactogenesis II. The second phase of lactogenesis begins following delivery of the placenta. The hormones cortisol, prolactin, and insulin spike, causing milk production. During this time there is a major decrease in estrogen and progesterone. The endocrine system then drives the volume of milk production to rapidly increase from less than 100 mL to 500 mL 4 days after birth. About a week postpartum, the mother produces around 650 mL (1). The newborn begins to adjust to eating during its first days postpartum, as it is naturally equipped to suckle. During this stage of the newborn’s life, it is provided with a rare variety of nutrients it cannot obtain elsewhere. This is the most essential time in a child’s growth since exposure to the world outside the womb, and breast milk gives the infant plenty of benefits. The general composition of breast milk is rich in macronutrients. The largest component of breast milk is water, which counts for 87% of its structure. Fat makes up 3.8%, and lactose accounts for 7%. Protein makes up 1% of breast milk as either whey or

110

casein, which have fluctuating ratios depending on the stage. However, during the entirety of lactation, the content of whey is higher than casein (2). Considering breast milk has a unique microbiome, it has the potential to transfer healthy species from mother to child, which can positively impact the baby (1). Bacterial species are also introduced to the infant through breast milk; these are probiotic bacteria that promote digestive health. Breast milk also contains vitamins, minerals, enzymes, and hormones essential for growth and development. These provide diversification for the infant immune responses and help colonization of upward of 90% of the infant biome (2). Production of colostrum begins during lactogenesis I, but measurable quantities are released from the mammary gland within the first few days post-birth. It is yellow in color and produced in minimal volume. Colostrum intake is essential, as it begins to build immunity in the newborn and acts as a barrier in the digestive tract. Colostrum also helps the newborn produce their first bowel movement. Beginning between days 4 and 5 postpartum, transitional milk is formed. This is a light-yellow substance that provides the infant with more nutrition than colostrum. Transitional milk production varies from colostrum production because it increases in quantity. The more the newborn feeds, the higher the production rate (3). Anywhere from 4 to 6 weeks postpartum is when milk is considered mature. Mature milk is white and opaque and will remain in this phase until the mother ends lactation. Mature milk is recommended to be consumed until the infant is 6 months to a year old, considering its high nutrient content (4). During a feeding session, the composition of milk transitions from foremilk to hindmilk. Foremilk is released first upon suckling. It is thin and rich in lactose and satisfies the newborn’s thirst. Hindmilk is expressed following foremilk. Hindmilk has a high-fat content and is denser than foremilk, further satisfying the baby’s nutritional necessities (2). The composition, texture, smell, and amount of breast milk vary upon the day and nursing session. Mothers can generally expect to produce enough to fulfill the baby's daily intake; however, some days a mother may produce more or less than others. This is natural during the process of lactation, as a set amount of milk is not expected. For a child to get the necessary amount of milk, it is best if the mother is relaxed and allows for the newborn’s instincts to kick in. Optimal positioning includes the mother sitting straight with support for her back while supporting the newborn’s body in her arms facing her breast. For proper attachment to the nipple, it is best if the newborn has their chin touching the breast with the head slightly tilted back, and the mouth open with the lower lip turned outward. By gently brushing the nipple between the upper lip and nose (not between upper and lower lip), the baby should naturally latch (5).

Short- and Long-Term Outcomes of Breastfeeding on Children’s Mental and Physical Health

Within the first year of a child’s life, a healthy infant will grow to be triple its weight from birth. During this time, the baby needs to obtain proper nourishment through food. Over the duration of the first 6 months of life it is highly recommended that the infant is given breast milk, as it is a foundational source of protection for the child. Even if a mother herself lacks adequate nutrition, breast milk can still sufficiently provide a newborn with all the nutrients it needs to grow (2). Breast milk has evolved to perfectly suit a baby as it adapts through the earliest stages of life. It contains all the essentials a child needs, whereas formula lacks the microbiome that is provided solely from the mother. With that being said, most mothers today are faced with the choice between breast and bottle feeding. While both have their positives and negatives, breast milk by far provides both mother and child a myriad of lifetime benefits that go beyond formula. Children who gain their nutrition through breastfeeding will have long-lasting positive effects shown in several studies following the beneficial aspects of breast milk and growth and development in a child.

Discussion Infant mortality rates Often in the developing world, mothers do not have the resources that can teach them to properly breastfeed their babies. When proper breastfeeding technique is instilled in a mother, the immunity of the child increases, along with the prevention of infection (4). Upon properly receiving breast milk children had lower mortality rates than those who weren’t breastfed. The children who did not receive breast milk also had an increased risk of infections (4). Measured rates of infection along with measured mortality rates provide an accurate assessment of the ways children obtain their nutrients and its effect on them. The study compared exclusive breastfeeding to partial breastfeeding and bottle-fed children. It took place in India, which classifies as an underdeveloped country. Infants from developing countries are 6 to 10 times more likely to die during the first stages of life if they are not breastfed (5). Patterns in the developing world continue to show a higher mortality rate in formula-fed children. In 2010, more than 7.7 million children passed away before their fifth birthday, with 98% of these children coming from developing countries (6). A common factor with these children was that they lived in areas with low percentages of breastfeeding. Research done in Libya reported how mothers attempted to feed and compared them to a list with criteria providing the best ways to feed a baby. A mother with less access to maternal guidance was far more likely to improperly feed her child, and it was recommended that first-time mothers should be monitored for the infant’s positioning and attachment to the breast (5). With the upkeep of suitable breastfeeding techniques, a contribution to a lower child mortality rate in developing countries is achievable and strongly urged (6). Following Roberts’ study in 2016, Victora et al. observed that if children were universally breastfed, 823,000 lives had the potential to be saved (7). By switching to exclusive breastfeeding, a significant number of babies would have the potential to make it past their first birthday in the developing world.

General statistics of breastfeeding Globally, 38% of infants are breastfed exclusively, meaning there is no outside source of formula introduced to the diet. Some mothers choose to provide both for their child to ensuring there is always food available; however, not all mothers are fortunate enough to interchange between the two. While the World Health Organization (WHO) recommends breastfeeding for a minimum of 6 months, formula is the more widely used choice (2). Especially in low-income areas and underdeveloped countries where breastfeeding may be the only option available. But why is it that formula is more heavily used if breast milk is significantly cheaper? Even in low and middle-income countries, the rate stayed the same for exclusive feeding (7). In one study that surveyed 650 mothers averaging age 27 on exclusive breastfeeding, the rate of breastfeeding was higher than 92%. However, after 5 days, it declined to 82%, then 44% by the end of the first month (8). This study had a higher rate of mothers who engaged in exclusive breastfeeding at 44% compared to the global statistic of 38%. The reason is that the study did not focus on socioeconomic status or diverse background, as all of the women were from the same small city in Iran. Socioeconomics plays a major role in the reason behind choosing breastfeeding or bottle feeding. For most mothers, breast milk is easier to obtain than formula, regardless of income statistics, because it is continually and naturally produced. For many, it is their only option. It is generally known that formula is expensive in most areas of the world, where breast milk costs virtually nothing. Both mother and baby benefit from breast milk, whereas only the child can benefit from the formula. Considering the financial benefits of breast milk, this may appeal more to mothers who cannot afford a luxury item such as formula, which can save potentially thousands of dollars a year that can be put towards her nutrition which doubly benefits her child. Breastfeeding is considered one of the most cost-effective health interventions for children urged (6). There is a general trend that finds a correlation between high-income areas and high percentages of breastfeeding and low-income areas with minimal breastfeeding. In a study that examined babies born in Avon, England (considered high income), and babies from Pelotas, Brazil (middle and low income), most of the children from England were raised on breast milk while in Brazil studied there were no ties between social class and breastfeeding (9). This can be considered an outlier because as mentioned in Roberts’ study, primarily focused on multiple low-income areas of Africa, exclusive breastfeeding was as low as 3% (6). Depending on the area where a woman lives, she may be persuaded into deciding how her baby should be fed. Low-income areas are often affiliated with a lack of resources and therefore little to no knowledge of the benefits of breastfeeding. Culture seemed to also be a heavy influence on their choice in feeding methods. Organizations such as the WHO and United Nations International Children’s Emergency Fund (UNICEF) have implemented education for mothers with children and have found that the sooner they educate the mother postpartum, the more likely she is to maintain breastfeeding (6). With that being said, high-income areas are provided with more opportunities for breast and bottle education, considering most births take place in a hospital which is most often abundant in designated resources to guide a person through breastfeeding.

111

Short- and Long-Term Outcomes of Breastfeeding on Children’s Mental and Physical Health

Benefits for a breastfed child Infection and disease prevention are two of the main reasons why mothers choose to breastfeed. A common disease that is prevented through breastfeeding is diarrhea. This can be prevented in the early stages by promoting healthy gut flora, which follows a baby from infancy into their adult lives. Diarrhea can be deadly in young children, as it causes around 1.34 million deaths yearly between the ages of 0 and 59 months (10). It is the second leading cause of death for this age group, considering it causes dehydration which leads to fatality. Children who were breastfed were less likely to have diarrhea (1). This is due to the antibodies provided from the milk which actively decrease gastric and intestinal problems. Breast milk has been shown to inoculate the infant gut with protective bacteria, which is useful when protecting a child from diarrhea (1). With a 6-month minimum feeding, the chances of contracting diarrhea decrease, and if the mother continues to breastfeed for more than 6 months, the chances are still reduced (10). Lodge and colleagues (11) demonstrated that asthma and allergies are also less likely to be observed in a breastfed child. Children between the ages of 5 and 18 years were screened for both food and seasonal allergies which cause rhinitis and asthma while categorizing the children based on the method in which they were fed as children (either breast- or bottle-fed). The study found small correlations between rhinitis allergy decline and asthma decline in breastfed children compared to those who were formula-fed (11). It was also discovered that breastfed children were less likely to have eczema, suggesting that breast milk may prevent rashes. Food allergies were found to have no association with breastfeeding (11). Obesity and blood pressure have a strong correlation to breastfeeding. Following the study that was broken into high and low-income observances; in England (a high-income affiliate), children had healthy ranges of BMI, and their blood pressure levels were sufficient as well (9). There was little to no correlation between BMI and low-income breastfed children (Pelota, Brazil); however, many of the children were born with low birth weights. These children mainly had healthy blood pressure levels, although the levels were comparable in the children that were breastfed compared to those who were not (9). Mental health and intelligence are two areas of study that have been tested in breastfed children. As shown in Brion’s study (9), in both low-income and high-income breastfed children, the intelligence quotient (IQ) average increased the longer the duration of feeding. IQ was higher among the children who were exclusively breastfed as babies, even in low-income areas. This is unique, considering wealthier families typically have greater access to tools to enhance and foster their children’s learning compared to those living in poverty who do not have the funding to boost and strengthen their intelligence. In a separate study run in 2015, high IQs in children fed breast milk were also shown. It was observed that children with higher IQs have a long-term, high-performance streak during testing, which lasted through adolescence into adulthood (12). This is due to the presence of long-chain polyunsaturated fatty acids which are found in breast milk. These fatty acids are beneficial in brain development. Seventeen separate studies were analyzed, and

112

all 17 studies showed that IQ was higher in children who were breastfed as babies. The average score was 3.44 points higher than in bottle-fed babies (12). Maternal benefits Breastfeeding benefits in the mother are often overshadowed by the importance of the health and well-being of the baby. Mothers strongly benefit both long-term and short-term, which is why many mothers choose to continue lactating past the recommended 6-month period. Inhibiting lactation can take days to months depending on the duration of the feeding period. Many mothers form stronger bonds with their babies as skinon-skin contact is beneficial to parent and child. Many long-term health benefits arise from breastfeeding within the first year of the child’s life. This may suggest why women choose to breastfeed children they have after the first child they chose to breastfeed. Short-term benefits include a decrease in postpartum depression, which is caused by the dramatic drop of progesterone and estrogen. On average, 1 in every 7 women who give birth suffers from postpartum depression within the first year of her baby’s life. Chowdhury (13) suggests that breastfeeding may decrease the chances of postpartum depression symptoms, but it is not always the case. Postpartum depression symptoms vary among women, and they are not always consistent in timing. Some women see symptoms days after delivery, whereas with others, it may take months for symptoms to surface (13). Chowdhury’s study also found that women were having fewer postpartum weight fluctuations upon breastfeeding; however, there were minor inconsistencies in the data that led to the conclusion that further research had to be conducted. However, many women who did breastfeed had stability rather than fluctuation. It is unknown whether women who intentionally try to lose weight after pregnancy can do so by breastfeeding (13). Another short-term benefit of lactating is the absence of menstruation, known as lactational amenorrhea. This lasts, on average, anywhere from 3 to 6 months, and it can take up to a year to regulate menstruation. This is due to the suppression of ovarian activity, as breastfeeding frequently can allow for an infant to properly suckle. Upon doing so, gonadotropinreleasing hormone is inhibited along with follicle-stimulating and luteinizing hormone. Gonadotropin secretes both folliclestimulating and luteinizing hormone. Follicle-stimulating hormone and luteinizing hormone typically trigger ovulation when not inhibited. This decreases the chance of pregnancy during the first few months of the newborn’s life; however, it is not completely preventive of pregnancy. Women can still become pregnant while nursing (13). Chowdhury’s study from 2015 showed that breast cancer was reduced by 26% in women breastfeeding their child for more than a year, compared to those who bottle-fed their children. Women who have at least one pregnancy in their lifetime are already at a decreased risk for breast cancer compared to women who do not have children at all in their lifetimes. For every 12 months of breastfeeding, the risk of breast cancer decreased by 4.3% (13). If all children globally were breastfed exclusively, it is estimated that 20,000 annual deaths due to breast cancer in mothers would be prevented (7). Chowdhury et al. (13) also showed that breastfeeding decreased the risk of

Short- and Long-Term Outcomes of Breastfeeding on Children’s Mental and Physical Health

ovarian cancer. There was a 35% reduction in ovarian cancer in women who breastfed compared to women who did not breastfeed (13). In America, more than 100 million people have Type II diabetes or are pre-diabetic, meaning they are at an increased risk of contracting Type II diabetes due to varying lifestyle habits such as diet and inadequate physical activity. Women who breastfeed reduce their risk for Type II diabetes by 32% if otherwise healthy. Each year that they continue to breastfeed reduces their risk by 9% (13). Type II diabetes is one of the leading causes of death for adult Americans. By nursing a child for a year, a woman can significantly lower her chances of becoming diabetic. Negative aspects of breastfeeding While the benefits of breastfeeding outweigh the costs, certain factors may increase risk for the newborn and the mother. Environmental aspects may cause a woman to bottle feed her baby along with her diet. Although the diet a woman must follow while nursing has fewer restrictions than a diet followed while pregnant, some women want to return to their regular eating habits. An example of a remaining limitation carried over from pregnancy is seafood and shellfish. Women must avoid eating seafood due to mercury causing problems with the baby’s nervous system. Caffeine intake, while not eliminated, should be reduced. Alcohol intake is extremely harmful to the newborn, as trace amounts can contaminate the breast milk. If a woman wants to drink alcohol, it is best to refrain from providing her child with breast milk. Children who were breastfed while the mother drank postpartum showed lower verbal IQ scores and weighed less due to duct blockages in the mammary gland due to alcohol (14). If a mother is a recreational drug user, it is also extremely dangerous to take drugs while nursing. Studies have shown that if a mother consumes marijuana 1 month postpartum, there is a risk of a decreased motor development in the infant (15). While for many it is common knowledge to avoid substance use, some mothers may be inclined to continue, leaving the baby to suffer from the consequences in the long run. Smoking while breastfeeding increases the risk of secondhand smoke exposure to the child. While it is common knowledge that smoking is a severe health hazard for the consumer, it can also be dangerous for the child by exposing them to secondhand smoke and nicotine. Studies have shown that nicotine levels are triple the amount in breast milk than in plasma. Smoking has also been shown to reduce the concentration levels of the milk and to reduce the lactation period for the mother. A mother may have to choose to quit smoking, but many who quit for their pregnancy pick it up in the months following the end of their gestational period, which introduces secondhand smoke into the environment that the mother and her child share (16). Many women choose not to breastfeed due mainly to personal disadvantages. One study suggests timing was a major factor as to why women choose bottle feeding from the start. While most women breastfeed through the transitions from colostrum to mature milk, many stop during the early stages. Only 13% of women across the United States follow the recommended guideline of 6 months of exclusive breastfeeding (2).

A study from the Netherlands asked 98 Dutch and Flemish women between the ages of 18 and 45 their verbal opinions to gain perspective as to why this group of women were against breastfeeding and chose to bottle feed instead (17). For most of the women, making the time to breastfeed was not achievable. If a woman works a full-time job outside of the home, she may not have the ability to pump milk in the middle of the day, as many office facilities do not have a designated place for women to pump. Others said they had trouble being physically present for their baby at all times that the child is hungry. When bottlefeeding, these women were able to distribute responsibility to their significant others or other helpers in their lives rather than being the designated person to do the feeding (17). Not only is time extracted from a woman, but it takes significant energy from the body to lactate and to continuously provide milk. A woman must change her caloric intake to support her health along with her child. One study suggested that women who are nursing require an additional 500 kilocalories a day. This ideal number comes from the volume of breast milk produced on average per day (780 mL) along with the energy content of milk, which is 67 kilocalories per 100 mL (18). When a woman is pregnant, her body stores an extra 19,000 to 48,000 kilocalories in her tissues as fat to save for when she begins lactation. Without consuming excess calories, the body will use this energy source as an alternative. This is a reason why weight fluctuation is possible post-pregnancy; however, breastfeeding mothers lose an average of 2 pounds per month alone from breastfeeding (18). Nutrient intake must increase, which can often be challenging to track. Women should increase their protein intake by 25 grams per day, and prenatal vitamins should still be consumed postpartum (18). Any fat- and water-soluble vitamins will be secreted into milk, which can cause a mother to be vitamin deficient. She may choose to upkeep by increasing her dosage of daily vitamins, but for many this is a challenging task (18). HIV transmission If HIV-positive, a mother will have to make the forced choice to bottle feed her baby if she wants to follow what is best for her baby. HIV or human immunodeficiency virus attacks the immune system, which inhibits its performance. HIV can lead to AIDS which is fatal. In recent years, there have not been as many studies linked to HIV and breastfeeding, as most research was done in the 90s following the spike in cases in the 80s, when HIV was first classified as an epidemic. HIV transmission more likely happens in utero before the fetus is exposed to breast milk; however, HIV can be isolated in breast milk and can pass on the virus to an otherwise healthy baby (19). A clinical trial run between the years of 1992 and 1998 gathered 425 HIV positive women. Of the 425 HIV positive women, 212 were told to breastfeed their newborns and 213 were told to bottle feed. Formula-fed babies had 44% reduced risk for contracting HIV compared to the breastfed babies (20). Even though evidence showed a decreased risk, the WHO encourages HIV-positive women to breastfeed their children, in order to provide their child nourishment of the highest nutritional value. This leaves a woman conflicted as to whether she should increase the risk of transmitting HIV to her baby and follow the WHO guidelines if the child had not already contracted it during gestation.

113

Short- and Long-Term Outcomes of Breastfeeding on Children’s Mental and Physical Health

Conclusion During the first few months of life, nutrients are essential to kickstart a healthy future for an infant. Immediately after leaving the womb, nourishment intake is key for a baby’s good health. Breast milk contains personalized nutrients for the newborn, unique and modified for the baby to thrive. The infant will obtain all it needs for survival through the milk it drinks, including water, fats, proteins, vitamins, minerals, and healthy bacteria. This sets up the digestive tract for the baby and will be its first source of protection. Unlike formula, breast milk is produced with varying nutrients important to the age of the baby. As the mother transitions from colostrum to transitional to mature milk, the contents vary depending on what an infant needs at the stage it is in to continue to the next (3). From birth until 12 months old, a baby will grow to be three times its size at birth. To ensure a child makes it to their first birthday, it is recommended that the infant consumes from the breast for a minimal 6-month period (6). Considering that we do not understand the full composition of breast milk, what a child is missing if it switches to formula remains elusive. Thus, a child misses out on some of the core components that a mother will produce that a lab cannot recreate. Babies who are formula-fed run the risk of a higher chance of mortality. Often being seen in impoverished areas of the world, millions of babies do not reach the age of 5, a majority which are fed through means other than breast (6). When a mother is given the proper care and education on breastfeeding her child, it is far more likely to live a longer and healthier life. Scientists across the world strongly encourage breastfeeding over bottle feeding, and many have made strides in bringing resources to women who need them the most. Thousands of lives could be saved globally if women were to choose the breast over the bottle (7). At this current moment in time, only 38% of infants are exclusively breastfed, although strongly urged by highly regarded organizations such as the WHO. Breastfeeding is financially beneficial and for average lactating mothers the supply is abundant. However, it seems that women with the funding to supply their children with formula are the ones that have higher rates of breastfeeding. Women in financially stable environments are far more likely to receive proper learning skills to promote nursing a child along with having constant access to resources at all times (6). Along with being financially beneficial, children are set up with a foundation of benefits that carry far into their adult lives. Many children who are breastfed are less likely to contract asthma, allergies, and common skin rashes such as eczema (11). Others were shown to have healthier BMI and healthy blood pressure levels which can prevent other health disorders in their futures such as obesity and diabetes (9). It has even been seen that IQ is increased compared to children who were bottle-fed, which often allows for the child to establish a good future for themselves whether it be academically or vocationally. One of the biggest disadvantages to formula feeding is that the mother does not benefit in the slightest compared to how she does upon lactation and breastfeeding. Breastfeeding provides mutual benefits that a bottle cannot satisfy. When a woman chooses to breastfeed for a year, she drops the risk of weight

114

fluctuation, postpartum depression, breast and ovarian cancer, and Type II diabetes (13). The absence of a menstrual period is a benefit to many, as most are focused on their infant rather than preparing for their next. Nursing mothers still have to be careful if preventing pregnancy, as lactational amenorrhea is not foolproof; however, the mother can put herself at ease knowing she can dedicate more time and energy to her child and other aspects of daily life (13). While plenty of benefits are provided for mother and child, some women make the choice to abandon the breast for the bottle. Breastfeeding requires a lot of work on the mother, and it is safer to bottle feed for some, especially if she does not have the availability to upkeep with the work it requires. For many women it is tough to follow a limited diet and a substance-free lifestyle while nursing. A mother must refrain from caffeine, alcohol, and seafood due to the harm it may cause on the child (14). If a mother is a substance user, the infant will have further repercussions if breastfed. Breast milk is tainted with whichever substance the mother uses which can delay birth weight and other developmental issues (15). A newborn can easily have high exposure to secondhand smoke when latched to its mother (16). In scenarios as such it is not ideal to breastfeed a child as it can do more harm than good. After all, long-term benefits can be overruled by a period in a child’s life when exposed to trauma or harm. With this information, it is also important to stick to bottle feeding if the mother is ill. Short-term sicknesses, such as common colds, fevers, and vomiting, will not transmit through milk; however mothers who are sick should proceed with caution as infection can be spread through droplet or airborne contact to the baby. If a mother has a chronic disease such as HIV she should proceed with caution. Two decades ago, it was advised against breastfeeding if a mother has HIV (18), yet in recent years it has been recommended to continue to breastfeed due to the nutritional value having high importance. Ultimately, it is the mother's choice, and her choice alone if she chooses to breastfeed or not. Thankfully, bottle feeding like breast has minimal risks, but, when provided with the choice between the two, if a woman is able-bodied and lactation is fruitful for her it is an applicable choice to make to better the chances of the baby having a better quality of life. If more women were taught the significance of breastfeeding, babies and mothers alike would be healthier. What once was the only option to feed a child, breastfeeding has since rapidly declined in all areas of the world. As we move toward a more sustainable future, women may choose to breastfeed to reduce their waste habits, as formula packaging contributes to landfills and pollution far more than the breast. Many women simply don’t have the time to stop what they are doing to feed their child, and many are shamed out of breastfeeding in public. If the stigma around breastfeeding was reduced, women may be more likely to engage. Women face daily decisions regarding what is best for their child. Breastfeeding is one of the simplest ways a woman can set her child straight for healthy development. It takes a strong woman to choose to provide a life for her child while sacrificing her own. When providing her infant with breast milk, a mother can put herself at ease knowing she is making the choice for her and her baby to share a rare bond, and together grow stronger.

Short- and Long-Term Outcomes of Breastfeeding on Children’s Mental and Physical Health

References 1.

Lyons KE, Ryan CA, Dempsey EM, Ross RP, Stanton C. Breast milk, a source of beneficial microbes and associated benefits for infant health. Nutrients. 2020;12(4):1039.

Martin CR, Ling P-R, Blackburn GL. Review of infant feeding: Key features of breast milk and infant formula. Nutrients. 2016;8(5):279.

Ballard O, Morrow AL. Human milk composition: nutrients and bioactive factors. Pediatr Clin North Am. 2013;60(1):49–74.

Sankar MJ, Sinha B, Chowdhury R, Bhandari N, Taneja S, Martines J, et al. Optimal breastfeeding practices and infant and child mortality: a systematic review and metaanalysis. Acta Paediatr. 2015;104(467):3–13.

Goyal RC, Banginwar AS, Ziyo F, Toweir AA. Breastfeeding practices: Positioning, attachment (latch-on) and effective suckling - A hospital-based study in Libya. J Family Community Med. 2011;18(2):74–9.

Roberts TJ, Carnahan E, Gakidou E. Can breastfeeding promote child health equity? A comprehensive analysis of breastfeeding patterns across the developing world and what we can learn from them. BMC Med. 2013;11(1):254.

Victora CG, Bahl R, Barros AJD, França GVA, Horton S, Krasevec J, et al. Breastfeeding in the 21st century: epidemiology, mechanisms, and lifelong effect. Lancet. 2016;387(10017):475–90.

Koosha A, Hashemifesharaki R, Mousavinasab N. Breastfeeding patterns and factors determining exclusive breastfeeding. Singapore Med J. 2008;49(12):1002–6.

Brion M-JA, Lawlor DA, Matijasevich A, Horta B, Anselmi L, Araújo CL, et al. What are the causal effects of breastfeeding on IQ, obesity and blood pressure? Evidence from comparing high-income with middle-income cohorts. Int J Epidemiol. 2011;40(3):670–80.

16. Napierala M, Mazela J, Merritt TA, Florek E. Tobacco smoking and breastfeeding: Effect on the lactation process, breast milk composition and infant development. A critical review. Environ Res. 2016;151:321–38. 17. Van Acker F, Bakker E. A functional assessment of the impact of advantages and disadvantages on breastfeeding attitude. Psicologica (Valencia). 2012;33(3):533–45 18. Kominiarek MA, Rajan P. Nutrition recommendations in pregnancy and lactation. Med Clin North Am. 2016;100(6):1199–215. 19. Quintanilla K. Can HIV be transmitted through breast milk? Nurs Times. 1996;92(31):35–7. 20. Nduati R, John G, Mbori-Ngacha D, Richardson B, Overbaugh J, Mwatha A, et al. Effect of breastfeeding and formula feeding on transmission of HIV-1: a randomized clinical trial. JAMA. 2000;283(9):1167–74.

10. Lamberti LM, Fischer Walker CL, Noiman A, Victora C, Black RE. Breastfeeding and the risk for diarrhea morbidity and mortality. BMC Public Health. 2011;11 Suppl 3:S15. 11. Lodge CJ, Tan DJ, Lau MXZ, Dai X, Tham R, Lowe AJ, et al. Breastfeeding and asthma and allergies: a systematic review and meta-analysis. Acta Paediatr. 2015;104(467):38–53. 12. Horta BL, Loret de Mola C, Victora CG. Breastfeeding and intelligence: a systematic review and meta-analysis. Acta Paediatr. 2015;104(467):14–9. 13. Chowdhury R, Sinha B, Sankar MJ, Taneja S, Bhandari N, Rollins N, et al. Breastfeeding and maternal health outcomes: a systematic review and meta-analysis. Acta Paediatr. 2015;104(467):96–113. 14. May PA, Hasken JM, Blankenship J, Marais A-S, Joubert B, Cloete M, et al. Breastfeeding and maternal alcohol use: Prevalence and effects on child outcomes and fetal alcohol spectrum disorders. Reprod Toxicol. 2016;63:13–21. 15. Metz TD, Borgelt LM. Marijuana use in pregnancy and while breastfeeding. Obstet Gynecol. 2018;132(5):1198– 210.

115

Scholarly Research In Progress • Vol. 5, November 2021

An Examination of the Impact of COVID-19 on Black, Latino and Asian New York City Residents and the Factors Associated with the Social Determinants of Health Stephanie Ren1†‡, Naomi Francois1†‡, and Vicki T. Sapp1 ¹Geisinger Commonwealth School of Medicine, Scranton, PA 18509 † Doctor of Medicine Program ‡ Authors contributed equally Correspondence: sren@som.geisinger.edu

Abstract New York City (NYC) has become the epicenter of the current coronavirus pandemic known as COVID-19. As the pandemic has progressed, the health disparities experienced by populations of color have become more prevalent. Secondary data sets were taken from the Centers for Disease Control and Prevention (CDC), the U.S. Census Bureau and the New York City Department of Health. Information from these sources regarding H1N1 flu, demographic information in NYC and COVID-19 data were used to highlight the populations’ risk of being disproportionately affected by diseases not unique to the current pandemic. The preliminary results currently show that although Black and Hispanic communities make up 22% and 29% of the NYC population, they make up 28% and 34% of the total COVID-19 cases, respectively. Asians make up 14% and 7% of the total coronavirus cases seen in NYC. Prior pandemics and epidemics such as the swine flu pandemic of 2009 and the global HIV/AIDS epidemic have shown that when there is a health crisis because of factors such as health insurance, language, education and poverty, minority populations are continuously disproportionately affected. These disparities prevent patients from accessing healthcare and on a larger scale are negatively impacted by health crises.

Introduction In December 2019, an atypical respiratory disease of unknown cause occurred in Wuhan, China and rapidly spread to other countries. This respiratory disease was soon discovered to be caused by a novel coronavirus named SARS-CoV-2 and the disease caused by the virus was called COVID-19 (1). As of April 13, 2020, the U.S. represented roughly 30.0% and 20.0% of the world’s COVID-19 morbidity and mortality (2). An emerging trend that is seen amongst COVID-19 cases and deaths is that they are persons of color. Racial disparities seen in the U.S. are consistent with other developed countries such as the United Kingdom (UK) with its Black Caribbean population (2). Minorities in both the UK and the U.S. comprise a large proportion of essential workers in comparison to their White counterparts and are therefore more likely to be vulnerable to the economic impact of this pandemic (2). In major cities such as Chicago, African Americans comprising 30.1% of the total population but have 45.6% of the COVID-19 cases and 56.0% of deaths (3). NYC has emerged as an epicenter of this pandemic and the Black and Latino population have been disproportionately affected. Although Black and Hispanic communities make up 22.0% and 29.0% of the NYC population, they make up 28.0% and 34.0% of the total COVID-19 cases,

116

respectively. Asians make up 14.0% of the population and 7.0% of the total cases seen in NYC (5). There are factors that leave certain racial groups vulnerable and contribute to these disparities. Health disparities are persistent differences in disease rates and health outcomes seen amongst people of different races, ethnicities, and socioeconomic statuses (6). These health disparities are also contributed to by social determinants of health. These social determinants of health are conditions in the environment that people are born, live, learn, work, play, worship, and age (7). These conditions affect a wide range of health, function, quality-of-life outcomes, and risks. Within the scope of social determinants of health, this study focuses on education, employment, language, literacy, and access to health care. The purpose of this study was to examine and to provide empirical evidence of how the social determinants of health and COVID-19 have impacted the health of Blacks, Latinos, and Asians living in NYC as well as making recommendations for public health intervention strategies.

Methods In this study, secondary data sets were taken from three different sources. These websites include the Center for Disease Control and Prevention (CDC), the U.S. Census Bureau and the NYC Department of Health. The Center for Disease Control and Prevention was used to obtain H1N1 flu related hospitalization broken down by race and ethnicity from April 15 to August 31, 2009, and from September 1, 2009, to January 26, 2010 (8). The U.S. Census Bureau provided yearly information about NYC and its five boroughs: Bronx, Brooklyn, Manhattan, Queens and Staten Island (9). Specifically, population estimates by race and ethnicity as of July 2019 were taken from this website. Other data values taken from topics such as income and poverty, education, health, families and living arrangements reflect information from 2015 to 2019. Next, the NYC Department of Health was used to look at NYC HIV/AIDS Annual Surveillance Statistics of 2018 across the five boroughs (10). The surveillance provided reported HIV diagnoses, AIDS diagnoses, the percent of individuals living with HIV and the percent of deaths related to HIV/AIDS in NYC as of December 31, 2018. This information was categorized by race and ethnicity. Finally, the NYC Department of Health website was used for COVID-19 data. Values such as case count and death rates broken down by boroughs, race and ethnicity were all provided through this website.

An Examination of the Impact of COVID-19 on Black, Latino and Asian New York City Residents

As the pandemic is still ongoing, the data is continually changing. All COVID-19 values are current as of December 13, 2020. As the pandemic is still ongoing, these numbers are subject to change.

Results Table 1 shows the demographic data taken from the U.S. Census bureau. It shows the race & Hispanic origin of individuals, the percent of persons without health insurance, languages other than English spoken at home, education level and poverty level across the five boroughs of NYC. From the table, the largest percentage of Hispanics or Latinos and Black or African American live in the Bronx, the highest percentage of Asians live in Queens, and the highest percentage of white reside in Staten Island. It should be noted that in the Bronx, there is also the largest percentage of individuals above the age of five who speak a language other than English at home, the largest percentage of persons in poverty and the lowest education percentages for people who have a high school degree or a bachelor’s degree or higher. After gathering the demographic information, information for prior pandemics and epidemics such as H1N1 influenza and HIV/AIDS were taken from the NYC Department of Health and Centers for Disease Control and Prevention as shown in Figures 1 and 2. In Figure 1, the age-adjusted 2009 H1N1 related hospitalizations rates by race and ethnicity from the Centers for Disease Control and Prevention is represented as a bar graph. Hispanic and Blacks have the highest hospitalization rates compared to their white and Asian counterparts. Next in Figure 2, data from the NYC HIV/AIDS Annual Surveillance Statistics 2018 is shown, which represents the reported HIV diagnoses, concurrent AIDS diagnoses, new AIDS diagnosis, and people living with HIV (PLWH). The highest percentage of diagnoses, death and those living with HIV are in the Black community. The second highest is followed by Hispanics and Latinos. Finally, the New York Department of Health was used to look at the current coronavirus pandemic percentage of cases and deaths distributed across boroughs by race and ethnicity as of December 13, 2020. The percentages are represented in Figure 3 and Figure 4. In Brooklyn, the Bronx, Manhattan, and Queens, Hispanic/Latino and Black/African American represent the highest percentage of cases and deaths. In Staten Island, Hispanic/Latino once again represent the highest percentage of cases and deaths, but whites are third highest. In all five boroughs, Hispanic/Latino lead in percentage of coronavirus cases and Hispanic/Latino also lead in percentage of coronavirus-related deaths in Brooklyn and Queens. For the Bronx, Manhattan, and Staten Island, Black/African Americans have the highest percentage of coronavirus-related deaths.

Table 1. Demographics of NYC as of July 2019 from U.S. Census Bureau given as percentages. For health insurance, only individuals under the age of 65 years old was provided. For language, this applies to individuals older than the age of 5 from 2015 to 2019. For education, both categories applied to individuals greater than the age of 25 years old from 2015 to 2019.

Figure 1. Age-adjusted 2009 H1N1 related hospitalization rates by race and ethnicity from Centers for Disease Control and Prevention

Figure 2. Reported HIV/AIDS diagnoses and deaths in 2018 and reported persons living with HIV (PLWH) as of December 31, 2018, in NYC according to race and ethnicity. Taken from NYC HIV/AIDS Annual Surveillance Statistics 2018

117

An Examination of the Impact of COVID-19 on Black, Latino and Asian New York City Residents

In a borough where they make up less than a fifth of the population, they account for nearly half the number of deaths. These statistics demonstrate that while the virus may not discriminate in who gets it, there is something else that is affecting treatment and recovery.

Figure 3. Percentage of coronavirus cases distribution by race and ethnicity and borough of residence as of December 13, 2020, taken from NYC Department of Health

Unfortunately, the coronavirus is not the first pandemic to show that communities of color are adversely impacted more affected by a disease than their white counterparts. In the 2009–2010 H1N1 pandemic, Blacks and Hispanics made up the majority of hospitalizations at 26.7% and 30.7% from September 2009 to January 2010 compared to 16.3% for whites and 12.5% for Asian/Pacific Islanders (13). While not specific to NYC, this data shows that Blacks and Hispanics are not only impacted in a single city or state alone, but across the U.S. as a whole. One would imagine that there might be discrepancy on a smaller scale, but even on a larger scale, the story does not change. Studies continue to show that minorities tend to have higher morbidity and mortality from both acute and chronic illnesses (14). Even end-oflife care has differences among different racial and ethnic groups. In fact, a study conducted in Seattle with 2,850 participants and 1,290 family member surveys concluded that minority race/ ethnicity of families were associated with lower family ratings of quality of dying in the ICU (15). These researchers hypothesized that lack of communication for these patients may have contributed to the decision making in the health care setting but there was no association between educational attainment and a family’s quality of end-of-life ratings (16).

Due to these disparities, patients have less trust of health care providers which in turn leads to a breakdown in communication between the patients and the providers. This was depicted in a study that hypothesized that Black and Latino communities did not have access to HIV testing and that was the reason for the lack of testing. The researchers discovered that there were in fact many readily available facilities for patients to access but there was still a lack of testing (17). The study concluded that medical mistrust may have contributed to that lack of communication and testing (17). However, even with the lack of testing, from the HIV/AIDS Surveillance of 2018, Black communities make up the largest percentage of total HIV diagnosis, concurrent AIDS diagnosis, AIDS diagnosis, deaths and people living with HIV (10). The statistics show that the Black community contribute to nearly half of the total cases for each category while Hispanics/Latinos trail second in each category (10). Currently, Black and Latino communities have been disproportionately affected by COVID-19, with more than half the cases in NYC associated with these two groups. However, the story is different for Asians. Asian Americans consistently report the lowest percentage in each category, except for HIV diagnosis where the percent increased by 36% from 2010 to 2014 (18). While Asians have a low HIV prevalence, data has shown that they represent a majority of

Figure 4. Percentage of coronavirus death distribution by race and ethnicity and borough of residence as of December 13, 2020, taken from NYC Department of Health

Discussion With over 16 million cases and more than 300,000 deaths in less than a year, the ongoing coronavirus pandemic has had devastating impacts across the U.S. (2). NYC became the epicenter of the pandemic in the early months but has since transitioned and is no longer the state with the greatest number of cases. However, in a city with over 8 million people, there have been over 300,000 cases and the results show that there are certain groups of people who have been adversely affected by the virus than others, specifically the Black and Hispanic/ Latino communities. For the current pandemic, in Brooklyn where Hispanic/Latino population is 18.9%, they made up 29.8% of the coronavirus cases and 35% of the deaths compared to their white counterparts who make up nearly half of the population, but only 28.7% of the cases and 20.7% of the deaths. This is not the only borough to show this disparity. In Manhattan, where whites alone make up 64.6% of the population and Blacks make up 17.8% of the population, they represent 18.6% of cases and 33.2% of cases and 11.2% and 41% of deaths respectively (12). 118

An Examination of the Impact of COVID-19 on Black, Latino and Asian New York City Residents

those who are undiagnosed and slip under the radar when it comes to targeted prevention efforts. According to the CDC, approximately 22% of Asians living in the U.S. are unaware of their HIV diagnosis (18). While the current statistics do not support the impact of HIV/AIDS on the Asian American community, it does not mean they are affected any less. There is limited data collection regarding the health of Asians in the U.S. as well as incomplete race reporting overall and without proper data, disparities will continue to exist. Several studies that covered previous pandemics and epidemics describe how minorities have suffered at the hands of racial and ethnic disparities in health care, yet little has changed (13, 14, 19, 20). The data shared here could be used to inform policy makers so a public health intervention can be enacted. It is well documented that although there have been improvements in disease prevention, and medical care overall, health disparities still exist (15, 21, 22). Economically disadvantaged racial and ethnic minorities as well as persons with low socioeconomic status experience these disparities. Health insurance Individuals with a lack of health insurance have multiple barriers to health services which lead to a myriad of issues. Without health insurance, people are faced with the burden of the high cost of health care, which can lead to financial burdens. The fear of this financial burden can lead individuals to delay seeking out services and may be faced with preventable hospitalizations due to this lack of preventive services. This gap demonstrates the dangers of being without health care because individuals who delay seeking out care due to the lack of insurance are less likely to receive the care they need, are more likely to be diagnosed with illnesses much later and are more likely to die prematurely. Poverty Individuals who live in poverty are especially vulnerable not only during this pandemic but any other pandemic that can occur. Persons in poverty do not have the privilege to quarantine because they most likely make up the essential worker demographic. In addition to having to report to work every day for fear of being fired or not being able to provide for themselves or their family, they are not able to stay off public transportation. When living in an urban area like NYC, public transportation is not only the easiest, but often the only option for people to commute to work, which significantly exposes them to encountering the coronavirus. People living in poverty also do not have access to healthy foods, and the areas these individuals live in are often food deserts. Without access to nutritious foods, individuals have higher rates of comorbidities such as hypertension and diabetes, making them even more susceptible to dying from COVID-19. The determination of who meets the threshold for poverty is based on total family income and how many individuals make up that household. If a total family income is less than that threshold set up by the Census Bureau, all the individuals in that household are living in poverty. If the household income is equal to or greater than the poverty threshold, the family is not considered to be living in poverty. If a family of five makes $32,000 but the threshold is $31,275, that $725 difference means this family is not considered to be in poverty (23). The

family does not meet the criteria for being in poverty, which means they do not meet the criteria for receiving federal aid, including assistance with access to food and health care. The sheer lack of resources available to persons in poverty not only makes them vulnerable to contracting the virus but dying from it as well (24). Language Health literacy is the ability for individuals to understand, find and use information provided to them so that they can make informed decisions about their health care. For people to do this, information must be provided to them in a language they can understand. For those who do speak English, this is still a challenging task. Not everyone is educated in the same way. A medical student can differentiate a bacterial infection versus a viral one, but to the average individual, that is a foreign concept. How do we begin to explain that the coronavirus is a RNA virus instead of a DNA virus? How do we tell people that viruses are not able to be treated with antibiotics? Do people know what antibiotics are? These are all questions we must consider for English speakers. Difficulties are presented when the person is not an English speaker. In NYC, the Bronx has the highest percentage of those who speak a language other than English at home and Hispanics and Latinos make up 56.4% of the borough’s population (24). Hispanics and Latinos also make up the largest percentage of COVID-19 cases in the Bronx. We must consider: Is this because they are not aware of the information provided by the CDC on how to protect themselves against the virus? Are pamphlets provided in languages to educate the community on how to stop the spread, on how to wash their hands properly or how to get tested for the virus? It should be important to note that Spanish is the second most spoken language in NYC, yet Hispanics and Latinos who mostly speak Spanish are still affected. Most signs that are in a different language are usually in Spanish but are those translations accurate and in a simple enough language that the average person can understand. These are all considerations that need to be considered because if not, we are contributing to health disparities and widening the gap to health care. To truly provide good health care, it is important to make information readily available to those who need it and in a language they can understand. Education Education is also an important social determinant of health. Those who are not well educated, have fewer opportunities to obtain a job that would allow them access to good health insurance when they are sick and the ability to work from home during quarantine. As many know, the best way to avoid becoming ill is to stay at home and reduce interactions with the public, but those who are essential workers are not afforded this luxury. Essential workers such as those who work in restaurants, grocery stores and transportation are potentially exposed to those who have the virus daily. Many times, those who are exposed and might have the virus are not able to take the day off to go receive a COVID-19 test or take time off to recover because they might be the sole provider of their family. Difference in education status can lead to a plethora of problems that one has to consider.

119

An Examination of the Impact of COVID-19 on Black, Latino and Asian New York City Residents

There is also a bias in our society toward those who are not well educated. For example, those who are in a respected profession such as the medical field, corporate business, or law are more likely to be taken seriously when they bring up a concern compared to those who work a menial job that does not require a degree. Simply because people assume that if you are well educated, what you are saying must be fact. The MD or PhD that comes after your name automatically makes you more credible than someone who doesn’t have the title. It is easy to disregard a comment from someone who cooks for a living simply because you think they don’t understand what is happening. There is the presumption that your education defines how truthful and believable you are. This might explain why in Manhattan, where 60.8% of individuals who have a bachelor's degree or higher have fewer deaths from COVID-19 than the Bronx where only 19.8% of individuals have a bachelor's degree or higher (5). It is possible that those who are coming for treatment in Manhattan are being taken more seriously and treated better than those in the Bronx. In order to provide equal treatment, one has to be aware of this potential bias and of this factor that can contribute to a disproportionate number of cases. A possible way to try and overcome this is to increase public engagement. Using community-based participatory research allows community members to not be treated as objects of policy but as individuals taking part in the governance of their own society (25). This can be done by identifying a community partner, identify decisions requiring community input, preparing content for panel deliberation, and facilitating panels involving the community. This ensures that the research is relevant for and supported by the community members as well as strengthening the feasibility and effectiveness of the study (26). For example, a study found that Pacific Islander adults reported three times the national rate of major depression, yet rarely seek mental health services due to the incompatibility with their cultural conceptions of mental illness. To address this unmet need in this population, researchers in partnership with the community decided that storytelling would be crucial to the intervention process (27). Interventions that are specific and culturally relevant to the population that is most in need is important in addressing health disparities. The goal here was to identify targeted strategies to improve health equity, so engaging the community being affected by these disparities may be a way to remedy that. Another strategy could be to incorporate “big data” science to address minority health disparities. Big data is data that is generated in high volume and variety that accumulates quickly. For example, patient records can be considered a big data problem because it contains millions of records (27). The volume, variety, variability of big data can bring benefits to health and health care as it has with sectors of the economy. With this large pool of data, a guided extraction of information and knowledge can be gained (27). A major opportunity here is to incorporate a standardized collection and input of race/ ethnicity, SES, access to health care, education, and other social determinants of health measures in the database systems (28). Standardizing the way this data is collected will allow all groups to be included, ensuring that no groups are excluded and ensures targeted ways to improve quality care.

120

A recommendation when using big data is it is important to build trust with the community (27). Medical mistrust exists due to the history of brutality against minorities in research. Serving minority populations need to be done in a respectful and beneficial way, which can be potentially done through a collaboration between minority-serving institutions and research-intensive institutions (30). More importantly, ensuring that researchers report back to the communities to provide results and next steps. Transparency and a continuity in partnership with community members should be a common practice. Limitations In this study, we examined the impact of COVID-19 on the social determinants of health of Blacks, Latinos, and Asians living in NYC to provide some racial and culturally relevant recommendations. As the coronavirus is still an ongoing pandemic, the data is changing every day, making it difficult to say definitively what the difference in case totals and deaths are for each race and ethnicity. Additionally, the statistics that are provided only include those who have been tested. It does not include those who may have gotten the virus but were asymptomatic and did not receive a test. There is also an inconsistency between definitions when retrieving secondary data. For example, for the U.S. Census bureau, Asian Americans and Pacific Islanders were separated into two different categories whereas for the NYC Department of Health, Pacific Islanders were grouped with Asian Americans, which leads to an inaccurate representation of how COVID-19 has impacted different races and ethnicities (19, 24). Finally, in terms of CBPR, there are few studies that follow up to see how effective public health interventions are (26, 27). It is important not only to make suggestions, but to see, whether the suggestions have a positive effect in the community. These limitations all impact our analysis, but with more research and better parameters, it is possible to eliminate them.

Conclusion Time and time again, history tends to repeat itself. Whether it be the H1N1 pandemic, HIV/AIDS epidemic or COVID-19, minority populations are constantly the victims of a health crisis due to the social determinants of health. From their education to their poverty level and even their language barriers, they are disproportionately affected and are dying because health disparities that lead to inadequate care are not being highlighted. We need to step back and see the bigger picture before we can begin to tackle the gap in health care. In order to decrease health disparities, research studies and policies must be inclusive and nonbiased. Through the implementation of our suggested strategies, there can be success in expanding the foundation necessary to achieve health equity.

Acknowledgments We would like to thank Vicki T. Sapp, PhD, for all her support, mentorship and feedback throughout this research project.

Disclosures We have no disclosures to address.

An Examination of the Impact of COVID-19 on Black, Latino and Asian New York City Residents

References 1.

Yuki K, Fujiogi M, Koutsogiannaki S. COVID-19 pathophysiology: A review. Clin Immunol. 2020;215:108427.

Anyane-Yeboa A, Sato T, Sakuraba A. Racial disparities in COVID-19 deaths reveal harsh truths about structural inequality in America. J Intern Med. 2020;288(4):479-80.

Kim SJ, Bostwick W. Social Vulnerability and Racial Inequality in COVID-19 Deaths in Chicago. Health Educ Behav. 2020;47(4):509-13.

Abrams EM, Szefler SJ. COVID-19 and the impact of social determinants of health. Lancet Respir Med. 2020;8(7):659661. doi:10.1016/S2213-2600(20)30234-4

U.S. Census Bureau (2019, July 01). Quick Facts. Retrieved from [https://www.census.gov/quickfacts/fact/table/US/ PST045219].

Cogburn CD. Culture, Race, and Health: Implications for Racial Inequities and Population Health. Milbank Q. 2019;97(3):736-61.

Office of Disease Prevention and Health Promotion. (2018). Social Determinants of Health. Healthy People 2020. U.S. Department of Health and Human Services. https://www.healthypeople.gov/2020/topics-objectives/ topic/social-determinants-of-health

Information on 2009 H1N1 impact by Race and Ethnicity. (n.d.). Retrieved December 28, 2020, from https://www. cdc.gov/h1n1flu/race_ethnicity_qa.html U.S. Census Bureau (2019, July 01). Quick Facts: New York City, New York. Retrieved from [https://www.census.gov/ quickfacts/newyorkcitynewyork].

10. HIV/AIDS Surveillance and Epidemiology Reports. (n.d.). Retrieved December 28, 2020, from https://www1.nyc. gov/site/doh/data/data-sets/hiv-aids-surveillance-andepidemiology-reports.page 11. Andermann A; CLEAR Collaboration. Taking action on the social determinants of health in clinical practice: a framework for health professionals. CMAJ. 2016;188(17-18):E474-E483. doi:10.1503/cmaj.160177 12. New York State Department of Health. (2020). New York State Department of Health COVID-19 Tracker. Retrieved from https://covid19tracker.health.ny.gov/views/NYSCOVID19-Tracker/NYSDOHCOVID-19Tracker-Fatalities? %3Aembed=yes&%3Atoolbar=no&%3Atabs=n 13. Chandrasekhar R, Sloan C, Mitchel E, Ndi D, Alden N, Thomas A, . . . Lindegren ML. (2017). Social determinants of influenza hospitalization in the United States. Influenza and Other Respiratory Viruses, 11(6), 479-488. doi:10.1111/ irv.12483 14. Duarte R, Lönnroth K, Carvalho C, et al. Tuberculosis, social determinants and co-morbidities (including HIV). Pulmonology. 2018;24(2):115-119. doi:10.1016/j. rppnen.2017.11.003

15. Coats H, Downey L, Sharma RK, Curtis JR, Engelberg RA. (2018). Quality of Communication and Trust in Patients With Serious Illness: An Exploratory Study of the Relationships of Race/Ethnicity, Socioeconomic Status, and Religiosity. Journal of Pain and Symptom Management, 56(4). doi:10.1016/j.jpainsymman.2018.07.005 16. Lee JJ, Long AC, Curtis JR, Engelberg RA. (2016). The Influence of Race/Ethnicity and Education on Family Ratings of the Quality of Dying in the ICU. Journal of Pain and Symptom Management, 51(1), 9-16. doi:10.1016/j. jpainsymman.2015.08.008 17. Crawford ND, Dean T, Rivera AV, Guffey T, Amesty S, Rudolph A, . . . Fuller CM. (2016). Pharmacy Intervention to Improve HIV Testing Uptake Using a Comprehensive Health Screening Approach. Public Health Reports, 131(1_ suppl), 139-146. doi:10.1177/00333549161310s116 18. Lee JJ, Zhou Y. (2019) Facilitators and barriers to HIV testing among Asians in the United States: a systematic review, AIDS Care, 31:2, 141-152, DOI: 10.1080/09540121.2018.1533231 19. Abrams EM, Szefler SJ. COVID-19 and the impact of social determinants of health. Lancet Respir Med. 2020;8(7):659661. doi:10.1016/S2213-2600(20)30234-4 20. Bambra C, Riordan R, Ford J, Matthews F. The COVID-19 pandemic and health inequalities. J Epidemiol Community Health. 2020;74(11):964-968. doi:10.1136/jech-2020214401 21. Andermann A; CLEAR Collaboration. Taking action on the social determinants of health in clinical practice: a framework for health professionals. CMAJ. 2016;188(17-18):E474-E483. doi:10.1503/cmaj.160177 22. Crawford ND, Dean T, Rivera AV, Guffey T, Amesty S, Rudolph A, . . . Fuller CM. (2016). Pharmacy Intervention to Improve HIV Testing Uptake Using a Comprehensive Health Screening Approach. Public Health Reports, 131(1_ suppl), 139-146. doi:10.1177/00333549161310s116 23. U.S. Census Bureau (2020, February 20). Income and Poverty in the United States:2019. Retrieved from [https:// www.census.gov/topics/income-poverty/poverty.html] 24. Zylke JW, Bauchner H. Mortality and Morbidity: The Measure of a Pandemic. JAMA. Published online July 01, 2020. doi:10.1001/jama.2020.11761 25. Institute of Medicine (US) Committee on the Review and Assessment of the NIH’s Strategic Research Plan and Budget to Reduce and Ultimately Eliminate Health Disparities; Thomson GE, Mitchell F, Williams MB, editors. Examining the Health Disparities Research Plan of the National Institutes of Health: Unfinished Business. Washington (DC): National Academies Press (US); 2006. 2, Health Disparities: Concepts, Measurements, and Understanding. Available from: https://www.ncbi.nlm.nih. gov/books/NBK57052/

121

An Examination of the Impact of COVID-19 on Black, Latino and Asian New York City Residents

26. Morain SR. Engaging Community Members to Eradicate Health Disparities. Am J Public Health. 2020 Feb;110(2):143-144. doi: 10.2105/AJPH.2019.305493. PMID: 31913668; PMCID: PMC6951371. 27. Subica AM, Brown BJ. Addressing Health Disparities Through Deliberative Methods: Citizens' Panels for Health Equity. Am J Public Health. 2020 Feb;110(2):166-173. doi: 10.2105/AJPH.2019.305450. Epub 2019 Dec 19. PMID: 31855474; PMCID: PMC6951363. 28. Zhang X, Pérez-Stable EJ, Bourne PE, Peprah E, Duru OK, Breen N, Berrigan D, Wood F, Jackson JS, Wong DWS, Denny J. Big Data Science: Opportunities and Challenges to Address Minority Health and Health Disparities in the 21st Century. Ethn Dis. 2017 Apr 20;27(2):95-106. doi: 10.18865/ed.27.2.95. PMID: 28439179; PMCID: PMC5398183. 29. Sanchez-Pinto LN, Luo Y, Churpek MM. Big Data and Data Science in Critical Care. Chest. 2018 Nov;154(5):12391248. doi: 10.1016/j.chest.2018.04.037. Epub 2018 May 9. PMID: 29752973; PMCID: PMC6224705. 30. Yearby R. Structural Racism and Health Disparities: Reconfiguring the Social Determinants of Health Framework to Include the Root Cause. J Law Med Ethics. 2020;48(3):518-526. doi:10.1177/1073110520958876

122

Scholarly Research In Progress • Vol. 5, November 2021

The Natural History of Genu Valgum in the Pediatric Obese Patient Mark Mandel1†, Brandi Woo1†, Benjamin Wheatley2, Amanda Young2, Peter Fabricant3, and Mark Seeley2 ¹Geisinger Commonwealth School of Medicine, Scranton, PA 18509 ²Geisinger Medical Center, Danville, PA 17822 ³Hospital for Special Surgery, New York, NY 10065 † Doctor of Medicine Program Correspondence: mmandel@som.geisinger.edu

Abstract Background: Pediatric obesity is a growing epidemic in the United States. There is little information on the impact this increasing incidence of obesity in the younger population has on the developing musculoskeletal system. Genu valgum is a common lower extremity malalignment that is thought to remodel as a child grows. Our hypothesis is that obesity may hinder the natural remodeling potential of these patients, which will lead to a greater clinical impact and cost of care. Methods: A retrospective chart review of patients with genu valgum diagnoses over a span of 19 years was performed in a pediatric population. Inclusion criteria was diagnosis before age 18. Exclusion criteria was misdiagnosis, deceased before 18, or not evaluated by our institution’s physicians. Demographic information was collected, and the clinical impact of the disease was estimated by using surrogate markers. Comparisons were analyzed using Fisher’s exact or chi-square tests for categorical data and the Wilcoxon test for non-parametric continuous data. Results: Out of 604 charts reviewed, 554 children met inclusion criteria for the study. Thirty-nine patients required surgical intervention. A majority of the patients who had surgery were obese (63.6%). A comparison of body mass index (BMI) categories showed a higher number of visits in the overweight/ obese categories as compared to the normal weight group (11.6% vs. 2.4% of patients with greater than five visits, p = 0.0068). The number of surgeries was higher in the overweight/ obese group compared to the healthy weight group (p = 0.0268), and a higher number of patients had surgery in the overweight/obese categories than the healthy weight category (6.9% vs. 1.9%, p = 0.0188). Conclusion: Historically, pediatric genu valgum runs an indolent course without requiring surgical intervention. Our findings suggest that the clinical impact of this disease changes with increased bodyweight. The relationship between obesity and the severity of genu valgum in the pediatric population appears to be a prognostic indicator of a more challenging disease course complicated by an increased likelihood for surgical intervention. Level of Evidence: Level II – Prognostic Studies – Investigating the Effect of a Patient Characteristic on the Outcome of Disease, Retrospective Study

Introduction The increased incidence of obesity in the pediatric population is accompanied by an increase in incidence of lower extremity pathology and decreased physical activity levels, all of which

predispose to musculoskeletal pain and bone/joint dysfunction later in life (1–3). Recent studies have examined the relationship between increased body weight and musculoskeletal pathology in children; however, there is currently a dearth of research that investigates the relationship between obesity and the trajectory of genu valgum in children. Genu valgum, also called “knockknee” deformity, is a common lower extremity abnormality seen in children that makes the knees appear like they are touching while the ankles are apart (4). For children ages 2 to 7 years old, genu valgum is a physiologic process and most cases spontaneously resolve. If the genu valgum persists beyond age 8 or is extreme, surgical intervention may be required to allow for normal growth and alignment of the legs (4, 5). While most idiopathic cases of genu valgum follow a mild course, our clinical observations suggest that genu valgum in an overweight or obese child may present a more challenging trajectory for the patient. A study by Jankowicz-Szymanska and Mikaloajcyzk reported a higher prevalence of genu valgum in obese children compared to non-obese children (6). Another study by Walker et al. found direct correlations between obesity and genu valgum by examining radiographs and calculating intermalleolar distances (7). This study aimed to gain a deeper understanding of the relationship between obesity and genu valgum in the pediatric population. We hypothesized that children with increased bodyweight, as reflected in their body mass index (BMI), and genu valgum are more likely to experience long-term sequalae and the need for surgical intervention.

Methods Upon Institutional Review Board (IRB) approval, we performed a retrospective chart review of genu valgum diagnoses over a span of 19 years. Patients with a genu valgum diagnosis were determined by using the Epic WebI database searching for the ICD-9 code 736.41 (Genu valgum, acquired) and the ICD-10 codes M21.069 (Valgus deformity, not elsewhere classified) and Q74.1 (Congenital malformation of the knee). Patients were eligible for inclusion if the diagnosis of genu valgum was made before age 18. Exclusion criteria included patients who were either misdiagnosed, not seen within the hospital system, or deceased before age 18. Charts were reviewed for relevant demographic and clinical information. The initial date and age of when a patient was seen for knee concerns/complaints were determined from the patients’ charts. The date of initial diagnosis was documented, what the diagnosis/cause of the knee complaint was (e.g., congenital or acquired genu valgum, any preceding

123

The Natural History of Genu Valgum in the Pediatric Obese Patient

Table 2. Comparing demographic and clinical characteristics for those that had surgery specifically to correct their genu valgum to those who did not.

Results Table 1. Summary table for all patients with genu valgum diagnosis.

trauma injury), and who diagnosed the knee problem (PCP or orthopaedics). Additional data collected included how many orthopaedic visits were made for knee complaints, whether knee X-rays were taken, how many knee surgeries were performed, the patient’s age at initial surgery, type of surgery, the date of the last orthopaedic visit for a knee complaint, and length of follow-up. All demographic and clinical characteristics are fully described by reoperation status. Continuous variables are summarized using means and standard deviations for normally distributed data or median and interquartile range (IQR) for non-normally distributed data. Categorical variables are summarized using frequency and percentages. Demographic and clinical characteristics are compared using Fisher’s exact or chi-square tests for categorical data and the Wilcoxon test for nonparametric continuous data. Logistic regression techniques were used to calculate the area under the curve (AUC) for the receiver operating characteristics (ROC) curve. All analyses were performed using SAS v9.4 (SAS Institute Inc., Cary, NC, USA).

124

Our study reviewed 604 charts from pediatric patients covering visits from March 1999 to December 2018. Of the 604 charts, 554 children met inclusion criteria for the study. A total of 50 patients (8.2%) were excluded due to misdiagnosis, not being seen by an OMITTED physician, or deceased before age 18. Demographic and clinical characteristics are described for all patients with genu valgum (Table 1). The mean age at initial encounter for all patients diagnosed with genu valgum is 6.14 years with a standard deviation of 4.43. The sample was 49.8% female and 85% white. The BMI percentile categories were a combined underweight and healthy weight category (48.6%), an overweight category (15%), and an obese category (36.4%). A total of 69 (12.4%) of these patients had knee surgeries, of which 39 (7.0%) patients had surgery specifically to correct their valgus deformity and 30 (5.4%) patients had knee surgeries for other reasons with their valgus being a significant contributor. Of the patients who had surgery for their genu valgum, 63.6% of them were obese. The mean age at initial surgery was 12.50 years with a standard deviation of 2.89. The majority of those who had surgery for genu valgum had hemiepiphysiodesis surgery (74.4%). Results are reported in Table 2, which is a comparison of those who had surgery for genu valgum versus those who did not have surgery for genu valgum. Those who had surgery had significantly more X-rays (97.4% v. 40.6%, P < 0.0001) and significantly more visits for knee complaints (>5 visits for knee complaints, 84.6% v. 8.2%. p <0 .0001).

The Natural History of Genu Valgum in the Pediatric Obese Patient

The BMI percentiles were divided into two groups, an underweight/healthy weight group and an overweight/obese group. A comparison of the two groups shows a higher number of visits for knee complaints in the overweight/obese as compared to the healthy weight group (over 5 visits, 11.6% v. 2.4%, p=0.0068). The number of knee surgeries is significantly higher in the overweight and obese categories compared to healthy weight (p=0.0268). There was not a significant difference in cause of diagnosis between the two groups. The prevalence of surgery for correction of genu valgum was significantly higher in the overweight/obese group (p=0.019). The ROC curves were designed to assess how well age and BMI separately predict the outcome of whether or not surgery is required for genu valgum in these children (Figure 1). The area under the curve (AUC) for age at initial encounter is 0.7972 (95%CI: 0.7429-0.8516). In Figure 2, the ROC curve for BMI is displayed, and the AUC is 0.8969 (95%CI: 0.8532-0.9406).

Discussion

Figure 1. This receiver operating characteristic (ROC) curve was designed to assess how well age at initial encounter for genu valgum can independently predict the outcome of whether or not surgery is required. Area under the curve = 0.792 (95% CI 0.7429 – 0.8516)

While the negative effects of excess body weight on the musculoskeletal system are well known, to our knowledge the burden of disease and clinical impact of genu valgum in overweight and obese children has not yet been rigorously studied. Our study utilized historical chart and anthropometric data to better understand the epidemiology and clinical trajectory of genu valgum in overweight and obese children compared to their normal weight counterparts. A review of previous studies was helpful in establishing original associations between obesity and genu valgum. Several studies have reported that obesity can cause deformities in the lower limbs which specifically include genu valgum (2, 8, 9). A paper published by Wearing et al. helped to detail the biomechanical basis of the pathology, making note that children have a larger amount of collagen during their growth phase, which is susceptible to remodeling in response to higher load and therefore susceptible to plastic deformation (10). The chronically increased load due to their obesity can exacerbate this process, causing significant deformation and excess angulation leading to genu valgum. In addition to this potential cause, Espandar et al. suggest that ligamentous laxity may also be the cause of genu valgum, and this is congruent with the soft-tissue deformation theory of the disease (9). Combined with our findings, these investigations suggest that obesity may exacerbate genu valgum in pediatric patients and therefore increase the need for corrective surgery as well as increase the burden of disease. However, a more detailed investigation into the timing and progression of genu valgum as it relates to the onset of obesity in these children would be helpful in determining possible causality.

Figure 2. This receiver operating characteristic (ROC) curve was designed to assess how well BMI can predict whether the outcome of surgical intervention was required. Area under the curve = 0.8969 (95% CI 0.8532 – 0.9406)

The median age of genu valgum diagnosis in our study was 4 with IQR of 2-9 which is consistent with the known epidemiology of genu valgum, as it is a physiologic process in children between the ages of 2 and 7 years old (5). Pathologic genu valgum is more common in older children, as the physiologic valgum may persist or worsen and this is similarly reflected in our data as the median age for those who required surgery was 12.9 years old (IQR 10.8:14.1) (9,11).

125

The Natural History of Genu Valgum in the Pediatric Obese Patient

The increased median age of those who required surgery was consistent with our ROC curve, which showed that age could be used to predict whether a child would require surgery. In our study, we used the number of visits for knee complaints and/or pain, the number of X-rays taken, and the requirement for surgical intervention as the surrogate markers to track the severity and burden of the disease course in these patients. We did not use radiographic images to track intermalleolar distances, as the purpose of the study was to investigate how the trajectory of the disease differs between the overweight and obese children compared to healthy weight children. Due to the mild course of genu valgum and relatively rare need for corrective surgery, it is easy to observe deviations from this path. It is important to make note of the cause of the diagnosis and compare these among weight groups, as a traumatic cause of genu valgum could potentially lead to a more severe genu valgum presentation than an idiopathic cause and thus be a confounder. The cause of diagnosis was recorded as either non-traumatic/congenital deformity, acquired, or traumatic secondary to motor vehicle accident, fractures, iatrogenic, or missing. Our findings indicated that there was no difference among the three BMI percentile categories for cause of diagnosis. As was expected due to the naturally mild course of genu valgum, most patients in our sample did not require surgery. However, 39 of the 554 patients required surgery specifically to correct their valgus deformity while 30 other patients had knee surgery where their valgum was a contributing factor. The group that received surgery also had a significantly higher percentage of having X-rays taken, which is evidence of both the greater direct cost of care due to medical bills and the greater indirect cost of care such as time off work for the parent to take the child for the imaging. The percentage of surgical patients who were obese was nearly double that of the percentage of obese patients in the sample. This is an interesting finding and agrees with Wearing et al. finding that obesity in their sample of genu valgum patients was twice the rate of obesity in the general pediatric population (10). Once again, this further helps us understand the association between obesity and an increase in the severity of the disease and likelihood of surgery. Our ROC curves for BMI and surgery also showed that BMI can be used to predict surgical intervention for these patients. The interrelatedness between genu valgum and obesity is apparent, but the causal directionality is still slightly unclear. The increased burden of disease in overweight children with genu valgum was additionally evidenced by the higher number of visits and longer length of follow-up. This data helps estimate the clinical impact of the disease on these patients and secondarily the direct and indirect costs of health care. Although we did not measure and compare radiographic data, our study findings are able to serve as a link between prior quantitative genu valgum research findings and the clinical consequences and outcomes experienced by these pediatric patients. The single most important finding of the study is that the overweight and obese group had a significantly higher number of patients undergoing surgery for their genu valgum than the underweight and healthy weight group. This is a potentially actionable finding and should help guide treatment via early weight-loss intervention programs, which may help these pediatric patients

126

decrease the likelihood of requiring surgery. In addition to the increased burden of disease for the patient undergoing surgery and postoperative rehabilitation, their parents/guardians must undergo additional financial burden in the form of missing work to accompany the patient at these visits. Our study has several limitations, which include the retrospective nature of the study, the reliance on ICD coding and EMR data input, and that the study was not designed with power or sample size considerations. We believe that the effect of the chart review based on ICD diagnosis likely underrepresents the number of visits and severity of genu valgum in the overweight and obese patients, as musculoskeletal problems in this group tend to involve multiple joint locations. Along these lines, if a child presented with shoulder pain, it may not have been recorded that they also had knee complaints. This theory is backed by several papers previously published on musculoskeletal health and musculoskeletal pain in overweight and obese children that show high rates of simultaneous joint pain complaints (1, 3). With prevention being the best medicine, our study findings reinforce the previous research on the association of genu valgum and obesity and underscore the necessity of early weight-loss intervention in children to prevent the development of more severe musculoskeletal pathology. Our findings also suggest that although genu valgum appears likely to resolve in healthy weight patients, we should attempt to identify lower extremity issues earlier and consider being more aggressive with treatment and surveillance of the deformity in overweight and obese patients. Additionally, our study bolsters the importance of our colleagues’ prior findings on the anatomic and measurement-based relationship between genu valgum and obesity in children, but adds the additional dimension of clinical impact. Furthermore, we demonstrated that genu valgum in an obese child is associated with an increased risk for more clinic visits, more imaging studies, and the potential need for surgery. This increase in clinical impact and burden of disease further increases the cost of care for the patient and on the health care system. Further research is needed within this area, and we hope that our findings can facilitate a multicenter and multidisciplinary approach to better understand the effect of obesity on the prognosis of children with genu valgum diagnoses, with the ultimate goal of improving outcomes in these children and preventing the progression of disease into adulthood and its associated long-term comorbidities.

References 1.

Smith SM, Sumar B, Dixon KA. Musculoskeletal pain in overweight and obese children. Int J Obes. 2014 Jan;38(1):11–5.

Taylor ED, Theim KR, Mirch MC, Ghorbani S, TanofskyKraff M, Adler-Wailes DC, et al. Orthopedic complications of overweight in children and adolescents. Pediatrics. 2006 Jun;117(6):2167–74.

Bell LM, Curran JA, Byrne S, Roby H, Suriano K, Jones TW, et al. High incidence of obesity co-morbidities in young children: A cross-sectional study: Obesity co-morbidities in young children. Journal of Paediatrics and Child Health. 2011 Dec;47(12):911–7.

The Natural History of Genu Valgum in the Pediatric Obese Patient

Kids Health. Available from: https://m.kidshealth.org/ Nemours/en/parents/az-genu-valgum.html

Hatch, Daniel. OrthoBullets. Available from: https://www. orthobullets.com/pediatrics/4052/genu-valgum-knockedknees

Jankowicz-Szymanska A, Mikolajczyk E. Genu Valgum and Flat Feet in Children With Healthy and Excessive Body Weight: Pediatric Physical Therapy. 2016;28(2):200–6.

Walker JL, Hosseinzadeh P, White H, Murr K, Milbrandt TA, Talwalkar VJ, et al. Idiopathic Genu Valgum and Its Association With Obesity in Children and Adolescents: Journal of Pediatric Orthopaedics. 2019 Aug;39(7):347–52.

Jannini SN, Dória-Filho U, Damiani D, Silva CAA. Musculoskeletal pain in obese adolescents. J Pediatr (Rio J). 2011 Aug;87(4):329–35.

Espandar R, Mortazavi SM-J, Baghdadi T. Angular deformities of the lower limb in children. Asian J Sports Med. 2010 Mar;1(1):46–53.

10. Wearing SC, Hennig EM, Byrne NM, Steele JR, Hills AP. Musculoskeletal disorders associated with obesity: a biomechanical perspective. Obesity Reviews. 2006 Aug;7(3):239–50. 11. Ciaccia MCC, Pinto CN, Golfieri F da C, Machado TF, Lozano LL, Silva JMS, et al. Prevalence of genu valgum in public elementary schools in the city of Santos (SP), Brazil. Rev Paul Pediatr. 2017 Dec;35(4):443–7.

127

Scholarly Research In Progress • Vol. 5, November 2021

Review of Selected Contemporary Treatments for Ischemic Stroke Gwendolene K. Conteh1* ¹Geisinger Commonwealth School of Medicine, Scranton, PA 18509 *Master of Biomedical Sciences Program Correspondence: gconteh@som.geisinger.edu

Abstract Stroke is considered a serious medical emergency, and it is crucial to get immediate medical treatment. Early detection can limit brain damage and complications. Ischemic stroke (IS) is caused by atherosclerosis, which is the hardening of the arteries. The accumulation of fats and cholesterol on the wall of an artery causes narrowing that eventually leads to blockage or interruption of blood flow through that artery, thus reducing blood flow to tissues and organ perfusion. A reduction or a blockage in blood flow to the brain can cause neurological problems such as stroke or ischemic stroke. This review critically analyzes the safety and effectiveness of novel treatments for ischemic stroke. The literature search was conducted via PubMed and Google Scholar. Review articles show that statin therapy helps in preventing additional strokes from occurring in patients that had small-vessel occlusion, largeartery atherosclerosis, and strokes that are caused by multiple reasons. Results showed that high dosages of simvastatin proved to be the most effective treatment in improving stroke prognosis and ineffective in preventing mortality. Mesenchymal stem cell (MSC) therapy might also help by improving the damaged tissue of IS through various mechanisms, but research still needs to be done on the neurogenesis and angiogenesis pathways. MSC was successful in clinical use, but it still needs to be conducted in a large sample size. Focusing on the effectiveness of the safety of MSC and the timing and optimal dosage are the main challenges in its clinical use and need to be conducted before taking it into clinical trials. Monoclonal antibody (mAb) therapy showed a positive benefit in increasing the neuronal repair, and regeneration in the brain of the animal models. Further research needs to be done on human stroke patients in assessing the benefit of mAb therapy.

Introduction Stroke is defined as a sudden loss of brain function that occurs due to the interruption of blood flow to the brain. This causes the brain cells in the affected area to die (1). Stroke is the third leading cause of death, and it is also the leading cause of disability in the United States. Each year, 15 million people suffer from stroke worldwide (2). There are three types of strokes, which include: hemorrhagic (ICH), transient ischemic attack (TIA), and ischemic stroke (IS). ICH occurs when a blood vessel is ruptured, causing bleeding inside the brain. TIA, also known as a mini stroke, causes a disruption of blood flow to the brain for a short period of time. Having an episode of TIA indicates that it is more likely someone will have a stroke in the future. IS takes place when there is a blockage in an artery that causes a blood clot. Blood flow is being stopped from traveling to different parts of the brain, and that portion of the brain will become deprived of oxygen (1).

128

Ischemic stroke is the most common type among the three and accounts for 87% of all strokes (2). While a stroke can affect people of all ages, the elderly are more susceptible. As an individual's age increases the arteries steadily begin to narrow and develop plaque (3). Also at risk are individuals who have underlying health problems such as diabetes, high blood pressure, and high cholesterol (4). Individuals who show symptoms of stroke might experience a sudden loss of strength or numbness in the arms or legs. They can also have severe and unusual headaches, dizziness, and difficulty speaking (5). Though strokes have many underlying risk factors and symptoms, strokes can be prevented and treated. Detecting and receiving treatment for IS early reduces the risk of having permanent brain damage. The sooner the signs and symptoms of a stroke are diagnosed, the more effective the potential treatment will be. This review focuses on ischemic stroke, the most common type of stroke which typically has a better chance of survival. If left untreated, IS can then lead to hemorrhagic strokes, which are the deadliest and are highly difficult to treat (6). It is important to know about stroke and the effects it can have on the individual’s life and family. Educating people on the risk factors of developing stroke can help individuals to take precautions that can save many lives and reduce the number of cases per year. Having new treatments in combination with current treatments can help treat IS faster and better. A literature review was conducted to determine which of the novel treatments — statin therapy, mesenchymal stem cell therapy, or monoclonal antibody therapy — is most effective for slowing and treating ischemic stroke.

Methods A literature review of review and primary articles were conducted using electronic databases such as Google Scholar and PubMed to retrieve studies on novel treatments for ischemic stroke published between 2018 and 2020. Terms such as “ischemic stroke,” “current treatments,” “new treatments,” “novel treatments,” “emerging treatments,” and “upcoming treatments” were used to find the articles. Articles were selected according to the following eligibility criteria: randomized control trial, clinical trial, peer-reviewed, systematic review, meta-analysis, free full text, general geographical location, English language. Only articles that addressed current and new treatments for ischemic stroke were reviewed. Articles that were published in 2017 and earlier were excluded, as were articles that did not include or address the terms ischemic stroke and treatments. The first search produced 25,800 total articles from both Google Scholar and PubMed. After excluding articles that were not relevant to the topic, 3,555 remained for full-text reviewed articles. The articles were narrowed down to

Review of Selected Contemporary Treatments for Ischemic Stroke

25 using the inclusion and exclusion criteria. Ten articles were selected for final analysis.

Results The use of statin therapy in the treatment of ischemic stroke High-level serum cholesterol leads to a buildup of fatty particles in the arteries. When the arteries are blocked, it can increase the risk of cardiovascular disease and stroke. Statins are drugs that can help remove plaque from arteries. Statins are prescribed by doctors to help reduce cholesterol levels and help prevent patients from having a stroke or heart attack (7). Statins have been shown to be highly effective in reducing the risk of cardiovascular disease (CVD) up to 10% in primary prevention and 5% in secondary prevention (8). Statins help reduce the risk of CVD and prevent the onset of the disease (primary prevention). They are also effective in slowing the progression of CVD, which in turn reduces the risk of morbidity and mortality. Based on the CVD benefit, the use of statin drugs is being expanded to the treatment of other diseases. Since statins are important in helping stroke patients, it has become a vital breakthrough in stroke prevention. Clinical trials have shown positive results in analyzing the safety and effectiveness of statin treatment, which leads to using statin treatment more often on stroke patients (9). In addition to clinical trials, cohort studies have been conducted to help in understanding how the treatments can be used in everyday clinical practice (9). The authors Vitturi and Gagliardi conducted a prospective cohort study of patients who were diagnosed with ischemic stroke. Participants who were 18 years and older and had their first ischemic stroke were included. They excluded participants that were followed up less than 24 months and using other lipid-lowering drugs. Participants were grouped into four different categories: non-statin, simvastatin 20 mg per day, simvastatin 40 mg per day, and high-potency statin groups (atorvastatin 40 mg per day or rosuvastatin 10 mg per day). The high-potency statin groups were those groups that were given atorvastatin 40 mg per day or rosuvastatin 10 mg per day that was expected to reduce the LDL cholesterol level by more than 50%. Participants were followed up by phone interviews for 2 years or until they died. Patients that were excluded based on their death were determined through searches from the National Registry of Death to detect (9). The primary outcome was the post-stroke functional limitations, which were assessed using the modified Rankin scale (mRS). After the IS participant was admitted to the hospital, each of their post-stroke functional limitations was assessed during days 7, 30, 6 months, and 2 years. Participants who had an mRS score ≥3 were defined as unfavorable outcomes, while an mRS score ≤2 was considered a favorable outcome. Secondary outcomes were those participants that had recurrent strokes such as hemorrhagic or ischemic stroke, any major cardiovascular disease, and cause of mortality. The results showed no significant difference in NIH Stroke Scale/Score (NIHSS), atrial fibrillation, smoking, dyslipidemia, etc. Patients who were given a statin had a lower risk of developing CVD (OR=0.3; 0.1-0.7; p=0.01) (9). There were not any differences found in the mortality rate of using the statin drug (OR=0.6; 0.1–2.3; p=0.73) (9). Three hundred and seven participants had a favorable functional outcome with a medium mRS score of 2.

Results also showed that taking a statin at an early stage of IS resulted in a favorable neurofunctional outcome and appeared to prevent recurrence of stroke (p < 0.01). Participants who terminated their treatment or did not fully respond to the treatments had poorer functional outcomes and were more likely to have another stroke (p < 0.01). Participants who did not receive the statin drug had a higher risk of a second stroke and worse functional outcomes (p < 0.01). Based on the secondary prevention description of their study, statin therapy was more beneficial to those participants that were given the simvastatin 40 mg than those that were given the simvastatin 20 mg and high-intensity statin. Overall, patients that received the simvastatin 40 mg and the highintensity drug had the best functional recovery in a year after they had a stroke (p < 0.05) (9). Statin therapy was also successful in preventing another stroke from occurring in patients that had small-vessel occlusion, large-artery atherosclerosis, and strokes caused by multiple reasons (9). In summary, statin use was associated with a significant beneficial effect in patients with ischemic stroke such as reducing the risk of stroke recurrence and having better functional outcomes. Without the use of statin, patients did not see any change in their functional outcomes (9, 16). Mesenchymal stem cells (MSC) and their potential for neuroregeneration in ischemic stroke Mesenchymal stem cells help in repairing different disorders and diseases such as kidney, neurodegenerative, autoimmune diseases, and graft-versus-host diseases. They help replace damaged cells and wound remodeling (10). Mesenchymal stem cells (MSC) are cells that are found in the bone marrow that are important in making and repairing the tissues in bones and cartilage. MSC is considered a novel therapeutic agent in treating diseases and injuries because it helps in the process of self-renewal of tissues and can benefit in healing cells (11). MSC works by expressing a variety of cytokines and chemokines that help in the repair of the damaged tissue thereby restoring the tissue back to normal, and this can help in reducing inflammation. Based on its benefits and unique characteristics in treating different pathologies, MSC is most frequently used in regenerative medicine and is now in the study of treating brain repair after ischemic stroke because of its potential for neuroregeneration (11, 10). A study by Zhang et al. explained how mesenchymal stem cells are transplanted into damaged tissues, examined the mechanisms used to prevent and treat damaged cells, and provided a summary of other clinical trials that are using MSC to treat ischemic stroke (11). MSC releases paracrine factors that activate the modulation of microglia to differentiate into the glial cells and neurons to repair the damaged tissues. Studies showed that the paracrine action of MSC showed an improvement in neurotrophic effects and that paracrine signaling might be the main reason IS victims recover (11). Other mechanisms for MSC include cell migration, immunomodulation, neuroprotection, angiogenesis, and neural circuit reconstruction were also discussed by Zhang et al. (11). MSC is effective in treating cell migration by crossing the bloodbrain barrier in migrating to the damaged cells. Mesenchymal stem cells migrate to the damaged brain of IS through the response of chemokines signals such as the MIP-1a and the 129

Review of Selected Contemporary Treatments for Ischemic Stroke

MCP-1; also, the increase of neurogenin-1 can increase the effectiveness of MSC, which in turn improves the effectiveness of the engraftment in the IS area (11). The immunomodulation and neuroprotective effects of MSC lower injuries that may occur hours or days after IS. MSC subdues the activation of the microglia cell and detain the death of neurons. It aids immunosuppression by the regulation of cytokine expression. MSC inhibits apoptosis by promoting endogenous repair in the neuroprotection of the damaged brain. It also increases the expression of the neurotrophic factors and this, in turn, is a major benefit in treating IS (11). In angiogenesis, growth factors secreted by MSC such as IGF-1, BDNF, and bFGF increased angiogenesis in the ischemic core and the border zone. Angiogenesis also increases the flow of blood that is found in the brain tissue and this, in turn, benefits the endogenous neurogenesis and miRNA-210 promotes angiogenesis. In neural circuit reconstruction, MSC improves the neurological function after IS by the process of axonal plasticity and endogenous neurogenesis. When the axonal sprouting increases or enhances the connection between different cerebral areas, it repairs the connections between the neurons (11). In the study by Zhang et al. found that the middle cerebral artery occlusion (MCAO) model rats showed significant recovery when they tested with both allogeneic and heterogeneic MSC in the early preclinical trial. Twentyone percent of the bone marrow-derived MSC (BMSC) were administered in the middle cerebral artery territory after intracarotid arterial injection. The treated rats showed functional improvement compared with the control rats (11). These same methods were then used with ischemic stroke rats using the intravenous injection of human MSC and showed similar results such as significant functional recovery results as shown with the treated (MCAO) models rats. The MSC treatment increases the level of growth factors or receptors of BDNF and NGF in the ischemic tissue of the rats and a reduced level of apoptosis in the penumbral area. In clinical trials by Oy et al., autologous BMSC was introduced to 30 patients with ischemic stroke. They were divided into two groups (MSC and control). The MSC groups had 5 participants and received an intravenous infusion of 1 × 108 cells while the control group had 25 participants and did not receive MSC. BI and mRS were administered at regular intervals for up to 1 year after the onset of stroke. During their 12-month follow-up period, the Barthel index (p = 0.011, 0.017, and 0.115 at 3, 6, and 12 months, respectively) and modified Rankin Scale (mRS) score (p = 0.076, 0.171, and 0.286 at 3, 6, and 12 months, respectively) of the MSC group improved consistently compared to the control group (11, 14). This means that MSC transplantation is safe and showed an improvement in the patient’s neurological condition. A randomized controlled trial (RCT) by Assia et al. assessed the efficacy, feasibility, and safety of intravenous autologous bone marrow derived MSC in subacute ischemic stroke patients. Their study included 31 patients. Sixteen patients were treated with autologous MSC therapy, while 15 patients were placed in the control group. The MSC group showed significant improvements in motor National Institutes of Health stroke scale (NIHSS) (p = 0.004), motor Fugl-Meyer score (p = 0.028),

130

and task-related fMRI activity in MI-4a (p = 0.031) in the primary motor cortex during the 2-year follow-up (11, 15). This shows that MSC treatment for subacute ischemic stroke was safe and feasible. Motor performance and task-related MI activity results suggest that MSC improved functional recovery via sensorimotor neuroplasticity (15). Monoclonal antibody (mAb): an upcoming therapy in treating acute ischemic stroke (AIS) patients Monoclonal antibodies (mAb) are used to treat cancer by interacting with specific targets of the cancer cells by bringing the T cells closer to cancer cells and this, in turn, kills the cancer cells (12). Since mAb has successfully been able to treat cancer patients, mAb might be a possible treatment for IS patients. The monoclonal antibody is a substitute antibody that restores, mimics, and enhances the immune system. It recognizes a particular protein form and is useful in producing large quantities of an antibody from a single B-cell clone and is highly specific (12). Monoclonal antibody therapy is an emerging treatment that may be able to help in treating the stroke outcomes of AIS patients outside of the 4.5-hour therapeutic window. Monoclonal antibodies can be given to a larger number of patients since they have a lifespan of days to weeks. Once the AIS occurs in the patient, anti-inflammatory and proinflammatory cytokines are then activated (13). This is where mAb will be given to help stop the receptors, and pathways that are caused by the inflammatory cytokines (13). Monoclonal antibody therapy aids in neuronal repair and axonal growth by binding to receptors and surface markers; when this happens, it blocks the cells that inhibit neuronal cell growth (13). Blocking pro-inflammatory cytokine, and ion channels, and neurotransmitter receptors, enhancing anti-inflammatory cytokines and activating growth factors help mAb to increase the neuronal repair and regeneration to manage and treat the outcomes of AIS patients. Having a life span of more than an hour makes it beneficial and more effective in treatment options. Also, mAb was effective in reducing the infarct volume and neuronal performance in the animal model that had middle cerebral artery occlusion (13). Monoclonal antibody therapy is administered against the tolllike receptor-4 (TLR4) and adhesion molecules (α-4 integrin) to stop the signaling molecules of the cascades to help reduce inflammation in MCAO mice. MCAO mice that were treated with a mAb against the TLR4 had a reduced volume of inaction and brain swelling compared to the MCAO mice that did not receive the treatment. Also, similar patterns were administered in human models using the natalizumab antibody which formed against leukocyte adhesion molecules, specifically against α-4 integrin. The natalizumab antibody was not successful as there was no significant reduction of infarct volume (13). Other studies showed that mAb was directed against an ICAM-1 in the embolic model of stroke followed by thrombolysis with tissue plasminogen activator (tPA) was conducted in rabbits, and results showed a decrease in the infiltration of inflammation which then leads to the decrease in neurological damage (13). Since mAb has been effective on experimental stroke models, it is certain to conduct further clinical research to assess the efficacy of human stroke patients (13).

Review of Selected Contemporary Treatments for Ischemic Stroke

Discussion Statin therapy is important in secondary prevention therapy in treating stroke patients. In the study by Vitturi et al., participants that were given statin had a major benefit. The recurrence of stroke risk was reduced, and they also had better functional outcomes (9). According to this article and other previous studies, participants who were not given statin didn't have any better outcome; instead, they were at a higher risk of getting another stroke (9). Even though there is no optimal dosage of statin treatment for treating IS, results showed that administering a high dosage of simvastatin (40 mg) was effective in improving stroke prognosis. Participants who had the high-intensity statin did not show any major improvement and some reasons might be because of the expensive cost of those drugs which will allow the patient to stop or limit the amount of taking it because they cannot afford it (9). Both simvastatin 40 mg and high-intensity statins have proven to have a better benefit in IS patients and because of this, it might be administered to those patients that had a severe case. In short, statins are effective in preventing CVD in stroke patients. According to the authors Vitturi et al., since their studies were conducted in a hospital setting, and single-centered, instead of a community setting they encountered a higher risk of bias. They also involved all the consecutive cases and did not have any restrictions as to who to be admitted to the hospital. The benefit of MSC in the treatment of other diseases has been an interest in seeing how it will benefit IS. MSC improves the damaged tissue of IS through various mechanisms. The authors did not discuss any limitations for their studies, but they talked about further research that needed to be done to study the pathways of angiogenesis and neurogenesis. The study supported that MSC transplantation is safe and effective in patients with cerebral recovery after ischemic stroke. Also, future studies should focus on large clinical trials, the safety and effective benefit for MSC, optimal dosages for MSC transplantation, and observe any adverse events in the order they can be used in clinical trials (11). On the other hand, mAb therapy has the potential to increase neuronal repair and regeneration in the brain of animal models. Therefore, there should be further research in assessing the efficacy of mAb on human stroke patients (13).

Having a specific mAb that targets the signaling pathways of stroke will unveil a better possible treatment in stroke therapy. However, even though mAb was effective in experimental stroke models, clinical research is needed to be done on human patients to see the benefit of this therapy (13). All of the treatments help improve the prognosis of IS, but further research is needed. Finding an effective mechanism for targeting the damaged cells of IS will be a greater benefit in treating it.

Acknowledgments Thanks to the Geisinger Commonwealth School of Medicine Library for assistance with procuring articles for this review. Thanks to Reema Persad-Clem, PhD, Chari Cohen, DrPH, Christine Rittenhouse, Solange Nsang, and Jennifer Boardman, PhD, for guiding me in this process.

Disclosures The author has nothing to disclose.

References 1.

Stroke [Internet]. Mayo Clinic. Mayo Foundation for Medical Education and Research; 2020 [cited 2020Dec18]. Available from: https://www.mayoclinic. org/diseases-conditions/stroke/symptoms-causes/syc20350113

Ischemic Stroke [Internet]. Strokecenter.org. [cited 2020 Dec 18]. Available from: http://www.strokecenter.org/ patients/about-stroke/ischemic-stroke/

Stroke Statistics [Internet]. Strokecenter.org. [cited 2020 Dec 18]. Available from: http://www.strokecenter.org/ patients/about-stroke/stroke-statistics/

Stroke [Internet]. Nih.gov. [cited 2020 Dec 18]. Available from: https://www.nhlbi.nih.gov/health-topics/stroke

Ischemic Stroke [Internet]. Cedars-sinai.org. [cited 2020 Dec 18]. Available from: https://www.cedars-sinai.org/ health-library/diseases-and-conditions/i/ischemic-stroke. html

Stroke: MedlinePlus medical encyclopedia. (n.d.). MedlinePlus - Health Information from the National Library of Medicine. https://medlineplus.gov/ency/ article/000726.htm

Side effects of cholesterol-lowering statin drugs [Internet]. Webmd.com. [cited 2020 Dec 18]. Available from: https:// www.webmd.com/cholesterol-management/side-effectsof-statin-drugs

Pinal-Fernandez I, Casal-Dominguez M, Mammen AL. Statins: pros and cons. Med Clin (Barc). 2018;150(10):398– 402.

Vitturi BK, Gagliardi RJ. Effects of statin therapy on outcomes of ischemic stroke: a real-world experience in Brazil. Arq Neuropsiquiatr. 2020;78(8):461–7.

Conclusion Ischemic stroke is a leading cause of neurological disability and requires fast treatment. Finding the right treatment that minimizes the long-term ischemic stroke will limit the number of cases per year. This review critically analyzes the safety and effectiveness of novel treatments for ischemic stroke. Statin therapy helps treat IS patients by improving functional performances, preventing recurrent stroke, and CVD events in patients. More studies are needed to verify the effect of statins in the stroke community. Also, MSC therapy is known to be safe and effective. Although it is still under research to test on a large number of patients, it might be a new treatment for the recovery of the neurological problems that are encountered in IS patients (11).

131

Review of Selected Contemporary Treatments for Ischemic Stroke

10. Saeedi P, Halabian R, Imani Fooladi AA. A revealing review of mesenchymal stem cells therapy, clinical perspectives, and Modification strategies. Stem Cell Investig. 2019;6:34. 11. Li J, Zhang Q, Wang W, Lin F, Wang S, Zhao J. Mesenchymal stem cell therapy for ischemic stroke: A look into treatment mechanism and therapeutic potential. J Neurol [Internet]. 2020; Available from: http://dx.doi.org/10.1007/s00415020-10138-5 12. Monoclonal Antibodies [Internet]. Cancer.gov. 2019 [cited 2020 Dec 18]. Available from: https://www.cancer. gov/about-cancer/treatment/types/immunotherapy/ monoclonal-antibodies 13. Woods D, Jiang Q, Chu X-P. Monoclonal antibody as an emerging therapy for acute ischemic stroke. Int J Physiol Pathophysiol Pharmacol. 2020;12(4):95–106. 14. Bang OY, Lee JS, Lee PH, Lee G (2005) Autologous mesenchymal stem cell transplantation in stroke patients. Ann Neurol 57(6):874–882. https://doi.org/10.1002/ ana.20501) 15. Jaillard A, Hommel M, Moisan A, Zeffiro TA, Favre-Wiki IM, Barbieux-Guillot M, Vadot W, Marcel S, Lamalle L, Grand S, Detante O (2020) Autologous mesenchymal stem cells improve motor recovery in subacute ischemic stroke: a randomized clinical trial. Trans Stroke Res. https://doi. org/10.1007/s12975-020- 00787-z 16. Zhao W, Xiao ZJ, Zhao SP (2019, July 1) The benefits and risks of Statin therapy in ischemic stroke: A review of the literature. Neurol India. (2019, July 1). Neurology India: Free full text articles from Neurol India. https:// neurologyindia.com/article.asp?issn=0028-3886;year=20 19;volume=67;issue=4;spage=983;epage=992;aulast=Zh ao#ref77

132

Scholarly Research In Progress • Vol. 5, November 2021

The Role of Socioeconomic Factors in Influencing Tuberculosis Rates: A Comparison of New York State and New York City 2007-2016 Raskirth P. Singh1*‡, Rachael O. Oluwasanmi1*, Smita S. Bajgain1*, Matthew M. Bradley1*, Michael I. Abdool1*, and Kylar J. Harvey¹*‡ ¹Geisinger Commonwealth School of Medicine, Scranton, PA 18509 *Master of Biomedical Sciences Program ‡ Authors contributed equally Correspondence: rsingh04@som.geisinger.edu

Abstract Background: The chronic and preventable lung disease tuberculosis (TB) is caused by a bacterial species known as Mycobacterium tuberculosis. TB is a contagious disease that may lead to death if left untreated. Adverse health outcomes from TB range from excessive coughing to weight loss, and it is estimated to cost the United States healthcare system $123.4 million per year. TB is a preventable disease that disproportionately affects individuals in a low economic status; however, TB infection rates have decreased in low socioeconomic status (SES) populations between the years 2007 and 2016 without a known cause. Methods: Through a secondary analysis of the New York City Health Tuberculosis Annual Report, three SES factors (poverty rate, average median income, and percent of the population holding health insurance between 18 and 64 years of age) were explored as contributing factors to the decline of TB infection rates within the state of New York. We evaluated correlations between TB infections and SES factors of New York State (excluding New York City), the five boroughs of New York City, and New York City as a whole. Analysis of the data using a Pearson's product-moment correlation & Spearman's rank correlation found statistically significant relationships between each of the three SES factors and the rate of TB in New York State. Results: Each borough of New York City had at least one statistically significant correlation between TB rates and a SES factor. There was a significant correlation between the rate of TB and the poverty rate in the Bronx (r (4) = -0.76, p < 0.10), New York State (r (4) = -0.79, p < 0.01), Richmond (r (4) = -0.74, p < 0.10), and Queens (r (4) = -0.58, p < 0.10); but not in the other boroughs and New York City. The rate of TB and the median income in the borough of Kings (r (4) = -0.78, p < 0.01), Manhattan (r (4) = -0.61, p < 0.10), and New York State (r (4) = -0.64, p < 0.10) correlated significantly, but not in the other boroughs and New York City. Lastly, there was a significant correlation between the rate of TB and the health insured population aged 18-64 in the borough of Kings (r (4) = -0.63, p < 0.10), Manhattan (r (4) = -0.71, p < 0.10), New York state (r (4) = -0.64, p < 0.10), and Richmond (r (4) = -0.777, p < 0.01); but the other boroughs and New York City presented an insignificant relationship. Conclusion: Our study's findings suggest that SES factors influence TB infections rates, though each geographical area we looked at varied in which SES factors affect the rate of TB. Our study indicates that each community has different factors that

may contribute to preventable diseases. Our study provides a framework for future research and policymakers to investigate workable solutions to reducing avoidable diseases in low-SES populations by making policies that help a specific community rather than blanket policies covering a large geographical area.

Introduction Tuberculosis (TB) is a lung disease caused by Mycobacterium tuberculosis and accounts for 10.4 million new infections annually (1). TB may also affect other parts of the body, including the kidneys, spine, or brain (2). If a person with TB is not treated, they could be at increased risk for death (2). The disease causes symptoms of coughing, weakness, and weight loss (2). TB can spread person-to-person through coughing/ sneezing via air particles as a vector (2) According to the World Health Organization (WHO), about one-quarter of the world population is infected with M. tuberculosis and at risk of disease development (1). Difficulties in treatment have also arisen in the setting of multi-drug-resistant forms of the disease (1). In the last several decades, however, the United States (U.S.) has successfully reduced the prevalence of TB, with rates at an all-time low of 2019 (3). This success in the U.S. has been attributed to finding and treating infected patients and contact trace those potentially exposed to reduce secondary infection (3). Annual data supplied by the CDC (Centers for Disease Control and Prevention) that significantly higher rates of TB have been reported in several large cities such as New York (5). In the 2018 State and City TB Report, the CDC said that eight of the nine major cities tracked for significant TB rates were above the national average infection rate of 2.8 per 100,000 (5). For example, a notable finding was that Chicago, Philadelphia, and New York City were above the national average infection rate and above their respective state rates (5). This data supplies a rationale for considering that SES may play a role in the ability to be successfully treated for a TB infection. The lower SES factors such as poverty diminish the quality of life. These three cities have substantially lower SES populations respective to their state. Lower SES populations in the US continue to have higher rates of TB infections, possibly indicative of certain SES factors influencing TB infection rates (6). Previous literature has shown that people of low SES disproportionally have a greater risk of developing TB (7). We examined three socioeconomic factors — poverty rate, average median income, and percent of the population with health insurance from 18 to 65 years of age — to determine their potential influence on the decline

133

The Role of Socioeconomic Factors in Influencing Tuberculosis Rates

of TB infection rates in New York. These factors were chosen because they improved over the years 2007 to 2016. A secondary analysis was also performed to test the correlation of these factors with TB rates in New York and its boroughs. If addressed, changes in these SES factors could lead to increased funding and research, improved identification of infection and treatment, and realization of favorable health outcomes.

where data was only available for the years 2007 to 2016. This made the analysis more sensitive to detecting correlations and limited Type-2 errors during analysis. A 95% confidence interval was calculated for each test. Data tables were generated from JASP and copied into Microsoft Word. The individual graphs were generated from JASP and then compiled into figures using Origin Pro 2021 graphing and analysis.

Methods

Results

Procedure

Table 1 shows a statistically significant negative correlation and association between the poverty rate and TB in Bronx, New York. The correlation between the rate of TB per 100,000 and Bronx-poverty rates showed a -0.76 r-value and a ρ-value of 0.011 in the Pearson's correlation analysis, indicative of a statistically significant, moderate negative correlation. Regarding Spearman's correlation, the ρ value was -0.69 with a ρ-value of 0.026, meaning a moderate negative association between the poverty rates and the rate of TB per 100,000 of Bronx, New York. The rate of TB per 100,000 compared with median income portrayed an r-value of -0.04. This r-value is quite close to zero, so there may be little to no association between the two variables. Regarding Spearman's correlation, the ρ value was -0.14, while the ρ-values for both Spearman's ρ and Pearson's r value were significantly greater than 0.1. Lastly, the rate of TB per 100,000 versus health insured between 1864 portrayed a -0.5 r-value in Pearson correlation with a ρ-value of 0.142. Regarding Spearman's correlation, the ρ value was -0.71 with a ρ-value of 0.022, meaning a statistically significant moderate negative association (Table 1 and Figure 1).

A secondary analysis of the New York City Health Tuberculosis Annual Reports between 2007 and 2016 from the Department of Health and Mental Hygiene (DOHMH) was conducted to describe the relationship of TB and SES factors (8). These annual reports supply intimate surveillance data, summaries of program activities, and highlights. Rates of TB are based on the New York State Department of Health Bureau of Tuberculosis Control (BTBC) extrapolated from the United States Census Bureau interpolated intercensal population estimates from 2007 to 2016 (8). Data was collected for New York State TB rates, along with the rates of TB for New York City and its boroughs spanning from 2007 to 2016 using the New York City Annual Reports. The data mentions Richmond County, which was the old name for Staten Island borough. The data also says Kings County, which is referring to Brooklyn borough. The data utilizes these county names interchangeably with their respective borough. SES data on percent poverty, median income, and health insurance status for those aged 18 to 64 were collected from several sources. Data on poverty in each of the boroughs of NYC (New York City) and New York State was collected from the New York State Community Health Indicator Reports (CHRIS) (9). Additional information collected from the CHRIS database included median income and percent of those health-insured aged 18-64 for New York State and the boroughs of New York City (9). SES and TB rate data for New York City was obtained from the United States Census Bureau American Community Survey 1-year estimates in the years 2007 to 2016 using the filters "New York City" and "SES variable or TB rate" (10). Data of those with health insurance between the ages of 18 and 64 for all of New York City was derived from the New York City Department of Health EpiQuery (11). Data analysis Collected correlation data and monotonic association data were analyzed using an open-source statistics software from the University of Amsterdam called Jeffrey's Amazing Statistics Program (JASP). Each SES factor within a region was compared to the TB rate within that region to detect significant correlations. A Pearson's product-moment correlation analysis was used to examine if individual SES factors had a linear correlation to the Rates of TB. Then, a Spearman's rank correlation coefficient was calculated to figure out the statistical dependence of TB rates on SES factors. For both Spearman's rank correlation and Pearson's product-moment correlation, a ρ-value was calculated to find if the data obtained could reject the null hypothesis of the variables having no effect on each other. The alpha-value adopted in the study was 0.10 due to the 10-year limit within the CHRIS database for New York,

134

In Table 2, the TB rate and poverty rates of Kings, New York, have an r-value of -0.2, a ρ-value of 0.585, a ρ value of -0.09, and a ρ-value of 0.811. This suggests there was no correlation or association between the two variables. The r-value for the rate of TB and the median income for Kings, New York, was -0.78 with a ρ-value of 0.001, showing there is a statistically significant strong-negative correlation between these two variables. Spearman's ρ was found to be -0.89, with a ρ-value of 0.001 suggesting a strong association between the two variables. For the rate of TB and those who are health insured between the ages of 18-64, the r-value reported was -0.63 with a ρ-value of 0.039, signifying a statistically significant negative correlation. The Spearman's ρ correlation was also shown to supply information on whether the variables were connected. Lastly, for the rate of TB and health insured, the ρ value obtained was -0.66 along with a ρ-value of 0.039, showing a statistically significant association between TB rates and the percentage of the population insured (Table 2 and Figure 2). Manhattan's correlation data comparing the rate of TB with each SES factor are shown in Table 3 and Figure 3. The rate of TB versus poverty rate produced an r-value of -0.1 and a ρ value of 0.03, suggesting there is no correlation between the two. The rate of TB compared with median income showed a moderate negative correlation with an r-value of -0.61 and a moderate negative association indicated with a ρ value of -0.6. The comparison of the rate of TB and median income had a ρ-value of 0.63 and a ρ-value of 0.073 for the Pearson correlation and Spearman's ρ, respectively, indicating a moderate statistical significance. The rate of TB compared with the percentage of the population who carry health insurance between 18-64 showed a moderate negative correlation via Pearson's r-value

The Role of Socioeconomic Factors in Influencing Tuberculosis Rates

of -0.71 and a strong association indicated with a Spearman's ρ of -0.83. The percentage of the population of health insured to the rates of TB was shown to have a moderate statistical significance with a ρ-value of 0.924 regarding their correlation and a strong statistical significance with a ρ-value of 0.003 about their monotonic association (Table 3 and Figure 3). Table 4 refers to the correlation and association data for New York City's rate of TB with each SES factor. The r-value for the rate of TB against poverty was reported to be 0.092 with a Spearman's ρ value of -0.01. This indicates no correlation between these variables in New York City along with ρ-values greater than 0.10, meaning we cannot reject our null hypothesis. For the rate of TB and median income, the r-value was -0.04 with a Spearman's ρ value of 0.104, suggesting there is no correlation or monotonic association between the rate of TB and median income. The r-value for the rate of TB and health insured in New York City was 0.109 and a ρ value of -0.07, indicating no correlation or monotonic association between the two variables (Table 4). The results for New York State are shown in Figure 4 and Table 5. For the two variables, rate of TB and poverty rate, the r-value was found to be -0.79 with a ρ-value of 0.007, which signifies a strong negative correlation that is statistically significant. The ρ value for TB and poverty rate of -0.58 is closer to -1, showing a moderate negative monotonic relationship with moderate statistical significance. For the rate of TB and median income, the r-value was -0.64, with a ρ-value of 0.047 indicating a moderate correlation that is statistically significant. A Spearman's ρ of -0.85 with a ρ-value of 0.002 meaning there is a statistically substantial strong negative correlation. For the rate of TB and health insured, the r-value was -0.64, with a ρ-value of 0.046 showing a statistically significant moderate negative correlation. The Spearman's ρ for TB and health insured was -0.83 along with a ρ-value of 0.003, shows a strong, negative, monotonic relationship (Table 5 and Figure 4). In the county of Richmond, New York, the rate of TB versus poverty rate portrayed a statistically significant correlation with a Pearson r-value of -0.704 and a ρ-value of 0.023. Spearman's ρ was deduced to be -0.74 and a ρ-value of 0.014, indicating a moderate negative monotonic relationship. The rate of TB compared with median income was not found to have a statistically significant correlation or monotonic relationship, with ρ-values greater than 0.10 for both Pearson's r-value and Spearman's ρ value. The rate of TB versus health insured between the ages of 18-64 showed a statistically significant strong correlation with an r-value of -0.777 and ρ-value of 0.008 along with a statistically significant, strong, negative, relationship shown by a ρ value of -0.93 and ρ-value <0.001 (Table 6 and Figure 5). Comparing the rate of TB versus poverty rate within Queens, New York, produced an r-value of -0.58 with a ρ-value of 0.082 and a Spearman's ρ value of -0.53 with a ρ-value of 0.112, indicating a statistically significant moderate correlation, but no association with each other. Examining median income, a statistically significant correlation was not found. For those insured from the ages of 18 to 64, there was no statistically significant correlation or association found with values for Pearson's r and Spearman's ρ, both over 0.10 (Table 7 and Figure 6).

Discussion In the U.S., surveillance of the preventable disease TB has been underway since 1953, yet we still have a TB endemic 68 years later (13). To this day, there has not been a statistically significant decrease in deaths from TB infections since 2003 (14). As a preventable disease, it's alarming how the death rate for the U.S. has not had a statistically significant decline in 18 years. People of low SES in the U.S. are disproportionally affected by TB (15). It is often generalized in literature that low SES leads to a greater likelihood of development of TB; however, many studies have not broken down which socioeconomic factors may contribute more to the rates of TB. Identifying the socioeconomic factors that most significantly influence the incidence of TB may improve municipalities to develop better public policies that will help decrease TB incidence down to the zip code. In this study, we used a Pearson and Spearman correlation analysis to determine if the three socioeconomic factors, age-adjusted poverty rate, adjusted median income, and the percentage of adults 18 to 64 years of age holding health insurance, influence TB incidence rates. Based on the data analyzed, it isn't easy to summate all three socioeconomic factors and their correlations for each location we analyzed. Correlations between TB infection rates per 100,000 people and the investigated SES factors varied in their association and strength depending on the geographical area being examined. For example, we found that Kings, New York had a statistically significant negative correlation of TB incidence with median income and the percentage of adults 18 to 64 years of age who hold health insurance. In contrast, Queens, New York, did not have a statistically significant correlation in either. In fact, Queens, New York, only had a statistically significant correlation of TB incidence with poverty rate, while Kings, New York, did not. Our results indicate that there are differences in what influences the incidence of TB depending on the geographical area where cities may have boroughs and townships that have different needs than others within the same city limits. This variation between communities that are within the same cities most likely comes from public boundaries of municipalities. These boundaries have differing zoning, social support, funding, and other factors which may negatively contribute to the social disparities seen with TB incidence (16). Geographic boundaries do not have to be defined; they can be created through social dynamics where communities may be determined by factors such as poverty, race. A previous report demonstrates that policy implementation is delicate in regards to geography and size (16). This report is consistent with our findings that SES factors which influence the incidence rate of TB are not uniform everywhere. Based on the data we collected, we propose that preventable diseases like TB may be better controlled if health systems dialed in on the smaller communities and their needs and the factors that may be affecting them rather than focusing on counties and the country. The only way we believe to eradicate diseases that are easily preventable like TB, there needs to be a more personalized approach to developing public policies, especially in regard to health. Table 4 shows how analyzing a large area may be misleading if communities are not taken into account to at least the zip code. In table 4 there is no correlation with TB incidence and the three investigated SES factors. This was surprising until we went further into our

135

The Role of Socioeconomic Factors in Influencing Tuberculosis Rates

Table 1. Pearson's product-moment correlation & Spearman's rank correlation results between the rate of TB per 100,000 persons and three socioeconomic factors: poverty rate, average median income and percentage of the population who possess health insurance between 18 and 64 years of age in Bronx, NY.

Table 4. Pearson's product-moment correlation & Spearman's rank correlation results between the rate of TB per 100,000 persons and three socioeconomic factors: poverty rate, average median income and percentage of the population who possess health insurance between 18 and 64 years of age in New York City.

Table 2. Pearson's product-moment correlation & Spearman's rank correlation results between the rate of TB per 100,000 persons and three socioeconomic factors: poverty rate, average median income and percentage of the population who possess health insurance between 18 and 64 years of age in Kings, NY.

Table 5. Pearson's product-moment correlation & Spearman's rank correlation results between the rate of TB per 100,000 persons and three socioeconomic factors: poverty rate, average median income and percentage of the population who possess health insurance between 18 and 64 years of age in New York state.

Table 3. Pearson's product-moment correlation & Spearman's rank correlation results between the rate of TB per 100,000 persons and three socioeconomic factors: poverty rate, average median income and percentage of the population who possess health insurance between 18 and 64 years of age in Manhattan, NY.

136

Table 6. Pearson's product-moment correlation & Spearman's rank correlation results between the rate of TB per 100,000 persons and three socioeconomic factors: poverty rate, average median income and percentage of the population who possess health insurance between 18 and 64 years of age in Richmond, NY.

The Role of Socioeconomic Factors in Influencing Tuberculosis Rates

Table 7. Pearson's product-moment correlation & Spearman's rank correlation results between the rate of TB per 100,000 persons and three socioeconomic factors: poverty rate, average median income and percentage of the population who possess health insurance between 18 and 64 years of age in Queens, NY.

Figure 1. Depicts a scatterplot with a trend line to show the association of the rate of TB per 100,000 persons with the poverty rate in Bronx, NY.

Figure 2. Depicts scatterplots with a trend line to show the association of the rate of TB per 100,000 persons with, and percentage of the population who possess health insurance from 18 to 64 years of age (left) and average median income (right) Kings, NY.

Figure 3. Depicts scatterplots with a trend line to show the association of the rate of TB per 100,000 persons, average median income (left), and percentage of the population who possess health insurance from 18 to 64 years of age (right) Manhattan, NY.

137

The Role of Socioeconomic Factors in Influencing Tuberculosis Rates

Figure 4. Depicts scatterplots with a trend line to show the association of the rate of TB per 100,000 persons with the, and percentage of the population who possess health insurance from 18 to 64 years of age (top left), average median income (top right), poverty rate (bottom), New York State.

Figure 5. Depicts scatterplots with a trend line to show the association of the rate of TB per 100,000 persons with the poverty rate (left), and percentage of the population who possess health insurance from 18 to 64 years of age (right) Richmond, NY.

138

The Role of Socioeconomic Factors in Influencing Tuberculosis Rates

Figure 6. Depicts scatterplots with a trend line to show the association of the rate of TB per 100,000 persons with the poverty rate, Queens, NY.

investigation of New York City and analyzed the five boroughs of New York City. We found that all of them had at least one SES factor had a statistically significant correlation in regard to TB incidence. Of the seven locations of New York, we looked at four places, Bronx, Richmond, and Queens boroughs, and New York State excluding New York City, interestingly had a statistically negative correlation of TB incidence and poverty rate. These findings were the opposite of what we were expecting and what is found in the literature. This decline in TB found with higher poverty rates may be explained by New York City increasing the percentage of individuals who complete their TB treatment from 60% to greater than 90% in 2012 (18). This increase in TB treatment completion aided the city to bring the amount of TB outbreaks in hospitals close to zero (17). Three of the locations — Kings and Manhattan boroughs, and New York State — had a statistically significant negative correlation of TB incidence and average adjusted median income. This aligns with our expectations and with what is typically found in literature, where higher income usually is an indicator of a lower incidence of preventable diseases. The surprising part of the median income data concerning TB incidence is that the boroughs of Bronx, Richmond, and Queens all had a significant correlation with TB incidence and poverty did not correlate with TB incidence median income and vice versa. More research would have to be done to see what influences these differences. Each location has somewhat similar poverty rates except for the Bronx, New York, with an average poverty rate of about 28% (9). Kings, Manhattan, Richmond, and New York State had significant negative correlations with TB incidence and Health Insurance. Manhattan, Kings, and New York State all had a significant correlation with median income. We analyzed all seven locations showed a significant positive correlation with median income and the percentage of the adult population who hold health insurance 18 to 64 years of age. Together with our data pointing that median income and the percentage of adults who have health insurance, this data infer that median income influences rates of health insurance, and health insurance rates may influence TB incidence rates.

This study has several limitations, including the possibility of TB cases being underreported due to individuals who may be asymptomatic. The possibility regarding the underrepresentation of TB patients may also be because of individuals not realizing their symptoms and neglecting to seek help, specifically referred to as lack of contact investigation and case management. The lack of contact investigation may not have allowed for tracing the individuals exposed to TB, deterring the authenticity of reported TB cases. Because the study focused on New York State as one entity and New York City, there is a chance the data collected is not representative of the entire sample collected since the additional counties of the state were not individually investigated thus, limiting the generalizability of our findings. Though statistically significant results were found in the study, the data tested three total socioeconomic factors. Other socioeconomic factors will need to be investigated to fully understand how socioeconomic status contributes to TB rates and more pathologies. The parametric tests used may have excluded non-linear relationships, so there is a possibility that meaningful non-linear relationships may be present. Another limitation was individual-level data was not available, so we had to use generalized statistics from the NY CHRIS database and the U.S. Census, which may not fully represent the population. This research supplies insight into the decline of TB infection rates in New York, especially New York City. The data could help predict the declines of TB in other cities or even help cities to tailor better preventive measures to reduce TB infection rates. Our data further can be used as a basis for future research to look at specific SES factors that may contribute to higher rates of preventable diseases so public health workers can better hone in on a more personalized approach for a geographical area, thus, lowering the financial burden of the disproportionate disparities of health disease among low SES populations.

Conclusion This study investigated three socioeconomic factors-poverty, individuals insured between ages 18 and 65, and average median income-and their association with TB infection rates per 100,000. Our results indicate that three SES factors contributed to the decline of TB infection rates within the state of New York; however, socioeconomic factors that influence the incidence rate of TB were very sensitive to location. This study contributes to our understanding of what causes the decline of TB infections, which may influence governments to implement specific measures to help reduce TB or other diseases.

Acknowledgments We would like to thank our instructors, Brian J. Piper, PhD, MS, and Elizabeth Kuchinski, MPH, whose knowledge and experience were critical in developing the research questions and methodology. Your constructive input pushed us to improve our thought and raise the quality of our work.

Disclosures We have no financial relationship with a commercial entity.

139

The Role of Socioeconomic Factors in Influencing Tuberculosis Rates

References 1.

World Health Organization. Global tuberculosis report 2019. Genève, Switzerland: World Health Organization; 2019. Tuberculosis (TB) Fact Sheets [Internet]. Center for disease control and prevention: Tuberculosis (TB) fact sheets. 2011 [cited 2021 Jan 15]. Available from: https://www.cdc.gov/ tb/publications/factsheets/general/tb.htm Schwartz NG, Price SF, Pratt RH, Langer AJ. Tuberculosis - United States, 2019. MMWR Morb Mortal Wkly Rep. 2020;69(11):286–9.

Ethnic and racial minorities & socioeconomic status [Internet]. American Psychological Association. 2017 [cited 2021 Jan 26]. Available from: https://www.apa.org/pi/ses/ resources/publications/minorities

2018 state and city TB report [Internet]. Centers for Disease Control and Prevention. 2019 [cited 2021 Jan 15]. Available from: https://www.cdc.gov/tb/statistics/ indicators/2018/StateCity-TBReport.htm

World Health Organization. Equity social determinants and public health programmes. Genève, Switzerland: World Health Organization; 2010.

Reichman LB, Hershfield ES. Tuberculosis: A comprehensive international approach, second edition. 2nd ed. New York, NY: Marcel Dekker; 2000.

New York City Department of Health and Mental Hygiene Bureau of Tuberculosis Control. Tuberculosis: Historical reports - NYC health [Internet]. NYC Health. 2020 [cited 2021 Jan 14]. Available from: https://www1.nyc.gov/site/ doh/health/health-topics/tuberculosis-historical-reports. page

New York State Department of Health. New York State community health indicator reports (CHIRS) [Internet]. New York State Department of Health. 2019 [cited 2021 Jan 14]. Available from: https://www.health.ny.gov/ statistics/chac/indicators/index.htm

10. US Census Bureau. Poverty status in the past 12 months [Internet]. 2007-2016 American Community Survey 1-year estimates. 2016 [cited 2021 Jan 23]. Available from: https://data.census.gov/cedsci/table?q=new%20 york%20city&t=Poverty&tid=ACSST1Y2016. S1701&hidePreview=false 11. New York Department of Health and Mental Hygiene. Health care access and use [Internet]. EpiQuery Community Health Survey. 2007 [cited 2021 Jan 14]. Available from: https://a816-health.nyc.gov/hdi/epiquery /visualizations?PageType =ts&PopulationSource=CHS& Topic=2&Subtopic=15 12. The Big Picture: Private and public health insurance markets in new york [Internet]. NY State of Health. United Hospital Fund; 2012 [cited 2021 May 11]. Available from: https://info.nystateofhealth.ny.gov/resource/big-pictureprivate-and-public-health-insurance-markets-new-york

140

13. Stewart RJ, Tsang CA, Pratt RH, Price SF, Langer AJ. Tuberculosis — United States, 2017. MMWR Morb Mortal Wkly Rep. 2018 Mar 23;67(11):317–23. 14. Beavers SF, Pascopella L, Davidow AL, Mangan JM, HirschMoverman YR, Golub JE, et al. Tuberculosis mortality in the United States: Epidemiology and prevention opportunities. Ann Am Thorac Soc. 2018 Jun;15(6):683–92. 15. Dawson P, Perri BR, Ahuja SD. High tuberculosis strain diversity among New York City public housing residents. Am J Public Health. 2016 Mar;106(3):563–8. 16. Institute of Medicine. Guidance for the national healthcare disparities report [Internet]. Washington, DC: The National Academies Press; 2002. Available from: https://www.nap. edu/catalog/10512/guidance-for-the-national-healthcaredisparities-report 17. Macaraig M, Burzynski J, Varma JK. Tuberculosis control in New York City — A changing landscape. N Engl J Med. 2014 Jun 19;370(25):2362–5.

Scholarly Research In Progress • Vol. 5, November 2021

Environmental Influences on Childhood Asthma Prevalence in Philadelphia Raeva N. Mulloth1*, Alexander J Blackley1*, Peter J. Koszuta1*, Kaitlyn M. Nemes1*, Maddison M. Vail1*, and Monglin L. Zhang1* ¹Geisinger Commonwealth School of Medicine, Scranton, PA 18509 *Master of Biomedical Sciences Program Correspondence: rmulloth01@som.geisinger.edu

Abstract Background: In 2019 the American Lung Association found that for the second year in a row, the Philadelphia metro has deteriorated the surrounding area’s air quality due to worsening ozone smog. This spike in unhealthy air quality in Philadelphia has affected the health of the city’s population. Unhealthy air quality can be exacerbated by asbestos, which has been found in many Philadelphia elementary schools. Although asbestos usage is now highly regulated, it can still be found in consumer products and construction material today. Among the many factors contributing to asthma onset and other lung diseases, air pollution and dangerous air particles such as asbestos are important contributors. Children in these asbestoscontaminated schools were exposed and ultimately sick which led to their school’s eventual closure. Due to children’s immature and more vulnerable airways, this exposure may have led to increased cases of respiratory distress. Methods: This research study analyzed asbestos data from Asbestos Hazard Emergency Response Act (AHERA) reports from four Philadelphia elementary schools (Laura H. Carnell, James J. Sullivan, Clara Barton, and Thomas M. Peirce) from 2016 to 2018 to further understand the influence of asbestos particles on asthma in children. Secondary data analysis determined the levels of asbestos contamination in each elementary school and the severity of the condition for each school. This was compared to children’s asthma prevalence during the selected time period. Results: Asbestos was mainly found in the pipe insulation and tiles within each school. Between 0.06 and 1.18% asbestos damage was found in 2-6-inch pipe insulation in schools closed for asbestos abatement. An r² of 0.9997 was found when comparing the 2-6-inch pipe damage percentage and the newly friable material found in each school. Thomas M. Peirce Elementary was determined to be the highest concern, according to the analysis of the AHERA reports. Conclusion: Children exposed to asbestos in elementary schools and with a predisposition to asthma were more likely to suffer from respiratory distress due to the multiple contributing environmental factors.

Introduction The City of Philadelphia Department of Public Health Air Management Services has been tracking air quality since the 1980s and discovered an increase in ozone smog in recent years that has contributed to unhealthy air quality (1). Oxidative species exposure, such as ozone and particulate matter found in the air, can lead to tougher, more fibrous, and less efficient respiratory tissue. These environmental influences put those

with existing respiratory conditions and vulnerable populations, such as elderly and developing children, at a higher risk for respiratory distress (1, 3). Asthmatic children who have been exposed to ozone-polluted air can be more at risk for hospitalization, and as a consequence, live shorter lives (2, 4). Negative outcomes related to respiratory illness in Philadelphia can be a results of ozone levels as well as other air pollutants (4). Children are especially vulnerable to respiratory irritation and distress due to their behavior and physiology, specifically the immaturity of their airways (2). Evidence supports that air pollution in developing cells early in life leads to abnormal respiratory functions and growth, which leads to the development of respiratory diseases later on in life (2, 5). Another environmental factor that influences Philadelphia children’s propensity for respiratory diseases is their exposure to asbestos-contaminated schools. Asbestos is a naturally occurring mineral that is resistant to heat, electricity, and corrosion, making it an effective insulator (6). From 1866 to 1978, asbestos-containing building material was manufactured because it was considered to be the ideal material for all types of insulations. Although asbestos presented as a great insulator, it is highly toxic with prolonged exposure (6). Buildings built before asbestos was deemed toxic were all insulated with asbestos, including many of the elementary schools in Philadelphia (7). Data was collected from different elementary schools in Philadelphia from 2016 to 2018 that were built between 1908 and 1931. These schools have confirmed findings of asbestos in their pipe and duct installation, and floor tiles (7–10). The asbestos in these elementary schools potentially exposed a total of 2,700 students, which may be associated with increased incidence rates of asthma in these children due to the toxicity of the product (7–10). In addition to assessing asbestos within the identified elementary schools, other information related to air quality and respiratory illness in Philadelphia County and surrounding counties was collected. Data collected from the city of Philadelphia showed rates of asthma in specific neighborhoods regarding population density and location. Research from 2019 shows the breakdown of 46 different neighborhoods within Philadelphia and compares averages of many different health factors and health outcomes within each neighborhood and in Philadelphia (11). The data indicates neighborhoods in Philadelphia with higher asthma rates in children are in the Upper North, Lower Northeast, and West areas (11). The neighborhoods and surrounding areas of this study’s chosen elementary schools have an increased prevalence of childhood asthma compared to other Philadelphia neighborhoods (11). These neighborhoods also have reported differing levels of access to medical care and housing code violations, which are

141

Environmental Influences on Childhood Asthma Prevalence in Philadelphia

factors that could contribute to poorer health outcomes for children with asthma. Therefore, it is imperative to understand the harm that asbestos can cause in these areas. Our study characterized children from asbestos-contaminated schools in Philadelphia with their negative health outcomes (asthma prevalence).

Methods Participants This study used de-identified patient data from the Philadelphia Department of Public Health. This information was supplemental data about the overall health and asthma prevalence in Philadelphia neighborhoods where these schools are located. This secondary data analysis report used the Philadelphia Asbestos Hazard Emergency Response act (AHERA) collective reports on inspection details for asbestos management plans from different elementary schools. The schools examined were Laura H. Carnell Elementary School (“Carnell”), James J. Sullivan Elementary School (“Sullivan”), Thomas M. Peirce Elementary School (“Peirce”), and Clara Barton Elementary School (“Barton”). The time frame was between 2016 and early 2019, due to available data in AHERA reports (7–10). These schools served as a foundation to the study of asthma in children and could support the idea of asbestos exposure increasing incidence of asthma. Information was gathered about confirmed asbestos findings in the school's building materials through room-by-room inspection by a professional general contractor (9). From these inspection plans, this study extracted what building materials were of high asbestos content, noting location, type, amount present, amount damaged, damage score, and plan of action. The term friable is used to characterize the way the asbestos can be easily broken and crushed into powder when disturbed. Based on this data and the school’s average class sizes, the total number of students that had been exposed to asbestos was determined. The student population was estimated to be around 2,700 students, which was reported in the school’s demographic sheets. This study was determined by the Geisinger Institutional Review Board as exempt research that does not involve or interact with human subjects.

asbestos was found in these areas, where it was present, and the condition, was graphed. Incident rate reports from Philadelphia Department of Public Health for asthma in children, as well as the health outcomes and comparative asthma rates of the surrounding neighborhoods were analyzed. Environmental measures, such as volatile organic compounds (VOC) and ozone levels, were collected from data published by the Department of Public Health in the city of Philadelphia (1). Data points were analyzed in geographic groups around elementary schools and neighborhoods of metropolitan Philadelphia. Comparisons made between these geographic groups and demographics included the age of patients with signs of asthma, elementary school exposure to asbestos, and the effect of environmental air quality.

Results The CDC reported in its Behavioral Risk Factor Surveillance System publications that Pennsylvania has consistently had above-average rates of childhood asthma compared to nationally (14). Figure 1 shows the prevalence of childhood asthma in the United States (U.S.) and Pennsylvania (PA), as reported by the Philadelphia Department of Public Health. Since 2000, PA has had a higher number of children with asthma compared to the rest of the nation. As of 2016, the rate is steadily rising again in PA, while in the U.S. it was decreasing (Figure 1). Philadelphia has a higher presence of VOC in the city’s air. The schools analyzed that were closer to the inner city had an increase in these compounds since 2017, according to the Philadelphia Department of Public Health (15). Compared to Philadelphia County, surrounding areas have lower asthma rates (Figure 2). This showed that 25% of Philadelphia residents are diagnosed with asthma. Comparing different Philadelphia neighborhoods, Center City residents have the highest asthma rates (Figure 3). Our selected Philadelphia elementary schools were chosen on the basis of having been closed as part of the 2019 asbestos crisis (10). Figures 4A and 4B exhibit the total amount of confirmed asbestos damage in the elementary schools selected, depending on the building material. There was a plethora of tile and pipe that was damaged with asbestos (Figure 4A and 4B). Carnell elementary had the most asbestos present in tile (Figure 4A), while Sullivan elementary had the

Data analysis Secondary data and reports were analyzed to acknowledge the instance of poor air quality, high asbestos levels, and higher rates of asthma in children. Our data collection methods included examining AHERA data reports about the inspections documenting asbestos levels (7–9). Data was extracted from each report by noting the amount of material, amount of damage, damage potential, and response actions in each elementary school room. We analyzed where asbestos was found inside of the schools and determined if asbestos was in high-traffic areas. Classrooms, hallways, nurse’s office, large classroom closets, cafeteria, and kitchen were all areas included in the data analysis since children could be expected to interact in these environments. Exclusion criteria was identified as areas with the least amount of children contact, such as administrative offices. The asbestos could be found in either floor tiling, pipe insulation, or duct insulation. Utilizing PRISM, data on how much 142

Figure 1. Child Asthma Prevalence in Pennsylvania (PA) compared to United States (U.S.) average according to the CDC (17). The prevalence of childhood asthma in the U.S. and PA is shown. Since 2000, PA has had a higher number of children with asthma compared to the rest of the nation. As of 2016, the rate is steadily increasing in PA, while in the U.S. it is decreasing.

Environmental Influences on Childhood Asthma Prevalence in Philadelphia

Figure 2. Asthma Prevalence in Surrounding Philadelphia Counties According to Community Asthma Prevention Program (21). This figure shows that approximately ¼ of people in Philadelphia have asthma, according to the 2010 report. The mean was 17.85 and the standard deviation was 4.57.

most asbestos damage on pipe insulation (Figure 4B). The asbestos materials were all found in areas where students would have access to be exposed, such as classrooms and hallways. Figure 5 displays the confirmed percent damage from asbestos on pipe insulation for each of the schools. Peirce elementary had the worst asbestos damage on this building, and this led to its eventual closure in 2019 for asbestos remediation (10) (Figure 5). The damage on the 2-6-inch pipe insulation ranged from 0.06 and 1.18% (Figure 5). An important aspect of asbestos contamination is the particulates breaking off and mixing into the air. Figure 6 illustrates the amount of asbestos-damaged material designated as friable by the general contractor who investigated the schools and wrote the reports. Peirce elementary had the highest value of 11 friable materials (7). The AHERA reports designated a section to outline the conditions of the asbestos and tagged a management plan to these conditions. Figure 7 illustrates the range of conditions associated with asbestos damage. The higher the number and letter, the worse the condition, according to the general contractor’s reports. The 1/D value can be thought of as the “watch and wait” action plan. Although it is of lowest concern, the asbestos damage is still there and needs to be managed accordingly. Peirce was the only school to contain 6/A asbestoscontaining material, which of the highest concern as the condition is severe (Figure 7). The children in the schools would have daily exposure to this toxic material, which could lead to respiratory aggravation in the young students.

Discussion

Figure 3. Percent of Asthma Prevalence by Philadelphia Neighborhood According to Community Asthma Prevention Program (21). Figure 3 portrays the vast differences in the prevalence of asthma in the city of Philadelphia reported in 2010 from Community Health Database. Center City, Philadelphia, was found to have to highest asthma prevalence. The mean was 25.42 and the standard deviation was 6.70.

Figure 4. (A) Total asbestos damage in tile according to AHERA (7-10) (B) Total asbestos damage in pipe insulation according to AHERA (7-10). The amount of confirmed asbestos damage found in tile, measured in square feet and in pipe insulation, which was measured in linear feet, is shown. The data show high amounts of reported asbestos in Carnell and Sullivan elementary schools specifically.

According to the Philadelphia Department of Public Health, Philadelphia was reported to have the highest child mortality rate compared to other United States cities, with 71.6 deaths per 100,00 children (20). Contributing conditions cited were low birth weight, prematurity, and asthma, among many others. They reported that in 2018, there were 15,450 asthma-related visits, with hospital visits being the highest among children under the age of 6. This fact is also influenced by insurance status and racial/ethnic background (20). These statistics show the fundamental importance of our study. Asthma is a very serious problem and is exceptionally bad in Philadelphia due to multiple contributing factors. It is a great concern that there is so much asbestos material in elementary schools. Peirce and Sullivan elementary schools had the worst outcomes that we had examined, but all four schools were deemed to have hazardous conditions for exposure to children. In addition, the increasing asthma rates in children in Center City Philadelphia puts these children in increasing risk for negative health outcomes due to their place of residence and consequently the place of education. Younger children are at higher risk of suffering from asbestos exposure, especially those who have asthma. Philadelphia shows a higher risk for asthma than most areas of the United States, and there needs to be an improvement in the public health department to help control the high rates of asthma in Pennsylvania’s children. Currently, the Philadelphia Allies Against Asthma (PAAA) is working towards decreasing asthma-related mortality and morbidity in children through education programs to medical providers and community members (23).

143

Environmental Influences on Childhood Asthma Prevalence in Philadelphia

This study was conducted during the SARS-COVID-19 pandemic and correspondence with databases and other researchers presented challenges with slow communication. This caused gaps in our AHERA reports analysis and clarification was hard due to low correspondence with primary data collectors. Also, only having whole children population data and not age group data with the asthma rates is a limitation for this study. But the significance of this data is important to be examined, as there are multiple factors that have led to increasing asthma rates. In future studies, it would be beneficial to compile data on asthma prevalence for specific age groups to allow for better statistical comparison. More research needs to be done on the contributing factors to increasing volatile organic compounds in Philadelphia.

Conclusion Asthma is relatively prevalent in Philadelphia and may affect children through historic issues with asbestos in the schools they attended. We primarily looked at secondary data, including inspection data reports documenting asbestos levels in elementary schools around Philadelphia. The goal is to eventually finance the reconstruction of the many asbestoscontaminated schools in Philadelphia in order to provide better air quality for young children with asthma. We hope further work could include partnerships with Children’s Hospital of Philadelphia and Philadelphia Department of Health to survey more data and elicit better understanding of the severity of asthma in children in the city through asbestos and air quality data. Our data serves as a framework for future studies to examine other contributing factors to asthma in children and to bring attention to the need for community health research.

Acknowledgments Shaylyn Paolello, who was a contributor at the beginning of this research project. Professor Elizabeth Kuchinski, Professor Catherine Freeland, and Dr. Brian Piper, who helped advise through the research analysis process. Dr. James Church, who helped with statistical analysis. Community Asthma Prevention Program (CAPP) with Children’s Hospital of Philadelphia, who gave data to aid in this report.

Figure 5. Percent damage of asbestos in 2-6-inch pipe insulation according to AHERA (7-10). The data for the amount of 2-6-inch pipe insulation that had been damaged by asbestos accumulation is shown. Peirce and Carnell elementary schools had the highest amount of pipe insulation damage.

Figure 6. Amount of newly friable material found according to AHERA (7-10). Friable material is defined as an easily powdered, broken down material. This graph shows the amount of asbestos accumulation that was designated as friable in the AHERA reports, according to the general contractor who investigated the schools.

Figure 7. Asbestos condition and associated management plan according to AHERA (7-10). The amount of asbestos and associated hazardous conditions in each of our four schools, as stated in the AHERA reports is shown. The higher the number and letter, the worse the condition. 1/D represents undamaged asbestos, which was found in all of the schools. 6/A represents immediate concern and great risk, which occurred in Peirce elementary school.

Disclosures One disclosure of this study is potential bias in picking our chosen elementary schools. These were selected on the basis of being closed as part of the 2019 asbestos crisis. The only information obtained was the name of the school and the information has been corroborated elsewhere (22).

References 1.

City of Philadelphia Department of Public Health Air Management Services. Philadelphia’s air quality report 2018. 2018.

Physicians For Social Responsibility. How air pollution contributes to lung disease. Physicians For Social Responsibility. 2009.

144

Figure 8. Asthma rate in school counties versus newly friable material found. Data for the asthma rates for each school’s county (Peirce, Upper North; Sullivan, Lower Northeast; Carnell, Upper Northeast; Barton, Upper North) versus the newly friable material found in each school is shown. The p value is 0.6 and the r² is 0.1151.

Environmental Influences on Childhood Asthma Prevalence in Philadelphia

Kim D, Chen Z, Zhou L-F, Huang S-X. Air pollutants and early origins of respiratory diseases. Chronic Dis Transl Med. 2018 Jun 7;4(2):75–94.

19. Growing Up Philly: Department of Public Health [Internet]. City of Philadelphia. 2020 [cited 2021May13]. Available from: https://www.phila.gov/documents/growing-up-philly/

MacMunn A. Air quality in Philadelphia metro area again worsened for ozone smog, finds 2019 “State of the Air” report, had best ever results for year-round particle pollution [Internet]. American Lung Association. 2019. Available from: https://www.lung.org/media/pressreleases/air-quality-in-philadelphia

20. The Children's Hospital of Philadelphia. “Community Asthma Prevention Program (CAPP).” Children's Hospital of Philadelphia, The Children's Hospital of Philadelphia, 5 May 2014, www.chop.edu/centers-programs/communityasthma-prevention-program-capp.

Lipfert F. Daily mortality in the Philadelphia metropolitan area and size-classified particulate matter. J Air Waste Manag Assoc. 2017;50(8).

King D. Asbestos Insulation. Mesothelioma Center - Vital Services for Cancer Patients & Families.

The School District of Philadelphia. Asbestos Hazard Emergency Response Act “AHERA” Thomas M. Peirce Elementary School. 2018.

The School District of Philadelphia. Asbestos Hazard Emergency Response Act “AHERA” Laura H. Carnell Elementary School. 2019.

The School District of Philadelphia. Asbestos Hazard Emergency Response Act “AHERA” James J. Sullivan Elementary School. 2019.

21. “Asbestos in Schools: A Guide for Parents &amp; Administrators.” Mesothelioma Center - Vital Services for Cancer Patients & Families, www.asbestos.com/asbestos/ schools/. 22. Vikaas K. Asthma. PHILADELPHIA ALLIES AGAINST ASTHMA (PAAA). [cited 2021Jul18]. Available from: https://www.phmc.org/site/programs/environmentalhealth/asthma

10. The School District of Philadelphia. Asbestos Hazard Emergency Response Act “AHERA” Clara Barton Elementary School. 2019. 11. Farley T, Washington R, Whitley J. Health of the city 2019. 2019;52. 12. Drexel University Urban Health Collaborative, City of Philadelphia Department of Public Health. Close to home: The health of Philadelphia’s neighborhoods. 2019. 13. Pennsylvania Department of Health. Ambulatory Surgery Center Reports [Internet]. Department of Health. Available from: https://www.health.pa.gov:443/topics/ HealthStatistics/HealthFacilities/SurgeryCenterreports/ Pages/ambulatory-surgery-center-reports.aspx 14. Centers for Disease Control and Prevention. Asthma behavioral risk factor surveillance system [Internet]. 2020. Available from: https://www.cdc.gov/asthma/brfss/default. htm 15. City of Philadelphia Department of Public Health Air Management Services. 2019-2020 air monitoring network plan. 2019. 16. AirNow. AirNow Philadelphia, PA [Internet]. 2021. Available from: https://www.airnow.gov/ 17. Centers for Disease Control and Prevention. AsthmaStats: Uncontrolled asthma among children, 2012–2014 [Internet]. 2019. Available from: https://www.cdc.gov/ asthma/asthma_stats/uncontrolled-asthma-children.htm 18. Centers for Disease Control and Prevention. Asthma as the underlying cause of death. 2018. Available from: https:// www.cdc.gov/asthma/asthma_stats/asthma_underlying_ death.html

145

Scholarly Research In Progress • Vol. 5, November 2021

Declines and Pronounced Regional Disparities in Prescription Opioids in the United States Joshua D. Madera1†, Amanda E. Ruffino1†, Adriana Feliz1†, Kenneth L. McCall2, and Brian J. Piper1,3 Geisinger Commonwealth School of Medicine, Scranton, PA 18509 University of New England, Portland, ME 04103 3 Center for Pharmacy Innovation and Outcomes, Forty Fort, PA 18704 † Doctor of Medicine Program Correspondence: jmadera@som.geisinger.edu 1 2

Abstract

Introduction

Background: The United States (U.S.) opioid epidemic, including the misuse of prescription drugs, is a significant public health crisis due to the risk of addiction and overdoses. Despite guidelines and several policies that have been implemented to combat these issues, drug overdoses continue to increase. This study aimed to assess the changes in opioid prescriptions between 2010 and 2019. We hypothesized that a decrease in opioid prescriptions would be observed but that there would be substantial state level variability.

The current rate of opioid prescriptions coupled with the risk of opioid addiction and unintentional overdose is a persistent public health concern. The aforementioned, intertwined with the observed increase in opioid overdose-related deaths, promoted the public health concern to a nationwide United States (U.S.) epidemic (1). The current rate of opioid prescriptions may be attributed to an increased prevalence of chronic pain, limited knowledge regarding long-term opioid use, and pharmaceutical companies capitalizing on false claims regarding the non-addictive nature of opioids (2). Since the emergence of heightened opioid utilization, the number of opioid-related deaths has steadily risen from 1999 to 2019 across the U.S. (3). In 2019 alone, a total of 70,630 drug overdose deaths occurred, corresponding to an age-adjusted rate of 21.6 per 100,000 population and a 4.3% increase from the 2018 rate (20.7) (1). However, despite the increasing mortality rate, the opioid prescription per capita expanded 7.3% between 2007 and 2012 with marked prescription rate increases in family medicine, general practice, and internal medicine (4). Given the stable prescription rates and increased pervasiveness of adverse outcomes, a sizable intervention was necessary to control this problem. With the plan to trigger nationwide transformation, the U.S. Department of Health and Human Services declared the opioid epidemic a public health emergency in 2017 (5).

Methods: Data was obtained from the U.S. Drug Enforcement Administration’s Automation of Reports and Consolidated Orders System. Data from 2010 to 2019 was collected for 10 prescription opioids (codeine, fentanyl, hydrocodone, hydromorphone, meperidine, methadone, morphine, oxycodone, oxymorphone and tapentadol) and was analyzed. Results: The peak years for all 10 prescription opioids were identified individually between 2010 and 2013 except for codeine (2015). There was a -51.96% overall decrease in opioid distribution. The largest decrease was observed in Florida (-61.61%) and smallest in Texas (-18.64%). While the largest quantities of opioid distribution were observed in Tennessee (520.70 morphine milligram equivalent or MME per person) and Delaware (251.45) in 2011 and 2019, respectively, the smallest MMEs were observed in Nebraska (153.39) and Minnesota (90.49), respectively. The highest to lowest state ratio of opioid use, corrected for population, was 3.39-fold in 2011 to 2.78-fold in 2019. Moreover, the distribution between hydrocodone and codeine, and hydromorphone and codeine increased significantly from 2011 to 2019, p < 0.05. Conclusion: Following the peak year for each opioid, a steady decline in opioid distribution was revealed. However, the extent of prescription use varied 3.39-fold in 2011 and 2.78-fold in 2019 by state. Moreover, there was an overall 42.99% decline in total opioid distribution across the U.S. from 2011 to 2019. If evidence-based medicine was being practiced, we would have expected that two states with similar rates of acute pain (e.g., North and South Dakota) should have similar use patterns for prescription opioids. However, we observed pronounced regional variability. The observed regional differences may partially be due to variability in state-level legislative changes that were implemented to combat the opioid epidemic or other factors. Additional research focused on the relationship between policy changes and prescription opioids is warranted.

146

Following this declaration and the resulting shift in mentality regarding the devastating effects associated with the overprescription of opioids, a strategic nationwide plan to combat the epidemic remained unestablished. Instead, the task to develop strategies aimed at curbing the epidemic was left to the individual states, creating an opportunity for opioid legislative disparities nationwide. State legislators have independently created legislation guiding the prescribing of opioids for a multitude of conditions such as acute pain, chronic pain, cancer, terminal condition, palliative care, etc. For example, in Florida, House Bill 21 was enacted that would require certain registered practitioners to complete a specified boardapproved continuing education course to obtain authorization to prescribe controlled substances, adopt rules establishing certain guidelines for prescribing controlled substances for acute pain, provide requirements for pharmacists for the dispensing of controlled substances to persons not known to them, and authorize a pharmacist to dispense controlled substances upon receipt of an electronic prescription if certain conditions are met (6). Further, prescribers are asked to develop

Declines and Pronounced Regional Disparities in Prescription Opioids in the United States

a written plan for assessing each patient’s risk of aberrant drugrelated behavior, which may include drug testing, pill counting, and close monitoring from the physician (6). In 2019, Texas signed House Bill 2174 which states that a practitioner may not issue a prescription for an opioid in an amount that exceeds a 10-day supply or provide for a refill of an opioid for acute pain. The law does not limit how many times the 10-day supply may occur and does not apply to cancer, hospice, or chronic pain management (7). Nevada adopted the AB 474 and SB 59 in 2018 which has five major requirements; two units of CME per licensing cycle required for all licensed prescribers, Mandated Registry and Use of Prescription Monitoring Programs, New Prescriptions Requirements, New Prescribing Guidelines for Controlled Substances, and Overdose Reporting (8). North Carolina’s 2017 house bill 243 extends standing orders for an opioid antagonist to community health groups, requires supervising physicians to personally consult with physician assistants and nurse practitioners who prescribe certain Schedule II and III controlled substances for long-term use, requires electronic prescribing of certain Schedule II and III controlled substances, and establishes maximum limits for initial prescription of certain Schedule II and III controlled substances (9). Each state has implemented what they believe to be the best approach at combating the epidemic; whether these laws are being enforced as well as their direct effects on opioid prescribing patterns is debatable. What is clear, however, is the lack of and need for either collaborative nationwide opioid legislation or a governing body tasked with the oversight and enforcement of the opioid legislation created by the states. The observed variation in prescription opioid trends across the U.S. despite the implementation of new opioid legislation suggests that additional extraneous factors may be limiting the extent of its impact. Specifically, racial-ethnic disparities may partially explain the observed patterns, as emergency department physicians were less likely to prescribe opioids to non-Hispanic blacks for abdominal and back pain as compared with non-Hispanic whites due to the increased incidence of drug-seeking behaviors associated with this patient population (10). Moreover, the incidence of opioid-related overdose deaths documented in lower-income white regions in California may also support the involvement of socioeconomic status as an influential factor in prescription-opioid rates (11). Furthermore, women were more likely than men to receive opioid prescriptions relative to men (12). These factors, alongside other variables, may not only impact the efficacy of imposed legislation but may also influence the overall production of specific opioids that are more frequently prescribed or beneficial for patients with certain health conditions. Therefore, additional analyses focused on stratifying opioid prescription rates based on influential factors may be necessary to explain the observed trends. The morphine milligram equivalent (MME) per capita in 2016 in the U.S. was 1,124 (13) which is appreciably higher than other developed countries or the U.S. Territories (14). However, even this belies the pronounced variability between states. There was a 5-fold difference between the lowest (North Dakota = 485) and highest (Rhode Island = 2,624) states (13). Examination of individual opioids revealed a 3-fold state level difference for fentanyl (15), 18-fold for meperidine (16), and 20-fold for buprenorphine (17).

Utilizing the above knowledge, this study aimed to confirm previously observed declining trends in opioid prescriptions by analyzing 10 opioids between 2010 and 2019. Additionally, through the identification of the state(s) with the highest and lowest decline in prescription rates, regional disparities in efforts to combat the opioid epidemic were examined. Through a direct comparison, opioid prescribing patterns for each of the 10 opioid drugs were evaluated. Lastly, to assess the relationship between the supply and demand of the analyzed opioids, the drugs’ production quotas were compared to opioid prescription rates across the U.S. Through these means, this study expected to identify decreasing opioid prescription rates across the U.S. in the period of 2010 to 2019. The largest opioid prescription decline was hypothesized to be seen in regions that are most proactive in implementing and enforcing opioid prescribing reduction policies and identify risk factors that may predispose certain Americans to fall victim to the opioid epidemic.

Methods Data collection Opioid production data was obtained from the U.S. Drug Enforcement Administration’s Production Quotas System, and opioid distribution data was obtained using the U.S. Drug Enforcement Administration’s Automation of Reports and Consolidated Orders System (ARCOS) (U.S. Department of Justice, Drug Enforcement Administration (DEA), Division of Diversion Control, 2019). ARCOS is a publicly available reporting system mandated by the DEA that reports detailed and comprehensive drug information from manufacturing to distribution. Information on 10 opioid prescription drugs was collected: codeine, fentanyl base, hydrocodone, hydromorphone, meperidine, methadone, morphine, oxycodone, oxymorphone and tapentadol. Population data was obtained from the U.S. Census. All procedures were reviewed and approved by Geisinger Commonwealth School of Medicine’s and the University of New England’s Institutional Review Board. Data analysis The MME for each opioid drug was determined. These values were calculated using the appropriate multipliers: codeine 0.15, fentanyl base 75, hydrocodone 1, hydromorphone 4, meperidine 0.1, morphine 1, oxycodone 1.5, oxymorphone 3, and tapentadol 0.4 (13, 18). The multiplier for methadone was defined as both 8 and 12, based on whether it was utilized for narcotic programs (13, 18). Methadone utilized in narcotic programs was not included in the methadone prescription value. To isolate the peak opioid prescription year, the weight in grams of each opioid was obtained from ARCOS from 2010 to 2019. Weights for each opioid were converted to their MME values and were then summated for all 10 opioids across all 50 states. To assess changes in opioid prescriptions from 2011 to 2019, heat maps demonstrating the percent change across all 50 states per capita were produced using JMP. The per capita use of each opioid drug for each state was calculated using the respective population estimate according to the U.S. Census. Using these values, ratios for the 95th:5th and 70th:30th percentiles were calculated. Furthermore, Spearman correlations were calculated for all the opioid drugs. Lastly, to assess the relationship between the supply and demand of the analyzed opioids, the 147

Declines and Pronounced Regional Disparities in Prescription Opioids in the United States

drugs production quotas were compared to opioid prescription rates across the U.S. All figures were constructed utilizing the GraphPad Prism version 8.

Results Examination of total opioid production quotas from 2010 to 2019 revealed a gradual increase beginning in 2011 that remained relatively stagnant until 2016 when a pronounced decline was observed. Many opioids, including methadone and oxycodone, exhibited a similar trend in their production. However, others such as codeine remained consistent throughout this timeframe (Figure 1A). Analysis of opioid distribution during the same time period identified 2011 as the peak year for total opioid distribution in the U.S. (Figure 1B). The peak year for each opioid was 2010 for oxycodone and meperidine; 2011 for hydrocodone and oxymorphone; 2012 for morphine and tapentadol; 2013 for hydromorphone and fentanyl base; and 2015 for codeine. Since 2011, total opioid prescription for the ten pain medications analyzed has continued to steadily decline, despite the increase in total opioid production. A -45.4% decline from 2011 to 2019 and an -11.3% decline from 2018 to 2019 was observed. Comparison of total opioid prescriptions per capita across the U.S. in 2011 (Figure 2A) and 2019 (Figure 2B) revealed an overall decrease of -51.96% in opioid distribution, but also considerable state level differences. Tennessee (520.70), Nevada (491.38), Delaware (476.15) and Florida (452.12) had significantly (p < 0.05) increased levels of total opioid distribution relative to the national average in 2011 (Figure 2C). Alabama (251.45) and Delaware (238.71) had significantly increased distribution relative to the mean in 2019 (Figure 2D). Moreover, analysis of changes in individual opioid prescription rates revealed the largest increase for codeine, which had a 143% increase in per capita distribution from 2011 to 2019. In contrast, oxycodone demonstrated the largest decrease of 50.09% in per capita distribution from 2011 to 2019. Furthermore, there was a 3.31-fold difference between the highest (Florida = -61.61%) and lowest (Texas = -18.64%) state opioid prescription decline. Figure 2E demonstrates the observed percent difference for total opioid prescription from 2011 to 2019. The ratios between the 95th:5th percentiles were calculated for each of the Schedule II opioids for 2011 and 2019 (Figure 3). Analysis of the differences in the 95th:5th percentile ratio between 2011 and 2019 revealed an increase for codeine, hydrocodone, meperidine, and methadone, but a decrease for the remaining opioids. Interestingly, methadone was the only opioid to demonstrate a 95th percentile value increase from 2011 to 2019. The remaining opioids demonstrated decreases in their percentile ranks. Table 1 demonstrates the Spearman’s correlation for the various opioids. Interestingly, oxycodone had the most significant correlations with the other opioids in both 2011 and 2019. The strength of the correlation between oxycodone and morphine (0.50 and 0.51, respectively) and oxycodone and oxymorphone (0.50 and 0.61, respectively) increased from 2011 to 2019. However, between oxycodone and methadone (0.66 and 0.39, respectively) the strength of the correlation decreased. These trends were also observed for the Spearman’s correlation 148

Figure 1. (A) Total opioid production and (B) distribution in morphine milligram equivalents across the United States from 2010 to 2019 as reported to the Drug Enforcement Administration’s Automated Reports and Consolidated Orders System. Opioids analyzed were codeine, fentanyl, hydrocodone, hydromorphone, meperidine, methadone, morphine, oxycodone, oxymorphone, and tapentadol.

between the other opioids analyzed. Moreover, there was a moderate correlation revealed between hydrocodone and meperidine in 2011 (0.62) and 2019 (0.58). Furthermore, Fisher r to Z transformations were calculated to determine whether there were any changes in the strength of the correlations between opioids in 2011 and 2019. These analyses revealed a significant increase in the strength of the correlation between codeine and hydrocodone (p < 0.05), codeine and hydromorphone (p < 0.01), hydrocodone and methadone (p < 0.05), and methadone and oxycodone (p < 0.05). Additionally, a significant difference was identified between the correlation for methadone between 2011 and 2019 (p < 0.05).

Discussion There are several key findings to this report. This study identified 2011 as the peak year of opioid prescriptions in the U.S., and a consistent decline in prescription rates since that point. These findings are consistent with the literature assessing opioid prescription trends at earlier time periods

Declines and Pronounced Regional Disparities in Prescription Opioids in the United States

Table 1. Spearman’s correlations for opioid distribution per state corrected for population in 2011 and 2019. Boxes in blue and orange represent significant associations between the opioids of < 0.05 and < 0.01, respectively. * p < .05 for Fisher r to Z scores when comparing 2011 and 2019. ** t-test p < .05 versus 2011.

(13, 19, 20, 21). It did not escape notice that the prescribing peak in 2011 was 5 years before the CDC issued its opioid prescribing guidelines (March 2016) and 6 years before rising overdoses resulted in the classification of the situation as a “national emergency” (July, 2017). There is a myriad of reasons that may explain the decline in opioid prescriptions, including individual legislative amendments implemented across the U.S. (22), PDMPs (23), increased awareness of the addictive nature of opioids (24), increased rates of opioid-related overdose mortalities (2), and heightened appreciation of the differences in biochemical characteristics of these opioids (22). Comparison of the variation of opioid prescriptions across the U.S. between 2011 and 2019 based on each respective state’s population revealed a substantial increase in codeine prescriptions. This increase may be attributable to the substitution of codeine and tramadol for pain management as compared with the typical protocol of hydrocodone (2). In fact, hydrocodone prescriptions following Emergency Department visits decreased by approximately 12% between 2012 and 2017 (2). Moreover, a pronounced decrease in oxycodone prescriptions was also identified in this analysis, which may have also contributed to the elevated rates of codeine prescriptions

for pain management. The CDC reported a decline in oxycodone prescriptions between 2011 and 2017, statistics that are consistent with our findings (2). While the increase in codeine prescriptions may explain the opioid prescription trends observed in our study, additional research is necessary to substantiate our reasoning. Furthermore, the 95th percentile of methadone was the only opioid to demonstrate an elevated rate in 2019 as compared with the estimated rate in 2011. The observed increase in methadone prescriptions may be a result of the drug’s efficacy in treating pain and cost-effectiveness (25). To further elucidate the observed trend, efforts should be directed at stratifying opioid prescription distribution and use between 2011 and 2019. Drug overdose deaths have not homogeneously impacted the U.S. Although death determination is an extremely complicated process and data should be interpreted cautiously because of methodological differences between states (26), the eastern U.S. has had higher rates of overdoses (27). Similarly, opioid prescriptions were not homogeneous. Several positive associations between opioid prescriptions were identified using Spearman correlations, suggesting that prescribing a specific opioid may increase the likelihood of adding another

149

Declines and Pronounced Regional Disparities in Prescription Opioids in the United States

Figure 2. Heat maps demonstrating the total opioid prescription per capita across the United States in 2011 (A) and 2019 (B). Per capita distribution was organized from largest to smallest values for 2011 and 2019 in panels C and D, respectively. Panel E shows the percent change in total opioid prescription from 2011 to 2019.

150

Declines and Pronounced Regional Disparities in Prescription Opioids in the United States

Figure 3. 95:5 percentile ratio for each of the prescription opioids analyzed.

opioid to the patient’s pain management regimen. Interestingly, oxycodone was implicated in various of these associations in both 2011 and 2019. However, a specific reason for this finding remains unknown. This may be due to medical specialties having certain preferences for one opioid over another; for example, otolaryngologists commonly prescribe hydrocodone and oxycodone, while family medicine physicians tend to prescribe buprenorphine (28, 29, 30). It is necessary to continue the battle against the opioid epidemic to eliminate opioid misuse, overdose, and deaths. While recent efforts may have significantly contributed to the observed decline in prescription opioids, additional strategies may be necessary to permanently remedy the issue. It has been speculated that this goal can be achieved by increasing patient awareness and education, implementing institutionalized prescription guidelines, and establishing patient-bedside standardized protocols focused on pain management (31). Furthermore, having individual states create their own legislative policies to combat the epidemic has led to varying outcomes and is perhaps not the path we should continue on any longer. This divisive solution could be remedied by creating a national PDMP system, updated in real time, which accounts for electronic prescriptions of all Schedule II and III opioids, patient diagnoses, overdose risk number calculated based on this history, and guidelines for alternative pain management to be explored with the patient (13). Furthermore, a nationwide mandate for the use of the CDC’s Drug Overdose Surveillance and Epidemiology (DOSE) System alongside an increase in state funding for Enhanced State Opioid Overdose Surveillance (ESOOS) could prove effective (32, 33). Until that time comes, current potential solutions include individual states implementing limits to the high morphine equivalent daily dose prescription rate (22) and considering prescribing marijuana as an alternative for pain management (34). However, there is still room for improvement, and perhaps the U.S. could adopt and modify international guidelines like the UK to aid in curbing the epidemic further (35). While we were able to successfully identify and isolate trends in opioid prescriptions between 2011 and 2019, this

comprehensive database and our analysis has some caveats. We reported a decline in opioid prescriptions beginning in 2011 based on the distribution of the aforementioned opioids. However, other commonly prescribed opioids such as buprenorphine were not included in our analysis, and therefore may have influenced the peak year of opioid prescriptions. In addition, ARCOS data does not account for Schedule IV or V opioids (e.g., codeine and acetaminophen combination products) that could influence the total number of opioid prescriptions across the U.S. (13). Thus, future studies should analyze trends in opioid prescriptions by including information on these additional opioids collected from additional drug databases like PDMPs and electronic medical records. Moreover, racial, gender and socioeconomic disparities were not statistically accounted for in this study. However, the literature has demonstrated increased opioid use in female and Caucasian individuals (11, 36). Furthermore, current rates of opioid prescriptions are substantially increased in low-income white communities compared to low-income black communities, and black patients are less likely to be prescribed opioids than white patients (11, 37). Moreover, black patients who are prescribed opioids for chronic pain are tested more frequently for concomitant use of illicit drugs than white patients, which may reduce the risk of unintentional overdose for black patients due to drug interactions (38). These findings support the suggested propensity of low-income and racial disparities in opioid prescription rates. This may indicate a need to stratify the data based on these demographics to identify more specific trends in opioid prescriptions. Furthermore, given the extended period of comparison of opioid prescriptions, it is difficult to isolate even an association between implemented legislative amendments and the observed reduction in opioid prescriptions. To elucidate this relationship, future studies should be directed at examining trends in opioid prescriptions at specific periods based on legislative changes enacted since 2011 to determine their overall impact on the declining distribution. An economist might rightly argue that this descriptive study cannot differentiate between reduced demand for opioids by prescribers as well as patients versus reductions in supply imposed by the DEA’s production quotas (Figure 1A) and subsequent shortages, particularly of parental formulations of fentanyl, hydromorphone, and morphine (39).

Conclusion Overall, we have demonstrated a pronounced decline in opioid prescription across the U.S. from 2011 to 2019, as well as sizeable and stable regional disparities in opioid distribution. We speculated that this decline may at least be due in part to legislative amendments implemented by various states and novel insurance policies, as well as the DEA opioid production quotas, to mitigate the opioid epidemic. While individual states have enacted legislative amendments they deemed necessary to combat the epidemic, further analysis regarding specific modifications and regulations may be necessary to illuminate the etiology behind the observed variation in the decline of opioid prescriptions across the U.S. Through these analyses, we may be able to identify additional strategies aimed at controlling the opioid epidemic and further increasing the rate of decline of prescription opioids.

151

Declines and Pronounced Regional Disparities in Prescription Opioids in the United States

Acknowledgments We would like to thank Geisinger Commonwealth School of Medicine and the Center for Pharmacy Innovation and Outcomes for supporting this research.

Disclosures BJP is part of an osteoarthritis research team supported by Pfizer and Eli Lilly. The other authors have no disclosures.

References 1.

Mattson CL, Tanz LJ, Quinn K, Kariisa M, Patel P, Davis NL. Trends and geographic patterns in drug and synthetic opioid overdose deaths — United States, 2013–2019. MMWR Morb Mortal Wkly Rep 2021;70:202–207.

Jones GH, Bruera E, Abdi S, Kantarjian HM. The opioid epidemic in the United States – overview, origins and potential solutions. Cancer. 2018 Nov;124(22):4279-4286.

Centers for Disease Control and Prevention (CDC). The drug overdose epidemic: Behind the numbers [Internet]. cdc.gov. 2021 [cited 2021 Mar 23]. Available from: https:// www.cdc.gov/drugoverdose/data/index.html

Levy B, Paulozzi L, Mack KA, Jones CM. Trends in opioid analgesic-prescribing rates by specialty, U.S. Am J Prev Med. 2015;49:409-413.

Digital Communications Division (DCD). Opioid crisis statistics [Internet]. Hhs.gov. 2018 [cited 2020 Dec 17]. Available from: https://www.hhs.gov/opioids/about-theepidemic/opioid-crisis-statistics/index.html

The Florida Senate [Internet]. Chapter 456 Section 072 - 2017 Florida Statutes - The Florida Senate. [cited 2021May23]. Available from: https://www.flsenate.gov/ laws/statutes/2017/456.072

TMB Clarifies New 10-Day Limit on Opioids By Joey Berlin [Internet]. Texmed. [cited 2021May24]. Available from: https://www.texmed.org/TexasMedicineDetail. aspx?id=51407

Nevada State Medical Association. [cited 2021May24]. Available from: https://nvdoctors.org/practice-resources/ prescribing-opioids/

Strengthen Opioid Misuse Prevention Act of 2017, House Bill 243, Session Law 2017-74 Part II, (2017).

10. Singhal A, Tien YY, Hsia RY. Racial-ethnic disparities in opioid prescriptions at emergency department visits for conditions commonly associated with prescription drug abuse. PloS One. 2016;11(8):e0159224. 11. Friedman J, Kim D, Schneberk T, Bourgois P, Shin M, Celious A, Schriger DL. Assessment of racial/ethnic and income disparities in the prescription of opioids and other controlled medications in California. JAMA Internal Med. 2019;179(4):469-476. 12. Schieber LZ, Guy Jr GP, Seth P, Losby JL. Variation in adult outpatient opioid prescription dispensing by age and sex— United States, 2008–2018. MMWR Morb Mortal Wkly Rep 2020;69(11):298.

152

13. Piper BJ, Shah DT, Simoyan OM, McCall KL, Nichols SD. Trends in medical use of opioids in the U.S., 20062016. Am J Prev Med. 2018;54(5):652-660. 14. Cabrera FF, Gamarra ER, Garcia TE, Littlejohn AD, Chinga PA, Pinentel-Morillo LD, Tirado JR, Chung DY, Pande LJ, McCall KL, Nichols SD. Opioid distribution trends (2006–2017) in the US Territories. PeerJ. 2019 Jan 15;7:e6272. 15. Collins LK, Pande LJ, Chung DY, Nichols SD, McCall KL, Piper BJ. Trends in the medical supply of fentanyl and fentanyl analogues: United States, 2006 to 2017. Prev Med. 2019 Jun 1;123:95-100. 16. Boyle JM, McCall KL, Nichols SD, Piper BJ. Declines, and pronounced regional disparities, in meperidine use in the United States. medRxiv. 2020 Jan 1. 17. Azar AR, Lam W, Muchisky A, Kaleem S, Lockard L, Mandel M. Regional examination of increased buprenorphine in the US (2008-2017). Pharmacoep Drug Safety. 2019 Aug 1;28:328. 18. Portman R. CARA 2.0 Act of 2018. Apr 11, 2018. 19. Sairam Atluri G, Manchikanti L. Assessment of the trends in medical use and misuse of opioid analgesics from 2004 to 2011. Pain Physician. 2014;17:E119-E128. 20. Guy Jr GP, Zhang K, Bohm MK, Losby J, Lewis B, Young R, Murphy LB, Dowell D. Vital signs: changes in opioid prescribing in the United States, 2006–2015. MMWR. 2017;66(26):697. 21. Raji MA, Kuo YF, Adhikari D, Baillargeon J, Goodwin JS. Decline in opioid prescribing after federal rescheduling of hydrocodone products. Pharmacoep Drug Safety. 2018;27(5):513-519. 22. Heins SE, Frey KP, Alexander GC, Castillo RC. Reducing high-dose opioid prescribing: state-level morphine equivalent daily dose policies, 2007-2017. Pain Med. 2020;21(2):308-316. 23. State PDMP Profiles and Contacts [Internet]. Pdmpassist.org. Bureau of Justice Assistance (BJA); [cited 2020 Dec 17]. Available from: https://www. pdmpassist.org/State 24. McGinty EE, Stuart EA, Caleb Alexander G, Barry CL, Bicket MC, Rutkow, L. Protocol: mixed-methods study to evaluate implementation, enforcement, and outcomes of U.S. state laws intended to curb high-risk opioid prescribing. Implement Sci. 2018;13(1):37. 25. Kuehn BM. Methadone overdose deaths rise with increased prescribing for pain. JAMA. 2012 Aug 22;308(8):749-50. 26. Kaufman DE, Nihal AM, Leppo JD, Staples KM, McCall KL, Piper BJ. Opioid mortality following implementation of medical cannabis programs in the United States. Pharmacopsychiatry. 2021 Mar;54(02):91-5.

Declines and Pronounced Regional Disparities in Prescription Opioids in the United States

27. Kiang MV, Basu S, Chen J, Alexander MJ. Assessment of changes in the geographical distribution of opioid-related mortality across the United States by opioid type, 19992016. JAMA network open. 2019 Feb 1;2(2):e190040-. 28. Mark TL, Kassed CA, Vandivort-Warren R, Levit KR, Kranzler HR. Alcohol and opioid dependence medications: prescription trends, overall and by physician specialty. Drug Alcohol Dep. 2009;99(1-3):345-349. 29. Schwartz MA, Naples JG, Kuo CL, Falcone TE. Opioid prescribing patterns among otolaryngologists. Otolaryngology Head Neck Surg. 2018;158(5):854-859. 30. Weiner SG, Chou SC, Chang CY, Garner C, El Ibrahimi S, Hallvik S, Hendricks M, Baker O. Prescription and prescriber specialty characteristics of initial opioid prescriptions associated with chronic use. Pain Med. 2020;21(12):3669-3678. 31. Lovecchio F, Premkumar A, Stepan JG, Albert TJ. Fighting back: institutional strategies to combat the opioid epidemic: a systematic review. HSS J 2019;15(1):66-71. 32. Centers for Disease Control and Prevention (CDC). CDC’s Drug Overdose Surveillance and Epidemiology (DOSE) System [Internet].cdc.gov. 2020 [cited 2021 Mar 23]. Available from: https://www.cdc.gov/drugoverdose/data/ nonfatal/case.html 33. Centers for Disease Control and Prevention (CDC). Enhanced State Opioid Overdose Surveillance [Internet]. cdc.gov. 2020 [cited 2021 Mar 23]. Available from: https:// www.cdc.gov/drugoverdose/foa/state-opioid-mm.html 34. Voelker R. States move to substitute opioids with medical marijuana to quell epidemic. JAMA. 2018;320(23):24082410 35. Weisberg DF, Becker WC, Fiellin DA, Stannard C. Prescription opioid misuse in the United States and the United Kingdom: cautionary lessons. Int J Drug Policy. 2014;25(6):1124-1130. 36. Manubay J, Davidson J, Vosburg S, Jones J, Comer S, Sullivan M. Sex differences among opioid-abusing chronic pain patients in a clinical trial. J Addiction Med. 2015;9(1):46. 37. Becker WC, Starrels JL, Heo M, Li X, Weiner MG, Turner BJ. Racial differences in primary care opioid risk reduction strategies. Annals Fam Med. 2011 May;9(3):219-225. 38. Kashef Z. Study finds racial disparities in prescribing opioids for chronic pain [Internet] Yale.edu. 2018 [cited 2020 Dec 17]. Available from: https://news.yale.edu/2018/08/20/ study-finds-racial-dsiparities-prescribing-opioids-chronicpain 39. Davis MP, McPherson ML, Mehta Z, Behm B, Fernandez C. What parenteral opioids to use in face of shortages of morphine, hydromorphone, and fentanyl. Amer Jour of Hospice and Palliative Med. 2018 Aug;35(8):1118-22.

153

Scholarly Research In Progress • Vol. 5, November 2021

The Association Between Early Menarche and Gestational Diabetes: A Secondary Analysis Annabel S. Alfonseca1*, Uzoamaka V. Eziri1*, Anmol K. Kaur1*, Taylor S. Mewhiney1*, and Grace L.Tieko1* ¹Geisinger Commonwealth School of Medicine, Scranton, PA 18509 *Master of Biomedical Sciences Program Correspondence: tmewhiney@som.geisinger.edu

Abstract Over the last few decades, the prevalence of gestational diabetes mellitus (GDM) has increased. Yearly, 2–14% of pregnancies are affected by GDM. Early age of menarche, the first menstrual period, has been associated with an increased risk of GDM. Race/ethnicity, socioeconomic status, environmental stress, and genetics may contribute to this finding. Early menarche age ranges are neither definitive nor universal and are based on studied populations. For this investigation, early menarche was defined as ages 9 to 12. Little data has been collected regarding the long-term consequences associated with early age of menarche. With new knowledge of at-risk communities with predispositions to GDM, more standardized guidelines and preventive measures can be implemented to better manage the long-term consequences of gestational diabetes. This investigation utilized secondary data from the National Health and Nutrition Examination Surveys (NHANES) Reproductive Health Questionnaire to address the gap in knowledge of early onset of menarche and later development of GDM in women living in the United States. This may have significant public health implications, as maternal obesity rates are increasing and can result in pregnancy complications. We hypothesize that early age of menarche (ages 9 to 12) will be associated with a higher risk of developing GDM in the United States. Our analysis showed there was a significant association between age of menarche and diabetes diagnosis while pregnant in addition to age of gestational diabetes diagnosis and age when delivering a baby 9 pounds or greater. These results contradict previous findings in the literature and call for further investigation into the relationship between early menarche, GDM, and race/ethnicity.

Introduction The American Diabetes Association defines gestational diabetes mellitus (GDM) as any degree of glucose intolerance during pregnancy regardless of insulin or diet modification and whether the condition persists after pregnancy (1). As of November 2020, there were a total 330,571,917 people living in the United States, (2) with women making up 50.6% of the population (3). From 2000 to 2010, the prevalence of GDM increased by 56% (4), and every year 2–14% of pregnancies are affected by gestational diabetes. Data suggests GDM appears to be more prevalent in African American, American Indian, and Hispanic/Latina American women (5). For the United States, the public health implications of this are significant, as maternal obesity rates are increasing and can result in negative long-term outcomes for both the mother and the fetus (6). In addition to GDM, maternal obesity also increases the risk for several pregnancy complications,

154

including gestational hypertension, maternal hypertensive disease or pre-eclampsia, risk of emergency cesarean delivery, and prolonged delivery times (7). As a result of these various complications, the neonate has increased adiposity and is at risk for birth complications. Children of women with GDM have a greater prevalence of childhood obesity and glucose intolerance (8). Additionally, women with a history of GDM have long-term risks associated with development of Type 2 diabetes mellitus (9), hypertension (10), dyslipidemia (11), and cardiovascular disease (12). Menarche, the first menstrual period in a female, is another factor of women’s reproductive health that has been correlated with the development of diabetes (13). Menstruation is the monthly shedding of the functional layer of the uterus (13). The average age of menarche varies within different ethnic and racial groups. On average, Non-Hispanic Black and Mexican American women reached menarche earlier than other groups at 11.82 to 12.36 and 11.81 to 12.37 years, respectively, compared to the average age of 12.38 to 12.67 years (14). The definition of early age of menarche varies and is defined from ages 9 to 11.5 (15). This is explained by its relation to the distribution of the onset of menarche in studied populations (15). As a result, ranges of early menarche are neither definitive nor universal (15). According to an NIH study, less than 10% of young women in the United States experience their first period before 11 years (16). Some factors that may contribute to this early occurrence are nutrition, race/ethnicity, socioeconomic status, environmental stress, and genetics (17). Little data have been collected regarding the long-term consequences associated with early age of menarche. Research suggests that early age of menarche is associated with increased risk for Type 2 diabetes (18). The University of Queensland found that the age at which girls start menstruating could predict an increased risk of developing diabetes during pregnancy (19). Early menarche has been associated with both the development of Type 2 diabetes and GDM, but these associations are not well-studied (5). Due to the similarities in causative factors and the lack of research available, we aimed to investigate the correlation between the early onset of menarche and the later development of GDM in women living in the United States. We investigated these similarities using secondary data by utilizing secondary data from the National Health and Nutrition Examination Surveys (NHANES) Reproductive Health Questionnaire. Furthermore, identifying certain ethnic and racial groups associated with early menarche and GDM will reveal communities most at risk. With new knowledge about at-risk communities with a predisposition to GDM, standardized guidelines and preventative measures can be implemented to better manage the long-term consequences of gestational diabetes.

The Association Between Early Menarche and Gestational Diabetes: A Secondary Analysis

Methods Participants Questionnaires were used to obtain data. The topics of interest, assessed using a combination of scaled and dichotomous questions, included reproductive health, history of pregnancy and menstruation, hormone therapy and a myriad of other reproductive conditions in relation to the subject. To ensure accuracy, certain populations were oversampled to gather a representative sample of the civilian and noninstitutionalized population of the United States (21). For the Reproductive Health Questionnaire, females aged 12 years and older were included in the final data set. Variables in relation to hysterectomy and pregnancy at the time of survey were excluded from the data for women aged 12 to 19 and over 44 years old for disclosure purposes. The sample size is representative of certain populations in the United States. Procedures NHANES Reproductive Data and NHANES Demographic Data sets for years 2013 to 2018 were downloaded via the statistical software, SAS (Statistical Analysis System). The data were then imported into Microsoft Excel to be filtered. Code 777 (refused to answer) and code 999 (don’t know), were removed from the data. Answer choices of participants who experienced menarche after the age of 12 were also excluded. Additionally, women who had not started menarche at the time of the questionnaire were also excluded. Answer responses from male participants were removed from the data. Filtered data was imported into statistical software, SPSS (Statistical Product and Service Solutions) to perform data analysis. In SPSS, the data of each year interval was merged into a single data file. The following variables were analyzed from the NHANES Reproductive data: RHQ010, age of menarche; RHQ162, participant told they had diabetes while pregnant; RHQ163, participant age of gestational diabetes diagnosis; and RHQ173, age when delivered baby 9 pounds or more. The following variables were analyzed from questionnaire data: RIAGENDR, gender of participant; RIDRETH3, ethnicity of participants was set to a nominal measure. RHQ010, RHQ163, and RHQ173 were set to a scale measure, where RHQ162 was set to a nominal measure. When necessary, the code/value of each variable was labeled. Values that were missing or would confound the data analysis were entered via the missing tab. The data values that were specified as user-missing were flagged for special treatment and were excluded from most calculations. The NHANES Reproductive Health Questionnaire and Demographics Data from 2013 to 2018 was used. The target population was residents of the United States who ranged from 12 to 50 years of age. Inmates or individuals who were part of the United States Armed Forces were not questioned. The first method of data collection was at the Mobile Examination Center (MEC), where participants were interviewed by a trained MEC interviewer using questionnaires: ComputerAssisted Personal Interview (CAPI) Questionnaire and AudioComputer-Assisted Self-Interview (ACASI) Questionnaire. The second method of data collection was household interviews. For this method, a trained interviewer screened for eligible participants at their doorstep using the Screener Questionnaire

(20). Next, the following questionnaires were used: Relationship Questionnaire, Family Questionnaire, Sample Person Questionnaire. Both the MEC and household interviews utilized electronic questionnaire forms to record the data electronically and sent to the central survey database system (20). Every year an average of 5,000 people of all ages are recruited and interviewed (4). NHANES uses a complex, multistage probability design to sample the population in four stages through a top-down approach. First, Primary Sampling Units (PSUs) are selected, which are counties or groups of counties in the United States. Then, segments within PSUs are selected that make up a block or groups of blocks that contain a cluster of households. From those segments, individual households are chosen, and finally individuals from those households are selected for the interviews (20). Data analysis A cross tabulation was utilized to evaluate the categorical occurrence of gestational diabetes between the abovementioned ethnic groups in SPSS. An ANOVA test was used to compare the following factors: age of menarche between 9 and 12 compared to age of gestational diabetes diagnosis. A Pearson correlation was used to investigate the associations between age of menarche between 9 and 12 vs age of gestational diabetes diagnosis, and age when delivered baby 9 pounds or more in comparison to age of gestational diabetes. Data for age of menarche in relation to age of gestational diabetes diagnosis and age of menarche assessed against ethnicity was graphically depicted to understand the prevalence of early menarche and any association it had with the development of gestational diabetes between ethnic groups. Age of menarche when compared to age when participants delivered a baby weighing 9 pounds or more was not represented graphically to depict understanding but was represented with a bivariate correlation test. The potential relationship between age of menarche and GDM diagnosis was assessed using a Pearson correlation test. A cross tabulation was performed to evaluate the pattern among ethnicity and age of first menarche. Throughout the duration of 2013 to 2018, the occurrences of menarche between ages 9 and 12 per ethnicity were calculated as percentages. The percentages were imported into Prism to formulate a donut chart to graphically depict the prevalence of the different ratios to compare the data of ethnicity and age of menarche.

Results Figure 1 represents the comparison of age of menarche and mean age of GDM diagnosis. Figure 2 shows the comparison between age of menarche and ethnicity. The results depicted are for years 2013 to 2018. In Figure 2A, 27% of the 309 women who had their first menstrual period at age 9; most of the demographic was white women. In Figure 2B, 31% of 525 women experienced their first menstrual period at age 10. For Figure 2C at age 11, 34% of 1,363 women, and Figure 2D 36% of 2,488 women who were 12 years of age also showed that white women were most of the overall total. As shown in all figures, the demographic with the highest percentage was white women. Black and Mexican American women follow with the second and third highest percentages, respectively. 155

The Association Between Early Menarche and Gestational Diabetes: A Secondary Analysis

Figure 1. A one-way analysis of variance, ANOVA showed that the relationship between age of menarche and whether there was a GDM diagnosis at pregnancy was significant at the 0.05 p-level, (F (3,271) = 3.099, p = 0.027).

Discussion A one-way ANOVA test was conducted to compare the means of age of menarche and diabetes diagnosis and to determine whether there were any statistically significant differences between the means of the two variables. The null hypothesis of this test was there is no statistical significance between age of menarche and diabetes diagnosis while pregnant. After completing the ANOVA test, there is enough evidence to reject the null hypothesis. A p-value of 0.027 signifies the strength of the relationship between age of menarche and whether a diabetes diagnosis was given during pregnancy. This strengthens and emphasizes analyses that suggests that younger age at menarche is associated with higher risk of Type 2 diabetes (24). A Pearson correlation was computed to assess the relationship between age of menarche (RHQ010) and if the patient was told they had diabetes in pregnancy (yes/no) (RHQ162). The results of the Pearson correlation indicated a level of significance. There was a positive, but weak correlation between age of menarche and age of GDM diagnosis, (r = 0.042, n = 4,706, p = 0.022). Correlation is significant at the 0.05 level (2-tailed). The Pearson correlation of 0.042 is weak yet positive, so this suggests that there is a statistically significant correlation. Additionally, the p-value of 0.022 suggests significance and to reject the null hypothesis. This weak but positive correlation supports the hypothesis that as women experienced menarche at an earlier age, the more likely they were to develop GDM in pregnancy. This is due to the ANOVA test suggesting significance. Previous studies have observed an association between early age of menarche and increased risk of GDM (13). Race/ethnicity, socioeconomic status, environmental stress, and genetics may contribute to this association (17). However, the weak correlation may be due to the fact that survey data was utilized to perform this study. Participants enrolled in the study may not have answered accurately as some may not remember when they experienced menarche. Additionally, menarche is a taboo and sensitive subject in some cultures, which may influence how the women answered question RHQ010. Another Pearson correlation was computed to assess the relationship between age told they had diabetes while pregnant (RHQ163) and age when the participant delivered a baby

156

Figure 2. Ethnic representation of ages of menarche for 2013-2018 NHANES Data. Age (A) 9, (B) 10, (C) 11, and (D) 12.

The Association Between Early Menarche and Gestational Diabetes: A Secondary Analysis

9 pounds or more (RHQ173). The results of the Pearson correlation indicated a level of significance. There was a positive, but weak correlation between age of menarche and age of GDM diagnosis, (r = 0.621, n = 76, p = <0.001). Correlation is significant at the 0.01 level (2-tailed). The Pearson correlation of 0.621 suggests a strong positive association between the age women were told they had diabetes while pregnant and the age when delivering a baby 9 pounds or more. The results were expected based on other studies. This also depicts the significant public health implications of pregnant women diagnosed with GDM. Previous studies have noted that the risk of developing GDM is 2 and 4 times higher among overweight and obese women, respectively, when compared to normalweight pregnant women (24). Maternal obesity rates can result in negative lifetime outcomes for both the mother and the fetus. Maternal obesity also increases the risk for a number of pregnancy complications like gestational hypertension, maternal hypertensive disease or pre-eclampsia, risk of emergency cesarean delivery, and prolonged delivery times (7). As a result of these various complications, the neonate is at risk for macrosomia or increased body fat.

Social desirability is the inclination to present one's reality and experiences in a socially acceptable manner by embellishment or omission (22). Participants may alter the facts to portray themselves in a positive manner, either to appeal to the interviewer or to detract from revealing unfavorable details. Additionally, recall bias played a significant role in this analysis. Recall bias, also referred to as telescoping, is the phenomenon of misattributing occurrences to earlier or later dates (23). Mistakenly recalling an event as having happened more recently than it did is known as forward telescoping. Backward telescoping describes the opposite bias — erroneously remembering an event as having occurred earlier than it did (24). A participant’s skewed perception of time could lead to altered results when investigating monumental, age-based life changes, such as menarche. These limitations, while intractable, can be minimized by employing data collectors that reflect the ethnicities of the target population, instructing them to utilize neutral language, and have questions that assess different time in various ways, i.e., age in years, grade in school, or calendar year.

Additionally, early age of menarche and mean age of GDM diagnosis was also compared (Figure 1). The ANOVA statistical data showed that the relationship between age of menarche and whether there was a GDM diagnosis at pregnancy was significant at the 0.05 p-level, F (3,271) = 3.099, p = 0.027. The purpose of this comparison was to see if developing GDM diagnosis at a later age was associated with early age of menarche, as previous studies have shown that early age of menarche is linked to increased risk of GDM (26). The results shown in Figure 1 depict a significant positive correlation, suggesting that pregnant women who were diagnosed with GDM at a later age were likely to have experienced menarche at a younger age. This suggests that age at which GDM diagnosis occurred in pregnancy may be correlated with earlier age of menarche. These results were obtained from subjective survey data, which is a limitation. Future studies should examine and confirm this finding.

Conclusion

In Figure 2, early age of menarche (9–12) was compared to ethnicity for years 2013-2018. Results showed that a higher percentage of Non-Hispanic white women experienced early menarche for each age studied (9–12). This was followed by Non-Hispanic African American and Mexican American women. These results were not expected, as previous studies have shown that the age of menarche of non-Hispanic African American girls was significantly earlier than non-Hispanic white and Mexican American women (16). The limitations are also due to survey data being utilized. Future studies should focus on data that is collected in ways that minimizes bias and uncertainty. The main limitation of this study was the method of instrumentation. The data utilized in the analysis came from a survey database. The data collection relied entirely on self-reported information. The CDC employees screened all participants thoroughly to ensure eligibility in the study prior to providing them with the extensive survey. However, the process of selection is also based on subjective responses. From beginning to end, the method of instrumentation has the potential to introduce uncertainty through social desirability and recall bias.

The results of our study warrant future research, which may include looking deeper into the participant demographics from the NHANES sample surveys and other questions. Other questions on the reproductive survey may also be included as a comparison to the questions that related to gestational diabetes, such as family history of gestational diabetes, family history of early menarche, and diagnosis of obesity. To further dive into this topic, a different survey could be added in addition to the NHANES to introduce variety and new perspective to the mix of data already incorporated. Our analysis suggests that younger age at menarche is associated with higher risk of Type 2 diabetes, and as women experience menarche early on, they are more likely to develop GDM in pregnancy. There is a strong positive association between age women were diagnosed with diabetes while pregnant and age when delivering a baby 9 pounds or more; along with a higher percentage of NonHispanic white women experienced early menarche for each age studied.

Appendix NHANES variables: RHQ010, age of menarche RHQ162, participant told they had diabetes while pregnant RHQ163, participant age of gestational diabetes diagnosis RHQ173, age when delivered baby 9lbs or more RIAGENDR, gender of participant RIDRETH3, ethnicity of participants

Acknowledgments We thank Brian J. Piper, PhD, MS, along with Elizabeth Kuchinski BS, MPH, and the teaching assistants Jonique Depina, MS, and Yasmin Mamani, MS, for their time and devotion to our paper and for providing us with the necessary guidelines needed.

157

The Association Between Early Menarche and Gestational Diabetes: A Secondary Analysis

Disclosures There is no financial relationship between this paper’s authors and any health care related institutions mentioned.

References

12. Tobias DK, Stuart JJ, Li S, Chavarro J, Rimm EB, RichEdwards J, et al. Association of History of Gestational Diabetes With Long-term Cardiovascular Disease Risk in a Large Prospective Cohort of US Women. JAMA Intern Med [Internet]. 2017 Dec 1;177(12):1735–42. Available from: https://doi.org/10.1001/jamainternmed.2017.2790

Association AC. Gestational Diabetes Mellitus. Diabetes Care [Internet]. 2004 Jan 1;27(suppl 1):s88 LP-s90. Available from: http://care.diabetesjournals.org/ content/27/suppl_1/s88.abstract

U.S. and World Population Clock [Internet]. Population Clock. Available from: https://www.census.gov/popclock/

Countrymeters.info. United States of America (USA) Population [Internet]. Countrymeters. Available from: https://countrymeters.info/en/United_States_of_America_ (USA)

NHANES 2017-2018 Overview [Internet]. Centers for Disease Control and Prevention. Centers for Disease Control and Prevention; 2020. Available from: https:// wwwn.cdc.gov/nchs/nhanes/continuousnhanes/overview. aspx?BeginYear=2017

Hedderson MM, Darbinian JA, Ferrara A. Disparities in the risk of gestational diabetes by race-ethnicity and country of birth. Paediatr Perinat Epidemiol [Internet]. 2010 Sep;24(5):441–8. Available from: https://pubmed.ncbi.nlm. nih.gov/20670225

Leddy MA, Power ML, Schulkin J. The impact of maternal obesity on maternal and fetal health. Rev Obstet Gynecol [Internet]. 2008;1(4):170–8. Available from: https:// pubmed.ncbi.nlm.nih.gov/19173021

Farahvar S, Walfisch A, Sheiner E. Gestational diabetes risk factors and long-term consequences for both mother and offspring: a literature review. Expert Rev Endocrinol Metab [Internet]. 2019 Jan 2;14(1):63–74. Available from: https:// doi.org/10.1080/17446651.2018.1476135

Catalano PM. The impact of gestational diabetes and maternal obesity on the mother and her offspring. J Dev Orig Health Dis [Internet]. 2010 Aug;1(4):208–15. Available from: https://pubmed.ncbi.nlm.nih.gov/25141869

18. He C, Zhang C, Hunter DJ, Hankinson SE, Buck Louis GM, Hediger ML, et al. Age at Menarche and Risk of Type 2 Diabetes: Results From 2 Large Prospective Cohort Studies. Am J Epidemiol [Internet]. 2010 Feb 1;171(3):334– 44. Available from: https://doi.org/10.1093/aje/kwp372

Bellamy L, Casas J-P, Hingorani AD, Williams D. Type 2 diabetes mellitus after gestational diabetes: a systematic review and meta-analysis. Lancet [Internet]. 2009;373(9677):1773–9. Available from: https://www.sciencedirect.com/science/article/pii/ S0140673609607315

19. NHANES 2017-2018 Questionnaire Data Overview [Internet]. Centers for Disease Control and Prevention. Centers for Disease Control and Prevention; 2020. Available from: https://wwwn.cdc.gov/nchs/nhanes/ continuousnhanes/overviewquex.aspx?BeginYear=2017

10. Tobias DK, Hu FB, Forman JP, Chavarro J, Zhang C. Increased risk of hypertension after gestational diabetes mellitus: findings from a large prospective cohort study. Diabetes Care [Internet]. 2011/05/18. 2011 Jul;34(7):1582–4. Available from: https://pubmed.ncbi.nlm. nih.gov/21593289 11. Ajala O, Jensen LA, Ryan E, Chik C. Women with a history of gestational diabetes on long-term follow up have normal vascular function despite more dysglycemia, dyslipidemia and adiposity. Diabetes Res Clin Pract [Internet]. 2015;110(3):309–14. Available from: https://www.sciencedirect.com/science/article/pii/ S0168822715004040 158

13. Lacroix AE, Gondall H, Langaker MD. Physiology, Menarche. [Updated 2020 Aug 24]. In: StatPearls [Internet].Treasure Island (FL) : StatPearls Publishing; 2021 Jan-. Available from: https://www.ncbi.nih.gov/books/ NBK470216/ 14. Anderson SE, Must A. Interpreting the Continued Decline in the Average Age at Menarche: Results from Two Nationally Representative Surveys of U.S. Girls Studied 10 Years Apart. J Pediatr [Internet]. 2005 Dec 1;147(6):753– 60. Available from: https://doi.org/10.1016/j. jpeds.2005.07.016 15. Adrien Gaudineau, Virginie Ehlinger, Christophe Vayssiere, Beatrice Jouret, Catherine Arnaud & Emmanuelle Godeau. Factors associated with early menarche: results from the French Health Behaviour in School-aged Children (HBSC) study [Internet]. Biomedcentral.com. 2010 [cited 2021 Mar 22]. Available from:https://bmcpublichealth. biomedcentral.com/articles/10.1186/1471-2458-10175#:~:text=We%20defined%20early%20menarche%20 as,years%20%5B9%E2%80%9311%5D 16. Chumlea WC, Schubert CM, Roche AF, Kulin HE, Lee PA, Himes JH, et al. Age at menarche and racial comparisons in US girls. Pediatrics. 2003;111(1):110–3.https://pubmed. ncbi.nlm.nih.gov/12509562/ 17. Posner RB. Early Menarche: A Review of Research on Trends in Timing, Racial Differences, Etiology and Psychosocial Consequences. Sex Roles [Internet]. 2006;54(5):315–22. Available from: https://doi. org/10.1007/s11199-006-9003-5

20. Somers EC, Ganser MA, Warren JS, Basu N, Wang L, Zick SM, et al. Mercury exposure and antinuclear antibodies among females of reproductive age in the United States: NHANES. Environ Health Perspect. 2015;123(8):792–8 Available from: https://ehp.niehs.nih.gov/doi/pdf/10.1289/ ehp.1408751 21. Bergen N, Labonté R. “everything is perfect, and we have no problems”: Detecting and limiting social desirability bias in qualitative research. Qual Health Res. 2020;30(5):783–92. 22. El Haj M, Janssen SMJ, Antoine P. Memory and time: Backward and forward telescoping in Alzheimer’s disease. Brain Cogn. 2017;117:65–72.

The Association Between Early Menarche and Gestational Diabetes: A Secondary Analysis

23. Telescoping [Internet]. Sagepub.com. [cited 2021 Mar 29]. Available from: https://methods.sagepub.com/reference/ encyclopedia-of-survey-research-methods/n579.xml 24. Chu SY, Callaghan WM, Kim SY, Schmid CH, Lau J, England LJ, et al. Maternal Obesity and Risk of Gestational Diabetes Mellitus. Diabetes Care [Internet]. 2007 Aug 1;30(8):2070 LP – 2076. Available from: http://care.diabetesjournals.org/ content/30/8/2070.abstract 25. Diabetes.org. [cited 2021 Mar 30]. Available from: https:// www.diabetes.org/diabetes/gestational-diabetes/how-willthis-impact-my-b 26. Li H, Shen L, Song L, Liu B, Zheng X, Xu S, et al. Early age at menarche and gestational diabetes mellitus risk: Results from the Healthy Baby Cohort study. Diabetes Metab. 2017;43(3):248–52. Available from. https://pubmed.ncbi. nlm.nih.gov/28161369/

159

Scholarly Research In Progress • Vol. 5, November 2021

Review of Ketamine as a Rapid Antidepressant for Treatment-Resistant Depression Sankung X. Darboe1*‡, Peter J. Koszuta1*‡, Paul W. Lee1*‡, and Mannaa I. Mannaa1*‡ ¹Geisinger Commonwealth School of Medicine, Scranton, PA 18509 *Master of Biomedical Sciences Program ‡ Authors contributed equally Correspondence: mmannaa@som.geisinger.edu

Abstract The anesthetic agent ketamine has been under investigation as a potential antidepressant for two decades. Animal studies and subsequent randomized control trials for patients have provided extensive evidence of ketamine’s rapid antidepressant outcomes both structurally at the receptor level as well as behaviorally. A review of current literature hones in on the putative rapid mechanisms of action of ketamine, which place the drug ahead of nearly all current antidepressants for treatment-resistant depression (TRD) patients. Clinical investigations, safety, and ethical implications are highlighted along with the need for reduced treatment cost. Finally, to more fully understand the effects of long-term use, efficacy in other types of psychiatric disorders, and interactions with comorbidities, more extensive trials are warranted.

Introduction Major depressive disorder (MDD) is a major psychiatric problem worldwide, reportedly affecting 12% of males and 20% of females in the United States alone (1, 2). The etiology of MDD is not well known since there are various neurobiological causes, such as glutamatergic transmission defects in the central nervous system, as well as social and environmental factors that may lead to the stress levels of patients suffering from MDD (3, 4). In patients suffering from MDD, glutamatergic transmission is believed to be homeostatically downregulated by rapid-acting antidepressants (3). While several pharmacological treatments have proven successful in treating MDD, many patients with serious cases experience TRD with existing treatment approaches, with long-term effectiveness lagging after several trials, and 30% of patients continue to experience depression after a short duration of effective treatment (1, 4, 5). Usually, these patients are diagnosed with TRD after not successfully responding to at least two different antidepressant treatments (6). Most of the antidepressants used for MDDs have been shown to function in monoaminergic pathways, but it normally takes weeks to months before any meaningful therapeutic effect is generated (4, 7). Due to its rapid biological effects and effectiveness, ketamine has attracted attention in neuropharmacology over the past few decades. Ketamine was originally FDA-approved as a rapid-acting anesthetic in 1970 and evidence of its antidepressant action also began to emerge in the 1970s as preclinical studies with sub-anesthetic doses of the drug showed similar mechanisms of action to antidepressants at much rapid response rates (8). Ketamine is not approved by the FDA for depression, but its enantiomer, S-ketamine (esketamine), was FDA-approved in

160

2019 as it was found to be a more potent N-methyl-D-aspartate (NMDA) receptor antagonist with favorable adverse effects compared to (R)-ketamine and showed lower variability in pharmacokinetics and pharmacodynamics in various patients (8, 9, 10). Esketamine was a more potent mu-opioid receptor agonist and a less potent sigma-opioid receptor agonist (10). In addition to its rapid effects, ketamine has also been reported to rapidly reverse anhedonic behavior and synaptic defects in animal models studies with chronic unpredictable stress (CUS) (11). Although the effects of ketamine on anhedonia in humans remains to be determined, these findings highlight the efficacy advantages of ketamine compared to current antidepressants that are relied upon for treatment (11, 12). Moreover, this review is based on the hypotheses on the mechanism of action of ketamine and how sub-anesthetic doses of ketamine can modulate homeostatic plasticity in the treatment of MDD. Homeostatic plasticity modulates strong neuronal connectivity as a composite of global plasticity and synaptic scaling. In patients with MDD, achieving and maintaining homeostatic plasticity becomes paramount (13). MDD patients are identified as having downregulation of synapses of the prefrontal cortex (PFC) and hippocampal gray matter due to neuronal excitotoxicity resulting from prolonged stress (13). Inflammatory cytokines and neurotrophins are among factors that mediate this downregulation (14, 15). Prolonged stress-associated neuronal atrophy and synaptic depression have been linked to disruption of the glutamatergic system and downstream excess of extracellular glutamate (13, 16). Neuronal atrophy and synaptic depression are encompassed by dysfunctional synaptic strength, reduced dendritic spine density, retraction of spines and reduced dendritic branching of the prefrontal cortex (PFC) (17).

Methods The literature review was conducted by accessing multiple databases, including primarily but not limited to PubMed, PubPsych, and Google Scholar. Most data, results, and perspectives were obtained from recent publications of research journals, some a mix of primary and secondary data. Some data were obtained from clinicaltrials.gov, and some clinical perspectives were obtained by blogs of medical schools and physician-writers. The time frame was principally comprised of the ten years since 2011, but older sources were relevant and reliable, and therefore included. The search was conducted between January and March 2021. Search terms include “ketamine,” “esketamine,” “depression,” “major depression,” “suicidal ideation,” “pharmacology,” “pharmacodynamics,” “clinical trial,” “randomized controlled trial,” and others. Inclusion criteria included articles or reports published in reputable

Review of Ketamine as a Rapid Antidepressant for Treatment-Resistant Depression

sources, emphasizing the last decade of research, clinical use, any attitudes toward ketamine, and the effects of ketamine. Exclusion criteria included unverified authors, discernable bias, unsubstantial discussion of results, incompletely available articles, and data never achieving statistical significance.

Discussion Pharmacodynamics It has been hypothesized that the mechanism of action of sub-anesthetic dosing of ketamine lies in three primary actions followed by important downstream actions. It is believed that rapid improvement in homeostatic plasticity is accomplished by inhibition of NMDARs on GABAergic interneurons, extrasynaptic NMDAR inhibition, activation of post-synaptic a-amino-3-hydroxy-5-methyl-4-isoxazole-propionic acid receptors (AMPARs), and blockage of spontaneous (NMDAR) activation (18). Vital downstream actions include increased BDNF and protein translation by mechanistic target of rapamycin complex 1 (mTORC1) (19-21). The pharmacodynamics of ketamine show both its enantiomers (R)-ketamine and (S)-ketamine to influence antidepressant actions. (S)-ketamine is a more potent NMDA antagonist; however, (R)-ketamine shows more potent antidepressant effects in rodents and longer lasting action in neonatal dexamethasone-treated pediatric depression animal model, chronic social defeat stress animal model, and learned helplessness animal model of depression (22–25). (R)-ketamine also shows greatly reduced psychotomimetic and dissociative side effects (26, 27). Among ketamine’s primary actions, it is hypothesized that it preferentially inhibits NMDARs on GABAergic interneurons, a mechanism unique to ketamine (28). It also has a high affinity for forebrain inhibitory interneurons expressing GluN2D-NMDAR subunits (29–32). The resulting overall decrease of inhibition is proposed to lead to the disinhibition of pyramidal cells facilitating bursts of excitatory glutamatergic neurotransmission in the medial prefrontal cortex (mPFC) (33, 34). Two-photon imaging of prelabeled layer V medial PFC pyramidal neurons post-ketamine administration demonstrated that ketamine increases synapse and spine formation. They confirmed rapid increase in spine number and morphology in distal and proximal apical tufts at 24 hours post ketamine administration (35). The bursts of neurotransmission further evoke the release of glutamate facilitating synaptic glutamatergic neurotransmission. Activation of post-synaptic α-amino-3-hydroxy-5-methyl-4isoxazolepropionic acid receptors (AMPARs) via this evoked release is vital for synaptic potentiation and plasticity (36). AMPARs are transmembrane glutamatergic receptors working ionotropically for transduction of fast synaptic neurotransmission in the brain (37). Ketamine has been shown to enhance AMPAR synaptic transmission in the mPFC and hippocampus through upregulation of receptors and increased phosphorylation of at least the GluA2 subunit (38). Increased levels of the GluA1 unit of the AMPA receptor were noted two hours post ketamine administration (39). As AMPARs conduct Na+ and CA²+ into the cell, the resulting local increase of intracellular Ca²+ signals increases vesicular delivery of BDNF into the synaptic space. The downstream effect is activation of

mTOR up-regulation of protein synthesis and synaptic plasticity (40). The result was significant enhancement of synaptogenesis and connectivity in the hippocampus and PFC (41). Extra-synaptic GluN2B-NMDARs are believed to be specifically inhibited by ketamine which prevents extracellular ambient glutamate activation of the NMDAR receptors. Inhibition of extra-synaptic NMDARS shuts down the NMDAR-regulation of mTORC1 which results in de-suppression of protein synthesis important for synaptic homeostasis (42–45). This specific mTOR pathway activation is associated with increased synaptic spine density within the mPFC (46). Ketamine has been shown to block NMDAR-mediated miniature excitatory postsynaptic currents (mEPSCs) where spontaneous glutamate neurotransmission at rest regulates synaptic strength and suppresses protein synthesis (23). mEPSCs promote phosphorylation and inactivation of eukaryotic elongation factor 2 (eEF2). Ketamine blocks this transmission facilitating active eEF2 promotion of synaptic plasticity as well as potentiation in the CA1 region of the hippocampus resulting in behavioral antidepressant actions (23, 47, 48). In addition to the most prominent target, NMDA receptors, other low-affinity targets include γ-amynobutyric acid (GABA), dopamine, serotonin, sigma, opioid, and cholinergic receptors and others (49). Pharmacokinetics Ketamine can be administered via several routes. This is due to its water and lipid solubility. The most common routes are oral, inhalation, intramuscular, subcutaneous, intravenous, epidural, or intrathecal (50, 51). Bioavailability is up to 24% oral, approximately 93% intramuscular, approximately 30% rectal, approximately 45% intranasal, and 77% epidural (50). Due to its lipid solubility and low protein binding, distribution is 160–550 L/70 kg (50). Steady-state plasma concentration for anesthetic purposes is 2,200 ng/ml, awakening concentration range is 640 to 1,100 ng/ml (52). Ketamine is distributed to highly perfused tissues including the brain with 10% plasma protein binding, which facilitates distribution across the blood-brain barrier (50, 53). Plasma concentrations of both enantiomers are equal after one minute post IV administration (54). Cytochrome P450 enzymes play a major role in the metabolism of ketamine via N-demethylation to metabolite norketamine (R,S)-norKET (55). The main ketamine metabolizer, CYP2B6, demethylates both enantiomers equally. CYP3A4 demethylates (S)-ketamine at a higher rate than (R)-ketamine (55). (R,S)norKET hydroxylation is enacted by several enzymes including CYP2A6 or CYP2B6 to produce, among others, (2R,6R;2S,6S)hydroxynorketamine (HNK) and (R,S)-dehydroxonorketamine (DHNK). CYP2A6 is responsible for direct hydroxylation of ketamine to (2R, 6R;2S,6S)-HK (56). Peak concentrations in the plasma of (R)- and (S)-ketamine and norketamine, and (2R,6R;2S,6S)-HNK, were found to be 1.33 hours, and 3.83 hours respectively (57). High polymorphism of the main ketamine metabolizer, CYP2B6, has clinical significance (58). This is due to the diminished metabolic ketamine N-demethylation activity with CYP2B6 polymorphism. The order of effective metabolism based on genotype is wild-type CYP2B6.1, then CYP2B6.4, then CYP2B6.26, CYP2B6.19, CYP2B6.17, and CYP2B6.6. The variant CYP2B6.9 is up to 35% that of the wild type, and

161

Review of Ketamine as a Rapid Antidepressant for Treatment-Resistant Depression

CYP2B6.16 and CYP2B6.18 are inactive (58). Consequences for ketamine pharmacokinetics, therapeutic effects, and elimination must be assessed in those with variants other than wildtype (58). Administered antidepressant dose of ketamine shows no detectable plasma levels within 1 day. This is due to ketamine’s short elimination half-life for both (R)- and (S)-ketamine demonstrated at 155 minutes (49). Far shorter times have been noted for children with elimination half-lives 50% that of adults (53). However, (2R,6R;2S,6S)-HNK have been shown to remain in circulation up to 3 days post-dosing (57). Ketamine elimination is primarily (~80%) via urine and bile as glucuronic acid-labile conjugates of HK and HNK (59). The remainder is eliminated via urine as ketamine and norketamine both at 4%, and DHNK at 16% (53, 59-62). Clinical investigations The leading cause of disability worldwide is major depressive disorder (63, 64). Between 1999 and 2016, suicide rates rose over 30% in 25 U.S. states (64), and of patients with major depressive disorder (MDD), approximately a third do not respond to existing antidepressants despite enduring the weeks of medicating needed to see an effect (63, 65). A clear unmet need exists for efficacious treatment for depression, including suicidal ideation (SI) (63, 66, 67). Few therapies act within a week, but emergent SI and MDD demand quick and superior treatments (63, 68). Ketamine administered by IV is typically a racemic mixture of both enantiomers, and isolated esketamine (S-ketamine) is used as a nasal spray, both FDA approved as an anesthetic (64, 69). Multiple systematic reviews support the efficacy of ketamine against MDD, bipolar depression (BD), post-traumatic stress disorder (PTSD), acute suicidal ideation and treatmentresistant depression (TRD) (63, 7, 67, 70). Ketamine is primarily for use in MDD and TRD (63, 71, 72). Use as a rapid-acting antidepressant has been investigated for over a decade (73) and currently constitutes off-label use (68). TRD is characterized by lax symptom relief by traditional antidepressants such as SSRIs (63). Ketamine given by single IV dose (0.5mg/kg) results in 50%-70% response in TRD (63). Systematic reviews report significant relief in as little as 15 minutes (74,75) to 2 hours (63, 68, 69, 76) lasting up to 2 weeks from a single dose, 11% reporting relief on day 14 (63, 7, 66, 74). A randomized control trial confirmed significant effects of ketamine at 4, 24, 48, and 72 hours post administration, and was superior to midazolam at reducing SI at every interval (74). IV treatment lasts 2 hours (77). Notably, 75–80% of patients experiencing regular depression improve on ketamine, compared to 35–40% on traditional medications (77). SI on the hopelessness scale reduced 90.7% at 3 days post infusion (75). Use is documented for opioid-induced hyperanalgesia by reducing pain threshold (78). Ketamine works for an apparent shorter duration in BD (63, 71) but predicts efficacy in patients with high BMI, family history of alcohol use, and anxious depression (76). Intramuscular and sublingual administration is documented with lesser improvement of MDD (71). Importantly, it is not necessary to wash out other antidepressants (79). Proof of concept has been verified (79) and observed over a 3-year period (80). Emergency medicine

162

warrants the need for rapid and severe reduction in acute SI (70). In repeated administration, the magnitude of response after 6 infusions was predicted by response 4 hours after first infusion (80). Most patients responded prior to third infusion (68). Effects are durable over treatment (70), yet it is unclear how often ketamine should be administered (66) and how effects diminish following cessation of treatment (81). Ketamine is carried at hospitals and increasing numbers of outpatient clinics (66). The most effective alternative for TRD is electroconvulsive therapy (81). SSRIs are commonly associated with weight gain, but this is not a side effect of ketamine (77). Midazolam improved 30% of patients compared to over 55% of patients on ketamine in one single-blind study (74). Psychotherapy is less reliably effective and requires some patient cooperation, complicated by instances of suicidal ideation (67). Safety Not all patients respond to ketamine (68, 71). Off-label use for depression is due to piecemeal data with unverified accuracy (68). The main drawbacks are side effects (82). Patients primarily report disturbances, dissociations, or abnormal sensations (83). Other adverse effects include increased heart rate and blood pressure up to four hours following administration, dizziness, headache, nausea, dry mouth, and restlessness (82). Long-term use is associated with mild cognitive disturbances and urinary cystitis, yet may be confined to daily users (68). Transient effects typically subside within an hour (7). Anesthetic doses, which are higher than those for depression, can cause hepatotoxicity (84). Systematic reviews do not report mania but include mild talkativeness after ketamine (83). Ketamine may augment other antidepressants without increasing observed adverse reactions (66). Adverse reactions are twice as likely with intranasal esketamine versus placebo (66). Patients with cardiovascular illness should be carefully monitored (63). This drug should be avoided in patients with status epilepticus due to sympathomimetic activity (76). Rapid onset requires readiness of supportive measures if needed (72). Ethics Any new therapy encourages discussions of ethics surrounding efficacy and potentially uncertain risk and benefit analyses. Because ketamine is not officially approved for all disorders it may effectively treat, proper professional use is vital to maintain availability of ketamine for these cases, and not prohibited (81). Because physicians may want to appear progressive, and because patients may not be of sound mind when presented with the option to receive ketamine, physicians need to respectfully balance their autonomy and patient risks and benefits (68). Potential patients likely present with an inhibited decision capacity due to their condition, thereby relying on decisions of medical staff, which could lead to conflict with the recovered patient (68). Responsible ethical decision-making is essential. Despite being called a breakthrough, cost is still high, and a burden to some patients, particularly unwanted if ketamine was administered when patients were in altered mental states (85).

Review of Ketamine as a Rapid Antidepressant for Treatment-Resistant Depression

Limitations Many clinical investigations into ketamine for depression are small in scale and target specific clinical endpoints. Systematic reviews compile data into significant reports with reproducible conclusions, but robust experimental data is sparse, conferring uncertainty about specific anti-suicidal properties of ketamine (70), long-term effects (83), usage beyond 12 doses and habituation (86), addiction potential (66), if ketamine should be avoided in patients with psychoses (73) or head injury (76), effects in special populations (7), timing of adverse effects (81), and mechanism in humans (82). Ketamine is a generic drug, so investment in expensive trials is unlikely and unprofitable, though required for approval (68). Recreational use complicates dosing (80, 86). Professional use may suit military, cancer patients, Alzheimer’s patients (65), and end-of-life care, especially in patients with MDD and pain (87). Ketamine is not covered by insurance; each treatment costs $300 to $450, with estimates reaching thousands per session, though coverage may soon expand (85, 77). Patients and physicians should understand knowledge of ketamine is subpar, though great hope exists in the anti-suicidal properties of ketamine (68).

Conclusion The need for a more rapid response to TRD exists and is compounded by the imminent dangers associated with suicidal ideation, major depressive disorder, and bipolar depression, as well as the tendency for these conditions to recur (8890). Non-ketamine treatments can take up to 8 weeks to become fully effective, while ketamine treatments can reach full effectiveness within 24 hours (88, 89). Due to its rapid biological effects and effectiveness, ketamine has attracted attention in neuropharmacology over the past few decades to become the future of rapid pharmacotherapy for TRD patients (4, 7). However, despite the apparent advantages of ketamine treatments, there are still noteworthy concerns. Different studies have indicated different success rates for ketamine treatment against TRD, with two examples being 71% and 65% (91, 92). While these success rates could be considered strong, they still leave the need for other treatments for the remaining 29% and 35% of patients respectively (91, 92). Those taking ketamine also face a variety of safety concerns, commonly including dizziness, vertigo, nausea, vomiting, anxiety, numbness, and high blood pressure (90). The cost of treatment could also be an issue for many patients; the first month of esketamine under the brand Spravato costs $4,800–6,800 with $1,200–3,600 per month after as of mid-2020 (90). Furthermore, the mechanism of action for these treatments is uncertain. Several possibilities exist, mostly based on antagonism of NMDA receptors and/or enhancement of AMPA receptors (28, 38). With AMPA receptor enhancement, a signal cascade begins and results in the release of BDNF, leading to amplified synaptogenesis and connectivity in brain areas often lacking BDNF in depressed patients (38, 40, 41, 90). Those who suffer from depression are known to have lower amounts of BDNF in certain areas of the brain like the prefrontal cortex and hippocampus (90). On the other hand, NMDA receptor antagonism on GABAergic interneurons and inhibitory interneurons results in glutamatergic activity, ultimately leading to synaptic potentiation and plasticity (28–34, 36). Inhibition of NMDA receptors can also prevent NMDA receptor-mediated

inactivation of protein synthesis in multiple ways. First, NMDA receptors can block homeostatic protein synthesis through mTORC1 and inhibition of the NMDA receptors can prevent this (42–45). Second, eEF2 inactivation can be dependent on NMDA receptors, and by blocking those receptors, the inactivation can be blocked as well, allowing eEF2 to function in protein synthesis to promote plasticity and potentiation (23, 47, 48). Perhaps multiple mechanisms of action work in tandem but determining the precise mechanisms of action in the treatment of depression is a topic for future studies. Some of these mechanisms also had similar end effects, so another possibly beneficial area of future study would be discerning which effects are simply biological and which are therapeutic. Efforts should be made to determine whether other NMDA blockers are as, or more, effective than ketamine. Safety based on dosage should also be investigated since extant data suggests ketamine is tolerated well in low doses in depressed patients, but the same data does not exist for higher repeated doses (88). General safety over longer periods of time could also be a cause for concern due to an apparent lack of data regarding long-term toxicity (90). Different studies utilizing different routes of administration at different dosages have resulted in different efficacies, so discernment of the most feasible and effective administrations would be valuable (88). Finally, more and larger studies could provide more data to determine the intensity of interstudy variations and interactions with comorbidities like anxiety and trauma.

Acknowledgments Guidance on appropriate referencing and expanding content in addition to editing and final review was graciously provided by Brian Piper, PhD. Dr. Piper is an assistant professor of neuroscience at Geisinger Commonwealth School of Medicine.

References 1.

Ionescu DF, Bentley KH, Eikermann M, Taylor N, Akeju O, Swee MB, et al. Repeat-dose ketamine augmentation for treatment-resistant depression with chronic suicidal ideation: A randomized, double blind, placebo-controlled trial. J Affect Disord. 2019;243:516-24. DOI: https://doi. org/10.1016/j.jad.2018.09.037

Feifel D, Malcolm B, Boggie D, Lee K. Low-dose ketamine for treatment resistant depression in an academic clinical practice setting. J Affect Disord. 2017;221:283-8. DOI: https://doi.org/10.1016/j.jad.2017.06.043

Ren Z, Pribiag H, Jefferson SJ, Shorey M, Fuchs T, Stellwagen D, et al. Bidirectional homeostatic regulation of a depression-related brain state by gamma-aminobutyric acidergic deficits and ketamine treatment. Biol Psychiatry. 2016;80(6):457-68. DOI: https://doi.org/10.1016/j. biopsych.2016.02.009

Marcantoni WS, Akoumba BS, Wassef M, Mayrand J, Lai H, Richard-Devantoy S, et al. A systematic review and metaanalysis of the efficacy of intravenous ketamine infusion for treatment resistant depression: January 2009–January 2019. J Affect Disord. 2020. DOI: https://doi.org/10.1016/j. jad.2020.09.007

163

Review of Ketamine as a Rapid Antidepressant for Treatment-Resistant Depression

Chen MH, Li CT, Lin WC, Hong CJ, Tu PC, Bai YM, et al. Rapid inflammation modulation and antidepressant efficacy of a low-dose ketamine infusion in treatmentresistant depression: A randomized, double-blind control study. Psychiatry Res. 2018;269:207-11. DOI: https://doi. org/10.1016/j.psychres.2018.08.078 Correia-Melo FS, Leal GC, Vieira F, Jesus-Nunes AP, Mello RP, Magnavita G, et al. Efficacy and safety of adjunctive therapy using esketamine or racemic ketamine for adult treatment-resistant depression: A randomized, doubleblind, non-inferiority study. J Affect Disord. 2020;264:52734. DOI: https://doi.org/10.1016/j.jad.2019.11.086

Mandal S, Sinha VK, Goyal N. Efficacy of ketamine therapy in the treatment of depression. Indian J Psychiatry. 2019;61(5):480-85. DOI: 10.4103/psychiatry. IndianJPsychiatry_484_18.

Zanos P, Moaddel R, Morris PJ, Riggs LM, Highland JN, Georgiou P, et al. Ketamine and ketamine metabolite pharmacology: Insights Insights into therapeutic mechanisms. Pharmacol. Rev. 2018;70(3):621-60. DOI: https://doi.org/10.1124/pr.117.015198

Włodarczyk A, Cubała WJ, Szarmach J, Małyszko A, Wiglusz MS. Short-term ketamine administration in treatment-resistant depression patients: Focus on adverse effects on the central nervous system. Psychiatr Danub. 2019;31(Suppl 3):530-3.

10. Andrade C. Ketamine for depression, 3: Does chirality matter? J. Clin. Psychiatry. 2017;78(6):674-7. DOI: 10.4088/JCP.17f11681 11. Zhuo C, Tian H, Li G, Chen M, Jiang D, Lin X, et al. Effects of ketamine on circadian rhythm and synaptic homeostasis in patients with treatment-resistant depression: A protocol for mechanistic studies of its rapid and sustained antidepressant actions in humans. Brain Behav. 2019;9(11):e01423. DOI: https://doi.org/10.1002/ brb3.1423 12. Caffino L, Piva A, Giannotti G, Di Chio M, Mottarlini F, Venniro M, et al. Ketamine self-administration reduces the homeostasis of the glutamate synapse in the rat brain. Mol Neurobiol. 2017;54(9):7186-93. DOI: https://doi. org/10.1007/s12035-016-0231-6 13. Zanos P, Gould TD. Mechanisms of ketamine action as an antidepressant. Mol Psychiatry. 2018;23(4):801-11. 14. Autry AE, Adachi M, Nosyreva E, Na ES, Los MF, Cheng PF, et al. NMDA receptor blockade at rest triggers rapid behavioural antidepressant responses. Nature. 2011;475(7354):91-5. DOI: 10.1038/nature10130 15. Garcia LS, Comim CM, Valvassori SS, Réus GZ, Barbosa LM, Andreazza AC, et al. Acute administration of ketamine induces antidepressant-like effects in the forced swimming test and increases BDNF levels in the rat hippocampus. Prog Neuropsychopharmacol Biol Psychiatry. 2008;32(1):140-4. DOI 10.1016/j.pnpbp.2007.07.027 16. Yoshii A, Constantine-Paton M. Postsynaptic BDNF-TrkB signaling in synapse maturation, plasticity, and disease. Dev. Neurobiol. 2010;70(5):304-22. DOI: 10.1002/dneu.20765

164

17. Yang C, Yang J, Luo A, Hashimoto K. Molecular and cellular mechanisms underlying the antidepressant effects of ketamine enantiomers and its metabolites. Transl Psychiatry. 2019;9(280). DOI: 10.1038/s41398-019-0624-1 18. Zanos P, Gould TD. Mechanisms of ketamine action as an antidepressant. Mol Psychiatry. 2018;23(4):801-1. DOI: 10.1038/mp.2017.255 19. Autry AE, Adachi M, Nosyreva E, Na ES, Los MF, Cheng PF, et al. NMDA receptor blockade at rest triggers rapid behavioural antidepressant responses. Nature. 2011;475(7354):91-5. DOI: 10.1038/nature10130 20. Garcia LS, Comim CM, Valvassori SS, Reus GZ, Barbosa LM, Andreazza AC, et al. Acute administration of ketamine induces antidepressant-like effects in the forced swimming test and increases BDNF levels in the rat hippocampus. Prog Neuropsychopharmacol Biol Psychiatry. 2008;32(1):140-4. DOI 10.1016/j.pnpbp.2007.07.027 21. Yoshii A, Constantine-Paton M. Post-synaptic BDNF-TrkB signaling in synapse maturation, plasticity and disease. Dev Neurobiol. 2010;70(5):304–22. DOI: 10.1002/dneu.20765 22. Yang C, Yang J, Luo A, Hashimoto K. Molecular and cellular mechanisms underlying the antidepressant effects of ketamine enantiomers and its metabolites. Transl Psychiatry. 2019;9(280). DOI 10.1038/s41398-019-0624-1 23. Zhang JC, Li SX, Hashimoto K. R(-)-ketamine shows greater potency and longer lasting antidepressant effects than S(+)-ketamine. Pharm Biochem Behav. 2014;116:137-41. DOI 10.1016/j.pbb.2013.11.033 24. Yang C, Shirayama Y, Zhang JC, Ren Q, Yao W, Ma M, et al. Correction: R-ketamine: a A rapid-onset and sustained antidepressant without psychotomimetic side effects. Transl Psychiatry. 2020;10(1):295. [Erratum in: Transl Psychiatry 2015; 5:e632] 25. Zanos, P, Moaddel R, Morris PJ, Georgiou P, Fischell J, Elmer IE, et al. NMDAR inhibition-independent antidepressant actions of ketamine metabolites. Nature. 2016;533(7604):481-6. DOI: 10.1038/nature17998 26. Fukumoto K, Toki H, Iijima M, Hashihayata T, Yamaguchi J, Hashimoto K, et al. Antidepressant potential of (R)ketamine in rodent Models: Comparison with (S)-ketamine. J Pharmacol Exp Ther. 2017; 361(1):9-16. [Erratum in: J Pharmacol Exp Ther. 2017;362(1):1] 27. Hashimoto K, Kakiuchi T, Ohba H, Nishiyama S, Tsukada H. Reduction of dopamine D2/3 receptor binding in the striatum after a single administration of esketamine, but not R-ketamine: A PET study in conscious monkeys. Eur Arch Psychiatry Clin Neurosci. 2017;267(2):173-176. DOI: 10.1007/s00406-016-0692-7 28. Matveychuk D, Thomas RK, Swainson J, Khullar A, MacKay MA, Baker GB, et al. Ketamine as an antidepressant: overview of its mechanisms of action and potential predictive biomarkers. Ther Adv Psychopharmacol. 2020;10:1-21. DOI 10.1177/2045125320916657

Review of Ketamine as a Rapid Antidepressant for Treatment-Resistant Depression

29. Moghaddam B, Adams B, Verma A, Daly D. Activation of glutamatergic neurotransmission by ketamine: A novel step in the pathway from NMDA receptor blockade to dopaminergic and cognitive disruptions associated with the prefrontal cortex. J Neurosci. 1997;17(8):2921–7 DOI: 10.1523/JNEUROSCI.17-08-02921.1997 30. Homayoun H, Moghaddam B. NMDA receptor hypofunction produces opposite effects on prefrontal cortex interneurons and pyramidal neurons. J Neurosci. 2007;27(43):11496–500. DOI: 10.1523/ JNEUROSCI.2213-07.2007

41. Wang CC, Held RG, Chang SC, Yang L, Delpire E, Ghosh A, et al. A critical role for GluN2B-containing NMDA receptors in cortical development and function. Neuron. 2011;72(5):789–805. DOI: 10.1016/j. neuron.2011.09.023 42. Wang CC, Held RG, Hall BJ. SynGAP regulates protein synthesis and homeostatic synaptic plasticity in developing cortical networks. PloS One. 2013;8(12):e83941. DOI: 10.1371/journal.pone.0083941

31. Kotermanski SE, Johnson JW. Mg2+ imparts NMDA receptor subtype selectivity to the Alzheimer’s drug memantine. J Neurosci. 2009;29(9):2774-9. DOI: 10.1523/ JNEUROSCI.3703-08.2009

43. Gray JA, Shi Y, Usui H, During MJ, Sakimura K, Nicoll RA. Distinct modes of AMPA receptor suppression at developing synapses by GluN2A and GluN2B: single-cell NMDA receptor subunit deletion in vivo. Neuron. 2011;71(6):1085-101. DOI: 10.1016/j. neuron.2011.08.007

32. Khlestova E, Johnson JW, Krystal JH, Lisman J. The role of GluN2C-containing NMDA receptors in ketamine’s psychotogenic action and in schizophrenia models. J Neurosci. 2016;36(44):11151-7. DOI: 10.1523/ JNEUROSCI.1203-16.2016

44. Sutton MA, Taylor AM, Ito HT, Pham A, Schuman EM. Postsynaptic decoding of neural activity: eEF2 as a biochemical sensor coupling miniature synaptic transmission to local protein synthesis. Neuron. 2007;55(4):648-61. DOI: 10.1016/j.neuron.2007.07.030

33. Henley JM, Wilkinson KA. Synaptic AMPA receptor composition in development, plasticity and disease. Nat Rev Neurosci. 2016;17(6):337-50. DOI: 10.1038/nrn.2016.37

45. Nosyreva E, Szabla K, Autry AE, Ryazanov AG, Monteggia LM, Kavalali ET. Acute suppression of spontaneous neurotransmission drives synaptic potentiation. J Neurosci. 2013;33(16):6990-7002. DOI: 10.1523/ JNEUROSCI.4998-12.2013

34. Derkach VA, Oh MC, Guire ES, Soderling TR. Regulatory mechanisms of AMPA receptors in synaptic plasticity. Nat Rev Neurosci. 2007; 8(2):101–13. DOI: 10.1038/nrn2055 35. Li N, Lee B, Liu RJ, Banasr M, Dwyer JM, Iwata M, et al. mTOR-dependent synapse formation underlies the rapid antidepressant effects of NMDA antagonists. Science. 2010;329(5994):959-64. DOI: 10.1126/science.1190287 36. Zanos, P, Moaddel R, Morris PJ, Georgiou P, Fischell J, Elmer IE, et al. NMDAR inhibition-independent antidepressant actions of ketamine metabolites. Nature. 2016; 533(7604):481-6. DOI: 10.1038/nature17998 37. Aleksandrova LR, Phillips AG, Wang YT. Antidepressant effects of ketamine and the roles of AMPA glutamate receptors and other mechanisms beyond NMDA receptor antagonism. J Psychiatry Neurosci. 2017;42(4):222-9. DOI: 10.1503/jpn.160175 38. Duman RS, Aghajanian GK, Sanacora G, Krystal JH. Synaptic plasticity and depression: New insights from stress and rapid-acting antidepressants. Nat Med. 2016;22(3):238-49. DOI: 10.1038/nm.4050 39. Duman RS. Ketamine and rapid-acting antidepressants: A new era in the battle against depression and suicide. F1000Res. 2018;7:F1000. DOI: 10.12688/ f1000research.14344.1 40. Miller OH, Yang L, Wang CC, Hargroder EA, Zhang Y, Delpire E, et al. GluN2B-containing NMDA receptors regulate depression-like behavior and are critical for the rapid antidepressant actions of ketamine. Elife. 2014;3:e03581. DOI:10.7554/eLife.03581

46. Park SW, Lee JG, Seo MK, Lee CH, Cho HY, Lee BJ, et al. Differential effects of antidepressant drugs on mTOR signalling in rat hippocampal neurons. Int J Neuropsychopharmacol. 2014;17(11):1831-46. DOI: 10.1017/S1461145714000534 47. Chotiner JK, Khorasani H, Nairn AC, O’dell TJ, Watson JB. Adenylyl cyclase-dependent form of chemical longterm potentiation triggers translational regulation at the elongation step. Neuroscience. 2003;116(3):743-52. DOI: 10.1016/S0306-4522(02)00797-2 48. Park S, Park JM, Kim S, Kim JA, Shepherd JD, SmithHicks CL, et al. Elongation factor 2 and fragile X mental retardation protein control the dynamic translation of Arc/ Arg3.1 essential for mGluR-LTD. Neuron. 2008;59(1):7083. DOI: 10.1016/j.neuron.2008.05.023 49. Zanos P, Moaddel R, Morris PJ, Riggs LM, Highland JN, Georgiou P, et al. Ketamine and ketamine metabolite pharmacology: insights into therapeutic mechanisms. Pharmacol Rev. 2018;70(3):621-660. DOI: 10.1124/pr.117.015198. [Erratum in: Pharmacol Rev 2018;70(4):879] 50. Peltoniemi MA, Hagelberg NM, Olkkola KT, Saari TI. Ketamine: A review of clinical pharmacokinetics and pharmacodynamics in anesthesia and pain therapy. Clin Pharmacokinet. 2016;55(9):1059-77. DOI: 10.1007/ s40262-016-0383-6 51. Johnson BA. Addiction medicine: Science and Practice. London: Springer; 2011. 604 p.

165

Review of Ketamine as a Rapid Antidepressant for Treatment-Resistant Depression

52. Idvall J, Ahlgren I, Aronsen K, Stenberg P. Ketamine infusions: Pharmacokinetics and clinical effects. Br J Anaesth. 1980;24(6):378. DOI: 10.1093/bja/51.12.1167 53. Haas DA, Harper DG. Ketamine: A review of its pharmacologic properties and use in ambulatory anesthesia. Anesth Prog. 1992;39(3):61-8. 54. Geisslinger G, Hering W, Thomann P, Knoll R, Kamp HD, Brune K. Pharmacokinetics and pharmacodynamics of ketamine enantiomers in surgical patients using a stereoselective analytical method. Br J Anaesth. 1993;70(6):666-71. DOI: 10.1093/bja/70.6.666 55. Portmann S, Kwan HY, Theurillat R, Schmitz A, Mevissen M, Thormann W. Enantioselective capillary electrophoresis for identification and characterization of human cytochrome P450 enzymes which metabolize ketamine and norketamine in vitro. J Chromatogr A. 2010;1217:7942– 7948. DOI: 10.1016/j.chroma.2010.06.028 56. Adams JD Jr, Baillie TA, Trevor AJ, Castagnoli N Jr. Studies on the biotransformation of ketamine. 1-Identification of metabolites produced in vitro from rat liver microsomal preparations. Biomed Mass Spectrom. 1981;8(11):527-38. DOI: 10.1002/bms.1200081103 57. Zhao X, Venkata SL, Moaddel R, Luckenbaugh DA, Brutsche NE, Ibrahim L, et al. Simultaneous population pharmacokinetic modelling of ketamine and three major metabolites in patients with treatment-resistant bipolar depression. Br J Clin Pharmacol. 2012;74(2):304-14. DOI: 10.1111/j.1365-2125.2012.04198.x 58. Wang PF, Neiner A, Kharasch ED. Stereoselective ketamine metabolism by genetic variants of cytochrome p450 cyp2b6 and cytochrome p450 oxidoreductase. Anesthesiology. 2018;129(4):756-768. doi: 10.1097/ ALN.0000000000002371 59. Dinis-Oliveira RJ. Metabolism and metabolomics of ketamine: A toxicological approach. Forensic Sci Res. 2017;2(1):2-10. DOI: 10.1080/20961790.2017.1285219 60. Lin HR, Lua AC. Detection of acid-labile conjugates of ketamine and its metabolites in urine samples collected from pub participants. J Anal Toxicol. 2004;28(3):181-6. DOI: 10.1093/jat/28.3.181 61. Adamowicz P, Kala M. Urinary excretion rates of ketamine and norketamine following therapeutic ketamine administration: method and detection window considerations. J Anal Toxicol. 2005;29(5):376-82. DOI: 10.1093/jat/29.5.376 62. Karch SB, Drummer OH. Karch’s Pathology of Drug Abuse. 5th edition. Boca Raton, FL: CRC Press; 2001. 697–754 p. 63. Schwartz J, Murrough JW, Iosifescu DV. Ketamine for treatment-resistant depression: Recent developments and clinical applications. Evid.-Based Ment. Health. 2016;19:358. 64. Meisner RC. Ketamine for major depression: new tool, new questions. Harv Health. 2019. 65. Pribish A, Wood N, Kalava A. A review of nonanesthetic uses of ketamine. Anesthesiol Res Pract. 2020;5798285.

166

66. Thase M, Connolly KR. Ketamine and esketamine for treating unipolar depression in adults: administration, efficacy, and adverse effects [Internet]. Up To Date; 2020 Aug [cited 2021 Feb 6]. Available from: https://www. uptodate.com/contents/ketamine-and-esketamine-fortreating-unipolar-depression-in-adults-administrationefficacy-and-adverse-effects 67. Naval Medical Center Camp Lejeune. Initiating ketamine in acutely suicidal patients in the emergency department [Internet]. National Library of Medicine; 2020 Feb [cited 2021 Feb 6]. Available from: https://clinicaltrials.gov/ct2/ show/NCT04260607 68. Singh I, Morgan C, Curran V, Nutt D, Schlag A, McShane R. Ketamine treatment for depression: opportunities for clinical innovation and ethical foresight. Lancet Psychiatry. 2017;4(5):419-26. 69. Hashimoto K. Rapid-acting antidepressant ketamine, its metabolites and other candidates: A historical overview and future perspective. Psychiatry Clin Neurosci. 2019 Oct;73(10):613-27. 70. Wilkinson ST, Ballard ED, Bloch MH, Mathew SJ, Murrough JW, Feder A, et al. The effect of a single dose of intravenous ketamine on suicidal ideation: a systematic review and individual participant data meta-analysis. Am. J. Psychiatry. 2018;175(2):150-8. 71. Phillips JL, Norris S, Talbot J, Hatchard T, Ortiz A, Birmingham M, et al. Single and repeated ketamine infusions for reduction of suicidal ideation in treatmentresistant depression. Neuropsychopharmacology. 2020;45(4):606-12. 72. Reinstatler L, Youssef NA. Ketamine as a potential treatment for suicidal ideation: A systematic review of the literature. Drug Dev. 2015;15(1):37-43. 73. Schwartz N. Ketamine for suicidal patients: a chance for meaningful psychiatric care in the ED [Internet]. Emergency Physicians Monthly; 2018 [cited 2021 Feb 6]. Available from: https://epmonthly.com/article/ketaminefor-suicidal-patients-a-chance-for-meaningful-psychiatriccare-in-the-ed/ 74. Murrough JW, Soleimani L, DeWilde KE, Collins KA, Lapidus KA, Iacoviello BM, et al. Ketamine for rapid reduction of suicidal ideation: a randomized controlled trial. Psychol Med. 2015;45(16):3571-80. 75. Haapanen LH. Ketamine safe for acutely suicidal patients in the emergency department setting [Internet]. Psychiatry Advisor; 2019 [cited 2021 Feb 6]. Available from: https:// www.psychiatryadvisor.com/home/depression-advisor/ ketamine-safe-for-acutely-suicidal-patients-in-theemergency-department-setting/ 76. Jansen KL. A review of the nonmedical use of ketamine: use, users and consequences. J Psychoactive Drugs. 2000;32(4):419-33. 77. Chen J. Treating depression: an expert discusses risks, benefits of ketamine [Internet]. Yale Medicine; 2017 [cited 2021 Feb 18]. Available from https://www.yalemedicine. org/news/ketamine-the-new-miracle-drug

Review of Ketamine as a Rapid Antidepressant for Treatment-Resistant Depression

78. Willeford A, Atayee RS, Winters KD, Mesarwi P. The enigma of low-dose ketamine for treatment of opioid-induced hyperalgesia in the setting of psychosocial suffering and cancer-associated pain. J Pain Palliat Care Pharmacother. 2018;32(4):248-55. 79. Domany Y, Shelton RC, McCullumsmith CB. Ketamine for acute suicidal ideation. An emergency department intervention: a randomized, double-blind, placebocontrolled, proof-of-concept trial. Depress Anxiety. 2020;37(3):224-33.

91. Zarate CA, Singh JB, Carlson PJ, Brutche NE, Ameli R, Luckenbaugh DA, et al. A randomized trial of an N-methylD-aspartate antagonist in treatment-resistant major depression. Arch Gen Psychiatry. 2006;63(8):864-5. 92. Mather SJ, Shah A, Lapidus K, Clark C, Jarun N, Ostermeyer B, et al. Ketamine for treatment-resistant unipolar depression. CNS Drugs. 2012;26(3):189-204.

80. Murrough J. Ketamine for suicidal ideation [Internet]. National Library of Medicine; 2017 [cited 2021 Feb 17]. Available from: https://clinicaltrials.gov/ct2/show/ NCT01507181 81. Phillips JL, Jaworska N, Kamler E, Bhat V, Blier J, Foster JA, et al. A randomized, crossover comparison of ketamine and electroconvulsive therapy for treatment of major depressive episodes: a Canadian biomarker integration network in depression (CAN-BIND) study protocol. BMC Psychiatry. 2020;20(1):268. 82. Makin S. Behind the buzz: how ketamine changes the depressed patient’s brain [Internet]. Scientific American; 2019 [cited 2021 Feb 6]. Available from: https://www. scientificamerican.com/article/behind-the-buzz-howketamine-changes-the-depressed-patients-brain/ 83. Serafini G, Howland RH, Rovedi F, Girardi P, Amore M. The role of ketamine in treatment-resistant depression: A systematic review. Curr Neuropharmacol. 2014;12(5):44461. 84. Kelly T, Rosonke S. Ketamine's therapeutic value in suicidal patients [Internet]. Emergency Medicine Resident; 2019 [cited 2021 Feb 6]. Available from: http://www.emra.org/ emresident/article/ketamine/ 85. Otto AM. Ketamine tied to remission from suicidal ideation [Internet]. MDedge Psychiatry; 2019 [cited 2021 Feb 6]. Available from: https://www.mdedge.com/psychiatry/ article/202023/depression/ketamine-tied-remissionsuicidal-ideation 86. Dadiomov D, Lee K. The effects of ketamine on suicidality across various formulations and study settings. Ment Health Clin. 2019;9(1):48-60. 87. Jansen KL. A review of the nonmedical use of ketamine: use, users and consequences. J Psychoactive Drugs. 2000;32(4):419-33. 88. Corriger A, Pickering G. Ketamine and depression: A narrative review. Drug Des Devel Ther. 2019;13:3051–67. 89. Al-harbi KS. Treatment-resistant depression: Therapeutic trends, challenges, and future directions. Patient Prefer Adherence. 2012;6:369-88. 90. Salahudeen MS, Wright CM, Peterson GM. Esketamine: New hope for the treatment of treatment-resistant depression? A narrative review. Ther Adv Drug Saf. 2020;11:204209862093789.

167

Scholarly Research In Progress • Vol. 5, November 2021

The Relationship between Food Deserts and the Prevalence of Type 2 Diabetes in Communities in Southeastern Pennsylvania Stephen O. Adeniyi1*, Adewale O. Awoyemi1*, Daniel O. Ayodele1*, Cheryl A. Frazier1*, Yvette M. Johnson1*, Cathie-Allegra Z. Nkabyo1*, and Theodore J. Tucker1* ¹Geisinger Commonwealth School of Medicine, Scranton, PA 18509 *Master of Biomedical Sciences Program Correspondence: cnkabyo@som.geisinger.edu

Abstract Limited access to healthy food has been linked to an increase in diabetes, specifically Type 2 diabetes. Food deserts contribute to the limited access of healthy foods due to their lack of fresh produce and other healthy food options. We evaluated the relationship between food deserts and the rates of diabetes in southeastern Pennsylvania using the Community Health Database Analysis Tool from the Public Health Management Corporation (PHMC). For the purpose and focus of this research, the measured variables included prevalence of food deserts, socioeconomic status, and population size in the Southeastern Pennsylvania (SEPA) online data analysis tool. The PHMC data from 2018 included a sample size of 7,501 household interviews in the 5-county area of southeastern Pennsylvania. A higher prevalence of diabetes was expected in these communities due to limited access to healthy foods. Results revealed that non-poor communities in Bucks, Chester, Delaware, Montgomery, and Philadelphia counties have the highest prevalence of Type 2 diabetes, despite having increased access to supermarkets. Therefore, access to supermarkets and grocery stores had no statistically significant relation with the prevalence of diabetes. Other factors such as access to health care and education were not considered in this study. More research on this topic is needed to gain greater insight into diabetes and food desert relationships. Results from this study can be used to increase awareness and inform local government officials about the relationship between food deserts and the prevalence of diabetes.

Introduction Despite an abundance of supermarket retailers like ShopRite, Walmart, and Kroger in the United States (U.S.), lack of access to fresh and healthy foods in certain communities is a pressing issue. These communities are referred to as “food deserts” and defined by public health officials as “geographic areas that lack sufficient access to grocery stores…” (1). According to the United States Department of Agriculture (USDA) (2), 6,500 food deserts were identified by correlating data from the 2000 U.S. census and data of supermarket and grocery locations in 2006 (3). From a public health standpoint, it is important to identify and improve communities that still lack access to fresh produce to prevent diseases associated with unhealthy diets. Food deserts tend to be observed in low-income communities that experience racial and economic disparities (3). Currently, 32 million adults living in the U.S. have been diagnosed with diabetes (4). As defined by the CDC, social determinants of health (SDOH) are conditions in the places where people live, learn, work, and play that affect a wide range of health and quality of life risks and outcomes. Based on the 168

CDC SDOH guidelines, residents of Philadelphia County are 92.04% more likely to be diagnosed with diabetes, compared to New York, Los Angeles, and Cook (Illinois) counties — 63.93%, 76.82%, and 80.20%, respectively (4). In comparison to other counties in the U.S., Los Angeles County is 76.82% more vulnerable, New York County is 63.93% more vulnerable, and Cook County (Illinois) is 80.20% more vulnerable (4). Additionally, in Philadelphia County 11.6% of the population has been diagnosed with diabetes compared to 9.1% in Los Angeles County, 6% in New York County, and 9% in Cook County (4). A diagnosis of diabetes may lead to other chronic diseases and conditions like kidney disease, heart disease, vision loss, and amputation (6). Diabetes is the sixth leading cause of death in Philadelphia and affects those living in poverty twice as much as those not living in poverty (5). Other risk factors of diabetes are obesity, smoking, physical inactivity, high blood pressure and high cholesterol (5). Food deserts have been shown to be a significant risk factor associated with diabetes, along with other SDOH having an impact such as low income, unstable housing, substandard education, and unhealthy environmental conditions. Low-income residents in food deserts face a variety of disadvantages, reflective of the inadequacy of the built environment. These include lower annual median household income, unequal access to educational and economic opportunities, and housing instability. Thus, SDOH needs to be considered when striving toward improving diabetes health outcomes. This is in part because the existence and cohesion to an adequate diet are the cornerstones to diabetes prevention and management (7). Previous research has demonstrated that there is no monotonic relationship between deficient food environments and diabetes. Alternatively, areas of limited access, food deserts, as well as areas of high concentration of health-harming food outlets are both associated with Type 2 diabetes (8). Understanding food deserts and their effect on the health of lower income individuals is vital when attempting to comprehend why certain populations may encounter diabetes more frequently. People who live in food-insecure households are two to three times more likely to have diabetes than those who live in food-secure households (9). This may be related to the discrepancy in the number of resources such as safe housing and local food markets in low-income versus affluent neighborhoods to sustain a healthy lifestyle. In southeastern Pennsylvania lower-income areas, there are more likely to be an abundance of fast-food restaurants and a scarcity of grocery stores (10). Unfortunately, there are many gaps in research on food deserts. When discussing research methods, investigators have mostly

The Relationship between Food Deserts and the Prevalence of Type 2 Diabetes

focused on cross-sectional data which compares people living in different neighborhoods (11). Cross-sectional study design is essential in learning about the characteristics and trends of food deserts at a particular point in time further supporting the research claims. It can also be descriptive and explain disease burdens in a particular community. However, there is also a need for longitudinal evidence. Longitudinal evidence can relate changes in healthy food availability to changes in diet over time and uncover predictors of certain diseases (11). One example of this type of study suggested that people who lived in neighborhoods with higher healthy food availability had a 45% reduced incidence of developing diabetes (11). Although crosssectional research is improving, there are still difficulties in interpreting causal links. These links include determining which aspects of the local food environment are most relevant to health; how to measure the scale at which changes to the local food environment are most effective; and the critical periods in which to expect any effects to occur (11).

Methods

In addition to advancement in research methods concerning food deserts, there are solutions being developed to decrease the prevalence of food deserts (12). For example, studies have shown that local networking between farmers and food producers within a few hundred miles’ radius of the city has contributed to the presence of healthy food. This in turn allows healthy food to arrive much quicker, from a shorter distance, at a lower price (12). The previously stated solutions are good, yet still inaccessible (for geographic, financial, and social reasons) to low-income residents in food deserts. Alongside the local farmers' network solution, another solution is a coupon incentive initiative. This initiative looks to grant customers $2 coupons for every dollar spent which can only be redeemed by the purchase of fresh fruits and vegetables. This helps customers save money on groceries and aims to bring new customers to fresh food markets in low-income communities to increase fruit and vegetable consumption (10). CSAs (community supported agriculture); farmer's market collaboration with health systems, and Geisinger’s Fresh Food Farmacy are also feasible solutions to address healthy food accessibility.

Data collection

Additionally, one of the most effective solutions is nutrition education. These programs are essential for the eradication of food deserts and improvement of impoverished areas (12). Knowledge is one of the more efficient solutions because it makes citizens more conscious of their diet, provides them with the tools to make informed decisions, and in turn may lower the prevalence of diabetes. Another significant part of the solution are the organizations that help make quality, healthy foods accessible. These factors will start to shift the dynamics towards a healthier path. Our research aims to identify the impact of food deserts on the prevalence of diabetes in low-income communities of southeastern Pennsylvania. This study examined information on the relationship between food deserts and the prevalence of diabetes using data from the Public Health Management Corporation (PHMC) Community Health Database Household Health Survey and the SEPA Online Data Analysis Tool. Our analysis offers new insight and solutions to eliminate food deserts in the U.S.

Participants The population included adult Pennsylvania residents with and without diabetes as well as those with or without low access to grocery stores. The PHMC houses data from 2018 and reports a sample size of 7,501 household interviews in the 5-county area of southeastern Pennsylvania. The SEPA Online Data Analysis Tool permits acquisition of data about a specific geographic area and indicators for data analysis. The data set was coded for: time (data from 2018), age group (adults 18–60), health status (diabetes), and poverty status. Poverty level is defined as a measure calculated based on family size and household income below 100% of the federal poverty level (13). Populations used to calculate heat maps were extracted from the United States Census Bureau released in the American Community Survey in December of 2020.

We used the Community Health Data Base Analysis Tool from the Public Health Management Corporation (PHMC) as the secondary data source. This database includes counties of the southeastern regions of Pennsylvania, including Philadelphia, Bucks, Chester, Delaware, and Montgomery. The database consists of household health surveys conducted every 2 to 3 years on adults dating back to 2000. Data from 2018 was used in our analysis. We used a pre-existing public data set from the SEPA Online Data Analysis Tool which excludes any identifiable personal information of the participants within the studied areas of Pennsylvania. This study was reviewed by the Geisinger Commonwealth School of Medicine Institutional Review Board and deemed to be non-human subject research. Data analysis plan In this study, the descriptive analysis consists of analyzing 2018 data found in the SEPA Online Data Analysis Tool. This data is compiled by PHMC, which collects data from Pennsylvania residents. We used the prevalence of diabetes as the measured variable for this study. We utilized data from PHMC, which allowed us to examine the relationship between socioeconomic status and the prevalence of diabetes in southeastern Pennsylvania. We identified the prevalence of food deserts in each of the selected using data from the Food Access Research Atlas created by the USDA. Thus, a comparative T-test utilized variables based on the low access to supermarkets within 0.5 and 10 miles and 1 and 20 miles in Bucks, Chester, Delaware, Montgomery, and Philadelphia counties. Heat maps, created through Excel, highlighted the correlation between the prevalence of diabetes, food deserts, and socioeconomic status within those counties. The analyses compared the prevalence of diabetes in poor communities versus non-poor communities in southeastern Pennsylvania which was taken from the SEPA Online Analysis Tool. When comparing the data regarding the low access to supermarkets, an assessment of whether a supermarket was 0.5 miles and 10 miles or 1 and 20 miles away through the USDA Food Access Research Atlas was used. Five counties were evaluated: Philadelphia, Montgomery, Bucks, Chester, and Delaware. 169

The Relationship between Food Deserts and the Prevalence of Type 2 Diabetes

Results

using the SEPA Online Data Analysis tool, food deserts are very prevalent in southeastern Pennsylvania. Our analysis indicated that in Bucks, Chester, Delaware, Montgomery, and Philadelphia County low-income areas are impacted by disproportionate prevalence of Type 2 diabetes. Diabetes is still present in lowincome areas, but the prevalence is not driven by income alone. Diabetes prevalence can be influenced by several different factors such as income status, location of supermarkets, and fresh produce markets. For example, Bucks County has a very high prevalence of diabetes in non-poor communities while the poor communities have a much lower prevalence, suggesting income is not a driver for diabetes in that county.

Contrary to our initial hypothesis, our data in Figure 1 depicts that non-poor communities in Bucks, Chester, Delaware, Montgomery, and Philadelphia counties all had a higher prevalence of diabetes compared to poor communities. Certain zip codes exhibited low access to supermarkets, meaning that they are considered food deserts. There are 31 zip codes in Philadelphia County that have low access to supermarkets. As shown in Figure 2, 68% of Philadelphia County has low access to supermarkets, making it the county with the highest percentage. There are 11 zip codes in Montgomery County that are considered to have low access to supermarkets making up 20% of the population. In Bucks County, there are 8 zip codes that are considered to have low access to supermarkets making up 17% of the population (Figure 2). In Chester County, the 7 zip codes that are considered low access make up 22% of the population. In Delaware County, 36% of the population had low access to supermarkets which consists of 14 zip codes. Heat maps were created to depict Figure 1. Prevalence of diabetes in poor and non-poor communities. A table depicting the prevalence of diabetes within each county’s the prevalence of diabetes in poor and non-poor communities. zip codes (Figures 3A-D, 4A-D, 5A-D, 6A-D, and 7A-D) and t-tests were created to represent the prevalence of low access in each county (Figures 3E and F, 4E and F, 5E and F, 6E and F, and 7E and F).

Discussion According to the USDA and Economic Research Service (ERS), about 23.5 million people live in food deserts, and nearly half of them are also low-income (2). According to the data collected

Figure 2. Low access to supermarkets. A table depicting the percentage of zip codes in each county that had low access to supermarkets.

Figure 3A–F. Prevalence of diabetes and low access to supermarkets in Bucks County. Heat maps created to depict the prevalence of diabetes and its correlation to socioeconomic status in Bucks County, Pennsylvania. T-test depicting low access to supermarkets in Bucks County, Pennsylvania.

170

The Relationship between Food Deserts and the Prevalence of Type 2 Diabetes

Figure 4A–F. Prevalence of diabetes and low access to supermarkets in Chester County. Heat maps created to depict the prevalence of diabetes and its correlation to socioeconomic status in Chester County, Pennsylvania. T-test depicting low access to supermarkets in Chester County, Pennsylvania.

Figure 5A–F. Prevalence of diabetes and low access to supermarkets in Delaware County. Heat maps created to depict the prevalence of diabetes and its correlation to socioeconomic status in Delaware County, Pennsylvania. T-test depicting low access to supermarkets in Delaware County, Pennsylvania.

171

The Relationship between Food Deserts and the Prevalence of Type 2 Diabetes

Figure 6A–F. Prevalence of diabetes and low access to supermarkets in Montgomery County. Heat maps created to depict the prevalence of diabetes and its correlation to socioeconomic status in Montgomery County, Pennsylvania. T-test depicting low access to supermarkets in Montgomery County, Pennsylvania.

Figure 7A–F. Prevalence of diabetes and low access to supermarkets in Philadelphia County. Heat maps created to depict the prevalence of diabetes and its correlation to socioeconomic status in Philadelphia County, Pennsylvania. T-test depicting low access to supermarkets in Philadelphia County, Pennsylvania.

172

The Relationship between Food Deserts and the Prevalence of Type 2 Diabetes

Another statistic stated by the USDA ERS is that approximately 2.3 million people (2.2% of all US households) live in low-income, rural areas that are more than 10 miles from a supermarket (3). The data collected and analyzed indicated that in Philadelphia County, a large percentage of individuals with diabetes live in zip codes where there is low access to supermarkets. Contrary to the original prediction made, the data collected suggests that the economic status of individuals within a particular zip code does not influence the prevalence of diabetes within that area. Instead, the immediate access to major grocery stores and supermarkets shows a stronger relationship of the prevalence of diabetes in certain zip codes within Philadelphia and surrounding counties. In the Chester County heat maps shown in Figure 4A-D, the 19010 zip code was excluded due to no available data found on SEPA. Additionally, SEPA did not differentiate between Type 1 and Type 2 diabetes in their database; however, according to the CDC over 90% of cases of diabetes are Type 2 (5). Furthermore, the values used to calculate the heat maps and t-tests were rounded to whole numbers instead of using decimals.

Acknowledgments This group of researchers would like to acknowledge the help received from: Brian J. Piper, PhD, MS; Elizabeth Kuchinski MPH, BS; Tenzing K. Dolma; Ijeoma C. Ejiogu; Alex I. Greenstone; Timothy H. Lozier; Mariah W. Panoussi; Alivia L. Roberts; Megha K. Sarada; and Sonia Lobo, PhD.

References 1.

Karpyn AE, Riser D, Tracy T, Wang R, Shen YE. The changing landscape of food deserts. UNSCN Nutr. 2019 Summer;44:46–53.

Food Access Research Atlas [Internet]. Usda.gov. [cited 2021 Aug 5]. Available from: https://www.ers.usda.gov/ data-products/food-access-research-atlas/

Ver Ploeg Tracy Farrigan PDM. Characteristics and Influential Factors of Food Deserts [Internet]. Washington DC: USDA; 2012. Available from: https://www.ers.usda. gov/webdocs/publications/45014/30940_err140.pdf

Farley T. Health of the City 2019 [Internet]. Department of Public Health City of Philadelphia; 2017. Available from: https://www.phila.gov/media/20191219114641/Health_ of_City_2019-FINAL.pdf

Diabetes Data and Statistics [Internet]. Cdc.gov. 2021 [cited 2021 Aug 5]. Available from: https://www.cdc.gov/ diabetes/data/index.html

Berkowitz SA, Karter AJ, Corbie-Smith G, Seligman HK, Ackroyd SA, Barnard LS, et al. Food insecurity, food “deserts,” and glycemic control in patients with diabetes: A longitudinal analysis. Diabetes Care. 2018;41(6):1188–95.

Mezuk B, Li X, Cederin K, Rice K, Sundquist J, Sundquist K. Beyond access: Characteristics of the food environment and risk of diabetes. Am J Epidemiol. 2016;183(12):1129– 37.

Gucciardi E, Vahabi M, Norris N, Del Monte JP, Farnum C. The intersection between food insecurity and diabetes: A review. Curr Nutr Rep. 2014;3(4):324–32.

Young CR, Aquilante JL, Solomon S, Colby L, Kawinzi MA, Uy N, et al. Improving fruit and vegetable consumption among low-income customers at farmers markets: Philly Food Bucks, Philadelphia, Pennsylvania, 2011. Prev Chronic Dis. 2013;10(120356):E166.

Conclusion The purpose of this study was to examine the relationship between residents living in low-income communities and the prevalence of diabetes in Pennsylvania. Although we hypothesized that non-poor communities would be less likely to have diabetes than poor communities, using geographic imaging software, we observed that a higher prevalence of diabetes occurred in higher income communities of Pennsylvania, particularly the southeastern region. After further investigation of the southeastern region of Pennsylvania, encompassing Bucks, Chester, Delaware, Montgomery, and Philadelphia County, our hypothesis was not supported. We cannot exclusively state that the prevalence of diabetes is attributed to income status, location of supermarkets, and fresh produce markets. This southeastern region has 71 zip codes flagged as having low access to supermarkets within 0.5 and 20 miles. The SDOH provides insight on barriers hindering living a healthy lifestyle. The prevalence of diabetes in areas where one’s health outcomes are negatively impacted at the level of economic stability, as well as adequate neighborhood and/ or built environment, suggests additional research is required to characterize the relationship between food deserts and diabetes. From a public health standpoint, more strategies are needed to decrease the prevalence of diabetes regardless of income status and geographical location. We suggest that a longitudinal cohort study be completed to better understand the relationship of Type 2 diabetes, food deserts, and income status.

Disclosures The authors of this investigation do not have any conflicts of interest.

10. National Research Council (US). Summary. Washington, D.C., DC: National Academies Press; 2009. 11. Segal A. Food Deserts: A Global Crisis in New York City [Internet]. New York, NY: Columbia University Press; November 30 2015. Available from: https:// academiccommons.columbia.edu/doi/10.7916/ D8PG1RCG 12. Welcome to community health data base [Internet]. Phmc. org. [cited 2021 Aug 5]. Available from: http://chdb.phmc. org/

173

Scholarly Research In Progress • Vol. 5, November 2021

Characterizing the Behavioral and Cellular Effects of the R904S Variant of OPA1 as a Tourette Disorder Probable Risk Gene Kinza Abbas1† and Cara Nasello2 ¹Geisinger Commonwealth School of Medicine, Scranton, PA 18509 ²Rutgers University, Piscataway, NJ 08854 † Doctor of Medicine Program Correspondence: kabbas@som.geisinger.edu

Abstract Tourette disorder (TD) is a heritable neuropsychiatric and developmental disorder characterized by motor and vocal tics. TD has a complex heterogenous etiology which makes the identification of genes linked to the disorder challenging. However, previous studies have identified optic atrophy 1 (OPA1) as a probable risk gene for Tourette disorder through whole exome sequencing analyses of de novo mutations. The OPA1 protein localizes to the inner membrane of mitochondria in cells throughout the body, where it regulates mitochondrial morphology and function. The Tischfield lab has identified a TD patient with an OPA1 missense mutation, R904S, which we hypothesized would result in abnormal behavioral activity in mutant mice and/or affect the cellular function of mitochondria. We utilized several behavioral paradigms, including motion sequencing, the open field arena test, the marble burying task, and prepulse inhibition. These paradigms test various factors, such as behavioral dynamics, motor activity levels, and sensorimotor gating. To understand the cellular characterization, we observed whole brain expression of the protein, protein expression levels in different brain regions and mitochondria of cortical neurons. Preliminary results of mitochondria imaging show potential mitochondrial aggregates in heterozygous and homozygous mutant cortical neurons, which could indicate an increased preapoptotic state of mutant cells. Although the results of the remaining assays have been inconclusive, there are many variables that complicate the studies. For a better understanding of the variant, mitochondrial function should be tested in the future, such as oxidative phosphorylation activity, cytochrome c levels and reactive oxygen species levels. Cytochrome c levels and apoptosis assays especially should be conducted to better understand the nature of the observed mitochondrial aggregates. Identifying possible behavioral or cellular abnormalities of this variant is an important step in characterizing the mutation, determining its potential role in TD pathology, and developing potential clinical treatments for TD in the future.

Introduction Tourette disorder Tourette disorder (TD) is a heritable neuropsychiatric and developmental disorder characterized by waxing and waning motor and phonic tics, which are defined by the DSM-V as sudden, rapid, recurrent, semi-voluntary, and nonrhythmic motor movements or vocalizations. The condition is three to four times more likely to occur in males than females and according to the Centers for Disease Control and Prevention, TD diagnosis requires a person to exhibit 174

at least two motor tics, at least one vocal tic and experience persistent symptoms for at least 1 year, with an onset before the age of 18 (1). Therefore, although TD is a tic disorder, not all tic disorders can be classified as TD. The disorder is often comorbid with other neuropsychiatric disorders — most commonly obsessive-compulsive disorder (OCD) and attention deficit hyperactivity disorder (ADHD). However, autism spectrum disorders (ASD), depression, and anxiety disorders may also be present. This overlap in disorders implies a shared neurological background and a genetic influence. Additionally, family and twin studies have established a strong genetic influence behind TD (2). In studies of premonitory urges, which are often experienced by TD patients, it is reported that these urges were more bothersome than tics, allowed them to better suppress their tics and that their tics were a response to these urges (3). Altogether, these findings demonstrate the complexity of TD etiology, which appears to involve a combination of psychological, neurological, and motor circuitry. This makes it challenging to pinpoint exact pathways or genes contributing to the disease. Optic atrophy 1 (OPA1) Previous whole-exome sequencing studies have found significant evidence for de novo gene-disrupting variants in TD (4). This study was expanded and optic atrophy 1 (OPA1) was specifically identified as a de novo probable risk TD gene (5). It has been heavily studied for its role in dominant optic atrophy (DOA), which is a neuro-ophthalmic disorder characterized by degeneration of the optic nerve. The OPA1 protein is a dynamin GTPase that mediates inner membrane fusion and enzymatic reactions for membrane remodeling, which in turn regulate mitochondria energetics. A major feature of the inner membrane is the cristae, which are invaginations of the membrane that have important proteins embedded, such as the electron transport chain complexes. OPA1 proteins oligomerize in the cristae to maintain cristae tightness and to function as a regulatory gate. OPA1 proteins also complex between inner membranes of fusing mitochondria to carry out inner membrane fusion (6). Upon apoptosis signaling, OPA1 oligomerization alters cristae structure to allow cytochrome c release (7). The protein’s role in membrane dynamics in turn impacts mitochondrial bioenergetics and general functionality. Mitochondrial dynamics, or the balance between fission and fusion, mediates mitochondrial quality control (MQC). The organelle maintains homeostasis by enforcing the following controls: (a) shifting morphology between filamentous or fragmented according to the energetic demands of the cell, (b) fusing dysfunctional mitochondria to mitigate damage, (c) mitochondria fragmentation, membrane permeability and cytochrome c release signal apoptosis (8).

R904S Variant of OPA1 as a Tourette Disorder Probable Risk Gene

Although OPA1 is more prominently expressed in the retina and brain, the protein is ubiquitously expressed and well conserved. In fact, OPA1 knockout mice are not viable. The gene encodes 30 coding exons and consists of over 90 kb of DNA on chromosome 3 (9). It is translated in the cytosol after which it is imported into the mitochondria, where it undergoes proteolytic cleavage. The long OPA1-L isoform remains anchored to the inner membrane while the short OPA1-S isoform is released to the inner membrane space. A balance of the two isoforms is required for proper mitochondrial functioning (6). The R904S variant of OPA1 A TD patient with the OPA1 mutation, R904S, in which there is an arginine conversion to serine at position 904, has previously been identified by the Tischfield laboratory. The amino acid conversion results from a single nucleotide missense mutation and is likely gene disrupting. Arginine is basic, hydrophilic, charged, and large, while serine has a hydroxyl chain group, is neutral, uncharged, and has a small side chain. The gene has several functional regions (9), including a mitochondrial import sequence (MIS) and three dynamin family conserved regions: a GTPase domain, a middle domain, and a GTPase effector domain (GED) which contains a coiled-coil domain (CC2). As previously stated, the gene comprises of 30 coding exons. The R904S mutation is found in exon 27, which is part of the CC2 domain that stretches through exons 27-28. The domain participates in protein-protein interactions; thus, dysfunction of this region may alter the protein’s ability to form complexes (9).

detailed characterization of behavior modules and replaces tedious and unreliable human scoring of behavior. Although the system still relies on human perception and intuition, as the initial quantification of behaviors is specified by human observers, this method offers an improved methodology. Using this data, behaviors can be deconstructed into “syllables,” or basic motions, such as a head turn, a body turn, or a pause. The Moseq approach captures pose dynamics of mice — data which are then used to identify behavioral motifs. Computational models can recognize and learn these behavioral modules so comparison to altered structure of behavior after environmental, genetic or neural manipulation can be made (10, 11). Open field The open field chamber is square-shaped and wall-enclosed, separated into an outer and inner zone. Invisible infrared beams are emitted from photocell receptors along the perimeter of the chamber and a computer analyzer can record and analyze beam breaks resulting from subject movement. Numerous motor activity variables can be scored including total ambulatory distance, latency to enter the center and rearing. Latency to enter the center and rearing are considered exploratory behaviors and can be used to measure anxiety (12). Along with anxiety, the open field parameters can also be observed to identify differences in motor activity between subjects and genotypes. As OPA1 is a mitochondrial protein, deficits in the protein function could disrupt mitochondrial function, which could result in increased/decreased motor function of mice.

Hypothesis

Marble-burying task

Tourette disorder is not well studied, and due to its complex etiology, it is also difficult to study. With this project, we anticipated on identifying a behavioral or cellular assay for TD with the variant, which would be the first TD model with OPA1. We hypothesized that the R904S variant of the TD risk gene, OPA1, would result in behavioral or cellular abnormalities as compared to wildtype mice/cells. To test this, we employed a mouse model of the mutation generated with CRISPR technology. We utilized motion sequencing, marble burying task, open field arena test, and prepulse inhibition as our behavioral testing paradigms. After behavioral testing, we performed western blots to determine whether there were any significant differences in the level of whole brain OPA1 protein expression between mutant and wildtype mice. We then stained brain slices for OPA1 to determine localization and expression patterns of the mutant protein as compared to the wildtype in various brain regions. Lastly, we stained cortical neurons for mitochondria to determine localization, size and shape of mutant mitochondria. As there are several mitochondrial functions that can be impacted by the OPA1 protein, we began with broad, qualitative cellular tests to determine if there were detectable mitochondrial defects that could potentially pinpoint the mechanism of the mutation. In this way, we could test mitochondrial function according to the mechanism observed.

The marble-burying test utilizes the natural inclination of mice to dig in natural settings. A greater number of buried marbles indicates greater burrowing and digging activity. Altogether, this test is useful in studying overall mouse activity, and more specifically, obsessive and compulsive behavior (13).

Behavioral paradigms

Prepulse inhibition Patients with Tourette disorder exhibit deficits in sensorimotor gating, which is tested by prepulse inhibition (PPI). Sensorimotor gating is a mechanism for filtering information deemed irrelevant to the brain, such as redundant stimuli. This mechanism prevents sensory overload from constant input of unnecessary information. PPI measures sensorimotor gating by measuring startle reflex, which is an involuntary contraction elicited by unexpected stimuli. On average, the subject should have less of a startle response, or greater inhibition, to the primary pulse in the presence of a prepulse stimulus (and increasing inhibition with increasing magnitude of the prepulse stimulus). This would show evidence of functioning sensorimotor gating, which is deficient in TD patients (14). It has also been suggested that premonitory urges can indirectly be assessed by PPI. This is crucial as premonitory urges are a major psychological aspect of TD (15).

Methods

Motion sequencing

Animals

Motion sequencing (Moseq) utilizes the Microsoft Xbox 360 Kinect camera, which acquires spatial and temporal data of moving mice using 3D infrared imaging. This allows more

Our animal protocol was approved by IACUC at Rutgers University, New Brunswick. A total of 224 mice were bred during the span of the project, with 110 males and 114 females.

175

R904S Variant of OPA1 as a Tourette Disorder Probable Risk Gene

Fifty-eight mice were wildtypes, 113 were heterozygotes and 38 were homozygous mutants. Approximately 2-month-old C57BL/6 mice were used for the behavioral assays which were run during the morning. The mice were brought into the behavioral annex at least 48 hours before testing to allow environmental acclimation. The tests were run at least 24 hours apart with prepulse inhibition run as the last test, as it induces hearing impairment and increased stress for mice. Mice were allowed to acclimate in the testing room for 15 minutes prior to being tested. Behavioral testing Motion sequencing (Moseq) The Moseq test was administered by placing mice into circular, opaque black containers with the Microsoft XBox Kinect camera attached above. Mice were filmed for 20 minutes, after which another team obtained and interpreted the footage to extract quantifiable data. Containers were wiped down with bleach, alconox, and 70% ethanol to remove feces, urine, and scents. Marble-burying task The marble-burying task protocol (13) consisted of 20 marbles being uniformly placed in a 5 x 4 pattern on 2 inches of unscented bedding in a 8 in x 13 in housing cage (Figure 1 left). The sides of the cage were covered with paper to prevent distractions from surroundings. Mice were placed in the testing cage for 30-minute sessions (Figure 1 right depicts post-test cage), and the number of marbles buried during the 30-minute session were counted and averaged by three observers. Marbles were considered buried if two-thirds or more of the marble was buried. New bedding was used for each mouse, and marbles were washed with ethanol between different testing sessions to remove feces, urine, and scents. Open field test Mice were placed at the periphery of the open field arena so the data for their interaction with the center of the field would not be skewed. Each mouse was tested for 20-minute sessions. The testing arena was wiped with 70% ethanol between mice to remove feces, urine, and scents.

PPI PPI was used to determine sensorimotor gating deficits using an established protocol involving the SR-LAB system (14). This setup consists of a chamber that administers the auditory stimulus and a stabilimeter placed inside the chamber. The stabilimeter is a plexiglass cylinder in which the mice are held, mounted on a piezoelectric accelerometer that records the movements. A continuous background noise level of 65 dB was maintained to standardize measurements, reduce external noises, and allow comparisons with other literature. The testing session is separated into 4 blocks, each of which presents a different pattern of stimuli. The first and last blocks present stimuli of 120 dB alone for the purpose of calculating habituation throughout the session. Blocks 2 and 3 consist of three different trials presented in a randomized order, including a pulse-alone trial, a prepulse + pulse trial, and a no stimulation (NOSTIM) trial (although the subject’s movements are still recorded during the NOSTIM trial). The pulse-alone trials present a 40 msec stimuli of 120 dB while the prepulse + pulse trials consist of 20 msec prepulses presented before the primary pulse stimulus (the time interval between the two pulses randomly varies between 30 and 500 msec for each trial). The prepulse + pulse trials can also randomly vary in the magnitudes of prepulse used, and in this case, magnitudes of 6, 12, and 16 dB were used (magnitudes of 71, 77, and 81 dB were delivered as the prepulses have to be 6, 12, and 16 dB above the background noise of 65 dB). The percent PPI was calculated for each prepulse intensity using the equation %PPI = 100 × [(pulse-alone) – (prepulse + pulse score)]/pulse-alone score.

Data analysis GraphPad Prism software was utilized to conduct data analysis. Two-way analysis of variance (ANOVA) SD was used to determine the statistical differences between testing groups. Cellular testing Restriction digest to confirm genotypes of mice To acquire mouse genotypes, tissue was collected from the ear during ear tagging. DNA from these tissues was extracted by incubating the tissue in the Sigma-Aldrich tissue extraction solution (Sigma-Aldrich). A PCR reaction was then performed to amplify the obtained DNA, followed by a restriction digest reaction with Taq1 restriction enzyme (Thermofisher). This enzyme recognizes and cleaves after an AGC site, allowing for cleavage of the mutant variant which alters an arginine (AGG/ AGA) to a serine (AGC). Western blot to determine OPA1 localization and whole brain levels

Figure 1. The left image shows a marble-burying task setup prior to placing the mouse in for testing. Marbles were arranged atop unscented bedding in a 5 x 4 arrangement. The right image depicts a cage after a 30-minute session. The number of buried marbles were then counted by 3 observers and averaged. 176

Mice were first dissected and whole brains were extracted, from which the mitochondria were then isolated using the Abcam Mitochondria Isolation Kit for Tissue (Abcam). A Pierce BCA Protein Assay Kit (Thermofisher) was then utilized to determine protein concentration in both our separated mitochondrial fragment and mitochondrial excluded fragment samples. This protein concentration was needed to calculate and normalize the amount of protein loaded onto a 4–12% Bis-Tris protein gel for each sample. Equal amount of protein loading was crucial to allow comparison between the protein samples.

R904S Variant of OPA1 as a Tourette Disorder Probable Risk Gene

Samples were then run with gel electrophoresis followed by an electrotransfer of the gel onto polyvinylidene fluoride membrane (PVDF). As three proteins were being identified, OPA1, Hsp60, and GAPDH, the membrane was cut according to the size of each protein. The GAPDH monomer that the antibody recognizes is 36 kDa, Hsp60 is 60 kDa and the two OPA1 peptides are 80–100 kDA. The protein sizes were varying enough to allow us to section the membrane into three to incubate each section with its corresponding antibody. For GAPDH, we used the monoclonal Anti-GAPDH HRP conjugated antibody (Abcam), for Hsp60, the polyclonal antiHsp60 antibody (Abcam) was utilized and for OPA1, we utilized a polyclonal OPA1 antibody (NovusBio). For the secondary antibody, the HRP conjugated monoclonal secondary antibody (Genscript) was utilized for Hsp60 and OPA1. No secondary antibody was required for GAPDH as the primary antibody was already conjugated with HRP. After antibody incubation, the membranes were visualized using the Western Blot Detection Kit (KwikQuant), which uses an enhanced chemiluminescence (ECL) substrate for HRP detection. OPA1 and nuclear staining of whole brain slices To visualize the localization of the R904S variant protein in various brain regions, we obtained brains from mice, sliced them coronally, and stained them with an OPA1 primary antibody (NovusBio). To obtain the brain, we performed a perfusion fixation. The mice were first anesthetized, after which the thoracic cavity was opened to expose the still-beating heart. A saline solution was streamed through the vascular system by inserting a needle into the left ventricle, followed by 4% paraformaldehyde as the fixative solution. Efficient fixation was indicated by a cleared liver color. After fixing, the brains of the mice were extracted and placed in fixation solution for maintenance. The Vibratome was used to slice the brains into coronal sections. The brain slices were then incubated in 0.3% 1X phosphate-buffered saline, 0.1% Tween (PBST) for 10 minutes to permeabilize the slices, followed by a 45-minute incubation in 2% bovine serum albumin (BSA) in 0.1% PBST to block nonspecific antibody binding. Brain slices were then incubated in rabbit polyclonal anti-OPA1 primary antibody (NovusBio) overnight, washed 5 x 5 min in 0.5% BSA in 1X PBS, incubated in goat anti-rabbit 555 fluorescent secondary antibody (Thermofisher) for 1 hour, washed 5 x 5 min in 0.5% BSA in 1X PBS, and again washed 5 x 5 min in 1X PBS. Finally, the slices were incubated with blue Hoechst dye (Invitrogen) for 5 min and washed 3 x 5 min in 1X PBS to stain nuclei. Fluoromount-G mounting medium (Thermofisher) was then used to mount the brain slices onto glass microscope slides for visualization under a fluorescence microscope. MitoTracker staining in cortical neurons P2 mice were dissected and their brain cortices were extracted. The Pierce Primary Neuron Isolation Kit protocol (Thermofisher) was utilized to enzymatically digest the brain tissue with papain. Cells were counted using a Vi-CELL cell counter and a density of 5 x 105 cells/cm² were plated in a 24well plate with glass coverslips placed in the wells. Neuronal cell culture medium and supplemental medium were applied every 2 to 3 days to reduce non-neuronal cell contamination and promote neuron purity. On approximately day 9, the wells were

incubated with MitoTracker Red (Thermofisher) for 30 minutes. The cells were then fixed with a solution of 3% formaldehyde, after which the cells were incubated in Hoechst dye (Invitrogen) for 5 minutes and washed 3 x 5 min with 1X PBS. The glass coverslips were then mounted on glass microscope slides using Fluoromount-G Mounting Medium (Thermofisher) and sealed with clear nail polish. The neurons were then visualized under a fluorescence and confocal microscope.

Results Mendelian ratio births As previously stated, TD is 3 to 4 times more likely to occur in males than in females. For this reason, the sex and genotypes of mice were noted as they were born to track whether the mice were born according to Mendelian ratios (Table 1). Aberrant ratios of births of mutant mice would indicate a possible lack of viability or a genetic predisposition. There was no statistically significant difference in the number of mice born between sexes and genotypes. Behavioral paradigms PPI PPI tests startle response and sensorimotor gating. Subjects startle when presented a loud sound but should startle less if that loud sound is preceded by another sound (a prepulse), which is a concept referred to as sensorimotor gating. The louder the preceding sound is, the lesser the startle response there should be to the main sound. PPI is a TD paradigm as TD patients show a deficiency in sensorimotor gating ability. Thus, we tested PPI as a TD behavioral paradigm. However, there were no significant differences in the percent of prepulse inhibition between the mutant and wildtype mice according to ANOVA analysis (Figure 2). Open field There are several parameters that can be tested with the open field test but only the results for time spent in the vertical (rearing) (Figure 3), total distance traveled (Figure 4), and count of central zone entries (Figure 5) are discussed. The total ambulatory distance covered by the mice can provide insight on the activity levels of mice while rearing behavior and count of central zone entries can be indicative of anxiety-related behavior (wall-hugging behavior or remaining in the periphery suggests higher anxiety levels in mice). Total ambulatory distance is measured in centimeters and tracks the total distance traveled by the mouse. The results for distance traveled and count of central zone entries are also not significantly different with ANOVA analysis, indicating that heterozygous and homozygous mutant mice do not perform differently as compared to wildtype mice in these paradigms. Marble-burying task Mice have a natural inclination to burrow in their environment, which is taken advantage of in the marble burying task. Increased number of buried marbles could indicate increased activity levels, repetitive/obsessive behavior, and/or anxietyrelated behavior. However, the marble-burying task showed no statistically significant differences in the performance of the heterozygous and homozygous mutant mice as compared to the wildtype mice according to ANOVA analysis. 177

R904S Variant of OPA1 as a Tourette Disorder Probable Risk Gene

Table 1. Chart displaying number of births for each genotype and sex. Number in parenthesis denotes the percent that births are by chance and following Mendelian ratios. Discrepancy in the total number of births per sex is due to pups that have not been genotyped yet. Average litter size is 7.2.

Figure 2. Graphs depict prepulse inhibition of subjects. Mice should startle less with increasing decibels of the prepulse administered. Thus, there should be a greater percent of prepulse inhibition with increasing decibels of prepulse — as is shown. There was no significant difference in the prepulse inhibition between genotypes for both males and females. This indicates that mutant mice do not show a deficiency in sensorimotor gating (Females: (F[2,81]=.581, p=.561) +/+ n=10, +/R904S n=14, R904S/R904S n=9; Males: +/+ n=14, +/R904S n=15, R904S/R904S n=7). Figure 3. Graphs depict the amount of time (in minutes) that the subjects spent in vertical, or the amount of time they spent reared up. There was no significant difference in the time spent in the vertical between the genotypes for both males and females (Females: (F[2,41]=.078, p=.925) +/+ n=14, +/R904S n=21, R904S/R904S n=10; Males: (F[2,66]=2.43, p=.096) +/+ n=25, +/R904S n=32, R904S/R904S n=12).

Figure 4. Graphs depict the additive distance traveled by each genotype set for males and females. There is no significant difference in the distance traveled between the genotypes for both males and females (Females: +/+ n=14, +/R904S n=21, R904S/R904S n=10; Males: (F[2,272]=1.836, p=.162) +/+ n=25, +/ R904S n=32, R904S/R904S n=12).

Figure 5. Graphs depict the average number of times subjects of each genotype entered the central zone of the open field area. There was no significant difference in the count of entry into the central entry zone between the genotypes for both males and females (Females: +/+ n=14, +/ R904S n=21, R904S/R904S n=10 (F[2,42]=.67, p=.516); Males: (F[2,49] =.667, p=.518) +/+ n=25, +/R904S n=32, R904S/R904S n=12).

178

R904S Variant of OPA1 as a Tourette Disorder Probable Risk Gene

Cellular testing Western blot to determine whole brain OPA1 levels The first cellular test performed was a western blot to determine if there were any differences in whole brain protein levels of OPA1 between genotypes in both males and females (results shown in Figure 6). Mitochondria were separated from the brain lysate to determine whether OPA1 is localizing correctly to the mitochondria alone. Two fragments of OPA1 were observed at approximately 80–100 kDa as OPA1 is proteolytically cleaved to a short and long form in the mitochondria. GAPDH was used as a positive control as it is ubiquitously expressed throughout the brain, thus it was expected to be present in both lysates. The western blot for female mice showed OPA1 solely in the mitochondria and equally expressed in wildtype and homozygous mutant brain lysates. The western blot for males was performed afterwards and includes heterozygous mouse brain lysates and a mitochondrial specific control, Hsp60. Again, OPA1 appeared to be equally expressed in brain lysates of all three genotypes. The western blot for females was subject to human error. First, there appeared to be Hsp60 and OPA1 signals present in the mitochondria exclusion fragment (whole brain lysate without the mitochondria). This could be due to either contamination between the two groups or insufficient separation of mitochondria from the exclusion fragment. There also appeared to be a nonspecific band on the membrane section incubated with Hsp60, which could be due to nonspecific binding, excessive incubation in Hsp60, or insufficient blocking incubation. GAPDH was also not present in every lane as expected, which could be due to excessive washing of that membrane section after incubation. Although there appeared to be light bands of OPA1 in the mitochondria exclusion fragment, we presumed that to be a result of either the contamination or insufficient separation of the mitochondria. However, there appeared to be no difference in the level of OPA1 protein expression between genotypes, which was our primary determination objective in carrying out the western blot. Brain slice staining to determine localization of OPA1 We obtained 10x objective tile images of wildtype and homozygous mutant mice

Figure 6: Excl: mitochondria excluded fragment, Mito: isolated mitochondria. This figure depicts the western blot of male and female subjects. Two fragments of OPA1 are seen between 80–100 kDa because OPA1 is proteolytically cleaved to a short and long form. GAPDH served as a positive control, as it is present ubiquitously throughout the brain. Hsp60 served as a positive control for the mitochondrial fragment as it is exclusive to the mitochondria. There are slight Hsp60 and OPA1 bands in the male exclusion fragments, indicating that there was slight contamination of mitochondria in the male exclusion fragments. However, OPA1 clearly appears to be equally strong in the male mitochondrial fragments between the genotypes. There is also a strong nonspecific band appearing in the male Hsp60 fragment of the gel, which can be a result of excessive Hsp60 incubation. Overall, the OPA1 protein bands appear to be equally intense in both wildtype and homozygous mutant of both males and females, indicating that there is no significant difference in whole brain expression between the two genotypes.

Figure 7. 10x objective tile images of brain slices of wildtype and homozygous mutant mice. Stained for nuclei with Hoechst dye (blue) and OPA1 using an OPA1 primary antibody (NovusBio) and Alexa Fluorescence 555 secondary antibody (Invitrogen).

179

R904S Variant of OPA1 as a Tourette Disorder Probable Risk Gene

brain slices for nuclei with Hoechst dye (blue) (Invitrogen), OPA1 using an OPA1 primary antibody (red) (NovusBio) and Alexa Fluorescence 555 secondary antibody (Invitrogen) (Figure 7). Brain slices were stained for OPA1 to determine whether there were brain regions with varying expression in homozygous mutant mice as compared to wildtype mice. We struggled to acquire a strong signal for OPA1. Initially, we utilized Alexa Fluorescent 488 and 657 secondary antibodies (Invitrogen), which were not successful due to low signal output in comparison to the Hoechst stain. This made imaging difficult, thus when we incubated with the 555 Alexa Fluorescent secondary antibody (Invitrogen), we decreased our wash steps to minimize washing away our secondary antibody. This was a preliminary experiment, where our primary goal was to optimize our procedure and imaging process. Therefore, although these results appear to indicate increased OPA1 levels in the homozygous mutant slices, this cannot be supported. There are not enough brain slice comparisons, the brain slices may not be sufficiently aligned to compare between the wildtype and homozygous mutant slices, and the imaging setup may not be optimal. However, this preliminary experiment allowed us to identify the 555 Alexa Fluorescent secondary antibody (Invitrogen) as capable of producing a strong signal that can be visualized with the Hoechst stain. Cortical neurons stained with MitoTracker to observe mitochondria morphology 63x and 100x objective images utilizing oil on a confocal microscope were taken of cortical neuron nuclei stained blue with Hoechst dye (Invitrogen) and mitochondria stained red with MitoTracker Red (Thermofisher) (Figure 8). The mitochondria appeared to localize correctly throughout the cytoplasm in all genotypes. However, we observed potential mitochondrial aggregates in heterozygous and homozygous mutant neurons that do not seem to be present in wildtype neurons (marked with white arrows). Mitochondrial aggregation may be caused by aberrant OPA1 functioning. A prior study found that siRNA mediated downregulation of OPA1 resulted in an aggregated mitochondria phenotype (16). The R904S mutation could be functioning to downregulate the protein expression by an amount that is not detectable through whole brain western blotting. Previous studies have also found that mitochondrial aggregation could indicate a proapoptotic state as aggregation has been found to be an event that precedes cytochrome c release (17). Thus, the R904S mutant neurons may also be more prone to apoptosis.

Discussion OPA1 has previously been identified as a probable risk gene for TD (5), which is a heritable, neuropsychiatric and developmental disorder characterized by motor and vocal tics. OPA1 is a protein involved in the fusion and membrane dynamics of the inner mitochondrial membrane. After the Tischfield laboratory identified a TD patient with the R904S mutant of OPA1, we hypothesized that the mutation would result in behavioral and/ or cellular abnormalities as compared to wildtype cells/mice. Essentially, we were searching for a visible phenotype caused by the mutation that could be characterized as a TD model phenotype. This is an important endeavor because TD is not well studied and OPA1’s involvement with the disorder has not been studied previously. 180

Figure 8. Cortical neurons of homozygous mutant, heterozygous and wildtype mice stained for nuclei with Hoechst dye (blue) and mitochondria with MitoTracker Red. White arrows distinguish possible mitochondrial aggregates that appear to be present in homozygous mutant and heterozygous neurons, while absent in wildtype neurons. Images were taken under 63x and 100x objective utilizing oil on a confocal microscope.

The behavioral paradigms tested were motion sequencing, open field arena test, marble-burying task, and prepulse inhibition. No statistically significant differences have been found in the performance of the mice among these tests. This can be because the mechanism of the variant is not addressed by these assays. Because we cannot know what phenotype the mutation induces, we cannot know which assay will identify it. Although we are aware of the symptoms of TD in humans, we are not certain how tics and other symptoms of TD manifest in mice. Additionally, TD is more than a motor disorder — there are also psychological and neurological aspects which place TD symptoms on a spectrum of severity. Thus, behavioral assays alone cannot encompass the complexity and dynamic nature of the disorder. Cellular testing is crucial because understanding the mechanism of the variant can help to pinpoint how the mutation may be affecting the mouse phenotypically. This in turn can help identify behavioral assays that address the behaviors that may be affected by the mutation. We utilized western blotting to observe whole brain protein expression, which showed no visible differences in expression levels. We were also working to identify certain brain regions with higher or lower OPA1 expression levels to potentially identify regions of interest. A better understanding of the protein’s expression throughout the brain can also provide a better understanding of the potential pathways that the protein may be involved in. Although we obtained preliminary results, they were not sufficient to conclude the nature of OPA1 level in different brain regions, as we did not have an appropriate amount of brain slices for comparison and our staining procedure was not optimized.

R904S Variant of OPA1 as a Tourette Disorder Probable Risk Gene

Lastly, we visualized mitochondria in cortical neurons and observed potential mitochondrial aggregation in heterozygous and homozygous mutant neurons that was not present in wildtype neurons. Studies have shown that mitochondrial aggregation can be a result of reduced OPA1 expression and that it may suggest a preapoptotic state as aggregation precedes cytochrome c release (16, 17). Thus, it is possible that the R904S mutation either encourages apoptosis or works in downregulating the protein’s expression. As there were no visible decreases in protein expression with the western blot, it is possible that the downregulation is not drastic enough to be qualitatively visible, or the mutation functions in inactivating the protein rather than altering its expression. This project is the first attempt at identifying a behavioral or cellular assay for TD with OPA1. There are several potential explanations for inconclusive results in the assays done thus far. As previously stated, it is possible that the assays done were not relevant to the mutation. For instance, the mutation may be affecting mitochondrial function rather than its appearance and localization. It is also possible that OPA1 is not a gene that contributes to TD, or that the TD patient identified with the R904S mutation has other TD risk genes that are contributing to their disorder. Thus, it is also possible that the TD patient found with the R904S OPA1 mutation may have other genetic mutations that may be contributing to their TD diagnosis. Lastly, the testing thus far has been done on mice between the ages of P2 to approximately 3 months old — thus it is possible that a phenotype manifests in older mice, which is not something that we have investigated yet. So far, we have tested behavior and protein/mitochondrial characteristics, thus additional future directions would be to test mitochondrial function. Due to the observed mitochondrial aggregates, the cristae and its functions specifically should be investigated, such as detailed cristae structure. A previous study has found little to no cristae formation in OPA1-KO mitochondria with transmission EM imaging and loss of cristae tightness in GTPase-defective OPA1 mutant mitochondria (6). Cytochrome c levels and apoptosis potential should also be investigated as mitochondrial aggregation is thought to play a role in apoptosis signaling. Additionally, oxidative phosphorylation and mitochondrial energetics should be studied as there are cases of patient cells with oxidative phosphorylation defects that have no apparent change in mitochondrial morphology (18). Lastly, dOPA1 mutations in Drosophila have been found to cause elevated reactive oxygen species (19). As increased levels of reactive oxygen species indicates stressful conditions in the cell or mitochondria, health of the mitochondria and cell can be tested for using MitoSox and CellRox, which are fluorescent dyes that identify reactive oxygen species in the mitochondria and whole cell respectively. TD is not well studied because of its complex etiology. Understanding the R904S mutation of OPA1 would increase our understanding of Tourette disorder pathology, which may help to develop therapeutic methods to target and ameliorate the debilitating symptoms of the disorder.

Acknowledgments I would like to acknowledge all members of the Tischfield laboratory for supporting me in my research endeavors. I

express my deepest gratitude to my mentors Jay Tischfield, and Cara Nasello for their advice and support throughout the project. I would also like to thank Mikaela Hufnell for her work with maintenance of mouse colonies and her assistance in the animal facility, Sam Iofel and Luis Ramirez for aiding in genotyping mice, Noriko Goldsmith, for her guidance of imaging equipment, Josh Thackray for his technical support, and Yurdiana Garcia for her assistance in perfusion methods.

References 1.

Diagnosing Tic Disorders | CDC [Internet]. [cited 2021 May 1]. Available from: https://www.cdc.gov/ncbddd/ tourette/diagnosis.html

Georgitsi M, Willsey AJ, Mathews CA, State M, Scharf JM, Paschou P. The genetic etiology of tourette syndrome: Large-scale collaborative efforts on the precipice of discovery. Front Neurosci [Internet]. 2016 Aug 3 [cited 2021 May 1];10(AUG):351. Available from: /pmc/articles/ PMC4971013/

Cavanna AE, Black KJ, Hallett M, Voon V. Neurobiology of the premonitory urge in tourette’s syndrome: Pathophysiology and treatment implications. J Neuropsychiatry Clin Neurosci [Internet]. 2017 Mar 1 [cited 2021 May 1];29(2):95–104. Available from: /pmc/articles/ PMC5409107/

Willsey AJ, Fernandez T V., Yu D, King RA, Dietrich A, Xing J, et al. De Novo Coding Variants Are Strongly Associated with Tourette Disorder. Neuron [Internet]. 2017 May 3 [cited 2021 May 1];94(3):486-499.e9. Available from: / pmc/articles/PMC5769876/

Wang S, Mandell JD, Kumar Y, Sun N, Morris MT, Arbelaez J, et al. De Novo Sequence and Copy Number Variants Are Strongly Associated with Tourette Disorder and Implicate Cell Polarity in Pathogenesis. Cell Rep [Internet]. 2018 Sep 25 [cited 2021 May 1];24(13):3441-3454.e12. Available from: /pmc/articles/PMC6475626/

Lee H, Yoon Y. Mitochondrial membrane dynamics— functional positioning of OPA1. Antioxidants [Internet]. 2018 Dec 1 [cited 2021 May 1];7(12). Available from: / pmc/articles/PMC6316456/

Guo R, Gu J, Zong S, Wu M, Yang M. Structure and mechanism of mitochondrial electron transport chain [Internet]. Vol. 41, Biomedical Journal. Elsevier B.V.; 2018 [cited 2021 May 1]. p. 9–20. Available from: /pmc/articles/ PMC6138618/

Alavi MV, Fuhrmann N. Dominant optic atrophy, OPA1, and mitochondrial quality control: Understanding mitochondrial network dynamics [Internet]. Vol. 8, Molecular Neurodegeneration. BioMed Central; 2013 [cited 2021 May 1]. p. 32. Available from: http:// molecularneurodegeneration.biomedcentral.com/ articles/10.1186/1750-1326-8-32

Belenguer P, Pellegrini L. The dynamin GTPase OPA1: More than mitochondria? Vol. 1833, Biochimica et Biophysica Acta - Molecular Cell Research. Elsevier; 2013. p. 176–83.

181

R904S Variant of OPA1 as a Tourette Disorder Probable Risk Gene

10. Wiltschko AB, Johnson MJ, Iurilli G, Peterson RE, Katon JM, Pashkovski SL, et al. Mapping Sub-Second Structure in Mouse Behavior. Neuron [Internet]. 2015 [cited 2021 May 1];88(6):1121–35. Available from: /pmc/articles/ PMC4708087/ 11. Datta SR. Understanding Action [Internet]. 2018 [cited 2021 May 1]. Available from: http://datta.hms.harvard.edu/ research/behavioral-analysis/ 12. Seibenhener ML, Wooten MC. Use of the open field maze to measure locomotor and anxiety-like behavior in mice. J Vis Exp [Internet]. 2015 Feb 6 [cited 2021 May 1];(96):52434. Available from: /pmc/articles/ PMC4354627/ 13. Angoa-Pérez M, Kane MJ, Briggs DI, Francescutti DM, Kuhn DM. Marble burying and nestlet shredding as tests of repetitive, compulsive-like behaviors in mice. J Vis Exp [Internet]. 2013 [cited 2021 May 1];82(82):50978. Available from: /pmc/articles/PMC4108161/ 14. Geyer MA, Dulawa SC. Assessment of Murine Startle Reactivity, Prepulse Inhibition, and Habituation. Curr Protoc Neurosci [Internet]. 2003 Jul 1 [cited 2021 May 1];24(1):8.17.1-8.17.15. Available from: https:// onlinelibrary.wiley.com/doi/10.1002/0471142301. ns0817s24 15. Yael D, Israelashvili M, Bar-Gad I. Animal models of tourette syndrome-from proliferation to standardization. Front Neurosci [Internet]. 2016 Mar 31 [cited 2021 May 1];10(MAR):132. Available from: /pmc/articles/ PMC4814698/ 16. Kamei S, Chen-Kuo-Chang M, Cazevieille C, Lenaers G, Olichon A, Bélenguer P, et al. Expression of the Opa1 mitochondrial protein in retinal ganglion cells: Its downregulation causes aggregation of the mitochondrial network. Investig Ophthalmol Vis Sci [Internet]. 2005 Nov 1 [cited 2021 May 1];46(11):4288–94. Available from: http:// www.ncbi.nlm.nih.gov/Genbank; 17. Haga N, Fujita N, Tsuruo T. Mitochondrial aggregation precedes cytochrome c release from mitochondria during apoptosis. Oncogene [Internet]. 2003 Aug 28 [cited 2021 May 1];22(36):5579–85. Available from: www.nature.com/ onc 18. Mishra P, Carelli V, Manfredi G, Chan DC. Proteolytic cleavage of Opa1 stimulates mitochondrial inner membrane fusion and couples fusion to oxidative phosphorylation. Cell Metab [Internet]. 2014 Apr 1 [cited 2021 May 1];19(4):630–41. Available from: /pmc/articles/ PMC4018240/ 19. Tang S, Le PK, Tse S, Wallace DC, Huang T. Heterozygous mutation of Opa1 in Drosophila shortens lifespan mediated through increased reactive oxygen species production. PLoS One [Internet]. 2009 Feb 16 [cited 2021 May 1];4(2). Available from: /pmc/articles/PMC2637430/

182

Scholarly Research In Progress • Vol. 5, November 2021

Why is There a Ten-Fold Variation between States in Clozapine Usage among Medicaid Enrollees in the United States? Rizelyn A. Benito1*‡, Michael H. Gatusky1†‡, Mariah W. Panoussi1*‡, Anisa S. Suparmanian2, and Brian J. Piper1 ¹Geisinger Commonwealth School of Medicine, Scranton, PA 18509 ²The Wright Center for Graduate Medical Education, Scranton, PA 18505 *Master of Biomedical Sciences Program † Doctor of Medicine Program ‡ Authors contributed equally Correspondence: mpanoussi@som.geisinger.edu

Abstract Background: Clozapine was the first atypical antipsychotic for treating schizophrenia, with a long history of controversy over its usage. Guidelines currently recommend clozapine for patients diagnosed with refractory schizophrenia. However, prescribers are underutilizing clozapine because of the costs associated with close monitoring of its adverse effects, particularly agranulocytosis. This is unfortunate because clozapine has demonstrated greater effectiveness compared with other antipsychotics. It is essential to examine clozapine usage to determine if it is being adequately utilized in the treatment of schizophrenia in the United States. Methods: Medicaid data, including the number of quarterly clozapine prescriptions and the number of Medicaid enrollees in each state from 2015 to 2019, was collected and used to evaluate clozapine use over time. Data analysis and figures were prepared with Excel. Results: The number of prescriptions, corrected for the number of enrollees in Medicaid, was generally consistent over time. However, average prescriptions per quarter were markedly lower in 2017 compared with other years, decreasing by 44.4% from 2016 average prescriptions per quarter. From 2015 to 2019, states from the upper Midwest and Northeast regions of the country had the highest average clozapine prescriptions per 10,000 Medicaid enrollees (ND: 190.0, SD: 176.6, CT: 166.2). States from the Southeast and Southwest had much lower average rates (NV: 17.9, KY: 19.3, MS: 19.7). Overall, there was a 10-fold difference in clozapine prescriptions between states from 2015-2019 (2015 = 19.9-fold, 2016=11.4 fold, 2017=11.6 fold, 2018=13.3 fold, and 2019=13.0 fold). There was a moderate correlation of (r(48)=0.50, p < 0.05) between prescriptions per 10,000 enrollees and the Medicaid spending per enrollee in each state in 2019. Conclusion: Clozapine is an important pharmacotherapy for refractory schizophrenia. Overall, clozapine use tends to be highest among the upper Midwest and Northeast states. Further research is ongoing to better understand the origins of the 10-fold regional disparities in clozapine use.

Introduction Clozapine is a second-generation antipsychotic (SGA) and is also generally known as an atypical antipsychotic (1). Current clinical guidelines state that clozapine is an effective treatment for those with treatment-resistant schizophrenia (TRS) (2). Presently, the clinical and research criteria are failure of two trials of non-clozapine antipsychotics of standard dose and

duration (3). However, current evidence demonstrates a lack of clozapine utilization by providers (4). Clozapine has been known to cause numerous side effects that range from hyperglycemia and weight gain to more serious reactions such as seizures and myocarditis (1). One particular adverse effect correlates with drug induced agranulocytosis in patients which leads to increased susceptibility and death from infectious diseases (5). Another factor that further complicates this controversy over clozapine is the high cost associated with close monitoring for agranulocytosis (2). To put this into perspective, the financial burden schizophrenia imposes annually on patients is around $23 billion, while patients diagnosed with TRS cost $34 billion annually (6). Direct healthcare costs in the United States (U.S.) are estimated to be 3- to 11-fold higher for TRS patients, which includes multiple hospitalizations (3). Nevertheless, recent research has revealed clozapine along with other SGAs as more cost effective (7). Besides adverse effects, other impediments influencing clozapine underutilization include complications when administration and registry changed from individual pharmaceutical companies to a single clozapine program (8). The new single registry has new requirements and procedures that have hindered patient initiation on clozapine (8). For example, it has interfered with provider collaborations specifically when transitioning a patient’s care (8). Another aspect is inadequate centralized resources to benefit patients and their families with services such as patient education and adherence monitoring (8). Additionally, failure to account for benign ethnic neutropenia when considering clozapine for African American patients as well as the scarcity of adequate psychiatric services in correctional facilities are obstacles leading to clozapine underutilization (8). There's also the interference of physicians' knowledge, perspectives, and attitudes affecting clozapine usage in various countries (9–10). Recent surveys have shown that 40.5% and 64% of physicians prefer other antipsychotics or combine two antipsychotics before considering clozapine (9, 11). One research survey discovered that 66% of psychiatrists stated that their patients were less satisfied when treated with clozapine compared to those treated with other atypical antipsychotics (11). Specifically in the U.S., researchers had 295 providers complete a questionnaire where 33% of providers reported prescribing clozapine after three or more unsuccessful antipsychotics have been tried (12). There was also unanimous physician reluctance with prescribing clozapine because of inadequate knowledge or experience with clozapine and concerns of patient compliance (8–10, 12). 183

Clozapine Usage among Medicaid Enrollees in the United States

The barriers described above are causing a visible variation in clozapine use throughout the U.S. states. One study noted that overall, 4.8% of schizophrenic patients were on clozapine with a slight decline during 2001 to 2005 (13). This analysis revealed that clozapine was used least by Deep South states, while states in New England, the Rocky Mountain region, and Washington had frequent use (13). A more recent review in 2016 confirmed this previous data by unveiling that 15.6% of Medicaid recipients with schizophrenia in South Dakota received clozapine compared to 2% in Louisiana (13–14). Therefore, there is a need to further understand the underlying cause for the variation of clozapine use among U.S. states. Past studies have endeavored to uncover possible reasons for this variation, yet it still remains ambiguous (13–14). Thus, the aim of our research was to conduct a secondary analysis of Medicaid data from 2015 to 2020 to further investigate the underlying cause(s) of clozapine usage variation throughout the U.S.

Methods The data was obtained from the data.medicaid.gov database for the years of 2015 to 2020 (15) which focused on the nationwide drug use of clozapine. Data was collected for clozapine and the brand name counterparts, Versacloz, Clozaril, and Fazaclo. The number of prescriptions per 10,000 Medicaid enrollees was calculated to find a standardized prescription rate for each state using Excel. Bar graphs and heat maps were used to analyze the overall trends of clozapine use by each state. Line graphs were used to visualize clozapine prescription rates in each quarter from 2015 to 2020. The 2019 Medicaid data was used to calculate the dollars spent per Medicaid enrollee, and the Pearson correlation was subsequently determined between dollars spent per enrollee and clozapine prescription rates. Additionally, a pie chart comparing generic clozapine to its brand name counterpart was generated. Prism was used to create these figures. Ninety-five percent confidence intervals were also calculated for the prescriptions per 10,000 enrollees. The fold difference was calculated between the states with the highest and lowest average prescriptions per 10,000 from 2015 to 2019. Pearson correlation coefficients were calculated for amount spent per enrollee, percent white population, and percent rural population in each state compared with prescription rates.

Results Upon calculating the prescriptions of clozapine in each state per 10,000 Medicaid enrollees, there were substantial differences among states’ prescription rates. Figure 1 shows this nationwide comparison for 2015 to 2019. North Dakota was shown to have the highest rate of prescribing clozapine, with 190 prescriptions per 10,000 Medicaid enrollees. The lowest prescribing state was Nevada, with 17.9 prescriptions per 10,000 Medicaid enrollees. The average rate was 80.4 prescriptions per 10,000 Medicaid enrollees with a standard deviation of 44.7. The 95% confidence interval was from -7.2 to 168 prescriptions per 10,000 enrollees with North and South Dakota above the range of the confidence interval. Finally, there was a 10.6-fold difference between the lowest- and highest-prescribing states, Nevada and North Dakota, respectively.

184

Figure 1. State use of clozapine, ranked, per 10,000 Medicaid enrollees for 2015 to 2019. Gray states were significantly elevated (p < 0.05) relative to the national mean (80.4 prescriptions with a standard deviation of 44.7, 95% Confidence interval = -7.2 to 168.0).

Figure 2A and 2B represents the nationwide usage of clozapine per 10,000 Medicaid enrollees per state in 2018 and 2019 respectively. As shown in Figure 2A, in 2018 North Dakota had the highest clozapine prescription rate with a value of 221.9 prescriptions per 10,000 Medicaid enrollees, while the state with the lowest rate was Nevada, with the value of 16.7 prescriptions per 10,000 enrollees. Figure 2B shows that South Dakota had the highest clozapine prescription rate in 2019 with a value of 191.6 prescriptions per 10,000 Medicaid enrollees. Arkansas had the lowest rate with an average value of 14.8 prescriptions per 10,000 enrollees. Clozapine prescriptions remained relatively constant from 2015 through the first half of 2020 with a considerable drop in total Medicaid prescriptions in 2017. Figure 3 shows the overall number of U.S. clozapine prescriptions per quarter during this timeframe, including a 44.4% drop from 2016 average prescriptions per quarter to 2017 prescriptions per quarter. The highest amount of clozapine being prescribed was during the third quarter of 2016, with a value of 173,087 prescriptions. The lowest amount of clozapine prescribed was during the third quarter of 2017, with a value of 80,070 prescriptions. We also found a stark difference between generic and brand name clozapine prescriptions. Figure 4 illustrates that the vast

Clozapine Usage among Medicaid Enrollees in the United States

Figure 2. Heatmap of clozapine usage per 10,000 Medicaid enrollees for 2018 (A) and 2019 (B).

Figure 3. Number of Medicaid prescriptions per quarter for clozapine during 2015 to quarter 2 of 2020.

Figure 4. Prescriptions of generic (579,875) and brand name (5,342, Fazaclo, Versacloz, and Clozaril) clozapine in 2019 in Medicaid.

majority of clozapine prescriptions in 2019 were generic. In total 579,875 prescriptions or 99.09% of nationwide Medicaid prescriptions were generic. The remaining 0.91% of clozapine prescriptions were brand name — Fazaclo, Versacloz, and Clozaril — which were only prescribed 5,342 times in 2019. Additionally, we found a significant correlation between Medicaid spending and the clozapine prescription rates of states. Figure 5 demonstrates this positive correlation between clozapine prescriptions per 10,000 enrollees and the Medicaid spending per enrollee in each state in 2019 (r(48)=0.50, p < 0.05).

Figure 5. Scatterplot between clozapine prescriptions per 10,000 enrollees and the Medicaid spending per enrollee in each state (r(48) = 0.50, p < 0.01).

Pearson coefficients were also calculated between states’ rural population percentage according to the 2010 Census and clozapine prescriptions per 10,000 Medicaid enrollees. The correlations were minimal at 0.072, 0.071, -0.004, -0.034, and -0.021 from 2015 through 2019, respectively. Small positive Pearson coefficients were calculated between states’ percent white population and clozapine prescriptions per 10,000. The coefficients were 0.36, 0.44, 0.34, 0.32, and 0.38 (all p < 0.05) from 2015 through 2019 respectively.

185

Clozapine Usage among Medicaid Enrollees in the United States

Discussion This study identified pronounced (10-fold) regional disparities in clozapine prescriptions in the Medicaid database throughout the years of 2015 to 2019. States in the Southeast and Southwest U.S. tended to have lower clozapine prescription rates, with Nevada, Kentucky, Mississippi, New Mexico, and Florida having the lowest average rates per 10,000 from 2015 to 2019. Furthermore, several low-prescription states in the Southeast and Southwest tended to have higher non-white populations compared with higher-prescription rate states such as North and South Dakota. Of the five states with the lowest prescription rates, all but Kentucky were in the top third of states with the highest non-white populations in the U.S. according to Kaiser Family Foundation estimates for 2019 (16). Conversely, North and South Dakota were in the bottom third for nonwhite population. Additionally, our analysis revealed small to moderate (0.32 to 0.48) significant positive correlations in all years between percent white population in a state and clozapine prescription rates. Overall, the 10-fold variation among the states in clozapine prescriptions aligns with past data showing that being white was a factor significantly associated with clozapine initiation (17). Furthermore, states in the Southeast have among the highest Black populations in the U.S. There is also data showing that Black Americans are less likely to be prescribed clozapine than white Americans (18). In addition to factors such as prescriber bias and the anticipation of nonadherence, the presence of benign ethnic neutropenia may impact prescription rates among Black patients (18–19). Benign ethnic neutropenia is an unexplained neutrophil count of < 1.5 × 109/L, which does not confer an increased risk of infection and is commonly seen in individuals of African and Afro-Caribbean descent, as well as some Middle Eastern ethnic groups (8, 18). Although the U.S.’s threshold for clozapine discontinuation is lower than that of the United Kingdom and FDA monitoring criteria were revised in 2015, there is still evidence that clozapine is underprescribed in Black populations and that Black patients are more likely to discontinue clozapine (8, 18–19). Practitioners may have been slow to adopt these new recommendations due to a lack of product labeling and concerns about using lower neutrophil count thresholds (8). Furthermore, benign ethnic neutropenia is a diagnosis of exclusion and is likely underdiagnosed (18). Past studies cited rural geographic location as an influence on both the prevalence of treatment resistant schizophrenia and on the availability of psychiatric treatment (3, 4, 20). Investigators found that living in a rural region was a predictor for treatmentresistant schizophrenia, suggesting that states with higher rural populations may have higher clozapine prescription rates (3). Conversely, the Health Resources and Services Administration, a part of the Department of Health and Human Services, found that rurality was the most common marker of counties with primary care health professional shortages across the U.S. (20). This suggests that counties with higher rural populations may also lack access to psychiatric care and thus receive less clozapine prescriptions proportionally. However, our analysis showed no significant correlation between percent rural population per state and clozapine prescriptions per state. This data analysis led to speculation that Medicaid expansion during the early 2010s would be linked with the upward trend

186

in clozapine usage. Surprisingly, there is a sizable decrease in clozapine prescriptions during 2017. One possible explanation for this sudden decrease could be related to the addition of long-acting injectable (LAI) antipsychotics to Medicaid’s preferred list in some states. For example, the Pennsylvania Department of Human Services Preferred Drug List now has haloperidol lactate syringe as a preferred agent (21). Currently, there are researchers that have unveiled that LAIs improve patient adherence and reduce healthcare costs (22–25). Three of these studies used Medicaid data to review the effectiveness of LAIs and cost, and each concluded that these drugs improved patient adherence and overall costs were reduced due to decreased hospitalizations (23–25). Thus, LAIs are being seen as more beneficial due to their efficiency in managing schizophrenia (22, 24). Interestingly, one research group stated that oral antipsychotics were similar in cost effectiveness to LAIs, while another found LAIs to reduce health care costs by half when compared to oral antipsychotics (23, 25). Hence, further investigation is required to fully determine if there is a difference between LAIs and SGA oral antipsychotic’s efficiency in managing TRS and schizophrenia along with their related health care costs. A limitation of this study is that Medicaid data for 2020 was incomplete at the time of data collection and analysis. There is also limited knowledge regarding the specific reasons for the drop in clozapine usage during 2017. Further studies that include electronic medical records will be required to narrow down accurate reasons to comprehend this sudden change in the clozapine trend. Future directions of this research also include an evaluation of the trends in utilization of generic and brand name clozapine. This would be invaluable knowledge, because of concerns with the safety and efficacy of generic clozapine (26). It is also problematic because as shown in Figure 4, 99.09% of prescriptions in 2019 were for generic clozapine. Evidence has shown that when some patients are switched from Clozaril (brand name) to a generic form, they exhibit relapse or exacerbation of schizophrenic symptoms (26–27). Therefore, research in this area would clarify if generic clozapine is harmful and if so, could lead to safer and efficacious versions that are less costly. Additionally, it would be valuable to research the costs of medication between different types and states in the future. It may also be possible that costs and rates of reimbursement for clozapine prescriptions, as well as the lab work necessary to monitor possible agranulocytosis, may have influenced prescription rates across the U.S. For instance, psychiatrists in states with lower clozapine costs or higher rates of reimbursement for hematological monitoring may be more likely to prescribe it. Lastly, the effects of step therapy to clozapine use throughout the years may affect the rates of clozapine prescription and should be analyzed in future research. Step therapy requires that patients undergo a series of treatments (generally lower cost) prior to prescription of another drug such as clozapine (28). It is likely that clozapine prescription rates will differ significantly in states where clozapine prescription is restricted by step therapy.

Clozapine Usage among Medicaid Enrollees in the United States

Conclusion There was a significant 10-fold difference between the states’ clozapine usage. There were lower average clozapine prescription rates in the Southeast and Southwest, which can be attributed to numerous factors. Significant associations with Medicaid spending and race were identified. Factors that are possibly influencing clozapine prescription numbers are physician reluctance, serious adverse effects (e.g., agranulocytosis), LAIs added to Medicaid’s preferred drug list, and more. Therefore, future investigation is required to uncover the true influential factors that cause clozapine usage to be underutilized in some regions of the U.S.

Acknowledgments We would like to thank the Biomedical Research Club for their support. We would also like to thank Kenneth L. McCall, PharmD, with the Department of Pharmacy Practice at the University of New England in Portland, ME, USA for his feedback and insight on this subject.

Disclosures Brian J. Piper, PhD, is involved with osteoarthritis research supported by Eli Lilly and Company and Pfizer.

References 1.

Warnez S, Alessi-Severini S. Clozapine: a review of clinical practice guidelines and prescribing trends. BMC Psychiatry 2014; 14: 102. Crilly J. The history of clozapine and its emergence in the US market: A review and analysis. Hist Psychiatry 2007; 18: 39–60.

Smart SE, Kępińska AP, Murray RM, et al. Predictors of treatment resistant schizophrenia: a systematic review of prospective observational studies. Psychol Med; 51: 44–53.

Farooq S, Choudry A, Cohen D, et al. Barriers to using clozapine in treatment-resistant schizophrenia: systematic review. BJPsych Bulletin 2019; 43: 8–16.

Alvir JM, Lieberman JA, Safferman AZ, et al. Clozapineinduced agranulocytosis. Incidence and risk factors in the United States. N Engl J Med 1993; 329: 162–167.

Desai PR, Lawson KA, Barner JC, et al. Estimating the direct and indirect costs for community-dwelling patients with schizophrenia. Journal of Pharmaceutical Health Services Research 2013; 4: 187–194.

Phanthunane P, Vos T, Whiteford H, et al. Costeffectiveness of pharmacological and psychosocial interventions for schizophrenia. Cost Eff Resour Alloc 2011; 9: 6.

Kelly DL, Freudenreich O, Sayer MA, et al. Addressing barriers to clozapine underutilization: A national effort. Psychiatr Serv 2018; 69: 224–227.

Whiskey E, Barnard A, Oloyede E, et al. An evaluation of the variation and underuse of clozapine in the United Kingdom. Acta Psychiatr Scand; 143: 339–347.

10. Tungaraza TE, Farooq S. Clozapine prescribing in the UK: views and experience of consultant psychiatrists. Ther Adv Psychopharmacol 2015; 5: 88–96. 11. Nielsen J, Dahm M, Lublin H, et al. Psychiatrists’ attitude towards and knowledge of clozapine treatment. J Psychopharmacol (Oxford) 2010; 24: 965–971. 12. Daod E, Krivoy A, Shoval G, et al. Psychiatrists’ attitude towards the use of clozapine in the treatment of refractory schizophrenia: A nationwide survey. Psychiatry Res 2019; 275: 155–161. 13. Olfson M, Gerhard T, Crystal S, et al. Clozapine for schizophrenia: State variation in evidence-based practice. Psychiatr Serv 2016; 67: 152. 14. Torrey EF. Why clozapine use varies by state. Psychiatr Times; 33, https://www.psychiatrictimes.com/view/whyclozapine-use-varies-state (2016, accessed May 6, 2021). 15. Data.Medicaid.gov. Drug Utilization 2020 - National Totals, https://data.medicaid.gov/State-Drug-Utilization/DrugUtilization-2020-National-Totals/y9i9-8pq3 (accessed November 9, 2020). 16. KFF estimates based on the 2008-2019 American Community Survey, 1-Year Estimates. Kaiser Family Foundation, https:// www.kff.org/other/state-indicator/distribution-by-racee thnicity/?currentTimeframe=0&sortModel=%7B%22col Id%22:%22Location%22,%22sort%22:%22asc%22%7D (accessed May 21, 2021). 17. Stroup TS, Gerhard T, Crystal S, et al. Geographic and clinical variation in clozapine use in the United States. Psychiatr Serv 2014; 65: 186–192. 18. Mijovic A, MacCabe JH. Clozapine-induced agranulocytosis. Ann Hematol 2020; 99: 2477–2482. 19. Kelly DL, Kreyenbuhl J, Dixon L, et al. Clozapine underutilization and discontinuation in African Americans due to leucopenia. Schizophr Bull 2007; 33: 1221–1224. 20. Streeter RA, Snyder JE, Kepley H, et al. The geographic alignment of primary care Health Professional Shortage Areas with markers for social determinants of health. PloS One 2020; 15: e0231443. 21. Pennsylvania Department of Human Services. Pennsylvania Medical Assistance Statewide Preferred Drug List (PDL). Pennsylvania Department of Human Services, https://papdl. com/sites/default/files/ghs-files/Penn%20Statewide%20 PDL%2001.05.21.pdf (2021, accessed May 19, 2021). 22. Kamat SA, Offord S, Docherty J, et al. Reduction in inpatient resource utilization and costs associated with long-acting injectable antipsychotics across different age groups of Medicaid-insured schizophrenia patients. Drugs Context; 4. Epub ahead of print March 17, 2015. DOI: 10.7573/dic.212267. 23. Pilon D, Tandon N, Lafeuille M-H, et al. Treatment patterns, health care resource utilization, and spending in Medicaid beneficiaries initiating second-generation long-acting injectable agents versus oral atypical antipsychotics. Clin Ther 2017; 39: 1972-1985.e2.

187

Clozapine Usage among Medicaid Enrollees in the United States

24. Bera R, Offord S, Zubek D, et al. Impact on healthcare resource usage and costs among Medicaid-insured schizophrenia patients after initiation of treatment with long-acting injectable antipsychotics. J Med Econ 2013; 16: 522–528. 25. Joshi K, Lafeuille M-H, Kamstra R, et al. Real-world adherence and economic outcomes associated with paliperidone palmitate versus oral atypical antipsychotics in schizophrenia patients with substance-related disorders using Medicaid benefits. J Comp Eff Res 2018; 7: 121–133. 26. Turkmani K, Marwa ME, Ahmad B, et al. Clozapine (generic versus branded) for people with schizophrenia. Cochrane Database Syst Rev. Epub ahead of print January 24, 2019. DOI: 10.1002/14651858.CD013248. 27. Kluznik JC, Walbek NH, Farnsworth MG, et al. Clinical effects of a randomized switch of patients from clozaril to generic clozapine. J Clin Psychiatry 2001; 62 Suppl 5: 14–7; discussion 23. 28. Gifford K, Winter A, Wiant L, et al. How State Medicaid Programs are Managing Prescription Drug Costs: Results from a State Medicaid Pharmacy Survey for State Fiscal Years 2019 and 2020. Kaiser Family Foundation, https://www. kff.org/report-section/how-state-medicaid-programs-aremanaging-prescription-drug-costs-cost-containment-andutilization-control-strategies/ (April 29, 2020, accessed May 20, 2021).

188

Scholarly Research In Progress • Vol. 5, November 2021

The Changing Use of Opioids in the U.S. from 2017 to Early 2020 Rachel Gifeisman1†‡, Aneesha Morris1†‡, Bianca Sanchez1†‡, Kenneth L McCall2, and Brian J Piper1,3 ¹Geisinger Commonwealth School of Medicine, Scranton PA 18509 ²Department of Pharmacy Practice, University of New England, Portland, ME 04103 3 Center for Pharmacy Innovation and Outcomes, Geisinger Precision Health Center, Forty Fort, PA 18704 † Doctor of Medicine Program ‡ Authors contributed equally Correspondence: rgifeisman@som.geisinger.edu

Abstract Background: The U.S. has been experiencing an opioid epidemic for over 3 decades, with previous data showing that prescription opioid use had been decreasing from 2006 to 2016. The COVID-19 pandemic has been a large disturbance in American health care, and public health scientists are curious if it will impact the opioid epidemic. This study provided a nationwide examination of opioid prescription trends in the 2 years leading up to the pandemic and in the first 2 quarters during it. Methods: Data were acquired from the U.S. Drug Enforcement Administration Automation of Reports and Consolidated Orders System (ARCOS) for 2017 through Q1 and Q2 of 2020. Analysis included amounts by mass of 10 opioids legally dispensed nationwide. Data were converted to morphine milligram equivalents per person, values for all 10 opioids were summed, and percent change between each of the consecutive years was calculated. Results: While total opioid prescription decreased from 2017 to 2018 (- 6.1%) and 2018 to 2019 (-2.1%), it increased from Q1 of 2019 to Q2 of 2020 (+2.4%). A paired t-test found a significant difference between the mean percent change from 2018 to 2019 and 2019 to 2020 (p<0.0005). Out of 29 states that had been decreasing opioid prescriptions in 2018 and 2019, 19 saw an increase in 2020. Among these 19 states, those with the largest reverse in prescription pattern included Arkansas, Montana, Oklahoma, and Tennessee. Conclusion: Overall, there appears to be an increase in opioid prescription from 2019 to 2020 following a decreasing trend in previous years. More research is needed to determine whether this was due to the COVID-19 pandemic.

Introduction It quickly became evident how the COVID-19 pandemic infiltrated every facet of our lives and demonstrated the glaring faults in our health care system and way of life, targeting mostly underrepresented groups of people. This includes those suffering from opioid use disorder (OUD), as treatment clinics had to change protocols or shut down entirely in order to adhere to the pandemic guidelines. In 2018, prescription opioid deaths in America fell for the first time in 25 years, with an overall decrease in opioid prescriptions from 2018 to 2019 (1). Using the Drug Enforcement Administration’s (DEA) Automated Reports and Consolidated Ordering System (ARCOS), our data showed increasing rates of opioid prescription, particularly in Kentucky, from 2017 to 2020. It became increasingly difficult to adjust to the pandemic, especially for those struggling with addiction to comply with

their treatment programs. In conjunction with the rise in opioid prescriptions, likely attributed to the ease the discomfort of patients with acute respiratory distress syndrome (ARDS) (2), there could be a new wave of demand for opioid prescriptions to manage the potentially chronic pains the disease may induce. Additionally, the COVID-19 pandemic may have disproportionately impacted those struggling with OUD as treatment clinics shut down, enabling patients to partake in opioid consumption in unsafe, coronavirus-unfriendly environments. In turn, this study aimed to track opioid prescription trends in the U.S. in order to better understand how the COVID-19 pandemic specifically affected the treatment and abuse of pharmacological pain management.

Methods Data sources National prescription quantities of 10 opioids were obtained from the Drug Enforcement Administration’s (DEA) Automated Reports and Consolidated Ordering System (ARCOS) for 2017 through the first and second quarter (June) of 2020. The 10 opioids included were buprenorphine, codeine, fentanyl, hydrocodone, hydromorphone, meperidine, morphine, oxycodone, oxymorphone, and tapentadol. Data were collected for all 50 states and Washington D.C.; data were unavailable for the U.S. territories of Guam, Puerto Rico, American Samoa, and the Virgin Islands. ARCOS is a comprehensive drug reporting system created as a result of the 1970 Controlled Substances Act. It reports on controlled substances in Schedules I to III distributed by hospitals, pharmacies, practitioners, and narcotic treatment programs (3). ARCOS data files consist of three different reports. Report two (“grams retail drug distribution by state within drug code”) was used for analysis. Population data for each of the 50 U.S. states was obtained from the annual American Community Survey and U.S. Census Bureau and used to normalize the state prescription data acquired from ARCOS (4). This study was deemed exempt from review by the Geisinger Institutional Review Board. Data analysis The total morphine milligram equivalent (MME) was calculated (in milligrams) for each of the 10 opioids per state per quarter of each year. To account for the relative potency of each agent, MME conversions were performed with drug-specific multipliers: buprenorphine 10, codeine 0.15, fentanyl 75, hydrocodone 1, hydromorphone 4, meperidine 0.1, morphine 1, oxycodone 1.5, oxymorphone 3, tapentadol 0.4. MME values

189

The Changing Use of Opioids in the U.S. from 2017 to Early 2020

Figure 1. Percent change in opioid prescription per capita for 10 opioids (buprenorphine, codeine, fentanyl, hydrocodone, hydromorphone, meperidine, morphine, oxycodone, oxymorphone, and tapentadol) as reported by the Drug Enforcement Administration from (A) 2017 to 2018 (B) 2018 to 2019 and (C) 2019 to mid-2020.

were divided by the populations of the respective states to obtain the MME per person. Then the values for all 10 drugs were added together to obtain the sum opioid prescription per state per person per quarter. The sum MME per capita data was used to calculate the percent change ([later year – earlier year]/earlier year) per state

190

between the first quarter of 2017 and the first quarter of 2018, the first quarter of 2018 and the first quarter of 2019, the first quarter of 2019 and the second quarter of 2020, and the first quarter of 2017 and the second quarter of 2020. Percent changes for all 50 states were averaged to obtain a mean percent change between each of the years. Paired T-tests were performed to compare the percent change between 2017 and

The Changing Use of Opioids in the U.S. from 2017 to Early 2020

Figure 2. (A) Percent change in opioid prescription per capita per 3 months from 2017 to mid-2020. (B) Comparison of total opioids in MME per state in 2017 vs mid-2020, with arrows pointing to states with the largest increase in prescription: Indiana, Kentucky, Maine, and West Virginia.

2018 to the percent change between 2018 and 2019 and the percent change between 2018 and 2019 to the percent change between 2019 to 2020. States with large reversal in prescription patterns, as well as states with the largest increase in total MME of opioids prescribed between 2017 and 2020, were identified. The fraction of buprenorphine was calculated by dividing the total amount of buprenorphine by the total amount of all summed opioids in MME prescribed in the U.S. per quarter. Excel and GraphPad Prism 9.1.0 were used for data analysis, graph generation, and statistical analysis.

Results From 2017 to 2018, the total amount of opioids prescribed decreased by 6.1% (Figure 1A) and from 2018 to 2017 opioid prescription decreased by -2.1% (Figure 1B). By comparison, the prescriptions from 2019 to mid-2020 increased by +2.4% (Figure 1C). A paired t-test found a significant difference between the mean percent change from 2018 to 2019 and 2019 to 2020 (p < 0.0005). Out of the 29 states that had decreasing opioid prescriptions in 2018 and 2019, 19 saw an increase in 2020. Among those 19 states, those with the largest

reverse in prescription pattern included Arkansas (from -7.79% to +3.75%), Montana (from -1.65% to +10.48%), Oklahoma (from -8.18% to +4.15%), and Tennessee (from -5.44% to +7.62%). From 2017 to mid-2020, the total percent change calculated was -0.05%. States with the largest increases in amounts of opioids prescribed were Maine (+0.31%), Kentucky (+0.15%), Indiana (+0.15), West Virginia (+0.12), and Washington (+0.12) (Figure 2a). Maine prescribed 317.26 MME in early 2017 and 416.33 MME in mid-2020. Kentucky increased from 383.68 MME prescribed in the first quarter of 2017 to 441.46 MME in the second quarter of 2020, West Virginia increased from 403.600 MME to 450.82 MME, and Washington from 191.28 MME to 213.67 MME (Figure 2b). Adding together all of the opioids prescribed in a given quarter and plotting the MME against time demonstrates a decrease in the amount prescribed throughout 2017, a plateau in 2018 and the first two quarters of 2019, and an increasing trend in the final two quarters of 2019 and the first two quarters of 2020. A line of best fit for the data from the first quarter of 2018 through the second quarter of 2020 was increasing with an R2 of 0.6244 (Figure 3A). Of the total opioids prescribed,

191

The Changing Use of Opioids in the U.S. from 2017 to Early 2020

Figure 3. Total U.S. opioids prescribed summed across states and drug type plotted (A) per quarter with a line of best fit for data from all four quarters of 2018 and 2019 and the first two quarters of 2020 and (B) as a comparison of totals with and without buprenorphine included in the calculation. (C) Percentage that buprenorphine is of total U.S. opioid prescription calculated and plotted against all four quarters of 2017, 2018, and 2019 and the first two quarters of 2020.

buprenorphine made up a large portion every year (Figure 3B). The portion increased every quarter from the first quarter of 2017 (40.3% buprenorphine) to the second quarter of 2020 (63.3% buprenorphine) (Figure 3C).

Discussion After a previous decline in opioid prescriptions across the U.S. from 2006 to 2016, our data showed an increase from 2017 to early 2020 with marked increase in Kentucky, Maine, Indiana, West Virginia, and Washington. The biggest reversal in opioid prescription rates were seen in Tennessee, Oklahoma, Montana, and Arkansas. In this study, we tracked a slowly increasing percent change in opioid prescriptions from 2017 to 2018, which included more states showing a positive percent change from 2018 to 2019 and an even steeper increase in the positive percent change among states from 2019 to early 2020.

192

Tennessee, which showed the greatest increase in percent change, did not see a positive percent change until 2019. However, Kentucky, which had the largest overall increase in opioid prescriptions, saw a positive percent change in 2017. Multiple conditions likely contributed to the increase in prescription rates seen across the U.S., with COVID-19 potentially exacerbating this upsurge. Additionally, the route of delivery was not tracked by ARCOS, therefore confounding the data, as it is unclear whether the increase was solely for take-home prescriptions or for hospital-based prescriptions. However, with the data showing a marked increase in most states from 2019 to early 2020, it can be inferred that the COVID-19 pandemic contributed to the change. As patients diagnosed with SARS-COV-2 conditions’ worsened, it was found that awake proning delayed the need for ventilation and better oxygenated patients than those in supine position (2). The

The Changing Use of Opioids in the U.S. from 2017 to Early 2020

need for awake proning, in addition to the cough and myalgia experienced by the patients, made positional treatment difficult and necessitated the prescription of IV or parenteral fentanyl, morphine, and hydromorphone to ease patients’ discomfort (2, 5). The need to begin opioid use among patients with COVID-19 could increase the population risk for addiction and aggravate the already worsening opioid pandemic. In Ontario, Canada, between January and September of 2020 there was a 108% increase in fentanyl use among OUD treatment patients, which was largely attributed to the disruption of treatment, medication diversion, and isolation during the pandemic (6) . This study demonstrated that the persistent increase of fentanyl exposure decreased the effectiveness of opioid agonist treatment (6). While OUD treatment in the U.S. is in clear need of expansion for greater availability of FDA-approved medications (7), it is possible that the pandemic-induced protocol changes to COVID-friendly practices in OUD treatment programs led to an increase in patient relapse. However, a case report of two OUD patients receiving buprenorphine/naloxone treatment via telemedicine in combination with street outreach during the pandemic provided alternative, lifesaving treatment options (8). In support of this, our data showed a steady increase in buprenorphine prescriptions over the years, with a noticeable uptick between Q4 2019 and Q1 2020. Further research is required to determine if one or both of these probable causes are implicated in the increase of opioid prescriptions from 2019-2020. As seen in emergency departments across the country in response to the pandemic, Kentucky reported a marked increase in opioid overdose emergency medical service (EMS) responses after declaring a state of emergency while also witnessing a decrease in all other EMS runs (9).This, in conjunction with our data showing that Kentucky had the largest increase in opioid prescriptions, implies the pandemic exacerbated the opioid epidemic. While more research needs to be done to determine the overlap of opioid overdoses to opioid prescription rates, it is clear there is a potential link between the coronavirus pandemic and the worsening opioid epidemic.

Disclosures B.J.P is supported by Fahs-Beck Fund for Research and Experimentation and supported by Pfizer and Health Resources Services Administration. The other authors have no relevant disclosures. The funders had no role in the design of the study; collection, analysis, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

References 1.

Harris M, Johnson S, Mackin S, Saitz R, Walley AY, Taylor JL. Low Barrier Tele-Buprenorphine in the Time of COVID-19: A Case Report. J Addict Med. 2020 May 20;10.1097/ ADM.0000000000000682.

Herndon KT, Claussen KS, Braithwaite JJ. A Novel Clinical Consideration to Conserve Parenteral Fentanyl During the COVID-19 Pandemic. Anesth Analg. 2020 Nov;131(5):1355–7.

ARCOS Retail Drug Summary Reports [Internet]. [cited 2021 Aug 1]. Available from: https://www.deadiversion. usdoj.gov/arcos/retail_drug_summary/index.html

Bureau UC. Data [Internet]. Census.gov. [cited 2021 Aug 1]. Available from: https://www.census.gov/data

Jones CM, Byrd DJ, Clarke TJ, Campbell TB, Ohuoha C, McCance-Katz EF. Characteristics and current clinical practices of opioid treatment programs in the United States. Drug Alcohol Depend. 2019 Dec 1;205:107616.

MacKinnon L, Socías ME, Bardwell G. COVID-19 and overdose prevention: Challenges and opportunities for clinical practice in housing settings. J Subst Abuse Treat. 2020 Dec;119:108153.

Manchikanti L, Vanaparthy R, Atluri S, Sachdeva H, Kaye AD, Hirsch JA. COVID-19 and the Opioid Epidemic: Two Public Health Emergencies That Intersect With Chronic Pain. Pain Ther. 2021 Jun;10(1):269–86.

Morin KA, Acharya S, Eibl JK, Marsh DC. Evidence of increased Fentanyl use during the COVID-19 pandemic among opioid agonist treatment patients in Ontario, Canada. Int J Drug Policy. 2021 Apr;90:103088.

Slavova S, Rock P, Bush HM, Quesinberry D, Walsh SL. Signal of increased opioid overdose during COVID-19 from emergency medical services data. Drug Alcohol Depend. 2020 Sep 1;214:108176.

Conclusion During a time when the volatile social and economic environment of the U.S. continues to impact the opioid epidemic, our results shed light on the rapidly developing state of opioid distribution. These findings indicate that the previous trend of decreasing opioid prescriptions nationwide may have reversed in the second half of 2019 and the first half of 2020. Buprenorphine accounted for the largest fraction of total opioids prescribed, increasing consistently over the last 3 years. Thus, the recent reversal in opioid prescription trends might be due in part to the increased use of buprenorphine for the treatment of overdose and addiction recovery. This study suggests that prescribers are taking a more active role in managing overdose and addiction by increasing prescriptions of buprenorphine.

193

Scholarly Research In Progress • Vol. 5, November 2021

Examining Health Literacy and Health Outcomes Among United States Immigrants and Non-Immigrants Jason L. McLeod1*‡, Muna M. Ahmed1*‡, Darin M. Chhing1*‡, Sami R. Hasan1*‡, Teresa N. James1*‡, and Yashoda T. Khatiwoda1*‡ ¹Geisinger Commonwealth School of Medicine, Scranton, PA 18509 *Master of Biomedical Sciences Program ‡ Authors contributed equally Correspondence: jmcleod@som.geisinger.edu

Abstract Background: Research on immigrant health is essential toward addressing the implications and needs of one of the largest growing demographics of the United States (U.S.) population. Immigrants face several challenges, including health literacy and access to health care. Many studies have shown that low health literacy is often associated with poor health outcomes. The goal of this study is to examine health literacy and health outcomes among immigrants who came to the U.S. between 1930 and 2011. Methods: A secondary analysis was completed using data from the Health Information National Trends Survey (HINTS) collected from 2011 to 2012. Several health literacy questions from the survey were used to score health literacy in our sample of 502 immigrants, composed of different races, who immigrated to the U.S. between 1930 and 2011. The health literacy questions included comprehension of the following topics: over the counter and prescription drugs, nutrition, and medical devices. To assess health outcomes in our samples, two questions about overall health and confidence in caring for oneself were used. Results: No statistically significant difference present between the health literacy score means of non-immigrant sample (n=1,646) and immigrant sample (n=203). A statistically significant difference was identified between the health literacy score means of Black/African American non-immigrants (n=274) and Black/African American immigrants (n=24) (t[25] = 2.493, p = 0.020). For both immigrant and non-immigrant samples, mean health literacy scores in association with results of health outcome assessments were relatively similar. Conclusion: Overall, high health literacy scores were noted among survey respondents. High health literacy was associated with high health rating and high health confidence among both immigrant and non-immigrant sample populations. NonImmigrant and immigrant health literacy score means were found to be relatively similar. Black/African American nonimmigrants had higher health literacy score means than Black/ African American immigrants. Additional analyses via collection of primary data and randomized surveys can emphasize and further elucidate the findings within this study. These strategies, combined with examination of poor health literacy, are anticipated to further reveal the influence of health literacy and socioeconomics on immigrant and non-immigrant populations.

Introduction Health literacy is the ability to comprehend and make decisions upon information regarding one’s overall health and/or the 194

health of close friends and family members (1). It is a critical factor that represents the degree to which individuals have the capacity to obtain, process, and comprehend basic health information and utilize the services needed to make appropriate and informed health decisions (2). Health literacy remains one of the major contributors to health disparities in the United States (U.S.), as approximately 80 million adults have limited or low health literacy (3). Factors that contribute to health literacy in the U.S. include, among others, a person’s level of education and their proficiency in the English language, along with one’s cultural background and environment (4). The U.S. is unique, in that it is composed of approximately 46 million immigrants, more than any other country in the world (5). The health of U.S. immigrants, and their access to health care services, features wide variability by aspects such as race, ethnicity, legal status, and citizenship (6, 7). Longstanding challenges faced by immigrant populations include difficulty navigating the U.S. health care system and accessing health information in their preferred language (8). In recent years, intensifying rhetoric discouraging immigrant-use of public health insurance programs, paired with aggressive enforcement of immigration practices and rescission of protections for those at risk for deportation, have worsened access and utilization of health care services among immigrants in the U.S. (9). Comprehensive understanding of the interrelation between health literacy and the health outcomes of immigrants to the U.S. may provide key insight into the development of strategies that seek to improve the health status of this growing demographic (10). Researchers have demonstrated findings indicating immigrants with both low health literacy and limited English proficiency as being more likely to exhibit poor health outcomes (2). Limited English proficiency can be a significant barrier between health care providers and immigrant patients, as these patients may experience difficulty in expressing their concerns, voicing questions, or following health care instructions (2, 11). Few studies have analyzed more recently published datasets from national samples to further elucidate the connection between health literacy and health outcomes among immigrant and non-immigrant individuals in the U.S. (12). The purpose of our investigation was to perform a secondary data analysis on participant responses from the Health Information National Trends Survey (HINTS) to examine health literacy and health outcomes in non-immigrant populations in comparison to those who immigrated to the U.S. Discussion of the results obtained through statistical techniques may suggest how health literacy interventions can be an effective tool in improving health comprehension, combating health disparities, and generating

Examining Health Literacy and Health Outcomes Among United States Immigrants and Non-Immigrants

health benefits through broad implementation across diverse immigrant populations.

Methods HINTS is sponsored by the Health Communication and Informatics Research Branch (HCIRB) of the Division of Cancer Control and Population Science. HINTS is leveraged to routinely accumulate nationwide data representative of the American public’s use of cancer-related information. The survey is also employed to monitor changes in the reception of health communications in individuals over the age of 18 by assessing usage of a variety of channels, including the internet, to obtain crucial health information for themselves and loved ones (13). Our study consisted of a secondary data analysis using data from the first cycle of the fourth edition of HINTS (HINTS 4 Cycle 1). The data to be analyzed was originally collected from October 25, 2011, through February 21, 2012. Only participants meeting specific sampling criteria were included in this study. Responses from survey question K6: “Were you born in the United States?” were used to determine inclusion of participant data. The resulting sample consisted of 3,310 participants who were born in the U.S. and 502 participants who were born abroad and immigrated to the U.S. between 1930 and 2011. There were no instances of U.S. citizens born abroad to report. All other respondents, including missing values, were excluded (n=147). The HINTS 4 Cycle 1 data was collected utilizing a survey that respondents received in the mail. In order to increase participation, the survey was translated into the Spanish language for households that were considered Spanishspeaking by the U.S. Census Bureau. A total of 12,385 addresses provided by Marketing Systems Group underwent a random sample method and were stratified such that census block groups with a population proportion of Hispanics or African Americans equaling or exceeding 40% were assigned to the high-minority stratum (n=6,730), while the remaining addresses were assigned to the low-minority stratum (n=5,475) (13). One-hundred and eighty addresses of the low minority group were from central Appalachia. The survey asked participants questions regarding topics such as: access to health information, health care opinions, current health status, health literacy, cancer, and demographics. Responses were in various formats, including dichotomous, non-Likert, and Likert scales. This information was highlighted in the HINTS 4 Cycle 1 methodology report (13). Our study employed secondary data analysis to better understand health literacy and health outcomes in our samples (Immigrant vs. Non-immigrant). Selected statements pertaining to health literacy and health outcomes from the HINTS 4 Cycle 1 survey instrument were identified for inclusion into this secondary analysis. A definition of health literacy from the National Library of Medicine, “The degree to which individuals have the capacity to obtain, process, and understand basic health information and services needed to make appropriate health decisions,” was adopted to guide the selection of relevant statements from the survey, as well as to help indicate the qualities of health literacy (14). Twelve statements from the HINTS survey

pertaining to health literacy were selected for inclusion in this study (Table 1), focusing primarily on over-the-counter drug consumption and nutrition. The responses “agree” or “disagree” were originally captured in the HINTS survey and were applied to operationalize a scale of health literacy for the secondary data analysis. The response “no opinion” has been excluded from the health literacy scale, as it does not add any benefit to the analysis. Any participants from either sample who had more than 16.67% missing data in the health literacy scale were excluded. Those participants with less than 16.67% missing data were included, and missing responses were imputed via the series mean method. Although the series mean method can potentially produce biased results, we mitigated this by imputing values only for participants with a very small amount of missing data. Two health outcome questions from the HINTS 4 Cycle 1 survey were selected for inclusion in this secondary analysis. Participants were asked, “In general, would you say your health is…,” (n=1,639) in addition to the question, “Overall, how confident are you about your ability to take good care of your health?” (n=1,638). Responses to these two questions were in Likert scale form in the original survey. Questions and responses are listed in Table 2. Any participants, from either sample, that had missing data from the two health outcomes questions were excluded (n=31). The Statistical Package for the Social Sciences (SPSS) software stored on a private hard drive both recorded and managed the resulting data to facilitate assessment of the health literacy and health outcomes in our samples. Personal identifiers were not provided in the HINTS 4 Cycle 1 dataset. An Institutional Review Board (IRB) committee determined that our research did not fall under human subjects research under the federal Common Rule, 45 CFR Part 46.102(d). Both descriptive and inferential statistics were employed to analyze the data and provide a statistical summary of immigrant (Figure 1) and non-immigrant (Figure 2) samples. The Kuder-Richardson Formula 20 test was performed to measure internal consistency of the 12-point health literacy scale. Independent samples t-tests were used to demonstrate if the mean health literacy score for each sample was different, and if this difference was statistically significant. In addition, independent samples t-tests were conducted to determine potential differences in mean health literacy scores when examining self-identified race between the non-immigrant and immigrant samples. Ultimately, the combined use of these tests helped characterize health literacy and health outcomes in our samples. Resulting data from the independent samples t-tests were reported on bar graphs.

Results Descriptive statistics were used to understand the generalization of health literacy scores between both immigrant (Figure 1) and non-immigrant samples (Figure 2). The mean health literacy score for the immigrant and non-immigrant sample was 8.82 (SD=3.44) and 9.02 (SD=3.22), respectively. The Kuder-Richardson Formula 20 test applied on the 12-question health literacy scale (Table 1) resulted in α=.877. This finding is an indicator of “good” internal consistency. All

195

Examining Health Literacy and Health Outcomes Among United States Immigrants and Non-Immigrants

Table 1. Health Literacy Measurement Statements Selected from the Health Information National Trends Survey (HINTS 4 Cycle 1)

Table 2. Health Outcomes Measurement Questions Selected from the Health Information National Trends Survey (HINTS 4 Cycle 1)

Figure 1. Frequency of health literacy scores from overall immigrant sample. questions used to determine the health literacy scale appeared to be meaningful in this test. Deletion of one or more questions from this analysis would cause a negligible change in alpha value. An independent sample t-test was performed to determine if the means of the health literacy scale were significantly different between our samples, non-immigrant (n=1646) vs. immigrant (n=203) (Figure 3). The results of this test suggested acceptance of the null hypothesis, indicating that there was not a statistically significant difference in the health literacy scale means between our sample of non-immigrants and immigrants (t[1847] = .864, p = 0.388). Further independent samples t-tests were performed to determine if there were differences in health literacy means when examining self-identified race between the non-immigrant 196

Figure 2. Frequency of health literacy scores from overall nonimmigrant sample. and immigrant samples (Figure 4). Notably, the data suggested a statistically significant difference in health literacy means (mean difference=1.97) between Black/African American nonimmigrants (n=274) and Black/African American immigrants (n=24), (t[25] = 2.493, p = 0.020). All other races were determined to be not statistically significant. The mean health literacy scores and standard deviations for immigrant and non-immigrant populations, in association with the health outcome question: “In general, would you say your health is…,” were calculated and reported in Table 3, and subsequently graphed in Figure 5. Likewise, the mean health literacy scores and standard deviations for non-immigrant and immigrant populations, in relation to the health outcome

Examining Health Literacy and Health Outcomes Among United States Immigrants and Non-Immigrants

Table 3. Health literacy score means for non-immigrant and immigrant populations in association with the health outcome question “In general, would you say your health is...”

Figure 5. Health literacy scores in relation to the health outcome question “In general, would you say your health is...” in non-immigrant and immigrant samples. Table 4. Health literacy score means for non-immigrant and immigrant populations in association with the health outcome question “Overall, how confident are you about your ability to take good care of your health?”

Figure 6. Health literacy scores in relation to the health outcome question “Overall, how confident are you about your ability to take good care of your health?” in non-immigrant and immigrant samples. Figure 3. Overall health literacy score means for non-immigrant (n=1646) and immigrant (n=203) samples (t[1847] = .864, p = 0.388).

Figure 4. Health literacy score means for Black/African American non-immigrants (n=274) and Black/African American immigrant (n=24) samples (t[25] = 2.493, p = 0.020).

question: “Overall, how confident are you about your ability to take good care of your health?” were calculated and reported in Table 4 and subsequently graphed in Figure 6.

Discussion The HINTS survey is considered a robust survey tool used to gauge the health and health communication in communities and has helped the National Cancer Institute gather information since 2003 (15, 16). Our selection of 12 questions from the HINTS Survey (Table 1) was determined to have good internal consistency for the purposes of this study. The list of frequently

asked questions included with the cover letter for each survey was intended to put the voluntary participants at ease with the survey methodology (17). Among other aspects to the study, participants were made aware of efforts to ensure their privacy, including the manner in which resulting data would be used. Stated under the Privacy Act, responses to survey questions could not be associated with participant names. In addition, the completed surveys would be securely stored throughout the duration of data analysis, and all versions destroyed upon study completion. It was also made known that the data was intended solely for the purposes of improving public health (17). In both the immigrant sample and non-immigrant sample, most participants appeared to have relatively high health literacy scores (refer to Figure 1 and Figure 2, respectively). This was expected, as participants in the voluntary survey may tend to have higher levels of health literacy (18). Although no relationship was identified for the health literacy score means between non-immigrants and immigrants, it is of value to note that the health literacy score mean was only slightly higher

197

Examining Health Literacy and Health Outcomes Among United States Immigrants and Non-Immigrants

for non-immigrants (9.02) than for immigrants (8.82), with the mean values differing only by 0.20. Observing a relatively similar mean health literacy score for immigrants and non-immigrants in our study may be attributed to the phenomenon known as the “Healthy Immigrant Effect.” In the Healthy Immigrant Effect, it is proposed that immigrants arrive with a health status above that of the country of destination. After 10 to 20 years of being in the destination country, immigrant health status resembles that of native residents (19). Economic stability, education, health care, one’s neighborhood, community, and environment are examples of determinants of health that may promote or hinder health literacy (20). These determinants of health are intimately influenced by the country of residence. Longer time spent in the U.S. can cause immigrants to have a similar health literacy score to non-immigrants, reflective of health care access, but also the amount of cultural assimilation (21). Though length of time in the country was not explored in our study, researchers may benefit from including time of residence as an element of future studies. African Americans have been found to have among the lowest health literacy in the U.S. (4). We report that a statistically significant difference was identified between the health literacy score means of Black/African American immigrants and Black/African American non-immigrants in our samples. In this study, the mean health literacy score of Black/African American non-immigrants (9.60) was higher than that of Black/ African American immigrants (7.60). The lower health literacy with Black immigrant populations can potentially be attributed to less exposure to the U.S. health care system and language barriers (22). There is diversity among Black immigrants to the U.S. that is also to be considered, with 54% and 34% being from the Caribbean and Africa respectively (23). The findings in our study are supportive of the knowledge that Black immigrants to the U.S. have been noted to face barriers to overall health, such as accent-based and race-based discrimination, physician inexperience with African and Caribbean cultural attitudes in regard to health, and not having health insurance or adequate access to language interpreters (23). Capturing the country of origin, along with the level of proficiency with the English language, could help provide more insight on the impact of these factors on low health literacy. Overall, for the immigrant and non-immigrant sample groups, the more health-literate the participants, the higher they rated the quality of their health, describing a possible correlation between health literacy and self-evaluated health quality between the two sample groups. It has been established that low health literacy is linked to poor rating of one’s health, poor self-management skills, and poor health outcomes (24). Our finding coincides with high health literacy correlating with high self-evaluated health quality. Health literacy is again affirmed as a fundamental and critical tool in construing health and health outcomes (24). Mean health literacy scores in association with results of the health outcome question, “In general, would you say your health is…” for both immigrant and non-immigrant samples were relatively comparable (Table 3). While non-immigrant health literacy ranged from 7.66 (“Poor” health rating) to 9.87 (“Excellent” health rating), immigrant health literacy had a more pronounced range of 5.39 (“Poor” health rating) to 10.62 (“Excellent” health rating). The non-immigrant mean health 198

literacy scores were greater than those of the immigrant group in all categories, with the exception of the mean health literacy score related to “Excellent” health status. The similar health literacy scores between both immigrant and non-immigrant samples for the health outcome question asking respondents to rate their general health may also potentially be ascribed to facets of the Healthy Immigrant Effect (19). As previously discussed, after some time in the country of destination, immigrants become similar in their health behaviors to their non-immigrant counterparts (19). Furthermore, since it is more common that healthier individuals would migrate, this period of assimilation is preceded by a period of time immediately following their migration in which immigrants have better health behaviors than non-immigrants (10). Barriers such as segregation, poor quality of health care, incompatibility with the new culture, lack of access to health care, insurance or familiarity with the health care system, and lack of English language fluency can lead to immigrants reporting lower health quality than non-immigrants (10). Conversely, the healthier individuals that migrate may tend to have higher education and greater economic stability than non-immigrants and may consequently report higher health ratings (10). This duality provides rationale behind the polarity in the health literacy scores among the immigrant sample, with immigrants having lower health literacy score means than non-immigrants for “Poor” health rating and higher health literacy score means than non-immigrants for “Excellent” health rating. Mean health literacy scores in association with results of the health outcome question “Overall, how confident are you about your ability to take good care of your health?” for both immigrant and non-immigrant samples were also relatively comparable (Table 4). Non-immigrant health literacy ranged from 7.05 (“A little confident”) to 9.88 (“Completely confident”), whereas immigrant health literacy exhibited a relatively similar range of 6.99 (“A little confident”) to 9.83 (“Completely confident”). The greatest difference in mean health literacy scores among the two samples was for the response “Somewhat confident,” with the non-immigrant sample exhibiting a mean (8.36) that was 0.75 higher than the immigrant sample mean (7.61). The interplay between health literacy scores and health confidence levels as seen in our data between immigrants and non-immigrants was as expected (25, 26, 27). Higher health literacy was associated with higher health confidence levels in both sample populations. In regard to the immigrant sample group, our findings may also be attributed to the Healthy Immigrant Effect (19). The results of both groups emphasized the important role that health literacy can play in contributing to health outcomes. Health confidence levels are impacted by the significance of health literacy, made evident as patients attempt to meet the stipulations of carrying out self-care, planning for care coordination, and navigating the intricacies of the health care system (25). Health literacy scores between immigrants and non-immigrants were found to be relatively similar within this study. In addition, it is important to note that for both immigrants and non-immigrants, the health literacy scores for all health confidence levels were above 50%, meaning that health literacy was relatively good for both sample populations. Aside from the Healthy Immigrant Effect explanation for our

Examining Health Literacy and Health Outcomes Among United States Immigrants and Non-Immigrants

findings among the immigrant sample, several factors may have contributed to health literacy in both sample groups, particularly socioeconomic factors (19, 10). Linking health literacy with additional elements assessed within the HINTS survey will require further exploration. Future analysis should investigate potential confounding factors that may falsely exhibit an apparent relationship between two variables or mask a true association. In addition, while HINTS does not include psychometric evaluation, this information would be relevant to collect with self-reported data in order to better address variation among the data sets, such as from small sample sizes. Furthermore, research might include examination of data for possible confounding effect or association between health literacy and the following factors, independently or in combination: length of time residing in a country, self-evaluation of overall health quality, level of confidence in caring for one’s personal health, age of immigrant and non-immigrant survey participants, mean age of sample, level of English fluency among immigrant and non-immigrant samples, education, and income.

Moreover, inspection of data and evaluation of the causal factors contributing to poor health literacy may be beneficial in determining an approach to minimizing health disparities. Primary data analysis may serve to provide broader flexibility to tailor analyses and critically examine socioeconomic factors, generating more conclusive findings.

Acknowledgments We would like to thank Brian J. Piper, PhD, Elizabeth Kuchinski, and Catherine Freeland for their assistance throughout our research. We would also like to thank Christine Nguyen for their participation in this project.

Disclosures There are no known conflicts of interest regarding this project.

References 1.

Sørensen K, Van den Broucke S, Fullam J, Doyle G, Pelikan J, Slonska Z, et. al. Health literacy and public health: A systematic review and integration of definitions and models. BMC Public Health. 2012;12-80.

Lee HY, Rhee TG, Kim NK, Ahluwalia JS. Health literacy as a social determinant of health in Asian American immigrants: Findings from a population-based survey in California. J. Gen. Intern. Med. 2015;30(8):1118-24.

Hickey KT, Creber RM, Reading M, Sciacca RR, Riga TC, Frulla AP, Casida JM. Low health literacy: Implications for managing cardiac patients in practice. J. Nurse Pract. 2018;43(8):49.

Rikard RV, Thompson MS, McKinney J, Beauchamp A. Examining health literacy disparities in the United States: A third look at the National Assessment of Adult Literacy (NAAL). BMC Public Health. 2016;16(1):1-1.

Khullar D, Chokshi DA. Challenges for immigrant health in the USA—the road to crisis. Lancet. 2019 25;393(10186):2168-74.

Engelman M, Ye LZ. The immigrant health differential in the context of racial and ethnic disparities: The case of diabetes. Adv Med Sociol. 2019;19:147-171.

Ku L. Waidmann T. How race/ethnicity, immigration status and language affect health insurance coverage, access to care and quality of care among the low-income population. Kaiser Family Foundation. 2003.

Johnson RM, Shepard L, Van Den Berg R, Ward-Waller C, Smith P, Weiss BD. A novel approach to improve health literacy in immigrant communities. HLRP: Health Lit. Res. Pract. 2019;4;3(3):S15-24.

Callaghan T, Washburn DJ, Nimmons K, Duchicela D, Gurram A, Burdine J. Immigrant health access in Texas: Policy, rhetoric, and fear in the Trump era. BMC Health Serv Res. 2019;19(1):1-8.

Conclusion Health literacy refers to the capability of individuals to achieve, process, and discern the essential information required to make applicable health decisions (2). Assessment of health literacy in our study was shown to be similar among both immigrant and non-immigrant U.S. samples. Through secondary data analysis, we obtained results characterizing how health literacy may impact the health outcomes of non-immigrant populations in juxtaposition to those who migrated to the U.S. Based on results of an independent samples t-test, the p value of 0.388 suggested acceptance of the null hypothesis, indicating no statistically significant differences in health literacy score means between immigrant and non-immigrant samples. Independent samples t-tests indicated a statistically significant difference (p = 0.020) in health literacy means between Black/ African American immigrants (7.60) and Black/African American non-immigrants (9.60). While inconclusive for other races, future studies may consider expanding upon health literacy and its influence on health outcomes among American Indian/ Alaska Native, Asian, Native Hawaiian/other Pacific Islander, and white respondents in subsequent HINTS reports. The results of this secondary data analysis demonstrated that most of the individuals who participated in the survey possess high health literacy. There were no definitive relationships between immigrants and non-immigrants. Irrespective of immigrant or non-immigrant designation, those with higher health literacy tended to rate their health status as higher. Mean health literacy scores in association with results of two health outcome questions for both immigrant and non-immigrant samples were relatively similar. Healthy Immigrant Effect is a phenomenon that helps describe how migrants to the U.S. tend to take on a health status that is similar to non-immigrants after residing in the country for 10 to 20 years (16). Health literacy was generally good overall for both immigrants and non-immigrants, indicated by the confidence intervals of 50% and above. Further research into the health of U.S. immigrants, who are deeply interwoven across a dynamic social and political landscape, is imperative for primary data analysis. The combination of primary data collection and randomized surveys should be pursued for more discrete results.

10. Prins E, Monnat S. Examining associations between selfrated health and proficiency in literacy and numeracy among immigrants and US-born adults: Evidence from the Program for the International Assessment of Adult Competencies (PIAAC). PloS One. 2015;10(7):e0130257. 199

Examining Health Literacy and Health Outcomes Among United States Immigrants and Non-Immigrants

11. Fields A, Abraham M, Gaughan J, Haines C, Hoehn KS. Language matters: Race, trust, and outcomes in the pediatric emergency department. Pediatr. Emerg. Care 2016;1;32(4):222-6. 12. Mantwill S, Monestel-Umaña S, Schulz PJ. The relationship between health literacy and health disparities: A systematic review. PLoS One. 2015;10(12):e0145455. 13. National Cancer Institute at the National Institute of Health. HINTS 4, Cycle 1- National Trends Survey [Internet]. 2012 [revised 1 Sept. 2020; Cited 2020 Aug 6] Available from: https://hints.cancer.gov/default.aspx 14. Somers SA, Mahadevan R. Health literacy implications of the affordable care act. Center for Health Care Strategies, Incorporated; 2010. 15. National Cancer Institute [Internet]. Rockville (Maryland): [cited 2021 Feb 28]. Available from: https://hints.cancer. gov/docs/Hints_Factsheet.pdf 16. National Cancer Institute [Internet]. Rockville (Maryland): [cited 2021 Feb 28]. Available from: https://hints.cancer. gov/about-hints/learn-more-about-hints.aspx 17. National Cancer Institute. Cyle 1 Methodology Report: June 2012 [Internet]. Rockville (Maryland): cited 2021 Jun 28]. Available from: https://hints.cancer.gov/docs/ methodologyreports/HINTS_4_Cycle_1_Methodology_ Report.pdf 18. Kripalani S, Heerman WJ, Patel NJ, Jackson N, Goggins K, Rothman RL, Yeh VM, Wallston KA, Smoot DT, Wilkins CH. Association of health literacy and numeracy with interest in research participation. J. Intern. Med. 2019;34(4):544-51. 19. Feinberg I, O'Connor MH, Owen-Smith A, Ogrodnick MM, Rothenberg R. The Relationship Between Refugee Health Status and Language, Literacy, and Time Spent in the United States. HLRP: Health Lit. Res. Pract. 2020;14;4(4):e230-6. 20. HealthyPeople.gov. Office of Disease Prevention and Health Promotion. United States Department of Health and Human Services [Internet]. (Washington D.C.): [cited 2021 Feb 27]. Available from: https://www.healthypeople. gov/2020/topics-objectives/topic/social-determinantshealth/interventions-resources/health-literacy 21. Antecol H, Bedard K. Unhealthy assimilation: why do immigrants converge to American health status levels?. Demography. 2006;43(2):337-60. 22. Ahad FB, Zick CD, Simonsen SE, Mukundente V, Davis FA, Digre K. Assessing the likelihood of having a regular health care provider among African American and African Immigrant women. Ethn Dis. 2019;29(2):253. 23. Wafula EG, Snipes SA. Barriers to healthcare access faced by black immigrants in the US: Theoretical considerations and recommendations. J. Immigr. Minor. Health 2014;16(4):689-98. 24. Baskaradoss JK. Relationship between oral health literacy and oral health status. BMC Oral Health. 2018;18(1):1-6.

200

25. Hersh L, Salzman B, Snyderman D. Health Literacy in Primary Care Practice. Am Fam Physician. 2015;92(2):11824. 26. Levy H, Janke A. Health literacy and access to care. J Health Commun. 2016;21 Suppl 1:43–50. 27. Sørensen K, Pelikan JM, Röthlin F, Ganahl K, Slonska Z, Doyle G, et al. Health literacy in Europe: comparative results of the European health literacy survey (HLS-EU). Eur J Public Health. 2015;25(6):1053–8.

Scholarly Research In Progress • Vol. 5, November 2021

Electroconvulsive Therapy Uses and Its Ability to Induce Neurogenesis: A Literature Review Kylar J. Harvey1*, Gwyneth J. Harris1*, Alexander I. Greenstone1*, Catherine L. Falzone1*, and Sami R. Hasan1* ¹Geisinger Commonwealth School of Medicine, Scranton, PA 18509 *Master of Biomedical Sciences Program Correspondence: shasan@som.geisinger.edu

Abstract Electroconvulsive therapy (ECT) has been used historically to treat depression, seizures, and many other psychiatric symptoms. With the investigation of the mechanisms of action of ECT, physiological changes were noted in patients treated with ECT where neurogenesis was consistently linked to reductions in symptoms of many psychiatric illnesses. Neurogenesis is a prolific area of clinical neuroscience research with a wealth of current research and possibilities for future research. Also linked to the therapeutic benefits of ECT are neuronal activation and endothelial proliferation in the midhypothalamic nuclei. The hypothalamus is a key regulator of endocrine function, so further research investigating the implications of ECT-induced mid-hypothalamic changes on endocrine functions of patients is warranted and may illuminate some potential therapeutic pathways for ECT. In addition, these hypothalamic changes do not represent the sum of all neurological changes following ECT, so interactions between the mid-hypothalamic changes and other neuro-anatomical changes after ECT should be considered to develop a more complete model of ECT’s therapeutic action in treating depression and epilepsy. ECT has promise for diseases of aging as, with aging, drug metabolism is altered, making pharmacotherapeutic intervention less consistent within this population. With an understanding of ECT comes more confident use of the treatment in the clinical setting and could help physicians decide when ECT is appropriate as the first line of treatment.

Introduction The generation of new cells from stem cells is a lifelong process sustained in a multitude of cells. In the case of neurons, they arise from a resident population of progenitors throughout adulthood via neurogenesis, proliferation, and differentiation of adult stem cells (1, 2). Neurogenesis is heavily studied in rodents due to ethical concerns and its conservation in mammals. There are three classes of neural stem cells and progenitor cells in rodent nervous systems: neuroepithelial cells, radial glial cells, and basal progenitors. With each cell type, there is a respective type of division-symmetric, proliferative division; asymmetric, neurogenic division; and symmetric, neurogenic division (3). Symmetric proliferation is defined as producing two daughter cells of the same fate, while asymmetric proliferation generates a single daughter cell that is identical to the mother cell and a second nonidentical cell. Asymmetric divisions, interestingly, may continue to replenish the stem cell pool but lack the ability to regulate adult neurogenesis (4). Neuroepithelial cells are considered true

stem cells due to their ability to differentiate and self-renew, while the radial glial cell and basal progenitors are restricted to a single cell fate and are unable to self-renew (3). The type of division and proliferative or differentiation direction is determined by many factors relating to specific epithelial cell characteristics: the apical-basal polarity and cell cycle length (1, 5). Neural progenitors may be influenced by their microenvironments, mainly through the influence of distal and proximal neurons, making them subject to extrinsic regulation (3). Gamma-aminobutyric acid (GABA) signaling is essential for both early neuron depolarization and mature neuron hyperpolarization; the latter is crucial for initiating proper reception of glutamatergic inputs (6). The subgranular zone (SGZ) progenitors, namely the dentate granule cells, receive excitatory glutamatergic and inhibitory GABAergic signals from the local interneurons while also being influenced by different neurotransmitters from distal brain areas (7). Cell fate determination is influenced by neurotransmitters, specifically GABA, glutamate, and nitric oxide (NO), prior to neurogenesis. Recent observations have indicated that GABA and glutamate also play a role in the control of neurogenesis (6, 8, 9). Neurogenesis can occur in the hippocampus and olfactory bulb of adult mammals, whereas it was previously thought to not take place outside of embryonic and early postnatal periods (1, 10). The process occurs in the SGZ of the dentate gyrus within the hippocampus, where it is relevant for some forms of learning and memory, and the subventricular zone (SVZ) of the lateral ventricles where ependymal cells are suspected of being the resident adult NSCs (5, 10, 11). Just like the rest of the body, neurogenesis is influenced by aging through a loss of homeostasis in stem cells, including telomere shortening, DNA damage, and cell cycle interruptions, to name a few pathways (4, 12). Neurogenesis is heavily influenced by both positive and negative factors, including epigenetic components of hippocampal neurogenesis that also need to be considered (13). Positive factors may include exercise and environmental enrichment (mating, diverse foods) while negative factors may be generalized as acute and chronic stress (14–19). Hippocampal neurogenesis can be enhanced by hormones, growth factors, drugs, physical exercise, and neurotransmitters and suppressed by aging, glucocorticoids, and stimuli that activate the pituitary/adrenal axis (1, 20, 21). While many models convey that neurogenesis may be considered a cellreplacement method, newer models show appreciation for the neuroplasticity that ensues from the continuous addition of new neurons, also emphasizing the structural plasticity contributions that result (21). In this review, we examined the effects between neurogenesis and electroconvulsive therapy (ECT).

201

Electroconvulsive Therapy Uses and Its Ability to Induce Neurogenesis: A Literature Review

Methods An examination of the literature was conducted in search of ECT’s uses and ability to generate neurogenesis. References were obtained from Google Scholar, Academic Search Ultimate, EBSCOhost, APA PsycArticles, and ScienceDirect. References were considered acceptable and reliable for inclusion as they were drawn from reputable databases and include peer-reviewed research articles. Articles' publish dates range from 1985 to 2020. In addition, National Center for Biotechnology Information StatPearls online textbook pages were referenced, copyright 2021. Keywords screened for included: electroconvulsive therapy, ECT, neurogenesis, seizure, depression, hypothalamic-pituitary axis, neuroendocrine disorder. Peer-reviewed articles cited are dated from 1985 to 2020 and include research articles as well as StatPearls online textbooks.

Discussion The link between neurogenesis and ECT ECT is psychiatry’s oldest behavioral psychiatric treatment, having been around since the 16th century, though its first documented use in a controlled clinical setting was in 1938 for general psychosis (22–25). ECT is a procedure administered with electrodes that are placed on the head to stimulate a specific portion of the brain using electric current applied through the electrodes via sine-wave current which has alternating frequencies (23, 24). The first medically documented use of ECT for a specific psychiatric illness was in 1941 to treat schizophrenia, depression, and seizures along with the use for general psychosis (23, 25). Today, ECT is administered to over 1,000,000 patients around the world each year for schizophrenia, depression, seizure disorders, and other psychotic symptoms (23). The use of ECT is not a widely adopted approach due to concerns of the possible side effects of memory impairment and possible brain damage (26). These concerns are noted due to mixed results from ECT experiments in rodent models that were conducted without the proper safeguards currently upheld today on human subjects (26). The concerns regarding memory impairment and the potential for brain damage are negated by the American Psychiatric Association (APA) 2001 guidelines along with the Royal College of Psychiatrists’ guidance for the use of a brief pulse-wave current instead of a continuous current (23). This brief pulse stops the damaging potential that is associated with the overproduction of glutamate caused by prolonged applications of constant electric current (22, 23). How ECT helps with the management of psychiatric diseases is suggested to be the mechanism of neurogenesis within the dentate gyrus of the hippocampus and the SVZ, as well as an increase in neuroplasticity (25, 27–35). Many mental health disorders are noted to exhibit decreased volume of the hippocampus and the SVZ, indicating there is impaired neurogenesis in these two regions (28, 31, 32, 35–38). ECT can modify monoamine transporters, promote increased neurogenesis, increase neuroplasticity, and aid in regulating the hypothalamic-pituitary-adrenal (HPA) axis (34). Neuroplasticity and neurogenesis were shown to be positively correlated in many scientific experiments and reviews suggesting some common factors that ECT may modulate neuroendocrine 202

responses including, angiogenesis, epigenetics, ATP release, immune response, increased number of immature neurons, and many other factors (27–30, 39). A large focus has been put on the modulating effects of ECT on neuroendocrine responses, as they have been shown to increase efficacy of medication and reduce need for medication in patients suffering from epilepsy and depression. (27, 28, 31, 33, 40–42). How ECT can cause neurogenesis to treat seizures For the past decade, ECT has been looked at as a potential treatment for reducing the effects of epilepsy (43–47). Long exposure epilepsy is shown to disrupt hippocampal granule cells and their production. Initially, there is an increase in the rate of granule cell neurogenesis of the hippocampus in the early stages of epilepsy development (43, 44). The onset of chronic seizures is thought to be caused by too little neurogenesis resulting in prolonged recurring seizures (43, 44). The contribution of ECT to amplifying neurogenesis is the targeted factor that is deemed critical for the reduction in the severity and duration of chronic seizures. The area of the brain that is of neurogenesis interest is the hippocampal dentate granule cells, as this area is seen to be diminished in epilepsy (45). This, in return, strongly affects stem cell-associated plasticity in the dentate gyrus (46). The importance of dentate granule cells is their involvement in the regulation of relayed information to the hippocampus (47). Chronic temporal lobe epilepsy is associated with neurodegeneration and inhibition in the hippocampal regions (44). As observed in animal models and rodents, an initial response to the development of epilepsy is an increase in neurogenesis, but because of long chronic exposures, a decrease in neurogenesis is observed (48, 49). Furthermore, persistent chronic seizures were shown to deplete progenitor cells, leading to reduced neurogenesis (48). As ECT is shown to induce a brief seizure, it demonstrated a correlation between how short, induced seizures may increase neurogenesis and ultimately help treat epilepsy (50, 51). The mechanism of action for ECT is still not yet understood, but according to a systematic review and meta-analysis, animal models appeared to demonstrate that ECT induces neuroplasticity, which ultimately increases the hippocampal volume (52). To increase the volume and neurogenesis of the hippocampal region as well as producing an anticonvulsant effect, ECT was administered to bypass the need for any medications (53, 54). As seen in many patients, chronic seizures and depression may sometimes be triggering risk factors for one another (55). As seen in specific cases, such as treatment-resistant depression and epilepsy, the reduction of these occurrence frequencies may be tackled by the utilization of ECT as it is seen to help increase the neurogenesis in many hippocampal subfields and the amygdala (55, 56). These areas are correlated with the occurrence of depression and anxiety (55). Some of the anticonvulsant effects in the improvement of these regions include decreasing seizure duration, increasing the seizure threshold, and decreasing spontaneous seizure frequencies after multiple ECT sessions (53, 54, 57–59). In some cases, ECT was used to reduce the frequency of seizures in patients with seizure disorders who did not optimally respond to antiepileptic drugs (58). ECT as a form of therapy may be favorable for those who are in the refractory status epilepticus (SE). Refractory SE patients are shown to be partially resistant to treatment medications for epilepsy and therefore need an

Electroconvulsive Therapy Uses and Its Ability to Induce Neurogenesis: A Literature Review

alternative form of treatment. SE patients experience inordinate recurring and prolonged seizures that affect their quality of life. Although there is a lack of clinical studies on ECT for refractory SE patients, there were 11 SE patients who reported being treated with ECT. These patients were assessed for their efficacy and were shown to succeed in decreasing epileptic episodes and, in some cases, a cessation in seizures (59). Long exposure high-intensity ECT provided temporary cessation of refractory SE for a few months from the point of the last ECT session (60). The main take from these findings is that ECT treatment has been shown to alter the neurogenesis frequency in the hippocampal regions (61). Graph theory is a quantitative analysis of complex networks to study the brain network organization, represented in a graph-like manner (61, 62). This theory in hand with the utilization of functional and structural MRI can be used to observe the neuroimaging of ECT test subjects to confirm the pre-ETC and post-ETC networking differences. To obtain a vast beneficial database on treating epilepsy with ECT, there needs to be more targeted clinical studies addressing its potential therapeutic improvements and effects.

Figure 1. This figure illustrates a very limited set of examples of interaction within the thalamus that influence downstream effects via the hypophyseal portal system.

Neurogenesis generated by ECT in the treatment of depression ECT has been a major topic of interest in the therapeutic treatment of depression. The key issue is to what degree structural changes can be viewed as trait-dependent, indicative of susceptibility to depression, or state-dependent, and therefore an important therapeutic target. Preclinical experiments have demonstrated that the initiation of pathways contributing to improved hippocampal plasticity is part of the anti-depressive therapeutic mechanism of action, indicating a potential state-dependent structural-level counteracting mechanism (63). A further area of interest is the placement of the electrodes (25) and the possible combination of pharmacotherapy in neurogenesis (25, 64). High-dose right unilateral (RUL) ECT was found to have higher efficacy than bilateral ECT when above the threshold level (65, 66). Regarding the duration of a pulse, ultra-brief RUL ECT had a greater efficacy on remission in those who experience psychotic features (67). ECT has been found to be a more effective measure over pharmacological interventions (68), because it acts directly on the central nervous system (69). ECT is usually not used as the first line of treatment, but one study suggests that if ECT was used earlier in those with long-term chronic depression, there could be an increase in remission (70). Despite these suggestions, a randomized controlled trial study found no significant difference between ECT and pharmacotherapy (71).

Late-life depression therapies have a different effect because of the biological changes associated with aging (72). ECT may be superior to pharmacological interventions for depression in the elderly, given that drugs experience decreased absorption, increased amount of distribution, decreased metabolism, and diminished excretion with aging (72). Also, patients can undergo age-related increases in drug susceptibility in later life. Elderly patients may have pharmacodynamic modifications that render them more sensitive to anticholinergic and noradrenergic side effects owing to age-related receptor vulnerability and age-related alterations in cholinergic and monoaminergic neurotransmission (72). However, such pharmacotherapies have been shown to act on amplifying neural progenitors (25) and provide protection against volume reductions caused by diseases (69). Historically there has been a lack of evidence that ECT may produce higher rates of remission than drug therapy (73, 74). More recently, several studies provide evidence suggesting an increase in the volume of gray matter with ECT among elderly patients and did not cause a significant difference in rates of remission (74, 75). Contrary to this, one study of elderly patients found no difference in brain volume between those treated with RUL and those treated with bitemporal ECT (76). Though high-dosage RUL has been found to be just as effective as bilateral ECT and is associated with a decrease in long-term amnesia (73), additional analysis is merited to define appropriate formula-based dosages for RUL ECT in elderly patients (77). While most studies have found a correlation between neurogenesis and remission of depressive episodes (69, 78, 79), there is a strong indication of relapse after 6 months (63, 76, 80). A better understanding of how ECT can achieve neurogenesis to treat depression may lie in optimized neuroplasticity in the HPA axis (81).

203

Electroconvulsive Therapy Uses and Its Ability to Induce Neurogenesis: A Literature Review

ECT and neuroendocrine responses ECT is associated with the activation of neurons and proliferation of endothelial cells (82) in three mid-hypothalamic nuclei: the paraventricular nucleus (PVN), supraoptic nucleus (SON), and ventromedial nucleus (VMN) (83). The PVN controls autonomic function via numerous inputs and projections (84), and is influenced by oxytocin, vasopressin, and corticotropinreleasing hormone (CRH) (84). The SON synthesizes oxytocin and vasopressin and communicates with the medial preoptic nucleus of the thalamus (85). The SON also communicates with the PVN (86). Both the PVN and the VMN are responsive to oxytocin and orexin (87–89). The VMN generates aggression (90), regulates satiety (91), and mediates sexual behavior (92), and secretes neuropeptide Y (93), which is known to activate CRH neurons in the PVN (94). Given the extensive crosstalk between these three nuclei, their communication is essential in coordinating a joint output signal. This signal is relayed via "hypophyseal portal vessels'' to the anterior pituitary (95), which releases adrenocorticotropic hormone, growth hormone (GH), prolactin, thyroid-stimulating hormone, folliclestimulating hormone, and luteinizing hormone into circulation (96) (Figure 1). These hormones are indicators of anterior pituitary and hypothalamic function and have been used to study the endocrine consequences of ECT-mediated mid-hypothalamic cell proliferation. Prolactin and GH levels increase with repeated ECT (97). ECT recipients given naloxone, a putative opioid receptor antagonist (98), had no significantly different response in prolactin or GH, suggesting an opiate-independent mechanism for ECT-stimulated hormone release (97). Thyroid hormone, known to improve mood, memory, and executive function (99), may be decreased with ECT, as low thyroid hormone patients have poor recall accuracy after ECT, while ECT patients receiving adjunctive triiodothyronine have statistically significant improvements to memory (100). These endocrine responses to ECT may be partly explained by neuronal activation and epithelial proliferation in the PVN, SON, and VMN. However, neurological effects of ECT — for instance, activation of the mesocorticolimbic dopamine system and frontotemporal glutamate-GABA processes (101) — may also contribute to ECT response. Characterization of ECT’s neuroendocrine effects requires further research into both neurological and endocrine impacts of the treatment (102). More complete knowledge of these mechanisms may in turn improve how ECT is used to treat seizure patients, depressed patients, and other diseases in the future.

Conclusion ECT has been used since 1941 to treat depression, seizures, and many other psychiatric symptoms, and the treatment has improved with time. As ECT has been explored, it was discovered that neurogenesis was repeatedly linked to decreases in symptomologies of many psychiatric illnesses. Neurogenesis is a prolific area of clinical neuroscience that applies to current and future studies of many psychiatric pathologies. A key area of ECT research is neuronal activation and endothelial proliferation in the mid-hypothalamic nuclei. Future research could characterize the implications of these mid-hypothalamic changes for endocrine balance. Additionally,

204

interactions between the mid-hypothalamic changes and other neuro-anatomical changes after ECT should be considered to develop a more complete model of ECT’s therapeutic action in treating depression and epilepsy. Due to the propensity of ECT to circumvent pharmacodynamics, especially in older patients, future clinical interventions using ECT should be explored as the first line of treatment. In efforts to improve ECTgenerated remission, studies could investigate how to sustain neurogenesis. This sustained neurogenesis may lead to new insights into minimizing relapse.

Disclosures The authors have nothing to disclose.

References 1.

Götz M, Huttner WB. The cell biology of neurogenesis. Nat Rev Mol Cell Biol. 2005;6(10):777–88.

Zhao C, Deng W, Gage FH. Mechanisms and functional implications of adult neurogenesis. Cell. 2008;132(4):645– 60.

Sahay A, Hen R. Adult hippocampal neurogenesis in depression. Nat Neurosci. 2007;10(9):1110–5.

Eisch AJ, Petrik D. Depression and hippocampal neurogenesis: A road to remission? Science. 2012;338(6103):72–5.

Wang H, Warner-Schmidt J, Varela S, Enikolopov G, Greengard P, Flajolet M. Norbin ablation results in defective adult hippocampal neurogenesis and depressive-like behavior in mice. Proc Natl Acad Sci. 2015;112(31):9745–50.

van Wijngaarden P, Franklin RJM. Ageing stem and progenitor cells: Implications for rejuvenation of the central nervous system. Development. 2013;140(12):2562–75.

Berg DA, Belnoue L, Song H, Simon A. Neurotransmittermediated control of neurogenesis in the adult vertebrate brain. Development. 2013;140(12):2548–61.

Duman RS, Aghajanian GK. Synaptic dysfunction in depression: Potential therapeutic targets. Science. 2012;338(6103):68–72.

Obernier K, Alvarez-Buylla A. Neural stem cells: Origin, heterogeneity and regulation in the adult mammalian brain. Development. 2019;146(4):dev156059.

10. Jacobs B. Adult brain neurogenesis and depression. Brain Behav Immun. 2002;16(5):602–9. 11. Seib DRM, Martin-Villalba A. Neurogenesis in the normal ageing hippocampus: A mini-review. Gerontology. 2015;61(4):327–35. 12. Drapeau E, Nora Abrous D. Stem cell review series: Role of neurogenesis in age-related memory disorders. Aging Cell. 2008;7(4):569–89. 13. Kumar A, Pareek V, Faiq MA, Ghosh SK, Kumari C. Adult neurogenesis in humans: A review of basic concepts, history, current research, and clinical implications. Innov Clin Neurosci. 2019;16(5-6):30–7.

Electroconvulsive Therapy Uses and Its Ability to Induce Neurogenesis: A Literature Review

14. Tang SW, Helmeste D, Leonard B. Is neurogenesis relevant in depression and in the mechanism of antidepressant drug action? A critical review. World J Biol Psychiatry. 2012;13(6):402–12.

28. Giacobbe J, Pariante CM, Borsini A. The innate immune system and neurogenesis as modulating mechanisms of electroconvulsive therapy in pre-clinical studies. J Psychopharmacol. 2020;34(10):1086–97.

15. Akers KG, Chérasse Y, Fujita Y, Srinivasan S, Sakurai T, Sakaguchi M. Concise review: Regulatory influence of sleep and epigenetics on adult hippocampal neurogenesis and cognitive and emotional function: Sleep, epigenetics, and adult neurogenesis. Stem Cells. 2018;36:969–76.

29. Ueno M, Sugimoto M, Ohtsubo K, Sakai N, Endo A, Shikano K, et al. The effect of electroconvulsive seizure on survival, neuronal differentiation, and expression of the maturation marker in the adult mouse hippocampus. J Neurochem. 2019;149(4):488–98.

16. Christian KM, Song H, Ming G. Functions and dysfunctions of adult hippocampal neurogenesis. Annu Rev Neurosci. 2014;37(1):243–62.

30. Ma DK, Marchetto MC, Guo JU, Ming G, Gage FH, Song H. Epigenetic choreographers of neurogenesis in the adult mammalian brain. Nat Neurosci. 2010;13(11):1338–44.

17. Lieberwirth C, Pan Y, Liu Y, Zhang Z, Wang Z. Hippocampal adult neurogenesis: Its regulation and potential role in spatial learning and memory. Brain Res. 2016;1644:127– 40.

31. Perera TD, Dwork AJ, Keegan KA, Thirumangalakudi L, Lipira CM, Joyce N, et al. Necessity of hippocampal neurogenesis for the therapeutic action of antidepressants in adult nonhuman primates. PLoS One. 2011;6(4):e17600.

18. Spurny B, Seiger R, Moser P, Vanicek T, Reed MB, Heckova E, et al. Hippocampal GABA levels correlate with retrieval performance in an associative learning paradigm. Neuroimage. 2020;204(116244):116244.

32. Malberg JE. Implications of adult hippocampal neurogenesis in antidepressant action. J Psychiatry Neurosci. 2004;29(3):196–205.

19. Elder GA, De Gasperi R, Gama Sosa MA. Research update: Neurogenesis in adult brain and neuropsychiatric disorders. Mt Sinai J Med. 2006;73(7):931–40. 20. Drew M, Hen R. Adult hippocampal neurogenesis as target for the treatment of depression. CNS Neurol Disord Drug Targets. 2007;6(3):205–18. 21. Bergmann O, Spalding KL, Frisén J. Adult neurogenesis in humans. Cold Spring Harb Perspect Biol. 2015;7(7):a018994. 22. Sienaert P. What we have learned about electroconvulsive therapy and its relevance for the practising psychiatrist. Can J Psychiatry. 2011;56(1):5–12. 23. Leiknes KA, Schweder LJ, Høie B. Contemporary use and practice of electroconvulsive therapy worldwide. Brain Behav. 2012;2(3):283–344. 24. Koga K, Furue H, Rashid MH, Takaki A, Katafuchi T, Yoshimura M. Selective activation of primary afferent fibers evaluated by sine-wave electrical stimulation. Mol Pain. 2005;1:13. 25. Rotheneichner P, Lange S, O’Sullivan A, Marschallinger J, Zaunmair P, Geretsegger C, et al. Hippocampal neurogenesis and antidepressive therapy: Shocking relations. Neural Plast. 2014;2014:723915. 26. Dwork AJ, Christensen JR, Larsen KB, Scalia J, Underwood MD, Arango V, et al. Unaltered neuronal and glial counts in animal models of magnetic seizure therapy and electroconvulsive therapy. Neuroscience. 2009;164(4):1557–64. 27. Nuninga JO, Mandl RCW, Froeling M, Siero JCW, Somers M, Boks MP, et al. Vasogenic edema versus neuroplasticity as neural correlates of hippocampal volume increase following electroconvulsive therapy. Brain Stimulat. 2020;13(4):1080–6.

33. Jun H, Mohammed Qasim Hussaini S, Rigby MJ, Jang M-H. Functional role of adult hippocampal neurogenesis as a therapeutic strategy for mental disorders. Neural Plast. 2012;2012:854285. 34. Xiao W, Zhan Q, Ye F, Tang X, Li J, Dong H, et al. Elevated serum vascular endothelial growth factor in treatmentresistant schizophrenia treated with electroconvulsive therapy: Positive association with therapeutic effects. World J Biol Psychiatry. 2019;20(2):150–8. 35. Wang J, Wei Q, Bai T, Zhou X, Sun H, Becker B, et al. Electroconvulsive therapy selectively enhanced feedforward connectivity from fusiform face area to amygdala in major depressive disorder. Soc Cogn Affect Neurosci. 2017;12(12):1983–92. 36. Kranaster L, Blennow K, Zetterberg H, Sartorius A. Electroconvulsive therapy does not alter the synaptic protein neurogranin in the cerebrospinal fluid of patients with major depression. J Neural Transm. 2017;124(12):1641–5. 37. Svensson M, Grahm M, Ekstrand J, Höglund P, Johansson M, Tingström A. Effect of electroconvulsive seizures on cognitive flexibility: ECS and cognitive flexibility. Hippocampus. 2016;26(7):899–910. 38. Singh A, Kar SK. How Electroconvulsive Therapy works?: Understanding the neurobiological mechanisms. Clin Psychopharmacol Neurosci. 2017;15(3):210–21. 39. Gosnell SN, Meyer MJ, Jennings C, Ramirez D, Schmidt J, Oldham J, et al. Hippocampal volume in psychiatric diagnoses: Should psychiatry biomarker research account for comorbidities? Chronic Stress. 2020;4:2470547020906799. 40. Maruyama S, Boku S, Okazaki S, Kikuyama H, Mizoguchi Y, Monji A, et al. ATP and repetitive electric stimulation increases leukemia inhibitory factor expression in astrocytes: A potential role for astrocytes in the action mechanism of electroconvulsive therapy. Psychiatry Clin Neurosci. 2020;74(5):311–7.

205

Electroconvulsive Therapy Uses and Its Ability to Induce Neurogenesis: A Literature Review

41. Inta D, Lima-Ojeda JM, Lau T, Tang W, Dormann C, Sprengel R, et al. Electroconvulsive Therapy induces neurogenesis in frontal rat brain areas. PLoS One. 2013; 8(7):e69869.

57. Griesemer DA, Kellner CH, Beale MD, Smith GM. Electroconvulsive therapy for treatment of intractable seizures: Initial findings in two children. Neurology. 1997;49(5):1389–92.

42. Becker S, Wojtowicz JM. A model of hippocampal neurogenesis in memory and mood disorders. Trends Cogn Sci. 2007;11(2):70–6.

58. Shah N, Pande N, Bhat T, Murke M, Andrade C. Maintenance ECT as a therapeutic approach to medication-refractory epilepsy in an adult with mental retardation: Case report and review of literature. J ECT. 2012;28(2):136–40.

43. Cho KO, Lybrand ZR, Ito N, Brulet R, Tafacory F, Zhang L, et. al. Aberrant hippocampal neurogenesis contributes to epilepsy and associated cognitive decline. Nat. Commun. 2015;6(1):6606. 44. Hattiangady B, Shetty AK. Implications of decreased hippocampal neurogenesis in chronic temporal lobe epilepsy. Epilepsia. 2008;49(5):26–41. 45. Hester MS, Danzer SC. Hippocampal granule cell pathology in epilepsy - a possible structural basis for comorbidities of epilepsy? Epilepsy Behav. 2014;38:105–16. 46. Jessberger S, Parent JM. Epilepsy and adult neurogenesis. Cold Spring Harb Perspect Biol. 2015;7(12):a020677. 47. Danzer SC, Kotloski RJ, Walter C, Hughes M, McNamara JO. Altered morphology of hippocampal dentate granule cell presynaptic and postsynaptic terminals following conditional deletion of TrkB. Hippocampus. 2008;18(7):668–78. 48. Danzer SC. Neurogenesis in epilepsy: Better to burn out or fade away? Epilepsy Curr. 2016;16(4):268–9. 49. Varma P, Brulet R, Zhang L, Hsieh J. Targeting seizure induced neurogenesis in a clinically relevant time period leads to transient but not persistent seizure reduction. J Neurosci. 2019;39(35):7019–28. 50. Salik I, Marwaha R. Electroconvulsive Therapy. StatPearls, Treasure Island (FL): StatPearls Pub; 2021. 51. Parent JM, Lowenstein DH. Seizure-induced neurogenesis: Are more new neurons good for an adult brain? Prog Brain Res. 2002;135:121–31. 52. Wilkinson ST, Sanacora G, Bloch MH. Hippocampal volume changes following electroconvulsive therapy: A systematic review and meta-analysis. Biol Psychiatry Cogn Neurosci Neuroimaging. 2017;2(4):327–35. 53. Regenold WT, Weintraub D, Taller A. Electroconvulsive therapy for epilepsy and major depression. Am J Geriatr Psychiatry.1998;6(2):180–3. 54. Lunde ME, Lee EK, Rasmussen KG. Electroconvulsive therapy in patients with epilepsy. Epilepsy Behav. 2006;9(2):355–9. 55. Gryglewski G, Baldinger-Melich P, Seiger R, Godbersen GM, Michenthaler P, Klöbl M, et. al. Structural changes in amygdala nuclei, hippocampal subfields and cortical thickness following electroconvulsive therapy in treatmentresistant depression: Longitudinal analysis. Br J psychiatry. 2019;214(3):159–67. 56. Abbott CC, Jones T, Lemke NT, Gallegos P, McClintock SM, Mayer AR, et. al. Hippocampal structural and functional changes associated with electroconvulsive therapy response. Transl Psychiatry. 2014;4(11):e483. 206

59. Lambrecq V, Villéga F, Marchal C, Michel V, Guehl D, Rotge JY, et. al. Refractory status epilepticus: Electroconvulsive therapy as a possible therapeutic strategy. Seizure. 2012;21(9):661–4. 60. Schneegans H, Stetefeld H, Dohmen C, Onur OA, Lehnhardt FG. Successful treatment of super-refractory status epilepticus with high-intensity electroconvulsive therapy – a case report and review of the current literature. J. Epilepsy Res. 2019;9(1):76–82. 61. Li Q, Liu S, Guo M, Yang CX, Xu Y. The principles of electroconvulsive therapy based on correlations of schizophrenia and epilepsy: A view from brain networks. Front Neurol. 2019;10:688. 62. Bullmore E, Sporns O. Complex brain networks: Graph theoretical analysis of structural and functional systems. Nat Rev Neurosci. 2009;10(3):186–98. 63. Nordanskog P, Larsson MR, Larsson EM, Johanson A. Hippocampal volume in relation to clinical and cognitive outcome after electroconvulsive therapy in depression. Acta Psychiatr Scand. 2014;129(4):303–11. 64. Bessa JM, Ferreira D, Melo I, Marques F, Cerqueira JJ, Palha JA, et. al. The mood-improving actions of antidepressants do not depend on neurogenesis but are associated with neuronal remodeling. Mol Psychiatry. 2009;14(8):764–73. 65. Sackeim HA, Prudic J, Devanand DP, Nobler MS, Lisanby SH, Peyser S, et. al. A prospective, randomized, double-blind comparison of bilateral and right unilateral electroconvulsive therapy at different stimulus intensities. Arch Gen Psychiatry. 2000;57(5):425–34. 66. Heikman P, Kalska H, Katila H, Sarna S, Tuunainen A, Kuoppasalmi K. Right unilateral and bifrontal electroconvulsive therapy in the treatment of depression: A preliminary study. J ECT. 2002;18(1):26–30. 67. Loo CK, Mahon M, Katalinic N, Lyndon B, Hadzi-Pavlovic D. Predictors of response to ultrabrief right unilateral electroconvulsive therapy. J Affect Disord. 2011;130(12):192–7. 68. Pagnin D, de Queiroz V, Pini S, Cassano GB. Efficacy of ECT in depression: A meta-analytic review. J ECT. 2008;6(1):155-62. 69. Joshi SH, Espinoza RT, Pirnia T, Shi J, Wang Y, Ayers B, et. al. Structural plasticity of the hippocampus and amygdala induced by electroconvulsive therapy in major depression. Biol Psychiatry. 2016;79(4):282–92.

Electroconvulsive Therapy Uses and Its Ability to Induce Neurogenesis: A Literature Review

70. Dombrovski AY, Mulsant BH, Haskett RF, Prudic J, Begley AE, Sackeim HA. Predictors of remission after electroconvulsive therapy in unipolar major depression. J Clin Psychiatry. 2005;66(8):1043–9.

83. Jansson L, Wennström M, Johanson A, Tingström A. Glial cell activation in response to electroconvulsive seizures. Prog Neuropsychopharmacol Biol Psychiatry. 2009;33(7):1119–28.

71. Kellner CH, Knapp RG, Petrides G, Rummans TA, Husain MM, Rasmussen K, et. al. Continuation electroconvulsive therapy vs pharmacotherapy for relapse prevention in major depression: A multisite study from the consortium for research in electroconvulsive therapy (CORE). Arch Gen Psychiatry. 2006;63(12):1337-44.

84. Ferguson AV, Latchford KJ, Samson WK. The paraventricular nucleus of the hypothalamus: A potential target for integrative treatment of autocrine dysfunction. Expert Opin Ther Targets. 2008;12(6):717–27.

72. Stanton L, Kohn R. Depression and the aging brain. Med Health R I. 2012;95(7):210–1 73. Sackeim HA. Electroconvulsive therapy in late-life depression. Clinical geriatric psychopharmacology. 4th ed. Baltimore: Lippincott Williams & Wilkins. 2004;385 p. 74. Spaans HP, Sienaert P, Bouckaert F, van den Berg JF, Verwijk E, Kho KH, et. al. Speed of remission in elderly patients with depression: electroconvulsive therapy v. medication. Br J Psychiatry. 2015;206(1):67–71. 75. Bouckaert F, De Winter FL, Emsell L, Dols A, Rhebergen D, Wampers M, et. al. Grey matter volume increase following electroconvulsive therapy in patients with late life depression: A longitudinal MRI study. J Psychiatry Neurosci. 2016;41(2):105. 76. Bouckaert F, Dols A, Emsell L, De Winter FL, Vansteelandt K, Claes L, et. al. Relationship between hippocampal volume, serum BDNF, and depression severity following electroconvulsive therapy in late-life depression. Neuropsychopharmacology. 2016;41(11):27418. 77. Bjølseth TM, Engedal K, Benth JŠ, Dybedal GS, Gaarden TL, Tanum L. Clinical efficacy of formula-based bifrontal versus right unilateral electroconvulsive therapy (ECT) in the treatment of major depression among elderly patients: A pragmatic, randomized, assessor-blinded, controlled trial. J Affect Disord. 2015;175:8–17. 78. Perera TD, Coplan JD, Lisanby SH, Lipira CM, Arif M, Carpio C, et. al. Antidepressant-induced neurogenesis in the hippocampus of adult nonhuman primates. J Neurosci. 2007;27(18):4894901. 79. Dukart J, Regen F, Kherif F, Colla M, Bajbouj M, Heuser I, et. al. Electroconvulsive therapy induced brain plasticity determines therapeutic outcome in mood disorders. Proc Natl Acad Sci U S A. 2014;111(3):1156–61. 80. Jelovac A, Kolshus E, McLoughlin DM. Relapse following successful electroconvulsive therapy for major depression: A meta-analysis. Neuropsychopharmacology. 2013;38(12):2467–74. 81. Schloesser RJ, Orvoen S, Jimenez DV, Hardy NF, Maynard KR, Sukumar M, et. al. Antidepressant-like effects of electroconvulsive seizures require adult neurogenesis in a neuroendocrine model of depression. Brain Stimul. 2015;8(5):862–70. 82. Bolwig, TG et. al. How does electroconvulsive therapy work? Theories on its mechanism. Can J Psychiatry. 2011;56(1):13–8.

85. Crespo D, Ramos J, Gonzales C, Fernandez-Viadero C. The supraoptic nucleus: A morphological and quantitative study in control and hypophysectomised rats. J Anat. 1990;169:115–23. 86. Ježová D, Michajlovskij N, Kvetńanský R, Makara GB. Paraventricular and supraoptic nuclei of the hypothalamus are not equally important for oxytocin release during stress. Neuroendocrinology. 1993;57(5):776–81. 87. Noble EE, Billington CJ, Kotz CM, Wang C. Oxytocin in the ventromedial hypothalamic nucleus reduces feeding and acutely increases energy expenditure. Am J Physiol Regul Integr Comp Physiol. 2014;307(6):R737–R45. 88. Liu L, Wang Q, Liu A, Lan X, Huang Y, Zhao Z, et. al. Physiological implications of orexins/hypocretins on energy metabolism and adipose tissue development. ACS Omega. 2020; 5(1):547–55. 89. Yuko M, Shinichi T, Katsuya T, Seiichi T, Yoichi U, Kenju S. Orexin action on oxytocin neurons in the paraventricular nucleus of the hypothalamus. Neuroreport. 2017;28(6):330–366. 90. Hashikawa Y, Hashikawa K, Falkner AL, Lin D. Ventromedial hypothalamus and the generation of aggression. Front Syst Neurosci. 2017;11:94. 91. Gaur A, Pal GK, Ananthanarayanan PH, Pal P. Role of ventromedial hypothalamus in high fat diet induced obesity in male rats: Association with lipid profile, thyroid profile and insulin resistance. Ann Neurosci. 2014;21(3):104–7. 92. Nomoto K, Lima SQ. Enhance male-evoked responses in the ventromedial hypothalamus of sexually receptive female mice. Curr Biol. 2015;25(5):589–94. 93. Chee MJ, Myers MG, Price CJ, Colmers WF. Neuropeptide Y suppresses anorexigenic output from the ventromedial nucleus of the hypothalamus. J Neurosci. 2010;30(9):3380– 90. 94. Dimitrov EL, DeJoseph RM, Brownfield MS, Urban JH. Involvement of neuropeptide Y Y1 receptors in the regulation of neuroendocrine corticotropinreleasing hormone neuronal activity. Endocrinology. 2007;148(8):3666–73. 95. Smith SM, Vale WW. The role of the hypothalamicpituitary-adrenal axis in neuroendocrine responses to stress. Dialogues Clin Neurosci. 2006;8(4):383–95. 96. El Sayed SA, Fahmy MW, Schwarz J. Physiology, pituitary gland. StatPearls, Treasure Island (FL): StatPearls Pub. 2021. 97. Haskett RF, Zis AP, Albala AA. Hormone response to repeated electroconvulsive therapy: Effects of naloxone. Biol Psychiatry. 1985;20(6):623–33.

207

Electroconvulsive Therapy Uses and Its Ability to Induce Neurogenesis: A Literature Review

98. US Department of Health and Human Services. Naloxone: Medical countermeasures database [Internet]. Maryland USA: National Libraries of Medicine; 2014 [updated 2020 Dec 21; cited 2021 Feb 20]. Available from https://chemm. nlm.nih.gov/countermeasure_naloxone.htm. 99. Samuels MH. Psychiatric and cognitive manifestations of hypothyroidism. Curr Opin Endocrinol Diabetes Obes. 2014;21(5):377–83. 100. Stern RA, Tremont G. Minimizing the cognitive effects of lithium therapy and electroconvulsive therapy using thyroid hormone. Int J Neuropsychopharmacol. 2000;3(2):175-86. 101. Read J, Fosse R. Electroconvulsive treatment: Hypotheses about mechanisms of action. Front Psychiatry. 2013;4:94. 102. Yu, J. Endocrine disorders and the neurologic manifestations. Ann Pediatr Endocrinology Metab. 2014;19(4):184–90.

208

Scholarly Research In Progress • Vol. 5, November 2021

Improving the Future of the Opioid Epidemic: Methocinnamox Colleen G. Jordan1*, Amy L. Kennalley1*, Tenzing Dolma1*‡, Kaitlyn M. Nemes1*‡, and Alivia L. Roberts1*‡ ¹Geisinger Commonwealth School of Medicine, Scranton, PA 18509 *Master of Biomedical Sciences Program ‡ Authors contributed equally Correspondence: cjordan02@som.geisinger.edu

Abstract The opioid epidemic is an ongoing public health crisis, and the United States health system is overwhelmed with increasing numbers of opioid-related overdoses. Methocinnamox (MCAM) is a novel mu-opioid receptor antagonist with an extended duration of action and potential to reduce the burden of the opioid epidemic through overdose rescue that could treat opioid use disorder (OUD) long-term. We compared the efficacy and effects of MCAM to the current treatments available to treat OUD, such as naloxone and naltrexone, which have their own limitations. A literature review was conducted using PubMed and Google Scholar databases. MCAM’s novel properties open a new avenue for treating the opioid crisis. The known therapeutic effect of MCAM can reduce the number of opioid deaths and reduce the number of relapse events in those with OUD. MCAM could be used as both a rescue and long-term treatment for opioid misuse. This is due to its pseudo-irreversible antagonism of the mu opioid receptor, abnormally long duration of action of nearly 2 weeks, and the possibility of using kappa or delta opioid receptor agonists for pain management during OUD treatment. Current studies in animal models show promise for this drug’s potential in humans to reduce opioid misuse and the impacts of the crisis, although further research is still needed.

Introduction Opioid addiction and misuse remain a prevalent issue in the United States (U.S.) (1). In 2019 alone, more than 70,000 deaths were attributed to opioid overdose in the U.S. (2). Opioids were originally discovered from poppy plants and were used to reduce pain sensation ranging from acute to severe, but they have become more accessible for recreational use outside of pain relief therapy (3, 4). The intended use of opioid pharmacologics was for the reduction of pain sensation by agonizing the opioid receptors located in the central nervous system (CNS) (5). There are three major opioid receptor types, mu (μ), delta (δ), and kappa (κ), but the mu-receptor is the main target of exogenous opioids (5, 6). In the past, many people turned to opioids to relieve daily suffering from chronic pain, and the drugs easily became addictive and created dependence (1). Today, synthetic opioids commonly make treatment for overdose and addiction exceptionally difficult due to altered administration, uptake, elimination rates, and their ability to bind to receptors (3). The heavy use of opioids and the addiction to these drugs in the U.S. have exacerbated the strain on resources in hospitals, emergency rooms, and on first responders as they try to save lives with the limited resources currently available (1). Naloxone is the only drug available to treat opioid overdose to be approved by the U.S. Food and Drug Administration (FDA) in the last 50 years, and the opioid users are younger and experimenting with synthetic opioid use beyond pain relief (7,

8). Naloxone is a competitive mu opioid antagonist with a high affinity for the mu-receptor used to reverse respiratory and CNS depression in those enduring an opioid overdose (9–11). Although naloxone is important to combat opioid overdose, it is not a long-term fix because it has the potential for dependence (10). Naloxone does not help decrease future use of opioids, and the use of synthetic opioids will require higher doses of naloxone, which could increase adverse effects such as tachycardia and hypertension (10, 12, 13). For these reasons, there is a dire need for a new opioid overdose intervention (14). Methocinnamox (MCAM) is a novel drug candidate that is a pseudo-irreversible antagonist for the mu opioid receptor (MOR), thereby preventing other opioid agonists from binding for a 2-week period (14–16). Due to the long-lasting effect of MCAM, it can be a safer and more effective alternative medication for the misuse of opioids (17, 18). MCAM has the potential to change the course of opioid misuse and help prevent relapse after administration (19, 20). This paper will explore how MCAM’s unique function could be useful in reducing the opioid crisis burden through its use in overdose rescue, long-term OUD (opioid use disorder) treatment, potential to be used with kappa and delta opioid receptor agonists, and nearly 2-week duration of action on a single dose.

Methods A literature review was conducted using PubMed and Google Scholar databases utilizing the following key terms: methocinnamox, MCAM, naloxone, naltrexone, buprenorphine, buprenorphine-naloxone, methadone, opioid overdose, opioid crisis, opioid abuse, mu-receptor, kappa-receptor, deltareceptor, inverse agonist, and naloxone and placebo. These terms and pharmacologics were included due to their relevance to the opioid epidemic, opioid overdose, and MCAM. Chemical structure made in ChemDraw version 19.0.0. No date range or journal exclusion was applied.

Discussion Opioid epidemic In the 1800s, the medicinal benefits of opiates were widely marketed as a safe and effective form of pain alleviation (21). Consequently, the absence of federal regulation on frequent opioid prescription and use drew widespread concern, which eventually led to the enactment of the 1914 Harrison Narcotic Control Act (22). While this prompted nationwide stigmatization of opioid use for non-cancer chronic pain management, it was later followed by a drastic shift in public attitude that advocated for the recognition of pain as a “fifth vital sign” in 1995 (23). As a result, several entities such as the Institute of Medicine, the Federation of State Medical

209

Improving the Future of the Opioid Epidemic: Methocinnamox

Boards, and the Drug Enforcement Agency, pushed for fewer regulations over opioid prescriptions, thereby encouraging health care providers to provide adequate pain relief for patients. Additionally, in 1995 the FDA approved an extendedrelease oxycodone formulation as a safer opioid alternative to the fast-release version because of its slow and sustained release of medication (24). Pressure by pharmaceutical companies, patients, and federal funding requirements further contributed to the overaggressive prescription of opioid analgesics that ultimately led to the rise of the opioid epidemic (25, 26). The Centers for Disease Control reported that in 2016, more than 42,000 Americans died from an opioid overdose, marking a 27% increase from 2015 (27). In 2017, the rate increased by 45.2%, indicating the increased prevalence of opioid misuse (28). More than 11.5 million Americans misuse opioids and roughly 2.1 million were formally diagnosed with an OUD (29). Later that year, the U.S. Department of Health and Human Services declared the opioid epidemic a public health emergency. Despite concerted efforts by medical practitioners to reduce opioid prescriptions, opioid overdose continues to rise, with the illegal manufacturing of fentanyl and its analogs as the leading cause. This can be attributed to fentanyl’s high potency, with a strength that is 30 to 50 times greater than that of heroin, its rapid onset of action, long duration of desired effect, and low production costs. While the rate of heroinrelated overdose deaths has started to stabilize, synthetic opioid-related deaths, such as those caused by fentanyl and its analogs, increased by 88% from 2013 to 2016 (25, 30). Public health experts agree that tackling the opioid epidemic will require interdisciplinary collaboration between medical providers, social service agencies, federal regulation, and community support (8, 31, 32). The lingering effects of the epidemic are rampant in low-income communities, predominantly African American and Hispanic communities, and are currently exacerbated by the social and health impacts of COVID-19 (33, 34). Potential proposed solutions include increasing harm-reduction programs, educating medical providers on safe opioid prescribing, eliminating stigma around OUDs, as well as finding safer alternatives to pain management (2, 3, 32, 35). Current OUD therapeutics include methadone, buprenorphine, naloxone, and extended-release naltrexone, all which function by reducing opioid withdrawal symptoms and cravings (1, 5). However, these medications can only successfully resolve the opioid crisis by working in tandem with public health efforts that include both prevention and harmreduction approaches (36). MCAM presents potential for a new avenue of OUD treatment. Pain receptors MCAM is a long lasting, pseudo-irreversible (non-covalent), potent, MOR antagonist that has no known interaction with nociceptors and reversibly binds kappa opioid receptors (KOR) and delta opioid receptors (DOR). Thus, kappa and delta agonists could be provided concomitantly for pain relief during treatment for OUD (15, 17–19, 37, 38). This unique pharmacodynamic mechanism of MCAM contributes to its long-lasting effects; the need for new MORs to induce the euphoric and depressive effects of opioid receptor agonists as receptor turnover is what limits the duration of action (14, 19). This is crucial because MOR agonists can not only induce the G protein-coupled receptor (GPCR) pathway, but can also induce 210

β-arrestin activation, leading to side effects such as respiratory depression (39, 40). The opioid receptors: MOR, KOR, and DOR, belong to the largest membrane receptor family called the trimeric GPCR superfamily, with opioids activating the inhibitory (Gi) signaling pathway to initiate analgesic functions (41–44). The GPCRs are known for their trimeric subunits consisting of alpha (Gα), beta (Gβ), and gamma (Gγ), (45). After the opioid agonist (endogenous or exogenous) binds, a signal stimulates Gα to migrate and suppress adenylate cyclase activity, thereby reducing cyclic AMP production (45). The Gβγ acts as a modulator for the signaling pathway, resulting in reduced neurotransmitter release and membrane hyperpolarization (45). Since GPCRs are so widespread, it is the target for 50% of marketed pharmacological therapeutics, revolving around the common amino-terminal peptide sequence, Tyr-Gly-Gly-Phe, which is referred to as the “opioid motif” as it directly interacts with the opioid receptor (46). The MOR agonists mentioned are oxycodone, fentanyl, heroin, morphine, and methadone. Buprenorphine is a partial MOR agonist and KOR antagonist (47, 48). MOR antagonists include naloxone, MCAM, and naltrexone. It is believed the activation of KORs antagonize MOR mediated analgesia; activation of the KOR hyperpolarizes neurons that are active indirectly by the MOR (7). Pain is multidimensional and dependent on subjective thresholds; chronic pain, which may be concurrent with anxiety, may be associated with neuroplastic changes in the amygdala, which may heighten the emotional and affective consequences of pain (49, 50). Analgesics on the pharmaceutical market are highly effective in most cases, but the desired effects mediated by the opioid receptor family may lead to craving, addiction, or dependence as a result of neurological changes (51–54). Repetitive opioid use will thus increase the threshold for analgesic effects secondary to compensatory upregulation of vesicular calcium content while developing opiate tolerance and may decrease one’s quality of life (42, 55, 56). Current OUD treatments The current pharmacological treatments for opioid overdose and misuse are administration of methadone, buprenorphine, naloxone, and naltrexone (9, 11, 13, 57–59). Methadone and buprenorphine are opioid agonists which may prevent withdrawal symptoms in those recovering from OUD, but pose risk for opioid overdose (57, 60, 61). However, naltrexone and naloxone are opioid antagonists, the latter being the only emergency rescue for opioid overdose and opioid induced respiratory symptoms (11, 62). Naltrexone is used to treat OUD and opioid dependency, usually post-opioid cessation, whereas naloxone can be used concomitantly with prescribed opioids such as buprenorphine (57–59, 63). Other pharmacological uses have been identified for naloxone and naltrexone such as treatment for alcohol dependence, and possible treatments for internet sex addiction, and Hailey–Hailey disease, but studies show these medications are not effective for smoking cessation (64–67). Administration of methadone or buprenorphine significantly reduces opioid-related deaths caused by nonfatal opioid overdose over a 12-month follow-up period by 59% and

Improving the Future of the Opioid Epidemic: Methocinnamox

38%, respectively (68). The abrupt discontinuation of opioids does not show great success rates and may result in relapse (8, 57). Use of these drugs in conjunction with psychosocial therapy are the best for treatment success in those with OUD. (8, 57, 59). While both methadone and buprenorphine are synthetic derivatives of opiates and used in medication assisted treatment (MAT) of OUD, they possess different mechanisms of action and adverse drug reactions (ADRs). Methadone is a longacting full agonist that binds the MOR, preventing withdrawal symptoms such as nausea and vomiting for at least 24 hours, while conferring analgesia and reducing opioid cravings (69, 70). Conversely, buprenorphine is a partial agonist at the MOR, making it less potent than methadone with decreased ADRs (70, 71). Additionally, buprenorphine’s “ceiling effect” reduces the risk of misuse or overdose by preventing the increase of opioid effects, or euphoria, beyond a designated threshold (70). However, buprenorphine may precipitate withdrawal, a condition that occurs without an adequate detoxification period from opioid drugs, due to its high affinity for the MOR (70, 71). An additional drug that is also used in MAT programs is naloxone, which is often paired with buprenorphine in an oral tablet form to prevent strong withdrawal symptoms and block the euphoric effects induced by other opioids (58). The mechanistic action of naloxone is by competitive binding to the MOR as a high affinity antagonist, and some researchers suggest it acts as an inverse agonist (11, 72). Administration of naloxone intravenously, intramuscularly, subcutaneously, intranasally, and even inhalation through endotracheal tube for intubated patients, during an opioid crisis competitively binds the opioid receptors to reverse respiratory and CNS depression (73). The time to decrease fentanyl occupancy at the receptor after 2 mg intramuscular naloxone administration is 3 minutes, after 25 ng/ml and 10 minutes after 50 ng/ml, but there is an increased dose requirement of naloxone due to larger doses of self-administered opioids (13). After a 13 μg/kg dose of naloxone, 50% of the receptors in the brain are occupied, but due to the rapid association and successive dissociation of naloxone from the receptors, toxicity reversal may be insufficient and the patient may experience renarcotization requiring subsequent doses (13, 73, 74). Although regarded as exceedingly safe, ADRs for naloxone can include tachycardia, hypertension, gastrointestinal upset, hyperthermia, cravings, nausea, vomiting, and rarely severe cardiovascular events (9, 13, 73). Naloxone also blocks the descending pain control system, thus diminishing the placebo pathway for pain perception by interfering with the coupling between the rostral anterior cingulate cortex and the periaqueductal gray area structures in the brain (75, 76). Though not the first-line treatment for opioid overdose, as it is not as efficacious as naloxone, naltrexone is used to reduce opioid use in those with OUD (59, 62). Naltrexone is an opioid receptor antagonist, some suggest an inverse agonist, that is prescribed to reduce opioid use in those who are attempting to practice abstinence from opioids but suffer from OUD (59, 62). Interventions for opioid misuse involving naltrexone, rather than receptor agonist like buprenorphine, have been successful when paired with behavior intervention and are a promising alternative treatment for opioid misuse in pregnant women (77, 78). In contrast to naloxone’s associated acute withdrawal symptoms, naltrexone reduces symptoms of withdrawal for

patients and even lowers the risk for overdose with the use of buprenorphine as an OUD treatment with no significant ADRs (63). Also having a considerable safety profile like naloxone, the potential side effects of naltrexone include mild to moderate injection site reaction, nausea, and gastrointestinal upset (79). With the intervention limitations mentioned above, MCAM may prove beneficial as a treatment to combat the opioid crisis. Methocinnamox MCAM, shown in Figure 1, was first mentioned in a publication in 2000 by researchers from the University of Michigan Medical School and the University of Bristol, but was initially discarded because it was believed to be useful only for MOR research purposes (15, 80). However, it is currently being studied for its promise in the opioid crisis as a long-term OUD treatment with funding from the National Institute of Health and the National Institute on Drug Abuse (81, 82). In animal models, a single subcutaneous dose of MCAM rescues a subject from acute opioid overdose and prevents subsequent overdose for up to 2 weeks with minimal adverse effects (16, 17, 80, 83, 84). Currently, the only known possible adverse effect for MCAM in non-dependent individuals is hyperventilation upon rescue (84). Interestingly, one study noted a slightly increased response to food (a non-drug alternative) several days

Figure 1. The chemical structure of methocinnamox is shown. Molecular Formula: C30H32N2O4, PubChem CID: 46877713, IUPAC name: (E)-N-(4R,4aS,7aR,12bR)-3-(cyclopropylmethyl)-9-hydroxy-7-oxo2,4,5,6,7a,13-hexahydro-1H-4,12-methanobenzofuro3,2-eisoquinolin4a-yl-3-(4-methylphenyl)prop-2-enamide (compound/Methocinnamox) (91).

211

Improving the Future of the Opioid Epidemic: Methocinnamox

following a single injection (17). Some studies have shown no statistically significant adverse effects nor potential ADRs with benzodiazepines and alcohol (14, 17, 85). MCAM has not been shown to cause a decrease in response to food or alter heart rate, blood pressure, body temperature, or social and physical activity and no indication of developing tolerance nor physical dependence (14, 17). MCAM is currently the most potent and selective MOR antagonist and shows no agonistic effects, even at high concentrations, with longest duration and highest potency when injected subcutaneously over other methods of administration (15, 84). Naltrexone and naloxone injections become ineffective in less than a single day with durations of action lasting 1 to 2 hours (73, 74). A single injection of MCAM has a duration of action of 13 days, reaching peak concentration 15 to 45 minutes after injection with a half-life of roughly 70 minutes (14, 19). MCAM’s exact mechanism of action is currently unknown, but the effectiveness at very low plasma levels suggests the pharmacodynamic properties play a significant role in its long-lasting effects rather than pharmacokinetic factors (14). Repeated administration of MCAM every 12 days in rodents remained effective for over 2 months without altering the duration of opioid withdrawal with no major ADRs and no decrease in effectiveness, suggesting positive potential for long-term OUD treatment (14, 37, 84). Naltrexone, naloxone, and MCAM are effective for acute reversal and prevention of respiratory depression and other overdose symptoms due to their effects on opioid receptors, but only MCAM prevents renarcotization in the hours and days following emergency intervention (14, 37, 84, 86, 87). Naltrexone and naloxone bind competitively, meaning higher amounts of an agonist will overcome their intended effects requiring a higher dose of either therapy to reverse initial and subsequent overdoses post-antagonist-injection (86). MCAM binds non-competitively, making it insurmountable and therefore more effective at blocking effects of opioids in the short and long term (14, 19, 87). MCAM can act as a preventative therapy for opioid misuse, indicating possible use at discharge from treatment facilities following a detoxification period, as well as use during ongoing therapeutic intervention negating the need for, and misuse of, buprenorphine and methadone (17, 19, 37, 87). Its prolonged dosing interval is hypothesized to relatively prevent patient noncompliance that is seen with extended-release naltrexone for outpatient treatment, including eliminating the possibility of an individual removing an implant (17, 19). In cases where a duration of action lasting roughly 5 days or less is needed, such as preventing renarcotization in the hours and few days following an overdose but not for long term treatment of OUD, intravenous administration of MCAM would be preferable (84, 86). There is discussion of creating an oral pill form of MCAM, an extended-release form, and faster-acting intranasal and intramuscular formulations, but further study of the drug is needed before these will be created (14, 17, 86). MCAM also blocks the physiological and behavioral effects of MOR agonists such as unfavorable impacts of sensitivity to mechanical stimulation, gastrointestinal motility, appetite, and memory and other cognition, suggesting the adverse effect profile is encouraging although no testing has been conducted in humans (16, 17, 20, 85, 88).

212

Conclusion The increased prevalence of OUD cases and opioid related deaths are an ongoing public health crisis in the US. While opioid antagonists naltrexone and naloxone are essential drugs used to treat OUD and reverse the effects of an overdose, respectively, they have risks that pose considerable limitations to their efficacy. These risks include withdrawal, poor patient compliance, short durations of action, lack of concurrent antinociceptive treatment, ability to surmount opioid receptor blockade, and potentially dangerous drug-drug interactions, especially for those with comorbid addictions (14, 17). The demand for novel therapeutics to decrease the misuse and overuse of opioid drugs and resulting overdoses provides an opportunity for MCAM to make a positive impact. By retaining the safety benefits of naltrexone and naloxone and providing a longer duration of action with a novel mechanism, MCAM is a promising pharmacological addition. Using non-MOR agonists such as the KOR agonist spiradoline concomitantly with MCAM also presents a potential intervention method allowing for antinociceptive effects during the withdrawal process and OUD treatment (37, 88, 89, 90). The preclinical phase of MCAM drug development began in 2005 with testing in mice, rats, and nonhuman primates; researchers aim to begin phase I clinical trials in 2022 (16, 37, 80, 83–85, 87, 88). MCAM has the potential to transform the future of OUD treatment, thereby reducing the health care and societal burden caused by the opioid epidemic and improving the lives of millions.

Acknowledgments The authors extend gratitude to Brian Piper, PhD, for his assistance during the research and writing phases of this manuscript. The authors also extend gratitude to Iris Johnston in Library Services for her assistance during the research process.

Disclosures The authors have no conflicts of interest to disclose.

References 1.

Coussens NP, Sittampalam GS, Jonson SG, Hall MD, Gorby HE, Tamiz AP, et al. The opioid crisis and the future of addiction and pain therapeutics. J Pharmacol Exp Ther. 2019; 371:396–408.

France CP, Ahern GP, Averick S, Disney A, Enright HA, Esmaeli-Azad B, et al. Countermeasures for preventing and treating opioid overdose. Clin Pharmacol Ther. 2021; 109:578-590.

Bechara A, Berridge KC, Bickel WK, Morón JA, Williams SB, Stein JS. A neurobehavioral approach to addiction: Implications for the opioid epidemic and the psychology of addiction. Psychol Sci Public Interest. 2019; 20:96–127.

Torralva R, Janowsky A. Noradrenergic mechanisms in fentanyl-mediated rapid death explain failure of naloxone in the opioid crisis. J Pharmacol Exp Ther. 2019; 371:453– 475.

Improving the Future of the Opioid Epidemic: Methocinnamox

Hoffman KA, Ponce Terashima J, McCarty D. Opioid use disorder and treatment: Challenges and opportunities. BMC Health Serv Res. 2019; 19:884.

19. Townsend EA. The lasting impact of methocinnamox on opioid self-administration. Neuropsychopharmacology. 2020; 45:1963–1964.

Raehal KM, Bohn LM. Mu opioid receptor regulation and opiate responsiveness. AAPS J. 2005; 7:E587–91.

Pan ZZ. mu-Opposing actions of the kappa-opioid receptor. Trends Pharmacol Sci. 1998; 19:94–98.

20. Minervini V, France C. Long-term antagonism of mu opioid receptors by methocinnamox (MCAM) without adverse effects on memory. FASEB J. 2020; 34:1.

Tofighi B, Williams AR, Chemi C, Suhail-Sindhu S, Dickson V, Lee JD. Patient barriers and facilitators to medications for opioid use disorder in primary care: An in-depth qualitative survey on buprenorphine and extended-release naltrexone. Subst Use Misuse. 2019; 54:2409–2419. Skolnick P. On the front lines of the opioid epidemic: Rescue by naloxone. Eur J Pharmacol. 2018; 835:147–153.

10. Ding H, Ko M-C. Translational value of non-human primates in opioid research. Exp Neurol. 2021; 338:e113602. 11. Ryan SA, Dunne RB. Pharmacokinetic properties of intranasal and injectable formulations of naloxone for community use: A systematic review. Pain Manag. 2018; 8:231–245. 12. Hammerslag LR, Hofford RS, Kang Q, et al. Changes in fentanyl demand following naltrexone, morphine, and buprenorphine in male rats. Drug Alcohol Depend. 2020; 207:107804. 13. Moss RB, Carlo DJ. Higher doses of naloxone are needed in the synthetic opioid era. Subst Abuse Treat Prev Policy. 2019; 14:6. 14. Maguire DR, Gerak LR, Sanchez JJ, Javors MA, Disney A, Husbands SM, et al. Effects of acute and repeated treatment with methocinnamox, a mu opioid receptor antagonist, on fentanyl self-administration in rhesus monkeys. Neuropsychopharmacology. 2020; 45:1986– 1993. 15. Broadbear JH, Sumpter TL, Burke TF, Husbands SM, Lewis JW, Woods JH, et al. Methocinnamox is a potent, long-lasting, and selective antagonist of morphinemediated antinociception in the mouse: Comparison with clocinnamox, beta-funaltrexamine, and betachlornaltrexamine. J Pharmacol Exp Ther. 2000; 294:933– 940. 16. Gerak LR, Gould GG, Daws LC, France CP. Methocinnamox produces sustained antagonism of the antinociceptive effects of morphine and persistent decreases in [3H] DAMGO binding to μ opioid receptors in rat cortex. FASEB J. 2020; 34:1. 17. Maguire DR, Gerak LR, Woods JH, Husbands SM, Disney A, France CP. Long-lasting effects of methocinnamox on opioid self-administration in rhesus monkeys. J Pharmacol Exp Ther. 2019; 368:88–99. 18. Zamora JC, Kotipalli V, Jennings EM, Disney A, Husbands S, Winger G, et al. Methocinnamox (MCAM) is a selective, long acting antagonist at mu opioid receptors in vitro. FASEB J. 2019; 33:498.

21. Lyden J, Binswanger IA. The United States opioid epidemic. Semin Perinatol. 2019; 43:123–131. 22. Courtwright DT. Preventing and treating narcotic addiction — A century of federal drug control. New Eng J Med. 2015; 373:2095–2097. 23. Levy N, Sturgess J, Mills P. “Pain as the fifth vital sign” and dependence on the “numerical pain scale” is being abandoned in the US: Why? Br J Anaesth. 2018; 120:435– 438. 24. Van Zee A. The promotion and marketing of OxyContin: Commercial triumph, public health tragedy. Am J Public Health. 2009; 99:221–227. 25. Jones MR, Viswanath O, Peck J, Kaye AD, Gill JS, Simopoulos TT. A brief history of the opioid epidemic and strategies for pain medicine. Pain Ther. 2018; 7:13–21. 26. Singh VM, Browne T, Montgomery J. The emerging role of toxic adulterants in street drugs in the US illicit opioid crisis. Public Health Rep. 2020; 135:6–10. 27. Seth P, Scholl L, Rudd RA, Bacon, S. Overdose deaths involving opioids, cocaine, and psychostimulants - United States, 2015-2016. MMWR Morb Mortal Wkly Rep. 2018; 67:349–358. 28. Stoicea N, Costa A, Periel L, Uribe A, Weaver T, Bergese SD. Current perspectives on the opioid crisis in the US healthcare system: A comprehensive literature review. Medicine (Baltimore). 2019; 98:e15425. 29. Hagemeier NE. Introduction to the opioid epidemic: The economic burden on the healthcare system and impact on quality of life. Am J Manag Care. 2018; 24:S200–S206. 30. Han Y, Yan W, Zheng Y, Khan MZ, Yuan K, Lu L. The rising crisis of illicit fentanyl use, overdose, and potential therapeutic strategies. Transl Psychiatry. 2019; 9:1–9. 31. McIver JS. Seeking Solutions to the Opioid Crisis. P T. 2017; 42:478. 32. Saloner B, McGinty EE, Beletsky L, Bluthenthal R, Beyrer C, Botticelli SG, et al. A public health strategy for the opioid crisis. Public Health Rep. 2018; 133:24S-34S. 33. Caton L, Cheng H, Garneau HC, Fisher T, Harris-Mills B, Hurley B, et al. COVID-19 adaptations in the care of patients with opioid use disorder: A survey of California primary care clinics. J Gen Intern Med. 2021;1–8. 34. Drake J, Charles C, Bourgeois JW, Daniel ES, Kwende M. Exploring the impact of the opioid epidemic in Black and Hispanic communities in the United States. Drug Sci Policy Law. 2020; 0:1–11. 35. Kolodny A, Courtwright DT, Hwang CS, Kreiner P, Eadie JL, Clark TW, et al. The prescription opioid and heroin crisis: A public health approach to an epidemic of addiction. Annu Rev Public Health. 2015; 36:559–574. 213

Improving the Future of the Opioid Epidemic: Methocinnamox

36. Volkow ND, Blanco C. The changing opioid crisis: Development, challenges and opportunities. Mol Psychiatry. 2021; 26:218–233. 37. Gerak LR, Latham EA, Woods JH, Disney A, Husbands SM, France CP. Methocinnamox: Sustained antagonism of the antinociceptive effects of morphine and not spiradoline in rats. FASEB J. 2019; 33:498. 38. Jennings EM, Disney A, Husbands S, Winger G, Berg KA, Clarke WP, et al. Methocinnamox (MCAM) is an effective long-term antagonist of peripheral mu, but not kappa or delta opioid receptors in vivo. FASEB J. 2019; 33:498. 39. Mafi A, Kim S-K, Goddard WA. Mechanism of β-arrestin recruitment by the μ-opioid G protein-coupled receptor. Proc Natl Acad Sci U S A. 2020; 117:16346–16355. 40. Manglik A, Lin H, Aryal DK, McCorvy JD, Dengler D, Corder G, et al. Structure-based discovery of opioid analgesics with reduced side effects. Nature. 2016; 537:185–190.

51. Al-Eitan LN, Rababa’h DM, Alghamdi MA. Genetic susceptibility of opioid receptor genes polymorphism to drug addiction: A candidate-gene association study. BMC Psychiatry. 2021; 21:5. 52. Liu SS, Pickens S, Burma NE, Ibarra-Lecue I, Yang H, Xue L, et al. Kappa opioid receptors drive a tonic aversive component of chronic pain. J Neurosci. 2019; 39:4162– 4178. 53. Massaly N, Copits BA, Wilson-Poe AR, Hipólito L, Markovic T, Yoon HJ, et al. Pain-induced negative affect is mediated via recruitment of the nucleus accumbens kappa opioid system. Neuron. 2019; 102:564-573. 54. Rosenblum A, Marsch LA, Joseph H, Portenoy, RK. Opioids and the treatment of chronic pain: Controversies, current status, and future directions. Exp Clin Psychopharmacol. 2008; 16:405–416. 55. Mercadante S, Arcuri E, Santoni A. Opioid-induced tolerance and hyperalgesia. CNS Drugs. 2019; 33:943–955.

41. Gibula-Tarlowska E, Kotlinska JH. Crosstalk between opioid and anti-opioid systems: An overview and its possible therapeutic significance. Biomolecules. 2020; 10:1376.

56. Kosten TR, George TP. The neurobiology of opioid dependence: Implications for treatment. Sci Pract Perspect. 2002; 1:13–20.

42. Al-Hasani R, Bruchas MR. Molecular mechanisms of opioid receptor-dependent signaling and behavior. Anesthesiology. 2011; 115:1363–1381.

57. Srivastava A, Kahan M, Nader M. Primary care management of opioid use disorders: Abstinence, methadone, or buprenorphine-naloxone? Can Fam Physician. 2017; 63:200–205.

43. Corder G, Castro DC, Bruchas MR, Scherrer, G. Endogenous and exogenous opioids in pain. Annu Rev Neurosci. 2018; 41:453–473. 44. Mondal D, Kolev V, Warshel A. Exploring the activation pathway and Gi-coupling specificity of the μ-opioid receptor. PNAS. 2020; 117:26218–26225. 45. Smrcka AV. G protein βγ subunits: Central mediators of G protein-coupled receptor signaling. Cell Mol Life Sci. 2008; 65:2191–2214. 46. Shenoy SS, Lui F. Biochemistry, endogenous opioids [Internet]. Florida: StatPearls Publishing; 2020 [cited 2021 March 2] Available from: http://www.ncbi.nlm.nih.gov/ books/NBK532899/ 47. Mancher M, Lechner AI. The effectiveness of medicationbased treatment for opioid use disorder [Internet]. United States: National Academies Press; 2019 [cited 2021 March 2]. Available from: http://www.ncbi.nlm.nih.gov/books/ NBK541393/ 48. Lipiński PFJ, Kosson P, Matalińska J, Roszkowski P, Czarnocki Z, Jarończyk M et al. Fentanyl family at the mu-opioid receptor: Uniform assessment of binding and computational analysis. Molecules; 24:1-30. 49. Navratilova E, Ji G, Phelps C, Qu C, Hein M, Yakhnitsa V, et al. Kappa opioid signaling in the central nucleus of the amygdala promotes disinhibition and aversiveness of chronic neuropathic pain. Pain. 2019; 160:824–832. 50. Zhou W, Li Y, Meng X, Liu A, Mao Y, Zhu, et al. Switching of delta opioid receptor subtypes in central amygdala microcircuits is associated with anxiety states in pain. J Biol Chem. 2021; 296:100277.

214

58. Heo Y-A, Scott LJ. Buprenorphine/naloxone (Zubsolv®): A review in opioid dependence. CNS Drugs. 2018; 32:875– 882. 59. Sullivan MA, Bisaga A, Pavlicova M, Carpenter KM, Choi CJ, Mishlen K, et al. A randomized trial comparing extended-release injectable suspension and oral naltrexone, both combined with behavioral therapy, for the treatment of opioid use disorder. Am J Psychiatry. 2019; 176:129–137. 60. Chen KY, Chen L, Mao J. Buprenorphine-naloxone therapy in pain management. Anesthesiology. 2014; 120:1262– 1274. 61. Kapur BM, Hutson JR, Chibber T, Luk A, Selby P. Methadone: A review of drug-drug and pathophysiological interactions. Crit Rev Clin Lab Sci. 2011; 48:171–195. 62. Jarvis BP, Holtyn AF, Subramaniam S, Tompkins DA, Oga A, Bigelow GE, et al. Extended-release injectable naltrexone for opioid use disorder: A systematic review. Addiction. 2018; 113:1188–1209. 63. Morgan JR, Schackman BR, Leff JA, Linus BP, Walley AY. Injectable naltrexone, oral naltrexone, and buprenorphine utilization and discontinuation among individuals treated for opioid use disorder in a United States commercially insured population. J Subst Abuse Treat. 2018; 85:90–96. 64. Sudakin D. Naltrexone: Not just for opioids anymore. J Med Toxicol. 2016; 12:71–75. 65. David SP, Lancaster T, Stead LF, Evins AE, Prochaska JJ. Opioid antagonists for smoking cessation. Cochrane Database Syst Rev. 2013; 6:CD003086.

Improving the Future of the Opioid Epidemic: Methocinnamox

66. Bostwick JM, Bucci JA. Internet sex addiction treated with naltrexone. Mayo Clin Proc. 2008; 83:226–230. 67. Garayar Cantero M, Canseco Martín M, Aguado García Á, Ruiz-Sánchez D, Valtueña J, Manchado Lopez, P. Use of low-dose naltrexone in the treatment of severe HaileyHailey disease: One case report. Dermatol Ther. 2019; 32:e12892. 68. Larochelle MR, Bernson D, Land T, Stopka TJ, Wang N, Xuan Z, et al. Medication for opioid use disorder after nonfatal opioid overdose and association with mortality: A cohort study. Ann Intern Med. 2018; 169:137–145. 69. Ferrari A, Coccia CPR, Bertolini A, Sternieri E. Methadone-metabolism, pharmacokinetics and interactions. Pharmacol Res. 2004; 50:551–559. 70. Whelan PJ, Remski K. Buprenorphine vs methadone treatment: A review of evidence in both developed and developing worlds. J Neurosci Rural Pract. 2012; 3:45–50. 71. Welsh C, Valadez-Meltzer A. Buprenorphine. Psychiatry (Edgmont). 2005; 2:29–39. 72. Cruz SL, Villarreal JE, Volkow ND. Further evidence that naloxone acts as an inverse opiate agonist: Implications for drug dependence and withdrawal. Life Sci. 1996; 58:381– 389. 73. Rzasa Lynn R, Galinkin J. Naloxone dosage for opioid reversal: Current evidence and clinical implications. Ther Adv Drug Saf, 2018; 9:63–88. 74. Schumacher MA, Basbaum AI, Naidu RK. Basic & clinical pharmacology [Internet]. 14th ed. New York: McGraw-Hill Education; 2017 [cited 2021 March 3]. Available from: https://accessmedicinemhmedical-com.gcsom.idm.oclc.org/content. aspx?sectionid=175220393&bookid=2249#175220593 75. Gracely RH, Dubner R, Wolskee PJ, Deeter WR. Placebo and naloxone can alter post-surgical pain by separate mechanisms. Nature. 1983; 306:264–265.

80. Mertens J. YOUR HEALTH: Something better than Naloxone. WQAD [Internet]. 2020 [cited 2021 February 14]; Your Health:[about 1 p.]. Available from: https://www. wqad.com/article/news/health/your-health/your-healthmcam/526-cda2622a-69a3-4962-885a-a2185e9507d6 81. National Institute On Drug Abuse. A Novel Opioid Receptor Antagonist For Treating Abuse and Overdose. [Internet]. United States: HHS TAGGS; [updated 2021; cited 2021 February 15]. Available from: https://taggs.hhs.gov/Detail/AwardDetail?arg_ AwardNum=R01DA048417&arg_ProgOfficeCode=114 82. National Institute of Health. Funded Projects. [Internet]. United States: NIH HEAL Initiative; [updated 2020 March 18; cited 2021 February 15]. Available from: https://www. heal.nih.gov/funding/awarded 83. Sansom W. Long-lasting treatment drug could slow the opioid crisis. UT Health San Antonio Newsroom [Internet]. 2020 March 9 [cited 2021 15 February]; News Releases:[about 1 p.]. Available from: https://news.uthscsa. edu/long-lasting-treatment-drug-could-slow-the-opioidcrisis/ 84. Jimenez VM, Castaneda G, France CP. Methocinnamox (MCAM) reverses and prevents fentanyl-induced ventilatory depression in rats. J Pharmacol Exp Ther. 2021; 377:29-38. 85. Minervini V, Disney A, Husbands SM, France CP. Methocinnamox (MCAM) antagonizes the behavioral suppressant effects of morphine without impairing delayed matching-to-sample accuracy in rhesus monkeys. Psychopharmacology (Berl). 2020; 237:3057–3065. 86. Gerak LR, Maguire DR, Woods JH, Husbands SM, Disney A, France CP. Reversal and prevention of the respiratorydepressant effects of heroin by the novel μ-opioid receptor antagonist methocinnamox in rhesus monkeys. J Pharmacol Exp Ther. 2019; 368:229–236.

76. Eippert F, Bingel U, Schoell ED, Yacubian J, Klinger R, Lorenz J, et al. Activation of the opioidergic descending pain control system underlies placebo analgesia. Neuron. 2009; 63:533–543.

87. Peckham EM, Barkley LM, Divin MF, Cicero TJ, Traynor JR. Comparison of the antinociceptive effect of acute morphine in female and male Sprague-Dawley rats using the long-lasting mu-antagonist methocinnamox. Brain Res. 2005; 1058:137–147.

77. Wachman EM, Saia K, Miller M, Valle E, Shrestha H, Carter G, et al. Naltrexone treatment for pregnant women with opioid use disorder compared with matched buprenorphine control subjects. Clin Ther. 2019; 41:1681– 1689.

88. Gerak LR, Minervini V, Latham E, Ghodrati S, Lillis KV, Wooden J, et al. Methocinnamox produces long-lasting antagonism of the behavioral effects of µ-opioid receptor agonists but not prolonged precipitated withdrawal in rats. J Pharmacol Exp Ther. 2019; 371:507–516.

78. Tran TH, Griffin BL, Stone RH, Vest KM, Todd TJ. Methadone, buprenorphine, and naltrexone for the treatment of opioid use disorder in pregnant women. Pharmacotherapy. 2017; 37:824–839.

89. Vonvoigtlander PF, Lewis RA. Analgesic and mechanistic evaluation of spiradoline, a potent kappa opioid. J Pharmacol Exp Ther. 1988; 246:259–262.

79. Krupitsky E, Nunes EV, Ling W, Illeperuma A, Gastfriend DR, Silverman BL. Injectable extended-release naltrexone for opioid dependence: A double-blind, placebo-controlled, multicentre randomised trial. Lancet. 2011; 377:1506– 1513.

90. Wadenberg M-LG. A review of the properties of spiradoline: A potent and selective kappa-opioid receptor agonist. CNS Drug Rev. 2003; 9:187–198. 91. PubChem. Compound Summary Methocinnamox. [Internet]. United States: National Library of Medicine, [updated 2021 May 15; cited 2021 May 18]. Available from: https://pubchem.ncbi.nlm.nih.gov/ compound/46877713

215

Scholarly Research In Progress • Vol. 5, November 2021

Treatment of Pediatric Lisfranc Injuries: A Systematic Review and Introduction of a Novel Treatment Algorithm Samuel Paek1† and Grant D. Hogue2,3 ¹Geisinger Commonwealth School of Medicine, Scranton, PA 18509 ²Harvard Medical School, Boston, MA 02115 ³Orthopedic Department at Boston Children’s Hospital, Boston, MA 02115 † Doctor of Medicine Program Correspondence: spaek@som.geisinger.edu

Abstract Background: Pediatric Lisfranc injuries (PLI) are rare injuries that have few studies published about their occurrence and treatment in pediatric populations. Due to this lack of information, the diagnostic criteria and surgical or non-surgical methods for treatment have not been clearly established within the pediatric orthopedic literature. The objective of this study was to review the published literature related to treatment options for PLI in order to develop a concise stepwise treatment algorithm for pediatric patients presenting with Lisfranc injuries. Methods: A systematic literature review was conducted using PubMed to find studies reporting treatment of PLI with followup of long-term outcomes. Data collection involved the number of feet treated, primary diagnostic method, treatment methods employed, and reported post-treatment complications. Results: Ten articles reporting on 114 Lisfranc injuries treated were found. In summation, 42% (49/115) of feet were treated with open reduction internal fixation (ORIF) using Kirschner wires (K-wires) and/or screws, 29% (33/115) required no reduction, 17% (20/115) using cast immobilization, 5% (6/115) using closed reduction, 3% (4/115) using suture button fixation, and 3% (3/115) with percutaneous fixation. Conclusion: There were two main limitations to this study. First, there are few published studies with longitudinal outcomes of PLI treatment. Second, some case series did not disclose which procedure a patient with post-treatment complications underwent. Therefore, an overall analysis of success and failure rates with associated complications of each procedure could not be conducted. In conclusion, we found that a stepwise approach to evaluating conservative and surgical treatment options based on the presentation of the PLI should be utilized to optimize long-term outcomes.

Introduction Pediatric Lisfranc injuries (PLI) are an uncommon injury of the tarsometatarsal (TMT) joint complex that primarily affects the first and second metatarsals (MT) and their connection to the medial and middle cuneiforms but can also have effects across the entirety of the TMT joint. The instability of this joint complex can be attributed to the absence of a stabilizing ligament between the first and second MTs. The mechanism of injury for PLI is commonly caused by axial loading in a plantarflexed foot position due to disruption of the Lisfranc ligament’s connection between the proximal base of the second MT to the medial cuneiform (1). Due to the rarity of published literature on this injury in pediatric cases, evidence-based treatment guidelines

216

for PLI are limited. This contrasts with adult Lisfranc injuries, which have standardized treatment protocols and less variability in their management due to the closure of physeal growth plates. This is an important area of concern when considering surgical versus non-surgical options for Salter-Harris type fractures because exposure of the physes or joint surface during operative management may induce premature physeal closure and/or fusion as a result (7, 18). Furthermore, the intraarticular nature of PLI increases the risk of developing posttraumatic degenerative arthritic changes in the TMT joint complex (19). Bone deformities of the midfoot region can develop during remodeling from physeal involvement which can create further complications for the patient (7, 18). There are many treatment options with varying outcomes, but no consensus on the most optimal management protocol for this injury (1, 2, 4–7, 8–11, 13). Nonetheless, the majority of studies included in this systematic review have achieved successful post-operative outcomes using surgical intervention to achieve reduction and fixation in treating PLI (2, 4–7, 8–11, 13). The purpose of this study is to define standardized diagnostic criteria and novel treatment guidelines based on the classification of injury to pair the best management protocols with optimal long-term outcomes for future patients receiving treatment for PLI.

Methods Search strategy A systematic review was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) research methods and reporting guidelines (3). A digital search of the online medical literature database MEDLINE (PubMed) was done between January 4, 2021, and January 18, 2021. The search strategy included the following keyword terms: “pediatric” OR “adolescent” OR “child” AND “lisfranc” OR “tarsometatarsal.” All potential studies were stored to Zotero (zotero.org), an open-source software program used for bibliographic citation management, to facilitate evaluation of the studies used. Study selection The published literature selected for this study involves therapeutic measures for PLI. Studies involving adult-only cases were excluded. Studies that included a mixture of pediatric and adult cases, but that did not distinguish data extracted from pediatric and adult patients were excluded. Case reports without appropriate post-operative follow-up were excluded. Studies involving cadaveric specimens were excluded. Studies published in languages other than English were excluded.

Treatment of Pediatric Lisfranc Injuries: A Systematic Review and Introduction of a Novel Treatment Algorithm

Figure 1. Flow chart of literature search identifying articles screened, assessed, included, and excluded

Data extraction Ten studies met criteria for an in-depth review. These articles were examined for the number of patients/feet included in the study, diagnostic indicators for PLI including imaging and physical exam findings, the number of patients undergoing either surgical or non-surgical therapy, instrumentation used (if applicable), and their respective outcomes. Each outcome was taken into consideration when determining the novel treatment guidelines. A PLI was determined resolved if there was complete resolution of pain and restored function of the foot. A PLI was determined to have a failure of resolution if there were posttreatment complications involving either pain, instrumentation malfunction (if applicable), or impaired function. Possible biases were considered when extracting data from case series studies and limitations were noted. Some case series studies did not differentiate which treatment option a patient had undergone even if they had a resultant complication; thus, the sum of complications for each specific treatment option was unable to be obtained from the case series studies.

Results There were 290 articles identified via a PubMed search. Duplicate articles identified through other sources and articles written in languages other than English were excluded. The remaining total articles were screened based on their titles, leaving 46 total full-text articles to be assessed for eligibility

using the previously defined criteria. Papers were excluded if they only included studies with adult patients, did not have post-operative follow-up, involved treatments for non-PLI, or included cadaveric studies. This left a total of 10 full text studies (2, 4–11, 13) to be used in qualitative synthesis for this systematic review. Figure 1 details this search and selection process for narrowing the initial 290 searched articles to the 10 full-text studies used in this systematic review. There were five PLI included from case reports and 109 feet from case series studies that summated to a total amount of ‘114’ feet included in the data from the 10 studies. One foot from a case report (7) was treated with both closed reduction and ORIF, bringing the total number of treatment procedures performed on feet to 115. Treatment methods were as follows: 43% (49/115) open reduction internal fixation (ORIF) with K-wires and/or screws, 29% (33/115) did not require reduction, 17% (20/115) with cast immobilization, 5% (6/115) closed reduction, 3% (4/115) suture button non-rigid fixation, and 3% (3/115) with percutaneous fixation. A summary of diagnostic imaging, treatment methods, and post-treatment complications are summarized for case reports in Table 1 and case series studies in Table 2. Buoncristiani (8) saw eight patients with a mean age of 6.6 years with an age range of 3 to 13 years old at the time of evaluation. The symptomatic presentation included tenderness over the TMT joint complex and mild foot edema in all patients; six of them also had ecchymosis over the dorsal midfoot. 217

Treatment of Pediatric Lisfranc Injuries: A Systematic Review and Introduction of a Novel Treatment Algorithm

Table 1. Summary of diagnostic tools, therapies performed, and outcomes from single case report studies.

Table 2. Summary of diagnostic tools, therapies performed, and outcomes from case series studies.

218

Treatment of Pediatric Lisfranc Injuries: A Systematic Review and Introduction of a Novel Treatment Algorithm

Each patient had initial radiographs that were taken, which revealed fractures of metatarsals and cuneiforms, cuneiform avulsion, or negative radiograph interpretations. All eight patients were treated non-operatively with a short leg walking cast and followed at 3- and 6-week intervals. Immobilization ranged from 3 to 7 weeks and was discontinued after the patients had resolution of pain at the TMT joint complex and were able to ambulate without complications. Final radiographs were taken at their last follow-up, which averaged 32 months. Seven out of eight patients treated using the SLWC had complete resolution of their foot pain at rest and during physical activity. Radiographs confirmed healing of the injury and revealed maintenance of the TMT joint complex. These seven patients had a midfoot functional rating (MFR) of 100. One out of the eight patients continued to have “nonlimiting” midfoot pain and edema after 5 minutes of physical activity but absent at rest; no pain medication management required; and a MFR of 87. This patient’s physical examination of the TMT joint complex was benign, and tenderness was not reproduced with palpation or pronation-abduction stress. The radiograph revealed a healed cuneiform fracture, but also had residual narrowing of joint space consistent with posttraumatic degenerative joint disease in the TMT joint complex. Veijola (9) saw seven patients with a mean age of 14.7 years with an age range of 13 to 16 years old at the time of evaluation. Symptomatic presentation of six patients included edema of the foot and one patient with erythema of the foot. All patients had an initial radiograph taken, which revealed Lisfranc dislocations in all seven patients, and three patients had a computerized tomography (CT) image taken prior to beginning operative treatment. All seven patients underwent ORIF using cannulated screws and K-wires within 24 hours to 7 days after the initial injury. Patients had the respective foot immobilized for an average of 5 weeks and hardware removal at a range of 3 to 7 weeks and an average of 5 weeks after the operation. Post-operative radiographs graded anatomic reduction as “excellent” in six patients and “good” in one patient. However, radiographs taken during outpatient follow-up of two patients with “excellent” grade postoperative anatomic reduction were graded as having “slight” degenerative changes. With regards to functional outcomes, three patients reduced their baseline activity level, and two patients were unable to participate in the same activities as prior to their injury. A possible limiting factor of this study is that the two patient cases with “slight” degenerative changes did not have their post-operative cast immobilization times recorded. Hill (10) conducted a retrospective review of 56 patient cases treated for PLI from 2003 to 2014 with an average age of 14.2; the age range was not provided. 59% (33/56) of patients did not require anatomic reduction to treat their injury; this cohort consisted of 52% (17/33) ligamentous Lisfranc injuries and 48% (16/33) Lisfranc fracture injuries. 9% (5/56) of patients were treated by closed anatomic reduction which included a short leg cast and/or air boot cast; this cohort consisted of 100% (5/5) Lisfranc fracture injuries. 32% (18/56) of patients required ORIF treatment to correct their injuries; this cohort consisted of 100% (18/18) Lisfranc fracture injuries. Of the patients with Lisfranc fractures, 26% (10/39) had physeal involvement. From this cohort, three patients

underwent ORIF, five patients were managed with closed reduction and castings, and two patients did not require reduction. It is also worth noting that from the patients with Lisfranc fractures, 51% (20/39) of them were treated non-operatively. All patients were followed during a range of 11 to 30 weeks with an average of 19.5 weeks to evaluate their recovery and ability to weight bear. The average time to weight bear as tolerated differed significantly between patients who incurred fractures, 14.5 weeks, compared to those who received conservative management, 6.5 weeks. Two specific patients incurred posttraumatic changes due to the severity of damage to their Lisfranc joint complex. One patient with a complete displacement of the first MT Salter-Harris fracture had physeal arrest at age 12 and the other with a severely displaced Lisfranc joint complex had unresolved pain with weight-bearing due to a broken pin that remained in his cuboid. Patients with closed physes were more likely to be selected for operative management compared to patients with open physes. Cheow (11) conducted a retrospective study of eight patients with a mean age of 13.6 years that were treated between 2009 to 2014 and had post-operative follow-up for greater than 12 months, with a mean duration of 3.8 years. Symptomatic presentation of all eight patients included tenderness and edema of the feet examined: neurovascularly intact. All patients were treated within 7 days of their injury onset. Three out of eight patients underwent closed reduction, with two patients undergoing screw fixation and one patient undergoing K-wire fixation. Five out of eight patients underwent ORIF, with three patients undergoing screw fixation and two patients undergoing K-wire fixation. The average time of hardware removal was 3 to 5 months for screw fixation and 4 to 5 weeks for K-wire fixation. The average time to return to full weight-bearing was 3 months in all patients. Post-operative complications were observed in four patients. One out of those four patients had an inadequately maintained anatomic reduction after undergoing ORIF using K-wires, despite being non-weight bearing before hardware removal, and continued to have mild intermittent foot pain. Another of those four patients revealed an appropriately maintained closed screw reduction during his 6-year follow-up examination but had screw breakage that was left in situ which was the suspected cause of his daily moderate foot pain. Screw breakage has been attributed to premature weight bearing on the injured foot before hardware removal took place that was also observed in another patient; however, this patient did not report midfoot pain during follow-up evaluations. These two out of four patients with screw breakages observed in follow-up imaging were reported to have initiated weight bearing before removal of implants. The last of those four patients with postoperative complications had closed screw fixation with loss of anatomic reduction. This may be attributed to the failure of the screw to reach across the Lisfranc joint in this patient that was seen in her post-operative radiograph. Subsequent imaging done after 6 months revealed an intermetatarsal distance between the first and second MT of 3.49 mm. Appropriate distance between the first and second MT should be less than 3 mm and was specifically measured to have a median distance of 1.0 mm for a child of this patient’s age (13 years old) (12).

219

Treatment of Pediatric Lisfranc Injuries: A Systematic Review and Introduction of a Novel Treatment Algorithm

A limiting factor of this study is the lack of information detailing time to partially weight bear and compliance with weightbearing before hardware removal as this may have caused some of the postoperative complications seen. The presence or absence of posttraumatic degenerative changes was also not discussed in the interpretations of radiographs taken at each patient’s follow-up examination. Kushare (13) conducted a retrospective study of 30 patients with an average age of 13.6 years, with a range of 8 to 17 years) and an average follow-up time of 36 weeks. Symptomatic presentation of all patients included tenderness and edema of the midfoot, and ecchymosis in 13 cases. All patients were screened using radiographs, but only five patients did not have further imaging. Twenty-five patients received CT or magnetic resonance imaging (MRI) to confirm or further examine their injuries; two cases required MRI due to negative radiographs. Sixty-one percent (19/30) of patients were managed operatively with screw fixation being used in 14 cases, interosseous suture-button technique fixation being used in three cases, and K-wire fixation being used in two cases. Hardware removal occurred on an average of 28.5 days with a range of 6 to 65 days. Surgical indications included >2 mm displacement or diagnoses of avulsion fractures on radiographs, CT, or MRI. Post-operative follow-up examinations averaged 36 weeks with a range of 12 to 176 weeks. Patients with the suture-button fixation did not require additional surgery for hardware removal. 31% (11/30) patients were managed conservatively for non-displaced injuries using a short leg cast (8/11) or CAM boot (3/11)

Figure 2. Step-wise diagnostic and treatment algorithm for PLI

220

and given instructions to limit weight-bearing. Indications for conservative management were non-displaced Lisfranc injuries. The only post-treatment complication reported was a patient that developed was an Achilles tendon contracture and was managed with physical therapy. There were no cases that reported broken hardware or implants left in situ. There were no obvious differences in the outcome of treatment if it took place less than 4 weeks of injury onset. Additionally, there were no significant differences in the demographics, clinical presentation, or treatment outcomes between the patients treated surgically and conservatively. A possible limitation of this study is the absence of details outlining the medical rationale for the selected method of treatments in patients even after accounting for the lack of significant differences in presentation. The summary of all results with the number of feet treated, diagnostic imaging used, treatment methods, and posttreatment complications are outlined in Tables 1 and 2. 42% (49/115) of feet were treated with ORIF using K-wires and/ or screws, 29% (33/115) required no reduction, 17% (20/115) using cast immobilization, 5% (6/115) using closed reduction, 3% (4/115) using suture button fixation, and 3% (3/115) with percutaneous fixation using K-wires and/or screws.

Discussion Currently, there is no literature with an agreed upon consensus for the best method to treat PLI, due to a shortage of papers discussing the diagnosis, treatment options, and long-term outcomes (1). This is in part due to the rarity of its presentation

Treatment of Pediatric Lisfranc Injuries: A Systematic Review and Introduction of a Novel Treatment Algorithm

in orthopedic settings and difficulty of diagnosis, as it is often missed (1, 10, 13). Differing opinions on treatment management are employed at different orthopedic centers and may further contribute to the lack of consensus (8-11, 13). However, synthesizing the information seen across the included studies has allowed us to generate a stepwise diagnostic and treatment algorithm for PLI that is outlined in Figure 2. Understanding the risks and benefits associated with initiating surgical intervention as the primary treatment modality should be carefully considered due to the increased risk for adverse longterm outcomes such as posttraumatic osteoarthritis associated with intraarticular Lisfranc injuries (7, 19). Nonetheless, the primary aim of treatment should be to maintain an anatomic reduction of less than 2 mm to preserve the best long-term functional outcome for the patient. When including PLI as a possible differential diagnosis, two points of reference that must be included are physical exam findings and weight-bearing radiographs (18). PLI involving the TMT joint complex typically manifests with midfoot pain, swelling, decreased ability to weight bear, and plantar ecchymosis on physical exam (20, 21). Initial weight-bearing radiographs should be taken of both the injured and non-injured foot to have a baseline comparison. If radiographs reveal a distance greater than 2.0 mm between the first and second MTs or between the medial cuneiform and proximal base of second MT, loss of alignment between the second MT and the middle cuneiform, Fleck sign indicating avulsion within the Lisfranc joint complex, diastasis and/or instability of the Lisfranc joint complex may be present (1, 13, 23). If weightbearing radiographs are inconclusive, advanced imaging is the next appropriate step using CT first and if necessary, a subsequent MRI. CT is chosen first because of its superior ability to detect non-displaced fractures, avulsion fractures, and minimal osseous subluxations (24). A systematic review that surveyed diagnostic imaging for Lisfranc injuries revealed that CTs diagnosed up to 60% more MT fractures, double the number of tarsal fractures and joint malalignments compared to plain radiographs (24). If both non-weight-bearing radiographs and CT are inconclusive, then an MRI is used for its accuracy in detecting ligamentous injuries. After diagnosis, conservative treatment should be considered if the PLI reveals displacement of less than 2.0 mm and maintains a stable Lisfranc joint complex (12). Kushare (13) found that there was no significant difference between clinical presentations and outcomes of patients with PLI treated surgically and conservatively. Positive outcomes were reported in patients who had received delayed treatment up to 4 weeks from injury onset, compared to those who received treatment shortly afterward. From this cohort, post-treatment surveys revealed that positive outcomes were consistent with patients receiving delayed treatment up to 4 weeks from injury onset. Therefore, our algorithm suggests initial conservative management with follow-up radiographs to evaluate the degree of displacement as a viable option by immobilization and limited weight bearing using a short leg walking cast or controlled ankle motion boot for 6 weeks (6, 8, 13). Length of immobilization may vary between patients depending on adherence to weight-bearing guidelines provided by the surgeon, which may range from 11

to 14 weeks for surgically treated PLI and up to 7 weeks for conservatively treated PLI. Surgical intervention has been used for PLI as a primary treatment modality for PLI but given our study’s findings we believe it should be used as the primary treatment only if the Lisfranc joint is unstable due to disruption of the Lisfranc ligament, has displacement greater than 2.0 mm, or if the patient has closed physes (1, 7, 8, 10). Surgeons should avoid large subperiosteal dissections and unnecessarily exposing open physes, because this may induce premature physeal closure and joint fusion as a result, leading to posttraumatic osteoarthritis and impaired function (7, 10, 25, 26). A surgical alternative to rigid hardware fixation using the suture-button technique has emerged with excellent results in adults and pediatric patients (2, 7, 27). Tzatzairis (2) and Kushare (13) published a case report and case series study, respectively, documenting the successful treatment of a PLI using the suture button technique, also known as the TightRopeTM Syndesmosis Device, to maintain non-rigid fixation across the Lisfranc joint. An adult case series study specifically focusing on suture-button fixation revealed exceptional outcomes (27). Advantages with this technique include increased healing that was shown through earlier time to full weight-bearing without requiring a cast at 6 weeks post-operatively and lack of a second procedure for hardware removal. Earlier weight-bearing and mobilization may be attributed to enhanced healing from the non-rigid fixation; all patients were reported returning to full weight-bearing status without requiring a cast by 6 weeks (27). A cadaveric study using the suture-button technique across the Lisfranc joint found evidence of similar stability as screw fixation (22), reaffirming the original goal of maintained anatomic reduction. This has potentially more promising long-term outcomes and decreased risks than rigid fixation with screws such as the risk for screw breakage, prolonged decreased range of motion affecting healing, cartilage damage, and a second procedure requiring screw removal (14). Additional studies are required to increase the generalizable efficacy of these results in treating PLI before a conclusion can be made.

Conclusion The results and analysis of this systematic literature review are subject to limitations of retrospective review, varying and sometimes short follow-up periods (months compared to years), small sample size, non-randomized selection, and inconsistent patient adherence to prescribed weight limiting status. To truly determine the efficacy of each treatment modality available, a randomized controlled study is necessary with an appropriately sized sample pool to assess the viability of each option. The goal of this study is to create guidelines for the treatment of PLI using the currently available literature based on the published data from many orthopedic centers. A stepwise treatment algorithm was created using the evidence cited from this study is available for surgeons to use in Figure 2 when considering both non-surgical and surgical options. We believe that initial treatment should include conservative options with appropriate follow-up to assess the healing of the PLI before surgical treatment is initiated.

221

Treatment of Pediatric Lisfranc Injuries: A Systematic Review and Introduction of a Novel Treatment Algorithm

Disclosures None declared.

References 1.

Halai M, Jamal B, Rea P, Qureshi M, Pillai A. Acute fractures of the pediatric foot and ankle. World J Pediatr. 2015;11(1):14–20.

Tzatzairis T, Firth G, Parker L. Adolescent Lisfranc injury treated with TightRopeTM: A case report and review of literature. World J Orthop. 2019;10(2):115–22.

Liberati A, Altman DG, Tetzlaff J, Mulrow C, Gøtzsche PC, Ioannidis JPA, et al. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate healthcare interventions: explanation and elaboration. BMJ. 2009;339(jul21 1):b2700.

Della Valle CJ, Su E, Nihal A, Rosenberg ZS, Trepman E: Acute disruption of the tarsometatarsal (Lisfranc’s) joints in a ballet dancer. J Dance Med Sci 2000;4:128-131. Carrasco A, Perez C, Caminero F. Lisfranc injury in a 14-year-old patient. A case report and literature review. Arch Clin Exp Surg. 2019;(0):1. Kriz P, Rafferty J, Evangelista P, Van Valkenburg S, DiGiovanni C. Stress fracture of the second metatarsal and sprain of lisfranc joint in a pre-professional ballet dancer. J Dance Med Sci. 2015;19(2):80–5. Lesko G, Altman K, Hogue G. Midfoot degenerative arthritis and partial fusion after pediatric Lisfranc fracture-dislocation. J Am Acad Orthop Surg Glob Res Rev. 2018;2(3):e004.

Buoncristiani AM, Manos RE, Mills WJ. Plantar-flexion tarsometatarsal joint injuries in children. J Pediatr Orthop. 2001;21(3):324–7.

Veijola K, Laine H-J, Pajulo O. Lisfranc injury in adolescents. Eur J Pediatr Surg. 2013;23(4):297–303.

10. Hill JF, Heyworth BE, Lierhaus A, Kocher MS, Mahan ST. Lisfranc injuries in children and adolescents. J Pediatr Orthop B. 2017;26(2):159–63. 11. Cheow X, Lam KY. Midterm functional outcomes in operatively treated adolescent Lisfranc injuries. J Pediatr Orthop B. 2018;27(5):435–42. 12. Knijnenberg LM, Dingemans SA, Terra MP, Struijs PAA, Schep NWL, Schepers T. Radiographic anatomy of the pediatric Lisfranc joint. J Pediatr Orthop. 2018;38(10):510– 3. 13. Kushare I, Wunderlich N, Elabd A, Attia E. Pediatric and adolescent Lisfranc injuries - Presentation, treatment and outcomes. Foot (Edinb). 2021;46(101737):101737. 14. Bansal A, Carlson DA, Owen JR, Cheatham SA, Williams DSB 3rd, Kates SL, et al. Ligamentous Lisfranc injury: A biomechanical comparison of dorsal plate fixation and transarticular screws: A biomechanical comparison of dorsal plate fixation and transarticular screws. J Orthop Trauma. 2019;33(7):e270–5.

222

15. Mayr JM, Seebacher U, Lawrenz K, Pesendorfer P, Berghold A, Baradaran S. Bunk beds--a still underestimated risk for accidents in childhood? Eur J Pediatr. 2000;159(6):440–3. 16. Wiley JJ. The mechanism of tarso-metatarsal joint injuries. J Bone Joint Surg Br. 1971;53-B(3):474–82. 17. Rosendahl K, Strouse PJ. Sports injury of the pediatric musculoskeletal system. Radiol Med. 2016;121(5):431–41. 18. Denning JR. Complications of pediatric foot and ankle fractures. Orthop Clin North Am. 2017;48(1):59–70. 19. Thomas AC, Hubbard-Turner T, Wikstrom EA, PalmieriSmith RM. Epidemiology of posttraumatic osteoarthritis. J Athl Train. 2017;52(6):491–6. 20. Pommering TL, Kluchurosky L, Hall SL. Ankle and foot injuries in pediatric and adult athletes. Prim Care. 2005;32(1):133–61. 21. Ross G, Cronin R, Hauzenblas J, Juliano P. Plantar ecchymosis sign: a clinical aid to diagnosis of occult Lisfranc tarsometatarsal injuries. J Orthop Trauma. 1996;10(2):119– 22. 22. Panchbhavi VK, Vallurupalli S, Yang J, Andersen CR. Screw fixation compared with suture-button fixation of isolated Lisfranc ligament injuries. J Bone Joint Surg Am. 2009;91(5):1143–8. 23. Grewal US, Onubogu K, Southgate C, Dhinsa BS. Lisfranc injury: A review and simplified treatment algorithm. Foot (Edinb). 2020;45(101719):101719. 24. Sripanich Y, Weinberg MW, Krähenbühl N, Rungprai C, Mills MK, Saltzman CL, et al. Imaging in Lisfranc injury: a systematic literature review. Skeletal Radiol. 2020;49(1):31–53. 25. Dubois-Ferrière V, Lübbeke A, Chowdhary A, Stern R, Dominguez D, Assal M. Clinical outcomes and development of symptomatic osteoarthritis 2 to 24 years after surgical treatment of tarsometatarsal joint complex injuries. J Bone Joint Surg Am. 2016;98(9):713–20. 26. Marín-Peña OR, Viloria Recio F, Sanz Gómez T, Larrainzar Garijo R. Fourteen years follow up after Lisfranc fracturedislocation: functional and radiological results. Injury. 2012;43 Suppl 2:S79-82. 27. Brin YS, Nyska M, Kish B. Lisfranc injury repair with the TightRopeTM device: A short-term case series. Foot Ankle Int. 2010;31(7):624–7.

Scholarly Research In Progress • Vol. 5, November 2021

The Relationship Between Treatment Center Services and Number of Opioid-related Deaths in the United States Before and After a Declaration of a National Opioid Crisis Brittany N. Davis1*‡, Courtney L. Hatton1*‡, Mahamed A. Jama1*‡, and Nidha S. Samdani1*‡ ¹Geisinger Commonwealth School of Medicine, Scranton, PA 18509 *Master of Biomedical Sciences Program ‡ Authors contributed equally Correspondence: mjama@som.geisinger.edu

Abstract Background: Opioid-related deaths are a national problem that has increased over the past two decades. Multiple policy interventions have been enacted to decrease opioid misuse and expand treatment. The Comprehensive Addiction and Recovery Act (CARA) was passed in July 2016, just before declaring the opioid epidemic a national emergency in 2017. CARA was enacted to combat the opioid epidemic by providing more funding yearly for items including but not limited to prevention, treatment, and opioid overdose reversal. Methods: To evaluate the impact of these policy changes, we carried out secondary data analysis for the period 2011–2019 using the Centers for Disease Control’s Wideranging Online Data for Epidemiologic Research and National Survey of Substance Abuse Treatment Services databases. We hypothesized that increased funding in the form of services offered by opioid treatment facilities was associated with a decreased opioid-related death rate at the state level. Research variables included: a comparison of the 50 states across the 2011–2019 timeframe as an interval, the number of opioid treatment centers per 100,000 inhabitants, the percentage of government funding for facilities per state, percentage of opioid treatment facilities which offer free/low-income services and the opioid death rate per 100,000 inhabitants. We also assessed differences in low-income access to opioid treatment services by comparing Medicaid expansion states versus nonMedicaid expansion states. Results: While both the number of treatment facilities per state and opioid death rates nearly doubled during this time, there was little to no association between them (R2 ranging from 0.094–0.188 for years 2013–2019). Additionally, our research suggests that while state-level differences in opioid use disorder treatment facility characteristics related to access to care, they were only weakly associated with opioid-related deaths. However, Medicaid expansion states had higher heroinspecific overdose death rates from 2014 to 2017 (p-values: 0.0007, 0.0017, 0.0358, and 0.0370). Conclusion: This analysis may be used in the planning of subsequent actions against the national opioid epidemic and invites further inquiry into the impact of state Medicaid expansion on drug-specific opioid usage and morbidity.

Introduction The rising opioid epidemic has become a significant public health crisis over the last decade and a half. This trend has been observed with a peak in opioid prescription dispensing in 2011,

a substantial increase in opioid-related deaths since 2000, and an increase in the point prevalence of opioid use disorder (OUD) (1). There have been multiple public policy interventions to address the epidemic at the state and federal levels. These include restrictions on prescribing opioids, along with law enforcement crackdowns on negligent prescribing (2, 3). Additionally, there has been the widespread implementation of prescription drug monitoring programs (PDMPs) in 48 states by 2014, increasing from 11 in 2007 (4). A recent retrospective study on this topic found an association between state implementation of PDMPs and decreased opioid-related death rates from 1999 to 2013. The strength of this effect was moderated by drug program characteristics (4). Other policy changes have targeted treatment and funding. The passage and implementation of the Affordable Care Act in the United States (U.S.) in 2010 has been a driving force in expanding access and treatment for substance use disorder (SUD) and OUD (5). It expanded insurance coverage to millions of Americans through Medicaid, established parity in requirements for SUD treatment for patients covered under Medicaid, reduced preauthorization requirements for OUD treatment, and added coverage for initial screenings of SUD (5–7). In addition, state Medicaid expansion has theoretically increased access to OUD treatment. However, the adoption of Medicaid expansion has been staggered and incomplete (8). In addition, the opioid overdose epidemic was declared a U.S. public health emergency in 2017, just after the Comprehensive Addiction and Recovery Act (CARA) was signed into law in 2016, both of which provided further funding for OUD treatment, prevention, and opioid-overdose reversal (9). The opioid epidemic has disproportionately affected people living in poverty and created an economic crisis for people with OUD (9). There has been limited research on how increased funding from CARA and Medicaid expansion for OUD treatment facilities and their services have affected the opioid death rate at the state level. Nevertheless, once a person is addicted to opioids, treatment facilities are the primary option for reducing OUD and preventing opioid-related deaths (8). Combating the increase in OUD prevalence is multifaceted and warrants interventions at both the prescriber and user-level (9). To investigate the state-level impact of CARA on OUD treatment facilities and opioid-related deaths, we analyzed publicly available data from the Centers for Disease Control’s (CDC) Wide-ranging Online Data for Epidemiologic Research (WONDER) and National Survey of Substance Abuse 223

The Relationship Between Treatment Center Services and Number of Opioid-related Deaths

Treatment Services (N-SSATS) databases from 2011 to 2019 for all 50 states (10, 11). We compared differences in opioidrelated death rates based on state Medicaid expansion status and other low-income access markers such as the percentage of opioid treatment centers receiving government funding and the portion of facilities per state offering free services or accepting Medicaid payment. Hypothesis one was that increased funding in the form of services offered by treatment facilities was associated with a decreased opioid-related death rate at the state level. Hypothesis two was that those states with a higher rate of facilities providing low-income access would have decreased opioid-related deaths compared to states with less access.

Methods

Table 1. N-SSATS Facility Respondents. Values displayed represent the total number of eligible facilities within the N-SSATS database from 2011 to 2019, including the total number of respondents, nonrespondents, and percentage of non-respondents. The average percent of non-respondents is 7.73%.

Procedures We used data from the CDC WONDER database and the National Survey Substance Abuse Treatment Services (N-SSATS) (10, 11). The CDC WONDER Database data set is collected and updated yearly; the most recent data available currently is from 2019. The N-SSATS data set is compiled by the Substance Abuse and Mental Health Services Administration (SAMHSA) from a voluntary annual census of substance treatment centers throughout the U.S. It includes facilityspecific responses to treatment utilization, funding types, type and the number of services offered, and other facility characteristics. We aggregated facility data to the state level. We used data from 2011 through 2019 in a zero-correlation study design with archival data. We chose this approach to determine a relationship between increased funding of substance treatment facilities and opioid-related deaths. The N-SSATS database was used to identify treatment facilities that provided OUD treatment services specifically. The percentage of those facilities in each state received government funding (state, federal, or local) and which percentage offered free services or accepted Medicaid. We compared this state-level data to the CDC WONDER database to determine if there is a correlation with opioid-related deaths per 100,000 inhabitants. Opioid-related deaths and the number of opioid treatment facilities were adjusted to reflect a per capita number. This enabled us to compare the numbers corresponding to each state and control differences in state population size. The N-SSATS survey had a variable non-response rate from facilities from year to year (Table 1) (11). For our purposes, missing data were omitted. We believe this study should be relatively resistant to outlier data because, as seen above, the missing data accounts for ~8% of the data. Typically research with secondary data can have up to 10% missing data without affecting it statistically (12). We chose and adjusted variables to reduce confounding variables within secondary data-analysis limitations. For determining if a facility is an opioid treatment center, we counted all facilities that provided one or more of the following opioid addiction treatment medications: methadone, naltrexone including extended-release, buprenorphine with or without naloxone, buprenorphine implants or extended-release pharmaceuticals, lofexidine, and/or clonidine. For determining opioid deaths, ICD10 codes X40-44, X60-64, X85, Y10-14 were used. This included all intentional suicides, unintentional overdoses, undetermined overdoses, and homicide via 224

Table 2. Government spending on substance abuse treatment and prevention as reported by the SAMHSA from 2011 to 2019 (16-26). The values represent the total amount spent per year, the total amount allocated for treatment, the total amount allocated for prevention, and the total amount allocated for opioid-specific treatment and prevention. The total amount of government spending per year increased nearly two-fold by 2019, with funds spent on opioid-related resources. Dollar amounts are in thousands.

overdose. The categories were further sub-coded into T40 codes; T40.1 identifies overdoses from heroin. Prescription opioid deaths were designated by T40.2-4 and T40.6; these represent other opioids, methadone, other synthetic narcotics, and other and unspecified narcotics. Because opioid addiction can be to prescription opiates or heroin, we combined the totals, except where explicitly stated otherwise. The N-SSATS survey includes a question that asks whether the facility receives funding from any federal, state, or local government; the answers were either yes or no. The total number of facilities answering yes to this question was used to determine the percentage of facilities in each state that receive government funding. However, the dollar amounts could not be determined. Similarly, the survey asks if the facility provides all services for free and if they accept Medicaid as payment. The number of facilities answering yes to either of these questions was totaled to calculate the percentage of facilities in each state that provide free services. Data analysis A correlation was used to examine the association between the number of opioid-related deaths to the number of opioid treatment facilities, the percentage of facilities in each state that receives government funding, and the percentage of facilities

The Relationship Between Treatment Center Services and Number of Opioid-related Deaths

in each state that provide free services or accept Medicaid. In addition, we used multiple unpaired t-tests to assess significant differences between death rates in Medicaid expanded states vs. non-expanded states. We generated all statistics and figures in GraphPad Prism version 9.1.0 (216) for macOS (13), Microsoft Excel version 16.50(21061301) for macOS (14), and JMP version 16.0.0 (512340) for macOS (15).

Results Between 2011 and 2019, the number of opioid treatment facilities doubled (4,280 to 8,437 facilities), and the number of national opioid deaths nearly tripled (23,617 to 64,404 deaths). State-specific trends in these metrics were visualized by heatmap distributions (Figures 1 and 2). Qualitatively, the increase in opioid overdose deaths was most significant in northeastern and Appalachian states (Figure 2), while the most significant increase in treatment facilities was in the Southwest. Changes in the mean distribution from year to year were substantial in both the opioid deaths and the number of facilities (p-values = <0.0001), showing an increasing number nearly every year except for 2018, being slightly lower than 2017. Comparison of the amount of money allocated to SAMHSA toward substance abuse treatment and prevention (Table 2) during this timeframe and the opioid deaths depicts a similar increasing trend (Figure 3) (16–26).

Linear regression of per capita opioid treatment centers and opioid deaths showed no to very little correlation with R² values ranging from 0.018 to 0.192 in each year (Figure 4), and the data model was significant from 2013 to 2019. Other variables that could have relevance in reducing opioid deaths, such as the percentage of facilities in each state that receive government funding either from the federal, state, or local government, or the percentage of facilities in each state that provide free care to all or accept Medicaid, were examined. Each state's range of facilities receiving government funding has remained consistent through the years, ~25–80%, respectively. The range of facilities that provide free care or accept Medicaid has increased from ~7–90% to ~30–90%. This suggests that more facilities provide greater access to their services, but neither of these variables had strong correlations with the opioid death rates (Figures 5 and 6). There is no or very low correlation between opioid overdose death rates and the percentage of facilities receiving government funding or the percentage of facilities providing free services or accepting Medicaid. The government funding data was not significant, but the percentage of facilities providing free care or accepting Medicaid was significant in every year except 2011 and 2012 (P-values: 2011=0.3561, 2012=0.2464, 2013=0.0015, 2014=0.0037, 2015=0.0028, 2016=0.0101, 2017=0.0016, 2018=0.0040, and 2019=0.0150).

Figure 1. State-level heatmap of the OUD treatment facilities per 100,000 people from 2011 to 2019. The total number of OUD treatment facilities increased two-fold from 4,280 to 8,437.

225

The Relationship Between Treatment Center Services and Number of Opioid-related Deaths

Figure 2. State-level heatmap of the opioid overdose deaths per 100,000 people from 2011 to 2019. The total number of opioid overdose deaths increased nearly three-fold from 23,617 to 64,404.

Figure 3. Comparison of SAMHSA allocated funding on substance abuse services and treatment versus opioid overdose deaths between 2011 and 2019. During this time, both the total amount of allocated funds for substance abuse services and opioid overdose deaths increased.

226

The Relationship Between Treatment Center Services and Number of Opioid-related Deaths

The other factor that we hypothesized would affect the opioid overdose death rates was whether a state had expanded Medicaid or not. A comparison of the opioid death rates was made between non-Medicaid expanded states and expanded states. We determined a state was considered expanded if the expanded Medicaid plan took effect at any point in the year. Using publicly available data, we created lists of expanded states for each year (27). After comparing opioid overdose death rates between Medicaid expanded states and non-expanded states, we determined that there was no significant mean difference between the two groups for any of the years (unpaired t-test, p-value>0.05) (Figure 7A). Medicaid expansion comparisons used in Figures 4, 5, and 6 were not statistically significant but were added for visualization. We then evaluated pharmaceutical opioid overdose deaths and heroin overdose deaths separately, comparing Medicaid expanded and unexpanded states (Figures 7B and 7C). Like the total opioid overdose deaths, pharmaceutical opioid overdose deaths showed no statistically significant difference in means. However, in the heroin overdose death rates, the Medicaid expanded states had higher means, and they were statistically significant in 2014, 2015, 2016, and 2017 (p-values as appears: 0.0007, 0.0017, 0.0358, and 0.0370). In the Medicaid expanded states, though, there seemed to be a leveling of deaths after 2017. Additionally, the heroin deaths in the Medicaid expanded states seemed to be getting lower. Again, it was not statistically significant and could be due to multiple missing data points that could not be determined if it was non-reported or reported zero.

Finally, we qualitatively visualized the effect of state-level Medicaid expansion using the percent change in opioid overdose deaths per 100,000 people (Figure 8). States with Medicaid expansion appear to have decreased percent changes in recent years, suggesting that though Medicaid expanded states still have more deaths, deaths decreased in those states.

Discussion The results of our research did not support our first hypothesis that a higher number of facilities providing low-income access would decrease opioid-related deaths versus states which do not. Overall, our data showed that the number of opioid-related deaths has almost tripled from 2011 to 2019. The results did not indicate any pattern between opioid-related deaths and the number of facilities, nor increased funding and opioid-related deaths. However, the results showed an increase in opioidrelated deaths and an increase in the number of facilities per capita. Although our data showed an increase in the number of treatment centers and opioid-related death rates, there seems to be a question about the efficiency of these current treatment centers. The ACA aims to address the opioid crisis through funding, ultimately leading to increasing treatment facilities (28). However, increased opioid treatment centers do not correlate to the increased treatment availability for all users. Despite the increase in facilities, only 20% of opioid users receive treatment (28).

Figure 4. Correlation of facility crude rate numbers and opioid overdose death crude rates. Medicaid expansion was added for visualization. Linear regression of per capita opioid treatment centers and opioid deaths with R² values ranging from 0.018 to 0.192 each year. 227

The Relationship Between Treatment Center Services and Number of Opioid-related Deaths

Figure 5. Linear regression comparisons by year for opioid deaths per 100,000 people and percent of facilities in each state receiving government funding. Data for 2014 was unavailable. There is no or very low correlation between opioid overdose death rates and the percentage of facilities receiving government funding (R2 values <0.03 and p-values>0.2).

Figure 6. Linear regression comparisons by year for opioid deaths per 100,000 people and the percent of facilities in each state providing free or accepting Medicaid services. The range of facilities that provide free care or accept Medicaid has increased from ~7–90% to ~30–90% 228

The Relationship Between Treatment Center Services and Number of Opioid-related Deaths

Figure 7. (A) Yearly comparison of opioid overdose deaths per 100,000 people in Medicaid expanded states versus non-expanded states. There was no significant mean difference between the two groups for any years (unpaired t-test, p-value>0.05). (B) Pharmaceutical opioid overdose death rates comparison between Medicaid expanded states and non-expanded states. Pharmaceutical opioid overdose deaths showed no statistically significant difference in means (unpaired t-test, p-value>0.05). (C) Heroin overdose death rates comparison between Medicaid expanded states and non-expanded states. The Medicaid expanded states had higher means, statistically significant in 2014, 2015, 2016, and 2017 (p-values as appears: 0.0007, 0.0017, 0.0358, and 0.0370).

229

The Relationship Between Treatment Center Services and Number of Opioid-related Deaths

Figure 8. Percent change in opioid overdose death rates per 100,000 people in each state with the indication of when each state expanded Medicaid.

In addition, there are barriers for opioid users seeking treatment, such as being placed on waitlists for programs that can last days, up to months. This can lead to continuous unguided drug usage behavior, which can contribute to opioidrelated deaths. A method needs to be implemented to optimize these treatment centers for those currently on the waiting lists to address facilities’ efficiency. There are current efforts of a technology-assisted interim dosing regimen (29). This includes creating methods to monitor patients through computerized buprenorphine dispensing and contacting patients through phone communication and random callbacks. This method has demonstrated efficacy in reducing opioid and drug use behavior and decreasing psychiatric distress during the waitlist process. Another issue is opioid users need a holistic, tailored treatment based on their history and opioid usage behavior. Vermont is one of the states that has developed a “hub-and-spoke system” which allows patients to start more intensive and custom treatment through in-person counseling, urine toxicology testing, and other medical management (29). In addition, Vermont aims to provide effective treatment by implementing a brief screening questionnaire to help match the most appropriate treatment to each patient (28). This information helps us better understand why increased treatment centers do not decrease opioid-related deaths. There needs to be an internal evaluation of current treatment facilities by addressing their waitlist patients, creating personalized treatments, and providing a screening method to match patients to the 230

appropriate treatment type. In addition to facility efficacy, another factor affecting the correlations between facility number and opioid-related deaths could be due to how we calculated whether a facility was an opioid treatment facility or not. We classified a facility as an opioid treatment facility if their response to the N-SSATS survey indicated that they provided one or more of the following opioid addiction treatment medications: methadone, naltrexone including extended-release, buprenorphine with or without naloxone, buprenorphine implants or extended-release pharmaceuticals, lofexidine, and/or clonidine. This was the best method of interrupting the data. Still, it likely included facilities that supply the pharmaceuticals but do not provide other necessary treatment services, such as mental and social health services. Additionally, because the N-SSATS survey is only given to physical facilities, it excludes the 37,000+ providers that SAMHSA provides waivers and certifications to, allowing them to prescribe opioid treatment medications (24). The opioidrelated deaths are not declining, but perhaps if there were a better way of determining the number of patients seen in facilities and private providers, there would be a better way to determine a correlation between treatment and opioidrelated deaths. The lack of significance between Medicaid expanded and unexpanded states and the fact that expanded states have higher means could be explained by larger states having

The Relationship Between Treatment Center Services and Number of Opioid-related Deaths

expanded sooner than smaller states. Even though the data was adjusted based on population size, states with larger populations would have a more significant load on the Medicaid system, so the effect of Medicaid expansion might be less visible initially. Additionally, since we only looked at whether a state expanded Medicaid and not how it expanded, there are likely variables within the difference expansions that could act as confounding factors. Likewise, we did not investigate the coverage available before the expansion and what coverage is available in non-expanded states. Since each state individually decides on its Medicaid expansion coverage, it is not a uniform variable and therefore makes the evaluation of its effectiveness difficult. An additional factor that we did not assess in this study but is a topic for future research is the use of opioid antagonists. SAMHSA has been promoting the training and use of opioid antagonists, such as naloxone, which can save a person’s life who is actively overdosing on opioids. Since these medications can make the difference between life and death, their use and implementation across the country could be a better predictor of opioid death rates. In a 2018 Morbidity and Mortality weekly report, the CDC investigated the emergency use of naloxone and found that its use rose by 75.1% from 2012 to 2016 (30). The authors used the emergency medical service records to collect this data. As a future addition to this study and our study, the use of naloxone should be investigated to determine its correlation with opioid-related deaths.

Kennedy-Hendricks A, Richey M, McGinty EE, Stuart EA, Barry CL, Webster DW. Opioid overdose deaths and Florida’s crackdown on pill mills. Am J Public Health. 2016;106(2):291–7. DOI: 10.2105/AJPH.2015.302953.

Schieber LZ, Guy GP, Seth P, Young R, Mattson CL, Mikosz CA, et al. Trends and patterns of geographic variation in opioid prescribing practices by state, United States, 2006-2017. JAMA Netw Open. 2019;2(3):e190665. DOI: 10.1001/jamanetworkopen.2019.0665.

Patrick SW, Fry CE, Jones TF, Buntin MB. Implementation of prescription drug monitoring programs associated with reductions in opioid-related death rates. Health Aff. 2016;35(7):1324–32. DOI: https://doi.org/10.1377/ hlthaff.2015.1496.

Abraham AJ, Andrews CM, Grogan CM, Pollack HA, D’Aunno T, Humphreys KN, et al. The Affordable Care Act transformation of substance use disorder treatment. Am J Public Health. 2017;107(1):31–2. DOI: 10.2105/ AJPH.2016.303558.

Abraham AJ, Smith BT, Andrews CM, Bersamira CS, Grogan CM, Pollack HA, et al. Changes in state technical assistance priorities and block grant funds for addiction after ACA implementation. Am J Public Health. 2019;109(6):885–91. DOI: 10.2105/AJPH.2019.305052.

Andrews CM, Pollack HA, Abraham AJ, Grogan CM, Bersamira CS, D’Aunno T, et al. Medicaid coverage in substance use disorder treatment after the affordable care act. J Subst Abuse Treat. 2019;102(Apr):1–7. DOI: 10.1016/j.jsat.2019.04.002.

Grogan CM, Andrews C, Abraham A, Humphreys K, Pollack HA, Smith BT, et al. Survey highlights differences in Medicaid coverage for substance use treatment and opioid use disorder medications. Health Aff. 2016;35(12):22892296. DOI: 10.1377/hlthaff.2016.0623.

Hagemeier NE. Introduction to the opioid epidemic: the economic burden on the healthcare system and impact on quality of life. Am J Manag Care. 2018 May;24(10):S200–6. PMID: 29851449.

Conclusion This study evaluated multiple ways in which increased access to OUD treatment services influenced opioid-related death rates. There are state-level differences in OUD treatment facility characteristics associated with opioid-related mortality, but these are only weakly correlated with opioid-related death rates. Medicaid expansion was hypothesized to impact decreasing opioid-related deaths because it would allow for more access to treatment. Still, facility limitations are likely the bottleneck in treatment, not financial access. Further research is needed to draw definitive epidemiological conclusions on the impact of CARA and Medicaid expansion. However, the exploratory analysis carried out in this study can help inform future investigation and public policy aimed at addressing the opioid epidemic.

Acknowledgments The authors would like to thank Brian Piper, PhD, MS, and Reema Persad-Chem, PhD, MPH, for their invaluable guidance and assistance throughout the research process.

Disclosures The authors have no conflicts of interest to disclose.

References 1.

Schieber LZ, Guy GP, Seth P, Losby JL. Variation in adult outpatient opioid prescription dispensing by age and sex — United States, 2008–2018. MMWR Morb Mortal Wkly Rep. 2020;69(11):298–302. DOI: http://dx.doi.org/10.15585/ mmwr.mm6911a5.

10. Centers for Disease Control and Prevention, National Center for Health Statistics. WONDER: Underlying cause of death 1999-2019 [Internet]. 2020 [cited 2021 Jan 29]. Available from: http://wonder.cdc.gov/ucd-icd10.html 11. U.S. Department of Health and Human Services, Substance Abuse and Mental Health Services Administration. National survey of substance abuse treatment services (N-SSATS) 2019-2011 [Internet]. 2019 [cited 2021 Jan 29]. Available from: https://www.datafiles.samhsa.gov/ study-series/national-survey-substance-abuse-treatmentservices-n-ssats-nid13519. 12. Dong Y, Peng CY. Principled missing data methods for researchers. Springerplus. 2013 May 14;2(1):222. doi: 10.1186/2193-1801-2-222. PMID: 23853744; PMCID: PMC3701793. 13. GraphPad Prism version 8.4.3 (471) for macOS. GraphPad Software. La Jolla, California USA. 2020 Jun 10. Available from: www.graphpad.com

231

The Relationship Between Treatment Center Services and Number of Opioid-related Deaths

14. Microsoft Excel version 16.50(21061301) for macOS. Microsoft Corporation. Redmond, Washington. 15. JMP version 16.0.0 (512340) for macOS. SAS Institute. Cary, North Carolina. Available from: https://www.jmp.com 16. Department of Health and Human Services. Fiscal year 2011, Substance abuse and mental health services administration, justification for estimates for appropriations committees. [Internet] Place of publication: Substance Abuse and Mental Health Services Administration. 2011. Available from: https://www.samhsa. gov/sites/default/files/samhsa_fy11cj.pdf 17. Department of Health and Human Services. Fiscal year 2012, Substance abuse and mental health services administration, justification for estimates for appropriations committees. [Internet] Place of publication: Substance Abuse and Mental Health Services Administration. 2012. Available from: https://www.samhsa. gov/sites/default/files/samhsa-fy12cj.pdf 18. Department of Health and Human Services. Fiscal year 2013, Substance abuse and mental health services administration, justification for estimates for appropriations committees. [Internet] Place of publication: Substance Abuse and Mental Health Services Administration. 2013. Available from: https://www.samhsa. gov/sites/default/files/samhsafy2013cj.pdf 19. Department of Health and Human Services. Fiscal year 2014, Substance abuse and mental health services administration, justification for estimates for appropriations committees. [Internet] Place of publication: Substance Abuse and Mental Health Services Administration. 2014. Available from: https://www.samhsa. gov/sites/default/files/samhsa-fy2014-budget_0.pdf 20. Department of Health and Human Services. Fiscal year 2015, Substance abuse and mental health services administration, justification for estimates for appropriations committees. [Internet] Place of publication: Substance Abuse and Mental Health Services Administration. 2015. Available from: https://www.samhsa. gov/sites/default/files/fy-2015-budget-cj.pdf 21. Department of Health and Human Services. Fiscal year 2016, Substance abuse and mental health services administration, justification for estimates for appropriations committees. [Internet] Place of publication: Substance Abuse and Mental Health Services Administration. 2016. Available from: https://www.samhsa. gov/sites/default/files/samhsa-fy2016-congressionaljustification_2.pdf 22. Department of Health and Human Services. Fiscal year 2017, Substance abuse and mental health services administration, justification for estimates for appropriations committees. [Internet] Place of publication: Substance Abuse and Mental Health Services Administration. 2017. Available from: https://www.samhsa. gov/sites/default/files/samhsa-fy-2017-congressionaljustification.pdf

232

23. Department of Health and Human Services. Fiscal year 2018, Substance abuse and mental health services administration, justification for estimates for appropriations committees. [Internet] Place of publication: Substance Abuse and Mental Health Services Administration. 2018. Available from: https://www.samhsa. gov/sites/default/files/samhsa-fy-2018-congressionaljustification.pdf 24. Department of Health and Human Services. Fiscal year 2019, Substance abuse and mental health services administration, justification for estimates for appropriations committees. [Internet] Place of publication: Substance Abuse and Mental Health Services Administration. 2019. Available from: https://www.samhsa. gov/sites/default/files/sites/default/files/samhsa-fy-2019congressional-justification.pdf 25. Department of Health and Human Services. Fiscal year 2020, Substance abuse and mental health services administration, justification for estimates for appropriations committees. [Internet] Place of publication: Substance Abuse and Mental Health Services Administration. 2020. Available from: https://www.samhsa. gov/sites/default/files/about_us/budget/samhsa_fy_2020_ cj_submission_031919_508_final.pdf 26. Department of Health and Human Services. Fiscal year 2021, Substance abuse and mental health services administration, justification for estimates for appropriations committees. [Internet] Place of publication: Substance Abuse and Mental Health Services Administration. 2021. Available from: https://www.samhsa. gov/sites/default/files/about_us/budget/fy-2021-samhsacj.pdf 27. Where the states stand on Medicaid expansion [Internet]. Place of publication: Advisory Board. 2020 Available from: https://www.advisory.com/en/daily-briefing/resources/ primers/medicaidmap 28. Blevins CE, Rawat N, Stein MD. Gaps in the substance use disorder treatment referral process: Provider perceptions. J Addict Med. 2018;12(4):273–7. Available from: https:// www.ncbi.nlm.nih.gov/pmc/articles/PMC6066414/ 29. Sigmon SC. Innovations in efforts to expand treatment for opioid use disorder. Prev Med. 2019;128(105818):105818. Available from: https://www.ncbi.nlm.nih.gov/pmc/articles/ PMC7081076/#R2 30. Cash RE, Kinsman J, Crowe RP, Rivard MK, Faul M, Panchal AR. Naloxone Administration Frequency During Emergency Medical Service Events — United States, 2012– 2016. MMWR Morb Mortal Wkly Rep 2018;67:850–853. DOI: http://dx.doi.org/10.15585/mmwr.mm6731a2

Scholarly Research In Progress • Vol. 5, November 2021

Racial Differences in Insurance Type between Diabetes Mellitus Type 2 Patients in the United States Mannaa I. Mannaa1* ¹Geisinger Commonwealth School of Medicine, Scranton, PA 18509 *Master of Biomedical Sciences Program Correspondence: mannaaonly@yahoo.com

Abstract Background: The Centers for Disease Control and Prevention (CDC) categorizes diabetes as an epidemic in the United States (U.S.). It further stands among health disparities with a disproportionate distribution among minorities. The expansion of Medicaid access under the Affordable Care Act (ACA), which serves the medical needs of disabled, chronically ill, and low-income residents, has not been enough to bridge the gap of health and health care disparities. This is in part due to Medicaid expansion being optional for states, leaving 2.2 million residents in the coverage gap. Past analysis of the annual National Hospital Ambulatory Medical Care Survey (NHAMCS) has elucidated the burden of diabetes on the health care system. It has also gauged how expansion of Medicaid has armed an appreciable number of working-age residents with health insurance. NHAMCS stands in a position to potentially highlight continued inequities in health care despite Medicaid expansion. Methods: Differences in insurance type between diabetes mellitus Type 2 patients based on race in the U.S. was analyzed in the 2018 NHAMCS. Results: Medicaid coverage as well as the lack of insurance coverage was found to be significantly higher among Black (42.3%, 9.9% respectively) compared to White patients (30.2%, 7.8% respectively) visiting Emergency Departments (EDs). Conclusion: Analyses of NHAMCSs stand as checkpoints as the U.S. strives in progression toward health and health care equities. They have revealed the burden of diabetes on the health care system. Further, they have illuminated success of Medicaid expansion where it has been adopted. Finally, disproportionate reliance upon Medicaid between Black as compared to White patients, along with unequal distribution of lack of health insurance between these races, have been made apparent.

Introduction The World Health Organization (WHO) identified 1.6 million deaths directly caused by diabetes in 2016 (1). While WHO classifies the disease as an epidemic globally, the CDC also confirms diabetes as an epidemic in the U.S. (1, 2). One in 10 Americans suffer from the disease while 1 in 3 are considered pre-diabetic and, for the most part, are unaware (2). Type 2 diabetes, once considered adult-onset and appearing over 45 years of age, is now appearing among American youths between 10 and 19. (2, 3) This deadly disease, a major cause of serious health issues including blindness, kidney failure, heart attack, stroke, and lower limb amputation, was counted by WHO as among the top 10 causes of death globally as of December 2020 (4). It also stands among health disparities with a disproportionate distribution among minorities in the U.S. In

2017 to 2018, the CDC found the prevalence of diabetes to be 11.7% for adult non-Hispanic blacks versus 7.5% for adult non-Hispanic Whites (1). Contributors to this disproportionate distribution including obesity and differences in neighborhood, psychosocial, socioeconomic, and behavioral factors (5). The NHAMCS is a national probability sample survey of outpatient visits to hospitals, emergency departments (EDs), and hospital-based ambulatory surgery centers (6). Analysis of NHAMCS from 1993 to 2005 revealed an average of 380,000 visits/year for hypoglycemia, a serious complication among diabetic patients (7). Demographic disparities were found for hypoglycemia-related ED visits by age, sex, race, ethnicity, and region. Thirty-four per 1,000 visits were by diabetic patients (7). Diabetes patient visit rates were significantly higher for female, Black, and Hispanic patients as compared to male, White, and non-Hispanic patients (7). In 2015, the epidemic of diabetes accounted for 92 per 1,000 ED visits by patients 45 and older, about 24% of all ED visits (8). Despite having this data available, health care utilization and cost for diabetes in the U.S. may be underestimated, and by extension, the severity of the disease underrealized. The American Diabetes Association was found to have utilized NAMCS and NHAMCS surveys to estimate utilization and costs in a 2012 report after quantifying diabetes-related ED visits using providers’ diagnosis codes and medication lists to quantify those visits (9). However, evaluations of NHAMCS data from 2006 and 2010 determined that solely using provider’ diagnosis codes and medication lists to identify diabetic patient visits would fail to identify approximately one quarter of outpatient visits by patients with diabetes (9). Regarding insurance coverage for diabetes patients, 24% of the diabetes ED visits for those age 45 to 64 had Medicare as the primary expected payment source, versus 14% of the ED visits for patients age 45 to 64 without diabetes (8). There was a significant increase from 66.0% to 71.8% in the percentage of working-age adult ED patients who had at least one form of health insurance in the first 2 years following ACA implementation. This was almost entirely due to Medicaid coverage increase, the expansion of which has been associated with improvements in self-reported access to health care and self-reported diabetes management (10, 11). Recent trends reveal the long-time inequities that have plagued African Americans and other minorities in the U.S., keeping many from fair access to income, education, neighborhoods, and health care equal to that of White Americans (12). As a result, minorities have more chronic or serious health conditions and have significantly lower financial resources than White Americans (13). Given these facts and the considerable role that Medicare has among diabetic patients as shown previously, it would be worth exploring other insurance trends among 233

Racial Differences in Insurance Type between Diabetes Mellitus Type 2 Patients in the United States

diabetic ED visits. Regarding diabetes-associated ED visits in the NHAMCS, it was investigated if there was a statistically significant difference in insurance types between diabetes mellitus (DM) Type 2 patients based on race in the U.S. in 2018.

Methods Participants This study was an analysis of data from the 2018 NHAMCS. The sample hospital emergency room visits were from noninstitutional general and short-stay hospitals as well as hospital-associated ambulatory surgery centers, none of which were associated with federal, military or Veterans Administration Hospitals (13). The total number of visits was 20,291. This sample was examined for ED visits associated with diabetes mellitus Type 2 was 1,215 (N=1,215, 67.3% White, 26.5% Black, 6.2% Other). Procedure The annual NHAMCS is based on sample visits from randomly assigned hospitals and outpatient departments. The 2018 NHAMCS analysis was limited to visits associated with patients having the valid diagnosis code for DM Type 2 which was assigned following affirmation to the question “Does the patient have diabetes mellitus Type 2?” Of note, as of 2005, many disease diagnoses, including diabetes, are input into the Patient Record Form of the NAMCS and NHAMCS survey by means of a checkbox (9). The checkbox enables surveyors to indicate a particular diagnosis of the patient regardless of the reason for the ED visit. This yields a far more accurate quantification of ED visits by patients with a particular disease. Previously, ED visit primary diagnosis codes and medication lists were solely used to make a diabetes diagnosis determination. This was problematic as medications could go unreported. Further, the Patient Record

Form only allowed three diagnosis codes, which could also lead to underreporting of diabetes if the patient visit was associated with multiple primary diagnosis codes for that particular ED visit. Therefore, while it was possible that diabetic patients went underreported in surveys previous to 2005, the checkbox increased accuracy of reporting (9). In total, 1,215 patient visits were found, representing 7.9 million encounters, after appropriate weighting. Because NHAMCS datasets are publicly available and de-identified, this study was deemed exempt by the Institutional Review Board at Geisinger Commonwealth School of Medicine. Data analysis The 2018 NHAMCS data was imported into SPSS so that diabetes mellitus Type 2 relevant visits could be captured. As instructed by the NHAMCS documentation, data was first weighted in SPSS for patient visits (“PATWT”) to facilitate national representation of the sample data (14). Next, all visits associated with diabetes mellitus Type 2 patients were aggregated, stratified, and summed by race and expected payment type via IBM SPSS (14). Of note, PAYTYPER uses a hierarchy of 8 payment types/categories to allow visits to be associated with a primary expected source of payment (14). Crosstabulations were then performed by race and payment type for these visits, Table 1. The rationale was to look for any patterns related to health insurance between races. Four of the 9 categories of expected sources of payment in the crosstabulations table were merged into 2 categories to reduce variables. Expected sources of payment “Unknown” and “Blank” were merged into “Unknown” and expected sources of payment “Other” and “No Charge/Charity” were merged into “Other.” The crosstabulations were exported to excel for 2x2 tables for Chi square testing. Statistical differences with p<0.05 were considered significant. This was performed for 3 of the 9 categories which correspond to actual insurance as well as uninsured (Private, Medicaid, Uninsured). Data for these tables were imported to GraphPad Prism for creation of figures.

Results Emergency department visits associated with diabetes mellitus Type 2 patients were stratified by race and primary expected sources of payment. Statistical significance was found for private insurance versus Medicaid between White and Black (p=0.001) and for private insurance versus uninsured between White and Black (p<0.006, Figure 1). There was no difference for private insurance versus Medicaid between races White and Other (p=0.879) or for private insurance versus uninsured between races White and Other (p=0.637).

Discussion

Table 1. Comparison of count and percentage by race of ED visits associated with diabetes mellitus Type 2 patients stratified by source of payment from the National Hospital Ambulatory Care Survey public use files, 2018.

234

Statistical significance was found for private insurance versus Medicaid and for private insurance versus uninsured between races White and Black. Therefore, Medicaid coverage as well as the lack of insurance coverage was found to be significantly higher among Black (42.3%, 9.9% respectively) compared to White patients (30.2%, 7.8% respectively) visiting the ED. Chi Square testing was performed for 2x2 tables where primary expected sources of payment of private insurance, Medicaid, and uninsured for White race was tested against the same

Racial Differences in Insurance Type between Diabetes Mellitus Type 2 Patients in the United States

Figure 1. Comparison of insurance types between diabetes mellitus Type 2 patients visiting the ED based on race in the U.S. in 2018, National Hospital Ambulatory Medical Care Survey. Private insurance versus Medicaid between races white and Black (p<0.001) and for private insurance versus uninsured between races white and Black (p<0.006). **p < 0.01.

for Black race and again for Other race. When comparing emergency department visits by White patients with diabetes mellitus Type 2 to Black patients with the disease, there were significantly more visits associated with private insurance among White patients. Additionally, there were significantly more visits associated with Medicaid or a lack of insurance among Black compared to White patients. Analysis of NHAMCS points to the need for heightened efforts to make health insurance and appropriate health care equally accessible to all. NHAMCS has been analyzed on other occasions to gauge growing health crises or to trend cost of care. Surveys spanning 1999 through 2005 were screened for ED visits associated with severe hypoglycemia and revealed about 5.0 million (380,000 per year) related ED visits. (7) The increased number of hypoglycemia visits to the ED did not equate to an increase in rate and was attributed to the increased overall prevalence of diabetes and intensive glucose control through insulin (7, 15, 16). This NHAMCS analysis further found disparities in hypoglycemia related ED visits in age, sex, ethnicity, race, and region (7). Visit rates among the diabetic population for female, Black and Hispanic patients were higher than those for male, White, and non-Hispanic patients (p<0.001) (7). A retrospective study of NHAMCS between 1999 and 2013 reported longitudinal trends in opioid-related ED visits along with resource utilization to gauge stress to emergency care systems. ED encounters increased 170% during that time (17). One-third of the visits arrived by EMS, nearly one-third required imaging studies, and there was a 250% increase in hospital admissions accounted for during that time (17). The results of this data were consistent with other data, showing drastic upward trends of opioid-related ED visits and cost related to care (18). Using the National Survey on Drug Use and Health (NSDUH), cost analysis from a prevalence-based estimation approach confirmed NHAMCS and other data showing that in 2007 abuse of prescription opioids cost workplace productivity, the judiciary system, and health care about $55.7 billion (19).

In addition to the disparities of diabetes and hypoglycemia, analyses of NHAMCS either solely or in conjunction with National Ambulatory Medical Care Survey have also pointed to other medical conditions disproportionately affecting subpopulations. These include findings that asthma-related ED visits disproportionately affect more children than adults, more Blacks than Whites, and more Hispanics than Whites (20). Other findings in 2003–2005 NHAMCS ED visits and NAMCS outpatient visits showed that hepatitis C virus-related visits were more than twice as likely to occur among non-White than White patients and more than three times as likely to occur among Medicaid than non-Medicaid patients (21). Finally, a study of NHAMCS and NAMCS visits between 1999 and 2004 looked at dermatophyte and cutaneous yeast infections and associated high cost of care (22). For tinea capitis, there was an average of 433,690 visits per year and the prevalence among the Black population was 12 times that of the White population (22). Moreover, of all the tinea capitis cases, 85.6% occurred among children less than 15 years old, making Black school-aged children disproportionately impacted (22). Another significant finding was that Medicare covered visits for the conditions that predominantly affected children at that time (tinea capitis, 56.9%; tinea corporis, 34.7%; Candida of the skin and nails, 43.6%) (22). All findings above have had or continue to have a level of public health significance. NHAMCS has had an important role in identifying these conditions. Limitations include the NHAMCS may be subject to selection bias, errors in the medical record, and errors during the data abstraction process. Medicare was not included for statistical significance evaluation as this insurance can become primary for retired persons, and factors including differences in life expectancy between races would potentially confound the data in this category. Worker’s compensation was also not tested because it is considered temporary coverage which can only be obtained because of work-related injuries. Currently, the ACA continues to give states the option to expand Medicaid. This coverage was found to have a significant impact on insurance status of working-age adult ED patients as previously brought out, yet there are still 12 states which have not adopted the expansion (10, 23). Eight of the 12 states not expanding Medicaid are southern states which comprise 92% of the 2.2 million people in the coverage gap (24). These eight have higher Black populations, which are among the most likely to be uninsured compared to other populations (25). If Medicaid expansion were to be adopted by more states in the future, it would be worth reexamining NHAMCS data for a direct impact on visits associated with Medicaid or a lack of insurance among Black compared to White patients.

Acknowledgements Thank you to Brian Piper, PhD, for invaluable feedback throughout the secondary analysis.

Disclosures Mannaa I. Mannaa has no conflicts of interest to report.

235

Racial Differences in Insurance Type between Diabetes Mellitus Type 2 Patients in the United States

References 1.

Diabetes. World Health Organization; 2020 Jun 8. Available from: https://www.who.int/news-room/factsheets/detail/diabetes

Diabetes. Centers for Disease Control and Prevention; 2020 Feb 11. Available from: https://www.cdc.gov/ diabetes/library/features/diabetes-stat-report.html

Diabetes. Centers for Disease Control and Prevention; 2019 May 30. Available from: https://www.cdc.gov/ diabetes/basics/type2.html

Top 10 causes of death. World Health Organization; 2020 Dec 9. Available from: https://www.who.int/news-room/ fact-sheets/detail/the-top-10-causes-of-death

NIH Research Matters. National Institutes of Health; 2018 Jan 9. Available from: https://www.nih.gov/newsevents/nih-research-matters/factors-contributing-higherincidence-diabetes-black-americans

National Center for Health Statistics. Centers for Disease Control and Prevention; 2019 Sep 6. Available from: https://www.cdc.gov/nchs/ahcd/about_ahcd. htm#NHAMCS

Ginde AA, Espinola JA, Camargo CA. Trends and disparities in U.S. emergency department visits for hypoglycemia, 1993-2005. Diabetes Care. 2008; 31(3):511-513. DOI: DOI: 10.2337/dc07-1790

National Center for Health Statistics. Centers for Disease Control and Prevention; 2018 Feb. Available from: https:// www.cdc.gov/nchs/products/databriefs/db301.htm

Asao K, McEwen LN, Lee JM, Herman WH. Ascertainment of outpatient visits by patients with diabetes: The National Ambulatory Medical Care Survey (NAMCS) and the National Hospital Ambulatory Medical Care Survey (NHAMCS). J Diabetes Complications. 2015;29(5):650-8. DOI: 10.1016/j.jdiacomp.2015.03.019.

10. Orgel GS, Weston RA, Ziebell C, Brown LH. Emergency department patient payer status after implementation of the Affordable Care Act: A nationwide analysis using NHAMCS data. Am J Emerg Med. 2019;37(9):1729-1733. DOI: 10.1016/j.ajem.2018.12.031 11. Lee J, Callaghan T, Ory M, Zhao H, Bolin JN. The impact of Medicaid expansion on diabetes management. Diabetes Care 2020;43(5):1094-1101. DOI 10.2337/dc19-1173 12. David, BA. Discrimination: a social determinant of health inequities. 2020 Feb 25. In: Health Affairs. Health Affairs Blog. Available from: https://www.healthaffairs.org/ do/10.1377/hblog20200220.518458/full/ 13. African Americans and Medicare. National Committee to Preserve Social Security and & Medicare; 2021 Jan 28. Available from: https://www.ncpssm.org/documents/ medicare-policy-papers/medicare-medicaid-importantafrican-americans/ 14. NHAMCS Micro-Data File Documentation. Centers for Disease Control and Prevention. 2018. Available from: https://www.cdc.gov/nchs/ahcd/datasets_documentation_ related.htm

236

15. Johnson ES, Koepsell TD, Reiber G, Stergachis A, Platt E. Increasing incidence of serious hypoglycemia in insulin users. J Clin Epidemiol. 2002;5(3):253-259. DOI: 10.1016/ S0895-4356(01)00479-6 16. Bulsara MK, Holman CD, Davis EA, Jones TW. The impact of a decade of changing treatment on rates of severe hypoglycemia in a population-based cohort of children with type 1 diabetes. Diabetes Care. 2004 10;27(10):2293-8. 17. Salzman M, Jones CW, Rafeq R, Gaughan J, Haroz R. Epidemiology of opioid-related visits in the UD emergency departments, 1999-2013: a retrospective study from the National Hospital Ambulatory Medical Care Survey. Am J Emerg Med. 2020;38(1):23-27. DOI: 10.1016/j. ajem.2019.03.052 18. Hsu DJ, McCarthy EP, Stevens JP, Mukamal KJ. Hospitalizations, costs and outcomes associated with heroin and prescription opioid overdoses in the United States 2001–12. Addiction. 2017;112: 1558– 1564. DOI: 10.1111/add.13795 19. Birnbaum HG, White AG, Schiller M, Waldman T, Cleveland JM, Roland CL. Societal costs of prescription opioid abuse, dependence, and misuse in the United States. Pain Med. 2011: 12(4):657–667. DOI: 10.1111/j.15264637.2011.01075.x 20. Qin XT, Zahran HS, Malilay J. Asthma-related emergency department (ED) visits and post-ED visit hospital and critical care admissions, National Hospital Ambulatory Medical Care Survey, 2010-2015. J Asthma. 2020; Early access available from: http://apps.webofknowledge. com.ezaccess.libraries.psu.edu/InboundService.

Scholarly Research In Progress • Vol. 5, November 2021

Pronounced Declines in Licit Fentanyl Utilization and Changes in Prescribing and Reimbursement Practices in the United States, 2010-2019 Raymond A. Stemrich1†, Jordan V. Weber1†, Kenneth L. McCall2, and Brian J. Piper1,3 ¹Geisinger Commonwealth School of Medicine, Scranton, PA 18509 ²University of New England, Portland, ME 04103 ³Center for Pharmacy Innovation and Outcomes, Forty Fort, PA 18704 † Doctor of Medicine Program Correspondence: rstemrich@som.geisinger.edu

Abstract

Introduction

Background: Fentanyl and its derivatives are highly utilized opioid-based analgesics provided to patients in various formulations. The evolving opioid crisis over the past decade has increased pressure on the health care industry to monitor opioid production. Different policies and procedures implemented to contain the crisis from a controlled substance standpoint have changed the utilization patterns of some opioids. A prior report identified 3-fold differences between states in fentanyl use. This study explored fentanyl distribution patterns from 2010 and 2019.

The opioid crisis continues to plague healthcare in the United States (U.S.) despite legislation enacted by the government, scrutiny placed on pharmaceutical companies, and limitations restricting opioid prescribers. Fentanyl, a synthetic μ-receptor agonist, is employed as an analgesic administered in several formulations including injectable and transdermal patches (1). Fentanyl and its derivatives alfentanil, sufentanil, and remifentanil are most often utilized as intraoperative analgesics and in the management of chronic pain, particularly related to cancer and chemotherapies (1, 2). The potency of fentanyl, the diversity of administrative routes, and the low cost of its production have resulted in it becoming a frequently misused substance (2, 3).

Methods: The amount of fentanyl base distributed from 2010 to 2019 was obtained from the Drug Enforcement Administration. Sufentanil, alfentanil, and remifentanil were also analyzed from 2010 to 2017, the most recent year reported. Prescriptions, units, and reimbursement for 2010 and 2019 were obtained from Medicaid and fentanyl prescriber specialty from Medicare Part D. Results: There was a 65.5% decrease in the milligrams of fentanyl per person distributed when correcting for population. From a regional perspective, Ohio had the greatest decrease (-79.3%) while Mississippi saw the smallest (-44.5%). There was a 6.8-fold regional difference in the quantity of fentanyl distributed in 2019 from hospitals (South Dakota = 775.2, Alabama = 113.2 ug/person). The regional difference was also sizeable for pharmacies (6.2-fold, Mississippi = 1,025.3, Washington, D.C. = 165.6). Medicaid reimbursement in 2019 was $165 million for over eight hundred-thousand prescriptions with the majority for generic (99.7%) and injectable (77.6%) formulations. Interventional pain management and anesthesia were overrepresented, and hematology/oncology significantly underrepresented for fentanyl in Medicare. Conclusion: The production and distribution of fentanyl-based substances decreased, although not uniformly, in the United States over the last decade. Additionally, the most prescribed formulations of fentanyl have transitioned away from transdermal, potentially in an effort to regulate its availability. Although impactful, overdose deaths attributed to synthetic opioid deaths continue to increase, highlighting the need for targeted public health interventions beyond the pharmaceutical and medical communities.

Synthetic opioids pose risks beyond dependence. Commonly prescribed in chronic respiratory conditions (e.g., COPD), opioids treat pain, insomnia, and refractory respiratory symptoms (4). Potent derivatives like fentanyl increase risks of severe adverse effects contributing to the increased mortality due to respiratory complications such as pneumonia observed in this population (4). Additionally, opioid use during pregnancy has been linked to teratogenic effects and neonatal abstinence syndrome (NAS) (5). Estimated to have impacted one newborn per hour in 2009 in the U.S., the syndrome causes disruptions in brain development and is associated with cardiac defects, spina bifida, and gastroschisis (5–7). The U.S. has a history of battling illicit substances and prescription drug misuse, but since the early 2000s the opioid crisis has been evolving. The steep increase in drug-related mortalities has been linked to two major factors: the overprescription of opioids and the illegal manufacturing of fentanyl and fentanyl analogues (8–13). Over-prescribing began in the 1990s and accelerated in the 2000s. This led to an increase in opioid dependence among patients and increased the diversion and misuse of prescribed opioids in the street markets. The extent of this crisis gained national attention in 2016 when roughly 11 million people were estimated to have misused prescription opioids (14). The initial focus of combating this growing epidemic was directed at regulating prescription fentanyl, specifically related to the apparent over-prescribing of opioids. Some of these measures included state legislation limiting the amount of

237

Pronounced Declines in Licit Fentanyl Utilization and Changes in Prescribing and Reimbursement Practices

opioids prescribed, utilization of drug monitoring programs, and new prescribing guidelines (14–16). Despite some success, fentanyl-related deaths have continued to increase with a larger proportion related to illicit fentanyl. Fentanyl distribution throughout the country has changed according to reports using similar methods to this study and others using associations between overdose deaths and underlying causative substances (9,11,16–18). One study noted an 18% reduction in licit fentanyl distribution in 2016; however, less is known about more recent use or trends over the last decade. A 3.5-fold regional difference in prescription fentanyl per state was identified between Oregon (1,718 μg/person) and Alaska (418.2 μg/person) (16). Between 2011 and 2017, states in the northeastern U.S. experienced an increase in fentanylrelated overdoses with an associated increase in seizures of diverted substances (9,11,18). New England states reported that 60–90% of overdose deaths were attributed to fentanyl (9,11). In contrast, states west of the Mississippi River reached a plateau in fentanyl deaths during that same period (18). States like Ohio, West Virginia, and Wisconsin reported that only 30–55% of deaths were associated with fentanyl (9,11). Despite these apparent trends, it was concluded that the distribution of fentanyl across the U.S. was significantly variable (18). The objective of this study was to explore the trends in prescription fentanyl and select fentanyl derivatives distribution throughout the U.S. over the last decade (2010–2019); further, to examine prescriber preferences for fentanyl formulations in Medicare during the same period.

Methods Procedures Distributions of fentanyl base and select fentanyl derivatives in the U.S. were obtained from the Drug Enforcement Administration’s (DEA) Automated Reports and Consolidated Ordering System (ARCOS) from 2010 to 2019. As a result of the 1970 Controlled Substance Act, this program mandates that the federal government track the distribution of controlled substances in grams by pharmacies, hospitals, providers, and treatment programs. This database has been used in previous research analyzing trends in controlled substance distribution (16). The last year where the fentanyl derivatives were reported by ARCOS was 2017. The total amounts for the drugs of interest were reported on the annual summary reports in grams from each of the U.S. states and territories. To normalize the data across the different states and years, population data were obtained from the American Community Survey and U.S. Census Bureau. The type of fentanyl or fentanyl citrate formulation and number of prescriptions for each of the formulations was obtained from Medicaid (19, 20). Formulations were categorized by National Drug Codes as generic versus brand and by route of administration. Fentanyl prescribers reporting to Medicare Part D were examined, specifically the number of prescribers in each specialty and the number of claims (including refills) per specialty (21). The methods used in this study using claims were similar to a previous study (22). Procedures were approved as exempt by the IRB of the University of New England.

238

Data analysis The following analyses were completed: (1) total distributed amounts of fentanyl base and select fentanyl derivatives (i.e., alfentanil, remifentanil, and sufentanil) for each state; (2) the percent change of the distributed amount from 2010 to 2019 for each state per person; (3) ratio of highest to lowest in population corrected use per state; (4) a 95% confidence interval with states outside this range interpreted as significantly different from the mean; (5) the total number of prescriptions and total reimbursement for Medicaid in 2010 and 2019; (6) ratio of prescribers in specific specialties compared to the total number of prescribers to the number of claims per specialties in Medicare Part D in 2018. Data analysis and figures were completed using Microsoft Excel and GraphPad Prism, Version 8.4.2. Heat maps were generated with JMP Statistical Software version 15.2.1.

Results From 2010 to 2019, the amount of fentanyl distributed declined from 1,689.9 μg/person to 583.9 μg/person, which was a -65.5% overall decrease across the U.S. (Figure 1). Further examination was completed by business activity. Hospitals showed a -63.4% decrease which was similar (-64.5%) among pharmacies (Figure 2). The heat map shows that all states had a reduced utilization of prescription fentanyl with the greatest overall decrease in Ohio (-79.3%), whereas Mississippi had the smallest decline (-44.4%, Figure 3). These regional differences were further explored by business activity. The states with the largest pharmacy reduction were Ohio (-80.7%), Oregon (-75.7%), and Nevada (-74.4%). In contrast, Idaho (-49.1%), Kansas (-48.9%), and Mississippi (-41.9%) experienced the smallest declines (Figure 4). Among hospitals, Vermont (-84.5%), South Dakota (-84.0%), and Connecticut (-80.5%) experienced the greatest fentanyl

Figure 1. Distribution of fentanyl base and fentanyl derivatives (sufentanil, remifentanil, and alfentanil) in micrograms per person as reported by the United States Drug Enforcement Administration’s Automated Reports and Consolidated Orders System from 2010 to 2019. Values in parentheses indicated percent change from 2010.

Figure 2. Fentanyl base (kilograms) distributed to pharmacies and hospitals across the United States from 2010 to 2019 as reported by the Drug Enforcement Administration’s Automated Reports and Consolidated Orders System.

Pronounced Declines in Licit Fentanyl Utilization and Changes in Prescribing and Reimbursement Practices

Table 1. Medicaid utilization of fentanyl formulations in 2010 and 2019.

reduction. Nevada (-47.5%), Delaware (-43.6%) and Alabama (-31.6%) had the smallest reductions (Figure 5, see also Figures 6–7). Further assessing potential regional differences, pharmacies in Mississippi distributed a significantly greater amount of fentanyl than average while Washington, D.C., distributed significantly less. However, regional differences in 2019 were non-significant (Figure 6). In contrast to the pharmacies, fentanyl distribution to hospitals was significantly greater in South Dakota, Montana, and North Dakota in 2010. Similarly, hospitals in North Dakota and Montana in 2019 were elevated relative to the mean (Figure 7). Alfentanil had a -19.5% decrease relative to 2010 (Figure 1). In contrast, sufentanil had a modest (+10.9%) change while remifentanil had an appreciable (+73.6%) increase. It is important to note that alfentanil, sufentanil, and remifentanil were not reported by ARCOS beyond 2017. Next, Medicaid prescriptions and expenditures were examined. Table 1 shows that the intravenous/intramuscular (IV/IM) formulations accounted for one-third of the approximately 700,000 fentanyl prescriptions in 2010, but three-quarters of those in 2019. Conversely, transdermal formulations fell from two-thirds to less than one-quarter of prescriptions. Tablets became much more common (0.2% to 3.0%), and lozenges doubled (0.2% to 0.5%). Brand name formulations were responsible for 13.8% of prescriptions in 2010, and this declined to 0.3% in 2019. Units of fentanyl from 2010 to 2019 decreased by 30% while enrollment increased by 38.1%. Finally, analyses were completed on fentanyl formulations in Medicare. A ratio of fentanyl (generic and brand name) prescribers to all claims was created with values greater than one indicating overrepresentation and values less than 1 were underrepresentation. Figure 8 shows that interventional pain management (2.85), anesthesiology (2.07), pain management (2.04), and physical medicine and rehabilitation (1.66) were among those specialties significantly overrepresented, while family practice (0.80), internal medicine (0.78), and hematologyoncology (0.64) were underrepresented in their prescriptions of fentanyl.

Figure 3. Percent decrease in distributed fentanyl base (μg/person) from 2010 to 2019 as reported by the United States Drug Enforcement Administration’s Automated Reports and Consolidated Orders System (blue: smallest reduction; red: largest reduction, * p < 0.05 versus state average).

Discussion Fentanyl base experienced a pronounced and consistent year-over-year decline in distribution throughout the U.S. over the last decade (2010–2019) according to the DEA’s ARCOS. This 65.5% decrease is congruent with and expands upon past research which found that the U.S. reduction in fentanyl was significantly greater than that of hydrocodone, morphine, or oxycodone (16). This decline was likely the result of efforts directed at one aspect of the opioid crisis. The over-prescribing of opioids including fentanyl by the medical community and in rare cases the criminal “pill-pushing” behaviors of certain prescribers, were deemed the primary culprits for the increased drug-related mortalities at this time (15). Similarly, alfentanil, a synthetic opioid with one-eighth the potency of fentanyl, also experienced a reduction. However, the extent of these recent

239

Pronounced Declines in Licit Fentanyl Utilization and Changes in Prescribing and Reimbursement Practices

Figure 4. Percent decrease (-40%, green, to -80%, red) in fentanyl base (μg/person) distributed to pharmacies in the U.S. from 2010 to 2019 as reported by the Drug Enforcement Administration’s Automated Reports and Consolidated Orders System.

reductions cannot be fully appreciated, because the DEA stopped reporting analog distribution in 2017 (23). In contrast to these fentanyl analogues, the amount of remifentanil and sufentanil increased during this period but were very modest relative to fentanyl. The rise in these derivatives — especially sufentanil, which is 10-fold more potent than fentanyl — is concerning and if diverted in appreciable quantities, could contribute to the death toll of the crisis (24). From a regional perspective, Ohio generated the greatest decline in fentanyl (-79%) while Mississippi had the smallest decrease (-44.4%). Macroscopically, these two states have many similarities including the breakdown of their insured population and the demographics of the population impacted by opioid overdoses. However, other state socioeconomic differences may influence the percentages of decrease between the two states. The median annual household income for Ohio in 2019

240

Figure 5. Percent decrease (-30, green, to -85%, red) in fentanyl base (μg/person) distributed to hospitals in the United States from 2010 to 2019 as reported by the Drug Enforcement Administration’s Automated Reports and Consolidated Orders System.

was $58,642 while it was $45,692 in Mississippi. Healthcare expenditures per capita in 2014 (the most recent reported data) was $8,712 in Ohio and $7,646 in Mississippi (25). Other notable differences include the states’ total Medicaid and Medicare spending. In 2019, Ohio’s total spending for Medicaid and Medicare respectively was roughly $24 million and $14 million, whereas Mississippi’s was approximately $6 million and $5 million. Furthermore, Mississippi’s percentage of uninsured patients is almost double that of Ohio’s at 6.7% compared to 12.9% (25). The ability to address a state’s opioid crisis is impacted by the risk factors affecting their patients and the vulnerability of their socioeconomic infrastructure; nevertheless, how the specific socioeconomic factors influence the opioid crisis is an area of ongoing research. Microscopically, there are key distinctions between the states that may explain the differences in their abilities to mitigate the opioid epidemic.

Pronounced Declines in Licit Fentanyl Utilization and Changes in Prescribing and Reimbursement Practices

Figure 6. (A) Fentanyl (μg/person) distributed to pharmacies in 2010 as reported by the DEA’s Automated Reports and Consolidated Orders System (green: lowest; red: highest). (B) Fentanyl (μg/person) distributed to pharmacies in 2019 as reported by the DEA’s Automated Reports and Consolidated Orders System. Note that the range is lower relative to that in Figure 4A (green: lowest; red: highest).

Ohio had one of the highest numbers of opioid-related deaths compared to any state at a rate of 32.9 deaths/100K with fentanyl resulting in a death rate of 21.7 deaths/100K (26). To combat these alarming numbers, Ohio’s Opiate Action Team doubled their spending from $10/person in 2017 to $19/ person in 2018 (27). This increased funding created more treatment and recovery programs, established preventative measures, and improved prescribing practices for opioids and pain management (27, 28). These expansions also addressed the illegal street opioids to an extent which could have augmented the decrease (28). Unlike Ohio, Mississippi had the smallest decrease of fentanyl (-44.4%). In 2018, fentanyl accounted for 2.9 deaths/100K and the rate of prescription opioid-related deaths was 1.4 deaths/100K. However, the more alarming number in 2018 was the 76.8 opioid prescriptions for every 100 persons (29). This represents an approximately 40% greater use compared to the national average in 2018, but also represents one of

the lower annual prescribing rates in the state’s recent history (29). To date, it appears that Mississippi is one of a handful of states still attempting to pass or just recently passed any type of legislation to combat fentanyl and other opioid misuse. The Opioid Crisis Response Act of 2018 attempts to reduce the trafficking of fentanyl and other opioids, improve prescribing practices, and increase programs for prevention, treatment, and recovery for those struggling with addiction (30). Parker’s Law is still under review by the full House, but it would impose stricter criminal penalties on individuals trafficking fentanyl, heroin, and other substances, including life in prison if an illegally distributed substance leads to an overdose death (31). The delayed approach to mitigating the crisis may explain the limited decrease in fentanyl distribution and stagnant death rate figures. Opioid-prescribing laws generally showed only modest benefits unless the legislation included fiscal penalties for non-adherence (16). The substantial region inhomogeneity in fentanyl (6.2-fold in pharmacies, 6.8-fold in hospitals) may warrant continued attention to characterize the epidemiological

241

Pronounced Declines in Licit Fentanyl Utilization and Changes in Prescribing and Reimbursement Practices

Figure 7. (A) Fentanyl (μg/person) distributed to hospitals in 2010 as reported by the Drug Enforcement Administration’s Automated Reports and Consolidated Orders System (green: lowest; red: highest). (B) Fentanyl (μg/person) distributed to hospitals in 2019 as reported by the Drug Enforcement Administration’s Automated Reports and Consolidated Orders Systems (green: lowest; red: highest).

differences in nociceptive and non-nociceptive factors responsible for these differences. Ohio’s reduction in fentanyl to pharmacies over the last decade (-80.7%) was almost twice as large as that of Mississippi’s (-41.9%). The effects of evolving policies and legislation attempting to subdue the impact of the opioid crisis may have penetrated further than just the amounts of fentanyl produced and distributed to also impact prescribing preferences. Using Medicaid, the 10 most prescribed fentanyl formulations showed pronounced changes. There was an 44.9% decrease in transdermal patch prescriptions and a 45.2% increase in injectable prescriptions from 2010 until 2019. This apparent change in prescribing preferences is potentially linked to the improved prescribing guidelines encouraged by state and federal laws. The impact in the change of prescribing preferences is further highlighted by the change in reimbursement from 2010 to 2019 which saw a transition away from transdermal and toward injectable formulations.

242

Transdermal patches, compared to other legally produced formulations, are uniquely susceptible to use and misuse. Even after administration, some patches contain 28–84% of the initial dose, which can be drawn out of the reservoir and misused. This type of diversion was documented in nursing homes and assisted living facilities (32, 33). Patches can also come in contact with non-patients accidentally, leading to absorption through the skin resulting in toxicity and death (2, 33–35). For these reasons, there are FDA and manufacturer warnings for health care professionals, FDA public health advisories, product labeling changes, and increased monitoring and disposal practices of patches have been adopted by many facilities to protect their patients, employees, and others in the environment (36–38). Additionally, the IV/IM formulations, relative to the sizable declines observed in their use in Medicaid, may be more preferred. In 2019, 47,511 prescribers from 80 different specialties and industries reported prescribing both generic and brand name

Pronounced Declines in Licit Fentanyl Utilization and Changes in Prescribing and Reimbursement Practices

and Medicare programs provides some insights into patient populations, but further research with the rich information provided by electronic medical records to characterize which pharmacological, or non-pharmacological, agents are substituting for fentanyl in the U.S. is needed.

Figure 8. Ratio of prescribers per specialty to fentanyl prescription claims per specialty for 2018 as reported by Medicare (*chi-square p < 0.00001).

fentanyl to Medicaid. A large percentage of claims (41.6%) were submitted by general practitioners in family practice (FP) and internal medicine. Although these specialties comprised the greatest percentages of prescribers and claims, they were both deemed to be underrepresented in the overall analysis of fentanyl prescribers. In contrast to the general specialties, anesthesiology, pain management (i.e., interventional pain management and general pain management), and physical and rehabilitative medicine were overrepresented. These results were congruent with previous studies that concluded that specific specialties are not solely responsible for the opioid crisis (22, 39). Neurology, a specialty underrepresented in fentanyl prescriptions, had a claim to prescriber ratio of 38:1 while FP had a ratio of 31:1. This is interesting because neurology only comprised 1.3% of prescribers and 1.2% of claims compared to 29.3% and 23.5% for FP. A previous study found that 8 neurologists prescribed more controlled medications than 141 emergency medicine and urgent care prescribers combined, which suggests an interesting pattern within the specialty and a potential area for future investigation (40). The employment of Prescription Drug Monitoring Programs has improved the surveillance and communication among physicians and specialties, but a broad effort addressing the prescribing practices of each discipline might improve the variations seen in opioid prescribing practices (39, 40). While this manuscript employed three complementary databases, no study is without limitations. A concern with ARCOS is reporting by drug weight instead of more standard units like prescriptions as Medicaid does. However, there was an excellent concordance (r = 0.985) between ARCOS and a state prescription drug monitoring program for another opioid (41). Examination of fentanyl use among the Medicaid

Despite the progress, the opioid crisis remains untamed and continues to amass an escalating death count (41). The majority of these deaths are attributed to the illegal drug trades ravaging the streets. Persons with substance abuse disorder want to avoid fentanyl, but the illegal industry utilizes fentanyl in so many ways that it is hidden even to the most seasoned user (42). Most of the successful efforts to mitigate the crisis have targeted the legal production of fentanyl. This may contribute to the shortage of anesthetic drugs across the country during the COVID-19 pandemic. This shortage led the DEA to increase production and imports to treat those patients on ventilators (43, 44). Unethical and illegal practices by one company manufacturing fentanyl, InSys, have resulted in clear consequences including prison sentences, fines, and bankruptcy (45). Future attempts at addressing the opioid crisis must regulate the legal entities of the problem but cannot ignore the uncontrollable nature of the illegal street market. Social programs increasing routes to treatment and recovery, educating users and misusers about the dangers of fentanyl, and providing fentanyl detection methods to street users are all approaches currently under investigation (10, 14).

Acknowledgments Thanks to Iris Johnston for technical assistance. Software used in this research was provided by the National Institute of Environmental Health Sciences (T32ES007060-31A1).

Disclosures BJP is part of an osteoarthritis research team supported by Pfizer and Eli Lilly. He receives research support from the Health Research Services Administration (D34HP31025). The other authors have no conflicts of interest to declare.

References 1.

Schumacher MA, Basbaum AI, Naidu RK. Opioid agonists & antagonists. In: Katzung BG, editor. Basic & Clinical Pharmacology [Internet]. 14th ed. New York, NY: McGraw-Hill Education; 2017 [cited 2020 Oct 24]. Available from: accessmedicine.mhmedical.com/content. aspx?aid=1148437600

Stanley TH. The fentanyl story. J Pain. 2014 Dec;15(12):1215–26.

Green TC, Gilbert M. Counterfeit medications and fentanyl. JAMA Intern Med. 2016;176(10):1555–7.

Vozoris NT, Wang X, Fischer HD, Bell CM, O’Donnell DE, Austin PC, et al. Incident opioid drug use and adverse respiratory outcomes among older adults with COPD. Eur Respir J. 2016 Sep;48(3):683–93.

Sutter MB, Leeman L, Hsi A. Neonatal Opioid Withdrawal Syndrome. Obstet Gynecol Clin North Am. 2014 Jun;41(2):317–34.

243

Pronounced Declines in Licit Fentanyl Utilization and Changes in Prescribing and Reimbursement Practices

Patrick SW, Schumacher RE, Benneyworth BD, et al. Neonatal Abstinence Syndrome and associated health care expenditures: United States, 2000-2009. JAMA. 2012 May 9;307(18):1934–40.

20. State Drug Utilization Data | Medicaid [Internet]. [cited 2020 Apr 27]. Available from: https://www.medicaid.gov/ medicaid/prescription-drugs/state-drug-utilization-data/ index.html

Broussard CS, Rasmussen SA, Reefhuis J, et al. Maternal treatment with opioid analgesics and risk for birth defects. Am J Obstet Gynecol. 2011 Apr;204(4):314.e1-314.e11.

Schifano F, Chiappini S, Corkery JM, Guirguis A. Assessing the 2004-2018 fentanyl misusing issues reported to an international range of adverse repoting systems. Front Pharmacol. 2019;10(46).

21. Medicare Provider Utilization and Payment Data: 2018 Part D Prescriber | Data.CMS.gov [Internet]. [cited 2021 Jan 6]. Available from: https://data.cms.gov/MedicarePart-D/Medicare-Provider-Utilization-and-PaymentData-201/mhdd-npjx

Donnell JK, Halpin J, Mattson CL, et al. Deaths involving fentanyl, fentanyl analogs, and U-47700: 10 states, JulyDecemeber 2016. US Department of Health and Human Services/Centers for Disease Control and Prevention; 2017 Nov p. 1197–202. (Morb Mortal Wkly Rep).

10. Barry CL. Fentanyl and the evolving opioid epidemic: what strategies should policy makers consider? Psychiatr Serv. 2018 Jan;69(1):100–3. 11. Mattson CL, O’Donnell J, Kariisa M, et al. Opportunities to prevent overdose deaths involving prescription and illicit opioids, 11 states, July 2016–June 2017. Morb Mortal Wkly Rep. 2018;67(34):7.

22. Chen JH, Humphreys K, Shah NH, Lembke A. Distribution of opioids by different types of Medicare prescribers. JAMA Intern Med. 2016 Feb;176(2):259–61. 23. ARCOS Retail Drug Summary Reports [Internet]. [cited 2020 Apr 27]. Available from: https://www.deadiversion. usdoj.gov/arcos/retail_drug_summary/ 24. Home - MICROMEDEX [Internet]. [cited 2020 Apr 29]. Available from: https://www-micromedexsolutionscom.gcsom.idm.oclc.org/micromedex2/librarian/ CS/7E5043/ND_PR/evidencexpert/ND_P/evidencexpert/ DUPLICATIONSHIELDSYNC/D6B0C8/ND_PG/ evidencexpert/ND_B/evidencexpert/ND_AppProduct/ evidencexpert/ND_T/evidencexpert/PFActionId/ pf.HomePage?navitem=topHome&isToolPage=true

12. Seth P, Rudd RA, Noonan RK, Haegerich TM. Quantifying the epidemic of prescription opioid overdose deaths. Am J Publ Health. 2018 Apr;108(4):500–2.

25. Kaiser Family Foundation. State health facts [Internet]. KFF. 2019 [cited 2021 Jul 12]. Available from: https://www. kff.org/statedata/

13. NIDA. Research on the use and misuse of fentanyl and other synthetic opioids [Internet]. National Institute on Drug Abuse. 2017 [cited 2020 Oct 30]. Available from: https://www.drugabuse.gov/about-nida/legislativeactivities/testimony-to-congress/2017/research-on-theuse-and-misuse-of-fentanyl-and-other-synthetic-opioids-

26. Centers for Disease Control and Prevention. Drug overdose deaths. 2019.

14. Manchikanti L. Reframing the prevention strategies of the opioid crisis: Focusing on prescription opioids, fentanyl, and heroin epidemic. Pain Phys. 2018 Jul 15;1(21;1):309–26. 15. Beletsky L, Davis CS. Today’s fentanyl crisis: Prohibition’s Iron Law, revisited. Int J Drug Policy [Internet]. 2017 Aug [cited 2020 Oct 30];(46). Available from: https://reader.elsevier.com/reader/sd/pii/ S0955395917301548?token=A2B93174A8314170D2C 032F81ED0215C3D4E448D7413C67C2BAFB8F62702 3B4DF9430EA611761CECD3060ACA5A314906

27. Hoagland WG, Parekh, A, Swope T, Bailie M, Hart N, LaBelle R, et al. Tracking federal funding to combat the opioid crisis [Internet]. 2019 Mar. Available from: https:// bipartisanpolicy.org/wp-content/uploads/2019/03/ Tracking-Federal-Funding-to-Combat-the-Opioid-Crisis. pdf 28. Governor’s Cabinet Opiate Action Team. New strategies to fight the opiate and fentanyl crisis in Ohio (2016-2017). 2018. 29. National Institute on Drug Abuse. Mississippi: Opioidinvolved deaths and related harms. 2020 Apr. 30. Paton A. U.S. Senators pass opioid legislation [Internet]. 2018 [cited 2020 Nov 14]. Available from: https://www. hydesmith.senate.gov/us-senators-pass-opioid-legislation

16. Collins LK, Pande LJ, Chung DY, et al. Trends in the medical supply of fentanyl and fentanyl analogues: United States, 2006 to 2017. Prev Med. 2019 Jun;123:95–100.

31. Baker, Paden, Hines. House Bill 867 [Internet]. [cited 2020 Nov 14]. Available from: http://billstatus.ls.state.ms.us/ documents/2019/html/HB/0800-0899/HB0867CS.htm

17. Hedegaard H, Bastian BA, Spencer MR, Warner M, Trinidad JP. Regional differences in the drugs most frequently involved in drug overdose deaths: United States, 2017. Natl Vital Stat Rep. 2019 0ctober;68(12).

32. Pimentel CB, Gurwitz JH, Tjia J, Hume AL, Lapane KL. New initiation of long-acting opioids in long-stay nursing home residents. J Am Geriatr Soc. 2016 Sep;64(9):1772–8.

18. Zoorob M. Fentanyl shock: The changing geography of overdose in the United States. Int J Drug Policy. 2019 Aug;70:40–6. 19. Bureau UC. American Community Survey Data [Internet]. The United States Census Bureau. [cited 2020 Apr 27]. Available from: https://www.census.gov/programs-surveys/ acs/data.html 244

33. Nelson L, Schwaner R. Transdermal fentanyl: Pharmacology and toxicology. J Med Toxicol. 2009 Dec;5(4):230–41. 34. Gardner-Nix J. Caregiver toxicity from transdermal fentanyl. J Pain Symptom Manage. 2001;(21):447–8. 35. Schauer CKMW, Shand JAD, Reynolds TM. The Fentanyl Patch Boil-Up – A novel method of opioid abuse. Basic Clin Pharmacol Toxicol. 2015;117(5):358–9.

Pronounced Declines in Licit Fentanyl Utilization and Changes in Prescribing and Reimbursement Practices

36. Brummond P, Chen D, Churchill W, et al. ASHP guidelines on preventing diversion of controlled substances. Am J Health Syst Pharm. 2017;74:325–48. 37. International Medication Safety Network. Medication incidents related to the use of fentanyl transdermal systems: An international aggregate analysis. [Internet]. [cited 2020 Dec 31]. Available from: https://www. intmedsafe.net/imsn-advocacy/medication-safety-issues/ 38. Center for Drug Evaluation and Research. Post-market drug safety information for patients and providers: fentanyl transdermal system (marketed as Duragesic) information. FDA [Internet]. 2018 Nov 3 [cited 2020 Dec 31]; Available from: https://www.fda.gov/drugs/postmarket-drug-safetyinformation-patients-and-providers/fentanyl-transdermalsystem-marketed-duragesic-information 39. Nataraj N, Zhang K, Guy GP, Losby JL. Identifying opioid prescribing patterns for high-volume prescribers via cluster analysis. Drug Alcohol Depend. 2019 Apr;197:250– 4. 40. Lev R, Lee O, Petro S, Lucas J, Castillo EM, Vilke GM, et al. Who is prescribing controlled medications to patients who die of prescription drug abuse? Am J Emerg Med. 2016 Jan;34(1):30–5. 41. Vital Statistics Rapid Release - Provisional Drug Overdose Data [Internet]. 2020 [cited 2020 Dec 22]. Available from: https://www.cdc.gov/nchs/nvss/vsrr/drug-overdose-data. htm 42. Mars SG, Rosenblum D, Ciccarone D. Illicit fentanyls in the opioid street market: Desired or imposed? Addiction. 2019 May;114(5):774–80. 43. Drug Enforcement Administration. DEA takes additional steps to allow increased production of controlled substances used in COVID-19 care [Internet]. DEA. 2020 [cited 2020 Nov 16]. Available from: https://www.dea.gov/ press-releases/2020/04/07/dea-takes-additional-stepsallow-increased-production-controlled 44. Brennan Z. FDA reports shortage of sedation drug used for putting Covid-19 patients on ventilators [Internet]. Endpoints News. 2020 [cited 2020 Nov 16]. Available from: https://endpts.com/fda-reports-shortage-of-sedationdrug-used-for-putting-covid-19-patients-on-ventilators/ 45. Thomas K. Insys founder gets 5½ years in prison in opioid kickback scheme. The New York Times [Internet]. 2020 Jan 23 [cited 2021 Jan 6]; Available from: https://www.nytimes. com/2020/01/23/health/opioids-insys-kapoor-prison.html

245

Scholarly Research In Progress • Vol. 5, November 2021

A Novel Approach to Chest Wall Reconstruction Following Intrathoracic Scapular Dislocation Michael S. Pheasant1† and Shazad Shaikh1,2,3 ¹Geisinger Commonwealth School of Medicine, Scranton, PA 18509 ²Geisinger Community Medical Center, Scranton, PA 18510 ³Geisinger Orthopaedics and Sports Medicine, Scranton, PA 18510 † Doctor of Medicine Program Correspondence: mpheasant@som.geisinger.edu

Abstract Intrathoracic scapular dislocation is a rare complication of rib resection, with only five cases previously identified in the literature. We report here a case of a 47-year-old female who presented with recurrent intrathoracic scapular dislocation secondary to first, second, and third rib resection in the setting of lung cancer resection. A successful chest wall reconstruction was achieved through a novel approach utilizing an Achilles tendon allograft to close the posterolateral chest wall defect and support scapular motion. Postoperatively, the patient reported no pain or shoulder instability and demonstrated a significant restoration of shoulder function and range of motion.

Introduction Intrathoracic scapular dislocation is a rare complication of thoracic surgery involving rib resection (1-4). Subluxation of the inferior border of the scapula can occur following rib resection associated with cancer resection or lung transplant. To date there have been very few cases of intrathoracic scapular dislocations reported in the literature, and only five which have occurred following rib resection (1-4). Here we report a case of intrathoracic scapular dislocation which involved two intraoperative closed reductions followed by reconstruction using a novel method which utilized an Achilles tendon allograft to reconstruct the posterolateral chest wall.

Case Presentation The patient is a 47-year-old woman who had undergone right superior sulcus tumor resection. The procedure involved resection of the posterolateral aspects of the first, second, and third ribs and was performed without complication. Approximately 3 months postoperatively, the patient presented with severe right shoulder pain, which she reported noting after swinging a softball bat. Imaging and workup concluded that the inferior border of the right scapula was impinged in the thoracic cavity. Closed reduction was performed successfully. It was decided that chest wall reconstruction would be postponed due to her ongoing radiation therapy. Three months later, the patient again presented with dislocation, and chest wall reconstruction was planned. On the day of the operation the patient was febrile and feeling ill, so reconstruction was postponed, and a second closed reduction was performed. One week later, the patient reported a third dislocation, and chest wall reconstruction was subsequently rescheduled. In addition to attempting chest wall reconstruction, the patient was presented with the option of scapulothoracic fixation. The 246

patient declined fixation and opted for reconstruction due to her understanding of the permanent loss of scapular motion that scapulothoracic fixation would cause. The patient was placed in the right decubitus position with the right arm supported on a padded arm support. The inferior border of the right scapula was identified and was mobilized and removed from the internal surface of the fifth and sixth ribs. The teres major muscle was partially transected along the inferior medial margin of the scapula to allow the scapula to be retracted posteriorly. This freed the tissue underlying the scapula and allowed anchor points for the installation of dual mesh to be identified. The anchor points for the dual mesh were, the intercostal facia of the previously resected fourth rib, superiorly, the medial edge of the pectoralis facia, medially, the body of the fifth rib, inferiorly, and the facia of the erector spinae, posteriorly. 2-0 Ethibond sutures were placed along each anchor point with pledgets which anchored a 6 x 15 patch of dual surface Gore-Tex mesh. Once the mesh was secured, a cadaveric Achilles allograft was placed along the superior border of the remaining rib and laid flat with the Achilles tendon fanned out over the lateral aspect coursing medially. The allograft was sutured to the pleural fascia and the periosteum from the rib superiorly, inferiorly, medially and laterally (Figure 1). The farthest lateral aspect of the graft and the most inferior portion were secured using a fiber tape suture in a locking Krackow fashion. The most superior aspect of the graft was secured to the pleural fascia with a running #2 FiberWire suture. The previously placed anchor point sutures were then passed through the Achilles allograft which was spanned across the defect. A mesh graft was then placed over the allograft and sewn into position. The remaining distal end of the Achilles allograft was then folded back over the mesh graft and sutured into position with FiberWire suture (Figure 2). Prior to closure the scapula was noted to glide over the allograft surface without dislocation. The procedure was completed without complication. Fourteen months postoperatively, the patient reported no pain and demonstrated significant range of motion, including forward flexion: 125 degrees; external rotation: 75 degrees; abduction: 105 degrees.

Discussion In this case, we describe a patient who presented with recurrent intrathoracic scapular dislocation following lung cancer resection. Previously there have been only five cases of intrathoracic scapular dislocation reported following an operation which involved rib resection. Of the five cases, one involved a patient who was able to reduce the dislocated scapula himself, and two required scapulothoracic fusion (1, 2). Another case utilized Marlex mesh to close the chest wall

A Novel Approach to Chest Wall Reconstruction Following Intrathoracic Scapular Dislocation

capsular reconstruction, and anterior and posterior cruciate ligament reconstruction (5, 6, 7). In this case, utilization of Achilles tendon allograft was ideal for posterior chest wall reconstruction. Utilization of the fanned-out portion of the allograft which covered the dual mesh, provided a supportive subscapular surface in addition to closure of the chest wall defect. The narrow end of the allograft once folded back and secured over the fanned-out portion, recreated a subscapular plane similar to the native posterolateral chest wall over which the scapula could glide without subluxation into the thoracic cavity. Based on the clinical outcome, including significant range of motion and functional mobility, we contend that this method should be considered in the management of intrathoracic scapular dislocations which require posterolateral chest wall reconstruction. Figure 1. Intraoperative photograph showing the patient in the right decubitus position with the scapula retracted posteriorly, and placement of the fanned-out portion of the Achilles tendon allograft over the previously secured mesh, providing closure of the posterolateral chest wall defect from the previous rib resection.

Acknowledgments Thank you to JMJ and Mary Pheasant OTRL for assistance with editing.

Disclosures No related conflicts of interest to disclose.

References

Figure 2. Intraoperative photograph showing the scapula retracted posteriorly, and the distal end of the Achilles allograft folded back and secured over the fanned-out portion of the allograft completing reconstruction of the subscapular surface on the posterior thoracic wall.

defect (3). Most recently, Tomita et al. described a chest wall reconstruction utilizing titanium plate fixation of the partially resected fifth rib (4). In this case we report a novel method of chest wall reconstruction utilizing a cadaveric Achilles tendon allograft to reconstruct the rib bed. Due to its tensile strength and versatility, Achilles allograft provides remarkable utility in a variety of surgical applications.

Conclusion Currently, Achilles allograft is routinely used in superior

Ward, W. G., Weaver, J. P., & Garrett, W. E. (1989). Locked scapula. A case report. The Journal of Bone & Joint Surgery, 71(10), 1558-1559. doi:10.2106/00004623-19897110000016

Endrizzi, D. P., Shubert, D. J., & White, R. R. (2015). Scapulothoracic fusion for low-energy intrathoracic scapula dislocation: a report of two cases. Journal of Shoulder and Elbow Surgery, 24(4), e91-e95. doi:10.1016/j. jse.2014.12.004

Eguchi, T., Kondo, R., Shiina, T., & Yoshida, K. (2011). An intrathoracic scapular prolapse with hemorrhagic shock after a thoracotomy. Interactive CardioVascular and Thoracic Surgery, 12(2), 326-327. doi:10.1510/icvts.2010.248690

Tomita, M., Iwasaki, A., Ayabe, T., Maeda, R., & Nakamura, K. (2018). Intrathoracic scapular dislocation following lung cancer resection. Journal of Surgical Case Reports, 2018(7). doi:10.1093/jscr/rjy178

Marx RG, Hetsroni I. Surgical technique: medial collateral ligament reconstruction using Achilles allograft for combined knee ligament injury. Clin Orthop Relat Res. 2012 Mar;470(3):798-805. doi: 10.1007/s11999-011-1941-8. PMID: 21660595; PMCID: PMC3270177.

Campbell MP, Barnes BJ, Vap AR. Reconstruction of Chronic Patellar Tendon Injury with Achilles Tendon Allograft: A Case Report. JBJS Case Connect. 2020 JulSep;10(3):e1900619. doi: 10.2106/JBJS.CC.19.00619. PMID: 32910605.

Mardani-Kivi M, Karimi-Mobarakeh M, Hashemi-Motlagh K. Simultaneous arthroscopic ACL and PCL reconstruction using allograft Achilles tendon. J Clin Orthop Trauma. 2019 Oct;10(Suppl 1):S218-S221. doi: 10.1016/j. jcot.2019.01.001. Epub 2019 Jan 2. PMID: 31695286; PMCID: PMC6823795. 247

Scholarly Research In Progress • Vol. 5, November 2021

Differences in Multiple and Single-Drug Arrests by the Maine Diversion Alert Program (DAP) Maaz Siddiqui1, John Piserchio2†‡, Misha Patel2†‡, Jino Park2†‡, Michelle Foster3, Clare E. Desrosiers3, John Herbert3, Stephanie D. Nichols4, Kenneth McCall4, and Brian J. Piper2,5 ¹The University of Scranton, Scranton, PA 18510 ²Geisinger Commonwealth School of Medicine, Scranton, PA 18509 ³Diversion Alert, Houlton, Maine 04730 4 University of New England, Portland, Maine 04103 5 Center for Pharmacy Innovation and Outcomes, Forty Fort, PA 18704 † Doctor of Medicine Program ‡ Authors contributed equally Correspondence: siddiquim40@gmail.com

Abstract

Introduction

Background: Much of the blame of the increasing death toll by drug overdoses has justifiably been attributed to the United States’ current opioid epidemic. However, nearly 80% of overdoses related to opioids involve another drug substance or alcohol. The objective of this study was to elucidate overrepresentation of drugs in polypharmacy arrests by identifying drugs that were more likely to be found in conjunction with other substances, using the drug arrest data provided by the Maine Diversion Alert Program (DAP).

Deaths attributed to drug overdose in the United States (U.S.) have continued to rise over the last 20 years, resulting in over 70,000 deaths in 2019 alone. Although much of the focus has rightfully been placed on the opioid drug class, deeper investigations into the patterns of drug misuse have exposed a multitude of other drug classes that contribute to overdose-related deaths, including stimulants such as cocaine and methamphetamine, depressants such as benzodiazepines, and even antidepressants (1). In fact, toxicology studies of all synthetic opioid-related overdose deaths in 2016 showed that nearly 80% of the death involved another drug or alcohol (2). Polysubstance drug use is becoming increasingly common in the US. An analysis of 2,244 opioid overdose-related deaths in Massachusetts showed that 36% of the deaths involved a stimulant and 46% involved a non-stimulant substance (3). Only 17% of the 2,244 deaths involved solely an opioid. In 2018, there were over 16,000 drug overdose deaths in 24 states and Washington D.C., a third of which included both opioids and stimulants. Polysubstance misuse and addiction continue to impact millions of American’s every day and new information is needed to develop tools to combat this pervasive public health crisis.

Methods: Single drug arrest and multiple drug arrest totals reported to the DAP from June 2013 to early 2018 were examined. Drugs involved in the arrests were classified by Drug Enforcement Administration Schedule (I–V or noncontrolled prescription) and categorized into five drug families: hallucinogens, opioids, sedatives, stimulants, and miscellaneous. Multiple drug arrest totals were compared to single drug arrest totals to create a multiple-to-single ratio (MSR) specific to each drug family and each drug. Chi-square was used to determine statistical significance through GraphPad’s 2x2 contingency tables. Results: Over three-fifths (63.8%) of all arrests involved a single drug. Opioids accounted for over half (53.5%) of single arrests, followed by stimulants (27.7%) and hallucinogens (7.7%). Similarly, nearly two-fifths (39.6%) of multiple arrests were opioids, followed by stimulants (30.8%) and miscellaneous (13.0%). Miscellaneous family drugs were recorded with the highest multiple-to-single ratio (1.51), followed by sedatives (1.09), stimulants (0.63), opioids (0.42), and hallucinogens (0.35). Carisoprodol (8.80), amitriptyline (6.34), and quetiapine (4.69) had the highest MSR values and therefore were the three most overrepresented drugs in polysubstance arrests. Conclusion: The abuse of opioids, both alone and in conjunction with another drug, deserves continued surveillance in public health. In addition, common prescription drugs with lesserknown misuse potential, especially carisoprodol, amitriptyline, and quetiapine, require more attention by medical providers for their ability to enhance the effects of other drugs or to compensate for undesired drug effects.

248

The Maine Diversion Alert Program (DAP) was an informational resource that facilitated communication about drug arrests between the criminal justice and healthcare fields (4–10). An annual DAP report made the counterintuitive finding that non-controlled pharmaceuticals were more likely to be involved in arrests that involved other drugs (11). The objective of this study was to elucidate overrepresentation of drugs in polypharmacy arrests by identifying the specific drugs that were more likely to be found with other substances. The DAP previously identified some drugs not commonly thought to have pronounced addictive profiles, such as gabapentin, which is now considered a controlled substance in some states (12). Substance use disorders are complex, and we hypothesized that a variety of agents, perhaps even those less commonly thought to contribute to substance use disorders (e.g., amitriptyline, carisoprodol, quetiapine), would be found in combination with other more common illicit drugs like heroin and cocaine. This hypothesis was based on the premise that

Differences in Multiple and Single-Drug Arrests by the Maine Diversion Alert Program

drug users experiment and discover ways to manipulate the pharmacodynamics or pharmacokinetics of their primary drug by using multiple drugs concurrently to enhance euphoria or hallucinogenic properties and control adverse effects. By investigating polysubstance arrest patterns, we hoped to contribute a greater understanding of substance use disorders and the role polysubstance misuse plays in overdose-related mortality, but at an earlier juncture where corrective measures could be implemented.

Methods Subjects Subjects included individuals reported to the Maine DAP for drug arrests (N = 9,216, 31.38% female; age = 32.98, SD = 9.85, min = 18, max = 83, a minority 246 (2.67%) of individuals did not have age information). Additional information concerning sex, town of residence, county, offense, drug(s), drug category, federal schedule, Maine drug schedule, number of drugs, date and year of charge, law enforcement agency, and agency category were submitted by local, county, state, and federal law enforcement personnel. This study was approved as exempt by the Geisinger Institutional Review Board. Procedures Drug arrest data was documented through the Maine DAP during its course of statewide operation from June 2013 to early 2018 when it ceased operations due to insufficient funding. Drugs (N = 170) involved in the arrests were classified by Drug Enforcement Administration Schedule as I–V or non-controlled prescription in accordance with the US 1970 Controlled Substances Act and updates. Substances were then categorized into drug families as: hallucinogens (e.g., lysergic acid diethylamide [LSD] and methylenedioxymethamphetamine [MDMA]); opioids (e.g., morphine, methadone); sedatives (e.g., carisoprodol, zolpidem); stimulants (e.g., amphetamine, methylphenidate); and miscellaneous pharmaceuticals including substances not classified under the other drug families that are typically manufactured by pharmaceutical companies (e.g., amitriptyline, quetiapine). Additional information about the families may be found elsewhere (4–11). Unspecified or unknown drugs at the time of arrest (N = 1,108) or drug paraphernalia (N = 85) were excluded from further analysis.

values did not exert an undue impact on the MSR. Chi-square approximations and two-tailed p-values were used to determine statistical significance through GraphPad’s 2x2 contingency tables. An alpha of <0.05 was considered statistically significant.

Results Multiple and single drug arrestees were similar demographically. Multiple (31.51% female) and single (31.35% female) were equivalent in terms of sex. Multiple (33.36 ± 9.56 years) and single (32.91 ± 9.91 years) also did not differ appreciably based on age. Over three-fifths (63.8%) of all drug arrests involved a single drug. Opioids (e.g., heroin, oxycodone, and buprenorphine) accounted for over half (53.5%) of single arrests, followed by stimulants (27.7%, e.g., cocaine, crack, and methamphetamine) and hallucinogens (7.7%, e.g., marijuana, bath salts, and MDMA). Specifically, the five most common drugs in single drug arrests were heroin (27.15%), oxycodone (10.48%), cocaine (10.36%), buprenorphine (9.05%), and crack (7.72%). Similar to the single drug arrest data, nearly two-fifths (39.6%) of multiple arrests were opioids, followed by stimulants (30.8%) and miscellaneous (13.0%). Likewise, the five most common drugs in multiple drug arrests were heroin (17.05%), cocaine (11.37%), crack (9.70%), buprenorphine (8.17%), and oxycodone (5.63%). Figure 1 shows the overall MSR of each drug family. The drug families that had the highest overall MSR values were miscellaneous (1.51) and sedatives (1.09) which were significantly overrepresented for arrests. Stimulants (0.63), opioids (0.42) and hallucinogens (0.35) were significantly underrepresented for arrests. Individual agents that were found in both single and multiple drug arrests (N = 34) were explored further. Figure 2 shows the MSR for each of the 34 drugs involved in both single and multiple drug arrests. Sedatives had the highest MSR, followed by miscellaneous, opioids, stimulants, and hallucinogens. Firstly, carisoprodol (8.80), zolpidem (2.93), and diazepam (2.86) had the three highest MSRs within the sedative drug family, while alprazolam (1.38) was the lowest. Secondly,

Data analysis Summons and indictments were categorized for simplicity as “arrests.” Single (N = 6,441) and multiple (N = 3,660) drug arrests totals were compared. For the assessment of multiple drug arrests, each unique drug reported per offense was analyzed. Figures were created using GraphPad Prism, version 9.1.0. Multiple drug arrest totals were compared to single drug arrest totals to create a multiple-to-single ratio (MSR) specific to each drug family and each drug. For example, a substance that accounted for 10% of multiple drug arrests and 5% of single drug arrests would have a MSR of (10%)/(5%) or 2.0. An MSR > 1.0 indicated overrepresentation and an MSR < 1.0 indicated underrepresentation. Agents that were involved in less than 10 multiple arrests were not reported in Figure 2 so that smaller

Figure 1. Multiple-to-single ratio by drug families for arrests reported to the Maine Diversion Alert Program. **p < 0.005 versus all other families.

249

Differences in Multiple and Single-Drug Arrests by the Maine Diversion Alert Program

Polysubstance misuse in the U.S. is commonly associated with notable drug combination patterns, such as the concurrent use of stimulants with opioids to create a more tolerable adverse effect profile of each drug. Illicitly manufactured drugs such as heroin, fentanyl, cocaine, and methamphetamine are known to have a large impact on polysubstance overdose-related mortality (13), however, our understanding of how prescription drugs impact polysubstance substance use disorders is still poorly understood. Interestingly, the drugs that were most underrepresented in our analysis (MSR < 1.0) were methamphetamine, heroin, bath salts, oxycodone, and marijuana. While drugs with a higher index of nonmedical use and comorbidity such as heroin and methamphetamine are well known to result in morbidity and mortality, our data suggests that there are other commonly prescribed drugs that may be more pervasive in the development of polysubstance use disorder and subsequent enhanced drug misuse behavior. What follows is a discussion that reviews the pharmacology of carisoprodol, amitriptyline, and quetiapine, and current theories of how each drug may contribute to polysubstance use disorder.

Figure 2. Multiple-to-single ratio for specific drugs grouped by drug families as reported by the Maine Diversion Alert Program. Each drug’s multiple and single drug arrests are in parentheses, respectively. *p < 0.05, **p < 0.005

amitriptyline (6.34), quetiapine (4.69), and clonidine (2.84) had the three highest MSRs within the miscellaneous drug family, while bupropion (1.48) was the lowest. Thirdly, morphine (3.06), methadone (1.64), and fentanyl (1.35) had the three highest MSRs within the opioid drug family, while oxycodone (0.54) was the lowest. Fourthly, amphetamine (2.27), methylphenidate (1.97), and lisdexamfetamine (1.29) had the three highest MSRs within the stimulant drug family, while methamphetamine (0.44) was the lowest. Lastly, LSD (1.94), MDMA/MDA (1.16), and hashish (0.71) had the three highest MSRs within the hallucinogen drug family, while α-PVP/bath salt (0.34) was the lowest.

Discussion We examined drugs that were overrepresented in multiple drug arrests relative to single drug arrests as reported by the Maine DAP. The substances with the highest, and significant, overrepresentation (MSR > 1.0) in our analysis were carisoprodol (8.80), amitriptyline (6.34), and quetiapine (4.69).

250

The tricyclic antidepressant (TCA) amitriptyline also has analgesic properties that lead to its use in treating depression, neuropathic pain, migraines, and fibromyalgia (14). TCAs exert their effect by inhibiting the reuptake of norepinephrine and serotonin in adrenergic and serotonergic neurons. As one of the first antidepressant classes to be discovered, TCAs have more recently fallen out of favor in treating depression due to their antimuscarinic, neurologic, and cardiac adverse effect profile and their high risk of overdose fatality. However, they remain one of the more commonly prescribed drugs for their efficacy in treating pain that is unresponsive to opioids. Despite a lack of addictive chemical components, amitriptyline’s sedative and anti-anxiolytic effects have led to misuse of the drug. Case reports of nonmedical use of amitriptyline describe feelings of “relaxation, giddiness, and contentment” as well as “a buzz… numbed up… calming” effects in patients who were found to have elevated levels of amitriptyline and no other substance in the emergency department (15, 16). Patients with a history of trauma are more likely to seek out methods of depersonalization and dissociation, including the nonmedical use of medications. Dosages of amitriptyline of 100–200 mg/ day are sufficient to cause euphoric effects and potentiate nonmedical use (17). Further, due to the nature of the diseases that TCAs treat, many patients who use amitriptyline have other comorbid conditions such as diabetes that necessitate polysubstance treatment or mental illness that foster higher risk for nonmedical use of mixed drugs. Although amitriptyline is not listed as a controlled substance by the US Drug Enforcement Administration, it is frequently one of the drugs involved in suicidal and accidental multi-drug overdoses, due to its narrow therapeutic index (18). Interestingly, among 346 people enrolled in a methadone maintenance program, 25% admitted to taking amitriptyline in order to achieve feelings of euphoria, not only indicating that amitriptyline misuse is not uncommon, but also suggesting that it may have synergistic effects with opioids (19). Altogether, this data warrants further investigation as to whether individuals were intentionally co-administering amitriptyline to enhance the sedating effects of opioids or benzodiazepines.

Differences in Multiple and Single-Drug Arrests by the Maine Diversion Alert Program

The muscle relaxer carisoprodol is indicated for use for acute musculoskeletal pain with guidelines suggesting a maximum prescription length of 2 to 3 weeks. Withdrawal is a concern with long-term or high-dose users. In the last two decades, nonmedical use of carisoprodol increased to the point of requiring scheduling as a controlled substance in January 2012. The exact mechanism of action is not known; however, research suggests that it causes a depressant effect at GABA receptors. This effect is often augmented with opioids, benzodiazepines, or alcohol. A 2012 national survey found that millions of individuals report using carisoprodol for non-medical reasons (20). Between 2011 and 2016, carisoprodol ranked amongst the top 15 drugs involved in drug overdose deaths. Despite the increased protections surrounding this Schedule IV drug, higher vigilance and judicious prescriptions from clinicians will benefit patients. The second-generation antipsychotic quetiapine has a diverse mechanism of action. Quetiapine shows greater affinity for the histamine (H1), adrenergic (α1), muscarinic (M1/3) and serotonergic (5-HT2A/C) receptors as well as the norepinephrine transporter relative to the dopamine (D2) receptor (21). Quetiapine is normally prescribed to treat psychotic and mood conditions, including schizophrenia, bipolar disorder, and as adjunctive therapy in major depressive disorder. There is a growing concern that quetiapine is being commonly prescribed and used as off-label prescriptions for dementia-related behavioral problems, insomnia, anxiety, and post-traumatic stress disorder (PTSD) and may be contributing to adverse outcomes in vulnerable populations, such as those with mental health illnesses (22). Quetiapine is also the most documented antipsychotic bought and sold illicitly which is most likely due to the drug’s relatively low risk for the development of dystonia and extrapyramidal symptoms due to its low affinity for D2 receptors when compared to other antipsychotics. It is also a common medication implicated in nonmedical use in jails and prisons (23). Due to quetiapine’s growing use with other drugs, its role in nonmedical use of drugs and polysubstance use has been investigated. Quetiapine has unique sedative and anxiolytic effects due to its antagonistic effect on H1 receptors and serotonin receptors, respectively. This drug effect profile has resulted in quetiapine’s frequent use with stimulants such as amphetamines and cocaine to reduce the feelings of anxiety that come with using stimulants (24). By controlling the adverse effects associated with stimulant use, we suspect this can lead to higher drug doses and higher risk of mortality. Additionally, studies on rodents have shown that administration of antihistamines increases the release of dopamine in the ventral striatum similar to other addictive psychoactive substances such as cocaine, and thus, quetiapine may also have reward properties that contribute to drug misuse and addiction. Alternatively, quetiapine has been often found in forensic autopsies and is thought to be used in combination with opioids to reinforce the sedative effects of these drugs (25). Our evidence suggests that polysubstance drug use is characterized by a pattern of drug use that works to enhance the effects of other drugs (carisoprodol and amitriptyline) or to compensate for undesired drug effects (quetiapine). While much attention in polysubstance literature focuses on the morbidity and mortality related to scheduled drug use, such as heroin and

cocaine, it is important to understand how polysubstance use may develop through the use of commonly prescribed drugs. In our investigation of DAP’s database of individuals arrested and charged with possession of drugs, specific drugs such as carisoprodol, amitriptyline, and quetiapine were each found to be most likely used with other drugs. The basic pharmacology and current understanding of each drug’s addictive potential indicates that these drugs ca contribute to polysubstance use disorder and a pattern of drug use that encourages the use of more potent and riskier drugs. In addition to pharmacological factors, legal practices and the broader sociocultural environment also contribute to the pattern of substances which were over- and underrepresented in multiple drug arrests. Data was collected between 2013 and 2018. The largest city in the state, Portland, decriminalized marijuana in 2013 and a statewide referendum legalizing recreational marijuana was passed in 2016. Law enforcement officers have some discretion in making arrests and therefore bias may have been present in some cases. This descriptive dataset cannot distinguish whether some individuals found in the possession of only a single non-scheduled prescription drug were not arrested or whether charges for carisoprodol, amitriptyline, or quetiapine were more likely to be issued when “harder” illicit substances were also discovered by officers. Some other limitations should be noted. Although DAP was a public health tool unique to Maine, information was submitted by law enforcement with varied backgrounds in pharmacology and may be based on information available at the time of arrests (e.g., “spot tests” with questionable validity) (26). This endeavor involved running multiple statistics and many findings were not hypothesized a priori. Findings that only met a more liberal alpha cutoff (0.05) should be viewed with caution. In conclusion, when analyzing multi-drug arrests reported by Maine’s Diversion Alert Program (DAP) between 2013 and 2018, we observed an overrepresentation of commonly prescribed drugs such as amitriptyline, carisoprodol, and quetiapine and an underrepresentation of drugs with a higher index of nonmedical use such as methamphetamine and heroin. These findings, in addition to amitriptyline, carisoprodol, and quetiapine’s pharmacological plausibility of manipulating illicit drug use, supports the notion that polysubstance use with commonly prescribed pharmaceutical medications may contribute to riskier illicit drug use tendencies. Further investigation is warranted to better understand the role polysubstance use plays in the development of substance use disorders, specifically with commonly prescribed drugs, and how healthcare professionals can identify these patterns in clinical practice to reduce drug misuse morbidity and mortality.

Disclosures BJP is part of an osteoarthritis research team supported by Pfizer and Eli Lilly. This research was supported by the FahsBeck Fund for Research and Experimentation. Diversion Alert (MF, CED, JF) was supported by a grant from Eastern Maine Medical Center. The other authors have no disclosures.

251

Differences in Multiple and Single-Drug Arrests by the Maine Diversion Alert Program

Acknowledgments The contributions of Stephanie D. Nichols, PharmD, Kevin J. Simpson, MD, Matthew T. Moran, MD, and Dipam T. Shah, MD, are gratefully acknowledged. Jove Graham, PhD, provided valuable input regarding the data analysis.

References

13. Huang X, Keyes KM, Li G. Increasing prescription opioid and heroin overdose mortality in the United States, 19992014: An age-period-cohort analysis. Am J Public Health. 2018;108(1):131-136. 14. Bryson HM, Wilde MI. Amitriptyline. A review of its pharmacological properties and therapeutic use in chronic pain states. Drugs Aging. 1996;8:458-476.

National Institute on Drug Abuse. Overdose death rates. National Institutes of Health. 2020;2018.

15. Delisle JD. A case of amitriptyline abuse. Am J Psychiatry. 1990; 147:1377-8.

Jones CM, Einstein EB, Compton WM. Changes in synthetic opioid involvement in drug overdose deaths in the United States, 2010-2016. JAMA 2018;319(17):18191821.

16. Wohlreich MM, Welch W. Amitriptyline abuse presenting as acute toxicity. Psychosomatics. 1993; 34:191-193.

Barocas JA, Wang J, Marshall BDL, LaRochelle MR, Bettano A, Bernson D, et al. Sociodemographic factors and social determinants associated with toxicology confirmed polysubstance opioid-related deaths. Drug Alcohol Depend. 2019;200:59-63.

18. Prahlow JA, Landrum JE. Amitriptyline abuse and misuse. Am J Forensic Med Pathology. 2005;26(1):81-86.

Piper BJ, Desrosiers CE, Fisher HC, McCall KL, Nichols SD. A new tool to tackle the opioid epidemic: Description, utility, and results from the Maine Diversion Alert Program. Pharmacotherapy. 2017;37(7):791-798.

McCall K, Nichols SD, Holt C, Ochs L, Cattabriga G, Tu C. Prescription Monitoring Program trends among individuals arrested in Maine for trafficking prescription drugs in 2014. Pharmacotherapy. 2016;36(6):585-589.

Martin SL, Desrosiers C. Diversion Alert: 1-year evaluation across Northern New England, 2013-2014. Prev Chronic Dis. 2016;13:e159. Piper BJ, Desrosiers CE, Lipovsky JW, Rodney MA, Baker RP, McCall KL, et al. Use and misuse of opioids in Maine: Results from pharmacists, the Prescription Monitoring, and the Diversion Alert Programs. J Stud Alcohol Drugs. 2016;77(4):556-65. Piper BJ, Suarez MJ, Piserchio JP, Shah DT, Simoyan OM, McCall KL, et al. Illicit and prescription drug misuse as reported to the Maine Diversion Alert Program. Forensic Sci Int. 2018;285:65-71. Simpson KJ, Moran MT, Foster ML, Shah DT, Chung DY, Nichols SD, et al. Descriptive, observational study of pharmaceutical and non-pharmaceutical arrests, use, and overdoses in Maine. BMJ Open. 2019;9(4): e027117.

10. Simpson KJ, Moran MT, McCall KL, Herbert J, Foster ML, Simoyan OM, et al. Increasing heroin, cocaine, and buprenorphine arrests reported to the Maine Diversion Alert Program. Forensic Science Int. 2019;303:109924. 11. Hyde TF, Bekoe-Tabiri AD, Kropp Lopez AK, Devia LG, Gutierrez BD, Lara MC, et al. County and demographic differences in drug arrests and controlled substance use in Maine. Journal of the Maine Medical Center, in press. 12. Campbell L, Coomer L, Jacob G, Lenz R. Gabapentin controlled substance status. J Am Pharmacists Association. 2021; S1544-3191(21)00027-3.

252

17. Sein Anand J, Chodorowski Z, Habrat B. Recreational amitriptyline abuse. Przegląd lekarski. 2005;62(6).

19. Cohen MJ. Abuse of amitriptyline. JAMA 1978;240(13): 1372-3. 20. Sun C, Hollenbach KA, Cantrell FL. Trends in carisoprodol abuse and misuse after regulatory scheduling: A retrospective review of California poison control calls from 2008 to 2015. Clin Tox. 2018;56(7):653-655. 21. Maan J, Ershadi M, Khan I, Saadabadi A. Quetiapine. Treasure Island: StatPearls; 2020. 22. Kim S, Lee G, Kim E, Jung H, Chang J. Quetiapine misuse and abuse: Is it an atypical paradigm of drug seeking behavior? J Research Pharmacy Pract. 2017;6(1):12-15. 23. 23.Pinta ER, Taylor RE. Quetiapine addiction? Am J Psychiatry. 2007 Jan;164(1):174-5 24. Piróg-Balcerzak A, Habrat B, Mierzejewski P. Misuse and abuse of quetiapine. Psychiatria Polska. 2015;49. 25. Andersen FD, Simonsen U, Andersen CU. Quetiapine and other antipsychotics combined with opioids in legal autopsy cases: A random finding or cause of fatal outcome? Basic Clin Pharmacol Toxicol. 2021;128:66-79. 26. Gabrielson R, Sanders T. How a $2 roadside drug test sends innocent people to jail. New York Times, July 10, 2016, MM34. Available from http://www.nytimes. com/2016/07/10/magazine/how-a-2-roadside-drug-testsends-innocent-people-to-jail.html?_r=0. Accessed May 26, 2021.

Scholarly Research In Progress • Vol. 5, November 2021

Investigating Potential Conflicts of Interest Among UpToDate and DynaMed Content Contributors SooYoung H. VanDeMark1*, Mia R. Woloszyn1†, Laura A. Christman1*, Michael Gatusky1†, Warren S. Lam1†, Stephanie S. Tilberry1†, and Brian J. Piper1 ¹Geisinger Commonwealth School of Medicine, Scranton, PA 18509 † Doctor of Medicine Program *Master of Biomedical Sciences Program Correspondence: svandemark@som.geisinger.edu

Abstract Background: Financial conflicts of interest among physicians have the potential to negatively impact patient care. Physicians contribute content to two popular, evidence-based websites, UpToDate and DynaMed; while other physicians use these websites to influence their clinical decision making. Each website maintains a conflict-of-interest policy, and contributors are required to self-report a disclosure status. This research investigated the occurrence for potential conflicts of interest among the self-reported statuses of UpToDate and DynaMed content contributors. Methods: An initial list of contributors for each website was compiled using the Centers for Disease Control and Prevention’s 2017 Leading Causes of Death. The top 50 causes were used to determine a relevant article with clinical implications from each database. All named authors and editors of those articles comprised our list of investigated contributors. Contributor disclosure status was then compared with public records of financial remuneration as reported in the Open Payments database maintained by the Centers for Medicare and Medicaid Services and ProPublica’s Dollar for Docs website from 2013 to 2018. Descriptive analysis and Fisher’s exact tests were performed on the data. Results: Of 76 UpToDate contributors, 57.9% reported nothing to disclose but had a record of receiving a financial payment on Open Payments, which was found to be statistically significant (p = 0.0002). Of DynaMed’s 42 contributors who reported nothing to disclose, 83.3% had an entry on Open Payments. However, this was not statistically significant. The sum total of industry payments between 2013-2018 made to UpToDate contributors was $68.1 million. The top 10 UpToDate contributors who received the most financial remuneration earned approximately $56.1 million (82.4% of all UpToDate money), were all male, and only one had a nothing-to-disclose status. The sum total of money reported for the discordant UpToDate contributors between 2013 and 2018 was approximately $4.81 million (or 7.07% of the total monies reported to UpToDate contributors.) In that same time frame, DynaMed contributors received a sum total of $9.58 million from industry, while the top 10 DynaMed contributors earned $8.88 million (or 92.8%) of that. The top 10 DynaMed contributors were 80% male and 20% female, and six individuals reported nothing to disclose, yet had an Open Payments entry. The sum total of money reported for all discordant DynaMed contributors between 2013 and 2018 was approximately $2.79 million (or 29.2% of the total monies reported to DynaMed contributors).

Conclusion: While this research does not ascertain that a conflict of interest or other such unethical behavior has occurred, it does provide evidence that there is a significant difference between having an Open Payment entry among those who did versus those who did not disclose a conflict of interest. Websites such as UpToDate and DynaMed should consider implementing a more stringent conflict of interest policy and/or employ an unbiased team to verify self-reported disclosure statuses among its content contributors. Similarly, physicians who use such informational websites to inform their clinical decision making should look beyond a contributor’s self-reported disclosure status and check for relevant financial remuneration from the healthcare industry via Open Payments or Dollars for Docs.

Introduction Per the Institute of Medicine’s Committee on Conflict of Interest in Medical Research, Education, and Practice, a conflict of interest (COI) is defined as a set of “circumstances that create a risk that professional judgments or actions regarding a primary interest will be unduly influenced by a secondary interest” (1). This research is focused on investigating potential COIs that could occur when a physician writes or edits medical content (the primary interest) yet is influenced by financial gain (the secondary interest) in the form of payments made by health care manufacturers to the author. Undisclosed COIs may not only affect the credibility of a health resource, but also other clinicians’ ability to provide quality care. Prior bioethical research has found potential COIs among physician authors of biomedical textbooks (2), pharmacology textbooks (3), psychiatry’s Diagnostic and Statistical Manual of Mental Disorders-5 (4), as well as within clinical guidelines (5). However, there is limited research on the potential COIs among authors and editors of online clinical resources, such as UpToDate and DynaMed. UpToDate and DynaMed are two online, subscription-based products used by physicians to assist in clinical decision making. Both UpToDate and DynaMed promote their websites as an evidence-based resource for physicians to improve patient health outcomes at the point-of-care (6, 7). The content on these websites is written, edited, and overseen by various health care professionals. UpToDate and DynaMed each maintain a publicly available COI policy (8, 9). In an effort to improve transparency and as part of the Physician Payments Sunshine Act, the Centers for Medicare and Medicaid Services (CMS) discloses financial payments from “drug, device, biological, and medical supply” manufacturers to U.S.-based

253

Investigating Potential Conflicts of Interest Among UpToDate and DynaMed Content Contributors

physicians through its online and publicly accessible database, Open Payments (OP) (10). Another online, publicly accessible database is Dollars for Docs (DFD), which is maintained by the nonprofit, investigative journalism organization, ProPublica. DFD also allows users to look up payments made from relevant manufacturers to U.S.-based physicians as reported to CMS (11). OP and DFD only report payments in amounts greater than $10. As mentioned, there has been limited research on potential COIs among authors of evidence-based, point-of-care websites. Researchers reviewed authors (n = 31) of the online medical resource, Medscape, for potential COIs and found 19 discordant authors (61.3%) who self-reported a nothing to disclose status but had an OP entry (12). Additionally, their study (12) found a strong correlation between the reported payment amounts on OP and DFD, consistent with prior research (3). A small study focused on UpToDate and DynaMed authors and editors, reviewed six articles on each site and found no COIs among DynaMed contributors and “numerous” potential COIs among UpToDate contributors (13). We investigated potential COIs among content contributors for DynaMed and UpToDate by cross-checking their self-reported disclosure statuses with financial records available from OP and DFD. Additionally, a descriptive analysis of UpToDate and DynaMed content contributors’ disclosure status, financial compensation, and gender was performed, with further evaluation of each website’s top 10 earners.

Methods Procedures Using the Centers for Disease Control’s Top 50 Causes of Mortality (14), each cause of death was searched for on UpToDate’s and DynaMed’s website. The research team, with individuals assigned to specific diseases, selected comparable articles from the first page of search results on each site. Only 42 causes of death were used for data collection, resulting in a total of 84 articles reviewed. Diseases were excluded from data collection, if they provided no clinically relevant search result on UpToDate or DynaMed (e.g., “operations of war and their sequelae.”) Content contributors, per this research, are defined as the individuals listed on a given UpToDate or DynaMed article page — specific titles for contributors depend on the database, but include author, deputy editor, section editor, recommendations editor, and American College of Physicians (ACP) reviewer. Each article’s listed content contributors, regardless of title, composed our initial list of physician contributors. Contributors who were of unknown or international origin were removed from this study, as OP and DFD only report on U.S.-based physicians. On both websites, under each article, a contributor states nothing to disclose or discloses the companies from which they have received payment. This status was collected as part of our data. All articles and contributor names were compiled between November 30, 2020, and December 7, 2020. Each unique contributor (n=179, 28.5% female) was then searched for in OP and DFD. If an entry was found, the financial information for 2013 to 2018 was recorded in accordance with each website’s categorization method. For example, money

254

reported to OP was categorized as general payments, research, associated research and owner/investment. Contributor location and gender was verified or cross-checked and documented using the National Provider Identifier registry website (15). The research team performed a randomized check on 11.2% of the data collected to ensure accuracy. No individuals were found to have contributed content to both UpToDate and DynaMed. Although some contributors were listed on multiple articles, no variability in disclosure status was found between articles. After data collection and prior to analysis, contributors were given a unique ID to avoid potential bias from the research team. Contributors were identified as discordant if they reported nothing to disclose but were found to have an OP or DFD entry. Contributors were identified as concordant if they made a disclosure and had an OP or DFD entry. Statistical analysis Two-sided Fischer exact tests were performed, due to the study’s small sample sizes, using Prism 9.1.0 (16). This study was deemed exempt from review by the Geisinger Institutional Review Board.

Results Of the 179 U.S-based physician content contributors, associated with the 84 articles, 128 were from UpToDate and 51 were from DynaMed (Figure 1). The sum total of reported financial compensation to OP for the UpToDate contributors between the years 2013 and 2018 was $68,085,233, of which the top 10 earners accounted for $56,083,923, or 82.4%. The sum total reported to OP for the DynaMed contributors within that same time frame was $9,576,109, of which the top 10 earners accounted for $8,882,249, or 92.8%. The majority of UpToDate (59.4%) and DynaMed (82.4%) contributors did not disclose any COIs related to their article topic (Figure 1). However, of those UpToDate contributors who did not disclose a COI, 57.9% had an OP entry. This discordance — self-reporting nothing to disclose yet having an OP entry — among UpToDate contributors was statistically significant (p = 0.0002). The 44 discordant UpToDate contributors accounted for $4,811,760, or 7.07% of UpToDate’s sum total as reported by OP. Although 83.3% of DynaMed’s nothing-to-disclose contributors had an OP entry, it was found to be not statistically significant. The 35 discordant DynaMed contributors accounted for $2,793,708, or 29.2% of DynaMed’s sum total as reported by OP. Six content contributors for UpToDate were found to be neither discordant nor concordant. They had made a disclosure, but no OP nor DFD record was found. Each of the nine DynaMed content contributors who disclosed were found to be concordant. All the top 10 earners within the UpToDate content contributors were male, of whom only one contributor was found discordant (Figure 2). There were six discordant contributors among the top 10 earners of DynaMed contributors, of whom eight were male and two were female (Figure 2). Figure 2 also shows a breakdown of the reported compensation (in millions) to the top ten earners of each website respectively in the four OP categories, with

Investigating Potential Conflicts of Interest Among UpToDate and DynaMed Content Contributors

Associated Research, which is defined as “funding for a research project or study where the physician is named as the principal investigator”10 dominating the payments. Further investigation of the seven discordant top 10 earners found only one contributor who received payment(s) from a manufacturer for drug(s) and/or medical device(s) that was specifically mentioned by brand name in the article to which the contributor was assigned. The financial renumeration to the physician contributor was $4,695. Among all UpToDate content contributors, 94 (73.4%) were male and 34 (26.6%) were female. A Fischer’s exact test found a statistically significant difference in that male UpToDate contributors (47.9%) were more likely to report a disclosure than females (20.6%, p =0.0076). Among the DynaMed contributors, 34 (66.7%) were male and 17 (33.3%) were female. There was no statistically significant difference in how the genders disclosed among DynaMed contributors. Another Fischer’s exact test found no statistically significant difference in discordance based on the contributor’s gender for either website.

Discussion

Figure 1. Flow chart of UpToDate (left) and DynaMed (right) content contributors, showing disclosure status and Open Payment (OP) entry status.

This study identified appreciable financial COIs among pointof-care contributors with potential room for improvement in self-reported disclosures. In contrast to prior research (12), this study found potential COIs among contributors for both online resources, UpToDate and DynaMed. This is likely due to a difference in methodologies. Our findings of discordance in UpToDate and DynaMed (57.9% and 83.3% respectively) are similar to a prior finding of 61.3% discordance among Medscape’s contributors (n=31) (11). Such high discordance rates suggest a need for further research to fully illuminate the issue, as well as follow-up remediation by these online clinical resources.

Figure 2. Financial Remuneration (in millions) by category as reported to Open Payments (OP) for the top 10 earners among UpToDate (left) and DynaMed (right) content contributors. NTD = Nothing to disclose. Female symbol designates the two female contributors.

255

Investigating Potential Conflicts of Interest Among UpToDate and DynaMed Content Contributors

Based on the high percentage of discordance found among content contributors for both UpToDate and DynaMed, and the large sums of remuneration among those discordant (over $7 million) in this study, there is a strong likelihood for there to be potential COIs among physician content contributors who self-report. Our own investigation into a subgroup of our sample contributors found a suggestive potential COI. However, it is important to note that the objective of this research was not to ascertain any specific contributor’s COI. Our gender analysis revealed an interesting difference in disclosure status between males and females. However, there are multiple ways to interpret this — perhaps male physicians disclose more because they are more often the beneficiaries of industry dollars and in a position requiring disclosure; or they are more ethical than female physicians; or perhaps the difference was due to the underrepresentation of female physician content contributors, which has been confirmed in prior studies regardless of the source being a printed text or online website (2, 3, 12). One argument against the findings of this research may be that as physicians contribute content to reputable online point-ofcare websites, such as UpToDate and DynaMed, they might then be hired by the healthcare industry for their expertise and financially compensated, all within the same calendar year. In this plausible example, no direct COI has occurred. Although OP and DFD provide specific dates for the payments reported to a physician, no such timeline is provided by UpToDate or DynaMed on when the content was initially published — each site does provide a last revised date, but not when the contributor declared their disclosure status. In order to work within these confines, the research team opted to focus on disclosure status and existence of an OP/DFD entry rather than timeline. Additional limitations of this study include the relatively small sample of contributors and some noted inconsistencies in the results. The research team could not ascertain the number of contributors for all the content on either UpToDate or DynaMed; thus, the possibility exists that our sample did not accurately reflect the characteristics of all the contributors in question. Similarly, our data is only as accurate as the financial reporting provided by OP and DFD. Given that a small number from the total number of contributors (3.35%) were neither discordant nor concordant is of concern.

Research in the area of medical authorship and COIs needs to continue, with a particular emphasis placed on online medical resources. Physicians, other health care providers, and by extension their patients, should be confident in knowing that the evidence-based medical information they receive is free from outside influences.

Acknowledgments The research team would like to thank Kaitlyn E. Sternat and Amalie Kropp Lopez for their contributions to this project.

Disclosure BJP is part of an osteoarthritis research team funded by Pfizer/Eli Lilly. All other authors declare no relevant conflicts of interest.

References 1.

Field MJ, Lo B, editors. Conflict of interest in medical research, education, and practice. Washington (DC): National Academies Press (US); 2009. [cited 2021 Jun 05]. Available from: https://pubmed.ncbi.nlm.nih. gov/20662118/ doi: 10.17226/12598

Piper BJ, Lambert DA, Keefe RC, Smukler PU, Selemon NA, Duperry ZR. Undisclosed conflicts of interest among biomedical textbook authors. AJOB Empirl Bioeth. 2018 Apr 3;9(2):59-68

Piper BJ, Alinea AA, Wroblewski JR, Graham SM, Chung DY, McCutcheon LR, Birkett MA, Kheloussi SS, Shah VM, Szarek JL, Zalim QK. A quantitative and narrative evaluation of Goodman and Gilman’s Pharmacological Basis of Therapeutics. Pharmacy. 2020 Mar;8(1):1

Cosgrove L, Krimsky S. A comparison of DSM-IV and DSM5 panel members' financial associations with industry: a pernicious problem persists. PLoS Med. 2012 Mar 13;9(3):e1001190.

Ngo-Metzger Q, Moyer V, Grossman D, Ebell M, Woo M, Miller T, Brummer T, Chowdhury J, Kato E, Siu A, Phillips W. Conflicts of interest in clinical guidelines: update of US Preventive Services Task Force policies and procedures. Am J Prev Med. 2018 Jan 1;54(1):S70-80.

UpToDate. About UpToDate [Internet]. Wolters Kluwer N.V; 2021 [cited 2021 Jun 05]. Available from: https:// www.wolterskluwer.com/en/solutions/uptodate/about

DynaMed. Mission [Internet]. EBSCO Industries, Inc; 2020 [cited 2021 Jun 05]. Available from: https://www.dynamed. com/about/mission/

UpToDate. Conflict of Interest Policy [Internet]. UpToDate Inc; 2021 [updated 2019 Nov 19; cited 2021 Feb 13]. Available from: https://www.wolterskluwer.com/en/ solutions/uptodate/policies-legal/conflict-of-interest-policy

DynaMed. Conflict of Interest [Internet]. EBSCO Industries, Inc; 2020 [cited 2021 Feb 13]. Available from: https://www.dynamed.com/about/conflict-of-interest/

Conclusion Our recommendations for evidence-based, point-of-care websites are two-pronged. First, the disclosure information provided for each contributor could be made more robust by providing a “verified” date or timeframe for a nothing-todisclose status by hyperlinking to the OP and DFD pages for those who have entries and by displaying a monetary range of financial compensation (e.g., $5,000 – $10,000) for those who have OP and/or DFD entries; such changes would offer transparency to the website user who consumes the information. Second, current COI policy should be reviewed and updated annually, and a verified “no-COI” editorial team should be established to crosscheck physician contributors at random against OP and DFD; such a policy might result in content contributors erring on the side of caution and disclosing more openly and completely. 256

Investigating Potential Conflicts of Interest Among UpToDate and DynaMed Content Contributors

10. Centers for Medicare & Medicaid Services. Open Payments Data - CMS [Internet]. Baltimore: Centers for Medicare & Medicaid Services, 2021 [cited 2021 Jan]. Available from: https://openpaymentsdata.cms.gov 11. ProPublica. Dollars for Docs [Internet]. Pro Publica Inc; 2019 [updated 2019 Oct 17; cited 2021 Jan] Available from: https://projects.propublica.org/docdollars/ 12. Chopra AC, Tilberry SS, Sternat KE, Chung DY, Nichols SD, Piper BJ. Quantification of conflicts of interest in an online point-of-care clinical support website. Sci Eng Ethics. 2020 Apr;26(2):921-30. 13. Amber KT, Dhiman G, Goodman KW. Conflict of interest in online point-of-care clinical support websites. J Med Ethics. 2014 Aug 1;40(8):578-80. 14. National Vital Statistics System. Deaths, percent of total deaths and rank order for 113 selected causes of death and Enterocolitis due to Clostridium difficile, by race and hispanic origin, and sex: United States, 2015-2017 [Internet]. Centers for Disease Control and Prevention; [updated 2017 Oct 23; cited 2020 May]. Available from: https://www.cdc.gov/nchs/nvss/msortality/lcwk6_hr.htm 15. National Plan and Provider Enumeration System. Search NPI Records [Internet]. Baltimore: Centers for Medicare & Medicaid Services; [cited June 07] Available from: https:// npiregistry.cms.hhs.gov/ 16. GraphPad (2021). Prism 9.1.0 [Computer software]. Retrieved from https://www.graphpad.com/

257

2022 Summer Research Immersion Program Each summer the Geisinger Commonwealth School of Medicine Summer Research Immersion Program (SRIP) brings together first-year medical students for an opportunity to gain research experience in basic science, clinical science, public/community health, behavioral health, or medical education under the guidance of a research mentor. The summer research experience includes a $2,500 educational stipend. At the end of the program, students present their research in a poster session.

Program dates:

In addition to research, SRIP students participate in a variety of complementary enrichment activities:

For more information, contact:

• GCSoM and Geisinger faculty research seminars • GCSoM Grand Rounds and clinical seminars at our hospital partners • Special events or conferences related to your research topic • Clinical exposure • Scientific writing & communication workshops

SRIP program goals: • To provide GCSOM medical students with an in-depth research experience under the guidance of a mentor • To enhance students’ knowledge of the scope and types of research relevant to improving health in the region, nationally, and globally • To provide research opportunities that span the translational continuum from laboratory based biomedical studies to clinical and public health research conducted with community partners • To provide opportunities for students to engage in peer learning and networking • To enhance students’ skills in oral and written scholarship

258

SRIP 2022 will be an eight-week program held June 6 – July 29, 2022.

Program deadlines: Application release date: Dec. 3, 2021 Application submission deadline: Feb. 7, 2022

Sonia Lobo, PhD SRIP Program Administrator Associate Dean for Research & Scholarship slobo@som.geisinger.edu Elizabeth Kuchinski, MPH SRIP Director ekuchinski@som.geisinger.edu

Finding Your Way: Opportunities for Student Funding You can find assistance in searching for funding opportunities specifically designed for students at the Office of Research and Scholarship. Funding opportunities may include support for fellowships, internships, research, programming, and collaboration. The Office of Research and Scholarship can help you locate and qualify funding opportunities as well as assist in application preparation, budgeting, and editing. We are here to help you every step of the way! School policy requires that student applications are submitted by our office, so call or stop by early in the proposal development process and we can work with you to meet your deadline.

Funds cannot be requested for stipends, tuition, travel, or wages for the student or faculty mentor. Indirect costs to the sponsoring institution are not allowed. SRAs are intended to foster student scholarship and lead to a tangible deliverable such as an abstract for submission to a regional/national meeting or a manuscript for publication in SCRIP and/or a peer-reviewed journal. SRA applications will be due May 2, 2022, at 11:59 PM ET. If you are interested in applying, contact Tracey Pratt, MPH, or StudentResearch@som.geisinger.edu.

Contact Information:

The Office of Research and Scholarship is also available to provide general guidance on topics like proposal writing basics and the fundamentals of grant management.

GCSoM Student Research Awards (SRAs) The Office of Research and Scholarship is pleased to announce the availability of funds for the 2022 academic year to support student research projects in the areas of basic or clinical science, public/community health, behavioral health, and medical education research. The proposed project must be under the supervision of a faculty mentor and be endorsed by the Office of Research and Scholarship. The proposed project period must be no longer than 6 months and conclude by Dec. 1, 2022. The maximum award for each project is $2,000.

Tracey Pratt, MPH Grants Specialist Office of Research and Scholarship Phone: 570-558-3955 Internal extension: 5335 Email: tpratt@som.geisinger.edu

259

Through the MRHP, you will:

Medical Research Honors Program Current first-year medical students are eligible to join the Medical Research Honors Program (MRHP). With a mentor’s guidance, you will drive this long-term, thesis-focused research experience. By completing the requirements while remaining in good academic standing, you’ll graduate with an honors distinction.

Application packet must include:

Through the MRHP, you will:

• Acknowledgment of mentor’s expectations

• Advance fundamental scientific knowledge

• MRHP application form • Letter of support from research mentor • CV • Project proposal: Title, specific aims, hypothesis, background, preliminary data (if available)

• Stand apart in competitive residency application fields • Refine scholarly communication

Be a mentor

• Gain a mindset of continual growth and learning

Submit a mentor data form to: mrhp@som.geisinger.edu

To complete this 4-year program, you must submit a research project proposal, write a thesis and deliver an oral defense. You will also write abstracts, present posters and publish findings while building toward their thesis defense. Your research experience is guided by a research mentor, a thesis advisory committee and the program director. We encourage you to participate in the Summer Research Immersion Program as well.

Application deadline: Friday, April 15, 2022

Be sure to indicate your willingness to commit time, facilities and resources to a student as needed throughout the program.

Questions about the MRHP program or mentoring? Contact: Sonia Lobo, PhD Associate Dean for Research and Scholarship slobo@som.geisinger.edu Adam Blannard, MS Program Manager ablannard@som.geisinger.edu

260

Cover art submissions This year’s call for SCRIP cover art yielded several creative and noteworthy submissions from our talented students which are showcased below. The image chosen for the SCRIP cover was submitted by Elana Stains, MD Class of 2025, and features the snake plant, which is valued for its ability to remove toxic air pollutants such as benzene, formaldehyde, xylene, and toluene.

Sandybell Anorga, MBS Class Cohort 3 Right: The picture is a digital work created with a drawing tablet in watercolor form on the female human skeleton. Learning anatomy has been one of my favorite subjects to learn, especially the skeleton system and drawing concepts at home.

My Nguyen, MD Class of 2025 Below left: This is the two-photon image of the presynaptic input from the basal forebrain to PV interneurons in the frontal cortex. Two-photon imaging is a cutting-edge research technique to investigate neuronal activities for various purposes, such as impaired decision-making and compulsive behaviors in substance use disorder. This represents the combined beauty of both science and art, as neurons appear like flickering stars in the universe of neural circuits.

Kinza Abbas, MD Class of 2025 Below middle: 10x objective tile images of brain slices of wildtype and homozygous mutant mice. Stained for nuclei with Hoechst dye (blue) and OPA1 using an OPA1 primary antibody (NovusBio) and Alexa Fluorescence 555 secondary antibody (Invitrogen).

Tice R. Harkins, MD Class of 2024 Below right: This image of a bovine blood clot was taken using a super resolution structured illumination microscope (SIM). Platelets were fluorescently tagged green and fibrin was tagged red.

261

525 Pine St. Scranton, PA 18509

570-504-7000

geisinger.edu/gcsom StudentResearch@som.geisinger.edu

From left: Mark Brown, Bianca Sanchez, and Wanyan Ma received Excellence in Research Awards for their outstanding abstract submissions at the 2021 Summer Research Symposium. 31055-11/21-HDAV/SL

SCRIP: Scholarly Research In Progress 2021

Articles inside

Cover art submissions

Finding your way

Medical Research Honors Program

Summer Research Immersion Program

Investigating Potential Conflicts of Interest

Differences in Multiple and Single-Drug Arrests by the Maine Diversion Alert Program (DAP)

A Novel Approach to Chest Wall Reconstruction Following Intrathoracic Scapular Dislocation

Pronounced Declines in Licit Fentanyl Utilization and Changes in Prescribing and Reimbursement Practices in the United States, 2010-2019

Racial Differences in Insurance Type between Diabetes Mellitus Type 2 Patients in the United States

The Relationship Between Treatment Center Services and Number of Opioid- related Deaths in the United States Before and After a Declaration of a National Opioid Crisis

Treatment of Pediatric Lisfranc Injuries: A Systematic Review and Introduction of a Novel Treatment Algorithm

Electroconvulsive Therapy Uses and Its Ability to Induce Neurogenesis: A Literature Review

Improving the Future of the Opioid Epidemic: Methocinnamox

Examining Health Literacy and Health Outcomes Among United States Immigrants and Non-Immigrants

The Changing Use of Opioids in the U.S from 2017 to Early 2020

Characterizing the Behavioral and Cellular Effects of the R904S Variant of OPA1 as a Tourette Disorder Probable Risk Gene

Why is There a Ten-Fold Variation between States in Clozapine Usage among Medicaid Enrollees in the United States?

The Relationship between Food Deserts and the Prevalence of Type 2 Diabetes in Communities in Southeastern Pennsylvania

Review of Ketamine as a Rapid Antidepressant for Treatment- Resistant Depression

The Association Between Early Menarche and Gestational Diabetes: A Secondary Analysis

Environmental Influences on Childhood

Declines and Pronounced Regional Disparities in Prescription Opioids in the United States

The Role of Socioeconomic Factors

The Natural History of Genu Valgum in the Pediatric Obese Patient

Review of Selected Contemporary Treatments for Ischemic Stroke

An Examination of the Impact of COVID-19 on Black, Latino and Asian New York City Residents and the Factors Associated with the Social Determinants of Health

Effects of Medical Cannabis on Patients with Charcot-Marie-Tooth Disease

Short- and Long-Term Outcomes of Breastfeeding on Children’s Mental and Physical Health

Hunger vs. Heart Failure Can Food Insecurity Screening Prevent CHF Exacerbations?

Measuring the Change in Use of Generic Oxcarbazepine (OXC) Versus Brand Names for Medicaid Enrollees Throughout the United States in 2018-2019

Conflict of Interest Disclosure Accuracy

Exploring the Impact of Medicaid

A Review of the Literature: How Intestinal Microbiota Shape the Immune System and the Related Impact on Human Health and Autoimmune Disease

Chronic Traumatic Encephalopathy A Literature Review

Primary Ectopic Breast Carcinoma of the Vulva: A Case Report

Trends and Regional Differences in the Diversion of Stimulants in the United States, 2015–2019

Changes in Morphine Distribution in the United States

Hydroxychloroquine, Azithromycin, and Chloroquine Prescribing Patterns in Medicaid

Battling Trainee Biases and Reconstructing Perceptions in Global Neurology

Virtual Newborn Nursery Rounds: An Alternative Experience During the COVID-19 Pandemic

Ethics and Current Climates Surrounding HPV Vaccination

Rat-Bite Fever in a 14-Year-Old Male

Recent Trends in Gabapentin Usage Among Medicaid Patients

Geographic Variability in Antibiotic Prescribing Rates in Medicaid

Assessment and Improvement of Sepsis Bundle Compliance at Geisinger Community Medical Center

Conflicts of Interest Differ Among Male and

Preventability Associated with Elevated Maternal Mortalities Among Black Women

Declining National Codeine Distribution in United States Hospitals and Pharmacies

Pelvic Examinations Under Anesthesia (EUA) Informed Consent Policy