wenz iD - proefschrift Crystel M. Gijsberts

Page 1


Coronary Artery Disease: Ethnicity, Sex and Risk Prediction Copyright © Crystel M. Gijsberts 2015 ISBN: 978-90-393-6437-6 Cover illustration: Mart Veeken, INK strategy, www.inkstrategy.com Lay-out and Design:
wenz iD | www.wenzid.nl Printed by: Boxpress, www.proefschriftmaken.nl Financial support by the Dutch Heart Foundation for the publication of this thesis is gratefully acknowledged. Financial support by the Heart & Lung Foundation Utrecht for the publication of this thesis is gratefully acknowledged Additional financial support by Research ICT UMC Utrecht, Furore B.V., Chipsoft B.V., Pfizer B.V. and de Stichting Cardiovasculaire Biologie for the publication of this thesis is also gratefully acknowledged.


Coronary Artery Disease Ethnicity, Sex and Risk Prediction

Coronairlijden Etniciteit, Sekse en Risicopredictie

(met een samenvatting in het Nederlands)

Proefschrift ter verkrijging van de graad van doctor aan de Universiteit Utrecht op gezag van de rector magnificus, prof. dr. G.J. van der Zwaan, ingevolge het besluit van het college voor promoties in het openbaar te verdedigen op donderdag 17 december 2015 des middags te 2.30 uur door

Crystel MeriĂŤlle Gijsberts geboren op 11 maart 1987 te Nijmegen


Promotoren:

Prof. dr. D.P.V. de Kleijn Prof. dr. F.W. Asselbergs

Copromotoren: Dr. I.E. Hoefer Dr. H.M. den Ruijter


Table of Content Chapter 1

General Introduction

9

PART ONE Ethnicity Chapter 2

Biomarkers of Coronary Artery Disease Differ between Asians and Caucasians in the General Population Global Heart. 2015 Mar 7. Pii: S2211-8160(14)02671-4

21

Chapter 3

Race/Ethnic Differences in the Associations of the Framingham Risk Factors With Carotid IMT and Cardiovascular Events Plos One. 2015 Jul 2;10(7):E0132321

57

Chapter 4

Ethnicity Modifies Associations between Cardiovascular Risk Factors and Disease Severity in Parallel Dutch and Singapore Coronary Cohorts Plos One. 2015 Jul 6;10(7):E0132278

75

Chapter 5

Inter-Ethnic Differences in Quantified Coronary Artery Disease Severity and All-Cause Mortality Among Dutch and Singaporean Percutaneous Coronary Intervention Patients Plos One. 2015 Jul 6;10(7):E0131977

95

Chapter 6

The Ethnicity-Specific Association of Biomarkers With the Angiographic Severity of Coronary Artery Disease Accepted For Publication in the Netherlands Heart Journal

113

Chapter 7

Ethnic Differences in QRS Prolongation and its Association with Ejection Fraction and Outcomes in Heart Failure In Preparation

133

PART TWO Sex Differences Chapter 8

Severity of Stable Coronary Artery Disease and its Biomarkers Differ between Men and Women Undergoing Angiography Atherosclerosis. 2015 Jul;241(1):234-40

151


Chapter 9

Women Undergoing Coronary Angiography For Myocardial Infarction or who Present with Multivessel Disease have a Poorer Prognosis Than Men Angiology. 2015 Sep 7. Pii: 0003319715604762

169

Chapter 10

Sex Differences in Health-Related Quality of Life in Patients Undergoing Coronary Angiography Open Heart. 2015 Aug 27;2(1):E000231

191

Chapter 11

The Sex-Specific Relation of Health-Related Quality of Life With Adverse Events and Mortality in Peripheral and Coronary Artery Disease Patients In Preparation

215

PART THREE General Risk Prediction Chapter 12

Non-Response to Questionnaires Independently Predicts Mortality of Coronary Angiography Patients Int J Cardiol. 2015 Jul 2;201:168-170

233

Chapter 13

Routinely Analyzed Leukocyte Characteristics Improve Prediction of Mortality After Coronary Angiography Submitted

239

Chapter 14

Hematological Parameters Improve Prediction of Mortality and Secondary Adverse Events in Coronary Angiography Patients: A Longitudinal Cohort Study Accepted for publication in Medicine

261

Chapter 15

The Value of Hematological Parameters Exceeds High-Sensitivity Troponin I and NT Pro-BNP for Mortality Prediction in Coronary Angiography Patients In Preparation

285


Summary and Discussion Chapter 16

Summary and General Discussion

305

Chapter 17

Nederlandse Samenvatting

319

Appendix Review Committee Author Affiliations List of Publications Dankwoord Curriculum Vitae

330 331 336 339 344



Chapter 1 General Introduction Parts of this text have been published in the Netherlands Heart Journal.1


Chapter 1

Cardiovascular disease, with atherosclerosis as the underlying syndrome, is the major contributor to mortality worldwide (Figure 1).2 Atherosclerosis is the process of accumulation of lipids and inflammatory cells causing plaque formation in arteries supplying blood to the heart (Figure 2), brain or other organs.3 When plaques become so large that they limit blood flow, or when vulnerable plaques rupture or plaque erosion leads to a thrombotic occlusion, this leads to clinical events due to ischemia such as myocardial infarction (heart attack) or stroke. Ischemic heart disease ranks first among the cardiovascular diseases claiming 7.4 million deaths in 2012 worldwide.4 Patients complaining of chest pain or other symptoms raising suspicion of coronary artery disease (CAD) often undergo coronary angiography to evaluate stenoses or blockages of the coronary arteries (Figure 3). Despite improving non-invasive imaging modalities, coronary angiography remains the gold standard for visualizing and diagnosing CAD. CAD used to be a disease of White men. But with increasing Westernization and women adopting risky life style habits5 the epidemic has spread to people of all ethnicities and both sexes. In order to justly target the efforts of (secondary) prevention and treatment, accurate risk predictors and algorithms are key. Therefore, in this thesis we evaluated the characteristics of CAD in people of different ethnic backgrounds and compared CAD characteristics between the sexes. Furthermore, we sought for patient characteristics that could improve the prediction of risk of adverse events and mortality.

Figure 1. Global burden of deaths due to cardiovascular disease, as proportion of total number of non-communicable diseases. Reproduced from the World Health Organization with permission.

10

Figure 2. The heart is supplied with blood by the coronary arteries. Narrowing or blockages of the coronary arteries leads to ischemic heart disease.

Figure 3. Coronary angiography images. The top panel shows a stenosis in the left anterior descending coronary artery. The bottom panel shows a blocked right coronary artery.


Introduction

PART ONE Ethnicity Historically CAD has been the number one cause of death in the Western world.6 With improved prevention strategies and increasing treatment options the upward trend of CAD deaths has come to a halt in the US and mortality rates might even be declining.7 However, the incidence of CAD in developing parts of the world is high (Figure 4) and rising.2 The World Health Organization (WHO) predicted an increase in cardiovascular deaths of 10% by the year 2030 in Africa, Eastern Mediterranean regions and South-East Asia.2,8 For South-East Asia, the WHO projected 5 million cardiovascular deaths in the year 2020.2

Figure 4. World map showing age-adjusted numbers of death due to ischemic heart disease. Reproduced from the World Health Organization with permission.

While an Asian CAD epidemic is thus forthcoming, research on CAD is still dominated by predominantly White study populations and ethnicity-specific research is sorely needed.9 It is known that the prevalence of cardiovascular risk factors differs between certain ethnic groups. For example South Asians (i.e. Asians of Asian Indian ethnicity, e.g. from India, Pakistan or Bangladesh) are known to have an extremely high prevalence of diabetes10 and dyslipidemia11,12 and they develop CAD at very young ages.13,14 The incidence of CAD among South Asians exceeds that of other Asian ethnic groups.15 In contrast, Chinese are considered to have a more benign risk factor profile, with lower rates of diabetes and more favorable lipid profiles than Whites.16 Also, the prevalence of CAD is lower in Chinese than Whites.17

11


Chapter 1

Ethnic differences in the severity of CAD have been very sparsely addressed in the literature. The majority of studies addressing CAD severity used computed tomography coronary artery calcification (CT calcium) scores and not coronary angiography, which is considered the gold standard. Population-based studies have demonstrated that community-dwelling South Asians had higher CT calcium scores than Chinese.18,19 And South Asian angina patients had more severe CT calcium than Whites with similar risk factors.20 One study directly compared coronary angiography data between Chinese and Australians, concluding that Chinese have less severe CAD as quantified by the Gensini score.21 A major literature gap exists in the evaluation of angiographic CAD severity between the Asian and White coronary angiography population. As a consequence of differing risk factor profiles and CAD severity, biomarkers of CAD are bound to differ between Asians and Whites. Generally accepted biomarkers of CAD such as high sensitivity C-reactive protein and Cystatin C are bluntly applied to other populations than Whites, in which they were discovered and validated. It is unknown but unlikely that biomarker research derived from White populations can be generalized to other ethnicities without any modifications. As a consequence of better survival of acute CAD events (i.e. myocardial infarction) heart failure prevalence has surged over the last decades.22 Especially in Asia an alarmingly high prevalence of heart failure is seen. In Malaysia and Singapore 6.7% and 4.5% of the population, respectively, live with heart failure, compared to 1-2% in Europe and 1.9% in the US.23 An important prognostic marker in heart failure is QRS duration on the electrocardiogram, which is related to left ventricular function and shown to be predictive of mortality in a White heart failure population.24 It is unknown whether QRS duration has the same relation with left ventricular function in Asians as observed in Whites and whether QRS prolongation is predictive of future adverse events in a similar manner. Singapore provides an outstanding ground for multi-ethnic research1 on heart disease due to its abundance of Asian ethnicities (Chinese, Indian and Malay) and a Western health care system.25 In the first part of this thesis we aimed to fill gaps in the current literature by comparing Asian and White CAD and heart failure patients. First, in chapter 2 we reviewed the literature for ethnic differences in CAD biomarker levels in the general population. Then, in chapter 3 we examined whether the effect of established risk factors in the general population differs by ethnicity. In chapter 4 we continue with a description of the United Coronary Biobank (UNICORN) cohort, a collaborative study between the Netherlands and Singapore. In this study we describe the ethnicity-specific association of risk factors with the angiographic severity of CAD. More in-depth, in chapter 5, we assessed ethnic differences in the quantified burden of CAD on coronary angiography by means of the highly detailed SYNTAX score26 in Dutch and Singaporean populations of percutaneous coronary intervention patients. In chapter 6 we sought for ethnicity-dependent associations of CAD biomarker levels with the severity of CAD and evaluated biomarker level cut-offs among the ethnic groups. Finally, in chapter 7 we looked into the ethnicityspecific relation of QRS duration on electrocardiography with the impairment of left ventricular ejection fraction in heart failure patients.

12


Introduction

PART TWO Sex differences In Europe, more women than men die of cardiovascular causes each year.2,27 However, women are older when CAD unfolds, assumedly due to hormonal protection until menopause.28 Recently, female-specific risk factors related to pregnancy and hormonal status have emerged in order to improve risk prediction for women.29 But traditional risk factors also contribute to CAD risk in women. Some risk factors, such as smoking30 and diabetes31 even confer more risk in women than in men. Women develop more “stable” atherosclerosis compared to men32 and are more likely to have plaque erosion as compared to plaque rupture33 as the underlying substrate for sudden death and myocardial damage. Microvascular disease - atherosclerosis of the microvasculature of the heart - is a common condition in women with angina34, as opposed to macrovascular or epicardial disease which is more often found in men. These differences in CAD phenotype are likely to be reflected in biomarker patterns as well. Established biomarkers: N-terminal pro-Brain Natriuretic Peptide, associated with ventricular dilatation and pressure overload35–37; high-sensitivity C-Reactive Protein38,39, a marker of inflammation; Cystatin C40–43, a marker of renal dysfunction; myeloperoxidase, linked to both inflammation and oxidative stress44–46; high-sensitivity Troponin I47–49, reflecting myocardial ischemia and von Willebrand factor50, a coagulation factor might well have different relations with the angiographic severity of CAD in women and men. Recently in the US, a concerning increase in mortality is described for women of relatively young age.51,52 Among these women, increasing rates of diabetes and obesity possibly nullify the effects of the reduced smoking prevalence and improved hypertension treatment.53 Contrary to the US however, in the European Union luckily, no increase in mortality rates has been observed among women; only plateauing of these rates occurred in a minority of the EU countries among younger individuals of both sexes.53 In the Netherlands, sex differences in prognosis after coronary angiography have not recently been examined. In addition to the biological sex differences in CAD, there are striking sex differences in the psychological and social impact of CAD on patients’ lives. Women who undergo percutaneous coronary intervention54,55 or coronary artery bypass grafting56 report lower health-related quality of life (HRQOL) - an indicator of general well-being - than men. As poor HRQOL is related to higher health care expenditure57, determinants of poor HRQOL need to be elucidated. Risk factors, such as obesity58,59, diabetes60 and smoking61 have been linked to a diminished HRQOL. But it is unknown whether their relation with HRQOL is similar in men and women. Perhaps, female-specific risk factors could explain in part the lower HRQOL reported by women with CAD. Also, we do not know if the difference in HRQOL between men and women depends on the severity of CAD and whether the low reported HRQOL is related to a worse prognosis. In chapter 8 we assessed the severity of CAD between men and women in detail and evaluated the sex-specific association of CAD biomarkers with the severity of CAD.

13


Chapter 1

In chapter 9 we describe the occurrence of adverse events and mortality in men and women after coronary angiography. Then in chapter 10, we looked into sex differences in HRQOL between men and women undergoing coronary angiography. In chapter 11 we continue with the evaluation of HRQOL as a risk factor for adverse events and mortality in men and women with peripheral arterial disease and CAD. In this chapter we also examined female-specific risk factors as possible determinants of poor HRQOL in women.

PART THREE Risk prediction Risk predictors come in many shapes and sizes. Any characteristic attributable to a person could serve as a predictor, as long as it is related to the outcome of interest. In the third part of this thesis we discuss two very different entities of predictors: patient responsiveness to questionnaires and hematological parameters. A substantial part of cardiovascular research is performed through patient questionnaires. Patients who do not respond to these questionnaires are logically excluded from the study. In a cohort of coronary angiography patients we evaluated the impact of nonresponsiveness to HRQOL questionnaires on mortality. The results of this study are presented in chapter 12. Several hematological parameters have been mentioned in the context of mortality prediction in CAD patients. For example, white blood cell counts62,63, neutrophil counts64–66, monocyte counts65 and lymphocyte percentages65 have been reported to associate with CAD presence and severity and/or death. And recently, high red blood cell distribution width (RDW), a measure of the variation of red blood cell size has emerged as a biomarker of atherosclerosis progression67, CAD severity68 and mortality69. Modern automated blood cell analyzers (e.g. the Abbott Sapphire70) not only report the parameters requested by the physician, but all hematological parameters that the machine is capable of measuring. For example, when a physician requests a hemoglobin measurement, the analyzer also automatically determines the platelet count and leukocyte differentiation. Although these values are not reported to the clinician, the analyzer stores the data. In the University Medical Centre in Utrecht data from hematology analyzers is stored in the Utrecht Patient Oriented Database 71. Thus, much more information on hematological characteristics is available than is reported back to the clinic and these routine measurements might serve as potent prognostic biomarkers of cardiovascular disease.72 In the final part of this thesis we evaluated leukocyte parameters in chapter 13 and a panel out of all hematological parameters (chapter 14) as prognostic markers for adverse events and mortality. In the final chapter, chapter 15, we compare the prognostic performance of hematological parameters with general CAD biomarkers: high-sensitivity Troponin I and N-Terminal pro-Brain Natriuretic Peptide.

14


Introduction

References 1.

Wang JW, Gijsberts CM, Seneviratna A, de Hoog VC, Vrijenhoek JEP, Schoneveld a H, Chan MY, Lam CSP, Richards a M, Lee CN, Mosterd A, Sze SK, Timmers L, Lim SK, Pasterkamp G, de Kleijn DP V. Plasma extracellular vesicle protein content for diagnosis and prognosis of global cardiovascular disease. Neth Heart J. 2013;21:467–71.

2.

Alwan A. Global status report on noncommunicable diseases 2010. 1st ed. Geneva, Switzerland: 2011.

3.

Ross R. Atherosclerosis--an inflammatory disease. N Engl J Med. 1999;340:115–26.

4.

World Health Organization. Causes of Death 2012 [Internet]. 2012;Available from: http://www.who.int/ mediacentre/factsheets/fs310/en/

5.

Shaw LJ, Bairey Merz CN, Pepine CJ, Reis SE, Bittner V, Kelsey SF, Olson M, Johnson BD, Mankad S, Sharaf BL, Rogers WJ, Wessel TR, Arant CB, Pohost GM, Lerman A, Quyyumi AA, Sopko G. Insights from the NHLBI-Sponsored Women’s Ischemia Syndrome Evaluation (WISE) Study: Part I: gender differences in traditional and novel risk factors, symptom evaluation, and gender-optimized diagnostic strategies. J Am Coll Cardiol. 2006;47:S4–S20.

6.

Lloyd-Jones D, Adams RJ, Brown TM, Carnethon M, Dai S, De Simone G, Ferguson TB, Ford E, Furie K, Gillespie C, Go A, Greenlund K, Haase N, Hailpern S, Ho PM, Howard V, Kissela B, Kittner S, Lackland D, Lisabeth L, Marelli A, McDermott MM, Meigs J, Mozaffarian D, Mussolino M, Nichol G, Roger VL, Rosamond W, Sacco R, Sorlie P, Stafford R, Thom T, Wasserthiel-Smoller S, Wong ND, Wylie-Rosett J. Heart disease and stroke statistics--2010 update: a report from the American Heart Association. Circulation. 2010;121:e46–e215.

7.

Ford ES, Ajani U a, Croft JB, Critchley J a, Labarthe DR, Kottke TE, Giles WH, Capewell S. Explaining the decrease in U.S. deaths from coronary disease, 1980-2000. N Engl J Med. 2007;356:2388–98.

8. 9.

WHO. World Health Statistics 2011. WHO Libr Cat Data. 2011;1:170. Goff DC, Lloyd-Jones DM, Bennett G, Coady S, D’Agostino RB, Gibbons R, Greenland P, Lackland DT, Levy D, O’Donnell CJ, Robinson J, Schwartz JS, Shero ST, Smith SC, Sorlie P, Stone NJ, Wilson PWF. 2013 ACC/ AHA Guideline on the Assessment of Cardiovascular Risk: A Report of the American College of Cardiology/ American Heart Association Task Force on Practice Guidelines. Circulation. 2013;

10. Tziomalos K, Weerasinghe CN, Mikhailidis DP, Seifalian AM. Vascular risk factors in South Asians. Int J Cardiol. 2008;128:5–16. 11. Klomp M, Damman P, Beijk M a M, Tan KH, Balian V, de Luca G, Tijssen JGP, Silber S, de Winter RJ. Differences in cardiovascular risk factors and clinical outcomes between Western European and Southeast Asian patients treated with the Genous Bio-engineered R stent: an e-HEALING worldwide registry substudy. Coron Artery Dis. 2012;23:271–7. 12. Enas EA, Garg A, Davidson MA, Nair VM, Huet BA, Yusuf S. Coronary heart disease and its risk factors in firstgeneration immigrant Asian Indians to the United States of America. Indian Heart J. 48:343–353. 13. Khan N, Grubisic M, Hemmelgarn B, Humphries K, King KM, Quan H. Outcomes after acute myocardial infarction in South Asian, Chinese, and white patients. Circulation. 2010;122:1570–1577. 14. Wong CP, Loh SY, Loh KK, Ong PJL, Foo D, Ho HH. Acute myocardial infarction: Clinical features and outcomes in young adults in Singapore. World J Cardiol. 2012;4:206–10. 15. Mak K-H, Chia K-S, Kark JD, Chua T, Tan C, Foong B-H, Lim Y-L, Chew S-K. Ethnic differences in acute myocardial infarction in Singapore. Eur Heart J. 2003;24:151–60. 16. Zheng Y, Ma W, Zeng Y, Liu J, Ye S, Chen S, Lan L, Erbel R, Liu Q. Comparative study of clinical characteristics between Chinese Han and German Caucasian patients with coronary heart disease. Clin Res Cardiol. 2010;99:45–50.

15


Chapter 1

17. Anand SS, Yusuf S, Vuksan V, Devanesen S, Teo KK, Montague PA, Kelemen L, Yi C, Lonn E, Gerstein H, Hegele RA, McQueen M. Differences in risk factors, atherosclerosis, and cardiovascular disease between ethnic groups in Canada: the Study of Health Assessment and Risk in Ethnic groups (SHARE). Lancet. 2000;356:279–84. 18. Kanaya AM, Kandula NR, Ewing SK, Herrington D, Liu K, Blaha MJ, Srivastava S, Dave SS, Budoff MJ. Comparing coronary artery calcium among U.S. South Asians with four racial/ethnic groups: the MASALA and MESA studies. Atherosclerosis. 2014;234:102–7. 19. Bild DE, Detrano R, Peterson D, Guerci A, Liu K, Shahar E, Ouyang P, Jackson S, Saad MF. Ethnic differences in coronary calcification: the Multi-Ethnic Study of Atherosclerosis (MESA). Circulation. 2005;111:1313–20. 20. Koulaouzidis G, Nicoll R, Charisopoulou D, McArthur T, Jenkins PJ, Henein MY. Aggressive and diffuse coronary calcification in South Asian angina patients compared to Caucasians with similar risk factors. Int J Cardiol. 2013;167:2472–6. 21. Jiang S, Lv L, Juergens CP, Chen S, Xu D, Huang Z. Racial differences in coronary artery lesions: a comparison of coronary artery lesions between mainland Chinese and Australian patients. Angiology. 2008;59:442–7. 22. Bui AL, Horwich TB, Fonarow GC. Epidemiology and risk profile of heart failure. Nat Rev Cardiol. 2011;8:30–41. 23. Lam CSP. Heart failure in Southeast Asia: facts and numbers. ESC Hear Fail. 2015;2:46–49. 24. Lund LH, Jurga J, Edner M, Benson L, Dahlström U, Linde C, Alehagen U. Prevalence, correlates, and prognostic significance of QRS prolongation in heart failure with reduced and preserved ejection fraction. Eur Heart J. 2013;34:529–539. 25. World Health Organization. Health systems: Improving performance. In: The world health report 2000. Geneva, Switzerland: 2000. 26. Sianos G, Morel M, Kappetein AP, Morice M, Colombo A, Dawkins K, van den Brand M, Van Dyck N, Russell ME, Mohr FW, Serruys PW. The SYNTAX Score: an angiographic tool grading the complexity of coronary artery disease. EuroIntervention. 2005;1:219–27. 27. Appelman Y, van Rijn BB, Ten Haaf ME, Boersma E, Peters S a E. Sex differences in cardiovascular risk factors and disease prevention. Atherosclerosis. 2015;1–8. 28. Harvey RE, Coffman KE, Miller VM. Women-specific factors to consider in risk, diagnosis and treatment of cardiovascular disease. Womens Health (Lond Engl). 2015;11:239–57. 29. Sattar N, Greer IA. Pregnancy complications and maternal cardiovascular risk: opportunities for intervention and screening? BMJ. 2002;325:157–160. 30. Grundtvig M, Hagen TP, German M, Reikvam A. Sex-based differences in premature first myocardial infarction caused by smoking: twice as many years lost by women as by men. Eur J Cardiovasc Prev Rehabil. 2009;16:174–179. 31. Huxley R, Barzi F, Woodward M. Excess risk of fatal coronary heart disease associated with diabetes in men and women: meta-analysis of 37 prospective cohort studies. BMJ. 2006;332:73–78. 32. Hellings WE, Pasterkamp G, Verhoeven BAN, De Kleijn DP V, De Vries J-PPM, Seldenrijk KA, van den Broek T, Moll FL. Gender-associated differences in plaque phenotype of patients undergoing carotid endarterectomy. J Vasc Surg. 2007;45:289–96; discussion 296–7. 33. Virmani R, Burke AP, Farb A, Kolodgie FD. Pathology of the vulnerable plaque. J Am Coll Cardiol. 2006;47:C13–8. 34. Pepine CJ, Anderson RD, Sharaf BL, Reis SE, Smith KM, Handberg EM, Johnson BD, Sopko G, Bairey Merz CN. Coronary Microvascular Reactivity to Adenosine Predicts Adverse Outcome in Women Evaluated for Suspected Ischemia. Results From the National Heart, Lung and Blood Institute WISE (Women’s Ischemia Syndrome Evaluation) Study. J Am Coll Cardiol. 2010;55:2825–2832.

16


Introduction

35. Abdullah SM, Khera A, Das SR, Stanek HG, Canham RM, Chung AK, Morrow DA, Drazner MH, McGuire DK, de Lemos JA. Relation of coronary atherosclerosis determined by electron beam computed tomography and plasma levels of n-terminal pro-brain natriuretic peptide in a multiethnic population-based sample (the Dallas Heart Study). Am J Cardiol. 2005;96:1284–9. 36. Goyal BM, Sharma SM, Walia M. B-type natriuretic peptide levels predict extent and severity of coronary artery disease in non-ST elevation acute coronary syndrome and normal left ventricular function. Indian Heart J. 2014;66:183–7. 37. Kurtul A, Yarlioglues M, Murat SN, Duran M, Acikgoz SK, Sensoy B, Ergun G, Cetin M, Ornek E. N-Terminal Pro-Brain Natriuretic Peptide Level is Associated With Severity and Complexity of Coronary Atherosclerosis in Patients With Acute Coronary Syndrome. Clin Appl Thromb Hemost. 2014; 38. Wensley F, Gao P, Burgess S, Kaptoge S, Di Angelantonio E, Shah T, Engert JC, Clarke R, Davey-Smith G, Nordestgaard BG, Saleheen D, Samani NJ, Sandhu M, Anand S, Pepys MB, Smeeth L, Whittaker J, Casas JP, Thompson SG, Hingorani AD, Danesh J. Association between C reactive protein and coronary heart disease: mendelian randomisation analysis based on individual participant data. BMJ. 2011;342:d548. 39. Arroyo-Espliguero R, Avanzas P, Quiles J, Kaski JC. Predictive value of coronary artery stenoses and C-reactive protein levels in patients with stable coronary artery disease. Atherosclerosis. 2009;204:239–43. 40. Dandana A, Gammoudi I, Chalghoum A, Chahed H, Addad F, Ferchichi S, Miled A. Clinical utility of serum cystatin C in predicting coronary artery disease in patients without chronic kidney disease. J Clin Lab Anal. 2014;28:191–7. 41. Imai A, Komatsu S, Ohara T, Kamata T, Yoshida J, Miyaji K, Shimizu Y, Takewa M, Hirayama A, Deshpande GA, Takahashi O, Kodama K. Serum cystatin C is associated with early stage coronary atherosclerotic plaque morphology on multidetector computed tomography. Atherosclerosis. 2011;218:350–5. 42. Koc M, Batur MK, Karaarslan O, Abali G. Clinical utility of serum cystatin C in predicting coronary artery disease. Cardiol J. 2010;17:374–80. 43. Luc G, Bard J-M, Lesueur C, Arveiler D, Evans A, Amouyel P, Ferrieres J, Juhan-Vague I, Fruchart J-C, Ducimetiere P. Plasma cystatin-C and development of coronary heart disease: The PRIME Study. Atherosclerosis. 2006;185:375–80. 44. Buffon A, Biasucci LM, Liuzzo G, D’Onofrio G, Crea F, Maseri A. Widespread coronary inflammation in unstable angina. N Engl J Med. 2002;347:5–12. 45. Wainstein R V, Wainstein M V, Ribeiro JP, Dornelles L V, Tozzati P, Ashton-Prolla P, Ewald IP, Vietta G, Polanczyk CA. Association between myeloperoxidase polymorphisms and its plasma levels with severity of coronary artery disease. Clin Biochem. 2010;43:57–62. 46. Baldus S, Heeschen C, Meinertz T, Zeiher AM, Eiserich JP, Münzel T, Simoons ML, Hamm CW. Myeloperoxidase serum levels predict risk in patients with acute coronary syndromes. Circulation. 2003;108:1440–5. 47. Saunders JT, Nambi V, de Lemos JA, Chambless LE, Virani SS, Boerwinkle E, Hoogeveen RC, Liu X, Astor BC, Mosley TH, Folsom AR, Heiss G, Coresh J, Ballantyne CM. Cardiac troponin T measured by a highly sensitive assay predicts coronary heart disease, heart failure, and mortality in the Atherosclerosis Risk in Communities Study. Circulation. 2011;123:1367–76. 48. Ndrepepa G, Braun S, Schulz S, Mehilli J, Schömig A, Kastrati A. High-sensitivity troponin T level and angiographic severity of coronary artery disease. Am J Cardiol. 2011;108:639–43. 49. Laufer EM, Mingels AM a, Winkens MHM, Joosen I a PG, Schellings MWM, Leiner T, Wildberger JE, Narula J, Van Dieijen-Visser MP, Hofstra L. The extent of coronary atherosclerosis is associated with increasing circulating levels of high sensitive cardiac troponin T. Arterioscler Thromb Vasc Biol. 2010;30:1269–75.

17


Chapter 1

50. Van Loon JE, Kavousi M, Leebeek FWG, Felix JF, Hofman a, Witteman JCM, de Maat MPM. von Willebrand factor plasma levels, genetic variations and coronary heart disease in an older population. J Thromb Haemost. 2012;10:1262–9. 51. Mosca L. Fifteen-Year Trends in Awareness of Heart Disease in Women: Results of a 2012 American Heart Association National Survey. AHA. 2013;127:1–21. 52. Mosca L, Barrett-Connor E, Wenger NK. Sex/gender differences in cardiovascular disease prevention: what a difference a decade makes. Circulation. 2011;124:2145–54. 53. Nichols M, Townsend N, Scarborough P, Rayner M. Trends in age-specific coronary heart disease mortality in the European Union over three decades: 1980-2009. Eur Heart J. 2013;34:3017–27. 54. Norris CM, Spertus J a, Jensen L, Johnson J, Hegadoren KM, Ghali W a. Sex and gender discrepancies in health-related quality of life outcomes among patients with established coronary artery disease. Circ Cardiovasc Qual Outcomes. 2008;1:123–30. 55. Bakhai A, Ferrières J, James S, Iñiguez A, Mohácsi A, Pavlides G, Belger M, Norrbacka K, Sartral M. Treatment, outcomes, costs, and quality of life of women and men with acute coronary syndromes who have undergone percutaneous coronary intervention: results from the antiplatelet therapy observational registry. Postgrad Med. 2013;125:100–7. 56. Kendel F, Dunkel A, Müller-Tasch T, Steinberg K, Lehmkuhl E, Hetzer R, Regitz-Zagrosek V. Gender differences in health-related quality of life after coronary bypass surgery: results from a 1-year follow-up in propensitymatched men and women. Psychosom Med. 2011;73:280–5. 57. Harrison PL, Pope JE, Coberley CR, Rula EY. Evaluation of the relationship between individual well-being and future health care utilization and cost. Popul Health Manag. 2012;15:325–30. 58. Hlatky MA, Chung SC, Escobedo J, Hillegass WB, Melsop K, Rogers W, Brooks MM. The effect of obesity on quality of life in patients with diabetes and coronary artery disease. Am Heart J. 2010;159:292–300. 59. Oreopoulos A, Padwal R, McAlister FA, Ezekowitz J, Sharma AM, Kalantar-Zadeh K, Fonarow GC, Norris CM. Association between obesity and health-related quality of life in patients with coronary artery disease. Int J Obes (Lond). 2010;34:1434–41. 60. Uchmanowicz I, Loboz-Grudzien K, Jankowska-Polanska B, Sokalski L. Influence of diabetes on healthrelated quality of life results in patients with acute coronary syndrome treated with coronary angioplasty. Acta Diabetol. 2013;50:217–25. 61. Stafford L, Berk M, Jackson HJ. Tobacco smoking predicts depression and poorer quality of life in heart disease. BMC Cardiovasc Disord. 2013;13:35. 62. Madjid M, Awan I, Willerson JT, Casscells SW. Leukocyte count and coronary heart disease: implications for risk assessment. J Am Coll Cardiol. 2004;44:1945–56. 63. Lee C Do, Folsom AR, Nieto FJ, Chambless LE, Shahar E, Wolfe DA. White blood cell count and incidence of coronary heart disease and ischemic stroke and mortality from cardiovascular disease in African-American and White men and women: atherosclerosis risk in communities study. Am J Epidemiol. 2001;154:758–64. 64. Guasti L, Dentali F, Castiglioni L, Maroni L, Marino F, Squizzato A, Ageno W, Gianni M, Gaudio G, Grandi AM, Cosentino M, Venco A. Neutrophils and clinical outcomes in patients with acute coronary syndromes and/ or cardiac revascularization: A systematic review on more than 34,000 subjects. Thromb Haemost. 2011;106:591–599. 65. Kato A, Takita T, Furuhashi M, Maruyama Y, Kumagai H, Hishida A. Blood monocyte count is a predictor of total and cardiovascular mortality in hemodialysis patients. Nephron Clin Pract. 2008;110:c235–43.

18


Introduction

66. Gillum RF, Mussolino ME, Madans JH. Counts of neutrophils, lymphocytes, and monocytes, cause-specific mortality and coronary heart disease: the NHANES-I epidemiologic follow-up study. Ann Epidemiol. 2005;15:266–71. 67. Lappegård J, Ellingsen TS, Vik A, Skjelbakken T, Brox J, Mathiesen EB, Johnsen SH, Brækkan SK, Hansen J-B. Red cell distribution width and carotid atherosclerosis progression. Thromb Haemost. 2015;113:649–654. 68. Sahin O, Akpek M, Sarli B, Baktir AO, Savas G, Karadavut S, Elcik D, Saglam H, Kaya MG, Arinc H. Association of Red Blood Cell Distribution Width Levels with Severity of Coronary Artery Disease in Patients with Non-ST Elevation Myocardial Infarction. Med Princ Pract. 2014;24:178–183. 69. Perlstein TS, Weuve J, Pfeffer M a, Beckman J a. Red blood cell distribution width and mortality risk in a community-based prospective cohort. Arch Intern Med. 2009;169:588–594. 70. Müller R, Mellors I, Johannessen B, Aarsand AK, Kiefer P, Hardy J, Kendall R, Scott CS. European multi-center evaluation of the Abbott Cell-Dyn sapphire hematology analyzer. Lab Hematol. 2006;12:15–31. 71. Ten Berg MJ, Huisman A, van den Bemt PML a, Schobben AF a M, Egberts ACG, van Solinge WW. Linking laboratory and medication data: new opportunities for pharmacoepidemiological research. Clin Chem Lab Med. 2007;45:13–9. 72. ó Hartaigh B, Bosch J a., Thomas GN, Lord JM, Pilz S, Loerbroks A, Kleber ME, Grammer TB, Fischer JE, Boehm BO, März W. Which leukocyte subsets predict cardiovascular mortality? From the LUdwigshafen RIsk and Cardiovascular Health (LURIC) Study. Atherosclerosis. 2012;224:161–169.

19



PART ONE Ethnicity

Chapter 2 Biomarkers of Coronary Artery Disease Differ between Asians and Caucasians in the General Population Global Heart, 2015 March 27, Epub ahead of print

Crystel M. Gijsberts, Hester M. den Ruijter, Folkert W. Asselbergs, Mark Y Chan, Dominique P.V. de Kleijn, Imo E. Hoefer


Chapter 2

ABSTRACT Coronary artery disease (CAD) markers have not been thoroughly investigated among Asians. The incidence of CAD, however, is rising rapidly in Asia. In this review, we systematically discuss publications that compare CAD biomarkers between Asians and Caucasians in the general population. A PubMed search yielded 5,570 hits, containing 59 articles describing 47 unique cohorts that directly compare Asians with Caucasians. Ten biomarkers were taken into account for this review; total cholesterol, triglycerides, HDL-cholesterol, LDL-cholesterol, C-reactive protein, glucose, insulin, glycated hemoglobin, fibrinogen and plasminogen activator inhibitor-1. Triglycerides were 1.13-fold higher in South Asians than in Caucasians and insulin levels 1.33-fold higher. In Japanese and Chinese lower CRP levels were reported; 0.52 and 0.36-fold, respectively. Ethnicity-specific prognostic measures of CAD biomarkers were rarely reported. CAD biomarker levels differ between Asians and Caucasians and among Asian ethnic groups in population-based cohorts. The ethnicity-specific prognostic value of CAD biomarkers is yet to be determined.

22


Systematic Review: CAD biomarkers in Asians and Caucasians

INTRODUCTION Coronary artery disease (CAD) is the number one cause of death worldwide and will remain so for the next decades. In contrast to the declining numbers in the Western world1, its incidence is expected to increase in other parts of the world, predominantly in Asia.2 In our efforts to tackle this upcoming epidemic in Asia, primary prevention will have to play a central role, requiring reliable and robust predictive biomarkers. To date, most CAD biomarker research has been conducted solely in Caucasian subjects. For example, the pivotal heart disease study, the Framingham Heart Study, consists almost exclusively of European descendants.3 Based on little evidence, current risk markers such as C-reactive protein (CRP), LDL-cholesterol (LDL) or HDL-cholesterol (HDL) are generally accepted to be valid for other ethnic groups than Caucasians, in which the biomarkers were validated. In fact, little is known about the actual generalizability of these markers towards other ethnic groups. However, previous studies already indicated significant differences in the traditional risk factors and CAD prevalence between Asians and Caucasians, making it unlikely that current biomarkers can simply be generalized to other ethnicities.4–6 Probably, different cut-off values and prognostic measures apply to the various populations and should be implemented accordingly. The pressing need for ethnicity-specific CVD research has recently been underlined by the American Heart Association.7 Work has been done by, for example Tillin et al.8 on recalibration of clinical risk scores as to serve other ethnic groups apart from Caucasians, but the value of CAD biomarkers in the Asian ethnic groups is largely unknown. In an attempt to gather the available data on differences in CAD biomarkers between Asians and Caucasians at the general population level, we hereby provide a comprehensive overview on the current knowledge on biomarker levels and (when available) their predictive value for CAD in these populations. Adaptation to a Western lifestyle, a phenomenon known as acculturation has a major effect on the cardiovascular risk of ethnic groups.9,10 Indicating that there is an important environmental factor, on top of a possible true biological component, explaining interethnic differences. In order to reduce bias by environmental differences, the main results in this review are based on data from Asians in diaspora countries only (encompassing the majority of the available literature).

METHODS Search criteria This review was conducted in accordance with the PRISMA guidelines.11 We created a search syntax based on Asian and Caucasian ethnicity, biomarkers and CAD, and their synonyms (table 1). PubMed MEDLINE was systematically searched for publications meeting our inclusion criteria: adult population-based cohorts including at least one Asian ethnic group and a Caucasian ethnic group. Also, at least one blood-derived

23


Chapter 2

biomarker had to be measured in a context of CAD. Titles and abstracts were screened (by CG) and publications providing levels of CAD biomarkers in Asians and Caucasians were selected for this review. Only studies that directly compared Asian individuals with Caucasian individuals were selected to ensure comparability of the measured biomarker levels within the study. Articles that solely report the presence or absence rather than the actual level of a particular risk factor or biomarker were excluded. Furthermore, studies that did not determine ethnicity on an individual basis and studies concerning children, CAD patients, or case-control studies were excluded, as they do not reflect the general population. The Asian ethnic groups could be further divided into South Asians (i.e. Asians originating from India, Pakistan, Bangladesh or Sri Lanka), Chinese and Japanese. Extraction of reported measures From all included articles we extracted the cohort characteristics, the biomarker levels, and (when available) the prognostic measures of a biomarker per ethnic group. When one specific cohort was described in more than one article -as is frequently the case for large cohorts- the most extensive dataset on a particular biomarker was used, i.e. the publication with the largest number of subjects or (when dataset size was the same) the most recent publication. To create a measure of comparison relative differences between Asians and Caucasians are calculated within the studies, with the biomarker levels for Caucasians set at 1, For example, when the Glucose level in Caucasians is 8 and in South Asians it is 12, the relative difference will be 12/8=1.5. The average relative risks across the publications selected in this review are calculated by dividing the sum of the number of Asian and Caucasian individuals within the cohort Ă— the relative risk in that particular cohort by the total number of individuals compared across the different cohorts. Hereby, the relative differences are weighted for cohort size. Also, we calculated biomarker means per ethnic group, per comparison. Mean levels for Caucasians were thus calculated three times, as they are derived from different cohorts, comparing with different Asian ethnic groups.

RESULTS Search The search syntax was entered in the PubMed database on August 14th 2014 (table 1). This search retrieved 5,570 hits. These 5,570 publications were screened by title and abstract, which initially yielded 111 publications. After considering the 111 full text articles, 64 articles were excluded for the following reasons: only patients with established CAD were included (n=14), actual biomarker levels were not provided (n=16), no specific context of CAD (n=16), cohort was discussed more extensively in another article (n=12), article concerned only children (n=5), individual ethnicity was not reported (n=1).

24


Systematic Review: CAD biomarkers in Asians and Caucasians

Table 1. Search syntax as entered in PubMed on August 14th, 2014. Coronary artery disease

#1: (atherosclero*[Title/Abstract] OR coronar*[Title/Abstract] OR cardiovascular[Title/Abstract] OR cardiac[Title/Abstract] OR myocardi*[Title/Abstract])

Biomarker

AND

#2: (biomarker[Title/Abstract] OR biomarkers[Title/Abstract] OR marker[Title/Abstract] OR markers[Title/Abstract] OR level[Title/ Abstract] OR levels[Title/Abstract] OR concentration[Title/ Abstract] OR concentrations[Title/Abstract] OR enzyme[Title/ Abstract] OR enzymes[Title/Abstract] OR peptide[Title/Abstract] OR peptides[Title/Abstract] OR protein[Title/Abstract] OR proteins[Title/Abstract] OR metabolite[Title/Abstract] OR metabolites[Title/Abstract] OR diagnos*[Title/Abstract] OR prognos*[Title/Abstract] OR lipid[Title/Abstract] OR lipids[Title/ Abstract] OR risk factor*[Title/Abstract]) OR predict*[Title/ Abstract])

AND

#3: ((asia*[Title/Abstract] OR chin*[Title/Abstract] OR indi*[Title/ Abstract] OR Pakistan*[Title/Abstract] OR Banglades*[Title/ Abstract] OR Japan*[Title/Abstract]) AND (Caucasian*[Title/ Abstract] OR white*[Title/Abstract] OR europe*[Title/Abstract]))

Ethnicity

Search result

#1: 1,061,901 hits #2: 8,251,453 hits #3: 109,637 hits #1 AND #2 AND #3: 5,570 hits

The search process (figure 1) resulted in 47 titles that provided biomarker levels in community population Asians and Caucasians in the context of CAD. Reference checking did not retrieve any additional articles. Reported biomarkers In the 47 articles retrieved from the search, 52 distinct biomarkers were reported. Out of these, 31 biomarkers were reported only once, seven were reported twice, four were reported 3 times and ten biomarkers were reported more than five times. Only biomarkers that were reported five or more times were taken into account for this review. These were; triglycerides (TG) from 27 cohorts, HDL from 28 cohorts, total cholesterol (TC) from 27 cohorts, glucose from 22 cohorts, LDL from 17 cohorts, insulin from 18 cohorts, CRP from 11 cohorts, fibrinogen from 8 cohorts, glycated haemoglobin (HbA1c) from 8 cohorts and plasminogen activator inhibitor-1 (PAI-1) from 7 cohorts. Three articles that fulfilled the inclusion criteria12–14 reported none of the biomarkers that were reported more than five times and are therefore not further discussed. Table 2 summarizes the 44 articles (from 33 unique cohorts) that were selected for this review with their cohort characteristics. CAD prevalence The 33 cohorts have vastly heterogeneous properties, as displayed in table 2. One important parameter is the prevalence of established CAD or cardiovascular disease (CVD), because the presence of (sub)clinical CAD might be reflected in the biomarker

25


Chapter 2

levels. Of the 33 cohorts, 21 solely included persons free of CAD. Among the 12 remaining cohorts CAD/CVD prevalence was significantly different between the ethnic groups in three cohorts (described in four articles4,18,31,32). In the SHARE-cohort4 the prevalence of CVD has been described to be significantly higher in South Asians (10.7%) as compared to Caucasians (5.4%) and Chinese (2.4%). In the Middlesex cohort of Chowdhury et al.31 the prevalence of CVD (complaints) was higher in South Asians than in Caucasians (4.8% vs. 3.1%) and in the Tower Hamlets cohort of Chowdhury et al.32 the prevalence of CAD was also higher in South Asians than in Caucasians (14.1% vs. 7.2%, p<0.001).

Figure 1. Flow diagram depicting the literature search process.

26


Systematic Review: CAD biomarkers in Asians and Caucasians

In five cohorts19,20,23,33,46 the prevalence of CAD/CVD was not significantly different between Asians and Caucasians (all 5 cohorts compared South Asians to Caucasians). Four cohorts did not exclude persons with previous CAD/CVD and omitted to report prevalence.21,36,47,54 The CAD prevalence per cohort is summarized in the supplemental table 1. Ethnic groups Most frequently (29 cohorts), South Asians were compared with Caucasians. Chinese and Japanese were compared with Caucasians far less often, i.e. in five and three cohorts, respectively. Noteworthy, only four cohorts described biomarker levels of Asians residing in their country of origin. One article discussed South Asians living in India, and one article discussed Chinese living in China, one article discussed Japanese living in Japan and one article discussed South Asians and Chinese living in Singapore. The vast majority of biomarker levels in Asians were derived from immigrants of Asian origin in the UK and the US. Biomarkers Among the ten biomarkers discussed in this review, four were related to lipid metabolism (TC, HDL, LDL, TG), three to glucose metabolism (glucose, insulin, HbA1c), two to coagulation (fibrinogen, PAI-1) and CRP is a general inflammation marker. In almost every comparison the number of Caucasian individuals outnumbered the number of Asian individuals. Differences in biomarker levels are presented in relative differences, which are calculated from the data in the reviewed articles (as explained in detail in the methods section of this paper). The average relative differences across the reviewed articles are presented in table 3, for all cohorts and stratified by migrant status of the Asian ethnic group. An overview of the relative difference per article is presented in the supplemental table 2. Also, relative differences from cohorts reporting on immigrated Asians (the majority of cohorts) are visualized in figure 2. Differences in biomarker levels that are discussed in the results refer only to studies on migrated Asians. For the cohorts discussing migrants, the mean levels of the biomarkers per ethnic group, stratified by comparison, are shown in table 4. Do note that the ratio between the Caucasian and Asian values from table 4 does not correspond directly to the relative differences presented in table 3, because those are weighted by total size of the comparison (N Caucasians + N Asian ethnic group), while the means presented in table 4 are weighted by N in each ethnic group separately. Lipids Lipid levels were compared between South Asians and Caucasians in 13 (LDL) to 21 or 22 cohorts (HDL, TC, TG), containing 11,167 (LDL) to 22,248 (TC) South Asian individuals. The lipid-derived biomarkers show a mixed pattern for South Asians with - relatively to Caucasians - a slightly lower HDL level (0.92-fold), a higher TG level (1.13-fold), a slightly lower LDL level (0.94-fold) and similar TC levels (0.99-fold).

27


Chapter 2

Figure 2. Bubble plot of all cohorts that are discussed in the review, one bubble stands for one group of Asian individuals from one article. The relative difference for Caucasians is set at one and represented by the blue line at y=1. The area size of the bubble corresponds to the number of Asian individuals in the cohort. Abbreviations: TG= triglycerides, TC= total cholesterol, HDL= high-density lipoprotein, LDL= low-density lipoprotein, HbA1c= glycated haemoglobin, CRP= C-reactive protein, PAI1= plasminogen activator inhibitor-1. The relative difference for PAI-1 levels in South Asians from the article by Raji (2001) falls outside the y-axis range (relative difference is 4.96) and is therefore not shown.

28


Systematic Review: CAD biomarkers in Asians and Caucasians

For Chinese no striking differences concerning lipid profiles were seen when compared to Caucasians, besides a slightly higher TG level (1.08-fold). All four lipid biomarkers were measured in approximately 1,364 Chinese individuals. For Japanese a relative difference of 1.11 was noted for the HDL levels when compared to Caucasians, possibly indicating a favourable lipid profile (with other lipids showing comparable levels). The number of Japanese individuals was much lower (maximum 270 persons). Glucose metabolism Glucose metabolism biomarkers were mostly measured in cohorts comparing South Asians with Caucasians, with the number of South Asian individuals ranging from 3,858 (HbA1c) – 5,417 (glucose). The difference in insulin levels between South Asians and Caucasians is striking; a 1.33-fold higher, possibly indicating more insulin resistance in South Asians than in Caucasians. Also, HbA1c and glucose levels were slightly higher (both 1.06-fold) when compared to Caucasians. Glucose metabolism biomarkers were measured in one (HbA1c) to four (glucose, insulin) cohorts containing Chinese, with the number of Chinese individuals ranging from 306 (HbA1c) -1,351 (glucose). Overall these markers show comparable levels between Chinese and Caucasians with relative differences of 1.03 - 1.05. The number of Japanese individuals was 248 for both glucose and insulin levels. HbA1c was not compared between Japanese and Caucasians in any of the reviewed articles. A lower insulin level (relative difference 0.89) was noted in Japanese as compared to Caucasians, although this result is derived from only one cohort containing 248 Japanese individuals. CRP CRP levels were found to be a 1.22-fold higher in South Asians when comparing a total of 1,456 South Asians to 1,468 Caucasians from eight cohorts. CRP was markedly lower in Chinese and Japanese, with relative differences of 0.52 and 0.36 respectively. However, these relative differences are derived from very few articles (three containing Chinese and one containing Japanese). Coagulation In South Asians, fibrinogen levels are similar to those in Caucasians (1.01-fold), however PAI-1 levels were markedly higher (1.31-fold), which might indicate a less favourable pro-coagulant state. These markers were measured in ~800 South Asian individuals derived from four/five cohorts. For Chinese, comparable levels of coagulation biomarkers were found (fibrinogen 0.98-fold, PAI-1 0.95-fold). In Japanese, fibrinogen levels are somewhat lower (0.87-fold), with slightly lower PAI-1 levels (0.96-fold), possibly displaying a more favourable anti-coagulant state.

29


Chapter 2

Table 2. Summary of the characteristics of the cohorts included in this systematic review. Articles printed in grey are articles that describe a cohort that is described more extensively in another article. They add biomarker details but contain fewer persons or are from a less recent publication. UK: United Kingdom, US: United States of America, SG: Singapore, SA: South Africa, NR: not reported.

Author, year

Cohort name/city, N

Follow-up

No CVD

Community sampled

Women only

Men only

Diabetics only

Non-diabetics only

Non-smokers only

Other ethnic groups Country, N, age

No lipid lowering drugs

Cohort characteristics

Adler, 199815

UKPDS, 4591

Ajjan, 200716

West Yorkshire, 490

Kain, 200217

West Yorkshire, 200

Anand, 20004

SHARE, 985

Anand, 200418

SHARE, 951

Bathula, 201019

LOLIPOP, 300

Bellary, 201020

UKADS, 1978

Bhalodkar, 200421

Framingham Offspring Study, 1895

Bhopal, 200522

Newcastle Heart Project, 1877

Reynolds, 200623

Newcastle Heart Project, 100

Cappuccio, 200224

Wandsworth Heart and Stroke, 922

Cook, 200125

Wandsworth Heart and Stroke, 912

Miller, 200326

Wandsworth Heart and Stroke, 476

Miller, 200927

Wandsworth Heart and Stroke, 125

Chambers, 199928

London, 44

Chambers 200129

London, 1025

Chandalia, 200330

Dallas, 137

Chowdhury, 200231

London (Middlesex), 292

Chowdhury, 200632

London (Tower Hamlets), 2074

Forouhi, 200633

Southall & Brent, 3207

Gama, 200234

Wolverhampton, 1000

Chatha, 200235

Wolverhampton, 191

Heald, 200336

Manchester, 272

Iso, 199637

Akita & ARIC, 300

Kamath, 199938

Chicago, 83

Kanaya, 201439

MESA/MASALA, 4228

Bild, 200540

MESA, 3422

Veeranna 201241

MESA, 3113

Kelley-Hedgepeth, 200842

SWAN, 2010

Matthews, 200543

SWAN, 1879

Liew, 200344

Singapore, 30

Lyratzopoulos, 200545

Stockport, 71367

Miller, 198846

London, 143

Mulukutla, 200847

HeartSCORE & IndiaSCORE, 1298

Nagi, 199648

London/Arizona, 575

Okamura, 201349

ERA JUMP/ Post-WWII, 607

30


Systematic Review: CAD biomarkers in Asians and Caucasians

Triglycerides

HDL-cholesterol

Total cholesterol

Glucose

LDL-Cholesterol

Insulin

CRP

Fibrinogen

HbA1c

PAI-1

Biomarkers measured in cohort

UK, 4101, 52.2

UK, 245, 41

UK, 245, 41

UK, 100, 51

UK, 100, 52

Canada, 324, 49.4 Canada, 317, 47.4

Canada, 326, 51.2

Canada, 323, 49.4 Canada, 306, 47.7

Canada, 322, 51.3

UK, 151, 62.1

UK, 149, 62.6

UK, 1486, 57.0

UK, 492, 64.8

US, 211, 50.0

US, 1684, 51.6

UK, 576, NR

UK, 1301, NR

UK, 56, 53.5

UK, 44, 54.3

UK, 447, 48.9

UK, 475, 49.7

UK, 224, 49.2

UK, 248, 49.9

UK, 215, NR

UK, 261, NR

UK, 63, NR

UK, 62, NR

UK, 26, 46.7

UK, 18, 45.8

UK, 518, 49.0

UK, 507, 49.4

US, 82, 31

US, 55, 29

UK, 165, 33.5

UK, 127, 35.3

UK, 1162, 56.8

UK, 912, 52.1

UK, 1420, 51.2

UK, 1787, 52.9

UK, 223, 49

UK, 787, 52

UK, 70, 55.7

UK, 121, 52.1

UK, 130, 49.9

UK, 142, 51.9

US, 150, 57

US, 44, 29.7

US, 803, 62.0

US, 2622, 62.5

US, 803, 62.9

US, 2619, 62.4

US, 751, 62.6

US, 2362, 62.4

South Asians Country, N, age

Chinese Country, Japanese N, age Country, N, age

UK, 490, 47

Japan, 150, 58 US, 39, 27.5 US, 803, 57.1

Caucasians Country, N, age

US, 244, 46.7

US, 270, 46.8

US, 1496, 46.1

US, 231, 46.6

US, 248, 46.7

US, 1400, 46.4

SG, 10, 27.1

UK, 880, 43.9

UK, 70487, 45.6

UK, 75, 49.6

UK, 68, 50.2

India, 205, 46.1

US, 720, 60.0

UK, 138, 49.8

UK, 129, 52.6

Japan, 310, 45.1 US, 297, 45.0

SG, 10, 26.4

SG, 10, 24.7

31


Chapter 2

Table 2. Continued Sekikawa, 200750

ERA JUMP/ Post-WWII, 493

Patel, 200751

Houston, 70

Raji, 200152

Massachusetts, 24

Raji, 200453

Massachusetts, 40

Rapeport, 201354

CEPHEUS SA, 2005

Smith, 200655

Montreal, 165

Zhang, 199356

Taiyuan/Belfast, 353

Zoratti, 200057

London, 92

DISCUSSION This systematic review summarizes the published data on CAD biomarkers measured in both Asians and Caucasians in general population-based studies. By far, most evidence on CAD biomarker levels in Asians is derived from South Asians, mostly South Asian migrants to Western environments; data from Chinese and Japanese populations are available to a much lesser extent. Taken together, the available data show that South Asians display an unfavourable CAD biomarker pattern (with higher TG levels, higher insulin levels, higher CRP levels and higher PAI-1 levels) in comparison to Caucasians. Chinese have a slightly favourable biomarker profile, with lower CRP levels and comparable levels for the other biomarkers. Japanese have the most favourable biomarker profile with lower insulin levels, lower CRP levels, lower fibrinogen levels and otherwise comparable biomarker levels. That is, if we assume the biomarkers to have the same predictive values in the Asian ethnic groups as in Caucasians. Theoretically, the differences in the absolute levels of the CAD biomarkers could be explained by differences in the prevalence of CAD across the ethnic groups, as biomarker levels are probably different in people with established CAD. Therefore we checked for inequality of CAD prevalence across the ethnic groups within the cohorts. We did not find clear evidence for unequal CAD prevalences, with the prevalence of CAD being significantly different in only three cohorts. We found no articles that compare predictive thresholds of CAD biomarkers between Asians and Caucasians and it remains to be elucidated whether the absolute biomarker levels have the same predictive value for the Asian ethnic groups as they do for Caucasians. Critically, accurate risk stratification in Asian populations requires predictive thresholds in each specific ethnic group. Otherwise, there is potential for misclassification of risk and inappropriate use of pharmacotherapies in primary and secondary prevention. Only five of the 33 reviewed cohorts reported on biomarker levels of Asians actually living in their country of origin. Of the remaining 28 cohorts that included Asians in other countries, only two4,50 provided immigration details, such as years since immigration or offspring generation. Hence, this information could not been taken into account in this review, although the moment of

32


Systematic Review: CAD biomarkers in Asians and Caucasians

US, 35, 30.1

US, 35, 30.8

US, 12, 34

US, 12, 35

US, 25, 35

US, 15, 36

SA, 620, 61.4

SA, 1385, 57.3

Canada, 82, 42.9

Canada, 83, 39.5

Ireland, 202, 54.5

UK, 31, 53.8

Japan, 250, 45.2 US, 243, 45.1

China, 151, 48.9 UK, 31, 48.5

immigration may significantly affect biomarkers levels. For example, it has been shown recently that carotid intima-media wall thickness (IMT) and plaque size, increase in Japanese with every generation after immigration.58 Biomarker levels The quality of the blood samples differed by study protocols, the majority of the studies drew blood after overnight fasting, however not all, or nothing was mentioned. Also, the methods of analysing biomarker levels (assays/machines used) were not described in detail in all articles. These factors (described in supplemental table 3) add to the heterogeneity of the biomarker levels described in the articles. South Asia South Asians have a higher burden of CAD when compared to Caucasians or other Asian ethnic groups. The general consensus is that this difference is multifactorial, but that the higher prevalence of diabetes and lipid abnormalities in South Asians are at least partly responsible for the high CAD incidence in South Asia.4 In line with this, we found an unfavourable lipid profile throughout literature for South Asians, characterized by higher TG and lower HDL when compared to Caucasians. However, we also found a slightly lower LDL-cholesterol in eight of 13 cohorts, which supposedly reduces the risk of CAD. Normal LDL levels have been described previously in people with increased risk of CAD, in which there is an increased fraction of small dense LDL, associated with a higher risk of CAD.59 Zoratti et al. showed indeed lower LDL particle size in South Asians compared to Caucasians57, a trend which is already observable in adolescence.60 None of the reviewed articles compared the predictive values for lipid levels between South Asians and Caucasians, although Forouhi et al.33 did calculate hazard ratios (HRs) for South Asians and Caucasians separately showing HRs in the same order of magnitude for TG, TC and HDL, with comparable confidence intervals. This would imply that lipid levels have similar predictive values in South Asian men as compared to Caucasian men. In the MESA and MASALA39,40 cohort the association of lipid levels with coronary artery calcium (CAC) has been described per ethnic group, showing that lipid levels are not associated with CAC in South Asians while LDL and HDL levels are associated with CAD in Whites.

33


Chapter 2

Table 3. Average relative differences in biomarker levels across the reviewed articles as compared to Caucasians. South Asians

Chinese

Migrant status

Biomarker

All

HDL

0.92

23

0.97

5

1.13

3

TC

0.99

22

0.97

5

1.01

3

TG

1.13

22

1.08

5

1.01

3

Glucose

1.06

19

1.05

4

1.02

2

Insulin

1.33

14

1.03

4

0.84

2

LDL

0.94

13

0.94

4

0.97

3

CRP

1.22

8

0.52

3

0.37

2

HbA1c

1.06

8

1.04

1

-

0

PAI-1

1.31

5

0.95

2

0.92

2

Fibrinogen

1.01

4

0.98

3

0.86

3

HDL

0.92

22

0.98

3

1.11

1

TC

0.99

21

0.98

3

1.02

1

TG

1.13

21

1.08

3

1.04

1

Glucose

1.06

18

1.05

3

1.01

1

Insulin

1.33

13

1.03

3

0.89

1

LDL

0.94

13

0.96

3

0.99

1

CRP

1.22

8

0.52

3

0.36

1

HbA1c

1.06

8

1.04

1

-

0

PAI-1

1.31

5

0.95

2

0.96

1

Fibrinogen

1.01

4

0.98

3

0.87

1

HDL

0.85

1

0.83

2

1.18

2

TC

1.09

1

0.71

2

0.99

2

TG

1.53

1

1.06

2

0.96

2

Glucose

1.07

1

1.02

1

1.06

1

Insulin

1.86

1

1.29

1

0.68

1

LDL

-

0

0.63

1

0.93

2

CRP

-

0

-

0

0.39

1

HbA1c

-

0

-

0

-

0

PAI-1

-

0

-

0

0.73

1

Fibrinogen

-

0

-

0

0.85

2

Migrant

Non-migrant

N Articlesยง

Relative* difference

Japanese

Relative* difference

N Articlesยง

Relative* difference

N Articlesยง

HDL: high-density lipoprotein cholesterol; TC: total cholesterol; TG: triglycerides; LDL: low-density lipoprotein cholesterol; HbA1c: glycated hemoglobin, CRP: C-reactive protein, PAI-1: plasminogen activator inhibitor-1. ยง Number of articles comparing the biomarker between Caucasians and the Asian ethnic group. *Relative difference is calculated by dividing the biomarker level of the Asian ethnic group by the level of Caucasians. Figures presented here are the average of the relative difference across the reviewed studies, taking into account study size.

34


Systematic Review: CAD biomarkers in Asians and Caucasians

Another important component of the increased CAD risk in South Asians could be a disturbed glucose metabolism. We found clearly higher insulin levels with slightly higher glucose and HbA1c levels in South Asians compared to Caucasians, indicating more insulin resistance in South Asians, which has been described for this ethnic group for over two decades now.61 Only one of the reviewed articles compared prognostic measures of glucose metabolism biomarkers between South Asians and Caucasians. Forouhi et al.33 performed a Cox regression analysis on South Asians and Caucasians separately, which revealed a higher HR for South Asians; 3.28 than for Caucasians; 2.18 (with overlapping confidence intervals) suggesting a stronger effect of insulin resistance for South Asians, although significance of this difference was not calculated. In the MASALA cohort39 glucose and insulin levels were not associated with CAC, unfortunately these associations were not calculated for Caucasians. Within the INTERHEART study, the relative risk of diabetes for acute myocardial infarction in South Asia, Western Europe and North America were calculated, showing odds ratios with similar directions (OR 2.48, 4.29 and 1.75 respectively, with overlapping confidence intervals).62 If disturbed glucose metabolism has a similar impact in the South Asian ethnic group compared to other ethnicities (which has not yet been described in literature, apart from the abovementioned regional comparison by the INTERHEART study) this will be an important target for primary prevention. With the extremely high prevalence of diabetes in South Asians, a major health burden reduction might be achieved by improved glycaemic regulation. While reviewing literature we found PAI-1 levels to be higher and fibrinogen to be slightly higher (relative difference 1.04) in South Asians than in Caucasians. Both of these biomarkers are involved in the coagulation system and might point to a more procoagulant state, which might also be an important component of the unfavourable risk pattern of South Asians. This phenomenon has been described before, mainly concerning ischemic stroke.63,64 The predictive values of fibrinogen and PAI-1 levels in South Asians were unfortunately not described in the reviewed articles, so it remains to be seen whether these markers actually influence the risk of CAD in South Asians. CRP levels were higher in South Asians, which might indicate a higher level of low-grade inflammation in South Asians in comparison to Caucasians. Although CRP is not a causal factor of CVD65, it has been found to be predictive for CVD events66,67. Whether the predictive value of CRP for CVD differs between South Asians and Caucasians remains unclear, as none of the reviewed cohorts investigated this. China and Japan Data on CAD biomarker levels for Chinese and Japanese is much scarcer than for South Asians. Similar to South Asians, TG levels in Chinese were marginally higher and LDLcholesterol levels are slightly lower than in Caucasians. The Asia Pacific Cohort study, which compared geographic regions, but not individual ethnicity, reported that TG levels were more strongly associated with CAD in Australia and New Zealand than in China and

35


36

1.6

292.8

7.7

8.4

1.2

CRP (mg/L)

Fibrinogen (mg/dL)

Glucose (mmol/L)

HbA1c (%)

HDL (mmol/L)

11,167

LDL (mmol/L)

7,076

8,009

282

3,053

3,997

8,076

3,858

5,417

1,121

1,456

N

21

21

3

13

13

22

8

18

4

8

N articles

1.7

5.0

18.7

3.0

7.0

1.3

5.4

5.1

312.7

2.6

4,444

4,441

1,400

4,544

4,344

4,444

322

4,348

4,088

4,084

1.6

4.9

17.1

2.9

7.7

1.3

5.7

5.4

308.9

1.5

1,364

1,364

231

1,364

1,340

1,364

306

1,351

1,299

1,288

Chinese N

3

3

1

3

3

3

1

3

3

3

N articles

Comparison Caucasians vs. Chinese Caucasians N

* Not all available data on PAI-1 levels could be converted to ng/mL, only the convertible data is shown.

1.8

1.5

TG (mmol/L)

13,123

22,248 5.3

5.7

TC (mmol/L)

16.5

272

3.1

10.9

1.2

8.1

6.5

296.5

1.9

South Asians

PAI-1 (ng/mL)* 9.6

3.4

9,747

Insulin (ÎźU/mL) 9.6

16,865

6,988

11,862

1,092

1,468

Caucasians N

Biomarker

Comparison Caucasians vs. South Asians

0.1

5.0

18.7

2.9

7.6

1.4

-

5.0

282.0

1.4

1,496

1,496

1,400

1,596

1,400

1,496

0

1,400

1,400

1,400

Caucasians N

0.1

5.0

18.0

2.9

6.8

1.6

-

5.1

246.0

0.5

Japanese

270

270

248

270

248

270

0

248

248

248

N

1

1

1

1

1

1

0

1

1

1

N articles

Comparison Caucasians vs. Japanese

Table 4. Mean biomarker levels for each comparison, weighted for N per ethnicity for each comparison. Only data from cohorts presenting data of migrated Asian ethnic groups were taken into account. These data are presented for each comparison separately, because not all articles compare all ethnic groups.

Chapter 2


Systematic Review: CAD biomarkers in Asians and Caucasians

Japan. This could mean that the higher TG levels in Chinese and Japanese have less predictive value68 and might explain at least partly why Chinese experience less CAD than Caucasians.4 HDL levels were higher in Japanese as compared to Caucasians. Possibly, this indicates a more favourable lipid profile in Japanese, composed of a relatively large proportion of HDL69,70 and lower LDL levels. The predictive values of lipids for the occurrence of CAD in Japanese and Chinese were not compared to Caucasians in the reviewed articles. However, in the MESA cohort the ethnicity-specific association of LDL and HDL cholesterol with CAC was not significant for Chinese, while it was significant in Caucasians.40 But these differences could be attributed to differences in statistical power between Chinese (n=789) and Caucasians (n=2,575). The INTERHEART study showed a higher odds ratio for lipid disturbances for myocardial infarction in South East Asians than in Europeans.62 However, the South East Asian group consists of very diverse Asian countries, encompassing multiple Asian ethnicities. Glucose levels were comparable to Caucasians in Chinese and Japanese, with comparable insulin levels in Chinese and lower levels in Japanese. This difference might be truly determined by ethnicity or by a confounder, for example immigration level. Insulin levels were measured in only two Japanese cohorts, of which one included people living in Japan50, showing a more extreme difference than reported when comparing American Japanese to American Caucasians43. Insulin levels in Chinese were only measured in Chinese living in the US or Canada, who may have adapted to a Western lifestyle, which has been shown to increase insulin levels (up to levels comparable to Caucasians).58 No prognostic measures for biomarkers of glucose metabolism were reported in the reviewed articles. But, in the Asia Pacific Cohort Study similar hazard ratios of diabetes for death from CAD were found for Australia and New Zealand; HR 1.80 and Asia (mainly China and Japan); HR 1.84.68 Fibrinogen levels were lower in all publications comparing Japanese to Caucasians, possibly contributing to their favourable risk profile as higher fibrinogen levels convey a higher risk of CVD.71 Evidence on the differences in the predictive value of fibrinogen levels across the ethnic groups is scarce and inconclusive. In women from the SWAN cohort43, fibrinogen levels correlated with the Framingham risk score in similar ways among Caucasians, Chinese and Japanese (r= 0.19, 0.16 and 0.22 respectively). Sekikawa et al. described a significant association of fibrinogen levels with CAC both for Japanese and Caucasian men, although absolute fibrinogen levels were higher in Caucasians, but no significant trend with carotid intima-media thickness.50 Within the MESA cohort41, a significant HR of 3.05 (p=0.02) for CVD events was calculated for Caucasians but not for Chinese (HR 2.39, NS), suggesting less influence of fibrinogen levels in Chinese than in Caucasians. CRP levels have consistently been reported to be lower in Chinese and Japanese compared with Caucasians. Within the MESA cohort, CRP did not appear to influence CVD events at all in Chinese (HR 0.88, NS), but it was a significant predictor in Caucasians; HR 1.23 (p=0.01). Also in Japanese men there was no relation of CRP with coronary artery

37


Chapter 2

calcium or carotid intima-media thickness. However, in this study this also was not the case for Caucasian men.50 In the women’s SWAN cohort there was a notably lower correlation coefficient of CRP with Framingham score in Japanese than in Chinese and Caucasians (r= 0.17, 0.26 and 0.24 respectively).43 This might indicate that CRP levels play a different - possibly prognostically less important - role in Chinese and Japanese than in Caucasians for predicting CVD events. Limitations A true meta-analysis of the collected data was unfortunately not possible due to the vast heterogeneity across the cohorts, as can be appreciated from the varying cohort characteristics shown in table 2 and supplemental table 2 and 3.As we only discuss cross-sectional studies, there is no way to distinguish between environmental factors and biological factors influencing biomarker levels among ethnic groups. For the purpose of this review South Asians were considered one ethnic group, however, differences have been shown to exist among specific South Asian subgroups.72 Data on South Asians were pooled in this review, because of the small number of articles that discuss the subgroups separately. The average relative differences, as reported throughout this review were weighted by cohort size. We were however unable to adjust or stratify for any other study characteristics (e.g. age or gender of study participants) due to the heterogeneity among the studies. Also, it is unclear what should be the cut-off to determine a clinically relevant difference. The true clinical implication of biomarker levels can only be assessed when their ethnicity-specific prognostic values have been established. In this review we specifically focused on CAD as an outcome measure. However, for primary prevention of CVD (beyond CAD alone) also stroke and other cardiovascular end points could be taken into account, as included for outcome measures in the Framingham risk score.73 Life style (and acculturation specifically in migrants), socio-economic status and education are influencers of the risk of CAD. Within this review we were unable to determine the effect of these factors on biomarker levels across the ethnic groups. Conclusion This review summarises differences in CAD biomarker levels between Caucasians and the Asian ethnic groups at a population-level. We highlight that the Asian population cannot be seen as one homogenous ethnic group, as different biomarker patterns are evident among the different Asian ethnicities. The predictive values of the most widely used CAD biomarkers is understudied in Asians. Longitudinal studies including multiple ethnic groups are therefore urgently needed to test the predictive thresholds of biomarkers, as the thresholds identified in Caucasians may not correctly assess risk in Asians. Accurate risk stratification and risk-appropriate primary prevention require ethnicity-specific CAD biomarker thresholds.

38


Systematic Review: CAD biomarkers in Asians and Caucasians

Acknowledgements This work was financially supported by the Royal Netherlands Academy of Art and Sciences via a strategic grant to the Interuniversity Cardiology Institute of the Netherlands (ICIN) and Dominique de Kleijn.

39


Chapter 2

References 1.

Ford ES, Ajani U a, Croft JB, Critchley J a, Labarthe DR, Kottke TE, Giles WH, Capewell S. Explaining the decrease in U.S. deaths from coronary disease, 1980-2000. N Engl J Med. 2007;356:2388–98.

2.

Mathers CD, Loncar D. Projections of global mortality and burden of disease from 2002 to 2030. PLoS Med. 2006;3:e442.

3.

Kannel WB, McGee D, Gordon T. A general cardiovascular risk profile: the Framingham Study. Am J Cardiol. 1976;38:46–51.

4.

Anand SS, Yusuf S, Vuksan V, Devanesen S, Teo KK, Montague PA, Kelemen L, Yi C, Lonn E, Gerstein H, Hegele RA, McQueen M. Differences in risk factors, atherosclerosis, and cardiovascular disease between ethnic groups in Canada: the Study of Health Assessment and Risk in Ethnic groups (SHARE). Lancet. 2000;356:279–84.

5.

Zheng Y, Ma W, Zeng Y, Liu J, Ye S, Chen S, Lan L, Erbel R, Liu Q. Comparative study of clinical characteristics between Chinese Han and German Caucasian patients with coronary heart disease. Clin Res Cardiol. 2010;99:45–50.

6.

Joshi P, Islam S, Pais P, Reddy S, Dorairaj P, Kazmi K, Pandey MR, Haque S, Mendis S, Rangarajan S, Yusuf S. Risk factors for early myocardial infarction in South Asians compared with individuals in other countries. JAMA. 2007;297:286–94.

7.

Goff DC, Lloyd-Jones DM, Bennett G, Coady S, D’Agostino RB, Gibbons R, Greenland P, Lackland DT, Levy D, O’Donnell CJ, Robinson JG, Schwartz JS, Shero ST, Smith SC, Sorlie P, Stone NJ, Wilson PWF, Jordan HS, Nevo L, Wnek J, Anderson JL, Halperin JL, Albert NM, Bozkurt B, Brindis RG, Curtis LH, DeMets D, Hochman JS, Kovacs RJ, Ohman EM, Pressler SJ, Sellke FW, Shen W-K, Tomaselli GF. 2013 ACC/AHA guideline on the assessment of cardiovascular risk: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines. Circulation. 2014;129:S49–73.

8.

Tillin T, Hughes AD, Whincup P, Mayet J, Sattar N, McKeigue PM, Chaturvedi N. Ethnicity and prediction of cardiovascular disease: performance of QRISK2 and Framingham scores in a U.K. tri-ethnic prospective cohort study (SABRE--Southall And Brent REvisited). Heart. 2014;100:60–7.

9.

Kandula NR, Diez-Roux A V, Chan C, Daviglus ML, Jackson S a, Ni H, Schreiner PJ. Association of acculturation levels and prevalence of diabetes in the multi-ethnic study of atherosclerosis (MESA). Diabetes Care. 2008;31:1621–8.

10. Diez Roux A V, Detrano R, Jackson S, Jacobs DR, Schreiner PJ, Shea S, Szklo M. Acculturation and socioeconomic position as predictors of coronary calcification in a multiethnic sample. Circulation. 2005;112:1557–65. 11. Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. PLoS Med. 2009;6:e1000097. 12. Cappuccio FP, Bell R, Perry IJ, Gilg J, Ueland PM, Refsum H, Sagnella GA, Jeffery S, Cook DG. Homocysteine levels in men and women of different ethnic and cultural background living in England. Atherosclerosis. 2002;164:95–102. 13. Siezenga MA, Chandie Shaw PK, van der Geest RN, Mollnes TE, Daha MR, Rabelink TJ, Berger SP. Enhanced complement activation is part of the unfavourable cardiovascular risk profile in South Asians. Clin Exp Immunol. 2009;157:98–103.

40


Systematic Review: CAD biomarkers in Asians and Caucasians

14. Shamaei-Tousi A, Stephens JW, Bin R, Cooper JA, Steptoe A, Coates ARM, Henderson B, Humphries SE. Association between plasma levels of heat shock protein 60 and cardiovascular disease in patients with diabetes mellitus. Eur Heart J. 2006;27:1565–70. 15. Adler A, Neil A, Stratton I, Holman R. Ethnicity and cardiovascular disease. The incidence of myocardial infarction in white, South Asian, and Afro-Caribbean patients with type 2 diabetes (U.K. Prospective Diabetes Study 32). Diabetes Care. 1998;21:1271–1277. 16. Ajjan R, Carter A, Somani R. Ethnic differences in cardiovascular risk factors in healthy Caucasian and South Asian individuals with the metabolic syndrome. J Thromb …. 2007;754–760. 17. Kain K, Blaxill JM, Catto AJ, Grant PJ, Carter AM. Increased fibrinogen levels among South Asians versus Whites in the United Kingdom are not explained by common polymorphisms. Am J Epidemiol. 2002;156:174–179. 18. Anand SS, Razak F, Yi Q, Davis B, Jacobs R, Vuksan V, Lonn E, Teo K, McQueen M, Yusuf S. C-reactive protein as a screening test for cardiovascular risk in a multiethnic population. Arterioscler Thromb Vasc Biol. 2004;24:1509–15. 19. Bathula R, Hughes a D, Panerai R, Potter J, Thom S a M, Francis DP, Shore a C, Kooner J, Chaturvedi N. Indian Asians have poorer cardiovascular autonomic function than Europeans: this is due to greater hyperglycaemia and may contribute to their greater risk of heart disease. Diabetologia. 2010;53:2120–8. 20. Bellary S, O’Hare JP, Raymond NT, Mughal S, Hanif WM, Jones A, Kumar S, Barnett AH. Premature cardiovascular events and mortality in south Asians with type 2 diabetes in the United Kingdom Asian Diabetes Study - effect of ethnicity on risk. Curr Med Res Opin. 2010;26:1873–9. 21. Bhalodkar NC, Blum S, Rana T, Bhalodkar A, Kitchappa R, Kim K-S, Enas E. Comparison of levels of large and small high-density lipoprotein cholesterol in Asian Indian men compared with Caucasian men in the Framingham Offspring Study. Am J Cardiol. 2004;94:1561–3. 22. Bhopal R, Fischbacher C, Vartiainen E, Unwin N, White M, Alberti G. Predicted and observed cardiovascular disease in South Asians: Application of FINRISK, Framingham and SCORE models to Newcastle Heart Project data. J Public Health (Bangkok). 2005;27:93–100. 23. Reynolds RM, Fischbacher C, Bhopal R, Byrne CD, White M, Unwin N, Walker BR. Differences in cortisol concentrations in South Asian and European men living in the United Kingdom. Clin Endocrinol (Oxf). 2006;64:530–4. 24. Cappuccio FP, Oakeshott P, Strazzullo P, Kerry SM. Application of Framingham risk estimates to ethnic minorities in United Kingdom and implications for primary prevention of heart disease in general practice: cross sectional population based study. BMJ. 2002;325:1271. 25. Cook DG, Cappuccio FP, Atkinson RW, Wicks PD, Chitolie A, Nakandakare ER, Sagnella G a, Humphries SE. Ethnic differences in fibrinogen levels: the role of environmental factors and the beta-fibrinogen gene. Am J Epidemiol. 2001;153:799–806. 26. Miller M a, Sagnella G a, Kerry SM, Strazzullo P, Cook DG, Cappuccio FP. Ethnic differences in circulating soluble adhesion molecules: the Wandsworth Heart and Stroke Study. Clin Sci (Lond). 2003;104:591–8. 27. Miller M a, McTernan PG, Harte AL, Silva NF Da, Strazzullo P, Alberti KGMM, Kumar S, Cappuccio FP. Ethnic and sex differences in circulating endotoxin levels: A novel marker of atherosclerotic and cardiovascular risk in a British multi-ethnic population. Atherosclerosis. 2009;203:494–502. 28. Chambers JC, McGregor a, Jean-Marie J, Kooner JS. Abnormalities of vascular endothelial function may contribute to increased coronary heart disease risk in UK Indian Asians. Heart. 1999;81:501–4.

41


Chapter 2

29. Chambers JC, Eda S, Bassett P, Karim Y, Thompson SG, Gallimore JR, Pepys MB, Kooner JS. C-reactive protein, insulin resistance, central obesity, and coronary heart disease risk in Indian Asians from the United Kingdom compared with European whites. Circulation. 2001;104:145–50. 30. Chandalia M, Cabo-Chan A V, Devaraj S, Jialal I, Grundy SM, Abate N. Elevated plasma high-sensitivity C-reactive protein concentrations in Asian Indians living in the United States. J Clin Endocrinol Metab. 2003;88:3773–6. 31. Chowdhury T a, Lasker SS. Complications and cardiovascular risk factors in South Asians and Europeans with early-onset type 2 diabetes. QJM. 2002;95:241–246. 32. Chowdhury T a, Lasker SS, Mahfuz R. Ethnic differences in control of cardiovascular risk factors in patients with type 2 diabetes attending an Inner London diabetes clinic. Postgrad Med J. 2006;82:211–215. 33. Forouhi NG, Sattar N, Tillin T, McKeigue PM, Chaturvedi N. Do known risk factors explain the higher coronary heart disease mortality in South Asian compared with European men? Prospective follow-up of the Southall and Brent studies, UK. Diabetologia. 2006;49:2580–8. 34. Gama R, Elfatih AB, Anderson NR. Ethnic differences in total and HDL cholesterol concentrations: Caucasians compared with predominantly Punjabi Sikh Indo-Asians. Ann Clin Biochem. 2002;39:609–11. 35. Chatha K, Anderson NR, Gama R. Ethnic variation in C-reactive protein: UK resident Indo-Asians compared with Caucasians. J Cardiovasc Risk. 2002;9:139–41. 36. Heald AH, Anderson SG, Ivison F, Laing I, Gibson JM, Cruickshank K. C-reactive protein and the insulin-like growth factor (IGF)-system in relation to risk of cardiovascular disease in different ethnic groups. Atherosclerosis. 2003;170:79–86. 37. Iso H, Koike K a, Folsom a R, Shimamoto T, Sato S, Lida M, Komachi Y. Lipoprotein(a) and its correlates in Japanese and U.S. population samples. Ann Epidemiol. 1996;6:324–30. 38. Kamath SK, Hussain E a, Amin D, Mortillaro E, West B, Peterson CT, Aryee F, Murillo G, Alekel DL. Cardiovascular disease risk factors in 2 distinct ethnic groups: Indian and Pakistani compared with American premenopausal women. Am J Clin Nutr. 1999;69:621–631. 39. Kanaya AM, Kandula NR, Ewing SK, Herrington D, Liu K, Blaha MJ, Srivastava S, Dave SS, Budoff MJ. Comparing coronary artery calcium among U.S. South Asians with four racial/ethnic groups: the MASALA and MESA studies. Atherosclerosis. 2014;234:102–7. 40. Bild DE, Detrano R, Peterson D, Guerci A, Liu K, Shahar E, Ouyang P, Jackson S, Saad MF. Ethnic differences in coronary calcification: the Multi-Ethnic Study of Atherosclerosis (MESA). Circulation. 2005;111:1313–20. 41. Veeranna V, Zalawadiya SK, Niraj A, Kumar A, Ference B, Afonso L. Association of novel biomarkers with future cardiovascular events is influenced by ethnicity: results from a multi-ethnic cohort. Int J Cardiol. 2013;166:487–93. 42. Kelley-Hedgepeth A, Lloyd-Jones DM, Colvin A, Matthews K a, Johnston J, Sowers MR, Sternfeld B, Pasternak RC, Chae CU. Ethnic differences in C-reactive protein concentrations. Clin Chem. 2008;54:1027–37. 43. Matthews K a, Sowers MF, Derby C a, Stein E, Miracle-McMahill H, Crawford SL, Pasternak RC. Ethnic differences in cardiovascular risk factor burden among middle-aged women: Study of Women’s Health Across the Nation (SWAN). Am Heart J. 2005;149:1066–73. 44. Liew C-F, Seah E-S, Yeo K-P, Lee K-O, Wise SD. Lean, nondiabetic Asian Indians have decreased insulin sensitivity and insulin clearance, and raised leptin compared to Caucasians and Chinese subjects. Int J Obes Relat Metab Disord. 2003;27:784–9.

42


Systematic Review: CAD biomarkers in Asians and Caucasians

45. Lyratzopoulos G, McElduff P, Heller RF, Hanily M, Lewis PS. Comparative levels and time trends in blood pressure, total cholesterol, body mass index and smoking among Caucasian and South-Asian participants of a UK primary-care based cardiovascular risk factor screening programme. BMC Public Health. 2005;5:125. 46. Miller GJ, Kotecha S, Wilkinson WH, Wilkes H, Stirling Y, Sanders T a, Broadhurst a, Allison J, Meade TW. Dietary and other characteristics relevant for coronary heart disease in men of Indian, West Indian and European descent in London. Atherosclerosis. 1988;70:63–72. 47. Mulukutla SR, Venkitachalam L, Marroquin OC, Kip KE, Aiyer A, Edmundowicz D, Ganesh S, Varghese R, Reis SE. Population variations in atherogenic dyslipidemia: A report from the HeartSCORE and IndiaSCORE Studies. J Clin Lipidol. 2008;2:410–7. 48. Nagi DK, Knowler WC, Hanson RL, Ali VM, Yudkin JS. Plasminogen activator inhibitor (PAI-1) and non-insulindependent diabetes in Pima Indians, south Asians and Europeans. Populations at varying risk of NIDDM and coronary artery disease. Thromb Haemost. 1996;75:921–7. 49. Okamura T, Sekikawa A, Sawamura T, Kadowaki T, Barinas-Mitchell E, Mackey RH, Kadota A, Evans RW, Edmundowicz D, Higashiyama A, Nakamura Y, Abbott RD, Miura K, Fujiyoshi A, Fujita Y, Murakami Y, Miyamatsu N, Kakino A, Maegawa H, Murata K, Horie M, Mitsunami K, Kashiwagi A, Kuller LH, Ueshima H. LOX-1 ligands containing apolipoprotein B and carotid intima-media thickness in middle-aged community-dwelling US Caucasian and Japanese men. Atherosclerosis. 2013;229:240–5. 50. Sekikawa A, Ueshima H, Kadowaki T, El-Saed A, Okamura T, Takamiya T, Kashiwagi A, Edmundowicz D, Murata K, Sutton-Tyrrell K, Maegawa H, Evans RW, Kita Y, Kuller LH. Less subclinical atherosclerosis in Japanese men in Japan than in White men in the United States in the post-World War II birth cohort. Am J Epidemiol. 2007;165:617–24. 51. Patel RT, Lev EI, Vaduganathan M, Guthikonda S, Bergeron A, Maresh K, Dong J, Kleiman NS. Platelet reactivity among Asian Indians and Caucasians. Platelets. 2007;18:261–5. 52. Raji A, Seely EW, Arky RA, Simonson DC. Body fat distribution and insulin resistance in healthy Asian Indians and Caucasians. J Clin Endocrinol Metab. 2001;86:5366–71. 53. Raji A, Gerhard-Herman MD, Warren M, Silverman SG, Raptopoulos V, Mantzoros CS, Simonson DC. Insulin resistance and vascular dysfunction in nondiabetic Asian Indians. J Clin Endocrinol Metab. 2004;89:3965– 72. 54. Rapeport N, Schamroth CL, Blom DJ. Gender and ethnic differences in the control of hyperlipidaemia and other vascular risk factors: insights from the CEntralised Pan-South African survey on tHE Under-treatment of hypercholeSterolaemia (CEPHEUS SA) study. Cardiovasc J Afr. 2013;24:238–42. 55. Smith J, Cianflone K, Al-Amri M, Sniderman A. Body composition and the apoB/apoA-I ratio in migrant Asian Indians and white Caucasians in Canada. Clin Sci (Lond). 2006;111:201–7. 56. Zhang W, Evans a E, Cambien F, Li P, Li X, Bard JM, Fruchart JC, McCrum EE, McMaster D. Distribution of lipid variables in subjects in Belfast, Northern Ireland and Taiyuan, P R China. Atherosclerosis. 1993;102:175– 80. 57. Zoratti R, Godsland IF, Chaturvedi N, Crook D, Stevenson JC, McKeigue PM. Relation of plasma lipids to insulin resistance, nonesterified fatty acid levels, and body fat in men from three ethnic groups: relevance to variation in risk of diabetes and coronary disease. Metabolism. 2000;49:245–52. 58. Egusa G, Watanabe H, Ohshita K, Fujikawa R, Yamane K, Okubo M, Kohno N. Influence of the extent of westernization of lifestyle on the progression of preclinical atherosclerosis in Japanese subjects. J Atheroscler Thromb. 2002;9:299–304.

43


Chapter 2

59. Lamarche B, Lemieux I, Després JP. The small, dense LDL phenotype and the risk of coronary heart disease: epidemiology, patho-physiology and therapeutic aspects. Diabetes Metab. 1999;25:199–211. 60. Raschke V, Elmadfa I, Bermingham MA, Steinbeck K. Low density lipoprotein subclasses in Asian and Caucasian adolescent boys. Asia Pac J Clin Nutr. 2006;15:496–501. 61. McKeigue PM. Coronary heart disease in Indians, Pakistanis, and Bangladeshis: aetiology and possibilities for prevention. Br Heart J. 1992;67:341–2. 62. Yusuf S, Hawken S, Ounpuu S, Dans T, Avezum A, Lanas F, McQueen M, Budaj A, Pais P, Varigos J, Lisheng L. Effect of potentially modifiable risk factors associated with myocardial infarction in 52 countries (the INTERHEART study): case-control study. Lancet. 2004;364:937–52. 63. Kain K, Catto AJ, Grant PJ. Clustering of thrombotic factors with insulin resistance in South Asian patients with ischaemic stroke. Thromb Haemost. 2002;88:950–3. 64. Kain K, Catto AJ, Young J, Bamford J, Bavington J, Grant PJ. Increased fibrinogen, von Willebrand factor and tissue plasminogen activator levels in insulin resistant South Asian patients with ischaemic stroke. Atherosclerosis. 2002;163:371–6. 65. Wensley F, Gao P, Burgess S, Kaptoge S, Di Angelantonio E, Shah T, Engert JC, Clarke R, Davey-Smith G, Nordestgaard BG, Saleheen D, Samani NJ, Sandhu M, Anand S, Pepys MB, Smeeth L, Whittaker J, Casas JP, Thompson SG, Hingorani AD, Danesh J. Association between C reactive protein and coronary heart disease: mendelian randomisation analysis based on individual participant data. BMJ. 2011;342:d548. 66. Ridker PM, Glynn RJ, Hennekens CH. C-Reactive Protein Adds to the Predictive Value of Total and HDL Cholesterol in Determining Risk of First Myocardial Infarction. Circulation. 1998;97:2007–2011. 67. Tracy RP, Lemaitre RN, Psaty BM, Ives DG, Evans RW, Cushman M, Meilahn EN, Kuller LH. Relationship of C-Reactive Protein to Risk of Cardiovascular Disease in the Elderly: Results From the Cardiovascular Health Study and the Rural Health Promotion Project. Arterioscler Thromb Vasc Biol. 1997;17:1121–1127. 68. Woodward M, Huxley H, Lam TH, Barzi F, Lawes CMM, Ueshima H. A comparison of the associations between risk factors and cardiovascular disease in Asia and Australasia. Eur J Cardiovasc Prev Rehabil. 2005;12:484–91. 69. Stampfer MJ, Sacks FM, Salvini S, Willett WC, Hennekens CH. A prospective study of cholesterol, apolipoproteins, and the risk of myocardial infarction. N Engl J Med. 1991;325:373–81. 70. Huxley RR, Barzi F, Lam TH, Czernichow S, Fang X, Welborn T, Shaw J, Ueshima H, Zimmet P, Jee SH, Patel J V, Caterson I, Perkovic V, Woodward M. Isolated low levels of high-density lipoprotein cholesterol are associated with an increased risk of coronary heart disease: an individual participant data meta-analysis of 23 studies in the Asia-Pacific region. Circulation. 2011;124:2056–64. 71. Danesh J, Lewington S, Thompson SG, Lowe GDO, Collins R, Kostis JB, Wilson AC, Folsom AR, Wu K, Benderly M, Goldbourt U, Willeit J, Kiechl S, Yarnell JWG, Sweetnam PM, Elwood PC, Cushman M, Psaty BM, Tracy RP, Tybjaerg-Hansen A, Haverkate F, de Maat MPM, Fowkes FGR, Lee AJ, Smith FB, Salomaa V, Harald K, Rasi R, Vahtera E, Jousilahti P, Pekkanen J, D’Agostino R, Kannel WB, Wilson PWF, Tofler G, Arocha-Piñango CL, Rodriguez-Larralde A, Nagy E, Mijares M, Espinosa R, Rodriquez-Roa E, Ryder E, Diez-Ewald MP, Campos G, Fernandez V, Torres E, Marchioli R, Valagussa F, Rosengren A, Wilhelmsen L, Lappas G, Eriksson H, Cremer P, Nagel D, Curb JD, Rodriguez B, Yano K, Salonen JT, Nyyssönen K, Tuomainen T-P, Hedblad B, Lind P, Loewel H, Koenig W, Meade TW, Cooper JA, De Stavola B, Knottenbelt C, Miller GJ, Bauer KA, Rosenberg RD, Sato S, Kitamura A, Naito Y, Palosuo T, Ducimetiere P, Amouyel P, Arveiler D, Evans AE, Ferrieres J, JuhanVague I, Bingham A, Schulte H, Assmann G, Cantin B, Lamarche B, Després J-P, Dagenais GR, Tunstall-Pedoe

44


Systematic Review: CAD biomarkers in Asians and Caucasians

H, Woodward M, Ben-Shlomo Y, Davey Smith G, Palmieri V, Yeh JL, Rudnicka A, Ridker P, Rodeghiero F, Tosetto A, et al. Plasma fibrinogen level and the risk of major cardiovascular diseases and nonvascular mortality: an individual participant meta-analysis. JAMA. 2005;294:1799–809. 72. Bhopal R, Unwin N, White M, Yallop J. Heterogeneity of coronary heart disease risk factors in Indian, Pakistani, Bangladeshi, and European origin populations: cross sectional study. Bmj. 1999;215–220. 73. D’Agostino RB, Vasan RS, Pencina MJ, Wolf PA, Cobain M, Massaro JM, Kannel WB. General cardiovascular risk profile for use in primary care: the Framingham Heart Study. Circulation. 2008;117:743–53.

45


46

Cohort name/city, N

UKPDS, 4591

West Yorkshire, 490

West Yorkshire, 200

SHARE, 985

SHARE, 951

LOLIPOP, 300

UKADS, 1978

Framingham Offspring Study, 1895

Newcastle Heart Project, 1877

Newcastle Heart Project, 100

Wandsworth Heart and Stroke, 922

Wandsworth Heart and Stroke, 472

Wandsworth Heart and Stroke, 476

Wandsworth Heart and Stroke, 125

Author, year

Adler 1998

Ajjan, 2007

Kain, 2002

Anand, 2000

Anand, 2004

Bathula, 2010

Bellary, 2010

Bhalodkar, 2004

Bhopal, 2005

Reynolds, 2006

Cappuccio, 2002

Cook, 2001

Miller, 2003

Miller, 2009

No CVD.

No CVD.

Doctor diagnosis of angina or MI

No CVD.

Numbers from Bhopal 1999. Rose questionnaire and ECG.

Numbers from Bhopal 1999. Rose questionnaire and ECG.

5.8% of entire cohort had CAD.

CAD, stroke or peripheral vascular disease.

CAD.

AP or MI pain by Rose questionnaire. Silent myocardial infarction (ECG), percutaneous coronary angioplasty, or coronary artery bypass graft surgery; or cerebrovascular disease, defined by self-report of a previous stroke confirmed by a physician.

Cardiovascular disease=history of myocardial infarction, angina, silent myocardial infarction, PTCA, CABG, or stroke.

No CAD.

Healthy.

No CAD.

Inclusion/exlusion criteria concerning CAD in article

Supplemental table 1. CAD/CVD prevalence per cohort.

10.7%

No CAD

No CAD

No CAD

18%

9.3%

No CAD

No CAD

4%

No CAD

AP: 6%, ECG: 2%

AP: 6%, ECG: 2%

No CAD

No CAD

3%

No CAD

AP: 3%, ECG: 6%

AP: 3%, ECG: 6%

Not Not mentioned mentioned

21%

10.7%

Not Not mentioned mentioned

5.4%

No CAD

No CAD

No CAD

Caucasian South Asian

Not mentioned

2.4%

Chinese

Japanese

Not mentioned

NS

NS

Not mentioned

NS

NS

Not mentioned

Ca vs. SA and Ch vs. SA: significantly different, Ca vs. Ch: NS

Difference

Chapter 2


London, 44

London, 1025

Dallas, 137

London (Middlesex), 292

London (Tower Hamlets), 2074

Southall & Brent, 3207

Wolverhampton, 1000

Wolverhampton, 191

Manchester, 272

Akita & ARIC, 300

Chicago, 83

MESA/MASALA, 4228

MESA, 3422

MESA, 3113

SWAN, 2010

SWAN, 1879

Singapore, 30

Stockport, 71367

London, 143

Chambers, 1999

Chambers, 2001

Chandalia, 2003

Chowdhury, 2002

Chowdhury, 2006

Forouhi, 2006

Gama, 2002

Chatha, 2002

Heald, 2003

Iso, 1996

Kamath, 1999

Kanaya, 2014

Bild, 2005

Veeranna, 2012

Kelley-Hedgepeth, 2008

Matthews, 2005

Liew, 2003

Lyratzopoulos, 2005

Miller, 1988

Supplemental table 1. Continued

Histories of angina pectoris and pain of possible myocardial infarction were elicited by standard questionnaire. ECG performed.

No CVD.

Healthy.

No CVD.

No CVD.

No CVD.

No CVD.

No CVD.

No CVD.

No CAD.

CAD not excluded.

No CVD.

No CVD.

CVD not excluded.

CAD.

Previous vascular disease.

No CAD.

Healthy males.

Healthy males.

14.1%

4.8%

No CAD

No CAD

No CAD

No CAD

No CAD

AP: 7%. Q-waves: 3%

No CAD

No CAD

No CAD

No CAD

No CAD

No CAD

No CAD

No CAD

No CAD

AP: 41% (probably overreporting). Q-waves: 4%

No CAD

No CAD

No CAD

No CAD

Not Not mentioned mentioned

No CAD

No CAD

Not Not mentioned mentioned

7.2%

3.1%

No CAD

No CAD

No CAD

No CAD

No CAD

No CAD

No CAD

No CAD

No CAD

No CAD

No CAD

No CAD

Q-waves: NS.

Not mentioned

Mentioned elsewhere: 8.5% SA, 8.2% Ca, NS.

P<0.001

p-value for difference: 0.02

Systematic Review: CAD biomarkers in Asians and Caucasians

47


48

London/Arizona, 575

ERA JUMP/ Post-WWII, 607

ERA JUMP/ Post-WWII, 493

Houston, 70

Massachusetts, 24

Massachusetts, 40

CEPHEUS SA, 2005

Montreal, 165

Taiyuan/Belfast, 353

London, 92

Nagi, 1996

Okamura, 2013

Sekikawa, 2007

Patel, 2007

Raji, 2001

Raji, 2004

Rapeport, 2013

Smith, 2006

Zhang, 1993

Zoratti, 2000

No CAD.

No CAD.

No CVD.

History of CHD.

No CVD.

No CVD.

Healthy.

No CVD.

No CVD.

No CAD.

Absence of known co-morbidities expected to limit life expectancy to less than 5 years.

No CAD

No CAD

No CAD

35.4%

No CAD

No CAD

No CAD

No CAD

No CAD

No CAD

No CAD

No CAD

42.1%

No CAD

No CAD

No CAD

No CAD

Not Not mentioned mentioned

No CAD

No CAD

No CAD

Not mentioned

Not mentioned

Abbreviations Ca: Caucasians, Ch: Chinese, SA: South Asians. NS: not significant. AP: angina pectoris. IHD: ischaemic heart disease. CHD: coronary heart disease. MI: myocardial infarction. PTCA: percutaneous transluminal coronary angioplasty. CABG: coronary artery bypass grafting. ECG: electrocardiogram. Articles printed in grey are articles that describe a cohort that is described more extensively in another article. They add biomarker details but contain fewer persons or are from a less recent publication.

HeartSCORE & IndiaSCORE, 1298

Mulukutla, 2008

Supplemental table 1. Continued

Chapter 2


Systematic Review: CAD biomarkers in Asians and Caucasians

Supplemental table 2. Overview of relative differences from all articles from each comparison between Asians and Caucasians. Caucasians (n)

South Asians (n)

Rel. difference SA vs. C

Adler, 1998

4101

490

1.00

Ajjan, 2007

245

245

1.25

Anand, 2000

326

342

1.20

Bathula, 2010

149

151

1.12

Bellary, 2010

1.25

TG

492

1486

Chambers, 1999

18

26

1.27

Chambers, 2001

507

518

1.25

Chowdhury, 2002

127

165

1.27

Fourouhi, 2006

1787

1420

1.21

Iso, 1996

150

Kamath, 1999

44

39

1.65

Kanaya, 2014

2622

803

1.08

Kelly-Hedgepeth, 2008

1496

Liew, 2003

10

10

1.53 0.79

Miller, 1988

68

75

Miller, 2003

261

215

1.21

Mulukutla, 2008

715

194

1.50

Nagi, 1996

129

138

0.81

Okamura, 2013

297

Patel, 2007

35

35

1.19

Raji, 2001

12

12

1.41

Raji, 2004

25

15

1.31

Rapeport, 2013

1385

620

0.99

Reynolds, 2006

44

56

1.50

Zhang, 1993

202

Zoratti, 2000

31

31

1.26

Caucasians (n)

South Asians (n)

Rel. difference SA vs. C

Adler, 1998

4101

490

1.03

Ajjan, 2007

245

245

0.92

HDL

Anand, 2000

326

342

0.87

Bellary, 2010

492

1486

0.87

Bhalodkar, 2004

1684

211

1.00

Bhopal, 2005

1301

576

0.84

Cappuccio, 2002

475

447

0.85

Chambers, 1999

18

26

0.75

Chambers, 2001

507

518

0.92

Chowdhury, 2002

127

165

0.77

Fourouhi, 2006

1787

1420

0.92

Gama, 2002

787

223

0.93

Iso, 1996

150

Kamath, 1999

44

39

0.88

Kanaya, 2014

2622

803

0.92

Kelly-Hedgepeth, 2008

1496

Liew, 2003

10

10

0.85

Miller, 1988

68

75

0.91

Chinese (n)

Rel. difference Ch vs. C

317

1.09

803

Rel. difference J vs. C

150

0.77

270

1.04

310

1.05

Japanese (n)

Rel. difference J vs. C

150

1.27

270

1.11

1.08

244

1.07

10

0.97

151

1.07

Chinese (n)

Rel. difference Ch vs. C

317

1.00

803

Japanese (n)

0.92

244

1.09

10

1.00

49


Chapter 2

Supplemental table 2. Continued Mulukutla, 2008

720

Okamura, 2013

297

Patel, 2007

35

Raji, 2001

12

12

0.83

Raji, 2004

15

25

0.89

Rapeport, 2013

205

0.67

35

0.92

1385

620

0.92

Smith, 2006

83

82

0.72

Zhang, 1993

202

Zoratti, 2000

31

31

0.94

Caucasians (n)

South Asians (n)

Rel. difference SA vs. C

Adler, 1998

4101

490

0.93

Anand, 2000

326

342

1.15

Bathula, 2010

149

151

0.94

Bellary, 2010

492

1486

1.09

Bhopal, 2005

1301

576

0.98

Bild, 2005

2619

Cappuccio, 2002

475

447

0.92 1.00

TC

Chambers, 1999

18

26

Chambers, 2001

507

518

1.00

Chowdhury, 2002

127

165

0.96

Chowdury, 2006

1162

912

1.06

Fourouhi, 2006

1787

1420

0.98

Gama, 2002

787

223

0.98

Iso, 1996

150

Kain, 2001

100

100

0.93

44

39

1.11

10

10

1.09

Kamath, 1999 Kelly-Hedgepeth, 2008 Liew, 2003 Lyratzopoulos, 2005

1496 8550

66

1.01

Miller, 1988

68

75

0.89

Mulukutla, 2008

720

205

0.76

Patel, 2007

35

35

0.94

Rapeport, 2013

1385

620

0.99

Sekikawa, 2007

243

Smith, 2006

83

82

1.06

Zhang, 1993

202

Zoratti, 2000

31

31

0.97

Caucasians (n)

South Asians ( n)

Rel. difference SA vs. C

Adler, 1998

4101

490

0.96

Ajjan, 2007

245

245

0.98

Anand, 2000

326

342

1.07

Bathula, 2010

149

151

1.02

Glucose

Chambers, 1999

18

26

1.14

Chambers, 2001

507

518

1.08

Chowdhury, 2002

127

165

0.99

Fourouhi, 2006

1787

1420

1.07

50

151

0.82

Chinese (n)

Rel. difference Ch vs. C

317

0.99

803

0.98

244

0.99

10

0.91

151

0.70

Chinese (n)

Rel. difference Ch vs. C

317

1.01

310

1.14

Japanese (n)

Rel. difference J vs. C

150

0.92

270

1.02

250

1.03

Japanese (n)

Rel. difference J vs. C


Systematic Review: CAD biomarkers in Asians and Caucasians

Supplemental table 2. Continued Kanaya, 2014 Liew, 2003 Matthews, 2005

2622

803

1.14

803

1.08

10

10

1.07

10

1.02

231

1.02

1400

Miller, 1988

68

Miller, 2003

261

215

1.01

Nagi, 1996

129

138

0.97

Patel, 2007

35

85

1.07

Raji, 2001

12

12

1.06

Raji, 2004

75

15

25

1.02

1385

620

1.20

Reynolds, 2006

44

56

1.19

Sekikawa, 2007

243 31

31

1.02

Caucasians (n)

South Asians (n)

Rel. difference SA vs. C

Adler, 1998

4101

490

0.92

Anand, 2000

326

342

1.04

Bhalodkar, 2004

1684

211

0.97

Chowdhury, 2002

127

165

0.94

Iso, 1996

150

Kamath, 1999

44

39

1.16

Kanaya, 2014

2622

803

0.93

Kelly-Hedgepeth, 2008

1596

Rapeport, 2013

Zoratti, 2000

LDL

Miller, 1988

68

75

0.89

Mulukutla, 2008

717

205

0.81

Okamura, 2013

297

Patel, 2007

35

35

0.90

Raji, 2001

12

12

1.24

Raji, 2004

15

25

1.06

Rapeport, 2013

1385

620

1.02

Zhang, 1993

202

Zoratti, 2000

31

31

0.96

Caucasians (n)

South Asians (n)

Rel. difference SA vs. C

Adler, 1998

4101

490

1.16

Ajjan, 2007

245

245

1.36

Anand, 2004

322

323

1.38

Bathula, 2010

149

151

1.41

Insulin

Chinese (n)

Rel. difference Ch vs. C

317

0.99

803

0.97

244

0.93

151

0.63

Chinese (n)

Rel. difference Ch vs. C

306

1.08

1.05

Chambers, 1999

18

26

1.89

Chandalia, 2003

55

82

1.49

Fourouhi, 2006

1787

1420

1.47

Kanaya, 2014

2622

803

1.35

803

10

10

1.86

10

1.29

231

0.97

Liew, 2003 Matthews, 2005

1400

Miller, 2003

261

251

1.53

Nagi, 1996

129

138

1.30

Raji, 2001

12

12

1.96

Raji, 2004

15

25

1.81

31

1.69

Sekikawa, 2007 Zoratti, 2000

243 31

248

1.01

250

1.06

Japanese (n)

Rel. difference J vs. C

150

0.83

270

0.99

310

0.98

Japanese (n)

Rel. difference J vs. C

248

0.89

250

0.68

1.13

51


Chapter 2

Supplemental table 2. Continued Caucasians (n)

South Asians (n)

Ajjan, 2007

245

245

1.31

Anand, 2004

322

323

1.29

Chambers, 2001

507

518

1.16

Chandalia, 2003

55

82

1.49

Chatha, 2002

121

70

0.96

142

130

1.00

63

1.73

25

1.11

South Asians (n)

Rel. difference SA vs. C 1.05

CRP

Heald, 2003 Matthews, 2005 Miller, 2009 Okamura, 2013 Raji, 2004 Veeranna, 2012

Fibrinogen

Rel. difference SA vs. C

1400 61 15 2362 Caucasians (n) 245

245

Anand, 2000

326

342

1.05

Cook, 2001

205

235

0.98

Matthews, 2005 Miller, 1988

68 243

Veeranna, 2012

2362

75

South Asians (n)

Rel. difference SA vs. C

Adler, 1998

4101

490

0.98

Anand, 2004

322

323

1.09

Bathula, 2010

149

151

1.10

Bellary, 2010

492

1486

1.14

Chowdhury, 2002

127

165

1.02

Chowdury, 2006

1162

912

1.06

Miller, 1988

68

75

1.08

Rapeport, 2013

567

256

1.25

Caucasians (n)

South Asians (n)

Rel. difference SA vs. C

Ajjan, 2007

245

245

1.54

Anand, 2000

326

342

1.13

Iso, 1996

150

Matthews, 2005

0.60

231

0.50

751

0.52

Chinese (n)

Rel. difference Ch vs. C

317

0.99

231

0.95

751

0.99

Chinese (n)

Rel. difference Ch vs. C

306

1.04

Chinese (n)

Rel. difference Ch vs. C

317

1.04

231

0.91

Japanese (n)

Rel. difference J vs. C

248

0.36

310

0.39

Japanese (n)

Rel. difference J vs. C

150

0.86

248

0.87

250

0.85

Japanese (n)

Rel. difference J vs. C

Japanese (n)

Rel. difference J vs. C

0.93

Caucasians (n)

PAI-1

306

150 1400

Sekikawa, 2007

HbA1c

Rel. difference Ch vs. C

297

Ajjan, 2007

Iso, 1996

Chinese (n)

1400

Nagi, 1996

129

138

1.00

Raji, 2001

12

12

4.96

Raji, 2004

15

25

1.44

150

0.73

248

0.96

Abbreviations: Rel. difference: Relative difference. SA: South Asians, C: Caucasians, Ch: Chinese, J: Japanese. TG: triglycerides, HDL: high-density lipoprotein, TC: total cholesterol, LDL: low-density lipoprotein, HbA1c: glycated hemoglobin, CRP: C-reactive protein, PAI-1: plasminogen activator inhibitor 1. References for the articles mentioned in this table can be found in the main review article.

52


Systematic Review: CAD biomarkers in Asians and Caucasians

Supplemental table 3. Quality samples and methods. Article

Quality Sample

Biomarker Methodology

Adler, 1998

Overnight fasting

Limited description of methods

Ajjan, 2007

Overnight fasting

Methodology described for only some markers

Anand, 2000

Overnight fasting

Machines and techniques indicated

Anand, 2004

Overnight fasting

Machines and techniques indicated

Bathula, 2010

Overnight fasting

Machines and techniques indicated

Bellary, 2010

Unknown

No description of methodology

Bhalodkar, 2004

Incomplete fasting

Machines and techniques indicated

Bhopal, 2005

Overnight fasting

Methodology described for only some markers

Bild, 2005

Overnight fasting

Methodology described for only some markers

Cappuccio, 2002

Overnight fasting

Machines and techniques indicated

Chambers, 1999

Overnight fasting

Machines and techniques indicated

Chambers, 2001

Overnight fasting

Methodology described for only some markers

Chandalia, 2003

Overnight fasting

Machines and techniques indicated

Chatha, 2002

Unknown

Machines and techniques indicated

Chowdhury, 2002

Overnight fasting

No description of methodology

Chowdury, 2006

Unknown

No description of methodology

Cook, 2001

Overnight fasting

Machines and techniques indicated

Fourouhi, 2006

Overnight fasting

Machines and techniques indicated

Gama, 2002

Unknown

Machines and techniques indicated

Heald, 2003

Overnight fasting

Machines and techniques indicated

Iso, 1996

Unknown

Machines and techniques indicated

Kain, 2001

Overnight fasting

Methodology described for only some markers

Kamath, 1999

Overnight fasting

Methodology described for only some markers

Kanaya, 2014

Overnight fasting

Machines and techniques indicated

Kelly-Hedgepeth, 2008

Overnight fasting

Machines and techniques indicated

Liew, 2003

Overnight fasting

Machines and techniques indicated

Lyratzopoulos, 2005

Unknown

Machines and techniques indicated

Matthews, 2005

Overnight fasting

Machines and techniques indicated

Miller, 1988

Overnight fasting

Machines and techniques indicated

Miller, 2003

Overnight fasting

Methodology described for only some markers

Miller, 2009

Overnight fasting

Methodology described for only some markers

Mulukutla, 2008

Overnight fasting

Machines and techniques indicated

Nagi, 1996

Overnight fasting

Machines and techniques indicated

Okamura, 2013

Overnight fasting

Methodology described for only some markers

53


Chapter 2

Supplemental table 3. Continued Patel, 2007

Overnight fasting

Methodology described for only some markers

Raji, 2001

Overnight fasting

Machines and techniques indicated

Raji, 2004

Overnight fasting

Machines and techniques indicated

Rapeport, 2013

Overnight fasting

No description of methodology

Reynolds, 2006

Overnight fasting

Methodology described for only some markers

Sekikawa, 2007

Overnight fasting

Methodology described for only some markers

Smith, 2006

Non-fasting

No description of methodology

Veeranna, 2012

Overnight fasting

Machines and techniques indicated

Zhang, 1993

Overnight fasting

Machines and techniques indicated

Zoratti, 2000

Overnight fasting

Methodology described for only some markers

54


55



Part One Ethnicity

Chapter 3 Race/Ethnic Differences in the Associations of the Framingham Risk Factors with Carotid IMT and Cardiovascular Events PLoS One. 2015 Jul 2;10(7):e0132321

Crystel M. Gijsberts¶, Karlijn A. Groenewegen¶, Imo E. Hoefer Marinus J.C. Eijkemans, Folkert W. Asselbergs, Todd J. Anderson, Annie R. Britton, Jacqueline M. Dekker, Gunnar Engström, Greg W. Evans, Jacqueline de Graaf, Diederick E. Grobbee, Bo Hedblad, Suzanne Holewijn, Ai Ikeda, Kazuo Kitagawa, Akihiko Kitamura, Dominique P. de Kleijn, Eva M. Lonn, Matthias W. Lorenz, Ellisiv B. Mathiesen, Giel Nijpels, Shuhei Okazaki, Daniel H. O’Leary, Gerard Pasterkamp, Sanne A.E. Peters, Joseph F. Polak, Jacqueline F. Price, Christine Robertson, Christopher M. Rembold, Maria Rosvall, Tatjana Rundek, Jukka T. Salonen, Matthias Sitzer, Coen D.A. Stehouwer, Michiel L. Bots, Hester M. den Ruijter These authors contributed equally to this work.


Chapter 3

Abstract Background Clinical manifestations and outcomes of atherosclerotic disease differ between ethnic groups. In addition, the prevalence of risk factors is substantially different. Primary prevention programs are based on data derived from almost exclusively White people. We investigated how race/ethnic differences modify the associations of established risk factors with atherosclerosis and cardiovascular events. Methods We used data from an ongoing individual participant meta-analysis involving 17 population-based cohorts worldwide. We selected 60,211 participants without cardiovascular disease at baseline with available data on ethnicity (White, Black, Asian or Hispanic). We generated a multivariable linear regression model containing risk factors and ethnicity predicting mean common carotid intima-media thickness (CIMT) and a multivariable Cox regression model predicting myocardial infarction or stroke. For each risk factor we assessed how the association with the preclinical and clinical measures of cardiovascular atherosclerotic disease was affected by ethnicity. Results Ethnicity appeared to significantly modify the associations between risk factors and CIMT and cardiovascular events. The association between age and CIMT was weaker in Blacks and Hispanics. Systolic blood pressure associated more strongly with CIMT in Asians. HDL cholesterol and smoking associated less with CIMT in Blacks. Furthermore, the association of age and total cholesterol levels with the occurrence of cardiovascular events differed between Blacks and Whites. Conclusion The magnitude of associations between risk factors and the presence of atherosclerotic disease differs between race/ethnic groups. These subtle, yet significant differences provide insight in the etiology of cardiovascular disease among race/ethnic groups. These insights aid the race/ethnic-specific implementation of primary prevention.

58


Ethnic Differences in Framingham Risk Factors

Introduction Cardiovascular disease (CVD) has historically been considered a disease of the developed world.1 However, CVD is rapidly becoming the largest contributor to morbidity and mortality in growing economies with diverse race/ethnic groups.2 Furthermore, due to globally increasing mobility large race/ethnic minority groups arise in developed countries. As primary prevention of CVD is key, adequate risk prediction models are necessary to identify and treat high-risk individuals. To date, most research on primary prevention and risk scores of CVD has been conducted amongst Whites. For example, the landmark Framingham Risk Score (FRS)3 and the European SCORE4 have been developed in a largely White population. Despite recalibrating risk scores for specific race/ethnic groups, it has been shown that existing scores - even scores with ethnicity as a covariate - perform inconsistently among different race/ethnic groups. This results in both under- and overestimating risk, seriously compromising their usefulness in diverse race/ethnic groups.5 Both QRISK2 and the FRS, identified only 10% to 24% of individuals to be at high risk of those who experienced cardiovascular (CV) events among African Caribbeans. One study recalibrated the FRS and additionally studied the value of the risk factors individually for Whites, Blacks and Mexican Americans; revealing differences in risk factor association with cardiovascular disease.6 For example, for CVD mortality the hazard ratio (HR) of age was significantly higher in Whites as compared to Blacks and Mexican Americans. Also, the HR for highdensity lipoprotein (HDL) cholesterol was significantly higher in Mexican Americans when compared to Whites. The prevalence of several established cardiovascular risk factors (systolic blood pressure, use of antihypertensive drugs, diabetes, smoking, total cholesterol and HDL-cholesterol)7 also differs between race/ethnic groups.8 For example, diabetes is more prevalent in Blacks and Hispanics than in Asians and Whites.9 But whether differences in absolute risk factor levels also entail race/ethnic differences in the associations with atherosclerosis and CVD has not yet been clarified. This gap in our knowledge has very recently been underlined by the American Heart Association guidelines on the assessment of cardiovascular risk.10 They strongly recommend “continued research to fill gaps in knowledge regarding short- and long-term atherosclerotic cardiovascular disease risk assessment and outcomes in all race/ethnic groups (...) Further research should include analyses of short- and long-term risk in diverse groups�. In order to fill this knowledge gap, a large multi-ethnic cohort with a sufficient number of CV events is needed. For this purpose we used the individual participant data metaanalysis USE-IMT cohort.11 In addition to analyzing the associations with CV events, this cohort also offers the opportunity to assess differences among race/ethnic groups in the association of risk factors to subclinical atherosclerosis measured by mean common carotid intima media thickness (CIMT).

59


Chapter 3

Methods Study population USE-IMT is an ongoing individual participant data meta-analysis of which the methods have been described in detail elsewhere.11 In short, general population cohorts were identified using literature search and expert suggestions. For inclusion in USE-IMT, cohorts were required to have available baseline data on age, sex, blood pressure, cholesterol fractions, smoking status, use of antihypertensive medication, diabetes mellitus and CIMT-measurements and follow-up information on occurrence of cardiovascular events. For the current analysis, we included the participating cohorts of which data on ethnicity on an individual level were available (n=9). Cohorts that did not have information on ethnicity were included if it was reasonable to assume that >95% of the participants belonged to one race/ethnic group due to either selection of participants or race/ethnic homogeneity of the source population (n=6). Ethnicity was recoded when applicable, to create uniform race/ethnic groups for analysis. Details concerning this recoding process and the original classification can be found in S1 Table. As our research is focused on the first-time events in asymptomatics, only individuals to whom the Framingham criteria 7 are applicable were included in the current analyses. Ethics statement This study was approved by the institutional review committee of the University Medical Centre Utrecht. Each individual cohort obtained approval from a local Ethical Review Board and written informed consent from all participants. All authors exchanged a material transfer agreement. This study conforms to the declaration of Helsinki. USE-IMT, CIMT and CV events Out of the 17 cohorts participating in USE-IMT, we included 15 cohorts12–26, consisting of 66,213 individuals. After excluding individuals of whom individual ethnicity was not known (n=285) or who had already experienced a cardiovascular event (n=5,717) 60,211 individuals were included for analyses. This group consisted of 46,788 Whites, 7,200 Blacks, 3,816 Asians and 2,407 Hispanics. Incomplete data on mean common CIMT, cardiovascular risk factors, and (time to) CV events, approximately 12% of total values, were imputed, as described previously.11 This imputation was done to increase power and to reduce omitted variable bias.27 Average mean common CIMT was calculated for each individual using the maximum set of carotid angles, near and/or far wall measurements, and left and/or right side measurements that were assessed within each cohort. Time to first fatal or non-fatal myocardial infarction or stroke (hemorrhagic or ischemic) was used as a primary endpoint in this analysis.

60


Ethnic Differences in Framingham Risk Factors

Statistical Analysis Baseline data are represented as means with standard deviations (SD) for continuous variables and as percentages for categorical variables. We describe our population both by their original cohort (Table 1) and by race/ethnic group (Table 2). All statistical analyses were performed in R (version 2.15.1). To examine the influence of ethnicity on associations of risk factors with mean common CIMT and CV events, we used linear regression models and Cox regression models, respectively. For both questions, we first fitted a model containing the Framingham risk factors (age, sex, systolic blood pressure (SBP), HDL-cholesterol, total cholesterol, current smoking, presence of diabetes and antihypertensive drug use)7 and ethnicity as independent variables. Mean common CIMT (linear regression-log transformed) and a combined endpoint of first-time stroke or myocardial infarction (Cox regression) were used as an outcome, respectively. To take between-study-variation into account, we took a random effects approach for each original cohort. We then extended this model by adding the interaction terms for ethnicity with each risk factor. Whites were the reference group in all analyses. We first compared the interaction model to the first model using a likelihood ratio test. Subsequently, we tested the significance of the interaction terms. Finally, we calculated hazard ratios and betas (with 95% confidence intervals) for each risk factor in each race/ ethnic group, and tested for significance of individual interaction terms using a Wald test. Due to the large number of tests performed, we took a conservative threshold for interaction at pinteraction=0.05 (commonly accepted p-values for interactions are 0.1 or 0.2). In these analyses, Whites were the reference group. Interaction terms were only considered significant when significantly adding to the model without interaction terms and when significantly different from Whites. No comparisons between ethnicities were performed, except for the comparisons to Whites. A two-sided significance level of 0.05 was used.

Results The White race/ethnic group was derived from 13 cohorts that originated from Canada, the United States of America (USA) and Europe. The Black group was derived from 7 cohorts, with the majority of individuals residing in the USA. The Asian group comprised of 8 cohorts. The majority of Asians (63.5%) originated from Japan; another substantial part (21% of total) of Asians was Chinese American. The Hispanics were derived from three (two American and one Canadian) cohorts. The characteristics of each cohort are described in Table 1. The mean age for the entire cohort was 59 years (SD 10), 51% were male. Median follow-up time was 9 years, with a total follow-up of 547,887 person years. During this time 4,730 CV events, i.e., fatal or non-fatal myocardial infarction or stroke, occurred.

61


62

1,939

USA

Japan

CHS14

USA

Japan

UK

Whitehall26

60,211

9,799 46,788

8,896

5,699

-

256

1,160

2,622

5,163

908

308

1,548

980

-

3,700

4,798

10,750

White (n)

7,200

354

-

-

295

9

1,893

-

-

-

10

-

-

661

-

3,978

Black (n)

3,816

549

-

484

33

3

803

-

-

-

1

-

1,939

4

-

-

Asian (n)

2,407

-

-

-

910

-

1,496

-

-

-

1

-

-

-

-

-

Hispanic (n)

59.0

61.2

59.3

65.5

68.9

60.8

62.2

57.5

51.2

68.7

49.4

68.9

65.5

72.4

49.4

54.0

Age (years)

51.3

67.1

47.2

49.6

40.0

46.7

47.2

40.5

100.0

48.1

99.8

49.1

75.7

38.6

48.1

43.1

Gender (% men)

0.75 (0.17)

0.78 (0.15)

0.78 (0.16)

0.87 (0.27)

0.73 (0.09)

0.83 (0.11)

0.76 (0.18)

0.77 (0.15)

0.76 (0.16)

0.85 (0.15)

0.72 (0.18)

0.77 (0.28)

-

0.87 (0.16)

0.72 (0.14)

0.65 (0.15)

Mean CIMT (mm, sd)

9.1

6.0

10.1

4.4

7.9

3.8

6.0

10.4

13.1

7.4

7.5

12.2

7.9

10.4

8.0

12.3

FU (years)

2,318

110

352

22

63

3

116

184

58

6

11

28

89

645

91

540

Stroke (n)

2,736

138

534

2

55

13

140

186

114

11

22

11

23

590

68

829

MI (n)

4,730

244

830

24

108

16

355

251

159

16

33

39

109

1,108

153

1,285

CV event (n)

Abbreviations. ARIC: Atherosclerosis Risk in Communities Study; CAPS: Carotid Atherosclerosis Progression Study; CHS: Cardiovascular Health Study; CIRCS: Circulatory Risk in Communities Study; EAS: Edinburgh Artery Study; FATE: The Firefighters and Their Endothelium Study; Hoorn: The Hoorn Study; KIHD: Kuopio Ischaemic Heart Disease Risk Factor Study; Malmö: Malmö Diet and Cancer Study, MESA; Multi-race/ethnic Study of Atherosclerosis; NBS: Nijmegen Biomedical Study 2; NOMAS: Northern Manhattan Study; OSACA2: Osaka Follow-Up Study for Carotid Atherosclerosis 2; Tromsø: Tromsø Study; Whitehall: Whitehall II Study; CIMT: mean common carotid intima media thickness; FU: follow-up duration; MI: myocardial infarction; USA: United States of America; UK: United Kingdom; NLD: The Netherlands.

Combined

Norway

Tromsø25

24

5,699

484

USA

NOMAS23

OSACA2

1,494

NLD

NBS22

21

1,172

6,814

Sweden

Malmö20

MESA

5,163

Finland

908

308

NLD

KHID19

Hoorn

18

1,560

Canada

FATE17

980

UK

4,798

EAS16

CIRCS

15

4,365

Germany

CAPS13

14,728

USA

ARIC12

Individuals (n)

Country

Cohort

Table 1. Details of participating USE-IMT cohorts.

Chapter 3


Ethnic Differences in Framingham Risk Factors

Asians and Hispanics were older than Blacks and Whites (64 versus 58 years). While sex in Whites was balanced, Blacks and Hispanics were mainly women (61% and 55%) and Asians were predominantly male (65%). Blacks and Hispanics had notably higher BMIs and a higher prevalence of diabetes mellitus as compared to Whites and Asians. Hispanics had lower rates of smoking (13.8%) compared to Blacks (23.7%), while the prevalence of smoking in Whites and Asians was comparable (respectively 19.6% and 21.3%). Mean LDL-cholesterol was highest in Whites (3.8 mmol/L). In Asians and Hispanics mean LDL-cholesterol was 3.2 mmol/L, in Blacks it was 3.4 mmol/L. Systolic blood pressure was 136 mmHg in Asians, 132 mmHg in Hispanics, 131 mmHg in Blacks and 130 mmHg in Whites. Baseline data for each race/ethnic group are presented in Table 2.

Table 2. Baseline properties per race/ethnic group Whites

Blacks

Asians

Hispanics

Total

Individuals (n)

46,788

7,200

3,816

2,407

60,211

Age, years

58.4 (10.1)

58.6 (9.9)

64.3 (7.3)

63.4 (10.0)

59.0 (10.0)

Gender, % men

52.4

39.3

64.5

44.5

51.3

CIMT, mm

0.74 (0.17)

0.74 (0.18)

0.78 (0.20)

0.74 (0.15)

0.75 (0.17)

Smoking, % yes

19.6

23.7

21.3

13.8

20.0

Diabetes, % yes

5.9

18.0

8.2

17.0

7.9

BMI, kg/m2

26.6 (4.4)

29.6 (6.0)

23.7 (3.3)

29.0 (5.0)

26.9 (4.8)

TC, mmol/L

5.8 (1.2)

5.4 (1.1)

5.3 (0.9)

5.2 (1.0)

5.7 (1.2)

HDL, mmol/L

1.4 (0.4)

1.4 (0.4)

1.4 (0.4)

1.2 (0.3)

1.4 (0.4)

LDL, mmol/L

3.8 (1.1)

3.4 (1.0)

3.2 (0.8)

3.2 (0.9)

3.7 (1.1)

TG, mmol/L

1.5 (1.0)

1.2 (0.8)

1.5 (0.9)

1.7 (1.0)

1.5 (0.9)

SBP, mmHg

130 (21)

131 (22)

136 (21)

132 (22)

131 (21)

DBP, mmHg

77 (12)

78 (12)

79 (12)

76 (12)

77 (12)

Stroke events (n)

1,732

398

125

63

2,318

MI events (n)

2,302

328

46

60

2,736

CV events (na)

3,780

665

168

117

4,730

Mean FU duration (years) 9.3

9.6

6.8

6.6

9.1

10-y event rateb

9.2

6.7

7.8

8.2

8.1

All data represent means (sd), unless stated otherwise. a Number of individuals with a CV event (first-time stroke or MI). b10-year event rate, estimated using Kaplan-Meier analysis.

Common CIMT Mean common CIMT was 0.74 mm in Blacks, Whites and Hispanics and 0.78 mm in Asians. The model with risk factors and interaction terms with ethnicity fitted the data better than the model with risk factors alone (likelihood ratio test p<0.001). The interaction terms of ethnicity with age, HDL-cholesterol, smoking and systolic blood pressure were statistically significant (p <0.05). We thus found that for specific ethnicities risk factor

63


Chapter 3

associations were significantly different from Whites. The magnitude of the association (i.e. betas of the linear regression model) of age with logCIMT was significantly smaller in Blacks and Hispanics as compared to Whites: 0.08/10 years (91% of the White beta) and 0.09/10 years (97% of the White beta), respectively. This means that every 10-year increase in age gives less increase in mean common CIMT in Blacks and Hispanics than it does in Whites. The beta for HDL-cholesterol per 1 mmol/L increase was -0.05 (144%) in Blacks (a stronger inverse association) as compared to the beta in Whites. The beta for smoking (yes vs. no) was 0.01 (30% of the beta in Whites). In Asians the beta for SBP was 0.02 (131% of the White beta). All betas are plotted in figure 1 and are presented in Table 3.

Figure 1. Association between risk factors and mean common CIMT, by ethnicity. Point estimates for betas, lines represent 95% confidence intervals.

64


0.04 (0.04 - 0.05)

0.02 (0.01 - 0.02)

0.04 (0.04 - 0.05)

-0.03 (-0.04 -0.03)

0.01 (0.01 - 0.02)

0.03 (0.02 - 0.03)

0.01 (0.01 - 0.01)

Diabetes

BP drug use

Gender

HDL

SBP

Smoking

TC

1.95 (1.86-2.05)

1.28 (1.21-1.35)

1.52 (1.45-1.59)

0.67 (0.58-0.76)

1.15 (1.14-1.17)

1.96 (1.89-2.04)

1.09 (1.07-1.12)

Diabetes

BP drug use

Gender

HDL

SBP

Smoking

TC 100%

100%

100%

100%

100%

100%

100%

100%

100%

100%

100%

100%

100%

100%

100%

100%

1.20 (1.13-1.26)a

1.70 (1.53-1.87)

1.18 (1.15-1.21)

0.64 (0.44-0.85)

1.35 (1.19-1.51)

1.50 (1.34-1.66)

2.13 (1.97-2.30)

1.52 (1.44-1.60)a

0.02 (0.01 - 0.02)

a

0.01 (-0.00 - 0.02)

0.01 (0.01 - 0.02)

-0.05 (-0.06 -0.04)a

0.05 (0.04 - 0.06)

0.01 (-0.00 - 0.01)

0.04 (0.03 - 0.05)

0.08 (0.08 - 0.09)a

Reference Black

110%

87%

103%

96%

89%

117%

109%

80%

140%

30%

98%

144%

118%

38%

100%

91%

difference %

0.95 (0.78-1.13)

1.59 (1.25-1.93)

1.12 (1.05-1.20)

0.88 (0.46-1.29)

1.82 (1.42-2.22)

1.47 (1.14-1.79)

2.62 (2.20-3.03)

1.75 (1.49-2.01)

0.01 (-0.00 - 0.02)

0.05 (0.03 - 0.08)

0.02 (0.01 - 0.02)a

-0.04 (-0.06 -0.02)

0.04 (0.02 - 0.06)

0.03 (0.01 - 0.04)

0.07 (0.04 - 0.09)

0.09 (0.08 - 0.10)

Asian

87%

81%

98%

132%

119%

115%

134%

92%

73%

205%

131%

110%

93%

166%

160%

97%

difference %

1.15 (0.96-1.34)

1.58 (1.09-2.08)

1.20 (1.12-1.29)

0.91 (0.33-1.49)

1.56 (1.17-1.95)

0.93 (0.53-1.32)

2.59 (2.20-2.98)

1.69 (1.48-1.90)

0.01 (0.01 - 0.02)

0.02 (-0.00 - 0.04)

0.01 (0.01 - 0.02)

-0.02 (-0.04 0.01)

0.04 (0.03 - 0.06)

0.01 (-0.01 - 0.03)

0.03 (0.01 - 0.05)

0.08 (0.07 - 0.09)a

Hispanic

105%

81%

104%

137%

102%

72%

133%

89%

105%

66%

91%

48%

100%

58%

70%

90%

difference %

The top half of the table displays betas and 95% confidence intervals of the Framingham risk factors for log CIMT for each race/ethnic group. The bottom half of the table displays hazard ratios and 95% confidence intervals of the Framingham risk factors for CV events for each race/ethnic group. The % difference columns express the percentage difference in effect size (beta or hazard ratio) as compared to the White race/ethnic group. Risk factors printed in bold have significantly different effect sizes among race/ ethnic groups (significant interaction). a Indicates significant difference as compared to Whites (p<0.05). b Cardiovascular events (first-time stroke or MI).

1.89 (1.86-1.93)

Age

CV events (HRs, 95% CI)

b

0.09 (0.09 - 0.09)

Age

IMT (betas, 95% CI)

White

Table 3. Effects of the Framingham risk factors in race/ethnic groups for the outcomes log CIMT (betas) and CV events (hazard ratios).

Ethnic Differences in Framingham Risk Factors

65


Chapter 3

CV events The 10-year event rate was 6.7% in Asians, 7.8% in Hispanics, 8.1% in Whites and 9.2% in Blacks. The model including the risk factors and the interaction terms with ethnicity fitted the data better than the model without interaction terms (likelihood ratio test p <0.001). Both age and total cholesterol had a significant interaction with ethnicity. The HR for CV events (first-time stroke or MI) per 10-year increase in age was lower in Blacks than in Whites (1.52 (1.44-1.60), 80% of the HR in Whites). The HR for a 1 mmol/L increase of total cholesterol was higher in Blacks (1.20 (1.13-1.26), 140%) compared to Whites. The hazard ratios of the Framingham risk factors per ethnic group are plotted in figure 2.

Figure 2. Association between risk factors and first-time stroke or myocardial infarction, by ethnicity. Point estimates for hazard ratios, lines represent 95% confidence intervals.

66


Ethnic Differences in Framingham Risk Factors

Discussion Our study shows that the associations of Framingham risk factors with subclinical atherosclerosis and CV events have similar directions across race/ethnic groups. However, the magnitude of the associations differs significantly for several risk factors among race/ ethnic groups. Prevalences of risk factors The prevalence of risk factors in our cohort was unevenly distributed among the race/ ethnic groups (Table 2). Race/ethnic differences in the prevalence of CVD risk factors have been previously described. Similar to our study, smoking was least prevalent in Asians.28 Also diabetes was more prevalent in Blacks and Hispanics than in Asians and Whites9, and the higher prevalence of diabetes was accompanied with higher BMI in these groups.29 Also in accordance with literature, TC and LDL-cholesterol levels were highest in Whites, HDL-cholesterol levels were lowest in Hispanics and triglyceride levels were lowest in Blacks.30 However, in contrast with these similarities between our data and published literature, blood pressure was highest in Asians, which is opposite to the findings in the Asia Pacific cohort.31 Possibly, this incongruence between our study and Asia Pacific can be explained by differences in age, as the Asians in our cohort are older than the Whites (in the Asia Pacific cohort the group from Australia and New Zealand is older than the Asians). The observations made above indicate that our cohort is fairly similar to other cohorts described in literature, thereby suggesting that our results might be generalizable. CIMT and CVD - Differences and similarities For Blacks, an increase in age was related to both an increase in mean common CIMT and a higher risk of CV events, in our analyses. The magnitude of these associations, however, was significantly smaller than in Whites. Although the direction of the association of age is the same in Blacks as in Whites, an increasing age has less effect on disease in Blacks than in Whites. This indicates that age is one of the factors that should be should weighted when developing a prediction model specifically for Blacks. HDL-cholesterol and smoking in Blacks, SBP in Asians and age in Hispanics showed a different magnitude of the association with mean common CIMT as compared to Whites. However, these differences in the magnitude with mean common CIMT were not detected in the analysis of CV events. This might be due to the difference in power between linear regression and Cox proportional hazards analysis (which is driven by the number of events), or a difference in biology of CIMT and the actual occurrence of CV events. The association between TC and CV event was significantly higher for Blacks as compared to Whites. The coefficient for TC on mean common CIMT was also higher than the coefficient in Whites, although this difference did not reach statistical significance. When it comes to geographical differences, some of these differences have been reported before. The Asia Pacific study reported lower HRs for triglycerides and SBP for

67


Chapter 3

respectively coronary heart disease and hemorrhagic stroke, when comparing Asian countries with Australia and New-Zealand.32 The INTERHEART study33 showed in an analysis of 52 countries across the globe that odds ratios for myocardial infarction were comparable, although (small) differences in effect sizes exist. These geographical differences are in agreement with our race/ethnicity-specific findings. Race/ethnicity-specific risk prediction should be exercised with discretion. Unwanted medical “discrimination� that could be perceived by patients should be avoided at all costs. On the other hand optimal risk estimation is of paramount importance for medical care. Adding race/ethnicity to risk prediction equations thus should be implemented with care and reasons to do so should be clarified to all parties, as to avoid undesirable situations in health care and society. Limitations Due to differences in ascertaining and recording ethnicity per cohort, misclassification might have occurred and may have diluted our results to some extent. While we are aware of the growing number of people from mixed backgrounds, people of mixed race/ ethnicity were excluded from the current analyses. Also, our Asian and Hispanic race/ ethnic groups were relatively small, which may have led to a lack of power. Therefore, our conclusions on these groups should be interpreted with caution. We were unable to take immigration status or acculturation into account, although time since immigration has been shown to influence mean common CIMT, risk factors levels and risk of cardiovascular events.34,35 Therefore we cannot determine in which way immigration status might influence our results, or whether the results are influenced at all. In our CV events analysis we were unable to take differences in prevention and treatment strategies into account, these might differ among the ethnic groups and influence the occurrence of CV events. Conclusion In conclusion, the associations of the Framingham risk factors with atherosclerosis (CIMT) and CVD had similar directions across race/ethnic groups. However, the magnitude of associations between risk factors and the presence of atherosclerotic disease differs between race/ethnic groups. These subtle, yet significant differences provide insight in the etiology of cardiovascular disease among race/ethnic groups. These insights aid the race/ethnic-specific implementation of primary prevention.

68


Ethnic Differences in Framingham Risk Factors

References 1.

Alwan A. Global status report on noncommunicable diseases 2010. 1st ed. Geneva, Switzerland: 2011.

2.

World Health Organization. World Health Statistics 2012. Geneva, Switzerland: 2012.

3.

Kannel WB, McGee D, Gordon T. A general cardiovascular risk profile: the Framingham Study. Am J Cardiol. 1976;38:46–51.

4.

Conroy RM, Pyörälä K, Fitzgerald AP, Sans S, Menotti A, De Backer G, De Bacquer D, Ducimetière P, Jousilahti P, Keil U, Njølstad I, Oganov RG, Thomsen T, Tunstall-Pedoe H, Tverdal A, Wedel H, Whincup P, Wilhelmsen L, Graham IM. Estimation of ten-year risk of fatal cardiovascular disease in Europe: the SCORE project. Eur Heart J. 2003;24:987–1003.

5.

Tillin T, Hughes AD, Whincup P, Mayet J, Sattar N, McKeigue PM, Chaturvedi N. Ethnicity and prediction of cardiovascular disease: performance of QRISK2 and Framingham scores in a U.K. tri-ethnic prospective cohort study (SABRE--Southall And Brent REvisited). Heart. 2014;100:60–7.

6.

Hurley LP, Dickinson LM, Estacio RO, Steiner JF, Havranek EP. Prediction of cardiovascular death in racial/ ethnic minorities using Framingham risk factors. Circ Cardiovasc Qual Outcomes. 2010;3:181–7.

7.

D’Agostino RB, Vasan RS, Pencina MJ, Wolf PA, Cobain M, Massaro JM, Kannel WB. General cardiovascular risk profile for use in primary care: the Framingham Heart Study. Circulation. 2008;117:743–53.

8.

Anand SS, Yusuf S, Vuksan V, Devanesen S, Teo KK, Montague PA, Kelemen L, Yi C, Lonn E, Gerstein H, Hegele RA, McQueen M. Differences in risk factors, atherosclerosis, and cardiovascular disease between ethnic groups in Canada: the Study of Health Assessment and Risk in Ethnic groups (SHARE). Lancet. 2000;356:279–84.

9.

Spanakis EK, Golden SH. Race/ethnic difference in diabetes and diabetic complications. Curr Diab Rep. 2013;13:814–23.

10. Goff DC, Lloyd-Jones DM, Bennett G, Coady S, D’Agostino RB, Gibbons R, Greenland P, Lackland DT, Levy D, O’Donnell CJ, Robinson JG, Schwartz JS, Shero ST, Smith SC, Sorlie P, Stone NJ, Wilson PWF, Jordan HS, Nevo L, Wnek J, Anderson JL, Halperin JL, Albert NM, Bozkurt B, Brindis RG, Curtis LH, DeMets D, Hochman JS, Kovacs RJ, Ohman EM, Pressler SJ, Sellke FW, Shen W-K, Tomaselli GF. 2013 ACC/AHA guideline on the assessment of cardiovascular risk: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines. Circulation. 2014;129:S49–73. 11. Den Ruijter HM, Peters S a E, Anderson TJ, Britton AR, Dekker JM, Eijkemans MJ, Engström G, Evans GW, de Graaf J, Grobbee DE, Hedblad B, Hofman A, Holewijn S, Ikeda A, Kavousi M, Kitagawa K, Kitamura A, Koffijberg H, Lonn EM, Lorenz MW, Mathiesen EB, Nijpels G, Okazaki S, O’Leary DH, Polak JF, Price JF, Robertson C, Rembold CM, Rosvall M, Rundek T, Salonen JT, Sitzer M, Stehouwer CD a, Witteman JC, Moons KG, Bots ML. Common carotid intima-media thickness measurements in cardiovascular risk prediction: a metaanalysis. JAMA. 2012;308:796–803. 12. Li R, Duncan BB, Metcalf PA, Crouse JR, Sharrett AR, Tyroler HA, Barnes R, Heiss G. B-mode-detected carotid artery plaque in a general population. Atherosclerosis Risk in Communities (ARIC) Study Investigators. Stroke. 1994;25:2377–83. 13. Lorenz MW, von Kegler S, Steinmetz H, Markus HS, Sitzer M. Carotid intima-media thickening indicates a higher vascular risk across a wide age range: prospective data from the Carotid Atherosclerosis Progression Study (CAPS). Stroke. 2006;37:87–92. 14. Fried LP, Borhani NO, Enright P, Furberg CD, Gardin JM, Kronmal RA, Kuller LH, Manolio TA, Mittelmark MB, Newman A. The Cardiovascular Health Study: design and rationale. Ann Epidemiol. 1991;1:263–276.

69


Chapter 3

15. Imano H, Kitamura A, Sato S, Kiyama M, Ohira T, Yamagishi K, Noda H, Tanigawa T, Iso H, Shimamoto T. Trends for blood pressure and its contribution to stroke incidence in the middle-aged Japanese population: the Circulatory Risk in Communities Study (CIRCS). Stroke. 2009;40:1571–7. 16. Price JF, Tzoulaki I, Lee AJ, Fowkes FGR. Ankle brachial index and intima media thickness predict cardiovascular events similarly and increased prediction when combined. J Clin Epidemiol. 2007;60:1067–75. 17. Anderson TJ, Charbonneau F, Title LM, Buithieu J, Rose MS, Conradson H, Hildebrand K, Fung M, Verma S, Lonn EM. Microvascular function predicts cardiovascular events in primary prevention: long-term results from the Firefighters and Their Endothelium (FATE) study. Circulation. 2011;123:163–9. 18. Henry RMA, Kostense PJ, Spijkerman AMW, Dekker JM, Nijpels G, Heine RJ, Kamp O, Westerhof N, Bouter LM, Stehouwer CDA. Arterial stiffness increases with deteriorating glucose tolerance status: the Hoorn Study. Circulation. 2003;107:2089–95. 19. Salonen R, Salonen JT. Determinants of carotid intima-media thickness: a population-based ultrasonography study in eastern Finnish men. J Intern Med. 1991;229:225–31. 20. Rosvall M, Ostergren PO, Hedblad B, Isacsson SO, Janzon L, Berglund G. Occupational status, educational level, and the prevalence of carotid atherosclerosis in a general population sample of middle-aged Swedish men and women: results from the Malmö Diet and Cancer Study. Am J Epidemiol. 2000;152:334–46. 21. Mora S, Szklo M, Otvos JD, Greenland P, Psaty BM, Goff DC, O’Leary DH, Saad MF, Tsai MY, Sharrett AR. LDL particle subclasses, LDL particle size, and carotid atherosclerosis in the Multi-Ethnic Study of Atherosclerosis (MESA). Atherosclerosis. 2007;192:211–7. 22. Holewijn S, den Heijer M, Swinkels DW, Stalenhoef AFH, de Graaf J. The metabolic syndrome and its traits as risk factors for subclinical atherosclerosis. J Clin Endocrinol Metab. 2009;94:2893–9. 23. Prabhakaran S, Singh R, Zhou X, Ramas R, Sacco RL, Rundek T. Presence of calcified carotid plaque predicts vascular events: the Northern Manhattan Study. Atherosclerosis. 2007;195:e197–201. 24. Kitagawa K, Hougaku H, Yamagami H, Hashimoto H, Itoh T, Shimizu Y, Takahashi D, Murata S, Seike Y, Kondo K, Hoshi T, Furukado S, Abe Y, Yagita Y, Sakaguchi M, Tagaya M, Etani H, Fukunaga R, Nagai Y, Matsumoto M, Hori M. Carotid intima-media thickness and risk of cardiovascular events in high-risk patients. Results of the Osaka Follow-Up Study for Carotid Atherosclerosis 2 (OSACA2 Study). Cerebrovasc Dis. 2007;24:35–42. 25. Stensland-Bugge E, Bønaa KH, Joakimsen O, Njølstad I. Sex differences in the relationship of risk factors to subclinical carotid atherosclerosis measured 15 years later : the Tromsø study. Stroke. 2000;31:574–81. 26. Halcox JPJ, Donald AE, Ellins E, Witte DR, Shipley MJ, Brunner EJ, Marmot MG, Deanfield JE. Endothelial function predicts progression of carotid intima-media thickness. Circulation. 2009;119:1005–12. 27. Moons KGM, Donders RART, Stijnen T, Harrell FE. Using the outcome for imputation of missing predictor values was preferred. J Clin Epidemiol. 2006;59:1092–101. 28. Centers for Disease Control and Prevention (CDC). Cigarette smoking among adults and trends in smoking cessation - United States, 2008. MMWR Morb Mortal Wkly Rep. 2009;58:1227–32. 29. Wei L, Wu B. Racial and ethnic differences in obesity and overweight as predictors of the onset of functional impairment. J Am Geriatr Soc. 2014;62:61–70. 30. Frank ATH, Zhao B, Jose PO, Azar KMJ, Fortmann SP, Palaniappan LP. Racial/ethnic differences in dyslipidemia patterns. Circulation. 2014;129:570–9. 31. Kengne AP, Patel A, Barzi F, Jamrozik K, Lam TH, Ueshima H, Gu DF, Suh I, Woodward M. Systolic blood pressure, diabetes and the risk of cardiovascular diseases in the Asia-Pacific region. J Hypertens. 2007;25:1205–13.

70


Ethnic Differences in Framingham Risk Factors

32. Woodward M, Huxley H, Lam TH, Barzi F, Lawes CMM, Ueshima H. A comparison of the associations between risk factors and cardiovascular disease in Asia and Australasia. Eur J Cardiovasc Prev Rehabil. 2005;12:484–91. 33. Yusuf S, Hawken S, Ounpuu S, Dans T, Avezum A, Lanas F, McQueen M, Budaj A, Pais P, Varigos J, Lisheng L. Effect of potentially modifiable risk factors associated with myocardial infarction in 52 countries (the INTERHEART study): case-control study. Lancet. 2004;364:937–52. 34. Lear SA, Humphries KH, Hage-Moussa S, Chockalingam A, Mancini GBJ. Immigration presents a potential increased risk for atherosclerosis. Atherosclerosis. 2009;205:584–9. 35. De Maio FG. Immigration as pathogenic: a systematic review of the health of immigrants to Canada. Int J Equity Health. 2010;9:27.

71


Chapter 3

Supplemental S1 Table. Recoding of race/ethnicity per cohort Cohort name

Original classification

Recoded for analysis

ARIC

Caucasian Black

Caucasian Black

CAPS

All Caucasian

Caucasian

CHS

Caucasian Black American Indian Asian or Pacific Islander Other

Caucasian Black Other Asian Other

CIRCS

All Japanesea

Asian

EAS

All Caucasiana

Caucasian

FATE

Caucasian Black Hispanic Asian or Pacific Islander Canaduab East Indian Other

Caucasian Black Hispanic Asian Other Asian Other

Hoorn

All Caucasiana

Caucasian

KIHD

All Caucasiana

Caucasian

Malmรถ

All Caucasian

Caucasian

MESA

Caucasian Chinese American Black Hispanic

Caucasian Asian Black Hispanic

NBS

Canada Surinam Austria Belgium Hungary Morocco Iraq Finland Greece Libya United States Luxembourg Indonesia Germany Netherlands Spain Great Britain Turkey Yugoslavia Mexico Netherlands Antilles Poland Russia Japan Uruguay Italy New Guinea Netherlands Indie

Caucasian Black Caucasian Caucasian Caucasian Other Other Caucasian Caucasian Other Caucasian Caucasian Asian Caucasian Caucasian Caucasian Caucasian Other Caucasian Hispanic Black Caucasian Caucasian Asian Other Caucasian Other Other

72

a


Ethnic Differences in Framingham Risk Factors

S1 Table. Continued Cohort name

Original classification

Recoded for analysis

NOMAS

Caucasian Chinese American Black Hispanic

Caucasian Asian Black Hispanic

OSACA2

Japanese

Japanese

Tromsø

All Caucasian

Caucasian

Whitehall

Caucasian South Asian Black Other

Caucasian Asian Black Other

a

Ethnicity was not available on an individual level, but it was reasonably assumable that the vast majority of the individuals in the cohort have this ethnicity. Individuals coded with ‘other’ ethnicity were excluded from the analysis. a

73



PART ONE Ethnicity

Chapter 4 Ethnicity modifies associations between cardiovascular risk factors and disease severity in parallel Dutch and Singapore coronary cohorts Plos One. 2015 Jul 6;10(7):E0132278

Crystel M. Gijsberts, Aruni Seneviratna, Leonardo P. de Carvalho, Hester M. den Ruijter, Puwalani Vidanapthirana, Vitaly Sorokin, Pieter Stella, Pierfrancesco Agostoni, Folkert W. Asselbergs, A. Mark Richards, Adrian F. Low, Chi-Hang Lee, Huay Cheem Tan, Imo E. Hoefer, Gerard Pasterkamp, Dominique P.V. de Kleijn, Mark Y. Chan


Chapter 4

Abstract Background In 2020 the largest number of patients with coronary artery disease (CAD) will be found in Asia. Published epidemiological and clinical reports are overwhelmingly derived from western (White) cohorts and data from Asia are scant. We compared CAD severity and all-cause mortality among 4 of the world’s most populous ethnicities: Whites, Chinese, Indians and Malays. Methods The UNIted CORoNary cohort (UNICORN) simultaneously enrolled parallel populations of consecutive patients undergoing coronary angiography or intervention for suspected CAD in the Netherlands and Singapore. Using multivariable ordinal regression, we investigated the independent association of ethnicity with CAD severity and interactions between risk factors and ethnicity on CAD severity. Also, we compared all-cause mortality among the ethnic groups using multivariable Cox regression analysis. Results We included 1,759 White, 685 Chinese, 201 Indian and 224 Malay patients undergoing coronary angiography. We found distinct inter-ethnic differences in cardiovascular risk factors. Furthermore, the associations of gender and diabetes with severity of CAD were significantly stronger in Chinese than Whites. Chinese (OR 1.3 [1.1-1.7], p=0.008) and Malay (OR 1.9 [1.4-2.6], p<0.001) ethnicity were independently associated with more severe CAD as compared to White ethnicity. Strikingly, when stratified for diabetes status, we found a significant association of all three Asian ethnic groups as compared to White ethnicity with more severe CAD among diabetics, but not in non-diabetics. Crude allcause mortality did not differ, but when adjusted for covariates mortality was higher in Malays than the other ethnic groups. Conclusion In this population of individuals undergoing coronary angiography, ethnicity is independently associated with the severity of CAD and modifies the strength of association between certain risk factors and CAD severity. Furthermore, mortality differs among ethnic groups. Our data provide insight in inter-ethnic differences in CAD risk factors, CAD severity and mortality.

76


Ethnic Differences in Risk Factors and CAD Severity

Introduction Coronary artery disease (CAD) affects diverse populations and has become a leading global cause of morbidity and mortality.1 The World Health Organization (WHO) reported 17 million cardiovascular deaths (30.5% of all deaths) in the year 2008 and this number is expected to rise to 23.32-253 million by the year 2030. While numbers of cardiovascular deaths are stabilizing or even declining in the Western world, numbers are rapidly increasing in other parts of the world.4 This rise is most pronounced in Africa, Eastern Mediterranean regions and South East Asia; in those regions an increase of more than 10% by 2030 is predicted.3,5 By 2020, the highest numbers of cardiovascular deaths are expected in the Western Pacific region (~6 million) and in South-East Asia (~5 million)5, these regions are defined by the WHO definitions6. Therefore, the largest part of cardiovascular deaths will be among people of Asian ethnicity. Furthermore, the probability of dying prematurely between 30 and 70 years of age from non-communicable disease, of which 48% is cardiovascular disease, is already much higher in these regions (>30%) as compared to Western Europe or North America (<20%).3 CAD research has been conducted predominantly among Whites, while multi-ethnic research is strongly endorsed by the American Heart Association.7 Sizable cohort studies including Asian CAD patients are few and mostly conducted among Asian immigrants living in Western countries. These studies comparing Asians and Whites have shown some clinically important differences in risk factor burden8,9, incidence10 and prevalence of cardiovascular disease11, suggesting that ethnicity influences cardiovascular risk factor burden and prevalence of cardiovascular disease. Furthermore, studies comparing regions around the globe demonstrate small, yet significant differences in the relationship between cardiovascular risk factors and cardiovascular outcomes.12,13 To date, there are few data directly comparing CAD risk factors and outcomes in 4 of the world’s most populous ethnic groups: Whites, Chinese, Indian and Malay (www. census.gov/popclock). We sought to identify inter-ethnic differences in CAD risk factor burden, the severity of CAD and CAD outcomes, comparing patients living in their region of origin in countries with comparable health care systems.

Materials and Methods Ethics statement The Medical Ethics Committees of both participating hospitals (Netherlands: UMC Utrecht Medical Ethics Board, Reference number: 11-183; Singapore: Domain Specific Review Boards, Office of Human Research Protection Program, Reference Number: C/10/323) approved the study and written informed consent was obtained from all patients. This study conforms to the Declaration of Helsinki.

77


Chapter 4

Study design The UNICORN cohort (clinicaltrials.gov NCT02126150) is a multi-ethnic prospective cohort study including patients undergoing either diagnostic coronary angiography and/ or percutaneous coronary intervention (PCI). The UNICORN cohort has been conducted in parallel in two countries - the Netherlands and Singapore - adhering to matched protocols. The two hospitals (UMC Utrecht in the Netherlands and National University Hospital in Singapore) are tertiary referral centers with large annual coronary angiography/ PCI volumes of >3,000 and >1,500 patients, respectively. At these sites we enrolled Whites and the three largest Asian ethnic groups: Chinese, Indians and Malays. Study population Consecutive patients, ≼21 years of age, undergoing coronary angiography and/or PCI for (suspected) stable or acute coronary heart disease were eligible. At inclusion, patient demographics were documented. In Singapore, trained staff recorded self-reported ethnicity as documented on state-issued identification cards using one of the following categories: Chinese, Malay, Indian and other. All Dutch patients were of self-reported White/Western-European descent. Documentation captured cardiovascular risk factors (body mass index, hypertension, diabetes, dyslipidemia, smoking); medication use at admission (renin-angiotensinaldosterone system (RAAS) inhibiting medication (angiotensin converting enzyme inhibitors, angiotensin II antagonists and aldosterone receptor blockers), statins, betablockers and platelet directed therapy (aspirin, clopidogrel, prasugrel or ticagrelor)); cardiovascular medical history, indication for coronary angiogram, coronary angiogram result and the treatment strategy for CAD. Angiography Both centers used Siemens machines for coronary angiography. The Xcelera program (Philips Medical Systems) was used in both centers to store and view the recorded angiograms. CAD severity was determined by the number of major epicardial vessels (left anterior descending coronary artery, circumflex artery and right coronary artery) with a stenosis of >50% diameter loss14 by visual assessment of the interventional cardiologist. A significant stenosis in the left main coronary artery equated to two diseased epicardial vessels. For the current analysis CAD severity was categorized as follows: no/minor CAD, single vessel disease, double vessel disease and triple vessel disease and analyzed as an ordinal variable. Risk factors Diabetes was defined as any type of diabetes (fasting glucose > 7mmol/L)15 in the medical history or during index admission requiring medical treatment by means of oral glucose regulating medication or insulin injections (impaired glucose tolerance is not considered as diabetes in this study).

78


Ethnic Differences in Risk Factors and CAD Severity

Hypertension is considered when mentioned in the patient’s medical history or when diagnosed during the index admission (systolic blood pressure > 140 mmHg or diastolic blood pressure >90 mmHg) and/or the use of one or more antihypertensiva.16 Dyslipidemia was defined by any dyslipidemia requiring treatment in the medical history or during index admission as recommended by the ESC/EAS17 guidelines. Smoking status was divided into three groups; current smoker, quit smoker (> 1 year since last smoke) and non-smoker. Advanced renal failure was defined as any renal disease requiring treatment with oral medication or any type of renal replacement therapy in the medical history. All-cause mortality The vital status of the Singaporean patients was extracted from state mortality registration and matched with individual patient data. In the Netherlands, follow-up is performed through annual patient follow-up questionnaires. When the patient did not respond, the general practitioner was contacted to obtain the patient’s vital status, which was subsequently added to the hospital registration. An extraction of the hospital registration was used for the current analysis. Statistical analysis Statistical analyses were performed using the R software[18] package (version 3.0.2, Vienna, Austria). The level of statistical significance was set at α <0.05. Continuous variables (as they were normally distributed) were compared using ANOVA with Bonferroni post-hoc testing as to correct for multiple comparisons. Proportional differences were tested using a chi-square test. In order to correct for multiple testing when comparing all ethnic groups with each other, which leads to 6 tests, the level of significance was set at α 0.05/6= 0.008 according to the Bonferroni method. Measures of association between ethnicity and CAD severity were derived from multivariable ordinal logistic regression analyses including covariates found to be significantly associated with CAD severity (p<0.05) by univariate analyses. Covariates included age, gender, BMI, diabetes, hypertension, dyslipidemia, smoking, prior acute coronary syndrome (ACS), indication for coronary angiogram and use of anti-platelet medication, statins, beta-blocker and RAAS medication. The selected covariates were coerced in the model for all analyses (“enter model”). Stratified analyses for the independent effect of ethnicity were performed for sex, age group (≤ 62 years, >62 years, split at median age of 62 years) and diabetes status. We tested for interactions of risk factors with ethnicity for CAD severity by adding appropriate interaction terms to the full model. The significance level for interaction was conservatively set at α<0.05. Crude survival rates were compared by Kaplan Meier analysis with log-rank testing. Ethnicity-specific mortality rates corrected for covariates were derived from multivariable Cox regression. Patients with an “other” indication for angiography were removed from the multivariable survival analyses. Covariates were included when they were univariably significantly associated with all-cause mortality. These were: age, ethnicity, gender,

79


Chapter 4

indication for angiography, angiographic CAD severity, diabetes, dyslipidemia, previous ACS, statin use, platelet inhibitor use, beta blocker use and RAAS-inhibiting medication use. We compared multivariably corrected mortality rates to mortality in Malays who had the highest corrected mortality rate. The authors had direct access to the data and take responsibility for its integrity. All authors have read and agree to the manuscript as written.

Results UNICORN cohort Enrollment commenced in September 2010 in Singapore and October 2011 in the Netherlands. By March 2014 the Netherlands had recruited 1,759 White patients and Singapore 1,110 patients (62% Chinese, 20% Malay and 18% Indians). Patient characteristics by ethnicity are shown in Table 1. The proportion of males was highest in the Chinese and Indian ethnic groups (82.9% and 83.1%, respectively, p for overall difference <0.001). The Indian patients were youngest at 54.6 years compared with Malays 55.9, Chinese 58.6 and Whites who were markedly older at 64.9 years (p for overall difference <0.001). CAD risk factors Diabetes was significantly more common in Malays and Indians (52.2% and 51.5%) than in Chinese and Whites (33.2% and 20.3%). Chinese had significantly lower BMI than the other ethnic groups (mean 26.1 kg/m2 versus >27 kg/m2). The prevalence of hypertension did not differ among any of the ethnic groups. Dyslipidemia was strikingly more prevalent among the three Asian ethnic groups (Chinese: 70.5%, Indian: 77.5%, Malay: 75.4%) as compared with Whites (48.1%). Higher percentages of current smokers were seen in the Asian ethnic groups, highest at 46.4% among Indians compared with 24% of White patients. Cardiovascular medical history A history of previous ACS, coronary artery bypass grafting (CABG), cerebrovascular accident, transient ischemic attack or peripheral arterial disease (PAD) was less common among the Asian ethnic groups than in Whites. A history of previous ACS or PCI was significantly less common in Chinese as compared to other ethnic groups. Advanced renal failure was more common in Chinese as compared with Whites (p<0.001). CAD severity, presentation and treatment The prevalence of triple vessel disease was strikingly high in Malays (31.6%), followed by Chinese with a prevalence of 23.8% and 23.2% in Indians, as compared to Whites (14.0%, p<0.001). The distribution of the severity of CAD among the ethnic groups is depicted in Figure 1.

80


Ethnic Differences in Risk Factors and CAD Severity

Table 1. Demographic and Clinical Characteristics of the UNICORN cohorts by Ethnicity. Baseline characteristics of the UNICORN cohort are displayed per ethnic group. Figures represent means ± standard deviation (sd) or percentages. White

Chinese

Indian

Malay

1759

685

201

224

Males (%)

73.7

82.9*

83.1**

78.6

Age (years, mean ± sd)

64.9±10.9

58.6±9.8*

54.6±9.4**$$

55.9±9.0$&

BMI (kg/m2, mean ± sd)

27.0±4.5

26.1±4.7*

27.3±4.7$$

28.4±5.4$

Diabetes (%)

20.3

33.2

51.5

52.2$&

Hypertension (%)

58.4

64.2

61.8

62.1

Dyslipidemia (%)

47.7

70.5*

77.5**

75.4$

Non-smoker (%)

51.0

47.4*

42.2**

42.3$

Quit smoker (%)

27.3

18.4

11.4

18.1

Current smoker (%)

21.7

34.2

46.4

39.6

Anti platelet (%)

54.0

54.7

54.7

52.2

Statin (%)

62.2

58.4

61.2

54.5

Beta blocker (%)

57.2

42.2*

42.3**

38.8$

RAAS (%)

46.6

33.7*

41.8

42.9

Previous ACS (%)

29.3

16.1*

30.5$$

27.1&

Previous PCI (%)

27.9

17.4

29.9

26.7&

Previous CABG (%)

10.3

6.0

*

7.0

5.4

CVA/TIA (%)

9.8

5.9*

7.5

6.8

PAD (%)

11.0

2.9*

4.5**

5.4

Advanced renal failure (%)

2.1

6.0*

4.0

5.9$

Stable CAD (%)

53.6

54.7*

47.8**

42.4$&

UA/NSTEMI (%)

20.1

33.1

42.3

47.3

STEMI (%)

11.3

7.6

7.5

8.0

Other (%)

14.9

4.5

2.5

2.2

Conservative (%)

34.1

47.0*

49.3**

50.0$

PCI (%)

59.7

44.4

46.3

39.7

CABG (%)

6.2

8.6

4.5

10.3

Follow-up time (days)

410

894

830

807

All-cause deaths (n)

56

36

5

15

Two-year mortality rate (%)

4.6

3.7

2.6

5.6

N

Risk factors *

**$$

Smoking

Medication

Medical history *

$$

Indication

Treatment

Mortality

Abbreviations: BMI body mass index, RAAS renin-angiotensin-aldosterone system, ACS acute coronary syndrome, PCI percutaneous coronary intervention, CABG coronary artery bypass grafting, UA unstable angina, NSTEMI non-ST-elevated myocardial infarction, STEMI ST-elevated myocardial infarction. P-values for ethnic differences were calculated with a chi-square test for categorical data and one-way ANOVA for continuous, normally distributed data. The level of significance for the interethnic comparisons has been set conservatively at a p-value of 0.05/6=0.008 in order to correct for multiple testing. * White vs. Chinese p<0.008. ** White vs. Indian p<0.008. $ White vs. Malay p<0.008. $$ Chinese vs. Indian p<0.008. & Chinese vs. Malay p<0.008.

81


Chapter 4

Approximately 11% of the White patients presented to the angiography laboratory with an ST-elevation myocardial infarction (STEMI), a significantly higher percentage as compared to Chinese (7.6%), Indians (7.5%) and Malays (8.0%). The combined category of non-ST-elevation myocardial infarction (NSTEMI) or unstable angina (UA) was more common in Chinese 33.1%, Indians 42.3% and Malays 47.3% than in Whites (20.1%). Within the entire cohort, conservative treatment, i.e. without revascularization but encompassing risk factor control and/or anti-anginal medications, was more common in Chinese 47.0%, Indians 49.3% and Malays 50.0% than Whites 34.1%. PCI was undertaken less frequently in Chinese 44.4%, Indians 46.4% and Malays 39.7% as compared with Whites (59.7%). No inter-ethnic differences were apparent for treatment by CABG. Medication use Prescription of anti-platelet medication and statins was equal across the ethnic groups (about 55%). Beta-blocker and RAAS inhibition use were most common among Whites. Because Whites more often had a history of cardiovascular disease in the UNICORN cohort we performed a specific sub-analysis (S1 Table) of preventive drug use in participants with no history of cardiovascular disease (no acute coronary syndrome, previous PCI, CABG, cerebrovascular accidents, transient ischemic attacks or peripheral arterial disease). In this subgroup we observed that the use of anti-platelet medication and statins was equal among the ethnic groups. The use of beta-blockers and RAAS inhibiting drugs in this subgroup analysis paralleled that in the overall cohort: highest in Whites (p<0.001 for difference across all ethnic groups).

Percentage of individuals per ethnic group

30

20

10

0

White No or minor CAD

Chinese Single vessel disease

Indian Double vessel disease

Malay Triple vessel disease

Figure 1. Severity of CAD by ethnicity. Bar chart depicting the distribution of CAD severity as the percentage of the total number of individuals per ethnic group. Triple vessel disease is significantly more common among Chinese, Indians and Malays than among Whites (p <0.001).

82


Ethnic Differences in Risk Factors and CAD Severity

Age (per 10 years increase)

*

Diabetes (yes vs. no)

Dyslipidemia (yes vs. no)

*

Male gender (vs. female gender)

Hypertension (yes vs. no)

Smoking (current vs. non−smoker) 0 2 4 6 Odds ratios with 95% confidence intervals (derived from multivariable model) White

Chinese

Indian

8

Malay

Figure 2. Odds ratios of risk factors for the severity of CAD by ethnicity. Odds ratios derived from multivariable ordinal regression analysis, depicting the strength of association between cardiovascular risk factors and CAD severity (categorized into no CAD, single vessel disease, double vessel disease and triple vessel disease). The point estimates and 95% confidence intervals are shown for each ethnic group. A larger odds ratio indicates a stronger association between the risk factor and CAD severity. The asterisks (*) indicate significant interactions (p<0.05) of the risk factor as compared to Whites.

UNICORN patients without previous CAD events Because Whites more often had a history of CAD than the other ethnic groups, we performed a stratified analysis of inter-ethnic differences among patients with and without a history of prior CAD (no ACS, PCI or CABG). As shown in the S1 Table, we observed the same inter-ethnic distribution of risk factor burden in both strata, indicating that a past history of CAD did not modify the relationship between ethnicity and risk factor burden.

83


Chapter 4

The impact of ethnicity on the association of risk factors with CAD severity The odds ratios (ORs) of specific cardiovascular risk factors for CAD severity differed between ethnic groups, as can be observed in Figure 2. The odds ratios in Figure 2 depict the ethnicity-specific odds for a given change in risk factor (for example diabetes yes or no or and increase in age of 10 years) to move up one CAD severity class (for example from single to double vessel disease, or from double to triple vessel disease). Male gender was more strikingly associated with more severe CAD in Chinese (OR 7.0 [4.0-12.6]) than in Whites (OR 2.2 [1.7-2.7], p for interaction <0.001) and diabetes had a stronger association with more severe CAD in Chinese (OR 3.3 [2.2-5.0]) than in Whites (OR 1.4 [1.0-1.8], p for interaction 0.001). There were no significant interactions of ethnicity with age or smoking with respect to the severity of CAD. Possible interactions of ethnicity with the relationship of dyslipidemia and hypertension to CAD severity were not tested, as these risk factors were not significantly associated with CAD severity in the multivariable model.

Chinese ethnicity

Indian ethnicity

Malay ethnicity

Total cohort

|

Men

|

|

Women

|

Age <=62 years

Age >62 years

|

Diabetics

|

Non−diabetics −0.5

−0.4

−0.3

−0.2

OR

−0.1

0.0

| 0

2

OR

4

6

0

2

OR

4

6

0

2

OR

4

6

Figure 3. The adjusted odds ratios of Chinese, Indian and Malay ethnicity for the severity of CAD in subgroups of the UNICORN cohort. The adjusted association (odds ratios plus confidence intervals) of Chinese, Indian and Malay ethnicity as compared to White ethnicity for CAD severity, depicted for the total cohort and subgroups of the UNICORN cohort. The displayed odds ratios are derived from a multivariable model containing: age, gender, diabetes, hypertension, dyslipidemia, smoking, BMI, prior acute coronary syndrome, indication for coronary angiogram and use of anti-platelet medication, statins, beta-blocker and RAAS medication.

84


0.98 0.96 0.94 0.92

Caucasian Chinese Indian Malay

0.90

Suvival probability (Cox proportional hazard)

1.00

Ethnic Differences in Risk Factors and CAD Severity

0

100

200

300

400

500

600

700

800

900

FU time (days) Figure 4. Adjusted survival probability from multivariable Cox regression analysis by ethnicity. Survival probability derived from multivariable Cox regression analysis. Ethnicity-specific curves are adjusted for: age, gender, indication for angiography, conclusion from angiography, diabetes, dyslipidemia, previous ACS, statin use, platelet inhibitor use, beta blocker use and RAAS-inhibiting medication use. White, Chinese and Indian ethnicity were significantly associated with a better survival as compared to Malay ethnicity (Whites: HR 0.4 [0.2-0.8], p=0.009, Chinese: HR 0.5 [0.3-0.98], p=0.044, Indians HR 0.4 [0.1-0.98], p=0.046).

Independent association of ethnicity with CAD severity From a multivariable ordinal logistic regression model containing ethnicity as a covariate, we obtained ORs for Chinese, Indian and Malay ethnicity as compared to White ethnicity for the angiographic severity of CAD in the total cohort, and in specific subgroups. The results are displayed in Figure 3. Within the total cohort, ORs for Chinese and Malay ethnicity were significantly higher (1.4 [1.1-1.7] and 1.9 [1.4-2.6], respectively) using Whites as the reference group. Indicating Chinese and Malay but not Indian ethnicity, were independently associated with more severe CAD within the total cohort. This finding was largely driven by a striking interaction between ethnicity and diabetes with respect to severity of CAD. Among diabetics all Asian ethnicities were independently associated with more severe CAD as compared to White ethnicity whereas in non-diabetics this independent association of ethnicity with the severity of CAD was not observed (Figure 3). In a sex-specific analysis, results remained similar to the total cohort among men. However, among women, Chinese ethnicity tended to be

85


Chapter 4

associated with less severe CAD (OR 0.6 [0.3-1.1]) as compared to White ethnicity, although not reaching statistical significance. Indian and Malay ethnicity were not significantly associated with CAD severity in women. The female subgroup, however, is small and power is markedly reduced in these analyses. This is especially the case for the Indian and Malay female subgroups, consisting of 34 and 48 women, respectively. Hence, the chance of a type II error is larger. The ORs of severe CAD in Chinese and Indian ethnicity were comparable to Whites in those 62 years (median age) or younger compared with those over 62. Ethnicity and All-cause Mortality Crude all-cause mortality rates (from Kaplan Meier analysis) showed lowest survival probability among Whites, and highest in Indians (Table 1, log-rank test for difference across the ethnic groups p=0.17). On correction for covariates by Cox regression analysis to assess the independent effect of ethnicity on all-cause mortality, survival in Malays fell below that in Chinese, Indians and Whites (Figure 4). White, Chinese and Indian ethnicity were significantly associated with a better survival as compared to Malay ethnicity (Whites: HR 0.4 [0.2-0.8], p=0.009, Chinese: HR 0.5 [0.3-0.98], p=0.044, Indians HR 0.4 [0.1-0.98], p=0.046).

Discussion In the UNICORN study, we compared four of the most populous ethnic groups in the world, living in two countries that are comparable in terms of development: Singapore and the Netherlands. Both countries are ranked within the top 20 on the human development index19 and have comparable health care systems.20 We defined ethnic differences in cardiovascular risk factors, the severity of CAD and allcause mortality in patients undergoing coronary angiography. Chinese and Malay ethnicity were independently associated with more severe CAD compared to White ethnicity. This finding was largely driven by a striking interaction between ethnicity and diabetes with respect to CAD severity. Ethnicity also interacted with male gender, modifying its association with CAD severity. Mortality was highest among Malays, this difference in all-cause mortality after coronary angiography persisted after adjustment for baseline differences. UNICORN characteristics Our results show clear ethnic differences in age at presentation. Previous studies have shown that Indians (South Asians) tend to incur CAD at a younger age, indicating a higher atherosclerotic burden earlier in life.10,21–23 With respect to other Asian ethnic groups; in the e-HEALING24 coronary stent registry of Asians from Singapore, Hong Kong and Malaysia, the mean age of Whites (from Western Europe) was 65.9, whilst the mean age of Asian registrants was 57.4 years, very much in line with our cohort. These prior reports

86


Ethnic Differences in Risk Factors and CAD Severity

highlighted key differences between White and Asian patients with CAD but Asians were typically classified as a single ethnic group. In the current study, we clearly demonstrate that significant and clinically relevant differences in age at presentation also exist among the Asian ethnic groups. Our data underscore the importance of delineating specific Asian ethnicities to avoid missing key inter-ethnic differences. Risk factor burden The existing literature, mainly derived from populations living within western communities, has mainly focused on the risk factor burden of South Asians (often residents of the UK25 or US9) as compared to Whites, and shows a higher burden in South Asians, which we also observe. Chinese have been found to display a slightly more benign risk factor pattern compared to Whites in population-based studies.26,27 However, in Chinese individuals with overt cardiovascular disease, a higher risk factor burden has been reported, corresponding to our findings.28,29 Risk factors in Malays as compared to Whites have been less well documented in the literature, but striking differences were encountered in this study. CAD severity In our cohort we find a higher prevalence of angiographic triple vessel disease in Chinese, Indians and Malays as compared to Whites. However, differences remained after adjustment for baseline differences in risk factors. This indicates that the differences in risk factor burden only partly explain the more severe CAD phenotype that is observed in Chinese, Indians and Malays. Apparently, ethnicity conveys an important independent (biological or life-style mediated) component that is not fully captured by the general patient characteristics or risk factor burden and warrants more detailed research. Importantly, these differences appeared to be largely driven by diabetes with risk of severe CAD clearly enhanced compared with non-diabetics to a greater degree in all three Asian ethnicities than the additional risk conferred by diabetes in Whites. Besides the independent association of ethnicity with CAD severity, we also find significant interactions of ethnicity with cardiovascular risk factors. It thus appears that ethnicity also modifies the effect of certain risk factors on CAD severity. This modifying effect has been rarely studied, although stronger associations of cholesterol levels and diabetes with carotid intima-media thickness have been reported by Chow et al.30 Their findings might indicate that the vascular wall of South Asians is more susceptible to glycemic and lipidemic disturbances than of Whites. The mechanisms underlying both the independent impact of ethnicity, as well as the modifying effect of ethnicity remain to be elucidated. Mortality South Asians have a higher incidence of CAD, but lower29,31,32 or comparable33,34 (coronary) mortality rates as compared to Whites. In Chinese survival similar28 or better29 than in Whites has been reported. Our results largely concur with existing literature, showing

87


Chapter 4

similar corrected survival probabilities for Indians, Chinese and Whites. However, to our knowledge, for Malays no comparison with Whites has been previously published. Although, a comparison among the Asian ethnic groups in Singapore, showed higher all-cause mortality in Malays as compared to Chinese and Indians among myocardial infarction patients.35 Accordingly, we found survival in Malays to be lower than in the other ethnic groups in both crude and corrected analyses, indicating that the high burden of risk factors and the more severe CAD are accompanied by higher mortality rates in Malays. Implications and future directions The most striking interaction we observed between ethnicity and risk factors on CAD severity was observed for diabetes. Specific focus might be granted to stricter glycemic control among Chinese in whom diabetes has the biggest impact on CAD severity. Earlier CAD screening might be appropriate among Chinese and Malay men, as Chinese and Malay ethnicity are independent predictors of more severe CAD in men, but not in women. Malays, with the heaviest burden of risk factors suffered the poorest survival. The proportion of conservative treatment was highest and use of preventive medications lowest among Malays whilst they carried the most severe CAD. More vigilant surveillance and more aggressive pharmacological and interventional treatment of CAD might be explored in this ethnic group. In order to elucidate inter-ethnic differences studies assessing dietary and life-style habits, as well as multi-ethnic biobanking initiatives offer valuable perspectives.36 Cultural and socioeconomic factors might influence risk of CAD and CAD severity among the ethnic groups. A careful approach in unraveling ethnicity-dependent patterns of diet, physical activity and other life-style habits might elucidate a substantial part of the risk that is conveyed by ethnicity. Although not available in our dataset, income levels are known to differ among the four examined countries and ethnic groups, being higher in Singapore than in the Netherlands: with a per capita GDP of US$ 55,182 and US$ 50,793 in the Netherlands (data for 2010-2014 from data.worldbank.org). Within Singapore the median monthly income differs markedly among the ethnic groups: S$5,100 for Chinese, S$3,844 for Malays and S$5,370 for Indians.37 The median monthly income in the Netherlands is estimated at â‚Ź2,39138 roughly corresponding to S$3571. While substantial differences are visible in the income levels among the ethnic groups, it might not necessarily mean that income affects CAD risk and severity in similar ways among the ethic groups. Per ethnic group, physical activity, dietary and health care utilizing behavior of poorer or richer individuals might be different. These habits must be elucidated in an ethnicity-specific manner in order to indicate possible targets for cardiovascular health improvement, which might best be done through qualitative research. Additionally, biomarkers related to CAD have been found to differ between Whites and certain Asian ethnic groups on a general population level.39 Biomarker differences, yet to be revealed among the ethnic groups, might guide us to biological pathways involved in the accelerated and more severe CAD observed in Indians and Malays.

88


Ethnic Differences in Risk Factors and CAD Severity

Limitations The largest limitation to our study is that we were unable to correct for dietary, lifestyle and socioeconomic factors40 as possible confounders. It is possible that these factors would influence our results, however we expect that by correcting for many other CAD risk factors the effect of life style and socioeconomic status has been covered for a large part.41 Although Singapore and the Netherlands are fairly comparable19,20 when it comes to development status and health care systems, we could not correct for possible differences in these factors. It is possible that differences in, for example, the referral habits of primary care and secondary care physicians between Singapore and the Netherlands have an impact on the ethnic differences we observe. However, if this would be the case one would not expect many differences within the Asian ethnic groups or between the sexes (within one ethnic group), which we do find. Also, clinical decision-making and patient preferences for invasive coronary investigations might differ between the countries and among the ethnic groups.42 Our results are exclusively applicable to patients who have undergone coronary angiography in a tertiary care center: the most symptomatic group of CAD patients. Caution should be exercised when extrapolating our results to the entire CAD population. Additional studies are needed to examine to which extent our results are applicable to other CAD populations. Due to a limitation in statistical power no sex-specific or other stratified analyses could be performed on the mortality data. Conclusion Striking differences were found between Whites, Chinese, Indians and Malays undergoing coronary angiography for suspected CAD. Ethnicity is independently associated with the severity of CAD and all-cause mortality after coronary angiography. Notably, Chinese ethnicity alters the strength of association between established cardiovascular risk factors (diabetes and male gender) and CAD severity. Our data gain insight in the characteristics of coronary artery disease among the ethnic groups and suggest that CAD management should account for the differential impact of risk factors on CAD severity among different ethnicities. Similar studies in the other cardiovascular disease populations would be useful. Acknowledgements We thank Ms. Jonne Hos for her excellent contribution to the organization of the Dutch part of the UNICORN cohort. From the Singapore site we thank Mrs. Fauziah Azizi for her outstanding work on the gathering of data.

89


Chapter 4

References 1.

Lloyd-Jones D, Adams RJ, Brown TM, Carnethon M, Dai S, De Simone G, Ferguson TB, Ford E, Furie K, Gillespie C, Go A, Greenlund K, Haase N, Hailpern S, Ho PM, Howard V, Kissela B, Kittner S, Lackland D, Lisabeth L, Marelli A, McDermott MM, Meigs J, Mozaffarian D, Mussolino M, Nichol G, Roger VL, Rosamond W, Sacco R, Sorlie P, Stafford R, Thom T, Wasserthiel-Smoller S, Wong ND, Wylie-Rosett J. Heart disease and stroke statistics--2010 update: a report from the American Heart Association. Circulation. 2010;121:e46–e215.

2.

Mathers CD, Loncar D. Projections of global mortality and burden of disease from 2002 to 2030. PLoS Med. 2006;3:e442.

3.

World Health Organization. World Health Statistics 2012. Geneva, Switzerland: 2012.

4.

Ford ES, Ajani U a, Croft JB, Critchley J a, Labarthe DR, Kottke TE, Giles WH, Capewell S. Explaining the decrease in U.S. deaths from coronary disease, 1980-2000. N Engl J Med. 2007;356:2388–98.

5.

Alwan A. Global status report on noncommunicable diseases 2010. 1st ed. Geneva, Switzerland: 2011.

6.

WHO. World Health Statistics 2011. WHO Libr Cat Data. 2011;1:170.

7.

Goff DC, Lloyd-Jones DM, Bennett G, Coady S, D’Agostino RB, Gibbons R, Greenland P, Lackland DT, Levy D, O’Donnell CJ, Robinson JG, Schwartz JS, Shero ST, Smith SC, Sorlie P, Stone NJ, Wilson PWF, Jordan HS, Nevo L, Wnek J, Anderson JL, Halperin JL, Albert NM, Bozkurt B, Brindis RG, Curtis LH, DeMets D, Hochman JS, Kovacs RJ, Ohman EM, Pressler SJ, Sellke FW, Shen W-K, Tomaselli GF. 2013 ACC/AHA guideline on the assessment of cardiovascular risk: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines. Circulation. 2014;129:S49–73.

8.

Anand SS, Yusuf S, Vuksan V, Devanesen S, Teo KK, Montague PA, Kelemen L, Yi C, Lonn E, Gerstein H, Hegele RA, McQueen M. Differences in risk factors, atherosclerosis, and cardiovascular disease between ethnic groups in Canada: the Study of Health Assessment and Risk in Ethnic groups (SHARE). Lancet. 2000;356:279–84.

9.

Meadows T a., Bhatt DL, Cannon CP, Gersh BJ, Röther J, Goto S, Liau C-S, Wilson PWF, Salette G, Smith SC, Steg PG. Ethnic Differences in Cardiovascular Risks and Mortality in Atherothrombotic Disease: Insights From the REduction of Atherothrombosis for Continued Health (REACH) Registry. Mayo Clin Proc. 2011;86:960–967.

10. Joshi P, Islam S, Pais P, Reddy S, Dorairaj P, Kazmi K, Pandey MR, Haque S, Mendis S, Rangarajan S, Yusuf S. Risk factors for early myocardial infarction in South Asians compared with individuals in other countries. JAMA. 2007;297:286–94. 11. Bild DE, Detrano R, Peterson D, Guerci A, Liu K, Shahar E, Ouyang P, Jackson S, Saad MF. Ethnic differences in coronary calcification: the Multi-Ethnic Study of Atherosclerosis (MESA). Circulation. 2005;111:1313–20. 12. Woodward M, Huxley H, Lam TH, Barzi F, Lawes CMM, Ueshima H. A comparison of the associations between risk factors and cardiovascular disease in Asia and Australasia. Eur J Cardiovasc Prev Rehabil. 2005;12:484–91. 13. Yusuf S, Hawken S, Ounpuu S, Dans T, Avezum A, Lanas F, McQueen M, Budaj A, Pais P, Varigos J, Lisheng L. Effect of potentially modifiable risk factors associated with myocardial infarction in 52 countries (the INTERHEART study): case-control study. Lancet. 2004;364:937–52. 14. Harris PJ, Behar VS, Conley MJ, Harrell FE, Lee KL, Peter RH, Kong Y, Rosati R a. The prognostic significance of 50% coronary stenosis in medically treated patients with coronary artery disease. Circulation. 1980;62:240–248. 15. World Health Organization. Definition and diagnosis of diabetes mellitus and intermediate hyperglycaemia. Geneva, Switzerland: 2006. 16. McManus RJ, Caulfield M, Williams B. NICE hypertension guideline 2011: evidence based evolution. BMJ. 2012;344:e181.

90


Ethnic Differences in Risk Factors and CAD Severity

17. Reiner Z, Catapano AL, De Backer G, Graham I, Taskinen M-R, Wiklund O, Agewall S, Alegria E, Chapman MJ, Durrington P, Erdine S, Halcox J, Hobbs R, Kjekshus J, Filardi PP, Riccardi G, Storey RF, Wood D. ESC/EAS Guidelines for the management of dyslipidaemias: the Task Force for the management of dyslipidaemias of the European Society of Cardiology (ESC) and the European Atherosclerosis Society (EAS). Eur Heart J. 2011;32:1769–818. 18. R Core Team. R: A Language and Environment for Statistical Computing. 2013; 19. United Nations Development Programme. Human development report 2013. New York, USA: 2013. 20. World Health Organization. Health systems: Improving performance. In: The world health report 2000. Geneva, Switzerland: 2000. 21. Silbiger JJ, Stein R, Roy M, Nair MK, Cohen P, Shaffer J, Pinkhasov A, Kamran M. Coronary artery disease in South Asian immigrants living in New York City: angiographic findings and risk factor burdens. Ethn Dis. 2013;23:292–5. 22. Zheng Y, Ma W, Zeng Y, Liu J, Ye S, Chen S, Lan L, Erbel R, Liu Q. Comparative study of clinical characteristics between Chinese Han and German Caucasian patients with coronary heart disease. Clin Res Cardiol. 2010;99:45–50. 23. Jones D, Rathod KS, Sekhri N, Junghans C, Gallagher S, Rothman MT, Mohiddin S, Kapur A, Knight C, Archbold A, Jain K, Mills PG, Uppal R, Mathur A, Timmis D, Wragg A. Case fatality rates for South Asian and Caucasian patients show no difference 2.5 years after percutaneous coronary intervention. Heart. 2012;98:414–9. 24. Klomp M, Damman P, Beijk M a M, Tan KH, Balian V, de Luca G, Tijssen JGP, Silber S, de Winter RJ. Differences in cardiovascular risk factors and clinical outcomes between Western European and Southeast Asian patients treated with the Genous Bio-engineered R stent: an e-HEALING worldwide registry substudy. Coron Artery Dis. 2012;23:271–7. 25. Bellary S, O’Hare JP, Raymond NT, Mughal S, Hanif WM, Jones A, Kumar S, Barnett AH. Premature cardiovascular events and mortality in south Asians with type 2 diabetes in the United Kingdom Asian Diabetes Study - effect of ethnicity on risk. Curr Med Res Opin. 2010;26:1873–9. 26. Chiu M, Austin PC, Manuel DG, Tu J V. Comparison of cardiovascular risk profiles among ethnic groups using population health surveys between 1996 and 2007. CMAJ. 2010;182:E301–10. 27. Zhang X, Patel A, Horibe H, Wu Z, Barzi F, Rodgers A, MacMahon S, Woodward M. Cholesterol, coronary heart disease, and stroke in the Asia Pacific region. Int J Epidemiol. 2003;32:563–72. 28. Khan N, Grubisic M, Hemmelgarn B, Humphries K, King KM, Quan H. Outcomes after acute myocardial infarction in South Asian, Chinese, and white patients. Circulation. 2010;122:1570–1577. 29. Quan H, Khan N, Li B, Humphries KH, Faris P, Diane Galbraith P, Graham M, Knudtson ML, Ghali WA. Invasive cardiac procedure use and mortality among South Asian and Chinese Canadians with coronary artery disease. Can J Cardiol. 2010;26:e236–e242. 30. Chow CK, McQuillan B, Raju PK, Iyengar S, Raju R, Harmer JA, Neal BC, Celermajer DS. Greater adverse effects of cholesterol and diabetes on carotid intima-media thickness in South Asian Indians: comparison of risk factor-IMT associations in two population-based surveys. Atherosclerosis. 2008;199:116–22. 31. Bansal N, Fischbacher CM, Bhopal RS, Brown H, Steiner MF, Capewell S. Myocardial infarction incidence and survival by ethnic group: Scottish Health and Ethnicity Linkage retrospective cohort study. BMJ Open. 2013;3:e003415. 32. Zaman MJS, Philipson P, Chen R, Farag A, Shipley M, Marmot MG, Timmis AD, Hemingway H. South Asians and coronary disease: is there discordance between effects on incidence and prognosis? Heart. 2013;99:729–36.

91


Chapter 4

33. Jones DA, Gallagher S, Rathod KS, Redwood S, de Belder MA, Mathur A, Timmis AD, Ludman PF, Townend JN, Wragg A. Mortality in South Asians and Caucasians After Percutaneous Coronary Intervention in the United Kingdom: An Observational Cohort Study of 279,256 Patients From the BCIS (British Cardiovascular Intervention Society) National Database. JACC Cardiovasc Interv. 2014;7:362–71. 34. Toor IS, Jaumdally R, Lip GYH, Pagano D, Dimitri W, Millane T, Varma C. Differences between South Asians and White Europeans in five year outcome following percutaneous coronary intervention. Int J Clin Pract. 2011;65:1259–66. 35. Mak K-H, Chia K-S, Kark JD, Chua T, Tan C, Foong B-H, Lim Y-L, Chew S-K. Ethnic differences in acute myocardial infarction in Singapore. Eur Heart J. 2003;24:151–60. 36. Wang JW, Gijsberts CM, Seneviratna A, de Hoog VC, Vrijenhoek JEP, Schoneveld a H, Chan MY, Lam CSP, Richards a M, Lee CN, Mosterd A, Sze SK, Timmers L, Lim SK, Pasterkamp G, de Kleijn DP V. Plasma extracellular vesicle protein content for diagnosis and prognosis of global cardiovascular disease. Neth Heart J. 2013;21:467–71. 37. Kim W. Census of population 2010 Statistical Release 2: Households and Housing. Department of Statistics, Ministry of Trade & Industry, Republic of Singapore; 2010. 38. CBS. Inkomen van particuliere huishoudens met inkomen naar kenmerken en regio [Internet]. 2014;Available from: http://statline.cbs.nl/StatWeb/publication/?VW=T&DM=SLNL&PA=80594NED&LA=NL 39. Gijsberts CM, den Ruijter HM, Asselbergs FW, Chan MY, de Kleijn DP V, Hoefer IE. Biomarkers of Coronary Artery Disease Differ Between Asians and Caucasians in the General Population. Glob Heart. 2015; 40. Nakamura Y, Moss AJ, Brown MW, Kinoshita M, Kawai C. Ethnicity and long-term outcome after an acute coronary event. Multicenter Myocardial Ischemia Research Group. Am Heart J. 1999;138:500–506. 41. Clark AM, DesMeules M, Luo W, Duncan AS, Wielgosz A. Socioeconomic status and cardiovascular disease: risks and implications for care. Nat Rev Cardiol. 2009;6:712–22. 42. Kressin NR, Petersen LA. Racial differences in the use of invasive cardiovascular procedures: Review of the literature and prescription for future research. Ann. Intern. Med. 2001;135:352–366.

92


Ethnic Differences in Risk Factors and CAD Severity

Supplementals Supplemental table. Baseline characteristics of the UNICORN participants with no history of CAD events. White

Chinese

Indian

Malay

P-value difference

N

914

516

121

142

Males (%)

69.1

81.6

78.5

76.1

<0.001

63.5 (11.4)

57.9 (9.8)

53.0 (9.5)

54.4 (8.7)

<0.001

Age (years, mean (sd)) Risk factors BMI (kg/m2, mean (sd))

26.6 (4.5)

26.1 (4.6)

27.7 (4.9)

28.7 (5.7)

<0.001

Diabetes (%)

15.6

28.9

48.3

48.6

<0.001

Hypertension (%)

53.6

61.2

53.8

57

0.048

Dyslipidemia (%)

38.0

67.0

67.5

68.3

<0.001

Current smoker (%)

52.6

50.1

43.6

47.3

<0.001

Quit smoker (%)

24.1

14.3

8.9

15.5

Non-smoker (%)

23.3

35.5

47.5

37.3

Smoking

Medication Anti platelet (%)

43.7

43.6

34.7

33.8

Statin (%)

46.2

48.8

43.8

40.1

0.044 0.283

Beta blocker (%)

42.9

30.0

22.3

26.8

<0.001

RAAS (%)

39.9

24.6

25.6

30.3

<0.001

Medical history CVA/TIA (%)

8.6

5.0

5.8

5.6

0.066

PAD (%)

8.0

2.3

1.7

3.5

<0.001

Advanced renal failure (%)

1.5

3.7

4.1

4.2

0.028

Stable (%)

44.2

55.2

51.2

41.5

<0.001

UA/NSTEMI (%)

21.9

30.2

33.9

45.1

STEMI (%)

17.4

9.3

11.6

10.6

Other (%)

16.5

5.2

3.3

2.8

30.8

32.6

29.8

27.5

1-Vessel Disease (%)

30.1

23.6

28.1

17.6

2-Vessel Disease (%)

24.0

23.1

24.8

24.6

3-Vessel Disease (%)

15.1

20.7

17.4

30.3

Indication

Angiographic finding No CAD (%)

<0.001

Treatment Conservative (%)

35.2

46.9

50.4

51.4

PCI (%)

58.0

44.0

45.5

38

6.9

9.1

4.1

10.6

CABG (%)

<0.001

Baseline characteristics of the UNICORN participants with no history of CAD events with p-values for the differences among the ethnic groups (derived from chi-square and ANOVA tests).

93



PART ONE Ethnicity

Chapter 5 Inter-Ethnic Differences in Quantified Coronary Artery Disease Severity and All-Cause Mortality among Dutch and Singaporean Percutaneous Coronary Intervention Patients PLoS One. 2015 Jul 6;10(7):e0131977

Crystel M. Gijsberts, Aruni Seneviratna, Imo E. Hoefer, Pierfrancesco Agostoni, Saskia Z.H. Rittersma, Gerard Pasterkamp, Mikael Hartman, Leonardo Pinto de Carvalho, A. Mark Richards, Folkert W. Asselbergs, Dominique P.V. de Kleijn, Mark Y. Chan


Chapter 5

Abstract Background Coronary artery disease (CAD) is a global problem with increasing incidence in Asia. Prior studies reported inter-ethnic differences in the prevalence of CAD rather than the severity of CAD. The angiographic “synergy between percutaneous coronary intervention (PCI) with Taxus and cardiac surgery� (SYNTAX) score quantifies CAD severity and predicts outcomes. We studied CAD severity and all-cause mortality in four globally populous ethnic groups: Caucasians, Chinese, Indians and Malays. Methods We quantified SYNTAX scores of 1,000 multi-ethnic patients undergoing PCI in two tertiary hospitals in the Netherlands (Caucasians) and Singapore (Chinese, Indians and Malays). Within each ethnicity we studied 150 patients with stable CAD and 100 with ST-elevated myocardial infarction (STEMI). We made inter-ethnic comparisons of SYNTAX scores and all-cause mortality. Results Despite having a younger age (mean age Indians: 56.8 and Malays: 57.7 vs. Caucasians: 63.7 years), multivariable adjusted SYNTAX scores were significantly higher in Indians and Malays than Caucasians with stable CAD: 13.4 [11.9-14.9] and 13.4 [12.0-14.8] vs. 9.4 [8.110.8], p<0.001. Among STEMI patients, SYNTAX scores were highest in Chinese and Malays: 17.7 [15.9-19.5] and 18.8 [17.1-20.6] vs. 15.5 [13.5-17.4] and 12.7 [10.9-14.6] in Indians and Caucasians, p<0.001. Over a median follow-up of 709 days, 67 deaths (stable CAD: 37, STEMI: 30) occurred. Among STEMI patients, the SYNTAX score independently predicted all-cause mortality: HR 2.5 [1.7-3.8], p<0.001 for every 10-point increase. All-cause mortality was higher in Indian and Malay STEMI patients than Caucasians, independent of SYNTAX score (adjusted HR 7.2 [1.5-34.7], p=0.01 and 5.8 [1.2-27.2], p=0.02). Conclusion Among stable CAD and STEMI patients requiring PCI, CAD is more severe in Indians and Malays than in Caucasians, despite having a younger age. Moreover, Indian and Malay STEMI patients had a greater adjusted risk of all-cause mortality than Caucasians, independent of SYNTAX score.

96


Ethnic Differences in Quantified CAD Severity and Mortality

Background Inter-ethnic differences in the prevalence of coronary artery disease (CAD) and cardiovascular risk factors such as diabetes1 and dyslipidemia2 are known. People of Indian (or South Asian) descent have been reported to have an unfavorable risk factor profile (e.g. higher prevalence of diabetes and dyslipidemia3,4) and a higher prevalence of CAD (as reported by the World Health Organization5) compared with Caucasians. Individuals of Chinese descent, on the other hand, have been reported to have a more favorable risk factor profile (e.g. low C-reactive protein levels2 and low insulin levels6) and lower prevalence of CAD (as assessed by coronary artery calcium (CAC) scoring).7 The World Health Organization has projected that the majority of the global population of patients with CAD will be of Asian descent by 2030.5 Yet, data on differences in the CAD burden among the individual Asian ethnic groups are sparse and predominantly based on Western (European) literature on Asian immigrants.8 As such, the American Heart Association has assigned a high priority to multi-ethnic research on the burden and outcomes of CAD.9 Studies assessing CAC scores have shown that CAC scores are higher among communitydwelling individuals of Indian descent as compared with those of Chinese descent.6,10,11 But, despite its sensitivity in detecting CAD, CAC scoring remains a screening tool that has limited specificity for the presence of underlying CAD. Coronary angiography remains the gold standard for assessing the presence and severity of CAD. Angiographic studies quantifying the severity of CAD are sparse; one study compared mainland Chinese with Australian Caucasians, showing less severe CAD in Chinese than in Caucasian coronary angiography patients as quantified by the Gensini score.12 In the context of significant multi-vessel CAD the angiographic synergy between percutaneous coronary intervention (PCI) with Taxus and cardiac surgery (SYNTAX) score has been developed.13 This score quantifies the anatomic extent and complexity of CAD over 16 anatomically defined coronary segments on coronary angiography. The SYNTAX score has been validated for predicting outcomes of patients undergoing PCI.14 Based on the available literature on inter-ethnic differences in risk factor burden and CAD prevalence, we hypothesized that the severity of angiographic CAD, as quantitatively measured by SYNTAX score, differs among Caucasians, Chinese, Indians and Malays, who constitute four of the largest ethnic groups in the world15. For this purpose we investigated PCI patients from two tertiary hospitals: the University Medical Center Utrecht, the Netherlands (enrolling Caucasian patients) and the National University Hospital, Singapore (enrolling Chinese, Indian and Malay patients). In two wellcircumscribed cardiologic patient groups: stable CAD and STEMI patients undergoing PCI, we investigated inter-ethnic differences in the severity of angiographic CAD by means of the SYNTAX score. Furthermore, we evaluated inter-ethnic differences in allcause mortality, adjusted for SYNTAX score.

97


Chapter 5

Methods Study population Patients were retrospectively, consecutively selected from the coronary angiography databases of two hospitals: the University Medical Center Utrecht (UMCU) in the Netherlands and the National University Hospital (NUH) in Singapore between December 2007 and October 2013. From the two sites, four ethnic groups were included in this study: Caucasians from the UMCU and Chinese, Indians and Malays from the NUH. From a power calculation on a preliminary cohort of 20 stable CAD and 20 STEMI patients per ethnic group we concluded that 150 stable CAD and 100 STEMI patients per ethnic group were needed for the current study in order to give us 80% power to detect a difference in SYNTAX score of 2.4 - 5.7 points (SD 1.5-1.6) between ethnic groups, considering a Bonferroni post-hoc Îą of 0.0125. For the current study, we selected 150 patients from each ethnic group who were diagnosed with stable CAD, defined as stable angina or angina equivalent, exertional dyspnea relieved either by rest or by nitroglycerin or silent myocardial ischemia.16 Only patients with stable CAD requiring PCI as a treatment16 were selected for the current study. Additionally, we selected 100 patients from each ethnic group with a clinical diagnosis of ST-elevated myocardial infarction (STEMI)17 and who underwent primary PCI. We selected patients from October 2013 backwards and went as far back as necessary (December 2007), to allow us to obtain the desired number of patients. Patients with a history of coronary artery bypass grafting (CABG) surgery were excluded, as the regular SYNTAX score13 which was used in this study is only applicable to the native coronary system. Ethics statement The institutional review boards of both participating hospitals exempted this retrospective database study from approval. The exempts are registered at the NUH (Domain Specific Review Board, DSRB) under: 2012/00971 and at the UMCU (Medical Ethics Committee, METC) under: 13/222. This study conforms to the declaration of Helsinki. No informed consent was necessary as data were analyzed anonymously. Ethnicity documentation In Singapore, trained staff recorded self-reported ethnicity as documented on stateissued identification cards using one of the following categories: Chinese, Malay, Indian and other. All Dutch patients were assumed to be Caucasian for this study (from a questionnaire among a group of 1,429 angiography study patients >94% were confirmed as Caucasian and only 2% were of Asian descent). Risk factors Diabetes was defined as any type of diabetes (fasting glucose >7mmol/L)18 in the medical history or during index admission requiring medical treatment by means of oral glucose

98


Ethnic Differences in Quantified CAD Severity and Mortality

regulating medication or insulin injections (impaired glucose tolerance is not considered to be diabetes in this study). Hypertension is considered when mentioned in the patient’s medical history or when diagnosed during the index admission (systolic blood pressure >140 mmHg or diastolic blood pressure >90 mmHg) and/or the use of one or more antihypertensiva.19 Dyslipidemia was defined by any dyslipidemia requiring treatment in the medical history or during index admission as recommended by the ESC/EAS20 guidelines. Smoking status was divided into three groups; current smoker, ex smoker (> 1 year since last smoke) and non-smoker. Advanced renal failure was defined as any renal disease requiring treatment with oral medication (phosphate-binding medications) or any type of renal replacement therapy in the medical history.21 Body mass index (BMI) was calculated by dividing weight (in kilograms) by squared height (in meters). Medication use The use of cardiovascular drugs known to lower cardiovascular risk22,23: anti-platelet medication, statins, beta-blockers and renin-angiotensin-aldosterone system (RAAS) inhibitors was assessed at the moment of admission for PCI. Anti-platelet medication comprises aspirin and all types of P2Y12 inhibitors. RAAS inhibiting medication comprises all types of angiotensin-converting enzyme (ACE) inhibitors, angiotensin receptor blockers and aldosterone antagonists. SYNTAX scoring The previously validated14,24 SYNTAX score considers stenotic lesions reducing the luminal diameter >50% in vessels of >1.5mm.13 In the total SYNTAX score, lesions are weighted depending on the anatomical position of the segment in which they occur. The more proximal in the coronary tree, the more points are assigned to a lesion. A SYNTAX score of >18 points corresponds to severe CAD and has been reported to be related to higher rates of adverse cardiac events.24 The SYNTAX score was measured by two independent observers (CG and AS), using SYNTAX score calculator25 version 2.11. The observers were blinded to ethnicity and other patient characteristics. The two observers employed quantitative coronary angiography26 (QCA) software (CAAS, Siemens) to measure the percentage of stenosis or the dimension of the vessel whenever they were uncertain about the angiographic significance of a lesion by visual estimation. QCA was performed for 95 cases in total (CG 67 cases, AS 47 cases). When the two observers were more than 5 SYNTAX points apart (60 cases), the case was discussed in order to reach consensus and QCA was performed to assess lesion significance. The average of the SYNTAX scores of each patient measured by both observers was used as a continuous outcome measure for the current analysis. All-cause mortality The vital status (alive or deceased) of the Singaporean patients was extracted from state mortality registration and matched with individual patient data. In the Netherlands,

99


Chapter 5

follow-up was performed through annual patient follow-up questionnaires. When the patient did not respond, the general practitioner was contacted to obtain the patient’s vital status, which was subsequently added to the hospital registration. An extraction of the completed hospital registration was used for the current analysis. Statistical analysis Continuous baseline variables are displayed as means with standard deviations (or confidence intervals for SYNTAX scores), and categorical variables are presented as percentages. The baseline characteristics were compared among the ethnic groups using ANOVA for continuous and chi-square testing for categorical data, respectively. Baseline characteristics that differed significantly across the four ethnic groups or that were significantly associated with SYNTAX score on univariable analysis, were added to the multivariable model. Covariates in the multivariable model included: age, BMI, diabetes, dyslipidemia, smoking, previous PCI, previous ACS, peripheral arterial disease, use of platelet inhibitor, use of statin and use of beta blocker. Inter-ethnic differences in crude SYNTAX scores were tested using ANOVA with Bonferroni post-hoc testing. Age-adjusted SYNTAX scores and SYNTAX scores adjusted for the covariates listed above were calculated with ANCOVA analysis. Differences in adjusted SYNTAX scores between the ethnic groups were tested using Tukey post-hoc testing. Inter-ethnic differences in all-cause mortality were analyzed using Kaplan Meier analysis and Cox regression analysis with adjustment for age, sex, SYNTAX score and diabetes. Also, we tested for interaction between ethnicity and SYNTAX score for all-cause mortality in the multivariable Cox model. The statistical analyses were performed using the R software package27 for statistical computing, version 3.1.2. The data used for the purpose of this study are provided in the Supporting Information (S1 File).

Results Patient characteristics The characteristics of the stable CAD and STEMI patients are presented in Table 1. Among the stable CAD patients (n=600) Indian patients were the youngest (mean age 56.8±9.5 years) and Caucasians were the oldest (63.7±10.5, p-value for difference across all ethnic groups <0.001). There was no significant difference in the proportion of men versus women among the different ethnic groups. Diabetes was significantly more common among Indians and Malays (58.0% and 52.7%, respectively) than in Caucasians and Chinese (23.5% and 36.0%, respectively), p<0.001. Dyslipidemia was more prevalent in all Asian ethnic groups (Chinese 77.3%, Indians 78.0%, Malays 75.3%) than among Caucasians (57.2%, p<0.001), and smoking was more common among Indians and Malays (40.0% and 47.0% vs. 24.6% in Caucasians, p=0.013).

100


Ethnic Differences in Quantified CAD Severity and Mortality

Table 1. Baseline characteristics and SYNTAX scores of stable CAD and STEMI patients. Figures represent percentages or means ± standard deviation. SYNTAX scores are presented as means with confidence intervals. Fully adjusted SYNTAX scores are adjusted for age, BMI, diabetes, dyslipidemia, smoking, previous PCI, previous ACS, peripheral arterial disease, use of platelet inhibitor, use of statin and use of beta blocker. Stable CAD patients N Males (%)

Caucasian

Chinese

Indian

Malay

150

150

150

150

p-value

83.3

81.3

77.3

78.0

0.52

Age (years, mean ± sd)

63.7±10.5

62.0±8.8

56.8±9.5

57.7±10.0

<0.001

BMI (kg/m2, mean ± sd)

28.0±4.4

26.5±4.8

27.4±5.1

29.2±4.9

<0.001

Diabetes (%)

23.5

36.0

58.0

52.7

<0.001

Hypertension (%)

64.0

78.0

69.3

71.3

0.06

Dyslipidemia (%)

57.2

77.3

78.0

75.3

<0.001

Current smoker (%)

24.6

28.8

40.0

47.0

0.013*

Ex smoker (%)

29.2

27.9

27.4

22.0

-

Non-smoker (%)

46.2

43.3

32.6

31.0

-

Previous PCI (%)

46.7

20.7

30.7

20.1

<0.001

Previous ACS (%)

33.3

17.4

26.8

18.0

0.002

8.1

10.0

8.0

8.7

0.92

10.7

2.7

3.3

2.0

<0.001

CVA/TIA (%) Peripheral arterial disease (%) Renal failure (%)

4.7

7.3

6.7

6.7

0.80

Anti platelet (%)

94.0

68.0

54.0

59.3

<0.001

Statin (%)

86.0

73.3

65.3

60.7

<0.001

Beta blocker (%)

74.7

40.0

47.3

52.7

<0.001

RAAS (%)

54.7

42.7

41.3

42.7

0.07

SYNTAX score (mean, 95% CI)

10.2 (9.1-11.3)

11.2 (10.2-12.1)

13.2 (12.0-14.4)

13.5 (12.4-14.6)

<0.001

Age adjusted SYNTAX score (mean, 95% CI)

10.1 (8.9-11.2)

11.1 (10.0-12.2)

13.3 (12.2-14.4)

13.6 (12.5-14.7)

<0.001

Fully adjusted SYNTAX score (mean, 95% CI)

9.4 (8.1-10.8)

11.8 (10.4-13.1)

13.4 (11.9-14.9)

13.4 (12.0-14.8)

<0.001

575

575

1,243

1,169

Median FU time (days) All-cause mortality (N) 1-year mortality estimate (%) STEMI patients

9

6

8

14

4.7

4.0

2.0

4.1

0.43

Caucasian

Chinese

Indian

Malay

p-value

N

100

100

100

100

Males (%)

79.0

82.0

85.0

88.0

0.35

Age (years, mean ± sd)

61.1±10.6

60.0±12.6

52.6±11.1

54.5±10.4

<0.001

BMI (kg/m2, mean ± sd)

27.2±4.1

25.1±5.3

26.0±5.0

26.9±4.1

0.009

13.0

35.0

46.0

41.0

<0.001

Hypertension (%)

39.4

54.0

42.0

47.0

0.17

Dyslipidemia (%)

32.3

57.0

66.0

64.6

<0.001

Current smoker (%)

44.7

53.2

69.1

79.5

<0.001*

Ex smoker (%)

22.3

14.3

10.3

8.4

-

Non-smoker (%)

33.0

32.5

20.6

12.0

-

Previous PCI (%)

7.0

7.0

12.0

10.0

0.53

Previous ACS (%)

6.0

10.0

11.1

10.0

0.62

CVA/TIA (%)

3.0

4.0

7.0

5.0

n/a

Peripheral arterial disease (%)

2.0

0.0

1.0

0.0

n/a

Renal failure (%)

0.0

0.0

3.0

2.0

n/a

Anti platelet (%)

42.0

9.0

11.0

9.0

<0.001

Statin (%)

27.0

21.0

23.0

21.0

0.71

Beta blocker (%)

27.0

13.0

10.0

11.0

0.002

Diabetes (%)

RAAS (%) SYNTAX score (mean, 95% CI)

25.0

18.0

18.0

11.0

0.08

14.0 (12.5-15.6)

18.5 (17.0-20.0)

16.1 (14.6-17.6)

18.6 (16.8-20.4)

<0.001

101


Chapter 5

Table 1. Continued Age adjusted SYNTAX score (mean, 95% CI)

13.4 (11.8-15.0)

18.0 (16.5-19.6)

16.8 (15.3-18.4)

19.0 (17.4-20.5)

<0.001

Fully adjusted SYNTAX score (mean, 95% CI)

<0.001

12.7 (10.9-14.6)

17.7 (159-19.5)

15.5 (13.5-17.4)

18.8 (17.1-20.6)

Median FU time (days)

689

589

1,050

696

All-cause mortality (N)

2

6

11

11

1-year mortality estimate (%)

2

5

9

11

0.05

* p-value for difference across all smoking groups, across all ethnic groups. n/a: chi-square test results are not robust due to <5 observations in a group. SD=standard deviation, CI=confidence interval. Significance of differences was tested with ANOVA for continuous measures, ANCOVA for adjusted SYNTAX scores and chi-square testing for proportional measures.

A history of previous PCI or ACS was significantly more common among Caucasians (p<0.001 and p=0.002), as was the use of medications recommended by international guidelines for the treatment of CAD (p<0.001). Among STEMI patients (n=400), Indians and Malays were younger (52.6±11.1 and 54.5±10.4 years) than Chinese and Caucasians (60.0±12.6 and 61.1±10.6), p-value for difference across all ethnic groups <0.001). A high proportion of males was observed among Malays (88% as compared to 79%, 82% and 85% in Caucasians, Chinese and Indians, respectively). This difference was not statistically significant (p=0.35). The prevalence of diabetes was two to three times higher among Chinese (35%), Indians (46%) and Malays (41%) than in Caucasians (13%), p-value overall <0.001. A history of PCI (7-12% across the ethnic groups) and history of ACS (6-11% across the ethnic groups) were equally common among the ethnic groups. SYNTAX scores in patients with stable CAD The crude and adjusted SYNTAX scores are depicted in Figure 1A. Crude SYNTAX scores (with 95% confidence intervals) were highest for Indians and Malays: 13.2 (12.0-14.4) and 13.5 (12.4-14.6), respectively. Age-adjusted mean SYNTAX scores for Caucasians, Chinese, Indians and Malays were 10.1 (8.9-11.2), 11.1 (10.0-12.2), 13.3 (12.2-14.4) and 13.6 (12.514.7). Even after multivariable adjustment, these ethnic differences in SYNTAX scores persisted: 9.4 (8.1-10.8), 11.8 (10.4-13.1), 13.4 (11.9-14.9) and 13.4 (12.0-14.8). Post-hoc testing revealed significantly lower SYNTAX scores in Caucasians when comparing with Indians (p=0.001) or with Malays (p<0.001) in the multivariable model. Other comparisons did not reveal significant differences. The crude and adjusted SYNTAX scores are presented in Table 1. SYNTAX scores in STEMI patients The crude and adjusted SYNTAX scores are depicted in Figure 1B. Crude SYNTAX scores for STEMI patients were highest in Chinese and Malays: 18.5 (17.0-20.0) and 18.6 (16.820.4), respectively, lower in Indians: 16.1 (14.6-17.6) and lowest in Caucasians: 14.0 (12.515.6). After adjustment for age, the scores of Caucasians and Chinese decreased to 13.4 (11.8-15.0) and 18.0 (16.5-19.6), respectively and the scores of Indians and Malays increased to 16.8 (15.3-18.4) and 19.0 (17.4-20.5). SYNTAX scores, fully adjusted for

102


Ethnic Differences in Quantified CAD Severity and Mortality

differences in baseline characteristics were: Malays 18.8 (17.1-20.6), Chinese 17.7 (15.919.5), Indians 15.5 (13.5-17.4) and Caucasians 12.7 (10.9-14.6). Post-hoc testing revealed that the fully adjusted SYNTAX scores of Caucasians differed from the scores of Chinese and Malays (p=0.002 and p<0.001). And the scores of Indians differed from the scores of Malays (p=0.046). Other post-hoc comparisons between ethnic groups did not yield a significant difference.

Figure 1. SYNTAX scores of stable CAD and STEMI patients, stratified by ethnicity. Panel A: SYNTAX scores of stable CAD patients (n=150 per ethnic group). Panel B: SYNTAX scores of STEMI patients (n=100 per ethnic group). Point estimates and error bars show the mean SYNTAX scores with 95% confidence intervals. Different transparencies present: crude mean SYNTAX scores (highly transparent), mean SYNTAX scores adjusted for age (lightly transparent) and multivariable adjusted mean SYNTAX scores (solid). P-values displayed in the figure are derived from multivariable (full model) ANCOVA, followed by Tukey post-hoc testing. The full model contains: age, body mass index, diabetes, dyslipidemia, smoking, previous PCI, previous acute coronary syndrome, peripheral arterial disease, platelet inhibitor, statin and beta-blocker use.

Mortality In patients with stable CAD, unadjusted all-cause mortality rates did not significantly differ between the four ethnic groups (Figure 2, left panel). Among STEMI patients, however, all-cause mortality was highest in Malays, reaching 11% at one year (p=0.053 for difference among the four ethnic groups). Among stable CAD patients, the multivariable Cox regression model including age, sex, SYNTAX score and diabetes as covariates, neither SYNTAX score nor ethnicity independently predicted all-cause mortality; whilst diabetes (HR 3.4, 1.6 to 7.3, p=0.001) and age (HR 1.6, 1.1 to 2.2, p=0.005) remained independent predictors of all-cause mortality. There was no interaction between SYNTAX score and ethnicity in the regression with all-cause mortality as the outcome. In STEMI patients (Figure 2, right panel), however, Indian and Malay ethnicity as compared to Caucasian ethnicity appeared to be significant independent predictors of all-cause mortality (HR 7.2, 1.5 to 34.7, p=0.01 and HR 5.8, 1.2 to 27.2, p=0.03, respectively) when

103


Chapter 5

adjusted for age, sex and diabetes. The SYNTAX score was an independent predictor of all-cause mortality (HR 2.5, 1.7 to 3.8, p<0.001 for every 10 to point increase) among STEMI patients, irrespective of ethnicity. There was no significant interaction between ethnicity and SYNTAX score for all-cause mortality in STEMI patients, indicating an equal predictive value of SYNTAX score across the ethnic groups.

Figure 2. Cox regression survival curves for up to 900 days of follow-up stratified by ethnicity. Cox regression survival curves for up to 900 days of follow-up stratified by ethnicity. The survival curves are adjusted for age, sex, SYNTAX score and diabetes. The left panel displays the stable CAD patients, the right panel the STEMI patients. No significant ethnic differences were found among stable CAD patients. Among the STEMI patients, mortality was significantly higher in Malays (HR 5.8) and Indians (HR 7.2) as compared to Caucasians.

Discussion In this comparison of CAD severity among four of the world’s most populous ethnic groups using a well-validated quantitative score of angiographic CAD severity (SYNTAX score), we observed clear ethnic differences in the severity of angiographic CAD among stable CAD and STEMI patients undergoing PCI. Indians and Malays with stable CAD had higher SYNTAX scores (reflecting more severe CAD) than Caucasians and Chinese. These differences were not attenuated when adjusting for confounders, indicating that ethnic differences in the prevalence of known traditional risk factors could not entirely explain the observed differences in CAD severity as quantified by the SYNTAX score.

104


Ethnic Differences in Quantified CAD Severity and Mortality

Indian and Malay ethnicity (in comparison to Caucasian ethnicity) were significantly associated with higher all-cause mortality in STEMI patients undergoing PCI. This remained so, after adjusting for age, SYNTAX score and diabetes. Among stable CAD patients undergoing PCI, no significant differences in mortality rates were found among the ethnic groups. No interactions were found between ethnicity and SYNTAX score for stable CAD and STEMI patients in the prediction of all-cause mortality, indicating that SYNTAX score has similar predictive properties among the ethnic groups. Ethnicity and the severity of CAD The association of Chinese ethnicity with higher SYNTAX scores among STEMI patients is intriguing, as Chinese ethnicity is often associated with a more favorable risk factor profile (e.g. lower cholesterol levels11) and lower cardiovascular mortality (2-year adjusted cardiovascular mortality rate of 1.8% vs. 4.5% among Whites28) than other ethnic groups. This pattern is also observed in Singapore.29 A lower angiographic burden of CAD has been described by Jiang et al.12 for Chinese (China) as compared to Caucasians (Australia), in a population with suspected CAD, in whom subsequent treatment was not specified. In contrast, a comparable burden of CAD between Chinese (China) and Caucasians (Germany) was reported by Zheng et al.30 They described patients with myocardial infarction or chest pain with a significant lesion upon coronary angiography. But, only part of this study population required revascularization, with more revascularization required in the Caucasian group than in the Chinese group, implying already a more diseased patient group in Caucasians than in Chinese. Our results could imply that at the point of developing symptoms of STEMI, Chinese patients have more advanced CAD than their Caucasian counterparts, as evidenced by significantly higher SYNTAX scores in the STEMI group (p=0.001). Indians (or South Asians) have repeatedly been described as a high risk group for early onset and severe CAD.31 For example, Vallapuri et al.32 described a coronary angiography cohort of patients with significant CAD, and found a higher prevalence of triple vessel disease in Asian Indians (India) as compared to Caucasians (USA) with an associated higher Gensini33 score. Compared with Caucasians, we observed high SYNTAX scores only in Indian patients with stable CAD but not in Indians presenting with STEMI. A key finding in our study is that patients of Malay descent have more severe CAD as compared to the other ethnic groups, regardless of whether they suffer from stable CAD or STEMI. Higher all-cause mortality in Malay STEMI patients has been described34 and our finding of greater CAD severity in patients of Malay descent may in part explain their poorer outcomes. Ethnicity, CAD severity and mortality For Indians, all-cause mortality from acute coronary syndromes has been described to be lower as compared to Caucasians35, which is in contrast with our results. However, in the study by Zaman et al.35 only 36% of the study population was comprised of STEMI patients, and an even lower percentage received primary PCI. Within the STEMI patients

105


Chapter 5

of their study, the mortality in Indians was higher compared to the Caucasians, which is in agreement with our findings. Chinese are known to have lower CAD-attributable mortality rates than Caucasians in the general population.36 However, the literature is inconclusive about mortality among Chinese CAD patients. Gasevic et al.37 and Khan et al.38 reported that after myocardial infarction Chinese are at higher risk of dying in the first 30 days after PCI than Caucasians. Long-term all-cause mortality did not differ significantly between Chinese and Caucasians in the analysis by Khan et al.38 In contrast, Qian et al.39 reported that Asian Americans (of whom the largest proportion is Chinese40) show better survival after myocardial infarction than Caucasians. In our current study with a median follow-up duration of 709 days, we did not find a difference between survival in Chinese and Caucasian patients. Due to the chance of a type II error and possible under-powering of the current study, we cannot exclude the possibility of an actual difference in all-cause mortality between Chinese and Caucasians, however our data indicate no striking numerical differences in mortality between Chinese and Caucasians. To date, no comparison between Caucasians and Malays has been reported in the literature. However, ten years ago, Mak et al.34 reported Malays to have a poorer prognosis after myocardial infarction than Chinese and Indians in Singapore (HR 1.26 as compared to Chinese ethnicity), supporting the results of high SYNTAX scores and high rates of all-cause mortality among Malays in the current study. Klomp et al.3 reported better survival in Asian patients undergoing PCI, mostly of Malay descent, than in Western Europeans undergoing PCI. This important earlier study, however, encompassed all indications for PCI, and elective PCI was significantly more common in the Asian ethnic group than in the Western European group; while survival analysis was not adjusted for PCI indication. In our study we find differences in all-cause mortality among STEMI patients, but not among stable patients. The STEMI patient group is a very homogeneous patient group with a clear phenotype. The stable CAD patient group, on the other hand, might encompass a wider range of patients, thereby masking a possible effect of ethnicity by subtle differences in stable indications for angiography among the ethnic groups. Possible mechanisms of inter-ethnic differences in CAD severity The association of ethnicity with SYNTAX score is likely to be of a multifactorial origin. It is conceivable that there are cultural lifestyle or country-dependent primary prevention strategies differences among the ethnic groups, which are not completely reflected by the cardiovascular risk factor profile. Reports suggest that CAD outcomes are strongly linked to a country’s economic prosperity.41 Remarkably, Singapore has a per capita GDP of US$ 55,182, higher than the Dutch GDP of US$ 50,793 (data for 2010-2014 from data. worldbank.org). On the other hand, there might be true genetic42,43 and biological differences among the ethnic groups of which only the tip of the iceberg is known. For example, a stronger association of cholesterol and diabetes with carotid intima-media thickness has been

106


Ethnic Differences in Quantified CAD Severity and Mortality

shown for Asian Indians as compared to Caucasians, suggesting higher susceptibility of the arterial vasculature to cholesterol and glucose levels.44 In our study we were only able to take the absence or presence of dyslipidemia and diabetes into account, and not the actual biomarker levels reporting on these diseases. It is possible that the actual cholesterol levels or glycated hemoglobin levels associate with SYNTAX score in different ways across the ethnic groups and thus explain part of the association of ethnicity with SYNTAX score that we find. Unraveling the underlying causes of these ethnic differences in CAD severity is critical when tackling the epidemic of CAD in Asia. When biological45 and epidemiological connections between ethnicity and CAD become more delineated, ethnicity-specific prevention and treatment strategies can be developed to improve survival in the ethnic groups with poorest prognoses. Limitations In this study we were unable to correct for lifestyle factors beyond those that were described in the baseline table. It is possible that dietary, socioeconomic factors obscure the true association of ethnicity with the angiographic severity of CAD into some extent. Furthermore, lifestyle and medication adherence may also explain differences in mortality following PCI. By specifically including PCI patients we created a homogenous patient group; however the choice for PCI might have been steered by patient preferences (refusing CABG). We were unable to take patient preferences for revascularization strategy into account and it is possible that patient preferences differ among ethnic groups. While the SYNTAX score is the most widely used scoring method to quantify the angiographic severity of CAD, it does have limitations. For example, SYNTAX score does not distinguish between a lesion of 50% and a lesion of 99% luminal stenosis.13 Plus, the assessment of significance of a lesion is performed by visual estimation and thus observer-dependent (although inter-observer correlation in our study was high, r=0.95). Due to a limitation in statistical power we could not extend our multivariable survival analyses with more covariates. Conclusion In both stable CAD and STEMI patients undergoing PCI, SYNTAX scores were higher in Malays and Indians as compared to Caucasians, also after adjustment for relevant covariates. After STEMI, mortality was significantly higher among Indian and Malay patients as compared to Caucasian patients, even after multivariable adjustment including for SYNTAX score. Future research should focus on understanding underlying intrinsic biological and environmental factors accounting for these differences in CAD severity and case-fatality, so as to identify better targets for early intervention.

107


Chapter 5

References 1. Chiu M, Austin PC, Manuel DG, Tu J V. Comparison of cardiovascular risk profiles among ethnic groups using population health surveys between 1996 and 2007. CMAJ. 2010;182:E301–10. 2.

Gijsberts CM, den Ruijter HM, Asselbergs FW, Chan MY, de Kleijn DP V, Hoefer IE. Biomarkers of Coronary Artery Disease Differ Between Asians and Caucasians in the General Population. Glob Heart. 2015;

3.

Klomp M, Damman P, Beijk M a M, Tan KH, Balian V, de Luca G, Tijssen JGP, Silber S, de Winter RJ. Differences in cardiovascular risk factors and clinical outcomes between Western European and Southeast Asian patients treated with the Genous Bio-engineered R stent: an e-HEALING worldwide registry substudy. Coron Artery Dis. 2012;23:271–7.

4.

Enas EA, Garg A, Davidson MA, Nair VM, Huet BA, Yusuf S. Coronary heart disease and its risk factors in firstgeneration immigrant Asian Indians to the United States of America. Indian Heart J. 48:343–353.

5.

Alwan A. Global status report on noncommunicable diseases 2010. 1st ed. Geneva, Switzerland: 2011.

6.

Kanaya AM, Kandula NR, Ewing SK, Herrington D, Liu K, Blaha MJ, Srivastava S, Dave SS, Budoff MJ. Comparing coronary artery calcium among U.S. South Asians with four racial/ethnic groups: the MASALA and MESA studies. Atherosclerosis. 2014;234:102–7.

7.

Anand SS, Yusuf S, Vuksan V, Devanesen S, Teo KK, Montague PA, Kelemen L, Yi C, Lonn E, Gerstein H, Hegele RA, McQueen M. Differences in risk factors, atherosclerosis, and cardiovascular disease between ethnic groups in Canada: the Study of Health Assessment and Risk in Ethnic groups (SHARE). Lancet. 2000;356:279–84.

8.

Shaw LJ, Shaw RE, Merz CNB, Brindis RG, Klein LW, Nallamothu B, Douglas PS, Krone RJ, McKay CR, Block PC, Hewitt K, Weintraub WS, Peterson ED. Impact of ethnicity and gender differences on angiographic coronary artery disease prevalence and in-hospital mortality in the American College of Cardiology-National Cardiovascular Data Registry. Circulation. 2008;117:1787–801.

9.

Goff DC, Lloyd-Jones DM, Bennett G, Coady S, D’Agostino RB, Gibbons R, Greenland P, Lackland DT, Levy D, O’Donnell CJ, Robinson J, Schwartz JS, Shero ST, Smith SC, Sorlie P, Stone NJ, Wilson PWF. 2013 ACC/ AHA Guideline on the Assessment of Cardiovascular Risk: A Report of the American College of Cardiology/ American Heart Association Task Force on Practice Guidelines. Circulation. 2013;

10. Koulaouzidis G, Nicoll R, Charisopoulou D, McArthur T, Jenkins PJ, Henein MY. Aggressive and diffuse coronary calcification in South Asian angina patients compared to Caucasians with similar risk factors. Int J Cardiol. 2013;167:2472–6. 11. Bild DE, Detrano R, Peterson D, Guerci A, Liu K, Shahar E, Ouyang P, Jackson S, Saad MF. Ethnic differences in coronary calcification: the Multi-Ethnic Study of Atherosclerosis (MESA). Circulation. 2005;111:1313–20. 12. Jiang S, Lv L, Juergens CP, Chen S, Xu D, Huang Z. Racial differences in coronary artery lesions: a comparison of coronary artery lesions between mainland Chinese and Australian patients. Angiology. 2008;59:442–7. 13. Sianos G, Morel M, Kappetein AP, Morice M, Colombo A, Dawkins K, van den Brand M, Van Dyck N, Russell ME, Mohr FW, Serruys PW. The SYNTAX Score: an angiographic tool grading the complexity of coronary artery disease. EuroIntervention. 2005;1:219–27. 14. Serruys PW, Onuma Y, Garg S, Sarno G, van den Brand M, Kappetein A-P, Van Dyck N, Mack M, Holmes D, Feldman T, Morice M-C, Colombo A, Bass E, Leadley K, Dawkins KD, van Es G-A, Morel M-AM, Mohr FW. Assessment of the SYNTAX score in the Syntax study. EuroIntervention. 2009;5:50–56. 15. Thompson JH, Potok N. Population Clock [Internet]. 2014;Available from: http://www.census.gov/popclock/

108


Ethnic Differences in Quantified CAD Severity and Mortality

16. Montalescot G, Sechtem U, Achenbach S, Andreotti F, Arden C, Budaj A, Bugiardini R, Crea F, Cuisset T, Di Mario C, Ferreira JR, Gersh BJ, Gitt AK, Hulot J-S, Marx N, Opie LH, Pfisterer M, Prescott E, Ruschitzka F, Sabaté M, Senior R, Taggart DP, van der Wall EE, Vrints CJM, Zamorano JL, Baumgartner H, Bax JJ, Bueno H, Dean V, Deaton C, Erol C, Fagard R, Ferrari R, Hasdai D, Hoes AW, Kirchhof P, Knuuti J, Kolh P, Lancellotti P, Linhart A, Nihoyannopoulos P, Piepoli MF, Ponikowski P, Sirnes PA, Tamargo JL, Tendera M, Torbicki A, Wijns W, Windecker S, Valgimigli M, Claeys MJ, Donner-Banzhoff N, Frank H, Funck-Brentano C, Gaemperli O, Gonzalez-Juanatey JR, Hamilos M, Husted S, James SK, Kervinen K, Kristensen SD, Maggioni A Pietro, Pries AR, Romeo F, Rydén L, Simoons ML, Steg PG, Timmis A, Yildirir A. 2013 ESC guidelines on the management of stable coronary artery disease: the Task Force on the management of stable coronary artery disease of the European Society of Cardiology. Addenda. Eur Heart J. 2013;34:2949–3003. 17. Thygesen K, Alpert JS, Jaffe AS, Simoons ML, Chaitman BR, White HD, Katus HA, Lindahl B, Morrow DA, Clemmensen PM, Johanson P, Hod H, Underwood R, Bax JJ, Bonow RO, Pinto F, Gibbons RJ, Fox KA, Atar D, Newby LK, Galvani M, Hamm CW, Uretsky BF, Steg PG, Wijns W, Bassand J-P, Menasché P, Ravkilde J, Ohman EM, Antman EM, Wallentin LC, Armstrong PW, Januzzi JL, Nieminen MS, Gheorghiade M, Filippatos G, Luepker R V, Fortmann SP, Rosamond WD, Levy D, Wood D, Smith SC, Hu D, Lopez-Sendon J-L, Robertson RM, Weaver D, Tendera M, Bove AA, Parkhomenko AN, Vasilieva EJ, Mendis S. Third universal definition of myocardial infarction. Circulation. 2012;126:2020–35. 18. World Health Organization. Definition and diagnosis of diabetes mellitus and intermediate hyperglycaemia. Geneva, Switzerland: 2006. 19. McManus RJ, Caulfield M, Williams B. NICE hypertension guideline 2011: evidence based evolution. BMJ. 2012;344:e181. 20. Reiner Z, Catapano AL, De Backer G, Graham I, Taskinen M-R, Wiklund O, Agewall S, Alegria E, Chapman MJ, Durrington P, Erdine S, Halcox J, Hobbs R, Kjekshus J, Filardi PP, Riccardi G, Storey RF, Wood D. ESC/EAS Guidelines for the management of dyslipidaemias: the Task Force for the management of dyslipidaemias of the European Society of Cardiology (ESC) and the European Atherosclerosis Society (EAS). Eur Heart J. 2011;32:1769–818. 21. Inker LA, Astor BC, Fox CH, Isakova T, Lash JP, Peralta CA, Kurella Tamura M, Feldman HI. KDOQI US commentary on the 2012 KDIGO clinical practice guideline for the evaluation and management of CKD. Am J Kidney Dis. 2014;63:713–35. 22. Law MR, Wald NJ, Rudnicka AR. Quantifying effect of statins on low density lipoprotein cholesterol, ischaemic heart disease, and stroke: systematic review and meta-analysis. BMJ. 2003;326:1423. 23. Law MR, Wald NJ, Morris JK, Jordan RE. Value of low dose combination treatment with blood pressure lowering drugs: analysis of 354 randomised trials. BMJ. 2003;326:1427. 24. Capodanno D, Di Salvo ME, Cincotta G, Miano M, Tamburino C, Tamburino C. Usefulness of the SYNTAX score for predicting clinical outcome after percutaneous coronary intervention of unprotected left main coronary artery disease. Circ Cardiovasc Interv. 2009;2:302–308. 25. SYNTAX Steering Committee. SYNTAX Score Calculator [Internet]. 2012;Available from: http://www. syntaxscore.com/calc/start.htm 26. Généreux P, Palmerini T, Caixeta A, Cristea E, Mehran R, Sanchez R, Lazar D, Jankovic I, Corral MD, Dressler O, Fahy MP, Parise H, Lansky AJ, Stone GW. SYNTAX score reproducibility and variability between interventional cardiologists, core laboratory technicians, and quantitative coronary measurements. Circ Cardiovasc Interv. 2011;4:553–61.

109


Chapter 5

27. R Core Team. R: A Language and Environment for Statistical Computing. Vienna, Austria: 2014. 28. Meadows T a., Bhatt DL, Cannon CP, Gersh BJ, Röther J, Goto S, Liau C-S, Wilson PWF, Salette G, Smith SC, Steg PG. Ethnic Differences in Cardiovascular Risks and Mortality in Atherothrombotic Disease: Insights From the REduction of Atherothrombosis for Continued Health (REACH) Registry. Mayo Clin Proc. 2011;86:960–967. 29. Ministry of Health: Epidemiology & Disease Control. National Health Surveillance Survey 2007. Singapore: 2007. 30. Zheng Y, Ma W, Zeng Y, Liu J, Ye S, Chen S, Lan L, Erbel R, Liu Q. Comparative study of clinical characteristics between Chinese Han and German Caucasian patients with coronary heart disease. Clin Res Cardiol. 2010;99:45–50. 31. Enas EA, Mehta J. Malignant coronary artery disease in young Asian Indians: thoughts on pathogenesis, prevention, and therapy. Coronary Artery Disease in Asian Indians (CADI) Study. Clin Cardiol. 1995;18:131–5. 32. Vallapuri S, Gupta D, Talwar K., Billie M, Mehta MC, Morise AP, Jain AC. Comparison of atherosclerotic risk factors in Asian Indian and American Caucasian patients with angiographic coronary artery disease. Am J Cardiol. 2002;90:1147–1150. 33. Austen WG, Edwards JE, Frye RL, Gensini GG, Gott VL, Griffith LS, McGoon DC, Murphy ML, Roe BB. A reporting system on patients evaluated for coronary artery disease. Report of the Ad Hoc Committee for Grading of Coronary Artery Disease, Council on Cardiovascular Surgery, American Heart Association. Circulation. 1975;51:5–40. 34. Mak K-H, Chia K-S, Kark JD, Chua T, Tan C, Foong B-H, Lim Y-L, Chew S-K. Ethnic differences in acute myocardial infarction in Singapore. Eur Heart J. 2003;24:151–60. 35. Zaman MJS, Philipson P, Chen R, Farag A, Shipley M, Marmot MG, Timmis AD, Hemingway H. South Asians and coronary disease: is there discordance between effects on incidence and prognosis? Heart. 2013;99:729–36. 36. Palaniappan L, Wang Y, Fortmann SP. Coronary heart disease mortality for six ethnic groups in California, 1990-2000. Ann Epidemiol. 2004;14:499–506. 37. Gasevic D, Khan N a, Qian H, Karim S, Simkus G, Quan H, Mackay MH, O’Neill BJ, Ayyobi AF. Outcomes following percutaneous coronary intervention and coronary artery bypass grafting surgery in Chinese, South Asian and White patients with acute myocardial infarction: administrative data analysis. BMC Cardiovasc Disord. 2013;13:121. 38. Khan N, Grubisic M, Hemmelgarn B, Humphries K, King KM, Quan H. Outcomes after acute myocardial infarction in South Asian, Chinese, and white patients. Circulation. 2010;122:1570–1577. 39. Qian F, Ling FS, Deedwania P, Hernandez AF, Fonarow GC, Cannon CP, Peterson ED, Peacock WF, Kaltenbach L a, Laskey WK, Schwamm LH, Bhatt DL. Care and outcomes of Asian-American acute myocardial infarction patients: findings from the American Heart Association Get With The Guidelines-Coronary Artery Disease program. Circ Cardiovasc Qual Outcomes. 2012;5:126–33. 40. Paisano E. We, the American Asians. 1st ed. Washington, DC: U.S. Government Printing Office; 1993. 41. Yusuf S, Rangarajan S, Teo K, Islam S, Li W, Liu L, Bo J, Lou Q, Lu F, Liu T, Yu L, Zhang S, Mony P, Swaminathan S, Mohan V, Gupta R, Kumar R, Vijayakumar K, Lear S, Anand S, Wielgosz A, Diaz R, Avezum A, Jaramillo PL, Lanas F, Yusoff K, Ismail N, Iqbal R, Rahman O, Lopez-Jaramillo P, Lanas F, Yusoff K, Ismail N, Iqbal R, Rahman O, Rosengren A, Yusufali A, Kelishadi R, Kruger A, Puoane T, Szuba A, Chifamba J, Oguz A, McQueen M, McKee M, Dagenais G. Cardiovascular Risk and Events in 17 Low-, Middle-, and High-Income Countries. N Engl J Med. 2014;371:818–827.

110


Ethnic Differences in Quantified CAD Severity and Mortality

42. Maitra A, Shanker J, Dash D, John S, Sannappa PR, Rao VS, Ramanna JK, Kakkar V V. Polymorphisms in the IL6 gene in Asian Indian families with premature coronary artery disease--the Indian Atherosclerosis Research Study. Thromb Haemost. 2008;99:944–50. 43. Tan JH-H, Low P-S, Tan Y-S, Tong M-C, Saha N, Yang H, Heng C-K. ABCA1 gene polymorphisms and their associations with coronary artery disease and plasma lipids in males from three ethnic populations in Singapore. Hum Genet. 2003;113:106–17. 44. Chow CK, McQuillan B, Raju PK, Iyengar S, Raju R, Harmer JA, Neal BC, Celermajer DS. Greater adverse effects of cholesterol and diabetes on carotid intima-media thickness in South Asian Indians: comparison of risk factor-IMT associations in two population-based surveys. Atherosclerosis. 2008;199:116–22. 45. Wang JW, Gijsberts CM, Seneviratna A, de Hoog VC, Vrijenhoek JEP, Schoneveld a H, Chan MY, Lam CSP, Richards a M, Lee CN, Mosterd A, Sze SK, Timmers L, Lim SK, Pasterkamp G, de Kleijn DP V. Plasma extracellular vesicle protein content for diagnosis and prognosis of global cardiovascular disease. Neth Heart J. 2013;21:467–71.

Supporting Information S1 File. Data. Data used for the purpose of this study. Available from: http://journals.plos. org/plosone/article?id=10.1371/journal.pone.0131977#sec026.

111



PART ONE Ethnicity

Chapter 6 The Ethnicity-specific Association of Biomarkers with the Angiographic Severity of Coronary Artery Disease Accepted for publication in the Netherlands Heart Journal

Crystel M. Gijsberts, Aruni Seneviratna*, Ingrid E.M. Bank*, Hester M. den Ruijter, Folkert W. Asselbergs, Pierfrancesco Agostoni, Jasper A. Remijn, Gerard Pasterkamp, Heng Chew Kiat, Mark Roest, Arthur Mark Richards, Mark Y. Chan, Dominique P.V. de Kleijn, Imo E. Hoefer * These authors contributed equally to this work.


Chapter 6

Abstract Background Risk factor burden and clinical characteristics of coronary artery disease (CAD) patients differ among ethnic groups. We related biomarkers to CAD severity in Caucasians, Chinese, Indians and Malays. Methods In the Dutch-Singaporean UNICORN coronary angiography cohort (n=2,033) we compared levels of five cardiovascular biomarkers: N-terminal pro-brain natriuretic peptide (NTproBNP), high-sensitivity C-reactive protein (hsCRP), Cystatin C (CysC), myeloperoxidase (MPO) and high-sensitivity Troponin I (hsTnI). We assessed ethnicityspecific associations of biomarkers with CAD severity, quantified by the SYNTAX score. Results Adjusted for baseline differences, NTproBNP levels were significantly higher in Malays than in Chinese and Caucasians (72.1 vs. 34.4 and 41.1 pmol/L, p<0.001 and p=0.005, respectively). MPO levels were higher in Caucasians than in Indians (32.8 vs. 27.2 ng/mL, p=0.026) and hsTnI levels were higher in Malays than in Caucasians and Indians (33.3 vs. 16.4 and 17.8 ng/L, p<0.001 and p=0.029) and hsTnI levels were higher in Chinese than in Caucasians (23.3 vs. 16.4, p=0.031). We found modifying effects of ethnicity on the association of biomarkers with SYNTAX score. NTproBNP associated more strongly with SYNTAX score in Malays than Caucasians (β 0.132 vs. β 0.020 per 100 pmol/l increase in NTproBNP, p=0.032). For MPO levels the association was stronger in Malays than Caucasians (β 1.146 vs. β 0.016 per 10 ng/mL increase, p=0.017). Differing biomarker level cut-offs for the ethnic groups were found. Conclusion When corrected for possible confounders we observe ethnicity-specific differences in biomarker levels. Moreover, biomarkers associated differently with CAD severity, suggesting that ethnicity-specific cut-off values should be considered.

114


Ethnicity-specific Association of Biomarkers with CAD Severity

Introduction Coronary artery disease (CAD) is highly prevalent worldwide but over the next few decades the majority of deaths due to CAD will occur in Asia.1 Evidence is accumulating that important differences exist between the Asian ethnic groups and Caucasians, who have been the main focus of cardiovascular research up until now. We know that risk factor levels differ markedly. Specifically, a high prevalence of diabetes has been described for South Asians.2 Also, the CAD phenotype appears to differ among the ethnic groups. Triple vessel disease is more common in South Asians than in Caucasians3, whilst the Chinese suffer from less severe CAD.4 Blood-derived biomarkers are non-invasive tools that can be indicative of CAD severity. Population means of biomarker levels differ among ethnic groups5. However, it is unknown whether biomarkers of CAD report similarly on underlying CAD severity among these groups.6 In the current study, we investigated patients undergoing coronary angiography for suspected CAD from four globally populous ethnic groups: Caucasians, Chinese, Indians and Malays, enrolled in two countries with high, comparable health care standards7: the Netherlands and Singapore. We evaluated five established biomarkers known to be affected by CAD: N-Terminal pro-brain natriuretic peptide8–10 (NTproBNP), high-sensitivity C-reactive protein11 (hsCRP), Cystatin C12,13 (CysC), myeloperoxidase14,15 (MPO) and high-sensitivity Troponin I16,17 (hsTnI). These biomarkers reflect pivotal CAD aspects and complications: cardiac hemodynamic load (NTproBNP), inflammation (hsCRP and MPO), kidney function (CysC) and cardiomyocyte damage (hsTnI). In this study we aimed to define inter-ethnic differences in the association of these biomarkers with CAD severity, quantified by the SYNTAX score18.

Methods This study was conducted using the parallel United CORoNary biobanks, the UNICORN cohort (clinicaltrials.gov ID: NCT02126150), consisting of consecutive patients undergoing coronary angiography recruited from two sites: the University Medical Center Utrecht, the Netherlands and the National University Hospital, Singapore. The institutional review boards of both centers approved of this study, which conforms to the declaration of Helsinki. Patients were enrolled between September 2010 and March 2013. Patients from four ethnic groups were enrolled, including Caucasians in the Netherlands and Chinese, Indians and Malays in Singapore. Blood was sampled from the arterial sheath, inserted at commencement of the coronary angiography procedure and immediately stored at -80°C. Clinical and angiographic characteristics Risk factor profiles were documented at or around the time of coronary angiography. Coronary angiograms were categorized into four groups: no/minor CAD, single vessel

115


Chapter 6

disease, double vessel disease and triple vessel disease (defined as number of epicardial vessels with 50% stenosis19 on visual assessment). In the latter three categories, a SYNTAX score was determined.18 The process of SYNTAX scoring has been described previously.20 Biomarker assays Plasma levels of NTproBNP, hsCRP, CysC and MPO were measured at the University Medical Center Utrecht, The Netherlands, using a semi-automated ELISA robot (Freedom EVO, Tecan, Switzerland). Commercial antibody combinations were used to quantify NTproBNP (15C4 and biotinylated 13G12, Hi-test Finland), hsCRP (Dy1707 duoset, R&D systems), CysC (Dy1196 duoset, R&D systems) and MPO (Dy3667 duoset, R&D systems). In brief, maxisorb plates were coated with mouse anti-human NTproBNP, hsCRP, CysC or MPO. Plates were blocked with 1% bovine serum albumin, and incubated with supernatants. Plates were washed with phosphate buffered saline (PBS) pH 7.4 with 0.05% Tween 20. Bound factors were detected with biotin coupled detection antibodies. Biotin coupled antibodies were bound with streptavidin-HRP, or goat-anti-human antibodies with rabbit-anti-goat HRP (DAKO, P0449). Detection was performed with SuperSignal West Pico Chemiluminescent substrate, and read with a luminometer. The intra-assay variation coefficient was: 10%. Levels of hsTnI were measured using the STATÂ High Sensitive Troponin-I assay on the clinically validated ARCHITECT i2000 analyzer (Abbott Laboratories, Lisnamuck, Longford, Ireland). All samples (Singaporean and Dutch) were randomly distributed across the plates for all analyses. Statistical analysis Figures are presented as means with standard deviations for normally distributed variables or medians with interquartile ranges for non-normally distributed variables. Baseline characteristics were compared among the ethnic groups using ANOVA for normally distributed variables, Kruskal-Wallis tests for non-normally distributed data and chi-square tests for categorical data. We evaluated inter-ethnic differences in biomarker levels, adjusted for baseline differences in age, sex, body mass index (BMI), diabetes, hypertension, dyslipidemia, smoking, indication for procedure, severity of CAD, anti-platelet medication use, beta blocker use, calcium antagonist use, renin-angiotensin-aldosterone system (RAAS) medication use and statin use. Using ANCOVA, adjusted biomarker levels were calculated for each ethnic group. Biomarker levels were positively skewed and therefore logtransformed for analysis when used as the dependent variable. The presented adjusted biomarker level means are antilogs (back-transformed after analysis). Biomarker levels were compared among the ethnic groups through Tukey post-hoc testing, thus adjusting for multiple testing. Also, we tested for interactions between ethnicity and biomarker levels for SYNTAX score in univariable and multivariable regression models. Direct comparison of biomarker coefficients between ethnicities was performed by testing an interaction term of

116


Ethnicity-specific Association of Biomarkers with CAD Severity

biomarker level with ethnicity; Caucasian ethnicity was the reference group. A conservative p-value of 0.05 for the interaction terms was deemed significant. Next, using receiver operating characteristic (ROC) analysis, we compared the optimal cut-off biomarker values by ethnicity, defined as the biomarker level at which sensitivity + specificity for a high SYNTAX score was largest. Also, areas under curves (AUC) for each biomarker in each ethnic group for association with a high SYNTAX score were calculated.21 The outcome SYNTAX score was dichotomized on a previously reported22 cut-off of 18 points. All analyses were performed using the R software package23. A two-tailed Îą of: 0.05 was considered statistically significant.

Results Baseline characteristics Marked baseline differences were observed among the ethnic groups (Table 1). Indian and Malay patients were younger (54.4 and 55.4 years, respectively) than the Chinese and Caucasian patients (59.1 and 64.7 years, respectively, p<0.001). Among Indian and Malay patients, diabetes (54.4% and 52.5%, respectively) and dyslipidemia (76.6% and 75.1%, respectively) were strikingly more common than in Caucasians (diabetes 22.6%, p<0.001, dyslipidemia 48.2%, p<0.001). The prevalence of current smokers was highest among Indians (42.9%). The indication for angiography differed among the ethnic groups (p<0.001). Stable CAD was slightly less common among Indians and Malays (58.8% and 54.8%, respectively, vs. 67.4% in Caucasians and 65.9% in Chinese). Unstable angina or non-ST-elevated myocardial infarction (UA/NSTEMI) was more frequently observed in all three Asian groups than in Caucasians (Chinese 31.6%, Indian 36.6%, Malay 42.9% vs. 22.5% in Caucasians, p<0.001). Also, triple vessel disease was more often diagnosed in the Asian ethnic groups than in Caucasians (Chinese 20.8%, Indian 22.0%, Malay 26.3% vs. 12.2% in Caucasians, p<0.001). Percutaneous coronary intervention (PCI) was more frequently performed on Caucasians than in the other ethnic groups, in whom a conservative strategy was opted for more frequently (p<0.001). Coronary artery bypass graft (CABG) surgery was performed most often in Malays (11.0% vs. 4.4% in Indians, 6.2% in Caucasians and 8.9% in Chinese). Biomarker levels Crude and multivariable-adjusted biomarker levels are displayed in Figure 1, stratified by ethnicity. Crude biomarker levels differed significantly among the ethnic groups for all of the examined biomarkers (table 1). In table 2, the crude biomarker levels are displayed for each ethnic group, stratified by indication for angiography and angiographic severity of CAD. All biomarkers differed among the ethnic groups in at least one indication group. hsTnI did not differ in any of the CAD severity groups, while the other biomarkers did differ significantly in at least one CAD severity group.

117


Chapter 6

Table 1. Baseline characteristics of UNICORN participants, stratified by ethnicity. Caucasian

Chinese

Indian

Malay

p-value

N

1132

562

158

181

Males (%)

72.5

81.0

80.4

79.6

<0.001

Age (mean (sd), years)

64.7 (11.3)

59.1 (9.9)

54.4 (9.5)

55.4 (9.4)

<0.001

BMI (mean (sd), kg/m2)

27.0 (4.3)

26.2 (4.7)

27.5 (5.0)

28.5 (5.4)

<0.001

22.6

34.0

54.4

52.5

<0.001

Diabetes (%) Hypertension (%)

56.1

65.1

63.9

64.6

0.001

Dyslipidemia (%)

48.2

69.6

76.6

75.1

<0.001

Smoking, current (%)

23.6

29.4

42.9

39.0

<0.001

Smoking, ex (%)

31.6

19.2

11.9

16.4

Renal failure (%)

3.1

6.0

3.8

5.6

0.034

Medications Platelet inhibitor (%)

64.6

58.4

57.6

55.8

0.016

Statin (%)

58.9

44.3

43.0

43.1

<0.001

Beta blocker (%)

25.4

26.7

24.7

19.3

0.259

RAAS (%)

51.4

35.2

40.5

43.6

<0.001

Calcium antagonist (%)

64.3

61.4

61.4

55.8

0.146

Previous PCI (%)

31.6

17.5

29.1

30.0

<0.001

Previous CABG (%)

13.0

6.2

6.3

6.1

<0.001

Previous ACS (%)

32.6

14.8

26.8

28.5

<0.001

History of CVA/TIA (%)

10.3

5.7

8.2

6.7

0.010

History of PAD (%)

11.5

2.8

4.4

5.0

<0.001

Stable CAD (%)

67.4

65.9

58.8

54.8

<0.001

UA/NSTEMI (%)

22.5

31.6

36.6

42.9

STEMI (%)

10.1

2.4

4.6

2.3

25.0

33.8

28.7

25.1

Medical history

Indication for coronary angiography

CAD severity No/minor CAD (%)

<0.001

Single vessel disease (%)

37.3

21.7

24.0

21.1

Double vessel disease (%)

25.5

23.6

25.3

27.4

Triple vessel disease (%)

12.2

20.8

22.0

26.3

Conservative (%)

35.0

53.6

58.2

55.2

PCI (%)

58.8

37.5

37.3

33.7

6.2

8.9

4.4

11.0

43.1 [10.5, 138.0]

37.8 [11.1, 135.2]

31.5 [8.2, 108.0]

63.9 [16.3, 201.8]

0.049

Treatment

CABG (%)

<0.001

Biomarker levels NTproBNP (median [IQR]) hsCRP (median [IQR])

1.5 [0.6, 3.8]

1.3 [0.6, 4.2]

2.3 [1.0, 6.7]

2.2 [0.8, 7.3]

<0.001

CysC (median [IQR])

0.8 [0.7, 1.1]

0.7 [0.6, 0.9]

0.8 [0.6, 0.9]

0.8 [0.5, 1.0]

<0.001

MPO (median [IQR])

28.1 [20.7, 44.8]

27.0 [20.6, 37.0]

26.8 [19.8, 35.1]

30.3 [23.5, 40.9]

0.001

hsTnI (median [IQR])

7.8 [3.8, 28.4]

8.1 [4.2, 48.1]

7.7 [4.2, 69.1]

12.2 [4.6, 209.5]

0.001

118


Ethnicity-specific Association of Biomarkers with CAD Severity

Table 1. Continued Caucasian

Chinese

Indian

Malay

p-value

SYNTAX Score (mean (sd))

11.6 (8.2)

15.2 (10.0)

14.5 (10.6)

15.6 (11.3)

<0.001

High SYNTAX score (n (%) ≥18 points)

147 (21.0)

116 (34.0)

34 (33.0)

47 (38.2)

823 [653, 988]

930 [734, 1062]

882 [714, 1044]

852 [649, 1045]

75

36

3

11

Outcome

FUtime (median [IQR], days) All-cause death (n)

<0.001 <0.001

Baseline characteristics of UNICORN patients stratified by ethnicity. Non-normally distributed continuous variables were compared using ANOVA across ethnic groups. Biomarker levels were compared by Kruskal-Wallis testing. Categorical data were compared using chi-square testing. Reported p-values refer to overall differences across the ethnic groups. Significant p-values are printed in bold. Abbreviations= BMI= body mass index, RAAS= renin-angiotensin-aldosteron system, PCI= percutaneous coronary intervention, CABG= coronary artery bypass grafting, ACS= acute coronary syndrome, CVA= cerebrovascular accident, TIA= transient ischemic attack, CAD coronary artery disease, UA= unstable angina, NSTEMI= non-ST-elevated myocardial infarction, STEMI= ST-elevated myocardial infarction, NTproBNP= N-terminal brain natriuretic peptide, hsCRP= high-sensitivity C-reactive protein, CysC= Cystatin C, MPO= myeloperoxidase, hsTnI= high-sensitivity troponin I, FUtime= follow-up time.

NTproBNP (pmol/L)

hsCRP (µg/mL)

p<0.001

CysC (µg/mL)

1.0

4

p=0.005

100

0.9

3 75

0.8

● 50

● 0.7

25

Chinese

Indian

Malay

MPO (ng/mL)

37.5

Caucasian

Chinese

Indian

Malay

60

27.5

p=0.031

Indian Malay

Malay

Chinese

40

30

Indian

Caucasian

● ●

Chinese

p=0.029

50

Caucasian

p<0.001

35.0

30.0

0.6

hsTnI (ng/L)

p=0.026

32.5

Caucasian

2

20

● 25.0

Caucasian

Chinese

Indian

Malay

10

Caucasian

Chinese

Indian

Malay

Figure 1. Biomarker levels by ethnicity, corrected for baseline differences (ANCOVA). The unadjusted means of biomarker levels (transparent) and adjusted means derived from ANCOVA (solid) are shown. The adjusted biomarker levels are corrected for: age, sex, BMI, diabetes, hypertension, dyslipidemia, smoking, indication for angiography, CAD severity, anti-platelet medication use, beta blocker use, calcium antagonist use, RAAS inhibiting medication use and statin use. CysC levels were additionally corrected for renal failure. P-values that are presented in the plots are the result from ANCOVA with post-hoc testing (p-values are corrected according to the Tukey method).

119


Chapter 6

When adjusted for baseline characteristics, certain differences in biomarker levels remained statistically significant. Post-hoc testing revealed that multivariable-adjusted NTproBNP levels were significantly higher in Malays than in Chinese and Caucasians (72.1 vs. 34.4 and 41.1 pmol/L, p<0.001 and p=0.005, respectively). MPO levels were higher in Caucasians than in Indians (32.8 vs. 27.2 ng/mL, p=0.026), hsTnI levels were higher in Malays than in Caucasians and Indians (33.3 vs. 16.4 and 17.8 ng/L, p<0.001 and p=0.029) and hsTnI levels were higher in Chinese than in Caucasians (23.3 vs. 16.4, p=0.031). hsCRP levels and CysC levels (additionally adjusted for renal failure) did not differ among the ethnic groups. Modifying effect of ethnicity on association between biomarker levels and SYNTAX score We tested interactions of ethnicity with biomarker levels for SYNTAX score (univariable and multivariable, table 3 and supplemental figure). We found a significantly higher beta for NTproBNP levels in Malays than in Caucasians (only in the multivariable model, β 0.132 vs. β 0.020 p=0.032), indicating a steeper increase in SYNTAX score with every 100-unit increase of NTproBNP in Malays than in Caucasians. Also, we found a significantly higher beta, in both the univariable (β 1.517 vs. β 0.101, p=0.002) and the multivariable

Figure 2 ROC curves with area under the curves of biomarker levels for high SYNTAX score (≥18 points) ROC curves and areas under the curves (AUCs) of biomarker level performance for each ethnic group. No significant differences between the AUCs were found for any of the biomarkers. The diagonal line represents an AUC of 0.5. The further a line deviates to the upper left, the better the discriminating properties (higher sensitivity and specificity) for a SYNTAX score of ≥18 points.

120


Ethnicity-specific Association of Biomarkers with CAD Severity

(β 1.146 vs. β 0.016, p=0.017) model, for MPO levels in Malays than in Caucasians, indicating that with 10-unit increase of MPO, the SYNTAX score increases more steeply in Malays than in Caucasians. For hsTnI levels (per 100-unit increase) we found a lower beta for Malays than for Caucasians (β 0.003 vs. β 0.044, p=0.010), however this difference was abolished when adjusting for baseline differences. Biomarker discrimination of high SYNTAX score From ROC analysis (Figure 2) we determined ethnicity-specific optimal biomarker cut-off values for the association with a SYNTAX score of ≥18. These cut-offs (Table 3) correspond to the biomarker level at which the sum of sensitivity + specificity is largest. We found markedly different biomarker cut-off values across the ethnic groups. The optimal NTproBNP cut-off for Indians, for example, was a 5-fold higher with 255.77 pmol/L than for Caucasians at 48.42 pmol/L. For hsCRP the cut-off was highest among Indians (8.95 μg/mL) and lowest in Chinese (1.69 μg/mL). The cut-offs for CysC and MPO were in the same order of magnitude among the ethnic groups. The optimal cut-off for hsTnI was strikingly higher for Caucasians than the other ethnic groups (116.65 ng/L vs. 60.75 in Chinese, 6.65 in Indians and 17.15 in Malays). This indicates that especially in Indians and Malays, lower levels of hsTnI concur with more severe CAD.

Discussion We observed inter-ethnic differences in biomarkers related to CAD. Also, the optimal cut-off levels at which these biomarkers offered discrimination of severe CAD varied substantially between ethnicities. Differences in NTproBNP levels in the general population have mainly been reported between Blacks and Caucasians, with lower levels reported for Blacks.24 In our cohort we find the highest fully adjusted levels for Malays and very high cut-off levels were found for Indians and Malays, indicating that in those ethnic groups NTproBNP levels correspond to less severe CAD than in Chinese and Caucasians. HsCRP levels are known to differ markedly among the ethnic groups in the general population with very low levels reported for Chinese and Japanese people.5,25 In a Singaporean cohort of individuals visiting the hospital for regular health checks, higher hsCRP levels were found in Indians as compared to Chinese, also when adjusted for confounders.26 Notably, within the general US population, hsCRP levels were not able to differentiate between US Chinese with and without a future cardiovascular event, while some predictive power was observed for Caucasians, indicating different discriminating properties.25 Similar to the general population cohorts, high adjusted hsCRP levels were found for Indians and Malays in the current study of CAD patients. In contrast to the general population cohorts however, we also found higher hsCRP levels in Chinese than in Caucasians in our cohort (albeit not significant after adjustment for confounders).

121


122

hsCRP (μg/mL)

NTproBNP (pmol/L)

CAD severity

hsTnI (ng/L)

MPO (ng/mL)

CysC (μg/mL)

hsCRP (μg/mL)

NTproBNP (pmol/L)

Indication

Biomarker

39.48 [7.31, 159.11]

Double vessel disease

1.20 [0.55, 3.11] 1.41 [0.62, 3.47] 1.47 [0.57, 3.84] 2.51 [0.80, 5.26]

No/minor CAD

Single vessel disease

Double vessel disease

Triple vessel disease

68.93 [22.53, 164.18]

40.3 [11.75, 123.41]

Triple vessel disease

39.83 [8.64, 132.79]

Single vessel disease

128.55 [26.43, 944.68]

No/minor CAD

STEMI

5.60 [3.30, 12.03] 26.10 [7.00, 283.1]

UA/NSTEMI

137.68 [62.42, 210.34]

Stable CAD

STEMI

24.93 [19.29, 33.05] 32.86 [22.94, 55.28]

0.75 [0.59, 0.90]

STEMI

UA/NSTEMI

0.83 [0.65, 1.07]

Stable CAD

0.84 [0.66, 1.08]

1.95 [0.60, 5.20]

STEMI

UA/NSTEMI

2.11 [0.74, 5.88]

UA/NSTEMI

Stable CAD

1.35 [0.55, 3.10]

Stable CAD

21.64 [3.00, 77.01]

UA/NSTEMI

STEMI

42.7 [13.41, 133.96] 48.86 [13.29, 162.27]

Stable CAD

Caucasian

1.89 [0.76, 8.28]

1.51 [0.60, 5.58]

1.05 [0.48, 3.01]

1.10 [0.51, 3.30]

70.98 [24.12, 233.65]

44.86 [19.34, 155.28]

27.39 [7.72, 68.81]

24.48 [3.00, 80.88]

78.70 [20.10, 2750.00]

102.65 [14.80, 1153.83]

5.60 [3.50, 11.50]

35.39 [28.93, 37.93]

30.74 [23.50, 41.80]

25.21 [19.41, 33.61]

0.76 [0.64, 0.87]

0.79 [0.65, 1.02]

0.70 [0.58, 0.87]

9.42 [1.39, 19.96]

3.40 [0.80, 11.36]

1.01 [0.46, 2.58]

122.87 [55.78, 446.12]

77.99 [25.47, 257.85]

25.71 [4.04, 60.78]

Chinese

36.35 [7.15, 229.38]

4.70 [3.50, 13.5]

27.89 [24.03, 34.51]

24.98 [20.14, 36.6]

26.8 [19.73, 32.45]

0.68 [0.62, 0.86]

0.80 [0.64, 1.05]

0.73 [0.57, 0.88]

28.46 [8.97, 61.99]

3.10 [1.31, 8.81]

1.92 [0.98, 3.74]

119.65 [103.66, 635.87]

49.85 [6.78, 255.63]

22.50 [6.95, 52.11]

Indian

2.70 [1.00, 9.81]

1.78 [0.85, 3.67]

2.58 [1.02, 4.51]

2.66 [1.31, 7.67]

61.61 [15.51, 397.78]

38.43 [16.08, 105.24]

34.16 [15.73, 84.92]

8.39 [3.00, 32.06]

4065.10 [1651.45, 21301.05]

Table 2. Biomarker levels stratified by indication for angiography and by CAD severity, for each ethnicity.

2.15 [0.90, 6.93]

2.74 [0.82, 8.29]

2.19 [1.29, 8.79]

1.93 [0.69, 6.17]

162.41 [35.96, 414.94]

66.45 [18.36, 311.30]

34.38 [5.00, 107.95]

22.71 [3.00, 97.30]

1206.25 [956.28, 2135.23]

171.30 [8.25, 1249.40]

8.30 [4.00, 18.8]

28.7 [22.35, 37.85]

35.7 [24.98, 45.22]

28.83 [22.51, 38.81]

0.55 [0.48, 0.75]

0.81 [0.59, 1.07]

0.76 [0.53, 0.98]

21.25 [12.46, 31.37]

3.65 [1.27, 10.33]

1.70 [0.70, 5.00]

274.25 [206.61, 2245.84]

70.92 [18.95, 289.93]

40.35 [15.81, 157.66]

Malay

0.822

0.235

0.001

0.002

0.079

0.315

0.351

0.002

0.005

0.002

0.034

<0.001

0.011

0.071

0.809

0.894

<0.001

0.002

0.070

<0.001

<0.001

0.041

<0.001

p-value

Chapter 6


18.30 [6.60, 228.30]

Triple vessel disease

7.20 [4.00, 54.8]

33.55 [7.88, 398.03]

12.25 [4.53, 109.1]

7.10 [3.70, 30.6] 10.50 [4.80, 39.85]

Double vessel disease

Single vessel disease

27.07 [21.13, 37.54] 29.57 [21.91, 40.13] 5.10 [3.30, 12.08]

Triple vessel disease

29.36 [21.40, 40.95]

25.34 [18.68, 33.80]

0.77 [0.63, 0.94]

0.74 [0.60, 0.95]

0.71 [0.56, 0.84]

0.70 [0.58, 0.88]

5.25 [2.80, 12.5]

27.87 [20.63, 46.29] 30.92 [22.88, 74.03]

Double vessel disease

No/minor CAD

28.01 [21.06, 42.76]

0.85 [0.69, 1.07]

Triple vessel disease

Single vessel disease

0.86 [0.68, 1.09]

Double vessel disease

26.00 [19.99, 37.21]

0.82 [0.64, 1.05]

Single vessel disease

No/minor CAD

0.80 [0.65, 1.02]

No/minor CAD

0.72 [0.62, 0.87]

29.10 [7.00, 299.30]

7.25 [4.60, 62.73]

14.75 [3.90, 123.50]

5.00 [3.10, 15.60]

25.80 [19.71, 35.89]

23.57 [20.05, 32.82]

27.53 [18.12, 37.14]

28.90 [21.14, 32.72]

0.86 [0.63, 1.19]

0.75 [0.62, 0.89]

0.68 [0.60, 0.85]

70.30 [10.50, 776.80]

13.80 [5.75, 825.80]

5.90 [3.50, 20.20]

9.15 [3.60, 19.53]

35.24 [25.66, 40.83]

28.19 [23.15, 40.79]

29.53 [24.11, 41.49]

29.33 [22.55, 38.88]

0.88 [0.60, 1.14]

0.87 [0.57, 1.08]

0.67 [0.54, 0.81]

0.63 [0.51, 0.93]

0.115

0.187

0.437

0.200

0.033

0.183

0.516

0.272

0.266

0.010

<0.001

<0.001

Medians and interquartile ranges of biomarker levels, stratified by indication for angiography and by angiographic CAD severity. P-values for comparison among the ethnic groups are from Kruskal-Wallis testing. Significant p-values are printed in bold. Abbreviations: CAD= coronary artery disease, UA= unstable angina, NSTEMI= non-ST-elevation myocardial infarction, STEMI= ST-elevation myocardial infarction, NTproBNP= N-terminal pro brain natriuretic peptide, hsCRP= high-sensitivity C-reactive protein, CysC= Cystatin C, MPO= myeloperoxidase, hsTnI= high-sensitivity troponin I.

hsTnI (ng/L)

MPO (ng/mL)

CysC (Îźg/mL)

Ethnicity-specific Association of Biomarkers with CAD Severity

123


124 0.035 ( 0.008 - 0.061) 0.022 (-0.023 - 0.067) 0.003 (-0.027 - 0.033)

Chinese

Indian

Malay

1.517 ( 0.451 - 2.583)

Malay 0.044 ( 0.026 - 0.062)

-0.077 (-1.314 - 1.160)

Indian

Caucasian

-0.008 (-0.485 - 0.469)

-0.168 (-0.923 - 0.586)

Malay

Chinese

0.022 (-0.449 - 0.492)

Indian

0.101 ( 0.015 - 0.187)

0.235 (-0.015 - 0.485)

Caucasian

0.162 (-0.015 - 0.340)

Chinese

-0.025 (-0.117 - 0.066)

Malay

Caucasian

0.009 (-0.095 - 0.112)

Indian

0.858

0.331

0.010

<0.001

0.006

0.902

0.974

0.022

0.660

0.928

0.066

0.073

0.584

0.870

0.243

0.012

0.035

0.255

0.003

0.059

p-value beta (univ.)

0.003 (-0.027 - 0.033)

0.050 (-0.004 - 0.104)

0.044 ( 0.013 - 0.074)

0.032 ( 0.012 - 0.052)

1.146 ( 0.023 - 2.269)

0.108 (-1.099 - 1.314)

-0.079 (-0.630 - 0.472)

0.016 (-0.095 - 0.127)

-0.165 (-0.935 - 0.605)

-0.172 (-0.700 - 0.356)

0.154 (-0.194 - 0.503)

0.081 (-0.131 - 0.293)

-0.040 (-0.148 - 0.069)

0.033 (-0.084 - 0.150)

0.035 (-0.025 - 0.095)

0.030 (-0.006 - 0.066)

0.132 (-0.020 - 0.284)

0.020 (-0.119 - 0.160)

0.104 (-0.013 - 0.221)

0.020 (-0.047 - 0.087)

Beta (CI, multivariable)

0.855

0.067

0.006

0.002

0.046

0.859

0.778

0.777

0.671

0.519

0.383

0.454

0.471

0.578

0.255

0.103

0.088

0.771

0.081

0.560

p-value beta (multiv.)

0.010

0.328

0.557

Ref.

0.002

0.743

0.633

Ref.

0.314

0.541

0.642

Ref.

0.113

0.505

0.663

Ref.

0.157

0.669

0.094

Ref.

p-value interaction (univ.)

0.351

0.958

0.164

Ref.

0.017

0.555

0.885

Ref.

0.457

0.753

0.371

Ref.

0.621

0.627

0.709

Ref.

0.032

0.361

0.115

Ref.

p-value interaction (multiv.)

The multivariable model adjusts for: age, sex, BMI, diabetes, hypertension, hyperlipidemia, smoking status, indication for coronary angiography, anti-platelet medication use, beta blocker use, calcium antagonist use, RAAS medication use and statin use. Interaction terms were tested in the univariable and the multivariable model. Ref.= reference group for interaction analysis. * Betas are displayed for a 100-unit increase in biomarker level, ** betas are given for a 10-unit increase in MPO levels. Abbreviations: CI= 95% confidence interval, NTproBNP= N-terminal pro brain natriuretic peptide, hsCRP= high-sensitivity C-reactive protein, CysC= Cystatin C, MPO= myeloperoxidase, hsTnI= high-sensitivity troponin I. Significant p-values are printed in bold.

hsTnI (ng/L)*

MPO (ng/mL)**

CysC (Îźg/mL)

0.029 (-0.020 - 0.077)

Chinese

0.152 ( 0.011 - 0.292)

Malay 0.041 ( 0.009 - 0.074)

0.088 (-0.065 - 0.241)

Indian

Caucasian

0.153 ( 0.053 - 0.254)

Chinese

hsCRP (Îźg/mL)

0.056 (-0.002 - 0.114)

Caucasian

NTproBNP (pmol/L)*

Beta (CI, univariable)

Ethnicity

Biomarker

Table 3. Regression coefficients (betas with 95% confidence intervals) of biomarker levels for SYNTAX score.

Chapter 6


Ethnicity-specific Association of Biomarkers with CAD Severity

While cut-offs for Caucasians, Chinese and Malays were quite comparable, the hsCRP cut-off for severe CAD in Indians was strikingly higher. Suggesting that high hsCRP levels correspond to less severe CAD in Indians than in the other ethnic groups. In our study we found no inter-ethnic differences in CysC levels among CAD patients whilst inter-ethnic differences have been described in the general population, with lower levels in Blacks than in Whites.27 To our knowledge, no study has evaluated the ethnicityspecific association of CysC with CAD. The Multi-ethnic Study of Atherosclerosis (MESA)28 has evaluated the association of CysC-estimated glomerular filtration rate with the incidence of coronary artery calcium and found a strong association, unfortunately no interactions of this association with ethnicity were tested. In another MESA sub-study29 however, the association of CysC levels with known biomarkers of CAD (CRP, interleukin-6, intercellular adhesion molecule-I and factor VIII) differed significantly by ethnicity. This leaves the ethnicity-specific role of CysC in relation to CAD severity unresolved. Table 4. Results from ROC analysis of biomarker levels for high SYNTAX score (≥18 points). Biomarker

Ethnicity

Optimal Cut-off

AUC (95% CI)

NTproBNP (pmol/L)

Caucasian

48.42

0.594 (0.543 - 0.645)

Chinese

120.92

0.596 (0.532 - 0.659)

Indian

255.77

0.584 (0.455 - 0.712)

Malay

154.41

0.621 (0.515 - 0.726)

Caucasian

2.40

0.589 (0.538 - 0.641)

Chinese

1.69

0.547 (0.481 - 0.613)

Indian

8.95

0.550 (0.425 - 0.675)

Malay

3.04

0.502 (0.395 - 0.608)

Caucasian

0.74

0.555 (0.501 - 0.609)

Chinese

0.86

0.527 (0.460 - 0.594)

Indian

0.85

0.602 (0.475 - 0.728)

Malay

0.60

0.571 (0.466 - 0.677)

Caucasian

30.17

0.601 (0.549 - 0.653)

Chinese

36.84

0.532 (0.467 - 0.598)

Indian

22.11

0.525 (0.405 - 0.644)

Malay

33.40

0.611 (0.508 - 0.715)

Caucasian

116.65

0.618 (0.566 - 0.671)

Chinese

60.75

0.605 (0.540 - 0.669)

Indian

6.65

0.614 (0.499 - 0.729)

Malay

17.15

0.711 (0.619 - 0.804)

hsCRP (μg/mL)

CysC (μg/mL)

MPO (ng/mL)

hsTnI (ng/L)

Results from ROC analysis, stratified by ethnicity. The optimal cut-off corresponds to the biomarker level at which the largest sum of sensitivity + specificity is found. The AUC is presented with its 95% confidence interval. When the confidence interval does not contain 0.5 (printed in bold), the cut-off of that biomarker is significantly associated with a SYNTAX score ≥18 points. Abbreviations: ROC= receiver operating characteristics, AUC= area under the curve, NTproBNP= N-terminal brain natriuretic peptide, hsCRP= high-sensitivity C-reactive protein, CysC= Cystatin C, MPO= myeloperoxidase, hsTnI= high-sensitivity troponin I.

125


Chapter 6

MPO levels have been reported to be related to coronary atherosclerosis in Blacks, but not in whites or Hispanics30, demonstrating an important modifying role of Black ethnicity on the association of MPO levels with CAD. No previous comparison of the association of MPO with CAD severity has been made between Asians and Caucasians. In our study we found a significantly stronger association of MPO levels with SYNTAX score among Malays than among Caucasians. However, comparable cut-off levels for all ethnic groups were calculated. The association of MPO with severity of CAD among Malay patients deserves the attention of further research. A striking difference was found for hsTnI levels, which were much higher in Malays (multivariably adjusted) than in Caucasians. Population levels have been reported to be comparable among Chinese, Indians and Malays in a Malaysian study31, thus the high levels we observe in Malays can probably not be explained by higher baseline levels. hsTnI levels can be elevated without actual myocardial necrosis.32 Experimental data suggest that hsTnI can be released from cardiomyocytes by release of proteolytic degradation products, which can already occur with mild ischemia. Mechanical stretch and ischemia have also been demonstrated to increase cellular wall permeability, leading to leakage of troponins from the cytosol. Lastly, cardiomyocyte-derived vesicles, shed upon ischemia might be involved in the release of hsTnI without cardiomyocyte necrosis. These phenomena might explain the high hsTnI levels in stable Malay patients, in the absence of necrosis (table 2). Cut-off levels of hsTnI for high SYNTAX score were much higher in Caucasians than in the other ethnic groups, indicating that cardiomyocyte damage is greater in Caucasians than in the other ethnic groups at a SYNTAX score of 18 points. One of the explanations for this feature could be preconditioning occurring in more severe, unstable CAD resulting in resilient myocardium releasing less troponin upon prolonged ischemia33. The Caucasian ethnic group most often presented with stable CAD, thus cardiomyocyte preconditioning would have occurred least in this group. In summary, our results suggest that inter-ethnic differences in biomarker levels exist and that the association of these biomarkers with the extent of disease differs by ethnicity. Prior to implementation of biomarkers into clinical practice, their performance should be evaluated in an ethnicity-specific manner.34 Limitations Biomarker levels in our study were measured in arterial blood; these levels may differ from venous levels. However, Martin et al. showed that arterial and venous levels of hsCRP and NTproBNP did not differ.35 Also, levels of NTproBNP, hsCRP, CysC and MPO were measured using an in-house ELISA method instead of clinically standardized assays. Therefore, generalizability of the biomarker levels described in our cohort is limited. Future studies should focus on examining inter-ethnic differences in biomarker levels measured in venous blood, measured by clinically validated assays. Although follow-up data for allcause death was available, an ethnicity-specific mortality analysis was not possible due to small numbers of fatalities in certain ethnic groups (e.g. n=3 in the Indian ethnic group).

126


Ethnicity-specific Association of Biomarkers with CAD Severity

Follow-up data on other adverse cardiovascular events were unfortunately not available. We have adjusted our results for baseline differences among the ethnic groups as much as possible. However, profound differences were observed at baseline, which might not have been completely abolished after adjustment. In larger multi-ethnic cohorts propensity-matched36 analysis might be considered in order to further eliminate confounding. Also, details on nutrition, lifestyle and socioeconomic factors were not available in this cohort; therefore we could not consider these (possibly confounding) factors in the analyses. Conclusion In a patient group undergoing coronary angiography we found inter-ethnic differences in levels of biomarkers, related to CAD. These differences persisted after correcting for baseline differences among ethnicities. Furthermore, certain biomarkers displayed interethnic differences in the relationship of biomarkers to severity of CAD, with cut-off values from ROC analyses varying over a five-fold range between ethnicities. Future research should focus on ethnicity-specific cut-off values of established CAD biomarkers. Acknowledgements We are grateful for the excellent contributions of ms. Jonne Hos (Utrecht) and mrs. Fauziah Azizi (Singapore) to the data gathering for the UNICORN project. Funding This work was supported by a Strategic grant from the Royal Netherlands Academy of Arts and Sciences to the Interuniversity Cardiology Institute of the Netherlands, ICIN (DdK); National University Singapore Startup grant to DdK, Singapore National Medical Research Council Centre Grant to MR, MC and DdK and the ATTRaCT, SPF 2014/003 grant BMRC to DdK and MR. These funding sources in no way influenced the analyses or the content of this manuscript.

127


Chapter 6

References 1. 2.

Alwan A. Global status report on noncommunicable diseases 2010. 1st ed. Geneva, Switzerland: 2011. Anand SS, Yusuf S, Vuksan V, Devanesen S, Teo KK, Montague PA, Kelemen L, Yi C, Lonn E, Gerstein H, Hegele RA, McQueen M. Differences in risk factors, atherosclerosis, and cardiovascular disease between ethnic groups in Canada: the Study of Health Assessment and Risk in Ethnic groups (SHARE). Lancet. 2000;356:279–84.

3.

Koulaouzidis G, Nicoll R, Charisopoulou D, McArthur T, Jenkins PJ, Henein MY. Aggressive and diffuse coronary calcification in South Asian angina patients compared to Caucasians with similar risk factors. Int J Cardiol. 2013;167:2472–6.

4.

Bild DE, Detrano R, Peterson D, Guerci A, Liu K, Shahar E, Ouyang P, Jackson S, Saad MF. Ethnic differences in coronary calcification: the Multi-Ethnic Study of Atherosclerosis (MESA). Circulation. 2005;111:1313–20.

5.

Gijsberts CM, den Ruijter HM, Asselbergs FW, Chan MY, de Kleijn DP V, Hoefer IE. Biomarkers of Coronary Artery Disease Differ Between Asians and Caucasians in the General Population. Glob Heart. 2015;

6.

Krenning BJ, Van der Heiden K. Should ethnicity be included in cardiovascular risk stratification? Netherlands Hear J. 2015;23:42–43.

7.

World Health Organization. Health systems: Improving performance. In: The world health report 2000. Geneva, Switzerland: 2000.

8.

Nishikimi T, Mori Y, Ishimura K, Tadokoro K, Yagi H, Yabe A, Horinaka S, Matsuoka H. Association of plasma atrial natriuretic peptide, n-terminal proatrial natriuretic peptide, and brain natriuretic peptide levels with coronary artery stenosis in patients with normal left ventricular systolic function. Am J Med. 2004;116:517–523.

9.

Sadanandan S, Cannon CP, Chekuri K, Murphy SA, Dibattiste PM, Morrow DA, De Lemos JA, Braunwald E, Gibson CM. Association of elevated B-type natriuretic peptide levels with angiographic findings among patients with unstable angina and non-ST-segment elevation myocardial infarction. J Am Coll Cardiol. 2004;44:564–568.

10. Weber M, Dill T, Arnold R, Rau M, Ekinci O, Müller KD, Berkovitsch A, Mitrovic V, Hamm C. N-terminal B-type natriuretic peptide predicts extent of coronary artery disease and ischemia in patients with stable angina pectoris. Am Heart J. 2004;148:612–620. 11. Arroyo-Espliguero R, Avanzas P, Quiles J, Kaski JC. Predictive value of coronary artery stenoses and C-reactive protein levels in patients with stable coronary artery disease. Atherosclerosis. 2009;204:239–43. 12. Imai A, Komatsu S, Ohara T, Kamata T, Yoshida J, Miyaji K, Shimizu Y, Takewa M, Hirayama A, Deshpande GA, Takahashi O, Kodama K. Serum cystatin C is associated with early stage coronary atherosclerotic plaque morphology on multidetector computed tomography. Atherosclerosis. 2011;218:350–5. 13. Niccoli G, Conte M, Della Bona R, Altamura L, Siviglia M, Dato I, Ferrante G, Leone AM, Porto I, Burzotta F, Brugaletta S, Biasucci LM, Crea F. Cystatin C is associated with an increased coronary atherosclerotic burden and a stable plaque phenotype in patients with ischemic heart disease and normal glomerular filtration rate. Atherosclerosis. 2008;198:373–80. 14. Baldus S, Heeschen C, Meinertz T, Zeiher AM, Eiserich JP, Münzel T, Simoons ML, Hamm CW. Myeloperoxidase serum levels predict risk in patients with acute coronary syndromes. Circulation. 2003;108:1440–5. 15. LaFramboise WA, Dhir R, Kelly LA, Petrosko P, Krill-Burger JM, Sciulli CM, Lyons-Weiler MA, Chandran UR, Lomakin A, Masterson R V, Marroquin OC, Mulukutla SR, McNamara DM. Serum protein profiles predict coronary artery disease in symptomatic patients referred for coronary angiography. BMC Med. 2012;10:157.

128


Ethnicity-specific Association of Biomarkers with CAD Severity

16. Ndrepepa G, Braun S, Schulz S, Mehilli J, Schömig A, Kastrati A. High-sensitivity troponin T level and angiographic severity of coronary artery disease. Am J Cardiol. 2011;108:639–43. 17. Saunders JT, Nambi V, de Lemos JA, Chambless LE, Virani SS, Boerwinkle E, Hoogeveen RC, Liu X, Astor BC, Mosley TH, Folsom AR, Heiss G, Coresh J, Ballantyne CM. Cardiac troponin T measured by a highly sensitive assay predicts coronary heart disease, heart failure, and mortality in the Atherosclerosis Risk in Communities Study. Circulation. 2011;123:1367–76. 18. Sianos G, Morel M, Kappetein AP, Morice M, Colombo A, Dawkins K, van den Brand M, Van Dyck N, Russell ME, Mohr FW, Serruys PW. The SYNTAX Score: an angiographic tool grading the complexity of coronary artery disease. EuroIntervention. 2005;1:219–27. 19. Harris PJ, Behar VS, Conley MJ, Harrell FE, Lee KL, Peter RH, Kong Y, Rosati R a. The prognostic significance of 50% coronary stenosis in medically treated patients with coronary artery disease. Circulation. 1980;62:240–248. 20. Gijsberts CM, Gohar A, Ellenbroek GHJM, Hoefer IE, de Kleijn DPV, Asselbergs FW, Nathoe HM, Agostoni P, Rittersma SZH, Pasterkamp G, Appelman Y, den Ruijter HM. Severity of stable coronary artery disease and its biomarkers differ between men and women undergoing angiography. Atherosclerosis. 2015;241:234–240. 21. Robin X, Turck N, Hainard A, Tiberti N, Lisacek F, Sanchez J-C, Müller M. pROC: an open-source package for R and S+ to analyze and compare ROC curves. BMC Bioinformatics. 2011;12:77. 22. Capodanno D, Di Salvo ME, Cincotta G, Miano M, Tamburino C, Tamburino C. Usefulness of the SYNTAX score for predicting clinical outcome after percutaneous coronary intervention of unprotected left main coronary artery disease. Circ Cardiovasc Interv. 2009;2:302–308. 23. R Core Team. R: A Language and Environment for Statistical Computing. Vienna, Austria: 2014. 24. Abdullah SM, Khera A, Das SR, Stanek HG, Canham RM, Chung AK, Morrow DA, Drazner MH, McGuire DK, de Lemos JA. Relation of coronary atherosclerosis determined by electron beam computed tomography and plasma levels of n-terminal pro-brain natriuretic peptide in a multiethnic population-based sample (the Dallas Heart Study). Am J Cardiol. 2005;96:1284–9. 25. Veeranna V, Zalawadiya SK, Niraj A, Kumar A, Ference B, Afonso L. Association of novel biomarkers with future cardiovascular events is influenced by ethnicity: results from a multi-ethnic cohort. Int J Cardiol. 2013;166:487–93. 26. Dalan R, Jong M, Chan S-P, Hawkins R, Choo R, Lim B, Tan ML, Leow MK. High-sensitivity C-reactive protein concentrations among patients with and without diabetes in a multiethnic population of Singapore: CREDENCE Study. Diabetes Metab Syndr Obes. 2010;3:187–95. 27. Groesbeck D, Köttgen A, Parekh R, Selvin E, Schwartz GJ, Coresh J, Furth S. Age, gender, and race effects on cystatin C levels in US adolescents. Clin J Am Soc Nephrol. 2008;3:1777–85. 28. Lamprea-Montealegre J a, McClelland RL, Astor BC, Matsushita K, Shlipak M, de Boer IH, Szklo M. Chronic kidney disease, plasma lipoproteins, and coronary artery calcium incidence: the Multi-Ethnic Study of Atherosclerosis. Arterioscler Thromb Vasc Biol. 2013;33:652–8. 29. Keller C, Katz R, Cushman M, Fried LF, Shlipak M. Association of kidney function with inflammatory and procoagulant markers in a diverse cohort: a cross-sectional analysis from the Multi-Ethnic Study of Atherosclerosis (MESA). BMC Nephrol. 2008;9:9. 30. Chen LQ, Rohatgi A, Ayers CR, Das SR, Khera A, Berry JD, McGuire DK, de Lemos JA. Race-specific associations of myeloperoxidase with atherosclerosis in a population-based sample: the Dallas Heart Study. Atherosclerosis. 2011;219:833–8.

129


Chapter 6

31. Sthaneshwar P, Jamaluddin FA, Fan YS. Reference value for cardiac troponin I in a multi-ethnic group. Pathology. 2010;42:454–456. 32. White HD. Pathobiology of Troponin Elevations**Editorials published in the Journal of the American College of Cardiology reflect the views of the authors and do not necessarily represent the views of JACC or the American College of Cardiology. J Am Coll Cardiol. 2011;57:2406–2408. 33. Jenkins DP, Pugsley WB, Alkhulaifi AM, Kemp M, Hooper J, Yellon DM. Ischaemic preconditioning reduces troponin T release in patients undergoing coronary artery bypass surgery. Heart. 1997;77:314–8. 34. Erqou S, Kip KE, Mulukutla SR, Aiyer a. N, Reis SE. Racial differences in the burden of coronary artery calcium and carotid intima media thickness between Blacks and Whites. Netherlands Hear J. 2015;23:44–51. 35. Martin J, Smith J, Bastien M, Cianflone K, Bussières J, Marceau S, Hould FS, Lebel S, Biertho L, Lescelleur O, Biron S, Bertrand F, Poirier P. Comparison between arterial and venous sampling of circulating hormones, substrates and peptides in severe obesity. Clin Investig Med. 2011;34. 36. Austin PC. An Introduction to Propensity Score Methods for Reducing the Effects of Confounding in Observational Studies. Multivariate Behav Res. 2011;46:399–424.

130


Ethnicity-specific Association of Biomarkers with CAD Severity

Supplemental 50

20

10

0

● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●

● ●● ● ●

● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ●● ●● ●● ● ● ● ● ● ● ●● ●● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ●● ● ●● ●●● ● ●● ● ● ● ● ● ● ●● ● ●● ● ● ● ●● ● ● ●●● ● ● ● ● ● ●● ● ●● ● ●●● ● ●● ● ● ● ● ●● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ●● ● ● ●● ● ● ●●●●●●● ● ● ● ●● ●● ● ●● ● ● ● ● ●● ● ●● ● ● ●● ● ● ●● ● ● ● ●● ● ●● ● ● ●● ● ● ●● ● ● ●● ● ● ● ● ● ● ● ● ●●● ● ●● ●● ● ● ● ● ●● ●●● ● ● ● ● ●● ● ●● ●● ● ● ● ● ● ● ●● ● ●●●● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ●● ●● ● ● ●● ● ● ● ● ● ●● ● ●● ● ●● ● ● ●● ● ● ●●● ● ● ●● ●● ● ●● ●●●● ●● ● ●● ●● ●● ● ● ● ● ●● ●●●●● ●● ● ● ●●● ● ●●● ● ● ● ● ● ●● ● ● ● ● ● ● ● ●● ● ● ● ●● ● ● ● ● ● ●●● ● ● ● ● ● ●● ● ●● ● ● ●●● ● ●● ● ● ● ● ● ●●● ●● ● ● ●● ● ●● ● ●● ●●●● ● ●●● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ●● ● ● ● ●● ● ●● ● ● ● ● ●● ● ● ● ●● ●● ● ●●● ●● ● ● ● ●● ● ●●● ●● ● ● ●●●● ● ●●●● ● ●● ●●●●● ●● ● ●● ●● ●●●●● ● ● ● ●●● ●● ● ●● ● ● ●●● ● ●● ●● ●●●● ● ●●● ● ●●● ● ● ● ● ● ● ●● ● ● ●● ●● ● ● ●●●●●● ●● ●● ●● ● ● ● ● ●● ●● ● ● ●● ● ● ● ● ● ●●● ● ● ● ● ● ● ●● ● ●●● ● ● ●●● ● ●●● ● ●● ●● ●● ●●●● ●● ● ●● ●● ●● ●● ●● ●●● ● ● ● ●●● ● ● ● ● ● ●● ● ● ● ● ● ●●● ●●● ●● ●● ●●●● ● ● ● ● ●● ●●●● ● ●●● ● ●● ● ● ●● ● ● ●● ● ● ● ● ● ● ● ● ● ●● ●●●● ● ● ●● ●● ● ●● ● ● ●●● ●●● ● ●● ● ●●●● ● ●●● ● ● ● ● ● ● ●●● ● ● ● ●● ●● ● ● ●● ● ●●● ●● ● ● ●● ●● ● ● ● ●●●● ●● ●● ●●● ● ●● ● ● ● ● ●●●● ● ● ●●● ● ●●● ● ●● ● ● ●●●● ● ● ● ●● ●●●●●●● ●● ● ● ● ●● ● ● ●●●● ●● ● ●●●● ● ● ● ● ●● ●● ● ●● ●● ●●● ● ● ● ●● ● ● ●● ●● ●● ● ● ●●● ● ● ●● ● ● ●● ●●● ●● ● ●●●● ●● ● ● ● ● ● ●●● ●● ● ●● ● ● ● ● ● ● ● ● ● ● ●●● ●● ●● ● ● ●●● ●● ● ● ● ● ● ● ● ●● ● ● ●

100

● ● ● ● ● ● ●

● ●

● ● ● ● ● ● ● ● ●

● ●

20 ●

●●

10 ● ●

0

SYNTAX Score

30

20

10

0

10

● ●

● ●

0

● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●

100

● ●

● ● ● ● ●● ● ● ● ● ● ● ●● ●● ● ●● ● ● ● ● ●●● ● ● ● ●● ● ● ● ● ● ● ●●●●●●● ● ● ● ●● ●● ●● ●● ● ● ●●● ● ●● ● ●●●● ● ● ● ●● ● ●● ●●● ● ● ●● ●● ●● ● ●● ● ● ● ● ● ● ● ●●● ● ●●● ● ● ● ● ● ● ● ● ● ● ●● ●● ●● ● ● ● ● ● ● ● ● ●● ● ● ● ● ●● ●● ● ● ● ●●● ●● ● ●● ● ●● ● ●●● ● ● ● ●● ●●● ● ● ● ●● ● ● ● ●●●● ● ● ●● ●● ●● ● ●● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ●● ●● ● ● ● ●●●●● ● ●●● ● ●● ● ●● ● ● ●● ●● ● ● ● ●●●● ●●●● ●● ●●● ● ●● ● ● ● ● ● ● ● ●● ●● ●● ●● ● ●●● ●● ●● ● ●●● ● ● ●● ●● ● ●● ● ● ●● ● ●● ● ● ●● ● ● ●●●● ●● ● ●● ●● ● ● ● ●● ● ● ●● ● ●●● ● ● ● ● ●●● ● ● ●●●● ● ● ●● ● ● ●● ● ●● ●● ● ●● ● ● ● ● ●● ● ● ●●● ● ●● ●●● ● ●● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ●● ● ● ● ● ● ●●● ●●●● ● ● ● ● ● ● ● ● ● ● ● ● ●●● ● ● ● ●● ●● ● ● ●● ● ● ● ● ●● ● ● ●● ● ● ●● ● ● ●● ● ● ● ●●● ● ●● ● ●●● ●● ● ● ● ● ● ● ● ●● ●●● ●● ●● ●● ●●●●●● ● ● ● ●● ● ●● ● ●● ● ● ● ● ●●● ● ● ●● ● ●● ● ●●● ●● ●● ● ● ● ● ● ●● ● ●● ●●● ●● ● ● ●● ● ●● ● ● ●● ● ●● ● ● ● ●● ● ●●● ● ●● ● ● ● ● ●● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ●● ● ●● ● ●● ● ● ●● ●● ●●● ● ● ●● ●● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●● ● ●●● ● ● ● ●●● ● ● ● ● ●● ● ● ● ● ● ●● ● ●● ● ● ●●● ● ● ● ● ●● ● ●●● ● ● ● ● ● ● ●● ● ●●●● ● ● ●● ●● ● ● ● ● ● ● ● ● ● ● ● ●● ● ●● ● ● ● ● ●● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ●● ● ● ●● ●●● ●● ● ●●● ●● ●● ● ● ●● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●● ● ● ● ● ●● ● ●● ● ● ● ●● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ●●● ● ● ●● ● ●●● ● ● ● ● ● ● ● ● ● ● ● ●●● ● ●● ● ● ●● ● ● ● ● ● ● ● ●● ● ● ●● ● ●● ● ● ● ●●● ● ●● ● ● ● ●● ● ● ● ● ● ●● ● ●● ● ● ●●● ●● ● ● ●● ●● ● ● ● ● ●● ● ● ● ●● ●● ● ● ● ●●●●● ● ● ● ● ●● ● ● ● ●● ●● ● ● ● ● ●● ● ●●● ● ● ●● ● ●● ● ● ●● ● ● ●●● ●●●● ● ● ●● ●● ● ● ● ● ● ●● ● ●● ●● ●● ● ● ●●● ●●● ● ●● ● ●● ● ● ● ● ● ● ● ●● ● ● ●

● ● ●

● ● ●

● ● ● ●

● ●

● ● ● ● ● ●

40

30

20 ●

10

0

1

10

MPO

100

● ●

● ●

● ●

● ●

● ●

● ●

●● ●

Caucasian

Chinese

Indian

Malay

● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ●● ● ● ●● ● ● ●● ● ● ● ●● ● ● ● ● ● ● ● ● ●●● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ●● ● ● ● ●● ● ● ●● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ●● ●● ● ●● ● ●●●● ●●● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ●● ●● ● ●● ●●● ● ● ●● ●● ● ● ● ●● ●● ● ● ● ●●● ●● ● ● ●● ● ●● ● ●●●●● ● ● ●● ●●● ● ● ● ●● ● ●● ●● ● ● ● ●● ● ● ● ●● ● ● ● ● ●●● ● ● ●● ●●● ● ● ● ● ● ● ● ●●● ●● ●● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ●● ● ● ● ●● ● ● ● ● ●●●● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ●● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ●● ● ● ● ●●● ● ●● ● ●● ● ● ●● ●● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ●●● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●● ●● ● ● ● ● ● ● ●● ●● ● ● ●● ●●● ●●● ● ● ● ●● ● ● ●● ● ●● ● ●● ● ●●● ● ●●● ● ● ● ● ●● ● ● ● ●● ●●● ● ● ● ● ● ●● ● ● ●● ●● ● ● ● ● ● ●● ● ● ●●● ● ● ● ● ●● ● ●● ● ● ● ●● ● ●● ●● ●● ● ● ●● ● ● ● ● ● ● ●● ●● ● ● ● ●● ● ● ●● ● ● ● ● ● ●● ● ● ● ●● ● ● ● ● ● ●● ●● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ●●● ● ●● ● ● ●● ●● ●● ● ● ●● ● ●● ● ● ● ● ● ●●● ●● ● ● ●●● ●● ●● ● ● ●● ● ● ● ●● ● ● ● ●●● ● ● ● ● ●●● ●● ●● ●● ● ●●●● ●● ● ● ●● ● ●●● ●●●● ● ● ●● ● ● ● ● ●●● ● ●● ●● ●●● ● ●● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ●●● ● ●● ● ● ● ●● ● ●●● ●● ●● ● ●●● ●● ● ● ● ● ● ● ● ● ● ●● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ●● ●● ● ●●●●● ●● ●●●● ● ● ●● ● ● ●● ● ●● ●● ●●● ● ● ● ● ● ● ●● ● ●● ●● ● ● ● ● ● ● ● ●● ● ● ● ●●● ● ● ● ● ●●●●● ● ● ●● ● ●● ●●●● ●● ● ● ● ● ● ●● ●● ●● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ●● ● ● ● ● ●● ●●● ● ●● ● ● ● ●● ● ●●●●● ●●●●● ●● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ●●● ● ●● ●●●● ● ● ● ● ● ●● ●●●●● ● ● ● ● ●● ● ●●● ●● ● ● ● ●● ● ● ● ●● ●●●● ● ● ● ● ● ● ● ●● ● ● ●● ●●●● ● ●● ● ●● ● ● ● ●●●● ● ●● ●●● ● ● ● ●●● ● ●● ● ● ●● ●● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ●● ●● ● ● ● ●● ● ● ● ●● ●●● ● ● ●● ● ● ●● ● ● ●● ● ● ● ● ● ●● ● ●● ● ●● ●●● ● ● ● ● ● ● ●● ●● ●● ● ●● ●● ● ● ● ●●● ● ● ● ●● ● ●●●● ● ● ● ●● ●●● ●● ● ●●● ●

1

● ● ●●

● ● ● ●● ●

● ●

20

10.0

● ●

40

● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ●● ● ● ● ●● ● ●● ● ● ●● ● ● ● ● ● ●● ● ●● ● ●● ● ●●● ● ● ● ● ● ● ● ● ● ● ●●● ● ●● ● ● ● ● ● ● ● ● ● ●●● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ●● ● ● ●● ●● ● ● ● ●●● ● ●● ● ●● ● ● ●● ● ●● ● ● ● ● ●● ● ●● ● ● ●● ● ● ● ● ●●● ●● ● ● ●● ●● ● ● ● ● ● ● ●● ● ●●●● ●● ●●●● ●● ● ● ● ● ● ● ● ● ●● ● ●●●●● ● ● ● ●● ●●● ●●● ● ● ●● ● ●● ● ● ● ● ● ●● ● ● ●● ●● ● ●●●●●● ● ● ● ● ●● ● ●● ● ●●● ●● ●● ● ●● ● ● ●● ●● ● ● ●● ●●●● ● ● ● ●● ● ● ● ● ● ● ●● ●●● ● ● ●●●● ●●●● ● ● ●●●● ● ● ●● ● ● ● ● ● ● ● ●● ●● ● ● ●● ● ●● ● ●● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ●● ● ●●● ● ●●● ●● ●● ●● ● ●●●● ●●● ●● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ●●● ● ● ● ● ● ● ● ● ● ● ●●●● ● ●●● ●● ●●● ●● ● ●● ● ● ● ● ● ● ●● ● ● ●● ● ●●● ●● ● ●● ●●● ● ● ● ● ● ● ● ●● ● ●●●● ● ● ● ●●●● ●●●● ● ●● ●● ●● ●● ●●●● ● ● ● ● ● ● ● ●●●● ● ● ● ● ● ● ● ● ● ● ● ● ●● ●● ●● ●● ● ● ● ● ●● ●●● ●●● ●● ● ●● ● ● ● ●● ●● ●● ●●●● ● ●● ● ●● ● ● ● ●●● ●● ●● ● ● ● ● ● ● ●● ● ● ● ● ● ●● ● ● ●● ●●●●● ● ● ●● ● ● ● ●●● ● ●●● ● ●●● ●●● ● ● ● ● ● ● ●●● ● ●●● ● ● ● ● ●● ●● ● ●● ●● ●●●●● ● ● ● ●● ● ●● ● ●● ● ● ●●● ● ● ● ● ● ● ● ●●● ● ● ●● ● ● ● ● ● ● ● ● ● ● ●● ● ●●● ●●●●● ● ●● ● ●● ●●● ● ●●●● ●● ● ● ● ● ●●● ●● ●● ●● ● ●●● ● ●● ●● ●● ● ●● ● ● ● ● ● ●● ● ● ●● ●● ●● ●● ● ● ●●● ● ●● ●● ● ●● ●● ●● ●● ●● ● ● ● ●●● ●● ●● ● ● ● ● ● ● ● ● ● ● ●● ● ●●●● ●● ●● ● ●●● ●● ● ● ● ●● ● ● ●● ● ●● ● ● ●●●● ● ● ● ● ●● ● ● ●● ● ●● ● ● ●● ●● ●● ●● ●● ● ●●● ●●●●● ● ●●● ● ●●●● ● ● ●● ● ● ● ●● ● ●● ●● ● ●●● ● ●● ●● ●●●● ●● ● ● ● ● ● ●●● ● ● ● ● ● ●●●●● ●● ● ● ● ● ● ●● ● ● ● ●● ● ● ● ●●● ●● ● ● ● ● ●●● ● ●●●●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ●●●● ●● ● ●● ●● ●● ●●● ●●● ● ●●●●● ●● ●● ●● ●● ● ● ●● ● ● ● ●●● ● ● ●● ● ●● ● ● ● ● ●● ● ● ● ● ●● ● ● ●● ● ●●●●● ●● ●●●●● ● ● ● ● ● ●● ● ● ●●●● ● ●● ●

1.0

50

● ● ●

● ●

● ●

50

CysC

● ●

● ●

● ●●●

● ● ● ● ● ●●● ●● ●● ● ● ●● ● ● ●● ● ● ● ●●● ● ● ● ● ●●●● ● ●● ●●● ● ● ●● ● ● ● ●● ● ●● ● ●● ● ● ● ● ● ●● ●● ● ● ●● ● ●●●●● ● ● ● ● ●●● ●● ● ● ●● ● ●● ●●● ● ● ● ● ● ● ● ● ●● ● ● ● ● ●● ●● ●● ●● ● ● ●● ● ● ● ● ● ● ● ● ●● ● ●● ●● ● ● ●●● ● ● ● ● ● ● ● ● ● ● ●● ● ●●● ● ● ● ● ● ● ●●● ● ● ● ●● ● ● ●● ● ●●●● ●● ● ● ●● ● ● ● ● ● ●● ● ●● ● ● ●●● ● ● ● ● ●● ●●● ● ●● ● ●● ● ● ● ● ●● ● ● ● ●●●● ● ●● ●● ● ● ● ●● ● ● ● ● ● ● ●● ● ● ● ● ●● ● ●●● ● ●● ●●●●● ●● ●●● ●● ● ● ●●● ●● ● ●● ● ● ● ● ● ●●● ● ● ● ● ● ● ● ● ●● ●● ● ●● ●●● ● ● ● ● ● ●● ● ● ● ● ●● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ●● ● ●● ● ● ● ●● ●● ● ● ● ●● ● ●● ● ● ●● ● ● ● ●● ● ●●● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ●● ● ● ● ●● ●●●● ●● ● ● ● ● ●● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ●● ●● ● ● ●● ● ●●● ● ● ● ●● ● ● ● ●● ● ● ● ●● ● ●● ● ● ● ● ● ●● ● ● ● ● ●●●● ●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ●●● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ●● ●●● ● ● ● ● ● ● ● ● ●● ●● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ●● ●● ● ●● ● ● ●● ● ● ● ● ● ● ●● ● ● ● ● ●● ● ● ●● ● ● ●● ●● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ●● ● ● ● ● ● ●● ●●●●●● ● ● ● ●● ● ● ● ● ● ● ● ● ● ●● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ●● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ●●● ●● ●● ●● ● ●● ● ● ●● ● ● ● ●●● ● ● ● ● ● ● ● ●● ● ● ● ● ●● ● ● ● ●● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ●● ● ●● ● ●● ● ●● ● ●● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ●● ●● ●●● ●● ● ● ● ● ● ●●●● ● ●● ● ● ● ● ● ● ● ● ●●●● ●● ●● ● ● ● ●●

0.1

CRP

1

● ●

30

● ●

● ●

30

40

1000

BNP ●

● ● ●

50

SYNTAX Score

● ●

10

40

SYNTAX Score

30

● ●

● ●

●● ●

SYNTAX Score

SYNTAX Score

● ●

● ● ●

● ●

40

50

● ●

100

TnI

10000

Supplemental figure. Ethnicity-specific association of biomarker levels with SYNTAX score from univariable regression model. Significant interactions (p<0.05) were found in a multivariable model for: NTproBNP with Malay ethnicity (p=0.032) and MPO with Malay ethnicity (p=0.017) as compared to Caucasian ethnicity.

131



PART ONE Ethnicity

Chapter 7 Ethnic Differences in QRS prolongation and its Association with Ejection Fraction and Outcomes in Heart Failure In preparation

Crystel M. Gijsberts, Lars H. Lund, Ulf Dahlstrรถm, David Sim, P.S. Daniel Yeo, Hean Yee Ong, Fazlur Jaufeerally, Gerard K.T. Leong, Lieng H. Ling, A. Mark Richards, Dominique P.V. de Kleijn, Carolyn S.P. Lam


Chapter 7

Abstract Background QRS prolongation is a prognostic marker indicating the need for device therapy in patients with heart failure (HF). However, QRS duration (QRSd) cutoffs were derived from predominantly White populations, and ethnic differences are poorly understood. Methods and Results We compared the association of QRSd with ejection fraction (EF) and outcomes between Singaporean Asian and Swedish White HF patients with preserved and reduced EF (HFPEF and HFREF) followed in prospective population-based HF studies. Among 903 Asians (aged 61Âą12 years, 23% women) and 9,576 Whites (aged 74Âą12 years, 39% women) with HF, lower EF was associated with longer QRSd (p<0.001), with a steeper association among Asians than Whites (pinteraction <0.001). In HFPEF Asians had shorter QRSd than Whites (median 90 vs. 94 ms; p=0.007); whereas in HFREF Asians had longer QRSd than Whites (median 102 vs. 100ms; p<0.001). These differences persisted after covariate adjustment. In HFREF, longer QRSd was associated with increased risk of HF hospitalization or death in both ethnicities, with a higher QRSd cutoff for increased risk (hazard ratio exceeding 1.0) in Asians compared to Whites (115ms vs. 105ms). Conclusions There is a stronger association between QRS prolongation and reduction in EF among Asians compared to Whites with HF. While QRS prolongation was related to adverse outcomes in both Asians and Whites with HFREF, the QRSd cutoff for increased risk may vary by ethnicity. Further studies are needed to determine whether ethnicity-specific cutoffs for clinical decision-making should be considered.

134


Ethnic Differences in QRS prolongation

Introduction QRS prolongation is an electrocardiographic characteristic of heart failure (HF) occurring in approximately 30% of patients.1 QRS duration (QRSd) is strongly related to reduction in left ventricular ejection fraction (EF) and to poor outcomes in HF.2 QRSd cutoffs are therefore used to guide device therapy in HF with reduced EF (HFREF).3 However, the vast majority of prior data on QRSd in HF was derived from studies in White patient populations.4 Ethnic differences in electrocardiographic characteristics, and specifically in QRSd, have been described among adults without HF5, but are unknown in HF. Furthermore, it is unknown whether the association of QRSd with EF and prognosis is modified by ethnicity. We sought to compare the association of QRSd with EF and outcomes between Asian and White patients with HF from population-based HF cohorts in Singapore6 and Sweden7 respectively. Given that general-population Asian adults have a narrower QRS complex compared to general-population White adults (possibly related to smaller heart sizes in Asians8), we hypothesized that among patients with HF, Asians would also have narrower QRS complexes for a given left ventricular EF. We further hypothesized that there could be important prognostic differences in QRSd between Asian and White patients with HF. Recognizing the pivotal role of QRSd in clinical decision-making for device therapy in HFREF9, we aimed to determine the threshold of QRSd associated with increased risk of adverse outcomes in Asian versus White patients with HFREF.

Methods Study population For the purpose of the current analysis two HF cohorts were combined: the Singapore Heart Failure Outcomes and Phenotypes (SHOP) cohort6 and the Swedish Heart Failure Registry10 (SwedeHF) - both contemporary population-based observational HF studies recruiting patients with HF regardless of EF, from either in-hospital (hospitalization with primary diagnosis of HF) or outpatient (visit related to HF management) settings. The SHOP cohort enrolled patients from 6 centers: National University Hospital, Khoo Teck Puat Hospital, National Heart Centre, Tan Tock Seng Hospital, Changi General Hospital and Singapore General Hospital. A total of 1,065 patients were enrolled between June 2010 and July 2014. The detailed SHOP study protocol has been described earlier.6 The protocol of SwedeHF has also been previously described11, recapitulating, enrollment of HF patients started in the year 2000 and is still ongoing. For the current analysis we applied the same time limits to the SwedeHF cohort as the SHOP cohort, and thus only included patients that were enrolled from 2010 onwards (excluding the first 30,987 patients enrolled from 2000 to 2009), n=20,112. Patients with inconsistent death dates were excluded from analysis (n=61 Whites, n=26 Asians).

135


Chapter 7

Patients who received pacing therapy (n=2,147 Whites and 42 Asians) and patients with LBBB (n=6,219 Whites and 92 Asians) were excluded from all analyses. QRSd was missing in 2,109 Whites and 2 Asians, leaving 9,576 Whites and 903 Asians with HF in the final study population. HFREF and HFPEF HFREF was defined as clinical features of HF with an echocardiographic EF lower than 50%. The EF of HFREF patients was further categorized into 40-49%, 30-39% and <30%. HF with preserved EF (HFPEF) was defined as clinical features of HF and an echocardiographic EF greater than or equal to 50%. Outcomes Outcomes were collected through standardized protocols in both cohorts. Mortality was obtained by checking the national death registry of Singapore and the Population Registry in Sweden Sweden. HF hospitalizations were obtained from ICD-10 code registrations in the main (first) position in the Patient Registry in Sweden and by follow-up visits or telephone follow-up in Singapore. Statistical analysis Baseline characteristics were stratified by ethnicity. Continuous, normally distributed data were presented as means ± standard deviations (sd) and compared using a t-test. Non-normally distributed data were presented as medians with interquartile ranges and compared using Kruskal-Wallis testing. Categorical data were reported as percentages per category and compared by chi-square testing. We first compared QRSd between the Asians and Whites by EF category using KruskalWallis (QRSd as a continuous variable) or chi-square testing (QRSd as a dichotomous variable <120ms or ≥120ms). We then examined whether ethnicity modified the association between QRSd and EF using interaction terms. Next, we determined the independent predictors of QRSd among HFPEF and HFREF patients using a backward stepwise linear regression analysis, with the initial model containing: age, sex, ethnicity (Asian vs. White), EF, height, weight, body mass index (BMI), systolic blood pressure (SBP), diastolic blood pressure (DBP), hypertension, diabetes, smoking, history of CAD, history of percutaneous coronary intervention (PCI), history of coronary artery bypass grafting (CABG), history of valve surgery, atrial fibrillation of flutter (AF), history of stroke, history of peripheral arterial disease (PAD), chronic obstructive pulmonary disease (COPD), depression, New York Heart Association (NYHA) class, N-terminal pro-brain natriuretic peptide (NTproBNP), estimated glomerular filtration rate (eGFR), hemoglobin, duration of HF (>6 vs. ≤6 months), heart rate, beta blocker use, angiotensin-converting enzyme (ACE)-inhibitor use, angiotensin II receptor blocker (ARB) use, diuretic use and statin use. Subsequently, we assessed the effect of adding each significant covariate identified from the above-mentioned stepwise regression analysis to the model containing the primary variable (ethnicity).

136


Ethnic Differences in QRS prolongation

Finally, we examined the ethnicity-specific association of QRSd with a composite endpoint of HF hospitalization and all-cause mortality in a Cox regression model for HFPEF and HFREF. The association of QRSd with outcome was tested in (1) a univariable model (plus interaction analysis); (2) in a model adjusting for age and sex; and (3) in a model including covariates derived from a backward stepwise Cox regression analysis, starting with the same covariates as the linear regression analysis. Additionally, we evaluated the non-linear relation of QRSd with outcome by adding a p-spline variable for QRSd with 4 degrees of freedom to the Cox regression model. From this model, ethnicity-specific QRSd cutoffs for outcome risk could be derived. Statistical power restricted us to analyze this only in HFREF patients. All analyses were performed using Rstudio12 and R software13 for statistical computing version 3.1.2. A p-value of <0.05 was considered to be statistically significant (also for interaction terms) and all p-values were 2-sided.

Results Baseline Characteristics by ethnicity Baseline characteristics by ethnicity are presented in Table 1. Asian HF patients were younger than Whites (61.2 vs. 73.5 years, p<0.001). Height, weight and body mass index (BMI) were significantly lower in (162 vs. 171cm, 69 vs. 80kg, and 26.3 vs. 27.1, all p<0.001). Asians had more severely impaired EF (EF was <30% in 45% of Asians vs. 24% of Whites). A history of CAD was more common among Asians (53% vs. 45%, p<0.001), but valve surgery was more common among Whites (5% vs. 1%, p<0.001). AF was more frequently observed in Whites than Asians (52% vs. 21%, p<0.001). Asians with HFREF had less severe NYHA classification and lower NTproBNP levels. Overall, QRSd was longer in Asians than Whites (100 vs. 98ms, p=0.002), but the proportion of patients with prolonged QRSd (≥120ms) did not differ (16.2% vs. 15.8%, p=0.796). QRS-duration in Asians and Whites with HFPEF and HFREF The distribution of QRSd between Asians and Whites stratified by EF is depicted in figure 1 (top panels). Among HFPEF patients, QRSd was shorter in Asians than Whites (median QRSd 90 vs. 94ms, p=0.007). Accordingly, the proportion of patients with prolonged QRS (≥120ms) was lower among Asians than Whites (4.5% vs. 12.0%, p=0.001). In contrast, in HFREF patients QRSd was longer in Asians than Whites (median QRSd 102 vs. 100ms, p<0.001). The proportion of patients with prolonged QRS (≥120ms) however, did not significantly differ (19.9% vs. 17.0% in Whites, p=0.068). The relationship of QRSd with EF differed by ethnicity (figure 1, bottom panel): in Asians a reduction in EF was related to a steeper increase of QRSd compared to Whites (p for interaction <0.001).

137


Chapter 7

Table 1. Baseline characteristics of HFREF patients, stratified by ethnicity. Asian

White

903

9576

61.2 ± 12.1

73.5 ± 12.4

<0.001

23.2

38.7

<0.001

Height (cm, mean ± sd)

162.3 ± 8.6

171.2 ± 9.9

<0.001

Weight (kg, mean ± sd)

69.3 ± 15.2

79.6 ± 19.0

<0.001

BMI (kg/m2, mean ± sd)

26.3 ± 5.3

27.1 ± 5.7

<0.001

SBP (mmHg, mean ± sd)

125.6 ± 22.0

128.2 ± 21.5

<0.001

DBP (mmHg, mean ± sd)

71.6 ± 12.8

73.7 ± 12.6

<0.001

Hypertension (%)

73.5

55.5

<0.001

Diabetes (%)

57.1

23.9

<0.001

4.4 ± 1.3

4.3 ± 1.1

n Age (years, mean ± sd) Sex (female %)

Total cholesterol (mmol/L, mean ±sd) Smoking (%)

0.295 <0.001

Current smoker

23.7

14.9

Ex smoker

29.6

42.3

Non smoker

p-value

46.8

42.8

53.0 ± 20.9

68.1 ± 34.0

<0.001

CAD (%)

52.7

44.7

<0.001

PCI (%)

20.4

17.1

0.015

CABG (%)

13.8

22.5

<0.001

eGFR (ml/min, mean ± sd) Medical History

Valve surgery (%)

1.0

5.2

<0.001

Atrial fibrillation/flutter (%)

21.3

52.4

<0.001

Stroke (%)

10.7

17.4

<0.001

PAD (%)

4.4

10.6

<0.001

COPD (%)

7.9

19.0

<0.001

Depression (%)

1.1

13.6

<0.001

39.6

38.4

Heart Failure Characteristics Duration of HF >6 months (%) EF (%) =>50

24.1

27.6

40-49

11.0

22.3

30-39

19.9

25.9

<30

45.0

24.2

26.4

12.2

NYHA (%) I

<0.001

II

57.7

49.1

III

14.6

34.2

IV Heart rate (bpm, mean ± sd) QRSd (ms, median [IQR]) QRSd >=120ms (%)

0.509 <0.001

1.2

4.5

75.5 ± 14.3

75.2 ± 16.3

0.572

100 [90, 110]

98 [88, 110]

0.002

16.2

15.8

0.796

2050 [814, 4346]

2690 [1190, 5899]

<0.001

12.9 ± 2.2

13.2 ± 1.8

<0.001

Beta-blocker (%)

93.0

86.9

<0.001

ACE-inhibitor (%)

64.2

64.2

1

ARB (%)

19.2

21.2

0.165

NTproBNP (pg/mL, median [IQR]) Hemoglobin (g/L, mean ± sd) Medication use

Diuretics (%)

94.1

77.8

<0.001

Statins (%)

89.3

47.5

<0.001

Values are shown in means ± sd for continuous values unless stated otherwise, categorical variables are presented as percentages per category. Abbreviations: BMI= body mass index, SBP= systolic blood pressure, DBP=diastolic blood pressure, eGFR=estimated glomerular fitration rate, CAD= coronary artery disease, PCI= percutaneous coronary intervention, CABG= coronary artery bypass grafting, PAD= peripheral arterial disease, COPD= chronic obstructive pulmonary disease, NYHA= New York Heart Association, EF= ejection fraction, bpm= beats per minute, NTproBNP= N-terminal pro-brain natriuretic peptide, ACE= angiotensin converting enzyme, ARB= angiotensin II receptor blocker.

138


Ethnic Differences in QRS prolongation

HFPEF

Proportion of patients

0.03

p−value 0.007

0.02

0.01

0.00

60

80

100

QRS duration (ms)

120

140

160

140

160

HFREF

Proportion of patients

0.03

p−value <0.001

0.02

0.01

0.00

60

80

100

QRS duration (ms)

120

QRS duration (ms)

110 105 100 95 90

HFPEF

HFREF Asian

White

Figure 1. Distribution of QRSd in Asians and Whites by EF-group. Asians are depicted in blue, Whites in pink. The top two plots are density plots of QRSd distribution between Asians and Whites, stratified by EF-group. The dotted lines display the median QRSd values per ethnicity. P-values are given for the non-parametric comparison of QRSd between Asians and Whites (Kruskal-Wallis test). In the bottom, the bar chart depicts the mean and standard error of the mean (errorbars) of QRSd for Asians and Caucasians, stratified by EF-group. The relation between EF and QRSd differed by ethnicity (p for interaction p<0.001); QRS prolongation with impaired EF is more pronounced in Asians than in Whites.

Predictors of QRSd in HFPEF and HFREF patients Predictors of QRSd for HFPEF and HFREF selected by backward stepwise linear regression are shown in Supplemental Table 1. Among HFPEF patients, male sex, higher BMI and a history of valve surgery were independently associated with longer QRSd. Among HFREF patients, older age, male sex, lower EF, lower heart rate, larger body size (higher BMI and

139


Chapter 7

height), HF duration >6 months, presence of diabetes, higher NTproBNP and Asian ethnicity (as compared to White ethnicity) were independently associated with longer QRSd; whereas the presence of AF and ARB use were independently associated with shorter QRSd. We evaluated the association of Asian vs. White ethnicity with QRSd upon cumulative addition of the abovementioned significant predictors of QRSd as covariates (Figure 2). When adjusted for the relevant covariates, Asian ethnicity remained significantly associated with shorter QRSd in HFPEF (-4.48ms [-7.15- -1.80] shorter than in Whites, p=0.001), and longer QRSd in HFREF (3.48ms [0.76- 6.21] as compared to Whites, p=0.012). HFREF Univariable + Age + Heart rate + EF + Sex + BMI + AF + HF duration >6 months + Height + DM + NTproBNP + ARB −8

−6

−4

−2

0

2

4

6

8

10

0

2

4

6

8

10

HFPEF Univariable + Sex + Valve surgery + BMI −8

−6

−4

−2

Beta for QRS duration

Figure 2. Association of Asian vs. White Ethnicity with QRSd, for HFPEF and HFREF The points and lines display the point estimate (beta) and 95% confidence intervals of the association of Asian vs. White ethnicity with QRSd, stratified by HFPEF and HFRPEF. The beta value represents the difference between Asian and White QRS-duration (shorter in HFPEF and longer in HFREF). The top estimates display the univariable association of Asian ethnicity with QRSd in both HFPEF and HFREF patients. Subsequently, the parameters (as displayed in table 3) that were significantly associated with QRSd (p-value <0.05) were added to the model. From each of those models the association of Asian ethnicity with QRSd is shown.

140


Ethnic Differences in QRS prolongation

Prognostic value of QRSd for HF Hospitalization and Mortality in HFPEF and HFREF patients The association of baseline QRSd (modeled as a continuous variable) with the composite end-point of HF hospitalization or all-cause death was assessed in both Asian and White HFPEF and HFREF patients (Table 2). In HFPEF, the unadjusted hazards ratio (HR) for each 10ms increase in QRSd was 1.06 in both Asians and Whites (reaching statistical significance only in Whites). Covariate adjustment increased confidence intervals but did not change the point estimate of the HR. In HFREF, QRSd was significantly related to outcomes in both Asians (HR per 10ms increase in QRSd 1.05 [1.00-1.11], p=0.047) and Whites (HR 1.10 [1.081.12], p<0.001). The HR of QRSd for Asians tended to be lower than for Whites (p for interaction 0.08). Addition of covariates attenuated the prognostic significance of QRSd in Asians but not in Whites, likely due to lower statistical power in the smaller Asian cohort. Similar results were found when assessing all-cause mortality as the only outcome measure (Supplemental Table 2). Due to limitations in statistical power in the smaller HFPEF cohorts, the following further analyses were limited to the HFREF cohorts: Penalized splines were constructed and assessed for QRSd cutoffs for increased risk of events in Asians vs. Whites (Figure 3). The QRSd at which the HR exceeded the value of 1.0 was 115ms for Asians and 105 for Whites. And a HR of 1.25 corresponded to QRSd cutoff values of 161ms for Asians and 120ms for Whites. The prognostic value of the standard clinical cutoff of ≼120ms did not differ significantly between Asians and Whites (p for interaction 0.16). The survival curves for the ethnicity-specific QRSd cutoffs from spline analyses (at HR 1.25) as well as the standard clinical cutoff of 120 ms are shown in Figure 4.

Table 2. Hazard ratios of QRS prolongation (per 10ms) for HF hospitalization or all-cause mortality per ethnicity in HFPEF and HFREF patients. Ethnicity

HF Type

Model

HR (95% CI)

Asians

HFPEF

UV

1.06 (0.93 - 1.21)

0.391

AS

1.04 (0.91 - 1.19)

0.546

MV

1.10 (0.90 - 1.35)

0.338

UV

1.05 (1.00 - 1.11)

0.047

AS

1.04 (0.98 - 1.09)

0.169

MV

1.07 (0.99 - 1.14)

0.073

UV

1.06 (1.03 - 1.09)

<0.001

AS

1.05 (1.02 - 1.09)

0.003

MV

1.08 (0.96 - 1.22)

UV

1.10 (1.08 - 1.12)

<0.001

AS

1.09 (1.07 - 1.11)

<0.001

MV

1.07 (1.02 - 1.13)

0.012

HFREF

Whites

HFPEF

HFREF

p-value

0.187

Univariable (UV), age and sex-adjusted (AS) and multivariable (MV) adjusted hazard ratios of QRSd (per 10ms increase) are presented. The multivariable analyses were corrected for age, EF (only in HFREF model), height, SBP, diabetes, history of CAD, history of valve surgery, history of PAD, COPD, depression, NYHA class, NTproBNP, eGFR, hemoglobin, HF duration (>6 months), heart rate, beta blocker use, ACE-inhibitor use, ARB use, diuretic use and statin use. No significant interactions were found between QRSd and ethnicity in the univariable model or the multivariable models, only a trend towards a difference was found for QRSd in HFREF in the univariable model (p for interaction 0.08).

141


Chapter 7

HFREF 2.00

Hazard Ratio

Asian White 1.50

1.25

1.00

0.80 80

100

120

140

160

180

QRSd (ms) Figure 3. Cox proportional hazard plot of spline estimates of QRSd for HF hospitalization and death Penalized spline plots of QRSd for HF hospitalization and all-cause death during follow-up. The QRSd at which the colored line crosses the reference line of HR=1 could be considered as a cutoff for increased risk of the composite end-point. This cutoff was 115ms for Asians with HFREF and 105ms for Whites. At a HR of 1.25 (dotted line) the cutoff was 120ms for Whites and 161ms for Asians. Clinical cut−off

Event−free survival

1.0

P−spline cut−off HR 1.25 1.0

HR Asian: 1.31 (1.00 − 1.72), p−value 0.047 HR White: 1.59 (1.45 − 1.75), p−value <0.001

0.8

0.8

0.6

0.6

0.4

0.4

0.2

0.2

Asian <120ms White <120ms Asian >=120ms White >=120ms

0.0 0

200

400

HR Asian: 1.92 (1.16 − 3.18), p−value 0.011 HR White: 1.59 (1.45 − 1.75), p−value <0.001

Asian <161ms White <120ms Asian >=161ms White >=120ms

0.0 600

Time (days)

800

1000

0

200

400

600

800

1000

Time (days)

Figure 4. Survival curves for QRSd≥120 and ethnicity-specific spline cutoffs for White and Asian HFREF patients Cox regression survival estimates are displayed for White (red lines) and Asian (blue lines) HFREF patients, using the established clinical cutoff of ≥120ms (left panel) and ethnicity-specific cutoffs derived from spline analysis at the thresholds of HR 1.25 (right panel). The cutoff at HR 1.25 (as can be observed from figure 3) was 120ms for Whites and 161ms for Asians.

142


Ethnic Differences in QRS prolongation

Discussion These are the first data comparing QRSd and its correlates and prognostic significance between Asian and White HF patients from parallel prospective population-based cohorts. Ethnicity significantly modified the relationship between EF and QRSd in HF, with a steeper association between QRS prolongation and EF reduction among Asians than Whites. In HFPEF, Asians had shorter QRSd than Whites, whereas in HFREF, QRSd was longer in Asians than Whites. These ethnic differences in QRSd were independent of clinical factors. While QRS prolongation was related to adverse outcomes in both Asians and Whites with HFREF, the QRSd cutoff for increased risk was lower in Whites (105ms) vs. Asians (115ms). Ethnic variation in QRSd QRS prolongation in HF patients has been examined among Whites, Blacks and Hispanics14, showing shorter QRSd in Blacks with HFREF.15 Nothing is known however, about QRSd in Asians versus Whites with HF or about the modifying role of HFREF vs. HFPEF. Contrary to our hypothesis, we found more severe QRS prolongation in Asians compared to Whites with HFREF, independent of clinical factors. The mechanism behind this phenomenon remains unclear. One intriguing consideration relates to ethnic differences in what may be considered “normal” QRSd in Asians vs. Whites. In general population adults without HF, a shorter QRSd has been described in Asians compared to Whites.8,16,17 This may relate to intrinsically smaller heart sizes in Asians18, and is consistent to our current observations in HFPEF patients. In our study, height, as a surrogate of heart size, was significantly related to QRSd (p<0.001). However, in stepwise backward linear regression height was not included in the final HFPEF model for the prediction of QRSd (figure 2). When height was forced into the model, the ethnic difference in QRSd was attenuated (data not shown), suggesting that the observed difference in QRSd between Asians and Whites with HFPEF could indeed be related to differences in body and heart dimensions. Another consideration relates to what is considered “normal” EF in Asians versus Whites. In the general population of the Multi-Ethnic Study of Atherosclerosis (MESA) cohort, cardiac magnetic resonance imaging showed a significantly higher EF of 72% for Chinese than Whites (EF 68%).19 Similarly, the EchoNormal study18 showed that “normal” EF is considerably higher in East Asians than in Whites (lower reference value for EF 56% vs. 50% in 50-year old men and 57% vs. 51% in 50-year old women). A given low EF measurement in HFREF (e.g. EF 30%) may therefore represent a greater EF decline in Asians than Whites, and may relate to the greater extent of QRSd-prolongation in Asians than Whites with HFREF. In other words, more severe QRS prolongation in Asians may be a reflection of the relatively larger impairment of EF. Ethnicity-specific association of QRSd with outcome QRS prolongation is known to be a predictor of mortality in HF, as shown in prior studies10 and confirmed in the current. We extend the prior data by showing that the cutoff values

143


Chapter 7

for QRSd associated with increased risk of adverse events may differ between Asians and Whites with HFREF, with higher cutoff values observed in Asians than Whites. One other study directly compared the prognostic value of QRSd among ethnicities albeit in a different population (healthy general population): the MESA study20 found no significant interaction between ethnicity (Whites, Blacks, Hispanics, Chinese) and QRSd in the general population without HF. While our findings are notable, we acknowledge that our observational data from two separate cohorts cannot provide conclusive evidence for different QRSd risk thresholds for Asians and Whites, and cannot be extrapolated to response to treatment such as cardiac resynchronization therapy (CRT). The opposing differences between HFPEF (shorter QRSd in Asians than Whites) and HFREF (longer QRSd in Asians than Whites) suggest that our results are not due to a systematic difference in QRSd measurements between cohorts. We had limited statistical power to assess outcomes in the HFPEF subgroups, and were unable to assess the ethnicity-specific relation of QRSd with outcome between the HF types (i.e. test for three-way interaction among ethnicity, QRSd and HF type). Further large prospective interventional studies may be needed to answer these questions. Conclusion These initial data comparing QRSd between Asian and White HF patients from parallel prospective population-based cohorts showed that ethnicity significantly modified the relationship between EF and QRSd in HF, with a steeper association between QRS prolongation and EF reduction among Asians than Whites. Independent of clinical factors, HFPEF Asians had shorter QRSd than Whites, whereas in HFREF, QRSd was longer in Asians than Whites. While QRS prolongation was related to adverse outcomes in both Asians and Whites with HFREF, the QRSd cutoff for increased risk appeared to vary by ethnicity. Further studies are needed to determine whether ethnicity-specific cutoffs for clinical decision-making should be considered. Funding This work was supported by grants to Lars Lund’s institution from Swedish Research Council [grant 2013-23897-104604-23]; Swedish Heart Lung Foundation [grants 20100419, 20120321]; Stockholm County Council [grants 20110120, 20140220]; Swedish Society of Medicine [grant 174111] and Boston Scientific. Acknowledgements We thank Lina Benson for assistance with data procurement.

144


Ethnic Differences in QRS prolongation

References 1.

Kashani A, Barold SS. Significance of QRS Complex Duration in Patients With Heart Failure. J Am Coll Cardiol. 2005;46:2183–2192.

2.

Sandhu R, Bahler RC. Prevalence of QRS prolongation in a community hospital cohort of patients with heart failure and its relation to left ventricular systolic dysfunction. Am J Cardiol. 2004;93:244–246.

3.

Brignole M, Auricchio A, Baron-Esquivias G, Bordachar P, Boriani G, Breithardt O-A, Cleland J, Deharo J-C, Delgado V, Elliott PM, Gorenek B, Israel CW, Leclercq C, Linde C, Mont L, Padeletti L, Sutton R, Vardas PE, Zamorano JL, Achenbach S, Baumgartner H, Bax JJ, Bueno H, Dean V, Deaton C, Erol C, Fagard R, Ferrari R, Hasdai D, Hoes AW, Kirchhof P, Knuuti J, Kolh P, Lancellotti P, Linhart A, Nihoyannopoulos P, Piepoli MF, Ponikowski P, Sirnes PA, Tamargo JL, Tendera M, Torbicki A, Wijns W, Windecker S, Blomstrom-Lundqvist C, Badano LP, Aliyev F, Bänsch D, Bsata W, Buser P, Charron P, Daubert J-C, Dobreanu D, Faerestrand S, Le Heuzey J-Y, Mavrakis H, McDonagh T, Merino JL, Nawar MM, Nielsen JC, Pieske B, Poposka L, Ruschitzka F, Van Gelder IC, Wilson CM. 2013 ESC Guidelines on cardiac pacing and cardiac resynchronization therapy: the Task Force on cardiac pacing and resynchronization therapy of the European Society of Cardiology (ESC). Developed in collaboration with the European Heart Rhythm Association . Eur Heart J. 2013;34:2281–329.

4.

Colvin M, Sweitzer NK, Albert NM, Krishnamani R, Rich MW, Stough WG, Walsh MN, Westlake Canary C a., Allen L a., Bonnell MR, Carson PE, Chan MC, Dickinson MG, Dries DL, Ewald G a., Fang JC, Hernandez AF, Hershberger RE, Katz SD, Moore S, Rodgers JE, Rogers JG, Vest AR, Whellan DJ, Givertz MM. Heart Failure in Non-Caucasians, Women, and Older Adults: A White Paper on Special Populations From the Heart Failure Society of America Guideline Committee. J Card Fail. 2015;21:674–93.

5.

Ambrosy AP, Gheorghiade M, Chioncel O, Mentz RJ, Butler J. Global Perspectives in Hospitalized Heart Failure: Regional and Ethnic Variation in Patient Characteristics, Management, and Outcomes. Curr Heart Fail Rep. 2014;11:416–427.

6.

Santhanakrishnan R, Ng TZEP, Cameron V a, Kui G, Leong TOH, Shuan POH, Yeo D. Clinical Trial : Methods and Design The Singapore Heart Failure Outcomes and Phenotypes ( SHOP ) Study and Prospective Evaluation of Outcome in Patients With Heart Failure With Preserved Left Ventricular Ejection Fraction ( PEOPLE ) Study : Rationale and . 2013;19:156–162.

7.

Jonsson A, Edner M, Alehagen U, Dahlström U. Heart failure registry: a valuable tool for improving the management of patients with heart failure. Eur J Hear Fail J Work Gr Hear Fail Eur Soc Cardiol. 2010;12:25–31.

8.

Tan E, Xu CF, Liang F, Santhanakrishnan R, Chan MM, Seow S-C, Ching CK, Richards M, Ng TP, Lim TW, Lam C. ASSOCIATION OF ETHNICITY, AGE AND BODY SIZE WITH ELECTROCARDIOGRAPHIC VALUES IN THE COMMUNITY. J Am Coll Cardiol. 2014;63:A1636.

9.

Zareba W, Klein H, Cygankiewicz I, Hall WJ, McNitt S, Brown M, Cannom D, Daubert JP, Eldar M, Gold MR, Goldberger JJ, Goldenberg I, Lichstein E, Pitschner H, Rashtian M, Solomon S, Viskin S, Wang P, Moss AJ. Effectiveness of cardiac resynchronization therapy by QRS morphology in the multicenter automatic defibrillator implantation trial-cardiac resynchronization therapy (MADIT-CRT). Circulation. 2011;123:1061–1072.

10. Lund LH, Jurga J, Edner M, Benson L, Dahlström U, Linde C, Alehagen U. Prevalence, correlates, and prognostic significance of QRS prolongation in heart failure with reduced and preserved ejection fraction. Eur Heart J. 2013;34:529–539. 11. Eklind-Cervenka M, Benson L, Dahlström U, Edner M, Rosenqvist M, Lund LH. Association of candesartan vs losartan with all-cause mortality in patients with heart failure. JAMA. 2011;305:175–182.

145


Chapter 7

12. Allaire J. RStudio: Integrated development environment for R. 2012; 13. R Core Team. R: A Language and Environment for Statistical Computing. 2013; 14. Hebert K, Quevedo HC, Tamariz L, Dias A, Steen DL, Colombo R a., Franco E, Neistein S, Arcement LM. Prevalence of conduction abnormalities in a systolic heart failure population by race, ethnicity, and gender. Ann Noninvasive Electrocardiol. 2012;17:113–122. 15. Ilkhanoff L, Soliman EZ, Ning H, Liu K, Lloyd-Jones DM. Factors associated with development of prolonged QRS duration over 20 years in healthy young adults: The coronary artery risk development in young adults study. J Electrocardiol. 2012;45:178–184. 16. Macfarlane PW, Katibi IA, Hamde ST, Singh D, Clark E, Devine B, Francq BG, Lloyd S, Kumar V. Racial differences in the ECG--selected aspects. J Electrocardiol. 2014;47:809–14. 17. Mansi I a, Nash IS. Ethnic differences in electrocardiographic intervals and axes. J Electrocardiol. 2001;34:303– 307. 18. The EchoNoRMAL Study. Ethnic-Specific Normative Reference Values for Echocardiographic LA and LV Size, LV Mass, and Systolic Function. JACC Cardiovasc Imaging. 2015;8:656–665. 19. Bahrami H, Kronmal R, Bluemke D a, Olson J, Shea S, Liu K, Burke GL, Lima J a C. Differences in the incidence of congestive heart failure by ethnicity: the multi-ethnic study of atherosclerosis. Arch Intern Med. 2008;168:2138–2145. 20. Ilkhanoff L, Liu K, Ning H, Nazarian S, Bluemke D a., Soliman EZ, Lloyd-Jones DM. Association of QRS duration with left ventricular structure and function and risk of heart failure in middle-aged and older adults: The Multi-Ethnic Study of Atherosclerosis (MESA). Eur J Heart Fail. 2012;14:1285–1292.

146


Ethnic Differences in QRS prolongation

Supplemental Supplemental table 1. Results from backward stepwise multivariable linear regression: estimates and 95% confidence intervals of associations with QRSd ordered by strength. Beta (95% CI)

P-value

-11.57 (-14.84- -8.29)

<0.001

10.77 (3.25- 18.28)

0.005

0.32 (0.02-0.61)

0.037

ACEi use

-3.14 ( -6.40-0.12)

0.059

Asian ethnicity vs. White ethnicity

-2.59 ( -6.12-0.94)

0.150

0.11 ( -0.04-0.26)

0.161

HFPEF Sex (Female vs. Male) History of valve surgery BMI (per 1 kg/cm2)

Age HFREF Age

0.29 ( 0.19- 0.39)

<0.001

-0.15 (-0.22--0.08)

<0.001

EF (<30% vs. 40-49%)

6.38 ( 3.44- 9.32)

<0.001

Sex (Female vs. Male)

-6.15 (-9.23--3.07)

<0.001

0.40 ( 0.18- 0.63)

<0.001

Heart rate (per 1-unit increase)

BMI (per 1 kg/cm2) AF

-4.19 (-6.55--1.82)

0.001

Duration of HF (>6 months vs. <6 months)

3.80 ( 1.53- 6.07)

0.001

Height (per 1 cm)

0.26 ( 0.10- 0.41)

0.001

Diabetes

4.00 ( 1.55- 6.46)

0.001

BNP (per 1000 pg/mL)

0.21 ( 0.05- 0.37)

0.011

Asian ethnicity vs. White ethnicity

4.38 ( 0.96- 7.79)

0.012

-3.71 (-7.19--0.23)

0.037

2.91 (-0.31- 6.13)

0.077

ARB use History of stroke SBP (per 1 mmHg)

-0.04 (-0.09- 0.01)

0.117

ACEi use

-2.55 (-5.78- 0.69)

0.122

1.89 (-0.93- 4.72)

0.189

EF (30-39% vs. 40-49%)

Results from a backward stepwise multivariable linear regression models are presented (separate analyses were conducted for HFPEF and HFREF). The variables in the initial model were: age, sex, ethnicity (Asian vs. White), EF (only in the HFREF analysis), height, weight, BMI, SBP, DBP, hypertension, diabetes, smoking, history of CAD, history of PCI, history of CABG, history of valve surgery, atrial fibrillation/flutter, history of stroke, history of PAD, COPD, depression, NYHA, BNP, eGFR, hemoglobin, duration of HF (>6 months), heart rate, beta blocker use, ACE-inhibitor use, ARB use, diuretic use and statin use.

147


Chapter 7

Supplemental table 2. Hazard ratios of QRS prolongation (per 10ms) for all-cause mortality per ethnicity in HFPEF and HFREF patients. Ethnicity

HF Type

Model

HR (95% CI)

Asians

HFPEF

UV

1.04 (0.80 - 1.33)

0.788

AS

0.97 (0.75 - 1.25)

0.820

MV

1.07 (0.71 - 1.62)

0.738

UV

1.09 (1.01 - 1.18)

0.030

AS

1.05 (0.97 - 1.14)

0.192

MV

1.08 (0.97 - 1.22)

0.168

UV

1.07 (1.03 - 1.11)

<0.001

AS

1.05 (1.01 - 1.09)

0.024

MV

1.23 (1.07 - 1.41)

0.004

UV

1.11 (1.09 - 1.14)

<0.001

AS

1.09 (1.06 - 1.11)

<0.001

MV

1.12 (1.03 - 1.21)

0.010

HFREF

Whites

HFPEF

HFREF

p-value

Univariable (UV), age and sex-adjusted (AS) and multivariable (MV) adjusted hazard ratios of QRSd (per 10ms increase) are presented. The multivariable analyses were corrected for age, EF (only for HFREF), height, SBP, diabetes, history of CAD, history of valve surgery, history of PAD, COPD, depression, NYHA class, NTproBNP, eGFR, hemoglobin, HF duration (>6 months), heart rate, beta blocker use, ACE-inhibitor use, ARB use, diuretic use and statin use. No significant interactions were found between QRSd and ethnicity in the univariable model or the multivariable models.

148


149



PART TWO Sex Differences

Chapter 8 Severity of Stable Coronary Artery Disease and its Biomarkers Differ between Men and Women Undergoing Angiography Atherosclerosis. 2015 Jul;241(1):234-40

Crystel M. Gijsberts, Aisha Gohar, Imo E. Hoefer, Dominique P.V. de Kleijn, Folkert W. Asselbergs, Hendrik M. Nathoe, Pierfrancesco Agostoni, Saskia Z.H. Rittersma, Gerard Pasterkamp, Yolande Appelman, Hester M. den Ruijter


Chapter 8

Abstract Background Coronary artery disease (CAD) affects both men and women. Cardiovascular biomarkers have been suggested to relate to CAD severity, but data on sex-specificity is scarce. Therefore, we investigated the association of established biomarkers with the severity of CAD in stable patients undergoing coronary angiography in a sex-specific manner. Methods We studied stable patients undergoing coronary angiography and measured CAD severity by SYNTAX score and biomarker levels (N-terminal pro-brain natriuretic peptide (NT pro-BNP), high-sensitivity CRP (hsCRP), cystatin C (CysC), myeloperoxidase (MPO), highsensitivity troponin I (hsTnI) and von Willebrand factor (VWF)). We tested for sex differences in SYNergy between percutaneous coronary intervention with TAXUS™ and cardiac surgery (SYNTAX) scores and biomarker levels using multivariable ANCOVA. We investigated the association of biomarker levels with SYNTAX score in a multivariable linear regression with interaction terms for sex. Results We analysed data on 460 men and 175 women. SYNTAX scores were significantly lower in women (9.99 points vs. 11.88 points). Univariably, hsCRP and hsTnI levels were significantly associated with SYNTAX scores (both β 2.5). In multivariable analysis only hsCRP associated with SYNTAX score (β 1.9, p=0.009). Sex did not modify the association of biomarkers with SYNTAX score. Conclusion CAD severity as quantified by SYNTAX score is lower in women than men based on coronary angiography. The association of biomarkers with CAD severity did not differ between the sexes.

152


Sex Differences in CAD Severity and Biomarkers

Introduction CAD is the leading cause of mortality in both men and women worldwide.1 Morbidity and death are attributed to the growth, destabilization or rupture of atherosclerotic plaques. Several mechanisms are implicated in the complex process of atherosclerosis; of most importance are inflammation2, endothelial dysfunction and myocardial ischemia. Several biomarkers relating to these processes have been studied and implemented as non-invasive tools for the diagnosis of CAD and for the prediction of future cardiovascular events in primary prevention. Established biomarkers include: NT pro-BNP, which is associated with ventricular dilatation and pressure overload3–5; hsCRP6,7, involved in the inflammatory process; CysC8–11, a marker of renal dysfunction; MPO, linked to both inflammation and oxidative stress12–14; hsTnI15–17, associated with myocardial ischemia and VWF18, which is known to be involved in coagulation. Sex-specific analyses on biomarkers for CAD may provide more insight into the underlying mechanisms of sex differences in CAD. Women represent less than 30% of the population included in cardiovascular research19, yet evidence is accumulating that women develop more “stable” atherosclerosis when compared to men20 and are more likely to have plaque erosion as compared to plaque rupture21 as the underlying substrate for sudden death and myocardial damage. For the purpose of this study we measured SYNTAX scores in men and women presenting with stable CAD (either stable angina, dyspnoea complaints or silent ischemia), undergoing coronary angiography. The SYNTAX score22 is currently the most widely used method to quantify the complexity and severity of CAD. Furthermore, the SYNTAX score is predictive of future cardiovascular events.23 We hypothesize that there are sex differences in established CAD biomarker levels and that they associate differently with the severity of CAD between men and women with stable complaints.

Methods Study population We analysed data from the UCORBIO cohort, a biobank of patients undergoing coronary angiography with or without coronary intervention in the University Medical Center in Utrecht, the Netherlands. From October 2011 to April 2013 we enrolled patients from the catheterization laboratories (n=1,030). For the current study only patients presenting with stable complaints (either stable angina, dyspnoea complaints or silent ischemia) were selected (n=635). Demographical data was collected at baseline (age, sex, cardiovascular risk factors, indication for angiography, treatment and medication use at the moment of angiography). All patients provided written informed consent. This study conforms to the declaration of Helsinki.

153


Chapter 8

CAD severity Angiographic data was collected and categorized into three categories: no CAD, minor CAD (wall irregularities, <50% stenosis) and significant CAD (at least one epicardial vessel with >50% stenosis) based on visual assessment. Two independent observers, using SYNTAX score calculator version 2.11, measured the SYNTAX scores. The SYNTAX score allows for the characterization of coronary vasculature with respect to the number of lesions involved, the location and complexity of the lesions. Lesions are only scored if they meet the required criteria (>50% stenosis and vessel diameter >1.5mm).22 Higher scores are allocated to the most complex lesions. The observers were blinded to the biomarker levels of the patient. The two observers had unlimited access to quantitative coronary angiography24 (QCA) software (CAAS, Siemens) to measure the percentage of stenosis or the dimension of the vessel if they were unsure about significance of a lesion by eyeballing. When the two observers were more than 5 SYNTAX points apart, the case was discussed in order to reach consensus and, if needed, QCA was performed in order to determine the significance of a lesion (>50% stenosis and vessel diameter >1.5mm). The average of the SYNTAX scores of the two observers was used for the current analysis. Patients, who the interventional cardiologists classified as having significant CAD, but ended up with a SYNTAX score of 0 (because of not meeting the criteria of >50% stenosis or vessel <1.5mm or only lesions in non-dominant right coronary artery) were discarded from the analysis, as this is not considered significant CAD in terms of the SYNTAX classification. Biomarkers Blood was drawn from the arterial sheath that was inserted for the angiographic procedure, before any procedure-related drugs were administered. The sample was immediately centrifuged and plasma was frozen at -80째C. Levels of NT pro-BNP, hsCRP, CysC, MPO and VWF were measured from thawed EDTA plasma using validated in-house sandwich ELISA assays performed in the University Medical Center Utrecht, the Netherlands. Quality controls were used in each plate. Inter- and intra assay coefficients of the assays are <10%. Levels of hsTnI were measured in the Gelre Ziekenhuis, Apeldoorn, the Netherlands using the clinically validated ARCHITECT STAT High Sensitive Troponin-I assay (Abbott Laboratories, Lisnamuck, Longford, Ireland). Statistics Differences in patient characteristics between men and women were tested with a t-test or Mann-Whitney U test for continuous variables and chi-square testing for categorical variables. The baseline characteristics that differed (p<0.20) by CAD severity (absence or presence of significant CAD in either men or women) were included in the analyses as covariates. These were: sex, age, body mass index (BMI), smoking, history of percutaneous coronary intervention (PCI), history of acute coronary syndrome (ACS), peripheral arterial disease (PAD), treatment of CAD, platelet inhibitor and statin use.

154


Sex Differences in CAD Severity and Biomarkers

We assessed sex differences in SYNTAX scores in univariable and multivariable analyses (ANCOVA), both in a model containing baseline differences and in a model containing baseline differences and biomarker levels. Biomarker levels were non-normally distributed and therefore log-transformed for analysis where needed. The log back-transformed biomarker means and the multivariably adjusted log back-transformed means were calculated through ANCOVA for men and women and sex differences were tested. The association of the biomarker levels for SYNTAX scores were tested using multivariable linear regression analysis. To determine whether the association of biomarkers with SYNTAX scores differed by sex we tested interaction terms of biomarker levels with sex. The level of significance for all analyses was set at Îą <0.05. All statistical analyses were performed using the R software25 package (version 3.1.2, Vienna, Austria).

Results Patient characteristics The patient characteristics are displayed in table 1, stratified by sex. We examined sex differences between men (n=460) and women (n=175) with stable CAD. We found that women were significantly older (67.0 vs. 64.8 years, p=0.01) and more likely to be nonsmokers (53.2% vs. 41%, p=0.003). Men, on the other hand, significantly more often had a history of ACS: 41.0% vs. 27.6%, PCI: 45.8% vs. 30.3% and CABG: 20.2% vs. 6.9%. Men more often were diagnosed with significant CAD (78.2% vs. 56.6%, p<0.001) and more commonly underwent PCI than women (59.6% vs. 43.4%, p=0.001). Also, men were more likely to be prescribed platelet-inhibitors (83.7% vs. 74.3%) and statins (82.1% vs. 70.3%), as expected given their higher prevalence of a history of CAD. Sex differences in severity of CAD Among stable CAD patients we specifically looked into sex differences in the severity of CAD. In patients with significant CAD we quantified the severity of CAD by SYNTAX scoring. These values are visualized in Figure 1. Only patients with a SYNTAX score higher than zero were compared between men (n=271) and women (n=85). We find that uncorrected SYNTAX scores are higher in men than in women, although not significantly (p=0.10). When we corrected the SYNTAX scores for baseline differences between men and women, we find a significantly higher SYNTAX score in men (11.88 points) than in women (9.99 points, p-value for difference between men and women 0.049). When further adjusted for biomarker levels no significant change was observed. The scores for men rose slightly to 11.93 points and for women decreased to 9.94 points (p-value for difference 0.046). In order to eliminate baseline differences in history of CVD between men and women we repeated the analyses for patients with no history of CVD (no ACS, PCI, CABG, CVA or PAD). This showed comparable results (depicted in the supplemental figure), with a

155


Chapter 8

Table 1. Patient charactistics by sex.

N Age (mean ± sd)

Men

Women

460

175

64.8 ± 10.1

67.0 ± 10.5

p-value

0.013

Risk factors BMI (mean ± sd)

27.4 ± 4.1

26.8 ± 5.1

0.171

Diabetes (%)

25.8

20.6

0.208

Hypertension (%)

61.2

64.9

0.437

Hypercholesterolemia (%)

57.9

53.2

0.336

19.7

22.4

0.003

39.3

24.4

Smoking (Current %) Quit Non-smoker

41.0

53.2

Family history (%)

53.5

60.9

0.183

0.003

Medical history History of ACS (%)

41

27.6

History of PCI (%)

45.8

30.3

0.001

History of CABG (%)

20.2

6.9

<0.001 0.463

History of CVA (%)

11.0

8.6

History of PAD (%)

15.9

10.9

0.144

Kidney failure (%)

3.0

2.3

0.805

COPD (%)

8.7

5.7

0.279

<0.001

Angiography CAD severity (no %)

4.0

15.4

Minor CAD

17.8

28.0

78.2

56.6

SYNTAX score (mean ± sd)

Significant CAD

12.3 ± 8.2

10.8 ± 6.7

0.148

Treatment (Conservative %)

35.7

52.0

0.001

59.6

43.4

4.8

4.6

PCI CABG Medication Platelet inhibitor (%)

83.7

74.3

0.010

Statin (%)

82.1

70.3

0.002

Beta blocker (%)

72.9

70.9

0.674

RAAS (%)

59.8

57.1

0.607

0.104

Biomarkers NT pro-BNP (pmol/L)

35.7 [7.7, 105.5]

42.7 [17.1, 105.6]

hsCRP (µg/mL)

1.2 [0.5, 2.8]

1.5 [0.7, 3.1]

0.017

CysC (µg/mL)

0.8 [0.7, 1.1]

0.8 [0.6, 1.0]

0.642

MPO (ng/mL)

24.5 [18.7, 32.9]

25.1 [19.8, 33.0]

0.471

5.3 [3.3, 10.6]

4.3 [2.6, 8.1]

0.004

13.4 [10.5, 17.4]

13.9 [10.4, 17.8]

0.891

hsTnI (ng/L) VWF (µg/mL)

Patient characteristics of men and women presenting with stable complaints. Continuous variables are presented in means ± standard deviation (sd). Categorical variables are presented in percentages. P-values are the result of ANOVA or chi-square testing. Biomarker levels are presented in medians with interquartile ranges in square brackets. Biomarkers were compared using a MannWhitney U test, as they were non-normally distributed. The SYNTAX score was only measured in people with significant CAD. Abbreviations: BMI: body mass index, ACS: acute coronary syndrome, PCI: percutaneous coronary intervention, CABG: coronary artery bypass grafting, CVA: cerebrovascular accident, PAD: peripheral arterial disease, COPD: chronic obstructive pulmonary disease, CAD: coronary artery disease, RAAS: renin-angiotensin-aldosteron system, NT pro-BNP: N-terminal pro-brain natriuretic peptide, hsCRP: high-sensitivity C-reactive protein, CysC: cystatin C, MPO: myeloperoxidase, hsTnI: high-sensitivity troponin I, VWF: von Willebrand factor.

156


Sex Differences in CAD Severity and Biomarkers

mean SYNTAX score of 11.31 among men and 10.25 among women. When adjusted for baseline differences the mean SYNTAX score for men is 11.20 and 10.44 for women and when biomarkers are added to the model the SYNTAX scores are 11.29 and 10.42, respectively. These differences, however, did not reach statistical significance, probably due to a large reduction in statistical power (only 89 men and 30 women were left for this analysis). Baseline differences

Baseline differences and biomarkers

p−value = 0.109

p−value = 0.049

p−value = 0.046

Men

Men

Men

Mean SYNTAX scores with 95% confidence intervals

Univariable

12

10

8

Women

Women

Women

Figure 1. SYNTAX scores of stable CAD patients, by sex. The bars display mean SYNTAX scores and confidence intervals, derived from univariable analysis, from a model containing baseline differences and from a model containing baseline differences plus biomarker levels. The covariates in the ANCOVAs were: age, sex (effect variable), BMI, smoking, history of PCI, history of ACS, history of PAD, treatment strategy for CAD, use of platelet inhibitor and use of statin (and biomarker levels of NT pro-BNP, hsCRP, CysC, MPO, hsTnI and VWF). The p-value represents the level of significance of the difference in SYNTAX score between men and women.

Sex difference in biomarker levels The biomarker levels of stable CAD patients are displayed in Figure 2, stratified by sex. We present the crude values (transparent) and the multivariably corrected values (nontransparent). The p-values from multivariable ANCOVA analysis are displayed at the top of each plot. Univariably, we found significantly higher levels of hsCRP in women than in men (p=0.02) and significantly lower levels of hsTnI in women than in men (p=0.004). When biomarker levels were corrected for baseline differences between women and

157


Chapter 8

men, we found similar differences as in the univariable analysis. Hs-CRP levels were higher in women than in men (1.65 (μg/mL) vs. 1.19 (μg/mL), p=0.13) and TnI levels were lower in women than in men (5.0 ng/L vs. 6.6 ng/L, p=0.018). The remainder of the biomarkers: NT pro-BNP, CysC, MPO and VWF did not differ between the sexes in the uncorrected analysis and in the multivariable corrected analysis.

p-value = 0.501

p-value = 0.013 2.00

NT pro-BNP (pmol/L)

55

hsCRP (μg/mL)

50

1.75

45 40

1.50

35

1.25

30

Men

Women

1.00

p-value = 0.837

Men

Women

p-value = 0.944

0.90

MPO (ng/mL)

CysC (μg/mL)

28

0.85 0.80

26

0.75

24

Men

Women

p-value = 0.018 7

Men

14.5

Women

p-value = 0.388

hsTnI (ng/L)

VWF (μg/mL)

14.0 13.5

6

13.0

5

12.5 4

Men

Women

Men

Women

Figure 2. Biomarker levels of stable CAD patients by sex, crude values (transparent) and corrected (nontransparent) values. The transparent values are the non-corrected means and confidence intervals of biomarker levels by sex. The non-transparent values are multivariable corrected means and confidence intervals from an ANCOVA model correcting for: age, sex (effect variable), SYNTAX score, BMI, smoking, history of PCI, history of ACS, history of PAD, treatment strategy for CAD, use of platelet inhibitor and use of statin. P-values printed on top of the plots represent p-values for sex difference in biomarker levels from the multivariable analysis.

158


Sex Differences in CAD Severity and Biomarkers

Association of biomarker levels with SYNTAX score The associations of biomarkers with SYNTAX scores are depicted in Figure 3. We examined the association of biomarker levels with SYNTAX score in a univariable model and a multivariable model. The multivariable model contained: age, sex, BMI, smoking, history of PCI, history of ACS, history of PAD, CAD severity, treatment strategy for CAD, use of platelet inhibitor and use of statin. The betas represent the increment in SYNTAX score for every 1-unit increase in log biomarker level (as biomarker levels were positively skewed). HsCRP levels associate with SYNTAX score in stable CAD patients, both in the univariable model and when corrected for baseline differences between men and women. The betas, which represent the increment in SYNTAX score for every 1 unit increase in log hsCRP level, were β 2.5 (p=0.001) and β 1.9 (p=0.009) for the univariable and multivariable model, respectively. HsTnI levels associate with SYNTAX score in both men and women, only in univariable analysis β 2.5 (p=0.004), suggesting that the association of hsTnI levels with SYNTAX score was confounded by baseline differences (multivariable analysis showed β 1.3, p=0.13). We tested interaction terms of biomarker levels with sex in the multivariable model, which yielded no significant interactions (all p>0.10), indicating that sex is not a significant modifier of the relation of biomarkers with SYNTAX score.

Discussion In patients undergoing coronary angiography for stable complaints, we found that women had significantly less severe angiographic CAD than men, as expressed by the SYNTAX score. In this same patient group we observed sex differences in biomarkers related to CAD. Only hsCRP and hsTnI were associated with more severe CAD (higher SYNTAX scores) and these associations were not modified by sex. CAD severity Our results show a disparity of CAD burden between men and women presenting with stable CAD, indicated by SYNTAX score. This difference was not attenuated, but in fact magnified by correction for baseline differences (Figure 1) and biomarker levels. It has been reported previously that women presenting with chest pain are more likely to have less severe CAD than men.26,27 In addition, it appears that CAD occurs in less important parts of the coronary vascular tree in women as opposed to men.28 The SYNTAX score takes the number of lesions, as well as the location of the lesion into account and thus quantifies the myocardium at risk of ischemia. A sex difference in SYNTAX scores has not been reported before and extends our knowledge on differences in the phenotypes of coronary atherosclerosis in men and women. For plaques in the carotid artery different phenotypes for men and women have been previously described20,29, indicating a more stable plaque phenotype in women. As atherosclerosis is a systemic disease, this implicates that the atherosclerosis phenotype might differ by sex throughout the body.

159


Chapter 8

Biomarkers HsCRP levels were significantly higher in women, both crude values and when corrected for baseline differences. High hsCRP levels are associated with increased cardiovascular risk30 and are predictive of future cardiovascular events in asymptomatic individuals.2 HsCRP has been previously shown to be higher in women with stable angina than men.31,32 These higher levels of hsCRP are remarkable, especially in view of less severe CAD, as there is a positive association of hsCRP levels with SYNTAX score. One explanation for this could be that a unit of increase in hsCRP is reciprocated by a similar rise in SYNTAX score both in men and women (as tested in this study: no interaction of sex with hsCRP

SYNTAX Score

30 20 10 0

● ●

● ●● ● ●● ● ● ●● ● ● ● ●● ● ● ●● ● ● ● ●● ●●●● ● ● ● ● ● ● ●●● ●●● ●● ● ● ● ●● ● ● ●● ● ●●● ● ● ● ●● ● ● ● ● ●● ● ●● ● ● ●● ● ●● ●● ● ● ● ● ● ●● ● ● ● ● ● ● ● ●● ● ●● ● ●● ● ● ●● ●● ● ● ● ● ●● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ●● ●● ●●● ● ●● ● ● ● ● ● ●● ● ● ●● ● ●●● ●●● ● ● ● ●● ●● ● ●● ●● ● ●●●● ●● ● ● ●●● ●● ● ● ● ●● ● ● ● ●●● ●● ● ● ● ● ● ● ●● ●● ● ●● ● ● ● ● ● ●● ● ● ●● ● ●●● ● ●● ● ●● ●●●● ● ●● ● ●● ● ● ●● ● ● ●● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ●●● ● ● ● ● ●

100

NTproBNP

30

● ● ●

10

20

● ● ●

10 ●

0

1000

● ● ● ● ● ●● ●●● ● ● ● ● ● ●●● ● ●●● ● ● ●●●● ●● ● ●● ● ● ● ● ● ●●● ●●● ●●● ● ● ● ● ● ● ●● ● ●● ●● ●● ● ● ● ●● ● ● ● ●● ●● ●●●● ● ● ● ●● ●●●● ●● ●● ● ●● ●●● ● ● ●●● ● ●● ● ● ● ●●●● ●● ●● ● ●● ●● ● ●● ●● ●● ● ●●● ● ● ● ●● ● ● ●● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ●● ●●●● ● ● ● ● ●● ●● ●●● ●● ● ● ●●● ●● ●● ● ●● ● ● ●●● ●● ● ● ●● ● ● ● ●● ● ● ●●●●● ● ●● ●● ●● ● ● ●●● ● ●● ● ●● ● ●● ● ●● ● ● ● ●●● ●● ●● ● ● ● ● ●●● ●● ● ●● ● ●● ●● ●●● ●

● ● ●

10 ●

0

1.0

CysC

● ●● ● ●● ●● ● ●● ●● ● ● ● ● ●● ● ● ●● ● ● ● ●●● ● ●● ●●●● ● ● ● ● ● ●● ●● ● ● ● ●● ●● ● ●●● ●● ● ● ●● ● ● ● ●● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ●●● ● ● ● ●● ● ● ●● ● ● ● ●● ● ● ● ● ● ● ● ● ● ●● ●● ●● ● ● ● ● ● ● ● ● ● ● ● ●● ●● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ●● ● ● ● ● ●●●● ● ● ●● ● ● ● ●● ● ●● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ●●● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ●● ● ● ●● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ●●● ● ● ●●● ●● ● ● ●● ●

30 20

10 0

10.0

1

10

MPO

20 10 ● 0

● ●● ● ●● ● ● ●●●● ●● ● ● ● ● ●● ●● ● ● ● ● ● ● ● ● ● ● ●●●●● ● ● ● ● ● ●● ● ●●●● ● ●● ● ●● ●● ● ●● ● ● ● ● ●●● ● ● ● ● ●● ● ●● ● ●● ●● ●●● ● ● ●●●● ● ● ● ● ●● ● ● ●● ● ● ● ● ● ●● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●● ● ● ●● ● ● ● ● ● ● ● ● ● ● ●● ● ●● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ●● ● ●● ● ●●● ●● ● ● ● ● ● ●●● ● ● ● ●● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ●● ● ● ● ● ● ●● ● ● ● ●●● ● ● ● ● ● ● ● ●●●● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ●● ● ● ●● ● ● ●● ●● ● ● ● ● ● ● ● ●●● ● ● ● ●●●● ●● ●●●

1

100

100

hsTnI

40

SYNTAX Score

SYNTAX Score

30

● ●

40

40

● ● ● ●●● ● ● ● ● ●●● ● ● ● ● ●● ● ●● ● ● ●● ● ●● ●●●● ●●● ● ● ● ● ●● ● ● ●●● ● ●● ● ● ● ●● ●● ● ●● ● ● ●● ●● ● ● ● ● ● ●● ● ● ● ● ●● ●● ● ●●●● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ●● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●● ● ● ●● ●● ●● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ●● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ●●● ● ●● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ●●● ● ● ●● ●● ● ● ● ● ● ● ● ● ●● ● ●● ●

0.1

100

SYNTAX Score

SYNTAX Score

● ●

hsCRP

● ●

● ●

20

1

40 30

40

SYNTAX Score

40

30 20

10000

10 0

●● ● ●●● ● ●●

● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ●● ● ●● ● ● ●● ● ● ● ● ● ● ● ●● ●●●● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ●● ●●● ● ●●●● ●● ● ●●●●● ●● ● ●● ●●●● ● ●● ● ●● ● ●● ● ●● ●● ● ●●●● ● ●●● ●● ● ●●● ● ● ● ●● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●●● ●● ● ● ● ● ● ● ●●● ●● ● ●● ●● ● ● ● ● ●● ●●●● ● ●● ● ● ●●● ●● ●● ●● ● ● ● ●● ● ●● ● ●●● ● ● ● ● ● ● ● ● ● ●●●● ●● ● ● ● ● ● ●● ● ● ● ●●●● ● ● ●● ● ●● ●● ● ● ●● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ●● ● ●●● ● ●● ● ● ● ● ●●●● ● ● ● ● ● ●● ●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ●

10

Men

VWF ●

100

Women

Figure 3. Association of biomarkers with SYNTAX score by sex, in stable CAD patients. Scatterplots of the biomarker levels (on a logarithmic x axis) and SYNTAX scores in stable CAD patients. Men are displayed in blue, women in red. Linear regression lines are displayed for each sex, however, no significant interactions found. HsTnI and hsCRP are univariably associated with higher SYNTAX scores. When adjusted for baseline differences only a significant association of hsCRP remained.

160


Sex Differences in CAD Severity and Biomarkers

for predicting SYNTAX scores, similar betas), but that the baseline levels actually differ by sex (different intercepts). Thus, similar hsCRP levels are associated with lower SYNTAX scores in women than they are in men (as can be observed in Figure 3). Besides these biomarker level differences, hsCRP levels appear to predict mortality in CAD patients equally well for men and women31, by providing important information in addition to the severity of CAD. This was also implied by Patel et al.33, who showed that higher hsCRP levels are not related to progression of atherosclerosis, but are associated with higher (cardiovascular) mortality in postmenopausal women. The exact biological cause of higher levels of hsCRP in women with less severe epicardial CAD remains to be elucidated. A possible explanation may lie in the fact that women suffer from microvascular CAD rather than epicardial CAD.34 Accumulating evidence is showing that microvascular disease is to be considered an inflammatory condition of the coronary endothelium; its presence has been linked to elevated levels of hsCRP.35–37 In contrast to hsCRP, women in our cohort showed lower levels of hsTnI than men, in the univariable as well as the multivariable analysis. High troponin (TnI or TnT) levels reflect the level of cardiac damage caused by ischemia. The association of troponin levels with actual cardiac ischemia may be less frank in women than in men. Lønnebakken et al. reported that women showed lower levels of TnT at a certain extent of myocardial ischemia and less severe CAD than men.38 This is in line with our findings; we find lower hsTnI levels in women, but the levels correspond with SYNTAX scores in a similar way for men as they do for women. From this we can conclude that at similar SYNTAX scores, women have lower hsTnI levels than men. Again, microvascular CAD could be an explanation for lower SYNTAX scores or even no evident epicardial disease in women presenting with complaints and with higher hsTnI levels. Women possibly suffer from more cardiac ischemia than one would expect based on the visualization of their epicardial vessels, as microvascular CAD is indeed known to be more prevalent among women than in men.27 The current focus of cardiologists on epicardial disease and consequential under recognition of microvascular CAD may explain the poorer outcome observed in women, compared to men with non-obstructive CAD.39,40 Investigation of microvascular CAD deserves specific attention in women presenting with chest pain complaints with no or minor epicardial CAD.41 Limitations This study is a cross-sectional single centre study, preventing the possibility of following patients up for the development of cardiovascular events. We were only able to analyse information from patients who provided informed consent, possibly introducing inclusion bias into the study. Patients selected for coronary angiography were strongly suspected of having CAD based on history, risk profile and/or ischemia detection. Specific details with respect to ischemia testing results were lacking in our data, unfortunately. However, our patient selection represents daily clinical practice and without selection differences between men and women.

161


Chapter 8

For the association of biomarkers with the severity of CAD we evaluated patients with a SYNTAX score of >0. Hereby we excluded patients who were classified by the interventional cardiologist to have “significant� CAD, but did not satisfy the SYNTAX criteria (of >50% stenosis in a vessel in >1.5mm). The exclusion of this patient group might pose a bias in our current patient selection (excluding 10 men and 3 women). We were unable to take the effect of menopause into account, as only 10 women of the stable CAD patients were younger than 50 years of age. Conclusion Among stable CAD patients undergoing coronary angiography women show less severe CAD than men, as quantified by SYNTAX score. We also find higher hsCRP and lower hsTnI levels in women than in men. The associations of biomarkers with SYNTAX scores did not differ by sex. This indicates that these established biomarkers are incapable of elucidating sex differences in the severity of CAD. In order to adequately prevent and treat CAD, extensive research is required to unveil the biological differences in the pathophysiology of CAD between men and women.

162


Sex Differences in CAD Severity and Biomarkers

References 1. 2.

Alwan A. Global status report on noncommunicable diseases 2010. 1st ed. Geneva, Switzerland: 2011. Vivona N, Bivona G, Noto D, Sasso B Lo, Cefalù AB, Chiarello G, Falletta A, Ciaccio M, Averna MR. C-reactive protein but not soluble CD40 ligand and homocysteine is associated to common atherosclerotic risk factors in a cohort of coronary artery disease patients. Clin Biochem. 2009;42:1713–8.

3.

Abdullah SM, Khera A, Das SR, Stanek HG, Canham RM, Chung AK, Morrow DA, Drazner MH, McGuire DK, de Lemos JA. Relation of coronary atherosclerosis determined by electron beam computed tomography and plasma levels of n-terminal pro-brain natriuretic peptide in a multiethnic population-based sample (the Dallas Heart Study). Am J Cardiol. 2005;96:1284–9.

4.

Goyal BM, Sharma SM, Walia M. B-type natriuretic peptide levels predict extent and severity of coronary artery disease in non-ST elevation acute coronary syndrome and normal left ventricular function. Indian Heart J. 2014;66:183–7.

5.

Kurtul A, Yarlioglues M, Murat SN, Duran M, Acikgoz SK, Sensoy B, Ergun G, Cetin M, Ornek E. N-Terminal Pro-Brain Natriuretic Peptide Level is Associated With Severity and Complexity of Coronary Atherosclerosis in Patients With Acute Coronary Syndrome. Clin Appl Thromb Hemost. 2014;

6.

Wensley F, Gao P, Burgess S, Kaptoge S, Di Angelantonio E, Shah T, Engert JC, Clarke R, Davey-Smith G, Nordestgaard BG, Saleheen D, Samani NJ, Sandhu M, Anand S, Pepys MB, Smeeth L, Whittaker J, Casas JP, Thompson SG, Hingorani AD, Danesh J. Association between C reactive protein and coronary heart disease: mendelian randomisation analysis based on individual participant data. BMJ. 2011;342:d548.

7.

Arroyo-Espliguero R, Avanzas P, Quiles J, Kaski JC. Predictive value of coronary artery stenoses and C-reactive protein levels in patients with stable coronary artery disease. Atherosclerosis. 2009;204:239–43.

8.

Dandana A, Gammoudi I, Chalghoum A, Chahed H, Addad F, Ferchichi S, Miled A. Clinical utility of serum cystatin C in predicting coronary artery disease in patients without chronic kidney disease. J Clin Lab Anal. 2014;28:191–7.

9.

Imai A, Komatsu S, Ohara T, Kamata T, Yoshida J, Miyaji K, Shimizu Y, Takewa M, Hirayama A, Deshpande GA, Takahashi O, Kodama K. Serum cystatin C is associated with early stage coronary atherosclerotic plaque morphology on multidetector computed tomography. Atherosclerosis. 2011;218:350–5.

10. Koc M, Batur MK, Karaarslan O, Abali G. Clinical utility of serum cystatin C in predicting coronary artery disease. Cardiol J. 2010;17:374–80. 11. Luc G, Bard J-M, Lesueur C, Arveiler D, Evans A, Amouyel P, Ferrieres J, Juhan-Vague I, Fruchart J-C, Ducimetiere P. Plasma cystatin-C and development of coronary heart disease: The PRIME Study. Atherosclerosis. 2006;185:375–80. 12. Buffon A, Biasucci LM, Liuzzo G, D’Onofrio G, Crea F, Maseri A. Widespread coronary inflammation in unstable angina. N Engl J Med. 2002;347:5–12. 13. Wainstein R V, Wainstein M V, Ribeiro JP, Dornelles L V, Tozzati P, Ashton-Prolla P, Ewald IP, Vietta G, Polanczyk CA. Association between myeloperoxidase polymorphisms and its plasma levels with severity of coronary artery disease. Clin Biochem. 2010;43:57–62. 14. Baldus S, Heeschen C, Meinertz T, Zeiher AM, Eiserich JP, Münzel T, Simoons ML, Hamm CW. Myeloperoxidase serum levels predict risk in patients with acute coronary syndromes. Circulation. 2003;108:1440–5.

163


Chapter 8

15. Saunders JT, Nambi V, de Lemos JA, Chambless LE, Virani SS, Boerwinkle E, Hoogeveen RC, Liu X, Astor BC, Mosley TH, Folsom AR, Heiss G, Coresh J, Ballantyne CM. Cardiac troponin T measured by a highly sensitive assay predicts coronary heart disease, heart failure, and mortality in the Atherosclerosis Risk in Communities Study. Circulation. 2011;123:1367–76. 16. Ndrepepa G, Braun S, Schulz S, Mehilli J, Schömig A, Kastrati A. High-sensitivity troponin T level and angiographic severity of coronary artery disease. Am J Cardiol. 2011;108:639–43. 17. Laufer EM, Mingels AM a, Winkens MHM, Joosen I a PG, Schellings MWM, Leiner T, Wildberger JE, Narula J, Van Dieijen-Visser MP, Hofstra L. The extent of coronary atherosclerosis is associated with increasing circulating levels of high sensitive cardiac troponin T. Arterioscler Thromb Vasc Biol. 2010;30:1269–75. 18. Van Loon JE, Kavousi M, Leebeek FWG, Felix JF, Hofman a, Witteman JCM, de Maat MPM. von Willebrand factor plasma levels, genetic variations and coronary heart disease in an older population. J Thromb Haemost. 2012;10:1262–9. 19. Melloni C, Berger JS, Wang TY, Gunes F, Stebbins A, Pieper KS, Dolor RJ, Douglas PS, Mark DB, Newby LK. Representation of women in randomized clinical trials of cardiovascular disease prevention. Circ Cardiovasc Qual Outcomes. 2010;3:135–42. 20. Hellings WE, Pasterkamp G, Verhoeven BAN, De Kleijn DP V, De Vries J-PPM, Seldenrijk KA, van den Broek T, Moll FL. Gender-associated differences in plaque phenotype of patients undergoing carotid endarterectomy. J Vasc Surg. 2007;45:289–96; discussion 296–7. 21. Virmani R, Burke AP, Farb A, Kolodgie FD. Pathology of the vulnerable plaque. J Am Coll Cardiol. 2006;47:C13–8. 22. Sianos G, Morel M, Kappetein AP, Morice M, Colombo A, Dawkins K, van den Brand M, Van Dyck N, Russell ME, Mohr FW, Serruys PW. The SYNTAX Score: an angiographic tool grading the complexity of coronary artery disease. EuroIntervention. 2005;1:219–27. 23. Serruys PW, Onuma Y, Garg S, Sarno G, van den Brand M, Kappetein A-P, Van Dyck N, Mack M, Holmes D, Feldman T, Morice M-C, Colombo A, Bass E, Leadley K, Dawkins KD, van Es G-A, Morel M-AM, Mohr FW. Assessment of the SYNTAX score in the Syntax study. EuroIntervention. 2009;5:50–56. 24. Généreux P, Palmerini T, Caixeta A, Cristea E, Mehran R, Sanchez R, Lazar D, Jankovic I, Corral MD, Dressler O, Fahy MP, Parise H, Lansky AJ, Stone GW. SYNTAX score reproducibility and variability between interventional cardiologists, core laboratory technicians, and quantitative coronary measurements. Circ Cardiovasc Interv. 2011;4:553–61. 25. R Core Team. R: A Language and Environment for Statistical Computing. 2013; 26. Smilowitz NR, Sampson BA, Abrecht CR, Siegfried JS, Hochman JS, Reynolds HR. Women have less severe and extensive coronary atherosclerosis in fatal cases of ischemic heart disease: an autopsy study. Am Heart J. 2011;161:681–8. 27. Han SH, Bae JH, Holmes DR, Lennon RJ, Eeckhout E, Barsness GW, Rihal CS, Lerman A. Sex differences in atheroma burden and endothelial function in patients with early coronary atherosclerosis. Eur Heart J. 2008;29:1359–69. 28. Khaki AA, Khaki A, Heydari M, Gholizadeh L. A description of gender differences in angiographic findings in a single-center Iranian hospital. J Vasc Nurs. 2010;28:11–3. 29. Vrijenhoek JEP, Den Ruijter HM, De Borst GJ, de Kleijn DP V, De Vries J-PPM, Bots ML, Van de Weg SM, Vink A, Moll FL, Pasterkamp G. Sex is associated with the presence of atherosclerotic plaque hemorrhage and modifies the relation between plaque hemorrhage and cardiovascular outcome. Stroke. 2013;44:3318–23.

164


Sex Differences in CAD Severity and Biomarkers

30. Calabrò P, Golia E, Yeh ETH. CRP and the risk of atherosclerotic events. Semin Immunopathol. 2009;31:79–94. 31. Ndrepepa G, Braun S, Cassese S, Fusaro M, King L, Kastrati A, Schmidt R. C-reactive protein and prognosis in women and men with coronary artery disease after percutaneous coronary intervention. Cardiovasc Revasc Med. 2013;14:264–9. 32. Garcia-Moll X, Zouridakis E, Cole D, Kaski JC. C-reactive protein in patients with chronic stable angina: differences in baseline serum concentration between women and men. Eur Heart J. 2000;21:1598–606. 33. Patel D, Jhamnani S, Ahmad S, Silverman A, Lindsay J. Discordant association of C-reactive protein with clinical events and coronary luminal narrowing in postmenopausal women: data from the Women’s Angiographic Vitamin and Estrogen (WAVE) study. Clin Cardiol. 2013;36:535–41. 34. Ong P, Athanasiadis A, Borgulya G, Mahrholdt H, Kaski JC, Sechtem U. High prevalence of a pathological response to acetylcholine testing in patients with stable angina pectoris and unobstructed coronary arteries. The ACOVA Study (Abnormal COronary VAsomotion in patients with stable angina and unobstructed coronary arteries. J Am Coll Cardiol. 2012;59:655–62. 35. Ong P, Carro A, Athanasiadis A, Borgulya G, Schäufele T, Ratge D, Gaze D, Sechtem U, Kaski JC. Acetylcholineinduced coronary spasm in patients with unobstructed coronary arteries is associated with elevated concentrations of soluble CD40 ligand and high-sensitivity C-reactive protein. Coron Artery Dis. 2015;26:126– 32. 36. Recio-Mayoral A, Mason JC, Kaski JC, Rubens MB, Harari O a., Camici PG. Chronic inflammation and coronary microvascular dysfunction in patients without risk factors for coronary artery disease. Eur Heart J. 2009;30:1837–1843. 37. Recio-Mayoral A, Rimoldi OE, Camici PG, Kaski JC. Inflammation and microvascular dysfunction in cardiac syndrome X patients without conventional risk factors for coronary artery disease. JACC Cardiovasc Imaging. 2013;6:660–7. 38. Lønnebakken MT, Nordrehaug JE, Gerdts E. No gender difference in the extent of myocardial ischemia in non-ST elevation myocardial infarction. Eur J Prev Cardiol. 2014;21:123–9. 39. Bugiardini R, Manfrini O, De Ferrari GM. Unanswered questions for management of acute coronary syndrome: risk stratification of patients with minimal disease or normal findings on coronary angiography. Arch Intern Med. 2006;166:1391–5. 40. Gulati M, Cooper-DeHoff RM, McClure C, Johnson BD, Shaw LJ, Handberg EM, Zineh I, Kelsey SF, Arnsdorf MF, Black HR, Pepine CJ, Merz CNB. Adverse cardiovascular outcomes in women with nonobstructive coronary artery disease: a report from the Women’s Ischemia Syndrome Evaluation Study and the St James Women Take Heart Project. Arch Intern Med. 2009;169:843–850. 41. Radico F, Cicchitti V, Zimarino M, De Caterina R. Angina pectoris and myocardial ischemia in the absence of obstructive coronary artery disease: practical considerations for diagnostic tests. JACC Cardiovasc Interv. 2014;7:453–63.

165


Chapter 8

Supplemental

Mean SYNTAX scores with 95% confidence intervals

Patients with no history of CVD 14

Univariable

Baseline differences

Baseline differences and biomarkers

12

10

8

p−value = 0.465

p−value = 0.657

p−value = 0.639

Men

Men

Men

Women

Women

Women

Supplemental figure. The bars display mean SYNTAX scores and confidence intervals, derived from univariable analysis, from a model containing baseline differences and from a model containing baseline differences plus biomarker levels. For this analysis only patients without a history of CVD were included (no prior ACS, CABG, PCI, CVA or PAD). The covariates in the ANCOVAs were: age, sex (effect variable), BMI, smoking, treatment strategy for CAD, use of platelet inhibitor and use of statin (and biomarker levels of NT pro-BNP, hsCRP, CysC, MPO, hsTnI and VWF). The p-value represents the level of significance of the difference in SYNTAX score between men and women.

166


167



PART TWO Sex Differences

Chapter 9 Women Undergoing Coronary Angiography for Myocardial Infarction or who have Multi-Vessel Disease have a worse Prognosis than Men Angiology. 2015 Sep 7. pii: 0003319715604762

Crystel M. Gijsberts, Bernadet T. Santema, Folkert W. Asselbergs, Dominique P.V. de Kleijn, Michiel Voskuil, Pierfrancesco Agostoni, Maarten J. Cramer, Ilonca Vaartjes, Imo E. Hoefer, Gerard Pasterkamp, Hester M. den Ruijter


Chapter 9

Abstract Background Coronary artery disease (CAD) affects both men and women. In this study we examine sex-specific differences in occurrence of major adverse cardiovascular events (MACE) after coronary angiography. Methods We analyzed data from the coronary angiography cohort UCORBIO (n=1,283 men, 480 women). Using Kaplan-Meier and multivariable Cox-regression, we tested for sex differences in MACE occurrence. Additionally, we compared mortality with an age- and sex-matched control group from the general Dutch population. Results During a median follow-up of 2.1 years (IQR 1.6-2.8) MACE occurred in 265 men and 103 women (20.7% vs. 21.3%, p=0.744). Women with myocardial infarction (MI) had significantly more MACE during follow-up than men (hazard ratio (HR) 1.66 for female sex, 95%CI 1.10-2.50, p=0.015), which was also the case for women who had multi-vessel disease (HR 1.41, 95%CI 1.03-1.94, p=0.031). During follow-up, mortality in women presenting with MI was higher than mortality of women in the general population; men did not show this disadvantage. Conclusion MACE occurred more often in women than in men who presented with MI or had angiographic multi-vessel disease upon coronary angiography. Clinical trial registration Clinicaltrials.gov identifier: NCT02304744. URL: https://clinicaltrials.gov/ct2/show/ NCT02304744.

170


Sex Differences in Prognosis After Coronary Angiography

Introduction Cardiovascular disease (CVD) is the leading cause of death worldwide with a total of 17.5 million deaths in 2012, of which 7.4 million were due to coronary artery disease (CAD).1 While ischemic heart disease has been labeled a men’s disease for decades2, in the United States (US), more women than men die of CVD each year.3,4 In the last decades, CAD incidence showed a tremendous decrease mainly as a result of better detection and control of major risk factors and more effective treatment options such as widely accessible percutaneous coronary intervention (PCI).2,5 Specifically for women, the American Heart Association launched campaigns in order to increase awareness of the risk of CAD and published women-specific guidelines, resulting in a substantial increase in awareness among women (from 30% in 1997 to 54% in 2009).6 Also, sex-specific research designated preeclampsia, gestational diabetes mellitus, early menopause and possibly estrogen replacement therapy7 as female-specific risk factors for CAD.2,4,8–10 In spite of the advances in knowledge about CAD in women, unfortunately recently in the US, a concerning increase in mortality rate is seen among women of relatively young age6,11. Among these women, increasing rates of diabetes and obesity possibly nullify the effects of the reduced smoking prevalence and improved hypertension treatment.12 Contrary to the US however, in the European Union (EU), no increase in mortality rates has been observed among women; only plateauing of these rates occurred in a minority of the EU countries among younger individuals of both sexes.12 Given changes in risk factors (higher prevalence of obesity and diabetes) and declining incidence rates of myocardial infarction (MI) in the EU, we investigated whether the previously reported sex differences in outcome still exist in a contemporary European cohort of coronary angiography patients and if so, whether sex differences are observed in certain CAD patient groups specifically (i.e. stable CAD or MI). In addition, we examined whether survival of men and women differed from age and sex matched samples of the general Dutch population. For this purpose, we evaluated the occurrence of all-cause mortality and major adverse cardiovascular events (MACE) among men and women from the Utrecht Coronary Biobank (UCORBIO) cohort, consisting of patients undergoing coronary angiography in the Netherlands. Based on current, mainly US, literature, we hypothesized that women from the UCORBIO cohort have a higher incidence of MACE and higher mortality than men, despite a lower burden of epicardial CAD13. This difference may be partly explained by differences in risk factor burden between men and women.

171


Chapter 9

Methods Study population We analyzed data from the UCORBIO cohort (clinicaltrials.gov identifier: NCT02304744), an observational cohort study of patients undergoing coronary angiography for any indication in the University Medical Center in Utrecht, the Netherlands. From October 2011 to December 2014 a total of 2,589 patients, >18 years of age, were enrolled from the catheterization laboratories. For the current study, patients presenting with MI (either ST- Segment Elevation MI [STEMI] and Non-ST-Segment Elevation MI [NSTEMI]), chest pain without release of cardiac enzymes (stable and unstable angina), dyspnea on exertion, silent ischemia or the need for preoperative screening (for non-cardiac surgery) were selected (n=2,390). All-cause mortality of these UCORBIO patients was compared to the general Dutch population, as explained below. Within the UCORBIO cohort, only patients who had been enrolled for more than one year and thus reached their first follow-up contact moment were analyzed (n=1,763). Figure 1 depicts the selection study patients. This study was approved by the Medical Ethics Committee of the UMC Utrecht (reference number 11-183). All patients provided written informed consent and this study conforms to the Declaration of Helsinki. Control group from general population In order to compare mortality risks of both men and women undergoing coronary angiography with the general population, an age- and sex-matched control group from the Dutch population registry14 was obtained. For every UCORBIO patient, four ageand sex-matched controls were randomly selected as registered on January 1st 2011 from the Dutch population registry. We compared data on these individuals’ survival until January 1st 2015 with survival in the UCORBIO cohort. Clinical data The investigators completed standardized electronic case report forms (eCRFs) at baseline containing age, sex, cardiovascular risk factors, indication for angiography, medication use before admission, angiographic findings and eventual treatment. Angiographic findings were categorized by the interventional cardiologists into four groups: no CAD, minor CAD (wall irregularities, <50% stenosis), single vessel disease (>50% stenosis15) and multi-vessel disease (containing both double and triple vessel disease). SYNTAX scores16 were calculated by two independent observers using SYNTAX score calculator version 2.1117 as described before13. Biomarkers Blood samples collected from the arterial sheath used for coronary angiography were immediately centrifuged and plasma was frozen at -80 °C. In 2013, N-terminal pro-brain natriuretic peptide (NT pro-BNP) levels were measured in the first 982 patients from

172


Sex Differences in Prognosis After Coronary Angiography

Figure 1. Flowchart of study population selection process. All-cause mortality was available for all study patients regardless of duration of follow-up or loss-to-follow-up. Therefore all 2,390 patients could be compared to the general Dutch population. For detailed study follow-up (occurring on a yearly basis, hence excluding patients who had not reached 1 year yet) 1,763 patients were available.

thawed EDTA plasma using a validated in-house sandwich ELISA assay. High-sensitivity troponin I (hsTnI) was measured in the first 936 patients using the clinically validated ARCHITECT STAT High Sensitive Troponin-I assay (Abbott Laboratories, Lisnamuck, Longford, Ireland). Follow-up On a yearly basis, patients received a questionnaire to check for hospital admissions and occurrence of major adverse cardiovascular events. When the patient did not complete or did not return the questionnaire, or reported a hospital admission suspect for MACE, the general practitioner or reported hospital was contacted for confirmation. In the case of a possible adverse event or death, medical records were obtained and details about the adverse event or death were determined. The occurrence of events was scored by CMG and BTS. When uncertain about the relevance or classification of an event (n= 46), the case was discussed by an expert panel of cardiologists (consisting of

173


Chapter 9

at least two of the following cardiologists: MJM, FWA, MV or PA). Additionally, all cases of possible in-stent restenosis (n=80) were scrutinized by an interventional cardiologist (PA or MV). The composite end-point MACE was defined as any and the first of the following clinical events: all-cause death, non-fatal MI, unplanned revascularization; both cardiac (PCI and CABG) and non-cardiac intervention, stroke and admission for heart failure. Statistical Analysis Baseline characteristics were reported as means and standard deviations for continuous variables and percentages for categorical variables. Sex-specific survival curves were plotted using Kaplan-Meier analysis; sex differences in survival were tested by means of a log-rank test. As to identify patient groups in which sex differences could be more pronounced, we further stratified the analysis by indication for angiography and angiographic CAD severity. A similar approach was used for the differences between our patient cohort and the general Dutch population. Hereafter, in order to correct for baseline differences between men and women, we performed Cox regression analysis for the patient groups in which sex differences were observed (patients with angiographic multi-vessel disease and patients presenting with MI). Significant baseline differences between men and women and factors that were associated with outcome in a univariable analysis (p-value <0.1) were selected as covariates in the multivariable Cox regression analysis. Consequently, Cox regression was performed with the following covariates: age, hypertension, hypercholesterolemia, diabetes mellitus, smoking, history of acute coronary syndrome (ACS), history of PCI, history of CABG, history of peripheral arterial disease (PAD), history of cerebrovascular accident (CVA), use of P2Y12 receptor antagonists, renin-angiotensin-aldosterone system (RAAS) inhibitors, statins or diuretics, angiographic CAD severity and treatment of CAD (conservative, PCI or CABG). NT pro-BNP levels (available in n= 982), hsTnI levels (n=936), left ventricular ejection fraction (LVEF, n=1,318) and SYNTAX scores (n=627) were not available in all patients and therefore not included in the Cox model. However, they were added to the model one by one in order to assess their effects on the difference between men and women. All statistical analyses were performed using the R software package (version 3.1.2, Vienna, Austria)18. A two sided p-value of <0.05 was considered statistically significant.

Results Patient characteristics We examined a total of 1,763 patients who underwent coronary angiography. The majority of patients were male (n=1,283, 72.8%) and 480 patients were female. The baseline characteristics are displayed in Table 1, stratified by sex. The most important baseline differences were age, with women being significantly older (66.5 Âą 11.0 years vs. 63.3 Âą 10.7 years, p<0.001) and having a higher prevalence of hypertension (63.1%

174


Sex Differences in Prognosis After Coronary Angiography

Table 1. Baseline Characteristics of UCORBIO patients stratified by sex. Male

Female

1,283

480

Age (mean ± sd)

63.3 ± 10.7

66.5 ± 11.0

<0.001

BMI (mean ± sd)

27.2 ± 4.1

26.9 ± 5.3

0.294

Diabetes (%)

22.2

22.1

1

Hypertension (%)

55.3

63.1

0.004

Hypercholesterolemia (%)

48.9

42.7

N

Smoking (%)

p-value

0.022 <0.001

Current smoker

25.5

Ex smoker

31.0

23.3 21.4

Non-smoker

43.5

55.3

History of ACS (%)

33.7

23.5

<0.001

History of PCI (%)

32.3

22.5

<0.001

History of CABG (%)

14.3

5.8

<0.001

History of CVA (%)

9.7

9.8

1

History of PAD (%)

12.3

9.0

0.059

Kidney failure (%)

3.0

1.9

0.274

COPD (%)

8.3

9.0

0.750

Normal

54.6

67.2

Mildly reduced

23.6

16.5

Moderately reduced

13.5

9.2

LVEF (%)

Severely reduced

<0.001

8.3

7.0

Aspirin (%)

59.6

60.2

0.862

P2Y12 (%)

25.8

19.4

0.006

RAAS (%)

51.3

51.0

0.970

Beta Blocker (%)

55.7

57.1

0.641

Statin (%)

64.2

56.9

0.006

Diuretics (%)

26.1

38.5

<0.001

54.2

55.8

Indication (%) Stable complaints

0.267

Unstable angina

10.2

9.6

Myocardial infarction

29.4

26.2

6.2

8.3

4.3

13.0

Minor CAD

15.4

20.7

Single vessel disease

35.0

31.0

Multi vessel disease

45.4

35.4

Conservative

30.4

41.6

PCI

64.1

53.8

Other CAD severity (%) No CAD

<0.001

Procedure (%)

CABG SYNTAXscore (median [IQR]) NT pro-BNP (median [IQR]) Troponin-I (median [IQR])

<0.001

5.6

4.6

11.0 [6.0, 17.5]

9.0 [5.0, 15.5]

38.0 [8.3, 113.3]

47.3 [16.1, 138.7]

0.027

8.2 [4.1, 30.0]

5.90 [3.1, 18.6]

0.002

0.022

Continuous variables are presented in means ± standard deviation (sd) when normally distributed and as medians with interquartile ranges when non-normally distributed. Categorical variables are presented in percentages. P-values are from t-test for normally distributed continuous data, from Kruskal-Wallis tests for non-normally distributed data and from chi-square testing for categorical data. Abbreviations: BMI: body mass index, ACS: acute coronary syndrome, PCI: percutaneous coronary intervention, CABG: coronary artery bypass grafting, CVA: cerebrovascular accident, PAD: peripheral arterial disease, COPD: chronic obstructive pulmonary disease, LVEF: left ventricular function, P2Y12: P2Y12 receptor antagonist, RAAS: renine-angiotensine-aldosteron system, CAD: coronary artery disease, PCI: percutaneous coronary intervention, NT pro-BNP: N-terminal pro-brain natriuretic peptide, IQR: interquartile range.

175


Chapter 9

vs. 55.3% in men, p=0.004). Also, women used diuretics significantly more often (38.5% vs. 26.1%, p<0.001). Men on the other hand had higher rates of hypercholesterolemia (48.9% vs. 42.7%, p=0.022) and a history of ACS (33.7% vs. 23.5%), PCI (32.3% vs. 22.5%) and CABG (14.3% vs. 5.8%). There were no major differences in indication to perform angiography between men and women, with stable complaints being the most frequent; 54.2% in men and 55.8% in women. MI (either STEMI or NSTEMI) was the indication in 29.4% of all men compared to 26.2% in women. Women were more likely to have a normal LVEF 67.2% vs. 54.6% in men, p<0.001. The prevalence of multi-vessel disease was higher in men (45.4% vs. 35.4%) whereas women more frequently had no coronary artery disease (13.0% vs. 4.3%, p for overall difference <0.001). Men, as a consequence, were more likely to undergo PCI than women (64.1% vs. 53.8%, p<0.001). SYNTAX score in patients with multi-vessel disease did not differ significantly, with 15 (IQR 10-21) for men and 13 (IQR 8-19.5) for women, p=0.220. Among patients presenting with MI, levels of hsTnI were not different between women (median 171.8, IQR 11.1-612.4) and men (median 94.5, IQR 12.6-813.3). NT pro-BNP levels were significantly higher in women (55.0 [IQR 9.8-161.8] vs. 91.3 [IQR 27.1-292.0]) in men, p=0.030. Sex differences in follow-up events The median duration of follow-up was 2.1 years for men (IQR 1.6-2.8) and 2.2 years for women (IQR 1.6-2.8). An overview of the number of events and 2-year Kaplan-Meier event rate estimates is displayed in Table 2. During follow-up of this study, a total of 74 men (6.4%) and 25 women (6.5%) died. Non-fatal MI occurred in 5.9% of all men and 6.6% of women (66 men vs. 28 women). The only significant difference in adverse event rate was the incidence of transient ischemic attack (TIA), which occurred in 9 women (2.1%), but only in 5 men (0.4%, p=0.002). The composite endpoint MACE, consisting of all-cause death, non-fatal MI, unplanned revascularization, stroke and admission for heart failure, occurred in 265 men and 103 women. Its overall incidence did not differ between men and women (20.7% in men vs. 21.3% in women, p=0.744). Stratification by indication for angiography We found a significantly higher occurrence of MACE in women who presented with MI, HR for female sex 1.66 (95%CI 1.10-2.50, p=0.015), Figure 2 (left panel). Among stable CAD patients however, women appeared to have a similar prognosis as men, (HR for female sex 0.80 (95%CI 0.58-1.10, p=0.163). A significant interaction was found between sex and indication for angiography for the occurrence of MACE (p=0.005) indicating that the impact of female sex on MACE occurrence significantly differed between the angiography indications stable CAD and MI. The sex difference in MACE occurrence was mainly driven by the higher occurrence of TIAs (3.8% vs. 0.5%, p=0.017) and rePCIs (8.7% vs. 5.6%, p=0.029) in women than men (supplemental table 1).

176


Sex Differences in Prognosis After Coronary Angiography

Table 2. Number of events and 2-year Kaplan-Meier estimates by sex. Event

N male

% male

N female

% female

p-value

MACE

265

20.7

103

21.3

0.744

Death

74

6.4

25

6.5

0.687

Cardiovascular death

35

2.8

9

2.4

0.346

Non-cardiovascular death

38

3.5

14

3.6

0.980

Non-fatal myocardial infarction

66

5.9

28

6.6

0.549

STEMI

11

1.1

5

0.8

0.712

NSTEMI

31

2.8

15

4.1

0.389 0.934

Unstable angina

26

2.1

10

2.4

108

9.4

36

7.6

0.563

New lesion

53

4.6

22

4.5

0.645

In-stent restenosis

60

5.1

20

4.5

0.650

CABG

11

0.8

3

0.6

0.622

Heart failure

38

3.0

15

3.7

0.859

CVA/TIA

20

2.0

16

3.6

0.019

CVA

15

1.6

8

1.8

0.408

Re-PCI

5

0.4

9

2.1

0.002

Heart or vascular intervention

59

5.0

18

4.6

0.439

Non-cardiac stent

25

2.3

6

1.8

0.328

Non-cardiac vascular surgery

17

1.5

6

1.4

0.902

6

0.4

1

0.2

0.438

TIA

Amputation due to PAD Hospital admission for PAD

1

0.1

0

0.0

0.541

Valve surgery

7

0.4

1

0.2

0.344

Percutaneous valve implantation

6

0.6

3

1.0

0.668 0.303

Device implantation

69

5.4

20

4.7

Heart rhythm disorder

47

3.9

11

2.4

0.154

Hemorrhagic event (extra cerebral)

18

1.5

13

2.3

0.064

Abbreviations: MACE= major adverse cardiovascular event, STEMI= ST-Segment Elevation Myocardial Infarction, NSTEMI= Non-ST-Segment Elevation Myocardial Infarction, PCI= Percutaneous Coronary Intervention, CABG= Coronary artery bypass graft, CVA= Cerebrovascular accident, TIA= Transient Ischemic Attack, PAD= peripheral arterial disease. Percentages: 2-year Kaplan Meier estimates. In 3 patients, cause of death was unknown. The sum of the number of subtypes of events (e.g. STEMI, NSTEMI, unstable angina) can exceed the number of “main� events (e.g. Non-fatal myocardial infarction), which can only be counted once in each patient. Only the first event of each patient is counted.

Stratification by severity of CAD When stratified for baseline angiographic severity of CAD, we found a significantly lower MACE-free survival probability in women than men with multi-vessel disease, HR for female sex 1.41 (95%CI 1.03-1.94, p=0.031), Figure 2 (right panel). This contrasted with women presenting with no, minor or single vessel CAD, where no difference in MACE between the sexes was observed, HR for female sex 0.85 (95%CI 0.61-1.19,

177


Chapter 9

p=0.347). The difference in MACE-free survival between men and women thus differed by angiographic CAD severity (p for interaction 0.031). The sex difference in MACE occurrence was mainly driven by the higher occurrence of heart failure admissions (9.1% vs. 2.7%, p=0.003) and CVA/TIAs (8.7% vs. 2.0%, p<0.001) in women than men (supplemental table 2).

MACE among Multi Vessel Disease patients

0

100

200

0.9 0.8

MACE-free Survival (Cox proportional hazard)

0.7

Men Women

0.5

Men Women

HR 1.43 [1.01-2.04], p=0.046 0.6

0.9 0.8 0.7 0.6

HR 1.61 [1.00-2.61], p=0.051

0.5

MACE-free Survival (Cox proportional hazard)

1.0

1.0

MACE among Myocardial Infarction patients

300

400

500

600

700

FU time (days)

800

900

1000

1100

1200

0

100

200

300

400

500

600

700

800

900

1000

1100

1200

FU time (days)

Figure 2. Multivariable corrected MACE occurrence from Cox regression analysis by sex among myocardial infarction patients and patients with angiographic multi-vessel disease. The left panel shows the multivariable adjusted sex differences in the occurrence of MACE among MI patients derived from Cox regression analysis. The right panel shows those differences among multi-vessel disease patients. MACE consists of all-cause mortality, MI, stroke, unplanned revascularization and admission for heart failure. The presented results were adjusted for: age, indication for angiography (not in MI analysis), angiographic CAD severity (not in multi-vessel disease analysis), hypertension, hypercholesterolemia, history of PCI, history of CABG, history of ACS, history of CVA, use of RAAS medication, history of PCI, history of CABG, history of ACS, history of CVA, use of RAAS medication, use of diuretics, use of P2Y12 inhibiting medication, use of statins, history of PAD, diabetes, kidney failure, treatment of CAD (conservative, PCI or CABG) and smoking status (non-smoker, ex smoker or current smoker).

Multivariable analysis The differences found with univariable analyses were adjusted for possible confounders. Hazard ratios for women presenting with MI remained 1.61 (95% CI 1.00-2.61, p= 0.051), for women with multi-vessel disease HR was 1.43 (95%CI 1.01-2.04, p=0.046). In order to further evaluate confounding factors we added NT pro-BNP levels, hsTnI levels, SYNTAX scores and LVEF measurements to the multivariable model one by one. With the addition of these parameters, statistical power decreased due to missing values, resulting in wider CIs and thus non-significant HR estimates. However, the point estimates of the HRs barely changed upon correction for any of these four parameters (HRs for female sex ranging from 1.53 to 1.83 among MI patients and from 1.33 to 1.56 among patients with multi-vessel disease) as can be observed from Figure 3.

178


Sex Differences in Prognosis After Coronary Angiography

Hazard ratio (95% CI) of Female Gender for MACE Univariable All Patients

*

Myocardial Infarction

*

Multi−vessel disease

Multivariable All Patients Myocardial Infarction Myocardial Infarction − hsTnI Myocardial Infarction − NT pro−BNP Myocardial Infarction − SYNTAX score Myocardial Infarction − LVEF

*

Multi−vessel disease Multi−vessel disease − hsTnI Multi−vessel disease − NT pro−BNP Multi−vessel disease − SYNTAX score Multi−vessel disease − LVEF

0

2

Reference is Male Gender (HR=1)

4

Figure 3. Hazard ratios with 95% confidence intervals are shown. The upper three estimates are from univariable Cox regression analysis. The remaining estimates are derived from multivariable Cox regression analysis adjusting for: age, indication for angiography (not in MI analysis), angiographic CAD severity (not in multi-vessel disease analysis), hypertension, hypercholesterolemia, history of PCI, history of CABG, history of ACS, history of CVA, use of RAAS medication, history of PCI, history of CABG, history of ACS, history of CVA, use of RAAS medication, use of diuretics, use of P2Y12 inhibiting medication, use of statins, history of PAD, diabetes, kidney failure, treatment of CAD (conservative, PCI or CABG) and smoking status (non-smoker, ex smoker or current smoker). Additionally, among MI patients and multi-vessel disease we added hsTnI, NT pro-BNP, SYNTAX score and LVEF, respectively to the model in order to observe changes in the HR estimate. * Indicates significant association of female sex with MACE (p<0.05).

179


Chapter 9

Comparison with the general population All-cause mortality in men presenting with complaints of stable CAD was higher than in the general population. The KM 2-year estimated all-cause mortality rate was 6.2% vs. 4.1% in the general population, p=0.014, Figure 4. For women with stable CAD, no difference in all-cause mortality was found (p=0.559). Women presenting with MI, on the other hand, showed a trend towards higher all-cause mortality than the general population group (KM 2-year estimated all-cause mortality rate 7.3% vs. 3.2%, p=0.071), whereas men with MI had similar survival rates as their counterparts from the general population (KM 2-year estimated all-cause mortality rate 4.6% vs. 4.1%, p=0.315). All-cause mortality rates for both men and women with no or minor CAD were comparable with the general population, (KM 2-year estimated all-cause mortality rate 5.2% vs. 4.1%, p=0.196 in men and 5.0% vs. 3.2%, p=0.111 in women). Both men and women with multivessel disease had a significantly worse survival than their general population counterparts (KM 2-year estimated all-cause mortality rate 9.2% vs. 4.1%, p=0.002 in men and 11.5% vs. 3.2%, p=0.004 in women). For the other indications and severities of CAD there was no difference in survival when compared to the general population.

Discussion Women presenting with MI or with multi-vessel disease at angiography had a higher occurrence of MACE as compared to men. Among women presenting with MI, mortality was higher than in the general population. Surprisingly, this was not the case for men. On the contrary, men with stable complaints had higher mortality compared to the general population, whereas women with stable CAD showed similar mortality. Myocardial infarction As long as sex-specific differences after MI have been studied, reports of women having a worse prognosis were published19–21. Even though the risk factor burden changed the last years, our study still reveals a more disadvantageous prognosis for women than for men. A persisting sex difference in delay among patients presenting with MI might be part of the cause. Both patient delay (time from symptom onset to first medical contact) and doctors delay (door-to-balloon time) have been reported to contribute to a poorer prognosis in women8,22,23. Several studies reported that the sex difference in delay can be as long as one hour20. Longer delays might result in greater loss of myocardium and consequently lower LVEF, higher hsTnI levels and higher NT pro-BNP levels. Therefore, we additionally adjusted the multivariable Correction for these factors in a multivariable Cox regression model (Figure 4) did not change our findings, suggesting that loss of myocardium did not account for the observed differences between men and women. Both laboratory tests have several drawbacks. Higher NT pro-BNP levels in women and elderly are common24,25 and hsTnI release depends on left ventricular mass.26 Therefore, the role of tissue loss and consequent diminished LVEF cannot entirely be excluded in explaining the worse prognosis in women.

180


Sex Differences in Prognosis After Coronary Angiography

Men with Stable CAD vs. General population

1.00

Women with General p

1.00

0.95

0.95

MACE-free survival

p = 0.014 0.90

0.90

0.85

0.85

0.80

0.80

Men, General Population

Women, General Population

Men, Stable CAD 0.75

General population

0

100

200

300

Women, Stable CAD 400

500

Time (days)

600

700

800

900

1000

6964

6919

6884

6857

6808

6765

6725

6689

6654

6619

6586

971

964

880

776

693

623

545

452

368

288

212

Stable CAD

0.75

6545General 6520 population 147Stable CAD 74

0

100

200

2575

2565

2555

2543

375

374

339

299

268

0.95

p = 0.014

p = 0.558 0.90

0.85

0.80

Women, General Population Women, Stable CAD 700

800

900

1000

6689

6654

6619

6586

452

368

288

212

0.75

6545General 6520 population 147Stable CAD 74

0

100

200

300

400

500

Time (days)

500

Time (d

25

2

Numb

Women with Stable CAD vs. General population

1.00

400

2588

Numbers at risk

D vs. on

300

600

700

800

900

1000

2588

2575

2565

2555

2543

2533

2518

2510

2490

2467

2450

375

374

339

299

268

243

203

178

136

117

87

Numbers at risk

Figure 4. Kaplan-Meier plots of all-cause mortality among stable CAD patients and the general age and sex matched Dutch population. On the upper panel men with stable CAD are depicted in dark blue, general population men in light blue (significant difference between the two groups, p=0.014). On the bottom panel women with stable CAD are depicted in dark red, women from the general population are in pink (no significant difference between the two groups, p=0.558).

181


Chapter 9

Multi-vessel disease Women with multi-vessel disease show higher mortality rates than men even after correcting for baseline differences. This does not seem to be a result of women in this group having more severe CAD, since SYNTAX scores (quantification scores of angiographic CAD severity) were not significantly different between the sexes among patients with multi-vessel disease: median SYNTAX score 15.0 [IQR 10.0-21.0] in men, 13.0 [IQR 8.0-19.5] in women, p=0.220. Women have been reported to have a smaller and stiffer vasculature in general, resulting in a reduced reserve capacity to supply the endangered myocardium when necessary4. Theories of women more often having microvascular dysfunction on top of epicardial CAD than men might explain the greater occurrence of events8,27,28. Hypothetically, it might even be so that with progression of epicardial disease, microvascular disease progresses as well in women and thereby even further increasing the myocardium at risk. While in men the disease might be more restricted to the larger coronary arteries, which are accessible for intervention. This hypothesis fits well with our data, as we show an increase in MACE occurrence with increasing severity of epicardial CAD in women but not in men. This theory warrants further investigation on the etiology, diagnostic tools and treatment of microvascular dysfunction. Another explanation might lie in a higher prevalence of diastolic heart failure among women.29 In our cohort, women with multi-vessel disease have a higher incidence of heart failure admissions after coronary angiography then men. At baseline, EF is more likely to be normal and NTproBNP levels are higher in women than in men, suggesting that heart failure with preserved EF30 might be the problem. Both for women with MI and women with multi-vessel disease an index event bias31 is not unlikely. Female sex is protective for the occurrence of a first cardiovascular event. However, when women do develop severe CAD, e.g. MI or multi-vessel disease, in our cohort their prognosis appears to be worse than men’s. Comparison with general population Interestingly, men with stable complaints have a lower survival probability than men from the general population, however, no such difference is observed among women. This may be due to the fact that men with stable complaints have angiographically more severe CAD than women, demonstrated by higher SYNTAX scores13. Women present with stable complaints more often than men, but these complaints are not necessarily of a cardiac origin. For women presenting with stable complaints 16% have no CAD, whereas this is only 4.5% in men. Patients with stable complaints as indication for coronary angiography not only comprised patients with stable angina, but also patients with diagnostic results suspicious for CAD. Noninvasive testing for CAD is of limited predictive accuracy in women, especially due to many false positive tests3,4,32, resulting in more invasive testing and consequently finding more women without significant epicardial CAD upon angiography than men. This phenomenon could also be applicable to our study population, and might be explain why women with no or

182


Sex Differences in Prognosis After Coronary Angiography

minor CAD have a similar prognosis compared to the general population, which is in contrast to results observed in the US where both men and women without obstructive CAD show a poorer prognosis8,33. To avoid unnecessary procedural risk of coronary angiography, a noninvasive diagnostic test with a high negative predictive value is needed. Possibly, coronary computed tomography34 can assist in this need. Strengths and limitations Our longitudinal observational cohort has a large sample size and a sufficient number of events during follow-up, enabling stratified analysis. Lost to follow-up was limited, comprising of only 4 patients (0.2%). Since patients with diverse indications were studied, our cohort presents a valuable reflection of daily clinical practice in a tertiary center. Since we only included patients who provided written informed consent, there is a possibility of selection bias in this study towards inclusion of less severe cases, who are more willing to participate in research. All-cause mortality in our cohort was compared with the Dutch general population; this endpoint cannot be misclassified. However, the composite endpoint MACE also comprised other cardiovascular events which could have been more subjected to the clinical judgment of the treating cardiologist, e.g. for the diagnosis of unstable angina. Nonetheless, the greater part of MACE involved all-cause death, MI and re-PCI, which are very relevant and less arguable cardiovascular events. Clinical implications Clinicians should be aware of a worse long-term prognosis for women presenting with MI or women who have multi-vessel disease at angiography. An accurate prediction of long-term prognosis for both patient and clinician is of great importance, since this could have implications for the eventual treatment. When it comes to medication therapy, men tend to be treated more aggressively than women, experience less side effects and tend to be more compliant to therapy28,35–38. A recent meta-analysis about statin compliance, notorious for its side-effects, showed that women were 10% less likely to be adherent to statin therapy than men39. Lower compliance could be a factor leading to higher adverse event rates. Since guidelines advise a conservative treatment in women more often than in men, e.g. low-risk women presenting with NSTEMI40, non-compliance to medication might have greater consequences for women. Medication therapy should be optimized in women and clinicians should stress the importance of medication compliance. Future perspectives In order to understand these sex-specific differences, further research enrolling large numbers of women is of utmost importance. Primary and secondary prevention should be optimized specifically in high-risk women in order to reduce their risk of recurrent adverse events. On the other hand, in order to protect the low-risk women from possible procedure-related complications of unnecessary diagnostic coronary

183


Chapter 9

angiography, biomarkers or other non-invasive tools indicative of the severity of CAD in women would be extremely helpful. Conclusion During a median follow-up duration of 2.1 years after coronary angiography, women presenting with MI or who had multi-vessel disease at angiography had a higher occurrence of MACE than men, also when adjusted for potential confounders. For women presenting with MI, all-cause mortality was higher as compared to the general population whereas men with MI did not differ from the general population. This was reversed for stable CAD patients, where men had higher mortality rates than the general population, but women did not. Funding This work was supported by a grant from the Netherlands Heart Foundation: 2013T084, Queen of Hearts: Improving diagnosis of CVD in women to Hester den Ruijter. Dominique de Kleijn is funded through a strategic grant from the Royal Netherlands Academy of Arts and Sciences to the Interuniversity Cardiology Institute of the Netherlands, ICIN, the National University Singapore Startup grant, the Singapore National Medical Research Council Centre Grant and the ATTRaCT, SPF 2014/003 grant BMRC. Folkert Asselbergs is supported by the UCL Hospitals NIHR Biomedical Research Centre and a Dekker scholarship-Junior Staff Member 2014T001 – Dutch Heart Foundation. These funding sources in no way influenced the analyses or the content of this manuscript. Acknowledgments We sincerely acknowledge the outstanding logistical support to the UCORBIO cohort by Ms Jonne Hos.

184


Sex Differences in Prognosis After Coronary Angiography

References 1.

Organization WH. Global status report on noncommunicable diseases 2014. Geneva, Switzerland: 2014.

2.

Appelman Y, van Rijn BB, Ten Haaf ME, Boersma E, Peters S a E. Sex differences in cardiovascular risk factors and disease prevention. Atherosclerosis. 2015;1–8.

3.

Shaw LJ, Bairey Merz CN, Pepine CJ, Reis SE, Bittner V, Kelsey SF, Olson M, Johnson BD, Mankad S, Sharaf BL, Rogers WJ, Wessel TR, Arant CB, Pohost GM, Lerman A, Quyyumi A a, Sopko G. Insights from the NHLBISponsored Women’s Ischemia Syndrome Evaluation (WISE) Study: Part I: gender differences in traditional and novel risk factors, symptom evaluation, and gender-optimized diagnostic strategies. J Am Coll Cardiol. 2006;47:S4–S20.

4.

Jacobs AK. Coronary intervention in 2009: are women no different than men? Circ Cardiovasc Interv. 2009;2:69–78.

5.

Mozaffarian D, Benjamin EJ, Go AS, Arnett DK, Blaha MJ, Cushman M, de Ferranti S, Després J-P, Fullerton HJ, Howard VJ, Huffman MD, Judd SE, Kissela BM, Lackland DT, Lichtman JH, Lisabeth LD, Liu S, Mackey RH, Matchar DB, McGuire DK, Mohler ER, Moy CS, Muntner P, Mussolino ME, Nasir K, Neumar RW, Nichol G, Palaniappan L, Pandey DK, Reeves MJ, Rodriguez CJ, Sorlie PD, Stein J, Towfighi A, Turan TN, Virani SS, Willey JZ, Woo D, Yeh RW, Turner MB. Heart Disease and Stroke Statistics-2015 Update: A Report From the American Heart Association. 2014.

6.

Mosca L. Fifteen-Year Trends in Awareness of Heart Disease in Women: Results of a 2012 American Heart

7.

Casanova G, Bossardi Ramos R, Ziegelmann P, Spritzer PM. Effects of low-dose versus placebo or

Association National Survey. AHA. 2013;127:1–21.

conventional-dose postmenopausal hormone therapy on variables related to cardiovascular risk: a systematic review and meta-analyses of randomized clinical trials. J Clin Endocrinol Metab. 2015;100:1028–37. 8.

Bairey Merz CN, Shaw LJ, Reis SE, Bittner V, Kelsey SF, Olson M, Johnson BD, Pepine CJ, Mankad S, Sharaf BL, Rogers WJ, Pohost GM, Lerman A, Quyyumi A a., Sopko G. Insights from the NHLBI-sponsored Women’s Ischemia Syndrome Evaluation (WISE) study. Part II: Gender differences in presentation, diagnosis, and outcome with regard to gender-based pathophysiology of atherosclerosis and macrovascular and microvascular cor. J Am Coll Cardiol. 2006;47.

9.

Den Ruijter H, Pasterkamp G, Rutten FH, Lam CSP, Chi C, Tan KH, van Zonneveld a J, Spaanderman M, de Kleijn DP V. Heart failure with preserved ejection fraction in women: the Dutch Queen of Hearts program. Neth Heart J. 2015;23:89–93.

10. Gierach GL, Johnson BD, Bairey Merz CN, Kelsey SF, Bittner V, Olson MB, Shaw LJ, Mankad S, Pepine CJ, Reis SE, Rogers WJ, Sharaf BL, Sopko G. Hypertension, menopause, and coronary artery disease risk in the Women’s Ischemia Syndrome Evaluation (WISE) Study. J Am Coll Cardiol. 2006;47:S50–8. 11. Mosca L, Barrett-Connor E, Wenger NK. Sex/gender differences in cardiovascular disease prevention: what a difference a decade makes. Circulation. 2011;124:2145–54. 12. Nichols M, Townsend N, Scarborough P, Rayner M. Trends in age-specific coronary heart disease mortality in the European Union over three decades: 1980-2009. Eur Heart J. 2013;34:3017–27. 13. Gijsberts CM, Gohar A, Ellenbroek GHJM, Hoefer IE, de Kleijn DP V, Asselbergs FW, Nathoe HM, Agostoni P, Rittersma SZH, Pasterkamp G, Appelman Y, den Ruijter HM. Severity of stable coronary artery disease and its biomarkers differ between men and women undergoing angiography. Atherosclerosis. 2015;241:234–240.

185


Chapter 9

14. Reitsma JB, Kardaun JWPF, Gevers E, De Bruin A, Van Der Wal J, Bonsel GJ. Mogelijkheden voor anoniem follow-uponderzoek van patiënten in landelijke medische registraties met behulp van de Gemeentelijke Basisadministratie: Een pilotonderzoek. Ned Tijdschr Geneeskd. 2003;147:2286–2290. 15. Harris PJ, Behar VS, Conley MJ, Harrell FE, Lee KL, Peter RH, Kong Y, Rosati R a. The prognostic significance of 50% coronary stenosis in medically treated patients with coronary artery disease. Circulation. 1980;62:240–248. 16. Sianos G, Morel M, Kappetein AP, Morice M, Colombo A, Dawkins K, van den Brand M, Van Dyck N, Russell ME, Mohr FW, Serruys PW. The SYNTAX Score: an angiographic tool grading the complexity of coronary artery disease. EuroIntervention. 2005;1:219–27. 17. SYNTAX Steering Committee. SYNTAX Score Calculator [Internet]. 2012;Available from: http://www. syntaxscore.com/calc/start.htm 18. R Core Team. R: A Language and Environment for Statistical Computing. Vienna, Austria: 2014. 19. Steg PG, Greenlaw N, Tardif J-C, Tendera M, Ford I, Kääb S, Abergel H, Fox KM, Ferrari R. Women and men with stable coronary artery disease have similar clinical outcomes: insights from the international prospective CLARIFY registry. Eur Heart J. 2012;33:2831–40. 20. Tomey MI, Mehran R, Brener SJ, Maehara A, Witzenbichler B, Dizon JM, El-Omar M, Xu K, Gibson CM, Stone GW. Sex, adverse cardiac events, and infarct size in anterior myocardial infarction: an analysis of intracoronary abciximab and aspiration thrombectomy in patients with large anterior myocardial infarction (INFUSE-AMI). Am Heart J. 2015;169:86–93. 21. Van der Meer MG, Nathoe HM, van der Graaf Y, Doevendans PA, Appelman Y. Worse outcome in women with STEMI: a systematic review of prognostic studies. Eur J Clin Invest. 2015;45:226–235. 22. Sullivan AL, Beshansky JR, Ruthazer R, Murman DH, Mader TJ, Selker HP. Factors associated with longer time to treatment for patients with suspected acute coronary syndromes: a cohort study. Circ Cardiovasc Qual Outcomes. 2014;7:86–94. 23. Ferrante G, Corrada E, Belli G, Zavalloni D, Scatturin M, Mennuni M, Gasparini GL, Bernardinelli L, Cianci D, Pastorino R, Rossi ML, Pagnotta P, Presbitero P. Impact of female sex on long-term outcomes in patients with ST-elevation myocardial infarction treated by primary percutaneous coronary intervention. Can J Cardiol. 2011;27:749–55. 24. Wang TJ, Larson MG, Levy D, Leip EP, Benjamin EJ, Wilson PW., Sutherland P, Omland T, Vasan RS. Impact of age and sex on plasma natriuretic peptide levels in healthy adults. Am J Cardiol. 2002;90:254–258. 25. Redfield MM, Rodeheffer RJ, Jacobsen SJ, Mahoney DW, Bailey KR, Burnett JC. Plasma brain natriuretic peptide concentration: impact of age and gender. ACC Curr J Rev. 2003;12:44. 26. Fernandez-Jimenez R, Silva J, Martinez-Martinez S, Lopez-Maderuelo MD, Nuno-Ayala M, Garcia-Ruiz JM, Garcia-Alvarez a., Fernandez-Friera L, Pizarro TG, Garcia-Prieto J, Sanz-Rosa D, Lopez-Martin G, FernandezOrtiz a., Macaya C, Fuster V, Redondo JM, Ibanez B. Impact of Left Ventricular Hypertrophy on Troponin Release During Acute Myocardial Infarction: New Insights From a Comprehensive Translational Study. J Am Heart Assoc. 2015;4:e001218–e001218. 27. Hemingway H, Langenberg C, Damant J, Frost C, Pyörälä K, Barrett-Connor E. Prevalence of angina in women versus men: a systematic review and meta-analysis of international variations across 31 countries. Circulation. 2008;117:1526–36. 28. Shaw LJ. Women and Ischemic Heart Disease: Evolving Knowledge. JACC. 2009;54:1561–1575. 29. Ruijter HM den, Haitjema S, W Asselbergs F, Pasterkamp G. Sex matters to the heart: A special issue dedicated to the impact of sex related differences of cardiovascular diseases. Atherosclerosis. 2015;241:205–7.

186


Sex Differences in Prognosis After Coronary Angiography

30. Colvin M, Sweitzer NK, Albert NM, Krishnamani R, Rich MW, Stough WG, Walsh MN, Westlake Canary C a., Allen L a., Bonnell MR, Carson PE, Chan MC, Dickinson MG, Dries DL, Ewald G a., Fang JC, Hernandez AF, Hershberger RE, Katz SD, Moore S, Rodgers JE, Rogers JG, Vest AR, Whellan DJ, Givertz MM. Heart Failure in Non-Caucasians, Women, and Older Adults: A White Paper on Special Populations From the Heart Failure Society of America Guideline Committee. J Card Fail. 2015;21:674–93. 31. Dahabreh IJ. Index Event Bias as an Explanation for the Paradoxes of Recurrence Risk Research. JAMA. 2011;305:822. 32. Lansky AJ, Hochman JS, Ward P a, Mintz GS, Fabunmi R, Berger PB, New G, Grines CL, Pietras CG, Kern MJ, Ferrell M, Leon MB, Mehran R, White C, Mieres JH, Moses JW, Stone GW, Jacobs AK. Percutaneous coronary intervention and adjunctive pharmacotherapy in women: a statement for healthcare professionals from the American Heart Association. Circulation. 2005;111:940–53. 33. Murthy VL, Naya M, Taqueti VR, Foster CR, Gaber M, Hainer J, Dorbala S, Blankstein R, Rimoldi O, Camici PG, Di Carli MF. Effects of sex on coronary microvascular dysfunction and cardiac outcomes. Circulation. 2014;129:2518–27. 34. Vogler N, Meyer M, Fink C, Schoepf U, Schönberg S, Henzler T. Predictive Value of Zero Calcium Score and Low-End Percentiles for the Presence of Significant Coronary Artery Stenosis in Stable Patients with Suspected Coronary Artery Disease. RöFo - Fortschritte auf dem Gebiet der Röntgenstrahlen und der Bildgeb Verfahren. 2013;185:726–732. 35. Blomkalns AL, Chen AY, Hochman JS, Peterson ED, Trynosky K, Diercks DB, Brogan GX, Boden WE, Roe MT, Ohman EM, Gibler WB, Newby LK. Gender disparities in the diagnosis and treatment of non-ST-segment elevation acute coronary syndromes: large-scale observations from the CRUSADE (Can Rapid Risk Stratification of Unstable Angina Patients Suppress Adverse Outcomes With Early Implementatio. J Am Coll Cardiol. 2005;45:832–7. 36. Brown MT, Bussell JK. Medication adherence: WHO cares? Mayo Clin Proc. 2011;86:304–14. 37. Radovanovic D, Erne P, Urban P, Bertel O, Rickli H, Gaspoz J-M. Gender differences in management and outcomes in patients with acute coronary syndromes: results on 20,290 patients from the AMIS Plus Registry. Heart. 2007;93:1369–75. 38. Vaughan CJ, Gotto AM. Update on statins: 2003. Circulation. 2004;110:886–92. 39. Lewey J, Shrank WH, Bowry ADK, Kilabuk E, Brennan T a, Choudhry NK. Gender and racial disparities in adherence to statin therapy: A meta-analysis. Am Heart J. 2013;165:665–678.e1. 40. Amsterdam E a, Wenger NK, Brindis RG, Casey DE, Ganiats TG, Holmes DR, Jaffe AS, Jneid H, Kelly RF, Kontos MC, Levine GN, Liebson PR, Mukherjee D, Peterson ED, Sabatine MS, Smalling RW, Zieman SJ. 2014 AHA/ ACC Guideline for the Management of Patients With Non-ST-Elevation Acute Coronary Syndromes: A Report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines. 2014.

187


Chapter 9

Supplemental Supplemental table 1. Number of events and 2-year Kaplan-Meier estimates by sex for patients undergoing coronary angiography for myocardial infarction. Event

N male

% male

N female

% female

MACE

67

18.1

35

28.8

0.014

Death

18

5.3

9

9.2

0.298

Cardiovascular death

p-value

7

1.8

5

4.6

0.178

Non-cardiovascular death

10

3.2

3

3.1

0.905

Non-fatal myocardial infarction

25

7.7

12

10.9

0.267

STEMI

6

2.0

2

1.6

0.991

NSTEMI

9

3.0

7

8.4

0.071

Unstable angina

11

2.9

5

3.9

0.568

Re-PCI

34

10.3

16

13.5

0.202

New lesion

17

5.6

12

8.7

0.029

In-stent restenosis

18

5.0

9

9.4

0.304

CABG

4

1.1

1

0.8

0.795

Heart failure

5

1.5

5

4.7

0.062

CVA/TIA

6

1.3

6

5.4

0.044

CVA

4

0.8

2

1.6

0.636

TIA

2

0.5

4

3.8

0.017

15

4.4

7

7.0

0.447

6

1.6

4

3.9

0.269 0.428

Heart or vascular intervention Non-cardiac stent Non-cardiac vascular surgery

3

1.0

2

2.3

Amputation due to PAD

2

0.5

0

0.0

0.413

Hospital admission for PAD

1

0.3

0

0.0

0.563

Valve surgery

1

0.5

0

0.0

0.571

Percutaneous valve implantation

2

0.5

1

1.5

0.734

10

3.3

2

2.3

0.508

9

2.4

3

3.1

0.998

10

2.7

4

2.4

0.769

Device implantation Heart rhythm disorder Hemorrhagic event (extra cerebral)

188


Sex Differences in Prognosis After Coronary Angiography

Supplemental table 2. Number of events and 2-year Kaplan-Meier estimates by sex for patients with multivessel disease on coronary angiography. Event

N male

% male

N female

% female

p-value

MACE

136

24.1

54

34.0

0.030

Death

35

7.3

11

8.5

0.875

Cardiovascular death

16

3.0

7

5.4

0.381

Non-cardiovascular death

19

4.3

3

2.2

0.303

Non-fatal myocardial infarction

42

8.2

11

6.8

0.726

STEMI

7

1.5

0

0.0

0.148

NSTEMI

22

4.3

9

6.0

0.388

Unstable angina

15

2.5

3

1.8

0.557

Re-PCI

63

12.4

16

10.8

0.594

New lesion

31

6.1

13

8.6

0.240

In-stent restenosis

35

6.5

7

5.1

0.336

CABG Heart failure CVA/TIA

7

1.3

3

1.8

0.571

16

2.7

13

9.1

0.003

9

2.0

12

8.7

<0.001

CVA

5

1.3

7

5.0

0.003

TIA

4

0.7

6

4.8

0.005

Heart or vascular intervention

31

5.6

12

8.4

0.380

Non-cardiac stent

12

2.4

5

3.8

0.510 0.742

Non-cardiac vascular surgery

9

1.6

2

1.2

Amputation due to PAD

4

0.5

1

0.6

0.914

Hospital admission for PAD

1

0.2

0

0.0

0.588

Valve surgery

2

0.0

1

0.6

0.583

Percutaneous valve implantation

4

1.0

1

1.0

0.868

Device implantation

25

3.9

5

3.8

0.437

Heart rhythm disorder

21

3.6

2

2.0

0.104

Hemorrhagic event (extra cerebral)

13

2.4

6

3.0

0.339

189



PART TWO Sex Differences

Chapter 10 Gender Differences in Health-related Quality of Life in Patients Undergoing Coronary Angiography Open Heart. 2015 Aug 27;2(1):e000231

Crystel M. Gijsberts, Pierfrancesco Agostoni, Imo E. Hoefer, Folkert W. Asselbergs, Gerard Pasterkamp, Hendrik Nathoe, Yolande E. Appelman, Dominique P.V. de Kleijn, Hester M. den Ruijter


Chapter 10

Abstract Background Health-related quality of life (HRQOL) reflects the general well-being of individuals. In coronary artery disease (CAD) patients HRQOL is compromised. Female CAD patients have been reported to have lower HRQOL. In this study we investigate gender differences in HRQOL and in associations of patient characteristics with HRQOL in coronary angiography (CAG) patients. Methods We cross-sectionally analyzed patients from the Utrecht Coronary Biobank, undergoing CAG. All patients filled in an HRQOL questionnaire (RAND-36 and EuroQoL) upon inclusion. RAND-36 and EuroQoL HRQOL measures were compared between the genders across indications for CAG, CAD severity and treatment of CAD. RAND-36 HRQOL measures were compared to the general Dutch population. Additionally, we assessed interactions of gender with patient characteristics in their association with HRQOL (EuroQoL). Results We included 1,421 patients (1,020 men and 401 women) with a mean age of 65 in our analysis. Women reported lower HRQOL measures than men (mean EuroQoL self-rated health grade 6.84Âą1.49 in men, 6.46Âą1.40 in women, p <0.001). The reduction in RAND36 HRQOL as compared to the general Dutch population was larger in women than in men. From regression analysis we found that diabetes, a history of cardiovascular disease and complaints of shortness of breath determined HRQOL (EuroQoL) more strongly in men than in women. Conclusions Women reported lower HRQOL than men throughout all indications for CAG and regardless of CAD severity or treatment. As compared to the general population, the reduction in HRQOL was more extreme in women than in men. Evident gender differences were found in determinants of HRQOL in patients undergoing CAG, which deserve attention in future research.

192


Sex Differences in HRQOL

Introduction As survival of coronary artery disease (CAD) patients keeps improving, their health-related quality of life (HRQOL) is of high relevance. A poor HRQOL is related to higher health care expenditure1, therefore it has become an increasingly important point of interest of physicians. More and more frequently HRQOL is used as an outcome measure in clinical research to assess, for example, the effect of treatments.2 HRQOL has been reported to be associated with several cardiovascular risk factors. For example obesity3,4, diabetes5 and smoking6 all have been linked to a diminished HRQOL, but up till now it is unclear whether this holds true for both men and women, as the majority of the cardiovascular disease (CVD) research has focused on men. However, evidence has emerged in primary care and primary prevention settings which shows differences in the association of risk factors with HRQOL between men and women (e.g. for obesity7, diabetes8 and smoking9). While CVD has long been considered a men’s disease, global mortality from CVD is equal between men and women.10 In the U.S. the CVD mortality rate for women even exceeds that of men.11 Furthermore, in population studies angina pectoris is more prevalent among women (6.7%) than men (5.7%)12 and women suffer from longer delays in the case of suspected acute coronary syndrome.13,14 Women who eventually undergo percutaneous coronary intervention (PCI)15,16 or coronary artery bypass grafting (CABG)17 report lower HRQOL. Up till now it is unknown whether all women who undergo coronary angiography (CAG) with or without PCI express the same low HRQOL scores, or if low HRQOL scores are found in particular subgroups of the female CAG patients. Increased acknowledgement and understanding of gender differences in patient characteristics that determine low HRQOL may lead to better treatment and more personalized care. Therefore, firstly, we investigated gender differences in reported HRQOL in a CAG population. Secondly, we examined gender differences in HRQOL across the indications for CAG, the angiographic severity of CAD and across the treatment strategies of CAD. Per gender, we also looked into differences in HRQOL scores across CAG indication, severity and treatment of CAD. Thirdly, we evaluated the difference in HRQOL between men and women undergoing CAG as compared to the general Dutch population, in order to evaluate gender discrepancies in the difference with the general population. Finally, we hypothesized that patient characteristics and angina complaints were associated with lower HRQOL in dissimilar ways between men and women undergoing CAG.

193


Chapter 10

Methods Patient selection We performed a cross-sectional study in the Utrecht Coronary Biobank (UCORBIO) cohort, registered under clinicaltrials.gov ID: NCT02304744. This ongoing biobank started enrolment in October 2011. All patients that were enrolled between October 2011 and March 2014 were included in the current analyses. All patients entering the catheterization laboratories of the University Medical Centre in Utrecht (the Netherlands) were asked to participate in this biobank. Hereby patients with all indications for CAG were included. The study has been approved by the medical ethical committee of the University Medical Centre Utrecht (registration code 11-183) and all patients provided written informed consent. The only exclusion criterion was age <18 years. Per study protocol, all patients who provided written informed consent were provided with an HRQOL questionnaire. Patients who returned the HRQOL questionnaire (described in more detail below) were considered for this study. The process of patient recruitment and selection is depicted in Figure 1. Between October 2011 and March 2014 3,405 patients underwent CAG, of whom 2,268 were asked to participate in UCORBIO. A total of 1,993 patients provided informed consent, of whom 1,421 returned the HRQOL questionnaire. Thus, the response rate was 71.6%. The response rate did not differ between men and women. Data collection HRQOL questionnaire Our HRQOL questionnaire contained the RAND-36 questionnaire version 1, consisting of 9 domains of HRQOL: physical functioning, mental functioning, social functioning, physical role limitations, emotional role limitations, pain, vitality, general health and health change. The internal consistency of this questionnaire has previously been established to be high with Cronbach’s alpha ranging from 0.71-0.92 across the domains.18 We extended the RAND-36 questionnaire with the self-rated health grading question derived from the EuroQoL19 questionnaire (“Please indicate how good or bad your health is on a scale from 0 (worst) to 10 (best)”). The questionnaire was handed to the patients directly after CAG. Patients were instructed to fill in the questionnaire according to their situation prior to CAG. Patients were allowed to take home and send back the questionnaire. The median time between CAG and return of the study number coded questionnaire to the research office (by mail) was 112 days. Medical records At baseline, demographical data, history of CVD (either previous acute coronary syndrome, previous PCI, previous CABG, cerebrovascular accident or peripheral arterial disease), medication use, cardiovascular risk factors (diabetes, smoking, BMI, hypertension, hypercholesterolemia and family history of CVD), and clinical data concerning the

194


!

Sex Differences in HRQOL Figure'1.'Flow!chart!of!patient!recruitment!and!selection!process.!

Eligible'angiography'pa/ents'between' 141042011'and'14342014:'' 3,405'

Approached'by'study'team'for' par/cipa/on:'' 2,268' Provided'wriFen'informed'consent:' 1,993' Returned'ques/onnaire:'' 1,421' Included'in'analysis'

Men:'' 1,020'

Women:'' 401'

!

Figure 1. Flow chart of patient recruitment and selection process.

indication for CAG, the angiographic severity of CAD and details from the procedure were collected from the medical records. The indication for catheterization was grouped into stable CAD (stable angina, dyspnea on exertion or silent myocardial ischemia), unstable angina, myocardial infarction (nonST-elevation (NSTEMI) or ST-elevation myocardial infarction (STEMI)) and other indications (mostly screening for valve surgery). The angiographic severity of CAD was determined by the number of epicardial vessels with an angiographic stenosis of >50% based on visual assessment or with a significant intracoronary fractional flow reserve (FFR) measurement (<0.75). The angiographic severity of CAD was grouped into four groups: normal coronaries (no or minor CAD with <50% stenosis), single vessel disease, double vessel disease and triple vessel disease. ! Anginal complaints questionnaire The characteristics of anginal complaints were obtained through a patient questionnaire, based on the Rose cardiovascular questionnaire.20 This questionnaire was combined with the HRQOL questionnaire and thus sent and returned at the same time as the HRQOL

195


Chapter 10

questionnaire. Six questions were asked: type of complaints (chest pain, dyspnea, no complaints or other), progression of complaints (yes or no), circumstances of complaints (exercise, cold, emotion, in absence of exercise, cold or emotion), start of complaints (more than 10 years ago, five to ten years ago, one to five years ago, less than one year ago), last occurrence of complaints (more than one year ago, one year to one month ago, one month to one week ago, last week) and limitations due to complaints (none, mild, severe, no activity possible). General Dutch population Data on HRQOL scores in the general Dutch population were derived from the Dutch manual on the RAND-36 questionnaire, presenting the scores of 1,036 randomly selected adult test subjects aged between 18-89 (mean 44.1). From the general population, separately, age-specific HRQOL means were reported and gender-specific means were reported by van der Zee et al.21 Therefore, we could only perform a gender-matched comparison to a general population sample that was younger than our study population. To our knowledge, no EuroQoL means are available for the general Dutch population. Computation of RAND-36 HRQOL scores For the interpretation of the RAND-36 data the SPSS syntax provided by the University of Groningen22 was used. This syntax calculates one score for each domain, composed of several domain items. The physical functioning domain consisted of 10 items each to be scored from 1 to 3. The social functioning domain consisted of 2 domains, to be scored from 1 to 5. The physical role functioning domain consisted of 4 items to be scored 1 or 2, emotional role functioning consisted of 3 items to be scored 1 or 2. Mental functioning comprised 5 questions to be scored 1 to 6, vitality comprised 4 questions to be graded 1 to 6. The pain domain consisted of 2 questions, the first to be scored 1 to 6, the second to be scored 1 to 5. The general health domain comprised 5 questions, to be rated 1 to 5 and health change was a single item to be scored 1 to 5. The total domain score was calculated when at least half of the items were answered. First, a mean score per domain was calculated. Subsequently, the mean score was transformed to a percentage of the highest possible score on each domain. Those percentages were used in the current analysis. Data analysis Data analysis was performed using IBM SPSS Statistics, version 20 and R software for statistical computing, version 3.1.223. Continuous data were presented as means ¹ standard deviation (SD) when normally distributed. Non-normally distributed continuous data were presented as median with interquartile range (IQR). Categorical data were presented as percentages per category and were compared using a chi-square test. Means were compared using a t-test for normally distributed data. Significance level was set at ι <0.05. Missings were deleted listwise. Questionnaire consistency was tested with Cronbach’s alpha for the RAND-36 questionnaire.

196


Sex Differences in HRQOL

Our primary analysis consisted of univariable comparison of HRQOL between men and women. These gender differences in the domains of RAND-36 HRQOL were compared using a Mann-Whitney U test, as these data were non-normally distributed. Also, in order to get a better idea of the gender differences in HRQOL (RAND-36 and EuroQoL selfrated health grade), we stratified the analyses by CAD severity, indication for CAG and treatment of CAD and looked into differenced in HRQOL scores across indication, severity and treatment of CAD using Kruskal-Wallis testing with bonferroni post-hoc pairwise comparisons. Plus, we compared our cohort to the general Dutch population HRQOL scores. Only means and SDs of the general population scores were available and thus these were used (inappropriately, as they are non-normally distributed) to perform a t-test between the general population and our patient sample. Our secondary analysis consisted of univariable and multivariable regression analysis of patient characteristics (cardiovascular risk factors and complaint characteristics) associated with EuroQoL HRQOL, with interaction terms for gender. We chose the EuroQoL self-rated health grade as an outcome measure for this analysis, as we assumed that it would provide the best overall representation of HRQOL. Gender-specific regression coefficients from univariable and multivariable analyses as well as p-values for interactions with gender were generated.

Results The response rates for the HRQOL questionnaire were similar for men (71.2%) and women (72.5%). In total, 1,020 men (comprising 71.8% of the study population) and 401 women were available for analysis. Patient characteristics of the responders are shown in table 1, stratified by gender. The mean age was 64.2 years for men and 66.8 years for women (p<0.001). Men had a higher BMI than women: 27.1 vs. 26.7, p=0.009. Women more often had a history of hypertension than men (62.1% vs. 56.0%, p=0.002). Hypercholesterolemia, a history of CVD and smoking were significantly more prevalent among men. There were no gender differences among the indications for CAG between men and women. CAG in women more often revealed normal coronaries than in men and women were subsequently more often treated conservatively than men. Self-reported anginal complaints Women more often complained of shortness of breath than men, also they experienced progressive complaints more often than men (table 2). Men significantly more often reported no complaints (17.5% vs. 9.7%) as compared to women. The triggers of complaints differed between men and women. In women emotion was more often a trigger than in men. Also women more often reported no triggers for their complaints when compared to men. The time since the start of the complaints did not differ between men and women, although slightly more men had a very long history of complaints (>10 years). Women were more likely to have complaints since 1 to 10 years.

197


Chapter 10

Table 1. Patient characteristics and reported HRQOL scores stratified by gender. Men Valid N

Women Valid N Test-value p-value

Demographics N (responders, % of total cohort) Age (years, mean ± sd) BMI (kg/m2, mean ± sd)

71.2

1020

72.5

401

64.2±10.7

1020

66.8±11.4

401

0.6 4.00

<0.001 0.009

27.1±4.2

1018

26.7±5.0

398

4.32

Diabetes (%)

21.7

1008

21.0

391

0.07

0.9

Hypertension (%)

56.0

996

62.1

390

4.31

0.002

Hypercholesterolemia (%)

50.8

985

44.6

383

3.99

0.05

Smoking (ever %)

54.8

899

42.2

344

16.03

<0.001

History of CVD (%)

56.2

1020

44.1

401

16.73

<0.001

64.9

1020

67.6

401

1.90

0.5

391

38.09

<0.001

397

19.91

<0.001

Indication for CAG Stable CAD (%) Unstable angina (%) Myocardial infarction (%) Other (%)

8.2

9.0

23.5

20.4

3.4

3.0

CAD severity Normal coronaries (%)

20.6

1-vessel disease (%)

35.1

30.7

2-vessel disease (%)

28.7

21.5

3-vessel disease (%)

15.6

11.5

1004

36.3

Treatment of CAD Conservative (%)

31.9

PCI (%)

61.8

50.1

6.3

5.3

CABG (%)

1014

44.6

HRQOL EuroQoL self-rated health grade (mean, sd)

6.84 (1.49)

997

6.46 (1.40)

386

4.32

<0.001

Physical functioning (median (%), IQR)

80 (55-95)

993

60 (35-80)

398

9.39

<0.001

1004 62.5 (37.5-87.5)

397

6.76

<0.001

25 (0-100)

375

6.24

<0.001

Emotional role limitations (median (%), IQR) 100 (66.7-100)

969 100 (33.3-100)

365

3.48

<0.001

Mental functioning (median (%), IQR)

80 (68-88)

998

72 (60-84)

388

6.29

<0.001

Vitality (median (%), IQR)

60 (45-75)

998

50 (35-65)

390

7.20

<0.001

1001 67.3 (44.9-100)

389

4.56

<0.001 <0.001

Social functioning (median (%), IQR) Physical role limitations (median (%), IQR)

Pain (median (%), IQR)

75 (62.5-100) 75 (0-100)

79.6 (57.1-100)

978

General health (median (%), IQR)

60 (40-70)

997

50 (25-75)

389

5.37

Health change (median (%), IQR)

50 (25-50)

1003

50 (25-75)

393

0.68

0.5

105 (21-324)

1020

125 (23-370)

401

1.69

0.19

Response delay in days (median, IQR)

P-values are derived from t-tests for normally distributed continuous data, and from non-parametric tests the non-normally distributed continuous data. Chi-square tests were performed on categorical data. The test value represents a t-value for continuous data and a Pearson chi-square value for categorical data. Abbreviations: BMI = body mass index, CAD = coronary artery disease, PCI = percutaneous coronary intervention, CABG = coronary artery bypass grafting, IQR = interquartile range.

198


Sex Differences in HRQOL

The time since last complaints (prior to CAG) was shortest in women, more women than men experienced their complaints in the week prior to CAG (32.1% vs. 24.9%). Indicating that women had complaints more recent to CAG than men. Significantly more men than women experienced no limitations due to their complaints. More women than men experienced mild limitations. Severe limitations or a state in which no activity is possible was equally common in men and women. RAND-36 questionnaire consistency In our study Cronbach’s alpha for internal consistency of the RAND-36 questionnaire was 0.87. Cronbach’s alpha for consistency among the items of each domain was 0.93 for the physical functioning domain, for social functioning it was 0.83, for physical role limitations 0.92, for emotional role limitations 0.90, for mental functioning 0.85, for vitality 0.84, for pain 0.90 and for general health 0.81. The health change domain consisted of a single item; therefore no consistency could be assessed. Overall, these values correspond to high consistency of the questionnaire in our cohort. Self-reported HRQOL The mean self-rated health grade (EuroQoL) was 6.46±1.40 for women, while men reported mean grades of 6.84±1.49 (t-value 4.32, p<0.001). Women also reported a significantly lower HRQOL in 8 out of 10 HRQOL measures as compared to men. Only the score for the RAND-36 domain health change was higher in women than in men. The RAND-36 domain general health did not differ between men and women. HRQOL measures per gender are shown in table 1. When we stratified by indication for CAG (Supplemental table), the EuroQoL self-rated health grade was significantly lower for women presenting with stable CAD (6.4±1.4 vs. 6.7±1.5, p<0.001) or myocardial infarction (6.6±1.4 vs. 7.0±1.4, p=0.039), but not among women who presented with unstable angina or ‘other’ indications. Women who presented with stable CAD complaints showed lower scores on 8/9 RAND36 domains, with unstable angina on 1/9 RAND-36 domain, with myocardial infarction on 6/9 RAND-36 domains and women presenting with ‘other’ indications showed lower scores on 1 RAND-36 domain. Stratified by angiographic CAD severity, the EuroQoL self-rated health grade was lower in women with no CAD (6.3±1.4 vs. 6.8±1.5, p=0.003) and women with single vessel disease (6.6±1.4 vs. 6.9±1.5, p=0.039), but did not differ significantly among patients with double or triple vessel disease. Significantly lower scores were found for women in 6/9 RAND-36 domains among patients with no angiographic CAD, 8/9 RAND-36 domains among patients with single vessel CAD, 7/9 RAND-36 domains among patients with double vessel CAD and in 4/9 RAND-36 domains for patients with triple vessel disease. When we looked into the treatment of CAD, both women treated conservatively and with PCI reported lower EuroQoL self-rated health grades (6.4±1.4 vs. 6.7±1.5, p=0.008 and 6.5±1.4 vs. 6.9±1.5, p=0.003, respectively) than men. No significant gender difference for the EuroQoL self-rated health grade was found for patients treated with CABG.

199


Chapter 10

Table 2. Self-reported anginal complaints characteristics, stratified by gender. Men

n

Women

n

Chi-square

p-value

Type of complaints Chest pain (%)

57.5

587

59.4

238

0.35

0.54

Shortness of breath (%)

32.7

334

44.1

177

16.23

<0.001

Other complaints (%)

19.7

201

17.7

71

0.74

0.39

No complaints (%)

17.5

178

9.7

39

13.28

<0.001

Progressive complaints (%)

52.0

427

60.7

134

7.34

0.007

Triggers of complaints Exercise (%)

57.9

591

58.6

235

0.05

0.82

Emotion (%)

12.7

130

19.5

78

10.36

0.001

Cold temperature (%)

14.5

148

12.7

51

0.77

0.38

No triggers (%)

25.0

255

30.4

122

4.34

0.04

More than 10 years (%)

21.2

173

14.9

51

3.24

0.07

Five to 10 years ago (%)

11.6

95

15.8

54

5.65

0.02

One to 5 years ago (%)

31.1

254

38.6

132

9.62

0.002

Less than 1 year ago (%)

36.1

295

30.7

105

0.63

0.43

Start of complaints

Last complaints More than 1 year ago (%)

17.8

143

13.8

46

1.28

0.26

One year to one month ago (%)

38.2

306

34.2

114

0.13

0.72

One month to one week ago (%)

19.1

153

19.8

66

0.64

0.42

Last week (%)

24.9

200

32.1

107

8.88

0.003

Limitations by complaints None (%)

20.5

168

13.1

45

5.36

0.02

Mild limitations (%)

42.5

348

47.5

163

5.63

0.02

Severe limitations (%)

30.5

250

30.9

106

0.80

0.37

6.5

53

8.5

29

2.41

0.12

No activity possible (%)

Self-reported anginal complaints characteristics, stratified by gender. The percentage and n display the proportion and number of positive responses. The p-value is derived from Pearson chi-square testing.

Women reported lower scores on 7/9 RAND-36 domains when treated conservatively, on 8/9 RAND-36 domains when treated with PCI and on 5/9 RAND-36 domains when treated with CABG. HRQOL scores of male myocardial infarction patients were significantly higher than those of male patients presenting with stable CAD, for 2/9 RAND-36 domains (physical functioning and general health), which was not seen for women. Also, among male patients presenting with an ‘other’ indication these domains were higher as compared to male stable CAD patients. In addition, general health was higher in women presenting with an ‘other’ indication as compared to women presenting with stable CAD.

200


84.5±22.3 70.9±26.1

88.4±19.6 72.7±27.0

81.5±33.6 56.2±44.4

87.3±29.3 80.5±35.9

79.4±17.3

69.5±20.5 60.0±21.2

83.2±23.8 76.4±26.5

71.4±23.3 56.9±21.4

52.6±18.3 46.1±26.4

Physical functioning

Social functioning

Physical role limitations

Emotional role limitations

Mental functioning

Vitality

Pain

General health

Health change

1003

997

1001

998

998

969

978

1004

993

Valid N

Men

-6.5

-14.5

-6.8

-9.5

-2.7

-6.8

-25.3

-15.7

-13.6

Δ Dutch population -UCORBIO

4.37

10.88

4.34

7.44

2.64

3.26

9.96

10.26

8.91

t-value

<0.001

<0.001

<0.001

<0.001

0.009

0.011

<0.001

<0.001

<0.001

p-value

Women

53.4±19.6 44.9±29.1

71.5±21.8 50.3±20.2

80.0±25.4 68.7±28.9

66.3±19.6 51.0±20.8

75.5±18.9 69.7±19.4

82.5±33.5 72.1±41.0

78.3±36.5 39.6±43.7

86.1±20.9 61.2±29.5

80.7±23.6 55.3±28.4

393

389

389

390

388

365

375

397

389

Dutch UCORBIO Valid N population N=691

-8.5

-21.2

-11.3

-15.3

-5.8

-10.4

-38.7

-24.9

-25.4

Δ Dutch population -UCORBIO

5.73

15.75

6.67

12.05

4.79

4.43

15.40

16.21

15.76

t-value

<0.001

<0.001

<0.001

<0.001

<0.001

<0.001

<0.001

<0.001

<0.001

p-value

-2.0

-6.7

-4.5

-5.8

-3.1

-3.7

-13.4

-9.2

-11.8

Difference in Δ between genders

All HRQOL scores are displayed as mean scores ± standard deviation. T-values and p-values are reported for gender-specific differences between our cohort and the general population. The most right column gives the differences between the delta (difference between general population and the UCORBIO cohort) of women and men, which is more negative in women in all nine domains of the RAND-36 questionnaire.

76.7±16.7

Dutch UCORBIO population N=372

Domains of RAND-36

Table 3. Comparison of RAND-36 HRQOL scores between the UCORBIO cohort and the general Dutch population, stratified by gender.

Sex Differences in HRQOL

201


Chapter 10

Remarkably, both men and women did not show a significant difference in HRQOL (EuroQoL) across the severities of CAD. General health was higher in men treated with PCI and CABG as compared to men who were treated conservatively. For women, a difference was found in health change when they were treated with CABG as compared to conservative treatment. Comparison with general population Subjects (men and women combined) in our cohort reported lower RAND-36 HRQOL scores than their age-matched counterparts who were randomly sampled from the Dutch population, as described by van der Zee et al.21 For the age category 55-64 years the average difference across the RAND-36 HRQOL domains was 8.5 points lower in our cohort than in the general population. For the age category 65-75 the average difference was 5.9 points lower than in the general population (data not shown). As shown in table 3, we calculated the difference in HRQOL scores between our cohort and the general Dutch population stratified by gender (all significantly lower in our cohort), and observed that the Δ was larger in women than in men. Women’s HRQOL scores were 2 to 13.4 points lower than men’s scores across the nine domains of the RAND-36, indicating a larger suppressing effect of CAD on HRQOL in women than in men. Women appeared to be more severely affected and restricted by chest pain and CAD than men, as reflected by greater differences in their HRQOL scores. Gender differences in associations of patient characteristics with HRQOL The associations of patient characteristics with the EuroQoL self-graded health grade are displayed in table 4, stratified by gender. These associations were tested in both a univariable manner and in a multivariable manner. Interactions of patient characteristics with gender were evaluated in both analyses. The betas in table 4 represent the change in the EuroQoL self-rated health grade for a given change in the patient characteristic. In the univariable models, significant gender interactions were found for diabetes, a history of CVD and shortness of breath. Diabetes was associated with lower HRQOL in men (β -0.46, p<0.001) but not in women (β 0.02, p=0.9, p-value for interaction 0.028). Having a history of CVD was also associated with lower HRQOL in men but not in women (β -0.60, p<0.001 for men, β -0.17, p=0.24 for women, p-value for interaction 0.013). Shortness of breath was associated with lower HRQOL in both men and women, but the effect size was significantly greater in men (β -0.87, p<0.001 vs. β-0.44, p=0.002 for women, p-value for interaction 0.013). In the multivariable model no significant gender interactions were found.

Discussion In this study we demonstrated that self-reported HRQOL differed by gender in patients undergoing CAG regardless of the indication for CAG, CAD severity and treatment of

202


Sex Differences in HRQOL

CAD (conservative, PCI or CABG). Women showed a larger difference in HRQOL as compared to the general population than men. Furthermore, gender differences were found in the associations of patient characteristics with HRQOL. HRQOL scores In our cohort we found that women consistently reported lower HRQOL scores than men for both the EuroQoL self-rated health grade and the RAND-36 domains. This has previously been described by Norris et al.15 in a cohort of patients with established CAD and also in patients with severe CAD undergoing CABG.17,24 In addition to existing literature we showed that gender differences in HRQOL are observable throughout all indications and severities of CAD, remarkably, also in patients in whom no significant epicardial CAD could be objectified by CAG and in whom no invasive treatment was undertaken. Gender differences in HRQOL can already be observed in the general Dutch population, where women report slightly lower HRQOL scores, although not consistently across all domains of the RAND-36 questionnaire (average difference across the nine domains 2.6 points).21 In our cohort the difference between men and women is on average 9.3 points, indicating that the baseline difference between men and women is amplified in a population in which CAG is indicated and that HRQOL apparently is more strongly affected in women than in men. To our knowledge, this phenomenon has not been reported before. Patient characteristics associated with HRQOL In our study we found significant interactions of gender with diabetes, history of CVD and shortness of breath. In all three cases the association of the patient characteristic with lower HRQOL was stronger in men than in women. Possibly, HRQOL in women is not so much determined by CVD risk factors, a history of CVD or other general patient characteristics but more by other factors such as hormonal status (menopause) and psychosocial factors that were not measured in this study. Menopause has previously been shown to have a negative impact on some domains of HRQOL scores.25 However, in our cohort we do not observe more extreme gender differences in HRQOL in the age group of 55-64 in which post-menopausal symptoms would occur (data not shown). Also, it has been shown that depression is related to HRQOL17, although depression and social support could also not completely explain the gender difference in HRQOL in a study by Norris et al.15 Unknown factors or factors that were not included in this study might better explain HRQOL in women. Socioeconomic status26, lifestyle and noncardiovascular comorbidities might account for a part of the unexplained variance, but were unfortunately not available in this study. Efforts should be pursued to elucidate the factors that determine the low HRQOL in women. When we know which factors determine low HRQOL in women, more targeted approaches can be sought in order to improve HRQOL in women.

203


Chapter 10

Table 4. Regression coefficients and interactions for the EuroQoL self-rated health grade (scaled 1 to 10) stratified by gender, derived from univariable and multivariable models Univariable Men

Women

Beta (95% CI) t-value Age (per 10 years)

-0.06 (-0.15- 0.03)

-1.31

p-value

Beta (95% CI) t-value

0.191

-0.04 (-0.17- 0.08)

-0.71

p-value

P Interaction

0.479

0.856

BMI (per 5 points)

-0.14 (-0.25--0.03)

-2.57

0.010

-0.06 (-0.20- 0.08)

-0.79

0.429

0.348

Diabetes

-0.46 (-0.69--0.24)

-4.03

<0.001

0.02 (-0.34- 0.37)

0.09

0.928

0.028*

Hypertension

-0.42 (-0.61--0.23)

-4.40

<0.001

-0.21 (-0.50- 0.08)

-1.41

0.160

0.248

Hypercholesterolemia

-0.32 (-0.51--0.14)

-3.38

0.001

-0.12 (-0.41- 0.17)

-0.81

0.420

0.252

Smoking

-0.26 (-0.46--0.06)

-2.55

0.011

-0.10 (-0.41- 0.21)

-0.64

0.524

0.409

History of CVD

-0.60 (-0.78--0.42)

-6.45

<0.001

-0.17 (-0.45- 0.11)

-1.18

0.240

0.013*

Stable CAD (vs. UA or infarction)

-0.25 (-0.44--0.05)

-2.48

0.013

-0.32 (-0.62--0.02)

-2.12

0.035

0.684

0.05 (-0.18- 0.28)

0.40

0.688

0.19 (-0.10- 0.48)

1.28

0.200

0.460 0.995

Significant CAD Conservative treatment (vs. PCI or CABG)

-0.15 (-0.36- 0.05)

-1.52

0.129

-0.15 (-0.44- 0.13)

-1.07

0.283

Chest pain

-0.36 (-0.54--0.17)

-3.76

<0.001

-0.29 (-0.58--0.01)

-2.02

0.044

0.711

Shortness of breath

-0.87 (-1.06--0.68)

-9.00

<0.001

-0.44 (-0.72--0.16)

-3.11

0.002

0.013*

Progressive complaints

-0.65 (-0.85--0.45)

-6.35

<0.001

-0.69 (-0.98--0.40)

-4.75

<0.001

0.838

Complaints with exercise

-0.35 (-0.54--0.17)

-3.72

<0.001

-0.18 (-0.46- 0.11)

-1.21

0.227

0.315

Complaints with emotion

-0.48 (-0.76--0.21)

-3.45

0.001

-0.38 (-0.73--0.03)

-2.13

0.033

0.655

Complaints with cold

-0.47 (-0.73--0.21)

-3.55

<0.001

-0.11 (-0.53- 0.31)

-0.51

0.614

0.160

0.56 ( 0.34- 0.77)

5.03

<0.001

0.36 ( 0.05- 0.67)

2.28

0.023

0.325

Start of complaints (short vs. long) Last complaints (recent vs. long)

-0.77 (-0.98--0.57)

-7.41

<0.001

-0.74 (-1.02--0.45)

-5.09

<0.001

0.849

Limitations due to complaints

-0.93 (-1.13--0.72)

-8.93

<0.001

-0.71 (-1.00--0.43)

-4.89

<0.001

0.250

Regression coefficients (betas) for the EuroQoL self-rated health grade (scale 1 to 10) stratified by gender, obtained from a univariable regression model (left part of table) and from a multivariable linear regression model containing: diabetes, hypertension, smoking, hypercholesterolemia, history of CVD, age, BMI, indication for CAG (stable CAD, unstable angina, myocardial infarction), treatment of CAD (conservative, PCI, CABG), angiographic significance of CAD, time since first complaints, time since last complaints, limitations due to complaints, chest pain, shortness of breath, triggers of complaints (exercise, emotion, cold) and progression of complaints. The beta is described for yes vs. no, unless indicated otherwise. CI=confidence interval. The significance of the interaction terms is given in the columns “P Interaction”, * indicates p-value for interaction <0.05. Significantly different univariable betas were found for diabetes, history of CVD and shortness of breath. No multivariable interaction terms were found to be significant.

Microvascular disease Surprisingly, patients in whom eventually no significant CAD could be objectified reported HRQOL scores that were equal to patients who were diagnosed with triple vessel disease (e.g. mean EuroQoL self-rated health grade 6.3 in both no CAD and triple vessel disease for women and 6.8 and 6.7 for men, respectively). Possibly, HRQOL is compromised in these patients with “healthy coronaries” in the same way as with significant CAD due to complaints of microvascular disease or spasms of the coronary arteries or microvascular system, which cannot be visualized on CAG. These conditions might give rise to complaints similar to macrovascular epicardial disease. As opposed to macrovascular disease, which is more common among men, microvascular disease has been reported to be equally common between the genders.27 As microvascular disease is more

204


Sex Differences in HRQOL

Table 4. Continued Multivariable Men

Women

Beta (95% CI)

t-value

p-value

Beta (95% CI) t-value

p-value

P Interaction

0.13 ( 0.03- 0.23)

2.53

0.012

0.04 (-0.11- 0.19)

0.50

0.615

0.510

0.06 (-0.06- 0.18)

0.92

0.359

-0.05 (-0.22- 0.12)

-0.62

0.537

0.933

-0.11 (-0.39- 0.16)

-0.82

0.415

0.06 (-0.35- 0.47)

0.30

0.762

0.342

-0.34 (-0.57--0.11)

-2.95

0.003

0.02 (-0.34- 0.38)

0.10

0.918

0.071

-0.06 (-0.29- 0.17)

-0.50

0.614

-0.03 (-0.36- 0.30)

-0.17

0.868

0.523 0.982

-0.17 (-0.38- 0.04)

-1.55

0.122

-0.22 (-0.56- 0.11)

-1.30

0.194

-0.33 (-0.57--0.09)

-2.71

0.007

-0.08 (-0.41- 0.26)

-0.45

0.651

0.236

0.05 (-0.19- 0.29)

0.40

0.688

-0.21 (-0.59- 0.17)

-1.08

0.283

0.464

0.25 (-0.14- 0.65)

1.25

0.212

0.07 (-0.53- 0.67)

0.23

0.819

0.988

0.00 (-0.35- 0.34)

-0.02

0.983

-0.14 (-0.73- 0.44)

-0.49

0.628

0.885

-0.19 (-0.41- 0.04)

-1.61

0.108

-0.32 (-0.66- 0.03)

-1.78

0.076

0.944

-0.48 (-0.71--0.26)

-4.24

<0.001

-0.14 (-0.48- 0.20)

-0.84

0.403

0.099

-0.23 (-0.45--0.02)

-2.11

0.035

-0.51 (-0.88--0.14)

-2.70

0.008

0.486

0.05 (-0.19- 0.29)

0.43

0.666

0.06 (-0.30- 0.42)

0.32

0.749

0.831

-0.21 (-0.51- 0.09)

-1.40

0.162

-0.24 (-0.61- 0.13)

-1.27

0.206

0.898

-0.25 (-0.52- 0.03)

-1.75

0.081

0.18 (-0.26- 0.63)

0.81

0.417

0.115

0.20 (-0.05- 0.44)

1.59

0.113

0.22 (-0.14- 0.57)

1.22

0.225

0.889

-0.58 (-0.79--0.36)

-5.31

<0.001

-0.67 (-0.99--0.34)

-4.04

<0.001

0.809

-0.53 (-0.76--0.31)

-4.60

<0.001

-0.42 (-0.77--0.07)

-2.37

0.018

0.671

complicated to diagnose, it is often unrecognized and thus undertreated.28–30 And most importantly, the presence of microvascular disease is associated with poor outcome and thus should not be trivialized.27,31 The predictors of microvascular disease in women may lie, at least in part, in female-specific risk factors e.g. estrogen deficiency.32 These factors, unfortunately, were not available for our analysis. Limitations A limitation of our study is that we could control the moment a patient decides to fill in and return their questionnaire. A considerable spread in delay was observed, ranging from 0 to 743 days with a median delay of 112 days (interquartile range: 22 to 338 days) between the date of CAG and the date of returning the questionnaire. There was a

205


Chapter 10

significant but very limited effect of delay of questionnaire return (per 100 days) with the EuroQoL self-rated health grade (β 0.06, p=0.007). Indicating that with every 100-day increase of delay the EuroQoL self-rated health grade was 0.06 points higher. In the analyses discussed in this paper we were unable to take a possible response bias into account. At baseline the responders and non-responders did not differ markedly (data not shown). In summary: they were younger, less often had a history of CVD and more often presented with myocardial infarction, all other baseline characteristics were equal. The indication group “otherâ€? might be different from the regular indications for CAG. Therefore, we performed a sensitivity analysis without these patients, which yielded similar results. Race/ethnicity-specific analyses could not be performed due to the predominantly Caucasian (93%) population that was studied. Implications Improving HRQOL is important because low HRQOL has been reported to lead to higher health care costs (hospital admission, emergency room and prescription expenditures).1 People with the lowest HRQOL scores utilize almost thrice the annual health care costs of people with the highest HRQOL scores (>5000USD vs. <2000USD). Diabetes, a positive history of CVD and complaints of shortness of breath were more strongly associated with HRQOL in men. Diabetes is treated with antidiabetic drugs, a strict regime of glucose checks and lifestyle adaptations, which can be a great burden for patients. Awareness of the benefit of strict glucose control might alleviate depressed HRQOL among diabetics.33 General strategies in order to prevent CVD in men could be beneficial in terms of HRQOL (albeit on the long term), as a history of CVD is associated with lower HRQOL in men. The characteristics of complaints were associated with HRQOL in both men and women. Specific complaint-focused treatment might be beneficial to improve HRQOL, especially in women, in whom no other factors associating with HRQOL could be determined. In the current guidelines on the treatment of stable angina, in patients with typical complaints but no epicardial CAD, it is advised to undertake further diagnostic tests to assess microvascular ischemic heart disease.34,35 But even when patients with microvascular disease are treated according to the guidelines, recurrent symptoms are common.36 Additional pain relieve interventions or coping programs should be considered in patients with refractory or microvascular anginal complaints in order to improve HRQOL.37 The difference in HRQOL was significant between men and women in our study, however clinical and personal relevance of this difference remains unclear. Future studies need to evaluate whether repressed HRQOL in men and women has similar consequences for e.g. hospitalizations, medication utilization and the ability to participate in work and social activities.

206


Sex Differences in HRQOL

Conclusion Women reported lower HRQOL than men throughout all indications for CAG and regardless of CAD severity and treatment. As compared to the general population, the reduction in HRQOL was more extreme in women than in men. Furthermore, there were evident gender differences in determinants of diminished HRQOL scores in patients undergoing CAG. These differences deserve attention in future research.

207


Chapter 10

References 1.

Harrison PL, Pope JE, Coberley CR, Rula EY. Evaluation of the relationship between individual well-being and future health care utilization and cost. Popul Health Manag. 2012;15:325–30.

2.

Marquis P, Caron M, Emery MP, Scott J, Acquadro C. The Role of Health-Related Quality of Life Data in the Drug Approval Processes in the US and Europe: A Review of Guidance Documents and Authorizations of Medicinal Products from 2006 to 2010. Pharmaceut Med. 2011;25:147–160.

3.

Hlatky MA, Chung SC, Escobedo J, Hillegass WB, Melsop K, Rogers W, Brooks MM. The effect of obesity on quality of life in patients with diabetes and coronary artery disease. Am Heart J. 2010;159:292–300.

4.

Oreopoulos A, Padwal R, McAlister FA, Ezekowitz J, Sharma AM, Kalantar-Zadeh K, Fonarow GC, Norris CM. Association between obesity and health-related quality of life in patients with coronary artery disease. Int J Obes (Lond). 2010;34:1434–41.

5.

Uchmanowicz I, Loboz-Grudzien K, Jankowska-Polanska B, Sokalski L. Influence of diabetes on healthrelated quality of life results in patients with acute coronary syndrome treated with coronary angioplasty. Acta Diabetol. 2013;50:217–25.

6.

Stafford L, Berk M, Jackson HJ. Tobacco smoking predicts depression and poorer quality of life in heart

7.

Müller-Nordhorn J, Muckelbauer R, Englert H, Grittner U, Berger H, Sonntag F, Völler H, Prugger C,

disease. BMC Cardiovasc Disord. 2013;13:35.

Wegscheider K, Katus HA, Willich SN. Longitudinal association between body mass index and health-related quality of life. PLoS One. 2014;9:e93071. 8.

Al Hayek AA, Robert AA, Al Saeed A, Alzaid AA, Al Sabaan FS. Factors Associated with Health-Related Quality of Life among Saudi Patients with Type 2 Diabetes Mellitus: A Cross-Sectional Survey. Diabetes Metab J. 2014;38:220–9.

9.

Coste J, Quinquis L, D’Almeida S, Audureau E. Smoking and health-related quality of life in the general population. Independent relationships and large differences according to patterns and quantity of smoking and to gender. PLoS One. 2014;9:e91562.

10. Mendis S, Puska P, Norrving B. Global Atlas on cardiovascular disease prevention and control. 2011. 11. Jacobs AK. Coronary intervention in 2009: are women no different than men? Circ Cardiovasc Interv. 2009;2:69–78. 12. Hemingway H, Langenberg C, Damant J, Frost C, Pyörälä K, Barrett-Connor E. Prevalence of angina in women versus men: a systematic review and meta-analysis of international variations across 31 countries. Circulation. 2008;117:1526–36. 13. Bruins Slot MHE, Rutten FH, van der Heijden GJMG, Doevendans PA, Mast EG, Bredero AC, Glatz JFC, Hoes AW. Gender differences in pre-hospital time delay and symptom presentation in patients suspected of acute coronary syndrome in primary care. Fam Pract. 2012;29:332–7. 14. Kaul P, Armstrong PW, Sookram S, Leung BK, Brass N, Welsh RC. Temporal trends in patient and treatment delay among men and women presenting with ST-elevation myocardial infarction. Am Heart J. 2011;161:91–7. 15. Norris CM, Spertus J a, Jensen L, Johnson J, Hegadoren KM, Ghali W a. Sex and gender discrepancies in health-related quality of life outcomes among patients with established coronary artery disease. Circ Cardiovasc Qual Outcomes. 2008;1:123–30. 16. Bakhai A, Ferrières J, James S, Iñiguez A, Mohácsi A, Pavlides G, Belger M, Norrbacka K, Sartral M. Treatment, outcomes, costs, and quality of life of women and men with acute coronary syndromes who have undergone

208


Sex Differences in HRQOL

percutaneous coronary intervention: results from the antiplatelet therapy observational registry. Postgrad Med. 2013;125:100–7. 17. Kendel F, Dunkel A, Müller-Tasch T, Steinberg K, Lehmkuhl E, Hetzer R, Regitz-Zagrosek V. Gender differences in health-related quality of life after coronary bypass surgery: results from a 1-year follow-up in propensitymatched men and women. Psychosom Med. 2011;73:280–5. 18. VanderZee KI, Sanderman R, Heyink JW, de Haes H. Psychometric qualities of the RAND 36-Item Health Survey 1.0: a multidimensional measure of general health status. Int J Behav Med. 1996;3:104–122. 19. EuroQoL group. EuroQol- a new facility for the measurement of health-related quality of life. Health Policy. 1990;16:220–33. 20. Rose GA, Blackburn H. Cardiovascular survey methods. Geneva, Switzerland: World Health Organization (Sales agent for U. K., H. M. Stationery Office); 1968. 21. Zee K Van de, Sanderman R. Het meten van de algemene gezondheidstoestand met de RAND-36: een handleiding. Rijksuniversiteit, Groningen)(ISBN 90–72156–60–9). 1993;28. 22. Groningen RU. SYNTAX file RAND-36 V1 [Internet]. Available from: https://www.umcg.nl/ SiteCollectionDocuments/research/institutes/SHARE/assessment tools/syntax_file_rand-36_v1_ withoutaggregatescores.pdf 23. R Core Team. R: A Language and Environment for Statistical Computing. 2013; 24. Martin LM, Holmes SD, Henry LL, Schlauch K a, Stone LE, Roots A, Hunt SL, Ad N. Health-related quality of life after coronary artery bypass grafting surgery and the role of gender. Cardiovasc Revasc Med. 2012;13:321–7. 25. Hess R, Thurston RC, Hays RD, Chang C-CH, Dillon SN, Ness RB, Bryce CL, Kapoor WN, Matthews K a. The impact of menopause on health-related quality of life: results from the STRIDE longitudinal study. Qual Life Res. 2012;21:535–44. 26. Degroote S, Vogelaers DP, Vermeir P, Mariman A, De Rick A, Van Der Gucht B, Pelgrom J, Van Wanzeele F, Verhofstede C, Vandijck DM. Socio-economic, behavioural, (neuro)psychological and clinical determinants of HRQoL in people living with HIV in Belgium: a pilot study. J Int AIDS Soc. 2013;16:18643. 27. Murthy VL, Naya M, Taqueti VR, Foster CR, Gaber M, Hainer J, Dorbala S, Blankstein R, Rimoldi O, Camici PG, Di Carli MF. Effects of sex on coronary microvascular dysfunction and cardiac outcomes. Circulation. 2014;129:2518–27. 28. Pries AR, Habazettl H, Ambrosio G, Hansen PR, Kaski JC, Schächinger V, Tillmanns H, Vassalli G, Tritto I, Weis M, de Wit C, Bugiardini R. A review of methods for assessment of coronary microvascular disease in both clinical and experimental settings. Cardiovasc Res. 2008;80:165–74. 29. Shaw LJ, Bugiardini R, Merz CNB. Women and ischemic heart disease: evolving knowledge. J Am Coll Cardiol. 2009;54:1561–75. 30. Bugiardini R, Bairey Merz CN. Angina with “normal” coronary arteries: a changing philosophy. JAMA. 2005;293:477–84. 31. Gulati M, Cooper-DeHoff RM, McClure C, Johnson BD, Shaw LJ, Handberg EM, Zineh I, Kelsey SF, Arnsdorf MF, Black HR, Pepine CJ, Merz CNB. Adverse cardiovascular outcomes in women with nonobstructive coronary artery disease: a report from the Women’s Ischemia Syndrome Evaluation Study and the St James Women Take Heart Project. Arch Intern Med. 2009;169:843–850. 32. Banks K, Lo M, Khera A. Angina in Women without Obstructive Coronary Artery Disease. Curr Cardiol Rev. 2010;6:71–81.

209


Chapter 10

33. Cox DJ. Blood Glucose Awareness Training: What Is It, Where Is It, and Where Is It Going? Diabetes Spectr. 2006;19:43–49. 34. Montalescot G, Sechtem U, Achenbach S, Andreotti F, Arden C, Budaj A, Bugiardini R, Crea F, Cuisset T, Di Mario C, Ferreira JR, Gersh BJ, Gitt AK, Hulot J-S, Marx N, Opie LH, Pfisterer M, Prescott E, Ruschitzka F, Sabaté M, Senior R, Taggart DP, van der Wall EE, Vrints CJM, Zamorano JL, Baumgartner H, Bax JJ, Bueno H, Dean V, Deaton C, Erol C, Fagard R, Ferrari R, Hasdai D, Hoes AW, Kirchhof P, Knuuti J, Kolh P, Lancellotti P, Linhart A, Nihoyannopoulos P, Piepoli MF, Ponikowski P, Sirnes PA, Tamargo JL, Tendera M, Torbicki A, Wijns W, Windecker S, Valgimigli M, Claeys MJ, Donner-Banzhoff N, Frank H, Funck-Brentano C, Gaemperli O, Gonzalez-Juanatey JR, Hamilos M, Husted S, James SK, Kervinen K, Kristensen SD, Maggioni A Pietro, Pries AR, Romeo F, Rydén L, Simoons ML, Steg PG, Timmis A, Yildirir A. 2013 ESC guidelines on the management of stable coronary artery disease: the Task Force on the management of stable coronary artery disease of the European Society of Cardiology. Eur Heart J. 2013;34:2949–3003. 35. Montalescot G, Sechtem U, Achenbach S, Andreotti F, Arden C, Budaj A, Bugiardini R, Crea F, Cuisset T, Di Mario C, Ferreira JR, Gersh BJ, Gitt AK, Hulot J-S, Marx N, Opie LH, Pfisterer M, Prescott E, Ruschitzka F, Sabaté M, Senior R, Taggart DP, van der Wall EE, Vrints CJM, Zamorano JL, Baumgartner H, Bax JJ, Bueno H, Dean V, Deaton C, Erol C, Fagard R, Ferrari R, Hasdai D, Hoes AW, Kirchhof P, Knuuti J, Kolh P, Lancellotti P, Linhart A, Nihoyannopoulos P, Piepoli MF, Ponikowski P, Sirnes PA, Tamargo JL, Tendera M, Torbicki A, Wijns W, Windecker S, Valgimigli M, Claeys MJ, Donner-Banzhoff N, Frank H, Funck-Brentano C, Gaemperli O, Gonzalez-Juanatey JR, Hamilos M, Husted S, James SK, Kervinen K, Kristensen SD, Maggioni A Pietro, Pries AR, Romeo F, Rydén L, Simoons ML, Steg PG, Timmis A, Yildirir A. 2013 ESC guidelines on the management of stable coronary artery disease: the Task Force on the management of stable coronary artery disease of the European Society of Cardiology. Addenda. Eur Heart J. 2013;34:2949–3003. 36. Shaw LJ, Bairey Merz CN, Pepine CJ, Reis SE, Bittner V, Kelsey SF, Olson M, Johnson BD, Mankad S, Sharaf BL, Rogers WJ, Wessel TR, Arant CB, Pohost GM, Lerman A, Quyyumi AA, Sopko G. Insights from the NHLBISponsored Women’s Ischemia Syndrome Evaluation (WISE) Study: Part I: gender differences in traditional and novel risk factors, symptom evaluation, and gender-optimized diagnostic strategies. J Am Coll Cardiol. 2006;47:S4–S20. 37. Lanza GA, Parrinello R, Figliozzi S. Management of microvascular angina pectoris. Am J Cardiovasc Drugs. 2014;14:31–40.

210


211


Chapter 10

Supplemental Supplemental table. HRQOL scores by CAD severity, by indication and by treatment, stratified by gender. Indication for CAG

Stable CAD

Unstable angina

Women

Men

Test statistic

Women

Men

6.4 (1.4)

6.7 (1.5)

3.72**

6.6 (1.2)

6.8 (1.5)

0.66

Physical functioning

55 (30-75)

75 (50-90)

57.36**

65 (34-80)

83 (55-95)

9.75*

Social functioning

63 (38-88)

75 (63-100)

36.07**

63 (38-88)

75 (50-100)

3.14

Physical role limitations

25 (0-100)

75 (0-100)

30.31**

50 (0-100)

75 (0-100)

0.35

100 (33-100)

100 (67-100)

11.76*

100 (33-100)

100 (67-100)

0.05

72 (60-84)

80 (68-88)

27.63**

72 (56-88)

80 (65-88)

3.05

EuroQoL self-rated health grade

Emotional role limitations Mental functioning Vitality Pain

50 (35-65)

60 (45-75)

37.91**

60 (40-70)

60 (45-79)

2.40

67 (45-100)

80 (57-100)

20.60**

88 (45-100)

80 (55-100)

0.00

General health

50 (35-65)

55 (40-70)

14.52**

50 (35-65)

60 (45-75)

1.73

Health change

50 (25-75)

50 (25-50)

0.53

50 (25-50)

50 (25-75)

0.95

6.3 (1.4)

6.8 (1.5)

2.99*

6.6 (1.4)

6.9 (1.5)

2.07*

Physical functioning

60 (40-80)

80 (55-95)

25.67**

60 (38-80)

80 (55-95)

28.92**

Social functioning

63 (50-88)

75 (63-100)

15.84**

63 (50-78)

75 (63-100)

16.05**

Physical role limitations

25 (0-100)

75 (0-100)

14.29**

25 (0-100)

75 (0-100)

10.80*

100 (67-100)

100 (100-100)

1.00

100 (33-100)

100 (83-100)

10.93*

76 (60-88)

80 (68-88)

9.42*

72 (60-84)

80 (68-88)

17.77** 16.71**

CAD severity EuroQoL self-rated health grade

Emotional role limitations Mental functioning Vitality

No CAD

Single vessel disease

50 (35-65)

65 (45-80)

15.87**

50 (35-65)

60 (45-75)

67 (45-100)

84 (57-100)

7.30*

78 (45-100)

80 (57-100)

6.24*

General health

50 (35-70)

55 (35-75)

3.67

50 (35-60)

60 (45-75)

21.14**

Health change

50 (25-50)

50 (25-50)

0.45

50 (25-63)

50 (25-50)

0.68

Pain

Treatment of CAD EuroQoL self-rated health grade

Conservative

PCI

6.4 (1.4)

6.7 (1.5)

2.64*

6.5 (1.4)

6.9 (1.5)

2.98*

Physical functioning

60 (40-75)

75 (50-95)

26.99**

55 (30-80)

80 (55-95)

51.98**

Social functioning

63 (50-88)

75 (63-100)

21.56**

63 (38-88)

75 (50-100)

23.08**

Physical role limitations

25 (0-100)

75 (0-100)

17.18**

25 (0-100)

75 (0-100)

19.19**

100 (67-100)

100 (100-100)

3.40

100 (33-100)

100 (67-100)

11.37*

76 (60-84)

80 (68-88)

15.49**

72 (56-84)

80 (68-88)

21.14** 22.76**

Emotional role limitations Mental functioning Vitality

50 (35-65)

60 (45-75)

19.86**

50 (35-70)

60 (45-75)

67 (45-100)

80 (57-100)

11.76*

78 (45-100)

80 (57-100)

7.74*

General health

50 (35-65)

55 (35-70)

5.21*

55 (35-65)

60 (45-75)#

11.78*

Health change

50 (25-50)

50 (25-50)

0.37

50 (25-50)

50 (25-75)

2.04

Pain

The test statistics represent chi-square values for the non-parametric tests (RAND-domains) and t-values for the EuroQol item. * indicates significance p<0.05, ** indicates p-value <0.001. # indicates a significant difference in HRQOL measurement for the different indications, CAD severities or treatments of CAD within men or women. No significant differences were found among CAD severities (all bonferroni corrected p-values >0.05). Differences in indications for CAD were only found between yocardial infarction and stable CAD and between ‘other’ indication and stable CAD. Among the treatment groups differences were found between CABG and conservative and PCI and conservative.

212


Sex Differences in HRQOL

Supplemental table. Continued Myocardial infarction

Other

Women

Men

Test statistic

Women

Men

6.6 (1.4)

7.0 (1.4)

2.08*

7.1 (1.0)

7.3 (1.5)

Test statistic 0.65

70 (40-80)

80 (65-95)#

19.67**

78 (61-89)

90 (75-95)#

4.19*

63 (50-88)

75 (50-100)

10.18*

81 (66-97)

88 (59-100)

0.00

0 (0-100)

75 (0-100)

8.12*

50 (6-100)

100 (19-100)

1.47

100 (33-100)

100 (67-100)

2.76

100 (75-100)

100 (100-100)

0.06

72 (59-84)

80 (64-88)

6.55*

72 (58-87)

84 (71-89)

1.89 0.55

53 (40-70)

65 (45-80)

9.11*

58 (46-79)

70 (50-80)

78 (45-100)

88 (59-100)

1.92

83 (60-97)

90 (75-100)

1.43

55 (45-65)

65 (50-78)#

8.05*

70 (58-80)#

70 (53-85)#

0.18

50 (25-56)

50 (25-50)

0.44

50 (25-50)

50 (25-50)

0.00

Double vessel disease

Triple vessel disease

6.5 (1.4)

6.8 (1.5)

1.90

6.3 (1.6)

6.7 (1.5)

1.10

55 (30-75)

75 (55-90)

31.61**

50 (25-80)

75 (51-90)

8.17*

63 (38-88)

75 (63-100)

10.02*

50 (25-88)

75 (50-100)

10.48*

0 (0-75)

50 (0-100)

16.21**

25 (0-100)

50 (0-100)

2.96*

100 (0-100)

100 (67-100)

9.04*

100 (50-100)

100 (67-100)

0.26

68 (52-84)

80 (64-88)

12.37**

76 (52-86)

76 (64-88)

2.15

45 (35-70)

60 (45-75)

12.48**

50 (33-68)

65 (41-75)

4.71*

72 (45-100)

80 (57-100)

5.85*

77 (45-100)

80 (57-100)

1.40

55 (38-65)

60 (40-70)

1.96

55 (35-65)

60 (40-70)

0.83

38 (25-75)

50 (25-75)

1.33

50 (25-75)

50 (25-50)

0.71

CABG 6.4 (1.5)

6.9 (1.5)

1.46

53 (28-75)

78 (65-90)

9.68*

63 (31-81)

75 (63-100)

5.23*

25 (0-50)

88 (0-100)

4.57*

100 (67-100)

100 (100-100)

0.81

76 (60-88)

84 (72-92)

2.60 7.49*

55 (40-65)

65 (50-80)

61 (45-100)

90 (67-100)

3.41

55 (43-65)

70 (50-80)#

6.14*

75 (25-100)#

50 (25-75)

1.76

213



PART TWO Sex Differences

Chapter 11 Sex Differences in Health-Related Quality of Life of Peripheral and Coronary Artery Disease Patients and its Relation with Adverse Events and Mortality In preparation

Crystel M. Gijsberts, Aisha Gohar, Saskia Haitjema, Gerard Pasterkamp, Dominique P.V. de Kleijn, Folkert W. Asselbergs, Michiel Voskuil, Gert-Jan de Borst, Imo E. Hoefer, Hester M. den Ruijter


Chapter 11

Abstract Background Women with cardiovascular disease are known to report lower health-related quality of life (HRQOL) than men; the reason for this is unknown. Here, we examine whether female-specific risk factors related to pregnancy disorders and hormonal changes explain part of the lower HRQOL reported in women. In addition, we examined whether HRQOL predicts outcome in peripheral arterial disease (PAD) and coronary artery disease (CAD) patients in a sex-specific manner. Methods PAD patients from the Athero-express cohort (525 women, 1351 men) and CAD patients from the UCORBIO cohort (315 women, 1105 men) completed the RAND-36 HRQOL questionnaire; the UCORBIO patients additionally reported a EuroQoL health grade. A summary HRQOL score was computed for analysis (HRQOLcomp). We evaluated femalespecific determinants of poor HRQOLcomp and assessed the prognostic value of HRQOLcomp in a sex-specific manner for predicting major adverse cardiovascular events (MACE) and all-cause death. The median follow-up duration for PAD patients was 3.0 years and for CAD patients 2.1 years. During this time 54 women and 179 men died. Results HRQOLcomp was lower in women than in men with CAD (median 6.4 vs. 6.8, p=0.002), but no sex-difference was found in PAD. In PAD and CAD women, cardiovascular risk factors and disease characteristics were associated with HRQOLcomp, but female-specific risk factors were not. HRQOLcomp was a significant predictor of MACE and all-cause death in men and women with PAD. Among CAD patients HRQOLcomp was less predictive of MACE (p interaction 0.21) and all-cause death (p interaction 0.030) in women than in men. Multivariable adjusted HRs for 1-point decreased HRQOLcomp for all-cause death in CAD men and women were 1.63 [1.36-1.95], p<0.001 and 1.10 [0.74-1.64], p=0.65, respectively. Conclusion Women with CAD reported lower HRQOLcomp than men whereas similar HRQOLcomp was found in PAD men and women. Lower HRQOLcomp was predictive of MACE and all-cause death in similar ways between PAD men and women. Among CAD patients, HRQOLcomp might be less predictive of MACE and all-cause death in CAD women than in CAD men.

216


Sex Differences in the Association of HRQOL with outcome

Introduction Health-related quality of life (HRQOL) is impaired in cardiovascular patients, such as peripheral arterial disease1 (PAD) and coronary artery disease2 (CAD) patients. A poor HRQOL is related to higher health care expenditure3 and increased risk of adverse cardiovascular events4 and mortality5,6 Sex differences in HRQOL impairment have been addressed particularly among CAD patients, in which women report markedly lower HRQOL than men.7–9 To date it is largely unclear which factors are associated with the lower HRQOL reported by women. Cardiovascular risk factors such as obesity10,11, diabetes12 and smoking13 are known to influence HRQOL, but the effects are modest among women2, not explaining their lower HRQOL. Female-specific risk factors for cardiovascular disease have emerged in the past decade. Hormonal and reproductive characteristics such as preeclampsia, gestational diabetes mellitus and hypertension, early menopause and hormone replacement therapy have been linked to cardiovascular risk.14–18 The relation of these factors with HRQOL in women has not been established yet. In the current study we evaluated HRQOL data from the Athero-express study19 of carotid and femoral artery endarterectomy patients and from the UCORBIO cohort of coronary angiography patients20. Per protocol, both studies administered the standardized RAND3621 questionnaire at the moment of enrolment and patients were followed-up for adverse events on a yearly basis. In light of lower baseline values of HRQOL in women than in men, we hypothesized that female-specific risk factors attribute to lower HRQOL in women. Furthermore, we hypothesized that HRQOL values provides sex-specific prognostic information in patients with overt cardiovascular disease.

Methods Study population Our study population consists of 5,376 patients with PAD (n=2,984) and CAD (n=2,392). The PAD patients were participants of the Athero-express biobank study, which has been described in detail before.19 Recapitulating, patients undergoing surgical carotid or femoral endarterectomy (CEA or FEA) are enrolled in this biobank and followed-up for the occurrence of adverse events. Enrolment has commenced in 2004 and is still ongoing. The CAD patients were selected from the UCORBIO cohort, which also has been described previously.22 In this cohort, patients undergoing coronary angiography for any indication were enrolled between October 2011 and December 2014 and followed for the occurrence of adverse events. For the current study patients with no history of CAD and no CAD upon angiography were excluded (n=406); leaving 1,967 CAD patients for analysis.

217


Chapter 11

Follow-up Both studies conducted follow-up by means of patient questionnaires on a yearly basis. When patients did not respond or when they reported events of interest, general practitioners and treating specialists were contacted for verification of the events. For the current analysis a composite end-point of major adverse cardiovascular events (MACE) and all-cause mortality were used. MACE consisted of: myocardial infarction, stroke, cardiovascular death, percutaneous coronary intervention, coronary artery bypass grafting, percutaneous transluminal angioplasty, peripheral arterial surgery and amputation due to arterial insufficiency. Health-related quality of life The protocols of the Athero-express study and the UCORBIO cohort both involved a HRQOL questionnaire at the moment of enrolment. In the Athero-express study the RAND-3623 questionnaire was filled in by the participants (response rate 63%) and in the UCORBIO cohort the RAND-36 and the EuroQoL24 self-rated health grade were obtained (response rate 73%). The RAND-36 questionnaire (version 1) consisted of nine domains of HRQOL: physical functioning, social functioning, physical role limitations, emotional role limitations, mental functioning, pain, vitality, general health and health change. In order to construct one “summary” measure of HRQOL in both cohorts we calculated the coefficients of the RAND-36 items in a linear regression model (with intercept set at zero) for the EuroQoL self-rated health grade, separately for men and women in the UCORBIO cohort, which had both RAND-36 and EuroQoL measurements. We applied these coefficients to the RAND-36 responses for all patients (also the CAD patients) and computed HRQOL, reported as HRQOLcomp. The formulae for the calculation of HRQOLcomp in men and women are provided in the supplemental methods. The HRQOLcomp measure resulting from these formulae has been used throughout this paper. Statistical approach R software for statistical computing, version 3.1.225 was used for all analyses. Normally distributed continuous data were presented as means ± standard deviation (SD); nonnormally distributed continuous data were presented as medians with interquartile ranges (IQR). Categorical data were presented as percentages per category and compared using a chi-square test. The level of statistical significance was set at α <0.05. Our primary analysis consisted of the comparison of HRQOLcomp (computed as described above) between men and women with PAD and CAD using a Mann-Whitney U test. Secondly, in women we examined the relation of cardiovascular risk factors, femalespecific risk factors and disease characteristics with HRQOLcomp using linear regression. Thirdly, we evaluated the relation of HRQOLcomp with MACE and all-cause mortality and tested for interactions between sex and HRQOLcomp for outcome in a Cox regression model. In all models, age and risk factors related to HRQOLcomp were added as covariates.

218


Sex Differences in the Association of HRQOL with outcome

The risk factors related to HRQOLcomp were BMI, history of coronary intervention, diabetes (for PAD patients only), hypertension (CAD patients only), location of PAD (PAD patients only) and history of PAD (CAD patients only).

Results Patient characteristics The characteristics of women and men are displayed in table 1. Women were older (67.2 vs. 66.1 years, p=0.005) and there were relatively more women among the PAD patients. BMI was lower (26.5 vs. 26.9, p=0.009), but hypertension was more prevalent among women (73.4% vs. 65.4%, p<0.001). Women less often had a history of myocardial infarction or coronary intervention, but more often had a history of cerebrovascular accident (CVA) (44.3% vs. 36.7%, p<0.001). Kidney function was poorer in women than men (estimated glomerular filtration rate 74.2 vs. 81.4 ml/min, p<0.001). Among PAD women, the location of PAD was more often carotid (as opposed to femoral) than in men, p=0.017. In female femoral PAD patients the Fontaine classification was more severe than in men. Women with carotid PAD were less often asymptomatic than men (10.3% vs. 15.4%, p=0.042). The presence of an occlusive contralateral stenosis was less common in women than men with carotid PAD. Women with CAD had less severe angiographic CAD; triple vessel disease was observed in 19.1% of men vs. 14.9% of women, p=0.009. Also, left ventricular ejection fraction (EF) was less impaired in women than men (normal EF 71.4% of women vs. 53.5% of men, p<0.001). Health-related Quality of Life Women (PAD and CAD combined) reported lower HRQOL in six of nine RAND-36 domains (table 2). The EuroQoL health grade was lower in women than men with CAD. The HRQOLcomp was also lower in women than in men (median 6.4 for women with PAD and CAD, 6.7 for men with PAD and 6.8 for men with CAD, p=0.007, shown in Figure 1). Female-specific risk factors The prevalence of female-specific risk factors is shown in table 3. Very few women were still menstruating; 5.6% of PAD women and 5.4% of CAD women. All-time use of oral contraceptive pill (OCP) was higher among PAD women than CAD women, 66.6% vs. 46.2%, p<0.001. Also, the duration of OCP use differed, p=0.026. The number of pregnancies was higher among PAD women than CAD women, although the median number of pregnancies was 2 in both groups (p=0.008). Gestational diabetes was significantly more common in CAD patients than PAD patients (8.3% vs. 4.1%, p=0.036).

219


Chapter 11

Table 1. Patient characteristics of PAD and CAD patients, stratified by sex. n Age (mean ±sd) Type of disease (%)

PAD CAD

Female

Male

840

2456

p-value

67.2 ±10.2

66.1 ±9.5

0.005

525 (62.5)

1351 (55.0)

<0.001

315 (37.5)

1105 (45.0)

BMI (mean ±sd)

26.5 ±5.0

26.9 ±3.8

Diabetes (%)

184 (21.9)

576 (23.5)

0.380

Hypertension (%)

613 (73.4)

1593 (65.4)

<0.001

501 (60.7)

1466 (60.5)

0.982

243 (30.1)

589 (24.9)

<0.001

Former smoker

357 (44.2)

1252 (52.9)

Active smoker

208 (25.7)

526 (22.2)

History of MI (%)

160 (19.1)

739 (30.2)

<0.001

History of CI (%)

193 (23.1)

865 (35.4)

<0.001

History of CVA (%)

370 (44.3)

896 (36.7)

<0.001

History of PAD (%)

224 (26.7)

726 (29.7)

0.112

eGFR (mean ±sd)

74.2 ±24.5

81.4 ±32.6

<0.001

Aspirin (%)

634 (75.6)

1767 (72.1)

0.056

P2Y12 inhibitors (%)

153 (18.2)

472 (19.2)

0.555

RAAS inhibitors (%)

434 (51.7)

1347 (54.8)

0.120

Beta-blocker (%)

438 (52.2)

1213 (49.4)

0.181

Statin (%)

612 (72.9)

1772 (72.2)

0.726

Diuretic (%)

330 (39.3)

762 (31.1)

<0.001

0.017

Hypercholesterolemia (%) Smoking (%)

Non smoker

0.009

PAD* Location of PAD (%)

Carotid

386 (76.3)

931 (70.5)

Femoral

120 (23.7)

389 (29.5)

2

133 (46.7)

442 (55.9)

3

107 (37.5)

235 (29.7)

4

45 (15.8)

113 (14.3)

Femoral patients Fontaine classification (%)

Ankle-brachial index (mean ±sd) Carotid stenosis (%)

0.021

0.7 ±0.3

0.7 ±0.2

0.660

392 (75.2)

937 (70.6)

0.050 0.042

Carotid patients Indication CEA (%)

Contralateral stenosis (%)

Asymptomatic

41 (10.3)

146 (15.4)

TIA

254 (63.8)

577 (61.1)

Stroke

103 (25.9)

222 (23.5)

0-50%

237 (59.1)

512 (51.7)

50-70%

56 (14.0)

126 (12.7)

70-99%

65 (16.2)

163 (16.4)

100%

43 (10.7)

190 (19.2)

169 (53.7)

626 (56.7)

36 (11.4)

97 (8.8)

90 (28.6)

326 (29.5)

20 (6.3)

55 (5.0)

0.001

CAD** Indication (%)

Stable CAD Unstable CAD Infarction Other

220

0.365


Sex Differences in the Association of HRQOL with outcome

Table 1. Continued Female

Male

p-value

39 (12.4)

84 (7.6)

0.009

Single vessel disease

137 (43.5)

434 (39.3)

Double vessel disease

92 (29.2)

374 (33.9)

Triple vessel disease

47 (14.9)

211 (19.1)

167 (71.4)

448 (53.5)

34 (14.5)

221 (26.4)

Impaired

22 (9.4)

104 (12.4)

Poor

11 (4.7)

64 (7.6)

70 (22.9)

217 (19.9)

219 (71.6)

800 (73.3)

17 (5.6)

74 (6.8)

Angiographic CAD severity (%) No or minor CAD

Ejection fraction (%)

Normal Mildly impaired

Procedure (%)

Conservative PCI CABG

<0.001

0.429

*PAD patients from the Athero-express biobank cohort. ** CAD patients from the UCORBIO cohort. Abbreviations: MI= myocardial infarction, CI= coronary intervention (PCI or CABG), CVA= cerebrovascular accident, GFR= glomerular filtration rate (MDRD formula).

Table 2. Quality of life stratified by sex for PAD and CAD patients. PAD

CAD Women vs. Men

Women

Men

Women

Men

525

1351

315

1105

Physical functioning

50 [30, 70]

65 [40, 85]

60 [35, 80]

80 [55, 95]

Social functioning

62 [38, 88]

75 [50, 88]

62 [50, 88]

75 [50, 100]

Role limitations, physical

75 [0, 100]

50 [0, 100]

25 [0, 100]

50 [0, 100]

0 [0, 100]

0 [0, 67]

100 [33, 100]

100 [67, 100]

0.442

68 [52, 84]

76 [64, 88]

72 [56, 84]

80 [64, 88]

<0.001

n

Role limitations, emotional Mental health Vitality

<0.001 <0.001 0.085

50 [35, 70]

60 [40, 75]

55 [35, 65]

60 [45, 75]

<0.001

67 [45, 100]

78 [45, 100]

67 [45, 100]

78 [55, 100]

<0.001

General health perception

60 [45, 70]

60 [50, 70]

55 [35, 70]

65 [40, 75]

<0.001

Health change

50 [25, 50]

50 [25, 50]

0.365

Pain

EuroQoL health grade HRQOLcomp

6.4 [5.1, 7.8]

6.7 [5.1, 7.9]

50 [25, 50]

50 [25, 50]

7.0 [6.0, 7.5]

7.0 [6.0, 8.0]

0.011

6.4 [5.2, 7.7]

6.8 [5.5, 7.9]

0.007

Medians and interquartile ranges are shown for the nine RAND-36 domains, the EuroQoL health grade (only CAD patients) and HRQOLcomp. Differences between women and men were tested with Kruskal-Wallis testing.

221


Chapter 11

Associations of patient characteristics with HRQOL in women Only BMI and a history of coronary intervention (PCI or CABG) were associated with lower HRQOLcomp in both PAD and CAD women (supplemental table 1). Among PAD women diabetes was associated with a 0.47 point lower HRQOLcomp (β -0.47, p=0.021). Also, location of PAD (femoral vs. carotid), higher Fontaine classification, indication for CEA and lower ankle-brachial index were associated with lower HRQOLcomp in PAD women. Among CAD women, hypertension (0.48 points lower) and a history of PAD (0.66 point lower) were associated with lower HRQOLcomp. No CAD-specific factors were associated with HRQOLcomp among CAD women. Of the female-specific risk factors only menstrual status (β -1.04, p=0.007) was significantly related to HRQOLcomp, only in PAD women. PAD women who were still menstruating (n=25) as compared to (post) menopausal women had a 1.04 point lower HRQOLcomp. When adjusted for age, this association was lost. In CAD women, none of the female-specific risk factors were related to HRQOLcomp.

PAD % of responses per sex

25%

p−value for sex−difference: 0.263

20% 15% 10% 5% 0%

0

1

2

3

4

5

6

7

8

9

10

6

7

8

9

10

CAD % of responses per sex

25%

p−value for sex−difference: 0.002

20% 15% 10% 5% 0%

0

1

2

3

4

5

HRQOLcomp Women

Men

Figure 1. Distribution of HRQOLcomp stratified by CAD/PAD and by sex The dashed lines represent the median HRQOLcomp values for men and women, which were significantly different between men and women among CAD patients but not among PAD patients. The p-values are derived from Kruskal-Wallis testing.

222


Sex Differences in the Association of HRQOL with outcome

Table 3. Female-specific risk factors in women with PAD and CAD.

Age (mean ±sd)

PAD

CAD

P-value

67.4 ±10.1

66.9 ±10.3

0.486

Menstruation Currently menstruating (%) Age menopause (%)

25 (5.6)

16 (5.4)

1

<45 years

103 (27.8)

64 (24.3)

0.076

45-48 years

77 (20.8)

61 (23.2)

49-52 years

112 (30.3)

99 (37.6)

>53 years

78 (21.1)

39 (14.8)

Hormonal medication OCP use (ever %) Years of OCP (%)

295 (66.6)

139 (46.2)

<0.001

<5 years

65 (26.4)

24 (17.9)

0.026

5-10 years

46 (18.7)

37 (27.6)

10-20 years

88 (35.8)

38 (28.4)

>20 years

47 (19.1)

35 (26.1)

32 (7.7)

26 (9.9)

0.387 0.556

Use of HRT (ever %) Years of HRT use (%)

<5 years

11 (37.9)

9 (37.5)

5-10 years

11 (37.9)

11 (45.8)

10-20 years

6 (20.7)

2 (8.3)

>20 years

1 (3.4)

2 (8.3)

Pregnancy Pregnancies (n)

2 [2, 3]

2 [1, 3]

0.008

Miscarriages (n)

0 [0, 0]

0 [0, 1]

0.190

Gestational DM (%)

17 (4.1)

21 (8.3)

0.036

Gestational HT (%)

n/a

62 (24.5)

n/a

Pre-eclampsia (%)

n/a

18 (7.1)

n/a

Abbreviations: OCP= oral contraceptive pill, HRT= hormone replacement therapy, DM= diabetes mellitus, HT= hypertension.

Relation of HRQOL with MACE and All-cause Mortality Among the PAD patients, during a median follow-up duration of 3.0 years (IQR= 2.5 to 3.1) 375 men and 126 women experienced a MACE and 138 men and 44 women died. CAD patients were followed-up for a median duration of 2.1 years (IQR= 1.3 to 2.9), during which 132 men and 45 women had a MACE and 41 men and 10 women died. The multivariable adjusted association of HRQOLcomp with MACE incidence and all-cause mortality in men and women with PAD and CAD are shown in Figure 2. The association of HRQOLcomp with MACE was similar for men and women with PAD (HR 1.16 [1.10-1.22], p<0.001 for each point decrease in HRQOLcomp and 1.21 [1.10-1.32], p<0.001, respectively), but differed slightly between men and women with CAD (HR 1.29 [1.16-1.42], p<0.001 vs. 1.12 [0.94-1.34], p=0. 20, respectively). The difference between the association of

223


Chapter 11

HRQOLcomp with MACE in men and women was not significant (p for interaction 0.21). Lower HRQOLcomp was also significantly associated with increased risk of all-cause mortality in men and women with PAD (HR 1.14 [1.04-1.25], p=0.005 and 1.23 [1.05-1.44], p=0.009, respectively) and men with CAD (HR 1.63 [1.36-1.95], p<0.001), but not in women with CAD (HR 1.10 [0.74-1.64], p=0.65). There was a significant difference in the association of HRQOLcomp with all-cause mortality in men and women (p for interaction 0.030). MACE, PAD

All−cause Mortality, PAD

3.0

3.0 Women

Women 2.5

Men Hazard Ratio

Hazard Ratio

2.5 2.0 1.5 1.0

Men

2.0 1.5 1.0

0.5

0.5 10

9

8

7

6

5

4

3

2

1

10

9

8

6

5

4

3

HRQOL (computed)

MACE, CAD

All−cause Mortality, CAD

3.0

2

1

2

1

3.0 Women

2.5

Women 2.5

Men Hazard Ratio

Hazard Ratio

7

HRQOL (computed)

2.0 1.5 1.0

Men

2.0 1.5 1.0

0.5

0.5 10

9

8

7

6

5

4

HRQOL (computed)

3

2

1

10

9

8

7

6

5

4

3

HRQOL (computed)

Figure 2. Multivariable adjusted hazard ratios of HRQOLcomp for MACE and all-cause mortality The hazard ratios (HRs) of HRQOLcomp (adjusted for age, BMI, diabetes (only PAD patients) hypertension (only CAD patients), location of PAD (only PAD patients), history of PAD (only CAD patients) and history of coronary intervention) are plotted for each of its values, for men (blue) and women (red) separately. The top panels display HRs for PAD patients for MACE (left) and all-cause mortality (right), the bottom panel shows HRs for CAD patients. HRQOLcomp was significantly associated with outcome in both sexes, except for MACE and all-cause death in CAD women. The association of HRQOLcomp with all-cause death was significantly stronger in men than in women with CAD (bottom right plot, p for interaction 0.030).

224


Sex Differences in the Association of HRQOL with outcome

Discussion Cardiovascular diseases such as CAD and PAD have a significant impact on HRQOL.1,2 Poor HRQOL is associated with an increased risk of adverse cardiovascular events3,4 and all-cause mortality6. Previous studies have consistently reported that women tend to report markedly lower HRQOL than men.7,26 Our results from this current study concur with these findings in that according to the HRQOLcomp, women with CAD suffer from a significantly poorer quality of life than their male counterparts. But HRQOLcomp in women with CAD did not predict the occurrence of MACE and all-cause death, while it did in men. Hypothesized reasoning regarding the sex differences in HRQOL has in the past included the (expected) role of masculine and feminine gender within society. One study computed a constructed gender role variable using structural equation modeling26, expressing a person’s gender in addition to biological sex. Norris et al. found that estimation of HRQOL was significantly improved upon addition of the “gender role” variable. Suggesting that the socially constructed “gender” might influence coping with cardiovascular disease and subsequent HRQOL more than biological “sex”. In addition, a possible explanation that has not been investigated in previous studies may relate to female-specific risk factors. In our study, among the women with PAD, HRQOL was associated with menstrual status but this association was lost upon adjustment for age. No association between female-specific risk factors and HRQOLcomp in women with CAD that could explain the sex difference in HRQOL. It is also important to consider the possibility that the women in our cohorts may be suffering from a greater prevalence of debilitating chronic long-term diseases such as osteoarthritis27 and fibromyalgia28, impacting emotional and physical functionality and subsequently HRQOL. This is of particular relevance with respect to our cohort of patients of advanced age, as these diseases tend to affect women more than men in the higher age group. Unfortunately, in our cohort these data were unavailable. Despite the sex difference in HRQOL in patients with CAD, we found that interestingly, this poorer perceived quality of life among women was not significantly related to poorer outcome. In our cohort, women with low HRQOL did not have an increased risk of MACE and all-cause mortality whereas men with a reported impaired quality of did have a worse outcome in terms of MACE and all-cause death. The difference in the occurrence of MACE was less pronounced and did not reach statistical significance (no significant interaction), while the difference in all-cause death did. Speculating, it might be that the sex difference in the relation of HRQOLcomp with outcome is more pronounced in noncardiovascular mortality, which was the case in 70% of CAD men and women who died. Unfortunately the information about comorbidities such as oncological diseases was too limited in our data to test this hypothesis. HRQOL has gained increasing attention over the years as an outcome measure of treatment and interventions of patients with established CVD.29 It is also now becoming an accepted part of routine clinical practice. It is a useful tool as it allows patients to

225


Chapter 11

inform healthcare professionals of harmful medication effects when treatments are effective so risks and benefits of such treatment can be weighed up. Also, it provides additional information when symptoms relating to diseases are only mild. Our results however, suggest that caution must be exercised when interpreting these results for women with CAD. Our results imply that women with CAD may be more likely to report poor quality of life. This has an important public health impact as low HRQOL is associated with increased healthcare expenditure3 due to healthcare seeking. Historically, compared with men, women tend to have the lion’s share of health care service use30, so women may actively seek out healthcare in an attempt to rectify their poor perceived HRQOL despite not necessarily having a poor cardiovascular outcome. This results in repeated hospitalizations and potentially unnecessary treatment and interventions subsequently further impacting quality of life in a detrimental manner. Limitations Due to a limited number of events in certain patient groups (particularly in CAD women), estimates of effect sizes are of limited accuracy. The non-significant associations of HRQOLcomp with outcome in CAD women should therefore be interpreted with caution. Larger cohorts are needed to evaluate the independent prognostic value of HRQOL in men and women with cardiovascular disease, further adjustment for possible confounders is needed. Conclusion Women with CAD reported lower HRQOLcomp than men; among PAD patients no sexdifferences were found. Female-specific risk factors could not explain lower HRQOLcomp in women. Lower HRQOLcomp was related to higher risk of MACE and all-cause death in PAD men and women. Lower HRQOLcomp was also predictive of MACE and all-cause death in men with CAD. This relation was weaker for CAD women. While women with CAD reported lower HRQOLcomp, the conferred risk of adverse events and mortality might be lower for women than men.

226


Sex Differences in the Association of HRQOL with outcome

References 1.

Haitjema S, Borst G De, Vries J De, Moll F, Pasterkamp G, Ruijter H Den. Health-related quality of life is poor but does not vary with cardiovascular disease burden among patients operated for severe atherosclerotic disease. IJCHV. 2014;4:53–58.

2.

Gijsberts CM, Agostoni P, Hoefer IE, Asselbergs FW, Pasterkamp G, Nathoe H, Appelman YE, Kleijn DPV de, Hester M HM den R. Gender differences in health-related quality of life in patients undergoing coronary angiography. Open Hear. 2015;Epub.

3.

Harrison PL, Pope JE, Coberley CR, Rula EY. Evaluation of the relationship between individual well-being and future health care utilization and cost. Popul Health Manag. 2012;15:325–30.

4.

Pedersen SS, Martens EJ, Denollet J, Appels A. Poor health-related quality of life is a predictor of early, but

5.

Gijsberts CM, den Ruijter HM. Non-Response to Questionnaires Independently Predicts Mortality of Coronary

not late, cardiac events after percutaneous coronary intervention. Psychosomatics. 2007;48:331–337.

Angiography patients. Int J Cardiol. 2015;Epub. 6.

Lenzen MJ, Scholte op Reimer WJM, Pedersen SS, Boersma E, Maier W, Widimsky P, Simoons ML, Mercado NF, Wijns W. The additional value of patient-reported health status in predicting 1-year mortality after invasive coronary procedures: a report from the Euro Heart Survey on Coronary Revascularisation. Heart. 2007;93:339–44.

7.

Norris CM, Spertus J a, Jensen L, Johnson J, Hegadoren KM, Ghali W a. Sex and gender discrepancies in health-related quality of life outcomes among patients with established coronary artery disease. Circ Cardiovasc Qual Outcomes. 2008;1:123–30.

8.

Bakhai A, Ferrières J, James S, Iñiguez A, Mohácsi A, Pavlides G, Belger M, Norrbacka K, Sartral M. Treatment, outcomes, costs, and quality of life of women and men with acute coronary syndromes who have undergone percutaneous coronary intervention: results from the antiplatelet therapy observational registry. Postgrad Med. 2013;125:100–7.

9.

Kendel F, Dunkel A, Müller-Tasch T, Steinberg K, Lehmkuhl E, Hetzer R, Regitz-Zagrosek V. Gender differences in health-related quality of life after coronary bypass surgery: results from a 1-year follow-up in propensitymatched men and women. Psychosom Med. 2011;73:280–5.

10. Hlatky MA, Chung SC, Escobedo J, Hillegass WB, Melsop K, Rogers W, Brooks MM. The effect of obesity on quality of life in patients with diabetes and coronary artery disease. Am Heart J. 2010;159:292–300. 11. Oreopoulos A, Padwal R, McAlister FA, Ezekowitz J, Sharma AM, Kalantar-Zadeh K, Fonarow GC, Norris CM. Association between obesity and health-related quality of life in patients with coronary artery disease. Int J Obes (Lond). 2010;34:1434–41. 12. Uchmanowicz I, Loboz-Grudzien K, Jankowska-Polanska B, Sokalski L. Influence of diabetes on healthrelated quality of life results in patients with acute coronary syndrome treated with coronary angioplasty. Acta Diabetol. 2013;50:217–25. 13. Stafford L, Berk M, Jackson HJ. Tobacco smoking predicts depression and poorer quality of life in heart disease. BMC Cardiovasc Disord. 2013;13:35. 14. Appelman Y, van Rijn BB, Ten Haaf ME, Boersma E, Peters S a E. Sex differences in cardiovascular risk factors and disease prevention. Atherosclerosis. 2015;1–8. 15. Jacobs AK. Coronary intervention in 2009: are women no different than men? Circ Cardiovasc Interv. 2009;2:69–78.

227


Chapter 11

16. Bairey Merz CN, Shaw LJ, Reis SE, Bittner V, Kelsey SF, Olson M, Johnson BD, Pepine CJ, Mankad S, Sharaf BL, Rogers WJ, Pohost GM, Lerman A, Quyyumi A a., Sopko G. Insights from the NHLBI-sponsored Women’s Ischemia Syndrome Evaluation (WISE) study. Part II: Gender differences in presentation, diagnosis, and outcome with regard to gender-based pathophysiology of atherosclerosis and macrovascular and microvascular cor. J Am Coll Cardiol. 2006;47. 17. Den Ruijter H, Pasterkamp G, Rutten FH, Lam CSP, Chi C, Tan KH, van Zonneveld a J, Spaanderman M, de Kleijn DP V. Heart failure with preserved ejection fraction in women: the Dutch Queen of Hearts program. Neth Heart J. 2015;23:89–93. 18. Gierach GL, Johnson BD, Bairey Merz CN, Kelsey SF, Bittner V, Olson MB, Shaw LJ, Mankad S, Pepine CJ, Reis SE, Rogers WJ, Sharaf BL, Sopko G. Hypertension, menopause, and coronary artery disease risk in the Women’s Ischemia Syndrome Evaluation (WISE) Study. J Am Coll Cardiol. 2006;47:S50–8. 19. Verhoeven BAN, Velema E, Schoneveld AH, De Vries JPPM, De Bruin P, Seldenrijk CA, De Kleijn DP V, Busser E, Van Der Graaf Y, Moll F, Pasterkamp G. Athero-express: Differential atherosclerotic plaque expression of mRNA and protein in relation to cardiovascular events and patient characteristics. Rationale and design. Eur J Epidemiol. 2004;19:1127–1133. 20. Gijsberts CM, Gohar A, Ellenbroek GHJM, Hoefer IE, de Kleijn DPV, Asselbergs FW, Nathoe HM, Agostoni P, Rittersma SZH, Pasterkamp G, Appelman Y, den Ruijter HM. Severity of stable coronary artery disease and its biomarkers differ between men and women undergoing angiography. Atherosclerosis. 2015;241:234–240. 21. Zee K Van de, Sanderman R. Het meten van de algemene gezondheidstoestand met de RAND-36: een handleiding. Rijksuniversiteit, Groningen)(ISBN 90–72156–60–9). 1993;28. 22. Gijsberts CM, Gohar A, Ellenbroek GHJM, Hoefer IE, de Kleijn DPV, Asselbergs FW, Nathoe HM, Agostoni P, Rittersma SZH, Pasterkamp G, Appelman Y, den Ruijter HM. Severity of stable coronary artery disease and its biomarkers differ between men and women undergoing angiography. Atherosclerosis. 2015;241:234–240. 23. VanderZee KI, Sanderman R, Heyink JW, de Haes H. Psychometric qualities of the RAND 36-Item Health Survey 1.0: a multidimensional measure of general health status. Int J Behav Med. 1996;3:104–122. 24. EuroQoL group. EuroQol- a new facility for the measurement of health-related quality of life. Health Policy. 1990;16:220–33. 25. R Core Team. R: A Language and Environment for Statistical Computing. 2013; 26. Norris CM, Murray JW, Triplett LS, Hegadoren KM. Gender roles in persistent sex differences in health-related quality-of-life outcomes of patients with coronary artery disease. Gend Med. 2010;7:330–9. 27. Zhang Y, Jordan JM. Epidemiology of Osteoarthritis. Clin Geriatr Med. 2010;26:355–369. 28. Wolfe F, Hawley DJ. Measurement of the quality of life in rheumatic disorders using the EuroQol. Br J Rheumatol. 1997;36:786–793. 29. Dyer MTD, Goldsmith K a, Sharples LS, Buxton MJ. A review of health utilities using the EQ-5D in studies of cardiovascular disease. Health Qual Life Outcomes. 2010;8:13. 30. Orfila F, Ferrer M, Lamarca R, Tebe C, Domingo-Salvany A, Alonso J. Gender differences in health-related quality of life among the elderly: the role of objective functional capacity and chronic conditions. Soc Sci Med. 2006;63:2367–2380.

228


Sex Differences in the Association of HRQOL with outcome

Supplemental methods Formulae for calculation of HRQOLcomp in men and women The formula for HRQOLcomp in men was: 0.0128 * physical functioning + 0.0084 * social functioning + -0.0053 * physical role limitations + -0.0015 * emotional role limitations + 0.0285 * mental functioning + 0.0121 * vitality + 0.0108 * pain + 0.0239 * general health + 0.0100 * health change. For HRQOLcomp in women it was: 0.0086 * physical functioning + -0.0023 * social functioning + -0.0005 * physical role limitations + -0.0046 *emotional role limitations + 0.0416 * mental functioning + 0.0156 * vitality + 0.0122 * pain + 0.0261 * general health + 0.0095 * health change.

229


230

History of coronary intervention

0.551

-0.23 (-0.47- 0.02) -0.05 (-0.97- 0.87)

N pregnancies (per 1)

N miscarriages (per 1)

Gestational diabetes

n/a n/a

Gestational hypertension*

Pre-eclampsia*

0.911

0.068

0.475 0.04 (-0.08- 0.15)

Duration HRT use (four categories)

0.658

-0.17 (-0.92- 0.58)

Use of HRT

-0.05 (-0.82- 0.73)

0.18 (-0.28- 0.64)

0.18 (-0.54- 0.89)

-0.12 (-0.27- 0.04)

0.04 (-0.08- 0.17)

-0.08 (-0.86- 0.70)

0.21 (-0.61- 1.03)

0.907

0.439

0.627

0.134

0.483

0.376

0.841

0.344

0.612

0.197

0.007

Menopausal age (four categories)

-1.04 (-1.81--0.28)

0.194

0.476

0.035

0.415

0.021

0.135

0.595

0.843

0.208

Menstruating (currently vs. stopped)

0.24 (-0.12- 0.61)

0.03 (-0.05- 0.10)

-0.66 (-1.28--0.05)

0.164 0.015

0.669

0.703

0.106

0.071

-0.27 (-0.92- 0.38)

-0.45 (-0.83--0.07)

0.034 0.585

-0.31 (-0.72- 0.10)

-0.13 (-0.60- 0.34)

0.05 (-0.43- 0.53)

-0.23 (-0.60- 0.13)

0.217

0.052

0.820

0.904

0.672 0.001

0.908

0.07 (-0.30- 0.45)

0.06 (-0.01- 0.13)

-0.31 (-0.75- 0.13) -0.48 (-0.87--0.10)

0.021 0.088

-0.31 (-0.49--0.13)

0.042

p-value

Duration OCP use (four categories)

OCP use (ever vs. never)

Hormonal

GFR (per 10 ml/min increase)

0.10 (-0.25- 0.44)

-0.47 (-0.91--0.04)

History of myocardial infarction

-0.31 (-0.65- 0.03)

-0.29 (-0.76- 0.17)

Smoking (Current vs. never)

History of PAD

-0.50 (-1.01- 0.00)

Smoking (Ex vs. never)

History of cerebrovascular accident

-0.02 (-0.38- 0.34) -0.05 (-0.50- 0.40)

Hypercholesterolemia

0.34 (-0.05- 0.73)

-0.47 (-0.86--0.07)

Diabetes

Hypertension

-0.17 (-0.34--0.01)

BMI (per 5 points increase)

Beta (95% CI) 0.04 (-0.14- 0.21)

0.548

p-value

Beta (95% CI) 0.05 (-0.11- 0.21)

Age (per 10 years increase)

CAD

PAD

Supplemental table. Associations of patient characteristics with HRQOLcomp in PAD and CAD women.

Chapter 11


0.17 (-0.21- 0.55)

Carotid stenosis 0.565

0.522

0.461

0.273

0.076

All betas represent the increase in HRQOLcomp for yes vs. no for each given patient characteristic, unless stated otherwise. The categories of variables with more than two categories are as described in table 1. *Only available for CAD patients. Abbreviations: BMI= body mass index, GFR= glomerular filtration rate, RAAS= renin-angiotensinaldosterone system, OCP= oral contraceptive pill, HRT= hormone replacement therapy, CEA= carotid endarterectomy, ABI= ankle-brachial index, EF= left ventricular ejection fraction.

Contralateral stenosis (four categories)

Procedure (three categories)

0.391

CAD severity (four categories)

0.041 0.020

Indication (four categories)

EF (four categories)

CAD specific

1.84 (0.29- 3.38)

<0.001

0.025

ABI operated leg (per 1 increase)

-0.44 (-0.82--0.06)

Indication CEA (stroke, TIA or asymptomatic)

Fontaine classification (2, 3 or 4)

PAD location (femoral vs. carotid)

PAD specific

Supplemental table. Continued

Sex Differences in the Association of HRQOL with outcome

231



PART THREE General Risk Prediction

Chapter 12 Non-Response to Questionnaires Independently Predicts Mortality of Coronary Angiography patients Int J Cardiol. 2015 Jul 2;201:168-170

Crystel M. Gijsberts, Hester M. den Ruijter


Chapter 12

Health-related Quality of Life (HRQOL) is known to have an impact on outcome in several medical fields. Specifically, it is predictive of mortality among coronary artery disease (CAD) patients1. HRQOL can be measured with several field-specific or general questionnaires, of which the EuroQoL2 is of frequent occurrence. On average approximately 80% of study participants will return their questionnaire3, the so-called “responders�. In the interpretation of questionnaire results, a high response rate is crucial4 in order to accurately assess its value. Logically however, only HRQOL data from responders can be analyzed. In the current study, we compared survival between EuroQoL HRQOL questionnaire responders and non-responders in a cohort of coronary angiography patients. For this purpose 2,392 coronary angiography patients, enrolled in the UCORBIO cohort5 between 1 October 2011 and 31 December 2014 were analyzed. This cohort is registered at clinicaltrials.gov under the identifier: NCT02304744. All patients provided written informed consent. The study protocol conforms to the ethical guidelines of the 1975 Declaration of Helsinki as reflected in a priori approval by the medical ethics committee of the University Medical Center Utrecht. All patients undergoing coronary angiography, >18 years were eligible for participation in this study. Directly after coronary angiography patients were provided with a HRQOL questionnaire consisting of the RAND-36 questionnaire6 and the self-rated health grade EuroQoL item which rates HRQOL on a scale of 1 to 10. Patients are followed-up for the occurrence of adverse events and allcause mortality over the course of 5 years, of which maximally 3.7 years have passed at the moment of writing. The study population as described in the current paper consisted of 1,886 responders and 506 non-responders (response rate 79%). The median follow-up duration was 1.7 years. Responders were divided in two groups, high HRQOL (EuroQoL score >=7 out of 10) and low HRQOL (EuroQoL score <7). The baseline characteristics are described in Table 1. In short, non-responders were younger (62.1 vs. 64.4 years, p<0.001), their body mass index was higher (27.5 vs. 27.0, p=0.029), more often they were current smokers (33.5% vs. 22.3%, p<0.001), they more often had a history of a cerebrovascular accident or transient ischemic attack (12.7% vs. 9.2%, p=0.028), they had more kidney failure (5.2% vs. 2.2%, p=0.001) and the indication for angiography more often was a myocardial infarction (33.6% vs. 25.5%, p=0.001) than among responders (low and high HRQOL combined). All-cause mortality in non-responders was significantly higher than in responders, as depicted in Figure 1. Kaplan-Meier estimates for 3-year all-cause mortality were 16.7% for non-responders vs. 8.5% for responders who reported low HRQOL and 4.2% for responders who reported high HRQOL (p<0.001). In a multivariable Cox regression analysis we adjusted for possible confounders: age, sex, BMI, smoking, diabetes, hypertension, hypercholesterolemia, history of acute coronary syndrome, history of percutaneous coronary intervention, history of coronary artery bypass grafting surgery, history of cerebrovascular accident, history of peripheral arterial disease, kidney failure, left ventricular ejection fraction, indication for angiography, angiographic severity of CAD

234


Non-response to Questionnaires Predicts Mortality

and treatment of CAD. From this analysis we found a hazard ratio (HR) of 2.69 (95% CI 1.64-4.42, p<0.001) for non-responders as compared to responders with low HRQOL and a HR of 4.67 (95% CI 2.62-8.33, p<0.001) as compared to responders with high HRQOL. Indicating that non-responders have significantly poorer survival than responders with low or high HRQOL, independent of differences in risk factors, cardiovascular medical history or findings upon coronary angiography. This phenomenon of nonresponse bias has been reported before, in other settings.7,8 Therefore, we conclude that results from questionnaire-based studies among coronary angiography patients should be interpreted with care. Data gathered by means of questionnaires cannot be generalized to the whole patient population due to a profound non-response bias.

All−cause Mortality by HRQOL response 1.0

0.9

p < 0.001

Survival

0.8

0.7

High HRQOL

0.6

Low HRQOL

Non−responder 0.5

0

100

200

300

400

500

600 700 Time (days)

800

900

1000

1100

1200

1300

High HRQOL 1024

1023

985

921

844

763

680

588

506

397

307

238

155

34

Low HRQOL 821

815

797

716

629

573

492

397

325

254

203

149

102

27

Non−responder 506

492

460

392

348

293

265

225

191

163

130

103

71

22

Numbers at risk

Figure 1. All-cause Mortality of Responders and Non-responders of HRQOL questionnaire. Kaplan-Meier survival curves for responders with high HRQOL (EuroQoL score ≥7), responders with low HRQOL (EuroQoL score <7) and non-responders (p<0.001).

235


Chapter 12

Table 1. Baseline characteristics of Coronary Angiography patients, stratified by HRQOL questionnaire response status.

N

Overall

Responders High HRQOL

Responders Low HRQOL

Nonresponders

p-value

2,392

1,024

821

506

63.9 ±11.0

64.5 ±10.4

64.3 ±11.0

62.1 ±12.2

72.9

76.5

69.3

71.9

0.618

27.1 ±4.4

26.8 ±4.0

27.4 ±4.7

27.5 ±4.8

0.029

Risk factors Age (mean ± SD) Sex (% males) BMI (mean ± SD)

<0.001

Diabetes (% yes)

22.5

17.6

26.2

25.4

0.083

Hypertension (% yes)

58.0

53.1

63.6

57.9

0.993

Hypercholesterolemia (% yes)

46.4

44.9

50.0

43.1

Smoking (%)

0.111 <0.001

Current smoker

24.7

20.5

24.3

33.5

Ex smoker

26.0

26.3

27.7

22.5

Non-smoker

49.3

53.2

48.0

44.0

History of ACS (% yes)

30.7

28.6

34.5

28.4

0.243

History of PCI (% yes)

28.3

24.4

34.0

26.4

0.332

History of CABG (% yes)

11.1

8.4

14.9

9.7

0.281

History of CVA/TIA (% yes)

9.9

7.2

11.1

12.7

0.024

History of PAD (% yes)

11.4

9.0

13.8

11.9

0.739

Kidney failure (% yes)

2.9

1.4

3.4

5.2

0.001

>50%

57.6

61.4

56.2

53.2

40-50%

22.6

22.0

21.7

25.0

30-40%

12.1

11.1

12.5

12.4

7.7

5.4

9.6

9.4

Aspirin (% yes)

58.4

58.0

57.8

59.6

P2Y12 inhibitor (% yes)

22.9

23.6

22.2

21.9

0.616

RAAS inhibitor (% yes)

50.6

46.6

56.8

48.0

0.204 0.082

Cardiovascular history

LVEF (%)

<30%

0.207

Medication 0.603

Beta Blocker (% yes)

55.1

54.1

58.4

51.6

Statin (% yes)

61.6

60.8

64.1

58.6

0.134

Diuretic (% yes)

29.1

26.1

34.0

25.9

0.087

56.5

57.0

60.3

48.9

Angiography Indication (%) Stable CAD Unstable angina

0.001 9.6

8.5

10.0

10.7

27.2

28.1

22.4

33.6

6.8

6.5

7.3

6.8

No or minor CAD

24.7

22.9

27.0

25.0

Single vessel disease

32.8

35.4

30.3

32.1

Double vessel disease

27.7

27.0

27.1

29.2

Triple vessel disease

14.8

14.7

15.7

13.7

Conservative

33.1

31.7

35.2

31.8

PCI

61.3

62.7

59.3

61.7

5.7

5.5

5.5

6.6

6.7 ±1.5

7.7 ±0.7

5.3 ±1.1

-

1.7 [0.9, 2.6]

1.9 [1.0, 2.6]

1.6 [1.0, 2.4]

1.7 [0.8, 2.7]

Myocardial infarction Other CAD severity (%)

0.756

CAD treatment (%)

CABG EuroQoL score (mean ± SD) FU Duration (years, median [IQR])

0.559

0.496

Continuous values are displayed as means ± standard deviations (SD) and compared between responders and non-responders using a t-test. Categorical variables are presented as percentages per group and compared between responders and nonresponders using a chi-square test. Follow-up time is presented in medians and interquartile ranges. Abbreviations: BMI= body mass index, ACS= acute coronary syndrome, PCI= percutaneous coronary intervention, CABG= coronary artery bypass grafting, CVA= cerebrovascular accident, TIA= transient ischemic attack, PAD= peripheral arterial disease, LVEF= left ventricular ejection fraction, RAAS= renin-angiotensin-aldosterone system, FU duration=follow-up duration.

236


Non-response to Questionnaires Predicts Mortality

References 1.

Lenzen MJ, Scholte op Reimer WJM, Pedersen SS, Boersma E, Maier W, Widimsky P, Simoons ML, Mercado NF, Wijns W. The additional value of patient-reported health status in predicting 1-year mortality after invasive coronary procedures: a report from the Euro Heart Survey on Coronary Revascularisation. Heart. 2007;93:339–44.

2.

EuroQoL group. EuroQol- a new facility for the measurement of health-related quality of life. Health Policy. 1990;16:220–33.

3.

Norris CM, Spertus J a, Jensen L, Johnson J, Hegadoren KM, Ghali W a. Sex and gender discrepancies in health-related quality of life outcomes among patients with established coronary artery disease. Circ Cardiovasc Qual Outcomes. 2008;1:123–30.

4.

Primatesta P, Allender S, Ciccarelli P, Doring A, Graff-Iversen S, Holub J, Panico S, Trichopoulou A, Verschuren WMM. Cardiovascular surveys: manual of operations. Eur J Cardiovasc Prev Rehabil. 2007;14 Suppl 3:S43–S61.

5.

Gijsberts CM, Gohar A, Ellenbroek GHJM, Hoefer IE, de Kleijn DPV, Asselbergs FW, Nathoe HM, Agostoni P, Rittersma SZH, Pasterkamp G, Appelman Y, den Ruijter HM. Severity of stable coronary artery disease and its biomarkers differ between men and women undergoing angiography. Atherosclerosis. 2015;241:234–240.

6.

Zee K Van de, Sanderman R. Het meten van de algemene gezondheidstoestand met de RAND-36: een handleiding. Rijksuniversiteit, Groningen)(ISBN 90–72156–60–9). 1993;28.

7.

Christensen AI, Ekholm O, Gray L, Glümer C, Juel K. What is wrong with non-respondents? Alcohol-, drugand smoking-related mortality and morbidity in a 12-year follow-up study of respondents and nonrespondents in the Danish Health and Morbidity Survey. Addiction. 2015;Epub.

8.

Batty GD, Gale CR. Impact of resurvey non-response on the associations between baseline risk factors and cardiovascular disease mortality: prospective cohort study. J Epidemiol Community Heal. 2009;63:952–955.

237



PART THREE General Risk Prediction

Chapter 13 Routinely Analyzed Leukocyte Characteristics Improve Prediction of Mortality After Coronary Angiography Under revision at European Journal of Preventive Cardiology

Crystel M. Gijsberts, Guilielmus H.J.M. Ellenbroek, Maarten J. ten Berg, Albert Huisman, Wouter W. van Solinge, Folkert W. Asselbergs, Hester M. den Ruijter, Gerard Pasterkamp, Dominique P.V. de Kleijn, Imo E. Hoefer


Chapter 13

Abstract Background Inflammation and leukocyte infiltration are hallmarks of atherosclerosis. Clinical routine hematology analyzers mostly perform an entire differential blood count by default irrespective of the requested parameter. We hypothesize that these normally unreported leukocyte characteristics associate with coronary artery disease (CAD) severity and can improve prediction of mortality in coronary angiography patients. Methods We studied coronary angiography patients suspected of CAD (n=1,015) from the UCORBIO cohort. Leukocyte characteristics were routinely assessed in blood drawn directly prior to angiography using an automated hematology analyzer and extracted from the UPOD database. Patients were followed-up for a median duration of 805 days, during which 65 patients deceased. We evaluated the association of leukocyte characteristics with SYNTAX score as measure of CAD severity, all-cause and cardiovascular mortality and major adverse cardiovascular events (MACE). In order to determine the improvement of risk prediction, we calculated continuous net reclassification improvement (cNRIs) and integrated discrimination improvement (IDIs). Results Monocyte % showed strong independent predictive value for all-cause mortality (HR 1.44 (1.19-1.74), p<0.001) and monocyte-to-lymphocyte ratio (MLR) performed best for cardiovascular mortality (HR 1.42 (1.11-1.81), p=0.005). The cNRI and IDIs of leukocyte characteristics for all-cause mortality confirmed improvement of mortality risk prediction. No significantly predictive leukocyte characteristics were found for MACE. Conclusion Readily available, yet unreported leukocyte characteristics from routine hematology analyzers, significantly improved prediction of mortality in coronary angiography patients on top of clinical characteristics.

240


Leukocyte Characteristics Predict Mortality

Introduction Inflammation is a key feature of coronary atherosclerosis.1 Several circulating inflammatory cell characteristics such as white blood cell counts3,4, neutrophil counts5,6, monocyte counts7 and lymphocyte percentages7 have been reported to associate with coronary artery disease (CAD) and/or mortality. Recently, the neutrophil-to-lymphocyte ratio (NLR), was introduced as a mortality predictor in cardiovascular disease8,9 and in oncologic patients10. Despite the large number of CAD patients, there are no specific tools to accurately predict survival. While such information is crucial for the treating physician for adjusting treatment and advising patients in an evidence-based manner. Leukocyte activation may be reflected by leukocyte counts11, percentages and leukocyte subtype morphology, which are routinely measured by modern hematology analyzers. Irrespective of the requested output, these analyzers perform an entire differential blood count using built-in classification algorithms based on characteristic light refraction and cell size. Thus, much more information can be made available and might serve as prognostic biomarkers.12 In this study, we evaluated the association of leukocyte characteristics with CAD severity and assessed their predictive value on top of patient characteristics and cardiovascular risk factors. We hypothesized that leukocyte characteristics improve mortality risk prediction and reclassification13 of coronary angiography patients.

Methods Study population We performed this study in the Utrecht Coronary Biobank (UCORBIO)14, a longitudinal biobanking study of coronary angiography patients (see online Supplemental Methods for details on study population, data collection, follow-up and statistical analysis). Leukocyte characteristics Parameters that were used in this study comprised 15 UPOD leukocyte parameters: leukocyte, neutrophil, monocyte and lymphocyte counts and percentages, neutrophil cell size (mean and coefficient of variation (CV)) and complexity (mean and CV) and lymphocyte cell size (mean and CV) and complexity (mean and CV). These numbers are derived from the Abbott Cell-Dyn15 machine which uses multi angle polarized scatter separation to classify cell properties. By shining an Argon laser on individual cells, cell size is directly correlated with the axial light loss at 0째 and cell complexity is associated with the intermediate angle forward scatter at 7째. Cell complexity refers to the cytoplasmatic granularity of leukocytes. The more complex the granules in the cytoplasm, the higher the intermediate angle scatter. Cell activation leads to changes in granule content and hence intermediate angle scatter. Simplified, granularity reflects cell activation.16

241


Chapter 13

Three leukocyte ratios were calculated: NLR9,10, monocyte-to-lymphocyte ratio (MLR)12,17,18 and lymphocyte-to-monocyte ratio (LMR)19,20, as these ratios had been previously reported in the context of (cardiovascular) mortality prediction.

Results Patient characteristics (n=1,015) are presented in table 1, stratified by outcome. Deceased patients were older, more frequently diabetic and more often had a history of previous cardiovascular events. The indication for angiography and the chosen treatment of CAD following angiography did not differ between alive and deceased patients. Remarkably, the highly detailed SYNTAX score quantifying CAD severity was higher in deceased patients (15.9 vs. 9.8, p=0.001), while categorical CAD severity did not significantly differ. Medication use at enrollment was fairly equal, although ACE-inhibitor and diuretic use were more common among deceased patients. hsCRP levels were significantly higher in the deceased patient group. Table 2 shows leukocyte characteristics at baseline, stratified by outcome. We found higher neutrophil %, lower lymphocyte count and %, higher monocyte count and %, higher neutrophil complexity CV, higher lymphocyte cell size and CV, higher NLR, higher MLR and lower LMR among deceased patients. Leukocyte characteristics and SYNTAX score We found significant univariable associations of NLR (β 1.18, p=0.010 i.e. a 1SD increase in NLR gives a 1.18 point rise in SYNTAX score), neutrophil % (β 1.15, p=0.011), lymphocyte % (β 0.78, p=0.013), neutrophil count (β 1.13, p=0.024) and MLR (β 1.12, p=0.042) with SYNTAX score (supplemental figure 1). In multivariable analysis, these associations were abolished (supplemental table 1). Mortality and Adverse Events During a median follow-up time of 805 days (2207 person years), 65 patients died, 29 of these from a cardiovascular cause. The majority of non-cardiovascular deaths were due to malignancies (20/36). MACE occurred in 154 cases. We multivariably assessed the predictive value of leukocyte characteristics that were significantly associated with mortality in a univariable analysis. Cox regression analysis was performed for the entire cohort (n=1,015) for all-cause mortality, cardiovascular mortality and MACE and specifically for the patients in whom a SYNTAX score could be measured (n=603, 27 deaths and 96 MACEs during 1,081 person years) for all-cause mortality and MACE. The results from these five analyses are shown in table 3 (analysis of cardiovascular deaths among SYNTAX patients was not sensible due to a low number of events, n=12). All leukocyte characteristics were transformed to multitudes of the SD, resulting in HRs representing the increase in mortality risk for each 1 SD increase in leukocyte characteristic.

242


Leukocyte Characteristics Predict Mortality

Table 1. Baseline characteristics of coronary angiography patients, stratified by outcome (all-cause mortality). Alive

Deceased

950

65

64.6 (11.2)

72.7 (10.4)

72.9

75.4

0.776

27.0 (4.3)

26.5 (4.1)

0.365

Diabetes (%)

22.3

35.4

0.024

Hypertension (%)

55.9

61.5

0.448

Hypercholesterolemia (%)

47.6

43.1

0.565

Smoking (ever %)

55.0

54.5

1.000

1.4 [0.6, 3.4]

3.6 [1.3, 8.4]

<0.001

ACS (%)

31.3

43.1

0.066

PCI (%)

31.4

27.7

0.631

9.7

20.0

0.015

N Age (mean (sd)) Sex (males %) BMI (mean (sd))

High-sensitivity CRP (Îźg/ml)

p-value <0.001

History

CABG (%) CVA/TIA (%)

12.7

23.1

0.029

PAD (%)

10.4

32.3

<0.001

2.5

13.8

<0.001

Indication (stable %)

66.0

70.8

0.447

UA/NSTEMI (%)

20.9

23.1

STEMI (%)

9.2

4.6

Other (%)

3.9

1.5

26.3

21.5

Single vessel disease

37.6

32.3

Double vessel disease

25.1

24.6

Triple vessel disease

11.0

21.5

35.9

33.8

57.9

60.0

6.1

6.2

9.8 [5.0, 15.5]

15.9 [9.0, 23.2]

0.001 1.000

Kidney failure (%) Angiography

CAD severity (no/minor CAD %)

Procedure (conservative %) PCI (%) CABG (%) SYNTAX score*

0.080

0.943

Medication Aspirin (%)

63.3

63.1

Clopidogrel (%)

23.5

20.0

0.621

Statin (%)

64.9

73.8

0.185 0.464

Beta blocker (%)

59.2

64.6

ACE-inhibitor (%)

36.2

50.8

0.027

ARB (%)

15.7

9.2

0.222 <0.001

Diuretics (%) Follow-up time (IQR days)* Follow-up time (min, max)

27.9

61.5

825 [670, 980]

331 [197, 556]

544, 1094

5, 904

Baseline characteristics of coronary angiography patients, stratified by outcome (all-cause mortality). P-values for comparison between the two groups. *Non-normally distributed measures, presented in medians with interquartile ranges. A non-parametric test was performed for comparison between the groups for SYNTAX score and hsCRP.

243


Chapter 13

Table 2. Leukocyte characteristics, stratified by outcome.

N Total leukocyte count* Neutrophil count* Neutrophil % Lymphocyte count* Lymphocyte % Monocyte count* Monocyte % Mean neutrophil cell size Mean neutrophil complexity

Alive

Deceased

950

65

Reference range

P-value

7.11 [5.89, 8.95]

7.01 [5.94, 8.77]

4.0 – 10.0

0.913

4.40 [3.46, 5.80]

4.68 [3.80, 5.72]

1.6 – 8.3

0.402

63.27 [56.29, 69.75]

65.32 [60.37, 72.19]

1.75 [1.37, 2.22]

1.37 [1.07, 1.74]

25.23 [19.30, 31.46]

20.47 [17.00, 25.48]

0.58 [0.46, 0.74]

0.69 [0.59, 0.82]

8.00 [6.71, 9.71]

0.047 0.8 – 4.0

<0.001 <0.001

0.2 – 0.8

0.001

9.40 [6.88, 11.45]

0.001

147.11 [142.49, 151.69] 147.78 [144.69, 151.59]

0.438

134.39 [130.29, 137.80] 134.95 [132.34, 138.62]

0.269

Neutrophil cell size CV

2.47 [2.26, 2.76]

2.47 [2.24, 2.79]

0.880

Neutrophil complexity CV

3.46 [3.18, 3.75]

3.58 [3.20, 3.98]

0.034

Mean lymphocyte cell size

100.21 [97.50, 102.66]

101.47 [98.23, 104.74]

0.018

Mean lymphocyte complexity

75.32 [73.92, 76.85]

76.03 [74.42, 77.31]

0.103

Lymphocyte cell size CV

4.89 [3.93, 5.87]

4.37 [3.51, 5.18]

0.001

Lymphocyte complexity CV

4.82 [4.22, 5.48]

4.58 [4.01, 5.14]

0.091

NLR

2.50 [1.79, 3.62]

3.17 [2.30, 4.14]

0.001

MLR

0.32 [0.25, 0.43]

0.44 [0.32, 0.65]

<0.001

LMR

3.11 [2.33, 4.03]

2.26 [1.53, 3.12]

<0.001

All values are presented as medians with interquartile ranges. *The cell counts are reported as the number of cells (x 109 cells/L), reference ranges are only available for routinely reported results. NLR: neutrophil-tolymphocyte ratio, MLR: monocyte-to-lymphocyte ratio, LMR: lymphocyte-to-monocyte ratio, CV: coefficient of variation, % percentage of total leukocyte count. P-values are given for difference between dead and alive patients.

For the entire cohort, higher monocyte % was independently and significantly associated with a higher risk of mortality (HR 1.44 [1.19-1.74], p<0.001), as was the MLR (HR 1.35 [1.14-1.59], p<0.001), lymphocyte cell size CV (HR 0.69 [0.53-0.91], p=0.007) and monocyte count (HR 1.39 [1.09-1.77], p=0.008). The significant and independent predictors of cardiovascular mortality were MLR (HR 1.42 [1.11-1.81], p=0.005), monocyte % (HR 1.36 [1.04-1.77], p=0.025) and monocyte count (HR 1.49 [1.01-2.22], p=0.047). No significant leukocyte predictors were found for MACE in the multivariable model. In patients in whom a SYNTAX score could be measured, neutrophil complexity CV (HR 1.67 [1.12-2.47], p=0.011) and lymphocyte cell size CV (HR 0.63 [0.40-0.99], p=0.047) were significantly associated with all-cause mortality. Again, no significant leukocyte predictors were found for the prediction of MACE in these patients. Figure 1 visualizes the discriminating properties of the best predictors in the entire cohort for all-cause mortality, cardiovascular mortality, and for all-cause mortality among patients with a measurable SYNTAX score.

244


Leukocyte Characteristics Predict Mortality

Monocyte % > 8.1

200

400

600

800

0.98 0.96

MLR > 0.32

1000

MLR ≤ 0.32 0

400

600

Cardiovascular death

Cardiovascular death

800

1000

800

1000

800

1000

0.995 0.990 0.985

0.985

0.990

0.995

1.000

FU time (days)

HR (CI): 1.42 (1.11 - 1.81)

HR (CI): 1.36 (1.04 - 1.77)

MLR > 0.32

Monocyte % > 8.1

200

Monocyte % ≤ 8.1

0.980

0.980

MLR ≤ 0.32 400

600

800

1000

0

200

400

600

FU time (days)

SYNTAX patients: all-cause death

SYNTAX patients: all-cause death

HR (CI): 1.67 (1.12-2.47) Neutrophil complexity CV > 3.47

0

200

400

0.99 0.98 HR (CI): 0.63 (0.40-0.99) Lymphocyte cell size CV > 4.86 Lymphocyte cell size CV ≤ 4.86

0.96

0.96

Neutrophil complexity CV ≤ 3.47

0.97

0.97

0.98

0.99

Survival probability (Cox proportional hazard)

1.00

FU time (days)

1.00

0

Survival probability (Cox proportional hazard)

200

FU time (days)

1.000

0

HR (CI): 1.35 (1.14 - 1.59)

0.92

Monocyte % ≤ 8.1

0.94

Survival probability (Cox proportional hazard)

0.98 0.96 0.94

HR (CI): 1.44 (1.19 - 1.74)

0.92

Survival probability (Cox proportional hazard)

1.00

All-cause death

1.00

All-cause death

600

FU time (days)

800

1000

0

200

400

600

FU time (days)

Figure 1. Cox proportional hazards plots of the best performing leukocyte characteristics. Survival curves are presented for leukocyte characteristic values above and below or equal to their median value. For each outcome measure the two best performing leukocyte characteristics are shown. The HRs printed in the plots represent the risk for a 1-SD increase of the leukocyte characteristic and are derived from the multivariable model.

245


246

Monocyte count

Neutrophil complexity CV

Lymphocyte cell size CV

mortality

SYNTAX patients

All-cause mortality

0.63 [0.40-0.99]

1.67 [1.12-2.47]

1.49 [1.01-2.22]

1.36 [1.04-1.77]

1.42 [1.11-1.81]

72.1

74.2

77.4

78.3

80.3

80.0

80.9

83.3

85.8

LLH ratio

0.01 [-0.01-0.06] 0.02 [-0.01-0.09]

0.047

0.01 [-0.01-0.09]

0.047 0.011

0.01 [-0.02-0.08]

0.02 [0.00-0.07]

0.008

0.025

0.01 [0.00-0.04]

0.007

0.02 [-0.01-0.08]

0.02 [0.00-0.05]

0.005

0.01 [0.00-0.07]

<0.001

2-year IDI (95% CI)

<0.001

p-value

0.213

0.365

0.179

0.385

0.086

0.013

0.080

0.033

0.066

p-value IDI

0.20 [-0.31-0.42]

0.25 [-0.12-0.48]

0.31 [-0.01-0.49]

0.28 [-0.16-0.50]

0.32 [-0.11-0.48]

0.26 [0.10-0.38]

0.18 [0.00-0.28]

0.19 [0.00-0.33]

0.21 [-0.04-0.34]

2-year cNRI (95% CI)

0.326

0.133

0.053

0.140

0.120

<0.001

0.040

0.047

0.086

p-value cNRI

The hazard ratios are presented for a 1-SD increase in the leukocyte characteristic level. The leukocyte characteristics are sorted by LLH (log-likelihood ratio), the larger LLH, the better the model can predict outcome. Covariates are as described in the statistical methods section. For MACE no significant leukocyte predictors were found.

Monocyte %

Cardiovascular

1.39 [1.09-1.77]

Monocyte count

MLR

0.69 [0.53-0.91]

Lymphocyte cell size CV

Entire cohort

1.35 [1.14-1.59]

MLR

All-cause mortality

1.44 [1.19-1.74]

Monocyte %

Entire cohort

HR (95% CI)

Leukocyte characteristics

Table 3. Hazard ratios, IDIs and cNRIs of significant leukocyte characteristics from multivariable Cox regression.

Model

Chapter 13


Leukocyte Characteristics Predict Mortality

Reclassification and Discrimination Predicted 2-year risks were calculated from the basic Cox regression model and the model extended with a leukocyte characteristic. The reclassification improvement by addition of a leukocyte characteristic is described by the cNRI and IDI, which are shown in Table 3. Supplemental Figure 2 shows the predicted 2-year all-cause mortality risk with and without MLR and with and without monocyte count (these leukocyte characteristics had significant IDIs and cNRIs). IDIs and cNRIs of monocyte count and MLR were significant for all-cause mortality, as was the cNRI of lymphocyte cell size CV. Among the deceased patients, the average predicted risk increased with 0.02 (from 0.18 to 0.20) when monocyte count was added and a similar increase was found for MLR addition, while the risk for people who were alive at 2-year follow-up (0.05) did not change. The addition of monocyte count and MLR correctly reclassified 26 and 19% of the patients (corresponding to cNRIs of 0.26 and 0.19). Monocyte count and MLR IDIs (0.01 and 0.021, respectively) and cNRIs (0.0.31 and 0.32, respectively) were in the same order of magnitude, but did not reach statistical significance for cardiovascular mortality.

Discussion In the current study, we tested the predictive value of leukocyte characteristics for mortality in coronary angiography patients. Monocyte count and MLR showed promising predictive properties. These leukocyte characteristics are readily available in most clinical laboratories; herewith they appear to be independent, potent and practically feasible biomarkers for all-cause and cardiovascular mortality in coronary angiography patients. Monocytes Monocytes are involved in plaque build-up, plaque rupture and ischemia-reperfusion injury.21 In our study, monocyte counts strongly associated with SYNTAX score, concurring with the role of monocytes in the atheroma accumulation in the vascular wall. Previously, Chapman et al.22 demonstrated that monocyte count correlated better with carotid intima-media thickness than CRP or interleukin-6, indicating that monocyte counts reflect more than generalized low-grade inflammation. Monocyte counts have been shown to predict mortality in hemodialysis patients7 and elective coronary angiography patients12, whereas no predictive value of monocyte counts was found in a healthy elderly population.23 Notably, Rogacev et al.24 demonstrated that a specific monocyte subset (CD14++ CD16+, but not other subsets) was predictive for cardiovascular events in patients undergoing elective coronary angiography. The same monocyte subset has been shown to be related to higher levels of lipoprotein (a)25 suggesting that only certain subsets of monocytes are involved in atherosclerosis progression and thus more useful for prediction.

247


Chapter 13

Lymphocytes The role of lymphocytes in atherosclerosis is less defined. In a healthy elderly population23 no predictive value was found, while some predictive value was found among coronary angiography patients12, rendering its role as a risk predictor inconclusive. T-cells form the largest part of lymphocytes. T-cell subsets with different roles in atherosclerosis have been identified.26 Potentially, the proportion of the anti-atherogenic regulatory T-cells (and possibly Th17) is diminished in diseased individuals (with consequently higher MLRs), providing a biological explanation for the strong relation of MLR with SYNTAX score and outcome. Unfortunately, we do not have data to further investigate this theory, however, it might provide a focus for future research. Leukocyte ratios In murine models, it has been shown that leukocyte cell differentiation changes upon ischemia11 and recently, leukocyte parameters have gained attention for cardiovascular studies. So far, leukocyte ratios have predominantly been reported to be predictive in oncologic20 and immunologic patients17. Within the cardiovascular field, the LMR has been shown to be associated with critical limb ischemia in PAD patients.19 Yue et al.27 recently reported MLR to be predictive of diabetic retinopathy and a report from the LUdwigshafen RIsk and Cardiovascular Health (LURIC) study group12, also demonstrated a predictive value of MLR for the prediction of cardiovascular mortality among stable and unstable (but not myocardial infarction) patients. However, no NRIs or IDIs were calculated for MLR. The NLR shows to be of limited predictive value in our cohort, while it has been reported to be predictive in the recent literature.28,29 This incongruence might be due to differences in the studied patient populations, as study populations range from individuals without any cardiovascular disease29 to STEMI patients28. Another explanation could be differing levels of adjustment for confounding among the reports. Leukocytes and SYNTAX score In our cohort, monocyte % and MLR were the most accurate predictors of all-cause and cardiovascular mortality. However, when adding SYNTAX score to the model to correct for CAD severity, their predictive properties were abolished. There are 2 potential explanations for this observation: 1) As we only analyzed patients in whom a SYNTAX score was measurable (603 of 1015 patients) the power was reduced: only 27 deaths occurred among the SYNTAX patients. 2) The SYNTAX score’s own predictive power leaves only little room for an added value of additional parameters such as leukocyte characteristics. This is supported by our observation that monocyte % and MLR significantly correlate with SYNTAX score (supplemental table). When SYNTAX score was added to the Cox regression model, the cNRIs and IDIs of neutrophil complexity CV and MLR were not significant, suggesting that their contribution to reclassification and discrimination is limited when a SYNTAX score is available.

248


Leukocyte Characteristics Predict Mortality

Mortality vs. MACE Leukocyte characteristics appeared to be predictive of mortality but not of MACE. This might be due to a difference in covariates (Supplemental Methods) or to actual differences between the relation of leukocyte characteristics with non-fatal cardiovascular events and mortality. Clinical implications Many novel biomarkers require specific laboratory assays and/or an elaborate work-up of blood samples. In contrast, leukocyte properties are available from automated hematology analyzers in most clinical laboratories. Leukocyte properties are routinely measured in whole blood differentiation, but normally not reported. These overlooked, but widely available, biomarkers are therefore interesting and feasible candidates for improving risk prediction in CAD patients and perhaps beyond. Improvement of risk estimation ameliorates the identification of high-risk patients who qualify for strict risk factor control30 and possibly even for novel therapies (e.g. proprotein convertase subtilisin/kexin (PCSK) type 9-inhibitors31). This might be particularly useful in selecting patients for clinical trials. Limitations The main limitation of this study is the limited number of (cardiovascular) deaths during follow-up. Although our prediction model fits well and generates hazard ratios with narrow confidence intervals (a sign of proper risk estimation), we realize power is limited. Therefore, we strongly recommend replication of our results. Limited power also restricted our possibilities to include covariates. In supplemental table 2 we present a very extensive Cox regression model, evaluating the predictive value of monocyte count in order to demonstrate that addition of more potentially relevant covariates did not affect our results. Arterial blood samples were used for the profiling of leukocyte parameters. While this is common practice for angiography patients (who already have arterial access through the arterial sheath), it is imaginable that venous samples provide slightly deviating results. However, the differences in blood counts between arterial and venous samples described in the literature are small32 and unlikely to significantly alter our findings. Conclusion Monocyte count and MLR are strongly and independently associated with all-cause and cardiovascular mortality in patients undergoing angiography and aids in the identification patients with a high mortality risk. The readily available data on leukocyte characteristics, in clinical laboratories may thus serve as clinically applicable biomarkers for risk prediction. Acknowledgements We gratefully acknowledge the significant contribution of Ms. Jonne Hos to the logistical support of the UCORBIO cohort.

249


Chapter 13

References 1. 2.

Hansson GK. Inflammation, atherosclerosis, and coronary artery disease. N Engl J Med 2005; 352: 1685–95. Buffon A, Biasucci LM, Liuzzo G, et al. Widespread coronary inflammation in unstable angina. N Engl J Med 2002; 347: 5–12.

3.

Madjid M, Awan I, Willerson JT, et al. Leukocyte count and coronary heart disease: implications for risk assessment. J Am Coll Cardiol 2004; 44: 1945–56.

4.

Lee C Do, Folsom AR, Nieto FJ, et al. White blood cell count and incidence of coronary heart disease and ischemic stroke and mortality from cardiovascular disease in African-American and White men and women: atherosclerosis risk in communities study. Am J Epidemiol 2001; 154: 758–64.

5.

Guasti L, Dentali F, Castiglioni L, et al. Neutrophils and clinical outcomes in patients with acute coronary syndromes and/or cardiac revascularization: A systematic review on more than 34,000 subjects. Thromb Haemost 2011; 106: 591–599.

6.

Gillum RF, Mussolino ME, Madans JH. Counts of neutrophils, lymphocytes, and monocytes, cause-specific mortality and coronary heart disease: the NHANES-I epidemiologic follow-up study. Ann Epidemiol 2005; 15: 266–71.

7.

Kato A, Takita T, Furuhashi M, et al. Blood monocyte count is a predictor of total and cardiovascular mortality in hemodialysis patients. Nephron Clin Pract 2008; 110: c235–43.

8.

Abe T, Kato S, Tsuruta Y, et al. Neutrophil/lymphocyte ratio as a predictor of cardiovascular events in incident dialysis patients: a Japanese prospective cohort study. Clin Exp Nephrol 2014.

9.

He J, Li J, Wang Y, et al. Neutrophil-to-lymphocyte ratio (NLR) predicts mortality and adverse-outcomes after ST-segment elevation myocardial infarction in Chinese people. Int J Clin Exp Pathol 2014; 7: 4045–56.

10. Inoue D, Ozaka M, Matsuyama M, et al. Prognostic value of neutrophil-lymphocyte ratio and level of C-reactive protein in a large cohort of pancreatic cancer patients: a retrospective study in a single institute in Japan. Jpn J Clin Oncol 2014; 1–6. 11. Haverslag R, De Groot D, Van Den Borne P, et al. Arterial occlusion induces systemic changes in leucocyte composition. Eur J Clin Invest 2011; 41: 943–950. 12. ó Hartaigh B, Bosch J a., Thomas GN, et al. Which leukocyte subsets predict cardiovascular mortality? From the LUdwigshafen RIsk and Cardiovascular Health (LURIC) Study. Atherosclerosis 2012; 224: 161–169. 13. Backer GD, Graham I, Cooney M-T. Do novel biomarkers add to existing scores of total cardiovascular risk? European Journal of Preventive Cardiology 2012; 19: 14–17. 14. Gijsberts CM, Santema BT, Asselbergs FW, et al. Women Undergoing Coronary Angiography for Myocardial Infarction or Who Present With Multivessel Disease Have a Poorer Prognosis Than Men. Angiology 2015. 15. Müller R, Mellors I, Johannessen B, et al. European multi-center evaluation of the Abbott Cell-Dyn sapphire hematology analyzer. Lab Hematol 2006; 12: 15–31. 16. Lam SW, Leenen LPH, van Solinge WW, et al. Comparison between the prognostic value of the white blood cell differential count and morphological parameters of neutrophils and lymphocytes in severely injured patients for 7-day in-hospital mortality. Biomarkers 2012; 17: 642–647. 17. Berens-Riha N, Kroidl I, Schunk M, et al. Evidence for significant influence of host immunity on changes in differential blood count during malaria. Malar J 2014; 13: 155.

250


Leukocyte Characteristics Predict Mortality

18. Naranbhai V, Hill AVS, Abdool Karim SS, et al. Ratio of monocytes to lymphocytes in peripheral blood identifies adults at risk of incident tuberculosis among HIV-infected adults initiating antiretroviral therapy. J Infect Dis 2014; 209: 500–9. 19. Gary T, Pichler M, Belaj K, et al. Lymphocyte-to-monocyte ratio: a novel marker for critical limb ischemia in PAOD patients. Int J Clin Pract 2014; 68: 1483–7. 20. Stotz M, Szkandera J, Stojakovic T, et al. The lymphocyte to monocyte ratio in peripheral blood represents a novel prognostic marker in patients with pancreatic cancer. Clin Chem Lab Med 2014; 1–8. 21. Ghattas A, Griffiths HR, Devitt A, et al. Monocytes in coronary artery disease and atherosclerosis: where are we now? J Am Coll Cardiol 2013; 62: 1541–51. 22. Chapman CML, Beilby JP, McQuillan BM, et al. Monocyte count, but not C-reactive protein or interleukin-6, is an independent risk marker for subclinical carotid atherosclerosis. Stroke 2004; 35: 1619–24. 23. Karino S, Willcox BJ, Fong K, et al. Total and differential white blood cell counts predict eight-year incident coronary heart disease in elderly Japanese-American men: The Honolulu Heart Program. Atherosclerosis 2014; 238: 153–158. 24. Rogacev KS, Cremers B, Zawada AM, et al. CD14++CD16+ monocytes independently predict cardiovascular events: a cohort study of 951 patients referred for elective coronary angiography. J Am Coll Cardiol 2012; 60: 1512–20. 25. Krychtiuk K a., Kastl SP, Hofbauer SL, et al. Monocyte subset distribution in patients with stable atherosclerosis and elevated levels of lipoprotein(a). J Clin Lipidol 2015; 9: 533–541. 26. Profumo E, Buttari B, Saso L, et al. T lymphocyte autoreactivity in inflammatory mechanisms regulating atherosclerosis. ScientificWorldJournal 2012; 2012: 157534. 27. Yue S, Zhang J, Wu J, et al. Use of the Monocyte-to-Lymphocyte Ratio to Predict Diabetic Retinopathy. Int J Environ Res Public Health 2015; 12: 10009–10019. 28. Arbel Y, Shacham Y, Ziv-Baran T, et al. Higher Neutrophil/Lymphocyte Ratio Is Related to Lower Ejection Fraction and Higher Long-term All-Cause Mortality in ST-Elevation Myocardial Infarction Patients. Can J Cardiol 2014; 30: 1177–82. 29. Shah N, Parikh V, Patel N, et al. Neutrophil lymphocyte ratio significantly improves the Framingham risk score in prediction of coronary heart disease mortality: insights from the National Health and Nutrition Examination Survey-III. Int J Cardiol 2014; 171: 390–7. 30. Kotseva K, Wood D, De Bacquer D, et al. EUROASPIRE IV: A European Society of Cardiology survey on the lifestyle, risk factor and therapeutic management of coronary patients from 24 European countries. Eur J Prev Cardiol 2015. 31. Shimada YJ, Cannon CP. PCSK9 (Proprotein convertase subtilisin/kexin type 9) inhibitors: past, present, and the future. Eur Heart J 2015; Epub: 1–12. 32. Yang ZW, Yang SH, Chen L, et al. Comparison of blood counts in venous, fingertip and arterial blood and their measurement variation. Clin Lab Haematol 2001; 23: 155–159.

251


Chapter 13

Supplemental Methods Patient population All patients >18 years undergoing coronary angiography for any indication (stable angina, unstable angina, myocardial infarction or other indication) at the University Medical Centre in Utrecht (UMCU) in the Netherlands were eligible to participate in this cohort. Patients in the current analysis were enrolled between October 2010 and April 2013. The study has been approved by the medical ethics committee of the UMCU and all patients provided written informed consent. UCORBIO is registered under the clinicaltrials.gov ID: NCT02304744. The study conforms to the Declaration of Helsinki. Data collection At the moment of inclusion, an arterial blood sample was drawn from the arterial sheath inserted for coronary angiography before any procedure-related drugs were administered. Differential blood counts were performed according to routine clinical practice and all hematological parameters were subsequently stored in the UPOD1. From there, we collected an extraction of the Abbott Cell-Dyn2 data. Additionally, demographical data, history of acute coronary syndrome (ACS), history of percutaneous coronary intervention (PCI), history of coronary artery bypass grafting (CABG), cerebrovascular accident/transient ischemic attack (CVA/TIA) or peripheral arterial disease (PAD), medication use, cardiovascular risk factors: diabetes, body mass index (BMI), hypertension, hypercholesterolemia, smoking and the indication for catheterization, the angiographic severity of CAD and procedural details were collected at the moment of enrollment. CAD severity CAD severity was categorized into no/minor CAD, single vessel disease, double vessel disease or triple vessel disease (according to the number of epicardial vessels with a stenosis of >50%, as visually assessed by the interventional cardiologist). In patients with significant CAD (i.e. one or more vessels with >50% stenosis) and without history of coronary artery bypass grafting (CABG), a SYNTAX score was measured (n=603). This was performed by two independent observers with unlimited access to quantitative coronary angiography3 analysis. When scores differed >5 points between the two observers (n=131), cases were discussed to reach consensus. The mean of the SYNTAX scores of the two observers was used for analysis. Follow-up Patients are followed-up for five years, of which three years had passed at the moment of writing. The vital status (deceased or alive) of patients was collected from the hospital administration system, updated by research staff. The cause of death was obtained from medical reports and blindly adjudicated by the first author of this article. When the cause of death was inconclusive, a panel of three cardiologists was consulted. If a patient died

252


Leukocyte Characteristics Predict Mortality

at home, the general practitioner was consulted to obtain the cause of death. Every year, patients received a questionnaire to ask for hospital admissions and occurrence of major adverse cardiovascular events (MACE). When patients reported a potential event or when the patient did not return the questionnaire, the general practitioner or the said hospital was contacted for confirmation. In case of hospitalization, medical records were requested and a panel of cardiologists adjudicated the event. The composite end-point MACE was defined as: cardiovascular mortality, non-fatal myocardial infarction, non-fatal stroke, unplanned coronary revascularization: percutaneous coronary intervention (PCI) or coronary artery bypass grafting (CABG). Statistical analysis All statistical analyses were performed using R version 3.1.2 (R Foundation for Statistical Computing, Vienna, Austria).4 Baseline characteristics are presented by outcome (alive or deceased). Continuous data are presented as mean with standard deviation (SD) or median with interquartile ranges. Categorical data are presented as percentages. Differences between the outcome groups were compared with a t-test or Kruskal-Wallis test for continuous data or chi-square testing for categorical data. In order to present the leukocyte characteristics estimates in a more universal manner, we transformed the actual values to multitudes of the SD. The associations of leukocyte characteristics with log10 SYNTAX scores were analyzed by univariable and multivariable linear regression analyses. The presented betas were back-transformed to ease interpretation. The associations of leukocyte characteristics with outcome (all-cause mortality, cardiovascular mortality and MACE) were assessed using multivariable Cox regression analysis. We analyzed the additive value of the leukocyte characteristics to a basic Cox regression model in the entire cohort (complete cases n=1,015) and specifically in patients in whom a SYNTAX score was measurable (complete cases n=603: no history of CABG and significant CAD). The basic Cox model contained baseline characteristics that were selected by backward stepwise Cox regression analysis (p-value <0.1 for relation with outcome). The considered covariates were: age, sex, BMI, diabetes, hypertension, hypercholesterolemia, smoking, history of ACS, history of PCI, history of CVA, history of CABG, history of PAD, kidney failure, indication for coronary angiography, treatment of CAD, severity of CAD, high-sensitivity C-reactive protein (hsCRP), aspirin use, clopidogrel use, beta blocker use, statin use, angiotensin converting enzyme (ACE)-inhibitor use, diuretic use and angiotensin II receptor blocker (ARB) use. For all-cause mortality age, history of ACS, history of PCI, history of PAD, kidney failure and diuretic use were selected as relevant covariates. For cardiovascular mortality age, history of PCI, history of CABG, history of PAD, kidney failure, indication for coronary angiography and diuretic use were selected. Covariates for the analysis of MACE consisted of BMI, history of PCI, history of CVA, treatment of CAD, angiographic severity of CAD and diuretic use. In the Cox regression analysis among patients with a SYNTAX score the covariates for all-cause mortality were: SYNTAX score, age, BMI, diabetes, smoking, history of CVA and

253


Chapter 13

kidney failure and for MACE they consisted of: SYNTAX score, history of PCI, history of PAD, aspirin use, statin use and diuretic use. First, we assessed the additive predictive information of the leukocyte characteristics to the basic model and tested whether they significantly added information to the basic Cox model by comparing the log-likelihoods of the basic model and the extended model (basic model plus leukocyte characteristic) using analysis of deviance. We sorted the characteristics by the likelihood (LLH) ratios of the multivariable Cox models (highest LLH ratio corresponds with best model). Then, we calculated the continuous Net Reclassification Improvement (cNRI)5 and the Integrated Discrimination Improvement (IDI)6 for the characteristics that significantly added to the basic Cox model. The cNRI corresponds to the percentage of patients that is correctly reclassified by the addition of a leukocyte characteristic to the basic Cox model and is calculated by adding the percentage of patients with an event who were correctly classified as higher risk in the new model to the percentage of patients without event who were correctly classified as lower risk in the new model.5 The IDI corresponds to the absolute change in predicted risk between the old and the new model, calculated by subtracting the difference in predicted risk between patients with event and without event in the old model from the difference in predicted risk between patients with event and without event in the new model.6 The cNRIs and IDIs were computed using the R software package ‘survIDINRI’ version 1.1-1 as described by Uno et al.7 The level of significance was set at Îą<0.05 (twotailed) for all analyses.

254


Leukocyte Characteristics Predict Mortality

References 1.

Ten Berg MJ, Huisman A, van den Bemt PML a, Schobben AF a M, Egberts ACG, van Solinge WW. Linking laboratory and medication data: new opportunities for pharmacoepidemiological research. Clin Chem Lab Med. 2007;45:13–9.

2.

Müller R, Mellors I, Johannessen B, Aarsand AK, Kiefer P, Hardy J, Kendall R, Scott CS. European multi-center evaluation of the Abbott Cell-Dyn sapphire hematology analyzer. Lab Hematol. 2006;12:15–31.

3.

Généreux P, Palmerini T, Caixeta A, Cristea E, Mehran R, Sanchez R, Lazar D, Jankovic I, Corral MD, Dressler O, Fahy MP, Parise H, Lansky AJ, Stone GW. SYNTAX score reproducibility and variability between interventional cardiologists, core laboratory technicians, and quantitative coronary measurements. Circ Cardiovasc Interv. 2011;4:553–61.

4. 5.

R Core Team. R: A Language and Environment for Statistical Computing. 2013; Pencina MJ, D’Agostino RB, Steyerberg EW. Extensions of net reclassification improvement calculations to measure usefulness of new biomarkers. Stat Med. 2011;30:11–21.

6.

Steyerberg EW, Vickers AJ, Cook NR, Gerds T, Gonen M, Obuchowski N, Pencina MJ, Kattan MW. Assessing the Performance of Prediction Models. Epidemiology. 2010;21:128–138.

7.

Uno H, Tian L, Cai T, Kohane IS, Wei LJ. A unified inference procedure for a class of measures to assess improvement in risk prediction systems with survival data. Stat Med. 2013;32:2430–2442.

255


Chapter 13

Supplemental figure 1. Scatterplots of the association of leukocyte characteristics with log10 SYNTAX score, only the significant associations (p<0.05) are displayed. SYNTAX score was transformed to a logarithmic scale, because the data were positively skewed. The blue line is the result from univariable linear regression analysis with its corresponding confidence intervals, the p-value displayed in the plot is the corresponding p-value of the association found.

256


Leukocyte Characteristics Predict Mortality

1.00

Predicted 2−year risk of all−cause mortality among deceased patients

1.00

Predicted 2−year risk of all−cause mortality among alive patients

Magnification: Alive patients with low predicted risk 0.15

0.75

Predicted risk with Monocyte count

Predicted risk with Monocyte count

Predicted risk with Monocyte count

0.75

0.10

0.50

0.50

0.05

0.25

0.25

0.00

0.00 0.00

1.00

0.25

0.50

0.75

Predicted risk without Monocyte count

1.00

Predicted 2−year risk of all−cause mortality among deceased patients

0.00 0.00

1.00

0.25

0.50

0.75

Predicted risk without Monocyte count

0.00

0.05

0.10

Predicted risk without Monocyte count

0.15

1.00

Predicted 2−year risk of all−cause mortality among alive patients

Magnification: Alive patients with low predicted risk 0.15

0.75

Predicted risk with MLR

Predicted risk with MLR

Predicted risk with MLR

0.75

0.10

0.50

0.50

0.05

0.25

0.25

0.00

0.00 0.00

0.25

0.50

Predicted risk without MLR

0.75

1.00

0.00 0.00

0.25

0.50

Predicted risk without MLR

0.75

0.00

0.05

0.10

Predicted risk without MLR

0.15

1.00

! Supplemental figure 2. Predicted 2-year risk of all-cause mortality for the model without monocyte count/ Supplemental!Figure!2.!Predicted!24year!risk!of!all4cause!mortality!for!the!model!without!monocyte! MLR (x-axis) and the model with monocyte count/MLR (y-axis), stratified by outcome. Predicted 2-year all-cause mortality risk is displayed for each individual, derived from the clinical model without monocyte count/MLR and the clinicalcount/MLR!(x4axis)!and!the!model!with!monocyte!count/MLR!(y4axis),!stratified!by!outcome.!! model plus monocyte count/MLR. The left panels show the predicted risk in deceased patients. The middle panels show the predicted risks in patients who were alive at 2-year follow-up. The magnifications in the Predicted(2*year(all*cause(mortality(risk(is(displayed(for(each(individual,(derived(from(the(clinical( right panels show the justly lowered risks of alive patients at 2-year follow-up in the lowest risk range. Deceased patients who shifted upwards (above the red dashed line) were correctly reclassified to higher risk by adding model(without(monocyte(count/MLR(and(the(clinical(model(plus(monocyte(count/MLR.(The(left( monocyte count/MLR to the model, alive patients who moved downwards (below the red dashed line) were appropriately reclassified to lower risk. panels(show(the(predicted(risk(in(deceased(patients.(The(middle(panels(show(the(predicted(risks(in( patients(who(were(alive(at(2*year(follow*up.(The(magnifications(in(the(right(panels(show(the(justly( lowered(risks(of(alive(patients(at(2*year(follow*up(in(the(lowest(risk(range.(Deceased(patients(who( shifted(upwards((above(the(red(dashed(line)(were(correctly(reclassified(to(higher(risk(by(adding( monocyte(count/MLR(to(the(model,(alive(patients(who(moved(downwards((below(the(red(dashed( line)(were(appropriately(reclassified(to(lower(risk.(!

257


Chapter 13

Supplemental table 1. Estimates (betas) of the association of leukocyte characteristics with SYNTAX score. Leukocyte characteristic NLR

Beta (per SD increase)

Univariable p-value

Multivariable p-value

1.18

0.010

0.168

Neutrophil %

1.15

0.011

0.097

Lymphocyte %

0.87

0.013

0.103

Neutrophil count

1.13

0.024

0.741

MLR

1.12

0.042

0.196

1.11

0.056

Lymphocyte cell size CV

0.92

0.133

Monocyte count

1.07

0.220

Mean neutrophil complexity

1.07

0.225

Neutrophil cell size CV

1.06

0.338

Mean lymphocyte cell size

1.06

0.342

Mean neutrophil cell size

0.95

0.379

Monocyte %

0.95

0.392

LMR

0.96

0.520

Neutrophil complexity CV

1.03

0.584

Lymphocyte count

0.97

0.621

Mean lymphocyte complexity

1.02

0.690

Lymphocyte complexity CV

0.98

0.724

Leukocyte count

Betas (back-transformed) and p-values for univariable regression analysis of leukocyte characteristics with log10 SYNTAX score. The latter column shows p-values for the univariably significant characteristics, which were additionally tested in a multivariable model (adjusting for: indication for angiography, sex, diabetes, previous PCI, history of CVA/TIA and history of PAD).

258


Leukocyte Characteristics Predict Mortality

Supplemental table 2. Results from multivariable Cox regression model for all-cause mortality including all presented covariates (enter model). Parameter

HR (95% CI)

p-value

Diuretics (no vs. yes)

0.30 (0.17-0.53)

<0.001

Monocyte % (per 1-sd)

1.49 (1.22-1.82)

<0.001

0.28 (0.13-0.59)

0.001

Kidney failure (no vs. yes) Age (per 1 year)

1.04 (1.02-1.08)

0.003

History of PAD (no vs. yes)

0.42 (0.23-0.76)

0.004

History of CVA (no vs. yes)

0.42 (0.22-0.79)

0.008

History of ACS (no vs. yes)

0.59 (0.33-1.04)

0.068

Hypercholesterolemia (no vs. yes)

1.52 (0.86-2.71)

0.152

ACE-inhibitor (no vs. yes)

0.70 (0.41-1.20)

0.195

CAD severity (Single vessel disease vs. no CAD)

0.59 (0.22-1.59)

0.293

CAD severity (Double vessel disease vs. no CAD)

0.57 (0.19-1.75)

0.329

History of CABG (no vs. yes)

0.73 (0.37-1.42)

0.352

BMI (per 1 kg/m2)

0.97 (0.91-1.04)

0.384

Statin (no vs. yes)

0.74 (0.37-1.47)

0.386

Beta Blocker (no vs. yes)

1.30 (0.71-2.36)

0.393

Sex (female vs. male)

0.79 (0.41-1.51)

0.475

Hypertension (no vs. yes)

1.23 (0.69-2.19)

0.487

Treatment (PCI vs. conservative)

1.25 (0.52-2.97)

0.617

Diabetes (no vs. yes)

0.92 (0.50-1.67)

0.776

High-sensitivity C-reactive protein (per 1 Îźg/ml)

1.00 (0.99-1.02)

0.781

CAD severity (Triple vessel disease vs. no CAD)

0.92 (0.28-3.07)

0.894

Treatment (CABG vs. conservative)

0.95 (0.24-3.73)

0.936

The results from multivariable Cox regression analysis are sorted by significance of the covariates. Abbreviations: HR= hazard ratio, CI= confidence interval, sd= standard deviation, BMI= body mass index, PAD= peripheral arterial disease, CABG= coronary artery bypass grafting, ACS= acute coronary syndrome, CVA= cerebrovascular accident, CAD= coronary artery disease, PCI= percutaneous coronary intervention, ACE= angiotensinconverting enzyme.

259



PART THREE General Risk Prediction

Chapter 14 Hematological Parameters Improve Prediction of Mortality and Secondary Adverse Events in Coronary Angiography Patients: a Longitudinal Cohort Study Accepted for publication in Medicine

Crystel M. Gijsberts, Hester M. den Ruijter, Dominique P.V. de Kleijn, Albert Huisman, Maarten J. ten Berg, Richard H.A. van Wijk, Folkert W. Asselbergs, Michiel Voskuil, Gerard Pasterkamp, Wouter W. van Solinge, Imo E. Hoefer


Chapter 14

Abstract Background Prediction of primary cardiovascular events has been thoroughly investigated since the landmark Framingham risk score was introduced. However, prediction of secondary events after initial events of coronary artery disease (CAD) poses a new challenge. Methods In a cohort of coronary angiography patients (n=1,760) we examined readily available hematological parameters from the UPOD (Utrecht Patient Oriented Database) and their addition to prediction of secondary cardiovascular events. Backward stepwise multivariable Cox regression analysis was used to test their ability to predict death and major adverse cardiovascular events (MACE). Continuous net reclassification improvement (cNRI) and integrated discrimination improvement (IDI) measures were calculated for the hematological parameters on top of traditional risk factors to assess prediction improvement. Results Panels of three to eight hematological parameters significantly improved prediction of death and adverse events. The IDIs ranged from 0.02-0.07 (all p<0.001) among outcome measures and the cNRIs from 0.11-0.40 (p<0.001 in five of six outcome measures). In the hematological panels red cell distribution width (RDW) appeared most often. The multivariable adjusted hazard ratio of RDW per 1 standard deviation (SD) increase for MACE was 1.19 [1.08-1.32], p<0.001. Conclusion Routinely measured hematological parameters, significantly improved prediction of mortality and adverse events in coronary angiography patients. Accurately indicating high-risk patients is of paramount importance in clinical decision-making.

262


Hematological Parameters Improve Prediction of Adverse Events

Introduction Improvements in cardiovascular health care have significantly increased survival of coronary artery disease (CAD) patients.1 Consequently, the number of patients at risk for secondary events has risen. Despite being generally considered as high risk, this patient group is far from homogeneous; the risk of developing secondary adverse events varies from very low to very high. The prediction of primary events has been studied for over half a century now, starting with the introduction of the landmark Framingham risk score.2 In addition to a clinical prediction model, many biomarkers have been evaluated for their ability to improve primary and secondary prediction, for example C-reactive protein3,4, Cystatin C5,6 and myeloperoxidase7,8 have shown to be associated with the risk of future events. Also, hematological parameters, mainly leukocyte-related parameters,9,10 have been reported to reflect the risk of primary cardiovascular events. More recently, high red blood cell distribution width (RDW), a measure of the variation of red blood cell size has emerged as a predictor for atherosclerosis progression11, CAD severity12 and mortality13. However, the secondary predictive value of traditional risk factors, e.g. body mass index14 (BMI) and of biomarkers used in primary risk prediction is limited or remains unclear. Accurately predicting secondary risk is of paramount importance to the patient and their treating clinician in order to optimize secondary preventive measures for those at need. This may include novel and/or expensive therapies, e.g. the proprotein convertase subtilisin/kexin (PCSK) type 9-inhibitors15. To date, reliable tools that discriminate between high or low risk patients with known CAD are lacking. In the current study we therefore sought to improve secondary risk prediction among coronary angiography patients. We did this by extending a clinical model containing risk factors, cardiovascular history and angiographic characteristics with routinely measured and readily available hematological parameters. For this purpose we used the Utrecht CORonary BIObank (UCORBIO) cohort16 in combination with hematological measurements from the Utrecht Patient-Oriented Database (UPOD)17 laboratory registrations.

Methods Study population We analyzed data from the UCORBIO cohort (clinicaltrials.gov identifier: NCT02304744), an observational cohort study of patients undergoing coronary angiography for any indication in the University Medical Center in Utrecht, the Netherlands. From October 2011 to February 2014, a total of 1,904 patients were enrolled. For the current study, adult (>18 years) patients presenting with myocardial infarction (either ST-Segment Elevation Myocardial Infarction [STEMI] or Non-ST-Segment Elevation Myocardial Infarction [NSTEMI]), chest pain without release of cardiac enzymes (stable or unstable angina), dyspnea on exertion, silent ischemia or screening for non-cardiac surgery were selected (n=1,760). Patients with other indications for coronary angiography (coronary anomalies, screening for cardiac surgery or heart transplant follow-up) were thus excluded (n=144).

263


Chapter 14

Ethics, consent and permissions All patients provided written informed consent and the study conforms to the Declaration of Helsinki. The institutional review board of the University Medical Centre Utrecht approved of this study (reference number 11-183). Data collection The investigators completed standardized electronic case report forms at baseline based on the patient’s medical files containing age, sex, cardiovascular risk factors, indication for angiography, medication use, angiographic findings and eventual treatment of CAD. The definitions used for the baseline variables were published previously in more detail.18 The angiographic findings were categorized into four groups by the treating interventional cardiologist: no CAD, minor CAD (wall irregularities, <50% stenosis), single vessel disease (one vessel with >50% stenosis19) and multi-vessel disease (two or three vessels with >50% stenosis). Hematological parameters The hematological parameters were obtained through complete blood count analysis at the moment of coronary angiography. The parameters that were used in this study comprised 56 routinely measured hematological parameters (listed in supplemental figure 1) from the UPOD database17. A feature of the automated blood cell analyzer is that it not only reports the parameters requested by the physician, but all hematological parameters that it is capable of measuring. For example, when a physician requests a hemoglobin measurement, the analyzer also automatically determines the platelet count. Although this platelet count is not reported to the clinician, the analyzer stores the data. Periodically, all data captured within the blood cell analyzers are downloaded to a database format, and are cleaned and checked for integrity, making the data available for research. The UPOD parameters contain information on red blood cell (RBC) numbers and characteristics, leukocyte numbers and characteristics and platelet numbers and characteristics. All hematological parameters are measured using the Cell-Dyn Sapphire20–22 hematology analyzer (Abbott Diagnostics, Santa Clara, CA, USA). This analyzer is equipped with an integrated 488-nm blue diode laser and uses spectrophotometry, electrical impedance, laser light scattering (multi-angle polarized scatter separation), and three-color fluorescent technologies to measure morphological parameters of leukocytes, RBCs and platelets for classification and enumeration. The morphological parameters entail the following five optical scatter signals for leukocytes: cell size (0˚ scatter, axial light loss), cell complexity and granularity (7˚ scatter, intermediate angle scatter (IAS)), nuclear lobularity (90˚ scatter, polarized side scatter (PSS)), depolarization (90˚ depolarized side scatter (DSS)) and viability (red fluorescence (FL-3), 630 ± 30 nm). For platelets, two optical scatter signals are measured: IAS scatter (7˚, cell size) and PSS scatter (90˚, granularity; internal structure). RBC parameters are measured or calculated based on the impedance measurement. Reticulocytes are optically

264


Hematological Parameters Improve Prediction of Adverse Events

measured using IAS scatter (7˚, cell size) and FL-1 fluorescence (RNA content). Throughout this paper, all values of hematological parameters are reported as multitudes of their standard deviation (SD) in order to ensure comparability of effect sizes among parameters with absolute values that vary strongly in their order of magnitude. Follow-up On a yearly basis, patients received a questionnaire to check for hospital admissions and occurrence of major adverse cardiovascular events (MACE). When the patient reported a hospital admission suspect for MACE or did not complete or return the questionnaire, the general practitioner or reported hospital was contacted for confirmation. In the case of hospitalization or death, medical records were obtained and the relevance of the event or the cause of death was determined. A panel of cardiologists adjudicated the occurrence of events. The composite end-point MACE was defined as any of the following clinical events: all-cause death, non-fatal myocardial infarction, unplanned revascularization; both cardiac (percutaneous coronary intervention (PCI) and coronary artery bypass grafting (CABG)) and non-cardiac intervention, stroke and admission for heart failure. Statistical Analysis This study is reported in accordance with the STROBE guidelines for observational research.23 Baseline characteristics were reported as means and standard deviations for continuous variables and percentages for categorical variables, for the entire cohort and separately for patients who experienced MACE during follow-up and who did not. First, we constructed a clinical risk prediction model. Covariates for this model were selected using a boosting technique for Cox regression models (R package “CoxBoost”24). The covariates considered were: age, sex, diabetes, hypertension, hypercholesterolemia, BMI, smoking, indication for angiography, angiographic CAD severity, treatment following angiography, history of PCI, history of CABG, history of acute coronary syndrome (ACS), history of cerebrovascular accident (CVA), history of peripheral arterial disease (PAD), kidney failure, use of ACE-inhibitor, use of beta-blocker, use of statin, use of P2Y12inhibitor (clopidogrel, prasugrel or ticagrelor) and use of diuretic. Age, sex, indication for angiography, angiographic severity of CAD and treatment following angiography were considered mandatory covariates. The variables additionally selected using a boosting procedure were: diabetes, history of PCI, history of ACS, history of PAD, kidney failure and use of diuretics. The coefficients of the clinical model parameters were re-fit for each outcome measure; the clinical model performed well in all outcome measures (AUCs ranging from 0.6810.884). For the identification of hematological parameters that could aid prediction of adverse events (total n= 56), first we evaluated mutual correlation of the parameters by means of a hierarchically clustered heatmap (supplemental figure 1). From each cluster of collinear parameters the parameter that showed the strongest relation with MACE was selected for the further analysis.

265


Chapter 14

The remaining parameters (n= 37) were entered in six backward stepwise Cox regression models, one for each outcome measure: all-cause death, MACE, cardiovascular death, non-cardiovascular death, re-PCI and myocardial infarction. From this procedure, the top ten significant parameters for each outcome were added to the clinical parameters (which were forced to stay in the model, i.e. mandatory covariates) and again backward stepwise Cox regressions were performed for the hematological parameters, rendering the final panels of hematological parameters for the six outcome measures while keeping the clinical parameters stable. For the panel of hematological parameters that appeared to be significantly related to adverse events and that were thus added to the clinical model, areas under the curve (AUCs) were compared to the clinical model alone using receiver operating characteristics (ROC) analysis. The R package “timeROC”25 was used for this purpose, which is based on the methods described by Chiang et al26. Furthermore, according to the most recent epidemiological recommendations, continuous net reclassification improvement (cNRI) and integrated discrimination improvement (IDI) measures were calculated using the “survIDINRI” package27,28 in order to assess the improvement of risk prediction of adverse events. Continuous NRI was chosen over categorical NRI due to the lack of established meaningful risk categories in secondary risk prediction.29 Of the hematological parameters, RDW appeared to be performing particularly well. Therefore, baseline characteristics were additionally summarized by quartiles of RDW. All statistical analyses were performed using Rstudio30 and the R software package (version 3.1.2, Vienna, Austria)31. A p-value of <0.05 was considered statistically significant. Missings were deleted listwise (<10%); no bias could be detected in terms of differing MACE occurrence between patients with and without missing covariates.

Results Patient characteristics The baseline results are presented for the entire cohort and stratified by the occurrence of MACE during follow-up (Table 1). On average, people with MACE were older (67.2 vs. 63.4, p<0.001) than those without. Diabetes and hypertension were also more prevalent in the MACE group. The cardiovascular medical history of people with MACE more often showed ACS, PCI, CABG, CVA and kidney failure. The indication for coronary angiography did not differ between the groups. The left ventricular ejection fraction (LVEF) was poorer and the angiographic burden of CAD was more severe in the MACE group. Consequently, the treatment was more invasive in the MACE group. The use of prasugrel, beta blockers, ACE-inhibitors, statins and diuretics was higher in the MACE group. During a median follow-up time of 779 days, 99 deaths and 368 MACEs occurred.

266


Hematological Parameters Improve Prediction of Adverse Events

Table 1. Baseline characteristics of UCORBIO patients, stratified by MACE during follow-up. n Age (mean ± SD) Sex (% males) BMI (mean ± SD)

All

No MACE

MACE

1760

1392

368

p-value

64.22 ±10.85

63.44 ± 10.71

67.18 ± 10.88

72.7

72.9

72.0

0.779

27.14 ±4.48

27.13 ± 4.45

27.18 ± 4.60

0.854

<0.001

Diabetes (% yes)

22.3

19.5

32.6

<0.001

Hypertension (% yes)

57.6

55.8

64.4

0.004

Hypercholesterolemia (% yes)

47.3

46.8

48.9

Smoking (%)

0.516 0.047

Smoker

24.9

24.8

25.4

Ex smoker

28.5

27.2

33.4

Non-smoker

46.6

48.0

41.2

History of ACS (% yes)

31.1

29.0

39.1

<0.001

History of PCI (% yes)

29.7

27.3

38.9

<0.001

History of CABG (% yes)

12.0

10.3

18.5

<0.001

History of CVA (% yes)

9.7

8.6

13.9

0.004

History of PAD (% yes)

11.4

8.2

23.6

<0.001

Kidney failure (% yes)

2.7

1.6

6.8

<0.001

58.0

60.7

48.3

Mildly impaired (40-50%)

21.7

21.2

23.4

Moderately impaired (30-40%)

12.3

11.1

16.6

8.0

6.9

11.7

Medical History

LVEF (%) Normal (>=50%)

Poor (<30%)

<0.001

Medication Aspirin (% yes)

59.6

59.7

59.5

1.000

Clopidogrel (% yes)

21.5

21.0

23.6

0.296

Ticagrelor (% yes)

2.0

1.8

3.0

0.221

Prasugrel (% yes)

0.7

0.4

1.9

0.010

Beta-blocker (% yes)

56.3

54.7

62.2

0.012

ACEi (% yes)

36.1

34.4

42.7

0.004

Statins (% yes)

62.2

60.7

67.7

0.017

Diuretic (% yes)

29.6

26.5

41.3

<0.001

54.9

54.8

55.2

Coronary Angiography Indication for angiography (%) Stable CAD

0.842

UAP

10.1

9.8

11.1

Infarction

28.6

28.9

27.7

6.4

6.5

6.0

6.5

7.2

3.9

Other Angiographic CAD severity (%) No CAD

<0.001

Minor CAD

16.7

17.7

12.7

Single vessel disease

34.0

34.8

30.9

Multi vessel disease

42.8

40.3

52.5

Conservative

33.1

34.8

26.7

PCI

61.6

59.6

68.9

5.3

5.6

4.4

779 [573, 1033]

753 [558, 1005]

886 [677, 1112]

Treatment of CAD (%)

CABG Follow-up (days, median [IQR])

0.005

Abbreviations: SD=standard deviation, BMI= body mass index, ACS= acute coronary syndrome, PCI= percutaneous coronary intervention, CABG= coronary artery bypass grafting, CVA= cerebrovascular accident, PAD= peripheral arterial disease, LVEF= left ventricular ejection fraction, ACEi= angiotensin-converting enzyme inhibitor, CAD= coronary artery disease, UAP= unstable angina pectoris, IQR= interquartile range.

267


Chapter 14

Hematological Parameters In Table 2, the baseline levels of the hematological parameters of interest (n=37) are displayed by the occurrence of MACE during follow-up. Sixteen parameters differed significantly between patients with and without MACE during follow-up: leukocyte count, monocyte count, eosinophil count, basophil count, lymphocyte %, hemoglobin, % RCBs larger than 120fL, RDW, mean corpuscular hemoglobin concentration (MCHC), mean platelet volume (MPV), mean neutrophil cell size, mean neutrophil granularity/lobularity, mean neutrophil red fluorescence, lymphocyte cell size coefficient of variation (CV), platelet granularity CV and reticulocyte hemoglobin concentration (CHCr). Risk prediction with hematological parameters For each outcome, the ten best predictive hematological parameters (derived from backward stepwise analysis as described in the methods) were added to the clinical model containing age, sex, diabetes, indication for angiography, angiographic CAD severity, history of PCI, history of ACS, history of PAD, kidney failure, treatment following angiography (conservative, PCI or CABG) and use of diuretics. The hematological parameters that remained significantly associated with the outcome of interest are displayed in Table 3. Panels of hematological parameters, sized between 3 and 8 parameters, were significantly predictive on top of the clinical model. For all outcomes except re-PCI (for which the panel only contained leukocyte parameters) the panels consisted of parameters from both leukocyte and RBC origin. In particular, RDW was abundant and appeared in four panels (for MACE, all-cause death, non-cardiovascular death and myocardial infarction), thus showing broadly applicable predictive properties. Improvement of Risk Prediction Measures of prediction improvement were calculated for the prediction models extended with hematological parameters as compared to the basic clinical model. The cNRIs and IDIs resulting from this comparison are presented in Table 4. Additionally, Figure 1 shows the result from traditional ROC analysis for MACE, all-cause death, cardiovascular death, non-cardiovascular death, myocardial infarction and re-PCI. Supplemental Figures 2 and 3 provides visual representations of the changes in predicted risk after addition of hematological parameters. For MACE, the IDI - indicating the change in the difference of the predicted risk between patients with events and patients without events in the model extended with hematological parameters as compared to the model without hematological parameters32 - was The cNRI for MACE - indicating the proportion of individuals that were justly reclassified to a higher or lower risk by the extended model33 - was 0.17 (95% CI: 0.08-0.23, p<0.001). Additionally, for all-cause death, cardiovascular death, non-cardiovascular death, myocardial infarction and re-PCI significant and substantial IDIs and cNRIs (except for non-cardiovascular death, p=0.059) were calculated, thus demonstrating to provide improvement of prediction for a diversity of adverse outcomes.

268


Hematological Parameters Improve Prediction of Adverse Events

Table 2. Baseline values of hematological parameters, stratified by the occurrence of MACE during follow-up. No MACE

MACE

1392

368

Leukocyte count (10^9/L)

7.31 [5.97, 9.13]

7.64 [6.38, 9.56]

Lymphocyte count (10^9/L)

1.84 [1.42, 2.31]

1.77 [1.37, 2.25]

0.141

Monocyte count (10^9/L)

0.60 [0.47, 0.76]

0.65 [0.52, 0.79]

0.001 0.038

n

p-value 0.017

Eosinophil count (10^9/L)

0.14 [0.08, 0.23]

0.16 [0.09, 0.26]

Basophil count (10^9/L)

0.04 [0.02, 0.05]

0.04 [0.02, 0.06]

0.026

25.96 [19.94, 31.74]

24.09 [18.03, 30.16]

0.002

Lymphocyte % (% of leukocyte count) Monocyte % (%of leukocyte count) Hemoglobin (g/dL) % RBC >120fL (%) RDW (% CV) MCHC (g/dL) Plateletcrit (ml/L) Mean platelet volume (MPV) (fL) Platelet distribution width (10(GSD)) Reticulocyte count (10^9/L)

8.07 [6.69, 9.77]

8.32 [6.88, 9.90]

0.207

13.88 [12.83, 14.81]

13.35 [12.32, 14.39]

<0.001

1.46 [0.90, 2.10]

1.63 [0.92, 2.54]

0.040

11.99 [11.55, 12.51]

12.41 [11.72, 13.46]

<0.001

0.34 [0.34, 0.35]

0.34 [0.34, 0.35]

0.004

0.17 [0.15, 0.21]

0.18 [0.16, 0.21]

0.054 0.002

7.77 [7.22, 8.42]

8.00 [7.34, 8.62]

16.17 [15.76, 16.61]

16.12 [15.71, 16.58]

0.119

65.57 [52.04, 80.02]

68.03 [53.96, 83.02]

0.097 0.020

Mean neutrophil cell size (AU)

145.94 [139.48, 151.33]

146.96 [142.31, 151.97]

Mean neutrophil complexity (AU)

135.65 [131.62, 139.01]

135.09 [130.94, 138.61]

0.052

Mean neutrophil granularity/lobularity (AU)

28.33 [26.29, 30.88]

27.87 [25.54, 30.10]

0.004

Mean neutrophil red fluorescence (AU)

70.34 [68.99, 71.58]

70.74 [69.52, 71.93]

0.001

2.55 [2.30, 2.87]

2.52 [2.27, 2.78]

0.094

CV of neutrophil cell size (%) CV of neutrophil complexity (%)

3.49 [3.20, 3.78]

3.43 [3.17, 3.74]

0.065

CV of neutrophil lobularity (%)

7.73 [6.00, 9.09]

8.01 [6.23, 9.33]

0.121

15.08 [14.08, 16.01]

15.24 [14.32, 16.02]

0.145

7.91 [6.95, 8.75]

7.77 [6.82, 8.62]

0.096

100.57 [98.07, 103.08]

100.52 [97.81, 103.40]

0.937

75.71 [74.17, 77.32]

75.87 [74.06, 77.57]

0.620

CV of neutrophil granularity/lobularity (%) CV of neutrophil red fluorescence (%) Mean lymphocyte cell size (AU) Mean lymphocyte complexity (AU) CV of lymphocyte cell size (%)

4.88 [3.97, 5.84]

4.62 [3.78, 5.77]

0.025

CV of lymphocyte complexity (%)

4.83 [4.21, 5.47]

4.81 [4.16, 5.48]

0.438

17.01 [16.40, 17.61]

17.06 [16.50, 17.64]

0.201

13.24 [12.72, 13.90]

13.54 [12.98, 14.21]

<0.001

Mean RBC 7째 scatter RETC (AU)

181.31 [178.59, 182.58]

181.06 [178.65, 182.97]

0.345

Mean RBC FL1, RNA, RETC (AU)

83.69 [81.20, 86.20]

84.00 [81.54, 86.38]

0.278

CV of RBC 7째 scatter RETC (%)

1.67 [1.54, 1.84]

1.68 [1.52, 1.83]

0.675

CV of RBC FL1, RNA, RETC (%)

11.16 [9.70, 12.71]

11.26 [9.87, 12.73]

0.534 <0.001

CV of platelet size and complexity (%) CV of platelet lobularity (%)

CHCr (fmol)

30.53 [29.58, 31.39]

30.09 [29.15, 30.96]

Hemoglobin distribution width RBCs (%)

7.23 [6.69, 7.90]

7.32 [6.72, 8.02]

0.144

Reticulated platelet count (% of platelet count)

1.93 [1.49, 2.50]

1.93 [1.52, 2.50]

0.728

All measurements are shown as medians and interquartile ranges. Abbreviations: MACE= major adverse cardiovascular event, RBC= red blood cell, RDW= red cell distribution width, MCHC= mean corpuscular hemoglobin concentration, CHCr= reticulocyte mean corpuscular hemoglobin, CV= coefficient of variation (corresponds to standard deviation), AU= arbitrary units, GSD= geometric standard difference, RETC= reticulocyte absolute concentration.

269


Chapter 14

Table 3. Multivariable adjusted hazard ratios derived from backward stepwise Cox regression. Outcome

Hematological parameter

Origin*

MACE

RDW

E

Basophil count CV of neutrophil red fluorescence All-cause death

CV-death

Non-CV death

MI

Re-PCI

HR (95% CI)

p-value

1.19 (1.08-1.32)

<0.001

L

1.13 (1.05-1.23)

0.002

L

0.87 (0.78-0.98)

0.020

Leukocyte count

L

1.33 (1.18-1.49)

<0.001

Mean RBC fluorescence

E

1.57 (1.19-2.07)

0.001

CV of neutrophil red fluorescence

L

0.69 (0.55-0.87)

0.001

Lymphocyte %

L

0.70 (0.56-0.87)

0.001

Mean RBC complexity

E

0.40 (0.21-0.76)

0.005

Monocyte %

L

1.28 (1.08-1.53)

0.005

Reticulocyte mean MCHC

E

0.66 (0.49-0.89)

0.006

RDW

E

1.25 (1.04-1.49)

0.016

CV of neutrophil red fluorescence

L

0.44 (0.31-0.63)

<0.001

Reticulocyte count

E

1.59 (1.20-2.11)

0.001

Lymphocyte %

L

0.57 (0.38-0.85)

0.006

Reticulocyte mean MCHC

E

0.62 (0.42-0.92)

0.017

Leukocyte count

L

1.22 (1.10-1.35)

<0.001

Large RBC % (>120fL)

E

1.19 (1.05-1.36)

0.006

Mean RBC fluorescence

E

1.52 (1.03-2.25)

0.034

RDW

E

1.24 (1.01-1.54)

0.043

RDW

E

1.43 (1.19-1.71)

<0.001

Basophil count

L

1.38 (1.10-1.75)

0.006

CV of neutrophil complexity

L

0.72 (0.56-0.93)

0.012

Monocyte count

L

2.06 (1.10-3.86)

0.024

Monocyte %

L

0.52 (0.28-0.97)

0.038

Mean lymphocyte complexity

L

1.29 (1.07-1.55)

0.007

CV of neutrophil cell size

L

0.75 (0.61-0.93)

0.009

CV of lymphocyte complexity

L

1.25 (1.05-1.50)

0.014

Eosinophil count

L

1.16 (1.02-1.33)

0.024

For six outcome measures the hematological parameters that could predict outcome significantly and independently are displayed. The hazard ratios are the results from backwards stepwise Cox regression analysis containing the ten best predictive hematological parameters, correcting for: age, sex, diabetes, indication for angiography, angiographic CAD severity, history of PCI, history of ACS, history of PAD, kidney failure, treatment following angiography (conservative, PCI or CABG) and diuretic use. *Hematological origin: L=leukocyte, E=erythrocyte. CV=coefficient of variation, RDW= red cell distribution width, RBC= red blood cell, MCHC= mean corpuscular hemoglobin concentration.

270


Hematological Parameters Improve Prediction of Adverse Events

0.8 0.6

Sensitivity 0.4

0.6

0.8

0.2 1.0

0.0

0.2

0.4

0.6

1−Specificity

Cardiovascular Death

Non−cardiovascular Death

0.8

1.0

0.8 0.6

Sensitivity

0.4

0.4

0.6

0.8

1.0

1−Specificity

AUC Clinical = 0.779 AUC Clinical + Hematology = 0.792 p−value 0.163

0.0

0.0

0.2

AUC Clinical = 0.884 AUC Clinical + Hematology = 0.917 p−value 0.027

0.2

0.4

0.6

0.8

1.0

0.0

0.2

0.4

0.6

1−Specificity

1−Specificity

Myocardial Infarction

Re−PCI

0.8

1.0

0.2

AUC Clinical = 0.740 AUC Clinical + Hematology = 0.760 p−value 0.615

0.0

0.0

0.2

AUC Clinical = 0.738 AUC Clinical + Hematology = 0.773 p−value 0.195

0.6

0.8

0.4

Sensitivity

0.6

0.8

1.0

0.2

1.0

0.0

0.4

Sensitivity

AUC Clinical = 0.809 AUC Clinical + Hematology = 0.849 p−value <0.001

0.0

0.0

0.2

1.0

0.0

Sensitivity

0.4

0.6 0.4

AUC Clinical = 0.681 AUC Clinical + Hematology = 0.696 p−value 0.007

0.2

Sensitivity

0.8

1.0

All−cause Death

1.0

MACE

0.0

0.2

0.4

0.6

1−Specificity

0.8

1.0

0.0

0.2

0.4

0.6

0.8

1.0

1−Specificity

Figure 1. ROC curves of models for a clinical model with and without hematological parameters. P-values are given for difference between the area under the curve (AUC) of the clinical model plus hematological parameters (black line) as compared to the clinical model only (grey line). The hematological parameters added to the model are as stated in table 3.

271


Chapter 14

Table 4. Measures of prediction improvement (IDIs and cNRIs) upon addition of hematological parameters. Outcome

IDI (95% CI)

p-value IDI

cNRI (95% CI)

p-value cNRI

MACE

0.02 (0.01-0.03)

All-cause death

0.07 (0.03-0.12)

<0.001

0.17 (0.08-0.23)

<0.001

<0.001

0.28 (0.18-0.42)

Cardiovascular death

<0.001

0.06 (0.01-0.15)

<0.001

0.40 (0.18-0.56)

<0.001

Non-cardiovascular death

0.04 (0.00-0.12)

<0.001

0.16 (-0.03-0.36)

0.059

MI

0.02 (0.01-0.06)

<0.001

0.11 (0.02-0.25)

<0.001

Re-PCI

0.03 (0.01-0.09)

<0.001

0.27 (0.13-0.36)

<0.001

The IDIs and cNRIs were calculated for the comparison of a clinical model with a clinical model plus hematological parameters. The clinical model comprised: age, sex, diabetes, indication for angiography, angiographic CAD severity, history of PCI, history of ACS, history of PAD, kidney failure, treatment following angiography (conservative, PCI or CABG) and diuretic use. This clinical model was extended with the significant hematological parameters as displayed in table 2.

Association of Patient Characteristics with RDW RDW was predictive of four of six outcome measures. In order to better understand the patient groups in which this parameter is elevated we evaluated baseline patient characteristics by quartiles of RDW (supplemental table 1). We found that RDW was positively associated with age, BMI, diabetes and hypertension prevalence, a history of CABG, PAD, kidney failure, use of beta-blocker and diuretics. RDW was negatively associated with LVEF. Multivariable adjusted survival by RDW quartile is depicted in Figure 2.

0.9 0.8 0.7

[10.5,11.6] (11.6,12.1] (12.1,12.7] (12.7,23.1]

0.6

Survival probability (Cox proportional hazard)

1.0

MACE−free Survival stratified by RDW quartile

0

100

200

300

400

500

600

700

800

900

1000 1100 1200

FU time (days)

Figure 2. Multivariable adjusted MACE-free survival by RDW quartiles. MACE-free survival plot by RDW quartiles. The results are derived from Cox regression analysis adjusting for age, sex, diabetes, smoking, indication for coronary angiography, angiographic severity of CAD, history of PCI, history of ACS, history of PAD, kidney failure, treatment of CAD and diuretic use.

272


Hematological Parameters Improve Prediction of Adverse Events

Discussion In this study, we showed that the addition of readily available hematological parameters to a clinical model could significantly improve prediction of death and adverse events in coronary angiography patients. Efforts should be pursued to translate our findings into a clinically applicable risk score. More accurate identification of high-risk patients can lead to improved follow-up of patients at highest risk and treatment of those who will benefit most, thereby lowering the burden of cardiovascular morbidity and mortality. Predictive Properties of Hematological Parameters Among the hematological parameters tested in our study for their predictive value, RDW was most abundant. The RDW is routinely measured by dividing the SD of the mean corpuscular volume (MCV) distribution by the mean of the MCV and then multiply it by 100 to provide a percentage.34 High RDW thus reflects a higher variation in RBC volumes, also referred to as anisocytosis. Traditionally, RDW is measured to aid differential diagnosis of anemias. However, ours and other studies have shown that higher RDW is associated with poorer outcome for traumatic injuries 35, sepsis36–38, stroke39,40, myocardial infarction12,41–43, PCI44–46, heart failure47–51 and in the general population13. In the current study we confirmed that RDW independently or in combination with other hematological parameters predicts mortality and secondary adverse events in a coronary angiography population. In addition to prior studies, we demonstrated that the addition of hematological parameters to clinical data can indeed improve risk prediction using modern statistical techniques (IDI32 and cNRI33). In addition to RDW, we found predictive potential for several leukocyte parameters; the CV of neutrophil red fluorescence (for MACE, all-cause death and MI), basophil counts (MACE and MI), lymphocyte % (for all-cause death and CV-death), monocyte % (for allcause death and MI), mean RBC red fluorescence (for all-cause death and non-CV death) and leukocyte count (for all-cause death and non-CV death). Some of these parameters: leukocyte, monocyte and lymphocyte counts have been described before,9,52 but the predictive values of the CV of neutrophil red fluorescence and basophil counts are largely uncovered in the current literature. To our knowledge, the CV of neutrophil red fluorescence has not been mentioned in the context of cardiovascular disease before. However, in a patient group with symptomatic PAD, basophil count was not an independent predictor of MACE53 and also among community-dwelling elderly, basophil count was not significantly associated with a history of cardiovascular disease (odds ratio 1.21 [0.98-1.50]).54 Possibly, these populations were too homogeneous for basophil counts to offer additive discriminative value. One can imagine that within the general population basophil counts are low, with little variation. The same could apply to a very sick population (like symptomatic PAD patients53), who would have high basophil counts with little variation. In our study population, patients with angiographic CAD severity ranging from no CAD or minor CAD to triple vessel disease are enrolled, thus representing a relatively heterogeneous population.

273


Chapter 14

Red Blood Cells and Cardiovascular Disease Several mechanisms relate cardiovascular disease to changes in RBC characteristics. First, atherosclerosis is hallmarked by oxidative stress. Upon oxidative stress, RBCs adopt a more irregular and heterogeneous conformation55. RBCs can encounter oxidative stress by passing through jeopardized tissues or microenvironments, such as atherosclerotic plaques.56 The oxidative changes can cause an increase in RBC degradation and turnover, resulting in a higher proportion of small RBCs and thus increased anisocytosis (higher RDW). Second, inflammation is a keystone of atherosclerosis and several proinflammatory cytokines (e.g. IL-657) have been related to increased RDW. Inflammatory cytokines such as interferon-Îł and tumor necrosis factor, which are elevated in CAD58, suppress erythropoiesis and stimulate phagocytosis of senescent RBCs, thereby increasing anisocytosis.59 Third, CAD is frequently accompanied by some degree of kidney function impairment60. Erythropoietin (EPO) is a hormone produced in the renal cortex promoting erythropoiesis and erythrocyte maturation. Disturbances in EPO production34 and responsiveness61 have been related to increased RDW. As EPO levels decrease upon inflammation62, a disturbed erythropoiesis and thereby increase in RDW can be observed63. Secondary Risk Prediction Improvement in Clinical Practice In the current study, we showed that risk estimation following coronary angiography can be significantly improved by addition of hematological parameters. These parameters are readily available in the vast majority of medical centers as they are measured with every differential blood count on automated hematology analyzers. Clinical risk prediction rules therefore might be effortlessly extended with a panel of hematological parameters, resulting in more accurate identification of high-risk individuals. These high-risk individuals need to be identified in order to justly provide expensive secondary prevention therapies with limited availability, such as the soon-to-come PCSK9 inhibitors. While our results are promising, external validation is warranted in order to establish the clinical usefulness of hematological parameters in the context of risk prediction. Limitations In our cohort lipid levels were not available. Therefore, established secondary risk prediction scores as the PROCAM, Framingham, SCORE or SMART-score64 could not be applied. Also, the duration of symptoms and the delay between acute onset of chest pain and the moment of coronary angiography might affect the levels of hematological parameters. However, the majority of our cohort consists of stable CAD patients (55%) without acute symptoms. Hence, such effects on RDW are very unlikely. In patients with >1 day delay between symptom onset and angiography, we investigated the possible correlation between the delay and RDW, which yielded no significant result (p=0.399). A study conducting repetitive measures would be necessary for evaluating changes of hematological parameters throughout the course of CAD, such as the BioMarcs65 program.

274


Hematological Parameters Improve Prediction of Adverse Events

In the current analyses we did not consider interaction terms for, for example, sex. It is possible, in the light of differing reference values for hematological parameters that different coefficients need to be applied to men and women. Future research has to evaluate the need for sex interaction terms in a clinically applicable risk prediction model. Conclusion Hematological parameters, particularly the RDW, can significantly improve the prediction of secondary adverse events in a coronary angiography population. This will help to identify high-risk patients more accurately and tailor secondary prevention based on individual risk. The clinical potential of a risk score extended with hematological parameters needs to be evaluated further. Declarations Funding This work was supported by a grant from the Netherlands Heart Foundation: 2013T084, Queen of Hearts: Improving diagnosis of CVD in women to Hester den Ruijter. Dominique de Kleijn is funded through a strategic grant from the Royal Netherlands Academy of Arts and Sciences to the Interuniversity Cardiology Institute of the Netherlands, ICIN, the National University Singapore Startup grant, the Singapore National Medical Research Council Centre Grant and the ATTRaCT, SPF 2014/003 grant BMRC. Folkert Asselbergs is supported by the UCL Hospitals NIHR Biomedical Research Centre and a Dekker scholarship-Junior Staff Member 2014T001 – Dutch Heart Foundation. These funding sources in no way influenced the analyses or the content of this manuscript. Competing Interests None of the authors have any competing interests. Acknowledgements We sincerely thank ms. Jonne Hos for her outstanding support to the UCORBIO cohort.

275


Chapter 14

References 1.

Nauta ST, Deckers JW, Akkerhuis M, Lenzen M, Simoons ML, van Domburg RT. Changes in clinical profile, treatment, and mortality in patients hospitalised for acute myocardial infarction between 1985 and 2008. PLoS One. 2011;6.

2.

DAWBER TR, MEADORS GF, MOORE FE. Epidemiological approaches to heart disease: the Framingham Study. Am J Public Health. 1951;41:279–281.

3.

Ridker PM, Buring JE, Shih J, Matias M, Hennekens CH. Prospective Study of C-Reactive Protein and the Risk of Future Cardiovascular Events Among Apparently Healthy Women. Circulation. 1998;98:731–733.

4.

Garcia-Moll X, Zouridakis E, Cole D, Kaski JC. C-reactive protein in patients with chronic stable angina: differences in baseline serum concentration between women and men. Eur Heart J. 2000;21:1598–606.

5.

Shlipak MG, Sarnak MJ, Katz R, Fried LF, Seliger SL, Newman AB, Siscovick DS, Stehman-Breen C. Cystatin C and the risk of death and cardiovascular events among elderly persons. N Engl J Med. 2005;352:2049–60.

6. Luc G, Bard J-M, Lesueur C, Arveiler D, Evans A, Amouyel P, Ferrieres J, Juhan-Vague I, Fruchart J-C, Ducimetiere P. Plasma cystatin-C and development of coronary heart disease: The PRIME Study. Atherosclerosis. 2006;185:375–80. 7.

Zhang R, Brennan ML, Fu X, Aviles RJ, Pearce GL, Penn MS, Topol EJ, Sprecher DL, Hazen SL. Association between myeloperoxidase levels and risk of coronary artery disease. JAMA. 2001;286:2136–42.

8.

Wong ND, Gransar H, Narula J, Shaw L, Moon JH, Miranda-Peats R, Rozanski A, Hayes SW, Thomson LEJ, Friedman JD, Berman DS. Myeloperoxidase, subclinical atherosclerosis, and cardiovascular disease events. JACC Cardiovasc Imaging. 2009;2:1093–9.

9.

Lee C Do, Folsom AR, Nieto FJ, Chambless LE, Shahar E, Wolfe DA. White blood cell count and incidence of coronary heart disease and ischemic stroke and mortality from cardiovascular disease in African-American and White men and women: atherosclerosis risk in communities study. Am J Epidemiol. 2001;154:758–64.

10. Gillum RF, Mussolino ME, Madans JH. Counts of neutrophils, lymphocytes, and monocytes, cause-specific mortality and coronary heart disease: the NHANES-I epidemiologic follow-up study. Ann Epidemiol. 2005;15:266–71. 11. Lappegård J, Ellingsen TS, Vik A, Skjelbakken T, Brox J, Mathiesen EB, Johnsen SH, Brækkan SK, Hansen J-B. Red cell distribution width and carotid atherosclerosis progression. Thromb Haemost. 2015;113:649–654. 12. Sahin O, Akpek M, Sarli B, Baktir AO, Savas G, Karadavut S, Elcik D, Saglam H, Kaya MG, Arinc H. Association of Red Blood Cell Distribution Width Levels with Severity of Coronary Artery Disease in Patients with Non-ST Elevation Myocardial Infarction. Med Princ Pract. 2014;24:178–183. 13. Perlstein TS, Weuve J, Pfeffer M a, Beckman J a. Red blood cell distribution width and mortality risk in a community-based prospective cohort. Arch Intern Med. 2009;169:588–594. 14. Widlansky ME, Sesso HD, Rexrode KM, Manson JE, Gaziano JM. Body mass index and total and cardiovascular mortality in men with a history of cardiovascular disease. Arch Intern Med. 2004;164:2326–2332. 15. Shimada YJ, Cannon CP. PCSK9 (Proprotein convertase subtilisin/kexin type 9) inhibitors: past, present, and the future. Eur Heart J. 2015;Epub:1–12. 16. Gijsberts CM, Gohar A, Ellenbroek GHJM, Hoefer IE, de Kleijn DPV, Asselbergs FW, Nathoe HM, Agostoni P, Rittersma SZH, Pasterkamp G, Appelman Y, den Ruijter HM. Severity of stable coronary artery disease and its biomarkers differ between men and women undergoing angiography. Atherosclerosis. 2015;241:234–240.

276


Hematological Parameters Improve Prediction of Adverse Events

17. Ten Berg MJ, Huisman A, van den Bemt PML a, Schobben AF a M, Egberts ACG, van Solinge WW. Linking laboratory and medication data: new opportunities for pharmacoepidemiological research. Clin Chem Lab Med. 2007;45:13–9. 18. Gijsberts CM, Seneviratna A, de Carvalho LP, den Ruijter HM, Vidanapthirana P, Sorokin V, Stella P, Agostoni P, Asselbergs FW, Richards AM, Low AF, Lee C-H, Tan HC, Hoefer IE, Pasterkamp G, de Kleijn DP V., Chan MY. Ethnicity Modifies Associations between Cardiovascular Risk Factors and Disease Severity in Parallel Dutch and Singapore Coronary Cohorts. PLoS One. 2015;10:e0132278. 19. Harris PJ, Behar VS, Conley MJ, Harrell FE, Lee KL, Peter RH, Kong Y, Rosati R a. The prognostic significance of 50% coronary stenosis in medically treated patients with coronary artery disease. Circulation. 1980;62:240–248. 20. Lam SW, Leenen LPH, Van Solinge WW, Hietbrink F, Huisman A. Evaluation of hematological parameters on admission for the prediction of 7-day in-hospital mortality in a large trauma cohort. Clin Chem Lab Med. 2011;49:493–499. 21. Groeneveld KM, Heeres M, Leenen LPH, Huisman A, Koenderman L. Immunophenotyping of posttraumatic neutrophils on a routine haematology analyser. Mediators Inflamm. 2012;2012. 22. Lam SW, Leenen LPH, van Solinge WW, Hietbrink F, Huisman A. Comparison between the prognostic value of the white blood cell differential count and morphological parameters of neutrophils and lymphocytes in severely injured patients for 7-day in-hospital mortality. Biomarkers. 2012;17:642–647. 23. Von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, Vandenbroucke JP. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) Statement: Guidelines for Reporting Observational Studies. PLoS Med. 2007;4:e296. 24. Binder H. CoxBoost: Cox models by likelihood based boosting for a single survival endpoint or competing risks. 2013. 25. Blanche P, Dartigues J-F, Jacqmin-Gadda H. Estimating and Comparing time-dependent areas under receiver operating characteristic curves for censored event times with competing risks. Stat Med. 2013;32:5381–5397. 26. Chiang C-T, Hung H. Non‐parametric estimation for time-dependent AUC. J Stat Plan Inference. 2010;140:1162–1174. 27. Uno H, Tian L, Cai T, Kohane IS, Wei LJ. A unified inference procedure for a class of measures to assess improvement in risk prediction systems with survival data. Stat Med. 2013;32:2430–2442. 28. Pencina MJ, D’Agostino RB, Steyerberg EW. Extensions of net reclassification improvement calculations to measure usefulness of new biomarkers. Stat Med. 2011;30:11–21. 29. Sniderman AD, D’Agostino RB, Pencina MJ. The Role of Physicians in the Era of Predictive Analytics. JAMA. 2015;314:25–6. 30. Allaire J. RStudio: Integrated development environment for R. 2012; 31. R Core Team. R: A Language and Environment for Statistical Computing. Vienna, Austria: 2014. 32. Steyerberg EW, Vickers AJ, Cook NR, Gerds T, Gonen M, Obuchowski N, Pencina MJ, Kattan MW. Assessing the Performance of Prediction Models. Epidemiology. 2010;21:128–138. 33. Pencina MJ, D’Agostino RB, Vasan RS. Statistical methods for assessment of added usefulness of new biomarkers. Clin Chem Lab Med. 2010;48:1703–11. 34. Salvagno GL, Sanchis-Gomar F, Picanza A, Lippi G. Red blood cell distribution width: A simple parameter with multiple clinical applications. Crit Rev Clin Lab Sci. 2014;8363:1–20. 35. Garbharran U, Chinthapalli S, Hopper I, George M, Back DL, Dockery F. Red cell distribution width is an independent predictor of mortality in hip fracture. Age Ageing. 2013;42:258–261.

277


Chapter 14

36. Karagöz E, Tanoglu A. Red blood cell distribution width : An emerging predictor for mortality in critically ill patients ? 2014;2014. 37. Kim CH, Kim SJ, Lee MJ, Kwon YE, Kim YL, Park KS, Ryu HJ, Park JT, Han SH, Yoo T, Kang S, Oh HJ. An Increase in Mean Platelet Volume from Baseline Is Associated with Mortality in Patients with Severe Sepsis or Septic Shock. PLoS One. 2015;10:e0119437. 38. Lorente L, Martín MM, Abreu-González P, Solé-Violán J, Ferreres J, Labarta L, Díaz C, González O, García D, Jiménez A, Borreguero-León JM. Red Blood Cell Distribution Width during the First Week Is Associated with Severity and Mortality in Septic Patients. PLoS One. 2014;9:e105436. 39. Kara H, Degirmenci S, Bayir A, Ak A, Akinci M, Dogru A, Akyurek F, Kayis SA. Red cell distribution width and neurological scoring systems in acute stroke patients. Neuropsychiatr Dis Treat. 2015;733. 40. Söderholm M, Borné Y, Hedblad B, Persson M, Engström G. Red Cell Distribution Width in Relation to Incidence of Stroke and Carotid Atherosclerosis: A Population-Based Cohort Study. PLoS One. 2015;10:e0124957. 41. Bekler A, Tenekecioglu E, Erbag G, Temiz A, Altun B, Barutcu A, Gazi E, Gunes F, Yilmaz M. Relationship between red cell distribution width and long-term mortality in patients with non-ST elevation acute coronary syndrome. Anadolu Kardiyol Dergisi/The Anatol J Cardiol. 2014;6–11. 42. Sun X, Chen W, Sun Z, Ding X, Gao X, Liang S, Zhao H, Yao D, Chen H, Li H, Li D. Impact of Red Blood Cell Distribution Width on Long-Term Mortality in Patients with ST-Elevation Myocardial Infarction. Cardiology. 2014;128:343–348. 43. Ephrem G. Red blood cell distribution width is a predictor of readmission in cardiac patients. Clin Cardiol. 2013;36:293–299. 44. Cavusoglu E, Chopra V, Gupta A, Battala VR, Poludasu S, Eng C, Marmur JD. Relation between red blood cell distribution width (RDW) and all-cause mortality at two years in an unselected population referred for coronary angiography. Int J Cardiol. 2010;141:141–146. 45. Tsuboi S, Miyauchi K, Kasai T, Ogita M, Dohi T, Miyazaki T, Yokoyama T, Kojima T, Yokoyama K, Kurata T, Daida H. Impact of red blood cell distribution width on long-term mortality in diabetic patients after percutaneous coronary intervention. Circ J. 2013;77:456–461. 46. Liu X, Ma C, Liu X, Du X, Kang J, Zhang Y, Wu J. Relationship between red blood cell distribution width and intermediate-term mortality in elderly patients after percutaneous coronary intervention. J Geriatr Cardiol. 2015;12:17–22. 47. Jung C, Fujita B, Lauten A, Kiehntopf M, Küthe F, Ferrari M, Figulla HR. Red blood cell distribution width as useful tool to predict long-term mortality in patients with chronic heart failure. Int J Cardiol. 2011;152:417–418. 48. Núñez J, Núñez E, Rizopoulos D, Miñana G, Bodí V, Bondanza L, Husser O, Merlos P, Santas E, Pascual-Figal D, Chorro FJ, Sanchis J. Red blood cell distribution width is longitudinally associated with mortality and anemia in heart failure patients. Circ J. 2014;78:410–8. 49. Dai Y, Konishi H, Takagi A, Miyauchi K, Daida H. Red cell distribution width predicts short- and long-term outcomes of acute congestive heart failure more effectively than hemoglobin. Exp Ther Med. 2014;600–606. 50. Van Kimmenade RRJ, Mohammed A a., Uthamalingam S, Van Der Meer P, Felker GM, Januzzi JL. Red blood cell distribution width and 1-year mortality in acute heart failure. Eur J Heart Fail. 2010;12:129–136. 51. Pascual-Figal D a., Bonaque JC, Redondo B, Caro C, Manzano-Fernandez S, Sánchez-Mas J, Garrido IP, Valdes M. Red blood cell distribution width predicts long-term outcome regardless of anaemia status in acute heart failure patients. Eur J Heart Fail. 2009;11:840–846.

278


Hematological Parameters Improve Prediction of Adverse Events

52. ó Hartaigh B, Bosch J a., Thomas GN, Lord JM, Pilz S, Loerbroks A, Kleber ME, Grammer TB, Fischer JE, Boehm BO, März W. Which leukocyte subsets predict cardiovascular mortality? From the LUdwigshafen RIsk and Cardiovascular Health (LURIC) Study. Atherosclerosis. 2012;224:161–169. 53. Haumer M, Amighi J, Exner M, Mlekusch W, Sabeti S, Schlager O, Schwarzinger I, Wagner O, Minar E, Schillinger M. Association of neutrophils and future cardiovascular events in patients with peripheral artery disease. J Vasc Surg. 2005;41:610–617. 54. Pinto EM, Huppert F a., Morgan K, Brayne C. Neutrophil counts, monocyte counts and cardiovascular disease in the elderly. Exp Gerontol. 2004;39:615–619. 55. Buttari B, Profumo E, Riganò R. Crosstalk between Red Blood Cells and the Immune System and Its Impact on Atherosclerosis. Biomed Res Int. 2015;2015:1–8. 56. Minetti M, Agati L, Malorni W. The microenvironment can shift erythrocytes from a friendly to a harmful behavior: Pathogenetic implications for vascular diseases. Cardiovasc Res. 2007;75:21–28. 57. Emans ME, van der Putten K, van Rooijen KL, Kraaijenhagen RJ, Swinkels D, van Solinge WW, Cramer MJ, Doevendans PAFM, Braam B, Gaillard CAJM. Determinants of red cell distribution width (RDW) in cardiorenal patients: RDW is not related to erythropoietin resistance. J Card Fail. 2011;17:626–33. 58. Hansson GK, Robertson A-KL, Söderberg-Nauclér C. Inflammation and atherosclerosis. Annu Rev Pathol. 2006;1:297–329. 59. Weiss G, Goodnough LT. Anemia of chronic disease. N Engl J Med. 2005;352:1011–1023. 60. Doganer YC, Rohrer JE, Aydogan U, Barcin C, Cayci T, Saglam K. Association of renal function, estimated by four equations, with coronary artery disease. Int Urol Nephrol. 2015;47:663–671. 61. Afsar B, Saglam M, Yuceturk C, Agca E. The relationship between red cell distribution width with erythropoietin resistance in iron replete hemodialysis patients. Eur J Intern Med. 2013;24:e25–9. 62. Jelkmann W. Proinflammatory cytokines lowering erythropoietin production. J Interferon Cytokine Res. 1998;18:555–559. 63. Patel HH, Patel HR, Higgins JM. Modulation of red blood cell population dynamics is a fundamental homeostatic response to disease. Am J Hematol. 2015;90:422–428. 64. Uthoff, Staub, Socrates, Meyerhans, Bundi, Schmid, Frauchiger. PROCAM-, FRAMINGHAM-, SCORE- and SMART-risk score for predicting cardiovascular morbidity and mortality in patients with overt atherosclerosis. Vasa. 2010;39:325–333. 65. De Mulder M, Umans VA, Stam F, Cornel JH, Oemrawsingh RM, Boersma E. Intensive management of hyperglycaemia in acute coronary syndromes. Study design and rationale of the BIOMArCS2 glucose trial. Diabet Med. 2011;28:1168–1175.

279


Chapter 14

Supplemental

Supplemental figure 1. Hierarchically clustered heatmap displaying correlations amongst hematological parameters. Heatmap of correlations between hematological parameters. Red areas indicate parameters that are positively associated, blue areas indicate a negative correlation. Light colored areas indicate no or weak correlation between parameters. Parameters printed italic were strongly correlated to another parameter; of that cluster the parameter printed in bold was selected for analysis (as were all normally printed parameters).

280


Hematological Parameters Improve Prediction of Adverse Events ● ● ●

0.3 0.2

● ●

0.1 0.0

● ●

● ● ● Mean risk: 0.19 ● ● ● ● ●

MACE with Hematology

● ● ● ● ● ●

0.5 0.4

● No event ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●

● ● ● ● ● ● ● ●

Mean risk: 0.29

Event

0.2 0.1

Mean risk: 0.18

0.5 0.4 0.3 0.2

No event ●

● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●

0.1 0.0

Event ●

● ● ●

● ● ● ● Mean risk: 0.06

Mean risk: 0.18

0.5 0.4

No event ●

Event ●

● ● ● ●

0.3 0.2 0.1 0.0

● ● ● ● ● ● ● ● ● ●

Mean risk: 0.02

Mean risk: 0.13

● No event

Event

● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●

● ●

0.3

0.0

● ● ●

Event

CV death without Hematology

0.4

No event

● ●

All−cause death without Hematology

0.5

● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●

Mean risk: 0.30

All−cause death with Hematology

MACE without Hematology

● ● No event ● ●

0.5 0.4 0.3 0.2 0.1 0.0

● ● ● ● ●

Event

● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●

Mean risk: 0.05

Mean risk: 0.25

No event

Event

CV death with Hematology

● ● ●

0.5 0.4 0.3 0.2 0.1 0.0

● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●

Mean risk: 0.02

Mean risk: 0.19

No event

Event

No event

0.3 0.2 0.1 0.0

Event

● ●

● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●

Mean risk: 0.03

● ● ● ● ● ●

0.5 0.4 0.3 0.2

● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●

● ●

● Mean risk: 0.05

Mean risk: 0.09

0.1 0.0

Mean risk: 0.09

0.5 0.4 0.3 0.2 0.1 0.0

● No event

Event

● ● ● ● ●

● ●

● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●

Mean risk: 0.03

0.5 0.4 0.3 0.2

● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●

0.1 0.0

Mean risk: 0.08

Mean risk: 0.14

No event ● ●

● Event

● ●

Mean risk: 0.13

0.5

MI with Hematology

Non−CV death with Hematology

rePCI without Hematology

0.4

0.4 0.3 0.2 0.1 0.0

● No event ●

Event

● ● ●

● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●

Mean risk: 0.05

● ● ● ●

Mean risk: 0.12

rePCI with Hematology

0.5

MI without Hematology

Non−CV death without Hematology

0.5

0.4 0.3 0.2

● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●

0.1 0.0

Mean risk: 0.08

Mean risk: 0.16

Supplemental figure 2. Predicted risks without and with addition of hematological parameters for patients who were deceased and who were alive. The boxplots show the median and interquartile ranges of the predicted risks for patients who experienced an event and those who did not during follow-up (all six outcome measures). For each outcome measure risk is predicted using the clinical model and the clinical model plus hematological parameters. The difference upon addition of hematological parameters in the slope between the mean predicted risk of patients with events and patients without represents the IDI (e.g. estimated to be 0.07 or 7% for all-cause death at 2 years of follow-up).

281


Chapter 14

Supplemental figure 3. Predictive properties of clinical model with and without hematological parameters for all-cause death and adverse events. The circles indicate patients who did not experience the outcome of interest at 2-year follow-up. The red triangles indicate patients who did experience the outcome of interest. Patients with an event who shifted upwards (above the dashed line) were correctly reclassified to higher risk by adding the hematological parameters to the model- patients without event who moved downwards (below the dashed line) were appropriately reclassified to a lower risk.

282


283



PART THREE General Risk Prediction

Chapter 15 The Value of Hematological Parameters Exceeds High-Sensitivity Troponin I and NT Pro-BNP for Mortality Prediction in Coronary Angiography Patients In preparation

Crystel M. Gijsberts, Hester M. den Ruijter, Dominique P.V. de Kleijn, Albert Huisman, Mark de Groot, Richard H.A. van Wijk, Folkert W. Asselbergs, Michiel Voskuil, Gerard Pasterkamp, Wouter W. van Solinge, Imo E. Hoefer


Chapter 15

Abstract Background High-sensitivity Troponin I (hsTnI) and N-terminal pro-brain natriuretic peptide (NT proBNP) have been reported to be predictive of mortality in patients with coronary artery disease. Recently, hematological parameters have emerged as potential predictors of mortality in several medical disciplines. In the current study we compared the predictive value of hematological parameters with hsTnI and NT pro-BNP in a coronary angiography population. Methods Coronary angiography patients (n=2,379) from the UCORBIO cohort were included in this study. Hematological parameters were extracted from UPOD. hsTnI and NT pro-BNP were measured at the moment of coronary angiography. We compared the predictive properties of these parameters using Cox regression, receiver operating characteristics, integrated discrimination improvement (IDI) and continuous net reclassification improvement (cNRI) analysis. Results During a median follow-up duration of 1.8 years 77 deaths occurred. The predictive value (on top of clinical characteristics) of hematological parameters (AUC 0.856, p<0.001, IDI 0.07, p<0.001, cNRI 0.37, p<0.001) was larger than hsTnI (AUC 0.818) and NT pro-BNP (AUC 0.834) separately or combined (AUC 0.834). hsTnI and NT pro-BNP could marginally increase the AUC of hematological parameters (AUC increased from 0.856 to 0.865, p=0.049) but did not improve prediction (IDI and cNRI non-significant). Conclusion Readily available hematological parameters were superior to hsTnI or NT pro-BNP either alone or in combination for prediction of mortality. Addition of hsTnI and NT pro-BNP to hematological parameters did not significantly improve prediction of mortality in coronary angiography patients.

286


Hematological Parameters vs. hsTnI and NT pro-BNP

Introduction In addition to their diagnostic properties, high-sensitivity Troponin I1 (hsTnI) and NT proB-natriuretic peptide2 (NT pro-BNP) have been shown to have prognostic value in stable heart failure3, acute myocardial infarction4, elective coronary angiography patients5, stable and unstable coronary artery disease (CAD)6–8 and percutaneous coronary intervention (PCI) patients9. Recently, a different and easily accessible type of biomarkers has emerged in the cardiovascular literature. Numerous blood cell characteristics, predominantly of leukocyte origin10–12 and red blood cell origin13–15 have been proposed as prognostic tools in a vast diversity of populations. These hematological parameters are widely available and measured on a routine basis. Modern automated hematology analyzers automatically perform a whole blood cell differentiation irrespective of the clinical request. The unrequested parameters are not routinely reported back to the physician, but can be stored for future reference. In the University Medical Center in Utrecht blood cell differentiation data from the Abbott Sapphire16 hematology analyzer have been stored in the Utrecht Patient Oriented Database (UPOD)17 for research purposes. Precise prediction of mortality in CAD patients is key for a patient-specific treatment policy and for accurate patient information on prognosis. In the current study we compared hematological parameters with hsTnI and NT pro-BNP as predictors of mortality in addition to a clinical model.

Methods Study population In this study we analyzed data from the UCORBIO cohort (clinicaltrials.gov identifier: NCT02304744), a cohort study of coronary angiography patients in the University Medical Center in Utrecht, the Netherlands. From October 2011 to December 2014, a total of 2,591 patients were enrolled. For the current study, adult (>18 years) patients presenting with myocardial infarction (either ST-Segment Elevation Myocardial Infarction or NonST-Segment Elevation Myocardial Infarction), chest pain without release of cardiac enzymes (stable or unstable angina), dyspnea on exertion, silent ischemia or screening for non-cardiac surgery were selected (n=2,379). Patients with other indications for coronary angiography (coronary anomalies, screening for cardiac surgery or heart transplant follow-up) were thus excluded (n=212). Complete data was available in 1,913 cases. Ethics, consent and permissions All patients provided written informed consent and the study conforms to the Declaration of Helsinki. The institutional review board of the University Medical Centre Utrecht approved of this study (reference number 11-183).

287


Chapter 15

Data collection The process of data collection in the UCORBIO cohort has been described before.18 In summary, standardized electronic case report forms were completed at baseline containing age, sex, cardiovascular risk factors, indication for angiography, medication use, angiographic findings and eventual treatment of CAD. The angiographic findings were categorized into four groups by the treating interventional cardiologist: no CAD, minor CAD (wall irregularities, <50% stenosis), single vessel disease (one vessel with >50% stenosis19) and multi-vessel disease (two or three vessels with >50% stenosis). Biomarkers Levels of hsTnI were measured using the STAT High Sensitive Troponin-I assay on the clinically validated ARCHITECT i2000 analyzer (Abbott Laboratories, Lisnamuck, Longford, Ireland). NT pro-BNP was measured using a semi-automated ELISA robot (Freedom EVO, Tecan, Switzerland, antibodies: 15C4 and biotinylated 13G12, Hi-test Finland). Hematological parameters The hematological parameters were obtained through complete blood count analysis at the moment of coronary angiography. The parameters that were used in this study comprised 34 routinely measured hematological parameters from the UPOD database17. The UPOD parameters contain information on the number and characteristics of red blood cells (RBC), leukocytes and platelets. All hematological parameters were measured using the Cell-Dyn Sapphire20–22 hematology analyzer (Abbott Diagnostics, Santa Clara, CA, USA). This analyzer is equipped with an integrated 488-nm blue diode laser and uses spectrophotometry, electrical impedance, laser light scattering (multi-angle polarized scatter separation), and three-color fluorescent technologies to measure morphological parameters of leukocytes, RBCs and platelets for classification and enumeration. The morphological parameters entail the following five optical scatter signals for leukocytes: cell size (0˚ scatter, axial light loss), cell complexity and granularity (7˚ scatter, intermediate angle scatter (IAS)), nuclear lobularity (90˚ scatter, polarized side scatter (PSS)), depolarization (90˚ depolarized side scatter (DSS)) and viability (red fluorescence (FL-3), 630 ± 30 nm). For platelets, two optical scatter signals are measured: IAS scatter (7˚, cell size) and PSS scatter (90˚, granularity; internal structure). RBC parameters are measured or calculated based on the impedance measurement. Reticulocytes are optically measured using IAS scatter (7˚, cell size) and FL-1 fluorescence (RNA content). Throughout this paper, all values of hematological parameters are reported as multitudes of their standard deviation (SD) in order to ensure comparability of effect sizes among parameters with absolute values that vary strongly in their order of magnitude. Statistical Analysis Baseline characteristics were reported as means and standard deviations for continuous variables and percentages for categorical variables, for the entire cohort and separately for patients who died during follow-up and who did not.

288


Hematological Parameters vs. hsTnI and NT pro-BNP

First, we constructed a clinical risk prediction model. Covariates for this model were selected using a backward stepwise Cox regression model for all-cause mortality and comprised: age, sex, diabetes, hypercholesterolemia, smoking status, indication for angiography, angiographic CAD severity, history of PCI, history of acute coronary syndrome (ACS), kidney failure and treatment following angiography. In order to determine a set of hematological parameters that could aid prediction of mortality, first we evaluated mutual correlation of the parameters (total n=56) by means of hierarchically clustered heatmap analysis. From each cluster of collinear parameters, the parameter that showed the strongest relation with all-cause mortality was selected for further analysis. The remaining parameters (n= 34) were entered in a backward stepwise Cox regression models. From this procedure the top 10 significant parameters were added to the clinical model (which was coerced to stay in the model) and again backward stepwise Cox regressions were performed for the hematological parameters, rendering the final set of hematological parameters for all-cause mortality. We assessed the prognostic value of addition of hsTnI values, NT pro-BNP values and the set of hematological parameters to the clinical model, using receiver operating characteristics (ROC) analysis. The clinical model was entered as a linear predictor to stabilize its predictive value. Next, we evaluated the prognostic value of adding hsTnI values, NT pro-BNP values or both to the panel of hematological parameters (all on top of the clinical model). The predictive properties of the set of hematological parameters were internally validated by means of post-estimation parameterwise shrinkage23 using the jackknife method. For this purpose we used the “shrink” package24 for R. Furthermore, continuous net reclassification improvement (cNRI) and integrated discrimination improvement (IDI) measures for the abovementioned comparisons were calculated using the “survIDINRI” package25,26 in order to assess the improvement of risk prediction. Continuous NRI was deemed preferable over categorical NRI due to the lack of established meaningful risk categories in secondary risk prediction.27 Additionally we created calibration plots, to assess the goodness of fit with Hosmer-Lemeshow testing. All statistical analyses were performed using Rstudio28 and the R software package (version 3.1.2, Vienna, Austria)29. A p-value of <0.05 was considered statistically significant.

Results Patient characteristics Patient characteristics are shown in table 1, separately for deceased and non-deceased patients. Patients who died were older than those who survived (72.5 years vs. 63.3 years, p<0.001). Diabetes was more prevalent (37.7% vs. 22.1%, p=0.002) among deceased patients. A history of acute coronary syndrome, cerebrovascular accident, kidney failure and impaired left ventricular function were all more common among patients who died. Also, the use of angiotensin-converting enzyme inhibitors (48.1% vs.

289


Chapter 15

Table 1. Baseline characteristics of UCORBIO patients. n Age (mean (sd)) Sex (male %) BMI (mean (sd))

Overall

Alive

Deceased

1913

1836

77

p

63.7 (10.9)

63.3 (10.8)

72.5 (8.7)

73.9

73.9

72.7

0.921

27.2 (4.5)

27.2 (4.5)

26.6 (4.8)

0.281

<0.001

Diabetes (%)

22.7

22.1

37.7

0.002

Hypertension (%)

59.0

58.8

63.6

0.464

Hypercholesterolemia (%)

48.0

48.4

39.0

Smoking (%)

0.133 0.068

Active smoker

25.6

25.9

Ex smoker

26.5

26.0

19.5 37.7

Non smoker

47.9

48.1

42.9

History of ACS (%)

49.7

50.5

29.7

History of PCI (%)

30.6

30.3

37.7

0.215

History of CABG (%)

27.7

27.8

23.4

0.468

History of CVA (%)

10.5

10.1

18.2

0.038

History of PAD (%)

10.5

10.2

16.9

0.091

Kidney failure (%)

11.7

11.3

20.8

0.018

Normal

57.0

58.2

32.4

Mildly impaired

23.2

23.1

25.0

Impaired

12.1

11.7

20.6

EF (%)

0.002

<0.001

7.6

6.9

22.1

Aspirin (%)

Poor

58.1

58.0

59.7

Clopidogrel (%)

20.8

20.9

18.2

0.658

Beta-blocker (%)

54.5

54.3

59.7

0.409

ACE inhibitor (%)

34.8

34.3

48.1

0.018

Statin (%)

61.6

61.5

63.6

0.802

Diuretic (%)

28.6

27.4

55.8

<0.001

56.5

56.5

55.8

0.859

Coronary Angiography Indication (%) Stable CAD UAP Infarction Other

0.222 9.7

9.8

6.5

28.4

28.4

27.3

5.5

5.3

10.4

6.3

6.3

6.5 19.5

Severity of CAD (%) No CAD

0.566

Minor CAD

15.2

15.0

Single vessel disease

33.7

33.9

27.3

Multi vessel disease

44.9

44.8

46.8

Conservative

31.3

31.1

36.4

PCI

62.6

62.8

58.4

6.1

6.1

5.2

7.4 [3.7, 30.1]

7.1 [3.6, 27.4]

22.3 [5.1, 65.9]

<0.001

NT proBNP (median [IQR])

86.4 [33.4, 210.5]

83.0 [32.6, 199.8]

260.5 [89.5, 598.4]

<0.001

FU in years (median [IQR])

1.8 [1.0, 2.6]

1.7 [1.0, 2.5]

2.2 [1.8, 2.9]

<0.001

Procedure (%)

CABG hsTnI (median [IQR])

0.613

Abbreviations: BMI= body mass index, ACS= acute coronary syndrome, PCI= percutaneous coronary intervention, CABG= coronary artery bypass grafting, CVA= cerebrovascular accident, PAD= peripheral arterial disease, EF= ejection fraction, ACE= angiotensin-converting enzyme, UAP= unstable angina pectoris, FU= follow-up.

290


Hematological Parameters vs. hsTnI and NT pro-BNP

34.3%, p=0.018) and diuretics was significantly more common in deceased patients (55.8% vs. 27.4%, p<0.001). Median hsTnI levels were significantly higher in deceased patients (22.3 vs. 7.1, p<0.001) as were NT pro-BNP levels (260.5 vs. 83.0, p<0.001). During a median follow-up duration of 1.8 years, 77 deaths occurred, of which 29 were of cardiovascular origin. Hematological parameters A panel of seven hematological parameters was derived from the selection process as described in the methods section. This panel consisted of the leukocyte count, the reticulocyte mean corpuscular hemoglobin concentration (MCHCr), red blood cell (RBC) red fluorescence, the % of neutrophils, the % large (greater than 120fL) RBCs, the % of monocytes and the coefficient of variation (CV) of neutrophil complexity. Hazard ratios (HRs) of these parameters (all significantly and independently predictive, p<0.05) derived from a multivariable model are displayed in table 2. The multivariable adjusted HR of hsTnI was 1.00 [0.93-1.08] per 1000 ng/mL increase, p=0.945 and for NT pro-BNP per 1000 pmol/L increase it was 1.27 [1.14-1.42], p<0.001.

Table 2. Characteristics of hematological parameters included in the prognostic set. Value Alive Leukocyte count (10^9 cells/L)

Value Deceased

p-value difference

HR (95% CI)

p-value

2.32 [1.91, 2.86]

2.42 [1.94, 3.11]

0.224

1.25 (1.12-1.39)

<0.001

MCHCr (mmol/L)

15.73 [15.28, 16.22]

15.24 [14.83, 15.68]

<0.001

0.65 (0.50-0.86)

0.003

RBC red fluorescence (AU)

17.52 [17.01, 18.03]

17.76 [17.30, 18.12]

0.011

1.51 (1.15-1.97)

0.003

% neutrophils (%)

5.97 [5.32, 6.60]

6.29 [5.74, 7.00]

0.001

1.37 (1.07-1.75)

0.012

% large* RBCs (%)

0.71 [0.46, 1.08]

1.03 [0.71, 2.00]

<0.001

1.17 (1.03-1.34)

0.019

% monocytes (%)

3.18 [2.65, 3.81]

3.58 [2.58, 4.19]

0.075

1.28 (1.04-1.59)

0.023

Neutrophil complexity CV (%)

7.27 [6.70, 7.89]

7.31 [6.65, 8.31]

0.377

1.31 (1.03-1.67)

0.026

Medians and interquartile ranges of hematological parameters are shown for alive and deceased patients. The multivariable adjusted hazard ratios for all-cause mortality are shown for each 1-SD increase the hematological parameter and derived from a model containing: age, sex, diabetes, hypercholesterolemia, smoking status, indication for angiography, angiographic CAD severity, history of PCI, history of ACS, kidney failure, treatment following angiography and the other hematological parameters shown in the table. Abbreviations: MCHCr= reticulocyte mean corpuscular hemoglobin concentration, RBC= red blood cell, AU= arbitrary units, CV= coefficient of variation. *>120fL.

Improvement of all-cause mortality prediction First, we evaluated the additive predictive value of hsTnI, NT pro-BNP and hematological parameters to a clinical model (Table 3 top) for the prediction of all-cause mortality. hsTnI did not improve prediction of mortality in addition to the clinical model (AUCincrease, IDI and cNRI all non-significant). The AUC increased from 0.818 for the clinical model to 0.834 upon addition of NT pro-BNP (p=0.019), as shown in Figure 1. NT proBNP slightly improved discrimination (IDI 0.02 [0.00-0.06], p=0.040) but reclassification was not improved (cNRI 0.03 [-0.14-0.22], p=0.625). Combining hsTnI and NT pro-BNP

291


Chapter 15

also gave an AUC of 0.834, p=0.016 as compared to the clinical model alone, but the IDI and cNRI were both non-significant. Hematological parameters however significantly improved discrimination (IDI 0.07 [0.03-0.14], p<0.001) and reclassification (cNRI 0.37 [0.19-0.49], p<0.001). The AUC increased to 0.856, p<0.001. Then, we assessed whether hsTnI, NT pro-BNP or the combination could improve prediction in addition to the clinical model plus hematological parameters (Table 3 bottom; Figure 2). While the AUC increased slightly upon addition of NT pro-BNP (AUC increased from 0.856 to 0.863, p=0.061) and upon addition of the combination of hsTnI and NT pro-BNP (AUC 0.865, p=0.049) hsTnI, NT pro-BNP or the combination could not significantly improve discrimination or reclassification (IDIs and cNRIs all non-significant). Calibration plots of the models are provided in Supplemental Figure 1. The HosmerLemeshow goodness of fit p-value increased upon addition of hematological parameters but did not further increase upon addition of hsTnI and NT pro-BNP, indicating improved model fit (a non-significant p-value indicates a small difference between predicted and observed risk).

0.4

Sensitivity

0.6

0.8

1.0

Addition to Clinical Model

0.2

AUC Clinical = 0.818 AUC Clinical + hsTnI = 0.818, p=0.357 AUC Clinical + NT pro−BNP = 0.834, p=0.019 AUC Clinical + hsTnI + NT pro−BNP = 0.834, p=0.016

0.0

AUC Clinical + Hematology = 0.856, p<0.001

0.0

0.2

0.4

0.6

0.8

1.0

1−Specificity

Figure 1. ROC plots of hematological parameters, hsTnI and NT pro-BNP in addition to a clinical model for the prediction of all-cause mortality. ROC plots of the clinical model and of the clinical model extended with hsTnI, NT pro-BNP (or both) and hematological parameters for the prediction of all-cause mortality during 2 years of follow-up.

292


Hematological Parameters vs. hsTnI and NT pro-BNP

Table 3. Measures of improvement of all-cause mortality prediction. In addition to clinical characteristics Hematology

IDI

p-value IDI

cNRI

p-value cNRI

0.07 (0.03-0.14)

<0.001

0.37 (0.19-0.49)

<0.001

hsTnI

0.00 (0.00-0.00)

0.817

-0.07 (-0.16-0.20)

0.970

NT pro-BNP

0.02 (0.00-0.06)

0.040

0.03 (-0.14-0.22)

0.625

hsTnI + NT pro-BNP

0.02 (0.00-0.07)

0.066

0.02 (-0.13-0.20)

0.671

Hematology + hsTnI + NT pro-BNP

0.09 (0.05-0.17)

<0.001

0.44 (0.24-0.53)

0.007

In addition to clinical and hematological parameters hsTnI

0.01 (0.00-0.02)

0.113

0.16 (-0.18-0.31)

0.292

NT pro-BNP

0.01 (-0.01-0.06)

0.186

-0.08 (-0.25-0.14)

0.777

hsTnI + NT pro-BNP

0.02 (0.00-0.06)

0.093

-0.03 (-0.23-0.27)

1

Abbreviations: IDI= integrated discrimination improvement, cNRI= continuous net reclassification improvement.

0.4

Sensitivity

0.6

0.8

1.0

Addition to Clinical Model + Hematology

0.2

AUC Clinical + Hematology = 0.856 AUC Clinical + Hematology + hsTnI = 0.857, p=0.520 AUC Clinical + Hematology + NT pro−BNP = 0.863, p=0.061

0.0

AUC Clinical + Hematology + hsTnI + NT pro−BNP = 0.865, p=0.049

0.0

0.2

0.4

0.6

0.8

1.0

1−Specificity

Figure 2. ROC plots of hsTnI and NT pro-BNP in addition to a clinical model plus hematological parameters for the prediction of all-cause mortality. ROC plots of the clinical model plus hematological parameters and of that model extended with hsTnI, NT pro-BNP or both for the prediction of all-cause mortality during 2 years of follow-up.

293


Chapter 15

Internal Validation The predicted all-cause mortality risk based on hematological parameters, adjusted for clinical parameters, was grouped into four quartiles (Q1 to Q4) and adjusted survival curves are shown in Figure 3. The HR for Q2 vs. Q1 was not significant, for Q3 vs. Q1 it was 6.5 [2.0-21.8], p=0.002; the HR for Q4 vs. Q1 was 11.8 [3.6-38.1], p<0.001. Because models derived from a single dataset tend to result in overfitting and overoptimistic estimates, the fit of this model was internally validated by means of postestimation shrinkage. After shrinkage the HR for Q2 vs. Q1 remained non-significant, the HR of Q3 vs. Q1 was 3.4 [2.1-5.5], p=0.010 and the HR for Q4 vs. Q1 was 6.1 [4.3-8.8], p<0.001. Survival by Hematology Quartiles

Survival adjusted for Clinical Characteristics

1.00

0.95

0.90

Q1 Q2 Q3 Q4

0.85 0

100

200

300

400

500

600

700

800

900

1000

1100

Time (days)

Figure 3. Survival plot showing survival by quartiles of predicted risk based on hematology, adjusted for clinical characteristics. The predicted risk of all-cause mortality based on hematological parameters, adjusted for clinical parameters, was grouped into four quartiles (Q1 to Q4). The adjusted survival in these four groups was plotted using Cox regression analysis. The HR for Q2 vs. Q1 was not significant, for Q3 vs. Q1 it was 6.5 [2.0-21.8], p=0.002; the HR for Q4 vs. Q1 was 11.8 [3.6-38.1], p<0.001.

294


Hematological Parameters vs. hsTnI and NT pro-BNP

Discussion We show for the first time superiority of hematological parameters over hsTnI and NT proBNP for prediction of mortality in cardiovascular disease patients. These parameters are a combination of leukocyte and RBC characteristics. Both leukocyte and RBC characteristics have been proposed as strong predictors of mortality in several medical fields. In cardiology, a direct comparison of hematological parameters with established biomarkers has been made for leukocyte characteristics and high-sensitivity C-reactive protein (hsCRP). Ă“ Hartaigh et al.10 showed in a cohort of stable coronary angiography patients undergoing coronary angiography that neutrophil count is superior to hsCRP for prediction of cardiovascular mortality. In addition to their study, we also included RBC and platelet characteristics as potential predictors of mortality. As high platelet reactivity has been reported to be an independent predictor of future adverse events in myocardial infarction patients30, we were surprised that none of the platelet characteristics added any predictive value to a clinical model in our study. This might be due to our study population, which mainly comprises stable CAD patients, where the risk of acute fatal thrombotic events due to high platelet reactivity is lower than in myocardial infartion patients. We also did not find any predictive value of hsTnI levels in our cohort. High-sensitivity Troponin T (hsTnT) has been reported by several other groups as a potential predictor of mortality, for example in the setting of stroke31, after cardiac surgery32 and after elective coronary angiography5. Possibly, this finding is due to differential prognostic properties of hsTnI and hsTnT, as reported by de Antonio et al.33 High BNP levels are predictive of adverse events in a population of stable CAD patients6 and NT pro-BNP has been shown to predict mortality in an unstable CAD population.78 Our study is in line with others, showing that the predictive value of NT pro-BNP for mortality is superior to hsTnI.34,35 Nevertheless, hematological parameters in our study outperformed both NT pro-BNP and hsTnI for prediction of mortality. Apparently, blood cell characteristics provide more prognostic information than the cardiac-specific biomarkers hsTnI and/or NT pro-BNP. The hematological parameters included in our panel are of leukocyte (total leukocyte count, % neutrophils, % monocytes and neutrophil complexity CV) and RBC origin (MCHCr, RBC red fluorescence and % large RBCs). Leukocyte36 and neutrophil counts37 have previously been described as prognostic markers for mortality in a population-based cohort. Neutrophil count was not included in our analysis due to its collinearity (r=0.68, p<0.001) with % neutrophils, which was included as it had a stronger association with mortality. Hematology vs. Cardiac-specific biomarkers The reason that blood cells convey more accurate prognostic information than CAD biomarkers could be related to the organ-specificity of hsTnI and NT pro-BNP. The endpoint in our study was all-cause mortality, meaning that not all deaths were due to cardiovascular disease or its consequences per se. Hematological parameters could

295


Chapter 15

provide a whole-body overview of an individual’s health status and subsequent prognosis. However, when we checked for cardiovascular death hematological parameters again outperformed hsTnI and NT pro-BNP (supplemental figure 2). Note however, that the number of cardiovascular deaths was low (n = 29) and these data should thus be interpreted cautiously. Hematological parameters, cardiovascular disease and mortality Monocytes are known to be causally involved in plaque build-up and plaque destabilization38 and monocytes have been reported to correlate with the presence of cardiovascular disease and with higher levels of interleukin-6, a marker of inflammation.39 The prognostic role of monocytes has been established before.12,40 Neutrophils have been reported to be causally involved in atherosclerotic plaque formation; Zernecke et al. demonstrated that plaque formation was attenuated upon depletion of neutrophils in mice 41 and have been shown to be present in coronary artery autopsy specimens of patients who died of myocardial infarction.42 In addition to sheer neutrophil numbers we found additional predictive value of the variation in neutrophil complexity. A higher variation in neutrophil complexity can be viewed as an indication of a ‘left shift’ or neutrophil activation as can be observed with infections.43 Possibly, a slight left shift concurs with chronic inflammation due to atherosclerotic disease or it might be a marker of subclinical ischemic events. Also, morphological changes in neutrophils have been described in an acute myocardial infarction in a porcine model. 44 Neutrophil numbers and morphology are apparently influenced by atherosclerosis (or vice versa41) and ischemia. The exact mechanisms relating these neutrophil characteristics to increased mortality risk remain to be elucidated. Recently, RBC characteristics have entered the stage of mortality prediction. Especially red cell distribution width45–47 (RDW), a marker of the variation in RBC volume, has been proposed as a powerful risk indicator of mortality. RBC volume is inversely related with its age; young cells are largest, senescent cells are smaller.48 In the current study we found RBC characteristics fitting with a higher proportion of young erythrocytes (closely related to RDW, r= 0.21, p<0.001), to be independently predictive of mortality, in particular the % of large (i.e. young) RBCs. RBC red fluorescence, associated with worse survival in our study, is a marker of RNA content of reticulocytes. Early reticulocytes contain high amounts of RNA, which is lost upon maturation.49 Markers of an immature RBC population with a low reticulocyte hemoglobin concentration were thus related to a higher risk of mortality in coronary angiography patients. Perspectives While we have internally validated the value of hematological parameters in the prediction of all-cause mortality in our cohort, our findings warrant external validation. Possibly, hematological parameters serve a wider use than coronary angiography patients alone. Cohorts comprising patients with other types of cardiac disease such as heart failure might be considered.

296


Hematological Parameters vs. hsTnI and NT pro-BNP

While information on the cause of death was available, we deemed the number of cardiovascular deaths too low (n=29) to perform thorough analysis with multiple biomarkers. Conclusion The predictive value of hematological parameters was superior to that of the established CAD biomarkers hsTnI or NT pro-BNP alone or in combination. hsTnI and NT pro-BNP also did not add significantly to clinical and hematological parameters. Panels composed of leukocyte and erythrocyte parameters performed best. Hence, readily available hematological parameters may provide a useful addition to current risk prediction algorithms for all-cause mortality in CAD patients.

297


Chapter 15

References 1.

Keller T, Zeller T, Peetz D, Tzikas S, Roth A, Czyz E, Bickel C, Baldus S, Warnholtz A, Fröhlich M, Sinning CR, Eleftheriadis MS, Wild PS, Schnabel RB, Lubos E, Jachmann N, Genth-Zotz S, Post F, Nicaud V, Tiret L, Lackner KJ, Münzel TF, Blankenberg S. Sensitive troponin I assay in early diagnosis of acute myocardial infarction. N Engl J Med. 2009;361:868–877.

2.

Hobbs FDR, Davis RC, Roalfe AK, Hare R, Davies MK, Kenkre JE. Reliability of N-terminal pro-brain natriuretic peptide assay in diagnosis of heart failure: cohort study in representative and high risk community populations. 2002.

3.

Grodin JL, Neale S, Wu Y, Hazen SL, Tang WHW. Prognostic Comparison of Different Sensitivity Cardiac Troponin Assays in Stable Heart Failure. Am J Med. 2015;128:276–282.

4.

Bonaca MP, O’Malley RG, Murphy SA, Jarolim P, Conrad MJ, Braunwald E, Sabatine MS, Morrow DA. Prognostic performance of a high-sensitivity assay for cardiac troponin I after non-ST elevation acute coronary syndrome: Analysis from MERLIN-TIMI 36. Eur Hear J Acute Cardiovasc Care. 2014;

5.

Hochholzer W, Valina CM, Stratz C, Amann M, Schlittenhardt D, Büttner HJ, Trenk D, Neumann F-J. Highsensitivity cardiac troponin for risk prediction in patients with and without coronary heart disease. Int J Cardiol. 2014;176:444–9.

6.

Schnabel R, Lubos E, Rupprecht HJ, Espinola-Klein C, Bickel C, Lackner KJ, Cambien F, Tiret L, Münzel T, Blankenberg S. B-type natriuretic peptide and the risk of cardiovascular events and death in patients with stable angina: Results from the AtheroGene study. J Am Coll Cardiol. 2006;47:552–558.

7.

Omland T, Persson A, Ng L, O’Brien R, Karlsson T, Herlitz J, Hartford M, Caidahl K. N-terminal pro-B-type natriuretic peptide and long-term mortality in acute coronary syndromes. Circulation. 2002;106:2913–2918.

8. James SK, Lindahl B, Siegbahn A, Stridsberg M, Venge P, Armstrong P, Barnathan ES, Califf R, Topol EJ, Simoons ML, Wallentin L. N-terminal pro-brain natriuretic peptide and other risk markers for the separate prediction of mortality and subsequent myocardial infarction in patients with unstable coronary artery disease: A GUSTO-IV Substudy. Circulation. 2003;108:275–281. 9.

Yildirir A, Acikel S, Ertan C, Aydinalp A, Ozin B, Muderrisoglu H. Value of peri-procedural B-type natriuretic peptide levels in predicting cardiac events after elective percutaneous coronary intervention. Acta Cardiol. 2008;63:47–52.

10. ó Hartaigh B, Bosch J a., Thomas GN, Lord JM, Pilz S, Loerbroks A, Kleber ME, Grammer TB, Fischer JE, Boehm BO, März W. Which leukocyte subsets predict cardiovascular mortality? From the LUdwigshafen RIsk and Cardiovascular Health (LURIC) Study. Atherosclerosis. 2012;224:161–169. 11. Lee C Do, Folsom AR, Nieto FJ, Chambless LE, Shahar E, Wolfe DA. White blood cell count and incidence of coronary heart disease and ischemic stroke and mortality from cardiovascular disease in African-American and White men and women: atherosclerosis risk in communities study. Am J Epidemiol. 2001;154:758–64. 12. Kato A, Takita T, Furuhashi M, Maruyama Y, Kumagai H, Hishida A. Blood monocyte count is a predictor of total and cardiovascular mortality in hemodialysis patients. Nephron Clin Pract. 2008;110:c235–43. 13. Azab B, Torbey E, Hatoum H, Singh J, Khoueiry G, Bachir R, McGinn JT, McCord D, Lafferty J. Usefulness of red cell distribution width in predicting all-cause long-term mortality after non-ST-elevation myocardial infarction. Cardiology. 2011;119:72–80. 14. Benedetto U, Angeloni E, Melina G, Pisano C, Lechiancole A, Roscitano A, Pooley M, Comito C, Codispoti M, Sinatra R. Red blood cell distribution width predicts mortality after coronary artery bypass grafting. Int J Cardiol. 2013;165:369–371.

298


Hematological Parameters vs. hsTnI and NT pro-BNP

15. Tonelli M, Sacks F, Arnold M, Moye L, Davis B, Pfeffer M. Relation between red blood cell distribution width and cardiovascular event rate in people with coronary disease. Circulation. 2008;117:163–168. 16. Müller R, Mellors I, Johannessen B, Aarsand AK, Kiefer P, Hardy J, Kendall R, Scott CS. European multi-center evaluation of the Abbott Cell-Dyn sapphire hematology analyzer. Lab Hematol. 2006;12:15–31. 17. Ten Berg MJ, Huisman A, van den Bemt PML a, Schobben AF a M, Egberts ACG, van Solinge WW. Linking laboratory and medication data: new opportunities for pharmacoepidemiological research. Clin Chem Lab Med. 2007;45:13–9. 18. Gijsberts CM, Seneviratna A, de Carvalho LP, den Ruijter HM, Vidanapthirana P, Sorokin V, Stella P, Agostoni P, Asselbergs FW, Richards AM, Low AF, Lee C-H, Tan HC, Hoefer IE, Pasterkamp G, de Kleijn DP V., Chan MY. Ethnicity Modifies Associations between Cardiovascular Risk Factors and Disease Severity in Parallel Dutch and Singapore Coronary Cohorts. PLoS One. 2015;10:e0132278. 19. Harris PJ, Behar VS, Conley MJ, Harrell FE, Lee KL, Peter RH, Kong Y, Rosati R a. The prognostic significance of 50% coronary stenosis in medically treated patients with coronary artery disease. Circulation. 1980;62:240–248. 20. Lam SW, Leenen LPH, Van Solinge WW, Hietbrink F, Huisman A. Evaluation of hematological parameters on admission for the prediction of 7-day in-hospital mortality in a large trauma cohort. Clin Chem Lab Med. 2011;49:493–499. 21. Groeneveld KM, Heeres M, Leenen LPH, Huisman A, Koenderman L. Immunophenotyping of posttraumatic neutrophils on a routine haematology analyser. Mediators Inflamm. 2012;2012. 22. Lam SW, Leenen LPH, van Solinge WW, Hietbrink F, Huisman A. Comparison between the prognostic value of the white blood cell differential count and morphological parameters of neutrophils and lymphocytes in severely injured patients for 7-day in-hospital mortality. Biomarkers. 2012;17:642–647. 23. Pavlou M, Ambler G, Seaman SR, Guttmann O, Elliott P, King M, Omar RZ. How to develop a more accurate risk prediction model when there are few events. Bmj. 2015;h3868. 24. Dunkler D, Heinze G. Package “shrink ” [Internet]. CRAN. 2014 [cited 2015 Oct 14];1–15. Available from: https://cran.r-project.org/web/packages/shrink/shrink.pdf 25. Uno H, Tian L, Cai T, Kohane IS, Wei LJ. A unified inference procedure for a class of measures to assess improvement in risk prediction systems with survival data. Stat Med. 2013;32:2430–2442. 26. Pencina MJ, D’Agostino RB, Steyerberg EW. Extensions of net reclassification improvement calculations to measure usefulness of new biomarkers. Stat Med. 2011;30:11–21. 27. Sniderman AD, D’Agostino RB, Pencina MJ. The Role of Physicians in the Era of Predictive Analytics. JAMA. 2015;314:25–6. 28. Allaire J. RStudio: Integrated development environment for R. 2012; 29. R Core Team. R: A Language and Environment for Statistical Computing. Vienna, Austria: 2014. 30. Campo G, Valgimigli M, Gemmati D, Percoco G, Tognazzo S, Cicchitelli G, Catozzi L, Malagutti P, Anselmi M, Vassanelli C, Scapoli G, Ferrari R. Value of platelet reactivity in predicting response to treatment and clinical outcome in patients undergoing primary coronary intervention: insights into the STRATEGY Study. J Am Coll Cardiol. 2006;48:2178–85. 31. Maoz A, Rosenberg S, Leker RR. Increased High-Sensitivity Troponin-T Levels Are Associated with Mortality After Ischemic Stroke. J Mol Neurosci. 2015; 32. Lurati Buse GAL, Bolliger D, Seeberger E, Kasper J, Grapow M, Koller MT, Seeberger MD, Filipovic M. Troponin T and B-type natriuretic peptide after on-pump cardiac surgery: prognostic impact on 12-month mortality and major cardiac events after adjustment for postoperative complications. Circulation. 2014;130:948–57. 33. De Antonio M, Lupón J, Galán A, Vila J, Zamora E, Urrutia A, Díez C, Coll R, Altimir S, Bayes-Genis A. Headto-head comparison of high-sensitivity troponin T and sensitive-contemporary troponin I regarding heart failure risk stratification. Clin Chim Acta. 2013;426:18–24.

299


Chapter 15

34. Nigro N, Wildi K, Mueller C, Schuetz P, Mueller B, Fluri F, Christ-Crain M, Katan M. BNP but Not s-cTnln Is Associated with Cardioembolic Aetiology and Predicts Short and Long Term Prognosis after Cerebrovascular Events. PLoS One. 2014;9:e102704. 35. Mayer O, Seidlerová J, Bruthans J, Vaněk J, Černá L, Wohlfahrt P, Filipovský J, Cífková R, Windrichová J, Topolčan O. The predictive potential of asymptomatic mild elevation of cardiac troponin I on mortality risk of stable patients with vascular disease. Clin Biochem. 2015;48:353–357. 36. Karino S, Willcox BJ, Fong K, Lo S, Abbott R, Masaki KH. Total and differential white blood cell counts predict eight-year incident coronary heart disease in elderly Japanese-American men: The Honolulu Heart Program. Atherosclerosis. 2014;238:153–158. 37. Guasti L, Dentali F, Castiglioni L, Maroni L, Marino F, Squizzato A, Ageno W, Gianni M, Gaudio G, Grandi AM, Cosentino M, Venco A. Neutrophils and clinical outcomes in patients with acute coronary syndromes and/ or cardiac revascularization: A systematic review on more than 34,000 subjects. Thromb Haemost. 2011;106:591–599. 38. Ghattas A, Griffiths HR, Devitt A, Lip GYH, Shantsila E. Monocytes in coronary artery disease and atherosclerosis: where are we now? J Am Coll Cardiol. 2013;62:1541–51. 39. Compté N, Bailly B, De Breucker S, Goriely S, Pepersack T. Study of the association of total and differential white blood cell counts with geriatric conditions, cardio-vascular diseases, seric IL-6 levels and telomere length. Exp Gerontol. 2015;61:105–12. 40. Madjid M, Fatemi O. Components of the complete blood count as risk predictors for coronary heart disease: in-depth review and update. Tex Heart Inst J. 2013;40:17–29. 41. Zernecke A, Bot I, Djalali-Talab Y, Shagdarsuren E, Bidzhekov K, Meiler S, Krohn R, Schober A, Sperandio M, Soehnlein O, Bornemann J, Tacke F, Biessen EA, Weber C. Protective role of CXC receptor 4/CXC ligand 12 unveils the importance of neutrophils in atherosclerosis. Circ Res. 2008;102:209–217. 42. Naruko T, Ueda M, Haze K, Van der Wal AC, Van der Loos CM, Itoh A, Komatsu R, Ikura Y, Ogami M, Shimada Y, Ehara S, Yoshiyama M, Takeuchi K, Yoshikawa J, Becker AE. Neutrophil infiltration of culprit lesions in acute coronary syndromes. Circulation. 2002;106:2894–2900. 43. Seebach JD, Morant R, Rüegg R, Seifert B, Fehr J. The diagnostic value of the neutrophil left shift in predicting inflammatory and infectious disease. Am J Clin Pathol. 1997;107:582–591. 44. Van Hout GPJ, De Jong R, Teuben M, Nijhoff F, Duckers H, Koenderman L, Stella P, Van Solinge W, Pasterkamp G, Hoefer I. P159 * Early changes in neutrophil morphology predict myocardial damage after myocardial infarction. Cardiovasc Res. 2014;103:S28–S28. 45. Tsuboi S, Miyauchi K, Kasai T, Ogita M, Dohi T, Miyazaki T, Yokoyama T, Kojima T, Yokoyama K, Kurata T, Daida H. Impact of red blood cell distribution width on long-term mortality in diabetic patients after percutaneous coronary intervention. Circ J. 2013;77:456–461. 46. Núñez J, Núñez E, Rizopoulos D, Miñana G, Bodí V, Bondanza L, Husser O, Merlos P, Santas E, Pascual-Figal D, Chorro FJ, Sanchis J. Red blood cell distribution width is longitudinally associated with mortality and anemia in heart failure patients. Circ J. 2014;78:410–8. 47. Perlstein TS, Weuve J, Pfeffer M a, Beckman J a. Red blood cell distribution width and mortality risk in a community-based prospective cohort. Arch Intern Med. 2009;169:588–594. 48. Patel HH, Patel HR, Higgins JM. Modulation of red blood cell population dynamics is a fundamental homeostatic response to disease. Am J Hematol. 2015;90:422–428. 49. Riley RS, Ben-Ezra JM, (ASCP) M, Tidwell A. Reticulocyte Enumeration: Past & Present. Lab. Med. 2001;32:599– 608.

300


Hematological Parameters vs. hsTnI and NT pro-BNP

Supplemental Clinical Model 0.15

Observed risk

0.10

● ●

0.05

Chi−square = 29.1 p−value Goodness of Fit = 0.047

● ● ●

0.00

●● ●●

0.00

0.05

Predicted risk

0.10

0.15

Clinical Model + Hematology 0.15

Observed risk

0.10

● ●

0.05

Chi−square = 18.7 p−value Goodness of Fit = 0.409

● ● ● ●

0.00

● ●●● ● ●●

0.00

0.05

Predicted risk

0.10

0.15

Clinical Model + Hematology + hsTnI + NT pro−BNP 0.15

Observed risk

0.10

0.05

● ●●●● ● ●●

0.00

Chi−square = 23.6 p−value Goodness of Fit = 0.167

0.00

0.05

Predicted risk

0.10

0.15

Supplemental figure 1. Calibration plots of clinical model, clinical plus hematology and clinical, hematology and cardiac biomarkers. The predicted risk of and observed all-cause mortality are plotted for the clinical model, clinical model plus hematology and for the clinical model plus hematology and hsTnI and NT pro-BNP. The goodness of fit improved upon addition of hematological parameters, but did not further increase upon addition of hsTnI and NT pro-BNP.

301


Chapter 15

0.4

Sensitivity

0.6

0.8

1.0

Addition to Clinical Model Cardiovascular Death

0.2

AUC Clinical = 0.845 AUC Clinical + hsTnI = 0.844, p=0.099 AUC Clinical + NT pro−BNP = 0.869, p=0.074 AUC Clinical + hsTnI + NT pro−BNP = 0.868, p=0.096

0.0

AUC Clinical + Hematology = 0.882, p=0.004

0.0

0.2

0.4

0.6

0.8

1.0

1−Specificity

Supplemental figure 2. ROC plots of hematological parameters, hsTnI and NT pro-BNP in addition to a clinical model for the prediction of cardiovascular death. ROC plots of the clinical model and of the clinical model extended with hsTnI, NT pro-BNP (or both) and hematological parameters for the prediction of cardiovascular death during 2 years of follow-up.

302


303



Summary and Discussion

Chapter 16 Summary and Discussion


Chapter 16

Cardiovascular disease is still the leading cause of death around the globe, despite major advances in prevention and treatment.1 In particular, coronary artery disease (CAD) has become a disease of men and women of all ethnicities. While historically, CAD was a disease of White men, Westernization of developing parts of the world transformed the Western epidemic into a global one.2 Especially in Asia3 with billions of inhabitants, large numbers of CAD deaths are projected. In the East-Asian region an increase in cardiovascular deaths of 20% is expected, contrasting with expected stable numbers in the European region.4 Also, the exclusivity of men has disappeared in the field of CAD. Due to increasing prevalence of cardiovascular risk factors in women5, CAD has become highly prevalent in women as well.6 In contrast to declining mortality of CAD in men7, mortality rates in women are merely stabilizing.8 Currently, more women than men die of CAD.9 Profound differences between ethnicities and between the sexes have been described for risk factor burden, CAD phenotype10,11 and outcome12–14 after events related to CAD. In order to effectively tackle this forthcoming epidemic affecting men and women of all ethnic groups, insight in the ethnicity- and sex-specific relation of cardiovascular risk factors with CAD, ethnicity- and sex-specific specific characteristics of CAD and outcomes is desperately needed. Accurate estimation of the effect of risk factors and total individual risk of cardiovascular events is key when applying prevention strategies. Risk factors with the largest impact on total risk should be the main focus of local prevention strategies and individuals with the highest estimated risk deserve the most aggressive preventive treatment.4 In this thesis we focused on differences in CAD risk factors, phenotype and outcome between Whites and Asians and men and women. Furthermore we evaluated hematological parameters as potential patient characteristics that could improve risk estimation.

PART ONE | Ethnicity First, in chapter 2 we reviewed the available literature for indications of ethnicitydependent differences CAD risk, reflected by levels of established CAD biomarkers in the general population. Doing this, we found a strikingly unfavorable biomarker pattern among South Asians, as compared to Whites. For instance, levels of C-reactive protein (CRP)15 and insulin levels were 1.22 and 1.33 times as high as in Whites, respectively. Indicating that low-grade inflammation and insulin resistance are more distinct in South Asians than Whites in the general population. Both CRP16 and insulin17 levels are associated with increased risk of cardiovascular events and mortality. Possibly, the relation of impaired glucose handling with functional outcome is even stronger in South Asians than Whites.18

306


Summary and General Discussion

In chapter 3 we evaluated the ethnicity-specific relation of the Framingham19 risk factors with carotid intima-media thickness (a measure of subclinical atherosclerosis and indicator of cardiovascular risk20) and with the occurrence of cardiovascular events in individuals without a history of cardiovascular disease. In this study we found small, yet significant differences in the strengths of risk factors. Similar differences were reported by Woodward et al. comparing Asian countries with Australia and New-Zealand.21 These seemingly small differences may have a large impact when applied on a global scale. Our findings indicate that ethnicity-specific prediction models may be preferred over non-ethnicity-specific models. Then in chapter 4, we looked into the effect of risk factors on the severity of CAD in patients undergoing coronary angiography for suspected CAD. For this study we used the Dutch-Singaporean United Coronary Biobank (the UNICORN) study, which was set up for the purpose of evaluating White-Asian differences in CAD and CAD biomarkers.22 In this study we find that the angiographic severity of CAD was worst in Malay, followed by Chinese and Indians. The least severe CAD was found in Whites. The observed ethnic differences were not abolished after adjustment for baseline differences. Also, we found that diabetes and male sex had a significantly stronger association with severe CAD in Chinese than Whites. Adjusted survival after coronary angiography was worst in Malays, fitting the scarce literature on this topic.12,23 Our results again imply that the effects of risk factors on CAD differ by ethnicity and that ethnicity-specific (secondary) prevention strategies should be considered. When looking into the severity of CAD in more detail in chapter 5, we find similar results to those described in chapter 4. The quantified severity of CAD (SYNTAX score24) in stable CAD patients undergoing percutaneous coronary intervention (PCI) was highest in Malays and Indians, followed by Chinese and lowest in Whites. In acute myocardial infarction patients undergoing PCI, however we found a surprisingly high SYNTAX score in Chinese. While Chinese in the general population typically have a lower risk of CAD25, apparently when myocardial infarction does occur underlying CAD is more extensive. However, survival of Chinese myocardial infarction patients is comparable to Whites, who had the least extensive CAD. Prognosis of Indians and Malays might be worse after myocardial infarction requiring PCI. Blood-derived biomarkers can be used as an estimator of risk of cardiovascular events.26 Most of these biomarkers associate with future risk via their underlying association with the severity of CAD. However these associations are almost exclusively studied in White CAD patients. Chapter 6 describes the ethnicity-specific association of biomarker levels with the severity of CAD (SYNTAX score). Within the UNICORN cohort we evaluated the ethnicity-specific association of established CAD biomarkers: NT pro-brain natriuretic peptide (NTproBNP)27, high-sensitivity (hs) CRP28, Cystatin C29, myeloperoxidase (MPO)30 and hsTroponin31 I (hsTnI). After adjustment for baseline differences NTproBNP levels were

307


Chapter 16

strikingly higher in Malays than in Whites or Chinese. MPO levels were markedly lower in Indians than in the other ethnic groups and hsTnI was highest in Malays. Also, for NTproBNP and hsTnI stronger associations with SYNTAX score were found in Malays than in Whites, meaning that at a given biomarker level SYNTAX scores are higher in Malays than Whites. Biomarker cut-off values for severe CAD differed over a great range among the ethnic groups, implying that ethnicity-specific biomarker cut-offs should be considered for prediction purposes. Unfortunately we were underpowered to perform analyses evaluating the prognostic value of these biomarkers across the ethnic groups. Such an analysis would greatly aid our understanding of the value of CAD biomarkers in less studied ethnic groups. Notably, in our study most biomarkers performed poorly in all ethnic groups for predicting severe CAD, illustrating the need for more potent CAD biomarkers.32 During the past decades, with improving survival of acute CAD events (i.e. myocardial infarction), heart failure (HF) has become a major source of morbidity.33 The duration of QRS (QRSd) on electrocardiography is known to relate strongly with the left ventricular ejection fraction (EF)34 and prognosis of HF patients.35 In chapter 7 we describe an interethnic comparison of QRSd between Asian and White HF patients with preserved EF (HFPEF) and reduced EF (HFREF). We find a remarkable and previously unknown modifying effect of ethnicity on the relation of EF with QRSd. Among HFPEF patients QRSd is shorter among Asians than Whites, while among HFREF patients QRS prolongation is more severe in Asians than Whites. The most probable explanation for this difference is the higher EF in healthy Asians than Whites as described in the EchoNormal36 study. While the relation with EF thus differed, QRSd was related to a worse outcome in both Asians and Whites with HFREF. However, higher cut-offs for increased risk of adverse events might need to be considered for Asians. As QRSd is one of the criteria for implantation of cardiac resynchronization therapy devices37, future studies should evaluate whether different cut-offs for QRSd should be implemented for Asians than for Whites. The explanation for ethnic differences in CAD is likely to be multifactorial. Life style habits that we could not adjust for might differ between Asians and Whites. For instance, the World Health Organization reported that in the Netherlands physical inactivity, a risk factor for cardiovascular disease3, is relatively rare (44%), while in Singapore this is a large problem (75% of the population has insufficient physical activity).38 Also, high-risk dietary habits, such as high fat39 and salt40 intake might differ between Asians and Whites and among Asian ethnic groups. But in addition to life style differences, genetic differences41,42 are probably also involved. It might even be so, that the combination of genetic predispositions and high-risk life style habits lead to the large observed differences. A study by Chow et al.43 showed a stronger association of glucose and cholesterol levels with carotid intima-media thickness in South Asians than in Whites, implying that arterial vasculature of South Asians is more susceptible to disturbances in lipid levels and glucose hemostasis than in Whites.

308


Summary and General Discussion

PART TWO | Sex differences While more women than men die of CAD9, women are underrepresented in cardiovascular research.44 Therefore, in chapter 8 we directly compared women with men, who underwent coronary angiography. In order to compare fairly, we only considered women and men who presented with stable complaints of CAD. The severity of CAD on the coronary angiograms of these women and men was quantified by means of the SYNTAX score and women were found to have less severe CAD, when adjusted for baseline characteristics. Moreover, hsCRP levels were higher in women than in men and hsTnI levels were lower, adjusted for baseline differences. The relationship of established CAD biomarkers with CAD severity however, did not differ between women and men. Higher hsCRP levels could be a reflection of underlying microvascular CAD, causing the chest pain complaints in women45 and known to be associated with inflammation of the coronary endothelium.46–48 Lower hsTnI levels might be a consequence of a smaller myocardial mass in women (left ventricular mass 191gr at age 50) than in men (263gr at age 50).36 Hereby, at a given proportion of myocardial tissue at jeopardy of ischemia, absolutely a larger myocardial mass is at risk in men, resulting in greater release of hsTnI. In chapter 9 we provide an up-to-date overview of the prognosis of women and men undergoing coronary angiography in the Netherlands. We found that prognosis of women who present with myocardial infarction or with triple vessel disease was markedly poorer than for men. Women, on a general population scale, have a lower chance of developing severe CAD events (e.g. myocardial infarction).49 However, the women who do develop severe CAD thus have poorer prognosis than men, suggesting that an indexevent bias is at play here.50 This means that the risk conveyed by a certain patient characteristic - in this case female sex - changes from protective to harmful after the occurrence of an initial event. Additionally we found to our surprise, that Dutch women with stable CAD or less severe CAD (no or minor CAD on the coronary angiogram) do not have a poorer prognosis than men or than women from the general Dutch population, as had been reported in the US.51 Possibly, higher rates of physical inactivity, obesity and diabetes (risk factors for microvascular CAD or microvascular dysfunction52) in US women as compared to Dutch women, contribute to this difference in prognosis between the Netherlands and the US.4 Then, in chapter 10 we focused on a whole different topic, namely the health-related quality of life (HRQOL) in women and men with CAD. Women reported lower HRQOL across all indications of coronary angiography and regardless of the severity of CAD. However, in chapter 11 we found that a decrease in HRQOL conveyed less risk of adverse events and mortality in women than in men. In other words, women experience poor HRQOL while objectively less adverse events happen to them. Potential reasons for this paradox which is also reflected in health service use53, involve the (expected) feminine social role compared to the masculine role.54 It is postulated by Norris et al.54 that the socially constructed “gender” as opposed to the biological “sex” might be a larger

309


Chapter 16

contributor to the observed differences in HRQOL and the discrepant conferred risk of adverse events. A differential stress hormone response to stressors between masculine and feminine gender roles55 further illustrates the intricate interaction between sex and gender and the way people process a stressor such as complaints of CAD. Lundberg described that men and women react differently to stressors, but when women adopt a masculine lifestyle (e.g. working in a masculine environment) they also adopt a response to stressors similar to males. Factors able to improve HRQOL in women, apart from improving physical complaints56, should be sought after in order to reduce psychological suffering and subsequent health care expenditure in women. We did not find evidence that female-specific risk factors associate with low HRQOL in women.

PART THREE | Risk prediction In the first chapter of this part of the thesis, chapter 12, we describe the prognostic difference between people who did and who did not return their HRQOL questionnaire. Logically, data from questionnaires that are never returned are left unanalyzed and disregarded when deriving conclusions. We show in our analysis that the patient group that never returned their questionnaire had a very poor prognosis, even beyond people who reported a poor HRQOL. Questionnaire non-responsiveness can thus be recognized as an indicator of high risk of mortality among coronary angiography patients. This phenomenon can have large consequences for questionnaire-driven research, rendering it non-generalizable to the patient group with the worst prognosis. In the last three chapters of this thesis easily measurable biological patient characteristics are considered as biomarkers for risk prediction. First, given the inflammatory nature of atherosclerosis57, in chapter 13 we evaluated the value of leukocyte subtypes and ratios as indicators of risk of adverse events and mortality. We found that of the tested leukocyte characteristics, the monocyte-to-lymphocyte ratio (MLR) performed best and could significantly improve the estimation of risk of adverse events and mortality. Then, motivated by increasing evidence on the prognostic value of red blood cell (RBC) characteristics58, we extended our search to the best hematological parameter. In this study we considered leukocyte, red blood cell and platelet characteristics that are routinely measured by hematology analyzers.59 When any hematological parameter is requested, modern hematology analyzers measure a whole blood differential. However, only the requested parameter is reported back to the clinician. Thus, much more data is available than known to the treating physician. In our center, the University Medical Center Utrecht, these data are stored in the Utrecht Patient Oriented Database60 and made available for research. Our analysis in chapter 14 showed us that panels of two to five hematological parameters could significantly improve risk prediction on top of

310


Summary and General Discussion

clinical risk estimation in coronary angiography patients. Five out of six panels consisted of a combination of leukocyte and RBC characteristics. Surprisingly, platelet characteristics such as mean platelet volume were not included in any of our panels, while they have been reported to be predictive of future events by other groups.61 Possibly, the predictive value of platelet characteristics is lost when adjusted for other hematological parameters such as leukocyte and RBC characteristics. In order to judge whether hematological parameters can actually better improve mortality risk estimation than established biomarkers of CAD, that are known to predict risk of adverse events in CAD patients: hsTnI62 and NTproBNP63, we performed a direct comparison in chapter 15. This comparison showed that NTproBNP could slightly improve risk estimation, but hsTnI did not improve risk estimation at all. Hematological parameters on the other hand, as already demonstrated in chapter 14, significantly improved risk estimation. HsTnI and NTproBNP could not further improve risk estimation when added to a model containing a panel of hematological markers, suggesting that most prognostic information is conferred by the hematological parameters. Hematological characteristics (e.g. hemoglobin level) are requested on a very regular basis. Consequently, for the majority of patients hematological data will be available and risk prediction improvement can be readily performed. The first sparks of evidence suggesting hematological parameters as potent risk predictors already appeared decades ago.64 While these analyses were not as extended as ours (improvement of risk estimation was not assessed), they already suggested clinical potential. Then, how come these apparently valuable prognostic biomarkers never became implemented into clinical practice? Easy and inexpensive applications might be considered. A model only taking age, sex and hematological parameters into account is almost as accurate at estimating risk as a model including all clinical characteristics (unpublished data). Therefore, a simple equation embedded in the hematology analyzer, which can obtain the age and sex of a patient through the electronic patient files, might prove to be of clinical value. Perspectives The last decade, the interest in tailored medicine has surged due to increasing numbers of people living with CAD and the increasing possibilities in medical care. The CAD population forms a heterogeneous patient group with diverse disease characteristics and differing risks of future adverse evens. In order to avoid over and under treatment, identifying the right individuals for the right therapies and follow-up is key in modern medicine. The results from this thesis stress the need for cardiovascular research stratified by ethnicity and sex. By gaining insight in the ethnicity and sex-specific characteristics and risks of CAD, (secondary) prevention strategies can be focused on the right individuals. Indian and Malay CAD patients, for instance, suffer from very severe CAD and particularly Malay CAD patients have a high mortality risk after coronary angiography. One might consider intensifying follow-up and secondary prevention in Malay coronary angiography

311


Chapter 16

patients. Genetic research will probably further elucidate ethnic differences in CAD phenotypes. Multi-ethnic coronary biobank initiatives, like the Dutch-Singaporean UNICORN study65, provide unique opportunities for such purposes. Regarding sex differences, women who undergo coronary angiography have less severe CAD than men. But women who had multi-vessel disease or who presented with myocardial infarction had a poorer prognosis than men. Stricter follow-up and secondary prevention regimens might benefit these women. Research on female-specific risk factors, pregnancy-related vascular disorders (e.g. preeclampsia) and research on the sex chromosomes66 might extend our knowledge on sex differences in the development, risk and phenotype of CAD. In addition to ethnicity and sex, improving risk prediction can be effectuated by addition of biomarkers. In this thesis, we critically evaluated the value of readily available hematological parameters and found that their clinical value could be substantial. Simple software additions to existing hematology analyzers can theoretically provide a mortality risk for every measured blood sample. Ways in which such a risk indicator should be incorporated into the daily clinic need to be evaluated. Possible applications might be the selection of high-risk patients for additional therapy with novel therapeutics, such as the LDL-cholesterol lowering PCSK9 inhibitors67. On the other hand, longer follow-up intervals might be considered in low-risk individuals.

312


Summary and General Discussion

References 1.

World Health Organization. World Health Statistics 2012. Geneva, Switzerland: 2012.

2.

Mathers CD, Loncar D. Projections of global mortality and burden of disease from 2002 to 2030. PLoS Med. 2006;3:e442.

3.

Alwan A. Global status report on noncommunicable diseases 2010. 1st ed. Geneva, Switzerland: 2011.

4.

Mendis S, Puska P, Norrving B. Global Atlas on cardiovascular disease prevention and control. 2011.

5.

Mosca L, Grundy SM, Judelson D, King K, Limacher M, Oparil S, Pasternak R, Pearson TA, Redberg RF, Smith SC, Winston M, Zinberg S. Guide to Preventive Cardiology for Women.AHA/ACC Scientific Statement Consensus panel statement. In: Circulation. 1999. p. 2480–2484.

6.

Leening MJG, Ferket BS, Steyerberg EW, Kavousi M, Deckers JW, Nieboer D, Heeringa J, Portegies MLP, Hofman A, Ikram MA, Hunink MGM, Franco OH, Stricker BH, Witteman JCM, Roos-Hesselink JW. Sex differences in lifetime risk and first manifestation of cardiovascular disease: prospective population based cohort study. BMJ. 2014;349:g5992.

7.

Ford ES, Ajani U a, Croft JB, Critchley J a, Labarthe DR, Kottke TE, Giles WH, Capewell S. Explaining the decrease in U.S. deaths from coronary disease, 1980-2000. N Engl J Med. 2007;356:2388–98.

8.

Bello N, Mosca L. Epidemiology of coronary heart disease in women. Prog Cardiovasc Dis. 2004;46:287–295.

9.

Appelman Y, van Rijn BB, Ten Haaf ME, Boersma E, Peters S a E. Sex differences in cardiovascular risk factors and disease prevention. Atherosclerosis. 2015;1–8.

10. Anand SS, Yusuf S, Vuksan V, Devanesen S, Teo KK, Montague PA, Kelemen L, Yi C, Lonn E, Gerstein H, Hegele RA, McQueen M. Differences in risk factors, atherosclerosis, and cardiovascular disease between ethnic groups in Canada: the Study of Health Assessment and Risk in Ethnic groups (SHARE). Lancet. 2000;356:279–84. 11. Bairey Merz CN, Shaw LJ, Reis SE, Bittner V, Kelsey SF, Olson M, Johnson BD, Pepine CJ, Mankad S, Sharaf BL, Rogers WJ, Pohost GM, Lerman A, Quyyumi A a., Sopko G. Insights from the NHLBI-sponsored Women’s Ischemia Syndrome Evaluation (WISE) study. Part II: Gender differences in presentation, diagnosis, and outcome with regard to gender-based pathophysiology of atherosclerosis and macrovascular and microvascular cor. J Am Coll Cardiol. 2006;47. 12. De Carvalho LP, Gao F, Chen Q, Hartman M, Sim L-L, Koh T-H, Foo D, Chin C-T, Ong H-Y, Tong K-L, Tan H-C, Yeo T-C, Yew C-K, Richards AM, Peterson ED, Chua T, Chan MY. Differences in late cardiovascular mortality following acute myocardial infarction in three major Asian ethnic groups. Eur Hear journal Acute Cardiovasc care. 2014; 13. Bucholz EM, Butala NM, Rathore SS, Dreyer RP, Lansky AJ, Krumholz HM. Sex differences in long-term mortality after myocardial infarction: a systematic review. Circulation. 2014;130:757–67. 14. Den Ruijter HM, Haitjema S, van der Meer MG, van der Harst P, Rouleau JL, Asselbergs FW, van Gilst WH. Long-term outcome in men and women after CABG; results from the IMAGINE trial. Atherosclerosis. 2015; 15. Arroyo-Espliguero R, Avanzas P, Quiles J, Kaski JC. C-reactive protein predicts functional status and correlates with left ventricular ejection fraction in patients with chronic stable angina. Atherosclerosis. 2009;205:319–324. 16. Calabrò P, Golia E, Yeh ETH. CRP and the risk of atherosclerotic events. Semin Immunopathol. 2009;31:79–94. 17. Fuller JH, Shipley MJ, Rose G, Jarrett RJ, Keen H. Coronary-heart-disease risk and impaired glucose tolerance. The Whitehall study. Lancet. 1980;1:1373–1376. 18. Park CM, Tillin T, March K, Ghosh AK, Jones S, Wright A, Heasman J, Francis D, Sattar N, Mayet J, Chaturvedi N, Hughes AD. Hyperglycemia has a greater impact on left ventricle function in South Asians than in Europeans. Diabetes Care. 2014;37:1124–1131.

313


Chapter 16

19. D’Agostino RB, Vasan RS, Pencina MJ, Wolf PA, Cobain M, Massaro JM, Kannel WB. General cardiovascular risk profile for use in primary care: the Framingham Heart Study. Circulation. 2008;117:743–53. 20. Lorenz MW, Markus HS, Bots ML, Rosvall M, Sitzer M. Prediction of clinical cardiovascular events with carotid intima-media thickness: A systematic review and meta-analysis. Circulation. 2007;115:459–467. 21. Woodward M, Huxley H, Lam TH, Barzi F, Lawes CMM, Ueshima H. A comparison of the associations between risk factors and cardiovascular disease in Asia and Australasia. Eur J Cardiovasc Prev Rehabil. 2005;12:484–91. 22. Wang JW, Gijsberts CM, Seneviratna A, de Hoog VC, Vrijenhoek JEP, Schoneveld a H, Chan MY, Lam CSP, Richards a M, Lee CN, Mosterd A, Sze SK, Timmers L, Lim SK, Pasterkamp G, de Kleijn DP V. Plasma extracellular vesicle protein content for diagnosis and prognosis of global cardiovascular disease. Neth Heart J. 2013;21:467–71. 23. Mak K-H, Chia K-S, Kark JD, Chua T, Tan C, Foong B-H, Lim Y-L, Chew S-K. Ethnic differences in acute myocardial infarction in Singapore. Eur Heart J. 2003;24:151–60. 24. Sianos G, Morel M, Kappetein AP, Morice M, Colombo A, Dawkins K, van den Brand M, Van Dyck N, Russell ME, Mohr FW, Serruys PW. The SYNTAX Score: an angiographic tool grading the complexity of coronary artery disease. EuroIntervention. 2005;1:219–27. 25. Chiu M, Austin PC, Manuel DG, Tu J V. Comparison of cardiovascular risk profiles among ethnic groups using population health surveys between 1996 and 2007. CMAJ. 2010;182:E301–10. 26. Van Holten TC, Waanders LF, de Groot PG, Vissers J, Hoefer IE, Pasterkamp G, Prins MWJ, Roest M. Circulating biomarkers for predicting cardiovascular disease risk; a systematic review and comprehensive overview of meta-analyses. PLoS One. 2013;8:e62080. 27. Weber M, Dill T, Arnold R, Rau M, Ekinci O, Müller KD, Berkovitsch A, Mitrovic V, Hamm C. N-terminal B-type natriuretic peptide predicts extent of coronary artery disease and ischemia in patients with stable angina pectoris. Am Heart J. 2004;148:612–620. 28. Arroyo-Espliguero R, Avanzas P, Quiles J, Kaski JC. Predictive value of coronary artery stenoses and C-reactive protein levels in patients with stable coronary artery disease. Atherosclerosis. 2009;204:239–43. 29. Imai A, Komatsu S, Ohara T, Kamata T, Yoshida J, Miyaji K, Shimizu Y, Takewa M, Hirayama A, Deshpande GA, Takahashi O, Kodama K. Serum cystatin C is associated with early stage coronary atherosclerotic plaque morphology on multidetector computed tomography. Atherosclerosis. 2011;218:350–5. 30. Baldus S, Heeschen C, Meinertz T, Zeiher AM, Eiserich JP, Münzel T, Simoons ML, Hamm CW. Myeloperoxidase serum levels predict risk in patients with acute coronary syndromes. Circulation. 2003;108:1440–5. 31. Ndrepepa G, Braun S, Schulz S, Mehilli J, Schömig A, Kastrati A. High-sensitivity troponin T level and angiographic severity of coronary artery disease. Am J Cardiol. 2011;108:639–43. 32. Hoefer IE, Steffens S, Ala-Korpela M, Bäck M, Badimon L, Bochaton-Piallat M-L, Boulanger CM, Caligiuri G, Dimmeler S, Egido J, Evans PC, Guzik T, Kwak BR, Landmesser U, Mayr M, Monaco C, Pasterkamp G, Tuñón J, Weber C, ESC Working Group Atherosclerosis and Vascular Biology. Novel methodologies for biomarker discovery in atherosclerosis. Eur Heart J. 2015; 33. Bui AL, Horwich TB, Fonarow GC. Epidemiology and risk profile of heart failure. Nat Rev Cardiol. 2011;8:30–41. 34. Luwaert RJ, Cosyns J, Rousseau MF, Brasseur LA, Detry JM, Brohet CR. Reassessment of the relation between QRS forces of the orthogonal electrocardiogram and left ventricular ejection fraction. Eur Heart J. 1983;4:103–109. 35. Lund LH, Jurga J, Edner M, Benson L, Dahlström U, Linde C, Alehagen U. Prevalence, correlates, and prognostic significance of QRS prolongation in heart failure with reduced and preserved ejection fraction. Eur Heart J. 2013;34:529–539.

314


Summary and General Discussion

36. The EchoNoRMAL Study. Ethnic-Specific Normative Reference Values for Echocardiographic LA and LV Size, LV Mass, and Systolic Function. JACC Cardiovasc Imaging. 2015;8:656–665. 37. Brignole M, Auricchio A, Baron-Esquivias G, Bordachar P, Boriani G, Breithardt O-A, Cleland J, Deharo J-C, Delgado V, Elliott PM, Gorenek B, Israel CW, Leclercq C, Linde C, Mont L, Padeletti L, Sutton R, Vardas PE, Zamorano JL, Achenbach S, Baumgartner H, Bax JJ, Bueno H, Dean V, Deaton C, Erol C, Fagard R, Ferrari R, Hasdai D, Hoes AW, Kirchhof P, Knuuti J, Kolh P, Lancellotti P, Linhart A, Nihoyannopoulos P, Piepoli MF, Ponikowski P, Sirnes PA, Tamargo JL, Tendera M, Torbicki A, Wijns W, Windecker S, Blomstrom-Lundqvist C, Badano LP, Aliyev F, Bänsch D, Bsata W, Buser P, Charron P, Daubert J-C, Dobreanu D, Faerestrand S, Le Heuzey J-Y, Mavrakis H, McDonagh T, Merino JL, Nawar MM, Nielsen JC, Pieske B, Poposka L, Ruschitzka F, Van Gelder IC, Wilson CM. 2013 ESC Guidelines on cardiac pacing and cardiac resynchronization therapy: the Task Force on cardiac pacing and resynchronization therapy of the European Society of Cardiology (ESC). Developed in collaboration with the European Heart Rhythm Association . Eur Heart J. 2013;34:2281–329. 38. World Health Organization. Prevalence of Physical Inactivity [Internet]. Available from: https://apps.who.int/ infobase/Indicators.aspx 39. Hooper L, Summerbell CD, Higgins JP, Thompson RL, Capps NE, Smith GD, Riemersma RA, Ebrahim S. Dietary fat intake and prevention of cardiovascular disease: systematic review. BMJ. 2001;322:757–763. 40. Strazzullo P, D’Elia L, Kandala N-B, Cappuccio FP. Salt intake, stroke, and cardiovascular disease: meta-analysis of prospective studies. BMJ. 2009;339:b4567. 41. Maitra A, Shanker J, Dash D, John S, Sannappa PR, Rao VS, Ramanna JK, Kakkar V V. Polymorphisms in the IL6 gene in Asian Indian families with premature coronary artery disease--the Indian Atherosclerosis Research Study. Thromb Haemost. 2008;99:944–50. 42. Tan JH-H, Low P-S, Tan Y-S, Tong M-C, Saha N, Yang H, Heng C-K. ABCA1 gene polymorphisms and their associations with coronary artery disease and plasma lipids in males from three ethnic populations in Singapore. Hum Genet. 2003;113:106–17. 43. Chow CK, McQuillan B, Raju PK, Iyengar S, Raju R, Harmer JA, Neal BC, Celermajer DS. Greater adverse effects of cholesterol and diabetes on carotid intima-media thickness in South Asian Indians: comparison of risk factor-IMT associations in two population-based surveys. Atherosclerosis. 2008;199:116–22. 44. Wenger NK. Coronary Heart Disease: The Female Heart Is Vulnerable. Prog. Cardiovasc. Dis. 2003;46:199– 229. 45. Ong P, Athanasiadis A, Borgulya G, Mahrholdt H, Kaski JC, Sechtem U. High prevalence of a pathological response to acetylcholine testing in patients with stable angina pectoris and unobstructed coronary arteries. The ACOVA Study (Abnormal COronary VAsomotion in patients with stable angina and unobstructed coronary arteries. J Am Coll Cardiol. 2012;59:655–62. 46. Ong P, Carro A, Athanasiadis A, Borgulya G, Schäufele T, Ratge D, Gaze D, Sechtem U, Kaski JC. Acetylcholineinduced coronary spasm in patients with unobstructed coronary arteries is associated with elevated concentrations of soluble CD40 ligand and high-sensitivity C-reactive protein. Coron Artery Dis. 2015;26:126–32. 47. Recio-Mayoral A, Mason JC, Kaski JC, Rubens MB, Harari O a., Camici PG. Chronic inflammation and coronary microvascular dysfunction in patients without risk factors for coronary artery disease. Eur Heart J. 2009;30:1837–1843. 48. Recio-Mayoral A, Rimoldi OE, Camici PG, Kaski JC. Inflammation and microvascular dysfunction in cardiac syndrome X patients without conventional risk factors for coronary artery disease. JACC Cardiovasc Imaging. 2013;6:660–7.

315


Chapter 16

49. Maas AHEM, Appelman YEA. Gender differences in coronary heart disease. Neth Heart J. 2010;18:598–603. 50. Dahabreh IJ. Index Event Bias as an Explanation for the Paradoxes of Recurrence Risk Research. JAMA. 2011;305:822. 51. Johnson BD, Shaw LJ, Pepine CJ, Reis SE, Kelsey SF, Sopko G, Rogers WJ, Mankad S, Sharaf BL, Bittner V, Bairey Merz CN. Persistent chest pain predicts cardiovascular events in women without obstructive coronary artery disease: Results from the NIH-NHLBI-sponsored Women’s Ischaemia Syndrome Evaluation (WISE) study. Eur Heart J. 2006;27:1408–1415. 52. Shaw LJ, Bugiardini R, Merz CNB. Women and ischemic heart disease: evolving knowledge. J Am Coll Cardiol. 2009;54:1561–75. 53. Mayor E. Gender roles and traits in stress and health. Front Psychol. 2015;6:779. 54. Norris CM, Murray JW, Triplett LS, Hegadoren KM. Gender roles in persistent sex differences in health-related quality-of-life outcomes of patients with coronary artery disease. Gend Med. 2010;7:330–9. 55. Lundberg U. Stress hormones in health and illness: The roles of work and gender. Psychoneuroendocrinology. 2005;30:1017–1021. 56. Abdallah MS, Wang K, Magnuson E a, Spertus J a, Farkouh ME, Fuster V, Cohen DJ. Quality of life after PCI vs CABG among patients with diabetes and multivessel coronary artery disease: a randomized clinical trial. JAMA. 2013;310:1581–90. 57. Ross R. Atherosclerosis--an inflammatory disease. N Engl J Med. 1999;340:115–26. 58. Mozos I. Mechanisms Linking Red Blood Cell Disorders and Cardiovascular Diseases. Biomed Res Int. 2015;2015:1–12. 59. Müller R, Mellors I, Johannessen B, Aarsand AK, Kiefer P, Hardy J, Kendall R, Scott CS. European multi-center evaluation of the Abbott Cell-Dyn sapphire hematology analyzer. Lab Hematol. 2006;12:15–31. 60. Ten Berg MJ, Huisman A, van den Bemt PML a, Schobben AF a M, Egberts ACG, van Solinge WW. Linking laboratory and medication data: new opportunities for pharmacoepidemiological research. Clin Chem Lab Med. 2007;45:13–9. 61. Vizioli L, Muscari S, Muscari a. The relationship of mean platelet volume with the risk and prognosis of cardiovascular diseases. Int J Clin Pract. 2009;63:1509–1515. 62. Hochholzer W, Valina CM, Stratz C, Amann M, Schlittenhardt D, Büttner HJ, Trenk D, Neumann F-J. Highsensitivity cardiac troponin for risk prediction in patients with and without coronary heart disease. Int J Cardiol. 2014;176:444–9. 63. Schnabel R, Lubos E, Rupprecht HJ, Espinola-Klein C, Bickel C, Lackner KJ, Cambien F, Tiret L, Münzel T, Blankenberg S. B-type natriuretic peptide and the risk of cardiovascular events and death in patients with stable angina: Results from the AtheroGene study. J Am Coll Cardiol. 2006;47:552–558. 64. Lee C Do, Folsom AR, Nieto FJ, Chambless LE, Shahar E, Wolfe DA. White blood cell count and incidence of coronary heart disease and ischemic stroke and mortality from cardiovascular disease in African-American and White men and women: atherosclerosis risk in communities study. Am J Epidemiol. 2001;154:758–64. 65. Gijsberts CM, Seneviratna A, de Carvalho LP, den Ruijter HM, Vidanapthirana P, Sorokin V, Stella P, Agostoni P, Asselbergs FW, Richards AM, Low AF, Lee C-H, Tan HC, Hoefer IE, Pasterkamp G, de Kleijn DP V., Chan MY. Ethnicity Modifies Associations between Cardiovascular Risk Factors and Disease Severity in Parallel Dutch and Singapore Coronary Cohorts. PLoS One. 2015;10:e0132278. 66. Ruijter HM den, Haitjema S, W Asselbergs F, Pasterkamp G. Sex matters to the heart: A special issue dedicated to the impact of sex related differences of cardiovascular diseases. Atherosclerosis. 2015;241:205–7. 67. Shimada YJ, Cannon CP. PCSK9 (Proprotein convertase subtilisin/kexin type 9) inhibitors: past, present, and the future. Eur Heart J. 2015;Epub:1–12.

316


317



Summary and Discussion

Chapter 17 Nederlandse Samenvatting


Chapter 17

Hart- en vaatziekten zijn nog steeds wereldwijd de nummer één doodsoorzaak, ondanks grote verbeteringen in de preventie en behandeling. In 2012 overleden 7,4 miljoen mensen aan deze aandoening. Voornamelijk coronairlijden, veroorzaakt door kransslagaderverkalking, is een ziekte geworden die mannen en vrouwen van alle etniciteiten treft. Aderverkalking (atherosclerose) is de onderliggende ziekte. Hierbij ontstaan er plaques in slagaders door ophoping van lipiden en ontstekingscellen. Wanneer de plaques zo groot worden dat ze de bloedstroom belemmeren of wanneer plaques scheuren, kan een hartinfarct optreden.

DEEL ÉÉN | Etniciteit Achtergrond Met name in Azië worden de komende decennia grote aantallen hart- en vaatziekten patiënten voorspeld door de Wereldgezondheidsorganisatie. In Oost-Azië wordt een toename van 20% voorzien, terwijl in Europa stabiele patiënten aantallen worden voorspeld. Hoewel een Aziatische epidemie van coronairlijden dus aanstonds is, wordt onderzoek naar coronairlijden gedomineerd door blanke studiepopulaties. Uit de bestaande literatuur is duidelijk geworden dat onderzoek gedaan op blanke populaties niet zomaar geëxtrapoleerd kan worden naar andere etnische groepen. Etnische verschillen in het risico en het voorkomen van coronairlijden zijn bekend. Voornamelijk onder Zuid-Aziaten (afkomstig uit India, Bangladesh, Pakistan, etc.) komen diabetes en een hoog cholesterol niveau vaak voor. Daarnaast is het bekend dat ZuidAziaten al op jongere leeftijd coronairlijden ontwikkelen dan blanke mensen. Daarentegen staan Chinezen er om bekend een gunstiger risicoprofiel te hebben met weinig diabetes en lagere cholesterolwaarden. Coronairlijden komt ook minder voor bij Chinezen dan bij blanke mensen. Etnische verschillen in de ernst van het coronairlijden zijn nog maar zelden onderzocht, zeker niet aan de hand van de gouden standaard: het coronairangiogram. Met duidelijke verschillen in de risicofactorprofielen tussen de etnische groepen is het aannemelijk dat er ook verschillen bestaan in de biomarkerprofielen en prognose tussen de etniciteiten. Ook hartfalen, een veel voorkomend gevolg van coronairlijden, begint een groot probleem in Azië te worden. Onderzoek naar etnische verschillen in hartfalen is schaars. Bevindingen Allereerst, in hoofdstuk 2, hebben we gekeken naar etnische verschillen in biomarkers van coronairlijden in de bestaande literatuur. Hierbij vonden we een opvallend onvoordelig biomarker profiel onder Zuid-Aziaten in de algemene populatie. In het bijzonder vielen hoge C-reactive protein (CRP) en hoge insulinewaarden op. Beide waarden zijn gerelateerd aan een hoger risico op hart- en vaatziekten en een hogere mortaliteit.

320


Nederlandse Samenvatting

In hoofdstuk 3 hebben we de etniciteitsspecifieke relatie van de traditionele Framingham risicofactoren met het optreden van hart- en vaatproblemen onderzocht. In deze studie vonden we een klein, maar mogelijk belangrijk verschil tussen de etniciteiten. Deze kleine verschillen kunnen een grote impact hebben, als predictiemodellen op een grote (wereldwijde) schaal worden toegepast. Vervolgens in hoofdstuk 4 hebben we etnische verschillen geëvalueerd in een patiëntengroep die een coronairangiogram ondergaat (de Nederlands-Singaporese UNICORN studie). Hierbij vinden we dat coronairlijden onder Maleisiërs het ernstigst is, gevolgd door Chinezen en Indiërs. Blanke mensen hadden het minst ernstige coronairlijden. Ook vonden we dat diabetes en mannelijk geslacht een sterkere relatie met de ernst van het coronairlijden hadden in Chinezen dan in blanken. De overleving na een coronairangiogram was het slechtst onder Maleisiërs. De resultaten van deze studie impliceren dat etniciteitsspecifieke (secundaire) preventieprogramma’s wellicht nuttig kunnen zijn. In hoofdstuk 5 hebben we de ernst van het coronairlijden van dotterpatiënten gekwantificeerd aan de hand van de SYNTAX-score. Daarbij vonden we een opvallend hoge SYNTAX-score bij Chinese dotterpatiënten die gedotterd werden in verband met een hartinfarct. Hoewel Chinezen een lager risico hebben op coronairlijden, is het coronairlijden dus ernstiger dan in andere etnische groepen wanneer ze een hartinfarct krijgen waarvoor een dotterbehandeling nodig is. De prognose na de dotter was daarentegen niet slechter voor Chinezen dan voor blanken. De prognose van Indiërs en Maleisiërs was mogelijk wel slechter. Etnische verschillen in biomarkers van coronairlijden onder patiënten die een coronairangiogram ondergaan hebben we onderzocht in hoofdstuk 6. Bij Maleisische patiënten waren NT pro-BNP (een marker voor hartfalen) en high-sensitivity Troponine I (een marker voor hartspierschade) significant hoger dan in blanke mensen, ook wanneer gecorrigeerd werd voor verschillen in risicofactoren en ziekte-eigenschappen van deze patiënten. Myeloperoxidase (een enzym geproduceerd door een bepaald type ontstekingscellen; neutrofielen) was significant lager in Indiërs dan in blanken. De relatie tussen biomarkers en de ernst van het coronairlijden leek erg te verschillen tussen de etniciteiten. Vooral de afkapwaarde van de biomarkers voor ernstig coronairlijden verschilde hevig. Dit impliceert dat etniciteitsspecifieke biomarkerafkapwaarden overwogen moeten worden. Helaas waren wij niet in de mogelijkheid om de etniciteitspecifieke relatie van biomarkers met prognose te onderzoeken. Een dergelijk onderzoek zou ons inzicht in de waarde van biomarkers in verschillende etnische groepen aanzienlijk vergroten. In hoofdstuk 7 hebben we naar etnische verschillen in hartfalenpatiënten gekeken. Hartfalen is een veelvoorkomend gevolg van coronairlijden. De QRS-duur op het elektrocardiogram is gerelateerd aan de functie van de linker hartkamer en een belangrijke voorspeller van de prognose van hartfalenpatiënten. In dit hoofdstuk beschrijven we een opvallend verschil in QRS-duur tussen Aziatische en blanke hartfalenpatiënten. Bij hartfalenpatiënten met een goed functionerende linker hartkamer is de QRS-duur korter

321


Chapter 17

in Aziaten dan in blanke patiënten. Echter, wanneer de functie van de linker hartkamer slechter werd, was de verlenging van de QRS-duur meer uitgesproken in Aziaten dan in blanke hartfalenpatiënten. Een mogelijke verklaring voor dit fenomeen kan een betere “normale” functie van de linker hartkamer in Aziaten dan in blanken zijn. Bij een bepaalde vermindering van de linkerkamerfunctie is de relatieve verslechtering dan groter bij Aziaten dan bij blanken. De ergere relatieve verslechtering van de linkerkamerfunctie gaat dan gepaard met een langere QRS-duur. Het lijkt erop dat QRS-duurverlenging zowel in Aziaten als in blanken een voorspeller is van toekomstige ziekenhuisopnames voor hartfalen en overlijden. Mogelijk moeten voor Aziaten en blanken wel andere afkapwaarden gebruikt worden voor het nemen van behandelbeslissingen.

DEEL TWEE | Sekse verschillen Achtergrond Hoewel hart- en vaatziekte vroeger werd gezien als een ziekte die voornamelijk bij mannen voorkwam, is het meer en meer een vrouwenziekte geworden. Het is momenteel zelfs zo dat er jaarlijks meer vrouwen dan mannen overlijden aan hart- en vaatziekten. Tussen mannen en vrouwen bestaan verschillen in de verschijningsvormen van coronairlijden. Vrouwen hebben bijvoorbeeld een stabielere plaque dan mannen en plaque-erosie, in tegenstelling tot plaqueruptuur, komt vaker voor bij vrouwen dan bij mannen. Mogelijk speelt bij vrouwen ook microvasculair lijden (ziekte van de kleine bloedvaatjes die het hart van bloed voorzien) een grotere rol dan bij mannen. In de afgelopen jaren zijn risicofactoren voor coronairlijden, die specifiek zijn voor vrouwen geïdentificeerd. Deze risicofactoren hebben te maken met zwangerschap en hormoonhuishouding en verklaren mogelijk deels de gevonden verschillen tussen mannen en vrouwen. Ook qua prognose bestaan er Sekse verschillen. In de Verenigde Staten is recentelijk een stijging vermeld in het aantal jonge vrouwen dat overlijdt aan coronairlijden, waarschijnlijk is dit het gevolg van een toename in roken en hoge bloeddruk. In Europa is deze stijgende lijn gelukkig nog niet ontdekt. Mannen en vrouwen lijken ook anders om te gaan met de psychologische gevolgen van coronairlijden. Dit kan gemeten worden aan de hand van een Kwaliteit van Leven (KvL)vragenlijst. Een lage KvL zorgt voor meer gebruik van de zorg en dus hogere kosten in de gezondheidszorg. Vrouwen die behandeld worden voor coronairlijden vermelden een lagere KvL dan mannen. Factoren die de KvL beïnvloeden in mannen en vrouwen met coronairlijden zijn grotendeels onbekend. Het is mogelijk dat vrouw-specifieke risicofactoren de lagere KvL van vrouwen deels verklaren. Tot op heden is het niet bekend of lagere KvL in vrouwen ook gepaard gaat met een slechtere prognose. Bevindingen In hoofdstuk 8 hebben we mannen en vrouwen vergeleken die een coronairangiogram ondergaan wegens stabiele pijn op de borst klachten. Wanneer we voor de verschillen

322


Nederlandse Samenvatting

in patiënt karakteristieken en risicofactoren corrigeren, zien we dat vrouwen minder uitgebreid coronairlijden hebben dan mannen. Hierbij vinden we lagere waarden van high-sensitivity Troponine I (een marker voor hartspierschade), passend bij een kleiner hart in vrouwen dan in mannen. Echter, het niveau van high-sensitivity C-reactive protein in het bloed (een marker voor ontstekingsreactie) was hoger in vrouwen dan in mannen. Mogelijk is dit een afspiegeling van ontsteking in de kleine vaatjes van het hart, microvasculair lijden, wat vaker wordt gezien in vrouwen. Vervolgens hebben we in hoofdstuk 9 verschillen in prognose na een coronairangiogram tussen mannen en vrouwen onderzocht. Vrouwen die bij het coronairangiogram meervatslijden hebben (meer dan één van de drie kransslagaderen is aangedaan) of die een coronairangiogram ondergaan in verband met een hartinfarct, hebben een slechtere prognose dan mannen. Hoewel “vrouw-zijn” beschermend is voor het krijgen van harten vaatziekten, lijkt het zo te zijn dat dit beschermende effect verloren gaat of zelfs schadelijk wordt als er ernstig coronairlijden bestaat. Dit wordt ook wel een “index-event bias” genoemd. In tegenstelling tot de Amerikaanse literatuur, hebben vrouwen met stabiele klachten en niet-ernstig coronairlijden in ons cohort geen slechtere prognose dan mannen. Een mogelijke oorzaak hiervoor zijn de verschillen in risicofactoren tussen Nederlandse en Amerikaanse vrouwen. In hoofdstuk 10 bekijken we coronairlijden van een hele andere kant. We beschrijven hier namelijk de Kwaliteit van Leven (KvL) van mannen en vrouwen die een coronairangiogram ondergaan. Vrouwen geven een slechtere KvL aan dan mannen, ongeacht de indicatie of de uitkomst van het coronairangiogram. Aansluitend, in hoofdstuk 11, onderzoeken we of vrouw-specifieke risicofactoren een oorzaak kunnen zijn voor de lagere KvL die in vrouwen wordt gezien. Dit lijkt niet zo te zijn. Verder bediscussiëren we de relatie tussen KvL en het optreden van hart- en vaatproblemen en overlijden ná een coronairangiogram. Bij mannen en vrouwen met perifeer vaatlijden (in de benen of in de halsslagader) is een lage KvL voorspellend voor het optreden van harten vaatproblemen en overlijden na een vaatoperatie. Er is geen verschil in deze relatie tussen mannen en vrouwen. Bij patiënten die een coronairangiogram hebben ondergaan, echter, lijkt de relatie tussen KvL en hart- en vaatproblemen en overlijden zwakker te zijn voor vrouwen dan voor mannen. Een lage KvL in vrouwen met coronairlijden lijkt niet sterk gerelateerd te zijn aan een slechtere prognose. Een mogelijke verklaring voor deze paradox kan gezocht worden in de psychologische omgang van mannen en vrouwen met klachten van coronairlijden. Het sociaal bepaalde “geslacht” van een persoon, in tegenstelling tot zijn/haar biologische “sekse”, lijkt een grote invloed te hebben op de manier waarop met stress omgegaan wordt. Factoren die de KvL kunnen verbeteren bij vrouwen, naast het verlichten van de lichamelijke klachten, moeten worden geïdentificeerd om het psychologische lijden en daaropvolgend gezondheidszorggebruik te verminderen.

323


Chapter 17

DEEL DRIE | Risicopredictie Achtergrond In aanvulling op etniciteits- en seksespecifieke schattingen, kan het voorspellen van risico op toekomstige hart- en vaatproblemen verbeterd worden door patiënteigenschappen of bloedwaarden toe te voegen aan een voorspellingsmodel. Eén van de patiënteigenschappen die we onderzoeken is de respons op de KvL-vragenlijst. Maar ook hematologische parameters, zoals witte en rode bloedceleigenschappen en eigenschappen van bloedplaatjes zijn vermeld in de literatuur als mogelijk voorspellers van toekomstige hart- en vaatproblemen. Een bijzonder voordeel van deze bloedwaarden is dat ze door moderne meetapparatuur standaard mee worden bepaald ook wanneer er maar één enkele parameter wordt aangevraagd. Van veel patiënten zijn deze waarden dus beschikbaar, zonder dat daar in de kliniek iets mee wordt gedaan. Bevindingen In het eerste hoofdstuk van dit deel van de thesis, hoofdstuk 12, beschrijven we het verschil in prognose tussen mensen die wel en niet hun KvL-vragenlijst retour hebben gestuurd. Logischerwijs kunnen de gegevens van patiënten die hun vragenlijst niet terug hebben gestuurd niet geanalyseerd worden en conclusies worden getrokken zonder deze gegevens in ogenschouw te nemen. In de analyse beschreven in dit hoofdstuk laten we zien dat patiënten die hun vragenlijst niet retour sturen een zeer slechte prognose hebben, zelfs slechter dan patiënten die een lage KvL vermelden. Non-respons op KvL-vragenlijsten kan dus gezien worden als een indicator voor een hoog risico op overlijden na een coronairangiogram. Non-respons kan grote gevolgen hebben voor de generaliseerbaarheid van vragenlijst-gedreven onderzoek. Het lijkt erop dat juist de patiëntengroep met de slechtste prognose hierbij buiten beschouwing gelaten wordt. In de laatste drie hoofdstukken van deze thesis gaan we in op het verbeteren van predictie door middel van het toevoegen van gemakkelijk te meten biologische patiënteigenschappen. Eerst bekijken we, in verband met de prominente rol van ontsteking in het ontstaan van aderverkalking, in hoofdstuk 13 de voorspellende waarde van witte bloedcellen (ontstekingscellen). Van de onderzochte witte bloedceleigenschappen, lijkt de monocytlymfocyt ratio het best de voorspelling van toekomstige hart- en vaatproblemen te verbeteren. Vervolgens, gedreven door het toenemende wetenschappelijke bewijs over de voorspellende eigenschappen van rode bloedceleigenschappen, hebben we de zoektocht naar de beste hematologische parameter uitgebreid. In hoofdstuk 14 onderzoeken we witte bloedcel-, rode bloedcel- en bloedplaatjeseigenschappen die routinematig worden gemeten in hematologische analyse apparatuur. In moderne analyzers wordt altijd een compleet bloedbeeld bepaald, wanneer één parameter wordt aangevraagd. Alleen de aangevraagde parameter wordt terug gerapporteerd aan de aanvrager. Er zijn dus veel meer data beschikbaar dan bekend wordt gemaakt aan de kliniek. In het UMC Utrecht

324


Nederlandse Samenvatting

worden deze data opgeslagen in de Utrecht Patient Oriented Database (UPOD) en zijn ze beschikbaar voor onderzoek. De analyse die we in dit hoofdstuk beschrijven laat zien dat panels van twee tot vijf hematologische parameters de voorspelling van toekomstige hart- en vaatproblemen en overlijden significant kunnen verbeteren. Vijf van de zes panels bestond uit een combinatie van witte bloedcel- en rode bloedceleigenschappen. Tot onze verbazing zaten bloedplaatjeseigenschappen, die in de literatuur vermeld worden als voorspellend, niet in onze panels. Mogelijk wordt de voorspellende waarde van bloedplaatjeseigenschappen overstemd door de eigenschappen van rode en witte bloedcellen. Om te bepalen of voorspelling door middel van hematologische parameters beter is dan bestaande biomarkers van coronairlijden, hebben we ze in hoofdstuk 15 vergeleken met high-sensitivity Troponine I (hsTnI, een marker voor hartspierschade) en met NT pro-BNP (een marker voor hartfalen). Deze vergelijking liet zien dat NT pro-BNP een iets toegevoegde waarde had, maar hsTnI helemaal geen. Hematologische parameters, daarentegen, zoals ook gedemonstreerd in hoofdstuk 14, konden de risicoschatting significant verbeteren. NT pro-BNP en hsTnI konden de voorspelling met behulp van hematologische parameters niet meer verder verbeteren, wat suggereert dat de meeste prognostische informatie afkomstig is van de hematologische parameters. Toekomstperspectief Het afgelopen decennium is de aandacht voor gepersonaliseerde geneeskunde (tailored medicine) gestegen door toegenomen patiëntenaantallen en toegenomen behandelmogelijkheden. Patiënten met coronairlijden vormen en heterogene groep met diverse patiëntkarakteristieken en hebben in verschillende mate risico’s op toekomstige hart- en vaatproblemen. Om over- én onderbehandeling te voorkomen, is het essentieel om de juiste individuen aan te wijzen voor de juiste therapieën. De resultaten gepresenteerd in deze thesis, benadrukken de noodzaak om onderzoek naar hart- en vaatziekten te stratificeren naar etniciteit en sekse. Door ons inzicht in etniciteits- en seksespecifieke karakteristieken en risico’s van coronairlijden te vergroten, kunnen (secundaire) preventiestrategieën gericht worden op de juiste personen. Indische en Maleisische patiënten met coronairlijden, bijvoorbeeld, hebben ernstiger coronairlijden op het coronairangiogram en een slechtere prognose nadien. Bij deze patiëntengroepen kan gedacht worden aan een strengere controle en secundair behandelregime. Genetisch onderzoek zal hopelijk helpen om etnische verschillen in coronairlijden verder te ontrafelen. Multi-etnische coronaire biobanken, zoals de Nederlands-Singaporese UNICORN-studie, bieden hiervoor unieke mogelijkheden. Wat betreft sekseverschillen, hebben vrouwen die een coronairangiogram ondergaan minder ernstig coronairlijden dan mannen. Maar vrouwen die zich presenteren met een hartinfarct of die meervatslijden op het angiogram hebben, hebben een slechtere prognose dan mannen. Deze vrouwen zouden baat kunnen hebben bij een striktere follow-up en zo mogelijk intensievere secundaire preventiemedicatie. Onderzoek naar vrouw-specifieke risicofactoren, zwangerschapsgebonden vaataandoeningen

325


Chapter 17

(bijvoorbeeld pre-eclampsie) en onderzoek naar de geslachtschromosomen kan onze kennis over sekseverschillen in de ontwikkeling, het risico en de ernst van hart- en vaatziekten uitbreiden. Naast etniciteit en sekse, kan risicopredictie verbeterd worden door toevoeging van biomarkers. In deze thesis hebben we kritisch gekeken naar de toegevoegde waarde van gemakkelijk te meten hematologische parameters. Daarbij vonden we dat de klinische waarde van deze biomarkers substantieel kan zijn. Simpele softwaretoevoegingen aan bestaande hematologieanalyzers kan er in theorie voor zorgen dat bij elk bloedmonster een risico op overlijden wordt meegegeven. Manieren waarop een dergelijke risicoindicator geïncorporeerd zou kunnen worden in de dagelijkse klinische praktijk moeten onderzocht worden. Een mogelijke toepassing zou de selectie van hoog-risico patiënten voor toediening van nieuwe, kostbare therapieën, zoals de PCSK9-inhibitoren kunnen zijn. Aan de andere kant kunnen langere intervallen tussen controles worden overwogen bij laag-risico patiënten.

326


327



Appendix


Appendix

Review Committee Prof. dr. A.W. Hoes, UMC Utrecht Prof. dr. F. Zijlstra, Erasmus MC Rotterdam Prof. dr. M.L. Bots, UMC Utrecht Prof. dr. P.A.F. Doevendans, UMC Utrecht A/Prof. C.S.P. Lam, Singapore General Hospital

330


Appendix

Author Affiliations Crystel M. Gijsberts

ICIN-Netherlands Heart Institute, Utrecht, The Netherlands Laboratory of Experimental Cardiology, University Medical Centre Utrecht, Utrecht, The Netherlands

Co-authors Affiliations A. Mark Richards Department of Cardiology, National University Singapore, Singapore, Singapore; Cardiovascular Research Institute (CVRI), National University Heart Centre (NUHCS), National University Health System, Singapore Adrian F.H. Low Cardiac Department, National University Heart Centre, National University Hospital, Singapore Ai Ikeda Osaka Medical Center for Health Science and Promotion, Osaka, Japan Aisha Gohar Laboratory of Experimental Cardiology, University Medical Centre Utrecht, Utrecht, The Netherlands Akihiko Kitamura Osaka Medical Center for Health Science and Promotion, Osaka, Japan Albert Huisman Department of Clinical Chemistry and Hematology, University Medical Center Utrecht, Utrecht, The Netherlands Annie R. Britton Department of Epidemiology and Public Health University College London, London, United Kingdom Aruni Seneviratna Cardiac Department, National University Heart Centre, National University Hospital, Singapore Bernadet T. Santema Laboratory of Experimental Cardiology, University Medical Centre Utrecht, Utrecht, The Netherlands Bo Hedblad Department of Clinical Sciences in Malmö, Lund University, Skåne University Hospital, Malmö, Sweden Carolyn S.P. Lam National Heart Centre Singapore, National University Hospital, Singapore; Duke-NUS, Singapore Christine Robertson Centre for Population Health Sciences, University of Edinburgh, United Kingdom Christopher M. Rembold Cardiology Division, Department of Internal Medicine, University of Virginia, Charlottesville, VA, United States of America Coen D.A. Stehouwer Department of Internal Medicine and Cardiovascular Research Institute Maastricht, Maastricht University Medical Center, Maastricht, The Netherlands Daniel H. O’Leary Department of Radiology, Tufts Medical Center, Boston, MA, United States of America David Sim National Heart Centre, Singhealth, Singapore

331


Appendix

Diederick E. Grobbee

Julius Center for Health Sciences and Primary Care, University Medical Center Utrecht, the Netherlands; University of Malaya Medical Center, Kuala Lumpur, Malaysia Dominique P.V. de Kleijn Laboratory of Experimental Cardiology, University Medical Centre Utrecht, Utrecht, The Netherlands Ellisiv B. Mathiesen Brain and Circulation Research Group, Department of Clinical Medicine, University of Tromsø, Tromsø, Norway Eva M. Lonn Department of Medicine, Division of Cardiology and Population Health Research Institute, McMaster University, Hamilton, Ontario, Canada Fazlur Jaufeerally Department of Cardiology, Singapore General Hospital, Singapore Folkert W. Asselbergs Department of Cardiology, University Medical Centre Utrecht, Utrecht, The Netherlands; Durrer Center for Cardiogenetic Research, ICIN-Netherlands Heart Institute, Utrecht, The Netherlands; Institute of Cardiovascular Science, faculty of Population Health Sciences, University College London, London, United Kingdom Gerard K.T. Leong Department of Cardiology, Changi General Hospital, Singapore Gerard Pasterkamp Laboratory of Experimental Cardiology, University Medical Centre Utrecht, Utrecht, The Netherlands Gert-Jan de Borst Department of Vascular Surgery, University Medical Center Utrecht, The Netherlands Giel Nijpels Institute for Health and Care Research, VU University Medical Center, Amsterdam, The Netherlands Greg W. Evans Department of Biostatistical Sciences and Neurology, Wake Forest School of Medicine, Winston-Salem, NC, United States of America Guilielmus H.J.M. Ellenbroek Laboratory of Experimental Cardiology, University Medical Centre Utrecht, Utrecht, The Netherlands Gunnar Engström Department of Clinical Sciences in Malmö, Lund University, Skåne University Hospital, Malmö, Sweden Hean Yee Ong Department of Cardiology, Khoo Teck Puat Hospital, Singapore Hendrik M. Nathoe Department of Cardiology, division Heart and Lungs, University Medical Centre Utrecht, The Netherlands Heng Chew Kiat Department of Paediatrics, National University of Singapore, Singapore Hester M. den Ruijter Laboratory of Experimental Cardiology, University Medical Centre Utrecht, Utrecht, The Netherlands Huay Cheem Tan Cardiac Department, National University Heart Centre,

332


Appendix

National University Hospital, Singapore Julius Center for Health Sciences and Primary Care, University Medical Center Utrecht, the Netherlands Imo E. Hoefer Laboratory of Experimental Cardiology, University Medical Centre Utrecht, Utrecht, The Netherlands Ingrid E.M. Bank Laboratory of Experimental Cardiology, University Medical Centre Utrecht, Utrecht, The Netherlands Jacqueline de Graaf Department of General Internal Medicine, Division of Vascular Medicine, Radboud University Nijmegen Medical Centre Jacqueline F. Price Centre for Population Health Sciences, University of Edinburgh, United Kingdom Jacqueline M. Dekker Institute for Health and Care Research, VU University Medical Center, Amsterdam, The Netherlands Jasper A. Remijn Clinical Chemistry and Hematology, Gelre Ziekenhuizen, Apeldoorn, the Netherlands; Clinical Chemistry and Hematology, University Medical Center Utrecht, Utrecht, the Netherlands Joseph F. Polak Department of Radiology, Tufts Medical Center, Boston, MA, United States of America Jukka T. Salonen MAS-Metabolic Analytical Services Oy Karlijn A. Groenewegen Julius Center for Health Sciences and Primary Care, University Medical Center Utrecht, the Netherlands Kazuo Kitagawa Osaka Medical Center for Health Science and Promotion, Osaka, Japan Lars H. Lund Department of Medicine and Department of Cardiology, Karolinska Institutet, Karolinska University Hospital, Stockholm, Sweden Leonardo Pinto de Carvalho Cardiac Department, National University Heart Centre, National University Hospital, Singapore Lieng H. Ling Yong Loo Lin School of Medicine, National University, Singapore; Cardiac Department, National University Health System, Singapore Maarten J. Cramer Department of Cardiology, division Heart and Lungs, University Medical Centre Utrecht, The Netherlands Maarten J. ten Berg Department of Clinical Chemistry and Hematology, University Medical Center Utrecht, Utrecht, The Netherlands Maria Rosvall Department of Clinical Sciences in Malmรถ, Lund University, Skรฅne University Hospital, Malmรถ, Sweden Marinus J.C. Eijkemans Julius Center for Health Sciences and Primary Care, University Medical Center Utrecht, the Netherlands Mark de Groot Department of Clinical Chemistry and Hematology, University Medical Center Utrecht, Utrecht, The Netherlands Ilonca Vaartjes

333


Appendix

Mark Roest Mark Y. Chan

Matthias Sitzer

Matthias W. Lorenz Michiel L. Bots Michiel Voskuil Mikael Hartman

P.S. Daniel Yeo Pierfrancesco Agostoni Pieter Stella Puwalani Vidanapthirana Richard H.A. van Wijk Ronald Chi Hang Lee Sanne A.E. Peters Saskia Haitjema Saskia Z.H. Rittersma Shuhei Okazaki Suzanne Holewijn Tatjana Rundek

334

Department of Clinical Chemistry and Hematology, University Medical Center Utrecht, Utrecht, The Netherlands Department of Cardiology, National University Singapore, Singapore, Singapore; Cardiovascular Research Institute (CVRI), National University Heart Centre (NUHCS), National University Health System, Singapore Department of Neurology, University Hospital, GoetheUniversity, Frankfurt am Main, Germany; Department of Neurology, Klinikum Herford, Germany Department of Neurology, University Hospital, GoetheUniversity, Frankfurt am Main, Germany Julius Center for Health Sciences and Primary Care, University Medical Center Utrecht, the Netherlands Department of Cardiology, division Heart and Lungs, University Medical Centre Utrecht, The Netherlands Department of Surgery, Yong Loo Lin School of Medicine, National University of Singapore; Saw Swee Hock School of Public Health, National University of Singapore Department of Cardiology, Tan Tock Seng Hospital, Singapore Department of Cardiology, division Heart and Lungs, University Medical Centre Utrecht, The Netherlands Department of Cardiology, division Heart and Lungs, University Medical Centre Utrecht, The Netherlands Department of Surgery, Yong Loo Lin School of Medicine, National University of Singapore, Singapore Department of Clinical Chemistry and Hematology, University Medical Center Utrecht, Utrecht, The Netherlands Cardiac Department, National University Heart Centre, National University Hospital, Singapore Julius Center for Health Sciences and Primary Care, University Medical Center Utrecht, the Netherlands Laboratory of Experimental Cardiology, University Medical Centre Utrecht, Utrecht, The Netherlands Department of Cardiology, division Heart and Lungs, University Medical Centre Utrecht, The Netherlands Stroke Center, Department of Neurology, Osaka University Graduate School of Medicine, Osaka, Japan Department of General Internal Medicine, Division of Vascular Medicine, Radboud University Nijmegen Medical Centre Department of Neurology, Miller School of Medicine, University of Miami, Miami, Fl, United States of America


Appendix

Todd J. Anderson Ulf Dahlstrรถm Vitaly Sorokin Wouter W. van Solinge Yolande Appelman

Department of Cardiac Sciences and Libin Cardiovascular Institute of Alberta, University of Calgary, Alberta, Canada Department of Cardiology and Department of Medical and Health Sciences Linkรถping University, Linkรถping, Sweden Department of Surgery, Yong Loo Lin School of Medicine, National University of Singapore Department of Clinical Chemistry and Hematology, University Medical Center Utrecht, Utrecht, The Netherlands Department of Cardiology, VU University Medical Center, Amsterdam, the Netherlands

335


Appendix

List of Publications Accepted 1. Gijsberts CM, den Ruijter HM, Asselbergs FW, Chan MY, de Kleijn DP V, Hoefer IE. Biomarkers of Coronary Artery Disease Differ Between Asians and Caucasians in the General Population. Global Heart. 2015; doi:10.1016/j.gheart.2014.11.004 2. Gijsberts CM*, Groenewegen KA*, Hoefer IE, Eijkemans MJC, Asselbergs FW, Anderson TJ, et al. Race/Ethnic Differences in the Associations of the Framingham Risk Factors with Carotid IMT and Cardiovascular Events. PLoS One; 2015;10: e0132321. doi:10.1371/journal.pone.0132321 3. Gijsberts CM, Seneviratna A, de Carvalho LP, den Ruijter HM, Vidanapthirana P, Sorokin V, et al. Ethnicity Modifies Associations between Cardiovascular Risk Factors and Disease Severity in Parallel Dutch and Singapore Coronary Cohorts. PLoS One. 2015;10: e0132278. doi:10.1371/journal.pone.0132278 4. Gijsberts CM, Seneviratna A, Hoefer IE, Agostoni P, Rittersma SZH, Pasterkamp G, et al. Inter-Ethnic Differences in Quantified Coronary Artery Disease Severity and All-Cause Mortality among Dutch and Singaporean Percutaneous Coronary Intervention Patients. PLoS One. 2015;10: e0131977. doi:10.1371/journal. pone.0131977 5. Gijsberts CM, Seneviratna A*, Bank IEM*, den Ruijter HM, Asselbergs FW, Agostoni P, et al. The Ethnicity-specific association of biomarkers with the angiographic severity of coronary artery disease. Accepted for publication in the Netherlands Heart Journal 6. Gijsberts CM, Gohar A, Ellenbroek GHJM, Hoefer IE, de Kleijn DPV, Asselbergs FW, et al. Severity of stable coronary artery disease and its biomarkers differ between men and women undergoing angiography. Atherosclerosis. 2015;241: 234–240. doi:10.1016/j.atherosclerosis.2015.02.002 7. Gijsberts CM, Santema BT, Asselbergs FW, de Kleijn DPV, Voskuil M, Agostoni P, et al. Women Undergoing Coronary Angiography for Myocardial Infarction or Who Have Multi-vessel Disease Have a Worse Prognosis Than Men. Angiology. 2015 Sep 7. pii: 0003319715604762 8. Gijsberts CM, Agostoni P, Hoefer IE, Asselbergs FW, Pasterkamp G, Nathoe H, et al. Gender differences in health-related quality of life in patients undergoing coronary angiography. Open Heart. 2015 Aug 27;2(1):e000231

336


Appendix

9. Gijsberts CM, den Ruijter HM. Non-Response to Questionnaires Independently Predicts Mortality of Coronary Angiography patients. Int J Cardiol. 2015; doi:10.1016/j. ijcard.2015.06.150 10. Gijsberts CM, den Ruijter HM, de Kleijn DPV, Huisman A, ten Berg MJ, van Wijk HA, et al. Hematological Parameters Improve Prediction of Mortality and Secondary Adverse Events in Coronary Angiography Patients: a Longitudinal Cohort Study. Accepted for publication in Medicine 11. Wang JW, Gijsberts CM, Seneviratna A, de Hoog VC, Vrijenhoek JEP, Schoneveld a H, et al. Plasma extracellular vesicle protein content for diagnosis and prognosis of global cardiovascular disease. Neth Heart J. 2013;21: 467–71. doi:10.1007/s12471013-0462-3 12. De Ruiter QMB, Moll FL, Gijsberts CM, van Herwaarden JA. AlluraClarity Radiation Dose-Reduction Technology In The Hybrid Operating Room During Endovascular Aortic Aneurysm Repair. Accepted for publication in the Journal of Endovascular Therapy 13. Van Hout GPJ, van Solinge WW, Gijsberts CM, Teuben MPJ, Leliefeld PHC, Heeres M, et al. Elevated mean neutrophil volume represents altered neutrophil composition and reflects damage after myocardial infarction. Accepted for publication in Basic Research in Cardiology 14. Bank IEM, Timmers L, Gijsberts CM, Zhang Y, Mosterd A, Wang JW, et al. The Diagnostic and Prognostic Potential of Plasma Extracellular Vesicles for Cardiovascular Disease. Accepted for publication in Expert Review of Molecular Diagnostics. Submitted and in preparation 15. Gijsberts CM, Lund L, DahlstrÜm U, de Kleijn DPV, Lam CSP. Ethnic Differences In QRS Prolongation and its Association With Ejection Fraction and Outcomes in Heart Failure. 16. Gijsberts CM, Gohar A, Haitjema S, Pasterkamp G, de Kleijn DPV, Asselbergs FW, et al. Sex Differences in Health-Related Quality of Life of Peripheral and Coronary Artery Disease Patients and its Relation With Adverse Events and Mortality. 17. Gijsberts CM, Ellenbroek GHJM, ten Berg MJ, Huisman A, van Solinge WW, Asselbergs FW, et al. Routinely Analyzed Leukocyte Characteristics Improve Prediction of Mortality After Coronary Angiography.

337


Appendix

18. Gijsberts CM, den Ruijter HM, de Kleijn DPV, Huisman A, de Groot M, van Wijk HA, et al. The Value of Hematological Parameters Exceeds High-Sensitivity Troponin I and NT Pro-BNP for Mortality Prediction in Coronary Angiography Patients. 19. Gijsberts CM*, Leunissen TC*, Wisman PP, Huisman A, ten Berg MJ, Asselbergs FW, et al. Lower Platelet Reactivity is Associated with Presentation of Unstable Coronary Artery Disease. 20. De Ruiter QM, Gijsberts CM, Schilham AM, Moll FL, van Herwaarden JA. Radiation Awareness for Endovascular Abdominal Procedures: an Instant Risk Chart for Daily Practice. 21. De Hoog VC, Lim SH*, Bank IEM*, Gijsberts CM, Ibrahim I, Kuan WS, et al. Ethnic Differences In Clinical Outcome Of Patients Presenting To The Emergency Department With Chest Pain. 22. De Hoog VC, Lim SH*, Bank IEM*, Gijsberts CM, Ibrahim I, Kuan WS, et al. HEART Score Performance In Asian And Caucasian Patients Presenting To The Emergency Department With Suspected Acute Coronary Syndrome. 23. Zhang Y, Vernooij F, Ibrahim I, Ooi S, Gijsberts CM, Schoneveld AH, et al. Extracellular Vesicle Proteins Associated with Systemic Vascular Events Correlate with Heart Failure: an Observational Study in a Dyspnoea Cohort.

* Indicates shared authorship.

338


Appendix

Dankwoord Promoveren doe je niet alleen. Graag wil ik hier een aantal mensen bedanken die direct of indirect een belangrijke bijdrage hebben geleverd aan de totstandkoming van dit proefschrift. Geachte promotor, prof. dr. D.P.V. de Kleijn, beste Dominique, graag wil ik je bedanken voor het vertrouwen dat je in mij hebt gehad om de samenwerking tussen de Utrechtse en Singaporese coronaire biobanken op te gaan zetten. Het op poten krijgen van de UNICORN was niet niks, maar we hebben het voor elkaar gekregen. Door je persoonlijke betrokkenheid in deze samenwerking heb nooit het gevoel gehad er alleen voor te staan. Je ongekende gastvrijheid in Singapore evenals de vele etentjes en wijntjes heb ik erg gewaardeerd. Geachte promotor, prof. dr. F.W. Asselbergs, beste Folkert, na kilo’s M&M’s en oneindig veel UCORBIO vergaderingen is UCORBIO dan toch echt gaan lopen. Zonder jouw inzet had UCORBIO niet bestaan. Ik wil je bedanken voor je enthousiasme, ambitie en je drijfkracht vanuit de klinische cardiologie om de coronaire biobank te maken tot wat hij is. Wat jij in je hoofd hebt, gebeurt; het liefst gisteren. Ik hoop dat ik ooit half zo efficiënt kan werken als jij. Geachte copromotor, dr. I.E. Hoefer, beste Imo, onze wekelijkse overleguurtjes begonnen altijd met even klagen over hoe traag alles gaat… Hoewel je in het begin nogal sceptisch was over de totstandkoming en voortgang van UCORBIO is het toch allemaal goed gekomen. Jij en Folkert, komisch duo, hebben Jonne en mij gecoacht bij het opzetten van UCORBIO. Ik ben erg blij dat je met het idee kwam om de UPOD aan UCORBIO te koppelen, daar zijn mooie stukken uit voortgekomen. Bedankt voor je begeleiding de afgelopen jaren. Geachte copromotor, dr. H.M. den Ruijter, beste Hester, ik vind het heel erg leuk dat je, laat in het traject, toch nog mijn copromotor bent geworden; ontzettend verdiend! EpiQueen van het lab, ook al wil je misschien van die reputatie af… Ik heb zoveel van je geleerd; met jouw to-the-point aanpak, kan geen enkel paper mislukken. Ik ben je dankbaar voor de red-de-promotie-van-Crystel projecten die we zijn gaan doen toen UCORBIO nooit van de grond leek te komen. Wie had gedacht dat het een heel deel van mijn boekje zou gaan beslaan?! Bedankt voor de gezellige samenwerking. Ik hoop dat we in de toekomst nog wat projecten samen kunnen doen. Geachte prof. dr. G. Pasterkamp, beste Gerard, ondanks dat je geen officieel lid meer bent van mijn promotieteam voelt dat toch wel zo. Ik wil je bedanken voor de prikkelende gesprekken, zeer kritische blik op de stukken en alle mogelijkheden die we binnen de experimentele cardiologie heb gekregen om UCORBIO op te zetten. Ik kan het woord

339


Appendix

“fishing expedition” niet meer horen, maar ik zal er altijd aan blijven denken als ik een paper schrijf! Leden van de beoordelingscommissie: prof. dr. A.W. Hoes, prof. dr. F. Zijlstra, prof. dr. M.L. Bots, prof. dr. P.A.F. Doevendans en prof. C.S.P. Lam. Bedankt voor het beoordelen van mijn manuscript. Geachte dr. M.J.M Cramer, beste Maarten Jan, bij jou heb ik tijdens mijn studie kennis gemaakt met het wetenschappelijk onderzoek. Ik wil je bedanken voor je grenzeloze enthousiasme en interesse, zonder jouw voorzet was dit proefschrift er niet geweest. Ik vind het erg leuk dat je coauteur bent op het follow-up paper! Geachte dr. P. Agostoni, beste Ago, bedankt voor de belangrijke rol die je hebt gespeeld op de HCK bij het opzetten van UCORBIO. Geachte dr. H.M. Nathoe, bedankt voor uw inzet in de opstartfase van UCORBIO. Geachte dr. M. Voskuil en dr. Z.H. Rittersma, Michiel en Saskia, bedankt voor het plaatsnemen in de eventbeoordelingscommissie van UCORBIO. Geachte dr. Y. Appelman, beste Yolande, bedankt voor je input voor het man-vrouw stuk. Lieve Jonne, samen begonnen we aan een traject waar we geen idee van hadden. Jij kwam uit de kunstwereld, ik had geneeskunde gestudeerd. Beiden hadden we geen benul wat een biobank was, laat staan hoe je dat zou moeten opzetten en runnen. Met het nodige vallen en opstaan hebben we er toch iets heel moois van gemaakt! Je weet het: dit boekje is net zoveel van jou als van mij. Je mag trots zijn op wat we de afgelopen jaren hebben neergezet. Bedankt voor een fantastische samenwerking; we hebben gelachen en gehuild en ik had me geen betere UCORBIO-partner-in-crime kunnen wensen. Ik ben onder de indruk van hoe jij je mannetje weet te staan tussen die haantjes van bazen. Je laat je niet onder de voet lopen en je weet waar je heen wilt en dat voor iemand die net 14 is ;)! Ik vind het erg fijn om te zien hoe je van een uitvoerder tot een coördinator bent uitgegroeid en hoe je kwaliteiten steeds beter op waarde worden geschat. Ik wens je een mooie carrière toe, waarin we hopelijk nog eens wat projectjes samen kunnen doen! Torenbuddies: Sander, Marten, Vince, Quirina, Ingrid, Bas, Joyce, Aisha, Saskia, Amir, Jelte en Geert. Bedankt voor de collegialiteit in de Toren. Ik heb genoten van de manier waarop we werk en after-werk hebben gecombineerd. Bedankt voor de gezelligheid op de vrimibo’s, ski-tripjes, wakeboardtripjes en het Tower Weekend. Ik kan over ieder van jullie een boek volschrijven met bedankjes, maar ik denk dat we dat op de borrel wel even bespreken.

340


Appendix

Vince, Ingrid en Aisha, bedankt voor de gezellige tripjes naar Singapore (en Borneo, Bali, Bintan, Australië en Dubai). Lieve Karlijn, ongekroonde R-Queen, ik heb ontzettend veel van je geleerd en genoten van onze samenwerking. Lieve Bernadet, je bent de eerste en enige student geweest die ik wetenschappelijk heb begeleid. Maar als ze allemaal zo goed waren als jij had ik er meer moeten hebben. In 3 maanden tijd stond er een paper, wat niet lang daarna gepubliceerd is. Ongekend, en het was ook nog eens heel gezellig! Ik wens je het allerbeste in je verdere carrière en hoop dat we elkaar nog eens tegenkomen. Alle overige coauteurs, bedankt voor de vruchtbare samenwerking! Alle collega-promovendi, stafleden en technici van de Experimentele en Klinische Cardiologie, bedankt voor de samenwerking, gezelligheid en interesse. Beste Arjan Schoneveld, bedankt voor de ondersteuning van UCORBIO vanuit het lab en hulp bij biobankzaken. Loes, Noortje en Julie bedankt voor het eindeloos verwerken van UCORBIO bloed. Beste Ineke, ontzettend bedankt voor het regelen van allerhande zaken. Maar bovenal ook bedankt voor je interesse in de voortgang van mijn promotie en persoonlijke interesses. Beste Marjolijn, ik heb je niet heel lang meer meegemaakt, maar nog bedankt voor de ondersteuning in het begin. Beste Irene, bedankt voor je ondersteuning bij UCORBIO. Loes en Merel, bedankt voor jullie rol in het includeren van UCORBIO patiënten en Joyce, bedankt voor je hulp bij het inhalen van de UCORBIO-achterstanden. Meer dan ik had kunnen bedenken, ben ik tijdens mijn promotie met research-ICT en datamanagement in aanraking gekomen. Graag wil ik een paar mensen bedanken die deze kennismaking zo prettig mogelijk hebben gemaakt en mij op weg hebben geholpen in de wondere wereld van (de achterkant van) EZIS en het SAS research data platform (RDP). Beste Aafke Jongsma, als projectleider, eerst vanuit Furore en vervolgens vanuit de Research ICT van het UMCU legde jij de fundering voor het datamanagement van UCORBIO. Ik kan me zo voorstellen dat het research ICT-technisch begeleiden van een project dat nog zo in de kinderschoenen staat niet heel aantrekkelijk is. Ik wil jou en je team bij de Research ICT van het UMCU, in het bijzonder Marc en Martine, bedanken voor het geduld met de data-newbies die we waren, fijne á-la-carte data-technische ondersteuning en Jip-en-Janneke uitleg over wat er technisch mogelijk is. Ook bedankt voor de vele keren dat ik “even” naar boven mocht komen als ik weer tegen een RDPprobleempje aanliep. De korte lijntjes heb ik erg gewaardeerd en zijn de productiviteit zeker ten goede gekomen. Beste Jan Martens, in het begin heb je wekelijks onze UCORBIO vergaderingen bijgewoond om de mogelijkheden in EZIS in kaart te brengen. Ontzettend bedankt voor

341


Appendix

je input vanuit dat perspectief en je gezellige aanwezigheid op de vergaderingen. Beste Jeroen Makkinje, bedankt voor het ondersteunen van het datamanagement van UCORBIO. As an important part of my PhD-time was spent in Singapore, I want to thank some people who have contributed to the research and the nice times outside of work over there. Dear Aruni, I couldn’t have wished for a better partner to start up the UNICORN study. I want to thank you for your tremendous efforts to get this project up and running and for the many hours that we spent SYNTAX-ing and data entering. You are an amazing person and researcher and I wish you all the best for the future. Dear dr. Mark Chan, thank you very much for leading the Singapore coronary biobank. Also, thank you for the large amounts of work you put in editing the papers. I am very happy with the results of our joint effort. I hope you are too! The co-authors on the UNICORN papers: prof. Richards, prof. Lee, dr. Low, prof. Tan, dr. Sorokin, dr. Vidanapthirana and last but not least dr. Leonardo de Carvalho; thank you all for the work you put in these manuscripts. Dear prof. C.S.P. Lam, Carolyn, in the last months of my PhD you introduced me to another project. As enthusiastic as you always are, together with dr. Lars Lund, we transformed this idea into an actual paper at incredible speed. I want to thank you for being so amazingly positive, efficient and inspirational. I hope you will think of me for future projects. All the members of Dominique’s lab group, thanks for the good times in Singapore. Naast collega’s wil ik ook mijn vrienden en familie bedanken. Zonder een gezonde basis is een promotie niet mogelijk. Lieve basisschoolvriendinnetjes: Anne, Anouk, Sarah, Aimée, Birgit en Eva. Ik besef me heel goed hoe bijzonder het is dat wij elkaar al zo lang kennen en nog steeds goed bevriend zijn. Lieve Nijmo’s: Berb, Vera, Mathia, Eva, Suus, Celmis, Dorus, Jachym, Jasper de L., Jasper F., Leon, Mart, René, Rick en Stefan. Bedankt voor jullie bijzondere vriendschap, ik hoop dat we tot we 100 zijn een jaarlijks Nijmo-weekend houden. Ik kan me geen gekkere vrienden wensen. Mart, bedankt voor de mooie voorkant van dit boekje! Lieve geneeskundevrienden: Kari, Rogier en vele anderen bedankt voor jullie vriendschap en interesse. Lieve paranimfen: Karin en Moniek. Kaar, in de eerste weken van geneeskunde merkten we dat we op 1 lijn lagen. Gedurende onze studietijd hebben we mooie reizen naar Australië en Nieuw-Zeeland gemaakt. Ik kan me maar weinig andere mensen voorstellen

342


Appendix

waar ik 2 maanden in een camper mee zou kunnen uithouden. Het lijkt soms wel telepathisch, de manier waarop wij hetzelfde kunnen denken. Ik hou van je heerlijke nerderige humor en gepassioneerde eigenwijsheid. Ik hoop dat we elkaar nooit uit het oog verliezen. Mo, in mijn tweede studiejaar kwam jij bij ons op de Domstraat wonen. Al snel waren we meer dan huisgenoten. Samen met Sarah en Margriet vormden we een hecht ‘gezinnetje’. Door de jaren heen zijn we steeds closer geworden en steeds meer dingen samen gaan doen. Ik kijk elke week weer uit naar het colaatje na de zwemles waarbij de levens weer even doorgesproken worden. Je bent een onvoorwaardelijke, goede vriendin die ik nooit meer kwijt wil. Ik ben ontzettend blij en vereerd dat jullie naast me staan. En natuurlijk bedankt voor het regelen van de boekjes-logistiek terwijl ik in Singapore zat. Lieve Papa en Mama, het mag duidelijk zijn dat ik zonder jullie nooit zover was gekomen. Wie er vandaag het trotst zijn weet ik wel! Bedankt voor de oneindige mogelijkheden die jullie ons geboden hebben. Geen hobby was te gek en alles kon altijd bij ons thuis. Als kind vond ik dat vanzelfsprekend, maar ik besef me steeds meer hoe bijzonder dat is. Tycho en Valéry, ik ben blij dat jullie mijn broertje en zusje zijn en hoop dat onze band altijd goed zal blijven. Lieve Ria, bedankt voor je bijdrage aan onze opvoeding. Lieve familie Hazen, bedankt voor het warme welkom in jullie midden. Lieve Wouter, ik ben trots op hoe ver jij het al geschopt hebt, als MDL-arts aan de bak in Tilburg. Ik hou van de manier waarop wij samen ons leven inrichten. Ik ben zo blij met jou en ik kijk uit naar onze toekomst samen.

343


Appendix

Curriculum Vitae van de Auteur Crystel Gijsberts werd op 11 maart 1987 in Nijmegen geboren als eerste kind van Jacques en Nina. Kort daarop werd het gezin aangevuld met Tycho en later ValĂŠry. Na een aantal jaar in Molenhoek te hebben gewoond, verhuisde het gezin in 1993 weer terug naar Nijmegen. In 2005 studeerde Crystel cum laude af aan het Stedelijk Gymnasium te Nijmegen. Vanwege haar interesse in de werking van het menselijk lichaam besloot zij vervolgens geneeskunde te gaan studeren aan de Universiteit Utrecht. Al vroeg tijdens de studie werd haar aandacht gegrepen door de cardiologie. Dit leidde tot wetenschappelijke keuzestages bij de echocardiografie afdeling van het UMC Utrecht gesuperviseerd door dr. Maarten Jan Cramer. Na het behalen van de artsenbul besloot zij promotieonderzoek te gaan doen bij de afdeling Experimentele Cardiologie van het UMC Utrecht onder supervisie van prof. dr. Dominique de Kleijn en prof. dr. Folkert Asselbergs. De onderzoeksprojecten bestonden voor een belangrijk deel uit een samenwerking met de National University Hospital in Singapore, waarbij een parallelle coronaire biobank werd opgericht, genaamd de UNICORN studie. Het resultaat van haar promotieonderzoek is gebundeld in dit boekwerk. Per 1 februari 2016 zal Crystel gaan werken als arts op de afdeling Cardiologie van het Radboud UMC in Nijmegen.

344




Turn static files into dynamic content formats.

Create a flipbook
Issuu converts static files into: digital portfolios, online yearbooks, online catalogs, digital photo albums and more. Sign up and create your flipbook.