How I do it - Series

“HOW I DO IT” Expandable esophageal stents AUTHORSHIP

How I do it: Expandable Esophageal Stents Chan-Sup Shim, MD Digestive Disease Center Soon Chun Hyang University Hospital Seoul Korea

Comment Jean-Francois Rey, MD Dept. of Hepatology and Gastroenterology Institut Arnault Tzanck, St. Laurent du Var, France

Summary Douglas O. Faigel, MD Oregon Health and Science University Portland, Oregon, USA

“HOW I DO IT” Expandable esophageal stents How I Do It Chan-Sup Shim

Introduction

Esophageal cancer is frequently unresectable at the time of diagnosis because of local invasion or metastatic disease. Therapy is therefore usually palliative in nature, with the major aims being relief of dysphagia, maintenance of nutrition, and occlusion of tracheoesophageal fistulas. Palliative surgery ultimately offers the best alleviation for esophageal obstruction symptoms and signs such as dysphagia and vomiting. However because of the poor prognosis, the short median survival time and the considerable morbidity and mortality rate for surgery, this approach cannot be justified in preference to less invasive nonsurgical techniques.

Recently various esophageal self-expanding metal stents (SEMSs) have been developed for palliation of malignant obstruction of the gastrointestinal tracts. The major impact of these stents relates to the ease of insertion and the potential for fewer complications compared with plastic stents. The physician’s perception of ease of stent placement is an important factor in choosing a SEMS. Because of the smallcaliber delivery system, a SEMS requires less dilation of the esophagus prior to placement.

This article mainly focuses on the technical aspects of esophageal SEMS placement, as well as considering what kind of patients are candidates, and what the physician needs to know before and after the procedure.

Specific indications and contraindications for esophageal SEMS placement

Indications The most common indication for placement of an esophageal selfexpanding metal stent (SEMS) is palliation of malignant dysphagia in patients with tumors of the esophagus and gastric cardia, that are judged to be inoperable because of extensive local or regional disease or poor functional status because of advanced age, comorbidity, or both. Dysphagia due to extraesophageal tumors such

OMED “How I Do It” Expandable esophageal stents 2

as lung cancer and malignant lymphadenopathy is also an indication. Another indication

for

placement

of

an

esophageal

SEMS

is

in

the

context

of

tracheoesophageal fistulas, which develop in patients with advanced esophageal and lung cancer and lead to continuous aspiration of saliva. Tracheoesophageal fistula is the only condition in which covered expandable metal stents may increase survival as compared with other therapies.

Contraindications Only a few years ago several clinical situations were considered to be contraindications for stent insertion, such as severe angulation of strictures, location of lesions at less than 2 cm from the upper esophageal sphincter (cervical esophagus), tumors lacking a proximal shelf to prevent migration, lesions having the risk of airway compression by the stent, luminal obstruction that could not be dilated, and horizontal orientation of a stricture at the gastroesophageal junction that would not allow good flow through a stent. However, this is a field of rapid technological and functional advances, and SEMSs have made a significant contribution to the management and palliation of dysphagia in a group of patients in whom treatment has traditionally proved extremely difficult. Currently, there is no consensus on absolute contraindications for esophageal SEMS placement, but careful patient selection is of utmost importance. Patients with a short life expectancy (less than 4 weeks), multiple metastatic disease, or peritoneal seeding should probably not be considered as candidates.

Benefits, risks and complications of esophageal SEMS placement

Informed consent should be obtained with all patients who undergo esophageal SEMS placement. Information should be given to the patient as well as the family on the expected benefits and risks, as well as on the short term and long term complications related to the procedure.

Benefits Improvement of quality of life is the most important objective of esophageal SEMS placement. Dysphagia has been shown to be effectively and reliably relieved after insertion of a SEMS. In recently published data the overall immediate success rate for palliation of dysphagia approached 90%. The ability of the patient to continue peroral nutrition is another benefit, which not only improves quality of life but possibly also the overall nutritional status of the patient. Though increased survival is not considered to be a general benefit of SEMS treatment, in the

OMED “How I Do It�

Expandable esophageal stents 3

context of tracheoesophageal fistulas the use of a covered SEMS may increase survival compared with other therapies. However, despite these benefits, it is important to consider the patient's life expectancy. Though SEMS may be more appropriate than chemotherapy or radiation therapy in patients with a life expectancy of more than 6 months, if a patient is not expected to live more than 1 month the merits of placing a SEMS are questionable.

Risks and complications Placement of an esophageal expandable metal stent can lead to several complications. Intraprocedural complications include those associated with conscious sedation, aspiration, malpositioning of the stent, and esophageal perforation. Early post-procedural complications may include chest pain, bleeding, and tracheal compression, with resultant airway compromise and respiratory arrest.

Late complications include distal stent migration, formation of an esophageal fistula, bleeding, perforation, and stent occlusion. Although most migrated stents can be retrieved endoscopically or will simply pass through the gastrointestinal tract, smallbowel obstruction develops in some patients.

Technique

Patient preparation Before stent placement, it may be helpful to obtain a barium esophagogram to define stricture location, length, angulations, and presence of tracheoesophageal fistulas. Sedation can be achieved by administering a sedative and analgesic (3–5 mg midazolam and 50 mg meperidine intravenously). It is imperative to have a gastrointestinal nursing assistant who is experienced in complex therapeutic endoscopic procedures with metal stent placement under fluoroscopy.

Stent selection The choice of stent for an individual is influenced by a variety of factors, including tumor length and position, and presence of a fistula, potential airway compromise, and personal preference of the individual inserting the stent.

The diameter and length of the stent should be determined after measuring and monitoring the length of the stricture using fluoroscopy or endoscopy. The length of the stent chosen should be at least 3 to 4 cm longer than the obstruction, to allow an adequate margin of stent on either side of the obstruction.

Many

different

types

of

stents

are

available,

each

with

slightly

different

characteristics (Table I). Commercially available esophageal stents include the Ultraflex (Microvasive, Boston Scientific, USA), the Z-stent (Wilson-Cook Medical, USA), the Polyflex (Boston Scientific), the Bonastent (Standard Sci-Tech, Korea), the

OMED “How I Do It�

Expandable esophageal stents 4

Choo stent (M.I.Tech, Korea), and the Niti-S stent (TaeWoong Medical, Korea) (Figure 1). All stents appear to be equally effective in palliating obstructive symptoms.

Placement of metal stents for tumors of the distal esophagus and gastric cardia is associated with specific problems, because the distal part of the stent projects freely into the fundus of the stomach and thus cannot fix itself to the wall. Uncovered stents are preferred for tumors at the cardia, as they are less likely to migrate [1].

However, covered stents are advocated for tumors with a high risk of fistula formation or when a fistula already exists. These are also used to avoid ingrowth of tumor through the metal mesh; this occurs in 20%–30% of patients who receive uncovered metal stents.

A SEMS deployed across the gastroesophageal junction leads to gastroesophageal reflux in most patients, causing significant morbidity [2]. Stents with an antireflux valve should be considered in this situation. Several stents with an antireflux function have been developed, including the Dua stent (Wilson-Cook Medical), which is a modified Z-stent with a polyurethane "windsock"-type valve; a modified Choo stent (M.I.Tech) with a long inner antireflux valve; and the Bonastent (Standard Sci-Tech) that has an S-type valve with "Shim's modification" (Figure 2). Though the Dua stent effectively prevents reflux without disrupting antegrade flow, this stent has some limitations. First, stent insertion is more difficult than with conventional stents.

In addition the antireflux valve can be inverted under high pressure gradients, for instance with belching and vomiting [3]. Recently we compared the dysphagia score and 24-hour ambulatory pH monitoring results for conventional SEMSs, the Dostent (M.I.Tech), and a newly developed S-type valve SEMS. The S-type antireflux valve was found to be best at preventing acid reflux according to 24-hour pH monitoring results [4].

Stents have also been developed that have an antimigration feature. The Flamingo Wallstent has a shift in the braiding angle, between the proximal and the distal part of the stent, which allows the distal part of the stent to stretch in response to peristalsis. The Ultraflex is available with proximal and distal uncovered segments which allow the normal mucosa above and below the tumor to project into the stent lumen. A newly developed covered esophageal SEMS with an uncovered proximal flange may helpful in the prevention of migration. This stent can be held in place by means of a silk thread attached from the edge of the proximal end of the stent to the patient ’ s ear via the nares. During a follow-up period of 7.5 months no stent

OMED “How I Do It” Expandable esophageal stents 5

migration was reported among 61 patients when this stent was used [5]. This design appears to prevent stent migration and to improve dysphagia in patients with malignant tumor stenosis at the esophagogastric junction, a short tumor stricture less than 5 cm in length, a soft tumor stenosis, or tracheoesophageal fistulas.

Procedure

The patient should be placed in a semi-oblique left lateral decubitus

position with both arms placed in front of the body. This position is advisable, on the one hand to provide good vantage points for marking the tumor position, and on the other hand to minimize the risk of aspiration. Stents should be inserted under endoscopic guidance with the aid of fluoroscopy.

Before stent insertion, the stenosis is pretreated by bougienage with a Savary– Gillard bougie dilator (Wilson-Cook Medical) until an endoscope with a minimum diameter of 9–10 mm can pass the stricture without resistance. This pre-dilation may increase the risk of perforation; however there is currently no consensus on how many sessions of pre-dilation can be performed before the risk of perforation increases. If the endoscope can be passed with minimal difficulty through the obstruction, this can be attempted without pre-dilation. If the stricture is so tight enough or growth of the tumor is so tortuous that the stiff guide wire for bougie dilation cannot be passed easily through the lesion, a hydrophilic biliary guide wire preloaded through a standard biliary catheter can be used to cannulate or traverse the stricture. Once the guide wire has been passed through the stricture with fluoroscopic monitoring, bougie dilation can be performed.

After the endoscope has been passed through the stricture, the distance of the tumor from the incisors and the tumor length are measured. Once the length of the stricture has been determined, the length of the stent can be chosen. The proximal and distal margins of the stricture can be shown by marking the skin, or endoscopically by tissue clips, or by the intramucosal injection of a radiopaque contrast agent. Of the various marking methods, injection with a 1 ml syringe of the lipid-soluble contrast agent lipiodol is preferable because it allows easy, accurate, and persistent marking. This allows radiopaque markers in the stent to be accurately positioned and deployed between the injected marks.

Once the upper and lower tumor borders are marked, the endoscope is withdrawn, with the guide wire, preferably a superstiff 0.038 Savary wire (or a 0.035 stiff Savary wire; Wilson-Cook Medical) being left across the stricture in the stomach.

In the case of most stents that are used at our institution (Choo stent, Bonastent, NITI-S stent), a premounted delivery device is then carefully advanced over the

OMED “How I Do It” Expandable esophageal stents 6

guide wire (Figure 3A) until the distal end of the stent is at least 2 cm below the distal margin of the tumor (Figure 3B). The stent is positioned under fluoroscopic control and then deployed by slowly retracting the outer sheath of the delivery device while maintaining the location of the inner shaft (Figure 3C). Once the stent is fully deployed, the delivery device and the guide wire are removed (Figure 3D). It is important that the inner shaft is held securely and not allowed to move, as pushing it will cause a misalignment of the stent. If retraction of the outer sheath is interrupted, the stent should be reloaded back into the outer sheath and the whole delivery device should be removed to perform the procedure from the beginning. If the ‘olive’ tip of the delivery device catches the distal part of the stent or the inside of the stent lumen due to the tight stricture, wait for the stent to expand to some extent or gently move the whole delivery device back over the ‘olive tip’ and gently remove the delivery device. If by mistake the stent is only partially deployed, along less than 50% of its length, from the delivery device, the stent can be repositioned by immobilizing the inner shaft and pushing the outer sheath gently until it reaches the distal marked end of the inner shaft. Thanks to their mechanical properties, most SEMS, whether fully or partially expanded, unexpanded, or migrated after release from the delivery device, are easy to reposition or remove endoscopically. Therefore, if a stent is completely deployed erroneously, deep down and far from the stricture, a forceps, inserted through the working channel of the endoscope can pull the lasso attached to the end proximal from the stricture (or in the case of stents with no lasso, the upper rim) for repositioning, causing the stent’s radial diameter to decrease.

To facilitate accurate deployment of certain stents (Ultraflex), the endoscope can be re-inserted to the proximal end of the stricture alongside the mounted but undeployed

stent

to

monitor

the

delivery

process

both

endoscopically

and

fluoroscopically. After deployment, the endoscope is passed into the proximal portion of the stent, but not through the stent, to assess stent position relative to the tumor while avoiding dislodgement.

Special care should be taken when stenting in the cervical esophagus. A placement that is too proximal may result in choking and/or aspiration. The patient may also feel an intolerable foreign body sensation especially if encroachment on to the cricopharyngeus occurs. In these circumstances, positioning under fluoroscopic control has been advocated. Recently a modified SEMS with a proximal funnel of shorter length has been developed for cervical stenting (M.I.Tech. Korea). Though data are as yet preliminary, the results appear to be promising [6].

OMED “How I Do It” Expandable esophageal stents 7

Post-procedure observation and care Following stent insertion, chest radiography should be carried out to verify the position of the stent and to check for signs of perforation. Stent expansion can best be confirmed by chest radiography. The following day, endoscopy can be done to ascertain the location of the stent. To prevent dislodgment, the endoscope should not be passed through the length of the stent at this time.

Following stent placement, patients can experience varying degrees of chest pain and discomfort. This is usually controlled with simple analgesia, and occasionally opiates. It is important to ascertain whether or not the chest pain is related to acid reflux.

Patients with esophageal stents must modify their diet to prevent large boluses of food from becoming impacted within the stent. Most patients will not be able to tolerate a solid diet immediately following stent insertion. Diet should be advanced in a stepwise pattern. Leafy or raw vegetables, which could result in stent occlusion, should be avoided. Patients should be counseled regarding nutrition and choosing food of appropriate consistency for avoiding food impaction.

If a stent without an antireflux valve is positioned across the gastroesophageal junction, strict antireflux precautions and aggressive acid suppression are needed to prevent gastroesophageal reflux and aspiration. Patients with such a stent should be placed on high dose proton pump therapy indefinitely. Additional precautions should be taken, such as elevating the head of the bed and avoiding recumbency within 3 hours after a meal.

Stent occlusion may result from an impacted food bolus, which can be dislodged endoscopically. Tissue-related stent occlusion may be due to tumor ingrowth, tumor overgrowth, or tissue hyperplasia. Treatment options to restore luminal patency include placing a new stent through the previous stent, ablative techniques such as argon plasma coagulation, and mechanical debridement.

Conclusion

Esophageal SEMSs are now fully established as a management for palliation of esophageal obstruction. With appropriate selection and deployment they can effectively reduce obstructive symptoms and improve the patient’s overall quality of life. However a multidisciplinary team approach must precede palliation. Patient

OMED “How I Do It” Expandable esophageal stents 8

selection and choice of device that are appropriate, stricture characterization, and communication of expectations among the physician, the patient and the patient’s family, are critical to successful endoscopic palliative therapy.

Over the past several years as innovative techniques and devices have been developed, esophageal metal stenting has become simpler and more convenient for the endoscopist, as well as more comfortable for the patient. I hope that this article may provide concise and useful information to physicians in this field.

OMED “How I Do It�

Expandable esophageal stents 9

Figures

Figure 1 Various esophageal stents. Left to right: Ultraflex, Z-stent, Polyflex, Niti-S stent, Choo stent, and Bonastent.

Figure 2

Stents with anti-reflux valves. A The Dua Z stent with a long ‘windsock’ type valve,

B the anti-reflux Choo stent with a S-type anti-reflux valve, C the BONASTENT with a long flexible and stable S-type anti-reflux valve.

Figure 3 Stent deployment.

A The delivery device is pre-mounted over the guide wire.

B The end of the stent is placed at least 2 cm below the distal margin of the tumor. outer sheath is retracted, the stent slowly expands.

C As the

D Once the stent is fully deployed, the

guide wire and delivery device are removed,.

OMED “How I Do It�

Expandable esophageal stents 12

References 1

Adam A, Ellul J, Watkinson AF et al. Palliation of inoperable esophageal

carcinoma: a prospective randomized trial of laser therapy and stent placement.

Radiology 1997; 202: 344–348 2

Acunas B, Rozanes I, Akpinar S et al. Palliation of malignant esophageal

strictures with self-expanding nitinol stents: drawbacks and complications.

Radiology 1996; 199: 648–652 3

Dua KS, Kozarek RA, Kim J et al. Self-expanding metal esophageal stent

with anti-reflux mechanism. Gastrointest Endosc 2001; 52: 603–613 4

Shim CS, Jung IS, Cheon YK. Management of malignant stricture of the

esophagogastric junction with a newly designed self-expanding metal stent with an antireflux mechanism. Endoscopy 2005; 37: 335–339 5

Shim CS, Cho YD, Moon JH et al. Fixation of a modified covered esophageal

stent: its clinical usefulness for preventing stent migration. Endoscopy 2001; 33: 843–848 6

Shim CS, Jung IS, Bhandari S et al. Management of malignant strictures of

the cervical esophagus with a newly designed self-expanding metal stent.

Endoscopy 2004;36: 554–557

OMED “How I Do It” Expandable esophageal stents 13

“HOW I DO IT” Expandable Esophageal Stents Comment Jean-Francois Rey

Self-expanding metal stents (SEMSs) are a major step in esophageal stenting. We have come a long way from the Celestin tube which, although very useful, is rather difficult to insert (having a 10% complication rate in the 1970s). Professor Shim’s contribution is an excellent reminder of the new possibilities.

Our practice in Europe is somewhat similar to the Korean experience except that it has been delayed by the slower development of fully covered metallic stents; for many years only partially covered metallic stents were available. Today we have a choice between completely covered metallic stents and Ultraflex devices where the last 2 cm at both ends are not covered. For most malignancies we are still using Ultraflex stents, as we think they embed more firmly and the risk of migration is lower.

We mostly use the fully covered Korean stent in the case of indications arising from benign stricture or fistula after gastric bypass. This latter indication will become more common with the development of this surgery. Thanks to the improvement in technique, there is now no real contraindication and we are able to stent even when only a limited space from the upper esophageal margin is available.

From the technical point of view, we rely only on endoscopic esophageal examination and the use of propofol sedation. In our own practice we have never used pre-stenting dilation in order to decrease migration, and we mark distal and proximal margins of the tumor with clips. Delivery of the stent is monitored fluoroscopically with direct endoscopic confirmation of the upper limit.

Patients are monitored the following day with endoscopy in order to verify correct positioning and the full deployment of the sent. In the case of benign fistula or benign stricture, the stent is removed after 3 months.

In summary, the use of SEMSs improves the endoscopic possibilities for esophageal treatment.

OMED “How I Do It” Expandable esophageal stents 14

“HOW I DO IT” Expandable Esophageal Stents Summary Douglas O. Faigel

Expandable esophageal stents provide the quickest and most reliable nonsurgical palliation of dysphagia from malignancy. Dr. Shim in his article does an excellent job of describing the current technology and the technique for stent insertion. He also gives us a glimpse of the future in his description of several new stents that are not yet widely available.

Dr. Rey provides additional insight in his comments. Similarly to Dr. Rey, it has been my practice not to perform a pre-dilation, to mark the stricture with metallic clips, and to use primarily the Ultraflex stent (Boston Scientific, USA).

While most authors use fluoroscopy, nonfluoroscopic monitoring during stent placement has been described by Dr. Todd Baron at the Mayo Clinic. In this technique, the proximal release Ultraflex stent is used and an endoscope is passed next to the insertion catheter. Using only endoscopic visualization, the proximal end of the stent is positioned above the stricture and continuously observed during deployment of the proximal release Ultraflex stent. Since only the proximal edge of the stent can be monitored, precise tumor measurements are required to select a stent of proper length and thus ensure adequate coverage of the stricture.

While palliation of malignant strictures accounts for the vast majority of stent placements, their use in benign strictures, previously considered a contraindication, has gained increasing acceptance. This has largely been due to the availability of an expandable plastic stent, the Polyflex stent (Boston Scientific), which has enhanced removability. This stent is approved by the US Food and Drug Administration for treatment of benign and malignant strictures as well as for its removability. The long-term efficacy of this stent for treatment of benign strictures has not been determined.

I thank Drs. Shim and Rey for their excellent articles and for their contributions to the field.

OMED “How I Do It” Expandable esophageal stents 15

“HOW I DO IT” Endoscopic mucosal resection (EMR) in the esophagus AUTHORSHIP

How I do it:

Endoscopic mucosal resection (EMR) in the esophagus

Horst Neuhaus, MD Department of Internal Medicine Evangelisches Krankenhaus Düsseldorf Germany

Comment Hiroyasu Makuuchi, MD Professor and Chairman Department of Digestive Surgery Tokai University School of Medicine Isehara Japan

Summary Kenneth K. Wang, MD Director of Advanced Endoscopy, Professor of Medicine Division of Gastroenterology and Hepatology Mayo Clinic. Rochester Minnesota USA

WEO “How I Do It” Endoscopic mucosal resection (EMR) in the esophagus 1

“HOW I DO IT” Endoscopic mucosal resection (EMR) in the esophagus How I Do It Horst Neuhaus

Indications Endoscopic mucosal resection (EMR) should be considered in patients with early esophageal neoplastic lesions, provided that these are not associated with lymph node metastasis. The risk mainly depends on the depth of tumor infiltration. According to Japanese studies of esophagectomy specimens, the incidence of lymph node metastasis is less than 2% if squamous cell cancer (SCC) is limited to the upper or mid layer of the mucosa (T1m1–2). Comparable data were reported from western countries for early Barrett’s adenocarcinoma. In addition EMR is a promising technique for removal of high grade intraepithelial neoplasia (HGIN) to prevent the progress to cancer which will otherwise occur in every third patient within a few years. In contrast to ablative techniques, EMR has the advantage of providing a specimen for histopathological diagnosis. Compared with biopsy studies this approach reveals more advanced tumor stages in terms of grading and vertical neoplastic infiltration in 10%–20% of patients. In the case of submucosal invasion, esophagectomy is indicated because of the risk of lymph node metastasis. In contrast, complete resection of SCC or Barrett’s adenocarcinoma/HGIN limited to the mucosa can be considered as a definitive treatment. Close follow-up is required for early diagnosis and treatment of residual, recurrent, or metachronous neoplastic lesions.

Because of the importance of histopathological diagnosis, we consider EMR for all suspicious lesions when biopsy studies have revealed HGIN or early adenocarcinoma, provided that they fulfil the criteria of early neoplasia (type 0) in the Paris classification. Endoscopic ultrasound (EUS) is performed prior to EMR to exclude pathological lymph nodes. We do not use high frequency EUS for discrimination between mucosal and submucosal tumor infiltration because of its limited accuracy. The best candidates for EMR are patients with a small area of early SCC or focal lesion in a short-segment Barrett’s esophagus. In these cases surveillance of non-neoplastic areas after EMR is easy, or EMR can be used for complete eradication of Barrett’s epithelium. Widespread EMR or total removal of WEO “How I Do It” Endoscopic mucosal resection (EMR) in the esophagus 2

Barrett’s epithelium is required in patients with multifocal lesions. This approach is technically more challenging in long segments and is associated with a considerable risk of esophageal stricture formation. As an alternative to extended resection, EMR can be combined with ablative methods such as photodynamic therapy (PDT), argon plasma coagulation (APC) or high power radiofrequency (the BARRX system).

Patients with residual areas of non-neoplastic Barrett’s epithelium after EMR of a focal lesion have to undergo long-term surveillance for early detection of metachronous neoplasia. Alternatively complete eradication of Barrett’s can be considered, but this approach will also require follow-up endoscopy because of limited data on the long-term outcome. Therefore patients with early neoplasia in a longsegment Barrett’s esophagus must be informed about the frequent need for repeat interventions, the risk of stricture formation and that information about long-term results is still limited. The risks and benefits of the endoscopic approach must be balanced against the results of subtotal esophagectomy.

Limited esophageal resection may be an alternative but lymph node resection can be incomplete. This is of concern in patients with an increased risk of lymph node metastasis due to deeper tumor infiltration not diagnosed, for example, by EMR prior to surgery.

Contraindications EMR of early esophageal neoplastic lesions should not be performed in the presence of: •

Obvious tumor infiltration into the submucosa or even deeper layers of the esophageal wall according to endoscopic appearance (lack of type 0 criteria) and/or EUS findings or

•

enlarged local lymph nodes not related to local inflammation (e.g. not locally inflamed due to reflux esophagitis); EUS guided FNA can be considered in undetermined cases.

Relative contraindications are widespread or multifocal lesions which require circumferential or extended longitudinal resections. Limited data for widespread EMR indicate a high risk of stricture formation. Strictures may require repeated dilation and may have a negative impact on surveillance due to difficulties of detection and treatment of residual or recurrent lesions.

Technique of EMR Several endoscopic techniques have been advocated for mucosal resection. In patients with early esophageal neoplasias, these should allow a safe and targeted resection of focal lesions as well as widespread removal of multifocal areas or even complete eradication of Barrett’s epithelium. These aims can be best achieved with the endoscopic mucosal resection cap (EMR-C) procedure or the ligateand-cut technique. Recent comparative trials have shown similar results for the two methods. Before WEO “How I Do It” Endoscopic mucosal resection (EMR) in the esophagus 3

EMR, electrocautery dots should be marked using the tip of a snare or a needle-knife around the circumference of focal lesions, with a safety margin of at least 2 mm. High resolution endoscopes and narrow-band imaging (NBI) or chromoendoscopy are helpful for a precise delineation of the tumor margins. Marking may be not required in the case of easily visible neoplastic nodules or planned widespread EMR. EMR-C is performed with a cap (Olympus, Tokyo, Japan) mounted on the tip of a forward-viewing endoscope. We usually prefer a hard cap with an oblique distal end. A large-bore soft cap provides a larger specimen but targeting of lesions can be more difficult, particularly if the esophageal lumen is narrow. To decrease the potential risk of perforation, saline solution is injected into the submucosal layer to separate the mucosa from the muscle layer. We add a small amount of indigo carmine 0.1% to give the solution a slightly blue color. The dye helps to color the submucosa which facilitates delineation of the lateral resection margins and discrimination from the red layer of the muscularis propria. The amount of solution injected depends on the diameter of the targeted area, the size of the cap and the degree of lifting. Intensive injection may reduce the risk of perforation, particularly when a large cap is used, but will reduce the size of the specimen. A specially designed thin snare is then “pre-looped” in the cap; for this purpose normal mucosa of the esophagus or the stomach is partially sucked into the outlet of the cap. The previously inserted snare is then opened and positioned into the ridge at the internal distal end of the cap. The targeted lesion can then be sucked into the cap until the cap has been completely filled with tissue, which is indicated by a “red-out” of the image. After firm strangulation by the snare wire, the suction is released to let the pseudopolyp thus created out of the cap. For subsequent resection we prefer the endocut mode (Vaio; Erbe, Tübingen, Germany) with level 3, cutting interval 6, and cutting duration 1. The specimen can then be easily retrieved by aspiration into the cap and withdrawal of the endoscope. The procedure can be repeated for piecemeal resection by positioning the cap laterally to the previous resection area of which only the outer margin should be sucked into the cap, to decrease the risk of perforation.

With the ligate-and-cut technique, an artificial polyp is created at the site of the lesion using an esophageal varices ligator. There seems to be no need for submucosal injection according to previous reports on this technique. A transparent cap loaded with a rubber band is used for suction of the targeted lesion into the cap with subsequent release of the band. A polypectomy snare is then positioned above or below the band for resection of the created pseudopolyp. We prefer to cut below the band to obtain a larger specimen. The Multiband Mucosectomy (MBM) kit (Duette system; Cook Medical, Bloomington, Indiana, USA) is useful for piecemeal resection. The snare of the kit can be kept in the accessory channel of the endoscope and it can be easily positioned over the ligated tissue. A single system facilitates the performance of up to six resections without the need to remove the endoscope. This technique is obviously faster and easier to perform than the cap procedure in particular WEO “How I Do It” Endoscopic mucosal resection (EMR) in the esophagus 4

for widespread EMR. However the visibility through the MBM cap is inferior to that for the fully transparent cap for EMR-C, which makes it more difficult to target focal lesions and avoid tissue bridges between multiple resection areas. In addition the size of the resected specimen is smaller for EMR done using the MBM system compared with EMR-C done using a large cap.

We repeat EMR in the same session until all coagulation markings for a focal lesion have disappeared or until all of the area targeted for widespread resection has been removed.

Management of complications Bleeding from small submucosal vessels may occur during EMR. We use the cap on the tip of the endoscope for compression and localization of the bleeding site. This approach usually makes it easy to grasp the vessel with a coagulation forceps (hot biopsy forceps or specially designed forceps for hemostasis) and to use diathermy (the soft mode of the Vaio system) for definitive hemostasis. During coagulation, we pull the entrapped tissue with the vessel towards the cap to decrease the risk of deeper diathermic damage to tissue. This technique nearly always achieves hemostasis. In the case of failure hemoclips can be used, but their application makes it more difficult to continue with piecemeal EMR. We never had a case of esophageal perforation which had been reported as rare and severe adverse event of EMR. If any small defect had occurred, we would have tried to close it with clips with confirmation by esophagography and a computed tomography (CT) scan to exclude a larger extraluminal fluid collection. Conservative management with antibiotic prophylaxis and total parenteral nutrition seems to be justified if there is no evidence of mediastinitis. However close collaboration with a surgical consultant is mandatory.

The risk of stricture formation depends on the extent of EMR and seems to be particularly high in the case of circumferential resection in a single session. Stepwise widespread EMR done in several sessions may reduce the incidence but EMR becomes more difficult if the esophageal lumen has become narrower due to scar formation. We manage symptomatic strictures caused by EMR by using bougienage in the same way as for peptic stenoses. Unfortunately this frequently has to be repeated due to recurrence, in some patients for example, up to five times at weekly intervals.

Follow-up After EMR the patients have an overnight hospital stay. We allow them to drink but not to eat, so that esophagogastroduodenoscopy (EGD) can be performed if there is any evidence of delayed bleeding.

Asymptomatic patients are discharged the morning after EMR with no further restrictions and with continuation of medication with proton pump inhibitors (PPIs) at a standard dosage twice daily.

WEO “How I Do It�

Endoscopic mucosal resection (EMR) in the esophagus 5

According to our follow-up protocol, EGD is repeated every 3 months in the first year after EMR and then every 6 months. We consider resection of neoplasia in Barrett’s to be complete if there is no macroscopic evidence of residual neoplastic tissue and histology confirms tumor-free margins of resected areas and neoplasia-free EMR scars. Residual Barrett’s epithelium is monitored by endoscopy, with biopsies from suspicious lesions and routine biopsies according to the Seattle protocol.

We are currently participating in a multicenter trial to study the efficacy and safety of high power radiofrequency ablation (BARRX) of residual Barrett’s epithelium after EMR of early neoplasia.

Residual, recurrent, or metachronous neoplastic lesions can be usually resected by EMR without changing the techniques. However the removal of specimens can become more difficult due to scar formation and insufficient lifting of the mucosa. Ablation by APC, PDT or BARRX can be done but we recommend that multiple biopsies of the targeted area should first be done, to exclude more advanced neoplastic tumor stages.

Conclusions EMR is a promising technique for local treatment of HGD or mucosal cancer in patients with Barrett’s esophagus or SCC. Several series demonstrate that different techniques of EMR are feasible and safe for this indication. The EMR-cap technique and the ligate-and-cut procedure seem to be equivalent. The selection depends on the size of the targeted area and the local expertise.

In contrast to ablative methods such as PDT, EMR allows histopathological evaluation and frequently reveals more advanced tumor stages. Limitations relate to malignant lesions or HGD which can not be endoscopically identified or to multifocal lesions. EMR is technically possible for large areas of Barrett epithelium with complete eradication but must be evaluated in appropriate trials. Alternatively EMR can be combined with ablation by PDT, APC or BARRX, but long-term data regarding this approach are limited.

EMR promises to be curative after complete local remission. The risk of recurrent, synchronous, or metachronous lesions has to be compared with the morbidity and mortality of esophagectomy or alternative nonsurgical methods. EMR with curative intention should be carried out only in appropriate prospective trials which include close follow-up examinations.

WEO “How I Do It” Endoscopic mucosal resection (EMR) in the esophagus 6

“HOW I DO IT” Endoscopic mucosal resection (EMR) in the esophagus Comment Hiroyasu Makuuchi

Introduction As a result of the advances in endoscopic instrumentation and the greater use of iodine staining, the detection of early stage esophageal carcinoma has been rapidly increasing. Endoscopic mucosal resection (EMR) techniques have gained popularity while endoscopic submucosal dissection (ESD) methods have also been introduced in some institutions. The acquisition of skills in early detection of esophageal carcinoma and in EMR techniques is now essential for endoscopists.

While EMR is regarded as the treatment of choice for early esophageal carcinoma, informed consent must be obtained, with the patient having been made aware of the possibilities of perforation, bleeding, stenosis, and other complications, as well as, in rare cases, lymph node recurrence, distant organ recurrence, local recurrence and metachronous metastases.

EMR yields an excellent prognosis, with a 97.1% 5-year survival rate in our experience.

Indications Cases in which EMR and ESD are indicated are those with endoscopically resectable lesions not associated with lymph node metastases or organ metastases. The best indications are high grade intraepithelial neoplasia (HGIN), lesions of carcinoma in situ (Tis), and those with a depth of invasion as far as the proper mucosal layer (T1a, m2), because lymph node metastases are unlikely in such lesions.

Since lymph node metastases are present in approximately 10% of cases with lesions reaching the muscularis mucosae (T1a, m3) and approximately 15% of cases with invasion up to the upper one third of the submucosa (T1b, sm1), these are considered to be relative indications for the procedures.

Histopathological examinations in EMR/ESD cases showed that when the depth of invasion is m3 or sm1, with lymph vessel invasion (ly(+)) or a significant infiltration (INF-γ), the probability of lymph node metastasis is even higher in 39% of cases. It is necessary to examine the depth of invasion and to assess the possibility of lymph node metastasis by diagnostic imaging using endoscopic ultrasound (EUS) or computed tomography (CT) before performing EMR or ESD. However, the diagnostic WEO “How I Do It” Endoscopic mucosal resection (EMR) in the esophagus 7

accuracy remains at only around 80%, suggesting the difficulty of assessment.

•

The indications are the same for adenocarcinoma in Barrett’s esophagus. The muscular layer below the columnar epithelium is an acquired muscular layer formed by glandular tissue and can be regarded as a basement membrane of squamous epithelium. It is best to determine indications based on the muscularis mucosae of the esophagus as standard.

•

The size of the lesion should be under 3 cm in longitudinal extent. Lesions of more than 5 cm in length would present greater technical difficulties and there will be some parts with deeper invasion and increased risk of lymph vessel invasion and lymph node metastasis. Of course the procedures are indicated if the entire lesion is Tis or HGIN.

•

Lesions should be no more than two-thirds of the circumference of the esophagus. Lesions with a greater extent around the circumference have a higher risk of postoperative stenosis.

Resection around the entire circumference should be avoided.

•

Many esophageal carcinomas have multiple lesions. It is possible to resect about three lesions in one day. When there are more, it is recommended that resection is carried out on separate days, in which case it is better to start from the lesions on the anal side. Resection from the oral side can cause stenosis, making it difficult to approach the remaining lesions.

Contraindications EMR and ESD are contraindicated in cases of: •

Massive tumor invasion in the submucosal layer: lymph node metastases are present in 40%– 50% of these cases.

•

Lymphadenopathy that is obviously metastasis according to EUS or CT findings. EMR and ESD are then contraindicated even when the depth of invasion is T1a, m3 or T1b, or sm1. In particular,

carcinomas

detected

by

EUS-fine-needle

aspiration

(FNA)

represent

contraindications.

•

Patients with a tendency to bleeding (e.g. patients with cirrhosis or esophageal varices).

Leukemia or hemophilia are contraindications. The procedures are contraindicated in patients on anticoagulant therapy but may be performed after halting the treatment.

•

Lesions inside or partially inside an esophageal diverticulum: these cases are susceptible to perforation.

•

Adhesion between the mucosa containing the lesion and the lamina propria mucosae, i.e. where there is no elevation of the mucosa when saline solution is injected. Careful manipulation WEO “How I Do It” Endoscopic mucosal resection (EMR) in the esophagus 8

by an experienced endoscopist is required as esophageal perforations occur easily in these cases.

•

Esophageal varices; but EMR or ESD may be possible after these have been treated with endoscopic injection sclerotherapy (EIS) or endoscopic variceal ligation (EVL). Esophageal varices generally represent contraindications, because the patients are prone to bleeding and have a higher risk of adhesion of mucosa and muscularis propria. Ablation may be performed carefully by an experienced surgeon.

•

Massive invasions into the submucosal layer (T1b, sm2/sm3). However EMR or ESD may be carried out in these patients if curative surgery or chemoradiotherapy is impossible due to their poor general condition, provided that no lymphadenopathy is found by EUS or CT.

Techniques of EMR and ESD The most common EMR methods are the endoscopic mucosal resection cap (EMR-C), endoscopic esophageal mucosal resection (EEMR) tube, and strip biopsy methods. ESD is a new means of ablating the submucosal layer by dissection of the mucosa around the lesion. The strip biopsy method often results in piecemeal resection. The ESD method requires a high level of skill.

With the EEMR method, the use of a soft cap rather than a hard cap reduces the risk of perforation.

With a larger lesion, a large-bore soft cap with an oblique distal end is useful but may be more difficult to pass through the orifice of the esophagus. We prefer the EEMR tube method (the “four-step” method).

However the scarce availability of the EEMR tube (Create Medic Co., Yokohama, Japan) outside of Japan is a problem. Briefly, the approach is as follows: 1

Identify the area of the lesion using iodine staining.

2

Inject saline solution with 0.02% indigo carmine into the submucosal layer just below the lesion until the entire lesion is elevated.

3

Insert the EEMR tube, mounted on an endoscope, and position it immediately in front of the lesion. Pass the snare through the side channel of the EEMR tube and open it above the targeted lesion.

4

Use the endoscope to apply aspiration inside the EEMR tube to bring the mucosa into the tube.

5

Tighten the snare and resect the mucosa using the endocut mode. Remove it together with the endoscope, using aspiration.

6

The procedure can be repeated for any residual lesion.

7

If the residual lesion is small, electrocautery can be applied. Insert the grasping forceps through WEO “How I Do It” Endoscopic mucosal resection (EMR) in the esophagus 9

the working channel of the endoscope and pass the snare through the side channel of the EEMR tube. Grasp the mucosa with the remaining lesion with the forceps. Close the snare and apply electrocautery.

8

Hot biopsy may be used to trim the resected margin or to remove the a residual lesion within the resected area. Argon plasma coagulation (APC) is another alternative.

Management of complications Esophageal perforation The most serious complication of EMR and ESD is esophageal perforation. With experienced staff the incidence rate is under 1%.

Prevention. The following points should be taken into consideration: •

EMR should not be carried out when there is no mucosal swelling upon injection of the saline solution.

•

Snaring of the ulcer base which has already been subjected to EMR should be avoided.

•

EMR should not be done on diverticular lesions.

Management. If esophageal perforation occurs: •

There is no need for urgent surgical intervention if there is no intrathoracic rupture. It is advisable not to try to close the perforation with clips.

(a) Administer Parenteral nutrition - the patient must not eat or drink. (b) Administer antibiotics. (c) Use intermittent aspiration with an indwelling nasogastric tube in the esophagus. The •

The perforation should close spontaneously after 10 to 14 days.

•

Continue careful surveillance, while considering possible surgical intervention including drainage if necessary.

•

Pneumothorax occurs when the mediastinal pleura is perforated causing intrathoracic rupture.

Surgical intervention is necessary in most of these cases. Perform chest drainage if it is possible to close the perforation using endoscopically placed clips. The clinician should bear in mind that a prolonged endoscopic treatment can cause tension pneumothorax.

Esophageal stenosis The risk of esophageal stenosis is particularly high when mucosal resection extends around more than WEO “How I Do It” Endoscopic mucosal resection (EMR) in the esophagus 10

three-quarters of the circumference of the esophagus. It will inevitably occur when resection extends to more than four-fifths of the circumference.

Prevention. Mucosal resection should be limited to two-thirds of the circumference to prevent esophageal stenosis.

Management. When esophageal stenosis occurs: •

Dilation should be repeated every week using a through-the-scope (TTS) balloon dilator, a mercuryfilled bougie or a tungsten-filled bougie. It should be repeated for up to 3 months in some patients.

•

Dilations must be accompanied by conservative management and care should be taken to prevent esophageal perforation.

•

We do not recommend the use of expandable metallic stents (EMSs) as there are risks of perforation upon removal, and of recurrence of stenosis after removal. If an EMS is inserted, it should be removed within 1 month.

Bleeding Arterial bleeding frequently occurs and is not preventable. For hemostasis when bleeding from arterioles is seen during EMR: 1

Grasp the bleeding site using hot biopsy forceps or hemostatic forceps and stop the bleeding using the soft coagulation mode.

2

If that is difficult, use the tip of the endoscope for compression or inject saline with epinephrine to reduce the amount of bleeding, and electrocoagulate.

Postoperative bleeding is extremely rare.

Aneurysmal hemorrhage should be treated with EIS, EVL or clipping.

Postoperative care After EMR/ESD, we allow patients to drink water but not to eat that day. A mucoprotective agent (sodium alginate, aluminum hydroxide) is administered but antibiotics are not usually used. Patients are started on a liquid diet the following day, followed by 50% liquid rice gruel (1:10–12 rice to water)) and regular rice gruel (1:5–7 rice to water). Patients are discharged after 3 or 4 days.

Follow-up

•

Endoscopic examination is repeated every 6 months. Perform EMR or APC if an area stains positively with iodine dye.

WEO “How I Do It” Endoscopic mucosal resection (EMR) in the esophagus 11

•

Endoscopic examination should be carried out with careful attention to the possibility of accompanying head and neck carcinoma.

•

In the case of T1a, m3 or T1b, or sm1 lesions, CT or EUS should be repeated every 4 months, taking care to look for any sign of lymph node recurrence.

•

Special attention should be paid to the possibility of metachronous multiple esophageal carcinomas, as these develop in 7%–8% of patients.

Conclusions

•

EMR and ESD have revolutionized the treatment of early-stage esophageal carcinoma. We believe we should try to expand the indications as much as is reasonably possible, for there is a huge difference between surgical intervention and endoscopic treatment.

•

EMR and ESD should be performed by skilled endoscopists with sufficient training.

•

Patients must be given adequate information prior to the treatment, so that properly informed consent can be given.

•

Periodic endoscopic examination using iodine staining is essential for the early detection of lesions for which EMR or ESD is indicated.

WEO “How I Do It” Endoscopic mucosal resection (EMR) in the esophagus 12

“HOW I DO IT” Endoscopic mucosal resection (EMR) in the esophagus Summary Kenneth K. Wang

Rationale and indications Endoscopic mucosal resection (EMR) and endoscopic submucosal dissection (ESD) are both directed towards complete removal of early esophageal carcinoma. They both offer the endoscopist the opportunity to examine the histology of the lesion being treated, unlike any other available ablative procedure. This distinction is critical since the depth of invasion of the esophageal cancer clearly defines the metastatic potential of the lesion, as well outlined by Dr. Makuuchi who has been a pioneer in this area. The exact depth at which metastatic risk becomes substantial has been debated. It has been disputed whether depths of invasion of m3 (lower third of the mucosa) and sm1 (top third of the submucosa) are associated with substantial risk of metastasis. A recent large Japanese study involving 464 patients with squamous cell cancers found that sm1 cancers have a metastasis potential of 53.1%. while m3 cancers also have a risk of approximately 18%. It is generally felt that if there is no evidence of lymphatic invasion, m3 lesions can be safely treated with mucosal resection. EMR can also be used as a diagnostic tool, since often (40% of cases), a nodule in Barrett’s esophagus can be misinterpreted as noncancerous rather than cancerous because of the limited depth of endoscopic biopsies. Thus the indications for EMR are diagnostic and therapeutic for esophageal cancer while ESD, because of its ability to completely resect the entire lesion, is primarily useful for its therapeutic benefits.

Lesions that are approachable with either technique should be confined to less than two-thirds of the esophageal circumference to avoid the formation of strictures after mucosal resection. It is important to note that strictures can form even after a single mucosa resection since the degree of scar formation is very individual. Although generally the lesion size should be less than 3 cm, it is permissible to resect larger areas so long as the lesion is felt to be superficial.

Tumor staging is important in the treatment of early esophageal neoplasia, with endoscopic ultrasound and computed tomography (CT) scans. EMR and especially ESD should not be performed in the setting of known metastatic esophageal cancer.

It is important to obtain a reasonable “lift” from submucosal injection. It is also more difficult to resect

WEO “How I Do It” Endoscopic mucosal resection (EMR) in the esophagus 13

areas of neoplastic tissue that have been previously resected because of the scarring that has occurred.

This situation occurs when the mucosal resection or even submucosal dissection specimen demonstrates neoplasia extending to the lateral margins of the resection.

Techniques of EMR and ESD The techniques of EMR and ESD have been performed quite differently depending on individual preferences. As indicated by Drs. Makuuchi and Neuhaus, the most commonly used approach in Europe and the United States uses the cap device (Olympus, Tokyo, Japan). The type of cap employed varies quite lot in terms of shape (oblique or flat tip), construction (soft or hard plastic), and diameter.

I find that for beginners, the ligating cap (Wilson Cook, Indianapolis, USA) is the easiest to use. This cap functions similarly to a six-band variceal ligation device and allows the endoscopist to band the lesion to form the pseudopolyp, prior to snare resection with a multiuse hexagonal snare which can be passed through the therapeutic channel of the ligating device. The ligating cap does have disadvantages, including decreased visibility, and does remove a slightly smaller piece of tissue. This device is simple to use, but is less desirable in situations where the mucosa is firm, such as after prior mucosal resection or uncontrolled reflux disease. Then a hard cap device seems to work better.

My personal preference is for the flat hard caps since the oblique caps will cut the mucosa when suctioning has been started. I prefer to suction the mucosa into the cap two or three times to be certain that the lesion is centered correctly in the specimen area and that the margins I wish to obtain can be resected. With the oblique cap, as soon as the mucosa has been lacerated, the tissue continues to enter the cap the same way and adjustments cannot be made. The soft caps gather more tissue but it is more difficult to regulate the amount of tissue coming into the cap.

ESD should only be performed by experts. It requires a good lift with a solution that can maintain the space between the mucosa and submucosa. We use hydroxypropyl methylcellulose, which is obtained as an ophthalmic solution and is sterile. We color this with indigo carmine to make the area of submucosa more apparent. This compound can maintain a cushion for at least 30 minutes. The dissection is performed with a variety of devices, amongst which our favorite is the triangle tip-knife that allows use of the points of the triangle for cutting and the flat end for cautery. It is important to initially circumscribe the lesion by cutting through the mucosa completely around the lesion, with a margin of about 2–3 mm from areas of known neoplasia. The lesion should then be lifted from the distal end using a cap (an EMR cap can be used), and dissection is done laterally in a side to side motion, taking care to stay within the submucosal space. The greatest challenge is to visualize the plane of dissection while carefully resecting the specimen in a proximal direction.

WEO “How I Do It” Endoscopic mucosal resection (EMR) in the esophagus 14

Complications Complications are not common with EMR and can generally be managed conservatively. Bleeding from small vessels is the most common adverse event. Cautery with a hot biopsy forceps works well for small vessels. The forceps should not be used to remove any tissue, simply to close and then apply cautery, then the tissue should be released. Sealing the resection site with multiple hemoclips is also often our practice after bleeding. Perforations are rare in EMR of the esophagus and I have not encountered one.

ESD is much more commonly associated with perforations, which are generally small and can be sealed with hemoclips in the majority of cases. If it is a large perforation that cannot be sealed, surgical management will be needed. Strictures usually occur within 2 weeks of the procedure and need to be managed with aggressive dilation.

Postoperative care Our EMR patients are managed as outpatients. They are monitored in the endoscopy laboratory for an hour after the procedure. They are instructed to return if they bleed or have increased pain. A clear liquid diet for 24 hours is recommended, with the instruction to advance to a normal diet as tolerated.

Often, bleeding complications appear 1–2 weeks after the procedure. ESD patients should be observed in hospital for 24 hours after the procedure to confirm that there is no perforation.

WEO “How I Do It” Endoscopic mucosal resection (EMR) in the esophagus 15

“HOW I DO IT�

Glue treatment of gastric varices AUTHORSHIP

How I do it: Glue treatment of gastric varices Ibrahim Mostafa, MD Professor of Gastroenterology and Hepatology Theodor Bilharz Research Institiute Egypt

Comment Norman E. Marcon, MD The Wellesley Hospital Division of Gastroenterology Toronto, Ontario Canada

Summary Tetsuya Mine, MD 4th Department of Internal Medicine Tokyo University Japan

“HOW I DO IT” Glue treatment of gastric varices How I Do It Ibrahim Mostafa

Indications Gastric varices are classified as: •

type I or II gastroesophageal varices, that are found as extensions of esophageal varices, or

•

type I or II isolated gastric varices.

Both are diagnosed endoscopically, either during active bleeding or in patients with a history of haematemesis and/or melaena.

Contraindications There are no specific contraindications, except for any contraindication to the upper endoscopy procedure itself.

However in patients awaiting liver transplant, the use of glue for gastric varices is contraindicated because of the risk of obliteration of the splenic vein, the portal vein or both.

Clinical scenario The patient presents at the emergency unit with haematemesis and or melaena.

The resuscitation of the patient and management of co-morbid conditions are the most important steps in dealing with gastric varices.

We carry out the following steps in patients with haematemesis: 1. The vital signs are checked.

2. We insert two wide-bore intravenous cannulae (with a central venous line if needed). Fluid resuscitation is done if necessary

WEO “How I Do It” Glue treatment of gastric varices 2

3. A complete blood evaluation is done, with, basic routine laboratory investigations. Blood transfusion is done if needed.

4. Start Intravenous infusion of Splanchnic circulation vasopressors: octreotide, terlipressin 5. Proton pump inhibitors (PPIs) are given by continuous intravenous infusion during the first 24 hours 6. A nasogastric tube is inserted and gastric lavage is done.

7. Endoscopy is done within 6 hours after resuscitation (anytime within 24 hours) as second episodes of haematemesis have a very high mortality rate compared with the first.

Informed consent: Including the risks of the procedure

Risks include aspiration, pulmonary embolism very rarely (once in last 20 years), or obliteration of the

splenic vein or the portal vein or both.

Sedation used Diprivan (propofol) may be used so that the patient is calm while the procedure is performed.

If the patient is in a very critical state, a cuffed endotracheal tube is inserted before the procedure.

I do not use topical anaesthesia.

Equipment used A prepared endoscope with channel size 2.8 mm or more to allow the passage of the injector needle is needed. Prepared here means sterilized, with the air supply connected and tested, and the suction function connected and tested.

Accessories used

•

21-gauge injection needles. A minimum of five needles must be ready in the endoscopy room, as the needle is usually blocked after the first injection.

•

23-gauge injection needles for esophageal varices

•

Band ligation sets for esophageal varices

•

All accessories used in the management of nonvariceal bleeding, such as argon plasma coagulation (APC) equipment, hemoclips, epinephrine, thermal coagulation equipment, etc, must be ready for use in the endoscopy room.

WEO “How I Do It” Glue treatment of gastric varices 3

Injection fluids These fluids are required:

• • •

Histoacryl (n-butyl cyanoacrylate) 0.5 ml per ampoule Lipiodol solution Distilled water.

The mixture of Histoacryl and lipiodol should be prepared before the procedure in patients with suspicion of gastric variceal bleeding, with each dose containing 1 ml of Histoacryl and 1 ml of lipiodol. Multiple doses should be prepared before beginning the procedure.

The procedure, step-by-step The endoscope is ready for use, with the suction, air, and oxygen functions prepared. All the accessories mentioned above are available inside the room. Facilities for endotracheal intubation are also available.

The patient lies in the left lateral position with oxygen supplied nasally.

The needle is checked and flushed with distilled water.

A complete diagnostic upper endoscopy is done as far as the second part of the duodenum (even the source of the bleeding is identified) The endoscope is retroverted to identify gastroesophageal extensions. A forward view is used to assess the best direction for injection of gastric varices.

To identify the the source of bleeding, if there is blood in the stomach, the patient is turned to the right lateral position and the endoscope is moved to get the best view without kinking the endoscope.

Assessment

of all gastric varices should be done without any help from the nurse in holding the

endoscope as any fine movement will alter the position of the endoscope during injection.

Gastroesophageal varices might be injected antegradely (at the cardia) or retrogradely. Varices at the lesser curvature are better injected retrogradely.

The gastric varix is probed to distinguish it from the mucosal folds, using the injection needle still within its sheath. One dose of the mixture is then injected, followed by 0.5–0.7 ml of lipiodol to flush out the residue of mixture inside the needle sheath.

The needle is removed from the vein, and then flushed with 2 ml of distilled water. The patency of the needle is checked.

If the needle is not blocked, the varix is probed, again using the injection needle inside its sheath, and if the varix is still soft, the injection procedure is repeated. If the needle is blocked it is changed for a new one.

The bleeding site is rinsed with water to ensure that there is no further bleeding.

If esophageal varices are present, band ligation or injection sclerotherapy is attempted in the same Session

but after obliteration of the gastric varices.

Notes

• •

There is no problem with suction of air while the needle is in the channel.

There is no damage to the endoscope.

WEO “How I Do It” Glue treatment of gastric varices 4

• • • • •

Up to three or four doses can be administered in the same session.

The patency of the needle must be checked before each injection.

If the needle is blocked at any time, it is changed for a new one.

The suction device must be turned off whilst the needle is changed.

All gastric varices must be obliterated in the first session.

Post-procedure observation, care and follow-up The patient is admitted into the gastrointestinal ward, but admission into the intensive care unit (ICU) is better for those who are hemodynamically unstable.

Splanchnic circulation vasopressors are continued for two more days (if already started before the procedure .

Endoscopic follow-up is done after 1 week. The gastric varices are again probed using the injection needle, still inside the sheath. If the varices are still soft, the injection procedure is repeated. The patient is discharged, with endoscopic follow-up after 3 weeks.

Post procedure, there is no need for radiography, or for PPIs or H2 blockers unless there is another indication.

WEO “How I Do It” Glue treatment of gastric varices 5

“HOW I DO IT” Glue treatment of gastric varices Comment Norman E. Marcon

Our approach to bleeding from gastric varices is similar, with some minor differences.

The injection catheter (21- or 23-gauge) is first flushed with lipiodol. The glue (0.5 ml Histoacryl) plus an equal amount of lipiodol is pushed into the catheter displacing out the lipiodol except for that in the distal 2 or 3 cm. The varix is then punctured and the glue mixture is injected into the vein with distilled water.

We do not reuse the needles unless there is a clear stream of water after the needle is removed. We have no prescribed limit as to the number of ampoules injected, and have used as many as eight in one session. Our hepatic surgeons are not concerned about thrombosis of the portal or splenic vein prior to transplant: we have no record of this happening.

Glue injection, however, does have some potential hazards. Embolization to the lungs, liver, and spleen is not uncommon but usually asymptomatic. We have encountered one female patient with cirrhosis and pulmonary-hepatic syndrome who developed multiple pulmonary infarcts, and two of our patients had symptomatic cerebral emboli. All these patients recovered.

Glue injection continues to be an important part of the management of bleeding related to portal hypertension.

WEO “How I Do It” Glue treatment of gastric varices 6

“HOW I DO IT” Glue treatment of gastric varices Summary Tetsuya Mine

Indications: gastric varices The Japanese classification of gastric varices is slightly different from that described by Professor Mostafa. In Japan, gastric varices are categorized into three types: •

Lg-c: limited to the cardiac region, and supplied mainly from the left gastric vein.

•

Lg-cf: located between the cardia and fornix, and supplied from the posterior and short gastric vein.

•

Lg-f: limited to the fornix, supplied from the posterior and short gastric vein and draining to the left renal vein.

Contraindications These include: •

Severe jaundice (total bilirubin greater than 4 mg/dl)

•

Severe hypoalbuminaemia (less than 2.5 g/dl)

•

Thrombocytopenia (platelet count less than 20000/μl)

•

Disseminated intravascular coagulation (DIC)

•

Massive ascites

•

Severe hepatic encephalopathy

•

Severe renal failure.

However, the patient’s clinical and social situation must be taken into account.

Treatment Sedation Local anaesthesia can be administered. Anticholinergic agents are given. The use of sedative drugs may depend on the patient’s condition.

WEO “How I Do It” Glue treatment of gastric varices 7

Endoscopic injection sclerotherapy (EIS) EIS can be done using 5% EO (monoethanolamineoleate) delivered by means of a balloon (15–30-ml) attached to the endoscope and a 23-gauge needle, using antegrade and retrograde approaches. In our department, direct injection into fundal varices using 1.0 ml Histoacryl (adding 0.8ml saline) may be done for varices of the lg-cf and lg-f type. In this case, a net volume of 1.0 ml of the Histoacryl solution with 0.4 ml saline might be injected, since about 0.4 ml saline might be left in the catheter. For preparing, first, 0.8 ml saline is sipped and subsequently, 1.0 ml of Histoacryl is sipped into the catheter with a 23gauge needle and connecting 2.5 ml syringe. In Japan, interventional radiology might be used for lg-f varices. The balloon-occluded retrograde transvenous obliteration (B-RTO) method was developed in Japan and has been used in cases of lg-f gastric varices, since most of these are associated with a splenorenal vein shunt which can be used for,angiographic sclerotherapy.

Post treatment care In the case of bleeding, upper gastrointestinal endoscopy will be done in the next 24 hours. In the course of the treatment, the patient may be fasted and will start eating thereafter. The patient may take antibiotics over several days, and also medication for hemostasis of post-treatment bleeding, and proton pump inhibitors (PPIs) or H2-receptor antagonists.

Competing interests: None

WEO “How I Do It” Glue treatment of gastric varices 8

“HOW I DO IT�

Detection of early gastric cancer using chromoendoscopy and other techniques AUTHORSHIP

How I do it: Detection of early gastric cancer using chromoendoscopy and other techniques Nobuhiro Sakaki Vice president, Ebara Hospital Tokyo Japan

Comment Guido Villa-Gomez Roig Director, La Paz WGO Training Center La Paz Bolivia

Summary Tetsuya Mine Director of Gastroenterological Center Professor and Chief Division of Gastroenterology Department of Internal Medicine University of Tokai School of Medicine Japan

“HOW I DO IT” Detection of early gastric cancer using chromoendoscopy and other techniques How I Do It Nobuhiro Sakaki

Introduction

Carcinoma of the stomach is the leading cause of cancer death worldwide. Treatment results are dependent on the stage of the disease, which is related to the extent of the tumor. In the pursuit of a cure, gastric cancer must be diagnosed in its early stage. Gastric cancer confined to the mucosa and submucosa is regarded as defining the early stage of the disease (early gastric cancer, EGC), due to its overall favorable prognosis: the 5-year survival rate is greater than 95%.

Most cases of ECG are asymptomatic and difficult to diagnose by X-ray examination tools such as the upper gastrointestinal series and computed tomography. Endoscopic examination is the only method by which EGC can be detected, and even then careful observation is needed, supported by accurate knowledge of the characteristics of EGC lesions.

This article briefly supplies a basic knowledge of EGC and supporting diagnostic methods including chromoscopy. The know-how needed to detect EGC is explained as basically and simply as possible. The detailed diagnostic procedure sometimes needed to assess the range of infiltration and depth of invasion of the carcinoma, which is always required before the appropriate therapy can be selected, is not presented here.

Basics of EGC

Definition

ECG is defined as carcinoma confined to the mucosa and submucosa

irrespective of lymph node involvement, and corresponds to a T1 tumor in the TNM classification [1]. The definition of EGC is slightly different in the West and in Japan due to the different terminological conceptions of the limit of the neoplasm in the gastric mucosa.

Despite such efforts as the Vienna classification [2], differences still exist in the criteria

WEO “How I Do It” Detection of early gastric cancer using chromoendoscopy and other techniques 2

used to distinguish high-grade dysplasia from intramucosal cancer by Western pathologists (who rely on invasion of the lamina propria) and those used by Japanese pathologists (who use cytological and architectural features). In this article, EGC is diagnosed and classified using the Japanese classification [3], except that Lauren’s classification [4] is used to show the characteristics of histological findings.

Endoscopic classification

The classification of EGC was established by the Japanese

Endoscopic Society (Fig. 1). Recently, it has been recorded as a T classification [3]. EGC is described as one of the following subtypes: 0; I: protruding type; IIa: superficial elevated type; IIb: flat type; IIc: superficial depressed type; and III: excavated type. In combined superficial types, the type occupying the largest area is described first, followed by the next most dominant type: e.g. IIc + III.

Figure 1

Japanese classification of early gastric cancer (the

schema carried in Stomach and Intestine (Tokyo) before the 1980s).

Supporting diagnostic methods

Chromoscopy

Chromoscopy, or chromoendoscopy, is an endoscopic dye-spraying

method. There are three different types of method: contrast, staining, and chemical reaction [5].

WEO “How I Do It” Detection of early gastric cancer using chromoendoscopy and other techniques 3

Contrast method The contrast method is a simple technique in which a blue pigment, such as 0.2% indigo carmine solution, is sprayed onto the mucosal surface [6]. The detailed morphological characteristics of the lesion are emphasized by the contrast effect of the blue- colored solution on the reddish gastric mucosa. Figure 2 shows how EGC suspected on the basis of color change in a localized area of abnormality (Fig. 2a) is diagnosed as type IIc EGC on the basis of typical chromoscopic findings with indigo carmine (Fig. 2b).

Figure 2a–c

Type IIc early gastric cancer.

a

Ordinary observation.

b

View with indigo carmine contrast method.

WEO “How I Do It” Detection of early gastric cancer using chromoendoscopy and other techniques 4

c

Magnified view with indigo carmine contrast method.

Staining and chemical reaction A staining method using methylene blue solution is mainly employed for detecting intestinal metaplasia of the gastric mucosa [7]. Here, 0.2% methylene blue is sprayed and then washed out with water. The area of metaplasia shows up as blue-stained spots. A chemical reaction method, the endoscopic Congo red test, shows up an acid-secreting area of the gastric mucosa by a color change from red to black [8]. Both methods are used essentially to assess the severity of atrophic gastritis, but they can be applied to detect EGC [9]. Methylene blue staining can reveal EGC as a nonstaining area surrounded by blue-stained atrophic mucosa, while the Congo red method is useful to demonstrate EGC located within a normal fundic area.

Safety of pigments Indigo carmine is a safe pigment that is used as a food additive and as an injectable dye for kidney function testing. Congo red, which is also used as a food additive, is considered safe. However, care is needed when using methylene blue, which is absorbed during chromoendoscopy and may have adverse effects [10, 11].

Magnifying observation

Magnifying endoscopic observation of the gastric mucosa was

first developed in the late 1970s. Fine gastric mucosal patterns formed by gastric pits and surrounding elevations were classified as A (dotted), B (short-linear), C (striped), and D (granular), and as mixed patterns AB, BC, and CD [12]. Intestinal-type EGC shows a characteristic striped C pattern, and diffuse-type EGC is characterized by an amorphous depression surrounded by an irregular granular D pattern. Nowadays, attempts to diagnose EGC by microvascular patterns are being developed in Japan using a new “narrow-band imaging” (NBI) system. The NBI system is based on narrowing the bandwidth of spectral transmittance of the red-green-blue optical filters. From magnifying observation combined with NBI, microvascular patterns of superficial depressed cancers have been classified into two types: a fine network pattern and a corkscrew pattern, which correspond to the intestinal type and the diffuse type of EGC, respectively [13]. Figure 2c shows the corkscrew pattern observed in a cancerous erosion using a magnifying endoscope (GIF-240Z, Olympus). Needless to say, specific endoscopes are needed for detailed magnifying observation, but the knowledge obtained by magnifying endoscopy can be applied for a diagnosis of EGC even using conventional endoscopes.

Incidence of EGC

In our experience from 1986 to 1990, 943 patients were diagnosed with gastric cancer WEO “How I Do It” Detection of early gastric cancer using chromoendoscopy and other techniques 5

and underwent surgical or endoscopic resection at Tokyo Metropolitan Komagome Hospital.

After the exclusion of 67 patients treated endoscopically and 38 cases with special lesions such as recurrent cancer, 838 patients were included in the survey to determine the characteristics of EGC. Based on the histopathological inspection of specimens from operated stomachs, 343 patients (41% of the total of 838 with gastric cancer) were diagnosed with EGC alone (Table 1) (90 patients had both advanced and early cancer, and the remaining 405 patients had advanced cancer alone). Of the patients with EGC, 100 (29% of the total of 343 with EGC) had multiple lesions. In terms of the relationship between macroscopic figures and histological types, almost all cases of elevated EGC were consistent with intestinal-type carcinoma, while cases of depressed EGC were related to all histological types.

Table 1

Early gastric cancer, confirmed histologically in operated stomachs

1986–1990, Tokyo Metropolitan Komagome Hospital Single cancer, 243 cases Histology:

Intestinal: 146 (60%)

Diffuse: 97 (40%)

Form:

Elevated: 55 (23%)

Depressed: 188 (77%)

Multiple cancer, 100 cases* Histology:

Intestinal: 76

Intest. + diff: 13

Diffuse: 11

Form:

Elevated: 16

Elev. + depr: 23

Depressed: 61

*Made up of: 60 patients with 2 lesions, 16 with 3, 11 with 4, 2 with 5, 4 with 6, 2 with 7, and 5 with 8 or more.

Endoscopic characteristics of EGC (Japanese classification)

I A pedunculated or semi-pedunculated lesion is classified as a type I EGC. Type I lesions consisting of papillary adenocarcinoma occasionally require a biopsy diagnosis to differentiate them from hyperplastic polyps. Almost all type I cancers are confined to the mucosa and are a good indication for polypectomy including the surrounding neck area.

IIa EGC is often detected as a sessile elevated lesion, type IIa. In general, type IIa is identified as a reddish flat elevation with a comparatively uneven surface (Fig. 3).

Submucosal invasion can be suspected on the basis of a central depression in the protrusion, which is described as type IIa + IIc. A whitish flat elevation smaller than 5 mm

WEO “How I Do It” Detection of early gastric cancer using chromoendoscopy and other techniques 6

is likely to be an adenoma or intestinal metaplasia.

IIb Definite cases of type IIb EGC are very rare, except for minute cancers discovered by a detailed histopathological search [14], as shown in Table 2. On the other hand, EGC with a surrounding flat area due to infiltration of cancer cells, IIc + IIb or IIa + IIb, was diagnosed in 21 of 816 EGC lesions in 552 patients. In the histology of the accompanying IIb area, most of the cases showed diffuse-type carcinoma. In these cases, endoscopic assessment of the cancerous area was very difficult. Figure 4 shows a IIc + IIb lesion surrounded by a signet-ring cell carcinoma occupying almost the total area of the antrum.

Figure 3

Type IIa early gastric cancer.

Figure 4

Type IIc + IIb early gastric cancer.

WEO “How I Do It� Detection of early gastric cancer using chromoendoscopy and other techniques 7

IIc

Type IIc lesions, characterized by a well-demarcated shallow depression with an

irregular margin, are easy to diagnose by endoscopic findings alone. A reddish shallow depression with flat cancerous mucosa surrounded by a saw-edged border is the characteristic form of intestinal-type carcinoma (Fig. 5), while typical diffuse-type carcinoma shows a discolored uneven depression with a sharp margin (Fig. 2).

Chromoendoscopy is most effective for diagnosing the type IIc cancer.

III, IIc + III A benign peptic ulcer occasionally occurs within a cancerous area. Type IIc lesions with a benign ulcer or a scar were observed in 127 of 816 EGC lesions in 552 patients. Depending on the healing process of a peptic ulcer surrounded by cancerous mucosa, macroscopic types can change from III (+ IIc), to IIc + III, to IIc (+ ulcer scar), serially. The healing and recurrence process of a peptic ulcer can occur even in cancerous mucosa, and in Japan is called a “malignant cycle”. If an active peptic ulcer is covering almost the entire cancerous area (III), the diagnosis is difficult. In our experience, 19% of type IIc + III cancers were not diagnosed initially.

For this reason, a follow-up

examination is very important for accurate diagnosis of a type IIc EGC with an ulcer or a scar.

Table 2

Endoscopic detection rate of minute early gastric cancer,

confirmed histologically in operated stomachs, according to size and macroscopic type (modified from [14]) Size (max. diameter), mm 1 2 3 4 5

Macroscopic type (Japanese classification)* IIa IIb IIc % n % n % – 0/0 0 0/12 50 0 0/2 0 0/27 11 0 0/1 17 1/6 13 50 1/2 0 0/7 50 57 4/7 0 0/3 21

n 1/2 1/9 1/8 2/4 4/19

* Diagnosed before operation/detected histologically. Surgical operations were performed 1986–1989 at Tokyo Metropolitan Komagome Hospital.

Endoscopy to find EGC

Screening To detect EGC, it is important to observe the gastric mucosa thoroughly. Even for detection of a small lesion, it is necessary to observe in a large view field at first. The posterior wall of the gastric body must be observed during careful insertion of the scope

WEO “How I Do It” Detection of early gastric cancer using chromoendoscopy and other techniques 8

with a relatively small amount of air. During pull-back of the scope after observation of the whole antrum and duodenum area, the gastric angle, body, and fornix should be observed continuously using a J-turn or U-turn reverse observation technique with a relatively large amount of air.

If a flat elevation, a discolored or reddish uneven area, or an irregular

depression is found, close-up observation with chromoscopy should be carried out to determine whether the lesion is neoplastic or not. If a lesion is suspected to be an EGC, detailed observation should be carried out to assess it accurately, including the depth of cancer invasion. Histological confirmation using biopsy specimens is necessary to diagnose EGC, which shows various forms. On the other hand, when an EGC has been detected, it is important always to search for other cancerous lesions, due to the possibility of multiple occurrence.

A case presentation A minute type IIc lesion detected in a routine endoscopy is shown in Fig. 5. A reddish, rough area located at the gastric angle was picked up during a bird’seye observation (Fig. 5a). After spraying with indigo carmine solution, a minute shallow depressed lesion with a saw-edged margin was revealed within the reddish area (Fig. 5b).

It could be diagnosed as an intestinal-type EGC confined to the mucosa on the basis of endoscopic findings alone. After histological diagnosis using a biopsy specimen, this lesion was resected by endoscopic submucosal dissection.

Figure 5a, b copy. a

A minute early gastric cancer detected during a routine endos

Ordinary view. b

View using the indigo carmine contrast method.

Background atrophic gastritis

For the endoscopic detection of EGC, it is useful to

observe the background gastric mucosa from the viewpoint of a Helicobacter pylori– gastritis–cancer sequence. Almost all EGC lesions are detected in the stomach, with atrophic gastritis diagnosed by the endoscopic findings of discoloration and transparency of blood vessels. Special attention should be paid to the atrophic mucosal area to detect an intestinal-type EGC. On the other hand, diffuse-type EGC can develop in various areas

WEO “How I Do It” Detection of early gastric cancer using chromoendoscopy and other techniques 9

of the stomach with background atrophic mucosa. If a diffuse type IIc lesion is surrounded by nonatrophic mucosa, it is easy to diagnose it as a discolored, well-demarcated depression. Accurate diagnosis of diffuse type IIc cancers which occur in atrophic gastric mucosa is occasionally difficult due to the IIb-like infiltration.

References 1. American Joint Committee on Cancer: AJCC Cancer Staging Manual. Philadelphia, Pa: Lippincott-Raven Publishers, 5th ed., 1997, pp 71–76 2. Schlemper RJ, Riddell RH, Kato Y et al: The Vienna classification of gastrointestinal epithelial neoplasia. Gut 2000; 47: 251– 255 3. Japanese Gastric Cancer Association: Japanese classification of gastric cancer – 2nd English edition. Gastric Cancer 1998; 1: 10– 24 4. Lauren P: The two histological main types of gastric carcinoma: diffuse and so-called intestinal-type carcinoma. Acta Pathol Microbiol Scand 1965; 64: 31– 49 5. Kawai K, Takemoto T, Suzuki S, Ida K: Proposed nomenclature and classification of the dye-spraying techniques in endoscopy. Endoscopy 1979; 11: 23– 25 6. Ida K, Hashimoto Y, Takeda S et al: Endoscopic diagnosis of gastric cancer with dye scattering. Am J Gastroenterol 1975; 63: 316– 320 7. Ito M, Haruma K, Kodama T et al: Helicobacter pylori eradication therapy improves atrophic gastritis and intestinal metaplasia: a 5-year prospective study of patients with atrophic gastritis. Aliment Pharmacol Ther 2002; 16: 1449– 1456 8. Tatsuta M, Saegusa T, Okuda S: Studies on gastritis in the upper portion of stomach by endoscopic Congo red test. Endoscopy 1973: 5: 61– 69 9. Tatsuta M, Okuda S, Taniguchi H: Endoscopic diagnosis of early gastric cancers of type IIb by endoscopic Congo red – methylene blue test. Gastroenterol Endosc 1981; 23: 487– 496 [Japanese with English abstract] 10. Sakaki N, Okita K, Watanabe S et al: Experimental study for side effects of methylene blue. Gastroenterol Endosc 1979; 21: 411– 418 (Japanese with English abstract) 11. Olliver JR, Wild CP, Sahay P et al: Chromoendoscopy with methylene blue and associated DNA damage in Barrett’s oesophagus. Lancet 2003; 362: 373– 374 12. Sakaki N, Iida Y, Saito M et al: New magnifying endoscopic classification of the fine gastric mucosal pattern. Gastroenterol Endosc 1980; 22: 377–383 [Japanese with English abstract] 13. Nakayoshi T, Tajiri H, Matsuda K et al: Magnifying endoscopy combined with narrow band imaging system for early gastric cancer: correlation of vascular pattern with histopathology (including video). Endoscopy 2004; 36: 1080–1084 14. Sakaki N, Kato H, Takizawa T et al: Limitation in the diagnosis of minute gastric cancers in the endoscopic observation using magnifying endoscopes. Stomach and Intestine (Tokyo) 1995; 30: 1239– 1248 [Japanese with English abstract] WEO “How I Do It” Detection of early gastric cancer using chromoendoscopy and other techniques 10

“HOW I DO IT” Detection of early gastric cancer using chromoendoscopy and other techniques Comment Guido Villa-Gomez Roig

This careful analysis of early detection of early gastric cancer (EGC) by Dr. Nobuhiro Sakaki of Tokyo’s Ebara Hospital carries the authority of Japanese medicine on this subject, for Japan is both the country with the highest incidence of gastric cancer in the world, and the originator of contributions that have managed to improve the early diagnosis and introduce innovative treatment techniques and significant changes in the prognosis of the disease in the Japanese population.

In 1962, the Japanese Gastric Cancer Association introduced a definition of EGC and a radiological and endoscopic classification for the early stage of the disease. For several decades this proposal was barely taken up in the West, and it was only in the year 2000 that the Vienna classification of gastrointestinal epithelial neoplasia came into being, and although agreements on several fronts have been achieved between experts, as mentioned Dr. Sakaki, there are still differences of opinion on distinguishing between highgrade dysplasia and intramucous cancer.

The classification of EGC into types 0, I (elevated), IIa (superficially elevated), IIb (flat), IIc (superficially depressed), and III (excavated), remains valid, unlike double-contrast radiographic studies, which have receded in importance compared to videoendoscopy. In the countries of South America the same trend is visible.

Dr. Sakaki discusses the various methods of chromoendoscopy, which are divided into three: contrast, staining, and chemical reaction. In South America, as in Japan, the contrast method using 0.2% indigo carmine solution is the best-known chromoendoscopic technique and it is being implemented with the following objectives:

• Primary goal: To improve the performance of endoscopic diagnosis and treatment of EGC. • Secondary goals: − To visualize more mucosal detail, providing better contrast of the margins against the surrounding normal mucosa WEO “How I Do It” Detection of early gastric cancer using chromoendoscopy and other techniques 11

− To allow complete endoscopic resection of the lesions

The article also analyzes the methods of magnification which improve the observation of fine gastric mucosal patterns formed by gastric pits and surrounding elevations, enabling their classification into four basic types with additional combinations, and distinguishing between the intestinal type and the diffuse type of EGC.

Narrow-band imaging (NBI) is described, with its ability to visualize better the fine mucous pattern and fine vascular pattern that, in combination with magnification techniques, allow the intestinal type to be differentiated from the diffuse type in cases of depressed superficial cancers. This technology was introduced in South America 2 years ago with evident benefits for the diagnosis of early lesions in the digestive tract and preneoplastic lesions. I believe that there is general and justifiable interest among specialists in the region to incorporate these new technologies into their practice.

Dr. Sakaki’s article highlights the differences in possibilities and performance of endoscopy for the diagnosis of EGC depending on the different types of lesion. The most difficult are types IIb, III, and IIc + III associated with benign ulcer, where development of a malignant cycle of the ulcer is possible and where endoscopic follow-up should be considered because of the possibility of false-negative initial assessments.

Finally, the article describes the sequence and procedure to be followed for complete observation of the stomach during videoendoscopy, and the need to consider multiple occurrences. I consider this description to be very important for

improving the

performance of early diagnosis of gastric cancer, because usually attention is not paid to all the details that impact on the final results of the endoscopic examination.

I feel that this article is practical, and describes in simple terms new and important aspects of the diagnosis of EGC based on the technological resources currently available and the broad experience of Japanese medicine in this field. It will be important in another article to focus on advances in currently proposed endoscopic therapy, the results, and the impact of these on patient prognosis.

WEO “How I Do It” Detection of early gastric cancer using chromoendoscopy and other techniques 12

“HOW I DO IT” Detection of early gastric cancer using chromoendoscopy and other techniques Summary Tetsuya Mine Dr. Nobuhiro Sakaki’s article on the detection of early gastric cancer (EGC) using chromoendoscopy and other techniques is well written and practical. He has conducted a careful analysis of this disease and the role of chromoendoscopy.

As Professor Guido Villa-Gomez Roig says, chromoendoscopy has two goals: first, to improve endoscopic diagnosis, leading to better treatment of EGC; and, second, to identify the detail of the margin between the lesion and surrounding normal gastric mucosa. This enables suspected lesions to be assessed for the depth and horizontal extent of cancerous invasion and improves the detection of any other sites of cancer in the stomach.

Magnifying endoscopy combined with chromoendoscopy provides a great deal of information about EGC as outlined above. The new technique of narrow-band imaging (NBI) has now been introduced into clinical use and may provide the same advantages as chromoscopy.

In the future, the knowledge obtained by chromoendoscopy and magnifying endoscopy, which partly includes NBI, will be applied in combination for the diagnosis of EGC.

Competing interests: None

WEO “How I Do It” Detection of early gastric cancer using chromoendoscopy and other techniques 13

“HOW I DO IT” Management of upper gastrointestinal bleeding AUTHORSHIP

How I do it: Management of upper gastrointestinal bleeding Joseph JY Sung MD, FRCP, FACG Endoscopy Center Prince of Wales Hospital, The Chinese University of Hong Kong Hong Kong

Comment David J. Bjorkman, MD MSPH Dean, University of Utah School of Medicine Salt Lake City, Utah USA

Summary K. R. Palmer Western General Hospital NHS Trust Edinburgh UK

WEO “How I Do It” Management of upper gastrointestinal bleeding 1

“HOW I DO IT” Management of upper gastrointestinal bleeding How I Do It Joseph JY Sung

Introduction The management of patients with upper GI hemorrhage requires a multidisciplinary approach mandating cooperation among medical and surgical gastroenterologists with access to skills in endoscopic and surgical hemostasis. Endoscopic therapy is often the first treatment in most management algorithms. Institution-specific protocols should be in place for the care of such patients.

Approximately 80%–85% of upper GI bleeding stops spontaneously and supportive therapy only is required [1]. In the remaining 15%–20% of cases bleeding continues or recurrent bleeding develops, and these patients constitute the high-risk group with substantially increased morbidity and mortality.

Early risk stratification of patients with upper gastrointestinal bleeding, based on clinical and endoscopic criteria, facilitates the delivery of the appropriate level of care to patients, thereby conferring important resource implications.

Initial management Patients with acute bleeding should be evaluated immediately on presentation. A rapid assessment should be performed to establish whether: 1) the airway is compromised, 2) there is active bleeding, and 3) the patient is hypovolemic. Orthostatic vital signs can help in the determination of lesser degrees of intravascular volume depletion. Postural hypotension is often present and indicates a 10%–20% reduction in blood volume. Fluid resuscitation takes priority in the presence of any of the above signs, and venous access via one or more wide-bore cannulas is essential.

Patients with severe acute bleeding require admission to a high-dependency or intensive care unit.

Those with significant cardiopulmonary disease are fragile patients and may require intense monitoring that includes central venous pressure measurements. Intravascular volume should be promptly replenished with crystalloid solutions to maintain organ perfusion, and supplemental oxygen should be administered to augment oxygen-carrying capacity of blood.

WEO “How I Do It” Management of upper gastrointestinal bleeding 2

Is a nasogastric tube required?

The routine insertion of a nasogastric tube prior to endoscopy may not be necessary. Nasogastric tube placement with lavage prior to endoscopy is often ineffective, as blood tends to form clots and may thus be difficult to aspirate. Bleeding lesions are often sited at the lesser gastric curvature, angular notch, gastric antrum, or bulbar duodenum while blood generally pools in the gastric fundus and corpus. An endoscopic examination is often possible and adequate with suction aspiration and patient positioning.

The insertion of a nasogastric tube is associated with a small risk of aspiration, in particular among obtunded patients. On occasions when the source of GI bleeding is uncertain, the insertion of a nasogastric tube can help to confirm an upper GI bleeding source, with the presence of gross blood or coffee ground-like substances in the gastric aspirate. In contrast, a clear nasogastric aspirate does not exclude upper GI bleeding inasmuch as bleeding may be episodic or the lesion might be located distal to the stomach in the presence of a competent pylorus, preventing retrograde reflux of blood from the duodenum. Nevertheless, the nature of nasogastric aspirates can serve as a prognostic indicator.

How early should endoscopy be arranged?

Early endoscopy is usually defined as performance of the examination within 24 hours of the patient’s admission. This procedure allows risk classification and hence safe and early discharge of patients classified as low-risk. In high-risk patients, it is logical to believe that the early control of bleeding reduces transfusion requirements and organ dysfunction and generally improves patient outcome.

Those with signs of ongoing bleeding after initial resuscitation, such as fresh hematemesis or fresh hematochezia and hypovolemic shock, should undergo urgent endoscopy to locate the source of bleeding and for possible endoscopic therapy. Otherwise, endoscopy within 24 hours is reasonable.

When should we apply endoscopic hemostasis?

The endoscopic features of an ulcer are important prognostic indicators. The presence of stigmata of recent bleeding in an ulcer confirms the source of bleeding. Three decades ago, Forrest & Finlayson categorized ulcers into those that were actively bleeding, those that showed stigmata of recent bleeding, and those that had a clean base [2]. The nomenclature has since been in widespread use. Forrest class I ulcers are those with active bleeding, which can be either spurting (Forrest class IA) or oozing (Forrest class IB). Stigmata of recent bleeding belong to Forrest class II: non-bleeding visible vessel, IIA; adherent clots, IIB; and flat pigmented spots, IIC. Ulcers with a clean base belong to Forrest III. Endoscopic treatment should be applied to both actively bleeding ulcers and ulcers with a visible vessel.

WEO “How I Do It” Management of upper gastrointestinal bleeding 3

Ulcers with flat pigmented spots and clean bases are associated with low risks of recurrent bleeding.

Endoscopic therapy in these cases is not warranted. There were some controversies about what to do with an ulcer with an adherent clot or protuberant vessel. Several randomized studies and a recent meta-analysis indicated that these ulcers should not be left to be managed with medical therapy only.

Endoscopic hemostasis may reduce the risk of recurrent bleeding [3].

Endoscopic stigmata should be interpreted along with clinical factors as well as other endoscopic features, including the size and the site of bleeding ulcers. Ulcers at the lesser curvature of the stomach or the posterior duodenal bulb belong to the high-risk category because of their proximity to the left gastric artery and the gastroduodenal artery complex, respectively.

Which endoscopic therapy?

Endoscopic hemostatic therapy is now widely accepted as the first-line therapy for upper GI bleeding.

Numerous clinical trials have been published confirming the efficacy of endoscopic therapy. The majority of these clinical trials demonstrated a reduction in recurrent bleeding and the need for surgical intervention with the use of endoscopic therapy. Endoscopic therapy can be broadly categorized into injection therapy, thermal coagulation, and mechanical hemostasis. Thermal devices can be further divided into contact and non-contact types. In spite of the large volume of published literature, no single solution for endoscopic injection has been established to be superior to another for achieving hemostasis. Similarly, no single method of endoscopic coaptive therapy is superior to others.

Mechanical hemostasis is as good as thermal approaches, if it can be applied successfully. Combined treatment with injection followed by thermal coagulation or hemoclips may provide the best results [4,5].

When should proton pump inhibitors be used?

Proton pump inhibitors (PPIs), when given at high doses, can shut down acid secretion in the stomach and help to stabilize blood clotting on peptic ulcers. A randomized study has shown that when intravenous PPIs are combined with endoscopic therapy, the rebleeding rate can be reduced to a minimal level [6]. In a systematic review including 9 trials (1829 patients) with either placebo or H2-receptor antagonists as control, the use of PPIs was associated with 50% reductions in the rate of recurrent bleeding and the need for surgery [7]. The use of PPIs also led to a trend of reduction in mortality. However, no study has compared high-dose PPI infusion with oral PPIs after endoscopic hemostasis. Intragastric pH control with oral PPI is suboptimal since the oral absorption of a PPI is not always reliable in critically ill patients.

Based on the above studies, it would appear that the optimal approach to the management of bleeding peptic ulcer should include early endoscopic treatment for patients with high-risk ulcers, followed by a high-dose PPI infusion to prevent recurrent bleeding.

PPIs given prior to endoscopy have been found to reduce the occurrence of active ulcer bleeding and therefore reduce the need for endoscopic therapy [8]. In settings where 24-hour emergency therapeutic WEO “How I Do It�

Management of upper gastrointestinal bleeding 4

endoscopy is not available, this should be considered as a stopgap treatment.

Is routine second-look endoscopy necessary?

Some endoscopists would routinely schedule patients for a second-look endoscopic examination the next morning, followed by the re-treatment of remaining stigmata of bleeding. A meta-analysis pooling several randomized trials showed that recurrent bleeding can be partially prevented [9]. However, the use of second-look endoscopy in a routine fashion cannot be recommended for the following reasons. First, the gain from a second-look endoscopy appears modest as shown from the pooled analysis. Second, most trials used epinephrine injection alone, which may represent a suboptimal index treatment; with an improved index therapy, the yield from a second treatment may diminish further. Third, repeat endoscopic treatment, especially with thermocoagulation, increases the risk of perforation. Finally, the use of adjunctive PPI infusion may obviate repeat endoscopy after endoscopic control. Nevertheless, it may be logical to consider selective re-endoscopy and re-treatment in cases in which the index endoscopic therapy may have been difficult and suboptimal and in patients with subtle signs of recurrent bleeding.

Is there a role for surgery?

Although endoscopic therapy has clearly reduced the need for surgical intervention, surgery retains an important role. Indications for emergency surgery include: (i) failure to secure active bleeding by endoscopic or angiographic means, (ii) inability to access a bleeding source due to anatomical reasons, (iii) rapid exsanguination and the inability to identify a bleeding lesion, and (iv) an endoscopic treatment complication, such as a perforation. Many would also recommend surgery after two episodes of recurrent bleeding after initial endoscopic control.

At the time of recurrent bleeding, the dilemma often faced by the managing physician is whether to once again attempt endoscopic treatment or to refer the patient directly to surgery. A randomized trial involving patients who re-bled after initial endoscopic control of their bleeding ulcers suggested that surgery is more successful in securing hemostasis, but at the cost of post-surgical complication [10]. A second attempt at endoscopic therapy may be a sensible option when the surgical risk is high.

What if endoscopy fails to stop bleeding and surgery is too risky?

Angiographic transcatheter embolization is a therapeutic option in patients who do not respond to endoscopic hemostasis and are otherwise poor candidates for surgery. Catheter and guideline technology have evolved significantly in recent decades. Superselective cannulation using 3- to 5-Fr catheters and micro-coils is now the standard for transcatheter embolization, with more durable hemostasis and reduced ischemic complications. Considerable skill and expertise are required. Angiographic embolization of a posterior bulbar duodenal ulcer, for instance, requires selective cannulation to the gastroduodenal artery via the celiac trunk and branches of the superior mesenteric artery. Retrograde filling via the inferior pancreaticoduodenal branch through confluence of the superior and inferior arterial arcades becomes WEO “How I Do It�

Management of upper gastrointestinal bleeding 5

possible if the ulcer is approached via the celiac axis alone. Endoscopic clipping of the bleeding point may help location of feeding vessels in upper gastrointestinal bleeding.

The management algorithm for upper gastrointestinal bleeding is shown in Figure 1.

References 1

Consensus Development Panel, ASGE. Consensus statement on therapeutic endoscopy and bleeding ulcers. Gastrointest Endosc 1990; 36: S62

2

Forrest JA, Finlayson ND, Shearman DJ. Endoscopy in gastrointestinal bleeding. Lancet 1974; 2: 394–397

3

Kahi CJ, Jensen DM, Sung JJY et al. Endoscopic therapy versus medical therapy for bleeding peptic ulcer with adherent clot: a meta-analysis. Gastroenterology 2005; 129: 855–862

4

Sung JJY, Tsoi KK, Lai LH, Wu JC, Lau JY. Endoscopic clipping versus injection and thermocoagulation in the treatment of non-variceal upper gastrointestinal bleeding: a meta-analysis. Gut 2007 ;56: 1364–1373

5

Marmo R, Rotondano G, Piscopo R et al. Dual therapy versus monotherapy in the endoscopic treatment of high-risk bleeding ulcers: a meta-analysis of controlled trials. Am J Gastroenterol 2007; 102: 279–289

6

Lau JYW, Sung JJY, Lee KKC et al. Effect of intravenous omeprazole on recurrent bleeding after endoscopic treatment of bleeding peptic ulcers. N Engl J Med 2000; 343: 310

7

Lau JYW, Leung WK, Wu JC et al. Omeprazole before endoscopy in patients with gastrointestinal bleeding. N Engl J Med. 2007; 356: 1631–1640

8

Zed PJ, Loewen PS, Slavik RS, Marra CA. Meta-analysis of proton pump inhibitors in treatment of bleeding peptic ulcers. Ann Pharmacother 2001; 35: 1528–1534

9

Marmo R, Rotondano G, Bianco MA et al. Outcome of endoscopic treatment for peptic ulcer bleeding: Is a second look necessary? A meta-analysis. Gastrointest Endosc 2003; 57: 62–67

10 Lau JYW, Sung JJY, Lam YH et al. Endoscopic retreatment compared with surgery in patients with recurrent bleeding after initial endoscopic control of bleeding ulcers. N Engl J Med 1999; 340: 751.

WEO “How I Do It” Management of upper gastrointestinal bleeding 6

Figure 1. Management algorithm for upper gastrointestinal bleeding.

Acute upper gastrointestinal bleeding

Initial resuscitation, airway, volume restoration, in-and-out nasogastric tube [?], empiric high-dose PPI

Signs of ongoing bleeding, such as shock

Stable patient

and fresh hematemesis or hematochezia

Endoscopy next morning

Urgent endoscopy

Epinephrine injection plus

Low-risk ulcers;

Thermocoagulation to actively bleeding ulcers and

flat pigments

ulcers with nonbleeding visible vessels

and clean base

Clots elevated and treatment to underlying vessel

Bleeding stopped

Consider early discharge

Unable to control bleeding, access or locate bleeding point in the presence of massive bleeding

High-dose intravenous

Second rebleed

infusion of PPI for 72 hours

Recurrent bleeding

Transcatheter arterial embolization

Surgery

Endoscopic confirmation of rebleeding and re-treatment

WEO “How I Do It�

Management of upper gastrointestinal bleeding 7

“HOW I DO IT” Management of upper gastrointestinal bleeding: Comment David J. Bjorkman Introduction Dr. Sung has provided an excellent and comprehensive outline of the clinical approach to nonvariceal upper gastrointestinal bleeding. He has provided an excellent flow chart to illustrate the appropriate approach. There is little that I can add to his description, but I will endeavor to highlight some points that Dr. Sung has made that we find particularly important.

Initial management As Dr. Sung has noted, the majority of nonvariceal upper gastrointestinal bleeding (80%) ceases spontaneously. The clinical outcome for these patients is determined by the volume of blood lost and their underlying illnesses. This makes the initial focus on the hemodynamic status of the patient the most important clinical activity for all patients with upper GI bleeding. Other diagnostic and therapeutic activities should not interfere with immediate assessment of the hemodynamic status of the patient and the replacement of intravascular volume. In patients who have had bleeding significant enough to produce tachycardia or orthostatic hypotension, intravenous access should be established promptly in two locations. The size of the intravenous catheter should be large enough to replace volume quickly and transfuse blood products, if clinically indicated.

Preparation for endoscopy Early endoscopy is indicated in all patients with clinically significant acute upper gastrointestinal bleeding. Preparation for endoscopy should include hemodynamic stabilization and assessment of the patient’s airway. If there is concern about possible airway compromise, an endotracheal tube should be placed prior to endoscopy. This may also help avoid aspiration in patients with vigorous ongoing bleeding.

Nasogastric aspiration can confirm that the bleeding lesion is in the upper GI tract and identify persistent bleeding. While a negative aspirate does not rule out an upper GI bleeding source, the presence of large amounts of fresh blood or clots suggests that more vigorous lavage of the stomach using an orogastric tube may be needed to allow adequate visualization of the entire gastric and duodenal mucosa. Some endoscopists have recommended that intravenous erythromycin should be given to patients with large clots in the stomach, to stimulate gastric activity and move the clots distally.

It may not be possible to remove all blood from the stomach prior to endoscopy. Spraying dilute

WEO “How I Do It” Management of upper gastrointestinal bleeding 8

hydrogen peroxide through a catheter at the time of endoscopy may cause the blood to blanch and aid visualization.

Timing of endoscopy There have been multiple studies of different scoring systems to predict the rebleeding rate and mortality for patients with nonvariceal upper GI bleeding. All have demonstrated that the best predictor for rebleeding is the endoscopic appearance of the ulcer or other lesion. The presence of active bleeding, a visible vessel or an adherent clot increases the likelihood of rebleeding. Endoscopic therapy is indicated in patients with these stigmata. On the other hand, patients with clean-base ulcers or flat pigmented spots are unlikely to have subsequent bleeding and do not require endoscopic therapy.

Scoring systems without endoscopy may identify some patients with small volume bleeds who are at very low risk. Ultimately all patients with upper GI bleeding should have endoscopic evaluation and possible treatment.

The decision on the optimal timing for endoscopy is based upon the ability of endoscopy to identify high-risk lesions that will benefit from immediate therapy and aggressive medical care in the hospital and also to identify low-risk patients who have a very low risk of rebleeding who may be treated as outpatients. In order for this triage to occur, endoscopy should ideally be performed prior to a decision for admission to the hospital. This is not possible in some institutions for logistical reasons, but efforts should be made to have endoscopy performed at the earliest possible time (within 24 hours) to allow prompt discharge of low-risk patients and more aggressive care for high-risk patients.

Recent data suggest that intravenous proton pump inhibitor therapy may be indicated in all patients with nonvariceal upper GI bleeding while they are waiting for endoscopy, as it may decrease the prevalence of high-risk lesions on the endoscopic examination. The intravenous therapy can be discontinued in low-risk patients and should be continued in patients with high-risk stigmata. As Dr. Sung points out, rebleeding rates are decreased by PPI therapy in the latter group of patients, but not in the low-risk group.

Endoscopic therapy Dr. Sung has provided an excellent overview of the options for endoscopic therapy. It is important to emphasize that injection of large volumes of saline or dilute epinephrine can provide temporary tamponade of bleeding lesions to aid subsequent therapy, but is not adequate to provide definitive hemostasis. Thermal (electrocautery, heater probe) or mechanical (clips) therapy should always follow injection therapy. Thermal treatment and clips are equally effective with and without simultaneous injection therapy. The selection of a specific technique should be based upon the endoscopist’s experience and the specific clinical situation.

In the setting of an adherent clot, which cannot be removed by vigorous flushing, injection of saline or WEO “How I Do It” Management of upper gastrointestinal bleeding 9

dilute epinephrine around the base of the clot should be performed prior to clot removal. It is possible to precipitate significant bleeding by dislodging the clot. Pretreatment with injection therapy may partially mitigate this bleeding and facilitate subsequent definitive endoscopic therapy.

When rebleeding occurs, a second application of endoscopic therapy may prevent further bleeding and avoid emergency surgery. Routine second-look endoscopy with repeated therapy has been advocated by some, but appears to have only a small marginal benefit in the reduction of bleeding. Most experienced endoscopists will have a sense of how effective a session of endoscopic therapy is likely to be. In some situations the therapy is performed easily with good effect on the bleeding lesions. In other situations the location of the lesion or retained blood may make the therapy technically more difficult. In situations where the initial therapy was felt to be suboptimal there may be a role for a scheduled second therapeutic endoscopy to make sure that the lesion has been adequately treated.

Failure of endoscopy and pharmacologic therapy Some patients may not respond to vigorous medical and endoscopic therapy. If repeated endoscopic therapy fails to control the bleeding, other modalities must be considered. Surgery is the most common approach to uncontrolled bleeding, as even large bleeding vessels can be directly visualized and ligated.

In situations where severe co-morbid illness prevents surgery, angiographic therapy may be possible.

Therapy after bleeding After the initial bleeding episode has been treated, a full course of oral PPI therapy should be completed to aid hemostasis and heal the ulcer. A search for the cause of the ulcer (Helicobacter pylori, nonsteroidal anti-inflammatory drugs [NSAIDS]) should identify ways to prevent the ulcer from recurring.

H. pylori should be treated, and NSAIDs should be discontinued. If NSAIDs or aspirin must be continued, chronic PPI therapy is indicated to reduce the risk of rebleeding. There are no data to support routine endoscopic examination after healing of a bleeding ulcer. Recurrent symptoms or evidence of recurrent bleeding, however, should prompt another endoscopic examination.

WEO “How I Do It�

Management of upper gastrointestinal bleeding 10

“HOW I DO IT” Management of upper gastrointestinal bleeding: Summary K. R. Palmer

Drs Sung and Bjorkman agree on the treatment algorithm for nonvariceal upper gastrointestinal bleeding and their comments are consistent with published guidelines.

Hematemesis and melena has become the province of the therapeutic endoscopist; surgery and interventional radiological procedures are now rescue therapies in patients not responding to endotherapy.

Deaths are almost entirely restricted to patients who have significant medical co-morbidities; bleeding may decompensate critical diseases affecting major organ systems leading to life-threatening events.

This is reflected in the observation that the mortality of acute bleeding amongst patients who are hospitalized for other medical or surgical diseases is much higher than that of patients who are admitted because of bleeding. A recent UK audit for example recorded a mortality of 26% for inpatients who bled compared to 7% amongst patients admitted because of bleeding. It follows that active management of co-morbidity is crucial to best outcome and that when individual patients present with acute gastrointestinal hemorrhage, co-morbidity is recognised and actively treated.

We should for example

be involving the acute physician, high-dependency unit doctors, and colleagues in other medical specialties in our initial assessment in order to get the patients into the best possible condition prior to endoscopy. If we are to significantly to improve outcome, it is clear that managing co-morbidity is crucial.

It is interesting to observe different approaches to management in the USA, Asia, and Europe.

In the

UK for example there has been a movement away from passage of a nasogastric tube in patients presenting with acute GI bleeding, for the reasons outlined by Dr Sung.

Nasogastric tubes are

unpleasant for patients and in my opinion have little diagnostic utility since all patients with major active bleeding will come to endoscopy; furthermore there is an impression that nasogastric tubes may confuse the diagnosis by causing erosions in the esophagus and could worsen variceal bleeding.

There is no evidence that they improve outcome.

The timing of endoscopy sometimes presents logistic problems, since (at least in the UK) not all units are in a position to provide optimal out-of-hours endoscopy. As both authors indicate, an endoscopy is no longer voyeurism, rather it is an important therapeutic modality and as such requires not only an expert endoscopist and optimal equipment but assistants who are familiar with endoscopic therapy.

For the majority of patients it is clear that endoscopy within 24 hours is optimal since the procedure will be done in an appropriate environment and by the right team. Clearly however, actively bleeding shocked patients will require emergency out-of-hours endoscopy following appropriate resuscitation.

WEO “How I Do It” Management of upper gastrointestinal bleeding 11

When this is done, in my personal view, it should be supported by anesthetic cover and in many patients this will entail endotracheal intubation at the time of the endoscopy.

This facilitates what is often the

most difficult of all endoscopic procedures – in the actively bleeding patient in whom access is difficult and in whom support of medical co-morbidity is absolutely crucial to best outcome.

It is very clear that dual endoscopic therapy is now standard of care and that in many cases secondlook endoscopy within 24 hours is required, particularly if endotherapy at the first session was less than satisfactory.

Both authors recommend administration of proton pump inhibitor drugs prior to endoscopy, yet this is something that I believe should be questioned. In 80% of cases bleeding stops spontaneously and there is no evidence to show that outcome is improved by acid suppression in this group. Whilst we know that major endoscopic stigmata are less evident in patients receiving PPIs prior to endoscopy, this has not translated into better rebleeding rates or mortality in those receiving these drugs. More important, however, is the message that endoscopic therapy, not acid-suppressant drugs, is the first-line requirement for patients with major GI bleeding; using PPIs as a stopgap or substitute is wrong. British guidelines have stated that PPIs are only required following endoscopic identification of major endoscopic stigmata, and we do not support their use prior to endoscopy.

Finally the message that the management of acute GI bleeding remains a multidisciplinary process is as cogent now as it was forty years ago when the primary treatment of such patients was surgical. Liaison between physicians and surgeons remains important but in addition we should now have greater liaison with other medical specialists. A very good example relates to the group of patients admitted with GI bleeding whilst taking aspirin and clopidogrel for severe vascular disease. Gastroenterologists tend to recommend stopping these drugs because of their adverse effects upon bleeding, yet whilst this may reduce the incidence of rebleeding, such a policy could actually increase overall mortality as more patients die from cardiac events. It is therefore important to discuss management with relevant expert colleagues to try to best define the risk–benefit balance of continuing or stopping antiplatelet drugs in individual bleeding patients.

WEO “How I Do It” Management of upper gastrointestinal bleeding 12

“HOW I DO IT” Endoscopic placement of percutaneous feeding tubes AUTHORSHIP How I do it: Endoscopic placement of percutaneous feeding tubes John C. Fang, MD Division of Gastroenterology, Hepatology, and Nutrition Department of Internal Medicine University of Utah Health Sciences Center Salt Lake City, Utah USA

Comment Suneil A. Kapadia, MD FRCP Department of Gastroenterology McHale Building Royal Wolverhampton Hospitals NHS Trust Wolverhampton UK

Summary Spiros D Ladas, MD Professor of Medicine & Gastroenterology 1st Department of Medicine-Propaedeutic “Laiko” General Hospital of Athens Medical School, University of Athens Athens Greece

“HOW I DO IT” Endoscopic placement of percutaneous feeding tubes How I Do It John C. Fang Introduction Enteral feeding is the preferred method of feeding in the presence of a functional gastrointestinal tract. Enteral nutrition support has been demonstrated to be superior to parenteral nutrition and no nutrition support. Numerous studies and meta-analyses have shown enteral nutritional support decreases complications and improves patient specific outcomes. Improved clinical outcomes include: enhanced wound healing, decreased infectious complications, and shorter hospital lengths of stay. Enteral nutrition may be viewed as a therapeutic tool that can alter a patient’s disease course towards a more favorable outcome.

Percutaneous enteral feeding tubes are indicated in patients requiring long-term (>4–6 weeks) enteral access. Percutaneous feeding tubes may be placed by endoscopic, fluoroscopic, or surgical methods. Ponsky & Gauderer first described percutaneous endoscopic placement of gastrostomy (PEG) tubes in 1980 [1]. Types of percutaneous feeding tubes include: PEG (percutaneous endoscopic gastrostomy), PEGJ (percutaneous endoscopic gastrojejunostomy) and DPEJ (direct percutaneous endoscopic jejunostomy). In this article we will review the indications, techniques, and complications of endoscopic placement of percutaneous feeding tubes.

Indications Percutaneous feeding tubes are indicated for patients who are unable to consume sufficient nutrients orally but have a functional gastrointestinal tract. The most common indications for PEG placement are impaired swallowing because of neurological events, oropharyngeal or esophageal cancer, dysphagia and severe facial trauma, and poor volitional intake. Placement for supplemental feedings in highly catabolic states including burn/trauma and cystic fibrosis is also appropriate. Gastrostomy placement can also be used for delivery of medications and hydration or for gastric decompression in cases of upper gastrointestinal obstruction or motility disorder.

Small-bowel delivery of enteral nutrition with either PEGJ or DPEJ is considered when a patient’s condition is complicated by gastroparesis, gastroesophageal reflux, gastric resection, pancreatitis, severe aspiration risk, or gastric feeding intolerance. PEGJ tubes may also be WEO “How I Do It” Endoscopic placement of percutaneous feeding tubes 2

preferred when enteral feeding into the small bowel with simultaneous gastric decompression is desired. If a patient has a pre-existing PEG tube, conversion to PEGJ does not require an additional skin puncture. The choice between PEGJ and DPEJ depends on local expertise and availability; however DPEJ tubes have longer functional capability but may require more expertise to place.

Contraindications Absolute contraindications to PEG placement are the same as those for upper gastrointestinal endoscopy as well as an inability to transilluminate the abdominal wall and oppose the anterior gastric wall. Relative contraindications to PEG placement include coagulopathy, gastric varices, morbid obesity, prior gastrointestinal surgery, ascites, chronic ambulatory peritoneal dialysis, and neoplastic, infiltrative, or inflammatory disease of the abdominal wall [2]. Similar contraindications apply to PEGJ and DPEJ tubes.

Technique: Patient preparation PEG is the second most common indication for upper gastrointestinal endoscopy and the procedure is performed using conscious sedation. PEG is considered a high bleeding risk procedure by the American Society of Gastrointestinal Endoscopy and checking preprocedural 3

coagulation parameters is recommended. Platelet count >50 000/mm and international normalized ratio (INR) of 1.4 are generally acceptable. Warfarin should be discontinued for 3–5 days and low molecular weight heparin 8 hours before the procedure. Discontinuation of antiplatelet agents 7–10 days prior to the procedure should be considered, depending on the risk of thromboembolism. Prophylactic antibiotics are also recommended to prevent peristomal infection.

Patient preparation is the same for PEGJ or DPEJ unless there is a pre-existing gastrostomy for a PEGJ procedure and no new skin incision is required. In this situation I do not give prophylactic antibiotics or check coagulation parameters.

Technique: Percutaneous gastrostomy (PEG) The most commonly performed methods for PEG placement are the Ponsky pull technique, which I use as described below, and the push technique. Two physicians or one physician and a GI assistant are required to perform either procedure. After advancement of the endoscope, the stomach is insufflated with air and an optimal site for PEG placement is determined by simultaneously transilluminating the gastric/abdominal wall and indenting the abdominal wall with a finger while visualizing that indentation endoscopically. Sterile techniques should be followed for the percutaneous component of the procedure. The abdominal wall and peritoneum are anesthesized by injection of 1% lidocaine. To help confirm that no loops of bowel are interposed

WEO “How I Do It” Endoscopic placement of percutaneous feeding tubes 3

between the stomach and abdominal wall, the “safe track” maneuver should be performed. The anesthetic needle (finder needle) is advanced into the stomach while aspirating with the plunger.

Simultaneously, the endoscopist confirms gastric puncture by visual inspection and with air aspiration in the syringe.

After the safe track has been confirmed, a small incision is made and a trocar is inserted through the abdominal wall into the stomach. A guide wire is passed through this trocar and grasped endoscopically. The guide wire is then drawn out through the mouth and a gastrostomy tube is affixed to it. Finally the guide wire is pulled back through the esophagus, stomach, and abdominal wall and held in place by the solid mushroom-type internal retention device and an external bumper. The external bumper should be placed ~1 cm or more from the abdominal wall with the dressing placed over (rather than under) the bumper to avoid wound compression, infection and tissue breakdown.

In the push technique, after the guide wire has been drawn out through the mouth, a PEG tube is pushed into place over the guide wire which extends from the mouth to the stoma site and the guide wire removed. The procedure is otherwise the same as the pull technique. Reported success rates for PEG placement range from 95% to 100%.

Technique: Percutaneous gastrojejunostomy (PEGJ) Multiple methods have been described for PEGJ placement, but we have found the following two over-the-wire methods to be the most efficient and successful. Commercial kits are available that include 9–12-F jejunal extension tubes that can be placed through 18–28-Fr PEG tubes. The jejunal extension tube of a PEGJ can be placed immediately following PEG placement or through an existing PEG/gastrostomy stoma created previously.

A guide wire is placed through an existing gastrostomy and is grasped endoscopically with a forceps or snare and carried into the jejunum. Using the forceps or snare to maintain the guide wire in the jejunum as the endoscope is withdrawn helps to prevent dislodgment of the guide wire. This technique requires the forceps to be pushed forward through the biopsy channel as the endoscope is slowly withdrawn. Once the endoscope is in the stomach, the guide wire can be released and the jejunal extension tube is then threaded over the guide wire into the small bowel.

Another technique uses an ultrathin endoscope (~5 mm diameter) passed through a 28-Fr PEG tube or mature gastrostomy tract into the small intestine [3]. A guide wire is fed through the endoscope deep into the small bowel and the endoscope removed. The jejunal tube is then passed over the wire into position and the wire removed (Figure 1).

With either method, we find wire placement and jejunal extension tube advancement can be done more easily with fluoroscopic guidance though it is not essential. Success rates for PEGJ range from 90% to 100%.

WEO “How I Do It” Endoscopic placement of percutaneous feeding tubes 4

Figure 1 Percutaneous endoscopic gastrojejunostomy (PEGJ) placement: ultrathin endoscope method. A Ultrathin endoscope advanced into jejunum through gastrostomy track.

B Endoscope removed, leaving wire in place. C Gastrojejunal feeding tube advanced over wire into place.

WEO “How I Do It� Endoscopic placement of percutaneous feeding tubes 5

Technique: Direct percutaneous jejunostomy (DPEJ) Direct percutaneous endoscopic jejunostomy (DPEJ) is performed in a manner similar to that of the PEG ‘pull’ technique. A pediatric colonoscope or dedicated push enteroscope is advanced to the small bowel and transillumination and finger indentation is performed over the jejunum (Figure 2). The site is most commonly found in the left mid/lower quadrant, but it may be anywhere on the abdominal wall as long as transillumination and finger indentation are adequate.

The site is stabilized by first snaring the anesthetic (finder) needle. Then the larger hollow trocar is passed next to the finder needle, and the finder needle is released and the trocar grasped with the snare. The insertion wire is passed through the trocar and grasped endoscopically. The safe track technique is used when passing both the finder needle and trocar, to ensure no bowel loop or stomach is interposed. The remainder of the procedure is as described for the PEG ‘pull’ technique.

Glucagon may be used to decrease small bowel peristalsis, and performing the procedure under general anesthesia may also increase success rate. Fluoroscopy has been reported to aid in locating an appropriate site though we have not found it useful. Procedure times are much longer than for PEG and comparable to those for PEGJ. Finding an appropriate site for stoma placement accounts for most of the additional time.

Direct percutaneous endoscopic jejunostomy placement is successful in 68%–100% of attempts [4].

Figure 2 (part 1, A–D) Direct percutaneous endoscopic jejunostomy placement (DPEJ).

A The enteroscope is advanced until transillumination is achieved. B A discrete indentation should be reproducible with direct depression at the site of transillumination. C The sounding/anesthesia needle is inserted at the site of depression/transillumination and advanced until it is seen to enter the jejunal lumen under endoscopic visualization. D The needle/cannula is inserted alongside the sounding needle.

Figure 2 (part 2, E–H) Direct percutaneous endoscopic jejunostomy placement (DPEJ).

E After the needle has been removed from the indwelling cannula, the insertion wire is advanced through the cannula and grasped with a snare that extends. F The scope is then removed, and the insertion wire is drawn out with it, so that one end of the insertion wire extends from the mouth and the other end extends from the abdominal wall. G, H The attachment loop of the pull-type gastrostomy feeding tube is tethered to the mouth end of the insertion wire and the assembly is pulled internally until the feeding tube has traversed the jejunal and abdominal walls and is snug against the wall.

WEO “How I Do It” Endoscopic placement of percutaneous feeding tubes 6

WEO “How I Do It” Endoscopic placement of percutaneous feeding tubes 7

Post-procedure care Feeding through the percutaneous tubes may begin after 3 hours. Decompression and medication administration can begin immediately post-procedure. The skin around tube enterostomies should be cleaned with mild soap and water, then rinsed and dried thoroughly. Use of irritant cleansers or full strength hydrogen peroxide should be avoided as they may lead to poor wound healing and leakage around the tube. Likewise, routine use of antibiotic ointments is not advised, and dressings at the tube insertion site are necessary only for the first few days after the procedure unless there is continued drainage at the site. Our patients are scheduled for a clinic ~2 weeks post-procedure, to assess for peristomal infection and stoma healing and to ensure familiarity and compliance with tube feeding on the part of patient and caregivers.

Complications Major and minor complications rates for PEG tubes range from 0.4% to 22.5% and from13% to 43%, respectively (Table 1). Procedure-related mortality is 0%–2%. The 30-day mortality varies from 6.7% to 26%, and is related to the underlying co-morbidities of this debilitated patient population [5]. Complications associated with PEGJ and DPEJ tubes are similar in nature and frequency to those observed for PEG tubes. PEGJ tubes are also complicated by frequent (53%– 84%) malfunction due to retrograde tube migration into the stomach or tube dysfunction caused by kinking, clogging and/ or occlusion of the smaller (8–12-Fr) jejunal extension tubes [6].

Additional complications associated with the DPEJ tube are rare jejunal volvulus and small-bowel perforation.

In comparison studies, DPEJ has been demonstrated to have greater longevity and less need for re-intervention compared with PEGJ [7]. This is likely due to the greater stability in the jejunum and larger diameter of DPEJ tubes compared with PEGJ feeding tubes. Despite expert opinion, the data are controversial as to whether more distal feeding with either type of jejunal tube significantly decreases a patient’s aspiration risk.

Conclusion Placement of percutaneous enteral feeding tubes is safe and effective by endoscopic methods.

Consideration of the appropriate device, level in the gastrointestinal tract, and insertion method are critical to ensure optimal outcomes. In addition, appropriate aftercare and monitoring with early recognition and treatment of any complications are crucial to the success of percutaneous enteral nutrition access. The expert endoscopist should be able to place and manage all types of percutaneous feeding tubes.

WEO “How I Do It” Endoscopic placement of percutaneous feeding tubes 8

Table 1 Major and minor complications of enterostomy tube placement.

Reported frequency, % Major complications

Minor complications

Aspiration

0.3–1.0

Hemorrhage

0–2.5

Peritonitis/necrotizing fasciitis

0.5–1.3

Procedure-related mortality

0–2.1

Peristomal infection

5.4–30

Peristomal leakage

1–2

Buried bumper

0.3–2.4

Inadvertent removal

1.6–4.4

Fistulous tracts

0.3–6.7

References 1

Ponsky J, Gauderer M. Percutaneous endoscopic gastrostomy: a non-operative technique for feeding gastrostomy. Gastrointest Endosc 1981; 27: 9–11

2

American Society for Gastrointestinal Endoscopy. Endoscopic enteral nutritional access devices. Gastrointest Endosc 2002; 56: 796–802

3

Adler D, Goustout C, Baron T. Percutaneous transgastric placement of jejunal feeding tubes with an ultrathin endoscope. Gastrointest Endosc 2002; 2002: 106–110

4

DiSario J, Baskin W, Brown R et al. Endoscopic approaches to enteral nutritional support.

Gastrointest Endosc 2002; 55: 901–908

5

Lynch C, Fang J. Prevention and management of complications of percutaneous endoscopic gastrostomy tubes. Pract Gastroenterol 2004; 28: 66–76

6

DeLegge M. Small bowel endoscopic enteral access. Gastrointest Endosc Clin N Am 2007; 17: 663–686

7

Fan A, Baron T, Rumalla A, Harewood G. Comparison of direct percutaneous endoscopic jejunostomy and PEG with jejunal extension. Gastrointest Endosc 2002; 56

WEO “How I Do It” Endoscopic placement of percutaneous feeding tubes 9

“HOW I DO IT” Endoscopic placement of percutaneous feeding tubes Comment Suneil A. Kapadia

Introduction Dr. Fang has provided a comprehensive review regarding placement of percutaneous endoscopic feeding tubes. Nevertheless there are some differences in emphasis and I will aim to highlight these as well as the similarities.

I would agree that enteral nutritional support is the preferred method of feeding where the gastrointestinal tract is not only functioning but also accessible. I also agree that many studies have demonstrated better clinical outcomes for enteral nutritional support compared with no nutrition or parenteral nutrition.

Enteral nutrition may take the form of enteral feed supplements, nasogastric tube feeding, or percutaneous enteral tube feeding. Percutaneous enteral tube feeding in the vast majority of cases is done with percutaneous endoscopic gastrostomy (PEG) tubes as described by Ponsky & Gauderer. As Dr. Fang has highlighted, these can also be placed surgically or radiologically. In addition, PEGs can be converted into percutaneous gastrojejunostomy (PEGJ) tubes or alternatively the jejunum can be accessed directly by a direct percutaneous endoscopic jejunostomy (DPEJ).

Indications It has been stated that percutaneous enteral feeding tubes are indicated for patients requiring enteral feeding for more than 4 weeks. It has to be acknowledged that this figure is arbitrary, and deciding which patients will fall into this category is fraught with difficulty. A multidisciplinary nutrition team taking into account all aspects of nutritional support is best placed to select appropriate patients for PEG insertion.

I would agree that the most common indication for PEG placement is impaired swallowing as a result of neurological events. This might be related to cerebrovascular causes or more chronically progressive neurological conditions such as motor neurone disease and multiple sclerosis. PEGs can also be placed electively for patients undergoing head and neck surgery where access to a functioning gastrointestinal tract is temporarily prevented. Concerns have been raised about the risk of PEG stoma metastases when the pull-through technique is used and one might wish to WEO “How I Do It” Endoscopic placement of percutaneous feeding tubes 10

consider the alternatives. I would also agree with Dr. Fang’s assertion that the delivery of medication, fluid for hydration, gastric decompression, and long-term enteral nutritional support for motility disorders are potential indications for PEG placement.

The rationale for PEGJ and DPEJ has been expertly and comprehensively explained by Dr Fang and I have nothing to add or dispute.

Contraindications In my view, appropriate patient selection is as important as absolute contraindication. The terminal phase of dementia or malignant disease is associated with failure of oral intake and a poor prognosis. Death is usually imminent and unaffected by nutritional intervention; therefore PEG placement under these circumstances is invariably inappropriate [1,2].

Provided the patient is able to have an upper gastrointestinal endoscopy there are very few contraindications to PEG placement. Tense ascites, an uncontrolled coagulopathy, and active systemic infection should be regarded as contraindications [3,4].

In our unit the relative contraindications described by Dr. Fang would be classified as significant co-morbidity. In these clinical scenarios the risk of potential complications is undoubtedly increased but has to be weighed against the benefits. In such situations, it is of crucial importance to obtain informed consent from the patient or their guardian.

Technique: Patient preparation Many patients undergoing PEG placement have significant co-morbidity which increases the risk of complications [2,4]. Therefore patient preparation needs to take this into account before the patient enters the endoscopy suite. The majority of patients have the procedure under conscious sedation and local anesthetic. However, by the time patients with chronic neurological disease present for PEG insertion, their ventilatory function is often significantly impaired. Under these circumstances I invariably enlist the help of an anesthetist to facilitate careful sedation and airway management. Where sedation is to be avoided at all costs, PEG tubes can be placed radiologically. As an alternative, I have placed these endoscopically, using only anesthetic throat spray and a local anesthetic into the abdomen. Speed and an experienced team are critically important from the patient’s point of view.

The British Society of Gastroenterology (BSG) recommend a platelet count of more than 100 000/mm3 and an INR less than 1.4 [4]. However, like Dr Fang I think it is reasonable to 3

consider the procedure in patients with a platelet count as low as 50 000/mm .

I would not stop all antiplatelet agents 7–10 days prior to the procedure. BSG guidelines suggest that clopidogrel should be stopped 1 week beforehand but continuance of aspirin is reasonable [5]. Before stopping clopidogrel I would ascertain the rationale for its prescription and WEO “How I Do It” Endoscopic placement of percutaneous feeding tubes 11

if necessary liaise with the patient’s cardiologist. Warfarin should be stopped 3–5 days before the procedure. If anticoagulation is for a high risk condition such as use of mechanical heart valves then heparin can be used as a substitute and stopped 6–8 hours before the procedure. Low molecular weight heparin should be avoided on the day of PEG placement. Warfarin can be recommenced on the day of the procedure and low molecular weight heparin on the day after [5].

Patients on peritoneal dialysis should have their dialysis stopped 2–3 days prior to PEG placement. Patients with portal hypertension can have a PEG placed with the aid of endoscopic ultrasound to avoid gastric varices.

I am in entire agreement with Dr. Fang’s comments on PEGJ and DPEJ. Technique: Percutaneous gastrostomy Dr. Fang’s technique for PEG insertion is essentially similar to mine with a few differences worth highlighting. He has not mentioned patient position at the time of endoscopy. Unlike routine upper gastrointestinal endoscopy where patients are in the left lateral position, I carry out this procedure with patients in a supine position. This facilitates transillumination and indenting the abdominal wall as already described. Where there is concern about patients lying on their back, PEG insertion can be done with the patient in the left lateral position. This position is generally a little more awkward for the person introducing the trocar but not for the endoscopist.

A comparison of 12-Fr and 20-Fr PEG tubes showed no difference regarding long-term patency or complications [6}. The 15-Fr PEG tube is one of the most widely used in the United Kingdom, in contrast to the larger sizes used in the United States [7]. I therefore find the incision size required is rarely more than 5 mm and is approximately the same as the width of the scalpel provided in the 15-Fr PEG kit we use. Once the trocar has penetrated the stomach I ensure that this is grasped with a snare prior to the introduction of the guide wire. This prevents the trocar’s being accidentally displaced if the patient were to cough or retch. The guide wire and gastrostomy tube should be pulled with gentle traction. (I liken this to pulling on the placenta – perhaps an experience not shared by many gastroenterologists!)

Once the PEG has been pulled into position it is important to ensure the internal retaining device is against the gastric mucosa. My preferred method is to check the external markings on the gastrostomy tube which provides an indication of the depth of the retaining device. For the majority of patients this is rarely more than 3 or 4 cm. When there is doubt one can repeat the gastroscopy to visually inspect the internal fixing device.

Unlike Dr Fang I have no strong views as to whether the dressing should be placed under or over the bumper. Indeed many would argue that a dressing is not necessary, particularly if there is a snug fit around the PEG. I would however be emphatic that the external device should not be more than 1 cm beyond the abdominal wall. In my view this helps reduce the risk of WEO “How I Do It” Endoscopic placement of percutaneous feeding tubes 12

pneumoperitoneum and leakage by keeping the gastric wall in close contact with the anterior abdominal wall and reducing lateral movement of the tube.

Prophylactic antibiotics are ideally given at the time of sedation to reduce the risk of early infectious complications [8].

Technique: Percutaneous gastrojejunostomy I would agree with the description given by Dr. Fang with some notable exceptions. These relate to differences between the UK and USA in the size of PEG tubes used. One of the most commonly used sizes in the United Kingdom is 15-Fr which will easily accommodate a 9-Fr jejunal extension tube. A 28-Fr PEG tube is rarely used and therefore the use of an ultrathin endoscope or bronchoscope equally uncommon.

Technique: Direct percutaneous jejunostomy Again I would agree with Dr Fang’s description for the placement of a DPEJ. However, we would tend to use hyoscine butylbromide instead of glucagon and I would not go as far as resorting to a general anesthetic or fluoroscopy to aid placement.

Post-procedure care Patients normally have nil by PEG or mouth and are observed hourly, for 4 hours. If the observations are satisfactory, feeding can be introduced via the PEG thereafter. Some of our patients, after a satisfactory 4-hour period of observation, return directly to the nonmedical institution from which they were referred. Under these circumstances I allow clear fluids to be infused via the PEG after 4 hours but defer feeding until the following day.

The PEG site should be treated as a surgical wound and the use of dressings should be discouraged and unnecessary after the first 48 hours. I agree with Dr Fang’s view on the avoidance of irritant skin cleansers and antibiotic ointments. Patients and their carers are given guidance about PEG care and appropriate dietetic and nursing support in the community.

Complications The true incidence of complications related to percutaneous enteral tube feeding is difficult to quantify. The differences in population and definitions largely account for the variation in complication rates. Complications can be classified according to when they arise and whether they are major or minor. Those arising within 48 hours of the procedure are categorized as immediate; those occurring within 2–4 weeks as early; and those occurring after 4 weeks, or when the fistula is fully formed, as late.

Directly procedure-related mortality rates are low and usually less than 1% but may be as high as

WEO “How I Do It” Endoscopic placement of percutaneous feeding tubes 13

2%. The 30-day mortality rates can vary from 10% to 28% depending on patient selection and underlying disease [4].

Major complications as tabulated by Dr. Fang occur in about 3% of PEG placements [4]. They are related to a number of factors including extreme old age, diabetes, and significant underlying comorbidity [2]. The minor complications upon which Dr. Fang comments are probably more common than is reported in the literature and occur in approximately 20% of patients.

I would agree that complications of DPEJ are similar to those of PEG but moderate or severe complication rates have been reported to be as high as 10% [9]. I also have the view that tube dysfunction in PEGJ is a particular problem due to a combination of length, narrow diameter, and poor care.

Conclusion The provision of a percutaneous enteral tube feeding service should be routine for most gastrointestinal units. Optimal outcomes can only be achieved through careful patient selection, pre-assessment, and post-procedural care, as part of a multidisciplinary team.

References 1

Murphy LM, Lipman TO. Percutaneous endoscopic gastrostomy does not prolong survival in patients with dementia. Arch Intern Med 2003; 163: 1351–1353

2

Johnston SD, Tham TCK, Mason M. Death after PEG: results of the National Confidential Enquiry into patient outcome and death. Gastrointest Endosc 2008; 68: 223–227

3

Figueiredo FA, da Costa MC, Pelosi AD et al. Predicting outcomes and complications of percutaneous endoscopic gastrostomy. Endoscopy 2007; 39: 333–338

4

O’Toole P. Complications associated with the placement of percutaneous endoscopic gastrostomy. In: Green J (editor). BSG guidelines, Complications of gastrointestinal endoscopy, Nov 2006; 26–30. Available at: http://www.bsg.org.uk/pdf_word_docs/complications.pdf

5

Veitch AM, Baglin TP, Gershlick AH et al. Guidelines for the management of anticoagulant and antiplatelet therapy in patients undergoing endoscopic procedures. Gut 2008; 57: 1322–1359

6

Duncan HD, Bray MJ, Kapadia SA et al. Prospective randomized comparison of two different sized percutaneous endoscopically placed gastrostomy tubes. Clin Nutr 1996; 15: 317–320

7

Merrick S, Harnden S, Shetty S et al. An evaluation of the “cut and push” method of endoscopic gastrostomy (PEG) removal. JPEN J Parenter Enteral Nutr 2008; 32: 78–80

8

Ahmad I, Mouncher A, Abdoolah A et al. antibiotic prophylaxis for percutaneous endoscopic gastrostomy – a prospective randomized double blind trial. Aliment Pharmacol Ther 2003; 15: 209–215

9

Maple JT, Petersen BT, Baron TH et al. Direct percutaneous endoscopic jejunostomy: outcomes in 307 consecutive attempts. Am J Gastroenterology 2005; 100: 2681–2688

WEO “How I Do It” Endoscopic placement of percutaneous feeding tubes 14

Endoscopic placement of percutaneous feeding tubes Summary Spiros D. Ladas Drs. Fang from the USA and Kapadia from the UK have excellently presented how they perform percutaneous endoscopic gastrostomy (PEG) and jejunostomy procedures. They both agree on the indications, contraindications and patient preparation. They describe in detail the technique of the PEG procedure, so that their text can be used by junior endoscopists as a reference guide. I am going, therefore, to focus on issues related to the avoidance of complications and the aftercare for PEG. Currently available endoscopic techniques include percutaneous endoscopic gastrostomy (PEG), jejunostomy (PEJ) and gastrojejunostomy (PEGJ). The commonest indications are provision of nutrition in patients who cannot swallow due to neuromuscular impairment (e.g. cerebral vascular accident [CVA], myopathies, or polyneuritis) or obstruction (e.g. head and neck cancer), in children with failure to thrive (e.g., because of cystic fibrosis), and gastric decompression (e.g., with intestinal pseudo-obstruction).

PEG is a major nonsurgical procedure. Patient selection is of paramount importance in minimizing complications and mortality. PEG feeding is justified for patients who need tube feeding for more than 30 days [1], because the 30-day mortality may be as high as 30% in hospitalized patients, mostly related to underlying comorbidities [2].

Informed consent processes Published studies have indicated that many of those making decisions (patients or relatives) are not satisfied with the quality of information given before informed consent. A survey of our institution showed that although 93% of the decision-makers considered that their opinion had been taken into account when the procedure was done, 25% said that they had not been adequately been informed about alternative methods, and 38% about the complications of the procedure [3]. It is therefore important to fully inform decision-makers before asking them to sign the informed consent document. This will avoid medicolegal consequences.

Selection of PEG tube size OMED “How I do it” descriptions are intended for use all over the world, i.e., in developed and developing countries, and the families of patients using PEG tube feeding could be of varying socioeconomic status; this has a bearing on the selection of PEG tube size.

WEO “How I Do It” Endoscopic placement of percutaneous feeding tubes 15

Two types of nutrients may be used for PEG tube feeding. The first type are commercially available nutrient solutions, which may be given with continuous drip or pump through the PEG tube. The daily cost of home enteral nutrition in most European countries is 10€–25€ and it is fully funded in many countries [4]. The second type is home-cooked food, which is blended, diluted with tapwater and injected using a 60-ml syringe through the PEG tube. However, the cost of commercial solutions is much higher than that of home-cooked food. The endoscopist should discuss with the caregiver which type of nutrient he/she prefers to use for the patient, and place a 15-Fr PEG tube for use with commercially available nutrient solutions, or an 18- or 20-Fr PEG tube for home-cooked food.

Safe PEG tube placement Transillumination of the abdominal wall and clear visualization of the indentation and movement of the stomach wall during ballottement is necessary to detect the best abdominal wall puncture site and to avoid puncturing neighboring organs such as the colon. Dr. Fang uses the “safe track” technique, i.e, he slowly advances a syringe, partially filled with saline, with the barrel of the syringe pulled back to create negative pressure. If air is aspirated before the needle appears in the gastric lumen, it is very likely that an interposed hollow viscus has been punctured.

Wound infection Wound infection is the most common complication of the PEG procedure, with a reported incidence of 4.3%–16%. In addition, 3%–8% of patients may develop local cellulitis or abscess formation. Necrotizing fasciitis is the most severe end of the spectrum of local infectious complications. Measures to prevent wound infection should include: •

a strictly aseptic insertion technique

•

skin incisions of adequate size to avoid skin ischemia and necrosis

•

avoidance of excessive tension of the retention discs which may also lead to edema and ischemia of the skin and gastric mucosa

•

use of prophylactic antibiotics, and perhaps

•

the use of mouthwash with antiseptic solutions

•

good post-insertion care of the PEG site.

A meta-analysis of randomized control trials confirmed that a single intravenous dose of a broadspectrum antibiotic, given approximately 30 minutes before PEG is effective in reducing the incidence of wound infection [5]. The most frequently used antibiotics are either intravenous amoxicillin + clavulanate 2.2 g or one dose of a second- or third-generation cephalosporin given

WEO “How I Do It” Endoscopic placement of percutaneous feeding tubes 16

30 minutes before the procedure.

Post-procedure care Accidental PEG removal Since the gastrostomy fistula heals completely within 24–48 hours after removal of the PEG tube, I advise caregivers that in case of accidental PEG tube removal, the family physician should introduce a Foley catheter into the gastrocutaneous fistula to keep it patent. This will give time to transfer the patient to hospital for a tube replacement on an outpatient basis the next day. If the fistulous tract is shrunk (stenosed) and a replacement PEG tube cannot be introduced, I dilate it with a Savary dilator over a guide wire introduced into the stomach through the fistulous tract.

PEG tube removal and replacement PEG tubes are frequently used for temporarily feeding patients in intensive care units or after a stroke, and therefore removal of the tube is often done, after the patient has recovered and can start oral feeding. After PEG tube removal the external orifice of the fistulous tract is covered with a sterile compress and the patient can immediately start oral feeding. Leak of gastric content is minimal and stops within 1–3 days.

The durability of a PEG tube is related to careful handling, and they may remain in situ for more than 1 year. However, PEG tubes need to be replaced when they become clogged from food or because the external catheter cap is broken or the tube’s integrity is altered. Tube blockage results from protein-enriched formulae, medications, and most importantly, from insufficient tube care. The smaller-sized tubes (e.g. 9 Fr) are prone to clogging. Polyurethane materials are recommended in preference to silicone PEGs as the latter are more susceptible to material deterioration and tube failure [7].

PEG tube replacement does not require a new endoscopy and it is done in the office setting. I usually remove a PEG tube as follows. The patient is placed supine and the abdominal skin round the PEG fistula is prepared with a topical antiseptic, e.g. povidone-iodine. The tube is then well lubricated and it is advanced into the stomach to lubricate the fistulous tract. I place my hand against the abdominal wall, near the fistula orifice, and with the other hand grasp the PEG tube and pull it out with steady, sudden tension. In this way the bumper (mushroom-type) of the PEG tube is deformed and comes out. A replacement gastrostomy tube or button (balloon-type) is then inserted through the fistulous tract.

An alternative way of removing the PEG tube is to cut away the external catheter and either remove the internal fixation bumper endoscopically by catching it with a snare or allow it to be passed naturally. In children, PEG tubes must be removed endoscopically. Certain types of PEG tube, e.g. the Freka, cannot be removed by external traction because the gastric bumper is not

WEO “How I Do It” Endoscopic placement of percutaneous feeding tubes 17

flexible. There are several reports of ileus and need for operation after the external catheter had been cut away, leaving the bumper to pass out naturally. However, recently published reports describe uneventful passing of the Freka internal bumper with feces [8].

References 1

American Society for Gastrointestinal Endoscopy. Role of endoscopy in enteral feeding.

Gastrointest Endosc 2002; 55: 794–797

2

Abuksis G, Mor M, Segal N et al. Percutaneous endoscopic gastrostomy: high mortality rates in hospitalized patients. Am J Gastroenterol 2000; 95: 128–132

3

Ladas SD, Triantafyllou K, Liappas I et al. Percutaneous endoscopic gastrostomy: adequacy and quality of information given to decision makers. Dig Dis 2002; 20: 289–292

4

Hebuterne X, Bozzetti F, Moreno Villares JM et al. Home enteral nutrition in adults: a European multicentre survey. Clin Nutr 2003; 22: 261–266

5

Jafri NS, Mahid SS, Minor KS et al. Meta-analysis: antibiotic prophylaxis to prevent peristomal infection following percutaneous endoscopic gastrostomy. Aliment Pharmacol Ther 2007; 25: 647–656

6

Bechtold ML, Matteson ML, Choudhary A et al. Early versus delayed feeding after placement of a percutaneous endoscopic gastrostomy: a meta-analysis. Am J Gastroenterol 2008; 103: 2919–24

7

Blacka J, Donoghue J, Sutherland M et al. Dwell time and functional failure in percutaneous endoscopic gastrostomy tubes: a prospective randomized-controlled comparison between silicon polymer and polyurethane percutaneous endoscopic gastrostomy tubes. Aliment Pharmacol Ther 2004; 20: 875–882

8

Kejariwal D, Bromley D, Miao Y. The “cut and push” method of percutaneous endoscopic gastrostomy tube removal in adult patients: the Ipswich experience. Nutr Clin Pract 2009; 24: 281–283

WEO “How I Do It” Endoscopic placement of percutaneous feeding tubes 18

“HOW I DO IT” Approach to lower gastrointestinal bleeding AUTHORSHIP

How I do it: Approach to lower gastrointestinal bleeding Roque Sáenz MD Eduardo Valdivieso MD, MSc The Latin-American Advanced Gastrointestinal Endoscopy Training Center Clínica Alemana, Universidad del Desarrollo Santiago, Chile

Comment David J. Bjorkman MD, MSPH University of Utah School of Medicine Salt Lake City, Utah USA

Summary Tamás Molnár MD, PhD Tibor Wittmann MD, PhD First Department of Medicine University of Szeged Szeged Hungary

“HOW I DO IT” Approach to lower gastrointestinal bleeding How I Do It Roque Sáenz MD & Eduardo Valdivieso Introduction Lower gastrointestinal bleeding (LGIB) ranges from trivial to massive life-threatening blood loss and is defined as bleeding that emanates from a source distal to the ligament of Treitz.

LGIB represents up to one third of the total cases of gastrointestinal bleeding. It is more frequent in males and older patients. The mortality rate is about 3%–6%, and the prognosis is worse when LGIB occurs during hospitalization [1].

The bleeding is from the small bowel in about 9% of cases and the source remains undiagnosed in 6% of patients with LGIB [2].

In practice, there are three common clinical scenarios related to LGIB: active bleeding, recent bleeding, and chronic bleeding. International nomenclature defines acute LGIB as a bleeding of less than 3 days’ duration associated with instability of vital signs, anemia, and/or need for blood transfusion. Chronic LGIB is any passage of blood per rectum that results from intermittent or slow loss of blood.

Although 80% of all LGIB stops spontaneously, and hemodynamic instability and the need for transfusion are rare, the identification of the bleeding source often remains challenging and rebleeding can occur in 25% of cases [3].

Clinical scenario Clinically, the most common presentation of LGIB is hematochezia, rectorrhagia, but melena, hemodynamic instability, anemia and abdominal pain can be seen.

Hematochezia is defined as gross blood seen either on toilet paper after defecation or mixed with stool. Rectorrhagia means the expulsion of fresh red bright blood without stools, while melena is defined as black stools resulting from the oxidation of hematin in the gut.

Hematochezia and rectorrhagia are frequent forms of LGIB, but melena could be present in the setting of constipation.

Because of rapid transit time or massive bleeding, approximately 10% of cases of rectorrhagia, are secondary to upper gastrointestinal bleeding (UGIB) [2].

Chronic bleeding can present as occult fecal blood, occasional episodes of melena, or even small quantities of rectorrhagia.

WEO “How I Do It” Approach to lower gastrointestinal bleeding

2

Unfortunately clinical signs are not reliable predictors of the source or severity of LGIB. As a result, the approach to a case of LGIB depends mainly on the clinical setting, suggesting either acute or chronic bleeding [4].

The hemodynamic stability of the patient and the rate of bleeding dictate the order in which diagnostic and therapeutic procedures should be undertaken.

An initial hematocrit of less than 35%, abnormal vital signs 1 hour after medical management and gross blood on initial rectal examination are all predictors of severe LGIB and adverse outcome [5].

Attention should be paid to anal diseases, which are a common cause of rectal bleeding. A history of anal pain, pruritus, and fresh red blood on the toilet paper is frequently the expression of anal fissures or hemorrhoids.

Anal inspection, anoscopy and rectal examination are mandatory, because patients with clear anal pathology will not require immediate colonoscopy or further investigations.

LGIB associated with an elevated blood urea nitrogen (BUN) or serum bilirubin denotes an increased probability of UGIB. Nasogastric tube aspiration can help identify an upper gastrontestinal bleeding source if it is positive, but a clear nasogastric tube aspirate does not rule out an upper source. The presence of bile in the aspirate suggests that duodenal contents have been sampled and further decreases the likelihood of an upper gastrointestinal source.

The presence of blood, clots or “coffee ground” material in the aspirate confirms UGIB and upper gastrointestinal endoscopy should follow immediately.

After aspiration of the gastric contents, the nasogastric tube can also be utilized to start bowel preparation with the instillation of 3–4 liters of polyethylene glycol (PEG) solution at a rate of 1 liter every 30 minutes. PEG, or phosphosoda (45 ml twice), is usually given orally. It usually provides adequate colon cleansing within 3–4 hours. Intravenous prokinetics may be used in order to avoid vomiting and to assure the purge progression.

The patient’s age affects the clinical approach to LGIB. In children and young patients, cow milk allergy, polyps, Meckel’s diverticulum, inflammatory bowel diseases, and anal diseases should be especially considered. Older patients are more likely to have diverticular disease, angiodysplasia, ischemic diseases, colorectal cancer, inflammatory bowel diseases, polyps, nonsteroidal anti-inflammatory drug (NSAID)-induced lesions, and lesions from prior radiotherapy, and may also have anal diseases. Other vascular malformations are infrequent but should occasionally be considered.

Young women could be affected by endometriosis, and when bleeding is episodic or chronic colonoscopy would be better scheduled for menstruation days.

Specific indications and contraindications for diagnostic and therapeutic procedures The diagnostic accuracy of colonoscopy ranges from 72% to 86% in the setting of LGIB. Cecal WEO “How I Do It” Approach to lower gastrointestinal bleeding

3

intubation is usually achieved in more than 95% of attempts. The yield from colonoscopy is greater when done earlier in the hospital stay, and patients who undergo colonoscopy for LGIB are reported to have a shorter length of stay compared with those who do not [6]. Colonoscopy leads to the greatest number of specific diagnoses and directed therapy in acute LGIB. For the vast majority of patients with LGIB, colonoscopy is the only intervention needed [7].

A negative technetium-labeled red blood cell nuclear scan rules out active bleeding. The test can detect bleeding at the rate of 0.1 ml/min or greater. However, the results of scintigraphy can be difficult to interpret, have poor accuracy in locating a bleeding site, and are poor predictors of subsequent angiogram results. A limited colonic resection should not be based on only a positive nuclear scan [8].

If the rate of ongoing arterial bleeding is at least 0.5 ml/min, angiography may show extravasations of contrast into the lumen and identify a bleeding site. Angiography is an invasive procedure, which can result in complications including contrast-induced renal failure, arterial injury, and mesenteric ischemia. The accuracy of this procedure can be quite variable; it detects only active bleeding and may miss lesions that bleed intermittently. Angiography can identify neoplastic or vascular disorders and their locations, and can also reveal vascular ectasias based upon the characteristic angiographic appearance. Angiographic therapy can be provided, by means of intra-arterial vasopressin or terlipressin infusions. Intra-arterial vasopressin infusion reduces the bleeding in 70%–80% but the effect is not permanent.

Polyvinyl alcohol particles or microcoils, can be placed to occlude vessels; these have a more persistent effect, but increase the risk of ischemia. Unfortunately, angiography is not readily available in all settings, particularly at night or weekends. The use of systemic somatostatin or octreotide infusion is controversial. These drugs reduce the splanchnic blood supply, permitting the spontaneous cessation of bleeding or facilitating local endoscopic treatment.

Computed tomography (CT) angiography is a helpful new diagnostic tool that permits the evaluation of the celiac axis as well as mesenteric arteries.

This technique may offer a very sensitive means to evaluate the source of acute LGIB, while avoiding some of the morbidity and resource-intensiveness of contrast angiography, and may provide unique morphologic information regarding the type of pathology. Investigation with the more rapid and more widely available multidetector CT (MDCT), followed by either directed therapeutic angiography or surgical management, may represent a reasonable algorithm for the early evaluation and management of acute LGIB in which an active bleeding source is strongly suspected [9].

Barium contrast studies have no role in the assessment of LGIB and should be avoided. These procedures have low accuracy and can interfere with subsequent evaluation by colonoscopy or angiography.

Occasionally, the LGIB source is in the small intestine. Push enteroscopy and capsule endoscopy have been used for the assessment of obscure bleeding [10]. New double- and WEO “How I Do It” Approach to lower gastrointestinal bleeding

4

single-balloon technologies are employed in an improved method, which can be considered for the assessment of acute LGIB; these single- or double-balloon enteroscopes have replaced the push enteroscope where they are available. Enteroscopes have a therapeutic capability, for instance for coagulation, polypectomy or mucosal resection, and for marking for subsequent surgical treatment. Patients with a suspected bleeding area, defined by angiography, scintigraphy or capsule, may be candidates for directed enteroscopic therapy; however, its utility as a primary diagnostic/therapeutic procedure needs further evaluation.

Capsule endoscopy is easier to perform, does not require sedation, and is becoming increasingly cheaper and more widely available. Capsule enteroscopy is considered the gold standard for evaluating patients with obscure gastrointestinal bleeding, but its cost-effectiveness is uncertain [11]. Capsule endoscopy identifies findings in 58% of patients compared with 28% with other imaging procedures [12].

In our experience, video capsule endoscopy is best utilized after a negative colonoscopy, but others recommend its earlier use as a first-step procedure for cases of midgut bleeding.

In exceptional conditions, an intraoperative endoscopy (possibly through an enterotomy) should be performed, because of the persistence of bleeding with a probable small-bowel origin that cannot be identified by other investigations.

Surgeons should be a part of the medical team from the very beginning, in case surgical resection is required. Surgery is indicated when more than six units have been transfused, or when hemodynamically significant persistent or recurrent bleeding occurs. Ideally surgery should be directed at a specific bowel segment based upon prior evaluation. Empiric subtotal colectomy should be reserved for life-threatening bleeding that cannot be located by vigorous diagnostic efforts. The overall surgical mortality is about 5%–10%.

Older patients or those with co-morbidities should be considered for earlier surgery, to avoid the complications of multiple transfusion, prolonged hemodynamic instability, cardiovascular impairment or coagulation disorders.

For all cases of confirmed or suspected LGIB, colonoscopy should be the diagnostic and therapeutic procedure of choice. The use of other modalities in the setting of continuous, clinically significant bleeding depends upon the availability of other diagnostic and therapeutic methods (angiography, scintigraphy, etc).

Algorithm for management of LGIB The very first steps must be to assure hemodynamic stability and to determine the need for hospitalization in an intensive care unit. Volume resuscitation should begin at presentation and continue through the diagnostic evaluation, as needed.

Adequate vascular access is essential. Coagulation tests, hemoglobin level, blood group information, and blood to be used for labeled red cells scintigraphy can be obtained when an

WEO “How I Do It�

Approach to lower gastrointestinal bleeding

5

intravenous line is placed.

As a guide, we propose the algorithm for management of acute LGBI that is shown in Figure 1.

Figure 1. Management algorithm for lower gastrointestinal bleeding (LGIB).

Acute lower gastrointestinal bleeding (LGIB)

Resuscitation and Evaluation Physical examination/Orthostatics Complete blood count (CBC)/Electrolytes/ Coagulation/Blood Type & Cross-match

Consider nasogastric tube

Negative aspirate or Aspirate not done

Oral purge (Polyethylene glycol [PEG] lavage)

Negative EGD

Colonoscopy

Massive bleeding

Positive aspirate or Risk factors for upper gastrointestinal bleeding (UGIB)

Surgical

Esophagogastroduodenoscopy (EGD)

Positive EGD: See UGIB algorithm

consultation

Colonoscopy

Successful

Observation

Successful angiography Embolization or intra-arterial injection

Observation

Persistent bleeding

Faster bleeding (≼0.5 ml/min): Angiography

Persistent bleeding

Slower bleeding (<0.5 ml/min): Vasoconstrictor drugs

Repeat colonoscopy

OME Surgery

6

We have also used another management algorithm for acute LGIB, which is based on colonoscopic findings; this is shown in Figure 2.

Figure 2. Management algorithm for lower gastrointestinal bleeding (LGIB), based on colonoscopy findings.

Colonoscopy

Source of bleeding identified

Unable to visualize colon endoscopically due to severity of bleeding

Endoscopic therapy Faster bleeding (>0.5 ml/min): Angiography Persistent recurrent bleeding Bleeding continues

Slower bleeding (>0.5 ml/min): Vasoconstrictor drugs

Repeat colonoscopy

Angiography Preoperative location

Bleeding stopped

Continued bleeding Consider repeat colonoscopy Surgery

Surgery

Evaluation of small bowel Esophagogastroduodeno-scopy (EGD) if not done previously Consider repeat colonoscopy

WEO â&#x20AC;&#x153;How I Do Itâ&#x20AC;?

Approach to lower gastrointestinal bleeding

7

Technique of colonoscopy in acute LGIB Restoration of hemodynamic stability is the most important first step, but bowel cleansing can begin as fluid resuscitation is being carried out. Ideally, colonoscopy should be performed as early as possible, within the first 12â&#x20AC;&#x201C;24 hours of admission.

Careful examination, meticulously clearing retained blood and stool, is important. If angiography or scintigraphy has been performed previously, special attention should be paid to areas identified as suspicious by those studies. Care should be taken to avoid excessive air insufflation during the procedure. Repeated aspiration of excess air can reduce the risk of subsequent abdominal distension.

Every effort should be made to visualize the terminal ileum during the colonoscopy. If no blood or clots are found in the small bowel the probable origin is restricted to the colon. A gentle and meticulous withdrawal, searching for active bleeding sources or stigmata, similar to the upper gastrointestinal process is required. Local rinsing in every segment is mandatory, to assure ptimal visualization.

Patient preparation The patient and relatives should be carefully informed about the clinical situation, the probable outcomes, the diagnostic/treatment plan, the associated risks and their solutions. Informed consent should be obtained for all necessary procedures with the patient having the opportunity to have all questions answered.

Sedation The specific approach to sedation for the procedure will depend upon the co-morbidities and hemodynamic stability of the patient. It may be appropriate to have the services of an anesthesiologist who can monitor and support the hemodynamic status of high-risk patients.

Equipment and accessories Wide-channel colonoscopes should be available. An experienced endoscopy assistant should be involved in the procedure. All the different devices that are potentially needed should be available at the start of the procedure. This includes, but is not limited to, injection needles and solutions, heater probes, electrocautery devices (including argon plasma coagulation, if available), clips, loops and bands.

Several 100-ml syringes and fluid for vigorous colonic lavage should be available, to flush through the biopsy channel, permitting a better view. Hydrogen peroxide solutions should be avoided because these produce blood bubbles, impairing the view.

WEO â&#x20AC;&#x153;How I Do Itâ&#x20AC;?

Approach to lower gastrointestinal bleeding

8

Hemostatic procedures Colonoscopic hemostasis techniques are based on the stigmata of hemorrhage that have been identified in the upper gastrointestinal tract as predictors of upper gastrointestinal bleeding, even though data are limited on the utility of endoscopic stigmata in predicting rebleeding in this setting [13]. Local endoscopic therapy should be reserved for those with active or recurrent bleeding.

In severe diverticulosis the location of the bleeding lesions may be indicated by the presence of fresh blood or clots.

If active bleeding is detected, local epinephrine injection, 1/10 000 or 1/20 000 solution, should be done into the area. In aged patients or patients with vascular impairment, epinephrine use is restricted, or more dilute solutions could be used. Diverticula should be examined carefully, as the bleeding vessel can sometimes be seen in the diverticular dome or on the lip of the diverticulum. The bleeding vessel can be treated with a clip. If a bleeding diverticulum is demonstrated but no vessel is visible, epinephrine injection should be done at the edge of the diverticulum.

The combination of cautery and injection therapy may be useful for visible vessels, according to uncontrolled studies. Clipping or banding could also be attempted according to the vessel characteristics. If an adherent clot is found, it can be guillotined and the underlying lesion treated more accurately. The area should be injected with dilute epinephrine prior to manipulating the clot to decrease the risk of precipitating vigorous bleeding.

Thermal therapies, using a heater probe, electrocautery, or argon plasma coagulation have similar efficacy. Bleeding lesions should be marked by a submucosal India ink tattoo, in order to facilitate location of the lesion in subsequent endoscopic or surgical therapy. In the case of angiodysplasia or angioma lesions, local epinephrine injection prior to thermal therapy is not advisable, because the lesions disappear with the vasoconstriction preventing subsequent therapy.

Submucosal saline injection can be used to produce a cushion that makes thermal treatment in the cecum and right colon less dangerous. It is also advisable to aspirate the excess air to make sure that the thickness of the bowel wall is maximized.

Some characteristics of LGIB •

The most frequent causes of LGIB are diverticular disease and angiodysplasias [14].

•

60% of bleeding patients have diverticula but frequently these are not the source of bleeding [15].

•

NSAIDs use increases the risk of LGIB from diverticula and NSAID-induced mucosal erosions.

WEO “How I Do It” Approach to lower gastrointestinal bleeding

9

•

It is unusual to find diverticulitis or local inflammation.

•

Diverticular bleeding usually presents with hematochezia, maroon material or bright red blood. It is not a common cause of occult bleeding.

•

When bleeding stops spontaneously (as occurs in 75%–80% of patients), 65%–75% of those patients do not rebleed.

•

Bleeding can occur from diverticula in the right colon.

•

Angiodysplasias are frequently seen in the cecum and ascending colon. If they are the source of bleeding or if this is the only finding in a patient with recurrent LGIB, they should be treated.

•

Angiodysplasia is present in 3%–40% of patients with occult bleeding. Colonoscopy detects 80%–90% of the cases and it is potentially therapeutic.

•

Vascular ectasias and Dieulafoy lesions could be the cause of the local bleeding, and also require local endoscopic treatment. Injection therapy plus coagulation or thermal treatment or mechanical endoscopic procedures (clipping, and banding), could be useful.

•

Sometimes if the vessel is difficult to treat, radiological intervention (embolization) or surgery is required.

•

Inflammatory bowel disease may be a cause of bleeding in 2%–6% of LGIB cases. It is more significant in Crohn’s colitis than in ulcerative colitis. Surgery is required in about 20%–30% of these patients. Bleeding is usually diffuse but occasionally a vessel in an ulcer could be clearly identified and locally treated.

•

Infectious colitis could be the cause of diffuse lesions that respond to specific treatment.

•

Ischemic colitis is another recognized cause of LGIB. It is more often diagnosed because of diffuse vascular disease, aging of the population, cocaine consumption, vigorous athletic activity, or cardiovascular surgery with hypotension and hormone therapies. It is more frequent in the splenic flexure, descending colon and sigmoid colon because of the characteristics of the blood supply.

•

Blue bubbles, ulcerated areas, and the presence of diffuse ectatic vessels and a relative absence of bleeding at the time biopsies are taken may be clues pointing to ischemic impairment. The findings may be asymmetric. A stenosis may be present if the ischemia is chronic or recurrent.

•

Dark blue, black, or green coloring of the mucosa suggests severe ischemia that requires urgent surgery in order to avoid perforation and peritonitis.

•

If a critical vascular stenosis is demonstrated, vascular surgery or radiological intervention for angioplasty and stenting could be necessary.

WEO“How I Do It” Approach to lower gastrointestinal bleeding

10

•

Polyps could be also a bleeding source, and can be treated with polypectomy.

•

Bleeding may also be seen at the site of a previous polypectomy or mucosal resection.

This can be avoided by careful attention at the time of the initial procedure. Large polyp stalks can be pretreated with clipping or looping. Sometimes adrenaline injection of the stalk produces vasoconstriction, permitting better coagulation or local inflammationof the artery that permits hemostasis. Careful attention to the cautery and coagulation settings, as well as to the snare technique can help reduce the risk of post-polypectomy bleeding.

•

Post-polypectomy bleeding that occurs immediately should be treated on the spot, by resnaring the stalk, adrenaline injection, electrocoagulation, looping or clipping.

•

Post-polypectomy bleeding can occur until the 17th day post procedure (mean 5th day).

About half of these require transfusions.

•

Aspirin, NSAIDs, antiplatelet drugs and anticoagulants can also increase the risk of postpolypectomy bleeding. The appropriate approach should be based upon the underlying need for clot prevention in the individual patient. The risk of bleeding must be balanced against potential thrombotic complications if the drug is stopped.

•

We have encountered (unpublished personal data) intestinal ischemia occurring after prolonged colonoscopy for multiple difficult procedures, and resultant overdistension of the bowel. Overinsufflation should be avoided, and excess air aspirated in these cases.

•

Colon cancer is also a cause of LGIB, and local treatment is indicated only for hemostasis, very early cancers, or palliation. Definitive surgical treatment is indicated in most cases of colon cancer.

•

Radiation colitis is also a common cause of LGIB. Prostate and gynecological cancers are still indications for radiotherapy. Vascular spiders are the cause of such bleeding. Local treatment could be achieved using argon plasma coagulation (APC) or multipolar coagulation. The lower rectum can be treated using scope retroflexion.

•

Inflammatory bowel diseases (IBDs) are also a common cause of LGIB. Colonoscopy is usually diagnostic. Ulcerative colitis is easily diagnosed due to its characteristic endoscopic features of confluent inflammation starting in the rectum and extending proximally. Crohn’s disease has a characteristic endoscopic appearance of segmentary inflammation and deep ulcers.

•

Hemodynamically significant LGIBs due to IBD are uncommon and are usually treated medically or surgically.

References 1

Longstreth GF. Epidemiology and outcome of patients hospitalized with acute lower gastrointestinal hemorrhage: a population-based study. Am J Gastroenterol 1997; 92: 419– 424

WEO “How I Do It” Approach to lower gastrointestinal bleeding

11

2

Peura DA, Lanza FL, Gostout CJ, Foutch PG. The American College of Gastroenterology Bleeding Registry: preliminary findings. Am J Gastroenterol 1997; 92: 924–928

3

Davila RE, Rajan E, Adler DG, et al. ASGE guideline: the role of endoscopy in the patient with lower-GI bleeding. Gastrointest Endosc 2005; 62: 656–660

4

Mant A, Bokey EL, Chapuis PH, et al. Rectal bleeding. Do other symptoms aid in diagnosis? Dis Colon Rectum 1989; 32: 191–196

5

Velayos FS, Williamson A, Sousa KH, et al. Early predictors of severe lower gastrointestinal bleeding and adverse outcomes: a prospective study. Clin Gastroenterol Hepatol 2004; 2: 485–490

6

Bounds BC, Friedman LS. Lower gastrointestinal bleeding. Gastroenterol Clin N Am 2003; 32: 1107–1125

7

Green BT, Rockey DC, Portwood G, et al. Urgent colonoscopy for evaluation and management of acute lower gastrointestinal hemorrhage: a randomized controlled trial. Am J Gastroenterol 2005; 100: 2395–23402

8

Levy R, Barto W, Gani J. Retrospective study of the utility of nuclear scintigraphic-labelled red cell scanning for lower gastrointestinal bleeding. ANZ J Surg 2003; 73: 205–209

9

Duchesne J, Jacome T, Serou M et al. CT-angiography for the detection of a lower gastrointestinal bleeding source. Am Surg 2005; 71: 392–397

10

de Leusse A, Vahedi K, Edery J et al. Capsule endoscopy or push enteroscopy for first-line exploration of obscure gastrointestinal bleeding? Gastroenterology 2007;132: 855–862

11

Mata A, Bordas JM, Feu F et al. Wireless capsule endoscopy in patients with obscure gastrointestinal bleeding: a comparative study with push enteroscopy. Aliment Pharmacol Ther 2004; 20: 189–194

12

Zuckerman GR, Prakash C, Askin MP, Lewis BS. AGA technical review on the evaluation and management of occult and obscure gastrointestinal bleeding. Gastroenterology 2000; 118: 201–221

13

Elta GH. Urgent colonoscopy for acute lower-GI bleeding. Gastrointest Endosc 2004; 59: 402–408

14

Jensen DM. Management of patients with severe hematochezia – with all current evidence available. Am J Gastroenterol 2005; 100: 2403–2406

15

Bloomfeld RS, Rockey DC, Shetzline MA. Endoscopic therapy of acute diverticular hemorrhage. Am J Gastroenterol 2001;9 : 2367–2372

WEO “How I Do It” Approach to lower gastrointestinal bleeding

12

“HOW I DO IT” Approach to lower gastrointestinal bleeding Comment David J. Bjorkman While of great clinical significance, lower gastrointestinal bleeding (LGIB) is a less common event than upper gastrointestinal bleeding. As a result, there are fewer prospective trials to guide our diagnostic and therapeutic approach to this condition.

The review of the causes and clinical approach presented by Drs. Sáenz and Valdivieso is excellent and comprehensive. I will not repeat the points they make, but will highlight areas where our approaches may vary slightly, and emphasize points of great importance.

As mentioned by Drs Sáenz and Valdivieso, the clinical approach to the patient will be different, depending upon whether the bleeding is acute, intermittent, or occult. I will focus on acute lower gastrointestinal bleeding (ALGIB), as they have addressed the issues of occult and intermittent bleeding which can be approached less urgently.

The causes of ALGIB vary with the patient population. Retrospective studies of the causes of LGIB have given different information in different settings and varying patient demographics. It is therefore difficult to generalize the frequency of specific causes of ALGIB to all patients.

Some common principles, however, appear to be universal. Elderly patients are more likely to bleed from diverticula and vascular lesions than younger patients. Patients taking nonsteroidal anti-inflammatory drugs (NSAIDs), antiplatelet drugs, or anticoagulants are more likely to have ALGIB. The incidence of inflammatory bowel disease and colon cancer in the local population will also affect the likelihood of these disorders causing the bleeding event.

Our approach to ALGIB is very similar to that described by Drs Sáenz and Valdivieso. The primary initial goal is to assess the hemodynamic status of the patient, and provide appropriate intravascular volume resuscitation. Tachycardia and orthostatic changes in blood pressure are indications for rapid volume replacement. Initial volume replacement is done with an isotonic solution (normal saline or lactated Ringer’s). The initial hematocrit or hemoglobin level may not be an accurate assessment of the volume of bleeding until intravascular volume has been replaced. Blood is transfused as indicated by the volume loss and underlying medical status of the patient.

After initial hemodynamic stabilization the next step is to prepare the patient for colonoscopy as soon as possible. Colonoscopy has been demonstrated to be the most accurate diagnostic test for ALGIB and can be therapeutic in many situations [1–3]. Early colonoscopy also allows for more rapid directed therapy and decreases the length of hospital stay [4].

WEO “How I Do It” Approach to lower gastrointestinal bleeding

13

Preparation for colonoscopy should begin as soon as the patient is hemodynamically stable.

This can be done be giving a polyethylene glycol (PEG) solution either orally or through a nasogastric tube. One liter of solution should be given every 30–45 minutes until the intestinal effluent becomes clear. We prefer this to sodium phosphate because of the ability to titrate the amount of cleansing solution needed and because of the potential electrolytic effects of sodium phosphate, particularly in elderly patients with impaired renal blood flow [5]. Prokinetic agents, such as metaclopramide, may facilitate the administration of the PEG solution.

Vigorous bleeding from the upper gastrointestinal tract can also present as bright red blood, accounting for 10%–15% of cases of presumed ALGIB. If the patient’s history suggests a possible upper gastrointestinal source of bleeding (NSAID use, history of peptic ulcers) an upper gastrointestinal endoscopy should be performed prior to starting the colonoscopy preparation. Prompt placement of a nasogastric tube allows for sampling of the gastric contents for blood, clears the stomach for an upper endoscopy, if needed, and facilitates administration of the PEG solution in preparation for colonoscopy.

The colonoscopy should be performed with anticipation of the need for additional flushing to remove blood and retained clots, even after a vigorous purge. Careful evaluation of the entire colonic surface is critical to identify possible bleeding sources. Stigmata of bleeding, similar to those used for upper gastrointestinal bleeding (active bleeding, visible vessels), should be treated at the time of colonoscopy. The method of treatment should be individualized according to the type of lesion seen and the location in the colon. As the colon wall is thinner in the cecum and right colon, I prefer to use dilute epinephrine injection followed by careful thermal therapy after decompressing the colon to maximize wall thickness in the right colon. When argon plasma coagulation (APC) is used the electrical current is transmitted to the colon wall via the argon gas. This makes thermal therapy easier in some situations, but the depth of penetration of the current and the tissue damage is difficult to control. For this reason I avoid using APC in the right colon; I find it to be useful in treating larger areas of radiation damage in the rectum because of the thicker rectal wall. Bleeding diverticula can be treated with thermal coagulation, with or without epinephrine injection, if the vessel can be seen on the lip of the diverticulum. If blood is coming from the opening without a visible vessel, or if the diverticulum is occluded by a clot, the mouth of the opening can be injected with dilute epinephrine or saline to cause local tamponade and promote clot formation. It can be helpful to tattoo the area of the bleeding lesion for later location if bleeding persists or recurs. This can guide subsequent endoscopic or surgical therapy.

In the event that a specific bleeding site cannot be identified and treated at the time of colonoscopy, one can obtain information that may help to locate the general segment that is bleeding by noting the location(s) of fresh blood. If blood is present in the descending colon but not in the transverse or ascending colon, the bleeding site is probably distal to the splenic flexure. It is particularly helpful to determine if blood is present in the terminal ileum, which suggests a small-bowel source. This information may be very useful if bleeding persists and WEO “How I Do It” Approach to lower gastrointestinal bleeding

14

requires subsequent angiography or surgery.

When colonoscopy does not identify the site of bleeding, other diagnostic approaches should be considered. There is no role for barium studies in the evaluation of ALGIB. They are unlikely to identify a bleeding source and they interfere with more valuable studies, such as endoscopy, angiography, and computed tomography (CT).

If bleeding has ceased and the patient is hemodynamically stable, a more deliberate approach to diagnosis can be used, considering capsule endoscopy, enteroscopy, or CT, as outlined by Drs Sáenz and Valdivieso for intermittent and occult bleeding.

If the active bleeding persists, the approach depends upon the volume of bleeding. Scintigraphy using labeled red blood cells, can detect a slower bleeding rate (0.5 ml/min). Our experience with scintigraphy has been disappointing. Often the bleeding slows to the point where it cannot be detected, or the location of the bleeding site is poorly defined by the study. It may provide information that guides subsequent angiographic diagnosis and therapy, allowing the radiologist to more selectively evaluate the suspected bleeding vessel. For this reason radiologists may request a bleeding scan before performing angiography.

Angiography allows both diagnosis and potential therapy when brisk bleeding persists. Local infusion of vasopressin, or embolic therapy can slow, or stop bleeding, but also carry a significant risk of causing local ischemia. Vasopressin also carries cardiac effects that may limit its use.

CT angiography with high-speed, high-resolution scanners can identify both vascular abnormalities that may be bleeding and soft tissue abnormalities that can be diagnostic. When this technology is available it may obviate the need for angiography.

Surgical intervention should be a last resort for the treatment of continued vigorous ALGIB. Every effort should be made to identify the bleeding segment of bowel prior to surgery, as empiric hemicolectomy has a significant risk of missing the bleeding lesions. When colonoscopy can define an area of bleeding, but cannot identify the specific lesion or provide treatment, the information can be used to determine the section of bowel to be resected. A submucosal tattoo left in the segment may guide subsequent urgent surgery and minimize the amount of bowel that needs to be removed. Tattoos should be left at the site of endoscopic therapy to guide subsequent surgery in case bleeding persists or recurs.

Finally, when a lower gastrointestinal source of bleeding cannot be found, the upper gastrointestinal tract should be evaluated, as it accounts for 10%–15% of ALGIB. While the approach outlined here applies to most patients presenting with ALGIB, the individual condition, situation and co-morbidities of the patient should ultimately guide the diagnosis and treatment of a specific patient

WEO “How I Do It” Approach to lower gastrointestinal bleeding

15

References 1

Elta GH. Urgent colonoscopy for acute lower-GI bleeding. Gastrointest Endosc 2004; 59: 402–408

2

Jensen DM, Machicado GA. Diagnosis and treatment of sever hematochezia. The role of urgent colonoscopy after purge. Gastroenterology 1988; 95: 1569–1574

3

Jensen DM, Machicado GA, Jutabha R, Kovacs TO. Urgent colonscopy for the diagnosis and treatment of severe diverticular hemorrhage. N Engl J Med 2000; 342: 78–82

4

Strate LL, Syngal S. Timing of colonoscopy: impact on length of hospital stay in patients with acute lower gastrointestinal bleeding. Am J Gastroenterol 2003; 98: 317–322

5

Beloosesky Y, Grinblat J, Weiss A, et al. Electrolyte disorders following oral sodium phosphate administration for bowel cleansing in elderly patients. Arch Intern Med 2003; 163: 803–808

WEO “How I Do It” Approach to lower gastrointestinal bleeding

16

“HOW I DO IT” Approach to lower gastrointestinal bleeding Summary Tamás Molnár Tibor Wittmann

Drs Sáenz, Valdivieso, and Bjorkmann have excellently summarized the causes, diagnostic methods, and endoscopic therapies of lower gastrointestinal bleedings (LGIBs). We would like to add some further practical notes based on our experience.

1

Endoscopic examination of the upper gastrointestinal tract is the first step in our practice in the case of any manifestation of gastrointestinal bleeding; clear nasogastric tube aspirate is not enough to exclude upper gastrointestinal bleeding (UGIB), especially in the case of duodenal origin.

2

Examination of the ileum is very important during colonosopy indicated by gastrointestinal bleeding, particularly if hemorrhage is also present in the cecum. However in such cases, blood can flow back from the colon to the ileum through Bauhin’s valve.

3

The major diagnostic problem is caused by gastrointestinal bleedings originating in the small bowel. Although capsule endoscopy can determine the source of bleeding, it is very difficult to locate the exact segment and therapeutic intervention cannot be performed.

Therefore positive capsule endoscopy often results in surgery, during which intraoperative endoscopy is essential.

4

The experience gained from the increasing number of single- and double-ballon enteroscopic examinations allows a more frequent review of the whole length of the small bowel and the use of therapeutic hemostasis, which could solve the problem of the diagnosis and endoscopic therapy of acute small-bowel bleedings.

WEO “How I Do It” Approach to lower gastrointestinal bleeding

17

“HOW I DO IT” Removing large or sessile colonic polyps AUTHORSHIP

How I do it: Removing large or sessile colonic polyps Brian Saunders MD FRCP St Mark’s Academic Institute Harrow Middlesex UK

Comment Gregory G. Ginsberg, MD University of Pennsylvania Health Systems Philadelphia USA

Summary David J. Bjorkman, MD, MSPH Dean, University of Utah School of Medicine Salt Lake City Utah USA

“HOW I DO IT” Removing large or sessile colonic polyps How I Do It Brian Saunders

Introduction

Endoscopic mucosal resection (EMR) has become the standard technique for resection of large sessile and flat colorectal lesions. Its simplicity is the key. By working with the natural tissue planes of the colonic wall, surprisingly large lesions can be removed without the need for heavy sedation or inpatient stay. The submucosa is composed of loose areolar tissue which can be filled with fluid, “ballooning” the mucosa away from the underlying muscularis propria and making polypectomy inherently safer and easier.

The term EMR encompasses several techniques, from simple saline injection for snaring a small sessile polyp through to widespread piecemeal excision of hemicircumferential 10-cm lesions. Good EMR technique ensures high levels of safety and complete endoscopic excision, offering a powerful tool for cancer prevention. It represents a major step towards the evolution of “colonoscopic surgery”, the ultimate form of minimally invasive surgery – an operation from within.

Basic EMR technique for sessile polyps 1–2 cm in size, or for small flat adenomas smaller than 1 cm, should be within the armamentarium of all colonoscopists. However, effective endoscopic removal of large or complex lesions by EMR can only be achieved by appropriate referral to expert endoscopists skilled in the technique, and all too often patients with lesions that could be removed endoscopically undergo surgery because there is a lack of an appropriate referral pathway. Surgery carries a greater immediate patient risk and invariably results in a loss of intestinal length and function. Conversely, the use of poor endoscopic technique by inexperienced endoscopists may be equally harmful, risking incomplete removal or major endoscopic complication. An excellent way of learning both basic and advanced EMR techniques is by means of the various animal models which have gained widespread approval and should be part of all training programmes.

Approximately 3%–6% of colorectal adenomas detected at colonoscopy are large sessile polyps and up to 20% of all polyps are flat or minimally elevated. The detection of these lesions is likely to increase with the introduction of population screening for colorectal cancer (CRC). Thus a significant number of WEO “How I Do It” Removing large or sessile colonic polyps 2

lesions are potentially suitable for removal by either basic EMR at routine colonoscopy or by piecemeal excision by an expert colonoscopist at a specialist clinic (Figures 1 and 2).

Indications I would consider using an EMR technique for any sessile polyp larger than 1 cm in size, anywhere in the colon and for any polyp in the right colon that is larger than 5 mm. Utilizing this strategy I have never encountered a polypectomy-related perforation in more than 10 000 procedures to date, including more than 400 large sessile polyp resections. True “depressed” (IIc) lesions are rare in the colon but should always be removed by EMR (if possible) regardless of size as these lesions may contain high grade dysplasia and are difficult to ensnare without submucosal lifting. Sessile or flat lesions larger than 2 cm are usually removed piecemeal, although large lesions can now be removed en bloc with the new technique of endoscopic submucosal dissection (ESD).

ESD involves using a viscous injection solution for sustained submucosal lifting, a diathermy knife, and a plastic hood to help retract the polyp as it is dissected away from the muscularis propria. Although feasible anywhere in the colon, currently this technique is technically challenging and time consuming and carries a relatively high rate of major complication. Detailed description of ESD is beyond the remit of this paper but at present I would only consider this technique for large, flat or minimally elevated lesions in the rectum or distal sigmoid colon. In the future, and with improved accessories, ESD may become the preferred method of resection for all large benign lesions and very early submucosally invasive cancers, due to its inherent advantages of dissecting the deep submucosal layer to produce clear lateral and deep resection margins and a more accurate, “oncologically correct” specimen for histological assessment.

Contraindications to EMR There are very few. If a polyp is located in an area of the colon where access and visibility is restricted, for instance in the sigmoid colon in the presence of diverticular disease, then submucosal injection with “ballooning” of the mucosa towards the opposite bowel wall, can make polypectomy more difficult due to decreased endoscopic access and visibility.

EMR should not be attempted if the polyp fails to lift with adequate submucosal injection. This is the “non-lifting sign” and indicates malignant invasion deep into the submucosal layer. In this situation biopsies should be taken and tattoos placed around the lesion for surgical identification. Non-lifting does not always indicate a malignant process if there has been a previous polypectomy attempt. In this situation, diathermy injury has caused scarring to the submucosal layer and lifting will either not occur or will only be partial. Complete endoscopic removal of a polyp can still be achieved in these circumstances, but often only with a combination of conventional piecemeal snare excision and thermal WEO “How I Do It” Removing large or sessile colonic polyps 3

ablation, followed by tattooing of the site and close endoscopic follow up at 6–8 weeks.

Clinical scenario Polyps suitable for EMR may be detected during any colonoscopic examination. Generally speaking all polyps smaller than 2 cm should be removed at the time of a routine diagnostic examination. However larger or more complex lesions, if potentially suitable for piecemeal EMR, should be scheduled for a therapeutic clinic carried out by an expert colonoscopist familiar with all aspects of EMR. In my own practice I have two exclusively therapeutic clinics per week, lasting 3.5 hours, with only two or three patients scheduled. Senior nursing staff familiar with the EMR equipment are allocated to these sessions and there is provision for an overnight hospital stay for elderly patients or those with significant co-morbidity.

Consent, sedation, and patient information Fully informed consent for the procedure is obtained from the patient. My explanation includes the following features: •

The patient’s polyp needs to be removed because if left it is likely to turn into a cancer. This sounds obvious but sets the tone for a procedure which should not be taken lightly. I describe it as “internal surgery” to make the distinction from just another endoscopy.

•

I explain that EMR is a good alternative to conventional surgery for most people as it avoids the need for an anaesthetic, a prolonged hospital stay, abdominal wounds, and the risks of a surgical anastomosis. It also preserves intestinal length and long-term function.

•

It is important that the patient appreciates that a piecemeal EMR procedure carries more risks than a routine colonoscopy and polypectomy – particularly of bleeding (for up to 2 weeks after the procedure) and of perforation, both of which could result in the need for surgery and, rarely, surgery with a stoma.

•

The patient should be aware that although the EMR may be successful in removing the polyp locally, if subsequent histological examination shows microscopic cancer then surgery might still be recommended.

•

An early repeat colonoscopy is necessary 3 months after piecemeal EMR to check for complete healing and any residual polyp. Further check colonoscopies will also be advised at intervals determined by findings. So the patient is committing him- or herself to several procedures and bowel preparations (often the part of the examination that is most disagreeable to the patient).

With regard to the procedure itself:

WEO “How I Do It” Removing large or sessile colonic polyps 4

•

I always give the patient a choice of sedation. Most patients have light conscious sedation with small doses (1–3 mg of midazolam plus 25–50 mg of pethidine) whilst some prefer to have no sedation. Deep propofol sedation or anaesthesia is rarely necessary, apart from in patients who are very anxious. The EMR procedure is actually more difficult and hazardous with a patient who is unresponsive under propofol medication, as repositioning the patient is difficult and there is no feedback regarding pain (see later). I always explain this to patients who request anaesthesia.

•

It is explained to the patient that the procedure can sometimes take over an hour, with the need to change the patient’s position several times. I encourage patients to watch the procedure on the monitor; most are so fascinated that the time flies!

•

I emphasize that if the patient experiences any sharp pain during the procedure they should let me know immediately. (Serosal irritation and hence pain may occur before perforation thereby warning the endoscopist to desist.)

•

Finally I always leave some time for reflection and ask the patient if they have any questions about the procedure.

Patient preparation before the procedure All patients attending for an EMR procedure should undergo full oral bowel preparation, even if the lesion is in the rectum. A clean bowel facilitates visualization and assessment prior to EMR and reduces the risk of explosive gases in the bowel. I recommend the following bowel preparation: •

48-h fibre-restricted diet

•

oral senna 22.5 mg, at 1400 p.m. the day before the procedure

•

1.5 sachets of magnesium citrate powder (Citramag; Sanochemia Diagnostics Ltd Bristol, UK) made up to 1.5 L with water and drunk slowly between 1600 and 2000 p.m. the day before the procedure

•

0.5 L magnesium citrate, drunk between 0600 and 0700 a.m. on the day of the EMR if the procedure is scheduled for 0930 a.m.–1230 p.m.. or taken at 0930 for a 1330–1630 p.m. procedure

This preparation is generally well tolerated, effective, and low cost, compared with other products. It is important that some of the bowel preparation should be administered on the morning of the procedure to ensure good cleansing in the proximal colon.

Patients due to undergo a wide piecemeal EMR are told that they may require hospital admission after WEO“How I Do It” Removing large or sessile colonic polyps 5

the procedure, and are asked to bring an overnight bag and to arrange an escort home after the procedure.

Patients are instructed to stop taking warfarin (with or without heparin cover) clopidogrel and aspirin; consultation with the patient’s cardiologist may be necessary. Iron should be stopped for at least 7 days prior to the procedure. Antibiotic prophylaxis is given when indicated.

Equipment required I always use carbon dioxide as the insufflation gas as this is more comfortable for the patient than air, during and after a prolonged procedure. CO2 is absorbed through the colonic wall and excreted via the lungs, therefore there is no prolonged gaseous distension post procedure. I always use an antispasmodic in the form of hyoscine (10–30 mg) or glucagon (500 micrograms–1 mg). Reducing colonic spasm improves the stability of access onto the polyp.

Before commencing a wide endoscopic resection, I check that the following equipment is available: •

50-ml syringe filled with water for direct washing

•

20-ml syringe with simethicone bubble breaker solution

•

20-ml syringe with 0.2% indigo carmine for surface dye application

•

Olympus variable-stiffness colonoscope with Scopeguide imager capability

•

160-cm paediatric variable-stiffness colonoscope

•

Olympus gastroscope

•

Olympus twin-channel colonoscope (CF 2T-200)

•

ERBE diathermy Unit (Vio-300D) or Olympus PSD (PSD-30)

•

ERBE argon plasma coagulation (APC) equipment

•

Snare-master Olympus snares both mini (1 cm) and large (2.5 cm)

•

Wilson Cook 260-cm 25-G injection needle (LDVI-25)

•

Submucosal injection solution: 1/200,000 adrenaline plus a few drops of methylene blue (60 ml solution in total in 3 × 20-ml syringes). The methylene blue stains the submucosa and exquisitely defines the edge of the polyp whilst dilute adrenaline provides an invariably “dry” field in which to work. Some bleeding ooze is inevitable if saline is used, and although this is rarely significant it reduces visibility and can delay the procedure. Very rarely I will use hyaluronic acid if a more sustained lift is required. This is expensive, and a very cheap and effective alternative that I have employed is hypermellose 0.5% (as used for “artificial WEO “How I Do It” Removing large or sessile colonic polyps 6

tears” for dry eyes) drawn up into 2.5-ml syringes and injected through a 21-G (wide-bore) needle.

•

5 ml of 1/10,000 adrenaline made up to 50 ml with sterile water provides a useful topical adrenaline solution to prevent oozing from the submucosal surface after EMR. A dry field makes APC more effective at the end of the procedure.

•

Endoscopic clips: at least six Olympus “quick-clips” available

•

Roth retrieval net (2.5 cm diameter).

EMR procedure Assessment of the lesion First and foremost, it is important to carefully assess the lesion to be removed. I will only attempt to resect a lesion endoscopically if I judge that it can be completely removed in a single procedure. Partial removal of a polyp is usually a disaster, leading to inevitable polyp regrowth over a “fixed” scar that may then prove impossible for resection at a later date.

Additional washing and patient repositioning is necessary to optimize visibility and access to the lesion.

Ideally access should be obtained with the patient in such a position that luminal fluid can seen to be pooling away from the lesion. This facilitates piecemeal EMR, as resected pieces will then fall away from the operating field due to gravity.

A visual assessment of the size and extent of the lesion must be made, and a plan for resection formulated. For lesions that pass over a haustral fold, optimal access to the proximal side might only be achieved in retroflexion, which sometimes necessitates a change of instrument to a paediatric colonoscope or even a gastroscope.

Ulceration, marked friability or major fold deformity alert the endoscopist to the possibility of an advanced malignancy unsuitable for EMR. Gentle palpation with the biopsy forceps may also provide useful information regarding the degree of fixation and hard (malignant) or soft (benign) feel of the lesion. Ultimately the lifting characteristics of the lesion (see above) will determine whether endoscopic resection is feasible and safe.

There is a small literature on the use of pit patterns (with magnification) and of miniprobe high frequency ultrasound for assessment of lesions, but in my opinion these techniques add very little in practical terms over commonsense assessment of the lesion visually and of its lifting characteristics. They are also time-consuming and add to the expense and complexity of the procedure, so I no longer use them.

Resection of the lesion Lesions extending to two-thirds of the bowel wall circumference can be removed with piecemeal EMR; WEO “How I Do It” Removing large or sessile colonic polyps 7

however there is a risk of significant stenosis. Very large or circumferential lesions are often better removed with laparoscopic bowel resection.

To delineate the margins and surface of flat or very subtle lesions I use indigo carmine dye injected directly down the biopsy channel.

Having decided to go ahead with EMR, I pass the injection catheter down the instrument channel and target the proximal side of the lesion to lift it forward into the field of view. The exception to this is where an exophytic lesion fills the lumen, obscuring the view onto the proximal side, and where retroflexion is not possible. In this situation I would inject the distal (leading) edge of the lesion, aiming to debulk the polyp and in so doing obtain access onto the proximal side.

Hitting the submucosal space is crucial. I aim for the needle to enter the mucosa at an angle of about 45° rather than at 90°, thereby allowing more submucosa to be encountered, as the needle passes in a more tangential direction into the bowel wall. Therefore there is a greater chance of hitting the desired space. For this purpose also, I ask the assistant to start fluid injection, with constant pressure, whilst the needle tip is still in the bowel lumen before it enters the bowel wall. This ensures that as soon as the submucosal layer is entered, fluid will course into the submucosal space and “ballooning” of the mucosa will occur, creating several millimeters of space between the underlying muscle and the mucosa to be resected.

Often it is necessary to jab the needle slightly to penetrate the relatively tough mucosa. If an effective lift is not occurring, it is likely that the needle is too deep and must be withdrawn, and the process must be started again. Injection of a sterile solution into the peritoneal cavity does not appear to cause any harm, although the patient may experience a sharp pain as the serosa is breached, which disappears immediately once the needle is pulled back.

If the lesion is obviously benign I will inject through it; however if there is any doubt, the injection should be started through adjacent normal mucosa to avoid the rare event of tumour “seeding”.

For very large lesions, injection may have to be done in stages, with the polyp being removed in segments. It is a mistake however to try to snare an area of the polyp that has not been adequately lifted, as the diathermy effect will obscure the tissue plane and make complete excision more difficult. I sometimes use 60–80 ml of injection solution for a single large polyp and there is no limit to the volume that may be used. If the decision has been made to remove the polyp piecemeal, then I am willing to resect multiple smaller pieces 1–2 cm in size rather than try to remove the polyp in three or four very large pieces; in my opinion this approach is inherently safer and equally effective.

Before the snare excision is attempted, the scope is rotated so that the lesion is at the 6 o’clock position.

It is kept in this position by an endoscopy assistant, who holds the shaft of the colonoscope thereby leaving the endoscopist’s right hand free to manipulate the relevant accessories.

WEO “How I Do It” Removing large or sessile colonic polyps 8

The snaring of the first piece is crucial: if it is too small or superficial, the submucosal plane is not correctly defined, whilst snaring too large a piece risks complication. It is always gratifying to see the methylene blue-stained submucosa after snaring, which means that the correct tissue plane has been reached.

When snaring polyp pieces, I carefully place the fully opened snare over the area to be resected, push the snare downwards and then gently aspirate luminal gas as the snare is closed by a second endoscopy assistant. This relaxes the bowel wall and allows a reasonably sized piece of polyp to be ensnared without much slippage. All snares used for EMR have a mark (usually a pen line) made on the snare handle which corresponds to the point at which the snare wire is almost fully closed onto the outer plastic catheter sheath. If the endoscopy assistant can close the snare easily to this mark without undue resistance, it is very unlikely that muscularis propria is caught in the snare. Once the snare has been closed to the mark, it is handed to me and I perform the cutting with my right hand, whilst angling the scope tip up and away from the bowel wall (left-hand controls).

When taking a large piece of polyp I often slightly relax the tension on the snare, which allows any trapped muscle or deep submucosa to slip away; I then tighten again back down to the mark.

To cut through the mucosa I use short bursts of diathermy current, applying increasing tension on the snare to control the rate at which the tissue is resected. With this approach low power coagulating current can be used (25–30 W with the ERBE system, or 15 W with the Olympus PSD), providing effective haemostasis and cutting using snare tension. If the patient develops pain at any time during resection, I stop immediately and reposition the snare, checking all diathermy connections.

Once the mucosa has been resected, the polyp piece falls away from the field of view, revealing the blue-stained submucosa which should be visualized carefully, after gas insufflation, to check for bleeding or perforation. The next area of the polyp to be snared can then be targeted, and the process is repeated until the entire polyp has been resected. Generally the polypectomy progresses from one side across the bowel wall, but the endoscopist must to some extent be opportunistic, resecting more easily accessed areas of the polyp in order to make room to work in.

It is practically never necessary to increase power settings or use a cutting current, as the rate of cut is determined by the squeeze pressure on the snare handle which is closely controlled by the endoscopist.

I liken this to a surgeon with a scalpel: light pressure will cause some cutting but harder pressure causes a deep cut. Using this technique the endoscopist is in charge, like the surgeon, rather than relying on an “intelligent” (perhaps!) diathermy machine or a less experienced endoscopy assistant.

In order to grasp smaller fragments of polyp, the snare often has to be exchanged for a mini-snare.

Even for larger polyps, if working space is restricted then it may be easier to place a mini-snare than a larger snare. Snaring can take place in forward or retroflexed positions and the same process is applied.

WEO “How I Do It” Removing large or sessile colonic polyps 9

Rarely, a very flat polyp proves difficult to grasp with the snare despite good positioning and adequate aspiration of air during snare closure. In this case I change instrument to a double-channel scope and utilize a grasping forceps down one channel with a snare down the other. The snare is first backloaded onto the extended grasping forceps which is then used to grab the polyp and pull it back into the open snare which is then closed to capture the polyp. Great care is required with this technique to ensure that the bowel wall is not ensnared.

I do not use barbed or spiked snares as I find these reduce “feel” and hence control during final snare closure and mucosal cutting.

During piecemeal EMR, additional injection of fluid may be necessary; and frequent washing to maintain a view and aspiration of excess gas are also required.

Once the polyp has been completely removed, I like to use the argon beamer to “touch up” the polypectomy margins and destroy any remaining visible polyp fragments (Figure 3). Settings of 30– 40 W in the right colon and 50–65 W in the left colon are appropriate. Newer APC delivery devices (ERBE VIO) are much more energy-efficient, and lower power settings of 20–25W in the right colon and 35–45 W in the left colon should be used. Be careful not to use diathermy on exposed areas of muscle – you will see muscle contraction if this occurs – and remember to aspirate excess gas.

At the end of the procedure, I always look carefully for any bleeding points and photodocument the end result. I have to be satisfied that clear visualization of the polypectomy site shows no remaining polyp tissue. Tattooing the site may also be important, with at least two large India ink tattoos, placed just distally to the lesion. Finally I retrieve all the polyp pieces with a Roth retrieval net. During withdrawal this can be held away from the colonoscope tip to allow an adequate inspection of the mucosa distal to the polypectomy site.

Complications of EMR

Bleeding Oozing from the polypectomy base can be effectively treated by topical administration of an adrenaline wash. Focal bleeding sites should be clipped or treated with APC. Severe arterial bleeding is rare during colonic EMR, but if it occurs the first priority is to maintain a view of the mucosa by changing the patient’s position so that blood pools, with gravity, away from the polypectomy site. Endoclips or APC can then be applied, the combination often being more effective than a single modality.

Delayed bleeding can occur for up to 2 weeks post procedure. Patients usually require nothing more than close observation as bleeding will stop spontaneously in the vast majority of cases. However, emergency colonoscopy after a rapid oral purge may be necessary, and always with surgical back-up as WEO “How I Do It” Removing large or sessile colonic polyps 10

a last resort.

Perforation Microperforations, if detected, can be closed with endoclips although it is better to not perforate in the first place! Delayed perforation is an absolute indication for laparotomy; however it is important not to confuse free perforation with post-polypectomy syndrome. In the latter situation, serosal irritation from a full thickness burn has caused localized peritonism, low grade fever, and raised inflammatory markers, without free gas in the abdomen. Conservative management in this case is usually successful, with intravenous antibiotics, bed rest and close clinical assessment.

Post-resection surveillance Most patients, even after a large piecemeal EMR, can be discharged the same day provided they are pain-free and fully ambulatory. I advise patients to avoid aspirin and proprietary nonsteroidal antiinflammatory (NSAID) drugs for 2 weeks post procedure. I do not impose any dietary or lifestyle restrictions other than to suggest that they do not undertake prolonged air travel and remain within reasonable access of modern medical facilities for 2 weeks. At the time of discharge patients are given a copy of their colonoscopy report which also contains emergency contact telephone numbers. I always stress to patients that should they get severe pain, fever, or bleeding then they must inform us or their local clinician. Delayed bleeding is the main risk, and patients need to be aware that they may be well for 10 days but then still get significant bleeding.

Standard adenoma surveillance intervals apply when a small sessile polyp is removed en bloc by single-snare EMR. Surveillance intervals are determined in the usual way by the number and size of the adenomas removed. Piecemeal excision however necessitates an early repeat assessment at 3 months to check for healing and the presence of residual polyp. If there is any doubt about completeness of resection of a large lesion, the patient must return even earlier, at 2 months, before any chance of a large recurrence. Any small area of recurrence can usually then be definitively destroyed at the second procedure with a further repeat check arranged for 3 monthsâ&#x20AC;&#x2122; time.

Whenever I perform a check examination post-EMR, I use indigo carmine dye to help highlight the scar and surrounding mucosa. Occasionally tiny areas of recurrence are only visible after dye application (Figure 4). APC is invaluable for treating small areas of recurrence at a previous polypectomy scar.

WEO â&#x20AC;&#x153;How I Do Itâ&#x20AC;?

Removing large or sessile colonic polyps 11

Figures Figure 1 Endoscopic mucosal resection (EMR) of 6-cm sessile tubulovillous adenoma (3/4 complete and showing “clean” resection plane through the submucosa.

Figure 2 EMR of flat (IIb) adenoma containing tiny focus of superficially submucosally invasive cancer (clear resection margin), A=before, B=after EMR. A

B

WEO “How I Do It” Removing large or sessile colonic polyps 12

Figure 3 EMR polypectomy site after treatment of margins with argon plasma coagulation (APC).

Figure 4 Tiny areas of recurrent adenoma in the polypectomy scar at 3 months.

WEO â&#x20AC;&#x153;How I Do Itâ&#x20AC;?

Removing large or sessile colonic polyps 13

Recommended reading Waye JD. Advanced polypectomy. Gastrointest Endosc Clin N Am 2005 Oct;15: 733–756 Conio M, Ponchon T, Blanchi S, Filiberti R. Endoscopic mucosal resection. Am J Gastroenterol 2006; 101: 653–663; Epub 2006 Feb 8 Uraoka T, Saito Y, Matsuda T et al. Endoscopic indications for endoscopic mucosal resection of laterally spreading tumours in the colorectum. Gut 2006; 55: 1592–15927; Epub 2006 May 8 Bories E, Pesenti C, Monges G et al. Endoscopic mucosal resection for advanced sessile adenoma and early-stage colorectal carcinoma. Endoscopy 2006; 38: 231–235 Su MY, Hsu CM, Ho YP et al. Endoscopic mucosal resection for colonic non-polypoid neoplasms. Am J Gastroenterol 2005; 100: 2174–2179 Conio M, Repici A, Demarquay JF et al. EMR of large sessile colorectal polyps. Gastrointest Endosc 2004; 60: 234–241 Tamura S, Nakajo K, Yokoyama Y et al. Evaluation of endoscopic mucosal resection for laterally spreading rectal tumors. Endoscopy 2004; 36: 306–312 Kudo S, Tamegai Y, Yamano H et al. Endoscopic mucosal resection of the colon: the Japanese technique. Gastrointest Endosc Clin N Am 2001; 11: 519–535 Kudo S, Kashida H, Tamura T et al. Colonoscopic diagnosis and management of nonpolypoid early colorectal cancer. World J Surg 2000; 24: 1081–1090 Matsuda K, Masaki T, Abo Y et al. Rapid growth of residual colonic tumor after incomplete mucosal resection. J Gastroenterol 1999; 34: 260–263 Kudo S, Kashida H, Nakajima T et al. Endoscopic diagnosis and treatment of early colorectal cancer.

World J Surg 1997; 21: 694–701 Suzuki N, Saunders BP, Brown G. Flat colorectal neoplasms: endoscopic detection, clinical relevance and management. Tech Coloproctol 2004; 8 Suppl 2: s261–s266 Brooker JC, Saunders BP, Shah SG, Williams CB. Endoscopic resection of large sessile colonic polyps by specialist and non-specialist endoscopists. Br J Surg 2002; 89: 1020–1024 Brooker JC, Saunders BP, Shah SG et al. Treatment with argon plasma coagulation reduces recurrence after piecemeal resection of large sessile colonic polyps: a randomized trial and recommendations. Gastrointest Endosc 2002; 55: 371–375

WEO “How I Do It” Removing large or sessile colonic polyps 14

“HOW I DO IT”

Removing large or sessile colonic polyps Comment Gregory G. Ginsberg

Large, flat, and otherwise laterally spreading adenomas make up the class of “defiant” polyps, i.e. those colonic polyps not resectable using a standard snare polypectomy technique. As Dr. Saunders delineates, however, the majority of these polyps may be resected with curative intent by expert endoscopists using adjunctive techniques. Dr. Saunders is a master colonoscopist and so I was gratified in reviewing his submission to find that we are largely in agreement with regard to strategies and techniques for the endoluminal resection of large sessile colonic polyps. Where the expertise is available, patients with large sessile colonic polyps should be afforded the opportunity for colonoscopic eradication in lieu of operative resection. I prefer the term “endoluminal resection” (ELR) over “endoscopic mucosal resection” because, in fact, the submucosa is resected as well.

Indications I incorporate submucosal injection to facilitate snare polypectomy for flat, sessile, and broad-based polyps greater than 1–2 cm in diameter, throughout the colon. The majority of large sessile colonic polyps that I encounter have been referred by another gastroenterologist or colorectal surgeon and so the expectations on the patient and physician sides are uniquely delineated. For the incidental large sessile polyp, I believe that it is within the scope of common practice to employ these techniques for resection of sessile polyps of up to about 2.5 cm, at the time of the index colonoscopy. However, for larger lesions, consideration should be given to photographic and written documentation of the lesion(s) including number, size, location, and configuration. Cold forceps biopsy should be performed for histological sampling. As Dr. Saunders implies, the use of electrosurgical energy should be avoided outside of a commitment to complete curative resection. Partial or incomplete thermal snare resection or “biopsy” results in a fibroinflammatory reparative response that tacks down the remnant margin of the lesion. This results in the “pseudo-non-lifting sign” and makes subsequent completion of resection more hazardous and more difficult. Tattooing to mark the lesion for subsequent recognition is usually welcomed. However, some tattooing agents (e.g. India ink) if placed too near the lesion may also promote a local fibroinflammatory response compromising subsequent ELR and so inert agents placed apart from the lesion are preferred.

Unlike Dr. Saunders, I have encountered polypectomy-related perforations, with two in approximately 500 large sessile polyp resections. These both occurred with moderate-size (~2 cm) sessile proximal WEO “How I Do It” Removing large or sessile colonic polyps 15

ascending colon polyps, during attempts to resect them en bloc rather than in a piecemeal fashion. I concur with Dr. Saunders that, short of applying ESD techniques, piecemeal resection is preferred for larger sessile lesions.

Contraindications Colonoscopic ELR should not be attempted when morphologic and/or tactile features indicate invasive carcinoma. These include ulceration, firm texture, and fixation. We routinely apply endoscopic ultrasound when considering rectal lesions for ELR and demur when there is evidence of invasion into the submucosa or beyond. Endoscopic ultrasound is not used in the evaluation of lesions proximal to the rectum. Except when there has been prior application of electrosurgical energy, the non-lifting sign is a contraindication to proceeding with attempted ELR. Beyond the rectum, sessile lesions that extend beyond more than 30% to 50% of the luminal circumference and those that extend beyond 7 cm in length should be considered for operative resection. Similarly lesions that are present within the appendiceal orifice and the iliocecal valve typically defy completion endoscopic resection.

Clinical scenario, consent, and sedation As stated above and in agreement with Dr. Saunders, large polyps potentially suitable for colonoscopic ELR should be biopsied and the patient rescheduled for dedicated ELR after a discussion of the options for management. Ideally, these procedures should be booked for a 30- to 60-minute block.

In addition to the points that Dr. Saunders emphasizes to patients, I quote an up to 20% risk of postpolypectomy bleeding that may be acute or delayed, and that delayed bleeding may occur anytime from 12 hours to 12 days after the procedure.

It is our standard practice to perform colonoscopic ELR as an outpatient procedure, using narcotic and benzodiazepine sedation. I advise that acute bleeding is treated endoscopically and may prompt a recommendation for overnight hospital observation. Most delayed bleeding is short-lived, self-limited, and not hemodynamically destabilizing. Patients with evidence of delayed bleeding are advised to report to their local hospital emergency department for evaluation and initial management. For the rare patient who requires directed therapy, we prefer to arrange transfer to our center, if feasible.

Patient preparation before the procedure I similarly recommend a full oral bowel preparation to best ensure adequate visualization of the lesion for resection. I simply employ a clear liquid diet on the day prior to the procedure and a 4-L polyethylene glycol oral purgative. I no longer adjust patientsâ&#x20AC;&#x2122; aspirin or other antiplatelet medications. Warfarin WEO â&#x20AC;&#x153;How I Do Itâ&#x20AC;?

Removing large or sessile colonic polyps 16

management is individualized.

Equipment We use a variable-stiffness, pediatric caliber colonoscope as our standard instrument, along with air insufflation. For distal colon and rectal lesions I use a therapeutic channel upper endoscope.

Antispasmodic agents are only rarely used. In addition to our standard electrosurgical generator, an argon plasma beam-capable system is also kept in readiness. Our assistants prepare: four 10-mL syringes with methylene blue-tinted normal saline solution, an injection needle, two standard polypectomy snares, a mini-snare, and a specimen retrieval net. For rectal lesions, the Olympus EMR kit is available in order to use the transparent aspiration cap and crescent snare, plus additional crescent snares (as these tend to deform after a single use). This latter approach is useful for very distal lesions that approach the anal verge and for those lesions displaying the pseudo-non-lifting sign. Acute bleeding is similarly treated with local injection of diluted epinephrine solution and clips are applied when this is insufficient.

EMR procedure I concur with Dr. Saunders’ detailed description of EMR techniques. Key to assessment and planning are satisfactory visualization and favorable orientation of the lesion. To achieve this, it is worth the expenditure of time and effort to reposition the patient, axially rotate the colonoscope, and/or visualize the lesion from the retroflexed position. I, too, preferentially inject around the base of the lesion, rather than directly into it when circumstances permit.

In contrast to Dr. Saunders’ standard technique, a trained gastrointestinal technician operates the opening and closing of the snare, while the endoscopist activates the electrosurgical current with a foot pedal and applies traction with a to-and-fro motion on the shaft of the snare. Training technical assistants for EMR requires interest, dedication, patience, and acceptance of an incremental approach.

We develop EMR specialists in that same way ERCP technicians were developed in the past. Like Dr. Saunders, we rely generally on a coagulation current from a standard electrosurgical generator.

Post-resection surveillance We have discontinued routine restrictions on aspirin and nonsteroidal anti-inflammatory drug (NSAID) use related to endoscopic procedures, in accordance with American Society for Gastrointestinal Endoscopy (ASGE) guidelines that indicate insufficient evidence of increased bleeding risk. However, these recommendations may be individualized. Patients are allowed to resume their diets.

When incomplete resection is suspected, 6 weeks is the minimum duration before follow-up WEO “How I Do It” Removing large or sessile colonic polyps 17

examination. This is sufficient for healing of the resection site in the vast majority of cases. In a patient returning sooner than this, the endoscopist is likely to encounter hyperplastic reparative changes that are indistinguishable in appearance from residual adenoma. Otherwise, all rectal lesions and colon lesions with high grade dysplasia (HGD) or ImCa dictate a 6-month follow-up and all others are followed up at 1 year. Subsequent surveillance follows established guidelines. Our experience indicates that increased lesion size and piecemeal resection are predictors of residual/recurrent adenoma. Given compliance with surveillance, residual/recurrent lesions are identified and effectively eradicated endoscopically.

WEO â&#x20AC;&#x153;How I Do Itâ&#x20AC;?

Removing large or sessile colonic polyps 18

“HOW I DO IT”

Removing large or sessile colonic polyps Summary David J. Bjorkman

Advances in endoscopic therapy have expanded the spectrum of lesions that can be excised without resorting to surgery, perhaps most commonly in the setting of neoplastic lesions of the colon. While it has been standard therapeutic practice to remove small and pedunculated polyps when they are discovered, large and sessile lesions have often been referred for surgical resection.

These two excellent and detailed summaries demonstrate that, with appropriate care and caution, many lesions previously referred for surgery can be resected endoscopically. The techniques described here are almost identical. The basic principle is to raise the lesion by injecting fluid into the submucosal space, then use electrocautery to snare portions of the lesion until it has been completely excised.

There are a few points made by both Dr. Saunders and Dr. Ginsberg that deserve emphasis. First, patient selection and preparation are critical. This approach has a higher risk of complications and patients must be willing to accept these. Second, the bowel must be thoroughly prepared to ensure that there is a clean field for the resection. A third important point is the extensive list of specialized endoscopes, accessories, medications/solutions, and trained support personnel outlined by both authors. This is not a procedure that can be performed with limited resources. One must be fully prepared to deal with any situation that may arise during the procedure. Finally, after beginning the excision, the endoscopist is committed to completing it, regardless of how long it may take. One cannot perform EMR in multiple procedures. Healing and fibrosis from a first partial excision would cause scarring that would prevent the required lifting of the lesion at subsequent procedures.

Both authors have given us very detailed instructions on how this procedure can be performed, but have also cautioned us that the endoscopist should know the limits of his or her resources and experience. When in doubt, these lesions should be referred to experienced endoscopists, such as Drs. Saunders and Ginsberg. The principles they have taught us can also be applied to our therapy of smaller sessile lesions and can improve our own endoscopic skills.

WEO “How I Do It” Removing large or sessile colonic polyps 19