Obesity is a worldwide epidemic, affecting more than 300

Transcription

1 CLINICAL GASTROENTEROLOGY AND HEPATOLOGY 2013;11: Endoscopic Management of Complications After Gastrointestinal Weight Loss Surgery NITIN KUMAR and CHRISTOPHER C. THOMPSON Division of Gastroenterology, Brigham and Women s Hospital, Boston, Massachusetts As more patients undergo bariatric surgery, gastroenterologists will increasingly encounter variant postsurgical anatomies and postoperative complications. We discuss the diagnosis and management of bleeding, ulcers, foreign bodies, stenoses, leaks, fistulas, pancreaticobiliary diseases, weight regain, and dilated outlets. Keywords: Weight Loss; Gastric Bypass; NSAIDs; Endoscopic Ultrasound. Obesity is a worldwide epidemic, affecting more than 300 million people and approximately 1 in 3 American adults. The prevalence of metabolic syndrome and diabetes in this population is 39.2% and 14.2%, respectively. 1 Lifestyle modification and pharmacologic therapy for weight reduction offer limited potential for consequential and sustained weight loss. Gastrointestinal weight loss surgery, in contrast, has been shown to be effective and is finding increased use. 2 An estimated 220,000 bariatric surgeries were performed in the United States and Canada during Laparoscopic or open Rouxen-Y gastric bypass (RYGB), laparoscopic adjustable gastric band (LAGB), vertical banded gastroplasty (VBG), sleeve gastrectomy (SG), biliopancreatic diversion (BPD), and duodenal switch (DS) comprise the vast majority of bariatric surgeries. 4 Of these, a meta-analysis of 136 studies found that RYGB is the most prevalent. Meta-analysis revealed that RYGB resulted in average percent excess weight loss of 56.7% 66.5% during 24 months after surgery. Diabetes resolved or improved in 86%, hypertension in 68%, obstructive sleep apnea in 81%, and hyperlipidemia in 97%; RYGB has also been found to have survival benefits. 2 Evaluation of the Symptomatic Patient Symptoms such as nausea, vomiting, and abdominal pain are common in patients after bariatric surgery. Accurate diagnosis requires knowledge of which surgery the patient underwent and when it was performed, whether subsequent revisions were performed, and whether the patient has habits or risk factors that predispose to postsurgical complications. Common presenting symptoms are not usually sufficient for diagnosis in this patient population. Endoscopic evaluation is warranted in symptomatic patients. We will describe the incidence, diagnosis, and management of bleeding, ulcer, foreign body complications, stenosis, fistula and leak, pancreaticobiliary disease, and weight regain in the patient after bariatric surgery. Upper Gastrointestinal Bleeding Upper gastrointestinal bleeding after bariatric surgery occurs more commonly after RYGB (1.9% of cases) than LAGB, SG, and VBG Bleeding can occur at multiple sites, including the pouch, the anastomoses, staple lines, the contiguous small intestine, the excluded stomach, or the bypassed small intestine. In addition, patients who have had LAGB can develop esophagitis because of altered anatomy. Laparoscopic RYGB is associated with a higher bleeding rate than open RYGB. 9 Early bleeding usually occurs within 24 hours postoperatively at the staple lines of the gastrojejunal anastomosis (GJA), gastric remnant, or jejunojejunal anastomosis. 10 A significant proportion of early bleeding is extraluminal; patients may develop hemodynamic instability, oliguria, and abdominal distention. 9 Late bleeding is often a result of anastomotic ulceration. Diagnosis The esophagus, pouch, and GJA are easily accessible with a standard gastroscope. Bleeding in the excluded stomach or at the jejunojejunal anastomosis presents a more significant challenge. In these cases, device-assisted enteroscopy may be required. In cases of early bleeding, endoscopy carries a risk of causing perforation of immature anastomoses and staple lines. 11 This risk is higher when forces are applied to the bowel during push enteroscopy or device-assisted enteroscopy. 12 If endoscopy is performed, air insufflation should be minimized, and carbon dioxide insufflation should be used. Management Use of endoclips is suggested to minimize tissue injury at the site of application and should be used in conjunction with epinephrine injection because dual therapy is preferred. Electrocautery should be avoided at fresh staple lines. Hemostatic powders may be another option in the near future, but there are currently no data on their use after bariatric surgery. Angiographic intervention can be considered; however, the resulting ischemia is a concern in patients with new anastomoses. Abbreviations used in this paper: APC, argon plasma coagulation; BPD, biliopancreatic diversion; DS, duodenal switch; ERCP, endoscopic retrograde cholangiopancreatography; GJA, gastrojejunal anastomosis; LAGB, laparoscopic adjustable gastric band; NSAID, nonsteroidal anti-inflammatory drug; RYGB, Roux-en-Y gastric bypass; SEMS, selfexpanding metal stents; SEPS, self-expanding plastic stents; SG, sleeve gastrectomy; VBG, vertical banded gastroplasty by the AGA Institute /$

2 344 KUMAR AND THOMPSON CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 11, No. 4 Figure 1. (A) Suture can entrap food, which can apply tension, causing intermittent pain, (B) linear ulceration, and (C) bezoar formation resulting in bowel obstruction; (D) suture should be extracted and cut with endoscopic tools. Ulcer Ulceration is the most common endoscopic finding in patients with abdominal pain. 13 Ulceration at the GJA is a common late complication after RYGB, occurring in approximately 20% of patients. 14 These most commonly develop in the first 3 months postoperatively but can occur at any time. Patients often present with epigastric pain, nausea, vomiting, food intolerance, and overt or occult bleeding. Ulcers can present on the gastric side or the jejunal side of the anastomosis. Ulcers on the jejunal side may be a result of ischemia, overproduction of acid, or coexisting gastrogastric fistula. Anastomotic ulcers may be a result of bile acid reflux, pouch orientation and size, tension on the Roux limb, Helicobacter pylori infection, nonsteroidal anti-inflammatory drugs (NSAIDs), alcohol, smoking, or foreign body such as nonabsorbable sutures Gastrogastric fistula and staple line disruption may result in ulceration caused by acid exposure. The jejunum is particularly vulnerable because the ph buffering function of the duodenum has been lost. Preoperative H pylori infection is associated with postoperative anastomotic ulceration; it is possible that preexisting mucosal damage may result in ulceration even after eradication. 18 An analysis for risk factors found that predictors of anastomotic ulcer include smoking (adjusted odds ratio, 30.6) and NSAID use (adjusted odds ratio, 11.5). 19 A multivariable analysis found that diabetes (odds ratio, 5.6) was also associated. 15 Surgical risk factors include use of a circular stapler technique rather than a hand-sewn or linear stapler. 20,21 Chronic irritation by staples or sutures may also play a role; absorbable sutures are associated with a significantly lower rate of ulceration. 22 Diagnosis Investigation for anastomotic ulcer in the first 2 weeks postoperatively may be performed with Gastrografin to avoid stomal disruption during endoscopy; however, endoscopy can be safely performed. Endoscopic visualization should include the gastric pouch, GJA, and proximal Roux limb. Size, depth, and potential etiologic factors should be noted for each ulcer. Aspiration of the fluid in the gastric pouch should be performed before entering the jejunum; if the ph is neutral or alkaline, sucralfate may be used, but acid suppression may not be helpful. 23 H pylori breath testing and pouch biopsies may not be reliable; serology may be better to detect infection and fecal antigen to confirm eradication. 24,25 Management Treatment of ulceration should be directed by suspected etiology. In patients with RYGB, anastomotic ulcers should be treated with soluble proton pump inhibitors or capsules broken open taken twice daily and tapered over 6 months. Sucralfate solution at 1g4times daily should be used concurrently when possible. 18 The tablet form is not effective. Bile reflux can be treated with bile acid binders such as cholestyramine or colestipol. Smoking cessation is critical. Control of diabetes should be optimized. NSAIDs should be discontinued if possible or combined with proton pump inhibitors or prostaglandin E 1 therapy if needed long-term. Although ulcers in patients with high Roux limb tension usually resolve spontaneously, those associated with large pouch warrant acid suppression. Gastrogastric fistulas should be closed when present, and patients should be maintained on proton pump inhibitors to decrease acid exposure. 26 Visible nonabsorbable sutures should be extracted when possible. 27,28 Patients with LAGB may benefit from partial deflation; if the band erodes, it must be removed. Healing should be assessed with repeat endoscopy. Persistent ulcers should prompt a search for gastrogastric fistula with upper gastrointestinal series because this may detect a fistula not seen on prior endoscopy. 14 Anastomotic ulceration may result in stricture formation at the GJA. Perforation is rare and often occurs in patients with active tobacco, NSAID, or steroid use. 29 Persistent deep ulcers may require surgical revision to prevent these complications. Foreign Body Complications Foreign material, such as sutures, staples, and bands, is often inserted during bariatric surgery. Embedded sutures and the resulting inflammatory response may result in pain, ulceration, bezoar formation, and obstruction (Figure 1). If a large amount of suture is seen, even in asymptomatic patients, it should be removed. A visible suture may be overlying an extensive embedded suture (Figure 2). Implanted foreign bodies, such as bands or mesh, can also erode or migrate. Management Patients with chronic pain after bariatric surgery should undergo endoscopic examination with removal of visible retained foreign material. Foreign material has been associated with pain, even when there is no adjacent visible inflammation. Traction on sutures or staples often reproduces pain. Ryou et al 28 demonstrated immediate symptom improvement in 71% of patients after foreign body removal, and Yu et al 30 reported resolution of symptoms in 69% and improvement in 14%. Patients with LAGB are subject to high rates of revision because many complications require band removal. 31 The most common complication is gastric prolapse through the band,

3 April 2013 ENDOSCOPIC MANAGEMENT OF BARIATRIC PATIENTS 345 Figure 2. (A) Visible suture may overlay significant embedded suture; (B) removal can result in improvement of symptoms. Figure 3. LAGB erosion. which presents with nausea, vomiting, food intolerance, and epigastric pain. Pouch dilation, gastric obstruction, and even gastric necrosis can result. This was more common with the perigastric technique (with incidence of 13.3%) than the currently recommended pars flaccida technique (with incidence of 1.8%). 32,33 Diagnosis can be made with endoscopy or imaging. Surgery is typically required for management; however, endoscopic removal has been reported. Erosion of the band through the gastric wall into the lumen is thought to occur as a result of tight placement or inclusion of excess tissue within the band (Figure 3). It has been reported in 3.8% of patients by using the perigastric technique, although it is much less frequent when the pars flaccida technique is used for placement. Symptoms include pain, vomiting, bleeding, and abscess. The initial symptom is often extension of infection along the connecting tube to the access port site, which may manifest as cellulitis. 34 Patients with VBG encounter band erosion in approximately 1% 3% of cases, but higher rates have been reported. 35,36 Symptoms include pain, loss of satiety, and subsequent weight regain. Endoscopic removal of eroded or partially embedded bands can be performed after transection. Several methods have been reported. Computed tomography imaging may be helpful before attempted endoscopic removal to confirm that the band is encapsulated. 37 Band cutting devices made of metallic wire have been used to transect the intragastric portion of migrated bands, which were then removed transorally. 38 A band migrated into the duodenum has been extracted after passing a Jag wire through it, transecting the band with a mechanical lithotripter, and then removing it transorally. 39 The port is removed externally in this hybrid endoscopic-laparoscopic procedure. Endoscopic removal should only be attempted if the band buckle is visible. Failures have been noted in patients in whom the band migrated into the jejunum, patients with port-site infection, and patients with bands stuck in the gastric wall. A partially eroded VBG band has been transected with endoscopic scissors. 40 Partially eroded Marlex mesh in VBG patients has been transected by using argon plasma coagulation (APC) at 1 L per minute and 80 W; the fragments were removed transorally. 41 The last 2 approaches would not be a viable approach for LAGB. Self-expanding stents have been used to facilitate endoscopic extraction of dysfunctioning LAGB and VBG bands. Temporary self-expanding plastic stents (SEPS) are placed to induce full intragastric band migration, and the stents and bands are then removed together after 6 8 weeks. The Deviere and Eubanks groups demonstrated removal of VBG rings. 42,43 The Deviere group subsequently demonstrated LAGB and VBG band removal in patients with pouch stenosis and dilation with or without intraluminal band erosion, with success in 12 of 13 cases and no major complications. 44 Stenosis Stenosis is a frequent complication after bariatric surgery. It can occur as a result of malfunction of prosthetic devices or as a result of stricture formation. Patients present with early satiety, nausea, vomiting, and dysphagia. 14 Postprandial retrosternal or abdominal pain may also occur. In patients with a history of RYGB, stenosis often occurs at the GJA; rates of 3% 5% have been reported after open RYGB and 5% 12% after laparoscopic RYGB Most stenosis at the GJA develops between 4 and 10 weeks after surgery. 49 In cases of stricture formation incited by ulcer or foreign material, presentation may be delayed for months or years. Medical factors predisposing patients to stricture formation include use of NSAIDs, smoking, and alcohol. 50 Surgical factors include method for anastomosis construction and mechanical tension or ischemia at the anastomosis. Gonzalez et al 21 reported a higher rate of stricture formation after using a circular stapler (31%) than a hand-sewn anastomosis (3%) or linear stapler (0%); the circular stapler is the most common technique currently used. 51 Twenty-five millimeter circular staplers have been shown to result in a lower stricture rate than 21-mm staplers with similar postoperative weight loss at 2 years. 49,52 Less common sites for stenosis after RYGB include the jejunojejunal anasto-

4 346 KUMAR AND THOMPSON CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 11, No. 4 mosis, sites of intestinal adhesions, and sites of passage through the mesocolon. Endoscopy is an excellent diagnostic tool because inciting pathology, such as foreign body, can be identified. Therapy can often be applied concurrently. Stomal stenosis is present if the standard 9.5-mm endoscope cannot traverse the anastomosis. Treatment of stomal stenosis can be performed with a throughthe-scope balloon, Savary dilator, or electrosurgical incision Balloon dilation is the most commonly used technique and is successful in more than 90% of cases. Some patients require 2 or 3 procedures, which can begin approximately 4 weeks postoperatively and be repeated every 2 3 weeks. 56 The balloon catheter should be advanced beyond the GJA with care to avoid entry into the blind limb; a guidewire and/or fluoroscopic visualization can be used when visualization of catheter advancement is suboptimal or resistance is encountered. Once the balloon is fully outside the endoscope, it can be inflated so that its midpoint applies radial pressure onto the stricture for at least 60 seconds or until waist dissipation on fluoroscopy. Dilation to 15 mm has been shown to be safe, even at the first procedure, and 20 mm has been reported as successful. 57 However, the smallest effective dilation is preferred. Perforation of the GJA or Roux limb is a risk. A gradual approach to dilation over several sessions can reduce perforation risk (reported to be 3% 5%) and decrease the possibility of overdilation and resultant weight regain. 54 Suture material at the GJA may need to be removed to achieve successful dilation. 58 A Silastic band or Mylar mesh may be present; these can make endoscopic dilation more challenging, if not impossible. Other types of bariatric surgery may be complicated by stenosis. Patients with a history of LAGB may have obstruction as a result of edema or excess tissue at the level of the band. 42 There may be a fibrous reaction to the band; in these cases, endoscopic dilation can be attempted if stenosis persists despite full band drainage. There may also be rotation away from the horizontal axis or adhesions causing pouch angulation. Band removal, band replacement, or conversion to RYGB may be considered. 55 If there is concern for gastric necrosis, surgical revision may be needed. 42 Patients with a history of SG may have stenosis at the gastroesophageal junction or at the incisura angularis. 59 Serial endoscopic balloon dilation up to 20 mm can be attempted, as well as temporary metal or plastic stent placement for up to 8 weeks. 60 Ultimately, conversion to RYGB may be required. Patients with VBG have a high rate of stomal stenosis, reaching 20% 33%. Endoscopic dilation, often requiring multiple sessions, has been reported successful in 68% of patients. Band migration may also occur, requiring removal. 35 Leaks and Fistulas Fistula is an abnormal communication originating in a visceral structure. Leak, one type of fistula, is defined as discontinuity of tissue apposition in the immediate postoperative period. The incidence of leak after bariatric surgery ranges from 1.7% 2.6% after open RYGB to 2.1% 5.2% after laparoscopic RYGB and reaches up to 5.1% after SG. 61 Improvements have been made over time with better surgical technique. 62,63 Leaks develop when intraluminal pressure exceeds tissue or suture line resistance. Leaks presenting within the first 2 days postoperatively are often mechanical, whereas ischemic leaks usually present at 5 6 days postoperatively. 64 Leaks are thought to be caused by tension on the anastomosis, suture or staple seepage, staple or stapler malfunction, tissue ischemia, distal obstruction, and hematoma. 65 Risk of leak may be increased in patients with impaired healing, infection, diabetes, hypertension, sleep apnea, age 55 years, male gender, and prior surgery After RYGB, leaks can occur at several sites; the most common sites are the GJA (68%), gastric pouch staple lines (10%), and jejunojejunal anastomosis (5%); an additional 14% involve multiple sites. 66 Other sites should be considered as well: the jejunal stump, excluded stomach, duodenal stump (in resectional bypass), and blind jejunal limb. 68 Some may be especially challenging to localize, such as leaks from the excluded stomach, because routine endoscopy and upper gastrointestinal series may be normal. Incidence of leak is highest in patients with divided RYGB. The risk of chronic gastrogastric fistula is highest when the pouch and excluded stomach are contiguous, as with the open surgical approach. 69 In patients with SG, most leaks occur in the proximal third of the stomach near the gastroesophageal junction (85.7%), with the remainder in the distal third. 70 It has been shown that fistula is associated with elevated intraluminal pressure and less sleeve distensibility; distal stenosis and delayed gastric emptying may play a role in fistula formation. 71 Leaks are associated with a mortality rate of 6% 14.7%. 72,73 One study found that mortality was 9% after leak at the GJA but 40% after leak at the jejunojejunal anastomosis. 72 Mortality is higher in patients with leak after open RYGB than laparoscopic RYGB. Other than pulmonary embolism, they are the most serious life-threatening complication after bariatric surgery. In addition to doubling the risk of mortality, leaks result in a 6-fold increase in hospital stay. 74 Patients who develop a leak are at increased risk for wound infection, sepsis, respiratory failure, renal failure, thromboembolism, internal hernia, and small bowel obstruction. Diagnosis Leaks often present without fever, leukocytosis, or pain. The most common reported symptom of leak is tachycardia, present in 72% 92% of patients. 73,75 Nevertheless, studies have reported nausea and vomiting (81%), fever (62%), and leukocytosis (48%), and one must have a high suspicion for leak in patients with any of these symptoms. Objective signs include increased drain output, as well as C-reactive protein elevated to greater than 22.9 mg/dl 2 days after surgery (sensitivity, 1.00). 76 This differs from chronic gastrogastric fistulas, which have a more indolent course and typically manifest with acid reflux, abdominal discomfort, and weight regain. Because gastric acid can flow into the pouch via gastrogastric fistula, patients with heartburn, acid reflux, or anastomotic ulcer should be evaluated for fistula. Diagnosis of leak can be made radiologically or endoscopically. Upper gastrointestinal series with water-soluble contrast and computed tomography scan have limited sensitivity because of body habitus but high positive predictive value Because many leaks form after imaging, series have reported a low rate of leak detection when upper gastrointestinal series are used routinely. 81 When leak is suspected clinically, sensitivity has been reported at up to 92%. Endoscopic diagnosis involves careful examination of the esophagus, stomach, or gastric pouch (including retroflexion) after suction of all fluid present and small intestine. Likely the most sensitive means of diagnosis is endoscopic examination under fluoroscopy by

5 April 2013 ENDOSCOPIC MANAGEMENT OF BARIATRIC PATIENTS 347 using a combination of maneuvers, including a bubble test (submerging the drain while performing endoscopic insufflation; bubbles indicate presence of a leak), and injection of contrast with methylene blue into an abdominal drain while looking endoscopically and fluoroscopically for evidence of leak. Management Management of patients with postoperative leak can be challenging. Local and systemic complications such as sepsis, multiorgan failure, abscess, and nutritional deficiency can present. Surgical management is associated with high morbidity (up to 50%) and mortality (2% 10%). 75,82 There is a high conversion rate to open surgery (48%), as well as long operative time and increased intraoperative blood loss As a result, initial management has moved toward conservative or endoscopic treatment. Conservative management begins with supportive care, suspension of oral intake, and institution of parenteral or distal enteral feeding, broad-spectrum antibiotics, and percutaneous drainage of collections. A study of nonoperative management of contained collections showed leak resolution in 33 of 40 cases at a median 17 days. 86 As endoscopic management techniques gain acceptance, they are being used earlier in the postoperative course. Dilation of distal stenoses should be performed. Exclusion techniques such as stent placement can occlude or bypass leaks. Leaks and fistulas can be closed with clips, suturing devices, or sealants. Finally, vacuum-assisted drainage has been investigated. These techniques have been used in combination for leaks and other types of fistulas. Each will be discussed in turn. Stent placement for exclusion of the leak from the gastrointestinal tract is the endoscopic technique supported by the most substantial body of evidence. Stent placement allows the leak to heal while enteral nutrition is resumed, potentially accelerating recovery and avoiding the need for parenteral nutrition. Peritoneal contamination is decreased, and improvement in abdominal pain may follow. 61 Both covered selfexpanding metal stents (SEMS) and SEPS have been used successfully. Stent placement is performed by using a forwardviewing endoscope and fluoroscopic guidance. The endoscope is advanced well into the Roux limb (in post-rygb patients) or to the third portion of the duodenum (in post-sg patients), and a guidewire is introduced. The leak site and squamocolumnar junction are marked with placement of an external marker under fluoroscopy, and the endoscope is removed. The stent delivery system is inserted under fluoroscopy (by using a stiff guidewire if acute angulation must be traversed). SEPS delivery systems are larger and more rigid, making deployment at acutely angulated sites more difficult. The stent should be deployed distally enough from the upper esophageal sphincter to avoid globus sensation. The distal end of the stent should not impact the enteral wall because bleeding, ulceration, or perforation may result. Stent deployment may begin proximally or distally, depending on the system. Adjustment of stent position can then be performed endoscopically by using forceps. Common complications include transient chest pain radiating to the back induced by stent expansion, nausea, and stent migration. Stents are usually left in place for 2 8 weeks, because longer indwelling periods can increase extraction difficulty. SEPS are more easily removed than SEMS, but migration risk is higher. Although use of partially covered stents can decrease migration rate because of tissue ingrowth into the stent, subsequent extraction becomes more difficult. If this occurs, APC can be used to fulgurate ingrown tissue. A largediameter SEPS can be placed inside the SEMS to induce pressure necrosis of the ingrown tissue; the stents can then be removed together a few days later. 87 A meta-analysis of stent placement for treatment of acute leaks after bariatric surgery by Puli et al 88 found a pooled proportion for successful leak closure, defined as radiographic evidence of leak closure after stent removal, of 87.8% (95% confidence interval, 79.4% 94.2%). Both SEMS and SEPS were used in 7 included studies. A majority of leaks closed with 1 treatment, but re-stenting was reported in 4 of 7 studies. Nine percent of patients had failure and required revision surgery. Stents were extracted between 4 and 8 weeks in the majority of studies. Stent migration was reported in 16.9% (95% confidence interval, 9.3% 26.3%). Endoscopic clips have also been used to close fistulas and leaks. Clips are used to approximate the tissue surrounding the defect to effect closure. The clip should be deployed perpendicular to the long axis of the defect. If needed, multiple clips can be placed sequentially, starting at either edge of a defect and meeting at the center. Thermal ablation or mechanical scraping of the tissue around the edges of the defect before clip deployment results in a more resilient seal. 89 Luminal distention and insufflation should be minimized before and after the procedure. The Over the Scope Clip (Ovesco Endoscopy AG, Tübingen, Germany) is a nitinol clip placed on a cap at the endoscope tip. An anchor clap and twin grasper instrument are available. 90 Unlike clips inserted through the endoscope, which appose mucosa, the Over the Scope Clip can perform full-thickness apposition. 91 Case series of gastrointestinal tract fistula closure have shown success rates of 72% 91% (Figure 4). 90,92,93 Sealants can be used to close fistulas and leaks. Fibrin sealant is a biodegradable compound with a long and varied history of surgical use. For endoscopic placement, any double-lumen catheter can be used except rapid-exchange catheters, in which the C-channel may allow sealant leakage within the endoscope. Care must be taken to insert the more viscous component via the larger lumen. Once the target mucosa has been excoriated, the fibrin can be applied, and a plug then forms. Multiple sessions may be needed. Lippert et al 94 reported a series of 52 patients with gastrointestinal fistula; 36.5% had sealing with fibrin alone, and 55.7% were cured with additional endoscopic therapy. Patients without infection had a higher cure rate. A series of 15 patients who failed conservative therapy for fistula was reported by Rabago et al. 95 Sealing of 86.6% of fistulas was achieved in a mean 16 days after a mean 2.5 sessions. Cure was more frequent in low-output fistulas than high-output fistulas. A randomized trial by Hwang et al 96 addressed 13 patients with persistent enterocutaneous fistulas after 2 4 weeks of conservative therapy. Closure time was 4 days in the patients randomized to fibrin glue vs 13 days in the control group. Wong et al 97 demonstrated treatment of postsurgical fistulas with insertion of fibrin glue during fistuloscopy by using a 5-mm choledochoscope. Other procedures were performed concurrently, including irrigation and debridement. All 9 fistulas sealed in a mean 18.7 days, without recurrence during the following year. An alternative sealant, cyanoacrylate, is associated with tissue necrosis and an inflammatory response. 98 It has been successfully

6 348 KUMAR AND THOMPSON CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 11, No. 4 Figure 4. Gastrogastric fistula (A) effectively closed after Overthe-Scope Clip placement (B). used as monotherapy to close gastrointestinal fistulas in small series. 99 Endoscopic suturing techniques have demonstrated effectiveness in fistula closure. However, device limitations combined with procedural complexity and need for specialized technical skill have limited adoption. The StomaphyX suturing system (EndoGastric Solutions, Redmond, WA) has been used by Overcash 100 to repair gastric leaks. A series of 2 cases demonstrated reduction of leak rate in 1 case and leak resolution in the other case. Fernandez-Esparrach et al 101 examined sutured gastrogastric fistula repair by using the Bard EndoCinch (C R Bard, Murray Hill, NJ). Initial success rate was 95%; durable success rate was 35%. None of the fistulas with diameter 20 mm remained sealed; fistula 10 mm had the best outcomes. There was 1 esophageal perforation and 1 occurrence of significant bleeding. It is possible that inadequate suture placement depth may have resulted in lack of successful closure. The Apollo OverStitch (Apollo Endosurgery, Austin, TX), which creates full-thickness plications, has shown early success (Figure 5). In an abstract presented by Watson and Thompson, 102 durable closure was achieved in 3 of 7 gastrogastric fistulas. No complications were noted. An emerging method for treatment of postsurgical leaks is vacuum-assisted sponge closure (Figure 6). The device consists of an open-cell sponge attached to external vacuum suction via tube. The sponge induces formation of granulation tissue, while vacuum suction improves perfusion and removes secretions. 103 A feeding tube is inserted intranasally and then orally exteriorized. A sponge, which is cut to a size smaller than the wound cavity, is fixed to the tip of the tube with suture. The sponge is grasped with endoscopic forceps and introduced into the fistula endoscopically. The feeding tube is then attached to continuous vacuum suction. The sponge is changed 2 or 3 times weekly. The system has been used successfully to close rectal anastomotic fistula. 104,105 It has also shown success in a prospective trial of treatment for intrathoracic anastomotic leaks. 106 Closure was achieved in 7 of 8 patients at a mean days after a median 7 endoscopies. Ahrens et al 107 reported use for gastroesophageal anastomotic leak. All 5 patients had leak closure at a median 42 days after a mean 9 sponge changes. Two patients developed stenosis requiring dilation, and 1 patient had hemorrhage after dilation. Surgisis anal fistula plugs (Cook Biotech, West Lafayette, IN) have been successfully used to treat enterocutaneous fistulas formed after bariatric surgery. 108 Surgisis is an acellular fibrogenic matrix; this material stimulates scar formation without foreign-body inflammatory reaction. 109 The plugs were developed for anorectal fistulas and have performed better than fibrin glue for that purpose. 110,111 Insertion begins by opacification of the fistula tract by contrast injection. A guidewire is inserted under fluoroscopic guidance from the percutaneous side, and the other end is grasped with a snare and pulled Figure 5. Gastrogastric fistula (A) after endoscopic repair (B). Figure 6. Endoscopic vacuum-assisted sponge device. Reprinted from Gastrointestinal Endoscopy; Wedemeyer J, Brangewitz M, Kubicka S, et al, Management of major postsurgical gastroesophageal intrathoracic leaks with an endoscopic vacuum-assisted closure system, 71(2): , Copyright 2010, with permission from Elsevier.

7 April 2013 ENDOSCOPIC MANAGEMENT OF BARIATRIC PATIENTS 349 through the mouth. The fistula tract is abraded over the wire until bleeding occurs. A snare is then attached to the guidewire and is passed through the fistula. This snare is used to grab the narrower end of the fistula plug. The plug is then pulled back into the fistula tract, and the snare is released. Larger-bore fistulas may require more plugs to achieve occlusion. In the series by Toussaint et al 108 of 5 enterocutaneous fistulas, 2 healed with one procedure, and 2 required a second procedure. One fistula did not seal. Notably, SEMS were also placed in some patients to cover the fistula tracts. Pancreaticobiliary Disease Endoscopic management of pancreatic and biliary disease with endoscopic ultrasound and endoscopic retrograde cholangiopancreatography (ERCP) presents a unique challenge in patients with altered anatomy. Unfortunately, rapid weight loss may induce a lithogenic state; nearly 50% of patients will develop gallstones or sludge after RYGB, and more than 25% may undergo cholecystectomy. 112 Before ERCP, preparation should include characterization of anatomy and pathology via cross-sectional imaging. Endoscopic ultrasound of the biliary system and the pancreas cannot be performed reliably after RYGB. Although patients with LAGB, SG, and VBG are usually amenable to ERCP with a side-viewing endoscope, patients with a history of RYGB and BPD DS often require special tools and procedures. Some patients are not amenable to a purely endoscopic approach. If surgery is planned, ERCP can be performed concurrently with laparoscopic assistance. In patients with BPD DS, ERCP may be possible via transanal approach; otherwise, laparoscopyassisted ERCP via a surgical gastrostomy or jejunostomy may be needed. 113,114 Patients with postoperative bile leaks likely need surgical intervention; ERCP is not safe immediately after surgery, and a radiologic approach is difficult because of an initial lack of intrahepatic dilation. 55 The post-rygb patient can present difficulty caused by a long intestinal limb, limited mobility of the mesentery, and acute angulation at the jejunojejunal anastomosis. An enteroscope or colonoscope can be used; alternatively, a forwardviewing endoscope can be used to introduce a stiff guidewire, and a duodenoscope can be advanced over the wire. 115 Single- or double-balloon enteroscopy can increase the depth of insertion, as can spiral overtube enteroscopy. 116 Once the papilla is reached, the balloon overtube and spiral overtube can assist with maintenance of endoscope tip position during cannulation. If a forward-viewing endoscope is used for cannulation, new challenges arise. The papilla must be cannulated in a forward-viewing configuration without use of an elevator. Tip angulation and torque may be compromised, complicating sphincterotomy; a protective stent may be helpful in this situation. Because force transmission is compromised when using a colonoscope or enteroscope, stone extraction and stent placement can become challenging. ERCP accessories may not be compatible with forward-viewing endoscopes. In these cases, specialized long accessories can be used. By using other points of enteral access, the continuity of the gastrointestinal tract can be temporarily restored for ERCP, allowing direct access to the papilla. This results in rapid, repeatable access to the papilla in a familiar orientation, and it allows use of conventional accessories. A surgically, endoscopically, or radiologically guided gastrostomy can be inserted into the excluded stomach, and the duodenoscope can be inserted percutaneously after tract dilation. 117 Alternatively, a doubleballoon enteroscope can be used to place a retrograde percutaneous gastrostomy tube into the excluded stomach. 118 This can also be dilated after maturation. If immediate ERCP is needed, laparoscopic access to the excluded stomach can be obtained via gastric trocar. 117 Weight Regain and Dilated Gastrojejunal Anastomosis Bariatric surgery is effective in achieving durable weight loss, but weight regain postoperatively is a common and significant problem. It reintroduces the risks of obesity-associated diseases and has significant impact on quality of life. 119 Although initial weight loss after bariatric surgery is often dramatic, a weight plateau is typically achieved in 1 2 years. 120 Approximately 20% of patients do not lose 50% of excess weight within 1 year of surgery; 30% of patients experience weight regain by 2 years postoperatively, and 63.6% regain within 4 years Because of the large number of patients undergoing bariatric surgery, demand for therapy to address weight regain will continue to increase. Weight regain can occur as a result of postoperative changes in neuroendocrine-metabolic regulation that result in a starvation response that induces increased appetite and energy conservation. 124,125 Decreased satiety may also occur if there is loss of restriction. It has been shown that larger pouch size and GJA diameter are associated with postoperative weight regain. 126,127 Gastrogastric fistula is another possible etiology. 128 Surgical revision to address these issues is problematic. Complication rates are prohibitively high, reaching more than 15%. 129 Procedures entail longer intraoperative time and greater blood loss. 83,84 Endoluminal therapy has shown promise in effectively addressing weight regain with lower morbidity. Injection and tissue plication techniques have been studied. Endoscopic sclerotherapy is performed by injecting morrhuate sodium around the GJA, with the goal of reducing aperture and tissue compliance. After injecting a 2-mL test dose of morrhuate sodium into the rim of the GJA to monitor for adverse reaction, morrhuate sodium is injected circumferentially around the GJA in 2-mL aliquots. A total of ml is injected. 130 The procedure is repeated every 3 6 months until the GJA measures less than 12 mm in diameter; 2 3 sessions are typically required. 131 A study of 28 patients in 2007 showed loss of 75% of regained weight in 64% of patients. 130 A subsequent study of 71 patients demonstrated weight maintenance or loss in 72% of patients at 1 year. 132 Patients with presclerotherapy GJA diameter greater than 30 mm did not appear to have successful outcomes. Morrhuate sodium is currently unavailable from commercial sources. Alternatives are under investigation to determine whether similar weight loss outcomes can be achieved. Endoscopic sutured revision of dilated GJA and gastric pouch has been studied on multiple platforms. Some of the devices discussed herein have been described above. The Bard EndoCinch can be used to decrease GJA diameter. After pretreating the rim of the GJA with APC, the device is used to place interrupted stitches across the anastomosis. 133 A randomized, double-blinded trial compared EndoCinch revision of dilated GJA with sham procedure in 77 patients with GJA diameter 20 mm. 134 GJA diameter was reduced to 10 mm in 89%,

8 350 KUMAR AND THOMPSON CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 11, No. 4 there were no perforations, and the adverse event rate was similar to the sham group. Ninety-six percent of revised patients had weight loss or stabilization in the following 6 months. Mean weight loss in the revised group was 4.7% 5.7% vs 1.9% 5.2% in the sham group (P.041) in a per-protocol analysis. The multichannel Incisionless Operating Platform (USGI Medical, San Clemente, CA) can reduce dilated gastric pouch and GJA with full-thickness plications via the Revision Obesity Surgery Endoscopic procedure. 135 The channels hold a 4.9-mm endoscope, tissue grasper, and tissue approximator. Tissue plications are created by using anchors. A prospective study of 20 patients with weight regain achieved technical success in 85% of patients, reducing pouch size by a mean 36% and GJA by 65% (to a mean 16 mm). 136 The result was mean weight loss of 8.8 kg after 3 months. A second-generation device capable of working in smaller pouches was subsequently developed; this achieved success in 5 of 5 patients with weight loss of 7.8 kg after 3 months. A prospective multicenter registry of 116 patients with weight regain in the setting of pouch and GJA dilation reported technical success in 112 patients (97%) and no procedural complications. 137 There was a mean 44% reduction in pouch length and 50% reduction in GJA diameter. After 6 months, patients lost a mean 32% of regained weight. When postprocedure GJA diameter of 10 mm was achieved, patients achieved 24% excess weight loss vs 10% for the rest of the cohort. 138 The Apollo OverStitch can place full-thickness sutures by using a cap-based suturing system that fits over a standard double-channel endoscope. 84 Only one hand is needed for suture deployment, allowing control of suture placement depth. Revision of dilated GJA has been demonstrated, and further study is in progress. Conclusion With the rising cumulative number of patients undergoing bariatric surgery, gastroenterologists will encounter both variant postsurgical anatomy and postoperative complications. Familiarity with the complications bariatric patients face and knowledge of effective methods to address these challenges will arm the gastroenterologist to deliver increasingly effective patient care. References 1. Nguyen NT, Magno CP, Lane KT, et al. Association of hypertension, diabetes, dyslipidemia, and metabolic syndrome with obesity: findings from the National Health and Nutrition Examination Survey, 1999 to J Am Coll Surg 2008;207: Sjöström L, Lindroos AK, Peltonen M, et al. Lifestyle, diabetes, and cardiovascular risk factors 10 years after bariatric surgery. N Engl J Med 2004;351: Buchwald H, Oien DM. Metabolic/bariatric surgery worldwide. Obes Surg 2009;19: Buchwald H, Avidor Y, Braunwald E, et al. Bariatric surgery: a systematic review and meta-analysis. JAMA 2004;13: Podnos YD, Jimenez JC, Wilson SF, et al. Complications after laparoscopic gastric bypass: a review of 3464 cases. Arch Surg 2003;138: Iannelli A, Dainese R, Piche T, et al. Laparoscopic sleeve gastrectomy for morbid obesity. World J Gastroenterol 2008;14: Papakonstantinou A, Terzis L, Stratopoulos C, et al. Bleeding from the upper gastrointestinal tract after Mason s vertical banded gastroplasty. Obes Surg 2000;10: Belachew M, Belva PH, Desaive C. Long-term results of laparoscopic adjustable gastric banding for the treatment of morbid obesity. Obes Surg 2002;12: Bakhos C, Alkhoury F, Kyriakides T, et al. Early postoperative hemorrhage after open and laparoscopic Roux-en-Y gastric bypass. Obes Surg 2009;19: Rabl C, Peeva S, Prado K, et al. Early and late abdominal bleeding after Roux-en-Y gastric bypass: sources and tailored therapeutic strategies. Obes Surg 2011;21: Nguyen NT, Rivers R, Wolfe BM. Early gastrointestinal hemorrhage after laparoscopic gastric bypass. Obes Surg 2003;13: Tang SJ, Rivas H, Tang L, et al. Endoscopic hemostasis using endoclip in early gastrointestinal hemorrhage after gastric bypass surgery. Obes Surg 2007;17: Marano BJ Jr. Endoscopy after Roux-en-Y gastric bypass: a community hospital experience. Obes Surg 2005;15: Keith JN. Endoscopic management of common bariatric surgical complications. Gastrointest Endosc Clin N Am 2011;21: Azagury DE, Abu Dayyeh BK, Greenwalt IT, et al. Marginal ulceration after Roux-en-Y gastric bypass surgery: characteristics, risk factors, treatment, and outcomes. Endoscopy 2011;43: Hedberg J, Hedenström H, Nilsson S, et al. Role of gastric acid in stomal ulcer after gastric bypass. Obes Surg 2005;15: Ramaswamy A, Lin E, Ramshaw BJ, et al. Early effects of Helicobacter pylori infection in patients undergoing bariatric surgery. Arch Surg 2004;139: Rasmussen JJ, Fuller W, Ali MR. Marginal ulceration after laparoscopic gastric bypass: an analysis of predisposing factors in 260 patients. Surg Endosc 2007;21: Wilson JA, Romagnuolo J, Byrne TK, et al. Predictors of endoscopic findings after Roux-en Y gastric bypass. Am J Gastroenterol 2006;101: Dallal RM, Bailey LA. Ulcer disease after gastric bypass surgery. Surg Obes Relat Dis 2006;2: Gonzalez R, Lin E, Venkatesh R, et al. Gastrojejunostomy during laparoscopic gastric bypass: analysis of three techniques. Arch Surg 2003;138: Vasquez JC, Wayne Overby D, Farrell TM. Fewer gastrojejunostomy strictures and marginal ulcers with absorbable suture. Surg Endosc 2009;23: Kaplan LM. Gastrointestinal management of the bariatric surgery patient. Gastroenterol Clin North Am 2005;34: Safatle-Ribeiro AV, Kuga R, Iriya K, et al. What to expect in the excluded stomach mucosa after vertical banded Roux-en-Y gastric bypass for morbid obesity. J Gastrointest Surg 2007;11: Gisbert JP, de la Morena F, Abraira V. Accuracy of monoclonal stool antigen test for the diagnosis of H. pylori infection: a systematic review and meta-analysis. Am J Gastroenterol 2006; 101: Gumbs AA, Duffy AJ, Bell RL. Management of gastrogastric fistula after laparoscopic Roux-en Y gastric bypass. Surg Obes Relat Dis 2006;2: Frezza EE, Herbert H, Ford R, et al. Endoscopic suture removal at gastrojejunal anastomosis after Roux-en Y gastric bypass to prevent marginal ulceration. Surg Obes Relat Dis 2007;3: Ryou M, Mogabgab O, Lautz DB, et al. Endoscopic foreign body removal for treatment of chronic abdominal pain in patients

9 April 2013 ENDOSCOPIC MANAGEMENT OF BARIATRIC PATIENTS 351 after Roux-en-Y gastric bypass. Surg Obes Relat Dis 2010; 6: Felix EL, Kettelle J, Mobley E, et al. Perforated marginal ulcers after laparoscopic gastric bypass. Surg Endosc 2008;22: Yu S, Jastrow K, Clapp B, et al. Foreign material erosion after laparoscopic Roux-en-Y gastric bypass: findings and treatment. Surg Endosc 2007;21: Chevallier JM, Zinzindohoue F, Douard R, et al. Complications after laparoscopic adjustable gastric banding for morbid obesity: experience with 1,000 patients over 7 years. Obes Surg 2004;14: O Brien PE, Dixon JB, Laurie C, et al. A prospective randomized trial of placement of the laparoscopic adjustable gastric band: comparison of the perigastric and pars flaccida pathways. Obes Surg 2005;15: Singhal R, Bryant C, Kitchen M, et al. Band slippage and erosion after laparoscopic gastric banding: a meta-analysis. Surg Endosc 2010;24: Abu-Abeid S, Keidar A, Gavert N, et al. The clinical spectrum of band erosion following laparoscopic adjustable silicone gastric banding for morbid obesity. Surg Endosc 2003;17: Moreno P, Alastrué A, Rull M, et al. Band erosion in patients who have undergone vertical banded gastroplasty: incidence and technical solutions. Arch Surg 1998;133: Westling A, Bjurling K, Ohrvall M, et al. Silicone-adjustable gastric banding: disappointing results. Obes Surg 1998;8: Regusci L, Groebli Y, Meyer JL, et al. Gastroscopic removal of an adjustable gastric band after partial intragastric migration. Obes Surg 2003;13: Lattuada E, Zappa MA, Mozzi E, et al. Band erosion following gastric banding: how to treat it. Obes Surg 2007;17: Offodile AC, Okafor P, Shaikh SN, et al. Duodenal obstruction due to erosion and migration of an adjustable gastric band: a novel endoscopic approach to management. Surg Obes Relat Dis 2010;6: Evans JA, Williams NN, Chan EP, et al. Endoscopic removal of eroded bands in vertical banded gastroplasty: a novel use of endoscopic scissors (with video). Gastrointest Endosc 2006; 64: Adam LA, Silva RG Jr, Rizk M, et al. Endoscopic argon plasma coagulation of Marlex mesh erosion after vertical-banded gastroplasty. Gastrointest Endosc 2007;65: Blero D, Deviere J. Removing foreign bodies in bariatric patient. Tech Gastrointest Endosc 2010;12: Eubanks S, Edwards CA, Fearing NM, et al. Use of endoscopic stents to treat anastomotic complications after bariatric surgery. J Am Coll Surg 2008;206: Blero D, Eisendrath P, Vandermeeren A, et al. Endoscopic removal of dysfunctioning bands or rings after restrictive bariatric procedures. Gastrointest Endosc 2010;71: Higa K, Ho T, Tercero F, et al. Laparoscopic Roux-en-Y gastric bypass: 10-year follow-up. Surg Obes Relat Dis 2011;7: Mathew A, Veliuona MA, DePalma FJ, et al. Gastrojejunal stricture after gastric bypass and efficacy of endoscopic intervention. Dig Dis Sci 2009;54: Smith SC, Edwards CB, Goodman GN, et al. Open vs laparoscopic Roux-en-Y gastric bypass: comparison of operative morbidity and mortality. Obes Surg 2004;14: Carrodeguas L, Szomstein S, Zundel N, et al. Gastrojejunal anastomotic strictures following laparoscopic Roux-en-Y gastric bypass surgery: analysis of 1291 patients. Surg Obes Relat Dis 2006;2: Nguyen NT, Stevens CM, Wolfe BM. Incidence and outcome of anastomotic stricture after laparoscopic gastric bypass. J Gastrointest Surg 2003;7: Wetter A. Role of endoscopy after Roux-en-Y gastric bypass surgery. Gastrointest Endosc 2007;66: Madan AK, Harper JL, Tichansky DS. Techniques of laparoscopic gastric bypass: on-line survey of American Society for Bariatric Surgery practicing surgeons. Surg Obes Relat Dis 2008;4: Cottam DR, Fisher B, Sridhar V, et al. The effect of stoma size on weight loss after laparoscopic gastric bypass surgery: results of a blinded randomized controlled trial. Obes Surg 2009;19: Fernández-Esparrach G, Bordas JM, Llach J, et al. Endoscopic dilation with Savary-Gilliard bougies of stomal strictures after laparoscopic gastric bypass in morbidly obese patients. Obes Surg 2008;18: Go MR, Muscarella P, Needleman BJ, et al. Endoscopic management of stomal stenosis after Roux-en-Y gastric bypass. Surg Endosc 2004;18: Obstein KL, Thompson CC. Endoscopy after bariatric surgery (with videos). Gastrointest Endosc 2009;70: Peifer KJ, Shiels AJ, Azar R, et al. Successful endoscopic management of gastrojejunal anastomotic strictures after Roux-en-Y gastric bypass. Gastrointest Endosc 2007;66: Espinel J, De-la-Cruz JL, Pinedo E, et al. Stenosis in laparoscopic gastric bypass: management by endoscopic dilation without fluoroscopic guidance. Rev Esp Enferm Dig 2011;103: Azagury DE, Lautz DB. Endoscopic techniques in bariatric patients: obesity basics and normal postbariatric surgery anatomy. Tech Gastrointest Endosc 2010;12: Sataloff DM, Lieber CP, Seinige UL. Strictures following gastric stapling for morbid obesity: results of endoscopic dilatation. Am Surg 1990;56: Schneider BE, Villegas L, Blackburn GL, et al. Laparoscopic gastric bypass surgery: outcomes. J Laparoendosc Adv Surg Tech A 2003;13: Morales MP, Miedema BW, Scott JS, et al. Management of postsurgical leaks in the bariatric patient. Gastrointest Endosc Clin N Am 2011;21: Filho AJ, Kondo W, Nassif LS, et al. Gastrogastric fistula: a possible complication of Roux-en Y gastric bypass. JSLS 2006; 10: Cucchi SG, Pories WJ, MacDonald KG, et al. Gastrogastric fistulas: a complication of divided gastric bypass surgery. Ann Surg 1995;221: Márquez MF, Ayza MF, Lozano RB, et al. Gastric leak after laparoscopic sleeve gastrectomy. Obes Surg 2010;20: Fernandez AZ, DeMaria EJ, Tichansky DS, et al. Experience with over 3000 open and laparoscopic bariatric procedures: multivariate analysis of factors related to leak and resultant mortality. Surg Endosc 2004;18: Ballesta C, Berindoague R, Cabrera M, et al. Management of anastomotic leaks after laparoscopic Roux-en-Y gastric bypass. Obes Surg 2008;18: Livingston EH, Huerta S, Arthur D, et al. Male gender is a predictor of morbidity and age a predictor of mortality for patients undergoing gastric bypass surgery. Ann Surg 2002;236: Csendes A, Burgos AM, Braghetto I. Classification and management of leaks after gastric bypass for patients with morbid obesity: a prospective study of 60 patients. Obes Surg 2012; 22: MacLean LD, Rhode BM, Nohr C, et al. Stomal ulcer after gastric bypass. J Am Coll Surg 1997;185:1 7.