Endoscopic techniques for the treatment of postoperative fistulae and leaks are rapidly developing. Conventional surgical therapy for postsurgical leaks and fistulae is associated with significant morbidity and mortality. Novel endoscopic therapies have demonstrated safety, despite the inherent challenges of intervention in this patient population, and are steadily building evidence for efficacy relative to surgical management. The article examines endoscopic therapy for leaks and fistulae after esophageal, gastric, bariatric, colonic, and pancreaticobiliary surgery.
Key Points
- •
Endoscopic techniques for the treatment of postoperative fistulae and leaks are rapidly developing.
- •
Conventional surgical therapy for postsurgical leaks and fistulae is associated with significant morbidity and mortality.
- •
Novel endoscopic therapies, such as endoscopic suturing, novel stent types, and vacuum-assisted sponge closure, have demonstrated safety and are building evidence for efficacy.
- •
We examine endoscopic therapy for leaks and fistulae after esophageal, gastric, bariatric, colonic, and pancreaticobiliary surgery.
Introduction
Endoscopic techniques for the diagnosis and therapy for gastrointestinal disease are rapidly developing. Fistulae are abnormal communications originating in a visceral structure. Postoperative leaks are 1 type of fistula, defined as discontinuity of tissue apposition in the immediate postoperative period. Conventional surgical therapy for postsurgical leaks and fistulae is associated with significant morbidity and mortality, and new endoscopic options offer significant benefits. This article examines endoscopic therapy for leaks and fistulae after esophageal, gastric, bariatric, colonic, and pancreaticobiliary surgery.
Introduction
Endoscopic techniques for the diagnosis and therapy for gastrointestinal disease are rapidly developing. Fistulae are abnormal communications originating in a visceral structure. Postoperative leaks are 1 type of fistula, defined as discontinuity of tissue apposition in the immediate postoperative period. Conventional surgical therapy for postsurgical leaks and fistulae is associated with significant morbidity and mortality, and new endoscopic options offer significant benefits. This article examines endoscopic therapy for leaks and fistulae after esophageal, gastric, bariatric, colonic, and pancreaticobiliary surgery.
Esophageal leaks and fistulae
Scope of the Problem
Postoperative esophageal leaks and fistulae can develop after esophagectomy or gastrectomy. The intrathoracic leak rate after esophageal resection has been reported at 7.9%, resulting in a 3-month mortality rate of 18.2% (OR 3.0). If treatment is delayed beyond the first 24 hours, a mortality rate has been reported at up to 50%. Esophageal leakage after gastrectomy occurs in 4% to 27%, and a mortality rate of 65% has been reported.
Past Management Options
Surgical therapy for postoperative leak historically consisted of surgical drainage and repair, nothing by mouth, parenteral nutrition, and antibiotics. The result was a mortality rate of up to 60%. Fever, systemic inflammatory response syndrome, and abnormal C-reactive protein, white blood cell count, and albumin are indicators of postoperative esophageal leak.
Current Status
Early diagnosis is important to prevent fulminant infection. Endoscopy is useful in diagnosis because the defect, integrity of surrounding tissue, and infection in the adjacent tissue can be assessed; these factors are important in choosing treatment modality. Fibrin glue or endoscopic clip placement can be considered for small defects, although patients with dehiscence of 30% to 70% of the esophageal circumference likely warrant stent placement. Larger defects are not amenable to endoscopic therapy.
Esophageal stent placement for postoperative fistula and leak has become well-established. Both covered self-expanding metal stents (SEMS) and self-expanding plastic stents (SEPS) have been studied. Stent placement excludes the defect to allow healing, prevents stricture formation, and may allow oral feeding. However, stent placement can be complicated by inadequate defect closure, stent migration, and difficult removal caused by tissue ingrowth or stricture. Leaks at an anastomosis where the esophagus is contiguous with small intestine or colon are challenging candidates for stent placement because of the risk of necrosis and migration, respectively. Freeman and colleagues recently reported that factors associated with failure of leak resolution after stent placement included leak of the proximal cervical esophagus, stent traversing the gastroesophageal junction, esophageal injury greater than 6 cm, and anastomotic leak associated with a more distal conduit leak.
New Endolumenal Options
Novel stents have been used to close postoperative esophageal leaks and fistulae. A covered biodegradable stent, which does not need to be removed after therapy completion or migration, has demonstrated leak closure in 4 of 5 patients. A mushroom-shaped metallic stent has been used successfully for durable treatment of gastroesophageal anastomotic fistula in 8 of 8 patients, with successful removal in within 48 days.
Endoscopic clips can be used to close fistulae and leaks. Clips are deployed perpendicular to the long axis of the defect to approximate the edges of the defect. Multiple clips can be placed beginning at either edge of a defect and meeting at the center. Mechanical scraping or thermal ablation of the defect edges before clip placement results in a more durable seal. Luminal distention should be avoided before and after clip placement. The Over the Scope Clip (OTSC; Ovesco Endoscopy AG, Tübingen, Germany) is a nitinol clip deployed from a cap at the endoscope tip after tissue is suctioned into the cap. The OTSC can perform full-thickness apposition, unlike through-the-scope clips. Case series have shown success rates of 72% to 91% for gastrointestinal tract fistula closure. Success has been reported for initial closure of postoperative esophagojejunal anastomotic leaks and fistulae.
Sealants have been used to treat postoperative esophageal leaks. N-butyl-2-cyanoacrylate (Histoacryl; B. Braun Dexon GmbH, Spangenberg, Germany) has been successfully used to repair an esophagojejunal anastomotic leak after 3 weeks of conservative therapy failed. Fibrin glue, which has a long history of surgical use, has been used to close an esophageal fistula via submucosal injection into the lateral walls of the fistula rather than injection into the fistula. The glue is injected into the submucosa until the lumen is occluded, ensuring that the fibrin plug does not become dislodged. No abrasion or deepithelialization of the fistula tract is required. Vicryl mesh (Ethicon, Hamburg, German) has been combined with use of fibrin sealant for closure of larger postoperative leaks or fistulae. After the mesh is placed over the defect, it is covered with 2 to 3 mL fibrin (Tissucol Duo, Baxter, Germany), and then fibrin is injected into the submucosa at the edges of the defect. The defect eventually epithelializes. Thirteen of 15 patients had long-term success with this technique. Surgisis Soft Tissue Graft (Cook, West Lafayette, IN), an acellular bioactive prosthetic matrix, has been used successfully in conjunction with the OTSC and SEMS ( Fig. 1 ).
Vacuum-assisted sponge closure (VAC) therapy has been investigated for closure of esophageal defects. Because negative intrathoracic pressure can draw fluid through even small defects with each inspiration, intrathoracic collections and infection can develop. VAC addresses this by occluding the defect and continuously withdrawing secretions. The device consists of an open-cell sponge attached to external vacuum suction via tube ( Fig. 2 ). The sponge induces the formation of granulation tissue, while vacuum suction improves perfusion and removes secretions. Small defects with well-perfused surrounding tissue and noncompartmentalized adjacent collections are amenable. VAC has some advantages over stent placement, allowing simultaneous closure of luminal defects, drainage of infectious foci, and granulation tissue formation in the wound cavity. To place the device, a feeding tube is inserted intranasally and then orally exteriorized. A sponge, 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. Ahrens and colleagues reported its use for gastroesophageal anastomotic leaks. 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.
A cardiac septal occluder (Amplatzer Occluder; AGA Medical Corp., Plymouth, MN) has been used to close a postoperative esophagotracheal fistula after failure of endoscopic clip, glue, and stent therapy. The device is a self-expanding nitinol wire mesh in the shape of 2 umbrellas linked together ( Fig. 3 ). Under fluoroscopic and endoscopic guidance, a guidewire is inserted via the esophagus through the fistula and then orally exteriorized via the trachea. The occluder is then inserted on the guidewire and released first on the tracheal side and then on the esophageal side. Durable results were noted at 8 months.
Bariatric leaks and fistulae
Scope of the Problem
Postoperative leaks have been reported in 1.7 to 5.2% of patients after Roux-en-Y gastric bypass (RYGB) and in 1.5% to 2.4% after sleeve gastrectomy (SG). Leak after RYGB can occur at the gastric pouch, gastrojejunal anastomosis, jejunal stump, jejunojejunal anastomosis, excluded stomach, duodenal stump (in resectional bypass), and blind jejunal limb. Leaks most commonly occur at the gastrojejunal (68%) or jejunojejunal (5%) anastomosis, at the gastric pouch staple lines (10%), or at multiple sites (14%). Leaks in the defunctionalized stomach may be especially challenging to diagnose. Most leaks in patients with SG occur in the proximal third of the stomach (85.7%). Distal stenosis may predispose to fistula formation. Leak is a major risk factor for mortality after bariatric surgery. The mortality rate after leak is 6% to 14.7%, with a mortality rate of 9% after leak at the gastrojejunal anastomosis but 40% after leak at the jejunojejunal anastomosis. After RYGB, chronic gastrogastric fistula occurs more commonly with the open approach, when the pouch is contiguous with the excluded stomach.
Past Management Options
Management of postoperative leak is complicated by infection, multiorgan failure, and nutritional deficiency. Surgical management results in a high conversion rate to open surgery (48%), morbidity (up to 50%), and mortality (2%–10%). Conservative management entails percutaneous drainage of collections, distal enteral or total parenteral feeding, and broad-spectrum antibiotics.
Current Status
Endoscopic techniques to address leak and fistula after bariatric surgery offer an excellent safety profile compared with surgical management. Endoscopic management should include dilation of distal stenoses if necessary. Closure can be performed with clips, sealants, and suturing devices. Stents can exclude defects to promote healing.
Leak source must be localized before therapy can be effectively performed. The procedure is started by performing a bubble test. The external drain is submerged underwater, and carbon dioxide is insufflated through the endoscope. The presence of bubbles indicates open communication with the drain. The next step is the injection of methylene blue with contrast into the drain with simultaneous endoscopic and fluoroscopic visualization of the suspected leak sites. Once the site is localized, the optimal therapy can be selected. If the leak is in the proximal pouch, a stent is the likely optimal therapy. If the leak is in a recessed area, such as the blind portion of the Roux limb, or in the distal bowel, a stent will not be effective; clips, sealants, or suturing should be considered.
Placement of both SEMS and SEPS for the treatment of leaks and fistulae is supported by considerable evidence. Enteral nutrition can be resumed during the healing process, and parenteral nutrition can be avoided. Extraluminal contamination is decreased, resulting in less infection and pain. A meta-analysis of 7 studies of stent placement (including both SEMS and SEPS) for the treatment of acute leak after bariatric surgery found a pooled proportion for radiographic evidence of leak closure after stent removal of 87.8% (95% CI 79.4–94.2%). Duration of stent therapy was between 4 and 8 weeks in most cases. Only 9% of patients proceeded to revision surgery. The pooled proportion of stent migration was 16.9% (95% CI 9.3–26.3%). Stent placement after SG has been shown effective in small case series. Future bariatric-specific stent designs may improve efficacy and migration rate.
SEMS and SEPS each have advantages. SEPS are more likely to migrate. The use of partially covered SEMS decreases migration risk, but tissue ingrowth may complicate removal. In these cases, large-diameter SEPS placed within the SEMS can cause pressure necrosis of the ingrown tissue, and both stents can be removed together a few days later. Alternatively, ingrown tissue can be fulgurated with argon plasma coagulation. Fully covered SEMS are effective for the treatment of leak and fistula, and migration can be prevented by securing the proximal end of the stent to the mucosa. Placement of 2 to 4 endoscopic clips on fully covered SEMS resulted in a migration rate of 13% versus 57% in the control group. Attachment of a polypectomy snare to a stent, with the catheter exteriorized nasally and attached to the earlobe, has also been reported.
Stents are placed using a forward-viewing endoscope and fluoroscopic visualization. The leak site and gastroesophageal junction should be marked externally using radio-opaque markers. The endoscope should be advanced into the Roux limb (in patients with RYGB) or the third portion of the duodenum (in patients with SG), and a guidewire should be introduced. The endoscope can then be removed and the stent delivery system inserted with fluoroscopic visualization. Stent length should be chosen so that the proximal end does not approach the upper esophageal sphincter (inducing globus sensation) and the distal end does not impact the enteral wall (inducing ulceration, bleeding, or perforation). A stiff guidewire can be used to assist traversion of acute angulation. Once the stent is deployed, its position can be adjusted endoscopically with forceps.
Sealants have shown efficacy in leak and fistula closure after bariatric surgery. The target mucosa should be ablated or excoriated to assist healing. A double-lumen catheter can be used to apply glue endoscopically, with the more viscous component inserted via the larger lumen (rapid-exchange catheters should not be used as they are prone to leakage). The fibrin is injected, forming a plug as it hardens. Fibrin sealant can also be used to fill residual cavities. Rábago and colleagues studied fistula closure with fibrin in a series of 15 patients after failure of conservative therapy; 86.6% sealed after a mean 2.5 sessions. High-output fistulae were less likely to seal. Wong and colleagues used a 5-mm choledochoscope to perform fistuloscopy in patients with postoperative gastrointestinal fistulae; fibrin glue was inserted, and some irrigation and debridement were concurrently performed in some patients. Nine of 9 fistulae were successfully sealed with no recurrence after 12 months. Cyanoacrylate, another sealant, is associated with tissue necrosis and inflammatory response; however, it is not inactivated by gastric or pancreatic enzymes. It has shown efficacy in closure of fistulae in small series.
Multimodality therapy is often required. Bége and colleagues prospectively evaluated an entirely endoscopic approach for management of fistulae after bariatric surgery in 27 patients. Clips, cyanoacrylate glue, and stents were used alone or in combination, often after endoscopic debridement. The first procedure was successful in 41%, but all patients eventually achieved resolution after a mean 4.4 endoscopies at a mean 86 days. Lippert and colleagues reported a success rate of 36.5% with fibrin alone and a 55.7% success rate with multimodal endoscopic therapy in a series of 52 patients with gastrointestinal fistula. Infection was a predictor of failure.
New Endoluminal Options
Surgisis AFP plugs (Cook Biotech, West Lafayette, IN), developed for surgical treatment of anal fistulae, have shown success in the treatment of enterocutaneous fistulae after bariatric surgery. The acellular nature of Surgisis, a fibrogenic matrix, means that it can stimulate fistula closure without inducing an inflammatory foreign-body reaction. Toussaint and colleagues demonstrated closure of 4 of 5 enterocutaneous fistulae in 1 or 2 procedures. SEMS were used concurrently in some patients. Surgisis AFP insertion is performed under fluoroscopic guidance. The fistula tract is opacified with contrast, and a guidewire is inserted into the percutaneous fistula opening. The other end of the guidewire is captured with a snare and orally exteriorized. The fistula tract is abraded over the guidewire. Next, a snare is passed through the fistula over the wire and used to grab the narrow end of the Surgisis plug. It is used to pull the plug into the fistula. Multiple plugs can be placed adjacently into a large-diameter fistula. Surgisis strips and AFP were used to treat gastrocutaneous fistulae in 25 post-RYGB patients. The AFP (5 patients) were deployed into fistulae as described earlier; strips (20 patients) were maneuvered into the fistula from the luminal end using a polypectomy snare. Successful closure was achieved in all patients treated with AFP and 75% of patients treated with strips.
Endoscopic suturing techniques are effective for fistula closure ( Fig. 4 ). However, device limitations and procedural complexity mean specialized technical skill is required. Fernandez-Esparrach and colleagues used the Bard EndoCinch (CR Bard, Murray Hill, NJ) to repair gastrogastric fistulae. Although initial success was achieved in 95% of patients, the durable success rate was 35%. Fistulae with a diameter less than or equal to 10 mm had the best outcomes; no fistula with a diameter greater than 20 mm remained sealed. Inadequate suture placement depth may have been responsible for the lack of durable closure. One patient had significant bleeding, and another had esophageal perforation. The StomaphyX suturing system (EndoGastric Solutions, Redmond, WA) was used by Overcash to repair 2 gastric leaks. One achieved a reduction of leak rate, and the other achieved leak resolution. The Apollo OverStitch (Apollo Endosurgery, Austin, TX), which creates full-thickness plications, achieved durable gastrogastric fistula closure in 3 of 7 cases in an abstract presented by Watson and Thompson. There were no procedural complications.
Related posts:
Endolumenal Therapy of Gastrointestinal Disorders
Endoscopic Therapy of Barrett Esophagus
Advances in Endoluminal Therapy for Esophageal Cancer
Small Bowel Polyps, Arteriovenous Malformations, Strictures, and Miscellaneous Lesions
Early Gastric Cancer and Dysplasia
Endoscopic Resection of Large Colon Polyps
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
