Endoscopic management of postoperative fistulas and leaks

Abstract

Introduction

In the current era of minimally invasive surgery for treatment of myriad of gastrointestinal (GI) conditions there has been a rise in postoperative complications, which include anastomotic leaks, strictures, acute bleeding and fistulae. The type of complication and time interval for development of complication after surgery depends on the type of surgery among which bariatric surgery is the most common. About 140,000 bariatric surgeries are performed in the US in the past decade and post-operative leaks and fistulae have increased in frequency too.1 Other surgical complications requiring endoscopic intervention include resection of GI tumors, liver transplant, pancreatico-biliary surgeries and postoperative bile leaks. Repeat surgery in these situations leads to prolonged hospitalization period, increase risk of infection and add significant burden to already increasing health care costs and may lead to increased morbidity and mortality. A multidisciplinary approach, frequently involving therapeutic endoscopy plays an important role in management by offering a minimally invasive modality of tackling these problems and thereby potentially minimizing the morbidity and mortality associated with these conditions and furthermore could decrease length of hospitalization and improved patient outcomes. There has been significant development of various techniques in the past decade including self-expanding stents, endoclips, fibrin glue and vacuum-assisted devices for management of fistula/leaks and balloon dilation and electrocautery incision in management of challenging strictures.2 In the hands of a trained therapeutic endoscopist, these procedures can be performed safely with high degree of success.1 We will review these techniques used in management of various surgical complications.

Endoscopic Clips

Fistula is an abnormal communication originating from a viscera. It can be classified as an internal fistula: from one viscera to another or an external fistula: from the viscera to the outside. Depending on the output can also be classified into high output (> 500 mL/day) or low output (< 500 mL/day). Postoperative leaks are defined as discontinuity of tissue apposition in the immediate postoperative period.3 Conventional surgical treatment for these types of complications carry significant conversion rate to open surgery (48%), leading to morbidity (up to 50%) and mortality (2%–10%).3–7 New endoscopic techniques have been used successfully to manage these complications. Endoscopic clips—both Through the Scope Clip (TTSC) and Over the Scope Clip (OTSC) can be used to treat fistulae and leaks. Feasibility of endoscopic long-term closure using clips and suturing compared to thoracoscopic closure was demonstrated in a randomized controlled study using an animal model.8

Multiple TTSCs can be deployed perpendicular to the long axis of the postoperative defect and the edges approximated to promote successful healing. Edges of the defect can be scraped mechanically or thermally to promote granulation tissue growth and lead to a more durable seal.9 After placement of the clips, luminal distension should be avoided to minimize the risk of dehiscence. TTS clips however cannot be reliably used to approximate a full thickness defect (involving all the wall layers) for which an OTSC can be successfully used.10 Other limitations of TTSC include smaller width of the clips arms (10–12 mm when completely opened) and limited tissue pressure from the clip which necessitates multiple clips to cover smaller defects there by increasing the cost of the procedure.11,12 These limitations could be overcome by using OTSC (Ovesco Endoscopy AG, Tübingen, Germany) which is a Nitinol clip that comes loaded over a transparent cap that can be fitted to the tip of the scope and deployed from the cap after the tissue surrounding the defect is suctioned into the cap.13 It can be used for full thickness defects and has high compression force on tissue.14 Multiple case series have shown good success rates ranging from 72% to 91%.13,15,16

Postoperative leaks after esophagectomy occur in up to 35% of cases17 and 4% to 27% after gastrectomy, carrying a mortality as high as 65%.18,19 Intrathoracic leak following esophagectomy carries higher mortality compared to cervical leak20 and if there is a delay in treatment for over 24 hours, mortality approaches as high as 50%.21 Endoscopic closure of esophageal leaks was first demonstrated by Schubert et al22 in which clips were used to treat smaller esophageal leaks. Qadeer et al23 reported successful closure of 10 mm esophageal perforations in a case series of 17 patients with varying etiologies. Though the role of clip closure has been very limited in complications after esophageal cancer resection, it has been used as an adjunctive to stent placement.24 Occasionally perforations are caused by drains placed for drainage of the inflammatory fluid collections and these can be managed by pulling the drain back and closing the defect with an OTSC (Fig. 1).

Figure 1

(A) Perforation with visible drain. (B) Perforation closed with Over the Scope Clip.

Bariatric surgery is associated with anastomotic leakage causing significant morbidity and mortality.25 While postoperative leaks occurs in 5.2% of Roux-en-Y gastric bypass26 patients, it can be as high as 9% in primary sleeve gastrectomy and 13% in patients undergoing revision surgeries.27–29 Leaks can occur at any site after bariatric surgery30 and although it occurs in only 5% at the jejuno-jejunal anastomosis, mortality rate can be as high as 40% which is substantially more compared to other sites.31,32 Due to increased morbidity and mortality and high conversion rate to open surgery with surgical management of complications,3–7 endoscopic management has continued to expand and endoscopic clips play an important part in management of postoperative leaks associated with bariatric surgery. Leak site may be localized prior to endoscopic therapy using a bubble test and confirmed with methylene blue injection.33 Leaks occurring in the proximal pouch respond to stent placement while for leaks in the Roux limb or distal bowel, clips are a better choice. OTSC device has been used for successful closure of larger leaks up to 20 mm in size.15 OTSC closure of fistulae has been demonstrated in multiple small studies and case series12,34–37; however, the largest series was an international multicenter study38 which included 188 patients (108 patients with fistulae, 48 patients with perforation and 32 patients with leaks) and measured long term success of the procedure for closure of GI defects. The rate of successful closure was much higher for perforation and leaks (90% and 73.3%, respectively) compared to fistula (42.9%, P < 0.05) and durable success was achieved in 60.2% of patients (median follow up of 146 days). When used as primary therapy, OTSC was more effective than when used as rescue therapy (69.1% vs 46.9%; P = 0.004). On univariate analysis, there were three predictors for long term success (1) type of defect (perforation and leaks compared to fistulae; odds ratio [OR], 35.08; P < 0.001 and OR, 5.36; P = 0.02, respectively), (2) primary vs rescue therapy (OR, 3.10; P = 0.003), and (3) chronicity of defects (≤ 30 days vs > 30 days; OR, 3.84; P < 0.0001). However, on multivariate analysis, only the type of defect continued to be significant (perforation and leaks compared to fistulae; OR, 51.4; P = 0.002 and OR, 8.36; P = 0.002, respectively). However, the study included patients with GI defects of all etiologies including bariatric surgery (20.4% fistulae and 28.1% leaks) and subgroup analysis was not performed.

The data for endoscopic clip closure for small bowel perforation is limited due to limitation of access with traditional endoscopy and most of the data on duodenal perforation is related to endoscopic retrograde cholangiopancreatography (ERCP). Perforation from ERCP could be either free wall perforation by the endoscope causing intraperitoneal leakage or a retroperitoneal perforation from a sphincterotomy. Successful treatment of such perforation with clips has been reported.39 Patient should be positioned appropriately so that the fluid pool is away from perforation thereby minimizing peritoneal contamination40 and perforation closed by placing clips on the superior portion of sphincterotomy thereby avoiding pancreatic and bile duct orifice injury.41

Anastomotic leak after colon surgery is a serious complication with incidence of 3% and mortality rate of 10.1%.42 Perforations less that 10 mm can be closed successfully with TTSC; however, larger defects might require OTSC.36,43–45 Although colon perforations have been successfully managed with OTSC, it was unsuccessful in the management of colon fistula.46 A subsequent small case series showed successful closure of colon fistula in 2 out of 3 patients using the OTSC clip.15 There is limited role for OTSC in large perforations with everted edges.47 Alternate methods like using tissue grasper with the OTSC or an overstitch device to approximate the everted edges can be employed.

Endoluminal Stents

Endoluminal stent placement is well established in management of fistula and perforation. Stent excludes the defect by interposing between the defect and lumen thereby promoting healing, minimizing risk of stricture formation and facilitate enteral feeding over parenteral nutrition. Stents are placed using a forward viewing scope and under fluoroscopic guidance with patient in supine position. The leak site, upper esophageal sphincter (UES) and gastroesophageal junction (GEJ) should be marked externally with radiopaque markers (paper clips). Stent length should be chosen appropriately so that the proximal end does not approach UES to avoid globus sensation and distal end does not impact gastric wall to avoid ulceration and bleeding. A measuring wire or a catheter with radiopaque marker at the tip may be used to estimate the length of the stent needed. The endoscope is advanced into the duodenum and guidewire is placed. Endoscope is then removed and the stent delivery system is introduced over the guidewire under fluoroscopy. After deployment, stent position can be confirmed with endoscope and readjusted if needed with a rat-tooth forceps. A stiff guidewire may be used to traverse acute angulation if needed.

Temporary placement (6–8 weeks) of a fully covered self-expanding metal stent (FCSEMS) or partially covered self-expanding metal stent (PCSEMS) or self-expanding plastic stent (SEPS) placement is effective in treating anastomotic leaks following gastro-esophageal and bariatric surgery.48–51 SEPSs can be used for managing fistula and leaks; however, they have they have lower friction co-efficient with the mucosa than metal stents,52 therefore have high incidence of migration33 and has fallen out of favor in the US.

Stents were initially designed to be placed in the esophagus for management of leaks. Postoperative leaks after esophagectomy and gastrectomy occurs in about 7%–8% of cases when done for malignancy18,53 with mortality approaching over 50% if diagnosis is delayed.18,33,53 First successful use of esophageal stents in patients with leaks following esophagectomy and gastrectomy was demonstrated by Langer et al54 by using Polyflex SEPS (Boston Scientific, Marlborough, MA, USA). Although initial success rate was good, stent migration occurred in about 50% of the patients. The use of SEPS grew steadily until 2010 when a number of studies reported successful use of metal stents.55–58 Since then, the use of SEPS declined not only due to high migration rate but other factors like need for pre-deployment assembly, stiffness of the stent and large size delivery system.59 Bakken et al55 was one of the first to report FCSEMS (AliMaxx-E; Alveolus, Charlotte, NC, USA) for benign esophageal leaks. The rate of stent migration was lower than SEPS at 28% and all stents were successfully retrieved. Of the 22 patients who had stents placed for esophageal leak, 32% achieved short term closure of defect but long term data was lacking. Senousy et al56 used the same stent and reported 86% short term success rate; however, the sample size was very small with only 7 patients. Buscaglia et al58 reported that with a different FCSEMS (Wallflex; Boston Scientific), the rate of stent migration was comparable to other studies and short term success was 80%. Recently, a more comprehensive study of 54 patients was published by El Hajj et al.60 Forty-four patients received treatment for fistula or leak and 10 for acute perforation, multiple stent types including SEPS, PCSEMS, and FCSEMS were used. Stent migration was 28%, primary closure (resolution of defect after index stent removal) was achieved in 74% of patients, including 10 patients with acute perforation and secondary closure (repeat endoscopic stenting performed after index stent removal) was achieved in 83% of patients.

Stents have been used in successful management of leaks and fistulae following bariatric surgery as well and supported by considerable evidence. Stent placement facilitates resuming enteral nutrition and prevents peritoneal contamination thereby decreasing abdominal pain and infection.26 In a meta-analysis done by Puli et al61 for treatment of acute leaks following bariatric surgery using stent placement (both SEMS and SEPS), the rate of complete radiological resolution of leak was 87.8%. Stents were left in place for 4–8 weeks. Overall stent migration was 16.9% and 9% received revision surgery. Stent placement is effective for leaks after sleeve gastrectomy as well.62,63

Stent placement for leaks in the lower GI tract in a challenging endeavor. The increased motility of lower GI tract causes stent migration both distally and proximally.59 This not only prevents fistula closure but causes obstruction, perforation and bleeding. Further, most commercially available stents are not delivered by through the scope technique which needs a wire guided placement under fluoroscopy which is quite challenging in the left colon. For these reasons, successful colonic stent placement is less commonly reported. Recently, successful closure of colonic fistula was demonstrated using biodegradable esophageal stent64; however, it is not available for clinical use in the US.

The major disadvantages of stents are stent migration and tissue ingrowth or overgrowth especially when left for a long time. Stent migration rate is higher with SEPS and FCSEMS while tissue ingrowth is common in PCSEMS. Tissue ingrowth is caused by local fibrosis with granulation tissue formation which can occur as early as 2 weeks and reduce stent migration and provide better watertight seal between defect and luminal contents facilitating fistula healing65 but causes difficulty in stent removal.66 In these cases, a large diameter SEPS or FCSEMS can be placed within the impacted stent which causes pressure necrosis of the ingrown tissue and the stents can be removed together later.67 Argon plasma coagulation can be used to fulgurate the tissue ingrowth and once the stent borders are freed, it can be removed by conventional way.33 Additional strategies to prevent migration include clipping the proximal end of the stent to the mucosa.68 Recent studies have favored endoscopic suturing to secure the stent to mucosa (Fig. 2).69,70 Six to eight weeks is the optimal time to remove the stent.2 A shorter interval can cause incomplete mucosal closure and longer interval can cause stent migration and tissue ingrowth. Prompt stent placement is of utmost importance since 100% success rate can be achieved with early intervention compared to 50% if stent placed a month after perforation.65

Figure 2

(A) Esophago-jejunal anastomotic dehiscence. (B) Fully covered self-expanding metal stent fixed with suture. (C) Stent on fluoroscopy.

Endoscopic Suturing Device

Endoscopic suturing once considered experimental has evolved considerably and is rapidly emerging as a tool of choice for gastroenterologists for a wide array of conditions. Although early generation endoscopic suturing device (Bard EndoCinch; CR Bard, Murray Hill, NJ, USA) offered excellent initial success rate, long term results were modest71 probably from the lack of depth of suture from the device. Overcash72 attempted closure of gastric leaks using StomaphyX suturing system (EndoGastric Solutions, Redmond, WA, USA) with similar results. T-shaped tissue anchors (TAS; Ethicon Endo Surgery, Cincinnati, OH, USA) were used to close gastro-gastric fistulae by Spaun et al73 and though 100% primary closure was obtained initially, durable closure at the end of 6 months could not be achieved in any of the patients. Apollo Overstitch (Apollo Endosurgery, Austin, TX, USA) offers full thickness plication and is the only U.S. Food and Drug Administration (FDA) approved endoscopic suture device currently available in the US. It is a disposable, single unit device which is attached to dual channel therapeutic endoscope and contains a tissue anchor, curved suturing arm and a cinch which enables intermittent or continuous suture application (Fig. 3). Watson and Thompson74 achieved durable closure of gastro-gastric fistula in 3 of their 7 patients using Overstitch device. Since the FDA approval in 2011, Overstitch has been successfully used in closure of iatrogenic esophageal perforations,75 stent fixation,69 endoscopic submucosal dissection (ESD) mucosal defect closure,76 mucosal defect closure after Peroral Endoscopic Myotomy (POEM)77 and stoma reduction post bariatric surgery.78,79

Figure 3

(A) Jejunal perforation with visible drain. (B) Closure with Overstitch device. (C) Perforation closed with suture with Overstitch.

Endoscopic Sealant

Endoscopic tissue adhesive has been successfully used in closure of leaks and fistulae associated with esophageal and bariatric surgery patients. Fibrin tissue adhesive applied under endoscopic guidance was used to close anastomotic dehiscence in patients with high-output entero-cutaneous fistula over 25 years ago.80 Since then, several reports and series have been published and both fibrin glue and cyanoacrylate have been successfully used in this regard.81–83 Cyanoacrylate has been associated with inflammation and tissue necrosis; however, it is not deactivated by gastric or pancreatic secretions.84,85 Sealants are mostly used in anastomotic leakage after esophagectomy and gastrectomy or after bariatric surgery59; however, it has also been used in moderate success to close fistulae associated with pancreatectomy in 8 of 12 patients.86 The mucosa around the fistula is typically de-epithelialized using cytology brush or APC at low power setting so that after the site is plugged, reactive inflammatory response occurs promoting complete sealing of the defect.59 Fibrin glue is formed of two components, glue and thrombin which form fibrin coagulum immediately upon contact. A double-lumen catheter is used through the endoscope for application as the components of the sealant react and form bond quickly upon contact and the wider channel is used for more viscous component.33 Rapid exchange/shortwire biliary catheters should be avoided since the sealant may leak out through the sides of the catheter and damage the scopes.

Pramateftakis et al81 successfully used N-butyl-2-cyano-acrylate (Histoacryl; B. Braun Dexon GmbH, Spangenberg, Germany) to close an esophago-jejunal anastomotic leak 3 weeks after surgery when conservative treatment failed. Fibrin glue has been used to close esophageal fistula through submucosal injection until the fistula closes82 and with this method, abrasion or de-epithelization of the fistula tract is not needed. Larger leaks and fistulae have been closed with good long term success using Vicryl mesh (Ethicon, Hamburg, Germany) combined with fibrin sealant.87 The mesh is placed over the defect and covered with 2–3 mL fibrin (Tissucol Duo, Baxter, Germany) and fibrin is injected on the edges which eventually epithelializes with time. Rábago et al88 reported a high success rate of 86.6% closure of fistulae with sealant after failure of conservative treatment at a favorable cost; however, only 55% of high-output fistulae achieved complete closure of defect. Wong et al89 demonstrated successful close of fistulae in all of his 9 patients by using 5 mm choledochoscope to perform fistuloscopy with irrigation and debridement and using fibrin glue without recurrence at 12 months. Multimodality treatment in combination of sealant with clips or stents appear to be more successful.34,83 In a prospective study of 27 patients with fistulae after bariatric surgery entirely managed endoscopically using clips, sealant and stent either alone or in combination, 41% of patient achieved complete closure of defect after first procedure. Eventually, all patients were treated successfully after a mean of 4.4 endoscopes at a mean of 86 days.84 Lippert et al90 demonstrated in a series of 52 patients 36.5% success rate using fibrin glue alone compared to 55.7% in multimodal endoscopic therapy.

Endoscopic Vacuum Assisted Closure Device

Vacuum assisted closure (VAC) device is an established treatment for management of cutaneous wounds.91 It works by applying negative pressure to the wound through a vacuum sealed sponge thereby promoting drainage of wound secretion, increasing vascularity, reducing tissue edema and promoting granulation which aids wound healing. It was introduced in 1990 and since then, its application has steadily increased. Endoscopic VAC (E-VAC) was successfully used in leakage of rectal anastomosis92–94 and has been applied to endoscopically accessible upper GI tract leaks. E-VAC is performed by placing polyurethane sponge into the leakage site and application of continuous negative pressure through transnasal gastric tube which is connected to the sponge thereby draining inflammatory fluid and promoting granulation and healing.95 The wound cavities were endoscopically debrided before placing the E-VAC device.

Esophageal leakage after esophagectomy varies widely from 1% to 30%96,97 and accounts for about 40% of all postoperative mortalities.98 Negative intrathoracic pressure can draw fluid through smallest of the defects with each inspiration and infection can develop.98 E-VAC occludes the defect and continuously suctions the secretions. It was first successfully demonstrated by Wedemeyer et al99 and Loske and Müller100,101 in treatment of esophageal leaks. Wedemeyer et al99 used a nasogastric tube (4 mm; Unomedical, Birkerod, Denmark) which was introduced through the nose and exteriorized through mouth. The distal end was inserted in a polyurethane foam (sponge, pore size 400–600 μm; KCL, Wiesbaden, Germany) after creating two side openings and fixed with a suture. The sponge was trimmed based on the wound size assessed endoscopically, grasped with grasping forceps and placed in the necrotic cavity endoscopically and connected to 125 mmHg of continuous suction. The sponge was changed twice a week. All authors have used the same technique with very little variation since then. Wedemeyer et al99 were the first to successfully demonstrate E-VAC assisted closure of esophageal leaks in 2 patients following esophagectomy and gastrectomy when surgical and conventional endoscopic interventions failed. Complete healing was achieved in both patients after a median of 5 endoscopies and median 15 days, no complication occurred. The same group later reported successful closure of major intrathoracic post-surgical leak in 7 of 8 patients (88%) after a median of 23 days and 7 endoscopies without major complications.102 Similarly, Ahrens et al97 reported E-VAC assisted closure of gastroesophageal leak in all 5 patients at a median 42 days and 9 procedures, 2 of them developed stenosis requiring endoscopic dilation. Recent study suggests E-VAC therapy may be superior to stent placement for management of esophageal leaks.103

About 3% of colonic surgery is complicated by anastomotic leak with an associated mortality of 10.1%.42 Glitsch et al92 demonstrated the first successful closure of rectal anastomosis leak in 16 of 17 patients, one patient ultimately needing Hartmann’s procedure. The mean closure time to closure was 53.1 days with 5.4 sponge exchanges and 10.7 endoscopies. Weidenhagen et al93 demonstrated complete healing of anastomotic leak from anterior resection of rectum in 28 of the 29 patients with median 11.4 endoscopic sessions and 34.4 days of Endovac therapy. van Koperen et al94 published a case series of 16 patients who had anastomotic leak after undergoing surgery for rectal cancer or ulcerative colitis. E-VAC therapy was successful in 6 of 8 patients (75%) with early intervention (within 6 weeks of surgery) compared to 3 of 8 patients (38%) when the treatment was started after 6 weeks; however, it was not clinically significant (P = 0.315).

Endoscopic Needle Knife Therapy

Sinus formation occurs as a complication in 2.8% to 8% of patients undergoing Ileal Pouch Anal Anastomosis (IPAA).104–106 Initial case reports of successful closure of chronic sinus complicating IPAA using a needle knife led to further larger studies.107,108 The largest case series of 65 patients with IPAA complicated by pouch sinus was reported by Wu et al.109 Though only 43.1% of patients achieved complete response after a median of 2 sessions at 1.1-year follow-up, 81.5% patients had functional pouch. Longer duration from colectomy to diagnosis of pouch sinus (OR, 0.85; 95% confidence interval [CI], 0.73–0.99; P = 0.033) and complex sinuses (OR, 0.17; 95% CI, 0.04–0.70; P = 0.014) were inversely associated with the healing of pouch sinuses on multivariate analysis and increased sessions of needle-knife therapy (OR, 1.36; 95% CI, 1.01–1.81; P = 0.041) was associated with improved the healing.

Novel Endoluminal Options

Surgisis Anal Fistula Plug (AFP; Cook Biotech, West Lafayette, IN, USA) was developed for anorectal fistula closure where they have better results compared to fibrin glue.110,111 Surgisis is an acellular biomaterial matrix obtained from porcine small intestinal submucosa which can stimulate fistula closure without inducing inflammatory response.112 Toussaint et al113 were the first to successfully use Surgisis AFP to close fistula associated with bariatric surgery. AFP was inserted into the fistula using “rendezvous procedure” which utilizes both percutaneous and endoscopic route. The endoscope is placed at the level of fistula and tract opacified with contrast, then a guidewire is inserted into the fistula with help of an ERCP catheter the other end of guidewire is grasped with a snare and exteriorized through the scope or percutaneously allowing oro-cutaneous access through the fistula tract. The tract is then abraded over the guidewire to facilitate complete closure. Next, the narrow end of the AFP is grasped with a snare over the guidewire and pulled into the fistula under endoscopic and fluoroscopic guidance with less than 1 cm of the thinner lumen on the luminal side and endoscopic loop or clips may be used to fix the plug to the mucosa. Multiple plugs may be placed adjacent to each other for larger defects. Four out of five patients achieved complete closure with 1 or 2 procedures and concomitant stents were used in 3 of 5 patients. Surgisis strips were used with reasonable success in patients with fistula post Roux-en-Y gastric bypass (RYGB).114 AFP was used in 5 patients and Surgisis strips in 20 patients. Successful closure was achieved in 30% patients after one session, 55% after 2 sessions and 15% required 3 sessions. Success rate was 100% with AFP and 75% with Surgisis strips.

Vicryl mesh (Ethicon) has been used in combination with fibrin glue (Tissucol Duo) for closure of larger postoperative leaks or fistulae.87 Vicryl mesh is made of Polyglactin 920 (90% glycolide and 10% L-lactide) with pores large enough to readily absorb fibrin glue. The mesh is placed over the defect and covered with 2–3 mL of fibrin and fibrin is injected into the surrounding submucosal layer which eventually epithelialize. Of 15 patients, 13 patients achieved durable closure. One patient failed immediately due to large direct broncho-esophageal fistula requiring surgery subsequently died. Percutaneous drain for abscess formation was needed on the other patient; however, the site healed spontaneously.

Cardiac septal occluder (Amplatzer Occluder; AGA Medical Corp., Plymoth, MN, USA) is a self-expanding, double umbrella shaped device made of nitinol mesh. A short connecting metal part connects the two umbrellas and compressed inside a loader catheter originally designed to occlude cardiac defects. Perretta et al115 successfully demonstrated its use to close gastrostomy site in a porcine model and first human use was reported by Repici et al.116 It was successfully used to close postoperative tracheoesophageal fistula in a patient who underwent esophagectomy for esophageal cancer and has failed endoscopic and surgical intervention. Under endoscopic and fluoroscopic guidance, a guide-wire is introduced through the fistula from esophageal side into the trachea and exteriorized orally so that both ends of guidewire come out through the mouth. Endoscope is removed with guide-wire intact and the occluder device is introduced over the guide-wire and then released, first on the tracheal side and then esophageal side. Immediate closure was demonstrated radiographically and persistent closure was demonstrated at the end of 8 months. Amplatzer Vascular Plug (AVP; St. Jude Medical, St. Paul, MN, USA) is a similar device that has been successfully used to achieve durable closure of postoperative esophago-bronchial fistula in a esophageal cancer resection patient.117

Conclusion

As the number of patients undergoing surgery and associated complications is on the rise, endoscopists should familiarize themselves with the altered anatomy and be cognizant of postoperative complications. Early recognition of these complications can lead to early interventions and better outcomes. Novel endoscopic tools and techniques have led to less invasive means of managing these complications leading to better clinical outcomes. The principles of management of these conditions include treating the infection with antibiotics and drainage of inflammatory fluid collections via external or internal drainage, debriding necrotic tissue and stimulating angiogenesis and tissue growth, diverting GI secretions away from the site of injury and relieving any downstream obstruction/high pressure zones using stents, and closure of the defects using established as well as novel techniques and tools. One technique or tool does not fit all conditions. Multiple factors including type of surgery, site of injury, chronicity of the injury and fistula output should be taken into account when determining the tools and techniques to be used. We propose an algorithm for the management of leaks (Fig. 4) and fistula (Fig. 5) based on the on the above principles. Further large prospective or randomized control trials addressing these specific issues are required to establish the safety and efficacy of different techniques and tools. A multidisciplinary approach is essential to formulate a treatment pathway for these complications which may require multiple procedures to achieve technical and clinical success. In conclusion, therapeutic endoscopy offers a safe and effective alternative for the management and treatment of postsurgical complications by following basic surgical principles and alleviates the need for high risk surgical repair, thereby improving patient outcomes.

Figure 4

Algorithm for management of leaks.

Figure 5

Algorithm for management of fistula.

Acknowledgments

Author contributions: Prathab Devaraj, preparation of the manuscript; Hemanth Gavini, preparation and critical revision of the manuscript for important intellectual content.

Article information

References