Closure of transmural defects in the gastrointestinal tract by methods other than clips and sutures

Techniques in Gastrointestinal Endoscopy 17 (2015) 141–150

Contents lists available at ScienceDirect

Techniques in Gastrointestinal Endoscopy journal homepage: www.techgiendoscopy.com/locate/tgie

Closure of transmural defects in the gastrointestinal tract by methods other than clips and sutures Juliana Yang, MD, David Lee, MD, MPH, Deepak Agrawal, MD, MPHn Department of Medicine, Division of Digestive and Liver Diseases, University of Texas Southwestern Medical Center, POB1, 520P, 5323 Harry Hines Blvd, Dallas, Texas 75390

a r t i c l e in fo

a b s t r a c t

Article history: Received 26 April 2015 Accepted 15 June 2015

Traditionally, the mainstay of therapy for transmural defects of the gastrointestinal tract has been surgical repair. However, in recent years, the spectrum of novel and innovative therapies available for the repair of such defects has been rapidly increasing, and patients now have a variety of nonsurgical options available to them for specific indications. In this article, we review the devices and techniques, other than clips and sutures, which have been developed for the closure of transmural defects. In this review, we include some well-known and commonly available interventions, such as tissue adhesives and endoscopic band ligation, as well as novel combinations of common techniques, such as the “clutching rose stems” technique and endoloop and endoclip closure. Additionally, we include a review of various innovative devices that have been explored, such as the AMPLATZER Septal Occluder, endoluminal vacuum therapy, and the T-tag tissue apposition system. Such devices and techniques represent a dynamic area of development currently, with many showing promising early results in treatment of transmural gastrointestinal defects. With further refinement, these devices and techniques may enter mainstream therapeutic use in the future. & 2015 Elsevier Inc. All rights reserved.

Keywords: Closure Transmural defect Fistula Perforation Tissue adhesives Fistula plugs Endoluminal vacuum therapy AMPLATZER Septal Occluder T-tag tissue apposition system

1. Definition and causes Transmural gastrointestinal (GI) defect is defined as a complete GI mucosal perforation resulting in a nonphysiological connection between the GI lumen and outermost layer of the GI wall due to natural, pathogenic, or iatrogenic causes. The greatest concern is the spillage of bacteria-laden GI contents into the sterile abdominal or thoracic cavity that may result in serious systemic inflammation and infection. Depending on the onset, perforations are classified as acute or chronic. Acute perforations are often unexpected and iatrogenic (eg, perforation during endoscopic dilation or polypectomy). It is reported that the incidence of iatrogenic endoscopic perforation is between 0.01% and 0.6% for diagnostic and between 0.6% and 5.5% for therapeutic endoscopies [1-5]. As we explore and expand the frontiers of endoscopic therapeutic options with techniques such as endoscopic mucosal resection (EMR), endoscopic submucosal dissection (ESD), peroral endoscopic myotomy, endoscopic fullthickness resection (EFTR), and others, the incidence of iatrogenic perforation is bound to increase [3,6-9]. Anastomotic leaks may complicate 5%-30% of esophageal anastomoses [10,11] and 5%-15% The author reports no direct financial interests that might pose a conflict of interest in connection with the submitted manuscript. n Corresponding author. E-mail address: [email protected] (D. Agrawal). http://dx.doi.org/10.1016/j.tgie.2015.06.005 0049-0172/& 2015 Elsevier Inc. All rights reserved.

of rectal anastomoses [12,13]. However, not all iatrogenic acute transmural defects are complications of the procedures. For example, endoscopic access from within the GI lumen to other structures or organs is established during placement of percutaneous enteral feeding tubes or endoscopic drainage of extraluminal cysts or abscesses. Acute perforations can be pathogenic in cases such as perforation of a diverticulum or a severely dilated colon. Chronic transmural defects occur when inflammation extends across the GI wall causing destruction of the tissue, resulting in fistula formation with connection of GI lumen to either skin or other lumen. Examples include fistulae in Crohn's disease and tracheoesophageal fistulae (TEFs) in esophageal malignancy. Owing to chronic inflammatory changes and scar tissue around the defect, intraluminal contents do not usually leak into surrounding structures such as the mediastinum or the peritoneum. Chronic cutaneous fistulae may develop at the site of external tube drainage, such as the formation of a gastrocutaneous fistula after removal of a gastric feeding tube or a pancreatocutaneous fistula after removal of tubes for drainage of pancreatic fluid collections.

2. Management of transmural defects Until a few years ago, surgical repair had been the mainstay of treatment of transmural defects, especially perforations and leaks.

142

J. Yang et al. / Techniques in Gastrointestinal Endoscopy 17 (2015) 141–150

However, surgical repair is associated with definite morbidity and mortality risks. With improvements in endoscopic technology— more flexible endoscopes, larger working channels, and innovative accessories—endoscopic methods of closure of transmural defects are being developed. Through-the-scope clips were the first commonly used method for closure of leaks and perforations. Since then, several devices and methods have been described such as ligation by endoloops and rubber bands, endoscopic suturing, sealing with fibrin glue (FG) or cyanoacrylate glue (CG), self-expanding metal stents, and more recently, over-the-scope-clipping device [4,6,14-16]. In this review, we discuss few novel techniques, other than clips and sutures (which have been reviewed in other articles) that can be used for closure of transmural defects. There are a few common requirements when endoscopy is performed for closure of transmural defects. First, there are no definite data yet, to recommend one technique over another, and the physicians should remain mindful of local expertise and technology and on-site surgical backup before attempting more challenging cases. The endoscopist may require fluoroscopy, as real-time imaging may facilitate identification and closure of transmural defects. Second, whenever possible, imaging studies should be obtained before endoscopy to determine the precise location and extent of leak. Upper GI contrast radiographies and fistulograms are preferred, as they may provide better information than the more commonly performed CT scans do. Third, carbon dioxide should be used for insufflation during endoscopy, as it is absorbed faster than room air is.

3. Tissue adhesives Tissue adhesives are compounds that bind to tissue and have been used for hemostasis, wound closure, or fistula repair. There are 4 categories of tissue adhesives: FG, CG, polyethylene glycol polymers, and albumin-based sealants, which differ in composition and the way they are used. Of these, only FG and CG have been used in endoscopy. FG has been described for closure of chronic fistulae and anastomotic leaks, whereas CG is primarily used for treatment of bleeding gastric varices. Endoscopic use of tissue adhesives has not been approved by the U.S. Food and Drug Administration and their use to close defects after EFTR has not been studied.

4. Fibrin glue Fibrin is the end product of the coagulation cascade and is formed when thrombin converts soluble fibrinogen to insoluble fibrin strands, which get cross-linked by Factor XIII. The basic principle of sealing a mucosal defect with fibrin is that the mixture of the 2 components (fibrinogen and thrombin) simulates the coagulation cascade in the fistulous tract while forming a matrix of fibrin. Fibrin then stimulates growth of fibroblasts, leading to scar formation, and fibrin is slowly replaced by collagen [17]. The technique of injecting FG involves endoscopic injection of fibrinogen and thrombin. If available, a double-lumen catheter is preferred, so that the 2 components meet only at the intended site (and not within the channel of the endoscope where they can aggregate). The larger lumen of the catheter should be reserved for the more viscous component. The rapid exchange or short-wire double-lumen endoscopic retrograde cholangiopancreatography (ERCP) catheters should not be used, because the sealant may leak out the sides of the catheter and damage the endoscope channel [16,18]. If a single-lumen catheter is used instead of a doublelumen probe for sealing, the catheter lumen should be flushed

between the instillation of fibrinogen and thrombin. The injection of the contents should begin distally from the orifice and then moved proximally to minimize fluid retention within the tract. Multiple endoscopic procedures (average: 2-5) with repeat injections may be required to achieve complete closure. The data on the success of FG in closing fistulae and leaks mainly come from small case series, case reports, and animal models. In a small, randomized controlled trial of 13 patients with persistent enterocutaneous fistulae, it was found that FG achieved closure of fistulae after a mean of 2 days compared with 13 days with conservative therapy (P o 0.01) [19]. In a retrospective series of 52 patients with GI fistulae or leaks, FG successfully closed fistulae as a sole therapy in 37% cases and as adjunct with other endoscopic modalities in 56% cases [20]. A median of 4 endoscopies per patient was needed to close the defects. However, 80% of the lesions in the colon and 50% involving the pancreas needed only 1 endoscopy, suggesting that these defects may be easier to close with FG. In a case series, FG was used in 5 patients with gastrocutaneous fistulae after gastrostomy tube removal [21]. The mean time to achieve total fistula closure was 7 days compared with 32 days in patients who were treated conservatively. Many other smaller case series or case reports have reported successful closure of enterocutaneous fistulae and gastrocutaneous fistulae after removal of feeding tubes or after bariatric surgeries [21-26] and esophageal [27] and duodenal perforations [28]. Efficacy of FG has also been studied for anal fistulae. In an RCT, patients with simple fistulae (low anal fistulae) and complex anal fistulae (high fistulae and low fistulae with compromised sphincters) were randomly assigned to either FG or conventional treatment (fistulotomy or seton insertion) [29]. The primary end point was fistula healing. No advantage was found for FG over fistulotomy for simple fistulae (50% vs 100%, P ¼ 0.06), but FG healed more complex fistulae than conventional treatment did (69% vs 13%, P ¼ 0.003). Another randomized trial comparing FG with observation only for patients with Crohn's disease with anal fistulae found higher closure rates in patients treated with FG (38% vs 16%, P ¼ 0.04) [30]. Prophylactic application of FG has been tried to prevent anastomotic leaks in high-risk patients, for example, esophageal resection and gastric bypass [31,32]. A recent systemic review suggested that the mechanism is likely due to mechanical sealing rather than improved healing per se [33]. FG has also been tried to prevent delayed perforation following ESD of superficial duodenal adenocarcinoma [34]. In this “tissue shielding” method, the authors first endoscopically placed small pieces of polyglycolic acid (NEOVEIL; Gunze Co., Kyoto, Japan) sheet over the mucosal defect and then used FG to fix the pieces together. Polyglycolic acid is an absorbable, suture enforcement material that has also been used to prevent pancreatic fistula formation after pancreaticoduodenectomy [35,36]. In a more recent study, the tissue shielding method was used in 45 patients after ESD of gastric lesions and in 41 historical controls. There was a significant difference in the post-ESD bleeding rate between the study group and the control group (6.7% vs 22%, P ¼ 0.04) [37].

5. Cyanoacrylate glue Cyanoacrylates are a class of synthetic glues that rapidly solidify on contact with weak bases, such as water and blood. The different kinds of CGs available for medical use have been reviewed by American Society for Gastrointestinal Endoscopy [17]. CGs were initially used for skin approximation and wound closure, but they have been used off-label in the United States to treat bleeding varices and less commonly, in the closure of

J. Yang et al. / Techniques in Gastrointestinal Endoscopy 17 (2015) 141–150

pancreatic, biliary, and GI fistulae. Use of CGs has not been described for acute perforations. The technique of CG injection is similar to injection of FG and has been described elsewhere [16,17,38]. A major concern with injection of CGs is inadvertent injection and solidification of glue within the working channel of the endoscope (much more so compared with FG). If this occurs, a cleaning brush usually cannot flush out the blockage, and the endoscope has to be sent out for costly repairs. Certain steps are advised to prevent this from happening. First, the injecting catheter or device should be completely outside the endoscope during delivery. Second, the catheter should be flushed with saline before pulling it back into the scope. Some advocate pulling the entire endoscope out (with the tip of catheter still exposed outside the endoscope) and cutting off the tip. Before injection of the glue, silicone oil or lipiodol is sometimes used to coat the tip and working channel of the endoscope to minimize the risk of glue adherence. Mixing CGs with lipiodol slows the rate of solidification, making injection easier but increasing the risk of glue migrating from the area of interest before it solidifies. Some cyanoacrylates (eg, Dermabond, Ethicon, Somerville, NJ, and GLUBRAN, Aspide Medical, La Talaudière, France) polymerize more slowly and thus do not require the use of lipiodol [17]. For treatment of mucosal defects with glue (FG or CG), proper patient selection is important to ensure high success rate and to avoid complications and unnecessary procedures. Approximately one-third of enterocutaneous fistulae close spontaneously within 4-6 weeks with conservative treatment [39]. However, the rate of spontaneous closure diminishes rapidly with time. Factors that can decrease success of glue include large fistula orifice ( 41 cm), high fistula output ( 4200 mL/d), infection, abscess, malignant tumors, radiation, and Crohn's disease [22]. Duodenal fistulae tend to close

143

slowly because of proximity to the digestive enzymes. It is important to treat the infection, drain the abscesses, and provide patients with adequate nutritional support. Plugging of the fistula with glue alone is often not sufficient to close the fistula completely. The orifice or the defect, especially if large, may have to be sealed with clips, sutures, or meshes. For smaller defects, throughthe-scope clips and over-the-scope clips are good options, but for larger defects, meshes may be tried. In a case series, 9 patients with anastomotic leaks or fistulae after surgery for upper GI cancers were treated by first filling the entrance to the cavity by a VICRLY mesh (2  1.5 cm, Ethicon, Brussels, Belgium) and then injecting FG over it [40]. The anatomy of the fistula tract or the defect is also important. Glue after injection needs to be localized until it hardens. When injected into a short, straight tract, glue may flow across the tract to the other side or structure. Placing vascular plugs or coils within the tract first and then injecting fibrin can help localize the fibrin.

6. Fistula plugs Fistula plugs were first developed for the treatment of anorectal fistulae [41,42]. Toussaint et al [43] used fistula plugs to close enterocutaneous fistulae in 5 patients in whom stenting and FG injection had failed. The fistula plug they (and others) used was the Surgisis anal plug (Cook Biotech, Inc). The conical plug is made from porcine small bowel submucosa that is inherently resistant to infection, does not cause a foreign body reaction, and is readily populated by the host's own cells. The technique used is illustrated in Figure 1. A self-expandable metal stent was also placed in 3 patients. The leaks healed in 4 of 5 (80%) patients, of which 2 closed with a single procedure and 2 required a second

Fig. 1. Closure of a GI cutaneous fistula with a fistula plug. (1) A guidewire is first passed into the fistula tract. (2) Snare is used to pull the guidewire through the fistula. (3) The snare is pulled out from the fistula and secured around the fistula plug. (4) The snare is retracted back to the endoscope till the fistula plug fits snugly.

144

J. Yang et al. / Techniques in Gastrointestinal Endoscopy 17 (2015) 141–150

procedure. In another study, Maluf-Filho et al [44] used a combination of strips and fistula plugs to treat gastrocutaneous fistulae in 25 patients after Roux-en-Y gastric bypass. The plugs were deployed in 5 patients and the strips in 20 patients. Closure was achieved in all patients with plugs and in 75% of patients treated with strips. A case of successful closure of posttraumatic enterocutaneous fistulae using the Surgisis anal plug has also been described [45]. A fistula plug specifically designed for enterocutaneous fistulae is now available by Cook Biodesign Advanced Tissue Repair (Cook Medical Inc., Bloomington, IN). It is a porcine submucosa–derived fistula plug that is designed to be placed by interventional radiology using a percutaneous approach. The reported complete fistula closure rate is 21% [46]. The fistula plugs, with either approach, have less chance of success if the fistulous tract is more than 6 mm, the internal opening of the tract is large (allowing leakage of fluids), the fistula output is high (4 200 mL), or there is an infection or abscess that has not been treated. The technique is relatively straightforward, but data are limited to make definite conclusions.

7. Endoluminal vacuum therapy In recent years, endoscopic endoluminal vacuum therapy (EVT) has emerged as a new endoscopic treatment option in anastomotic leakages. The vacuum-assisted closure system device was introduced in 1990s as a treatment modality for large, infected cutaneous wounds. Negative pressure is applied to the wound with a vacuum-sealed sponge, resulting in drainage of wound secretions, improved blood flow, reduction of edema, promotion of granulation, and healing by secondary intention [47]. In 2007, Weidenhagen et al [48] reported EVT to treat anastomotic leakages after anterior rectum resections. Encouraged by the results, its use was then expanded to esophageal anastomotic leaks and then to all esophageal perforations. The technique involves an endoscopic insertion of polyurethane sponges across the perforation into the abscess. The sponge is cut to a size slightly smaller than the wound cavity to promote collapse and subsequent closure. If the perforation is small, the sponge is left in the esophageal lumen. The sponge is fixed to a transnasal gastric tube, which is then placed on controlled, continuous negative pressure (Figure 2). The sponge is usually downsized and replaced approximately twice a week, until the cavity appears to be clean and smaller than 1 cm in diameter. After

completion of EVT, patients should be followed up endoscopically once weekly until complete healing of the defect. A more complete discussion of this technique is described in detail elsewhere [47]. In a retrospective analysis, Brangewitz et al [49] compared 39 patients who were treated with stents and 32 patients treated with endoscopic EVT for intrathoracic leakage. The overall closure rate was significantly higher in the EVT group (84.4%) compared with the stents group (53.8%). In another retrospective study of 62 patients with anastomotic leaks after esophagectomy, endoscopic EVT was preferred to stents placement or surgery [50]. The overall success of closing the leaks by EVT in different studies has been reported to be 90% (84%-100%) [47]. Interestingly, although all the initial studies of this technique had been from different institutions in Germany, recently, Liu et al [51] from China reported successful closure of postsurgical esophageal leaks with endoscopic EVT in all 5 patients it was attempted on, suggesting that this procedure is gaining wider acceptance internationally. Although most available data have described the use of EVT for postsurgical esophageal leaks, it has also been used for the treatment of acute iatrogenic perforations. Kronsbein et al described successful closure of a 6-cm transmural esophageal perforation following pneumatic dilation. The sponge was changed on day 7, and complete closure was achieved by day 14 [52]. The advantage of this procedure over other techniques that simply close the defect (clips, sutures, or stents) is the ability to maintain continuous, active drainage of the area and thus control this potential septic focus. The disadvantage is the need for multiple endoscopic procedures—on average 4-6 procedures per patient. However, this depends on the size of the cavity, size of perforation, and duration since perforation [53,54]. The main complications associated with EVT are stenosis after completed therapy owing to scarring. There has also been a reported case of development of an aortic anastomotic fistula [55]. Endoscopic EVT is a promising, emerging technique, especially when leakages are associated with an infected cavity.

8. Endoscopic band ligation In endoscopic band ligation (EBL), GI wall layers are suctioned into a cap attached to the end of the endoscope, and a rubber band located outside the cap is slipped over the suctioned GI wall. Band ligation was first introduced in 1951 and was used mainly for treatment of internal hemorrhoids [56]. Since then, its use has

Fig. 2. Endoluminal vacuum therapy. (1) The size of the mucosal defect is estimated endoscopically. A sponge is cut to a size slightly smaller than the defect or cavity. The sponge is inserted into the defect under direct endoscopic visualization (through overtube or alongside the endoscope). (2) A transnasal gastric tube attached to the sponge is placed on controlled, continuous negative suction. (3) Follow-up endoscopies are performed to check for healing and downsizing the sponge, as the cavity closes and heals by secondary intention.

J. Yang et al. / Techniques in Gastrointestinal Endoscopy 17 (2015) 141–150

145

Fig. 3. The Clutching Rose Stem technique. (1) A mucosal defect or perforation is endoscopically visualized. (2) Clips are placed circumferentially around the defect, followed by placement of an endoloop at the base around the periphery of the clips. (3) The endoloop is tightened to bring the clips together and approximate the edges. Fibrin glue may be injected over the defect to ensure complete sealing of the defect. (Color version of figure is available online.)

been expanded to treatment of varices, bleeding lesions such as postpolypectomy bleeding [57,58], arteriovenous malformations [58,59], Dieulafoy lesions [60,61], and diverticular bleeding [62]. Band ligation EMR was described in the mid-1990s [63] and was quickly adapted into clinical practice because of ease of performing the procedure, low adverse events, and successful mucosal resections, especially for esophageal, gastric, and rectal lesions. Use of EBL to close perforations involving the stomach, duodenum, and colon has been described in many case reports. In fact, studies have shown that EBL can be successful where clips have failed. Han et al [64] used EBL to close gastric perforations after EMR or ESD in 5 patients where clip placement had failed owing to tangential location of the defect or fibrosis. EBL was successful in all 5 patients. In another report, rescue band ligation of a sigmoid colon perforation was similarly performed [65]. EBL has also been used to close duodenal perforations after polypectomy [66] or after ERCP [67]. EBL offers several technical advantages over clipping in the closure of GI perforations. The ligator cap in EBL enables perpendicular placement of the endoscope, which is especially helpful in tangentially placed defects. EBL pulls (suctions) the defect rather than pushing during clipping, so it is easier to maintain the position of the tip of the endoscope. The suctioning during EBL also enables closure of perforations with diameters even larger than that of the suction cap. Some reports describe using EBL for only partial closure of large perforations. This brings the edges of the perforation closer and makes subsequent placement of clips for complete closure of the defect easier [68]. Finally, most endoscopists are familiar with the EBL technique, and the required equipment is available in most endoscopic centers. Few questions remain about EBL for closure of perforations: is the closure of the defect transmural, involving all layers? After all, the success and safety of band ligation EMR depends on the fact that only the mucosa and submucosa get suctioned. If true, is suction of just the mucosa and the submucosa sufficient to close a perforation? There have been a few animal and ex vivo studies to determine the number of GI layers that are suctioned into the cap during EBL (in the absence of varices). In these studies, bands were placed in the duodenum, ileum, and colon, and histologic examination was performed immediately or after a few days [62,69]. Bands placed in the left colon appeared to involve only the mucosa and the submucosa, but the results were mixed for the duodenum, ileum, and right colon where walls may be thinner, although most studies reported suctioning of the muscularis propria and serosa. Theoretically, this can cause delayed perforation, as bowel layers trapped in the band become ischemic and prone to perforation when bands slough off in 3-7 days. However, animal studies suggest that even when the muscularis propria is suctioned into the band, it is replaced by granulation tissue; hence, perforation is unlikely [70]. In a recent animal study on the use of EBL for GI wall perforations, Law et al [71] found that healing of the perforation occurred by primary intention. This suggests that when the edges of a perforation are suctioned into the cap, it is likely that all the layers are suctioned and there is serosa-to-serosa apposition.

9. Clutching rose stems As the name suggests, after closure with this technique, the appearance is that of somebody clutching rose stems—endoclips being the rose stems and endoloop being the clutching device. The technique involves placing endoclips at the perimeter of the defect edge that are subsequently brought together with an endoloop placed around the perimeter base of the clips (Figure 3). Samarasena et al [72] used this technique to close a post-EMR perforation in the duodenum. The defect was too large to be closed by endoclips alone. FG was also injected at the center of the clips to secure the closure. Mocciaro et al [73] used this technique to close a 2.5-cm lesion with 3 discrete perforations in the sigmoid colon. Clips were placed over each perforation and then brought together with 2 endoloops. The 2 borders of perforation that did not close completely were sealed by injecting FG. This technique has been used to close large mucosal defects after EMR or ESD, up to 4 cm in diameter. The authors described this as “lucky loop” and “a tobacco-pouch” suture technique.

10. Endoloop and endoclip closure This technique has been used in perforations in the duodenum, ampulla, colon, and rectum. It is similar to the clutching rose stem technique described earlier, as it uses an endoloop and endoclips. The technique was first described by Endo in 2004 to close large mucosal defects in 32 patients after EMR of gastric mucosal tumors [74]. It involves using a 2-channel endoscope. Through one of the working channels, an open endoloop is placed on the perforation site, and through the second channel, the upper and the lower arms of endoloop were fixed on the lower and the upper margin of the perforation area with 2 endoclips. The endoloop is then tightened slowly to bring the perforation edges closer (Figure 4). Additional endoclips may be placed as necessary to achieve full closure. The perforation edges may be cauterized by argon plasma coagulation before placement of an endoloop to stimulate the inflammatory reaction and local collagen synthesis. Variations of this technique have been described, such as attaching multiple endoclips to the endoloop or using more than one endoloop [75]. A few case reports have described use of this technique to close perforations on the duodenal lateral wall during ERCP [76-78]. This technique has also been used to close a rectal perforation that occurred following retroflexion [79,80].

The 8-ring technique This technique by Fujii et al [81] uses a double metal ring that is shaped like the number “8.” This metal ring is designed to obviate the need for using a double-channel endoscope and an endoloop. Of the 2 rings of the 8-ring, 1 is grasped with an endoclip beforehand and loaded inside the sheath. The endoclip preloaded with the 8-ring is attached to the normal mucosa on 1 side of the

146

J. Yang et al. / Techniques in Gastrointestinal Endoscopy 17 (2015) 141–150

Fig. 4. Endoloop and Endoclip Closure. (1) An endoloop is placed over the peripheral edges of the defect. (2) Clips are placed over the endoloop and tightened slightly to approximate the clips and the edges. (3) A second endoloop is placed over the remaining defect. (4) Then, 2 more clips are placed to secure the second endoloop. (5) The second endoloop is tightened to completely close the defect. (Color version of figure is available online.)

resection site. The remaining ring of the 8-ring is then grasped with another endoclip and attached to the normal mucosa on the other side of the resection site to close the defect. More secure closure can be accomplished by clips placed between the first 2 clips. This technique was used to close mucosal defects after EMR and not for perforations. Furthermore, the mean size of the defect was 16 mm. Further studies are needed to assess the feasibility and efficacy of this method.

11. Cardiac septal defect occluders The AMPLATZER Septal Occluder (ASO) (AGA Medical Group, Plymouth, MN) is a device that was initially developed for the occlusion of cardiac septal defects, but it has been used off-label for the closure of GI fistulae. The ASO device consists of a nitinol wire mesh shaped into 2 umbrellas linked by a short metal connecting portion (Figure 5). The discs are covered by a polyester fabric, favoring the growth of tissue over the mesh. It is constrained within a 70-cm delivery catheter, which is too short to be passed through an endoscope. To place the ASO, a guidewire is placed endoscopically in the area of interest and the endoscope is removed. The

catheter is then passed over the guidewire under fluoroscopic guidance. An endoscope may be passed alongside the wire for direct visualization. After implantation, the device apposes the wall on each side of the defect, mechanically occluding it and potentially creating a platform for subsequent tissue ingrowth (Figure 5). The first reported use of an ASO device in the GI tract was likely by Rabenstein et al [82] to close a bronchoesophageal fistula in an elderly woman. They describe using an endobronchial approach to place the device, with one umbrella on the esophageal side and the other on the bronchial side. The patient's symptoms of chronic cough were relieved after ASO placement. The reported complication was an episode of hemoptysis that responded to conservative treatment. Notably, at 1-year follow-up, the esophageal umbrella was noted in the mediastinum below the esophageal wall, but the fistula remained closed. The endobronchial approach has been successfully used in a few other reports to close bronchopleural fistulae [83,84], tracheomediastinal fistulae [85], and bronchoesophageal fistulae [86]. Boulougouri et al [87] described a cutaneous approach to place an ASO to close a cutaneous-duodenal fistula. There was significant output through the externally placed drains during the first 8 days, but this gradually stopped in 30 days. Repici et al [88] reported the

Fig. 5. The AMPLATZER Septal Occluder (ASO) for a tracheoesophageal fistula. (1) A tracheoesophageal fistula is identified on endoscopy. A guidewire is passed into the fistula. (2) The ASO device catheter is passed over the guidewire. The endoscope is passed alongside the delivery catheter for endoscopic visualization. (3) The ASO device is deployed into the tracheoesophageal fistula, one umbrella on each side of the fistula opening.

J. Yang et al. / Techniques in Gastrointestinal Endoscopy 17 (2015) 141–150

first endoscopic approach to close a TEF at the esophagogastric anastomosis site. Clips, stents, and glue and surgical placement of a patch between the anastomosis and the trachea had failed to close the fistula. A Gastrografin study immediately following the procedure confirmed closure of the fistula. In a recent report, ASO placed endoscopically successfully closed a 13-mm jejunocutaneous fistula, which had developed following removal of a jejunal feeding tube [89]. Coppola et al [90] reported a different experience when they tried to use ASO to close a benign TEF, which was due to an accidental ingestion of dental amalgam. The TEF had been previously treated unsuccessfully by the placement of a plastic covered stent, endoclips, and FG. The ASO was placed with one umbrella on the esophageal side and another on the tracheal side. At 2 months later, the fistula orifice had enlarged and the ASO migrated into the bronchial tree and had to be removed from the middle lobe bronchus. Experiences such as this are rarely reported but are a useful reminder that caution should be exercised during off-label use of a device, and patients should be made aware of unexpected outcomes. The various case reports published suggest that when ASOs are placed, it may take a few weeks before complete closure is achieved. As compared with vascular applications in which the ventricular septal defect occluder is completely covered by an endothelial layer after a very short time, this cannot be expected in GI or bronchial cases. Hence, some studies chose to inject glue within the waist of an ASO device. The ASOs come in different sizes and lengths, and it is important to choose the size that corresponds to the size of the defect. A long fistula may not be conducive for closure with an ASO, as the waist on the device is short and as the device tries to assume its natural shape, it may retract and get displaced. An ASO may also migrate if the mucosal walls are fragile.

147

of a small metal bar, or “T-tag,” with an attached suture mounted on a loading system, which consists of a needle passed through the working channel of the endoscope. The needle is used to puncture the bowel wall adjacent to the defect, and a push of the stylet dislodges the T-tag from the needle. Similarly, a second T-tag is deployed on the opposite side of the defect. The strings attached to the T-tags are then pulled tight to close the defect. A knot-tying device is then passed over the strings and tightened to secure the T-tags in place (Figure 6). Agrawal et al [94] used T-tag Tissue Apposition System (Ethicon Endo-Surgery, Cincinnati, OH) to close large postpolypectomy mucosal defects in the colon and were successful in 87.5% (7 out of 8 patients). The average polyp size was 3.3, and hospital stay decreased by 1 day. The major challenge with this technique was the inability to visualize structures behind the bowel wall that was being punctured. A 19-gauge needle, with the loaded T-tag, did not puncture the colon wall readily and sometimes caused excessive “tenting” of the colon wall before giving way suddenly. Laparoscopy was therefore essential to ensure that there were no viscera or blood vessels in the needle path. Experimental studies using T-tags for GI perforation repair show that needle damage of an organ outside the GI wall is not a rare event and continues to be the limiting factor for adopting this technique in clinical practice [95-97]. Investigators have tried different techniques such as use of endoscopic ultrasound [95] and special chambers. As an example, in a porcine model, Hashiba et al [98] evaluated the use of a closed cap (chamber) fitted at the end of an endoscope to prevent the needle from being pushed too far across the GI wall. The chamber had a lateral window that allowed for the placement of the gastric wall inside (Figure 7). T-tags were then placed on opposite ends of the defect and the suture tightened. This technique has not been reported in humans.

12. T-tag tissue apposition system 13. Conclusion T-tags were first used in endoscopy to treat gastroesophageal reflux disease [91]. It was later described in natural orifice translumenal endoscopic surgery to close visceral defects [92,93]. The device consists

In this review, we have described devices and techniques, other than clips and sutures, for the closure of transmural defects in the

Fig. 6. The T-tag Tissue Apposition System. (1) A needle is used to puncture the mucosa. Concurrent laparoscopy is performed to confirm that the needle does not damage structures outside the bowel wall. (2) The T-tag is deployed on the serosa side. (3) The same procedure is duplicated on the opposite side of the defect. (4) Both sutures are pulled upward to approximate the defect. (5) A knot-tying device is used to secure the T-tags.

148

J. Yang et al. / Techniques in Gastrointestinal Endoscopy 17 (2015) 141–150

Fig. 7. Closed chamber modification for the T-tag Tissue Apposition System [97]. (1) A closed cap (chamber) fitted at the end of the endoscope. (2) A needle is used to puncture the mucosa on one side of the defect, with the closed cap limiting the depth of perforation. (3) T-tags are deployed on either side of the defect and pulled upward. (4) The T-tags are pulled upward to approximate the mucosal defect. (5) A knot-tying device is used to secure the T-tags. (Color version of figure is available online.)

GI tract. Most of these devices and techniques have not been described for closing defects after EFTR. The glues (fibrin and cyanoacrylate) work well for closure of fistulae, but additional measures such as closing the orifice are generally needed. EBL has been used successfully to close fistulae and perforations in the stomach, duodenum, and colon, but concerns remain if it achieves true full-thickness closure and the role of band ligation remains to be seen. Endoscopic vacuum therapy is a promising technique for large esophageal defects, especially when there is concern about localized infection and abscesses, but US experience is limited. Innovative techniques using endoclips, endoloops, and glues can be useful on a case-by-case basis. The current T-tag anchoring system is an interesting concept, but the delivery tools need further refinement. References [1] Changela K, Virk MA, Patel N, et al. Role of over the scope clips in the management of iatrogenic gastrointestinal perforations. World J Gastroenterol 2014;20:11460–2. [2] Singhal S, Changela K, Papafragkakis H, et al. Over the scope clip: technique and expanding clinical applications. J Clin Gastroenterol 2013;47:749–56. [3] Kojima T, Parra-Blanco A, Takahashi H, et al. Outcome of endoscopic mucosal resection for early gastric cancer: review of the Japanese literature. Gastrointest Endosc 1998;48:550–4. [4] Al Ghossaini N, Lucidarme D, Bulois P. Endoscopic treatment of iatrogenic gastrointestinal perforations: an overview. Dig Liver Dis 2014;46:195–203. [5] Iqbal CW, Cullinane DC, Schiller HJ, et al. Surgical management and outcomes of 165 colonoscopic perforations from a single institution. Arch Surg 2008;143:701–6. [6] Disibeyaz S, Koksal AS, Parlak E, et al. Endoscopic closure of gastrointestinal defects with an over-the-scope clip device. A case series and review of the literature. Clin Res Hepatol Gastroenterol 2012;36:614–21. [7] ASGE Technology Committee, Kantsevoy SV, Adler DG, et al. Endoscopic mucosal resection and endoscopic submucosal dissection. Gastrointest Endosc 2008;68:11–8. [8] Paspatis GA, Dumonceau JM, Barthet M, et al. Diagnosis and management of iatrogenic endoscopic perforations: European Society of Gastrointestinal Endoscopy (ESGE) Position Statement. Endoscopy 2014;46:693–711.

[9] Inoue H, Tianle KM, Ikeda H, et al. Peroral endoscopic myotomy for esophageal achalasia: technique, indication, and outcomes. Thorac Surg Clin 2011;21: 519–25. [10] Miller JD, Jain MK, de Gara CJ, et al. Effect of surgical experience on results of esophagectomy for esophageal carcinoma. J Surg Oncol 1997;65:20–1. [11] Lang H, Piso P, Stukenborg C, et al. Management and results of proximal anastomotic leaks in a series of 1114 total gastrectomies for gastric carcinoma. Eur J Surg Oncol 2000;26:168–71. [12] Nesbakken A, Nygaard K, Lunde OC, et al. Anastomotic leak following mesorectal excision for rectal cancer: true incidence and diagnostic challenges. Colorectal Dis 2005;7:576–81. [13] Lee WS, Yun SH, Roh YN, et al. Risk factors and clinical outcome for anastomotic leakage after total mesorectal excision for rectal cancer. World J Surg 2008;32:1124–9. [14] Raju GS. Endoscopic clip closure of gastrointestinal perforations, fistulae, and leaks. Dig Endosc 2014;26(Suppl 1):95–104. [15] Kirschniak A, Subotova N, Zieker D, et al. The Over-The-Scope Clip (OTSC) for the treatment of gastrointestinal bleeding, perforations, and fistulas. Surg Endosc 2011;25:2901–5. [16] Willingham FF, Buscaglia JM. Endoscopic management of gastrointestinal leaks and fistulae. Clin Gastroenterol Hepatol 2015. http://dx.doi.org/10.1016/ j.cgh.2015.02.010. [17] ASGE Technology Committee, Bhat YM, Banerjee S, et al. Tissue adhesives: cyanoacrylate glue and fibrin sealant. Gastrointest Endosc 2013;78:209–15. [18] Kumar N, Thompson CC. Endoscopic therapy for postoperative leaks and fistulae. Gastrointest Endosc Clin N Am 2013;23:123–36. [19] Hwang TL, Chen MF. Randomized trial of fibrin tissue glue for low output enterocutaneous fistula. Br J Surg 1996;83:112. [20] Lippert E, Klebl FH, Schweller F, et al. Fibrin glue in the endoscopic treatment of fistulae and anastomotic leakages of the gastrointestinal tract. Int J Colorectal Dis 2011;26:303–11. [21] Gonzalez-Ojeda A, Avalos-Gonzalez J, Mucino-Hernandez MI, et al. Fibrin glue as adjuvant treatment for gastrocutaneous fistula after gastrostomy tube removal. Endoscopy 2004;36:337–41. [22] Rabago LR, Ventosa N, Castro JL, et al. Endoscopic treatment of postoperative fistulas resistant to conservative management using biological fibrin glue. Endoscopy 2002;34:632–8. [23] Barthelemy C, Audigier JC, Fraisse H. A non-tumoral esophago-bronchial fistula managed by isobutyl-2-cyanoacrylate. Endoscopy 1983;15:357–8. [24] Papavramidis ST, Eleftheriadis EE, Papavramidis TS, et al. Endoscopic management of gastrocutaneous fistula after bariatric surgery by using a fibrin sealant. Gastrointest Endosc 2004;59:296–300. [25] Papavramidis TS, Kotzampassi K, Kotidis E, et al. Endoscopic fibrin sealing of gastrocutaneous fistulas after sleeve gastrectomy and biliopancreatic diversion with duodenal switch. J Gastroenterol Hepatol 2008;23:1802–5.

J. Yang et al. / Techniques in Gastrointestinal Endoscopy 17 (2015) 141–150 [26] Cellier C, Landi B, Faye A, et al. Upper gastrointestinal tract fistulae: endoscopic obliteration with fibrin sealant. Gastrointest Endosc 1996;44:731–3. [27] Rabago LR, Castro JL, Joya D, et al. Esophageal perforation and postoperative fistulae of the upper digestive tract treated endoscopically with the application of Tissucol. Gastroenterol Hepatol 2000;23:82–6. [28] Mutignani M, Iacopini F, Dokas S, et al. Successful endoscopic closure of a lateral duodenal perforation at ERCP with fibrin glue. Gastrointest Endosc 2006;63:725–7. [29] Lindsey I, Smilgin-Humphreys MM, Cunningham C, et al. A randomized, controlled trial of fibrin glue vs. conventional treatment for anal fistula. Dis Colon Rectum 2002;45:1608–15. [30] Grimaud JC, Munoz-Bongrand N, Siproudhis L, et al. Fibrin glue is effective healing perianal fistulas in patients with Crohn's disease. Gastroenterology 2010;138:2275–81 [2281 e1]. [31] Sapala JA, Wood MH, Schuhknecht MP. Anastomotic leak prophylaxis using a vapor-heated fibrin sealant: report on 738 gastric bypass patients. Obes Surg 2004;14:35–42. [32] Silecchia G, Boru CE, Mouiel J, et al. Clinical evaluation of fibrin glue in the prevention of anastomotic leak and internal hernia after laparoscopic gastric bypass: preliminary results of a prospective, randomized multicenter trial. Obes Surg 2006;16:125–31. [33] Nordentoft T, Pommergaard HC, Rosenberg J, et al. Fibrin glue does not improve healing of gastrointestinal anastomoses: a systematic review. Eur Surg Res 2015;54:1–13. [34] Doyama H, Tominaga K, Yoshida N, et al. Endoscopic tissue shielding with polyglycolic acid sheets, fibrin glue and clips to prevent delayed perforation after duodenal endoscopic resection. Digestive endoscopy 2014;26(Suppl. 2):41–5. [35] Hiura Y, Takiguchi S, Yamamoto K, et al. Use of fibrin glue sealant with polyglycolic acid sheets to prevent pancreatic fistula formation after laparoscopic-assisted gastrectomy. Surg Today 2013;43:527–33. [36] Uemura K, Murakami Y, Hayashidani Y, et al. Combination of polyglicolic acid felt and fibrin glue for prevention of pancreatic fistula following pancreaticoduodenectomy. Hepatogastroenterology 2009;56:1538–41. [37] Tsuji Y, Fujishiro M, Kodashima S, et al. Polyglycolic acid sheets and fibrin glue decrease the risk of bleeding after endoscopic submucosal dissection of gastric neoplasms (with video). Gastrointest Endosc 2015;81:906–12. [38] Seewald S, Sriram PV, Naga M, et al. Cyanoacrylate glue in gastric variceal bleeding. Endoscopy 2002;34:926–32. [39] Visschers RG, van Gemert WG, Winkens B, et al. Guided treatment improves outcome of patients with enterocutaneous fistulas. World J Surg 2012;36: 2341–8. [40] Truong S, Bohm G, Klinge U, et al. Results after endoscopic treatment of postoperative upper gastrointestinal fistulas and leaks using combined Vicryl plug and fibrin glue. Surg Endosc 2004;18:1105–8. [41] Champagne BJ, O'Connor LM, Ferguson M, et al. Efficacy of anal fistula plug in closure of cryptoglandular fistulas: long-term follow-up. Dis Colon Rectum 2006;49:1817–21. [42] Johnson EK, Gaw JU, Armstrong DN. Efficacy of anal fistula plug vs. fibrin glue in closure of anorectal fistulas. Dis Colon Rectum 2006;49:371–6. [43] Toussaint E, Eisendrath P, Kwan V, et al. Endoscopic treatment of postoperative enterocutaneous fistulas after bariatric surgery with the use of a fistula plug: report of five cases. Endoscopy 2009;41:560–3. [44] Maluf-Filho F, Hondo F, Halwan B, et al. Endoscopic treatment of Roux-en-Y gastric bypass-related gastrocutaneous fistulas using a novel biomaterial. Surgical Endoscopy 2009;23:1541–5. [45] Satya R, Satya RJ. Successful treatment of an enterocutaneous fistula with an anal fistula plug after an abdominal stab wound. J Vasc Interv Radiol 2010;21: 414–5. [46] Piduru SM, Yamada K, Morales AF. Enterocutaneous fistula closure with porcine submucosa derived fistula plug: a novel minimally invasive closure technique. J Vasc Interv Radiol 2013;24:S145–6. [47] Mennigen R, Senninger N, Laukoetter MG. Novel treatment options for perforations of the upper gastrointestinal tract: endoscopic vacuum therapy and over-the-scope clips. World J Gastroenterol 2014;20:7767–76. [48] Weidenhagen R, Gruetzner KU, Wiecken T, et al. Endoscopic vacuum-assisted closure of anastomotic leakage following anterior resection of the rectum: a new method. Surg Endosc 2008;22:1818–25. [49] Brangewitz M, Voigtlander T, Helfritz FA, et al. Endoscopic closure of esophageal intrathoracic leaks: stent versus endoscopic vacuum-assisted closure, a retrospective analysis. Endoscopy 2013;45:433–8. [50] Schniewind B, Schafmayer C, Voehrs G, et al. Endoscopic endoluminal vacuum therapy is superior to other regimens in managing anastomotic leakage after esophagectomy: a comparative retrospective study. Surg Endosc 2013;27: 3883–90. [51] Liu YN, Yan Y, Li SJ, et al. Reliable management of post-esophagectomy anastomotic fistula with endoscopic trans-fistula negative pressure drainage. World J Surg Oncol 2014;12:240. [52] Kronsbein H, Etzold M, Fein M, et al. Endoscopic vacuum therapy for acute esophageal perforation following pneumatic dilation. Endoscopy 2014;46 (Suppl. 1):E485–6. [53] Loske G, Schorsch T, Muller C. Endoscopic vacuum sponge therapy for esophageal defects. Surg Endosc 2010;24:2531–5. [54] Kuehn F, Schiffmann L, Rau BM, et al. Surgical endoscopic vacuum therapy for anastomotic leakage and perforation of the upper gastrointestinal tract. J Gastrointest Surg 2012;16:2145–50.

149

[55] Ahrens M, Schulte T, Egberts J, et al. Drainage of esophageal leakage using endoscopic vacuum therapy: a prospective pilot study. Endoscopy 2010;42:693–8. [56] Trowers EA, Ganga U, Rizk R, et al. Endoscopic hemorrhoidal ligation: preliminary clinical experience. Gastrointest Endosc 1998;48:49–52. [57] Smith RE, Doull J. Treatment of colonic post-polypectomy bleeding site by endoscopic band ligation. Gastrointest Endosc 1994;40:499–500. [58] Abi-Hanna D, Williams SJ, Gillespie PE, et al. Endoscopic band ligation for nonvariceal non-ulcer gastrointestinal hemorrhage. Gastrointest Endosc 1998; 48:510–4. [59] Matsui S, Kamisako T, Kudo M, et al. Endoscopic band ligation for control of nonvariceal upper GI hemorrhage: comparison with bipolar electrocoagulation. Gastrointest Endosc 2002;55:214–8. [60] Norton ID, Petersen BT, Sorbi D, et al. Management and long-term prognosis of Dieulafoy lesion. Gastrointest Endosc 1999;50:762–7. [61] Valera JM, Pino RQ, Poniachik J, et al. Endoscopic band ligation of bleeding dieulafoy lesions: the best therapeutic strategy. Endoscopy 2006;38:193–4. [62] Farrell JJ, Graeme-Cook F, Kelsey PB. Treatment of bleeding colonic diverticula by endoscopic band ligation: an in-vivo and ex-vivo pilot study. Endoscopy 2003;35:823–9. [63] Fleischer DE, Wang GQ, Dawsey S, et al. Tissue band ligation followed by snare resection (band and snare): a new technique for tissue acquisition in the esophagus. Gastrointest Endosc 1996;44:68–72. [64] Han JH, Lee TH, Jung Y, et al. Rescue endoscopic band ligation of iatrogenic gastric perforations following failed endoclip closure. World J Gastroenterol 2013;19:955–9. [65] Han JH, Park S, Youn S. Endoscopic closure of colon perforation with band ligation; salvage technique after endoclip failure. Clin Gastroenterol Hepatol 2011;9:e54–5. [66] Fan CS, Soon MS. Repair of a polypectomy-induced duodenal perforation with a combination of hemoclip and band ligation. Gastrointest Endosc 2007; 66:203–5. [67] Li Y, Han Z, Zhang W, et al. Successful closure of lateral duodenal perforation by endoscopic band ligation after endoscopic clipping failure. Am J Gastroenterol 2014;109:293–5. [68] Ji JS, Cho YS. Endoscopic band ligation: beyond prevention and management of gastroesophageal varices. World J Gastroenterol 2013;19:4271–6. [69] Barker KB, Arnold HL, Fillman EP, et al. Safety of band ligator use in the small bowel and the colon. Gastrointest Endosc 2005;62:224–7. [70] Akimaru K, Suzuki H, Tsuruta H, et al. Eversion and ligation of a diverticulum: report of an inspirational case and subsequent animal study. J Nippon Med Sch 2008;75:157–61. [71] Law R, Deters JL, Miller CA, et al. Endoscopic band ligation for closure of GI perforations in a porcine animal model (with video). Gastrointest Endosc 2014;80:717–22. [72] Samarasena JB, Nakai Y, Park DH, et al. Endoscopic closure of an iatrogenic duodenal perforation: a novel technique using endoclips, endoloop, and fibrin glue. Endoscopy 2012; 44 Suppl. 2 e424-e425. [73] Mocciaro F, Curcio G, Tarantino I, et al. Tulip bundle technique and fibrin glue injection: unusual treatment of colonic perforation. World J Gastroenterol 2011;17:1088–90. [74] Endo M, Inomata M, Terui T, et al. New endoscopic technique to close large mucosal defects after endoscopic mucosal resection in patients with gastric mucosal tumors. Dig Endosc 2004;16:372–5. [75] Lee TH, Han JH, Park SH. Endoscopic treatments of endoscopic retrograde cholangiopancreatography-related duodenal perforations. Clin Endosc 2013;46:522–8. [76] Nakagawa Y, Nagai T, Soma W, et al. Endoscopic closure of a large ERCPrelated lateral duodenal perforation by using endoloops and endoclips. Gastrointest Endosc 2010;72:216–7. [77] Kim DH, Kwon CI, Chung JG, et al. Endoscopic hemostasis with multiple hemoclips and an endoloop for uncontrolled peptic ulcer bleeding. Endoscopy 2011;43(Suppl. 2):E3–4. [78] Shi Q, Chen T, Zhong YS, et al. Complete closure of large gastric defects after endoscopic full-thickness resection, using endoloop and metallic clip interrupted suture. Endoscopy 2013;45:329–34. [79] Katsinelos P, Lazaraki G, Chatzimavroudis G, et al. Closure of an iatrogenic rectal perforation with the endoloop/clips technique in a purse-string fashion. Ann Gastroenterol 2014;27:264. [80] Katsinelos P, Kountouras J, Chatzimavroudis G, et al. Endoscopic closure of a large iatrogenic rectal perforation using endoloop/clips technique. Acta Gastroenterol Belg 2009;72:357–9. [81] Fujii T, Ono A, Fu KI. A novel endoscopic suturing technique using a specially designed so-called 8-ring in combination with resolution clips (with videos). Gastrointest Endosc 2007;66:1215–20. [82] Rabenstein T, Boosfeld C, Henrich R, et al. First use of ventricular septal defect occlusion device for endoscopic closure of an esophagorespiratory fistula using bronchoscopy and esophagoscopy. Chest 2006;130:906–9. [83] Kramer MR, Peled N, Shitrit D, et al. Use of Amplatzer device for endobronchial closure of bronchopleural fistulas. Chest 2008;133:1481–4. [84] Feller-Kopman D, Bechara R, Garland R, et al. Use of a removable endobronchial valve for the treatment of bronchopleural fistula. Chest 2006;130:273–5. [85] Ranes JL, Budev MM, Murthy S, et al. Management of tracheomediastinal fistulas using self-expanding metallic stents. J Thorac Cardiovasc Surg 2006;131:748–9. [86] Green DA, Moskowitz WB, Shepherd RW. Closure of a broncho-toneoesophageal fistula using an Amplatzer Septal Occluder device. Ann Thorac Surg 2010;89:2010–2.

150

J. Yang et al. / Techniques in Gastrointestinal Endoscopy 17 (2015) 141–150

[87] Boulougouri K, Theodoropoulos E, Karydas G, et al. Combined endoscopic and percutaneous treatment of a duodenocutaneous fistula using an Amplatzer septal occluder. Cardiovasc Intervent Radiol 2009;32:356–60. [88] Repici A, Presbitero P, Carlino A, et al. First human case of esophagus-tracheal fistula closure by using a cardiac septal occluder (with video). Gastrointest Endosc 2010;71:867–9. [89] Kumbhari V, Tieu AH, Saxena P, et al. Closure of a large, persistent enterocutaneous fistula by use of a ventricular septal occluder. Gastrointest Endosc 2014;80:332. [90] Coppola F, Boccuzzi G, Rossi G, et al. Cardiac septal umbrella for closure of a tracheoesophageal fistula. Endoscopy 2010;42(Suppl. 2):E318–9. [91] Swain P, Park PO, Mills T. Bard EndoCinch: the device, the technique, and preclinical studies. Gastrointest Endosc Clin N Am 2003;13:75–88. [92] Rahmani E, Freeman L, Sherman S, et al. New method of gastric incision closure in NOTES: short term survival study. Gastrointest Endosc 2008;67:AB121. [93] Ikeda K, Tajiri H. Evaluation of bifurcated double T-bar suturing device for effective closure of gastric perforation in ex vivo porcine models—aimed

[94]

[95]

[96]

[97]

[98]

at clinical application of NOTES [abstract]. Gastrointest Endosc 2008;67: AB116. Agrawal D, Chak A, Champagne BJ, et al. Endoscopic mucosal resection with full-thickness closure for difficult polyps: a prospective clinical trial. Gastrointest Endosc 2010;71:1082–8. Fritscher-Ravens A, Mosse CA, Mills TN, et al. A through-the-scope device for suturing and tissue approximation under EUS control. Gastrointest Endosc 2002;56:737–42. Pham BV, Raju GS, Ahmed I, et al. Immediate endoscopic closure of colon perforation by using a prototype endoscopic suturing device: feasibility and outcome in a porcine model (with video). Gastrointest Endosc 2006;64:113–9. Bergstrom M, Ikeda K, Swain P, et al. Transgastric anastomosis by using flexible endoscopy in a porcine model (with video). Gastrointest Endosc 2006; 63:307–12. Hashiba K, Siqueira PR, Brasil HA, et al. Endoscopic treatment for gastric perforation using T-tag and a plastic protection chamber: a short-term survival study. Arq Gastroenterol 2011;48:159–62.