Exiting the lumen of the gut can be rather a frightening experience, at least for endoscopists who have been conditioned throughout conventional GI training to stay within the lumen, and that to do otherwise constitutes a perforation and therefore a complication. When exiting the lumen, one runs the risk of injuring a nearby organ or causing bleeding from vessels on the serosal side of a hollow organ that cannot be seen when the site of exit is selected. Every effort should be made to exit in a location and a manner that minimizes these risks. Therefore, certain landmarks should be sought and rules followed when exiting a natural orifice. For example, the “triangle of safety” can be used for transvaginal access [5]. We always attempt to exit the stomach or bowel on the antimesenteric border, where blood vessels are the fewest and smallest. Some exiting techniques were specifically designed with safety in mind.

This is the simplest but also the least safe of exiting techniques. A needle knife or other cutting device is used to incise the hollow organ in layers to provide a full-thickness defect through which the endoscope can be passed. The risk of injury to nearby loops of bowel and/or solid organs is not negligible. But this method is simple and quick. It is often used in nonsurvival animal experiments where perforation of a nearby loop of bowel is of little consequence. This type of exit is also the most difficult to close, essentially requiring endoscopic suturing or, if the defect is not too big, over-the-scope clips (OTSCs). This is yet another reason why this method is used in nonsurvival experiments, where closure is not attempted or at least is not critical because the animal is to be sacrificed immediately afterward. Some workers initially advocated the use of endoscopic ultrasound (EUS) to provide additional safety, but they now feel that this has little added value, and most do not use EUS in an attempt to make gastric puncture/incision safer [6].

This method comprises a blind puncture with a 19-ga EUS needle placed through the working channel of a straight endoscope followed by the passage of a guidewire into the abdominal or other cavity. Risk of puncturing another organ is low if the puncture is done smoothly and slowly. Other hollow viscera tend to float away from the needle. If solid organ anatomy is kept in mind, this can be done safely. Once a guidewire is advanced into the peritoneal cavity, the needle is removed leaving the wire in place. A standard 15- to 18-mm esophageal dilation balloon can then be advanced over the wire and used to dilate the tract. One then pushes the balloon through with the endoscope, whose tip follows it out into the abdominal cavity. An additional way of increasing the safety of this technique is to insufflate the abdomen with CO₂ prior to anterior wall gastric puncture. This helps to prevent injury to other viscera nearby at the time of the gastric puncture or incision [7].

An advantage to this method is that there is no cutting, so bleeding risk is minimized. Another advantage is that without cutting, the muscle layers stretched during the dilation tend to return to their original configuration once the endoscope is removed, and closure may be simplified. Indeed, as will be shown in the next section, Jagannath et al. did not even close the gastrotomy site (a needle knife rather than a needle was used to make the initial small puncture) in a porcine survival model after ligation and transection of the uterine horn to simulate appendectomy [8].

This is likely the safest but also the most complicated method of entry into the abdomen. Because the method utilizes techniques normally used in PEG placement, it can only be used for gastric exit into the peritoneal cavity. The method has been described in detail by Kantsevoy et al. [9]. In brief, the method comprises endoscopic insufflation of the stomach, transcutaneous needle puncture, and guidewire insertion under endoscopic viewing and then removal of the wire out through the mouth as would normally be done for PEG placement. The wire is then back loaded into the working channel and is captured and pulled out through the biopsy port. The scope is then reinserted and used to anchor the wire in the stomach. Next, capnoperitoneum is achieved with a transabdominal Veress needle and CO₂. Finally, the wire is pushed into the abdomen from the skin side, after fixing it in place on the stomach side with the tip of the endoscope. By pushing firmly on the skin side of the wire, it has no place else to go except to “knuckle” or flex into the inflated abdomen. A through-the-scope (TTS) balloon can then be advanced along the wire and pushed along the wire across the wall, much like a push PEG technique, and then used to dilate the puncture site to allow the scope to exit into the abdominal cavity.

The advantages to this method are that it is very safe, there is no cutting, thus minimizing bleeding risk, and closure is made easier. Disadvantages are that it is a bit time-consuming and complicated compared to just cutting one’s way out of the stomach, and also that only the anterior of the stomach can be exited (which may not be appropriate for some NOTES procedures). Because of its safety profile, we use the PEG technique exclusively in survival animal studies and would recommend it in human procedures as well.

Several groups have now reported on the creation of a submucosal tunnel proximal to the seromuscular incision used to exit the stomach or esophagus [10–12]. These are all variations on a theme. Briefly, saline is injected submucosally, lifting the mucosa from the muscle layers and expanding the submucosal potential space. A mucosal incision is made at one end of the saline lift. By blunt dissection with forceps or balloons, or by using electrocautery, a long submucosal tunnel is formed that allows passage of a cap-fitted endoscope. Further injections of saline and continued dissection are used to lengthen the tunnel. Then, the muscle and serosa are incised to allow the endoscope to pass into the mediastinum or peritoneal cavity. Once the extraluminal procedure is completed, the endoscope is withdrawn. The mucosotomy and myotomy are at distant sites; they do not overlap. The mucosal tunnel acts as a flap valve preventing luminal contents from exiting through the tunnel. Most investigators [10–15] will close the mucosotomy with endoscopic hemostatic clips as insurance against leakage.

An advantage to this technique is that it provides for a very easy and safe closure. In addition, the exit site can be targeted at a place on the stomach (for gastric exit) that might not be achievable with the PEG method. However, it is time-consuming, adding to the time of the procedure and therefore to its cost. It requires considerable endoscopic skill as well, and experience in endoscopic submucosal dissection (ESD) or POEM is a prerequisite.

Gastrotomy or other enterotomy closure has been an area of active research since the inception of NOTES. Various devices have been designed specifically for this purpose, and others such as hemostatic clips are adapted for closure and are essentially used off-label. The actual devices (rather than the techniques of closure) have been reviewed elsewhere [16]. Many surgeons and endoscopists emphasize the importance of adherence to proper surgical principles to prevent leaks. Leakage of luminal contents into the abdomen or mediastinum would be a potential postoperative disaster, so as much research attention has been paid to closure methods as to the surgical procedures being developed.

It is beyond the scope of this chapter to review the physiology of intraluminal gut pressures and what factors may contribute to the failure of an enterotomy closure, but a few salient points should be made. Surgical dogma has long held that enterotomy closures must be able to withstand “cough pressure” or the intra-abdominal pressure changes that occur when coughing. The mean intra-abdominal pressure can be as much as 165 cm H₂O (121 mm Hg) during cough [17]. However, our group has observed that extraluminal and intraluminal pressures are nearly identical during simulated cough and that pressure changes as a cause of leak are greatly overstated [18]. Shear forces, failure to appose the edges of a defect, and breakdown of the closure (dehiscence, device failure) are all more important than simple pressure changes in the abdomen because these pressure changes are the same both inside and outside the lumen, and the pressure differential or pressure gradient across the gut wall is near zero even during cough. Clearly, something else must influence leakage rather than pressure alone. Indeed, some surgeons do not even close the gastrotomy after uterine horn resection as will be shown in the next section.

We will now briefly review the types of closures most commonly encountered (no closure, hemostatic clips, OTSC, and T-tags). An exhaustive treatment of closure methods is beyond the scope of this chapter, and the interested reader is referred to several reviews on this subject [16, 19].

Jagannath et al. did not close the gastrotomy exit site in their survival porcine model of appendectomy (transgastric uterine horn resection) [8]. Exit was made by needle-knife puncture, but the tract was dilated with a balloon prior to pushing the endoscope across the gastric wall and into the abdominal cavity. The endoscope was simply withdrawn at the end of the procedure. The animals were fasted overnight and then given standard laboratory chow the next day. There were no leaks or infections. It is theorized that a small defect is easily tolerated (we know this from EUS/FNA) and that balloon dilation of the defect serves to spread the muscle fibers but not cut them. They spring back tonically, and the gastrotomy closes rapidly after endoscope removal. To our knowledge, no one has tried this yet in humans.

Clip closure was one of the first methods used in NOTES enterotomy closure. Many case reports had already been published describing effective closure of iatrogenic perforations with endoscopic hemostatic clips after EMR, polypectomy, or other endoscopic procedures. Much work was subsequently done in the animal laboratory with the closure of intentional perforations, and clips were shown to be effective. Work initially began with hemostatic clips [20–23], and later, work was done with OTSC clips [24–33].

Endoscopic hemostatic clips are titanium devices that are intended for endoscopic hemostasis of arterial hemorrhage (Fig. 2.1). Most have two arms; some can be rotated, some can be opened and closed repeatedly, and some can do both. They are useful for peptic ulcer bleeding, Dieulafoy’s lesions, postpolypectomy or postsphincterotomy bleeding, or for bleeding colonic diverticula and other lesions. However, they can also be used off-label to close a mural defect.

Raju et al. demonstrated that endoscopic hemostatic clips could be used to close small, full-thickness colon defects in a porcine model [20]. Merrifield et al. used hemostatic clips to close the gastrotomy after transgastric uterine horn resection in a survival porcine study [21]. However, 3 of the 5 experimental animals developed significant complications due to incomplete or failed gastric closure. Given this, it seems that clip closure can be risky, and strict attention must be paid to the integrity and strength of the closure. Fritscher-Ravens et al. have demonstrated effective closure of esophageal perforations in a swine NOTES model using hemostatic clips [22]. Tsunada et al. demonstrated that clip closure could be effective in human patients. Seven patients who suffered full-thickness gastric perforations after EMR had their defects closed successfully using endoscopic hemostatic clips [23]. No patient required laparotomy.

The most commonly used OTSC for both NOTES and closure of defects after ESD or EFTR (endoscopic full-thickness resection) is the OTSC^® Closure system (Ovesco Endoscopy, Tübingen, Germany). These clips are made of nitinol and are in the shape of overlapping jaws or “bear claws” (Fig. 2.2). The jaws are in the closed position when manufactured, but are flexed into the open position when mounted on the tip of an endoscope. When deployed by pulling back the inner collar of the deployment pod, the device will snap shut into the closed position due to the memory properties of nitinol. This entraps the gastrotomy between the jaws of the clip, thus sealing the defect.

The feasibility of OTSC closure of defects was noted as far back as 2008 [24]. This study used the Ovesco OTSC^® to close gastric defects made after needle-knife exit in a nonsurvival porcine model. Defects were primarily closed in 8 of 9 experimental subjects, but the ninth could not be closed effectively due to a 20-mm rent accidentally made in the gastric wall. In 3 of the remaining 8 animals, the closure did not withstand “burst pressures,” a fact that may not be clinically significant as noted earlier [18]. Nevertheless, the feasibility of using OTSC for closure had been established.

Matthes et al. used the Ovesco OTSC to close standardized defects in explanted porcine stomachs and then burst tested them with compressed air under water to assess for strength of the closure [25]. Many similar ex vivo studies, in vivo survival and nonsurvival animal studies, and human studies have shown them to be effective as closure devices for NOTES and other iatrogenic perforations in the stomach, duodenum, colon, and esophagus [26–29].

Although used less commonly, the Padlock Clip™ OTSC (Aponos Medical, Kingston, NH) has also been used for closure of persistent fistulas, as well as transgastric and transcolonic NOTES. Originally approved for endoscopic hemostasis, this device consists of a hexagonal nitinol clip mounted in a translucent cap that fits over the endoscope. It has 6 prongs which, when deployed, gather tissue into the center of the hexagon, thus tamponading a bleeding site or, in the case of NOTES, sealing the defect (Fig. 2.3).

Our group published some of the first animal work with this device. We demonstrated in 2009, in an explant study with burst pressures, that the Padlock Clip™ could provide a secure gastric closure for NOTES [30]. Later, we reported the use of the Padlock Clip™ for gastrotomy closure in a survival study [31]. Two pigs were survived for 2 weeks, and 2 were survived for 6 weeks. All animals did well, the device appeared to be easy to use, and it provided a secure closure. So and Adler published a case report of closure of a persistent tracheoesophageal fistula in a human patient using the Padlock Clip™ [32], and Guarner-Argente et al. reported successful colonic closure with this device [33].