Anastomotic (and long staple line) leak is one of the most feared complications of any gastrointestinal surgery. Gastric bypass and duodenal switch procedures incorporate two anastomoses at risk for postoperative leak (Fig. 8.1a, d). The gastrojejunostomy and duodenojejunostomy in particular can be under tension and have subsequent ischemia, leading to leak. Leak rates of these proximal anastomoses are typically higher than the distal enteroenterostomies and are less likely to be under tension. From an endoscopic perspective, this is fortunate since the distal anastomoses are more difficult to access. Sleeve gastrectomy entails removal of the greater curvature of the stomach, leaving behind a tubularized stomach with intact pylorus. This leaves a high-pressure system with a long staple line when compared to the other stapled procedures. Leaks along the sleeve staple line can become chronic and very slow to heal. Two areas particularly prone to being troublesome include the very proximal stomach at the angle of His which is at risk for leakage, and the distal stomach alongside the incisura which is at risk for functional obstruction due to folding resulting in upstream pressurization. Leaks typically manifest with tachycardia, fatigue, malaise, shortness of breath, and hypotension. An upper GI fluoroscopy study can help elucidate the site of the leak. However, a negative upper GI study does not completely rule out a leak, and if there is clinical concern for a leak it is appropriate to investigate further endoscopically or operatively.

Endoscopically placed stents (either tubular or pigtail) can be helpful in the immediate postoperative period. When used in an acute setting, stents can be highly effective as a solo treatment of leak [5]. Even in cases where another therapy is required, use of a tubular exclusion stent can allow patients to continue to take oral nutrition. In some settings, percutaneous drainage of intraabdominal abscess or laparoscopic drainage of extralumenal fluid may speed the rate of healing a leak [6, 7]. The use of pigtail stents facilitates internal drainage of a leak-related abscess in the same fashion as the management scheme used for pancreatic pseudocysts with formation of a cyst-gastrostomy . Use of pigtail stents is rarely associated with migration or erosion; however distal enteral access is usually needed for nutritional support to avoid ongoing outflow through the leak from proximal enteral intake.

Stents are not risk free. Tubular stents used to treat leaks and nonmalignant strictures in the USA are for the most part being used in an off-label (although often effective) manner. They are not perfectly designed to complete the task for which they are often being used. Tubular stent migration is a common nuisance, seen in nearly 50 % of stent placements [7]. Inserting a second “nested ” stent into the first creating a longer effective stent with distal abutment in the antrum can decrease the migration rate , and increase coverage of a sleeve staple line to ensure coverage of an entire staple line if necessary. Stents can also be sutured into place to decrease migration, either using an endoscopic suturing device or with laparoscopic assistance [8].

Other morbidities related to tubular stents include erosion and fistula formation due to the radial pressure applied by the stent. Radial pressure is crucial to maintain stent position and exclude the leak. These rates increase as stent duration increases, and it is advised to leave stents in no longer than 28 days before removal or exchange [9]. Use of tubular stents for short durations can be very useful, but the total duration of stent use should be limited to 2–3 months due to the potential concerns of erosion, stenosis, and fistula formation noted earlier [10].

When a leak persists over the course of weeks and becomes chronic, tubular stents are less likely to facilitate a nonoperative durable solution . In a study of stents placed for chronic leak, only 19 % had a successful closure with stent alone [7]. Thus, in this population, stents are likely most effective as a bridge therapy to definitive surgical closure. In this setting, they can help with control of sepsis and allow patients to resume enteral nutrition. Use of downstream jejunal feeding access, which can be placed percutaneously with endoscopic guidance, can also avoid parenteral nutrition in many circumstances.

Endoscopic clips are also useful to manage small leaks. Typically these work best if the leak is discovered in the early postoperative period. Clips are available in several types. Through-the-scope clips are available from multiple manufacturers. They are small, and can close mucosal defects; however they do not ordinarily provide full-thickness approximation of tissue . Over-the-scope clips, although more cumbersome and bulky, can provide full-thickness approximation of gastric tissue, and thus are the most helpful type of clip for treatment of leak.

Leak algorithm: Assess stability of patient and acute/chronic nature of leak. If acute and small, endoscopic closure with through-the-scope or over-the-scope clip can be performed. If acute and large, tubular stent with external drainage and distal enteral access should be performed. If chronic and small, pigtail stent ± distal enteral access can be performed. An over-the-scope clip can be attempted if tissue is pliable. If chronic and large either type of stent can be used for local control (alternate is using endoscopic/surgically placed endolumenal/external drain to create a controlled fistula). Distal enteral access is advised and plans should be made for surgical revision if the leak fails to close (Videos 8.1, 8.2, and 8.3).

Hemorrhage from staple lines and anastomotic sites is often intralumenal and well suited to endoscopic hemostasis . Acute bleeding manifests in the first 72 h after surgery with tachycardia, palor, fatigue, malaise, and/or a drop in hemoglobin. Many of the staple lines are easily accessible endoscopically. Epinephrine injection, argon plasma coagulation, heater probe cautery, endoscopic suturing, and endoscopic clipping are all potentially useful adjuncts to treat staple line bleeding . Through-the-scope clip application is particularly useful for fresh anastomoses, as it avoids thermal damage that may further compromise staple line healing.

Through-the-scope clips are effective for both gastrojejunostomy and sleeve gastrectomy bleeding. Over-the-scope clips are capable of compressing larger areas of tissue and thus larger vessels. Bleeding from the enteroenterostomy of a Roux-en-Y gastric bypass or duodenal switch is far enough downstream that it is unlikely to be managed endoscopically. Luckily, bleeding at that site is fairly rare and can often be managed expectantly or in severe cases with laparoscopic oversewing of the staple line.

Intraperitoneal bleeding can occur from staple lines, as well as from the short gastric vessels and mesentery and solid organs which can be inadvertently damaged during surgery (particularly the spleen and liver). These areas are less amenable to endoscopic hemostasis and require formal surgical treatment.

Bleeding algorithm: Assess stability of patient and acute/chronic nature of hemorrhage. If patient is vomiting blood the site is endolumenal and proximal for which endoscopy is ideal. Therapeutic endoscope with large suction channel is helpful for removing clot enabling precise placement of endoscopic clip or suture to achieve hemostasis. If patient is passing blood per rectum with no proximal symptoms the site is likely distal requiring laparoscopic assessment and potentially oversewing the staple line. If patient has no gastrointestinal symptoms but has decreasing hemoglobin/hematocrit and symptoms of shock, the site is likely extralumenal and requires surgical treatment.

The anatomy of a Roux-en-Y gastric bypass can lead to a marginal ulceration. Marginal ulceration can occur in up to 16 % of patients [11]. This ulcer, often seen in the jejunum just distal to the gastrojejunal anastomosis , can be multifactorial, Table 8.1 (Fig. 8.2). Ischemia is one common cause. Operative ischemia and microvascular ischemia are both culprits. Smoking in particular has been shown to cause microvascular ischemia that can lead to marginal ulceration, and patients who have had a gastric bypass are discouraged from smoking. The cardia cells in the gastric pouch are capable of acid production, and if an ulcer develops a trial of antisecretory therapy is appropriate to eliminate acid production that might cause the ulceration to develop or propagate. Foreign material can be a nidus for marginal ulceration, and the use of absorbable suture is helpful in avoiding this complication. If there is permanent suture present at the anastomosis, it can be removed endoscopically as long as there has been adequate time for anastomotic healing (typically 6 weeks, but can vary depending on each individual situation). Helicobacter pylori infection is another cause for ulceration, and all patients should be screened and treated for Helicobacter pylori preoperatively to avoid this. Nonsteroidal anti-inflammatory drugs can also exacerbate marginal ulcers, and patients should avoid their use if they have other risk factors for marginal ulceration.