Endoscopy after bariatric surgery

Anastomotic leak and fistula

Gastric leaks after bariatric operations represent one of the most serious complications with widely variable reported rates. According to a recent review, the incidence of leaks after open and laparoscopic RYGB ranges from 1.7 to 2.6% and from 2.1 to 5.2%, respectively. SG is complicated with leaks in up to 5.1% of cases [3]. Leaks usually present with tachycardia in 72-92% of patients and they are associated with mortality rates ranging from 6 to 14.5%, although rates as high as 40% have been reported [3,7,8].

Accumulating evidence supports the use of both self-expandable metallic (SEMS) and plastic stents (SEPS) along with conservative measures to treat post-surgery leaks. In 2006, Merrifield BF et al presented the endoscopic repair of gastric leaks in three patients after RYGB using Polyflex (Boston Scientific Corporation, Natick, MA, USA) esophageal stents. Durable fistula closure was achieved in all three patients, while one migrated stent had been retrieved endoscopically [9]. During the same year, Eubanks et al presented his experience using stents in 19 patients with staple line complications after bariatric surgery: 11 of them had acute leaks, 2 chronic fistula and 6 strictures. All patients were treated with endoscopic stent placement (11 SEMS, 23 SEPS). Treatment failure occurred in one patient in each group. Eventually, 58% of the stents migrated necessitating surgical removal in 3 of them [10]. Accordingly, partially-covered SEMS were introduced in 21 patients with large anastomotic leaks; 15 gastrocutaneous, two duodenocutaneous, three gastroperitoneal and one gastrobronchial fistulas, respectively. Overall success rate reached 81%. Best results were achieved in the RYGB and BPD groups (100%), while gastrocutaneous fistulas on a sleeve suture proved the most difficult to treat [11]. In another retrospective study, 18 patients (14 bariatrics) underwent endoscopic stent placement -successful in 13 of them- due to anastomotic complications; 13 anastomotic leaks, three strictures and two fistulas. Five patients required endoscopic or surgical intervention, while stent migration that occurred in four patients was managed endoscopically. There were two deaths unrelated to stent placement [12].

In a meta-analysis of 7 studies including 67 bariatric patients with evidence of anastomotic leaks, the proportion of successful leak closures using self-expandable stents was 87.8%. Of note, 9% of patients required revision surgery, while stent migration was reported in 16.9% of them [13]. The authors proposed the optimal time for stent removal to be between 6 to 8 weeks since that time period seems to prevent both incomplete closure, as well as stent migration or mucosal hypertrophy leading to increased difficulty of stent extraction or dysphagia that may require endoscopic treatment. However, there are currently no specific evidence-based recommendations regarding criteria and timing of stent removal. Finally, Murad et al reported on the stent treatment of 47 patients with acute leaks after bariatric surgery. 41 of them were healed with stent treatment alone and 5 of 6 persisting leaks required laparoscopic intervention; complication rate was 28.7% and there was no mortality [14].

With specific regard to laparoscopic SG, a recent large retrospective series and systematic analyses report a post-operative leak rate ranging from 1.5 to 2.8%. In terms of management, a combination of endoscopic stent placement and percutaneous drainage, antibiotics, as well as a short duration of parenteral nutrition usually provides very good results. Therefore, a re-operation is of no need in the vast majority of the cases [15-19].

As concluded from the aforementioned studies, leaks can be successfully and safely treated using endoscopic stents in the majority of cases. However, treatment failures as well as complications do occur, necessitating re-interventions. Gastrocutaneous fistulas are the most difficult to treat, thus necessitating surgical intervention. A major issue on leak treatment is stent migration that requires re-intervention and exposes patients to complication risks. The stents used in the presented studies vary from 16 to 22 mm in diameter and may reach up to 18 cm in length [9-14]. To prevent migration, larger stents (up to 25-28 cm) may be required [20]. These stents are not specifically designed for the surgically altered upper GI anatomy, they are not widely available in the market and they require sufficient expertise to be successfully placed. Open to discussion remains the selection of stent type (e.g. metallic versus plastic), since there are not comparative studies in the literature. Theoretically, metallic stents (Fig. 1) are easier to be placed than the non-preloaded SEPS and with lower risk of migration, while plastic stents are easier and safer to remove. Undoubtedly, these assumptions require high-powered controlled studies to be confirmed or rejected.

Figure 1 Use of a fully-covered self-expandable metallic stent (Hanarostent, M.I.Tech Co., Ltd., Seoul, Korea, length: 110 mm, diameter: 70 mm) to treat an anastomotic leakage following sleeve gastrectomy. (A) The arrow shows the anastomotic leak at the proximal aspect of the gastric sleeve; (B) Endoscopic image of the stent after full deployment; (C) Proximal aspect of the metallic stent into the lower esophagus proximal to the leak site; (D) Distal aspect of the metallic stent rested along the gastric antrum

Furthermore, either in combination with stent sealing or alone, endoscopically inserted nasocystic catheters, one or more double pig-tail stents as well as sponges connected transnasally to an external vacuum system have been used to drain leak-associated fluid collections. However, evidence in bariatrics is limited [21,22].

Beyond the above-mentioned endoscopic treatment methods for post-RYGB leaks, a couple of studies have highlighted the role of intra-operative endoscopy (IOE) to prevent such a complication [23,24]. Accordingly, IOE can readily detect anastomotic leaks thus indicating the need for subsequent operative maneuvers to reduce post-operative morbidity. Nevertheless, relevant randomized comparative studies are still lacking.

It should be finally mentioned, that regarding patients in the early (<4 weeks) post-operative period with suspected leaks, major scientific associations recommend the use of the rapidly-absorved carbon dioxide for insufflation and the choice of water-soluble contrast radiography as the initial diagnostic method [6].

Early postoperative hemorrhage

Upper GI bleeding occurs in up to 4% of patients after RYGB; it occurs less commonly after LAGB, SG and VBG [3,20]. Early GI bleeding is typically presented within 48 h after surgery, most commonly originating from the staple line of gastric pouch or gastric remnant, and the GJ or jejunojejunal anastomosis.

Esophagogastroduodenoscopy performed in 27 of 30 post-RYGB surgery early hemorrhage patients identified the bleeding site at the gastrojejunostomy staple line. Standard hemostatic techniques, such as heater probe cautery, epinephrine injection, and hemoclips were applied in 24 patients and bleeding control was achieved in all of them. Five re-bleeders required second endoscopic intervention; none required operation [26].

Stenosis of the GJ anastomosis following RYGB

RYGB can result in GJ anastomosis stenosis. Its incidence ranges from 2.9 to 23%; the rate of postoperative strictures following open RYGB is 0.67% being lower in comparison with laparoscopic RYGB (4.63%) [25,27-29]. Scanty data suggest that, similar to leaks, rates of stenosis can decrease by means of IOE [23,24]. Stoma stenosis diagnosis requires inability of a standard 9.5 mm diameter endoscope to pass through the anastomosis in patients presenting with dysphagia, vomiting and unhealthy rates of weight loss. GJ anastomotic stenosis can also present with worsening reflux symptoms that should always be endoscopically evaluated.

Treatment of anastomotic stenosis can be performed using Savary Gilliard bougies and through-the-scope (TTS) balloon dilations, the latter being the most commonly used technique with success rate exceeding 90%. Symptoms resolution is the primary sign of successful treatment and it is usually attained at a stoma diameter of 10-12 mm [30], though reports of successful and uneventful dilations to 18 and 20 mm have been published [31].

Table 1 summarizes the results presented in a large systematic review regarding endoscopic dilatation in patients with stenosis after RYGB [32]. From 1988 to 2010, 23 studies -the majority of them conducted in the USA- with 760 patients who underwent RYGB (open or laparoscopic) were identified [30,31,33-53]. 1298 dilatations were performed using TTS balloons in 16 studies (2 of them used additionally Savary-Gillard bougies), while two studies did not report the dilation method and in three studies non-specified balloon types was used. Overall, the success rate of the procedures exceeded 90% (older series excluded), while complications occurred in less than 2% of the patients [32]. Moreover, in a large cohort of RYGB patients (n=929), successful TTS balloon stricture dilation was related to the early -within 90 days after surgery- occurrence of the stricture since 98% and 61% of the early and late strictures were successfully treated [54].

Table 1 Endoscopic treatment of anastomotic strictures in post-bariatric surgery patients (from reference 22, modified)

Therefore, it can be concluded that TTS balloon dilation represents the primary means for postoperative anastomotic stenosis, while bougies dilation remains a complementary one. However, no head-to-head comparisons in terms of their safety and efficacy are available.

Dilated GJ stoma with enlargement of the gastric pouch following RYGB

A subset of patients regains weight long after RYGB; among other causes dilation of the GJ anastomosis also represents a possible factor that should always be endoscopically investigated. Endoscopic techniques used for GJ anastomotic tightening include sclerotherapy [55-59] and endoluminal suturing using the Endocinch suturing system (Bard Endoscopic Technologies, Billerica, MA, USA) [60-63], the Incisionless Operating Platform (USGI Medical Inc., San Clemente, CA, USA) [64-66], the StomaphyX system (Endogastric Solutions Inc., Redmond, WA, USA) [67-70], the OverStitch endoscopic suturing system (Apollo Endosurgery, Austin, TX, USA) [71,72], and, recently, serial argon plasma coagulation applications [73].

The published data regarding the endoscopic management of post-RYGB anastomotic dilation are extracted from small case series and no long-term (more than one-year follow up) results are available (Table 2). Only Thompson et al, reporting the 12-month follow up of their initial Incisionless Operating Platform treatment of stoma dilation, showed that the results were durable with no long-term adverse events and that stoma repair to a diameter of less than 10 mm was related with greater sustainable weight loss [74].

Table 2 Endoscopic treatment of dilated gastrojejunal stoma and enlargement of the gastric pouch following Roux-en-Y gastric bypass

It is noteworthy that nowadays the only commercially available device is the OverStitch Endoscopic Suturing System (Apollo Endosurgery, Austin, TX, USA) while the rest should be considered within clinical trials in reference centers. This system represents a cap-based device that adapts into a double-channel endoscope allowing a curve needle to deploy sutures under direct visualization. There are only two small series on the management of RYGB failure. In the former series, nine patients with dilated gastrojejunostomy after RYGB underwent OverStitch system plication achieving weight loss ranging from 4.9 to 9.5 kg in one month. Emesis and bleeding were reported in one patient, while another one required balloon dilation due to liquids intolerance [74]_. In the latter series, OverStitch System treatment in eight patients with stoma dilation after RYGB was associated with 6-8 kg weight loss in four of them, with no complications [72].

Marginal ulcerations in RYGB

Marginal ulcers appear in 20% of patients undergoing RYGB, more often during the first three months postoperatively. They commonly arise at the gastric side or the jejunal side of anastomosis. Many factors contribute to ulcer formation, such as bile acid reflux, ischemia, coexisting gastrogastric fistula, Helicobacter pylori infection, tension on the Roux limb, nonsteroidal anti-inflammatory drugs, nonabsorbable sutures, smoking and alcohol [75].

The patients with anastomotic ulcers usually present with persistent abdominal pain that indicates further endoscopic evaluation. Treatment includes proton pump inhibitors twice daily, tapered over 6 months. Eradication of Helicobacter pylori should be performed, if detected. Visible, non-absorbable sutures should be removed when possible to allow ulcer healing and treatment success should be evaluated with repeat endoscopy mainly in patients with persistent symptoms to rule out other complications [3,76].

Obscure GI bleeding

Investigation of iron deficiency anemia and obscure GI bleeding (OGIB) is challenging in surgically altered anatomy since potential bleeding sites such as the anastomotic site and the excluded stomach might not be accessible with conventional endoscopy. The American Society for Gastrointestinal Endoscopy recommendation for deep enteroscopy as the initial diagnostic evaluation in post bariatratric surgery patients with OGIB was based on experts’ opinion [77]. Currently, there is only one case series focused on the efficacy of device assisted enteroscopy to successfully identify and treat lesions related to OGIB in the operated stomach [78]. More precisely, double balloon enteroscopy identified the responsible for OGIB lesion in 10 of the 12 patients with altered anatomy; in 9 of them the bleeding site was at the anastomosis and in one at the afferent limb. Endoscopic treatment was applied in eight patients, four patients required repeat endoscopy due to bleeding recurrence and there was one death attributed to perforation [78].

Finally, access to the excluded portion of the stomach and/or Roux limb can be facilitated through a surgically created gastrostomy, when traditional endoscopic techniques fail [6].

Gastric band slippage and erosion

Literature indicates that gastric band slippage rates vary from 2 to 14% [79,80]. Patients commonly present with worsening epigastric pain, nausea/vomiting, acid reflux or even weight gain. The subsequent gastric prolapse is diagnosed with upper GI series and endoscopy and necessitates surgical intervention.

On the contrary, band erosion through the gastric wall develops in 0.2 to 33% of cases giving an overall incidence of 1.46% in a 2011 systematic review [81]. It occurs in a mean of 22.5 months after surgery [82] and its presentation involves pain, nausea, vomiting, hematemesis and failure to lose weight. Diagnosis is achieved by endoscopy which reveals the eroding part of the band into the lumen. Removal of the eroded band can be accomplished by either endoscopy (Fig. 2) or laparotomy. Several case reports/series have demonstrated feasibility, safety and efficacy of the endoscopic technique suggesting it as a first choice procedure in such cases [83-86].

Figure 2 Endoscopic management of an eroded gastric band by means of a gastric band cutter. (A) More than 75% of the gastric band has migrated into the stomach; (B) The cutting wire of the device (shown by the open arrow) is positioned into the stomach through the working channel of the gastroscope next to the gastric band and thereafter folded around the band using a snare to retrieve it; (C) A metallic tube (cutter) shown by the solid arrow is inserted into the stomach over the folded cutting wire (open arrow) and it is pushed against the gastric band under direct endoscopic view; (D) The outer edge of the cutter is inserted into the tourniquet of the handgrip. By twisting the handle of the device the band is strangulated, cut and easily removed by gentle traction

Endoscopic treatment of common bile duct stones

RYGB is associated with long Roux limb and multiple luminal angulations, therefore poses several technical difficulties in accessing the biliary tree. Most available data come from series including patients that undergone RYGB for indications other than morbid obesity. ERCP in this population is performed with variable results by either enteroscopy (push, balloon-assisted, spiral) or by advancing the duodenoscope through a gastrostomy tract into the excluded stomach and duodenum.

A few studies including patients undergone RYGB for morbid obesity are available. Accordingly, Gostout et al reported for the first time the use of a pediatric colonoscope to treat common bile duct stones in operated bariatric patients [87]. Lopes et al reported that laparoscopy-assisted ERCP (LA-ERCP) in such patients was safe and facilitated biliary cannulation in 90% of them [88], while technical success was universal (100%) in a recent study of similar design [89]. Moreover, “transgastrostomy” ERCP was successful in all patients after bariatric gastric bypass [90]. Two comparative studies in RYGB-treated obese patients have also come to light. Choi et al showed that ERCP via gastrostomy is more effective than double balloon enteroscopy ERCP in gaining access to the pancreatobiliary tree but it is hindered by the gastrostomy maturation delay and a higher morbidity [91]. On the other hand, LA-ERCP was found superior than balloon (either single- or double-) enteroscopy ERCP especially in patients with length of Roux limb plus biliopancreatic limb more than 150 cm [92]. Very recently, a novel method that overcomes the disadvantages of enteroscopy and laparoscopy has been proposed. In detail, endoscopic ultrasonography-facilitated access to the excluded stomach through the creation of a gastro-gastric fistula tract covered with a metallic stent that allowed the antegrade passage of a duodenoscope [93,94]. Furthermore, Law et al demonstrated a novel method to access the biliary tract with a duodenoscope that is advanced through a SEMS that was placed within an endoscopically-created gastrostomy tract [95]. Nevertheless, it should be noted that the optimal method to perform ERCP in RYGB patients remains unclear and the final choice should be based on endoscopist’s expertise and available equipment.