Keeping the fistula open: paradigm shift in the management of leaks after bariatric surgery?

Leaks associated with bariatric surgery occur in 1 % – 3 % of patients [1] [2]. In sleeve gastrectomy, over 90 % of leaks occur at the proximal margin of the staple line [3]. The pathomechanism behind leak formation in this location is thought to relate to ischemia secondary to take down of the short gastric arteries, mismatch in staple height selection due to the thinner wall of the gastric fundus, high intraluminal pressures resulting from a poorly complaint sleeve, or gastric stenosis at the incisura angularis. The management of leaks after bariatric surgery lacks a standard management algorithm owing to a paucity of prospective randomized trials. Regardless, it is universally accepted that endoscopy plays a pivotal role in the management of both acute and chronic leaks.

There is a myriad of endoscopic techniques that have been reported with varying efficacy for the treatment of leaks after bariatric surgery. Traditionally, treatment has been aimed at facilitating drainage (percutaneous or surgical), followed by endoscopic closure of the leak site [1]. The development of endoscopic techniques, such as over-the-scope clips and endoscopic suturing, allows for high rates of technically successful closure, although recurrence of the leak is not uncommon [4] [5]. Mechanisms behind the failed closure are thought to be secondary to: a) poor integrity of the tissue surrounding the leak as a result of ischemia and inflammation, b) difficulty obtaining a perpendicular endoscopic view of the leak for optimal repair, and c) failure to address high intraluminal pressures caused by downstream gastric stenosis. Therefore, primary endoscopic closure is often not durable despite being performed by highly skilled endoscopists using state-of-the-art technologies.

Diversion therapy using self-expandable metallic stents (SEMS) provides an alternative to primary closure, and is often a very useful adjunct to surgical washout and drainage in the management of early leaks [6]. Diversion therapy is appealing as it allows for early oral alimentation and addresses downstream gastric stenosis. However, there is a dichotomy that exists with the use of SEMS. If a fully covered SEMS is used, then it is mandatory to secure it (e. g. endoscopic suturing) in order to minimize migration, and diversion may be suboptimal as the seal is not “watertight.” Partially covered SEMS solve these two aforementioned problems, but stent removal is difficult and may require the use of a second stent insertion, which results in additional endoscopic procedures and higher costs. In addition, regardless of stent type, there is often occlusion of the distal end of the stent by tissue overgrowth, which increases intraprosthetic pressure leading to poor tolerability as well as inadequate diversion.

Endoscopists managing these patients recognize that closing the leak site with tissue apposition techniques or diversion therapy may not be the ideal treatment strategy. Could facilitating internal drainage by “keeping the fistula open” result in superior outcomes? The rationale behind endoscopic internal drainage (EID) is similar to that of cystgastrostomy for walled-off pancreatic necrosis, and is based on the pressure within the perigastric collection being higher than that in the gastric lumen. Therefore, EID as a sole strategy is not suitable if: the perigatric collection is not organized, there is high intragastric pressure secondary to downstream stenosis, or there is a gastropleural fistula. Once the contents of the collection have been drained, the cavity slowly contracts until it is obliterated, and therefore closure of the orifice of the fistula is unnecessary.

In this issue of Endoscopy, Bouchard et al. [7] report the outcomes of a single-center retrospective series of 33 patients who underwent EID using plastic double-pigtail stents predominantly through the pre-existing fistula. In their study, technical success was achieved in 100 %, and clinical success was achieved in 79 % at long-term follow-up. The authors included patients who underwent transgastric stents as primary therapy as well as those undergoing the procedure as salvage therapy, with similar outcomes in the two groups. Interestingly, scheduled removal of the transgastric stents was not included and this resulted in adverse events. Furthermore, there was an unusually small proportion of patients who had downstream gastric stenosis.

EID with transgastric stents has been reported previously by several groups, with Donatelli et al. reporting the largest series [8]. In his series of 67 patients, the clinical success rate was similar to that of Bouchard et al. at 75 %, however, there were differences in the patients and methods used. Donatelli et al. performed transgastric stenting through the fistula, even in patients with disorganized collections. We recommend against this, as when there is a disorganized or uncontained collection, the pressure within the collection may be less than that within the gastric lumen, thus preventing drainage of debris into the stomach. A potential explanation for the leak not being perpetuated in such patients in Donatelli’s study may be that the patients received aggressive treatment of downstream stenosis, which is likely to have reduced intragastric pressure. Another difference between the studies was that Donatelli et al. performed multiple procedures (mean 3, range 2 – 16) in order to facilitate drainage and remove the stents. The advantage of performing multiple procedures is that it allows the opportunity to perform necrosectomy and repeated dilation of downstream stenosis, which is frequently present. Finally, Bouchard et al. demonstrated that early oral feeding is safe, whereas Donatelli et al. preferred enteral feeding or parenteral nutrition in the early stages of therapy. We endorse early oral feeding, as oral contents should not enter the perigastric cavity because of the unfavorable pressure gradient. The avoidance of the complications associated with a feeding tube or intravenous alimentation, as well as the improved quality of life with oral intake, must be acknowledged.

Although EID with transgastric stents appears to be effective, controversies exist as to the optimal technique. In patients with infected solid material, we believe that necrosectomy expedites clinical improvement and should be performed at intervals of 2 – 4 weeks until the debris has been removed [8] [9]. In addition, lavage of the collection with either a percutaneous drain or nasocystic catheter has also been reported. However, external drains are not only associated with patient discomfort, but may in fact hinder resolution of the collection. External drains result in a reduction in pressure within the perigastric collection, which may hamper internal drainage of material into the gastric lumen. Therefore, if an external drain already exists within the collection, we believe that it should either: a) be removed at the time of internal drainage, or b) be used to intermittently flush the collection but be clamped at other times. The second controversy surrounds whether scheduled stent exchange and eventual removal of the transgastric stents are necessary. We do not see the value in routine stent exchange unless necrosectomy is also being performed. However, long-term stent dwell without endoscopic surveillance is dangerous. Bouchard et al. reported ulcerations, dysphagia, discomfort, and splenic hematoma [7]. Therefore, we recommend that stent removal should be performed after clinical and radiologic resolution, and we see no benefit in keeping stents in situ once this has been achieved.

Controversially, both Bouchard et al. and Donatelli et al. suggest that EID with transgastric stents is effective in patients regardless of leak classification [3]. Although this may be the correct conclusion drawn from their statistical analysis, 97 % of patients in Bouchard et al. and 87 % of patients in Donatelli et al., had acute or early leaks. Therefore, we cannot be confident that this technique has high efficacy in patients who have late or chronic leaks. In late or chronic leaks, endoscopic vacuum therapy and endoscopic septotomy may provide an alternative treatment strategy, with both methods also utilizing the concept of keeping the fistula open.

Endoscopic vacuum therapy using a sponge was recently reported in a small series of nine patients with proximal staple line leaks after sleeve gastrectomy [10]. Clinical success was reported in all patients. Despite the excellent outcomes, there are several disadvantages to this treatment strategy. Notably, a mean number of 10 (range 2 – 18) endoscopies were performed, as the sponge needs to be changed every 4 days, necessitating that patients live in close proximity to the hospital. In addition, because this technique results in negative pressure at the leak site, a feeding jejunostomy or total parenteral nutrition is required.

Endoscopic septotomy was described first in 2011 by Campos et al. [11] The rationale behind the technique relates to higher intraluminal pressures within the sleeve compared with the perigastric cavity, promoting flow of contents through the fistula. To equalize these pressures, the pseudoseptum is cut between the perigastric cavity and the gastric lumen to the base of the perigastric gastric cavity using argon plasma coagulation or a needle-knife [12]. This procedure is combined with aggressive dilation of downstream stenosis using Rigiflex pneumatic balloons (Boston Scientific, Marlborough, Massachusetts, USA) to further reduce intraluminal gastric pressure. Repeated therapy is performed as necessary, with more pseudoseptum being cut each time. In a series of 27 patients, fistulas closed in all patients, with a mean time to closure of 18 days [13]. A potential advantage of this procedure over the technique described by Bouchard et al. is that the fistula is opened to encourage drainage as opposed to simply keeping the existing tract patent.

There is no consensus on the management of patients who do not respond to endoscopic treatment strategies. Bouchard et al. report that 21 % of patients did not respond to endoscopic therapy, and endoscopic techniques are not without morbidity. The criteria for salvage surgery, such as conversion to Roux-en Y gastric bypass or fistulojejunostomy, are unknown. Surgical revision may have a lower morbidity than once thought, with one recent study of 21 patients reporting fistula control in all patients, with no mortality and a low morbidity of 5 % [14]. Therefore, a multidisciplinary approach is essential in managing these complex patients.

• References

• 1 Gagner M, Hutchinson C, Rosenthal R. Fifth International Consensus Conference: current status of sleeve gastrectomy. Surg Obes Relat Dis 2016; 12: 750-756
• 2 Varban OA, Cassidy RB, Sheetz KH et al. Technique or technology? Evaluating leaks after gastric bypass. Surg Obes Relat Dis 2016; 12: 264-272
• 3 Rosenthal RJ, Diaz AA. International Sleeve Gastrectomy Expert Panel et al. International Sleeve Gastrectomy Expert Panel Consensus Statement: best practice guidelines based on experience of >12,000 cases. Surg Obes Relat Dis 2012; 8: 8-19
• 4 Sharaiha RZ, Kumta NA, DeFilippis EM et al. A large multicenter experience with endoscopic suturing for management of gastrointestinal defects and stent anchorage in 122 patients: a retrospective review. J Clin Gastroenterol 2016; 50: 388-392
• 5 Haito-Chavez Y, Law JK, Kratt T et al. International multicenter experience with an over-the-scope clipping device for endoscopic management of GI defects (with video). Gastrointest Endosc 2014; 80: 610-622
• 6 Murino A, Arvanitakis M, Le Moine O et al. Effectiveness of endoscopic management using self-expandable metal stents in a large cohort of patients with post-bariatric leaks. Obes Surg 2015; 25: 1569-1576
• 7 Bouchard S, Eisendrath P, Toussaint E et al. Trans-fistulary endoscopic drainage for post-bariatric abdominal collections communicating with the upper gastrointestinal tract. Endoscopy 2016; 48: 809-816
• 8 Donatelli G, Dumont JL, Cereatti F et al. Treatment of leaks following sleeve gastrectomy by endoscopic internal drainage (EID). Obes Surg 2015; 25: 1293-1301
• 9 Lemmers A, Tan DM, Ibrahim M et al. Transluminal or percutaneous endoscopic drainage and debridement of abscesses after bariatric surgery: a case series. Obes Surg 2015; 25: 2190-2199
• 10 Leeds SG, Burdick JS. Management of gastric leaks after sleeve gastrectomy with endoluminal vacuum (E-Vac) therapy. Surg Obes Relat Dis 2016; (in press) DOI: http://dx.doi.org/10.1016/j.soard.2016.01.017.
• 11 Campos JM, Pereira EF, Evangelista LF et al. Gastrobronchial fistula after sleeve gastrectomy and gastric bypass: endoscopic management and prevention. Obes Surg 2011; 21: 1520-1529
• 12 Campos JM, Ferreira FC, Teixeira AF et al. Septotomy and balloon dilation to treat chronic leak after sleeve gastrectomy: technical principles. Obes Surg 2016; 26: 1992-1993
• 13 Baretta G, Campos J, Correia S et al. Bariatric postoperative fistula: a life-saving endoscopic procedure. Surg Endosc 2015; 29: 1714-1720
• 14 Chouillard E, Chahine E, Schoucair N et al. Roux-en-Y fistulo-jejunostomy as a salvage procedure in patients with post-sleeve gastrectomy fistula. Surg Endosc 2014; 28: 1954-1960