- Email: [email protected]
Classification, surgical management and outcomes of patients with gastrogastric fistula after Roux-En-Y gastric bypass
Author’s Accepted Manuscript Classification, surgical management and outcomes of patients with gastrogastric fistula after roux-en-y gastric bypass Lara Ribeiro-Parenti, Gregoire De Courville, Abdenasser Daikha, Konstantinos Arapis, Denis Chosidow, Jean-Pierre Marmuse www.elsevier.com/locate/buildenv
PII: DOI: Reference:
S1550-7289(16)30726-2 http://dx.doi.org/10.1016/j.soard.2016.09.027 SOARD2781
To appear in: Surgery for Obesity and Related Diseases Received date: 14 March 2016 Revised date: 14 September 2016 Accepted date: 20 September 2016 Cite this article as: Lara Ribeiro-Parenti, Gregoire De Courville, Abdenasser Daikha, Konstantinos Arapis, Denis Chosidow and Jean-Pierre Marmuse, Classification, surgical management and outcomes of patients with gastrogastric fistula after roux-en-y gastric bypass, Surgery for Obesity and Related Diseases, http://dx.doi.org/10.1016/j.soard.2016.09.027 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Title of the paper: Classification, Surgical Management And Outcomes Of Patients With Gastrogastric Fistula After Roux-En-Y Gastric Bypass.
Names and academic degree for each author: Lara Ribeiro-Parenti1, M.D, Gregoire De Courville, M.D, Abdenasser Daikha2, M.D, Konstantinos Arapis1, M.D, PhD, Denis Chosidow1, M.D, Jean-Pierre Marmuse1, M.D, Ph.D. Affiliation:
1 Service
de Chirurgie Générale et Digestive, Hôpital Bichat Claude
Bernard, Paris, France. 46 Rue Henri Huchard, 75018 Paris, France. 2 Service
de Radiologie, Hôpital Bichat Claude Bernard, Paris, France. 46
Rue Henri Huchard, 75018 Paris, France. Correspondence to: Lara Ribeiro Parenti Adress: Hôpital Bichat Claude Bernard, Service Chirurgie Générale et Digestive 46 Rue Henri Huchard, 75018 Paris, France Phone : +33140256926 Fax: +33140258839 Mail : [email protected]
Short title: Gastrogastric fistula: classification and management.
Abstract: Background: Gastrogastric fistula (GGF) is a known complication after Roux-en-Y gastric bypass (RYGB) that can lead to marginal ulceration (MU) and failure of weight loss. Objective: Describe our experience with GGF management and propose a classification of GGF based on its anatomic location. Setting: University hospital, France. Methods: After internal review board approval, data from all patients with a GGF were reviewed. GGF was classified as type 1 when located in the proximal part of the gastric pouch and type 2 when located near the gastrojejunostomy. Results: Nine patients developed a GGF (0.5%). GGF symptoms included epigastric pain (78%), vomiting (11%), gastrointestinal bleeding (11%) and weight regain (44%). Upper contrast study identified GGF in all patients. Upper endoscopy confirmed GGF in 6 patients, all with type 2. Eight patients required revisional surgery. Patients with type 1 GGF (n=3) had excision of the fistulous tract. Patients with type 2 GGF (n=5) had associated revision of the gastrojejunostomy. Mean operative time was significantly longer for type 2 GGF. The mean follow-up was 43 months, with no patient lost. One patient developed a recurrent MU requiring iterative revision. After that, all revisional patients were symptom free and the mean BMI was 35.3 ± 9.5 kg/m². Conclusion: Weight regain and epigastric pain with or without associated MU are the most common signs of GGF. Combining upper GI endoscopy and contrast study is the best method to confirm the diagnosis. Surgical treatment should be tailored to both GGF location and status of the gastrojejunostomy. Based on its anatomic location, GGF classification could serve as a working basis to compare different surgical approaches. Keywords: Gastrogastric fistula, marginal ulcer, weight regain, complications, gastric bypass, revisional surgery.
Manuscript: Introduction: Gastrogastric fistula (GGF) was a well-described complication after non-divided Roux-en-Y gastric bypass (RYGB), with a documented rate of up to 50% [1]. Gastric transection with isolation of the gastric pouch (divided RYGB) reduces but doesn’t eliminate this complication that still occurs with a reported incidence ranging from 0 to 6% [2-4]. Weight regain, epigastric pain and marginal ulcer (MU) are the most common symptoms [2, 5]. Some GGF can be managed conservatively or by endoscopic approach but surgery remains the standard of care. To date there are only a few studies reporting the incidence and the management of GGF after divided RYGB and several come from the same institution [2-4, 6-8]. Moreover the surgical management remains controversial with different favoured approaches including: excision of the fistula [3], anastomotic revision with fistulous tract excision, [8] and remnant gastrectomy with excision or exclusion of the fistula [8]. We report the incidence and our experience with GGF management after divided laparoscopic RYGB. We also suggest an anatomic classification of the GGF, which could serve for a consensus on the management of this uncommon complication. Patients and methods: From November 2000 to January 2015 all the patients who underwent laparoscopic RYGB were entered in a prospective database. After internal review board approval, we retrospectively reviewed data of all patients who were diagnosed with GGF. Data abstracted for analysis included patient characteristics, presenting symptoms, initial diagnostic studies, time from surgery to diagnosis and to revisional surgery. Moreover, we retrospectively searched all signs that could be symptomatic of a leak, during their postoperative course. According to endoscopic, radiologic and surgical findings GGF were classified as follow:
Type 1: located in the proximal part of the gastric pouch more than 2 cm above the gastrojejunostomy. Type 2: located in the distal part of the gastric pouch less than 2cm above the gastrojejunostomy. This cut-off was chosen because when GGF are located on the lateral side of the pouch and do not involve the gastrojejunostomy, with at least 1 or 2 cm of heathy well vascularized tissue between the gastrojejunostomy and the fistula, the pouch can be vertically transsected by firing the linear stapler to exclude the fistula from the pouch with no risk of anastomosis narrowing. In contrast when the fistula is too close to the gastrojejunal anastomosis a revision with creation of a new gastrojejunostomy must be considered to avoid anastomotic damage. Outcomes data assessed were length of hospital stay, early and late complications, weight loss after revision and body mass index (BMI) at last follow-up. Surgical revisional technique: (Fig 1). Two types of revisional procedures were performed: (a) simple resection of the fistula or (b) resection of the fistula with complete revision of the gastrojejunal anastomosis. In the first procedure the pouch was vertically transected medial to the GGF over a 36 French calibration tube. Then on the other side of the fistula the remnant was also vertically transected lateral to the GGF. The second procedure was performed in this series as described by Corcelles et al [8]. Here the fistula and the gastrojejunostomy were resected en bloc with care to leave enough gastric remnant to preserve acid remnant secretion and prevent a retained antrum syndrome. Then a new anastomosis was created with a 25-mm circular, transoral, endto-side stapled anastomosis (Fig 1). Whatever the procedure, a leak test was performed with methylene blue dye injected in the gastric pouch. No patient had gastrostomy or jejunostomy for postoperative enteral feeding. Statistical methods:
Categorical variables were reported as frequencies (%). Continuous variables were expressed as mean ± SD, range. Chi-square test or Fisher’s exact test were performed for categorical data. A p-value of less than 0.05 was considered statistically significant.
Results: During the study period, 1900 consecutive patients underwent divided RYGB; 1890 were completed laparoscopically and 10 converted to open. Nine patients developed a GGF (0.5%). Demographic data are detailed in Table 1. Among them, one used to smoke, one had previous surgery for craniopharyngioma and one had been converted to open RYGB. None had had preoperative Helicobacter Pylori infection and none used nonsteroidal anti-inflammatory drugs. One had developed a postoperative gastric leak managed nonoperatively. For the remaining 8 patients no leak was documented before discharge but none had strictly uneventful post-operative course. Three had a brief episode of fever, 4 a sustained tachycardia and 6 leukocytosis with elevated C reactive protein. In all of them septic work-ups (including CT and upper contrast study) failed to identify the source of their symptoms, which abated before discharge after a mean hospital stay of 8.6 days (range 6-11). GGF associated symptoms were: epigastric pain (n=7/9), vomiting (n=1/9), gastrointestinal bleeding (n=1/9) and weight regain, defined as more than 20% of excess weight loss, (n=5/9). For these 9 patients the mean nadir weight after RYGB was 91± 23 kg versus a mean weight of 107 ± 37 kg at the time of GGF diagnosis. Likewise the mean BMI was significantly higher in the 5 patients with weight regain compared to the others for; 45 ± 13 kg/m² versus 36 ± 9 kg/m². All these 9 patients underwent both upper gastrointestinal (UGI) contrast studies and upper endoscopy. UGI study identified GGF in all patients. In one case the first UGI study was negative and the second, 3 months later, showed contrast within the remnant. Upper endoscopy confirmed GGF in 6 patients. All of them had a GGF located near the
gastrojejunostomy and an associated MU. For the remaining 3 patients, upper endoscopy was negative for GGF. According to radiologic and endoscopic findings GGF was located in the upper pouch (type 1) in 3 patients and in the distal pouch (type2) in six (Table 2). Regardless of symptoms, all patients were treated with PPI, associated with sucralfate in case of concomitant MU. One is still on medical therapy because she denied revision as she maintained good weight loss without abdominal discomfort and 8 had revisional surgery. Indications for reoperation were weight regain in 5 patients, intractable epigastric pain in 2 and gastrointestinal bleeding in one. Five had resection of both the fistula and the gastrojejunostomy with anastomosis revision and 3 had only a fistula excision. Seven procedures (87.5%) were performed laparoscopically and one by laparotomy because of previous open procedures. The mean postoperative hospital stay was 8.1 ± 2.4 days. There were no deaths and no intraoperative complication. The only early complication was a leak leading to systemic sepsis which required reoperation on post-operative day (POD) 7. As the gastric pouch was too small to allow a secure second anastomotic revision we performed a pouch resection with an oesophagojejunostomy. Postoperative course was uneventful and the patient was discharged on POD 14. Late complication occurred in one patient who had a MU, revealed by epigastric pain, five months after surgical revision. He was initially treated by PPI combined with sucralfate but MU and epigastric pain persisted and a second revision was considered. Despite a normal fasting serum gastrin level as he had no recurrent GGF or dilated pouch on repeated imaging and endoscopic studies we hypothesized that he had a retained gastric antrum syndrome (RGA) due to an oversized remnant resection, leaving an excluded antrum, during the revision for GGF. Thus we performed a second revision with total remnant gastrectomy and creation of a new gastrojejunostomy. Postoperative course was uneventful and more than 5 years later, this patient was still asymptomatic without PPI therapy. During these 5 years
follow-up he underwent several endoscopic controls, two the first year and thereafter once a year, all negative for MU. No other patient had complication during follow-up. Early and late complications occurred only in patients who had anastomotic revision. No complications were observed in those who had fistula excision (Table 2). After revision, 4 patients lost additional weight, mean 20 kg (range 11-41 kg), 3 maintained their pre-revision weight and one presented weight regain related to persistent food compulsion after craniopharyngioma resection. The mean follow-up was 43 months (range 20-72) and at the time of study, all patients had recent office visit. At the last follow-up point, all were asymptomatic, no one used PPI, and the mean BMI for the entire group was 35.3 ± 9.5 kg/m².
Discussion: In our series of 1900 consecutives patients who had divided RYGB, 9 developed a GGF (0.5%) a relatively low incidence compared to others series [2-4]. However the true incidence in most study, including our own, are probably underestimated as asymptomatic patients can achieve good durable weight loss [9-11]. Pathogenesis of GGF after divided RYGB is not clearly understood. Various causes have been advocated; such failure of complete gastric transection, foreign body erosion, and distension of the gastric pouch related to gastric outlet obstruction [5, 7, 12, 13] but the main reported causes remain gastric leak and MU. Indeed GGF may be caused by a leak with abscess formation, which drains spontaneously into the distal stomach [5]. Of the 6 patients reported by Cucchi et al [5] 5 had developed a documented leak (83%) before diagnosis of the GGF. Similarly in the study of Carrodeguas et al [2], 4 of the 15 reported patients (27%) had had a postoperative leak. In our series, only 1 patient had a documented leak (11%). However, none of the remaining patients had a strictly uneventful postoperative primary course. Some had probably a small leak internally drained by the
development of a GGF before occurrence of a systemic sepsis. It has been also reported that MU and GGF are closely interconnected [1, 3, 8]. Usually, MU is considered as a consequence of the GGF because the acid from the gastric remnant spills over through the GGF and then contributes to MU formation. On the other hand some authors argue that the inclusion of parietal cell mass in the gastric pouch can lead to acid exposure, and then to a GGF in case of penetrating MU [14]. Patients with GGF may present nonspecific symptoms including vomiting, pyrosis, epigastric pain and hematemesis [2]. Others may be asymptomatic with only failure to lose weight or weight regain and others achieve good weight loss despite an obvious GGF [2, 10, 11]. These clinical presentations, varying from one patient to another, explain why the diagnosis may be challenging and delayed [1, 7, 15]. In our study, time to GGF diagnosis ranged from 1 to 87 months and the most common observed symptoms were epigastric pain, often related to associated MU, and weight regain [6, 15-17]. UGI contrast study allowed us to detect the GGF in all the cases. However negative UGI study does not rule out the diagnosis. For Corcelles et al [8], UGI series was falsely negative in 30% of the cases. In the same study endoscopy was also falsely negative in 27 % of the cases. In our study, upper endoscopy was falsely negative in 3 patients, all with a GGF located at the upper end of the staple line (type 1). By contrast, all the GGF in the distal pouch (type 2) were identified and all were associated with MU. Diagnosing a GGF is not per se an indication for prompt surgical revision but all symptomatic GGF must be initially treated by PPI therapy combined with sucralfate in case of associated MU. This may eliminate abdominal pain, may allow MU to resolve or prevent it and sometimes lead small GGF to close spontaneously [3]. Moreover if GGF’s symptoms are well controlled surgery may be postponed as long as patients achieve good weight loss [2, 10]. It should be noted however that most GGF don’t close spontaneously and that prolonged PPI use might have potential adverse effects [18]. To date, 4 years after the diagnosis, one of our
patients is still on medical therapy because she denied revision as she maintained good weight loss without abdominal discomfort despite persistent GGF. Several studies have shown that GGF can be managed by endoscopy and given its conservative nature it can be a good initial approach, especially for small GGF[15, 19, 21]. Different methods have been used including fibrin sealant, endoclips and endoscopic suturing systems, with often initial success but poor long-term results. Fernandez Esparrach et al. [19] achieved initial closure of GGF in 90 of 95 patients (94%) with endoscopic suturing system or clips, but after a mean follow-up of 395 days only 10 remain closed. They concluded that per-oral endoscopic repair of GGF is feasible and safe but with limited durability. TorresVillalobos et al [20] used an endoscope passed through the GGF to introduce, via a percutaneous trans-gastric route, laparoscopic instruments in the remnant to repair GGF in 3 patients. No complications were reported but the GGF recurred in two patients. Thus for most authors surgery remains the standard of care and the main reported indications are persistent symptoms despite an optimized medical management, weight regain and failure of endoscopic repair. Conversely there is no accepted surgical technique. Some performed an excision of the fistula [1, 3], others an anastomotic revision with fistulous tract excision [8] and others a remnant gastrectomy with excision or exclusion of the fistula [2, 4, 6, 7]. Most authors also highlight that these procedures are challenging, because of the extensive inflammation around the fistulous tract, with as consequence an increased operative time, an increased length of hospital stay and a high early major complications rates ranging from 17 to 29% [21, 22]
. In our unit we performed either a fistula excision or a resection of the fistula and the
gastrojejunostomy with anastomosis recreation. Our goal in both procedures was a complete resection of all involved tissues to well vascularized tissues with care to leave enough remnant to avoid a retained antrum [23, 24]. No patient had fistula exclusion as we have concerns with this procedure that left in situ a narrow cuff of poorly vascularized remnant
lateral to the GGF. The choice of the procedure was done during surgery according to the location of the GGF and the anastomotic status. For type 1 GGF, which allow GGF resection without anastomotic damage we performed a simple excision by stapling both ends of the fistula as they were not associated with persistent MU or anastomotic stenosis. For type 2 GGF we performed a complete revision of the gastrojejunostomy with fistula excision. Except one done by laparotomy all these procedures were successfully completed laparoscopically. As previously reported by others, these operations were challenging especially when it required anastomosis revision. In these cases the mean operative time, which is a surrogate marker of technical complexity, was significantly longer than for simple fistula excision; 204 ± 58 min versus 50 ± 9 min. No mortality was observed and the only early complication was a leak, which required reoperation. Postoperative course was uneventful in the others patients and the mean length of stay for the entire group was 8.1 ± 2.4 days. Five months after GGF excision with anastomosis revision one patient developed a MU. Medical therapy was ineffective, despite high PPI dose and as this patient had, no recurrent GGF, no dilated gastric pouch and no anastomotic stenosis we hypothesized that he had a retained gastric antrum syndrome (RGA) due to an oversized remnant resection, during GGF treatment procedure. This syndrome is a well-known complication of an old surgical technique, with transection of the stomach, end to side gastrojejunostomy and retention of the antrum in continuity with the duodenum which was called “antral exclusion” [25] . Performed for giant duodenal ulcer and popularized by Devine in 1925, this procedure was abandoned because it was followed by an unacceptably high incidence of anastomotic ulcers (around 40%) due to a gastrin-mediated acid hypersecretion[26]. Indeed when antrum remains excluded from the gastric acid stream the antral producing G cells will release gastrin without negative feed-back with as consequence hypergastrinemia and acid hypersecretion in the remaining stomach. Our patient had normal fasting serum gastrin level but in the literature up to 20% of patients with RGA
have normal levels of fasting gastrin [24, 27]. As a normal gastrin level does not exclude this diagnosis we performed a second revision, with total remnant gastrectomy and creation of a new gastrojejunostomy. Postoperative course was uneventful with immediate symptoms resolution. The iterative upper endoscopies performed during follow-up were normal and more than 5 years after antrectomy the patient is still symptom free without PPI therapy. To avoid this complication, we should have performed a total remnant gastrectomy instead of a too large partial remnant resection leaving an excluded antrum during the first revision. No other complications were observed during follow-up. Early and late complications occurred both after revision with creation of a new anastomosis. No complications were observed after fistula excision. After revision 4 patients (50%) lost additional weight mean 20 kg (range 1141 kg). The others maintained their preoperative weight except one who regained weight because she had persistent food compulsions after craniopharyngioma resection [28, 29]. No patient was lost to follow-up. At the last follow-up point (mean 43 months, range 20 to 72 months) all the patients were symptom free with no one PPI use. Although this study was limited by its retrospective design and included a relatively small number of patients, we have longer follow-up than other reported studies. Conclusion: Gastro-gastric fistula is a rare complication after divided RYGB. Weight regain and epigastric pain with or without marginal ulcer are the most common presenting symptoms. Diagnosis can be challenging but combining upper GI endoscopy and contrast study is the best method to confirm the diagnosis. Surgical treatment should be tailored to both location of the GGF and status of the gastrojejunostomy, but anastomotic revision is associated with higher complication rates. GGF classification, based on its anatomic location, could serve as a working basis to compare different surgical approaches.
Disclosures None of the authors have anything to disclose. References: 1.
Capella J.F, Capella R.F. Gastro-gastric fistulas and marginal ulcers in gastric bypass procedures for weight reduction. Obes Surg 1999; 9(1):22-7; discussion 28.
2.
Carrodeguas L, Szomstein S, Soto F, et al. Management of gastrogastric fistulas after divided Roux-en-Y gastric bypass surgery for morbid obesity: analysis of 1,292 consecutive patients and review of literature. Surg Obes Relat Dis 2005; 1(5) : 467474.
3.
Gumbs A.A, Duffy A.J, Bell R.L. Management of gastrogastric fistula after laparoscopic Roux-en-Y gastric bypass. Surg Obes Relat Dis 2006; 2(2):117-121.
4.
Salimath J, Rosenthal R.J, Szomstein S. Laparoscopic remnant gastrectomy as a novel approach for treatment of gastrogastric fistula. Surg Endosc 2009; 23(11): 2591-2595.
5.
Cucchi S G, Pories W.J, MacDonald K.G, Morgan E.J. Gastrogastric fistulas. A complication of divided gastric bypass surgery. Ann Surg 1995; 221(4): 387-391.
6.
Cho M, Kaidar-Person O, Szomstein S, Rosenthal R. J. Laparoscopic remnant gastrectomy: a novel approach to gastrogastric fistula after Roux-en-Y gastric bypass for morbid obesity. J Am Coll Surg 2007; 204(4): 617-624.
7.
Tucker O.N, Szomstein S, Rosenthal R.J. Surgical management of gastro-gastric fistula after divided laparoscopic Roux-en-Y gastric bypass for morbid obesity. J Gastrointest Surg 2007; 11(12)/ 1673-1679.
8.
Corcelles R, Jamal M.H, Daigle C. R., Rogula T, Brethauer S.A, Schauer P.R. Surgical management of gastrogastric fistula. Surg Obes Relat Dis 2015; 11(6):122732.
9.
Carucci L.R, Turner M.A. Radiologic evaluation following Roux-en-Y gastric bypass surgery for morbid obesity. Eur J Radiol 2005; 53(3) : 353-365.
10.
Gustavsson S, Sundbom M. Excellent weight result after Roux-en-Y gastric bypass in spite of gastro-gastric fistula. Obes Surg 2003; 13(3): 457-459.
11.
Stanczyk M, Deveney C.W, Traxler S.A, McConnell D.B, Jobe B.A, O'Rourke R.W. Gastro-gastric fistula in the era of divided Roux-en-Y gastric bypass: strategies for prevention, diagnosis, and management. Obes Surg 2006; 16(3): 359-364.
12.
Gomez C.A. Gastroplasty in morbid obesity. Surg Clin North Am 1979; 59(6): 11131120.
13.
Gould J.C, Garren M.J, Starling, J.R. Lessons learned from the first 100 cases in a new minimally invasive bariatric surgery program. Obes Surg 2004; 14(5): 618-625.
14.
Ito Y, Blackstone M.O, Hatfield G.E. A gastrogastric fistula linking two penetrating ulcers. Report of a case. Gastrointest Endosc 1978; 24(5): 247-248.
15
Pauli E.M, Beshir H, Mathew A. Gastrogastric fistulae following gastric bypass surgery-clinical recognition and treatment. Curr Gastroenterol Rep 2014; 16(9): 405.
16.
Brethauer S.A, Nfonsam V, Sherman V, Udomsawaengsup S, Schauer P.R, Chand B. Endoscopy and upper gastrointestinal contrast studies are complementary in evaluation of weight regain after bariatric surgery. Surg Obes Relat Dis 2006; 2(6): 643-648; discussion 649-650.
17.
Chandler R.C, Srinivas G, Chintapalli K.N, Schwesinger W.H, Prasad S.R. Imaging in bariatric surgery: a guide to postsurgical anatomy and common complications. AJR Am J Roentgenol 2008; 190(1): 122-135.
18.
Sheen E, Triadafilopoulos G. Adverse effects of long-term proton pump inhibitor therapy. Dig Dis Sci 2011; 56(4): 931-950.
19.
Fernandez-Esparrach G, Lautz D.B, Thompson C.C. Endoscopic repair of gastrogastric fistula after Roux-en-Y gastric bypass: a less-invasive approach. Surg Obes Relat Dis 2010; 6(3): 282-288.
20.
Bhardwaj A, Cooney R.N, Wehrman A, Rogers A.M, Mathew A. Endoscopic repair of small symptomatic gastrogastric fistulas after gastric bypass surgery: a single center experience. Obes Surg 2010; 20(8): 1090-1095.
21.
Spaun G.O, Martinec D.V, Kennedy T.J, Swanstrom L.L. Endoscopic closure of gastrogastric fistulas by using a tissue apposition system (with videos). Gastrointest Endosc 2010; 71(3): 606-611.
22.
Torres-Villalobos G, Leslie D.B, Hunter D, Buchwald H, Martin-del-Campo L.A, Ikramuddin S. Preliminary report: search for a transgastric approach for managing gastrogastric fistulas. World J Surg 2011; 35(2): 372-376.
23.
Nesset E.M, Kendrick M.L, Houghton S.G et al. A two-decade spectrum of revisional bariatric surgery at a tertiary referral center. Surg Obes Relat Dis 2007; 3(1): 25-30; discussion 30.
24.
Gagner M, Gentileschi P, De Csepel J et al. Laparoscopic reoperative bariatric surgery: experience from 27 consecutive patients. Obes Surg 2002; 12(2): 254-260.
25.
Lee C.H, P'Eng F K, Lui W.Y. The clinical aspect of retained gastric antrum. Arch Surg 1986; 121(10): 1181-1186.
26.
Gibril F, Lindeman R.J, Abou-Saif A et al. Retained gastric antrum syndrome: a forgotten, treatable cause of refractory peptic ulcer disease. Dig Dis Sci 2001; 46(3): 610-617.
27.
Ribeiro-Parenti L, Arapis k, Chosidow D, Marmuse JP. Comparison of marginal ulcer rates between antecolic and retrocolic laparoscopic Roux-en-Y gastric bypass. Obes Surg, 2015. 25(2): p. 215-21.
28.
Csendes A, Burgos AM, Altuve J, Bonacic S. Incidence of marginal ulcer 1 month and 1 to 2 years after gastric bypass: a prospective consecutive endoscopic evaluation of 442 patients with morbid obesity. Obes Surg, 2009. 19(2): p. 135-8.
29.
Devine HB. Basic principles and supreme difficulties in gastric surgery. Surg., Gynec. & Obst., 1925. 40: p. 1-16.
30.
McKittrick LS, Moore FD, Warren R. Complications and Mortality in Subtotal Gastrectomy for Duodenal Ulcer: Report on a Two-Stage Procedure. Ann Surg, 1944. 120(4): p. 531-54.
31
Burhenne HJ. The retained antrum, preoperative roentgenologic diagnosis of an iatrogenic syndrome. AJR , 1967. 101(2): p. 459-466.
32
Shyu JF, Chen TH, Shyr YM, Su Ch, Wu C, Lui W. Gastric body partition to avoid ulcerogenic risk and hypergastrinemia. Surgery, 2006. 140(1): p. 44-9.
33.
Skorzewska A, Lal S, Waserman J, Guyda H. Abnormal food-seeking behavior after surgery for craniopharyngioma. Neuropsychobiology 1989; 21(1) : 17-20.
34.
Rottembourg D, O'Gorman C.S, Urbach S et al. Outcome after bariatric surgery in two adolescents with hypothalamic obesity following treatment of craniopharyngioma. J Pediatr Endocrinol Metab 2009; 22(9): 867-872.
Figure 1 legend: a. Resection of the fistula, in patients with type 1 GGF (i.e. located at the proximal part of the gastric pouch) b. En-bloc resection of both gastrogastric fistula and gastrojejunostomy, in patients with type 2 GGF (i.e. located at the distal part of the gastric pouch, near the gastrojejunostomy).
Table 1. Demographics data Patients with GGF N (%) Mean age (years) Gender (M/F)
9/1900 (0.5%) 39 ± 5 3/6
Mean preoperative weight (kg) Mean preoperative BMI (kg/m2)
154 ± 42 56 ± 12
Mean nadir weight (kg) Mean nadir BMI (kg/m2)
92 ± 23 33 ± 6
Mean weight at reoperation (kg) Mean BMI at reoperation (kg/m2)
107 ± 37 39 ± 10
Symptoms epigastric pain (n) vomiting (n) bleeding (n) weight regain (n) GGF, gastrogastric fistula; M, male; F, female; BMI, body mass index;
7 1 1 5
Table 2. Clinical presentation, preoperative workup and results of GGF management according to GGF location.
Clinical presentation
Time from primary to diagnosis
Type 1
Type 2
(n=3)
(n=6)
Mild epigastric pain
Intractable epigastric pain
Weight regain
Recurrent bleeding
30 (1-87)
10 (1-20)*
3 (100) Upper part of the pouch
6 (100) Near the GJ
0 (0) 0 (0) 1 (33)
6 (100)* 6 (100)* 0 (0)
40.6 (25-62.5)
11 (1-36)*
3 0 0 3
5
Months, (range)
Upper contrast study GGF identification, n, (%) GGF location Upper endoscopy GGF identification, n, (%) Associated MU, n, (%) Associated gastritis n, (%) Time from surgery to revision months (range) Indications for surgery, n Abdominal pain, n Recurrent bleeding, n Weight regain, n Revisional surgery Laparoscopy, n Operative time, minutes Early morbidity, n Late morbidity, n Additional weight loss, n Mean BMI last follow-up point (kg/m2)
Fistulous tract resection 3 50 ± 9 0/3 0/3 2/3 41.3 ± 10.1
2 1 2 En bloc resection of GGF and GJ New GJ 4 204 ± 58* 1/5 1/5 2/5 33.6 ± 9.5
GGF, gastrogastric fistula; MU, marginal ulcer; GJ, gastrojejunostomy; * p<0.05, significant. Type 1, GGF located in the proximal part of the gastric pouch; Type 2, GGF located in the distal part of the gastric pouch
a. b. Figure 1. Surgical revisional technique according to gastrogastric fistula location
PII: DOI: Reference:
S1550-7289(16)30726-2 http://dx.doi.org/10.1016/j.soard.2016.09.027 SOARD2781
To appear in: Surgery for Obesity and Related Diseases Received date: 14 March 2016 Revised date: 14 September 2016 Accepted date: 20 September 2016 Cite this article as: Lara Ribeiro-Parenti, Gregoire De Courville, Abdenasser Daikha, Konstantinos Arapis, Denis Chosidow and Jean-Pierre Marmuse, Classification, surgical management and outcomes of patients with gastrogastric fistula after roux-en-y gastric bypass, Surgery for Obesity and Related Diseases, http://dx.doi.org/10.1016/j.soard.2016.09.027 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Title of the paper: Classification, Surgical Management And Outcomes Of Patients With Gastrogastric Fistula After Roux-En-Y Gastric Bypass.
Names and academic degree for each author: Lara Ribeiro-Parenti1, M.D, Gregoire De Courville, M.D, Abdenasser Daikha2, M.D, Konstantinos Arapis1, M.D, PhD, Denis Chosidow1, M.D, Jean-Pierre Marmuse1, M.D, Ph.D. Affiliation:
1 Service
de Chirurgie Générale et Digestive, Hôpital Bichat Claude
Bernard, Paris, France. 46 Rue Henri Huchard, 75018 Paris, France. 2 Service
de Radiologie, Hôpital Bichat Claude Bernard, Paris, France. 46
Rue Henri Huchard, 75018 Paris, France. Correspondence to: Lara Ribeiro Parenti Adress: Hôpital Bichat Claude Bernard, Service Chirurgie Générale et Digestive 46 Rue Henri Huchard, 75018 Paris, France Phone : +33140256926 Fax: +33140258839 Mail : [email protected]
Short title: Gastrogastric fistula: classification and management.
Abstract: Background: Gastrogastric fistula (GGF) is a known complication after Roux-en-Y gastric bypass (RYGB) that can lead to marginal ulceration (MU) and failure of weight loss. Objective: Describe our experience with GGF management and propose a classification of GGF based on its anatomic location. Setting: University hospital, France. Methods: After internal review board approval, data from all patients with a GGF were reviewed. GGF was classified as type 1 when located in the proximal part of the gastric pouch and type 2 when located near the gastrojejunostomy. Results: Nine patients developed a GGF (0.5%). GGF symptoms included epigastric pain (78%), vomiting (11%), gastrointestinal bleeding (11%) and weight regain (44%). Upper contrast study identified GGF in all patients. Upper endoscopy confirmed GGF in 6 patients, all with type 2. Eight patients required revisional surgery. Patients with type 1 GGF (n=3) had excision of the fistulous tract. Patients with type 2 GGF (n=5) had associated revision of the gastrojejunostomy. Mean operative time was significantly longer for type 2 GGF. The mean follow-up was 43 months, with no patient lost. One patient developed a recurrent MU requiring iterative revision. After that, all revisional patients were symptom free and the mean BMI was 35.3 ± 9.5 kg/m². Conclusion: Weight regain and epigastric pain with or without associated MU are the most common signs of GGF. Combining upper GI endoscopy and contrast study is the best method to confirm the diagnosis. Surgical treatment should be tailored to both GGF location and status of the gastrojejunostomy. Based on its anatomic location, GGF classification could serve as a working basis to compare different surgical approaches. Keywords: Gastrogastric fistula, marginal ulcer, weight regain, complications, gastric bypass, revisional surgery.
Manuscript: Introduction: Gastrogastric fistula (GGF) was a well-described complication after non-divided Roux-en-Y gastric bypass (RYGB), with a documented rate of up to 50% [1]. Gastric transection with isolation of the gastric pouch (divided RYGB) reduces but doesn’t eliminate this complication that still occurs with a reported incidence ranging from 0 to 6% [2-4]. Weight regain, epigastric pain and marginal ulcer (MU) are the most common symptoms [2, 5]. Some GGF can be managed conservatively or by endoscopic approach but surgery remains the standard of care. To date there are only a few studies reporting the incidence and the management of GGF after divided RYGB and several come from the same institution [2-4, 6-8]. Moreover the surgical management remains controversial with different favoured approaches including: excision of the fistula [3], anastomotic revision with fistulous tract excision, [8] and remnant gastrectomy with excision or exclusion of the fistula [8]. We report the incidence and our experience with GGF management after divided laparoscopic RYGB. We also suggest an anatomic classification of the GGF, which could serve for a consensus on the management of this uncommon complication. Patients and methods: From November 2000 to January 2015 all the patients who underwent laparoscopic RYGB were entered in a prospective database. After internal review board approval, we retrospectively reviewed data of all patients who were diagnosed with GGF. Data abstracted for analysis included patient characteristics, presenting symptoms, initial diagnostic studies, time from surgery to diagnosis and to revisional surgery. Moreover, we retrospectively searched all signs that could be symptomatic of a leak, during their postoperative course. According to endoscopic, radiologic and surgical findings GGF were classified as follow:
Type 1: located in the proximal part of the gastric pouch more than 2 cm above the gastrojejunostomy. Type 2: located in the distal part of the gastric pouch less than 2cm above the gastrojejunostomy. This cut-off was chosen because when GGF are located on the lateral side of the pouch and do not involve the gastrojejunostomy, with at least 1 or 2 cm of heathy well vascularized tissue between the gastrojejunostomy and the fistula, the pouch can be vertically transsected by firing the linear stapler to exclude the fistula from the pouch with no risk of anastomosis narrowing. In contrast when the fistula is too close to the gastrojejunal anastomosis a revision with creation of a new gastrojejunostomy must be considered to avoid anastomotic damage. Outcomes data assessed were length of hospital stay, early and late complications, weight loss after revision and body mass index (BMI) at last follow-up. Surgical revisional technique: (Fig 1). Two types of revisional procedures were performed: (a) simple resection of the fistula or (b) resection of the fistula with complete revision of the gastrojejunal anastomosis. In the first procedure the pouch was vertically transected medial to the GGF over a 36 French calibration tube. Then on the other side of the fistula the remnant was also vertically transected lateral to the GGF. The second procedure was performed in this series as described by Corcelles et al [8]. Here the fistula and the gastrojejunostomy were resected en bloc with care to leave enough gastric remnant to preserve acid remnant secretion and prevent a retained antrum syndrome. Then a new anastomosis was created with a 25-mm circular, transoral, endto-side stapled anastomosis (Fig 1). Whatever the procedure, a leak test was performed with methylene blue dye injected in the gastric pouch. No patient had gastrostomy or jejunostomy for postoperative enteral feeding. Statistical methods:
Categorical variables were reported as frequencies (%). Continuous variables were expressed as mean ± SD, range. Chi-square test or Fisher’s exact test were performed for categorical data. A p-value of less than 0.05 was considered statistically significant.
Results: During the study period, 1900 consecutive patients underwent divided RYGB; 1890 were completed laparoscopically and 10 converted to open. Nine patients developed a GGF (0.5%). Demographic data are detailed in Table 1. Among them, one used to smoke, one had previous surgery for craniopharyngioma and one had been converted to open RYGB. None had had preoperative Helicobacter Pylori infection and none used nonsteroidal anti-inflammatory drugs. One had developed a postoperative gastric leak managed nonoperatively. For the remaining 8 patients no leak was documented before discharge but none had strictly uneventful post-operative course. Three had a brief episode of fever, 4 a sustained tachycardia and 6 leukocytosis with elevated C reactive protein. In all of them septic work-ups (including CT and upper contrast study) failed to identify the source of their symptoms, which abated before discharge after a mean hospital stay of 8.6 days (range 6-11). GGF associated symptoms were: epigastric pain (n=7/9), vomiting (n=1/9), gastrointestinal bleeding (n=1/9) and weight regain, defined as more than 20% of excess weight loss, (n=5/9). For these 9 patients the mean nadir weight after RYGB was 91± 23 kg versus a mean weight of 107 ± 37 kg at the time of GGF diagnosis. Likewise the mean BMI was significantly higher in the 5 patients with weight regain compared to the others for; 45 ± 13 kg/m² versus 36 ± 9 kg/m². All these 9 patients underwent both upper gastrointestinal (UGI) contrast studies and upper endoscopy. UGI study identified GGF in all patients. In one case the first UGI study was negative and the second, 3 months later, showed contrast within the remnant. Upper endoscopy confirmed GGF in 6 patients. All of them had a GGF located near the
gastrojejunostomy and an associated MU. For the remaining 3 patients, upper endoscopy was negative for GGF. According to radiologic and endoscopic findings GGF was located in the upper pouch (type 1) in 3 patients and in the distal pouch (type2) in six (Table 2). Regardless of symptoms, all patients were treated with PPI, associated with sucralfate in case of concomitant MU. One is still on medical therapy because she denied revision as she maintained good weight loss without abdominal discomfort and 8 had revisional surgery. Indications for reoperation were weight regain in 5 patients, intractable epigastric pain in 2 and gastrointestinal bleeding in one. Five had resection of both the fistula and the gastrojejunostomy with anastomosis revision and 3 had only a fistula excision. Seven procedures (87.5%) were performed laparoscopically and one by laparotomy because of previous open procedures. The mean postoperative hospital stay was 8.1 ± 2.4 days. There were no deaths and no intraoperative complication. The only early complication was a leak leading to systemic sepsis which required reoperation on post-operative day (POD) 7. As the gastric pouch was too small to allow a secure second anastomotic revision we performed a pouch resection with an oesophagojejunostomy. Postoperative course was uneventful and the patient was discharged on POD 14. Late complication occurred in one patient who had a MU, revealed by epigastric pain, five months after surgical revision. He was initially treated by PPI combined with sucralfate but MU and epigastric pain persisted and a second revision was considered. Despite a normal fasting serum gastrin level as he had no recurrent GGF or dilated pouch on repeated imaging and endoscopic studies we hypothesized that he had a retained gastric antrum syndrome (RGA) due to an oversized remnant resection, leaving an excluded antrum, during the revision for GGF. Thus we performed a second revision with total remnant gastrectomy and creation of a new gastrojejunostomy. Postoperative course was uneventful and more than 5 years later, this patient was still asymptomatic without PPI therapy. During these 5 years
follow-up he underwent several endoscopic controls, two the first year and thereafter once a year, all negative for MU. No other patient had complication during follow-up. Early and late complications occurred only in patients who had anastomotic revision. No complications were observed in those who had fistula excision (Table 2). After revision, 4 patients lost additional weight, mean 20 kg (range 11-41 kg), 3 maintained their pre-revision weight and one presented weight regain related to persistent food compulsion after craniopharyngioma resection. The mean follow-up was 43 months (range 20-72) and at the time of study, all patients had recent office visit. At the last follow-up point, all were asymptomatic, no one used PPI, and the mean BMI for the entire group was 35.3 ± 9.5 kg/m².
Discussion: In our series of 1900 consecutives patients who had divided RYGB, 9 developed a GGF (0.5%) a relatively low incidence compared to others series [2-4]. However the true incidence in most study, including our own, are probably underestimated as asymptomatic patients can achieve good durable weight loss [9-11]. Pathogenesis of GGF after divided RYGB is not clearly understood. Various causes have been advocated; such failure of complete gastric transection, foreign body erosion, and distension of the gastric pouch related to gastric outlet obstruction [5, 7, 12, 13] but the main reported causes remain gastric leak and MU. Indeed GGF may be caused by a leak with abscess formation, which drains spontaneously into the distal stomach [5]. Of the 6 patients reported by Cucchi et al [5] 5 had developed a documented leak (83%) before diagnosis of the GGF. Similarly in the study of Carrodeguas et al [2], 4 of the 15 reported patients (27%) had had a postoperative leak. In our series, only 1 patient had a documented leak (11%). However, none of the remaining patients had a strictly uneventful postoperative primary course. Some had probably a small leak internally drained by the
development of a GGF before occurrence of a systemic sepsis. It has been also reported that MU and GGF are closely interconnected [1, 3, 8]. Usually, MU is considered as a consequence of the GGF because the acid from the gastric remnant spills over through the GGF and then contributes to MU formation. On the other hand some authors argue that the inclusion of parietal cell mass in the gastric pouch can lead to acid exposure, and then to a GGF in case of penetrating MU [14]. Patients with GGF may present nonspecific symptoms including vomiting, pyrosis, epigastric pain and hematemesis [2]. Others may be asymptomatic with only failure to lose weight or weight regain and others achieve good weight loss despite an obvious GGF [2, 10, 11]. These clinical presentations, varying from one patient to another, explain why the diagnosis may be challenging and delayed [1, 7, 15]. In our study, time to GGF diagnosis ranged from 1 to 87 months and the most common observed symptoms were epigastric pain, often related to associated MU, and weight regain [6, 15-17]. UGI contrast study allowed us to detect the GGF in all the cases. However negative UGI study does not rule out the diagnosis. For Corcelles et al [8], UGI series was falsely negative in 30% of the cases. In the same study endoscopy was also falsely negative in 27 % of the cases. In our study, upper endoscopy was falsely negative in 3 patients, all with a GGF located at the upper end of the staple line (type 1). By contrast, all the GGF in the distal pouch (type 2) were identified and all were associated with MU. Diagnosing a GGF is not per se an indication for prompt surgical revision but all symptomatic GGF must be initially treated by PPI therapy combined with sucralfate in case of associated MU. This may eliminate abdominal pain, may allow MU to resolve or prevent it and sometimes lead small GGF to close spontaneously [3]. Moreover if GGF’s symptoms are well controlled surgery may be postponed as long as patients achieve good weight loss [2, 10]. It should be noted however that most GGF don’t close spontaneously and that prolonged PPI use might have potential adverse effects [18]. To date, 4 years after the diagnosis, one of our
patients is still on medical therapy because she denied revision as she maintained good weight loss without abdominal discomfort despite persistent GGF. Several studies have shown that GGF can be managed by endoscopy and given its conservative nature it can be a good initial approach, especially for small GGF[15, 19, 21]. Different methods have been used including fibrin sealant, endoclips and endoscopic suturing systems, with often initial success but poor long-term results. Fernandez Esparrach et al. [19] achieved initial closure of GGF in 90 of 95 patients (94%) with endoscopic suturing system or clips, but after a mean follow-up of 395 days only 10 remain closed. They concluded that per-oral endoscopic repair of GGF is feasible and safe but with limited durability. TorresVillalobos et al [20] used an endoscope passed through the GGF to introduce, via a percutaneous trans-gastric route, laparoscopic instruments in the remnant to repair GGF in 3 patients. No complications were reported but the GGF recurred in two patients. Thus for most authors surgery remains the standard of care and the main reported indications are persistent symptoms despite an optimized medical management, weight regain and failure of endoscopic repair. Conversely there is no accepted surgical technique. Some performed an excision of the fistula [1, 3], others an anastomotic revision with fistulous tract excision [8] and others a remnant gastrectomy with excision or exclusion of the fistula [2, 4, 6, 7]. Most authors also highlight that these procedures are challenging, because of the extensive inflammation around the fistulous tract, with as consequence an increased operative time, an increased length of hospital stay and a high early major complications rates ranging from 17 to 29% [21, 22]
. In our unit we performed either a fistula excision or a resection of the fistula and the
gastrojejunostomy with anastomosis recreation. Our goal in both procedures was a complete resection of all involved tissues to well vascularized tissues with care to leave enough remnant to avoid a retained antrum [23, 24]. No patient had fistula exclusion as we have concerns with this procedure that left in situ a narrow cuff of poorly vascularized remnant
lateral to the GGF. The choice of the procedure was done during surgery according to the location of the GGF and the anastomotic status. For type 1 GGF, which allow GGF resection without anastomotic damage we performed a simple excision by stapling both ends of the fistula as they were not associated with persistent MU or anastomotic stenosis. For type 2 GGF we performed a complete revision of the gastrojejunostomy with fistula excision. Except one done by laparotomy all these procedures were successfully completed laparoscopically. As previously reported by others, these operations were challenging especially when it required anastomosis revision. In these cases the mean operative time, which is a surrogate marker of technical complexity, was significantly longer than for simple fistula excision; 204 ± 58 min versus 50 ± 9 min. No mortality was observed and the only early complication was a leak, which required reoperation. Postoperative course was uneventful in the others patients and the mean length of stay for the entire group was 8.1 ± 2.4 days. Five months after GGF excision with anastomosis revision one patient developed a MU. Medical therapy was ineffective, despite high PPI dose and as this patient had, no recurrent GGF, no dilated gastric pouch and no anastomotic stenosis we hypothesized that he had a retained gastric antrum syndrome (RGA) due to an oversized remnant resection, during GGF treatment procedure. This syndrome is a well-known complication of an old surgical technique, with transection of the stomach, end to side gastrojejunostomy and retention of the antrum in continuity with the duodenum which was called “antral exclusion” [25] . Performed for giant duodenal ulcer and popularized by Devine in 1925, this procedure was abandoned because it was followed by an unacceptably high incidence of anastomotic ulcers (around 40%) due to a gastrin-mediated acid hypersecretion[26]. Indeed when antrum remains excluded from the gastric acid stream the antral producing G cells will release gastrin without negative feed-back with as consequence hypergastrinemia and acid hypersecretion in the remaining stomach. Our patient had normal fasting serum gastrin level but in the literature up to 20% of patients with RGA
have normal levels of fasting gastrin [24, 27]. As a normal gastrin level does not exclude this diagnosis we performed a second revision, with total remnant gastrectomy and creation of a new gastrojejunostomy. Postoperative course was uneventful with immediate symptoms resolution. The iterative upper endoscopies performed during follow-up were normal and more than 5 years after antrectomy the patient is still symptom free without PPI therapy. To avoid this complication, we should have performed a total remnant gastrectomy instead of a too large partial remnant resection leaving an excluded antrum during the first revision. No other complications were observed during follow-up. Early and late complications occurred both after revision with creation of a new anastomosis. No complications were observed after fistula excision. After revision 4 patients (50%) lost additional weight mean 20 kg (range 1141 kg). The others maintained their preoperative weight except one who regained weight because she had persistent food compulsions after craniopharyngioma resection [28, 29]. No patient was lost to follow-up. At the last follow-up point (mean 43 months, range 20 to 72 months) all the patients were symptom free with no one PPI use. Although this study was limited by its retrospective design and included a relatively small number of patients, we have longer follow-up than other reported studies. Conclusion: Gastro-gastric fistula is a rare complication after divided RYGB. Weight regain and epigastric pain with or without marginal ulcer are the most common presenting symptoms. Diagnosis can be challenging but combining upper GI endoscopy and contrast study is the best method to confirm the diagnosis. Surgical treatment should be tailored to both location of the GGF and status of the gastrojejunostomy, but anastomotic revision is associated with higher complication rates. GGF classification, based on its anatomic location, could serve as a working basis to compare different surgical approaches.
Disclosures None of the authors have anything to disclose. References: 1.
Capella J.F, Capella R.F. Gastro-gastric fistulas and marginal ulcers in gastric bypass procedures for weight reduction. Obes Surg 1999; 9(1):22-7; discussion 28.
2.
Carrodeguas L, Szomstein S, Soto F, et al. Management of gastrogastric fistulas after divided Roux-en-Y gastric bypass surgery for morbid obesity: analysis of 1,292 consecutive patients and review of literature. Surg Obes Relat Dis 2005; 1(5) : 467474.
3.
Gumbs A.A, Duffy A.J, Bell R.L. Management of gastrogastric fistula after laparoscopic Roux-en-Y gastric bypass. Surg Obes Relat Dis 2006; 2(2):117-121.
4.
Salimath J, Rosenthal R.J, Szomstein S. Laparoscopic remnant gastrectomy as a novel approach for treatment of gastrogastric fistula. Surg Endosc 2009; 23(11): 2591-2595.
5.
Cucchi S G, Pories W.J, MacDonald K.G, Morgan E.J. Gastrogastric fistulas. A complication of divided gastric bypass surgery. Ann Surg 1995; 221(4): 387-391.
6.
Cho M, Kaidar-Person O, Szomstein S, Rosenthal R. J. Laparoscopic remnant gastrectomy: a novel approach to gastrogastric fistula after Roux-en-Y gastric bypass for morbid obesity. J Am Coll Surg 2007; 204(4): 617-624.
7.
Tucker O.N, Szomstein S, Rosenthal R.J. Surgical management of gastro-gastric fistula after divided laparoscopic Roux-en-Y gastric bypass for morbid obesity. J Gastrointest Surg 2007; 11(12)/ 1673-1679.
8.
Corcelles R, Jamal M.H, Daigle C. R., Rogula T, Brethauer S.A, Schauer P.R. Surgical management of gastrogastric fistula. Surg Obes Relat Dis 2015; 11(6):122732.
9.
Carucci L.R, Turner M.A. Radiologic evaluation following Roux-en-Y gastric bypass surgery for morbid obesity. Eur J Radiol 2005; 53(3) : 353-365.
10.
Gustavsson S, Sundbom M. Excellent weight result after Roux-en-Y gastric bypass in spite of gastro-gastric fistula. Obes Surg 2003; 13(3): 457-459.
11.
Stanczyk M, Deveney C.W, Traxler S.A, McConnell D.B, Jobe B.A, O'Rourke R.W. Gastro-gastric fistula in the era of divided Roux-en-Y gastric bypass: strategies for prevention, diagnosis, and management. Obes Surg 2006; 16(3): 359-364.
12.
Gomez C.A. Gastroplasty in morbid obesity. Surg Clin North Am 1979; 59(6): 11131120.
13.
Gould J.C, Garren M.J, Starling, J.R. Lessons learned from the first 100 cases in a new minimally invasive bariatric surgery program. Obes Surg 2004; 14(5): 618-625.
14.
Ito Y, Blackstone M.O, Hatfield G.E. A gastrogastric fistula linking two penetrating ulcers. Report of a case. Gastrointest Endosc 1978; 24(5): 247-248.
15
Pauli E.M, Beshir H, Mathew A. Gastrogastric fistulae following gastric bypass surgery-clinical recognition and treatment. Curr Gastroenterol Rep 2014; 16(9): 405.
16.
Brethauer S.A, Nfonsam V, Sherman V, Udomsawaengsup S, Schauer P.R, Chand B. Endoscopy and upper gastrointestinal contrast studies are complementary in evaluation of weight regain after bariatric surgery. Surg Obes Relat Dis 2006; 2(6): 643-648; discussion 649-650.
17.
Chandler R.C, Srinivas G, Chintapalli K.N, Schwesinger W.H, Prasad S.R. Imaging in bariatric surgery: a guide to postsurgical anatomy and common complications. AJR Am J Roentgenol 2008; 190(1): 122-135.
18.
Sheen E, Triadafilopoulos G. Adverse effects of long-term proton pump inhibitor therapy. Dig Dis Sci 2011; 56(4): 931-950.
19.
Fernandez-Esparrach G, Lautz D.B, Thompson C.C. Endoscopic repair of gastrogastric fistula after Roux-en-Y gastric bypass: a less-invasive approach. Surg Obes Relat Dis 2010; 6(3): 282-288.
20.
Bhardwaj A, Cooney R.N, Wehrman A, Rogers A.M, Mathew A. Endoscopic repair of small symptomatic gastrogastric fistulas after gastric bypass surgery: a single center experience. Obes Surg 2010; 20(8): 1090-1095.
21.
Spaun G.O, Martinec D.V, Kennedy T.J, Swanstrom L.L. Endoscopic closure of gastrogastric fistulas by using a tissue apposition system (with videos). Gastrointest Endosc 2010; 71(3): 606-611.
22.
Torres-Villalobos G, Leslie D.B, Hunter D, Buchwald H, Martin-del-Campo L.A, Ikramuddin S. Preliminary report: search for a transgastric approach for managing gastrogastric fistulas. World J Surg 2011; 35(2): 372-376.
23.
Nesset E.M, Kendrick M.L, Houghton S.G et al. A two-decade spectrum of revisional bariatric surgery at a tertiary referral center. Surg Obes Relat Dis 2007; 3(1): 25-30; discussion 30.
24.
Gagner M, Gentileschi P, De Csepel J et al. Laparoscopic reoperative bariatric surgery: experience from 27 consecutive patients. Obes Surg 2002; 12(2): 254-260.
25.
Lee C.H, P'Eng F K, Lui W.Y. The clinical aspect of retained gastric antrum. Arch Surg 1986; 121(10): 1181-1186.
26.
Gibril F, Lindeman R.J, Abou-Saif A et al. Retained gastric antrum syndrome: a forgotten, treatable cause of refractory peptic ulcer disease. Dig Dis Sci 2001; 46(3): 610-617.
27.
Ribeiro-Parenti L, Arapis k, Chosidow D, Marmuse JP. Comparison of marginal ulcer rates between antecolic and retrocolic laparoscopic Roux-en-Y gastric bypass. Obes Surg, 2015. 25(2): p. 215-21.
28.
Csendes A, Burgos AM, Altuve J, Bonacic S. Incidence of marginal ulcer 1 month and 1 to 2 years after gastric bypass: a prospective consecutive endoscopic evaluation of 442 patients with morbid obesity. Obes Surg, 2009. 19(2): p. 135-8.
29.
Devine HB. Basic principles and supreme difficulties in gastric surgery. Surg., Gynec. & Obst., 1925. 40: p. 1-16.
30.
McKittrick LS, Moore FD, Warren R. Complications and Mortality in Subtotal Gastrectomy for Duodenal Ulcer: Report on a Two-Stage Procedure. Ann Surg, 1944. 120(4): p. 531-54.
31
Burhenne HJ. The retained antrum, preoperative roentgenologic diagnosis of an iatrogenic syndrome. AJR , 1967. 101(2): p. 459-466.
32
Shyu JF, Chen TH, Shyr YM, Su Ch, Wu C, Lui W. Gastric body partition to avoid ulcerogenic risk and hypergastrinemia. Surgery, 2006. 140(1): p. 44-9.
33.
Skorzewska A, Lal S, Waserman J, Guyda H. Abnormal food-seeking behavior after surgery for craniopharyngioma. Neuropsychobiology 1989; 21(1) : 17-20.
34.
Rottembourg D, O'Gorman C.S, Urbach S et al. Outcome after bariatric surgery in two adolescents with hypothalamic obesity following treatment of craniopharyngioma. J Pediatr Endocrinol Metab 2009; 22(9): 867-872.
Figure 1 legend: a. Resection of the fistula, in patients with type 1 GGF (i.e. located at the proximal part of the gastric pouch) b. En-bloc resection of both gastrogastric fistula and gastrojejunostomy, in patients with type 2 GGF (i.e. located at the distal part of the gastric pouch, near the gastrojejunostomy).
Table 1. Demographics data Patients with GGF N (%) Mean age (years) Gender (M/F)
9/1900 (0.5%) 39 ± 5 3/6
Mean preoperative weight (kg) Mean preoperative BMI (kg/m2)
154 ± 42 56 ± 12
Mean nadir weight (kg) Mean nadir BMI (kg/m2)
92 ± 23 33 ± 6
Mean weight at reoperation (kg) Mean BMI at reoperation (kg/m2)
107 ± 37 39 ± 10
Symptoms epigastric pain (n) vomiting (n) bleeding (n) weight regain (n) GGF, gastrogastric fistula; M, male; F, female; BMI, body mass index;
7 1 1 5
Table 2. Clinical presentation, preoperative workup and results of GGF management according to GGF location.
Clinical presentation
Time from primary to diagnosis
Type 1
Type 2
(n=3)
(n=6)
Mild epigastric pain
Intractable epigastric pain
Weight regain
Recurrent bleeding
30 (1-87)
10 (1-20)*
3 (100) Upper part of the pouch
6 (100) Near the GJ
0 (0) 0 (0) 1 (33)
6 (100)* 6 (100)* 0 (0)
40.6 (25-62.5)
11 (1-36)*
3 0 0 3
5
Months, (range)
Upper contrast study GGF identification, n, (%) GGF location Upper endoscopy GGF identification, n, (%) Associated MU, n, (%) Associated gastritis n, (%) Time from surgery to revision months (range) Indications for surgery, n Abdominal pain, n Recurrent bleeding, n Weight regain, n Revisional surgery Laparoscopy, n Operative time, minutes Early morbidity, n Late morbidity, n Additional weight loss, n Mean BMI last follow-up point (kg/m2)
Fistulous tract resection 3 50 ± 9 0/3 0/3 2/3 41.3 ± 10.1
2 1 2 En bloc resection of GGF and GJ New GJ 4 204 ± 58* 1/5 1/5 2/5 33.6 ± 9.5
GGF, gastrogastric fistula; MU, marginal ulcer; GJ, gastrojejunostomy; * p<0.05, significant. Type 1, GGF located in the proximal part of the gastric pouch; Type 2, GGF located in the distal part of the gastric pouch
a. b. Figure 1. Surgical revisional technique according to gastrogastric fistula location