New endoscopic technique for the treatment of large gastric fistula or gastric stenosis associated with gastric leaks after sleeve gastrectomy

Author’s Accepted Manuscript New endoscopic technique for the treatment of large gastric fistula or gastric stenosis associated with gastric leaks after sleeve gastrectomy Lionel Rebibo, Sami Hakim, Franck Brazier, Abdennaceur Dhahri, Cyril Cosse, Jean-Marc Regimbeau www.elsevier.com/locate/buildenv

PII: DOI: Reference:

S1550-7289(16)30058-2 http://dx.doi.org/10.1016/j.soard.2016.04.026 SOARD2664

To appear in: Surgery for Obesity and Related Diseases Received date: 29 February 2016 Revised date: 4 April 2016 Accepted date: 25 April 2016 Cite this article as: Lionel Rebibo, Sami Hakim, Franck Brazier, Abdennaceur Dhahri, Cyril Cosse and Jean-Marc Regimbeau, New endoscopic technique for the treatment of large gastric fistula or gastric stenosis associated with gastric leaks after sleeve gastrectomy, Surgery for Obesity and Related Diseases, http://dx.doi.org/10.1016/j.soard.2016.04.026 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.

New endoscopic technique for the treatment of large gastric fistula or gastric stenosis associated with gastric leaks after sleeve gastrectomy. Lionel Rebibo MD1, Sami Hakim MD2, Franck Brazier MD2, Abdennaceur Dhahri MD1, Cyril Cosse PhD1, 4, Jean-Marc Regimbeau MD, PhD1, 3, 4.

1. Department of Digestive Surgery, Amiens University Hospital, Avenue René Laennec, F-80054 Amiens Cedex 01, France 2. Department of Gastroentrology, Amiens University Hospital, Avenue René Laennec, F-80054 Amiens Cedex 01, France 3. EA4294, Jules Verne University of Picardie, F-80054 Amiens Cedex 01, France 4. Clinical Research Center, Amiens University Hospital, Avenue René Laennec, F80054 Amiens Cedex 01, France

Running Title: Alternative management for gastric leaks after sleeve gastrectomy Correspondence: Professor J.-M. Regimbeau Service de chirurgie digestive, Hôpital Sud, CHU d'Amiens, Avenue René Laennec, F-80054 Amiens Cedex 01, France Phone: +33 322 668 301 Fax: +33 322 668 680 E-mail: [email protected]

Conflict of interest: None of the authors have any conflicts of interest to declare. 1

ABSTRACT Background. Covered stent (CS) is required when gastric leak (GL) after sleeve gastrectomy (SG) is combined with gastric stenosis (GS) or when a large (> 2 cm in diameter) gastric fistula is present (increasing the likelihood of double pigtail stent (DPS) migration). Objective. To compare the results of our previous endoscopic management of large GL or GS associated with GL (using CS only) with those of our new endoscopic treatment (using combined CS and DPS). Setting. University Hospital, France, public practice. Material & methods. Between January 2009 and June 2015, all patients treated for large GL or GS associated with GL after SG (n = 20 patients) were included. Our previous endoscopic management required CS placement (CS group), while our new endoscopic treatment required combined CS and DPS placement (CS + DPS group). The primary efficacy endpoint was the treatment duration after CS placement until closure of the GL. The secondary efficacy endpoints were the number of endoscopic procedures, the stent migration rate and the failure rate. Results. Nine patients were treated by CS only (CS group), whereas 11 patients were treated by both CS and DPS (CS+DPS group). The median time to GL closure after CS placement was 84 days (33-130) in the CS group and 32 days (26-89) in the CS+DPS group (p≤0.05). The median number of endoscopic procedures at the time of CS placement was 2 (1-3) in the CS group and 1 (1-2) in the CS+DPS group (p≤0.05). The stent migration rate after CS placement was 33.3% in the CS group and 0% in the CS+DPS group (p=0.21) and the failure rate was 11% and 0% (p=0.36). Conclusion. The combination of CS and DPS constitutes an effective treatment for large GL or GS associated with GL, allowing significantly fewer endoscopic procedures and a shorter treatment duration.

2

Key words: sleeve gastrectomy; gastric leak; double pigtail stent, covered stent. Abbreviations: gastric leak (GL), sleeve gastrectomy (SG), double pigtail stent (DPS), covered stent (CS).

INTRODUCTION Gastric leak (GL) is the main early postoperative complication after sleeve gastrectomy (SG) and, according to a recent meta-analysis, occurs in 2% of cases surgery in the case of early-onset and/or poorly tolerated GL

(1)

. GL may require further

(2)

followed by additional

treatment for GL closure, such as endoscopic placement of a covered stent (CS)

(3 - 5)

or

double pigtail stent (DPS) (6). No consensus has been reached concerning the choice of stent at the present time. In our experience, use of a DPS is associated with shorter hospital stay, lower pain levels and shorter treatment duration than CS

(6)

. CS can be used in cases of concomitant gastric stenosis or

large GL (with a high risk of DPS migration and insufficient drainage). However, in many cases, GL persists after removal of the CS and therefore requires DPS placement under more favourable conditions or replacement of the CS, which increases the treatment duration and the risk of GL treatment failure. We have therefore developed a novel technique comprising combined CS and DPS placement during the same endoscopic procedure (7). The objective of the present study was to compare the results of our previous endoscopic management of large GL or gastric stenosis associated with GL (using CS only) with those of our new endoscopic treatment (using combined CS and DPS).

3

MATERIAL AND METHODS Population We performed a retrospective analysis of prospective data (ACOS database) on a group of 20 patients with large GL (greater than 2 cm in diameter, as measured during endoscopy) or gastric stenosis associated with GL after primary or secondary SG performed between January 2009 and June 2015. From January 2009 to November 2012, we used a CS to treat large GL or gastric stenosis associated with GL. These patients formed the CS group. Our treatment procedure changed progressively since December 2012, with the use of combined CS and DPS placement in some patients (forming the CS+DPS group).

Our institution's surgical procedures for SG Our surgical procedures for primary and secondary SGs (8-10) have been described elsewhere. A 34-gauge French bougie was used when transecting the greater gastric curvature. Gastric resection was initiated 6 cm above the pylorus (in the antrum). For patients in whom SG was performed between January 2004 and December 2009, stapling was performed using Endo GIATM Universal XL 60 (COVIDEN France SAS, Elancourt, France), with two 4.8 mm green reloads and then four or five 3.5 mm blue reloads. For patients in whom SG was performed between January 2010 and December 2013, purple Tri-StapleTM reloads (COVIDEN France SAS, Elancourt, France) were used. In our institution, the staple line is not reinforced for firstline SGs. For SGs performed after 2010, the abdominal drain was not left in place. For second-line SG, we used black Tri-StapleTM reloads (COVIDEN France SAS) with Peri-Strips Dry with Veritas (Baxter France, Maurepas, France) for the last two staples (in cases of previous gastric banding or cases of gastric banding removal and SG in the same procedure). For repeat SG, we used black Tri-StapleTM reloads (COVIDEN France SAS) with Peri-Strips

4

Dry with Veritas (Baxter France, Maurepas, France) for all stapling. A methylene blue test was always performed at the end of the surgical procedure. These data was not available for SG performed in other institutions.

Definitions of GL and large GL The presentation, time to onset and staple line site of gastric leakage were classified according to the modified UK Surgical Infection Study Group definitions

(11, 12)

. The patient's clinical

presentation was further described in terms of systemic signs of inflammation (tachycardia (>100 bpm) and hyperthermia (>38°C)), peritonitis (diffuse abdominal tenderness), pulmonary symptoms (cough and expectoration) and intra-abdominal abscess (local abdominal tenderness). The time to onset after SG was used to differentiate between acute gastric leakage (from postoperative day (POD) 1 to POD 7) and other types of leakage (≥POD 8) according to Rosenthal’s classification

(13)

. Oral contrast-enhanced abdominal computed

tomography (CT) was used to distinguish between leakage from the upper third of the staple line and leakage from the lower third. As mentioned above, large GL was defined as a fistula orifice greater than 2 cm in diameter allowing passage of the endoscope through the leak orifice (as measured during endoscopy).

Definition of gastric stenosis Gastric stenosis after SG was defined as a narrowing of the gastric remnant. Gastric stenosis can be diagnosed by endoscopy or oral contrast agent (Gastrografin®, Bracco Imaging France SAS, Courcouronnes, France) study of the upper gastrointestinal demonstrating narrowing of the stomach.

5

Management of GL Surgery Our revisional surgery procedure for GL has been described elsewhere

(2)

. Our institution's

revisional surgery procedure for early-onset, poorly tolerated GL is as follows: collection of bacteriological and fungal samples, washing of the abdominal cavity, suture of the GL orifice (if possible), optimal drainage of the GL (using two drains) and feeding jejunostomy placement. Oral contrast-enhanced abdominal CT scan was performed 6 days after surgery to ensure that GL was adequately drained before endoscopy. Endoscopic treatment was performed on POD 7 after revisional surgery. Our procedures for well-tolerated, late-onset GL and secondary endoscopic management after immediate revisional surgery have been described elsewhere

(6)

. In patients with well-tolerated, late-onset GL, endoscopy was

performed on the day of discovery of the GL by endoscopists with extensive experience in the management of postoperative complications.

Endoscopic procedures A CS (HANAROSTENT®, Life Partners Europe, Bagnolet, France) and, in some cases, a DPS (Zimmon® Biliary Stent, Cook Ireland Ltd, Limerick, Ireland) were placed during the endoscopic procedure. After stent placement, patients were allowed to drink water. Refeeding was performed with a feeding jejunostomy for cases with revisional surgery or a nasojejunal tube for other cases.

Combined CS and DPS placement Our endoscopic procedure for combined CS and DPS placement has been described elsewhere

(7)

. Endoscopy was performed under general anesthesia. The first step consisted of

locating the large GL or gastric stenosis. The edges of the GL were visualized by contrast

6

agent injection (Figure 1). A CS was then placed to cover the GL and, in some cases, expand the associated gastric stenosis. Under radiological guidance, the endoscope was placed in front of the previously identified orifice of the GL. The GL was then catheterized to allow the placement of two DPS. The two DPS were placed via a transprosthetic approach, in order to drain the GL internally (Figure 2) and were attached to the CS in the absence of gastric stenosis.

Concomitant therapy In parallel with surgical and/or endoscopic procedures, patients received antimicrobial and antifungal therapy, somatostatin-12 and refeeding via a nasojejunal tube or feeding jejunostomy, depending on the case. After both types of endoscopic procedures, patients were allowed to drink water. An oral contrast-enhanced abdominal CT scan was performed 4 weeks after the endoscopic procedure. The stent was removed endoscopically (when the GL had closed) or changed (when the GL had not closed) on the day after abdominal CT scan.

Definition of GL closure GL closure was defined as oral refeeding in the absence of (i) surgical drainage or endoscopic stents, (ii) flow via a previous surgical drainage pathway (gastrocutaneous fistula) and (iii) collection on an abdominal CT scan (with or without oral contrast agent) close to the GL site.

7

Inclusion criteria Patients included in the study had to meet the following criteria: post-SG GL (visualized on abdominal CT scan, endoscopy or surgery), GL on the upper part of the gastric staple line, a combination of gastric stenosis and GL or large GL, and CS placement. Patients who had undergone bariatric procedures other than SG, those without associated gastric stenosis or large GL and those not requiring CS placement for GL closure were excluded from the study.

Endpoints and parameters recorded The primary efficacy endpoint was the time to GL closure after CS placement (time interval between CS placement and GL closure). The secondary endpoints were preoperative demographic data, early- and late-onset GL rates, endoscopic data (number of endoscopic procedures, number of stents used, and stent migration rate), mortality rate, length of hospital stay after CS placement, and treatment failure rate. The two study groups were compared in terms of the following parameters: -

Demographic data: age, gender, BMI, preoperative comorbidities, type of surgical procedure performed.

-

GL data: median time interval between primary SG and discovery of the GL, site of the GL, early- and late-onset GL rates, and revisional surgery rate.

-

Endoscopic data: diameter of GL orifice, mean number of endoscopic procedures, endoscopic procedures prior to CS placement, mean number of stents placed, type of stent placed and stent migration rate.

-

Outcomes: mortality rate, endoscopic treatment failure rate and mean time to complete GL closure.

8

-

Analysis for risk factor of faster GL closure: early reoperation versus no reoperation, acute GL versus delayed GL, type of prosthesis used (CS only versus CS + DPS).

Statistical analysis The patients’ baseline characteristics are expressed as mean ± standard deviation (SD) and median (interquartile range) for continuous data and number (frequency) for categorical data. Univariate analysis used Student's T-test for quantitative variables. The Mann–Whitney U-test was used for non-parametric variables. The limit for statistical significance was set at p<0.05. The time to GL closure was represented by a Kaplan-Meier curve, on which the time variable corresponded to the time interval between CS placement and the follow-up visit. All statistical tests were performed with SPSS software (version 15.0 for Windows, SPSS Inc., Chicago, IL, USA).

RESULTS Demographic data During the study period, 127 patients were treated for postoperative GL after SG or SGrelated procedures (SG with a history of gastric banding removal, simultaneous gastric banding removal and SG in the same procedure, or repeat SG). Twenty of the 127 atients were treated for large GL (n=16) or gastric stenosis associated with GL. (n=4). The CS group consisted of 9 patients (7 women (77.7%) with a median age of 34 years (range: 23-58). The median preoperative BMI was 42.8 kg/m2 (38.1 - 59). Two patients (22.2%) had a BMI ≥ 50 kg/m2. Seven patients had undergone first-line SG, whereas the other two had undergone SG with a history of gastric banding removal and repeat SG, respectively.

9

Comorbidities were diabetes (22.2%, n=2), hypertension (0%, n=0), dyslipidemia (11.1%, n=1) and obstructive sleep apnea syndrome (11.1%, n=1). The CS+DPS group consisted of 11 patients (8 women (72.7%) with a median age of 40 years (range: 21-56). The median preoperative BMI was 43.2 kg/m2 (37.3 - 62.5). Two patients (18.2%) had a BMI ≥ 50 kg/m2. Seven patients had undergone primary SG, whereas the other three had undergone gastric banding removal and SG in the same procedure (n=3) and repeat SG (n=1), respectively. Comorbidities were diabetes (27.2%, n=3), hypertension (9%, n=1), dyslipidemia (18.2%, n=2) and obstructive sleep apnea syndrome (36.3%, n=4) (Table 1).

Gastric leak data The median time to discovery of the GL was 6 days (1-53). Eleven patients had acute GL (55%). Fourteen patients were reoperated. The differences between the two groups are summarized in Table 1.

Endoscopic data The median (range) total number of endoscopic procedures per patient was 3 (2-4) in the CS+DPS group and 4 (3-6) in the CS group (p=0.42). In the CS group, CS placement was performed because of large GL in 7 cases (77.7%). In the CS+DPS group, combined CS and DPS placement was performed because of large GL in 9 cases (81.8%). The median diameter of GL orifice in case of large GL was 2.5 cm (2-3) in the CS group and 2.4 cm (2-3) in the CS+DPS group (p=0.32). In the CS group, first-line CS placement was performed in three patients, second-line CS placement was performed in 5 patients and third-line CS placement was performed in one patient. In all cases of second- and third-line CS placement, the procedure was performed because of failure of DPS-only endoscopic treatment with DPS migration in 6 cases. First-

10

line combined CS and DPS placement had been performed in one patient in the CS + DPS group prior to CS placement. Second-line CS + DPS placement was performed in 8 patients and third-line CS + DPS placement was performed in two patients. In 9 out of 10 cases, CS + DPS placement was performed because of failure of DPS-only endoscopic treatment with DPS migration in 9 cases. In the last case, CS + DPS placement was performed because of CS migration. The median number of endoscopic procedures per patient at the time of CS placement was 2 (1-3) in the CS group and 1 (1-2) in the CS+DPS group (p≤ 0.05). An additional endoscopic procedure was necessary in 8 patients (88.8%) in the CS group: DPS placement in 7 patients and CS replacement in one patient. An additional endoscopic procedure was necessary in three patients (27.7%) in the CS+DPS group: DPS placement in two patients and repeat CS+DPS in one patient. The stent migration rate after CS placement was 33.3% (n=3) in the CS group (p=0.21) and 0% in the CS+DPS group. No morbidity related to the endoscopic procedure was observed in either of the two groups. The only factor associated with faster closure of GL was the the endoscopic management using combined CS + DPS (Table 2).

Outcomes There was no mortality in either group. The endoscopic treatment failure rate was 11% (n=1) in the CS group and 0% in CS+DPS group. A 60 cm Roux-en-Y gastrojejunal anastomosis on the fistula orifice was required in this case. The median total length of hospital stay after discovery of GL was 29 days (8-51) in the CS group and 27 days (5-86) in the CS+DPS group (p=0.73). The median total time to complete GL closure was 100 days (52-234) in the CS group and 72 days (48-131) in the CS+DPS

11

group (p≤ 0.05). The median time between CS placement and GL closure was 84 days (33130) in the CS group and 32 days (26-89) in the CS+DPS group (p≤ 0.05) (Table 1 & Figure 3).

DISCUSSION In our center, we have progressively tried to standardize the management of postoperative GL by systematic reoperation in the case of acute GL

(14)

, perioperative care with nutritional

support, appropriate use of antifungal medications in some cases of GL presenting risk factors for yeast infection and systematic use of somatostatin-14 following the discovery of the GL. Most literature series on CS have reported good results apart from a high incidence of secondary reoperation for persistent GL despite endoscopic treatment (in 41% of cases) (5). In our institution, we prefer to use DPS because it is associated with a shorter hospital stay, shorter treatment duration

(6)

, lower pain levels and absence of the gastroesophageal reflux

disease that may be observed with CS. We have therefore progressively abandoned CS-only endoscopic treatment for GL and now prefer the additional use of DPS. Most literature series have focused on gastric stenosis in the absence of GL. Treatment of gastric stenosis alone is based on endoscopic balloon dilatation in the event of endoscopic treatment failure, seromyotomy

(17)

(15)

or CS placement

(16)

and,

or Roux-en-Y gastric bypass

(18). In cases of gastric stenosis associated with GL, balloon dilatation is not possible or may be dangerous and endoscopic treatment requires the use of CS. In the four cases of gastric stenosis with concomitant GL in our series, two required a CS and two required a CS and a DPS. However, the presence of a large GL or gastric stenosis and concomitant GL may lead to DPS migration. In such cases, DPS placement is not possible or is ineffective and CS placement is required, as highlighted in our series with a high DPS migration rate prior to CS+DPS

12

placement. We totally agree with Nedelcu et al

(19)

who reported that leak size and gastric

stenosis associated with GL were very important factors in the management of postoperative GL and can be associated with a higher endoscopic treatment failure rate when treatment is not adapted to the type of GL. We also agree with the letter by Manos report of Donatelli et al

(21)

(20)

commenting the

that leak size is very important and can explain some failures of

endoscopic treatment. For this reason, we have developed a novel endoscopic technique combining DPS and CS designed to combine the benefits of each stent during a single endoscopic procedure. In our series, stent replacement was necessary in 88.8% of cases in the CS group, which prolonged the treatment duration (relative to the CS+DPS group) with a median duration of 84 days after CS placement. Furthermore, some cases of CS migration were observed in the CS group (33.3%), but not in the CS+DPS group, which may have been due to better fixation of the CS when a DPS was placed via a transprosthetic approach. These elements prompted us to adopt the CS+DPS approach. A CS can be used to treat large GL (Figure 2) or gastric stenosis with concomitant GL in some cases, whereas a DPS allows simultaneous internal drainage of the GL. Despite the small sample size, use of this novel approach avoided the need for additional endoscopic procedures (other than stent removal) for all but one case in the CS+DPS group. This difference was significant (p≤ 0.05) because 80% of the patients in the CS group required additional endoscopic procedures for stent placement or replacement. Furthermore, the median time to complete GL closure in the CS+DPS group was shorter than that in the CS group (72 days and 100 days, respectively; p≤ 0.05), mainly due to a shorter median time interval between CS placement and GL closure (32 days in the CS+DPS group, p=≤ 0.05) and the absence of endoscopic treatment failure that would have required subsequent revisional surgery (Figure 3).

13

An important finding in this series is the high failure rate of DPS-only placement, which can be explained by an inappropriate indication for this type of stent. Although large GL can be suspected on oral contrast-enhanced abdominal CT or during reoperation for GL, most cases of large GL are diagnosed during endoscopy for stent placement. Similarly, upper GI endoscopy plays an important role in the diagnosis and management of gastric stenosis associated with GL. Our endoscopic procedure for the management of GL has now been standardized by systematically looking for GS during endoscopy and measuring the length of the GL orifice. In the light of our results, we propose CS + DPS placement as first-line endoscopic treatment for large GL (greater than 2 cm) and/or stenosis associated with GL.

CONCLUSION Combined use of CS and DPS constitutes an effective treatment for large GL or gastric stenosis with concomitant GL. This technique is associated with a significantly lower number of endoscopic procedures and shorter treatment duration. In view of these positive results, our institution has adopted this new approach for patients with large GL or gastric stenosis with concomitant GL.

14

REFERENCES [1] Parikh M, Issa R, McCrillis A, Saunders JK, Ude-Welcome A, Gagner M. Surgical strategies that may decrease leak after laparoscopic sleeve gastrectomy: a systematic review and meta-analysis of 9991 cases. Ann Surg. 2013;257:231-7. [2] Rebibo L, Dhahri A, Verhaeghe P, Regimbeau JM. Early gastric fistula after laparoscopic sleeve gastrectomy: surgical management. J Visc Surg. 2012;149:e319-24. [3] Sakran N, Goitein D, Raziel A, Keidar A, Beglaibter N, Grinbaum R, Matter I, Alfici R, Mahajna A, Waksman I, Shimonov M, Assalia A. GLs after sleeve gastrectomy: a multicenter experience with 2,834 patients. Surg Endosc. 2013;27:240-5. [4] Simon F, Siciliano I, Gillet A, Castel B, Coffin B, Msika S. Gastric leak after laparoscopic sleeve gastrectomy: early covered self-expandable stent reduces healing time. Obes Surg. 2013;23:687-92. [5] Moszkowicz D, Arienzo R, Khettab I, Rahmi G, Zinzindohoué F, Berger A, Chevallier JM. Sleeve gastrectomy severe complications: is it always a reasonable surgical option? Obes Surg. 2013;23:676-86. [6] Pequignot A, Fuks D, Verhaeghe P, Dhahri A, Brehant O, Bartoli E, Delcenserie R, Yzet T, Regimbeau JM. Is there a place for pigtail drains in the management of gastric leaks after laparoscopic sleeve gastrectomy? Obes Surg. 2012;22:712-20. [7] Rebibo L, Fumery M, Hakim S, Brazier F, Dhahri A, Regimbeau JM. Combined stents for the treatment of large gastric fistulas or stenosis after sleeve gastrectomy. Endoscopy. 2015;47 Suppl 1:E59-60. [8] Dhahri A, Verhaeghe P, Hajji H, Fuks D, Badaoui R, Deguines JB, Regimbeau JM. Sleeve gastrectomy: technique and results. J Visc Surg. 2010;147(5 Suppl):e39-46.

15

[9] Rebibo L, Mensah E, Verhaeghe P, Dhahri A, Cosse C, Diouf M, Regimbeau JM. Simultaneous gastric band removal and sleeve gastrectomy: a comparison with front-line sleeve gastrectomy. Obes Surg. 2012;22:1420-6. [10] Rebibo L, Fuks D, Verhaeghe P, Deguines JB, Dhahri A, Regimbeau JM. Repeat sleeve gastrectomy compared with primary sleeve gastrectomy: a single-center, matched case study. Obes Surg. 2012;22:1909-15. [11] Csendes A, Burdiles P, Burgos AM, Maluenda F, Diaz JC. Conservative management of anastomotic leaks after 557 open gastric bypasses. Obes Surg. 2005;15:1252-6. [12] Bruce J, Krukowski ZH, Al-Khairy G, et al. Systematic review of the definition and measurement of anastomotic leak after gastrointestinal surgery. Br J Surg. 2001;88:1157-68. [13] Rosenthal RJ; International Sleeve Gastrectomy Expert Panel, Diaz AA, Arvidsson D, Baker RS, Basso N, Bellanger D, Boza C, El Mourad H, France M, Gagner M, Galvao-Neto M, Higa KD, Himpens J, Hutchinson CM, Jacobs M, Jorgensen JO, Jossart G, Lakdawala M, Nguyen NT, Nocca D, Prager G, Pomp A, Ramos AC, Rosenthal RJ, Shah S, Vix M, Wittgrove A, Zundel N. 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. [14] Rebibo L, Bartoli E, Dhahri A, Cosse C, Robert B, Brazier F, Pequignot A, Hakim S, Yzet T, Delcenserie R, Dupont H, Regimbeau JM. Persistent gastric fistula after sleeve gastrectomy: an analysis of the time between discovery and reoperation. Surg Obes Relat Dis. 2015 Apr 24. [Epub ahead of print] [15] Shnell M, Fishman S, Eldar S, Goitein D, Santo E. Balloon dilatation for symptomatic gastric sleeve stricture. Gastrointest Endosc. 2013 Nov 9. [Epub ahead of print].

16

[16] Parikh A, Alley JB, Peterson RM, Harnisch MC, Pfluke JM, Tapper DM, Fenton SJ. Management options for symptomatic stenosis after laparoscopic vertical sleeve gastrectomy in the morbidly obese. Surg Endosc. 2012;26:738-46. [17] Vilallonga R, Himpens J, van de Vrande S. Laparoscopic management of persistent strictures after laparoscopic sleeve gastrectomy. Obes Surg. 2013;23:1655-61. [18] Burgos AM, Csendes A, Braghetto I. Gastric stenosis after laparoscopic sleeve gastrectomy in morbidly obese patients. Obes Surg. 2013;23:1481-6. [19] Nedelcu M, Manos T, Cotirlet A, Noel P, Gagner M. Outcome of leaks after sleeve gastrectomy based on a new algorithm adressing leak size and gastric stenosis. Obes Surg. 2015;25:559-63. [20] Manos T, Nedelcu M, Noel P, Gagner M. Pigtails Internal Drainage for 2-cm Gastric Leak After Sleeve Gastrectomy Prolongs Healing. Obes Surg. 2015;25:1261-2. [21] Donatelli G, Catheline JM, Dumont JL, Vergeau BM, Tuszynski T, Cereatti F, Fiocca F, Meduri B. Outcome of Leaks After Sleeve Gastrectomy Based on a New Algorithm Addressing Leak Size and Gastric Stenosis. Obes Surg. 2015;25:1258-60.

17

FIGURE LEGENDS Figure 1: Endoscopic procedure in the combined covered stent and double pigtail stent placement. A: Identification of the edges of the gastric leak (white arrow) by injecting contrast agent; B: Covered stent placement (black arrow); C: Placement of two double pigtail stents (white arrow) via a transprosthetic approach. Figure 2: Diagram of combined covered stent and double pigtail stent placement. A: Frontal view of combined covered stent and double pigtail stent placement via a transprosthetic approach (from

(7)

); B: Post-endoscopic appearance of the combined covered

stent and double pigtail stent placement. Figure 3: A Kaplan-Meier curve representing the time interval between covered stent placement and complete gastric leak closure in each of the two groups. CS: covered stent CS + DPS: covered stent and double pigtail stent

18

Table 1: Data concerning the GL, endoscopic procedures and outcomes for the CS+DPS group and the CS group.

Gastric fistula data Female gender (%) BMI (kg/m²) Median time to GL discovery (days) Acute GL (%) Early reoperation (%) Gastric stenosis (%) Large GL (%) Endoscopic data Number of endoscopic procedures (n) Number of endoscopic procedures after CS placement (n) Stent migration (%) Outcomes Failure rate (%) Time to GL closure (days) Time between CS placement and GL closure (days)

CS group (n=9)

CS+DPS group (n=11)

p

7 (77.7) 42.8 (38.1-59) 13 (3-53) 4 (44.4) 5 (55.5) 2 (22.3) 7 (77.7)

8 (72.7) 43.2 (37.3-62.5) 6 (1-13) 9 (81.2) 9 (81.2) 2 (18.2) 9 (81.8)

0.77 0.71 0.1 0.1 0.19 0.83 0.77

4 (3-6)

3 (2-4)

0.42

2 (1-3) 3 (33.3)

1 (1-2) 0 (0)

≤ 0.05 0.21

1 (11) 100 (52-234)

0 (0) 72 (48-131)

0.36 ≤ 0.05

84 (33-130)

32 (26-89)

≤ 0.05

BMI: body mass index GL: gastric leak CS: covered stent DPS: double pigtail stent CS+DPS: combined CS and DPS

19

Table 2: Analysis of different risk factors for earlier closure of GL.

Yes

No

p

Early reoperation Number of patients (%) Time to GL closure (days) Time between CS placement and GL closure (days)

14 (70) 72 (48 - 211)

6 (30) 77 (52 - 234)

0.84

41 (26 - 130)

64 (28 - 126)

0.64

Acute GL Number of patients (%) Time to GL closure (days) Time between CS placement and GL closure (days)

13 (65) 72 (48 - 211)

7 (35) 96 (52 - 234)

0.73

41 (26 - 130)

73 (28 - 126)

0.52

Combined CS and DPS placement Number of patients (%) Time to GL closure (days) Time between CS placement and GL closure (days)

11 (55) 72 (48-131)

9 (45) 100 (52-234)

≤ 0.05

32 (26-89)

84 (33-130)

≤ 0.05

GL: gastric leak CS: covered stent DPS: double pigtail stent

20

21

22

23