Original article
Endoscopic treatment of fistula after sleeve gastrectomy: results of a multicenter retrospective study
Authors
Dimitri Christophorou1, Jean-Christophe Valats1, Natalie Funakoshi1, Claire Duflos2, Marie-Chistine Picot2, Bruno Vedrenne3, Frédéric Prat4, Phillipe Bulois5, Julien Branche6, Sébastien Decoster6, Emmanuel Coron7, Antoine Charachon8, Guillaume Pineton De Chambrun1, David Nocca1, Paul Bauret1, Pierre Blanc1
Institutions
Institutions are listed at the end of article.
submitted 29. September 2014 accepted after revision 16. March 2015
Background and study aims: Fistula is the main complication of laparoscopic sleeve gastrectomy (LSG), for which healing is difficult to achieve. The aims of the study were to evaluate the efficacy of interventional endoscopy for post-LSG fistula treatment, to evaluate various endoscopic techniques used and identify their complications, and to identify predictive factors of healing following endoscopic treatment. Patients and methods: This retrospective multicenter study included patients with post-LSG fistula. Therapeutic procedures were evaluated, taking into account complications and healing times. Endoscopic procedures were considered to have promoted healing if no other surgical procedure was performed. Predictive factors of healing were identified by univariate and multivariate analysis. Results: A total of 110 patients were included, of whom 6 (5.5 %) healed spontaneously, 81 (73.6 %) healed following endoscopic treatment, and 19 (17.3 %) healed following surgery. Healing rates
following endoscopic treatment were 84.4 % in the first 6 months of treatment (65/77), 52.4 % for treatment lasting 6 – 12 months (11/21), and 41.7 % after 12 months of treatment (5/12). A drainage procedure (surgical, endoscopic, or percutaneous) was performed in 92 patients (83.6 %). A total of 177 esogastric stents were placed in 88 patients (80.0 %). Surgical debridement, clip placement, glue sealing, and plug placement were also performed. Multivariate analysis identified four predictive factors of healing following endoscopic treatment: interval < 21 days between fistula diagnosis and first endoscopy (P = 0.003), small fistula (P = 0.01), interval between LSG and fistula ≤ 3 days (P = 0.01), no history of gastric banding (P = 0.04). Conclusion: Endoscopic treatment facilitated healing of post-LSG fistula in 74 % of patients. Early endoscopic treatment increased the likelihood of success, and was most effective during the first 6 months of management. After this point, surgical treatment should be considered.
Introduction
peritoneal collections) and medical treatment (antibiotics, enteral or parenteral nutrition). Interventional endoscopic techniques are becoming more frequently used for the treatment of these patients. Techniques initially developed for the treatment of fistulas in cancerous lesions are now proposed for post-LSG fistula, such as placement of covered esogastric stents [10 – 20] or endoscopic clips [11, 17, 21, 22], application of biological glue [9, 11, 23], endoscopic drainage (nasocavitary drains or double-pigtail catheters) [24], and placement of fistula plugs [25]. Because of the variety of techniques available, management of these patients is heterogeneous, and no consensus has been reached concerning optimal treatment modalities. Evaluation of the role of interventional endoscopy has been limited to small case series and reviews [10 – 26].
Bibliography DOI http://dx.doi.org/ 10.1055/s-0034-1392262 Published online: 25.6.2015 Endoscopy 2015; 47: 988–996 © Georg Thieme Verlag KG Stuttgart · New York ISSN 0013-726X Corresponding author Pierre Blanc, MD Service d’HépatoGastroentérologie B Hôpital Saint Eloi CHU Montpellier 80 rue Augustin Fliche 34295 Montpellier Cedex 5 France Fax: +33-467-337575 [email protected]
!
Scan this QR-Code to watch the video comment.
Obesity is a major public health problem in industrialized countries, with an ever-increasing prevalence [1]. Obesity is associated with significant morbidity [2 – 4] and is recognized as an independent risk factor for increased mortality [5]. Bariatric surgery plays an important role in the treatment of morbid or severe obesity associated with co-morbidities. Laparoscopic sleeve gastrectomy (LSG) is reported to have been the most widely used technique in France in 2011 [6]. The main complication of LSG is the development of fistula as a result of staple line leak, which leads to significant morbidity. Treatment of post-LSG fistula is prolonged and difficult [7 – 9]. There are no standardized guidelines for the treatment, which most often involves a combination of a surgical procedure (irrigation and drainage of intra-
Christophorou Dimitri et al. Endoscopic treatment of fistula after sleeve gastrectomy … Endoscopy 2015; 47: 988–996
Downloaded by: University of Florida. Copyrighted material.
988
Original article
Patients and methods !
Patients and centers In this descriptive, observational, multicenter study, patients were included if they fulfilled the following criteria: 1) classified as obese or overweight (body mass index [BMI] > 25 kg/m2); 2) had undergone LSG between 1 August 2007 and 31 May 2012; 3) presented with fistula due to a staple-line leak; 4) had undergone at least one diagnostic or therapeutic upper gastrointestinal endoscopy. The diagnosis of fistula was confirmed either by barium swallow, computed tomography [CT] scan of the abdomen and pelvis with upper digestive opacification, or by visualization of the fistula orifice during upper gastrointestinal endoscopy with leakage of contrast agent identified at fluoroscopy. Patients were included retrospectively from 1 August 2007 to 1 September 2011, and prospectively from 1 September 2011 to 31 May 2012. Data were collected until 31 May 2013, which was the end point date of the study. All patients were treated at seven digestive endoscopy centers in France (Hôpital Saint Eloi, Montpellier; Clinique du Diaconat, Mulhouse; Hôpital Cochin, Paris; Hôpital Claude Huriez, Lille; Clinique La Louvière, Lille; Hôpital Hôtel Dieu, Nantes; Hôpital Henri Mondor, Paris). All endoscopic procedures were performed by experienced gastroenterologists, who were all members of the National French Society of Digestive Endoscopy (Société Française d’Endoscopie Digestive, SFED). A letter containing information on the study was sent to all patients, who were free to refuse participation. If this was the case, the patient was excluded from the study. The study protocol was approved by the local Institutional Review Board (CHU Montpellier), and by the Research and Development committee of the SFED.
▶ Use of enteral or parenteral nutritional support, for a duration greater than 1 month
▶ Mortality.
Endoscopic techniques Stent placement included placement of one or more stents, plastic or metallic, totally or partially covered. Fully covered, self-expanding, metal stents [FCSEMS] used were Niti-S Enteral Colonic Covered Stent (Taewoong Medical, Seoul, Korea), Wallflex (Boston Scientific Corp., Natick, Massachusetts, USA), Hanarostent (M.I. Tech., Seoul, Korea). The partially covered, self-expanding, metal stent [PSEMS] used was the Ultraflex (Boston Scientific). The self-expanding plastic stent [SEPS] used was the Polyflex (Boston Scientific). Clip placement involved the Quick Clip 2 and EZ Clip (Olympus, Tokyo, Japan), Resolution (Boston Scientific), Triclip (Cook Medical, Winston-Salem, North Carolina, USA), and over-the-scope clips (OTSC; Ovesco Endoscopy AG, Tübingen, Germany). Two types of glue were used for injection: resorbable biological glue derived from fibrin (Beriplast– CSL Behring, Marburg, Germany; Tissucol – Baxter, Deerfield, Illinois, USA) and biocompatible nonresorbable N-butyl-2-cyanoacrylate glues (Glubran – General Enterprise Marketing S.r.l., Viareggio, Italy; and Ifabond – Fimed, Quincié-en-Beaujolais, France). Bioprosthetic plug placement involved Surgisis AFP (Cook Biotech Inc., West Lafayette, Indiana, USA). Irrigation of abscess cavities was usually performed with diluted povidone – iodine solution through a catheter placed endoscopically. Debridement was performed on any nonresorbable material remaining after previous surgical or endoscopic procedures, which could delay healing if left. In some cases, biopsy forceps were used to induce bleeding of the cavity or of the edges of the fistula in order to promote inflammation and healing. For drainage, placement of one or more endoscopic drains through the fistula involved a double-pigtail catheter (Cook Medical) or nasocavitary catheter (Liguory Nasal Biliary Drainage Set; Cook Medical).
Healing of the fistula Data collected The following data were collected for each patient. ▶ Sex, age, BMI, history of bariatric surgery ▶ Operative report of the LSG, including application of peroperative application of glue on the staple line ▶ Characteristics of the fistula – time to occurrence: acute postoperative (Day 0 – 3), early (Day 4 – 7), intermediate (Day 8 – 42), and late ( > Day 42) according to the International Sleeve Gastrectomy Expert Panel Consensus Statement [27]) ▶ Repeat surgical procedures or percutaneous or surgical drainage procedures ▶ Endoscopy reports, including: number of endoscopies performed; size of the fistula orifice (“small” [≤ 1 cm] if the endoscope was not able to pass through the orifice; otherwise “large” [> 1 cm]); endoscopic techniques used (see below), including stent, clip, and plug placement, glue injection, irrigation and debridement of abscess cavities, placement of one or more endoscopic drains, and pneumatic dilation of stenosis of the gastric sleeve. ▶ The evolution of the fistula: time to healing (duration between the diagnosis of fistula and healing) and duration of treatment (time from the first therapeutic procedure to healing). ▶ Complications of endoscopic procedures
Healing was defined as the absence of leakage of contrast agent as shown by various imaging techniques (barium swallow, CT scan), or the absence of identification of a fistula orifice following meticulous examination at upper gastrointestinal endoscopy, with use of a 0.35-mm guidewire and opacification with fluoroscopy during the procedure. Moreover, the diagnosis of healing was retained if no leakage, no fistula, and no intermittent recurrent left upper quadrant abscess occurred during 1 year of followup. Healing was achieved following interventional endoscopy, surgery, or after medical treatment alone (spontaneous healing). Healing was considered to have been achieved following surgical treatment if one of the following definitive surgical procedures was performed: complete gastrectomy with esophagojejunostomy, gastric by-pass (when a Roux-en-Y limb was brought up to the proximal aspect of the previous sleeve) or fistulojejunal anastomosis (when the jejunum was brought up to the fistulous opening). Medical treatment included antibiotics, enteral or parenteral nutrition, and surgical or percutaneous drainage. Healing following endoscopic treatment was defined as the closing of the fistula following one or more endoscopic therapeutic procedures, in association with medical treatment. Endoscopic treatment was considered to have failed when surgical treatment was
Christophorou Dimitri et al. Endoscopic treatment of fistula after sleeve gastrectomy … Endoscopy 2015; 47: 988–996
Downloaded by: University of Florida. Copyrighted material.
The aims of this retrospective multicenter study were therefore to evaluate the efficacy of interventional endoscopy in the treatment of post-LSG fistula, to describe the various endoscopic techniques used and their associated complications, and to identify predictive factors of healing following endoscopic treatment.
989
Original article
necessary for healing. Deceased patients were included in the final analysis.
110 patients met inclusion criteria 6 patients were excluded from survival analysis, as they healed following medical treatment alone
Statistical analysis Data were collected on all included patients. Quantitative variables are represented as means with SDs and medians with ranges. Qualitative variables are represented as numbers and percentages. The healing rate according to treatment duration was calculated for the following time periods: 0 – 1 month, 1 – 3 months, 3 – 6 months, 6 – 12 months, and over 12 months. The starting point for evaluation was the date of the first therapeutic procedure. The efficacy of endoscopic treatment in patients who had undergone a therapeutic endoscopic procedure was evaluated. The time to endoscopic healing was described using Kaplan – Meier survival curves. The censored data were patients lost to follow-up, deceased patients, or patients who healed following surgery or patients who did not heal within 1 year after the first therapeutic procedure. Patients who healed spontaneously were excluded from the survival analysis describing time to endoscopic healing, as they underwent only diagnostic endoscopy with no therapeutic procedure. The effect of potential predictive factors on endoscopic healing was studied with semi-parametric Cox models, after assessment of linearity and proportional hazard hypothesis. All variables with a P value of < 0.25 in a univariate Cox model were included in a multivariate Cox model. After selection, only variables with a P value < 0.10 were maintained in " Table 5). Values presented for the effect of variables the model (● are hazard ratios (HR), their 95 % confidence intervals (CI), and the P value of Wald’s test. For all analyses, tests were considered to be significant if P < 0.05. The SAS statistical package was used (SAS Institute, Cary, North Carolina, USA).
104 patients were included in survival analysis of efficacy of endoscopic treatment
Fig. 1
Table 1
3 patients were lost to follow-up
Flow chart of patients included in the survival analysis.
Characteristics of patients and fistulas.
Patient characteristics
n = 110
Age at inclusion, years Mean ± SD
39.7 ± 11.8
Median (range)
39 (18 – 67)
Sex, male/female, n (%)
18 (16.4) / 92 (83.6 %)
Weight, kg Mean ± SD
120.3 ± 26.1
Median (range)
115 ± 77 – 247
Initial BMI, kg/m² Mean ± SD
44.4 ± 8
Median (range)
43.4 (27.3 – 83.5)
Medical history, n/N (%) Smokers History of gastric banding With complications Per-operative glueing
39/87 (44.8) 28/109 (25.7) 8/28 (28.6) 11/84 (13.1)
Fistula characteristics Time to occurrence, days Mean ± SD
40.6 ± 116
Median (range)
10 (1 – 803)
Time to occurrence, n (%)
Results !
Patients and centers A total of 110 patients were identified across seven French digestive endoscopy centers, 104 of whom underwent endoscopic " Fig. 1). Three patients were lost to foltherapeutic procedures (● low-up. The mean age of patients at inclusion to the study was 39.7 years, and the majority of patients were female (83.6 %) " Table 1). The median initial weight was 115 kg (range 77 – (● 247 kg) with a mean BMI of 44.4 kg/m². The median time to fistula occurrence was 10 days following LSG (range 1 – 803 days). The time to fistula occurrence after LSG was within the first 7 days in 45.4 % (acute post-operative in 20.9 %, early post-operative in 24.5 %, intermediate in 38.2 %, and late in 16.4 %). Among the 110 patients, 44 (40.0 %) presented a large fistula orifice, and 96 (87.3 %) underwent surgical, endoscopic or percutaneous " Table 1). drainage (●
Description of treatment modalities Of the 110 patients, 83 (75.5 %) underwent repeat surgery, in" Tacluding peritoneal irrigation and drainage following LSG (● ble 2). A jejunostomy feeding tube was placed in 32 patients (29.1 %). In all, 92 patients (83.6 %) underwent one or several drainage procedures for intra-abdominal collections: surgical (n = 80, 72.7 %), endoscopic (n = 30, 27.3 %), and/or percutaneous (n = 17, 15.5 %). A total of 519 endoscopies were performed, with a mean of 4.6 endoscopies per patient (range 1 – 31). The first endoscopy was performed in a median of 6 days following initial
Acute postoperative (Day 0 – 3)
23 (20.9)
Early (Day 4 – 7)
27 (24.5)
Intermediate (Day 8 – 42)
42 (38.2)
Late ( > Day 43)
18 (16.4)
Large ( > 1 cm) or allowing passage of the endoscope), n (%)
44 (40.0)
Drainage, n (%)
96 (87.3)
BMI, body mass index.
" Table 2). The medidiagnosis of the fistula (range 0 – 200 days) (● an time from the first therapeutic procedure to healing (duration of treatment) was 111 days (range 1 – 1315 days).
Efficacy of endoscopic treatment and healing rates according to duration of treatment Healing of the fistula was achieved following endoscopic treatment in 81 out of the 110 patients (73.6 %) and following surgical " Table 3). Six patients (5.5 %) treatment in 19 patients (17.3 %) (● healed with medical treatment alone (spontaneously). One patient died prior to healing of the fistula. One patient had fistula recurrence 13 months after endoscopic treatment and finally underwent a gastric by-pass, and a second patient had fistula recurrence 2 months after a gastric by-pass on the gastrojejunal anastomosis. The median time to healing was 113 days (range 5 – 1315 days). The six patients who healed spontaneously were excluded from the survival analysis describing time to endoscopic healing, as they underwent only diagnostic endoscopy and no
Christophorou Dimitri et al. Endoscopic treatment of fistula after sleeve gastrectomy … Endoscopy 2015; 47: 988–996
Downloaded by: University of Florida. Copyrighted material.
990
Original article
Table 2
991
Description of different treatment modalities.
Patients (%)
Mean ± SD
Median (range)
1.3 ± 1.1
1 (0 – 6)
Total procedures
n = 110 Repeat surgery
83 (75.5)
Drainage technique
92 (83.6)
Surgical
80 (72.7)
1 ± 0.8
1 (0 – 3)
110
Endoscopic
30 (27.3)
0.5 ± 1.2
0 (0 – 11)
51
Percutaneous
17 (15.5)
0.3 ± 0.7
0 (0 – 3)
28
Endoscopies
110 (100)
4.7 ± 3.6
4 (1 – 31)
519
Interval between diagnosis and first endoscopy, days
–
17.8 ± 30.9
6 (0 – 200)
Description of healing rates according to duration of treatment.
Treatment
Time to healing, months1 0–1
1–3
3–6
6 – 12
> 12
Medical, n (%)
4 (50.0)
1 (2.6)
1 (3.2)
Endoscopic, n (%)
3 (37.5)
36 (94.7)
26 (83.9)
11 (52.4)
0 (0) 10 (47.6)
Total
0 (0)
6 (5.5)
5 (41.7)
81 (73.6)
6 (50.0)
19 (17.3)
Surgical, n (%) 2
1 (12.5)
0 (0)
2 (6.5)
Deaths
0
1
0
0
0
1
Lost to follow-up
0
0
2
0
1
3
Total, n (%)
8 (7.3)
38 (34.5)
31 (28.2)
21 (19.1)
12 (10.9)
Downloaded by: University of Florida. Copyrighted material.
Table 3
146 189
110
1
The starting point (T0) for the evaluation of treatment efficacy was the date of the first therapeutic procedure. 2 Healing was considered to have been achieved following surgical treatment if one of the following definitive surgical procedures was performed: complete gastrectomy with esophagojejunostomy, gastric by-pass, or fistulojejunal anastomosis. Surgical drainage alone without definitive surgery was not considered to be a surgical treatment for the analysis of healing rates.
Description of endoscopic techniques On average, 2.5 endoscopic techniques per patient were used for fistula treatment. The four main techniques used were esogastric stent placement (80.0 %), endoscopic clip placement (40.9 %), fis-
1.00
Survival Distribution Function
therapeutic procedure. The Kaplan–Meier curve for time to healing after endoscopic treatment in the 104 patients is presented in ●" Fig. 2. Between 0 and 1 month of treatment, healing was achieved in eight patients (7.3 % of the total patient population), three following endoscopic treatment (37.5 %), one following surgery " Table 3). (12.5 %), and four spontaneous healings (50.0 %) (● Treatment lasting between 1 and 3 months resulted in healing in 37 patients (33.6 % of the total patient population), 36 following endoscopic treatment (97.3 %), and 1 spontaneous healing. Between 3 and 6 months of treatment, healing was achieved in 29 patients (26.4 %), 26 following endoscopic treatment (89.7 %), 2 following surgery, and 1 spontaneously. Between 6 and 12 months of treatment, fistula healing was achieved in 21 patients (19.1 %), 11 by endoscopy (52.4 %) and 10 by surgery (47.6 %). After 12 months of treatment, fistula healing was achieved in 11 patients (10.0 %), 5 by endoscopy (45.5 %) and 6 by surgery (54.5 %). After 3 months of treatment, fistula healing rates following endoscopic treatment tended to decrease, whereas surgical healing rates increased; rates of spontaneous healing were " Table 3). minimal throughout (● The probability of achieving fistula healing with endoscopic treatment decreased over time, from 76.4 % at 1 month (78/102), to 65.6 % at 3 months (42/64), 48.5 % at 6 months (16/33), and 41.7 % at 12 months (5/12). Conversely, the probability of achieving healing following surgical treatment increased over time, from 17.6 % at 1 month (18/102), to 28.1 % at 3 months (18/64), 48.5 % at 6 months (16/33), and 50.0 % at 12 months (6/12).
0.75
0.50
0.25
0.0 0
2
4
6 Time (months)
8
10
12
Fig. 2 Kaplan–Meier curve for time to healing after endoscopic treatment.
tula irrigation and debridement (39.1 %), and application of endo" Table 4). scopic glue (29.1 %) (● A total of 177 stents were placed in 88 patients: 154 FCSEMS in 82 patients, 14 PCSEMS in 14 patients, and 9 SEPS in 8 patients. Clips were placed in 45 patients during 71 procedures. A total of 14 patients had one or more OTSCs placed during 21 procedures. Irrigation and fistula debridement was performed 125 times in 43 patients. A total of 72 glue applications were performed in 32 patients, with an average of 2.25 sessions per patient. Resorbable biological glue derived from fibrin was used in 22 patients (46 procedures), and biocompatible nonresorbable acrylic glues were used in 12 patients (26 procedures). Ten patients (9.1 %)
Christophorou Dimitri et al. Endoscopic treatment of fistula after sleeve gastrectomy … Endoscopy 2015; 47: 988–996
Original article
Table 4
Description of the different endoscopic techniques used in the study.
Technique
Patients, n (%)
Procedures Mean ± SD
Median (range)1
Total
88 (80)
2 ± 1.5
1 (1 – 10)
177
FCSEMS
82 (74.5)
1.8 ± 1.4
1 (0 – 9)
154
PCSEMS
14 (12.7)
0.2 ± 0.4
0 (0 – 1)
14
8 (7.3)
0.1 ± 0.3
0 (0 – 2)
9
45 (40.9)
1.6 ± 2.1
1 (1 – 15)
71
OTSC
14 (12.7)
0.5 ± 0.9
1 (0 – 5)
21
Irrigation
43 (39.1)
2.8 ± 2.9
2 (1 – 15)
125
Stents
SEPS Clip placement
Glue
32 (29.1)
2.25 ± 1.8
1.5 (1 – 9)
72
Fibrin
22 (20.0)
1.4 ± 1.4
1 (0 – 5)
46
Acrylate
12 (10.9)
0.8 ± 1.5
0 (0 – 7)
26
Plugs
10 (9.1)
1.2 ± 0.4
1 (1 – 2)
12
Endoscopic drainage
30 (27.3)
1.7 ± 1.9
1 (1 – 11)
51
DPTC
20 (18.2)
0.9 ± 0.2
1 (0 – 6)
28
NCC
15 (13.6)
0.8 ± 1.1
0.5 (0 – 5)
23
1.8 ± 1
1.5 (1 – 3)
11
2.5 ± 0.9
2 (1 – 7)
Pneumatic dilation of the gastric sleeve Total endoscopic techniques
6 (5.5) 104 (94.5)
FCSEMS, fully covered self-expanding metal stent; PCSEMS, partially covered self-expanding metal stent; SEPS, self-expanding plastic stent; DPTC, double-pigtail catheter; NCC, nasocavitary catheter. 1 Among patients who underwent an endoscopic technique.
were treated by bioprosthetic plug placement during 12 procedures. A total of 30 patients underwent endoscopic drainage, in a total of 51 procedures. A double-pigtail catheter was placed in 20 patients, and a nasocavitary catheter was placed in 15 patients. At least two of these four main techniques were used in 60.9 % of patients (67/110). The most common therapeutic strategy was the combination of fistula drainage (surgical, percutaneous, or endoscopic) and esogastric metal stent placement (75 patients, 68.2 %).
Stent incarceration, caused by the entrapment of mucosa within the mesh of the stent, making stent removal impossible, occurred in 14 of the 177 stents placed (7.9 %), the majority of which were PCSEMS (13/14, 92.8 %). The remaining case involved an FCSEMS following erosion of the cover. Incarceration was shown to be a frequent complication of PCSEMS, occurring in 13 out of 14 patients in which these stents were used (92.8 %) compared with only 1 out of 154 patients who had FCSEMS (0.6 %; P < 0.000001).
Weight loss, nutritional support, and malnutrition Mortality, morbidity, and complications of endoscopic procedures Three deaths occurred among the 110 patients (2.7 %). The first death was due to duodenal perforation secondary to migration of an FCSEMS. The second death occurred in a patient who had achieved fistula healing following surgical treatment, and was due to hemorrhagic necrosis of an esophageal coloplasty (featuring an esophagocolic anastomosis and a colojejunal anastomosis). The third patient died from cardiogenic and septic shock during repeat surgery 7 days after a gastric by-pass. Endoscopic treatment-related morbidity was exclusively linked to esogastric stent placement. The most frequent complication was stent migration, which occurred in 42.9 % of cases (76/177). Stent migration was most frequently observed with FCSEMS (46.1 %, 71/154), but also occurred with PCSEMS (15.4 %, 2/13) and SEPS (25.0 %, 2 /8). Ulcers caused by stent impaction into the gastric or esophageal wall occurred in 73 out of the 177 stents placed (41.2 %), and involved 66 FCSEMS (90.4 %), 4 PCSEMS (5.5 %), and 3 SEPS (4.1 %). There was no significant difference in ulcer formation between the different stent types used (66 /154 [42.8 %] for FCSEMS, 4/14 [28.6 %] for PCSEMS, and 3/9 [33.3 %] for SEPS; P = 0.54). These ulcers led to further complications in some cases: three led to digestive bleeding, which was treated endoscopically, and eight led to perforation of the gastric or esophageal wall. Six of these perforations were self-sealing, and two perforated into the peritoneal cavity and required emergency surgery. These additional complications were observed only with FCSEMS.
The mean percentage of excess weight lost was 64.2 % at 6 months and 89.5 % at 12 months. A total of 97 patients received enteral and/or parenteral nutritional support, 83 of whom (75.5 %) required the support for over a month (46 enteral, 47 parenteral). Complications linked to central lines used for parenteral nutrition were observed in 21 patients (15 infections, 9 thrombosis, and 4 pulmonary embolisms). Finally, four patients with malnutrition presented Gayet Wernicke encephalopathy.
Factors associated with reduced healing times following endoscopic treatment " Table 5. On The results of univariate Cox models are presented in● multivariate analysis, factors associated with a shortening of healing duration following endoscopic treatment were: no history of gastric banding (HR = 0.49; P = 0.04), small size (≤ 1 cm) of fistula (HR = 0.51; P = 0.01), interval between LSG and fistula ≤ 3 days (HR = 2.1; P = 0.01), and a short interval (≤ 21 days) between fistula " Table 5). diagnosis and first endoscopy (HR = 0.38; P = 0.003) (●
Discussion !
The occurrence of a fistula following LSG is a severe complication. In the current series, it was associated with major morbidity and a mortality of 2.7 %. Although post-LSG fistulas are rare (2 % – 4 % of LSGs) [6, 28, 29], a combination of the increasing number of bariatric surgical procedures, the long times to healing (median 113 days [range 5 – 1315] in the current series), and the multiple
Christophorou Dimitri et al. Endoscopic treatment of fistula after sleeve gastrectomy … Endoscopy 2015; 47: 988–996
Downloaded by: University of Florida. Copyrighted material.
992
Original article
Variable
P value
n
HR [95 %CI]
Male sex
104
1.372 [0.765 – 2.459]
0.2887
Age 2
104
1.018 [0.996 – 1.040]
0.1091
BMI > 40 kg/m²
104
0.629 [0.390 – 1.017]
0.0585
81
1.186 [0.698 – 2.016]
0.5287
103
0.380 [0.200 – 0.723]
0.0032
Per-operative glueing
80
1.136 [0.560 – 2.304]
0.7238
Large fistula ( > 1 cm)
104
0.400 [0.245 – 0.652]
0.0002
Interval between LSG and fistula (> 3 days)
104
1.386 [0.798 – 2.407]
0.2464
Interval between diagnosis and first endoscopy (> 21 days)
104
0.435 [0.234 – 0.809]
0.0086
Univariate analysis 1
Tobacco History of gastric banding
993
Table 5 Univariate and multivariate analysis of times to healing following endoscopic treatment.
Multivariate analysis Age 2
1.019 [0.997 – 1.042]
0.0944
History of gastric banding
0.493 [0.246 – 0.989]
0.0463
Large fistula ( > 1 cm)
0.510 [0.299 – 0.872]
0.0139
Interval between LSG and fistula (> 3 days)
2.100 [1.176 – 3.750]
0.0122
Interval between diagnosis and first endoscopy (> 21 days)
0.381 [0.201 – 0.721]
0.0030
endoscopic procedures involved (mean 4.7 endoscopies per patient) is likely to result in these patients being increasingly seen in interventional endoscopy units. The management of these patients requires a multidisciplinary approach. The first treatment is often surgical, and repeat surgery was necessary in 83 patients (75.5 %) in the current study. This allows the irrigation and drainage of intra-abdominal collections, but suture of the orifice is rarely possible because of local inflammation and difficulty in exposing the orifice if attempted more than 3 days after the date of the original sleeve gastrectomy. Similarly, the treatment of the fistula with surgical glue or application of a bovine pericardial patch have not been shown to be very effective [30]. Subsequent treatment involves medical management (intensive care, antibiotics, and nutritional support), complementary percutaneous drainage procedures, and endoscopic procedures. The current series is the largest multicenter study including exclusively patients who had undergone LSG. At the end of the study, 107 patients out of 110 were healed and 3 patients were lost to follow-up. Overall, 79.1 % of patients were healed by medical and endoscopic treatment, without the need for specific surgical management. Among these patients, 5.5 % presented spontaneous healing, for which an endoscopy was performed without a therapeutic procedure, whereas 73.6 % healed following endoscopic treatment. These figures are in agreement with previous studies, which report healing rates for endoscopic treatment of 50 % – 100 % [9 – 20]. All six patients who healed spontaneously had very small fistulas (1 – 3 mm diameter), with no significant collection over the staple line leak, and no requirement for drainage. After the fistula was considered healed, patients were followed up for at least 1 year (until 31 May 2013) to monitor for leakage recurrence or for recurrent left upper quadrant abscesses even without demonstrated fistula. In retrospect, left upper quadrant abscesses were considered to be persistent intermittent fistulas. Recurrence of fistula appeared in two patients. One patient had fistula recurrence 13 months after endoscopic treatment and finally underwent a gastric by-pass, and and a second patient had fistula recurrence 2 months after a gastric by-pass on the gastrojejunal anastomosis. Success rates of endoscopic treatment correlated with the duration of treatment. Patients had a diminishing chance of healing
following endoscopic treatment as the treatment period lengthened. After 3 months of treatment, 65.6 % of unhealed patients were cured endoscopically, 28.1 % were healed by surgery, and 1.6 % healed spontaneously. These rates declined to 48.5 %, 48.5 %, and 0 %, respectively, when treatment was 6 months in duration, and to 41.7 %, 50.0 %, and 0 %, respectively, when treatment lasted for 12 months. These figures demonstrate that patients had a greater chance of healing following surgery than following endoscopic treatment when duration of treatment exceeded 6 months. It seems reasonable, therefore, to propose endoscopic treatment during the first 6 months, with a preference for surgical treatment after 6 months, as with time, the chances of healing following endoscopic treatment alone tend to dwindle. Several endoscopic techniques are used to obtain closure of a fistula. In this study, the three most common techniques used were stent placement, clip placement, and endoscopic glue application. Certain authors, based on experience with upper gastrointestinal cancers [31 – 37], have developed the use of covered stents for the treatment of digestive fistula following bariatric surgery (gastric by-pass or LSG) [9 – 20]. In the current study, the placement of a bridging esogastric stent was the most frequently used endoscopic treatment technique (177 stents in 88 patients). Stent placement was most often associated with an additional endoscopic technique (endoscopic irrigation and debridement, clip placement, glue application). A recent meta-analysis of seven studies, which included 67 patients who had endoscopic placement of self-expanding stents for the treatment of bariatric surgery leaks, revealed that stent placement is a minimally invasive, safe, and effective alternative in the management of leaks after bariatric surgery, with successful leak closures in 87.8 % of cases [38]. However, no randomized controlled trials exist in which stent placement is compared with either surgical treatment alone (drainage and debridement), medical treatment alone (with or without percutaneous drainage), or another endoscopic technique. Several different types of endoscopic clips are currently in use, particularly for the treatment of digestive perforations, iatrogenic or otherwise [39 – 46]. The use of endoscopic clips for the closure of small post-LSG fistula orifices was reported several years ago [11, 17]. More recently, some authors have proposed the use of
Christophorou Dimitri et al. Endoscopic treatment of fistula after sleeve gastrectomy … Endoscopy 2015; 47: 988–996
Downloaded by: University of Florida. Copyrighted material.
HR, hazard ratio; CI, confidence interval; BMI, body mass index; LSG, laparoscopic sleeve gastrectomy. 1 Cox model. 2 HR is estimated for an increase of 1 year.
Original article
OTSCs in this indication [20 – 22]. In the current series, several different types of clips were used (Quick Clip 2, EZ Clip, Resolution, and OTSCs). Clips were placed in 45 patients during 71 procedures. A total of 14 patients were treated with OTSCs, 9 of whom also had FCSEMS, whereas 2 also had glue application. Placement of OTSCs was generally one of the final endoscopic therapeutic procedures performed on small-sized fistulas. Of these 14 patients, 11 healed following endoscopic treatment. We believe that the OTSC seem to be particularly well adapted to the treatment of post-LSG fistula, because of its large opening. The current results concord with data published by Mercky et al., which showed successful treatment of gastrointestinal postoperative leakage with OTSCs in 30 patients. The success rate was 71.4 %, and the healing rate after LSG fistula was significantly higher than the overall rate (P = 0.01) [22]. Two types of glue were used to treat 32 patients during 72 procedures: resorbable biological glue derived from fibrin (Beriplast / Tissucol; 22 patients, 46 procedures), and biocompatible nonresorbable acrylic glues (Glubran, Ifabond; 12 patients, 26 procedures). The use of glue has been reported for postsurgical esophageal fistula [47 – 51], and for the treatment of digestive fistula following bariatric surgery [9, 11, 23]. Other techniques used include endoscopic drain placement and endoscopic plug placement. In order to effectively drain intra-abdominal collections, the placement of a nasocavitary catheter or a double-pigtail catheter are treatment options. This type of drainage, associated with enteral nutrition through a jejunostomy tube, has been reported in five patients by El Hassan et al. [52] with a favorable outcome. In the current series, 30 patients were treated by endoscopic drainage during 51 procedures. This type of treatment seems to be effective and it would be interesting to evaluate the efficacy of this technique compared with others in a prospective study. The use of plugs has been inspired by the treatment of anoperineal fistula in Crohn’s disease [53, 54]. Toussaint et al. published a study in which enterocutaneous fistulas following bariatric surgery were treated with plugs in five patients. Two patients healed rapidly, whereas additional endoscopic procedures were necessary in three. The healing rate was 80 %. In the current series, 10 patients were treated with a plug (9.1 %), in association with a FCSEMS [25]. In the current study, 110 patients underwent 519 endoscopies (mean 4.7 procedures per patient) until fistula healing was achieved. On average, 2.5 endoscopic techniques were used per patient. The heterogeneity of treatment management and the frequent combination of several different endoscopic techniques in the same patient meant that comparison of efficacy between the different techniques was not possible. However, the most common treatment strategy was the placement of a covered metallic stent (80.0 % of patients) following effective drainage of the fistula (83.6 %). Criteria that seemed to be important in the choice of endoscopic technique were the size of the fistula and the presence of effective drainage. Clip placement and glue application were more often used on small fistulas, whereas esogastric stents were mostly placed if large fistulas were present at the beginning of treatment. In case of undrained collections, percutaneous, endoscopic, or surgical drainage was necessary. The evolution of endoscopic management was similar in the various centers over time: glue injection and stent placement were predominant at the beginning of the study, whereas double-pigtail catheters and OTSC placement become more frequently used over time.
The morbidity of interventional endoscopy in the current study was exclusively due to esogastric stent placement. The main complications of these stents are migration, impaction and ulceration, digestive perforation, and incarceration. Intragastric migration is a frequent complication of esogastric stents, and occurred in 42.9 % of stents placed in the current study compared with 11 % – 58 % reported in medical literature for this indication [10 – 14]. When stents migrate they may impact into the wall of the digestive tract, creating a contact ulcer. In the current series, the impaction rate was 41.2 %, and involved 90.4 % of FCSEMS. Eight digestive perforations were also reported, and all were due to impaction of a metallic stent. To our knowledge, this is the first series to describe the occurrence of iatrogenic digestive perforations following stent migration in the treatment of post-LSG fistula. Stent incarceration occurred in 14 stents, 13 of which were PCSEMS (92.8 %) and one was a FCSEMS (7.1 %) following erosion of its cover. Incarceration into the wall of the digestive tract was the main complication of PCSEMS, with an incarceration rate of 92.8 % for PCSEMS compared with 0.6 % for FCSEMS in the current series (P < 0.001). Stent extraction was obtained by careful traction in 64.3 % of cases (9 /14), with the help of an SEPS left in place for 7 – 15 days in 28.6 % of cases (4/14), and by surgical extraction in 7.1 % of cases (1/14). According to Swinnen et al., the use of a PCSEMS is an effective technique for leakage healing after stent incarceration. Resolution of leakage was obtained in 84.2 % of cases and extraction of incarcerated stents was obtained by simple traction in 9.2 % of cases, with the help of SEPS placement in 78.9 % of cases, and surgically in 2.6 % of cases [10]. However, noncovered and partially covered metal stents are only indicated for malignant pathologies, and their constant incarceration is sometimes difficult to manage. Therefore, we cannot recommend the systematic use of this kind of stent. However, if a PCSEMS is placed, it should be left for only a short time not exceeding 2 – 3 weeks, otherwise incarceration is inevitable. In case of incarceration of a PSEMS, the placement of an SEPS could be an acceptable technique facilitating the extraction of both the PCSEMS and the SEPS. In the current series of patients presenting with fistulas, the percentage of weight lost was statistically higher than expected compared with noncomplicated LSG: at 6 months it was 64.2 % vs. 49 % (P = 0.001), and at 12 months it was 89.5 % vs. 59.5 % at 12 months (P < 0.001) [29]. These patients are particularly at risk of severe malnutrition and complications linked to infections and vitamin deficiencies. Indeed, the occurrence of four cases of Gayet Wernicke encephalopathy linked to vitamin B1 deficiency demonstrates that appropriate nutritional support is sometimes lacking. In multivariate analysis, the reduction of time to healing following endoscopic treatment was associated with no history of gastric banding (P = 0.04), small fistula (≤ 1 cm) (P = 0.01), an interval between LSG and fistula ≤ 3 days (P = 0.01), and an interval between diagnosis of the fistula and first endoscopy less than 21 " Table 5). The study shows that early digestive days (P = 0.003) (● endoscopy in the first 21 days following diagnosis shortens the time to endoscopic healing. This is an important message for bariatric surgeons. In the current series, the first endoscopy was performed on average 17.8 days following diagnosis of the fistula (median of 6 days), but this interval was up to 200 days for one patient. In light of this study, we recommend that patients with post-LSG fistula undergo early endoscopy, which is associated with shorter healing time.
Christophorou Dimitri et al. Endoscopic treatment of fistula after sleeve gastrectomy … Endoscopy 2015; 47: 988–996
Downloaded by: University of Florida. Copyrighted material.
994
Original article
Diameter 1 cm
Drained
Undrained
Surgical drainage
Undrained
OTSC
Double-pigtail catheter
Double-pigtail catheter or stent*
Double-pigtail catheter
Observation +/– OTSC
OTSC
OTSC
Fig. 3 Endoscopic treatment algorithm based on the size of the fistula and the existence of an effective drainage. LSG, laparoscopic sleeve gastrectomy; OTSC, over-the-scope clip.
This study is the largest multicenter series evaluating the efficacy of interventional endoscopy in the treatment of post-LSG fistula. However, there are some limitations to the study. First, despite exhaustive data collection, the retrospective design of the study may have resulted in information bias, as patients were heterogeneously distributed between the different centers. Second, the failure of endoscopic treatment was defined by the need for surgical treatment, and it must be noted that the timing of this decision was arbitrary in the absence of formal recommendations. This decision depends on several factors, such as the experience of the surgeon or gastroenterologist, and current practices at different centers. The current series demonstrated that stent placement was associated with high morbidity. FSEMS migrated with the risk of impaction ulcers and perforation, whereas PCSEMS seemed to be effective but had a very high rate of incarceration. Therefore, we cannot recommend the systematic use of stents for endoscopic management. If a stent is placed, an early endoscopic re-evaluation is necessary within 2 weeks. In our current practice, the use of esogastric stents becomes exceptional, whereas double-pigtail catheters and OTSC placement are frequently used. We propose " Fig. 3). a therapeutic approach based on our current practice (● In conclusion, interventional endoscopy was effective in the treatment of post-LSG fistula and facilitated healing in 73.6 % of cases. Drainage of extra-digestive collections and stent placement are still the most common treatments, but stent placement is linked to very high morbidity. Performing early endoscopy (before Day 21) increased the chances of success. The efficacy of endoscopic treatment was optimal in the first 6 months of management. After 6 months, surgical treatment should be considered, as the probability of healing following endoscopic treatment decreases. Competing interests: None
Institutions 1 Service d’Hépato-Gastroentérologie B, Hôpital Saint Eloi, CHU Montpellier, Montpellier, France 2 Département d’Information Médicale, CHU Montpellier, Montpellier, France 3 Service de Gastroentérologie et Endoscopie Digestive, Clinique du Diaconat, Mulhouse, France 4 Service de Gastroentérologie et endoscopie digestive, Hôpital Cochin, APHP Paris, Paris, France 5 Service d’Hépato-Gastroentérologie, Clinique La Louvière, Lille, France 6 Service d’Hépato-Gastroentérologie, Hôpital Claude Huriez, CHU Lille, Lille, France 7 Service d’Hépato-Gastroentérologie, Hôpital Hôtel Dieu, CHU Nantes, Nantes, France 8 Service d’Hépato-Gastroentérologie, APHP, Hôpital Henri Mondor, Paris, France
References 1 International Obesity Taskforce. National Study Nutrition Health, ENNS. 2006: http://www.iaso.org/iotf/ 2 World Health Organization. Prevention and management of the global epidemic of obesity. Report of a WHO consultation on obesity, 3–5 June 1997 Geneva: WHO; 1997: Available from: http://whqlibdoc. who.int/hq/1998/WHO_NUT_NCD_98.1_(p1-158).pdf 3 Enquête épidémiologique nationale sur le surpoids et l’obésité. Obépi. 2012: Available from: http://www.roche.fr/content/dam/corporate/ roche_fr/doc/obepi_2012.pdf 4 Buchwald H, Cowan GSM, Pories WJ. Surgical management of obesity. Philadelphia: Saunders; 2007 5 Organisation mondiale de la santé. Le défi de l’obésité dans la région européenne de l’OMS et les stratégies de lutte. Résumé Genève: OMS; 2007: Available from: http://www.euro.who.int/__data/assets/ pdf_file/0010/98245/E90159.pdf 6 Burgos AM, Braghetto I, Csendes A et al. Gastric leak after laparoscopicsleeve gastrectomy for obesity. Obes Surg 2009; 19: 1672 – 1677 7 Deitel M, Crosby RD, Gagner M. The First International Consensus Summit for Sleeve Gastrectomy (SG), New York City, October 25–27, 2007. Obes Surg 2008; 18: 487 – 496 8 Csendes A, Burdiles P, Burgos AM et al. Conservative management of anastomotic leaks after 557 open gastric bypasses. Obes Surg 2005; 15: 1252 – 1256 9 Casella G, Soricelli E, Rizzello M et al. Nonsurgical treatment of staple line leaks after laparoscopic sleeve gastrectomy. Obes Surg 2009; 19: 821 – 826 10 Swinnen J, Eisendrath P, Rigaux J et al. Self-expandable metal stents for the treatment of benign upper GI leaks and perforations. Gastrointest Endosc 2011; 73: 890 – 899 11 Bège T, Emungania O, Vitton V et al. An endoscopic strategy for management of anastomotic complications from bariatric surgery: a prospective study. Gastrointest Endosc 2011; 73: 238 – 244 12 Eisendrath P, Cremer M, Himpens J et al. Endotherapy including temporary stenting of fistulas of the upper gastrointestinal tract after laparoscopic bariatric surgery. Endoscopy 2007; 39: 625 – 630 13 Eubanks S, Edwards CA, Fearing NM et al. Use of endoscopic stents to treat anastomotic complications after bariatric surgery. J Am Coll Surg 2008; 206: 935 – 938 14 Tan JT, Kariyawasam S, Wijeratne T et al. Diagnosis and management of gastric leaks after laparoscopic sleeve gastrectomy for morbid obesity. Obes Surg 2010; 20: 403 – 409 15 Serra C, Baltasar A, Andreo L et al. Treatment of gastric leaks with coated self-expanding stents after sleeve gastrectomy. Obes Surg 2007; 17: 866 – 872 16 Nguyen NT, Nguyen X-MT, Dholakia C. The use of endoscopic stent in management of leaks after sleeve gastrectomy. Obes Surg 2010; 20: 1289 – 1292 17 Oshiro T, Kasama K, Umezawa A et al. Successful management of refractory staple line leakage at the esophagogastric junction after a sleeve gastrectomy using the HANAROSTENT. Obes Surg 2010; 20: 530 – 534 18 Iqbal A, Miedema B, Ramaswamy A et al. Long-term outcome after endoscopic stent therapy for complications after bariatric surgery. Surg Endosc 2011; 25: 515 – 520 19 Simon F, Siciliano I, Gillet A et al. Gastric leak after laparoscopic sleeve gastrectomy: early covered self-expandable stent reduces healing time. Obes Surg 2013; 23: 687 – 692 20 Hadj Amor WB, Bonin EA, Vitton V et al. Successful endoscopic management of large upper gastrointestinal perforations following EMR using
Christophorou Dimitri et al. Endoscopic treatment of fistula after sleeve gastrectomy … Endoscopy 2015; 47: 988–996
Downloaded by: University of Florida. Copyrighted material.
Fistula after LSG
995
Original article
21
22
23
24
25
26
27
28
29
30
31
32 33
34
35
36
over-the-scope clipping combined with stenting. Endoscopy 2012; 44: E277 – E278 Surace M, Mercky P, Demarquay JF et al. Endoscopic management of GI fistulae with the over-the-scope clip system (with video). Gastrointest Endosc 2011; 74: 1416 – 1419 Mercky P, Gonzalez JM, Aimore Bonin E et al. Usefulness of over-thescope clipping system for closing digestive fistulas. Dig Endosc 2015; 27: 18 – 24 Papavramidis TS, Kotzampassi K, Kotidis E et al. Endoscopic fibrin sealing of gastrocutaneous fistulas after sleeve gastrectomy and biliopancreatic diversion with duodenal switch. J Gastroenterol Hepatol 2008; 23: 1802 – 1805 Pequignot A, Fuks D, Verhaeghe P et al. Is there a place for pigtail drains in the management of gastric leaks after laparoscopic sleeve gastrectomy? Obes Surg 2012; 22: 712 – 720 Toussaint E, Eisendrath P, Kwan V et al. Endoscopic treatment of postoperative enterocutaneous fistulas after bariatric surgery with the use of a fistula plug: report of five cases. Endoscopy 2009; 41: 560 – 563 Sakran N, Goitein D, Raziel A et al. Gastric leaks after sleeve gastrectomy: a multicenter experience with 2,834 patients. Surg Endosc 2013; 27: 240 – 245 Rosenthal RJ, Diaz AA, Arvidsson D et al. International Sleeve Gastrectomy Expert Panel Consensus Statement: best practice guidelines based on experience of >12,000 cases. Surg Obes Relat Dis 2012; 8: 8 – 19 Aurora AR, Khaitan L, Saber A. Sleeve gastrectomy and the risk of leak: a systematic analysis of 4,888 patients. Surg Endosc 2012; 26: 1509 – 1515 Nocca D, Krawczykowsky D, Bomans B et al. A prospective multicenter study of 163 sleeve gastrectomies: results at 1 and 2 years. Obes Surg 2008; 18: 560 – 565 Consten ECJ, Gagner M, Pomp A et al. Decreased bleeding after laparoscopic sleeve gastrectomy with or without duodenal switch for morbid obesity using a stapled buttressed absorbable polymer membrane. Obes Surg 2004; 14: 1360 – 1366 Shimi SM. Self-expanding metallic stents in the management of advanced esophageal cancer: a review. Semin Laparosc Surg 2000; 7: 9 – 21 Sabharwal T, Morales JP, Salter R et al. Esophageal cancer: self-expanding metallic stents. Abdom Imaging 2005; 30: 456 – 464 Dormann AJ, Eisendrath P, Wigginghaus B et al. Palliation of esophageal carcinoma with a new self-expanding plastic stent. Endoscopy 2003; 35: 207 – 211 Dormann A, Meisner S, Verin N et al. Self-expanding metal stents for gastroduodenal malignancies: systematic review of their clinical effectiveness. Endoscopy 2004; 36: 543 – 550 Thompson AM, Rapson T, Gilbert FJ et al. Endoscopic palliative treatment for esophageal and gastric cancer: techniques, complications, and survival in a population-based cohort of 948 patients. Surg Endosc 2004; 18: 1257 – 1262 Conio M, Repici A, Battaglia G et al. A randomized prospective comparison of self-expandable plastic stents and partially covered self-expandable metal stents in the palliation of malignant esophageal dysphagia. Am J Gastroenterol 2007; 102: 2667 – 2677
37 Hünerbein M, Stroszczynski C, Moesta KT et al. Treatment of thoracic anastomotic leaks after esophagectomy with self-expanding plastic stents. Ann Surg 2004; 240: 801 – 807 38 Puli SR, Spofford IS, Thompson CC. Use of self-expandable stents in the treatment of bariatric surgery leaks: a systematic review and meta-analysis. Gastrointest Endosc 2012; 75: 287 – 293 39 Mangiavillano B, Viaggi P, Masci E. Endoscopic closure of acute iatrogenic perforations during diagnostic and therapeutic endoscopy in the gastrointestinal tract using metallic clips: a literature review. J Dig Dis 2010; 11: 12 – 18 40 Qadeer MA, Dumot JA, Vargo JJ et al. Endoscopic clips for closing esophageal perforations: case report and pooled analysis. Gastrointest Endosc 2007; 66: 605 – 611 41 Binmoeller KF, Grimm H, Soehendra N. Endoscopic closure of a perforation using metallic clips after snare excision of a gastric leiomyoma. Gastrointest Endosc 1993; 39: 172 – 174 42 Minami S, Gotoda T, Ono H et al. Complete endoscopic closure of gastric perforation induced by endoscopic resection of early gastric cancer using endoclips can prevent surgery (with video). Gastrointest Endosc 2006; 63: 596 – 601 43 Sebastian S, Byrne AT, Torreggiani WC et al. Endoscopic closure of iatrogenic duodenal perforation during endoscopic ultrasound. Endoscopy 2004; 36: 245 44 Tang S-J, Tang L, Gupta S et al. Endoclip closure of jejunal perforation after balloon dilatation. Obes Surg 2007; 17: 540 – 543 45 Magdeburg R, Collet P, Post S et al. Endoclipping of iatrogenic colonic perforation to avoid surgery. Surg Endosc 2008; 22: 1500 – 1504 46 Kirschniak A, Kratt T, Stüker D et al. A new endoscopic over-the-scope clip system for treatment of lesions and bleeding in the GI tract: first clinical experiences. Gastrointest Endosc 2007; 66: 162 – 167 47 Cellier C, Landi B, Faye A et al. Upper gastrointestinal tract fistulae: endoscopic obliteration with fibrin sealant. Gastrointest Endosc 1996; 44: 731 – 733 48 Akhras J, Tobi M, Zagnoon A. Endoscopic fibrin sealant injection with application of hemostatic clips: a novel method of closing a refractory gastrocutaneous fistula. Dig Dis Sci 2005; 50: 1872 – 1874 49 Lippert E, Klebl FH, Schweller F et al. Fibrin glue in the endoscopic treatment of fistulae and anastomotic leakages of the gastrointestinal tract. Int J Colorectal Dis 2011; 26: 303 – 311 50 Pross M, Manger T, Reinheckel T et al. Endoscopic treatment of clinically symptomatic leaks of thoracic esophageal anastomoses. Gastrointest Endosc 2000; 51: 73 – 76 51 Truong S, Böhm G, Klinge U et al. Results after endoscopic treatment of postoperative upper gastrointestinal fistulas and leaks using combined Vicryl plug and fibrin glue. Surg Endosc 2004; 18: 1105 – 1108 52 El Hassan E, Mohamed A, Ibrahim M et al. Single-stage operative management of laparoscopic sleeve gastrectomy leaks without endoscopic stent placement. Obes Surg 2013; 23: 722 – 726 53 O’Connor L, Champagne BJ, Ferguson MA et al. Efficacy of anal fistula plug in closure of Crohn’s anorectal fistulas. Dis Colon Rectum 2006; 49: 1569 – 1573 54 O’Riordan JM, Datta I, Johnston C et al. A systematic review of the anal fistula plug for patients with Crohn’s and non-Crohn’s related fistulain-ano. Dis Colon Rectum 2012; 55: 351 – 358
Christophorou Dimitri et al. Endoscopic treatment of fistula after sleeve gastrectomy … Endoscopy 2015; 47: 988–996
Downloaded by: University of Florida. Copyrighted material.
996
Endoscopic treatment of fistula after sleeve gastrectomy: results of a multicenter retrospective study
Authors
Dimitri Christophorou1, Jean-Christophe Valats1, Natalie Funakoshi1, Claire Duflos2, Marie-Chistine Picot2, Bruno Vedrenne3, Frédéric Prat4, Phillipe Bulois5, Julien Branche6, Sébastien Decoster6, Emmanuel Coron7, Antoine Charachon8, Guillaume Pineton De Chambrun1, David Nocca1, Paul Bauret1, Pierre Blanc1
Institutions
Institutions are listed at the end of article.
submitted 29. September 2014 accepted after revision 16. March 2015
Background and study aims: Fistula is the main complication of laparoscopic sleeve gastrectomy (LSG), for which healing is difficult to achieve. The aims of the study were to evaluate the efficacy of interventional endoscopy for post-LSG fistula treatment, to evaluate various endoscopic techniques used and identify their complications, and to identify predictive factors of healing following endoscopic treatment. Patients and methods: This retrospective multicenter study included patients with post-LSG fistula. Therapeutic procedures were evaluated, taking into account complications and healing times. Endoscopic procedures were considered to have promoted healing if no other surgical procedure was performed. Predictive factors of healing were identified by univariate and multivariate analysis. Results: A total of 110 patients were included, of whom 6 (5.5 %) healed spontaneously, 81 (73.6 %) healed following endoscopic treatment, and 19 (17.3 %) healed following surgery. Healing rates
following endoscopic treatment were 84.4 % in the first 6 months of treatment (65/77), 52.4 % for treatment lasting 6 – 12 months (11/21), and 41.7 % after 12 months of treatment (5/12). A drainage procedure (surgical, endoscopic, or percutaneous) was performed in 92 patients (83.6 %). A total of 177 esogastric stents were placed in 88 patients (80.0 %). Surgical debridement, clip placement, glue sealing, and plug placement were also performed. Multivariate analysis identified four predictive factors of healing following endoscopic treatment: interval < 21 days between fistula diagnosis and first endoscopy (P = 0.003), small fistula (P = 0.01), interval between LSG and fistula ≤ 3 days (P = 0.01), no history of gastric banding (P = 0.04). Conclusion: Endoscopic treatment facilitated healing of post-LSG fistula in 74 % of patients. Early endoscopic treatment increased the likelihood of success, and was most effective during the first 6 months of management. After this point, surgical treatment should be considered.
Introduction
peritoneal collections) and medical treatment (antibiotics, enteral or parenteral nutrition). Interventional endoscopic techniques are becoming more frequently used for the treatment of these patients. Techniques initially developed for the treatment of fistulas in cancerous lesions are now proposed for post-LSG fistula, such as placement of covered esogastric stents [10 – 20] or endoscopic clips [11, 17, 21, 22], application of biological glue [9, 11, 23], endoscopic drainage (nasocavitary drains or double-pigtail catheters) [24], and placement of fistula plugs [25]. Because of the variety of techniques available, management of these patients is heterogeneous, and no consensus has been reached concerning optimal treatment modalities. Evaluation of the role of interventional endoscopy has been limited to small case series and reviews [10 – 26].
Bibliography DOI http://dx.doi.org/ 10.1055/s-0034-1392262 Published online: 25.6.2015 Endoscopy 2015; 47: 988–996 © Georg Thieme Verlag KG Stuttgart · New York ISSN 0013-726X Corresponding author Pierre Blanc, MD Service d’HépatoGastroentérologie B Hôpital Saint Eloi CHU Montpellier 80 rue Augustin Fliche 34295 Montpellier Cedex 5 France Fax: +33-467-337575 [email protected]
!
Scan this QR-Code to watch the video comment.
Obesity is a major public health problem in industrialized countries, with an ever-increasing prevalence [1]. Obesity is associated with significant morbidity [2 – 4] and is recognized as an independent risk factor for increased mortality [5]. Bariatric surgery plays an important role in the treatment of morbid or severe obesity associated with co-morbidities. Laparoscopic sleeve gastrectomy (LSG) is reported to have been the most widely used technique in France in 2011 [6]. The main complication of LSG is the development of fistula as a result of staple line leak, which leads to significant morbidity. Treatment of post-LSG fistula is prolonged and difficult [7 – 9]. There are no standardized guidelines for the treatment, which most often involves a combination of a surgical procedure (irrigation and drainage of intra-
Christophorou Dimitri et al. Endoscopic treatment of fistula after sleeve gastrectomy … Endoscopy 2015; 47: 988–996
Downloaded by: University of Florida. Copyrighted material.
988
Original article
Patients and methods !
Patients and centers In this descriptive, observational, multicenter study, patients were included if they fulfilled the following criteria: 1) classified as obese or overweight (body mass index [BMI] > 25 kg/m2); 2) had undergone LSG between 1 August 2007 and 31 May 2012; 3) presented with fistula due to a staple-line leak; 4) had undergone at least one diagnostic or therapeutic upper gastrointestinal endoscopy. The diagnosis of fistula was confirmed either by barium swallow, computed tomography [CT] scan of the abdomen and pelvis with upper digestive opacification, or by visualization of the fistula orifice during upper gastrointestinal endoscopy with leakage of contrast agent identified at fluoroscopy. Patients were included retrospectively from 1 August 2007 to 1 September 2011, and prospectively from 1 September 2011 to 31 May 2012. Data were collected until 31 May 2013, which was the end point date of the study. All patients were treated at seven digestive endoscopy centers in France (Hôpital Saint Eloi, Montpellier; Clinique du Diaconat, Mulhouse; Hôpital Cochin, Paris; Hôpital Claude Huriez, Lille; Clinique La Louvière, Lille; Hôpital Hôtel Dieu, Nantes; Hôpital Henri Mondor, Paris). All endoscopic procedures were performed by experienced gastroenterologists, who were all members of the National French Society of Digestive Endoscopy (Société Française d’Endoscopie Digestive, SFED). A letter containing information on the study was sent to all patients, who were free to refuse participation. If this was the case, the patient was excluded from the study. The study protocol was approved by the local Institutional Review Board (CHU Montpellier), and by the Research and Development committee of the SFED.
▶ Use of enteral or parenteral nutritional support, for a duration greater than 1 month
▶ Mortality.
Endoscopic techniques Stent placement included placement of one or more stents, plastic or metallic, totally or partially covered. Fully covered, self-expanding, metal stents [FCSEMS] used were Niti-S Enteral Colonic Covered Stent (Taewoong Medical, Seoul, Korea), Wallflex (Boston Scientific Corp., Natick, Massachusetts, USA), Hanarostent (M.I. Tech., Seoul, Korea). The partially covered, self-expanding, metal stent [PSEMS] used was the Ultraflex (Boston Scientific). The self-expanding plastic stent [SEPS] used was the Polyflex (Boston Scientific). Clip placement involved the Quick Clip 2 and EZ Clip (Olympus, Tokyo, Japan), Resolution (Boston Scientific), Triclip (Cook Medical, Winston-Salem, North Carolina, USA), and over-the-scope clips (OTSC; Ovesco Endoscopy AG, Tübingen, Germany). Two types of glue were used for injection: resorbable biological glue derived from fibrin (Beriplast– CSL Behring, Marburg, Germany; Tissucol – Baxter, Deerfield, Illinois, USA) and biocompatible nonresorbable N-butyl-2-cyanoacrylate glues (Glubran – General Enterprise Marketing S.r.l., Viareggio, Italy; and Ifabond – Fimed, Quincié-en-Beaujolais, France). Bioprosthetic plug placement involved Surgisis AFP (Cook Biotech Inc., West Lafayette, Indiana, USA). Irrigation of abscess cavities was usually performed with diluted povidone – iodine solution through a catheter placed endoscopically. Debridement was performed on any nonresorbable material remaining after previous surgical or endoscopic procedures, which could delay healing if left. In some cases, biopsy forceps were used to induce bleeding of the cavity or of the edges of the fistula in order to promote inflammation and healing. For drainage, placement of one or more endoscopic drains through the fistula involved a double-pigtail catheter (Cook Medical) or nasocavitary catheter (Liguory Nasal Biliary Drainage Set; Cook Medical).
Healing of the fistula Data collected The following data were collected for each patient. ▶ Sex, age, BMI, history of bariatric surgery ▶ Operative report of the LSG, including application of peroperative application of glue on the staple line ▶ Characteristics of the fistula – time to occurrence: acute postoperative (Day 0 – 3), early (Day 4 – 7), intermediate (Day 8 – 42), and late ( > Day 42) according to the International Sleeve Gastrectomy Expert Panel Consensus Statement [27]) ▶ Repeat surgical procedures or percutaneous or surgical drainage procedures ▶ Endoscopy reports, including: number of endoscopies performed; size of the fistula orifice (“small” [≤ 1 cm] if the endoscope was not able to pass through the orifice; otherwise “large” [> 1 cm]); endoscopic techniques used (see below), including stent, clip, and plug placement, glue injection, irrigation and debridement of abscess cavities, placement of one or more endoscopic drains, and pneumatic dilation of stenosis of the gastric sleeve. ▶ The evolution of the fistula: time to healing (duration between the diagnosis of fistula and healing) and duration of treatment (time from the first therapeutic procedure to healing). ▶ Complications of endoscopic procedures
Healing was defined as the absence of leakage of contrast agent as shown by various imaging techniques (barium swallow, CT scan), or the absence of identification of a fistula orifice following meticulous examination at upper gastrointestinal endoscopy, with use of a 0.35-mm guidewire and opacification with fluoroscopy during the procedure. Moreover, the diagnosis of healing was retained if no leakage, no fistula, and no intermittent recurrent left upper quadrant abscess occurred during 1 year of followup. Healing was achieved following interventional endoscopy, surgery, or after medical treatment alone (spontaneous healing). Healing was considered to have been achieved following surgical treatment if one of the following definitive surgical procedures was performed: complete gastrectomy with esophagojejunostomy, gastric by-pass (when a Roux-en-Y limb was brought up to the proximal aspect of the previous sleeve) or fistulojejunal anastomosis (when the jejunum was brought up to the fistulous opening). Medical treatment included antibiotics, enteral or parenteral nutrition, and surgical or percutaneous drainage. Healing following endoscopic treatment was defined as the closing of the fistula following one or more endoscopic therapeutic procedures, in association with medical treatment. Endoscopic treatment was considered to have failed when surgical treatment was
Christophorou Dimitri et al. Endoscopic treatment of fistula after sleeve gastrectomy … Endoscopy 2015; 47: 988–996
Downloaded by: University of Florida. Copyrighted material.
The aims of this retrospective multicenter study were therefore to evaluate the efficacy of interventional endoscopy in the treatment of post-LSG fistula, to describe the various endoscopic techniques used and their associated complications, and to identify predictive factors of healing following endoscopic treatment.
989
Original article
necessary for healing. Deceased patients were included in the final analysis.
110 patients met inclusion criteria 6 patients were excluded from survival analysis, as they healed following medical treatment alone
Statistical analysis Data were collected on all included patients. Quantitative variables are represented as means with SDs and medians with ranges. Qualitative variables are represented as numbers and percentages. The healing rate according to treatment duration was calculated for the following time periods: 0 – 1 month, 1 – 3 months, 3 – 6 months, 6 – 12 months, and over 12 months. The starting point for evaluation was the date of the first therapeutic procedure. The efficacy of endoscopic treatment in patients who had undergone a therapeutic endoscopic procedure was evaluated. The time to endoscopic healing was described using Kaplan – Meier survival curves. The censored data were patients lost to follow-up, deceased patients, or patients who healed following surgery or patients who did not heal within 1 year after the first therapeutic procedure. Patients who healed spontaneously were excluded from the survival analysis describing time to endoscopic healing, as they underwent only diagnostic endoscopy with no therapeutic procedure. The effect of potential predictive factors on endoscopic healing was studied with semi-parametric Cox models, after assessment of linearity and proportional hazard hypothesis. All variables with a P value of < 0.25 in a univariate Cox model were included in a multivariate Cox model. After selection, only variables with a P value < 0.10 were maintained in " Table 5). Values presented for the effect of variables the model (● are hazard ratios (HR), their 95 % confidence intervals (CI), and the P value of Wald’s test. For all analyses, tests were considered to be significant if P < 0.05. The SAS statistical package was used (SAS Institute, Cary, North Carolina, USA).
104 patients were included in survival analysis of efficacy of endoscopic treatment
Fig. 1
Table 1
3 patients were lost to follow-up
Flow chart of patients included in the survival analysis.
Characteristics of patients and fistulas.
Patient characteristics
n = 110
Age at inclusion, years Mean ± SD
39.7 ± 11.8
Median (range)
39 (18 – 67)
Sex, male/female, n (%)
18 (16.4) / 92 (83.6 %)
Weight, kg Mean ± SD
120.3 ± 26.1
Median (range)
115 ± 77 – 247
Initial BMI, kg/m² Mean ± SD
44.4 ± 8
Median (range)
43.4 (27.3 – 83.5)
Medical history, n/N (%) Smokers History of gastric banding With complications Per-operative glueing
39/87 (44.8) 28/109 (25.7) 8/28 (28.6) 11/84 (13.1)
Fistula characteristics Time to occurrence, days Mean ± SD
40.6 ± 116
Median (range)
10 (1 – 803)
Time to occurrence, n (%)
Results !
Patients and centers A total of 110 patients were identified across seven French digestive endoscopy centers, 104 of whom underwent endoscopic " Fig. 1). Three patients were lost to foltherapeutic procedures (● low-up. The mean age of patients at inclusion to the study was 39.7 years, and the majority of patients were female (83.6 %) " Table 1). The median initial weight was 115 kg (range 77 – (● 247 kg) with a mean BMI of 44.4 kg/m². The median time to fistula occurrence was 10 days following LSG (range 1 – 803 days). The time to fistula occurrence after LSG was within the first 7 days in 45.4 % (acute post-operative in 20.9 %, early post-operative in 24.5 %, intermediate in 38.2 %, and late in 16.4 %). Among the 110 patients, 44 (40.0 %) presented a large fistula orifice, and 96 (87.3 %) underwent surgical, endoscopic or percutaneous " Table 1). drainage (●
Description of treatment modalities Of the 110 patients, 83 (75.5 %) underwent repeat surgery, in" Tacluding peritoneal irrigation and drainage following LSG (● ble 2). A jejunostomy feeding tube was placed in 32 patients (29.1 %). In all, 92 patients (83.6 %) underwent one or several drainage procedures for intra-abdominal collections: surgical (n = 80, 72.7 %), endoscopic (n = 30, 27.3 %), and/or percutaneous (n = 17, 15.5 %). A total of 519 endoscopies were performed, with a mean of 4.6 endoscopies per patient (range 1 – 31). The first endoscopy was performed in a median of 6 days following initial
Acute postoperative (Day 0 – 3)
23 (20.9)
Early (Day 4 – 7)
27 (24.5)
Intermediate (Day 8 – 42)
42 (38.2)
Late ( > Day 43)
18 (16.4)
Large ( > 1 cm) or allowing passage of the endoscope), n (%)
44 (40.0)
Drainage, n (%)
96 (87.3)
BMI, body mass index.
" Table 2). The medidiagnosis of the fistula (range 0 – 200 days) (● an time from the first therapeutic procedure to healing (duration of treatment) was 111 days (range 1 – 1315 days).
Efficacy of endoscopic treatment and healing rates according to duration of treatment Healing of the fistula was achieved following endoscopic treatment in 81 out of the 110 patients (73.6 %) and following surgical " Table 3). Six patients (5.5 %) treatment in 19 patients (17.3 %) (● healed with medical treatment alone (spontaneously). One patient died prior to healing of the fistula. One patient had fistula recurrence 13 months after endoscopic treatment and finally underwent a gastric by-pass, and a second patient had fistula recurrence 2 months after a gastric by-pass on the gastrojejunal anastomosis. The median time to healing was 113 days (range 5 – 1315 days). The six patients who healed spontaneously were excluded from the survival analysis describing time to endoscopic healing, as they underwent only diagnostic endoscopy and no
Christophorou Dimitri et al. Endoscopic treatment of fistula after sleeve gastrectomy … Endoscopy 2015; 47: 988–996
Downloaded by: University of Florida. Copyrighted material.
990
Original article
Table 2
991
Description of different treatment modalities.
Patients (%)
Mean ± SD
Median (range)
1.3 ± 1.1
1 (0 – 6)
Total procedures
n = 110 Repeat surgery
83 (75.5)
Drainage technique
92 (83.6)
Surgical
80 (72.7)
1 ± 0.8
1 (0 – 3)
110
Endoscopic
30 (27.3)
0.5 ± 1.2
0 (0 – 11)
51
Percutaneous
17 (15.5)
0.3 ± 0.7
0 (0 – 3)
28
Endoscopies
110 (100)
4.7 ± 3.6
4 (1 – 31)
519
Interval between diagnosis and first endoscopy, days
–
17.8 ± 30.9
6 (0 – 200)
Description of healing rates according to duration of treatment.
Treatment
Time to healing, months1 0–1
1–3
3–6
6 – 12
> 12
Medical, n (%)
4 (50.0)
1 (2.6)
1 (3.2)
Endoscopic, n (%)
3 (37.5)
36 (94.7)
26 (83.9)
11 (52.4)
0 (0) 10 (47.6)
Total
0 (0)
6 (5.5)
5 (41.7)
81 (73.6)
6 (50.0)
19 (17.3)
Surgical, n (%) 2
1 (12.5)
0 (0)
2 (6.5)
Deaths
0
1
0
0
0
1
Lost to follow-up
0
0
2
0
1
3
Total, n (%)
8 (7.3)
38 (34.5)
31 (28.2)
21 (19.1)
12 (10.9)
Downloaded by: University of Florida. Copyrighted material.
Table 3
146 189
110
1
The starting point (T0) for the evaluation of treatment efficacy was the date of the first therapeutic procedure. 2 Healing was considered to have been achieved following surgical treatment if one of the following definitive surgical procedures was performed: complete gastrectomy with esophagojejunostomy, gastric by-pass, or fistulojejunal anastomosis. Surgical drainage alone without definitive surgery was not considered to be a surgical treatment for the analysis of healing rates.
Description of endoscopic techniques On average, 2.5 endoscopic techniques per patient were used for fistula treatment. The four main techniques used were esogastric stent placement (80.0 %), endoscopic clip placement (40.9 %), fis-
1.00
Survival Distribution Function
therapeutic procedure. The Kaplan–Meier curve for time to healing after endoscopic treatment in the 104 patients is presented in ●" Fig. 2. Between 0 and 1 month of treatment, healing was achieved in eight patients (7.3 % of the total patient population), three following endoscopic treatment (37.5 %), one following surgery " Table 3). (12.5 %), and four spontaneous healings (50.0 %) (● Treatment lasting between 1 and 3 months resulted in healing in 37 patients (33.6 % of the total patient population), 36 following endoscopic treatment (97.3 %), and 1 spontaneous healing. Between 3 and 6 months of treatment, healing was achieved in 29 patients (26.4 %), 26 following endoscopic treatment (89.7 %), 2 following surgery, and 1 spontaneously. Between 6 and 12 months of treatment, fistula healing was achieved in 21 patients (19.1 %), 11 by endoscopy (52.4 %) and 10 by surgery (47.6 %). After 12 months of treatment, fistula healing was achieved in 11 patients (10.0 %), 5 by endoscopy (45.5 %) and 6 by surgery (54.5 %). After 3 months of treatment, fistula healing rates following endoscopic treatment tended to decrease, whereas surgical healing rates increased; rates of spontaneous healing were " Table 3). minimal throughout (● The probability of achieving fistula healing with endoscopic treatment decreased over time, from 76.4 % at 1 month (78/102), to 65.6 % at 3 months (42/64), 48.5 % at 6 months (16/33), and 41.7 % at 12 months (5/12). Conversely, the probability of achieving healing following surgical treatment increased over time, from 17.6 % at 1 month (18/102), to 28.1 % at 3 months (18/64), 48.5 % at 6 months (16/33), and 50.0 % at 12 months (6/12).
0.75
0.50
0.25
0.0 0
2
4
6 Time (months)
8
10
12
Fig. 2 Kaplan–Meier curve for time to healing after endoscopic treatment.
tula irrigation and debridement (39.1 %), and application of endo" Table 4). scopic glue (29.1 %) (● A total of 177 stents were placed in 88 patients: 154 FCSEMS in 82 patients, 14 PCSEMS in 14 patients, and 9 SEPS in 8 patients. Clips were placed in 45 patients during 71 procedures. A total of 14 patients had one or more OTSCs placed during 21 procedures. Irrigation and fistula debridement was performed 125 times in 43 patients. A total of 72 glue applications were performed in 32 patients, with an average of 2.25 sessions per patient. Resorbable biological glue derived from fibrin was used in 22 patients (46 procedures), and biocompatible nonresorbable acrylic glues were used in 12 patients (26 procedures). Ten patients (9.1 %)
Christophorou Dimitri et al. Endoscopic treatment of fistula after sleeve gastrectomy … Endoscopy 2015; 47: 988–996
Original article
Table 4
Description of the different endoscopic techniques used in the study.
Technique
Patients, n (%)
Procedures Mean ± SD
Median (range)1
Total
88 (80)
2 ± 1.5
1 (1 – 10)
177
FCSEMS
82 (74.5)
1.8 ± 1.4
1 (0 – 9)
154
PCSEMS
14 (12.7)
0.2 ± 0.4
0 (0 – 1)
14
8 (7.3)
0.1 ± 0.3
0 (0 – 2)
9
45 (40.9)
1.6 ± 2.1
1 (1 – 15)
71
OTSC
14 (12.7)
0.5 ± 0.9
1 (0 – 5)
21
Irrigation
43 (39.1)
2.8 ± 2.9
2 (1 – 15)
125
Stents
SEPS Clip placement
Glue
32 (29.1)
2.25 ± 1.8
1.5 (1 – 9)
72
Fibrin
22 (20.0)
1.4 ± 1.4
1 (0 – 5)
46
Acrylate
12 (10.9)
0.8 ± 1.5
0 (0 – 7)
26
Plugs
10 (9.1)
1.2 ± 0.4
1 (1 – 2)
12
Endoscopic drainage
30 (27.3)
1.7 ± 1.9
1 (1 – 11)
51
DPTC
20 (18.2)
0.9 ± 0.2
1 (0 – 6)
28
NCC
15 (13.6)
0.8 ± 1.1
0.5 (0 – 5)
23
1.8 ± 1
1.5 (1 – 3)
11
2.5 ± 0.9
2 (1 – 7)
Pneumatic dilation of the gastric sleeve Total endoscopic techniques
6 (5.5) 104 (94.5)
FCSEMS, fully covered self-expanding metal stent; PCSEMS, partially covered self-expanding metal stent; SEPS, self-expanding plastic stent; DPTC, double-pigtail catheter; NCC, nasocavitary catheter. 1 Among patients who underwent an endoscopic technique.
were treated by bioprosthetic plug placement during 12 procedures. A total of 30 patients underwent endoscopic drainage, in a total of 51 procedures. A double-pigtail catheter was placed in 20 patients, and a nasocavitary catheter was placed in 15 patients. At least two of these four main techniques were used in 60.9 % of patients (67/110). The most common therapeutic strategy was the combination of fistula drainage (surgical, percutaneous, or endoscopic) and esogastric metal stent placement (75 patients, 68.2 %).
Stent incarceration, caused by the entrapment of mucosa within the mesh of the stent, making stent removal impossible, occurred in 14 of the 177 stents placed (7.9 %), the majority of which were PCSEMS (13/14, 92.8 %). The remaining case involved an FCSEMS following erosion of the cover. Incarceration was shown to be a frequent complication of PCSEMS, occurring in 13 out of 14 patients in which these stents were used (92.8 %) compared with only 1 out of 154 patients who had FCSEMS (0.6 %; P < 0.000001).
Weight loss, nutritional support, and malnutrition Mortality, morbidity, and complications of endoscopic procedures Three deaths occurred among the 110 patients (2.7 %). The first death was due to duodenal perforation secondary to migration of an FCSEMS. The second death occurred in a patient who had achieved fistula healing following surgical treatment, and was due to hemorrhagic necrosis of an esophageal coloplasty (featuring an esophagocolic anastomosis and a colojejunal anastomosis). The third patient died from cardiogenic and septic shock during repeat surgery 7 days after a gastric by-pass. Endoscopic treatment-related morbidity was exclusively linked to esogastric stent placement. The most frequent complication was stent migration, which occurred in 42.9 % of cases (76/177). Stent migration was most frequently observed with FCSEMS (46.1 %, 71/154), but also occurred with PCSEMS (15.4 %, 2/13) and SEPS (25.0 %, 2 /8). Ulcers caused by stent impaction into the gastric or esophageal wall occurred in 73 out of the 177 stents placed (41.2 %), and involved 66 FCSEMS (90.4 %), 4 PCSEMS (5.5 %), and 3 SEPS (4.1 %). There was no significant difference in ulcer formation between the different stent types used (66 /154 [42.8 %] for FCSEMS, 4/14 [28.6 %] for PCSEMS, and 3/9 [33.3 %] for SEPS; P = 0.54). These ulcers led to further complications in some cases: three led to digestive bleeding, which was treated endoscopically, and eight led to perforation of the gastric or esophageal wall. Six of these perforations were self-sealing, and two perforated into the peritoneal cavity and required emergency surgery. These additional complications were observed only with FCSEMS.
The mean percentage of excess weight lost was 64.2 % at 6 months and 89.5 % at 12 months. A total of 97 patients received enteral and/or parenteral nutritional support, 83 of whom (75.5 %) required the support for over a month (46 enteral, 47 parenteral). Complications linked to central lines used for parenteral nutrition were observed in 21 patients (15 infections, 9 thrombosis, and 4 pulmonary embolisms). Finally, four patients with malnutrition presented Gayet Wernicke encephalopathy.
Factors associated with reduced healing times following endoscopic treatment " Table 5. On The results of univariate Cox models are presented in● multivariate analysis, factors associated with a shortening of healing duration following endoscopic treatment were: no history of gastric banding (HR = 0.49; P = 0.04), small size (≤ 1 cm) of fistula (HR = 0.51; P = 0.01), interval between LSG and fistula ≤ 3 days (HR = 2.1; P = 0.01), and a short interval (≤ 21 days) between fistula " Table 5). diagnosis and first endoscopy (HR = 0.38; P = 0.003) (●
Discussion !
The occurrence of a fistula following LSG is a severe complication. In the current series, it was associated with major morbidity and a mortality of 2.7 %. Although post-LSG fistulas are rare (2 % – 4 % of LSGs) [6, 28, 29], a combination of the increasing number of bariatric surgical procedures, the long times to healing (median 113 days [range 5 – 1315] in the current series), and the multiple
Christophorou Dimitri et al. Endoscopic treatment of fistula after sleeve gastrectomy … Endoscopy 2015; 47: 988–996
Downloaded by: University of Florida. Copyrighted material.
992
Original article
Variable
P value
n
HR [95 %CI]
Male sex
104
1.372 [0.765 – 2.459]
0.2887
Age 2
104
1.018 [0.996 – 1.040]
0.1091
BMI > 40 kg/m²
104
0.629 [0.390 – 1.017]
0.0585
81
1.186 [0.698 – 2.016]
0.5287
103
0.380 [0.200 – 0.723]
0.0032
Per-operative glueing
80
1.136 [0.560 – 2.304]
0.7238
Large fistula ( > 1 cm)
104
0.400 [0.245 – 0.652]
0.0002
Interval between LSG and fistula (> 3 days)
104
1.386 [0.798 – 2.407]
0.2464
Interval between diagnosis and first endoscopy (> 21 days)
104
0.435 [0.234 – 0.809]
0.0086
Univariate analysis 1
Tobacco History of gastric banding
993
Table 5 Univariate and multivariate analysis of times to healing following endoscopic treatment.
Multivariate analysis Age 2
1.019 [0.997 – 1.042]
0.0944
History of gastric banding
0.493 [0.246 – 0.989]
0.0463
Large fistula ( > 1 cm)
0.510 [0.299 – 0.872]
0.0139
Interval between LSG and fistula (> 3 days)
2.100 [1.176 – 3.750]
0.0122
Interval between diagnosis and first endoscopy (> 21 days)
0.381 [0.201 – 0.721]
0.0030
endoscopic procedures involved (mean 4.7 endoscopies per patient) is likely to result in these patients being increasingly seen in interventional endoscopy units. The management of these patients requires a multidisciplinary approach. The first treatment is often surgical, and repeat surgery was necessary in 83 patients (75.5 %) in the current study. This allows the irrigation and drainage of intra-abdominal collections, but suture of the orifice is rarely possible because of local inflammation and difficulty in exposing the orifice if attempted more than 3 days after the date of the original sleeve gastrectomy. Similarly, the treatment of the fistula with surgical glue or application of a bovine pericardial patch have not been shown to be very effective [30]. Subsequent treatment involves medical management (intensive care, antibiotics, and nutritional support), complementary percutaneous drainage procedures, and endoscopic procedures. The current series is the largest multicenter study including exclusively patients who had undergone LSG. At the end of the study, 107 patients out of 110 were healed and 3 patients were lost to follow-up. Overall, 79.1 % of patients were healed by medical and endoscopic treatment, without the need for specific surgical management. Among these patients, 5.5 % presented spontaneous healing, for which an endoscopy was performed without a therapeutic procedure, whereas 73.6 % healed following endoscopic treatment. These figures are in agreement with previous studies, which report healing rates for endoscopic treatment of 50 % – 100 % [9 – 20]. All six patients who healed spontaneously had very small fistulas (1 – 3 mm diameter), with no significant collection over the staple line leak, and no requirement for drainage. After the fistula was considered healed, patients were followed up for at least 1 year (until 31 May 2013) to monitor for leakage recurrence or for recurrent left upper quadrant abscesses even without demonstrated fistula. In retrospect, left upper quadrant abscesses were considered to be persistent intermittent fistulas. Recurrence of fistula appeared in two patients. One patient had fistula recurrence 13 months after endoscopic treatment and finally underwent a gastric by-pass, and and a second patient had fistula recurrence 2 months after a gastric by-pass on the gastrojejunal anastomosis. Success rates of endoscopic treatment correlated with the duration of treatment. Patients had a diminishing chance of healing
following endoscopic treatment as the treatment period lengthened. After 3 months of treatment, 65.6 % of unhealed patients were cured endoscopically, 28.1 % were healed by surgery, and 1.6 % healed spontaneously. These rates declined to 48.5 %, 48.5 %, and 0 %, respectively, when treatment was 6 months in duration, and to 41.7 %, 50.0 %, and 0 %, respectively, when treatment lasted for 12 months. These figures demonstrate that patients had a greater chance of healing following surgery than following endoscopic treatment when duration of treatment exceeded 6 months. It seems reasonable, therefore, to propose endoscopic treatment during the first 6 months, with a preference for surgical treatment after 6 months, as with time, the chances of healing following endoscopic treatment alone tend to dwindle. Several endoscopic techniques are used to obtain closure of a fistula. In this study, the three most common techniques used were stent placement, clip placement, and endoscopic glue application. Certain authors, based on experience with upper gastrointestinal cancers [31 – 37], have developed the use of covered stents for the treatment of digestive fistula following bariatric surgery (gastric by-pass or LSG) [9 – 20]. In the current study, the placement of a bridging esogastric stent was the most frequently used endoscopic treatment technique (177 stents in 88 patients). Stent placement was most often associated with an additional endoscopic technique (endoscopic irrigation and debridement, clip placement, glue application). A recent meta-analysis of seven studies, which included 67 patients who had endoscopic placement of self-expanding stents for the treatment of bariatric surgery leaks, revealed that stent placement is a minimally invasive, safe, and effective alternative in the management of leaks after bariatric surgery, with successful leak closures in 87.8 % of cases [38]. However, no randomized controlled trials exist in which stent placement is compared with either surgical treatment alone (drainage and debridement), medical treatment alone (with or without percutaneous drainage), or another endoscopic technique. Several different types of endoscopic clips are currently in use, particularly for the treatment of digestive perforations, iatrogenic or otherwise [39 – 46]. The use of endoscopic clips for the closure of small post-LSG fistula orifices was reported several years ago [11, 17]. More recently, some authors have proposed the use of
Christophorou Dimitri et al. Endoscopic treatment of fistula after sleeve gastrectomy … Endoscopy 2015; 47: 988–996
Downloaded by: University of Florida. Copyrighted material.
HR, hazard ratio; CI, confidence interval; BMI, body mass index; LSG, laparoscopic sleeve gastrectomy. 1 Cox model. 2 HR is estimated for an increase of 1 year.
Original article
OTSCs in this indication [20 – 22]. In the current series, several different types of clips were used (Quick Clip 2, EZ Clip, Resolution, and OTSCs). Clips were placed in 45 patients during 71 procedures. A total of 14 patients were treated with OTSCs, 9 of whom also had FCSEMS, whereas 2 also had glue application. Placement of OTSCs was generally one of the final endoscopic therapeutic procedures performed on small-sized fistulas. Of these 14 patients, 11 healed following endoscopic treatment. We believe that the OTSC seem to be particularly well adapted to the treatment of post-LSG fistula, because of its large opening. The current results concord with data published by Mercky et al., which showed successful treatment of gastrointestinal postoperative leakage with OTSCs in 30 patients. The success rate was 71.4 %, and the healing rate after LSG fistula was significantly higher than the overall rate (P = 0.01) [22]. Two types of glue were used to treat 32 patients during 72 procedures: resorbable biological glue derived from fibrin (Beriplast / Tissucol; 22 patients, 46 procedures), and biocompatible nonresorbable acrylic glues (Glubran, Ifabond; 12 patients, 26 procedures). The use of glue has been reported for postsurgical esophageal fistula [47 – 51], and for the treatment of digestive fistula following bariatric surgery [9, 11, 23]. Other techniques used include endoscopic drain placement and endoscopic plug placement. In order to effectively drain intra-abdominal collections, the placement of a nasocavitary catheter or a double-pigtail catheter are treatment options. This type of drainage, associated with enteral nutrition through a jejunostomy tube, has been reported in five patients by El Hassan et al. [52] with a favorable outcome. In the current series, 30 patients were treated by endoscopic drainage during 51 procedures. This type of treatment seems to be effective and it would be interesting to evaluate the efficacy of this technique compared with others in a prospective study. The use of plugs has been inspired by the treatment of anoperineal fistula in Crohn’s disease [53, 54]. Toussaint et al. published a study in which enterocutaneous fistulas following bariatric surgery were treated with plugs in five patients. Two patients healed rapidly, whereas additional endoscopic procedures were necessary in three. The healing rate was 80 %. In the current series, 10 patients were treated with a plug (9.1 %), in association with a FCSEMS [25]. In the current study, 110 patients underwent 519 endoscopies (mean 4.7 procedures per patient) until fistula healing was achieved. On average, 2.5 endoscopic techniques were used per patient. The heterogeneity of treatment management and the frequent combination of several different endoscopic techniques in the same patient meant that comparison of efficacy between the different techniques was not possible. However, the most common treatment strategy was the placement of a covered metallic stent (80.0 % of patients) following effective drainage of the fistula (83.6 %). Criteria that seemed to be important in the choice of endoscopic technique were the size of the fistula and the presence of effective drainage. Clip placement and glue application were more often used on small fistulas, whereas esogastric stents were mostly placed if large fistulas were present at the beginning of treatment. In case of undrained collections, percutaneous, endoscopic, or surgical drainage was necessary. The evolution of endoscopic management was similar in the various centers over time: glue injection and stent placement were predominant at the beginning of the study, whereas double-pigtail catheters and OTSC placement become more frequently used over time.
The morbidity of interventional endoscopy in the current study was exclusively due to esogastric stent placement. The main complications of these stents are migration, impaction and ulceration, digestive perforation, and incarceration. Intragastric migration is a frequent complication of esogastric stents, and occurred in 42.9 % of stents placed in the current study compared with 11 % – 58 % reported in medical literature for this indication [10 – 14]. When stents migrate they may impact into the wall of the digestive tract, creating a contact ulcer. In the current series, the impaction rate was 41.2 %, and involved 90.4 % of FCSEMS. Eight digestive perforations were also reported, and all were due to impaction of a metallic stent. To our knowledge, this is the first series to describe the occurrence of iatrogenic digestive perforations following stent migration in the treatment of post-LSG fistula. Stent incarceration occurred in 14 stents, 13 of which were PCSEMS (92.8 %) and one was a FCSEMS (7.1 %) following erosion of its cover. Incarceration into the wall of the digestive tract was the main complication of PCSEMS, with an incarceration rate of 92.8 % for PCSEMS compared with 0.6 % for FCSEMS in the current series (P < 0.001). Stent extraction was obtained by careful traction in 64.3 % of cases (9 /14), with the help of an SEPS left in place for 7 – 15 days in 28.6 % of cases (4/14), and by surgical extraction in 7.1 % of cases (1/14). According to Swinnen et al., the use of a PCSEMS is an effective technique for leakage healing after stent incarceration. Resolution of leakage was obtained in 84.2 % of cases and extraction of incarcerated stents was obtained by simple traction in 9.2 % of cases, with the help of SEPS placement in 78.9 % of cases, and surgically in 2.6 % of cases [10]. However, noncovered and partially covered metal stents are only indicated for malignant pathologies, and their constant incarceration is sometimes difficult to manage. Therefore, we cannot recommend the systematic use of this kind of stent. However, if a PCSEMS is placed, it should be left for only a short time not exceeding 2 – 3 weeks, otherwise incarceration is inevitable. In case of incarceration of a PSEMS, the placement of an SEPS could be an acceptable technique facilitating the extraction of both the PCSEMS and the SEPS. In the current series of patients presenting with fistulas, the percentage of weight lost was statistically higher than expected compared with noncomplicated LSG: at 6 months it was 64.2 % vs. 49 % (P = 0.001), and at 12 months it was 89.5 % vs. 59.5 % at 12 months (P < 0.001) [29]. These patients are particularly at risk of severe malnutrition and complications linked to infections and vitamin deficiencies. Indeed, the occurrence of four cases of Gayet Wernicke encephalopathy linked to vitamin B1 deficiency demonstrates that appropriate nutritional support is sometimes lacking. In multivariate analysis, the reduction of time to healing following endoscopic treatment was associated with no history of gastric banding (P = 0.04), small fistula (≤ 1 cm) (P = 0.01), an interval between LSG and fistula ≤ 3 days (P = 0.01), and an interval between diagnosis of the fistula and first endoscopy less than 21 " Table 5). The study shows that early digestive days (P = 0.003) (● endoscopy in the first 21 days following diagnosis shortens the time to endoscopic healing. This is an important message for bariatric surgeons. In the current series, the first endoscopy was performed on average 17.8 days following diagnosis of the fistula (median of 6 days), but this interval was up to 200 days for one patient. In light of this study, we recommend that patients with post-LSG fistula undergo early endoscopy, which is associated with shorter healing time.
Christophorou Dimitri et al. Endoscopic treatment of fistula after sleeve gastrectomy … Endoscopy 2015; 47: 988–996
Downloaded by: University of Florida. Copyrighted material.
994
Original article
Diameter 1 cm
Drained
Undrained
Surgical drainage
Undrained
OTSC
Double-pigtail catheter
Double-pigtail catheter or stent*
Double-pigtail catheter
Observation +/– OTSC
OTSC
OTSC
Fig. 3 Endoscopic treatment algorithm based on the size of the fistula and the existence of an effective drainage. LSG, laparoscopic sleeve gastrectomy; OTSC, over-the-scope clip.
This study is the largest multicenter series evaluating the efficacy of interventional endoscopy in the treatment of post-LSG fistula. However, there are some limitations to the study. First, despite exhaustive data collection, the retrospective design of the study may have resulted in information bias, as patients were heterogeneously distributed between the different centers. Second, the failure of endoscopic treatment was defined by the need for surgical treatment, and it must be noted that the timing of this decision was arbitrary in the absence of formal recommendations. This decision depends on several factors, such as the experience of the surgeon or gastroenterologist, and current practices at different centers. The current series demonstrated that stent placement was associated with high morbidity. FSEMS migrated with the risk of impaction ulcers and perforation, whereas PCSEMS seemed to be effective but had a very high rate of incarceration. Therefore, we cannot recommend the systematic use of stents for endoscopic management. If a stent is placed, an early endoscopic re-evaluation is necessary within 2 weeks. In our current practice, the use of esogastric stents becomes exceptional, whereas double-pigtail catheters and OTSC placement are frequently used. We propose " Fig. 3). a therapeutic approach based on our current practice (● In conclusion, interventional endoscopy was effective in the treatment of post-LSG fistula and facilitated healing in 73.6 % of cases. Drainage of extra-digestive collections and stent placement are still the most common treatments, but stent placement is linked to very high morbidity. Performing early endoscopy (before Day 21) increased the chances of success. The efficacy of endoscopic treatment was optimal in the first 6 months of management. After 6 months, surgical treatment should be considered, as the probability of healing following endoscopic treatment decreases. Competing interests: None
Institutions 1 Service d’Hépato-Gastroentérologie B, Hôpital Saint Eloi, CHU Montpellier, Montpellier, France 2 Département d’Information Médicale, CHU Montpellier, Montpellier, France 3 Service de Gastroentérologie et Endoscopie Digestive, Clinique du Diaconat, Mulhouse, France 4 Service de Gastroentérologie et endoscopie digestive, Hôpital Cochin, APHP Paris, Paris, France 5 Service d’Hépato-Gastroentérologie, Clinique La Louvière, Lille, France 6 Service d’Hépato-Gastroentérologie, Hôpital Claude Huriez, CHU Lille, Lille, France 7 Service d’Hépato-Gastroentérologie, Hôpital Hôtel Dieu, CHU Nantes, Nantes, France 8 Service d’Hépato-Gastroentérologie, APHP, Hôpital Henri Mondor, Paris, France
References 1 International Obesity Taskforce. National Study Nutrition Health, ENNS. 2006: http://www.iaso.org/iotf/ 2 World Health Organization. Prevention and management of the global epidemic of obesity. Report of a WHO consultation on obesity, 3–5 June 1997 Geneva: WHO; 1997: Available from: http://whqlibdoc. who.int/hq/1998/WHO_NUT_NCD_98.1_(p1-158).pdf 3 Enquête épidémiologique nationale sur le surpoids et l’obésité. Obépi. 2012: Available from: http://www.roche.fr/content/dam/corporate/ roche_fr/doc/obepi_2012.pdf 4 Buchwald H, Cowan GSM, Pories WJ. Surgical management of obesity. Philadelphia: Saunders; 2007 5 Organisation mondiale de la santé. Le défi de l’obésité dans la région européenne de l’OMS et les stratégies de lutte. Résumé Genève: OMS; 2007: Available from: http://www.euro.who.int/__data/assets/ pdf_file/0010/98245/E90159.pdf 6 Burgos AM, Braghetto I, Csendes A et al. Gastric leak after laparoscopicsleeve gastrectomy for obesity. Obes Surg 2009; 19: 1672 – 1677 7 Deitel M, Crosby RD, Gagner M. The First International Consensus Summit for Sleeve Gastrectomy (SG), New York City, October 25–27, 2007. Obes Surg 2008; 18: 487 – 496 8 Csendes A, Burdiles P, Burgos AM et al. Conservative management of anastomotic leaks after 557 open gastric bypasses. Obes Surg 2005; 15: 1252 – 1256 9 Casella G, Soricelli E, Rizzello M et al. Nonsurgical treatment of staple line leaks after laparoscopic sleeve gastrectomy. Obes Surg 2009; 19: 821 – 826 10 Swinnen J, Eisendrath P, Rigaux J et al. Self-expandable metal stents for the treatment of benign upper GI leaks and perforations. Gastrointest Endosc 2011; 73: 890 – 899 11 Bège T, Emungania O, Vitton V et al. An endoscopic strategy for management of anastomotic complications from bariatric surgery: a prospective study. Gastrointest Endosc 2011; 73: 238 – 244 12 Eisendrath P, Cremer M, Himpens J et al. Endotherapy including temporary stenting of fistulas of the upper gastrointestinal tract after laparoscopic bariatric surgery. Endoscopy 2007; 39: 625 – 630 13 Eubanks S, Edwards CA, Fearing NM et al. Use of endoscopic stents to treat anastomotic complications after bariatric surgery. J Am Coll Surg 2008; 206: 935 – 938 14 Tan JT, Kariyawasam S, Wijeratne T et al. Diagnosis and management of gastric leaks after laparoscopic sleeve gastrectomy for morbid obesity. Obes Surg 2010; 20: 403 – 409 15 Serra C, Baltasar A, Andreo L et al. Treatment of gastric leaks with coated self-expanding stents after sleeve gastrectomy. Obes Surg 2007; 17: 866 – 872 16 Nguyen NT, Nguyen X-MT, Dholakia C. The use of endoscopic stent in management of leaks after sleeve gastrectomy. Obes Surg 2010; 20: 1289 – 1292 17 Oshiro T, Kasama K, Umezawa A et al. Successful management of refractory staple line leakage at the esophagogastric junction after a sleeve gastrectomy using the HANAROSTENT. Obes Surg 2010; 20: 530 – 534 18 Iqbal A, Miedema B, Ramaswamy A et al. Long-term outcome after endoscopic stent therapy for complications after bariatric surgery. Surg Endosc 2011; 25: 515 – 520 19 Simon F, Siciliano I, Gillet A et al. Gastric leak after laparoscopic sleeve gastrectomy: early covered self-expandable stent reduces healing time. Obes Surg 2013; 23: 687 – 692 20 Hadj Amor WB, Bonin EA, Vitton V et al. Successful endoscopic management of large upper gastrointestinal perforations following EMR using
Christophorou Dimitri et al. Endoscopic treatment of fistula after sleeve gastrectomy … Endoscopy 2015; 47: 988–996
Downloaded by: University of Florida. Copyrighted material.
Fistula after LSG
995
Original article
21
22
23
24
25
26
27
28
29
30
31
32 33
34
35
36
over-the-scope clipping combined with stenting. Endoscopy 2012; 44: E277 – E278 Surace M, Mercky P, Demarquay JF et al. Endoscopic management of GI fistulae with the over-the-scope clip system (with video). Gastrointest Endosc 2011; 74: 1416 – 1419 Mercky P, Gonzalez JM, Aimore Bonin E et al. Usefulness of over-thescope clipping system for closing digestive fistulas. Dig Endosc 2015; 27: 18 – 24 Papavramidis TS, Kotzampassi K, Kotidis E et al. Endoscopic fibrin sealing of gastrocutaneous fistulas after sleeve gastrectomy and biliopancreatic diversion with duodenal switch. J Gastroenterol Hepatol 2008; 23: 1802 – 1805 Pequignot A, Fuks D, Verhaeghe P et al. Is there a place for pigtail drains in the management of gastric leaks after laparoscopic sleeve gastrectomy? Obes Surg 2012; 22: 712 – 720 Toussaint E, Eisendrath P, Kwan V et al. Endoscopic treatment of postoperative enterocutaneous fistulas after bariatric surgery with the use of a fistula plug: report of five cases. Endoscopy 2009; 41: 560 – 563 Sakran N, Goitein D, Raziel A et al. Gastric leaks after sleeve gastrectomy: a multicenter experience with 2,834 patients. Surg Endosc 2013; 27: 240 – 245 Rosenthal RJ, Diaz AA, Arvidsson D et al. International Sleeve Gastrectomy Expert Panel Consensus Statement: best practice guidelines based on experience of >12,000 cases. Surg Obes Relat Dis 2012; 8: 8 – 19 Aurora AR, Khaitan L, Saber A. Sleeve gastrectomy and the risk of leak: a systematic analysis of 4,888 patients. Surg Endosc 2012; 26: 1509 – 1515 Nocca D, Krawczykowsky D, Bomans B et al. A prospective multicenter study of 163 sleeve gastrectomies: results at 1 and 2 years. Obes Surg 2008; 18: 560 – 565 Consten ECJ, Gagner M, Pomp A et al. Decreased bleeding after laparoscopic sleeve gastrectomy with or without duodenal switch for morbid obesity using a stapled buttressed absorbable polymer membrane. Obes Surg 2004; 14: 1360 – 1366 Shimi SM. Self-expanding metallic stents in the management of advanced esophageal cancer: a review. Semin Laparosc Surg 2000; 7: 9 – 21 Sabharwal T, Morales JP, Salter R et al. Esophageal cancer: self-expanding metallic stents. Abdom Imaging 2005; 30: 456 – 464 Dormann AJ, Eisendrath P, Wigginghaus B et al. Palliation of esophageal carcinoma with a new self-expanding plastic stent. Endoscopy 2003; 35: 207 – 211 Dormann A, Meisner S, Verin N et al. Self-expanding metal stents for gastroduodenal malignancies: systematic review of their clinical effectiveness. Endoscopy 2004; 36: 543 – 550 Thompson AM, Rapson T, Gilbert FJ et al. Endoscopic palliative treatment for esophageal and gastric cancer: techniques, complications, and survival in a population-based cohort of 948 patients. Surg Endosc 2004; 18: 1257 – 1262 Conio M, Repici A, Battaglia G et al. A randomized prospective comparison of self-expandable plastic stents and partially covered self-expandable metal stents in the palliation of malignant esophageal dysphagia. Am J Gastroenterol 2007; 102: 2667 – 2677
37 Hünerbein M, Stroszczynski C, Moesta KT et al. Treatment of thoracic anastomotic leaks after esophagectomy with self-expanding plastic stents. Ann Surg 2004; 240: 801 – 807 38 Puli SR, Spofford IS, Thompson CC. Use of self-expandable stents in the treatment of bariatric surgery leaks: a systematic review and meta-analysis. Gastrointest Endosc 2012; 75: 287 – 293 39 Mangiavillano B, Viaggi P, Masci E. Endoscopic closure of acute iatrogenic perforations during diagnostic and therapeutic endoscopy in the gastrointestinal tract using metallic clips: a literature review. J Dig Dis 2010; 11: 12 – 18 40 Qadeer MA, Dumot JA, Vargo JJ et al. Endoscopic clips for closing esophageal perforations: case report and pooled analysis. Gastrointest Endosc 2007; 66: 605 – 611 41 Binmoeller KF, Grimm H, Soehendra N. Endoscopic closure of a perforation using metallic clips after snare excision of a gastric leiomyoma. Gastrointest Endosc 1993; 39: 172 – 174 42 Minami S, Gotoda T, Ono H et al. Complete endoscopic closure of gastric perforation induced by endoscopic resection of early gastric cancer using endoclips can prevent surgery (with video). Gastrointest Endosc 2006; 63: 596 – 601 43 Sebastian S, Byrne AT, Torreggiani WC et al. Endoscopic closure of iatrogenic duodenal perforation during endoscopic ultrasound. Endoscopy 2004; 36: 245 44 Tang S-J, Tang L, Gupta S et al. Endoclip closure of jejunal perforation after balloon dilatation. Obes Surg 2007; 17: 540 – 543 45 Magdeburg R, Collet P, Post S et al. Endoclipping of iatrogenic colonic perforation to avoid surgery. Surg Endosc 2008; 22: 1500 – 1504 46 Kirschniak A, Kratt T, Stüker D et al. A new endoscopic over-the-scope clip system for treatment of lesions and bleeding in the GI tract: first clinical experiences. Gastrointest Endosc 2007; 66: 162 – 167 47 Cellier C, Landi B, Faye A et al. Upper gastrointestinal tract fistulae: endoscopic obliteration with fibrin sealant. Gastrointest Endosc 1996; 44: 731 – 733 48 Akhras J, Tobi M, Zagnoon A. Endoscopic fibrin sealant injection with application of hemostatic clips: a novel method of closing a refractory gastrocutaneous fistula. Dig Dis Sci 2005; 50: 1872 – 1874 49 Lippert E, Klebl FH, Schweller F et al. Fibrin glue in the endoscopic treatment of fistulae and anastomotic leakages of the gastrointestinal tract. Int J Colorectal Dis 2011; 26: 303 – 311 50 Pross M, Manger T, Reinheckel T et al. Endoscopic treatment of clinically symptomatic leaks of thoracic esophageal anastomoses. Gastrointest Endosc 2000; 51: 73 – 76 51 Truong S, Böhm G, Klinge U et al. Results after endoscopic treatment of postoperative upper gastrointestinal fistulas and leaks using combined Vicryl plug and fibrin glue. Surg Endosc 2004; 18: 1105 – 1108 52 El Hassan E, Mohamed A, Ibrahim M et al. Single-stage operative management of laparoscopic sleeve gastrectomy leaks without endoscopic stent placement. Obes Surg 2013; 23: 722 – 726 53 O’Connor L, Champagne BJ, Ferguson MA et al. Efficacy of anal fistula plug in closure of Crohn’s anorectal fistulas. Dis Colon Rectum 2006; 49: 1569 – 1573 54 O’Riordan JM, Datta I, Johnston C et al. A systematic review of the anal fistula plug for patients with Crohn’s and non-Crohn’s related fistulain-ano. Dis Colon Rectum 2012; 55: 351 – 358
Christophorou Dimitri et al. Endoscopic treatment of fistula after sleeve gastrectomy … Endoscopy 2015; 47: 988–996
Downloaded by: University of Florida. Copyrighted material.
996
