Leakage after sleeve gastrectomy: Endoscopic stenting VS surgical intervention

Patients and method: Our study included 30 patients presented with post sleeve leaks discovered by routine postoperative imaging or during the follow up period. Patients were recruited from October 6th university hospital during the period from August 2017 to August 2019. Patients were divided to the following groups: 1) Endoscopy group: This included 15 patients with post sleeve leakage undergoing endoscopic stent insertion. 2) Surgery group: which included 15 patients with post sleeve leak age undergoing surgical management. This division was random.


Introduction
Laparoscopic sleeve gastrectomy (LSG) is getting more popularity for the treatment of morbid obesity. LSG is a safe procedure with a low complication rate. The complications encountered nevertheless can result in morbidity and even mortality. The most signi icant complications are staple-line bleeding, stricture, and staple-line leak [1].
Leaks are the major complication associated with LSG with a reported prevalence between 1.9% and 2.4%. Most leaks occur at the proximal part of the staple line, at the oesophagogastric junction [2].
Leaks can be classi ied based either on the time of onset (early, intermediate and late leaks), clinical presentation, site of leak, radiological appearance, or mixed factors. By clinical relevance and extent of dissemination, they de ined type I or subclinical leaks and Type II leaks [3]. presentation can vary from asymptomatic to severe septic shock. Usual symptoms may be of the septic nature: fever, tachycardia, tachypnea, leukocytosis, abdominal pain, and peritonitis, tachycardia > 120 beats per minute (bpm) may be the most diagnostic sign of a gastric leak. 1 Burgos, et al. reported 7 leaks in 214 patients (3.3%), of which 5 patients presented abdominal pain, fever, tachycardia, tachypnea and increased laboratory signs of infection. They observed that tachycardia is an initial sign of early leak [4].
Debates still exist on which diagnostic modality is the most sensitive and speci ic concerning the diagnosis of a post sleeve gastrectomy leakage, but all of them agree that early detection is associated with better outcome, and that a high index of suspicion is the cornerstone in the detection and diagnosis of leak [3].
The management of leak post sleeve gastrectomy imposes a lot of controversies and dif iculties in the adoption of a standard algorithm, classi ication for gastric leakage post sleeve may constitute the irst step in the establishment of such an algorithm or protocol, based on 3 characteristics: Time of appearance (early, intermediate and late); Location (proximal, mid or distal gastric); Severity or magnitude (type I and II) [3].
Treatment principles for leaks that occur after bariatric surgery include medical management (intravenous antibiotics and nutritional support), drainage of extra digestive soiling and stopping the leakage. Only once adequate drainage of any leak-associated luid collections has been performed closure can be considered [5].
Classic surgical management of leaks consists of early reintervention to close or patch the gastric or anastomotic defect. These surgeries are associated with high morbidity and mortality. Reinterventions such as gastrectomy, gastric bypass on a complicated sleeve, or istulojejunal anastomosis might be associated with less morbidity when performed on a more chronic defect [6].
Endoscopic management of postsurgical leaks offers the advantage of being less aggressive than surgery and include techniques to: cover (or exclude) the defect; close the defect; or temporarily maintain the defect in an open con iguration to provide for internal drainage of the luid collection [7].
Endoscopic options to treat leakage include: partially covered metallic self-expandable stent (PCSES) or fully covered stents, clipping of the defect, over-the-scope clips system (OTSCs), endoscopic insertion of a pigtail, ibrin sealant and suturing devices [8].
Over the last decade, there has been increasing use of selfexpanding metal stents (SEMS) for the treatment of sleeve leaks. The objective of stenting has been to divert gastric contents from the istula site and to bypass the distal stenotic portion if present. Casella and colleagues reported the use of endoscopic stents for sleeve leaks in three patients with 100 % success [9].
Surgicaloptions for leak management include primary repair and bowel, gastric or omentalpatching.
More de initive surgical options include: conversion of the LSG to a regular Roux-Y gastric bypass (RYGB) anastomosis of the jejunal Roux limb to the istula and total gastrectomy [11].

Aim of the work
To compare surgical intervention and endoscopic stenting for treatment of gastric leakage after sleeve gastrectomy.

Study design
This study will be conducted at October 6 university hospital during the period from August 2017 to August 2019 and will include 30 patients presented with post sleeve leaks discovered by routine postoperative imaging or during the follow up period. The study will include early hemodynamically stable type I leakage cases who failed to respond to conservative medical treatment. 15 patients will undergo endoscopic stent insertion and 15 patient will undergo surgical intervention. All patients were consented to participate in the study. Ethical approval from October 6 university hospital Ethical Committees was obtained.
Patient inclusion criteria: the study will include Type I leakage Hemodynamically stable patients.
Exclusion criteria: Unstable cases presented with septic shock or peritonitis.

Methods
Pre-operative assessment of a leaking patient included the following: History taking (age, sex, BMI, date of operation, onset of symptoms, history of acute or chronic illness). clinical examination including vital signs and local abdominal examination. Routine preoperative investigations (CBC, liver function tests, kidney functions test and electrolytes).

Management
1. Drainage of the extra digestive space is part of the initial treatment strategy for all cases.
2. Conservative management include: nutritional support, IV antibiotics and good monitoring.
3. Surgical management include: direct suturing for closure of the defect or patching of the defect.
4. Endoscopic management: insertion of partially covered metallic self-expandable stent (PCSES) for 6-8 weeks with the possibility of reinsertion of another stent after this period if leakage is not controlled (possible complications include: stent migration-failure to control-mucosal hyperplasia and ulceration).

Postoperative assessment
Follow up after management of leakage: to compare surgical management (group A) with endoscopic management (group B) will include: Follow up till patient is discharged from hospital by: Vital signs (pulse, temperature), drains output, early postoperative complications, post-operative hospital stay, CBC and CT scan for rate of resolution.
After hospital discharge: weekly follow up till leakage is controlled by: Evaluation of leak output and rate of control (drains, CT scan) and development of complications.
After leakage control: monthly follow up for 6 months for: Long term complications (recurrence, stricture).

Statistical analysis
An Excel spreadsheet was established for the entry of data. We used validation checks on numerical variables and option-based data entry method for categorical variables to reduce potential errors. The analyses were carried with SPSS software (Statistical Package for the Social Sciences, version 24, SSPS Inc, and Chicago, IL, USA). The normality of the data were assessed using Shapiro-Wilk Test. Numerical data were described as mean ± SD if normally distributed; or median and interquartile range [IQR] if not normally distributed. Frequency tables with percentages were used for categorical variables. Independent Student t-test and paired t-test were used to compare parametric quantitative variables; while Mann-Whitney tests and Wilcoxon matched pairs test were used to compare non-parametric quantitative variables. Chi-square test or McNemar-Bowker tests were used to analyze categorical variables. Multilinear logistic regression was undertaken to assess the predictors of mortality. A p -value < 0.05 is considered statistically signi icant.
Prior to start we got approval from the ethical committee in our faculty with a written informed consent from every patient. Table 1 shows that there was no statistically signi icant association between type of management and age (p = 0.88), BMI (p = 0.14), gender (p = 0.72), and comorbidities (p = 0.47). Table 2 shows that there was no statistically signi icant association between type of management and leak site (p = 0.48) and Interval between surgery and leak (p = 0.147). Table 3 shows that there was no statistically signi icant association between type of management and initial treatment (p = 0.47). Table 4 shows that there was no statistically signi icant association between type of management and outcomes (p = 0.51). In contrary, there was statistically signi icant association between type of management and interval between closure and leak control (p < 0.001) and hospital stay (p = 0.013). Patients underwent endoscopic management showed shorter hospital stay. -The interval means that the patient became stable and not in need of more hospitalization.

Discussion
Laparoscopic sleeve gastrectomy (LSG) is emerging to be one of the commonly performed bariatric procedures worldwide for patients with different degrees of obesity. This restrictive procedure has several advantages. It is technically simpler to perform without the need of an anastomosis. It induces a reduction in ghrelin causing appetite suppression, which adds to the effect of restriction. It has been reported to have a lower morbidity and mortality rate in comparison to Roux-en-Y gastric bypass or biliopancreatic diversion with or    without duodenal switch. It can be performed concomitantly with other procedures [12].
LSG can be associated with three signi icant complications, which include staple line gastric bleeding, staple line gastric leaks and gastric strictures. Of these, a gastric leak after sleeve gastrectomy is associated with signi icant and prolonged morbidity, remaining one of the most feared complications. Gastric leak can present as peritonitis, abscesses, cutaneous or other istulas, sepsis, organ failure and even death [13].
The treatment for leaks after sleeve gastrectomy varies and depends upon the extent of disruption, the extent of abdominal contamination, and the site of leak (proximal versus distal). Endoscopy is an excellent irst line tool and may be simultaneously diagnostic and therapeutic. Endoscopic stents were initially designed as a tool of palliation for obstructing esophageal, gastric, and colorectal cancer. Over the recent years, a growing body of evidence has shown that stent placement for gastro-jejunal leaks is a safe therapeutic option [14].
Nevertheless, there is a scarcity in the published literature regarding the ef icacy and safety of endoscopic stenting for management of leakage after LSG. Therefore, we conducted the present study in order to compare surgical intervention and endoscopic stenting for treatment of gastric leakage after LSG.
In the present study, the mean age of the included patients was around 35 years old and the majority of the patients were females. In addition, more than two-third of the patients had one or more comorbidities this was in agreement with, Juza and colleagues [15].
Leaks can be classi ied based either on the time of onset, clinical presentation, site of leak, radiological appearance, or mixed factors. Csendes, et al [16]. de ined early, intermediate and late leaks as those appearing 1 to 4, 5 to 9 and 10 or more days following surgery respectively The most frequent site of leak, in a vast majority of patients with SG, is proximal, near the gastroesophagal (GE) junction [14].
In line with these indings, Sakran and colleagues [17] performed a retrospective analysis by querying 2,834 patients who underwent LSG, 44 (1.5%) with gastric leaks were identi ied. Of these 44 patients, 30 (68%) were women. The patients had a mean age of 41.5 years Leaks were diagnosed at a median of 7 days postoperatively: early (0-2 days) in nine cases (20%), intermediately (3-14 days) in 32 cases (73%), and late (> 14 days) in three cases (7%). In 33 of the patients (75%), the leak site was found in the upper sleeve near the gastroesophageal junction.
Similarly, Vix and colleagues [18] aimed to report the leak rate and its management in 378 LSGs. The overall leak rate was 9/378 (2.38%). The leaks were mostly at GE junction and intermediate.
Regarding the primary outcomes of the present study, the endoscopic stents achieved leak closure in13 (86.7%) patients and only 2 patients required conversion to Roux en Y. The interval between closure of leak and leak control was 3.36 ± 5.4 days. The average hospital stay was 8.7 ± 4.1 days. The analysis showed that there was no statistically signi icant association between type of management and outcomes (p = 0.51). In contrary, there was statistically signi icant association between type of management and interval between closure and leak control (p < 0.001) and hospital stay (p = 0.013). Patients underwent endoscopic management showed shorter interval till leak control and hospital stay.
In concordance with our indings, Puli and colleagues [19] performed a systematic review and meta-analysis to evaluate the success of self-expandable stents (SESs) in the treatment of bariatric surgery leaks. A total of 189 relevant articles were reviewed of which 7 studies (67 patients with leaks) met inclusion criteria. The pooled proportion of successful leak closures by using SESs was 87.77% (95% CI, 79.39%-94.19%). The pooled proportion of successful endoscopic stent removal was 91.57% (95% CI, 84.22%-96.77%).
Similarly, Murino and colleagues [20] evaluated the effectiveness of a endoscopic stents for management of postbariatric surgery leaks in a large cohort of patients. Data from patients with anastomotic leaks after bariatric surgery endoscopically treated with partially covered SEMS between January 2006 and December 2012 were retrospectively reviewed. The stenting policy was successful to close the leak in 88% of the patients. Southwell and colleagues [21] described the experience of endoscopic management of post-LSG leaks. A total of 21 patients have received endotherapy for post-LSG leak management. Treatment included the deployment of primary SEMS across the leak site. A total of 20/21 (95%) patients had resolved leaks following a mean of 75 days of treatment (median 47, range 9-187).
While, Garofalo and colleagues [22] included a total of 872 LSGs, overall, 10 of 872 patients (1.1%) developed a gastric leak. Endoscopic istula closure at the gastroesophageal junction was achieved in 10 of 11 cases (90%) and the average time for closure was 9.9 (range: 4-24) weeks.
Moreover, Tsai and colleagues [23] evaluated the ef icacy and complications of SEMS in the treatment of post-bariatric surgical leak. seven patients underwent covered SEMS placement for leak after bariatric surgery, including LSG (n = 6) Among six patients, one patient who received stent placement one year after leak diagnosis failed to achieve leak closure, and ive patients with early stent placement achieved leak closure (83.3%).
Eubanks and colleagues [24] presented the outcomes of large series to date treating staple line complications after bariatric surgery with endoscopic covered stents. A retrospective evaluation was performed of all patients treated for staple line complications after bariatric surgery at a single tertiary care bariatric center. 19 patients were treated with a total of 34 endoscopic silicone covered stents. Resolution of leak after stent treatment occurred in 16 of 19 patients (84%).
Complications of stent insertion include early removal of the stent due to stent migration, bleeding, and obstruction due to kinking of the proximal portion of the stent. Stent migration is the main drawback of the technique requiring stent removal [25].
In the present study, the incidence of short-term complications in endoscopy group were, restenting due to failure of therapy (6.7%), stent migration (26.7%), and subphrenic abscess (6.7%). In long-term follow-up, 2 patients (13.4%) developed stent-related ulcer and a similar number of patients developed stricture.
In agreement with our indings, Campo and colleagues [26] described single-institution experience in managing SG leaks with endoscopic stents. Data for all patients who underwent endoscopic stent placement for an SG leak Twenty-four patients with SG staple-line leaks treated with covered endoscopic stents were identi ied. Migration occurred in 22% of all stent placements.
Botaitis, et al. [27] support primary surgical repair of the defect and immediate surgical intervention with wash out drainage and primary repair for patients with early leaks due to the fact that the surrounding tissues are still healthy. More de initive surgical options include: conversion of the LSG to a regular Roux-Y gastric bypass (RYGB), anastomosis of the jejunal Roux limb to the istula and total gastrectomy. Gomes suggested that converting a sleeve to a RYGB leads to decompression of the high intragastric pressure within the sleeve to a low pressure system. Also a Roux limb allows for better drainage than a sleeve, which can have functional disorders or stenotic areas. Conversion of the LSG to a RYGB may not be advisable in the presence of signi icant peritonitis [11] studies have shown that early re-suturing within the irst three days can result in successful closure versus resuturing leaks after the third day. Hence this is considered as a 'favorable' window period and attempt at early surgical closure of the defect may be performed when re-exploration is early and tissues are healthy. If possible re-sleeve of the istula site by stapling can be done with suture reinforcement [12].
In our study 7 patients were managed by early surgical intervention with drainage and direct suturing for the leak site as the tissues weren't friable and the repair was successful.
Despite the overall complications rates were comparable between endoscopy and surgery arms, there were notable more severe complications in the surgery group. DVT, wound infection, and severe vomiting can have devastating consequences in postoperative period. For example, the autopsy studies document that 50% of all patients dying in hospital have DVT. Such indings may explain the favorable outcomes of endoscopic stenting in terms of hospital stay and time to return to normal function.

Conclusion
Endoscopic management of post-sleeve gastrectomy leakage with stenting is recommended because it successfully manages the leaks and avoids invasive procedures with less risk. , with shorter hospital stay and early return of function.