Leak Closure Research Articles

Endoscopic Vacuum Therapy in the Management of Postoperative Leakage After Esophagectomy.

Postoperative leakage after esophagectomy is associated with significant life-threatening complications. Recently, endoscopic vacuum therapy (EVT) was introduced and has been successfully used as a new treatment option. The purpose of this study was to evaluate the safety and efficacy of EVT for the management of postoperative leakage after esophagectomy. A total of 22 patients were treated with either intraluminal or intracavitary EVT for the management of postoperative leakage from May 2012 to April 2018. The location of leakage was intrathoracic in 17 patients and cervical in five patients. The outcomes of EVT were analyzed retrospectively. Complete postoperative leakage closure was achieved in 19 of 22 patients. The median duration of EVT application was 14days (range 2-103), and a median number of three EVT systems (range 1-14) were used. In 19 patients who were successfully managed with EVT, oral feeding was possible a median of 15days after the first day of treatment. There were no cases of mortality related to postoperative leakage. EVT is a well-tolerated and effective therapeutic option for the treatment of various types of postoperative leakage after esophagectomy.

1835 Closure of a Large Gastroesophageal Anastomosis Leak With Endoscopic Suturing After the Failure of Stent Therapy

INTRODUCTION: Gastroesophageal anastomotic leaks remains a potentially life-threatening post-esophagectomy complication. Surgical repair or conservative therapy have been used traditionally with suboptimal outcomes. Endoscopic methods are now an effective alternative to surgery for the management of anastomotic leak. We report a case of closure of a large Gastroesophageal anastomotic leak using a combination of endoscopic suturing with OverStitch (Apollo Endosurgery, Austin, Texas, USA) and esophageal stent after patient failed to respond to esophageal stent alone. CASE DESCRIPTION/METHODS: A 49-year-old male with a history of adenocarcinoma of distal esophagus underwent distal esophagectomy and placement of a percutaneous gastrostomy tube. The postoperative course was complicated by sepsis and empyema. Broad-spectrum antibiotics were started, and a percutaneous drain was placed in the mediastinum for drainage of empyema. EGD was then performed by the surgical team, which led to the identification of dehiscence of the gastroesophageal anastomosis and leak. A fully covered, 23 mm by 105 mm esophageal metal stent was placed. The patient was not able to tolerate liquids orally and continued to complain of nausea and vomiting. He continued to have a large amount of drainage from the mediastinal percutaneous drain. Repeat esophagram 8 weeks post op revealed persistent leak at the anastomosis (Figure 1). Gastroenterology was then consulted for endoscopic closure of the leak. Endoscopy was performed, and the fully covered metal stent was removed. Endoscopic examination after stent removal revealed a gastroesophageal anastomosis defect about 20 mm in diameter with necrotic appearing mucosa at the site of the leak (Figure 2). The anastomotic defect was then successfully closed with approximating the healthy gastric and esophageal mucosa. A figure-of-8 pattern of endoscopic sutures was placed using the Apollo OverStitch device. Another layer of reinforcement sutures was also placed. One fully covered 23 mm by 105 mm esophageal stent was placed over the anastomosis. The patient was started on a clear liquid diet. Esophagram performed on day 5 (Figure 1) showed no anastomotic leak. Diet was then further advanced, and the patient was discharged home. DISCUSSION: Large gastroesophageal anastomotic leak remains a potentially life-threatening complication post-esophagectomy. Patients failing stent therapy usually with large leaks can be successfully treated with a combination of endoscopic suturing and stent placement.

872 Clip It and Forget It? A Natural History of 358 Over-the-Scope Clips

INTRODUCTION: Over-the-scope clips (OTSC) are used for closure of fistulas, leaks, and acute perforations, and for GI bleeding. Adverse events include perforation, migration, and ulceration. Little is know about the natural history of OTSCs and no data addressing the longevity of clips is available. We present an interim analysis of our center’s experience with OTSC focusing on natural history. METHODS: OTSCs used from 1/1/2011 to 1/1/2019 were compiled and retrospective chart review undertaken. Primary endpoint was clip retention time (time when the clip was placed to time the clip was last seen or was physically removed). Secondary endpoints were rate of complications and factors that influence clip retention time and clinical success. Cox proportional hazards were used to evaluate variables impacting clip retention. Univariate regression was used to assess potential risk factors for adverse events. RESULTS: A total of 358 OTSC were placed in 299 patients (55.8% male, n = 184) with mean age 64 + 15.3 years. Procedures and indications presented in Table 1. Clip placement was clinically successful in 81% of cases (N = 265). Median time to clip loss as assessed by Kaplan-Meier curve was 267 days (95% CI 196-406) (Figure 1). Adverse events occurred in 34 clip placements (10.7%) and included bleeding (n = 8, 24.2%), perforation (n = 5, 15.2%), ulceration (n = 3, 9.1%), migration (n = 15, 45.5%), and infection (n = 1, 3.0%). Univariate analysis (Table 2) revealed that age, gender, type of clip used (there are 3 types of OTSC), and performing endoscopist did not impact clip retention rate. A size 14 clip had marginally higher rates of retention compared to others. Clips placed in the rectum or esophagus were lost earlier compared to those in the small/large intestine and stomach. Clips placed for fistulas and those associated with adverse event were associated with earlier clip loss. In evaluation of clinical success, clips placed for fistulas had a relative risk of clinical failure of 2.08 (95% CI 1.29-3.36) compared to all other indications. CONCLUSION: OTSCs are retained for a median of 8.9 months with clip location and indication for placement being key factors in time to clip loss. Clips used for fistula closure were more likely to be lost compared to those placed for other indications, less likely to be clinically successful, and more likely to cause a complication. These data help guide decision-making when considering OTSC use across clinical indications.

IBS26.03 Surgical Technique of Pleurectomy/Decortication

IBS26.03 Surgical Technique of Pleurectomy/Decortication

1798 An Endoscopic Vacuum-Assisted Closure Technique Modified Through the Overtubes for Patients With Esophageal Anastomotic Leaks

INTRODUCTION: Endoscopic vacuum-assisted closure (E-VAC) therapy is safe and effective for esophageal anastomotic leaks. However, the repetitive exchange of sponge tubes is technically demanding. CASE DESCRIPTION/METHODS: We designed an E-VAC technique, modified through the use of overtube, to make E-VAC therapy safer and more comfortable for both endoscopists and patients. Intraluminal E-VAC therapies were used for the treatment of post-operative thoracic esophageal leaks in five cases of esophageal cancer. After overtube insertion, sponge tube was inserted through the overtube, while keeping biopsy forcep inside the nasogastric tube (NGT), and was positioned at intraluminal leakage site under endoscopic guidance. NGT was repositioned from the mouth to the nostrils using carrier tube and a controlled negative pressure was applied. Sponge tube was regularly changed twice a week until the complete closure of leaks. Technical and clinical success rates were analyzed to evaluate the safety and efficacy of this technique. The modified E-VAC therapy was applied in five selected cases for a mean of 13.8 days (range, 7-28 days) with a mean of 3.2 sponge tube changes (2-6 changes) for anastomotic leaks that were, on average, 1.0 cm (0.5-2 cm). The mean procedural time was 14.9 min (12-30 min) and no E-VAC therapy-related complications occurred. All patients were discharged, on average, after 44.4 days (range, 14-95 days) of hospitalization and two cases were treated with additional esophageal metal stents after six and two sponge tube changes, respectively. DISCUSSION: Our E-VAC therapy technique, modified via inserting sponge tubes through overtube, is effective and safe for patients with thoracic esophageal anastomotic leaks.

2513 Endoscopic Management With an Over-the-Scope Gastric Clip of a Staple-Line Leak Following Sleeve Gastrectomy

INTRODUCTION: Obesity has become a global epidemic with substantial socioeconomic and healthcare-related costs. Due to the limited utility of diet and behavior modification, there is tremendous interest in surgical weight loss strategies. Among those, sleeve gastrectomy (SG) has become the most commonly performed bariatric procedure. This is due to its efficacy and relative safety compared to other surgical options. However, post-SG leaks do occur. For example, staple-line leak (SLL) occurs in up to 3% of patients, most commonly at the proximal border of the staple line, near the esophagogastric junction. Usually, these leaks present in the early postoperative period, but there are increasing reports of SLL occurring months or even years after the initial surgery. Often, bariatric surgical complications are referred to tertiary centers for management. We present a case of a SLL managed at a community hospital. CASE DESCRIPTION/METHODS: A 38 year-old woman, with limited past medical history, presented 6 weeks after SG with a staple-line leak. Her immediate post-op period had been complicated by a peri-splenic abscess, which was managed with IV antibiotics (ABX) and percutaneous drainage (PD). Upper GI series (UGIS) at this time did not reveal a SLL. After clinical and radiographic resolution, the drain was removed. Several days later, the patient presented with nausea, vomiting, and non- bloody watery diarrhea. She was found to be febrile and tachycardic; labs, including a test for Clostridium dificile, were negative. A repeat CT scan confirmed an upper abdominal fluid collection. Subsequent UGIS revealed a SLL in the proximal stomach with a suture line fistula. After several weeks of NPO status, ABX and PD failed to lead to leak closure, the patient underwent endoscopy with successful placement of an Ovesco Over-the-Scope gastric clip (OTSC). Repeat UGIS confirmed fistula closure. After several weeks of additional conservative management, the PD was removed. Several months later, the patient remains stable. DISCUSSION: As the use of SG becomes more widespread, complications such as SLL will need to be addressed. Often, patients with SLL are referred to a tertiary center. We present a case successfully managed in a community hospital with the application of an OTSC. Practitioners who follow SG patients should be familiar with the indications and use of the OTSC.

P1001Safety and effectiveness of transcatheter leak occlusion with detachable coils following left atrial appendage closure

Abstract Introduction Incomplete left atrial appendage (LAA) closure resulting in residual, significant leak is common after LAA exclusion procedures, potentially hindering an effective thromboembolic (TE) prevention. Detachable coils have found a wide range of applications for transcatheter vascular occlusion/embolization procedures. Purpose We report for the first time the feasibility and efficacy of transcatheter leak closure with detachable coils in patients with incomplete LAA closure. Methods Thirtyconsecutive high TE-riskpatients with a clinically-relevant residual leak (mean age: 72±9 years; 73.3% males; CHA2DS2-VASc: 4.3±1.5; HAS-BLED: 3.4±1.1)underwent percutaneous closure of the LAA patency via detachable coils (Interlock-35 Fibered IDC Occlusion System, Boston Scientific Inc.; Concerto Helix Detachable Coil System, ev3 Endovascular Inc; Azur Embolization System, Terumo Interventional Systems). Trans-oesophageal echocardiography (TEE) was performed at 45±15 days post-procedure to assess for residual LAA patency. Results LAA closure had been previously attempted with a Watchman device in 25 patients, an Amulet in 2 and a LARIAT in 3. At pre-procedural TEE, a moderate/severe leak was documented in 28 (93.3%) patients, a mild one in 2 (6.7%) patients. Procedure andfluoroscopy times were 73±40min and 21±15min, respectively; the mean volume of iodinated contrast medium was 80±48mL. Overall, 114 coils were released into patents' LAA (median: 3 coils/patient; IQR: 2–4). Coil deployment was successful in 98.2% of cases (112/114). The overall complication-free rate was 93.3% (28/30). There were no instances of vascular injury, TE events, device dislodgment. There was 1 major adverse device-related event in an 87-yo male, who developed pericardial tamponade requiring a pericardial window. A small pericardial effusion without hemodynamic compromise was noted in a 78-yo woman and treated conservatively. Follow-up TEE in all patients revealed complete LAA sealing or a minimal-mild leak in 28 cases (93.3%; 22 with no residual leak, 6 with a minimal-mild residual one), and a moderate residual leak in 2 patients (6.7%). Conclusions Transcatheter LAA leak occlusion via detachable coils is a safe, effective and promising approach in high TE-risk patients with incomplete LAA closure.

TCT-256 Minimally Invasive Transradial Approach for Percutaneous Closure of Aortic Paravalvular Leaks: A Single-Center Experience

TCT-256 Minimally Invasive Transradial Approach for Percutaneous Closure of Aortic Paravalvular Leaks: A Single-Center Experience

Added Value of Practicing Cardiac Interventions Under Fluoroscopy Using Patient-Specific 3D Printed Cardiac Models

ABSTRACT Background: 3D printing technology has been used to practice cardiac structural interventions in a benchtop setting. We hypothesized that practice on such models under fluoroscopy using standard access and deployment techniques would be feasible and create more realistic simulations. Methods: Computed Tomography Angiography data was segmented with commercial software and models exported for 3D printing. The resulting models were fixed to a catheterization table and the planned procedures attempted under fluoroscopy. Operator success and failures were noted and compared with actual case outcomes. Results: Three cases investigating left atrial appendage ligation and paravalvular leak closure were successfully simulated, with associated outcomes in the actual cases paralleling the simulated experiences. Conclusion: Patient-specific 3D printed cardiac models are easily visualized under fluoroscopy, with an anatomic fidelity sufficient to permit simulation and practice of structural interventions. Further work is needed to prove the value of this technology for improving patient outcomes and optimizing the training of future operators.

Percutaneous valve‐in‐valve procedure and simultaneous paravalvular leak closure

Concern for early degeneration limits the use of bioprosthetic heart valves. A 77-year-old man who underwent surgical aortic valve replacement at age 70 for severe aortic stenosis (AoS) presented with premature bioprosthesis degeneration and AoS recurrence. Transthoracic echocardiography demonstrated severe AoS and aortic regurgitation, a 30% ejection fraction, and pulmonary hypertension. Transesophageal echocardiography revealed that the aortic regurgitation was due to a 5-mm paravalvular leak (PVL). A high EuroScoreII excluded surgical treatment. Simultaneous transcatheter aortic valve replacement and PVL closure with an Occlutech PLD Square 5 Twist PVL closure device were performed with good results and improved clinical status.

Multiple paravalvular leak closures

Introduction 70 year old man with paroxysmal atrial fibrillation and heart failure with preserved ejection fraction with previous bioprosthetic aortic and mitral valve replacement in 2003 and 2008 presented for fatigue and dyspnea. He was an unsuitable candidate for surgery due a EuroSCORE II of 7.27%. Preoperative TEE showed posterior mitral periprosthetic regurgitation (peak gradient 73.3 mmHg) for which he underwent a transcatheter closure using an Amplatzer Vascular Plug (AVP) 2 size 12 mm and 18 mm with improved regurgitation. Methods Arterial line was placed and general anesthesia induced. Using a 6 french and 14 french catheters, the left femoral artery and right femoral vein were accessed, respectively. The septum was crossed posteriorly in order to avoid the aortic valve. TEE demonstrated a 1.4cm by 0.6cm mitral leak – almost 1/3 the circumference of the biopresthesis, as well as a left atrial width of 5.8cm. After placement of device, TEE demonstrated coverage of 2/3 of the original paravalvular orifice with regurgitation decreased significantly, as well as S wave blunting on pulmonary vein flow. Results A month postoperatively, the patient returned with worsening fatigue, shortness of breath, and weakness. In addition to anemia, patient showed hemolysis. Hematology was consulted and they believed that his PVL was again the etiology. Thus, he underwent repeat attempt at closure of the mitral in addition to early closure of the aortic paravalvular leak. Of note, TEE showed previously closed leak measuring 9mm by 9mm. There were no immediate postoperative complications. Several days later however, the patient suffered decompensated heart failure leading to multisystem organ failure. Discussion Transcatheter leak closure results in significant decrease in regurgitation. Moreover, mortality rate within 30 days averages 1.4-2.0% and 3-year survival rate is 64.5%. When both aortic and mitral prosthesis are involved, it is generally accepted to close the most regurgitant lesion and return in about 6 months for closure of the second lesion. Followup TTE/TEE and labs with close attention to hemolysis is also warranted for possible recurrence.

Risk of thromboembolic events after percutaneous left atrial appendage ligation in patients with atrial fibrillation: Long-term results of a multicenter study

Percutaneous left atrial appendage (LAA) occlusion with Lariat has emerged as a viable alternative to oral anticoagulation (OAC) to prevent thromboembolic (TE) events in patients with atrial fibrillation. We evaluated the long-term TE risk in post-Lariat patients. Consecutive patients undergoing LAA ligation with the Lariat device at multiple centers with at least 1-year follow-up were included in the analysis. Transesophageal echocardiography (TEE) was performed at 4 weeks, 6 months, and 12 months to assess the completeness of LAA occlusion. OAC was discontinued if 4-week TEE revealed no device-related thrombus and complete closure of the appendage. Patients remained on 81 mg of aspirin per day after discontinuation of the blood thinner. A total of 306 patients were included in the study (mean age 68.8 ± 11.0 years; mean CHA2DS2-VASc score 3.6 ± 1.7). Four-week TEE revealed leaks in 81 patients (26.5%); all leaks were less than 5 mm in diameter. At 6-month TEE, spontaneous closure of the leak was demonstrated in 21 patients (25.9%), 26 patients (32%) underwent a successful leak closure procedure, and the remaining 34 (42%) patients were placed on OAC. At the median follow-up period of 15.9 ± 9.2 months, 9 TE events (2.9%) were reported: 7 with persistent leak and 2 without any detectable leaks on 2-dimensional TEE (P < .001). Complete occlusion of the LAA with the Lariat device was associated with the low rate of TE events at long-term follow-up. However, residual leaks were common after Lariat closure and the stroke rate was significantly higher in patients with incomplete occlusion, even with small leaks.

Successful endoscopic closure of an esophageal leak after endoscopic ultrasound-guided hepaticoesophagostomy by using n-butyl-2 cyanoacrylate

Successful endoscopic closure of an esophageal leak after endoscopic ultrasound-guided hepaticoesophagostomy by using n-butyl-2 cyanoacrylate

Clinical Utility of CT Fusion Imaging in Guiding Trans-catheter Paravalvular Leak Closure

Paravalvular leaks (PVL) occur in 2–12% of patients after mitral valve replacement and in 1–5% after aortic valve replacement 1,2 and can cause heart failure and/or hemolytic anemia. Redo surgery h...

Intradural non-calcified thoracic disc herniation causing spontaneous intracranial hypotension: a case report

BackgroundSpontaneous intracranial hypotension (SIH) is a rare pathology caused by a cerebrospinal fluid (CSF) leak. If intractable by conventional methods (i.e. bedrest, analgesics, or epidural blood patching) it may lead to the inability of the patient to cope with daily life and eventually to life-threatening complications. Recently, calcified discogenic microspurs or dorsal osteophytes were identified as a major cause for ventral CSF loss through vertical longitudinal dural slits. We report a rare case of intractable SIH due to an intradural disc herniation at the thoracolumbar junction (without signs of calcification) and its management.Case presentationA 46-year old woman suffered from orthostatic headache (sudden onset, no history of trauma) due to intractable SIH for over 2 month (without neurologic deficits). There was no clinical amelioration by conservative measures (analgesics, bedrest) and serial unspecific epidural blood patches (repeated for 3 times). She was diagnosed with an intradural disc herniation at the thoracolumbar junction causing a CSF leak. Surgical exploration by a translaminar and transdural approach with removal of the disc herniation and closure of the CSF leak was performed with immediate cessation of orthostatic symptoms. Histological workup revealed non-calcified intervertebral disc material. After 3 months of follow-up and no evidence for clinical relapse the patient returned to work.ConclusionsWe report the rare phenomenon of an intradural non-calcified disc sequester at the thoracolumbar junction as the cause of a ventral dural tear leading to a CSF leak with intractable SIH. This is of particular interest as the major cause of ventral dural leakage is thought to arise from calcified discogenic microspurs or dorsal osteophytes. Furthermore, we comprehensively describe a short and reasonable diagnostic and surgical approach of this rare pathology, which may particularly be of use in daily clinical routine in neurological wards and general surgical spine centers not facing such pathologies on a regular basis.

Lateral nasal wall extension of the nasoseptal flap for skull-base and medial orbital wall defects.

Sinonasal and skull-base tumors that previously required open resection can often be completely resected via an endonasal approach. The nasoseptal flap (NSF) is the workhorse vascularized tissue flap for the endoscopic reconstruction of large skull-base defects from tumor resections. The objective of the current article is to describe a novel modification of the NSF for simultaneous reconstruction of skull-base and medial orbital wall defects. An extension of the standard NSF to include mucosa of the lateral nasal wall was developed for closure of simultaneous skull-base and medial orbital wall defects. Outcomes including successful cerebrospinal fluid (CSF) leak closure, orbital edema, and postoperative cosmesis are reported. Eyelid edema was characterized according to the Surgeon Periorbital Rating of Edema and Ecchymosis (SPREE) scale. Three patients underwent reconstruction using the modified NSF (average age 75 years). The average defect size of the skull base was 3.6 ± 0.1 cm by 2.3 ± 0.2 cm. The average defect size of the medial orbit was 2.7 ± 0.1 cm by 2.6 ± 0.1 cm. All defects were successfully covered intraoperatively using the lateral nasal wall extension of the NSF. Two patients developed mild eyelid edema, whereas 1 individual had no noticeable swelling (SPREE classification 2, 2, and 1). All patients were successfully sealed at last clinical follow up (average 28 weeks). The modification of the NSF described here provides excellent coverage for reconstruction of large anterior skull-base defects and simultaneous medial orbital wall defects.

Progel Use is Not Associated with Decreased Incidence of Postoperative Air Leak after Nonanatomic Lung Surgery.

Progel Pleural Air Leak Sealant (CR Bard, Warwick, RI) is a US Food and Drug Administration-approved hydrogel designed for application to surgical staple lines to prevent air leak after lung surgery. This product has demonstrated efficacy in reducing intraoperative air leaks compared with standard air leak closure methods. However, the impact on chest tube duration and length of hospital stay has not been reported. To evaluate the effect on rates of postoperative air leak, chest tube duration, and hospital stay in surgical patients with and without use of Progel. Retrospective study of 176 patients aged 18 to 80 years who underwent video-assisted thoracoscopic wedge resections between 2014 and 2016. Eighty-four (48%) cases using Progel were included, as well as a representative sample of non-Progel cases (n = 92; 52%). Presence of postoperative lung air leak. No difference existed between the Progel and non-Progel groups in the rate of postoperative air leak (20/84, 23.81% Progel; 16/92, 17.39% non-Progel; p = 0.33). The length of time patients had a chest tube was similar (23.5 vs 23 hours, p = 0.721), as was percentage of patients with a less than 2-day hospitalization (77.17% non-Progel vs 82.14% Progel, p = 0.414). Our results suggest that Progel, used routinely in patients undergoing nonanatomic lung resection, does not have a significant impact on postoperative air leak, chest tube duration, or length of hospital stay. Further studies are warranted to evaluate the utility of Progel in reducing postoperative complications after thoracoscopic wedge resection in those treated for air leak or in the reduction of postoperative air leak in high-risk patients.

Minimally Invasive Transradial Percutaneous Closure of an Aortic Paravalvular Leak After Transcatheter Aortic Valve Replacement

Minimally Invasive Transradial Percutaneous Closure of an Aortic Paravalvular Leak After Transcatheter Aortic Valve Replacement

Endo-sponge Assisted Early Surgical Closure of Ileal Pouch-anal Anastomotic Leakage Preserves Long-term Function: A Cohort Study.

Background and AimsEndo-sponge [Braun Medical] assisted early surgical closure [ESC] is an effective treatment to control pelvic sepsis after ileal pouch-anal anastomosis [IPAA] leakage, and became standard treatment in our centre from 2010 onwards. The aim of this cohort study was to assess the long-term pouch function of ulcerative colitis [UC] patients treated with ESC or conventional management [CM] for anastomotic leakage after IPAA.MethodsConsecutive patients who underwent an IPAA for UC between 2002 and 2017 were included. Patients treated with ESC [2010–2017] or CM [2002–2009] for anastomotic leakage were compared with control patients without anastomotic leakage of the corresponding time period. Main endpoints were long-term pouch function on a 3-point scale and pouch failure, as measured with the validated pouch dysfunction score questionnaire.ResultsSome 280 of 334 patients [84%] returned the pouch dysfunction questionnaire, of whom 18 were treated with ESC and 22 with CM for anastomotic leakage. Control cohorts included 133 [2010–2017] and 107 patients [2002–2009]. Between ESC-treated patients and control patients, pouch function [p = 0.647] and pouch failure rates [0/18 versus 5/133, p >0.99] were similar. CM resulted in worse pouch function [p = 0.016] and a higher pouch failure rate [5/22 versus 5/107, p = 0.013] compared with control patients.ConclusionsESC, in contrast to CM, for IPAA leakage in UC patients is associated with preservation of pouch function and preclusion of pouch failure, probably due to early and effective treatment of pelvic sepsis.

189 – Technical Success is the Sole Predictor of Successful Endoscopic Closure of Post-Surgical Gastrointestinal Leaks: A Multicenter North American Cohort

189 – Technical Success is the Sole Predictor of Successful Endoscopic Closure of Post-Surgical Gastrointestinal Leaks: A Multicenter North American Cohort