Predictors and outcomes of leak after Roux-en-Y gastric bypass: an analysis of the MBSAQIP data registry

Background: Gastrointestinal leak is one of the most severe postoperative complications after Roux-en-Y gastric bypass (RYGB), occurring in up to 2% of all patients. This has led to adoption of simpler procedures, such as sleeve gastrectomy, which have improved safety profiles but potentially less effective long-term metabolic outcomes. Yet, in contrast to sleeve gastrectomy, a paucity of modern literature exists regarding predictors of leak for RYGB. Objectives: The purpose of this study was to examine gastrointestinal leak in patients undergoing RYGB using the Metabolic and Bariatric Surgery Accreditation and Quality Improvement data registry. Methods: We identified all Metabolic and Bariatric Surgery Accreditation and Quality Improvement patients who underwent RYGB in 2015 and 2016. Primary outcomes of interest include identifying the prevalence, impact, and predictors of leak in RYGB patients. Our secondary outcomes of interest include characterizing overall complication rates in RYGB patients. Univariate analysis of pre-, intra-, and postoperative variables was performed using X-2 tests for categoric data and independent sample t test for continuous data. A nonparsimonious multivariable logistic regression model was then developed to determine predictive factors for development of leak. Setting: All centers belonging to the Metabolic and Bariatric Surgery Accreditation and Quality Improvement data registry. Results: A total of 77,596 patients underwent RYGB from 2015 to 2016. The majority of patients were female (79.8%), white (75.9%), and underwent laparoscopic RYGB (89.7%). The mean age of patients was 45.2 years (standard deviation 11.9) with a mean body mass index of 46.3 kg/m(2) (standard deviation 8.17). Complication rates for RYGB were low with a mortality of .16% and a total complication rate of 7.5%. A total of 476 leaks were identified with an overall leak rate of .6% and a mortality of 1.5%. Leak was associated with a statistically significant increase in all complications as well as readmission, reoperation, and mortality rates at 30 days. Multivariable logistic regression analysis revealed the following statistically significant independent predictors of leak: body mass index, age, operative length, American Society of Anesthesiologists score >3, prior pulmonary embolus, and partially dependent functional status. Albumin was the only independent protective variable after adjusting for confounders and interactions. Conclusion: Using the robust Metabolic and Bariatric Surgery Accreditation and Quality Improvement database, we found RYGB to be a safe procedure with low morbidity and mortality. The overall leak rate was .6% with leak significantly increasing all other complications, readmission, reoperation, and mortality rates at 30 days. Logistic regression identified prior pulmonary embolus and partially dependent functional status as the 2 largest predictors of leak while increased albumin was the only protective factor. Optimizing preoperative nutrition and strength in these patients through structured multidisciplinary programs may therefore have a role in the ongoing improvement of outcomes after RYGB. (C) 2019 American Society for Bariatric Surgery. Published by Elsevier Inc. All rights reserved.