Risk stratification of serious adverse events after gastric bypass in the Bariatric Outcomes Longitudinal Database.
BACKGROUND: There is now sufficient demand for bariatric surgery to compare bariatric surgeons and bariatric centers according to their postsurgical outcomes, but few validated risk stratification measures are available to enable valid comparisons. The purpose of this study was to develop and validate a risk stratification model of composite adverse events related to Roux-en-Y gastric bypass (RYGB) surgery. METHODS: The study population included 36,254 patients from the Bariatric Outcomes Longitudinal Database (BOLD) registry who were 18-70 years old and had RYGB between June 11, 2007, and December 2, 2009. This population was randomly divided into a 50% testing sample and a 50% validation sample. The testing sample was used to identify significant predictors of 90-day composite adverse events and estimate odds ratios, while the validation sample was used to assess model calibration. After validating the fit of the risk stratification model, the testing and validation samples were combined to estimate the final odds ratios. RESULTS: The 90-day composite adverse event rate was 1.48%. The risk stratification model of 90-day composite adverse events included age (40-64, ≥ 65), indicators for male gender, body mass index (50-59.9, ≥ 60), obesity hypoventilation syndrome, back pain, diabetes, pulmonary hypertension, ischemic heart disease, functional status, and American Society of Anesthesiology classes 4 or 5. Our final gastric bypass model was predictive (c-statistic = .68) of serious adverse events 90 days after surgery. CONCLUSIONS: With additional validation, this risk model can inform both the patient and surgeon about the risks of bariatric surgery and its different procedures, as well as enable valid outcomes comparisons between surgeons and surgical programs.
Maciejewski, ML; Winegar, DA; Farley, JF; Wolfe, BM; DeMaria, EJ
Volume / Issue
Start / End Page
Electronic International Standard Serial Number (EISSN)
Digital Object Identifier (DOI)