Postoperative Risk Evolves in the Operating Room: Incremental Value of Routine Intraoperative Variables After Metabolic and Bariatric Surgery Publisher



Rezvani M ; Monshizadeh A ; Kiani FAD ; Nourzad SHH
Source: Obesity Surgery Published:2026

Abstract

Introduction: Most bariatric risk prediction tools rely on preoperative variables and do not incorporate intraoperative information that may reflect evolving operative complexity. We evaluated whether routinely recorded intraoperative variables provide incremental predictive value for 30-day postoperative risk assessment. Methods: We performed a retrospective analysis of 645,602 patients undergoing laparoscopic sleeve gastrectomy or Roux-en-Y gastric bypass identified from the 2020–2023 Metabolic and Bariatric Surgery Accreditation and Quality Improvement Program (MBSAQIP) database. Multivariable logistic regression models were developed using preoperative variables alone and then augmented with routinely documented intraoperative variables, including operative time, MBSAQIP-coded approach conversion, drain placement, robotic assistance, anastomosis or leak testing, and venous thromboembolism prophylaxis method. Models were developed using 2020–2022 data and temporally validated in a 2023 cohort. Model performance was assessed using discrimination (area under the receiver operating characteristic curve [AUROC]), calibration, and decision curve analysis (DCA). Results: Adding intraoperative variables was associated with consistent but modest improvements in discrimination for common outcomes (ΔAUROC approximately + 0.008 to + 0.044), including any complication, serious complication, readmission, reintervention, reoperation, surgical-site infection, gastrointestinal bleeding, and transfusion. Findings were consistent in temporal validation. Operative time provided the most consistent incremental signal, with increasing risk per 10-minute increase. DCA demonstrated improved net benefit for selected outcomes, particularly serious complications and readmission, while gains were limited for rarer outcomes. Calibration metrics were similar between models, with no meaningful improvement after inclusion of intraoperative variables. Conclusion: Routinely collected intraoperative variables provide modest but consistent improvements in short-term risk prediction beyond preoperative factors alone and may incrementally support postoperative risk stratification for selected outcomes and perioperative planning. © The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2026.