Background: Laparoscopic sleeve gastrectomy (LSG) has become popular due to its technical ease and excellent short-term results. Understanding the risk profile of LSG compared with the gold standard laparoscopic Roux-en-Y gastric bypass (LRYGB) is critical for patient selection.
Objectives: To use traditional regression techniques and random forest classification algorithms to compare LSG with LRYGB using the 2015 Metabolic and Bariatric Surgery Accreditation and Quality Improvement Data Registry.
Setting: United States.
Methods: Outcomes were leak, morbidity, and mortality within 30 days. Variable importance was assessed using random forest algorithms. Multivariate models were created in a training set and evaluated on the testing set with receiver operating characteristic curves. The adjusted odds of each outcome were compared.
Results: Of 134,142 patients, 93,062 (69%) underwent LSG and 41,080 (31%) underwent LRYGB. One hundred seventy-eight deaths occurred in 96 (.1%) of LSG patients compared with 82 (.2%) of LRYGB patients (P<.001). Morbidity occurred in 8% (5.8% in LSG versus 11.7% in LRYGB, P<.001). Leaks occurred in 1% (.8% in LSG versus 1.6% in LRYGB, P<.001). The most important predictors of all outcomes were body mass index, albumin, and age. In the adjusted multivariate models, LRYGB had higher odds of all complications (leak: odds ratio 2.10, P<.001; morbidity: odds ratio 2.02, P<.001; death: odds ratio 1.64, P<.01).
Conclusion: In the Metabolic and Bariatric Surgery Accreditation and Quality Improvements data registry for 2015, LSG had half the risk-adjusted odds of death, serious morbidity, and leak in the first 30 days compared with LRYGB.
Keywords: Laparoscopic Roux-en-Y gastric bypass; Laparoscopic sleeve gastrectomy; MBSAQIP; Perioperative complications.
Copyright © 2018 American Society for Bariatric Surgery. Published by Elsevier Inc. All rights reserved.