American Surgical Association

Back to 2011 Program


Predicting Risk for Serious Complications With Bariatric Surgery: Results from the Michigan Bariatric Surgery Collaborative
Jonathan F Finks, MD*1, Kerry Kole, MD*2, Yenumula Panduranga, MD*3, Wayne J English, MD*4, Kevin Krause, MD*5, Arthur Carlin, MD*6, Jeffrey Genaw, MD*6, John D. Birkmeyer, MD1, Nancy J Birkmeyer, MD*1
1University of Michigan, Ann Arbor, MI;2Harper University Hospital, Detroit, MI;3Michigan State University, Lansing, MI;4Marquette General Hospital, Marquette, MI;5William Beaumont Hospital, Royal Oak, MI;6Henry Ford Hospital, Detroit, MI

OBJECTIVE(S): To develop a risk prediction model for serious complications following bariatric surgery.
METHODS: The Michigan Bariatric Surgery Collaborative is a statewide consortium of hospitals and surgeons, which maintains an externally-audited prospective clinical registry. We analyzed data from 23,411 patients undergoing bariatric surgery between June, 2006 and August, 2010. Using a random split-sample design, we performed multivariate stepwise logistic regression with half of the sample to develop a predictive model of risk factors associated with serious complications (life threatening and/or associated with lasting disability) within 30 days of surgery. The predictive model was applied to the second half of the sample for the purpose of internal validation.
RESULTS: Overall, 588 patients(2.5%) experienced a serious complication. Significant risk factors(p<0.10) included: heart disease(OR 1.63); age category(per 10 years)[OR 1.12]; male gender(OR 1.29); VTE history(OR 2.37); musculoskeletal disorder(OR 1.63); smoking history(OR 1.31); pulmonary disease(OR 1.52) and procedure type (reference lap band): laparoscopic gastric bypass(OR 3.26); open gastric bypass(OR 2.62); sleeve gastrectomy(OR 2.81); duodenal switch(OR 11.25). When examined across terciles of predicted risk, the mean complication rates were similar in the study and validation populations: 0.8%, 2.4%, 4.4% and 1.0%, 2.3%, 4.2%, respectively. The c-statistic was 0.70 and the model was well-calibrated across deciles of predicted risk.
CONCLUSIONS: We have developed and validated a population-based risk scoring system for serious complications following bariatric surgery. We expect that this scoring system will improve the process of informed consent, facilitate the selection of procedures for high-risk patients, and enable evidence-based insurance approval guidelines.


Back to 2011 Program