Positive and Negative Independent Predictive Factors of Weight Loss After Bariatric Surgery in a Veteran Population

Introduction Factors predicting outcomes after bariatric surgery are yet to be elucidated. We aim to characterize patient-level factors that predict midterm weight loss. Methods A database of bariatric surgery at a Veterans Affairs (VA) hospital was retrospectively reviewed. Patient characteristics including age, race, sex, median zip code household income, and distance to the VA bariatric center were analyzed for relationships with percent excess body mass index loss (%EBMIL). Univariate and multivariate analyses were conducted to identify factors independently associated with weight loss after accounting for follow-up time, using stepwise variable selection. A multivariable mixed effects linear regression model was constructed with random intercepts for repeated measures by veteran and fixed effects for time, patient, and procedural characteristics, including comorbidities. Results A total of 1124 observations were analyzed for 340 bariatric patients. Most were male (77%), white (73%); mean age was 53.2 years and mean preoperative BMI was 43.9 kg/m(2). Follow-up ranged from 99% at 1 year, 54% at 5 years, and 24% at 10 years, with a mean of 6.9 years for Roux-en-Y gastric bypass (RYGB) and 3.5 years for laparoscopic sleeve gastrectomy (LSG). RYGB (p < 0.001) and female (p = 0.016) predicted greater %EBMIL up to 10 years after surgery. African American race and higher comorbidity burden predicted poorer %EBMIL (p = 0.008, p = 0.012, respectively). Analysis of individual comorbidities demonstrated that type 2 diabetes was most strongly associated with poorer %EBMIL (p = 0.048). Conclusion RYGB and female sex are independent predictors of greater midterm weight loss after bariatric surgery. African American race and a high burden of comorbidity are predictive of poorer weight loss. Neither zip code median income nor distance from bariatric center was associated with weight loss.