Survival among high-risk patients after bariatric surgery.

CONTEXT Existing evidence of the survival associated with bariatric surgery is based on cohort studies of predominantly younger women with a low inherent obesity-related mortality risk. The association of survival and bariatric surgery for older men is less clear. OBJECTIVE To determine whether bariatric surgery is associated with reduced mortality in a multisite cohort of predominantly older male patients who have a high baseline mortality rate. DESIGN, SETTING, AND PARTICIPANTS Retrospective cohort study of bariatric surgery programs in Veterans Affairs medical centers. Mortality was examined for 850 veterans who had bariatric surgery in January 2000 to December 2006 (mean age 49.5 years; SD 8.3; mean body mass index [BMI] 47.4; SD 7.8) and 41,244 nonsurgical controls (mean age 54.7 years, SD 10.2; mean BMI 42.0, SD 5.0) from the same 12 Veteran Integrated Service Networks; the mean follow-up was 6.7 years. Four Cox proportional hazards models were assessed: unadjusted and controlled for baseline covariates on unmatched and propensity-matched cohorts. MAIN OUTCOME MEASURE All-cause mortality through December 2008. RESULTS Among patients who had bariatric surgery, the 1-, 2-, and 6-year crude mortality rates were, respectively, 1.5%, 2.2%, and 6.8% compared with 2.2%, 4.6%, and 15.2% for nonsurgical controls. In unadjusted Cox regression, bariatric surgery was associated with reduced mortality (hazard ratio [HR], 0.64; 95% confidence interval [CI], 0.51-0.80). After covariate adjustment, bariatric surgery remained associated with reduced mortality (HR, 0.80; 95% CI, 0.63-0.995). In analysis of 1694 propensity-matched patients, bariatric surgery was no longer significantly associated with reduced mortality in unadjusted (HR, 0.83; 95% CI, 0.61-1.14) and time-adjusted (HR, 0.94; 95% CI, 0.64-1.39) Cox regressions. CONCLUSION In propensity score-adjusted analyses of older severely obese patients with high baseline mortality in Veterans Affairs medical centers, the use of bariatric surgery compared with usual care was not associated with decreased mortality during a mean 6.7 years of follow-up.