Background: Bariatric surgery has been reported to reduce long-term mortality in operated participants in comparison with nonoperated participants.

Methods: We performed a systematic review and meta-analysis of clinical trials published as full articles dealing with cardiovascular (CV) mortality, all-cause mortality (noncardiovascular), and global mortality (sum of CV and all-cause mortality). Pooled-fixed effects of estimates of the risk of mortality in participants undergoing surgery were calculated compared with controls.

Results: Of 44,022 participants from 8 trials (14,052 undergoing surgery and 29,970 controls), death occurred in 3317 participants (400 in surgery, 2917 in controls); when the kind of death was specified, 321 CV deaths (118 in surgery, 203 in controls), and 523 all-cause deaths (218 in surgery, 305 in controls) occurred. Compared with controls, surgery was associated with a reduced risk of global mortality (OR = 0.55, CI, 0.49-0.63), of CV mortality (OR = 0.58, CI, 0.46-0.73), and of all-cause mortality (OR = 0.70, CI, 0.59-0.84).Data of all-cause mortality were not heterogeneous; heterogeneity of data of CV mortality decreased when studies were grouped according to size (large vs small studies). The reduction of risk was smaller in large than in small studies (OR = 0.61 vs 0.21, 0.63 vs 0.16, 0.74 vs 0.35 for global, CV, and all-cause mortality, respectively). The effect of gastric banding and gastric by-pass (3797 vs 10,255 interventions) was similar for global and all-cause mortality (OR = 0.57 vs 0.55, and 0.66 vs 0.70, respectively), different for CV mortality (OR = 0.71 vs 0.48). At meta-regression analysis, a trend for a decrease of global mortality (Log OR) linked to increasing BMI appeared.

Conclusion: This meta-analysis indicates that (1) bariatric surgery reduces long-term mortality; (2) risk reduction is smaller in large than in small studies; and (3) both gastric banding and gastric by-pass reduce mortality with a greater effect of the latter on CV mortality.