Meta-analyses included 621 studies (n=135,246), among them were 29 RCTs, 49 non-randomised controlled trials, 60 comparative retrospective studies, 187 uncontrolled prospective case series, 266 single-arm retrospective studies, 25 observational studies and two case-control studies. Most included studies were classified as III or IV level of evidence. Most RCTs had a Jadad score of 1 to 3.
Bariatric surgery was significantly associated with an overall absolute weight loss of 38.5kg (95% CI 36.6 to 40.4; 300 treatment groups, n=23,380) and with excess body weight loss of 55.9% (95% CI 54.1 to 57.8; 319 treatment groups, n=34,329). Significant heterogeneity was observed in both outcomes (I2=99%).
Bariatric surgery led to an overall 78.1% of patients with resolution of clinical manifestations of diabetes. Diabetes was improved or resolved in 86.6% of patients. It was not possible to access the on-line report to gauge the 95% CIs for these values. Results of statistical heterogeneity assessment on these outcomes were not presented.
Sensitivity analyses did not materially affect the results. Subgroup analyses showed that weight loss and diabetes resolution were largest in patients who underwent biliopancreatic diversion/duodenal switch, next-largest for gastric bypass and least for banding procedures. Subgroup analyses that included diabetic patients only were reported.