Importance: Childhood obesity is an important public health problem with increasing prevalence. Because treatment often has limited success, new approaches must be identified.
Objective: To evaluate the effectiveness and safety of metformin for treating obesity in children aged 18 years and younger without a diagnosis of diabetes mellitus.
Evidence review: We included randomized clinical trials identified through searches of MEDLINE, the Cochrane Library, and ClinicalTrials.gov. Our primary outcome measure was change in body mass index (BMI, calculated as weight in kilograms divided by height in meters squared). We assessed study quality, pooled data using a random-effects model, and performed subgroup and sensitivity analyses.
Findings: Fourteen randomized clinical trials were eligible. For BMI, moderate-strength evidence indicated a reduction of -1.38 (95% CI, -1.93 to -0.82) from baseline compared with control at 6 months. A similar, if less dramatic, effect was observed in studies less than 6 months, but the pooled estimate from studies of 1 year of treatment was not statistically significant. Subgroup analyses indicated smaller, but significant, effects for those with baseline BMI below 35, those of Hispanic ethnicity, those with acanthosis nigricans, those who had tried and failed diet and exercise programs, and in studies with more girls or higher mean age (adolescents). Moderate-strength evidence indicated that with metformin, 26% reported a gastrointestinal event compared with 13% in control groups (relative risk, 2.05; 95% CI, 1.19-3.54), although there was no difference in discontinuations due to adverse events. No serious adverse events were reported.
Conclusions and relevance: Metformin provides a statistically significant, but very modest reduction in BMI when combined with lifestyle interventions over the short term. A large trial is needed to determine the benefits to subgroups or impacts of confounders. In the context of other options for treating childhood obesity, metformin has not been shown to be clinically superior.