Highlight
– A multicentre pragmatic randomized trial (MOBILITY Consortium) tested metformin plus lifestyle intervention versus lifestyle alone in 1,565 overweight/obese youth (aged 8–19) with bipolar spectrum disorders treated with second‑generation antipsychotics.
– Metformin produced a modest but statistically significant improvement in age- and sex‑normalized BMI Z-score at 6 months (standardized effect size 0.26; 95% CI 0.15–0.37) and at 24 months (effect size 0.11; 95% CI 0.00–0.22).
– Gastrointestinal adverse effects were 2–4 times more common with metformin; suicidality rates did not differ significantly between groups. The investigators conclude the benefits outweigh the risks for most patients.
Background: clinical context and unmet need
Second‑generation (atypical) antipsychotics (SGAs) are widely used in children and adolescents for acute and maintenance treatment of bipolar spectrum and related mood disorders. SGAs can be highly effective for mood stabilization but are commonly associated with substantial weight gain and adverse metabolic effects (insulin resistance, dyslipidaemia), which raise long‑term cardiovascular risk and contribute to morbidity and premature mortality.
In pediatric populations, SGA‑associated weight gain is particularly concerning because of the potential for early onset of metabolic disease, treatment nonadherence driven by body image concerns, and the prolonged duration of antipsychotic exposure over a lifetime. Despite evidence from smaller trials and meta‑analyses that metformin can attenuate antipsychotic‑related weight gain, routine use of metformin for this indication has not been widely adopted, partly because of limited data from pragmatic, real‑world settings and sparse long‑term follow‑up.
Study design and methods
DelBello and colleagues (the MOBILITY Consortium) conducted a multi‑site, open‑label, pragmatic parallel‑group randomized trial in the USA to evaluate the effectiveness of metformin added to a behavioral lifestyle program (LIFE) versus LIFE alone in youth aged 8–19 years with a diagnosis of a bipolar spectrum disorder who were overweight (BMI percentile 85th to <95th) or obese (≥95th) and treated with, or initiating, an SGA.
Key design features:
- Setting: 64 community and academic clinical sites with broad inclusion criteria to increase generalizability.
- Randomisation: 1:1 allocation to MET plus LIFE versus LIFE alone, stratified by baseline BMI category (overweight vs obese), SGA‑naïve versus continuing SGA, and sex assigned at birth; block randomisation with blocks of six.
- Interventions: both groups received a standardized healthy eating and physical activity program (LIFE); the intervention group additionally received open‑label metformin.
- Co‑primary outcomes: change in age‑ and sex‑normalized BMI Z‑score at 6 months and at 24 months, analysed in the intention‑to‑treat population.
- Patient and public involvement: people with lived experience of bipolar disorder participated in study design, conduct, and reporting.
- Registration and funding: Patient‑Centered Outcomes Research Institute funded; ClinicalTrials.gov NCT02515773.
Key findings
Enrollment and follow‑up
The trial randomized 1,565 participants between November 2015 and February 2022 (777 to MET plus LIFE; 788 to LIFE). At month 6 data were available for 1,252 participants (565 MET plus LIFE; 687 LIFE) and at month 24 for 1,299 participants (579 MET plus LIFE; 720 LIFE). The cohort had a mean age of 13.9 years (SD 2.9), 53% male, and was racially diverse (65% White, 19% Black).
Primary outcomes
Assignment to metformin plus LIFE produced greater reductions in BMI Z‑score compared with LIFE alone at both primary timepoints:
- 6 months: standardized effect size 0.26 (95% CI 0.15–0.37), p < 0.0001.
- 24 months: standardized effect size 0.11 (95% CI 0.00–0.22), p = 0.047.
These results indicate a moderate standardized benefit at 6 months with attenuation by 24 months, though still statistically significant. The trial reports effect sizes rather than absolute mean differences in BMI Z‑score in the summary; clinicians should interpret standardized effects in the clinical context of baseline BMI distribution and expected trajectory on SGAs.
Safety and adverse events
Gastrointestinal adverse events (nausea, diarrhea, abdominal discomfort) were 2–4 times more frequent in the metformin group — a known and expected profile for metformin. Importantly, there were no statistically significant differences in suicidality during randomized treatment between groups as measured by item 9 of the PHQ‑9 (MET plus LIFE: 42/519 [8%]; LIFE: 57/655 [9%]). Reported suicide attempt counts were numerically lower in the metformin group (12 participants attempted suicide once and one attempted twice) than in the control group (25 participants attempted suicide once and three attempted twice), but these counts should be interpreted cautiously given low overall incidence and the trial’s open‑label design.
Interpretation and clinical significance
The MOBILITY trial provides the largest and longest randomized evidence to date that adjunctive metformin reduces antipsychotic‑associated weight gain in youth with bipolar spectrum disorders treated with SGAs. Key points for clinicians:
- Magnitude of effect: the benefit is modest and attenuates over time but is reproducible across a large, pragmatic sample. Even modest BMI reductions may be meaningful in young people, potentially reducing progression to cardiometabolic disease and supporting treatment adherence.
- Safety profile: the adverse events were consistent with known effects of metformin — mainly gastrointestinal and generally manageable. No increase in suicidality was found.
- Practical interpretation: metformin should be considered an evidence‑based option to prevent or reduce SGA‑related weight gain in overweight/obese youth with bipolar spectrum disorders, used alongside lifestyle interventions and routine metabolic monitoring.
Biological plausibility and mechanism
Metformin’s metabolic actions provide biological plausibility for its effect on antipsychotic‑associated weight gain. Mechanisms include reduction of hepatic gluconeogenesis, improvement in insulin sensitivity, modest appetite suppression, altered gut hormone signalling, and potential effects on the gut microbiome. These actions can counteract insulin resistance and adiposity driven in part by SGAs.
Strengths and limitations
Strengths
- Large sample size and multi‑site, real‑world (pragmatic) design increase external validity and generalizability across community and academic settings.
- Long follow‑up (24 months) allows assessment of durability of effect.
- Patient involvement strengthened relevance to lived experience and pragmatic implementation.
Limitations
- Open‑label design could introduce expectation bias; however, BMI Z‑score is an objective outcome less vulnerable to subjective bias.
- The standardized effect sizes reported require careful translation into clinically meaningful absolute changes in BMI or weight; the overall clinical impact varies by baseline age and degree of adiposity.
- Heterogeneity in antipsychotic agents, doses, concomitant medications, and adherence could influence outcomes and limit causal inferences about specific SGAs.
- The trial included a behavioural lifestyle program for all participants, so the observed effect is incremental to lifestyle counselling; effects may differ where lifestyle support is unavailable.
Practical recommendations for clinicians
Based on the MOBILITY trial and prior evidence, clinicians caring for overweight/obese children and adolescents with bipolar spectrum disorders treated with SGAs should consider the following approach:
- Prevention and monitoring: discuss metabolic risks at treatment initiation, measure baseline weight, height, BMI percentile, blood pressure, fasting glucose and lipids, and repeat monitoring at regular intervals per local guidance.
- Lifestyle first: offer evidence‑based lifestyle interventions (nutrition, physical activity, sleep) for all patients; behavioural programs are a key component and were included in the trial for both arms.
- Consider metformin when: a young person is overweight/obese or has significant, sustained weight gain on an SGA and lifestyle measures are insufficient. Use shared decision‑making, discussing modest expected benefits, common gastrointestinal side effects, and rare risks (e.g., lactic acidosis in severe renal impairment).
- Monitoring during metformin: check renal function before initiation and periodically thereafter, as metformin is contraindicated in significant renal impairment. Consider baseline vitamin B12 level and periodic checks in long‑term use. Monitor weight/BMI and metabolic labs as above. Provide anticipatory guidance on GI side effects and strategies (dose titration, taking with food).
- Dosing and initiation: follow local pediatric dosing guidance and pharmaceutical labeling. Initiate at a low dose and titrate as tolerated; many pediatric regimens start at 500 mg once daily and increase gradually, balancing efficacy and tolerability.
Research and practice gaps
Remaining questions include optimal timing (pre‑emptive at SGA initiation versus reactive after weight gain), ideal dosing strategy in children and adolescents, differential responses by specific SGAs, and identification of subgroups most likely to benefit. Longer‑term outcomes beyond 24 months on hard cardiometabolic endpoints remain unknown. Implementation research to integrate metformin use with behavioural programs in routine community care settings would help translate trial findings.
Conclusion
The MOBILITY trial provides robust, pragmatic randomized evidence that adjunctive metformin yields a modest but statistically significant reduction in BMI Z‑score at 6 and 24 months in overweight and obese children and adolescents with bipolar spectrum disorders treated with SGAs. Given the scale of metabolic risk in this population and the manageable safety profile observed, clinicians should consider metformin as part of a comprehensive prevention and management strategy for SGA‑related weight gain, using shared decision‑making and appropriate monitoring.
Funding and registration
Funding: Patient‑Centered Outcomes Research Institute. Trial registration: ClinicalTrials.gov NCT02515773.
Key reference
DelBello MP, Welge JA, Klein CC, Blom TJ, Fornari V, Higdon C, Sorter MT, Kurtz B, Starr C, Smith A, Huang B, Chen C, Modi AC, Crimmins N, Correll CU; MOBILITY Consortium. Metformin for overweight and obese children and adolescents with bipolar spectrum and related mood disorders treated with second‑generation antipsychotics: a randomised, pragmatic trial. Lancet Psychiatry. 2025 Dec;12(12):893-905. doi: 10.1016/S2215-0366(25)00273-1. PMID: 41233082.
Author disclosures and acknowledgements
This article summarizes and interprets published trial data. Clinicians should consult the full trial report and current clinical guidelines when making treatment decisions. The author has no conflicts to declare related to this summary.
