Highlights
– In a large, multisite pragmatic randomized trial (n=1,565), adding metformin to lifestyle counselling (MET plus LIFE) produced a modest but statistically significant improvement in BMI Z‑score versus lifestyle counselling alone at 6 months (standardised effect size 0.26, 95% CI 0.15–0.37, p<0.0001) and at 24 months (effect size 0.11, 95% CI 0.00–0.22, p=0.047).
– The trial enrolled 8–19‑year‑olds with bipolar spectrum or related mood disorders who were overweight/obese and treated with second‑generation antipsychotics across 64 real‑world clinical sites in the USA, enhancing generalizability.
– Gastrointestinal adverse events were 2–4 times more common with metformin; there were no significant differences in suicidality by PHQ‑9 item 9. Trial investigators conclude benefits outweigh risks for most patients, and that clinicians should consider metformin in this population.
Background and disease burden
Second‑generation antipsychotics (SGAs) are commonly used in pediatric psychiatry for bipolar spectrum disorders and other mood and behavioral conditions. While effective for mood stabilization and symptom control, many SGAs cause clinically meaningful weight gain and adverse metabolic changes (insulin resistance, dyslipidaemia), particularly in youth. These changes increase lifetime cardiometabolic risk, reduce quality of life, and can worsen treatment adherence.
Although lifestyle modification is first‑line for antipsychotic‑associated weight gain, its effectiveness is often limited by illness, psychosocial context, and medication effects. Metformin, an insulin‑sensitizing agent with an established safety record in pediatric type 2 diabetes and mounting evidence in antipsychotic‑related weight gain, has been proposed as an adjunctive therapy. Prior randomized trials in smaller, often single‑site populations suggested benefit over short durations (typically ≤6 months). However, long‑term effectiveness in routine clinical practice and pragmatic evidence for broad implementation in young people with bipolar spectrum disorders has been limited—an evidence gap this study addresses.
Study design and methods
This multisite, open‑label, pragmatic parallel‑group randomized trial (MOBILITY Consortium) enrolled overweight or obese youth aged 8–19 years with a current or prior diagnosis of a bipolar spectrum disorder who were either starting or continuing an SGA. Recruitment and follow‑up occurred at 64 community and academic sites across the United States.
Participants (n=1,565) were randomized 1:1 to either a behavioural healthy eating and physical activity program (LIFE) or to metformin plus the same lifestyle program (MET plus LIFE). Randomization used eight strata defined by baseline BMI percentile (overweight vs obese), antipsychotic exposure at baseline (naive/starting vs continuing), and sex assigned at birth, with block randomisation.
The co‑primary endpoints were the change in age‑ and sex‑normalised BMI Z‑score at 6 months and at 24 months in the intention‑to‑treat population. The trial was pragmatic and open‑label by design to mimic real‑world practice; people with lived experience of bipolar disorders contributed to trial design and conduct. The trial registration is ClinicalTrials.gov NCT02515773.
Key findings and detailed results
Enrollment and follow‑up: Between November 2015 and February 2022, 1,633 individuals consented; after exclusions, 1,565 were randomized (777 to MET plus LIFE and 788 to LIFE). Data completeness was high for a pragmatic trial: 1,252 participants had data at month 6 and 1,299 at month 24. The cohort’s mean age was 13.9 years (SD 2.9); 53% were male; 65% were White and 19% Black.
Primary outcomes
At 6 months, adding metformin produced a greater reduction in BMI Z‑score than lifestyle counselling alone with a standardised effect size of 0.26 (95% CI 0.15–0.37), p<0.0001. This effect persisted but attenuated at 24 months (effect size 0.11, 95% CI 0.00–0.22, p=0.047). The authors describe the absolute differences as modest when translated into BMI Z‑score units; standardized effect sizes allow cross‑trial comparison but should be interpreted in clinical context (see Expert Commentary below).
Secondary and subgroup findings
The report indicates the trial collected clinically relevant secondary measures (metabolic indices, subgroup analyses by baseline obesity status and antipsychotic exposure), but the summary provided focuses on BMI Z‑score changes as co‑primary outcomes. The pragmatic nature allowed heterogeneous antipsychotic types and dosing, reflecting usual care rather than a tightly controlled pharmacologic trial.
Safety
Gastrointestinal adverse events (nausea, diarrhea, abdominal discomfort) were 2–4 times more common in the MET plus LIFE group and were the principal tolerability issue. No unexpected safety signal related to serious medical adverse events was reported in the provided summary. Regarding suicidality, among participants taking metformin there were 12 single suicide attempts and 1 participant with two attempts; among participants not taking metformin there were 25 single attempts and 3 participants with two attempts. When suicidality was assessed using PHQ‑9 item 9 during randomized treatment, proportions did not differ significantly between groups (MET plus LIFE: 42/519 [8%]; LIFE: 57/655 [9%]). The authors conclude that metformin’s benefits outweigh its risks for most patients in this population.
Expert commentary: interpretation, strengths, and limitations
Strengths: This trial is notable for its size, duration (24 months), and pragmatic multisite design across community and academic clinics, offering strong external validity. Involving people with lived experience strengthens relevance and feasibility. The randomized design and intention‑to‑treat analysis provide credible evidence that adjunctive metformin can mitigate SGA‑related weight gain in youth with bipolar spectrum disorders.
Clinical magnitude and interpretation: The improvements in BMI Z‑score are statistically significant but modest in magnitude. Translating standardized effect sizes to clinical outcomes depends on baseline BMI distribution and age; even small Z‑score reductions can be meaningful at the population level and may slow progression to more severe obesity and cardiometabolic dysfunction. Clinicians and families will need to weigh modest mean treatment effects against patient priorities, comorbidities, and tolerability.
Limitations and cautions: The open‑label design could influence behavioural components and reporting of subjective adverse events. The trial allowed heterogeneous SGA regimens, which increases generalizability but complicates assessment of whether metformin’s effect varies by specific antipsychotic agent, dose, or treatment duration. Adherence to metformin and to lifestyle counselling likely varied; detailed adherence data and per‑protocol analyses are important for interpretation but are not provided in the summary. The provided suicidality data do not suggest an excess risk with metformin, but psychiatric risk in this population requires ongoing vigilance. Finally, the trial focused on BMI Z‑score rather than hard cardiometabolic endpoints; whether metformin reduces long‑term cardiometabolic events remains untested.
Mechanistic plausibility
Metformin primarily improves peripheral insulin sensitivity and reduces hepatic gluconeogenesis; it may also modulate appetite and the gut microbiome. SGAs can promote weight gain via antagonism of histaminergic and serotonergic receptors, insulin resistance, and increased caloric intake. By improving insulin sensitivity and attenuating weight gain, metformin is biologically plausible as an adjunct to behavioural interventions in SGA‑treated youth.
Clinical implications and practical guidance
Given the trial’s findings, clinicians who manage children and adolescents with bipolar spectrum disorders treated with SGAs should consider metformin as a therapeutic option to mitigate weight gain when lifestyle interventions alone are insufficient. Key practical points for implementation include:
- Shared decision‑making: Discuss expected benefits (modest BMI Z‑score reduction) and common side effects (gastrointestinal symptoms), setting realistic expectations.
- Baseline assessment: document weight/BMI percentile, fasting glucose and insulin or HbA1c as indicated, lipids, liver and renal function where clinically appropriate, and screen for contraindications (e.g., significant renal impairment).
- Dosing and tolerability: Start low and titrate to improve tolerability. Monitor for gastrointestinal side effects and adjust titration pace or formulation (e.g., extended‑release preparations) when available and appropriate.
- Monitoring: periodic assessment of growth, pubertal development, metabolic parameters, and adherence. Maintain close psychiatric follow‑up given baseline suicidality risk in this population.
- Comprehensive care: Combine pharmacologic therapy with continued behavioural interventions, family support, and coordination with primary care for cardiometabolic risk management.
Research gaps and future directions
This trial advances the evidence base but leaves open questions: Which subgroups derive the largest benefit (e.g., those earlier in antipsychotic exposure, specific antipsychotic agents)? What is the optimal timing to start metformin relative to antipsychotic initiation? Can metformin’s modest effects be augmented by combined behavioural, pharmacologic, or digital interventions? Importantly, longer‑term studies linking early weight trajectories to adult cardiometabolic outcomes are needed to quantify the public health impact of metformin use in this context. Safety monitoring beyond two years and evaluation of metabolic biomarkers, quality of life, and functional outcomes should be priorities.
Conclusion and practical takeaways
This large, pragmatic randomized trial provides the strongest real‑world evidence to date that adjunctive metformin modestly reduces antipsychotic‑associated weight gain in overweight and obese children and adolescents with bipolar spectrum and related mood disorders. The effect is statistically robust at 6 months and persists, though attenuated, at 24 months. Gastrointestinal side effects are more frequent with metformin, but serious safety signals did not emerge in the reported data. For many patients who experience problematic SGA‑induced weight gain, metformin should be considered as part of a comprehensive, patient‑centred strategy that includes lifestyle modification and careful metabolic monitoring.
Funding and trial registration
Funding: Patient‑Centered Outcomes Research Institute. Trial registration: ClinicalTrials.gov NCT02515773.
References
1) 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.