Background:  Changes in weight and metabolic parameters have been commonly reported in patients with schizophrenia. Metformin has been evaluated in clinical studies to prevent or reduce weight gain and changes in metabolic parameters in non-diabetic subjects. We undertook a systematic review of the efficacy and safety of metformin in reducing weight gain and metabolic abnormalities in non-diabetic subjects with schizophrenia or bipolar disorder taking antipsychotic medication to establish if these data could potentially drive guideline development.

Methods:  Medical databases were searched using terms including ‘antipsychotic’, ‘atypical antipsychotic agent’, ‘antipsychotic agents’, ‘antipsychotic-drug’ and ‘metformin’ and ‘weight’. Studies reporting weight and/or metabolic outcomes in non-diabetic subjects with schizophrenia and bipolar disorder were included regardless of methodological type and subject age.

Results:  Nine randomised double-blind studies and two open cohort studies evaluating metformin and changes in weight in trials up to 16 weeks were identified. In all, 495 participants received antipsychotics (mostly olanzapine), and three studies were in subjects aged < 18 years. The adult studies predominantly utilised non-Caucasian subjects with chronic schizophrenia. Weight and lifestyle intervention programmes were provided to all cohorts in eight studies, which confounded interpretation of the data. In ten studies, the addition of metformin to antipsychotic treatment was associated with either significantly attenuated weight gain or weight loss compared with control groups. Nine studies measured various glucose parameters. In four studies, subjects prescribed metformin had significantly improved glucose parameters relative to controls. The two studies of metformin in patients with first-episode schizophrenia demonstrated the largest improvement in weight and glucose parameters.

Conclusions:  Metformin may have some value in reducing or preventing weight gain and changes in metabolic parameters during treatment with antipsychotic medication particularly in first-episode psychosis; however, it has been predominantly studied short-term and in non-Caucasian populations. A number of new trials are due to report data 2009–2013 to aid definitive interpretation of the role of metformin. Further longer-term studies are warranted before definitive guidelines can be established.