Utilizing Metformin in Treatment-Resistant Bipolar Depression

Adjunctive metformin in bipolar disorder. Researchers performed a randomized clinical trial of treating insulin resistance to improve clinical outcomes in treatment-resistant bipolar depression.

Research Update

Treatment-resistant bipolar depression (TRBD) is defined by non-remission after 2 8-week trials of recommended medications, including combination therapy, at therapeutic doses, and associated with increased psychiatric hospitalizations and disability, and poorer quality of life.1 The prevalence of type 2 diabetes and insulin resistance in patients with bipolar disorder is approximately 22% and 32%, respectively, and these conditions may worsen illness course.2

Case Vignette

“Mr Joseph” is a 60-year-old Caucasian male with a history of chronic bipolar I disorder, whose most recent episode included depression without psychotic features. He failed a previous trial of lithium due to adverse effects on renal function. He had improved mood with treatment with ziprasidone, titrated to 40 mg qAM and 80 mg qPM, although he has residual depressive symptoms. He gained weight with ziprasidone, with an increase in BMI from 28 to 32. He maintained a normal fasting glucose. He requested to continue ziprasidone, and 3 months ago was started on metformin, which was titrated to 1000 mg BID. At his most recent outpatient clinic visit, his BMI had decreased to 29.5, and he also noted mild improvements in mood and anxiety.

The Current Study

Calkin and colleagues hypothesized that patients with TRBD who met criteria for insulin resistance would exhibit a favorable response to metformin, which has insulin-sensitizing properties.3 They performed a randomized quadruple-masked (patient, investigator, outcomes assessor, statistician) trial of adjunctive metformin at 2 sites in Canada and the United States. They included adults aged 18 and older with DSM-5 bipolar disorder and unremitting depressive symptoms (Montgomery-Asberg Depression Rating Score [MADRS] score ≥15) for at least 4 weeks despite optimal treatment.

Subjects were screened for insulin resistance using the Homeostatic Model Assessment-Insulin Resistance (HOMA-IR) equation, which is based on fasting blood glucose and insulin. Subjects were excluded if they did not have insulin resistance (HOMA-IR <1.8), had manic symptoms (Young Mania Rating Scale score ≥15), had a suicidal ideation rating of 5 on the Columbia Suicide Severity Rating Scale, were already receiving or were allergic to metformin, or had liver function tests ≥3 times the upper limit of normal.

Subjects received metformin 500 mg (or placebo) twice daily for 1 week, then titrated to 1000 mg BID for up to 25 additional weeks. The minimum maintenance dose of metformin was 1500 mg daily. The MADRS was administered, and insulin resistance was measured at baseline, week 2, and every 4 weeks thereafter (primary outcomes). Secondary outcomes included the Clinical Global Impressions Scale (bipolar disorders version; CGI-BP) and the Global Assessment of Functioning (GAF) scale. Returned pill counts were used to determine study medication adherence. The primary hypothesis that metformin would reverse insulin resistance (conversion) was tested using Fisher’s exact test. Improvements in MADRS scores (and other outcomes) in converters were tested using mixed-effects linear models.

The authors consented and screened 88 subjects, of whom 50 were randomized. Efficacy and safety data were reported for 45 subjects (20 metformin and 25 placebo). Thirty-nine subjects (87%) completed the study to the week-14 primary outcome endpoint, and 28 subjects completed the full 26 weeks. Medication adherence was 97% in both treatment groups. Mean subject age was 48, 76% were women, and the mean illness duration was 26 years. There were no significant between-group differences in weight, BMI, glucose, insulin, or blood levels of lithium or lamotrigine.

Ten metformin-treated patients no longer met insulin resistance criteria (converters) at week 14, versus 1 placebo-treated patient (Fisher’s exact p=0.0009). Converters had a significantly greater reduction in MADRS total scores beginning at week 6 and continuing through week 26, with large effect sizes (1.04 to 1.17). There were concordant improvements in GAF, Hamilton anxiety scale, and, to a lesser extent, CGI-BP scores in converters. The most common adverse events were diarrhea and vomiting, which were expected in the metformin group. At week 14, metformin-treated patients lost 1.7 kg and placebo-treated patients gained 1.4 kg, and this difference was statistically significant.

Study Conclusions

The authors concluded that reversal of insulin resistance by metformin resulted in statistically and clinically significant sustained improvements in depression, with no changes in mania, in patients with TRBD. No serious adverse events occurred, and the most common adverse events were gastrointestinal symptoms. Study limitations included the modest sample size, and whether a higher rate of conversion from insulin resistance could be achieved with either higher doses or extended-release preparations of metformin.

The Bottom Line

Reversal of insulin resistance by metformin improves depression, anxiety, and general functioning in a large proportion of patients with TRBD. These findings warrant replication in a larger trial.


1. Fountoulakis KN, Yatham LN, Grunze H, et al. The CINP guidelines on the definition and evidence-based interventions for treatment-resistant bipolar disorder. In J Neuropsychopharmacol. 2020;23(4):230-256.

2. Calkin CV, Ruzickova M, Uher R, et al. Insulin resistance and outcome in bipolar disorder, 2015. Br J Psychiatry. 2015;206(1):52-57.

3. Calkin CV, Chengappa KNR, Cairns K, et al. Treating insulin resistance with metformin as a strategy to improve clinical outcomes in treatment-resistant bipolar depression (the TRIO-BD Study): a randomized, quadruple-masked, placebo-controlled clinical trial. J Clin Psychiatry. 2022;83(2):21m14022.