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

Abstract

Objective: Therapeutic options are limited for treatment-resistant bipolar depression (TRBD). Insulin resistance (IR) confers increased risk for TRBD. We investigated metformin, an insulin sensitizer, to reverse IR and improve clinical outcomes in TRBD. Methods: Using a random-assignment (1:1), intent-to-treat, 2-site, quadruple-masked, parallel-group (metformin to 2,000 mg/d or placebo) clinical trial design, patients with DSM-5 bipolar disorder (BD) type I or II and IR received study medication for 26 weeks (February 2016 to October 2019). The primary outcome was the change in depression rating scores (Montgomery-Asberg Depression Rating Scale [MADRS]) at 14 weeks between those who no longer met IR criteria (converters) and those who still did (non-converters). Additional outcomes included scores on the Global Assessment of Functioning (GAF); the Clinical Global Impressions Scale, Bipolar Disorders version (CGI-BP); and the Hamilton Anxiety Rating Scale (HAM-A) and maintenance of improved outcomes up to 26 weeks. Results: Forty-five BD patients were randomized to metformin (n = 20) or placebo (n = 25), and at 14 weeks or later, 11 subjects no longer met IR criteria (n = 10 with metformin, n = 1 with placebo; P= .0009). These converters experienced significant improvements in MADRS (P values ranged from .031 to .008) and GAF (P values ranged from .045 to .008) scores compared to non-converters beginning at week 6, sustained to week 26. HAM-A (P= .022 at week 14 and .019 at week 26) and CGI-BP change scores (P= .046 at 26 weeks) significantly favored converters over non-converters. Effect sizes were large for the MADRS and GAF (Cohen d > 1 at 14 and 26 weeks) and large for the HAM-A and CGI-BP at 26 weeks. Transient gastrointestinal side effects occurred under both treatment conditions. Conclusions: Pending replication, this early study suggests that reversal of IR by metformin offers a path out of TRBD. Further characterization of metformin converters with TRBD will prove informative.