Key Takeaways
Extended Takeaways
- In this highly refractory cohort, the metabolic target identified a clinically meaningful subgroup: 91.2% had failed drug trials from at least 3 psychotropic drug classes, 55.6% had failed drugs from all 4 drug classes, and mean lifetime failed medication trials for bipolar disorder was 8.6.
- The antidepressant signal tracked with insulin resistance conversion rather than simple assignment to metformin; among non-converters, 40% on metformin and 29.2% on placebo still met the ≥ 30% MADRS response criterion, but converters had higher response rates at week 14 at 81.8% versus 39.3% (Fisher exact P = .031).
- Clinical improvement emerged early enough to monitor in routine follow-up: converters showed significantly greater improvement in MADRS and GAF beginning at week 6, with depression effect sizes remaining large at week 14 (d = 1.17) and week 26 (d = 1.04).
- Metformin’s benefit was not limited to depressive symptoms; anxiety and functioning also improved, with HAM-A effect sizes of d = 0.61 at week 14 and d = 1.0 at week 26, and GAF effect sizes of d = 1.47 at week 14 and d = 1.58 at week 26.
- Tolerability was favorable in a medically and psychiatrically complex population: adherence was 97% in both treatment groups, no subjects withdrew because of side effects or noncompliance, and no serious adverse events occurred over 26 weeks.
- For psychiatrists weighing metabolic tradeoffs, metformin produced a modest short-term weight advantage without destabilizing mood: at week 14, metformin-treated patients lost 1.67 kg while placebo-assigned patients gained 1.35 kg, and YMRS mean scores remained low throughout the study.
