In a previous study by Dr. Calkin, the principal investigator of this study, persons with bipolar disorder and either type II diabetes or insulin resistance were found to experience more severe symptoms of bipolar illness and a lower response to treatment, compared to persons with bipolar disorder who did not have type II diabetes or insulin resistance. To further explore these findings, the investigators have developed this study to see if treating insulin resistance (using metformin, a drug used to improve the body's use of insulin) may also help improve the symptoms of bipolar illness.

This is a 26-week randomized, double-blind, parallel group prospective study of the effectiveness of treating insulin resistance (IR) to improve mood in patients with IR and treatment-resistant bipolar depression (TRBD). The investigators will compare the effects of treating IR (with metformin) versus placebo on outcome in each patient. The primary outcome will be change in Montgomery-Ǻsberg Depression Rating Scale (MADRS) scores. Patients' current optimized mood stabilizing treatment as usual (TAU, according to the Canadian Network for Mood and Anxiety Treatments [CANMAT] or American Psychiatric Association [APA] guidelines) must remain unchanged for a period of at least 4 weeks prior to and throughout the study. Patients will undergo a baseline assessment and then be randomized to treatment with metformin or placebo with titration to full dose after 2 weeks. Patients will remain on full treatment for 24 weeks thereafter (total trial duration of 26 weeks for each patient). In those patients with TRBD assigned to treatment with the insulin sensitizer metformin, a significant improvement in depression symptoms will be mediated by the conversion of IR to insulin sensitivity.

Subjects: We aim to enrol 110 subjects with IR and TRBD from 2 sites: the primary site in Halifax, Nova Scotia, Canada, and a second site in Pittsburgh, Pennsylvania, USA.

Inclusion Criteria:

1. 18 years of age or older
2. diagnosis of BD I or II
3. non-remitting BD as defined by the presence of mood symptoms of at least moderate severity, indicated by a MADRS score ≥ 15 despite being on optimal treatment according to the CANMAT/APA guidelines
4. HOMA-IR ≥ 1.8, indicating IR (subjects will have FPG and FSI testing done to determine whether they have IR or T2D)
5. current episode of depression 4 weeks or longer in duration
6. on a stable optimal dose of mood stabilizing treatment for at least 4 weeks prior to study entry

Exclusion Criteria:

1. Diagnoses of organic mood disorder, mood disorder not otherwise specified, alcohol dependence, T1D or T2D
2. presence of rapid cycling (by DSM-5 criteria), mania, (indicated by a Young Mania Rating Scale [YMRS] score > 15), or suicide ideation (current score of 5 on the Suicidal Ideation section of the Columbia-Suicide Severity Rating scale [C-SSRS])
3. patient receiving metformin < 2 weeks prior to study entry
4. metformin allergy or sensitivity
5. metformin contraindicated where liver function tests > three times the upper limit of normal, estimated glomerular filtration rate (eGFR) < 30, CBC revealing megaloblastic anemia or pre-existing untreated B12 deficiency
6. pregnancy or breastfeeding
7. lactose intolerance, diagnosed by a physician
8. chronic use of narcotic medications
9. patient lacks full capacity to consent to study participation.