Comparative efficacy and acceptability of drug treatments for bipolar depression: a multiple-treatments meta-analysis

Treatment of bipolar depression is complicated by variable response and risk of switch to mania. Guidance is informed by the strength of evidence rather than by comparative data.

We performed a multiple-treatments meta-analysis of randomised, double-blind, controlled comparisons of 4-16 weeks in adults in bipolar depression. The primary efficacy outcome was effect size. The primary acceptability outcome was 'switch to mania'. Secondary outcomes were likelihood of response and withdrawals from trials.

Twenty-nine studies were included (8331 participants). Olanzapine + fluoxetine and olanzapine performed best on primary outcome measure being ranked highest for effect size. Switch to mania was least likely with ziprasidone and then quetiapine. Olanzapine + fluoxetine was also ranked the highest for response with lurasidone second, but olanzapine + fluoxetine and olanzapine had the optimal effect on response and withdrawal from treatment when the two parameters were considered together. Several treatments [monoamine oxidase inhibitors (MAOIs), ziprasidone, aripiprazole and risperidone] have limited or no therapeutic activity in bipolar depression.

Olanzapine + fluoxetine should be first-line treatment. Olanzapine, quetiapine, lurasidone, valproate and selective serotonin re-uptake inhibitors are also recommended. Tricyclic antidepressants and lithium are worthy of consideration but lamotrigine (high risk of switching, less robust efficacy) and MAOIs, ziprasidone, aripiprazole and risperidone (no evidence of efficacy) should not be used.