Seventy-three studies were included in the review (n=114,521): 39 RCTs and 34 open trials.
Thirty-five studies (48 comparisons) provided sufficient data for meta-analyses.
Risk of mania was significantly increased among patients exposed to antidepressants compared with those not exposed: 12% versus 7.46% (RR 1.76, 95% CI 1.33 to 2.33; 48 comparisons, NNH=27, 95% CI 17 to 75). Rates of mania/hypomania were greater among patients with bipolar disorder with or without antidepressant treatment (15.3% with versus 13.8% without antidepressants) than in patients with major depressive disorder (5.97% versus 1.24%).
Elimination of one unusually large trial did not significantly alter the results. Similar results were found for RCTs, clinical cohort studies, patients with major depressive disorder and those who received tricyclic antidepressants. There was little difference in risk of mania with or without mood stabilisers. No statistically significant differences were found between antidepressants and no antidepressants in the risk of mania/hypomania in patients with bipolar disorder or those who received modern antidepressants or monoamine oxidase reuptake inhibitors.