Effectiveness of mood stabilizers and antipsychotics in the maintenance phase of bipolar disorder: a systematic review of randomized controlled trials

BACKGROUND: Bipolar disorder (BD) is a leading cause of disability. Systematic reviews of randomized trials for the treatment of the maintenance phase of BD are lacking. OBJECTIVES: To determine the efficacy and tolerability of mood stabilizers and antipsychotics in the maintenance treatment of BD. METHODS: We systematically reviewed randomized controlled trials of licensed medications for the treatment of any phase of BD. We included randomized controlled trials comparing a medication to placebo or another medication. Comprehensive searches of electronic databases were conducted to March 2005. Outcomes investigated were relapse due to mania, depression or any mood episode, and withdrawal due to any reason or due to an adverse event. Data were combined through meta-analysis. RESULTS: Fourteen studies (n = 2,526) met the inclusion criteria. Lithium, lamotrigine, olanzapine and valproate semisodium each demonstrated evidence to support long-term use. Compared with placebo, all medications were more effective at preventing relapse because of any mood episode. Hazard ratios (HR) were 0.68 95% confidence interval (CI) = 0.53-0.86 for lithium, 0.68 (95% CI = 0.55-0.85) for lamotrigine, and 0.82 (95% CI = 0.57-1.20) for valproate semisodium; for olanzapine, the risk ratio (RR) was 0.58 (95% CI = 0.49-0.69). Lithium and olanzapine significantly reduced manic relapses (HR = 0.53; 95% CI = 0.35-0.79 and RR = 0.37; 95% CI = 0.24-0.57, respectively). Lamotrigine and valproate semisodium significantly reduced depressive relapses (HR = 0.65; 95% CI = 0.46-0.91 and RR = 0.40; 95% CI = 0.20-0.82, respectively). Lithium significantly reduced manic relapses compared with lamotrigine (HR = 0.56; 95% CI = 0.34-0.92) and olanzapine significantly reduced manic relapses compared with lithium (RR = 1.69; 95% CI = 1.12-2.55). Withdrawal due to an adverse event was approximately twice as likely with lithium compared with valproate semisodium (RR = 1.81; 95% CI = 1.08-3.03) and lamotrigine (RR = 2.20; 95% CI = 1.31-3.70). There were few data for carbamazepine or medications given as adjunct therapy. CONCLUSIONS: Mood stabilizers have differing profiles of efficacy and tolerability, suggesting complementary roles in long-term maintenance treatment.