To evaluate the long-term safety and efficacy of adjunctive aripiprazole (ARI) to lithium (LI) or valproate (VAL) in delaying time to relapse in bipolar I disorder.

Bipolar I disorder subjects with a current manic or mixed episode received LI or VAL for at least 2 weeks; inadequate responders (YMRS score ≥ 16 and ≤35% decrease from baseline at 2 weeks) received adjunctive ARI. Subjects maintaining mood stability (YMRS and MADRS ≤ 12 for 12 consecutive weeks) were randomised 1:1 to double-blind ARI (10 to 30 mg/day) or placebo (PBO) plus LI or VAL. Relapse was monitored up to 52 weeks.

337 subjects were randomised to continuation of mood stabiliser plus adjunctive ARI or PBO; 61.3% and 52.7%, respectively, completed the study. Adjunctive ARI significantly delayed the time to any relapse, hazard ratio = 0.544 (95% CI: 0.33, 0.89, log-rank p = 0.014). Overall relapse rates at 52 weeks were 14.9% and 25.4% in ARI vs PBO subjects. A superior reduction in CGI-BP Mania Severity of Illness from baseline at 52 weeks was also observed (0.3 vs. 0.0, respectively, p = 0.01). Adverse events generally were as expected per known drug and illness profiles with no significant difference in mean change in body weight between adjunctive PBO (0.60 kg) and adjunctive ARI (1.07 kg) (p = 0.49 Week 52, LOCF).

Continuation of aripiprazole treatment increased time to relapse to any mood episode compared with placebo plus LI/VAL over 1 year, indicating a long-term benefit in continuing adjunctive aripiprazole to a mood stabiliser after sustained remission is achieved.