Relationship of open acute mania treatment to blinded maintenance outcome in bipolar I disorder

Background: Most bipolar treatment guidelines recommend that acutely effective therapy be continued as maintenance treatment, unless contraindicated. No published studies have statistically evaluated the relationship between the outcomes in maintenance treatment and the principal acute agent. Methods: Post-hoc analysis of prophylactic results obtained during a 12-month maintenance study in recently manic patients assessed the relationship between open treatment and randomized maintenance treatment. Patients were grouped according to mood stabilizer (if any) received on the last day of the open-label, stabilization phase (divalproex, lithium, or no mood stabilizer), and the mood stabilizer to which they were randomly assigned in the double-blind, maintenance phase (divalproex, lithium, or placebo). Results: Patients treated with open-phase divalproex and randomized to divalproex maintenance therapy had a significantly longer period before development of a mood episode compared to those randomized to either placebo (Log-rank, p = 0.05) or lithium (Log-rank, p = 0.04), and better tolerability than those randomized to lithium. Patients treated with open-phase lithium showed no treatment-related differences in time to a mood episode after being randomized to lithium, divalproex, or placebo and tolerability was comparable in all groups. Limitations: Open treatment was selected by treating physicians, contributing to possible selection bias. More patients were randomized to divalproex than to lithium or placebo. The study was therefore better powered to detect divalproex effects. Conclusions: Acute response to divalproex predicted both superior prophylactic response and tolerability to divalproex, than to placebo (efficacy) and to lithium (efficacy and tolerability). Results could have implications for both clinical practice and clinical trial design.