Patterns of response to aripiprazole, lithium, haloperidol, and placebo across factor scores of mania

Michael J Ostacher, Trisha Suppes, Alan C Swann, James M Eudicone, Wally Landsberg, Ross A Baker, Berit X Carlson

A previous factor analysis of Young Mania Rating Scale and Montgomery-Åsberg Depression Rating Scale items identified composite factors of depression, mania, sleep disturbance, judgment/impulsivity, and irritability/hostility as major components of psychiatric symptoms in acute mania or mixed episodes in a series of trials of antipsychotics. However, it is unknown whether these factors predict treatment outcome. Data from six double-blind, randomized, controlled clinical trials with aripiprazole in acute manic or mixed episodes in adults with bipolar I disorder were pooled for this analysis and the previously identified factors were examined for their value in predicting treatment outcome. Treatment efficacy was assessed for aripiprazole (n = 1,001), haloperidol (n = 324), lithium (n = 155), and placebo (n = 694) at baseline, days 4, 7, and 10, and then weekly to study end. Mean change in factor scores from baseline to week 3 was assessed by receiver operating characteristics curves for percentage factor change at day 4 and week 1. Subjects receiving aripiprazole, haloperidol, and lithium significantly improved mania factor scores versus placebo. Factors most predictive of endpoint efficacy for aripiprazole were judgment/impulsivity at day 4 and mania at week 1. Optimal factor score improvement for outcome prediction was approximately 40% to 50%. Early efficacy predicted treatment outcome across all factors; however, response at week 1 was a better predictor than response at day 4. This analysis confirms clinical benefits in early treatment/assessment for subjects with bipolar mania and suggests that certain symptom factors in mixed or manic episodes may be most predictive of treatment response.