Number needed to treat or harm analyses of olanzapine for maintenance treatment of bipolar disorder.

The number-needed-to-treat (NNT) or the number-needed-to-harm (NNH) analysis was performed on olanzapine and comparators for all known controlled clinical studies of olanzapine for bipolar maintenance treatment or relapse prevention to assess safety and efficacy. Studies compared olanzapine (n = 225) and placebo (n = 136) for 12 months, olanzapine (n = 217) and lithium (n = 214) for 12 months, and olanzapine plus lithium or valproate (n = 72) and placebo plus lithium or valproate (n = 64) for 18 months. For prevention of all-cause treatment discontinuation, the NNT was 7 to 8. For 9 of 11 efficacy and disposition measures examined, beneficial outcomes were more common with olanzapine than placebo. Beneficial outcomes were more common with olanzapine than lithium for 7 measures and more common for olanzapine plus lithium or valproate than placebo plus lithium or valproate for 1 measure. The NNHs of 5 to 8 for a weight gain of 7% or higher and 10 to 11 for the increase in body mass index category to overweight or obese during maintenance treatment indicated that these outcomes were more common for olanzapine or olanzapine plus mood stabilizers than for the comparators. All efficacy and disposition measures showing significant differences between groups for 12 to 18 months have NNTs favoring olanzapine or olanzapine plus lithium or valproate over placebo, lithium, or placebo plus lithium or valproate. However, the NNHs favor these comparators for avoidance of weight gain and of increase in body mass index category to overweight or obese. Clinicians should consider these and other potential benefits and risks in using maintenance treatments for patients with a bipolar disorder.