Olanzapine was licensed in the USA by the Food and Drug Administration in 2003 for the prevention of relapse in patients with bipolar disorder when the acute manic episode had responded to treatment with olanzapine. However, olanzapine is commonly used in clinical practice for preventing relapse in patients with bipolar disorder even when acute response has not been demonstrated. The aim of this systematic review and meta-analysis is to determine the effectiveness and acceptability of olanzapine in preventing recurrent mood episodes in bipolar disorder. MEDLINE, EMBASE, the Cochrane Collaboration Depression, Anxiety and Neurosis Controlled Trials Register and the Cochrane Central Register of Controlled up to July 2008 were accessed. Only randomised controlled trials comparing olanzapine with placebo or other active drugs for long-term treatment were included. Two reviewers independently extracted data. Authors were contacted to provide additional data. Of the five trials included in this review, four were conducted by Eli Lilly, the manufacturer of olanzapine. Olanzapine was more effective than placebo at preventing manic relapse, but there was no difference between olanzapine (alone or in combination with lithium or valproate) and placebo (alone or in combination with lithium or valproate) in terms of relapse into any mood episode, as defined as primary outcome by authors in each of the primary studies. We conclude that olanzapine may prevent further manic episodes only in patients who have responded to olanzapine in an acute manic or mixed episode and who have not previously had a satisfactory response to lithium or valproate.