OBJECTIVE: The purpose of this study was to evaluate the outcome of bipolar disorder in the context of maintenance pharmacotherapy. METHOD: Eighty-two bipolar outpatients were followed prospectively for a mean of 4.3 years (minimum of 2 years); symptom rating and psychosocial outcome scales were used, and pharmacotherapy was rated on a 5-point scale. RESULTS: Despite continual maintenance treatment, survival analysis indicated a 5-year risk of relapse into mania or depression of 73%. Of those who relapsed, two-thirds had multiple relapses. Relapse could not be attributed to inadequate medication. Even for those who did not relapse, considerable affective morbidity was observed. A measure of cumulative affective morbidity appeared to be a more sensitive correlate of psychosocial functioning than was the number of relapses. Poor psychosocial outcome paralleled poor syndromal course. Poor psychosocial functioning, especially occupational disruption, predicted a shorter time to relapse. Depressions were most strongly related to social and family dysfunction. CONCLUSIONS: Even aggressive pharmacological maintenance treatment does not prevent relatively poor outcome in a significant number of bipolar patients.