Patients with recurrent affective illness (N-122) were randomly assigned to lithium carbonate, imipramine, or placebo therapy for two years following discharge from hospitalization for acute depression. Patients were classified as bipolar or unipolar based on the presence or absence of a history of mania. With bipolar patients lithium carbonate was significantly more effective than imipramine or placebo in preventing affective episodes (ie, manic or depressive attacks severe enough to require hospitalization or use of nonstudy drugs). There was a relatively high incidence of manic episodes on imipramine.
Unipolar patients responded equally well to lithium carbonate or imipramine. Both treatments were significantly more effective than placebo. The difference between treatments was due primarily to depressive episodes; manic episodes occurred infrequently and were evenly distributed across treatment groups. There was no significant difference between lithium carbonate and imipramine among unipolar patients.