OBJECTIVE This study examined whether a selective serotonin reuptake inhibitor (paroxetine) had comparable efficacy but greater tolerability than a tricyclic antidepressant (imipramine) in depressed patients with HIV infection. METHODS Seventy-five HIV-positive patients (45% of whom had AIDS) were blindly and randomly assigned to receive paroxetine (N = 25), imipramine (N = 25), or placebo (N = 25) in a 12-week trial. The Hamilton Anxiety Rating Scale, the Hamilton Depression Rating Scale, the Clinical Global Impression scale, and the SAFETEE general inquiry (for safety and tolerability) were administered at weeks 2, 4, 6, 8, and 12. RESULTS Fifty-six (75%) of the 75 patients completed 6 weeks and 34 (45%) completed 12 weeks of the trial. The mean daily doses of both paroxetine (33.9 mg) and imipramine (162.5 mg) were significantly more effective than placebo; they were comparably effective at weeks 6, 8, and 12 according to the intent-to-treat analysis and at week 8 according to the analysis for the subjects who completed the trial (for them, only imipramine was superior to placebo at week 12). There were significantly more dropouts due to side effects from imipramine (48%) than from both paroxetine (20%) and placebo (24%). CONCLUSIONS Depressed patients with HIV infection responded to imipramine or paroxetine at a higher rate than to placebo irrespective of severity of immunosuppression. Because paroxetine was much better tolerated than imipramine, its overall effectiveness may be greater. However, because of the small study group and the high attrition rate, these findings cannot be generalized and may need replication in a larger study group.