Bupropion, a new antidepressant with a novel structure and neurochemical profile, was compared to placebo in a three-center evaluation. Seventy-five depressed inpatients, 48 on bupropion and 27 on placebo, participated for up to 28 days. At dosages of 300-600 mg/day, bupropion was significantly more active than placebo in reducing depressive and anxious symptomology, beginning at the 3-week evaluation. Bupropion was particularly efficacious in more severely ill and older patients. Symptoms reflecting cognitive disturbance, somatic anxiety, and psychomotor retardation were most improved; sleep disturbances showed less marked effects. Appetite and diurnal variation were not markedly influenced. Adverse effects were minimal, with no significant cardiovascular or clinical laboratory findings. Three patients developed a skin rash. There was a notable absence of sedative, anticholinergic, and cardiovascular-related side effects. Bupropion has demonstrated antidepressant effects and a spectrum of activity that holds promise for a difficult subgroup: the older depressed patient with retardation and cognitive disturbance.