A 6-week, double-blind, parallel group study compared the efficacy and tolerability of paroxetine and fluoxetine in 106 depressed geriatric outpatients (age, ≥65 years). Patients with an acute major depressive episode were randomized to receive either paroxetine (20 to 40 mg; N = 54) or fluoxetine (20 to 60 mg; N = 52) after a 3− to 7-day washout. Efficacy evaluations at weeks 1, 3, and 6 used the 21-item Hamilton Rating Scale for Depression (HAM-D). Cognitive function was assessed by use of the Mini-Mental State Examination (MMSE) and the Sandoz Clinical Assessment Geriatric Scale (SCAG). Tolerability was assessed by response to a nonleading question concerning adverse events. There were no significant differences between treatments at week 6 on the HAMD total, change from baseline. However, there was a statistically significant difference (p < 0.05) at week 3 in favor of paroxetine. Results from the MMSE and SCAG showed that, during treatment, patients in the paroxetine group were characterized by greater improvement in cognitive function than were those in the fluoxetine group. This result was statistically significant at week 3 for both scales (p < 0.05). Adverse events occurred most frequently within the gastrointestinal and nervous systems for both drugs, with no significant differences between treatments.