In a double-blind multicentre study of outpatients with DSM-III-R major depressive disorder, 129 sertraline and 129 placebo patients were evaluated over a 6-week period. Sertraline exhibited a significantly greater (P < 0.001) antidepressant effect compared to placebo as measured by the HAM-D, MADRS, CGI-S and CGI-I. In the subset of patients with severe depression (baseline HAM-D ≥ 25), sertraline was also significantly more effective than placebo (P < 0.05). Side effects were more commonly reported in sertraline (59%) compared to placebo (38%) patients; the most common being nausea, headache and insomnia. A subset of 107 patients (66 sertraline; 41 placebo) who were defined as responders (CGI-I of 1 or 2) after 6 weeks treatment were entered into a 20-week continuation phase. In this responder subset, there was continuing improvement in both groups of patients, but with no significant differences in mean HAM-D or MADRS between the groups. However, a higher number of sertraline patients were associated with a persistent pattern of improvement relative to placebo (P < 0.05). The incidence of side effects was similar in sertraline (52%) and placebo (49%) treated patients in the continuation period.