To compare the safety and antidepressant efficacy of paroxetine, imipramine, and placebo, data from six centres using the same protocol were pooled. A double-blind parallel-group design was used, with therapy lasting six weeks. From week 2 onwards, both the 240 paroxetine-treated and the 237 imipramine-treated patients were significantly different from the 240 placebo-treated patients, but no different from each other. Side-effects with paroxetine were less likely to lead to drop-out than with imipramine. Paroxetine had a possible earlier antidepressant effect than imipramine, and a possible earlier beneficial effect on anxiety symptoms associated with depression.