Fluvoxamine and imipramine were compared to placebo in an 8-week doubleblind randomized multicentre trial comprising of 148 outpatients between 19 and 57 years of age (mean: 35) with a DSM-III-R diagnosis of Panic Disorder. mean daily dose at endpoint was: fluvoxamine, 171.4 mg; imipramine 164.7 mg. The mean number of panic attacks per week at baseline were 10.9, 14.4 and 6.5 for fluvoxamine, imipramine and placebo, respectively. The intent-to-treat analysis of the change from baseline (difference score) of the number of panic attacks at endpoint revealed: a difference of 3.3 attacks (95% CI: -0.3, 6.8) between fluvoxamine and placebo and a difference of 6.0 attacks (95% CI: 1.5, 10.5) between imipramine and placebo. Treatment was stopped prematurely in 31 (62%) on fluvoxamine, 16 (33%) on imipramine and 29 (58%) on placebo. The number of patients withdrawing due to intolerance was 13 (26%) for fluvoxamine, 10 (21%) for imipramine and 4 (8%) for placebo. The number of patients withdrawing due to lack of efficacy was 10 (20%) for fluvoxamine, 4 (8%) for imipramine and 12 (24%) for placebo. Overall, this study demonstrated that fluvoxamine was not effective in the treatment of panic disorder but did show a strong effect for imipramine. A chance occurrence of significantly fewer number of panic attacks in the placebo group at baseline may limit the conclusions of this study.