Nineteen children (mean [+/- SD] age, 14.5 +/- 2.3 years) with severe, primary obsessive-compulsive disorder completed a ten-week, double-blind, controlled trial of clomipramine hydrochloride (mean dosage, 141 mg/day) or placebo, each of which was administered for five weeks. Half of the subjects had not responded to previous treatment with other tricyclic antidepressants. There was a significant improvement in observed and self-reported obsessions and compulsions that was independent of the presence of depressive symptoms at baseline. Improvement in obsessive-compulsive symptoms did not correlate significantly with plasma concentrations of the drug or its metabolites. Clomipramine appears to be effective in the treatment of children with obsessive-compulsive disorder and the treatment seems to be independent of an antidepressant effect.