Isocarboxazid and placebo were evaluated in 130 anxious depressives. Drug was superior to placebo on depression, anxiety, interpersonal sensitivity, and global measures, and on symptoms of hostility, anxiety, obsessiveness, and psychological-cognitive components of depression. There were no significant differences between treatment effects on psychomotor and typical vegetative symptoms. Isocarboxazid was more effective than placebo in major, but not in minor, depression. It was significantly more effective in depression classified as endogenous depression or melancholia by various diagnostic criteria. Drug was more effective than placebo in atypical depression with vegetative reversal and in Brief Psychiatric Rating Scale (BPRS)-derived profiles of anxious and hostile depression; there were no drug-placebo differences in atypical depression without vegetative reversal, or in BPRS retarded and agitated/excited depression. Interpersonal sensitivity emerged as an important drug-responsive dimension.