OBJECTIVE The authors investigated trazodone as a hypnotic for depressed patients who had persistent, exacerbated, or new insomnia while taking either fluoxetine or bupropion. METHODS Seventeen depressed patients who had insomnia while taking fluoxetine or bupropion were given either trazodone or placebo in a double-blind crossover trial. Sleep was assessed by self-report with the Pittsburgh Sleep Quality Index and the sleep items of the Yale-New Haven Hospital Depressive Symptom Inventory. RESULTS Improvement with trazodone, but not with placebo, was shown by the total Pittsburgh index scores and Yale-New Haven inventory total sleep scores and by the Pittsburgh index measures of sleep duration and Yale-New Haven inventory measures of early morning awakening, and there was a trend toward improvement in the Yale-New Haven inventory item regarding middle of the night awakenings. Subjective sleep quality and sleep latency also showed a trend toward improvement, but the Pittsburgh index measures of sleep efficiency and disturbances and the Yale-New Haven inventory item regarding difficulty falling asleep were unaffected by trazodone. One patient dropped out because of excessive daytime sedation with trazodone, and another dropped out because of nonresponse to placebo. Of the completers, 67% experienced overall improvement in sleep with trazodone according to a priori criteria, whereas only 13% experienced improvement with placebo. CONCLUSIONS Trazodone is an effective hypnotic for patients with antidepressant-associated insomnia.