Twenty-seven inpatients participating in a stroke rehabilitation program were randomized to receive either placebo or trazodone hydrochloride (Desyrel) beginning a mean (+/- SEM) of 44 +/- 4 days after stroke. The target dosage was 200 mg/d. Patients with either a clinical diagnosis of depression or abnormal Zung depression scores showed a consistent trend toward greater improvement in Barthel activities of daily living (ADL) scores with trazodone than with placebo. An abnormal dexamethasone suppression test result was associated with significant improvement in the Barthel ADL scores of patients receiving trazodone (38 +/- 6 vs 20 +/- 6 for placebo). Patients with stroke and evidence of depression are therefore likely to benefit from treatment with trazodone.