OBJECTIVE: The current study examines whether antidepressants, contrary to current thinking, are safe and effective treatments for generalized anxiety disorder (GAD) not complicated by depression or panic disorder. DESIGN: Randomized, double-blind, placebo-controlled, flexible-dose, 8-week treatment study comparing imipramine hydrochloride (mean maximum daily dose, 143 mg), trazodone hydrochloride (255 mg), and diazepam (26 mg). PATIENTS: Two hundred thirty patients with a DSM-III diagnosis of GAD in whom major depression and panic disorder has been excluded, and who had a Hamilton Anxiety Scale total score of at least 18. SETTING: Seventy-five percent of patients were treated in family practice settings in the community, with the remainder treated in psychiatric practices, either academic or private. RESULTS: Patients treated with diazepam showed the most improvement in anxiety ratings during the first 2 weeks of treatment, with somatic symptoms being most responsive. From week 3 through week 8 trazodone achieved comparable, and imipramine somewhat better, anxiolytic efficacy when compared with diazepam, with psychic symptoms of tension, apprehension, and worry being more responsive to the antidepressants. Among completers, moderate to marked improvement was reported by 73% of patients treated with imipramine, 69% of patients treated with trazodone, 66% of patients treated with diazepam, but only 47% of patients treated with placebo. Overall, patients treated with antidepressants reported a higher rate of adverse effects than diazepam-treated patients, but attention rates were the same across all treatments. CONCLUSIONS: The results of the study need replication, but suggest a potentially important role for antidepressants, particularly imipramine, in patients suffering from GAD.