• Five hundred four outpatients suffering from a major depressive episode were randomly assigned to receive either amitriptyline, doxepin, alprazolam, or placebo. The study was conducted in three treatment centers during a six-week period. All three active medications produced significantly more clinical improvement than did placebo, irrespective of the patient's initial anxiety, depression, and psychomotor retardation and irrespective of the patient's assignment to various subtypes of depression, including the DSM-III melancholia subtype. Compared with placebo, sedation was reported more frequently with all three medications, whereas anticholinergic effects were reported more frequently only for the two tricyclic antidepressants, but not for alprazolam.