At the end of a two-week single-blind placebo baseline, 43 patients with a diagnosis of panic disorder with agoraphobia without significant dysphoria-depression and with moderate to severe panic and phobic symptoms were assigned to, and 32 of them completed, a placebo-controlled (n = 7) dose-response study with three weight-adjusted imipramine hydrochloride dosages: 0.5 mg/kg/d (n = 10), 1.5 mg/kg/d (n = 9), and 3 mg/kg/d (n = 6). Eleven patients, three from the medium-dose and eight from the high-dose conditions, dropped out owing to side effects. No instructions or encouragement for self-directed exposure to phobic situations or other coping strategies with panic or fear were given throughout the trial. Compliance, as assessed by pill counts and by plasma tricyclic levels, was high. Results provided strong evidence for a positive dose-response relationship on panic and phobic symptoms and confirmed earlier suggestions (1) that imipramine without concurrent exposure possesses a significant antipanic and antiphobic effect, (2) that improvement correlates primarily with imipramine but not N-desmethylimipramine plasma levels, and (3) that side effects prevent optimum dose buildup in a substantial proportion of patients with this disorder.