OBJECTIVE: The authors' goal was to evaluate the relationship between plasma concentrations of alprazolam and both treatment response and side effects in patients with panic disorder and agoraphobia. METHOD: Ninety-six patients with panic disorder and agoraphobia were treated at three sites in a 6-week, fixed-dose, double-blind, placebo-controlled, dose-response study of 2 mg/day or 6 mg/day of alprazolam. Assessments were made of panic attacks, avoidance behavior, generalized anxiety, and global response. Blood samples were collected throughout the study and analyzed for alprazolam and other benzodiazepines. RESULTS: Patient compliance with the protocol was judged to be good on the basis of plasma concentrations. According to logistic regression analysis, the relationships between plasma alprazolam concentration and response, as reflected by number of panic attacks reported, phobia ratings, physicians' and patients' ratings of global improvement, and the emergence of side effects, were significant. However, there was no significant relationship between plasma alprazolam concentration and the degree of generalized anxiety symptoms. CONCLUSIONS: The authors conclude that plasma concentration of alprazolam is related to treatment response, particularly in panic attacks. The alprazolam concentration associated with treatment response or with emergence of a given side effect varied widely among individuals, highlighting the necessity for individualized dose adjustment to obtain optimal treatment response while minimizing side effects.