Fifty-five young adult patients completed a study comparing alprazolam, propranolol, and placebo in the treatment of panic disorder and agoraphobia with panic attacks. Twenty completed 5 weeks of treatment with alprazolam. No concomitant psychological treatment was administered. Plasma alprazolam levels were determined at baseline and at the end of the trial by means of automated gas chromatography. These levels were significantly correlated with dose (p= 0.001). Dividing the data into quartiles based on alprazolam concentration, no direct, linear relationship was found between alprazolam levels and response (as defined by a criterion of zero panic attacks). However, analysis of the combined middle two quartiles versus the combined upper and lower quartiles showed a positive trend toward a curvilinear relationship; i.e., there was a greater response rate within the 18–62 ng/ml concentration range than in the combined 0–17 plus 63–107 ng/ml range, (x2= 2.4;p= 0.12). The findings are very preliminary in nature. It remains to be seen if the results will reach significance with a larger sample size and a more tightly controlled study design.