The inhalation of 35% carbon dioxide has consistently been shown to provoke panic attacks in panic disorder patients. We aim to determine if an acute dose of clonazepam (2 mg) attenuates the panic attacks induced by an inhalation of 35% carbon dioxide in panic disorder. Twenty-two panic disorder patients who had been drug-free for 1 week participated in a carbon dioxide challenge test 1 h after a dose of either 2 mg of clonazepam or placebo with a randomized double-blind method. Also in a double-blind design during the tests the patients inhaled either atmospheric compressed air ('placebo control') or the carbon dioxide mixture. All patients participated in both tests which were done with a 20-min interval. Immediately before and after the inhalation, the anxiety levels and the symptoms of panic were always assessed. In the clonazepam group (n=11) two patients (18.2%) had a mild panic attack and in the placebo group (n=11) nine patients (81.8%) had a moderate to severe panic attack in the CO(2) challenge test. No patient had panic attacks during inhalation of atmospheric compressed air although anticipatory anxiety levels tended to be higher than in the CO(2) tests. After the CO(2) test anxiety levels were significantly greater in the CO(2) group (three-way ANOVA with Geisser-Greenhouse adjustments, F(31.92,1.86)=17.15, d.f.=7, P=0.013). Although a small sample was studied, the findings suggest the efficacy of an acute dose of clonazepam in attenuating panic attacks induced by carbon dioxide inhalation.