Background: Disordered breathing among patients,vith panic disorder, including hyperventilation during attacks and increased anxiogenic response to carbon dioxide (CO2) inhalation, is well established We wished to assess whether there is a change in the physiological response to CO2 after patients have undergone antipanic therapy with either tricyclic antidepressants or cognitive behavioral therapy (CBT). Methods: Twenty-nine patients with panic disorder underwent baseline CO2 sensitivity testing using the traditional Read rebreathing method and then received either antidepressant treatment (n = 21) or CBT (n = 8). After completing treatment, CO2 testing was repeated. A comparison sample of 14 normal volunteers also had two CO2 sensitivity tests, separated by an average of 21.6 (SD = 8.8) weeks. Results: Using a liberal standard, in which all CO2 sensitivity tests whose correlations between minute ventilation and end-tidal CO2 were at least .75 were used, patients, but not controls, demonstrated a significant reduction in CO2 sensitivity between the first and second rest. Using a more conservative .90 correlation standard reduced the sample size available and resulted in trend reduction in patients but no significant change in controls. There was a suggestion that the change was most pronounced in treatment responders, although the number of patient nonresponders is extremely small in this sample, Conclusions: These data indicate that treatment reduces CO2 sensitivity in patients with panic disorder, We speculate that manipulation of the serotonergic and noradrenergic neurotransmission systems, both known to play a role in the control of respiration may have a specific effect in reducing respiratory hype, activity in panic disorder. (C) 1997 Society of Biological Psychiatry.