Psychological parameters that are believed to affect estimations of cardiovascular fitness were examined in patients with panic disorder and nonclinical controls. Fifty-four participants [panic disorder patients (n = 27) and age- and sex-matched nonclinical controls (n = 27)] completed a cycle ergometer test and were compared on the basis of estimated VO2 max. Participants were randomly assigned to experimental conditions in which they received heart-rate feedback or no feedback during the test. Patients with panic disorder exhibited lower VO2 max and decreased exercise tolerance (i.e., were more likely to discontinue the test) than nonclinical controls. Furthermore, individuals with high anxiety sensitivity (i.e., a fear of autonomic arousal), but not a panic disorder diagnosis per se, achieved significantly lower VO2 max when provided with heart-rate feedback. Moreover, diagnostic status interacted with levels of anxiety sensitivity to predict VO2 max. Patients with panic disorder display poorer cardiovascular fitness after controlling for anxiety and other factors that underestimate performance during fitness testing.