A hyperreactive autonomic nervous system has been claimed for patients with panic disorder. The goal of this project was to test for an impaired endurance capacity as one reason for their episodic excessive sympathetic arousal in panic attacks. Additionally, the reversibility of fitness deterioration should be investigated and underlined by clinically meaningful parameters. 39 patients (PP) with panic disorder with or without agoraphobia (DSM III-R) and 11 healthy, untrained control subjects (CO) were tested bicycle-spiroergometrically. For a subgroup of 10 patients and CO these measurements were repeated after a 10 weeks' endurance training program (3 times a week 45 min running/walking). PP at baseline ranged clearly below average fitness values of the German population (Pmax / BW = 2,27 ± 0,74 W/kg, ˙VO2max = 31,0 ± 9,1 ml/min·kg, PLac4 = work load at a blood lactate concentration of 4 mmol/l). The comparison with CO (representing a low-fit subgroup) revealed significant differences (p < 0,05) for Pmax, Pmax / BW, PLac4, PLac4/BW and ˙VO2max in favour of CO. Endurance training was effective but no inter-group differences in fitness changes were detected (p > 0,05 forΔ Pmax, Δ Pmax/BW, Δ PLac4, Δ PLac4/BW,Δ ˙VO2max, ΔPWC150, ΔPWC150/BW). The HR and RER at a blood lactate concentration of 4 mmol/l decreased in PP only (Δ HRLac4 = -12/min ± 27,6/min vs. +2,7/min ± 10,7/min, p = 0,143; Δ RERLac4 = -0,03 ± 0,07 vs. + 0,01 ± 0,02, p = 0,006). Conclusions: 1. PP show an impaired endurance capacity presumably because of phobic avoidance of physical activity. 2. Their trainability is undisturbed. 3. After endurance training the known panic-inducing agent lactate generates a reduced hyperventilatory response - possibly an index for clinical improvement.