Endurance capacity was determined by bicycle spiroergometry in patients with panic disorder before (n = 38) and after (n = 10) a 10 week running program and compared to untrained healthy control subjects carrying out the same training (n = 11) and patients receiving clomipramine drug therapy (n = 7) or placebo (n = 7). Before the running program maximal oxygen uptake (VO2peak) and the workload corresponding to a lactate concentration of 4 mmol/l (PLAC4) were significantly reduced in panic patients compared to controls. Patients in the running program and healthy controls improved PLaC4 significantly by running. No significant differences in endurance gains were found between these groups, showing that patients and controls improved equally. At study termination (week 10) running was more efficient than placebo in improving panic symptoms (Bandelow PanicAgoraphobia Scale, Hamilton Anxiety Scale, Clinical Global Impression Scale). Clomipramine treatment was better than placebo (all scales) and running (only Clinical Global Impression). Endurance capacity did not correlate with anxiety scores at baseline, nor did improvement in fitness substantially correlate with changes in psychopathology measures during the study. Panic patients were shown to have a decreased endurance capacity which can be raised by training. Endurance training based on spiroergometric results gives rise to clinical improvement.