OBJECTIVE We investigated the mechanism involved with the initial drop and subsequent recovery of exercise capacity in the early postoperative period of thoracotomy patients. METHODS Sixteen patients (13 who had undergone lobectomy, 3 who had undergone pneumonectomy) underwent a routine pulmonary function test (PFT) and a cardiopulmonary exercise test preoperatively, within 14 postoperative days (POD; post-1; mean +/- SD, 9 +/- 2 POD), and after 14 POD (post-2; mean, 26 +/- 12 POD). RESULTS After surgery on post-1, PFT results of FVC, FEV(1), and maximum ventilatory volume (MVV) significantly decreased. Oxygen uptake (VO(2)) at a venous blood lactate level of 2.2 mmol/L (La-2. 2), which was adopted as the empirical anaerobic threshold, and maximum V O(2) (VO(2)max) decreased significantly to 88.2 +/- 7.9% and 73.1 +/- 15.4% of the preoperative values, respectively. La-2.2 min ventilation (VE)/ MVV and maximum VEmax)/MVV increased significantly from 0.36 +/- 0.08 to 0. 66 +/- 0.20 and from 0.58 +/- 0.14 to 0.80 +/- 0.09, respectively. On post-2, though La-2.2 VO(2) did not change, VO(2)max improved significantly to 81.5 +/- 19.7% of the preoperative values, in association with significant increases in maximal tidal volume and VEmax, which were produced by significant increases in the PFT results. La-2.2 VE/MVV also decreased significantly to 0.49 +/- 0.13, which indicated a sufficient recovery of respiratory reserve at submaximal exercise. CONCLUSIONS The initial drop of exercise capacity after lung resection seems to be derived from both circulatory and ventilatory limitations. Further, the subsequent recovery within 1 month seems to be produced by an improvement in ventilatory limitation, which was caused by the surgical injury to the chest wall.