A controversy exists over whether or not preoperative exercise testing can predict postthoracotomy complications. This study was designed to evaluate the usefulness of a presurgical exercise protocol in patients with lung disease, but no evidence of cardiac disease. Seventy patients underwent baseline pulmonary function testing and split function perfusion studies, when indicated, to calculate predicted postoperative pulmonary function. Noninvasive data were incrementally collected from 17 patients by using a treadmill exercise tolerance test that was designed to elicit maximal performance. Inhaled and exhaled gas flow and volume, the partial pressure of O2 and CO2, maximal O2 consumption (VO2max), and maximal minute ventilation (VE max) were measured. The breathing and heart rate reserves were calculated by standard formulae in an attempt to separate cardiac from pulmonary exercise limitation. Two patients had postoperative cardiopulmonary complications after thoracotomy and lung resection, and six patients had noncardiopulmonary complications. There was no significant prognostic relationship among VO2max, VE max, maximum O2 pulse, and the incidence of postoperative cardiopulmonary complications. The percentages of predicted VE max and predicted maximum heart rate were related to the occurrence of total complications, but not specifically to cardiopulmonary complications. The results emphasize the difficulty in attempting to exercise thoracotomy candidates with chronic lung disease to maximal performance. Excluding patients from further surgical consideration because of exercise limitation is not feasible based on these data.