Patients with impaired pulmonary function are at increased risk for the development of postoperative complications. Recently exercise testing and predicted postoperative (ppo) function have gained increasing importance in the evaluation of lung resection candidates. We prospectively evaluated an algorithm for the preoperative functional evaluation that was developed at our institution. This algorithm incorporated the cardiac history including an electrocardiogram (ECG), and the three parameters FEV1, diffusing capacity of the lungs for carbon monoxide (DLCO), and maximal oxygen uptake (VO2max), as well as their respective ppo values (FEV1-ppo, DLCO-ppo, and VO2max-ppo) calculated based on radionuclide perfusion scans. A consecutive group of 137 patients (mean age 62 yr; range 23 to 81; 102 males, 35 females) with clinically resectable lesions underwent assessment according to our algorithm. Five patients were deemed functionally inoperable, 132 passed the algorithm and underwent pulmonary resections with standard thoracotomy: 9 segmental or wedge resections, 85 lobectomies (inclusive 3 bilobectomies), and 38 pneumonectomies. All patients were extubated within 24 h. The mean stay in the ICU was 1.4 (+/- 1.8) d, and the mean hospital stay was 14.6 (+/- 5) d. Postoperative complications (within 30 d) occurred in 15 patients (11%), of whom two died (overall mortality rate 1.5%). In comparison to our previous series this meant a 50% reduction in complications whereas the percentage of inoperable patients remained unchanged (4% now, 5% before). We conclude that adherence to our algorithm resulted in a very low complication rate (morbidity and mortality), and excluded more rigorous patient selection as a bias for the improved results.