The aim of this study was to investigate whether invasive exercise testing with gas exchange and pulmonary haemodynamic measurements could contribute to the preoperative assessment of patients with lung cancer at a high-risk for lung resection. Sixty-five patients scheduled for thoracotomy (aged 66+/-8 yrs (mean+/-SD), 64 males, forced expiratory volume in one second (FEV1) 54+/-13% predicted) were studied prospectively. High risk was defined on the basis of predicted postpneumonectomy (PPN) FEV1 and/or carbon monoxide diffusing capacity of the lung (DL,CO) <40% pred. Arterial blood gas measurements were performed in all patients at rest and during exercise. In 46 patients, pulmonary haemodynamic measurements were also performed at rest and during exercise. Predicted postoperative (PPO) values for FEV1 and DL,CO were calculated according to quantitative lung scanning and the amount of resected parenchyma. There were four postoperative deaths (6.2% mortality rate) and postoperative cardiorespiratory complications developed in 31 (47.7%) patients. Patients with respiratory complications only differed from patients without or with minimal (arrhythmia) complications in FEV1,PPO. Peak O2 uptake and haemodynamic variables were similar in both groups. The four patients who died had a lower FEV1,PPO, a lower DL,CO,PPO and a greater decrease in arterial oxygen tension during exercise, compared with the remaining patients. In conclusion, the forced expiratory volume in one second, together with the extent of parenchymal resection and perfusion of the affected lung, are useful parameters to identify patients at greatest risk of postoperative complications among those at a high-risk for lung resection. In these patients, pulmonary haemodynamic measurements appear to have no discriminatory value, whereas gas exchange measurements during exercise may help to identify patients with higher mortality risk.