By analyzing expired gas with exercise testing, we examined the relationship between test results and postoperative cardiopulmonary complications. The analysis was evaluated for its accuracy and reliability in comparison with general pulmonary function tests including spirometry, flow volume curves, diffusing capacity and arterial blood gas analysis. Enrolled were 52 of patients admitted between a period from 1991 to 1993 for thoracic esophageal cancer, who underwent a radical operation of right thoracolaparotomy. Patients were divided into two groups according to postoperative cardiopulmonary complications; one with complications (group C, n = 11) and the other without complication (group NC = 41). The two groups were compared in maximum oxygen consumption (VO2 max)/m2, anaerobic threshold (AT)/m2, VC/m2, %VC, FEV1.0/m2, V25/m, V50/V25, %DLCO and PaO2. In the results, there were significant differences between the two groups only in VO2 max/m2 (p < 0.001), AT/m2 (p < 0.001), and V50/V25 (p < 0.05), while there were no significant differences between the two groups in VC/m2, %VC, FEV1.0/m2, FEV1.0%, V25/m, %DLCO or PaO2. The above results suggest that analysis of expired gas with exercise testing is able to measure easily both cardiac and pulmonary reserve forces, which conventional pulmonary function tests are unable to do, and is evaluating surgical indications for esophageal cancer patients and predicting postoperative cardiopulmonary complications.