Severity of congestive heart failure is usually graded by clinical and resting hemodynamic criteria, although patients are most often symptomatic during exertion. Upright bicycle exercise to symptomatic maximum and anaerobic threshold was performed by 44 male patients with clinical class II to IV chronic congestive heart failure, and in 16 similarly aged normal male volunteer subjects. Patients achieved a maximal oxygen consumption of 13.3 ± 0.8 (SEM) ml/kg/min compared to 21.9 ± 1.2 ml/kg/min in normal subjects (p < 0.001). At maximal exercise, patients had significantly lower heart rate, mean arterial pressure and cardiac index, but at comparable submaximal workloads only cardiac index was lower in patients, whereas oxygen consumption and other hemodynamic parameters were similar. Patients were arbitrarily classified into four groups according to maximal oxygen consumption achieved (in ml/kg/min): I = > 20, II = 14-20, III = 10-13, IV = < 10. These groups were clearly separated by exercise duration but not by hemodynamic parameters at peak exercise. Clinical classification of patients (New York Heart Association criteria) agreed with the exercise classification in only 16 of 44 patients. Resting heart rate, blood pressure, cardiac index and pulmonary wedge pressure also failed to correlate with exercise capacity. Thus, maximal exercise testing is preferred to grade the symptomatic severity of congestive heart failure. © 1979.