OBJECTIVE: Peak oxygen uptake (VO2) is a powerful prognostic index, but maximal exercise testing in heart transplant candidates has a number of disadvantages. It is unknown whether it is possible to predict peak VO2 from a comprehensive dataset with parameters of heart and lung function at rest. METHODS: One hundred adult patients in sinus rhythm and with either idiopathic or ischaemic heart failure performed a graded cycle ergometer test until volitional fatigue and underwent radionuclide ventriculography, heart catheterization, and lung function measurements at rest. RESULTS: Weight, height, age, gender and aetiology of heart failure explained 48% of the variance of peak VO2. On top of these anthropometric, demographic and clinical patient characteristics, 12% of the variance of peak VO2 was additionally explained by all resting measurements combined, i.e. radionuclide left ventricular ejection fraction, peak ejection rate, peak filling rate, cardiac frequency, mean right atrial pressure, pulmonary capillary wedge pressure, pulmonary artery pressures, cardiac output, forced vital capacity, forced expiratory volume in one second, and pulmonary diffusing capacity (cumulative R2 = 0.60); among these, pulmonary vascular resistance was the most important predictor (+6%; P < 0.001). Analyses in a subset of 43 male patients pointed out that systemic pressures and vascular resistance were not related to peak VO2. CONCLUSION: On the basis of resting left ventricular function, haemodynamics, and routine pulmonary measurements, it is unlikely to accurately predict exercise tolerance in the majority of heart transplant candidates, i.e. patients with either idiopathic or ischaemic heart failure and able to exercise until exhaustion.