The effects of captopril and placebo were compared in 18 patients with chronic heart failure and angina pectoris with use of a double-blind crossover trial design. Symptoms were assessed by patient treatment preference, visual analogue scores and nitroglycerin consumption. Exercise performance was assessed using two different treadmill protocols of different work intensity with simultaneous measurement of oxygen consumption and by supine bicycle exercise and simultaneous radionuclide ventriculography. Arrhythmias were assessed by 48 h ambulatory electrocardiographic monitoring. Patients generally preferred placebo to captopril, and this appeared to be due to an increase in symptoms of angina with captopril. Treadmill exercise time on a high intensity protocol was shorter with captopril than with placebo; on a low intensity protocol, angina became a more frequent limiting symptom even though overall exercise performance was not changed. The heart rate-blood pressure product was reduced, but largely because of a reduction in blood pressure rather than in heart rate. During supine bicycle exercise, no differences in symptoms, exercise performance, ejection fraction or changes in blood pressure were noted and ventricular arrhythmias were reduced. Captopril does not appear to be clinically useful in alleviating angina pectoris in patients with heart failure, and this effect may be related to a decrease in coronary perfusion pressure. Nonetheless, desirable metabolic effects, a reduction in arrhythmias and potential effects on survival require further study of captopril in patients with both angina and heart failure.