To elucidate the potential harmful effects of sodium nitroprusside (NP) on myocardial ischaemia, eleven patients with coronary artery disease were studied before and during two doses NP infusion. The infusion rates were adjusted to reduce systolic aortic pressure by 10-20 mmHg and to between 100 and 110 mmHg. Myocardial haemodynamic and metabolic measurements were undertaken in sinus rhythm and during pacing-induced angina pectoris. The procedure was repeated at the low and high dose NP using the same pacing rate. One patient did not develop ischaemia and was excluded. Low dose NP reduced ischaemia slightly, whereas the higher dose had no effect despite falling indices of cardiac work. The oxygen uptake was maintained except at high dose treatment during pacing due to an increased catecholamine stimulation. The transmural perfusion gradient was unchanged at low dose NP, but fell at the higher dose suggesting reduced perfusion of ischaemic tissue. High dose NP also decreased coronary arteriolar resistance considerably which may have induced a coronary 'steal' effect. Thus, NP may induce adverse haemodynamic and metabolic changes which counteract the beneficial actions of the drug on ischaemia when used in high doses and in patients without heart failure.