A consecutive series of 198 patients (148 men and 50 women, mean age 51 years, range 18 to 76) with pure, isolated, severe aortic regurgitation was retrospectively studied to determine the prevalence of angiographically significant coronary artery disease (CAD) and its relation to angina pectoris and coronary risk factors. Significant CAD (coronary diameter stenoses greater than 50%) was found in 28 patients (14%). Typical angina was present in 18% and atypical chest pain in 16%. Angina alone had a sensitivity of 57% to detect significant CAD. The predictive accuracy of a positive history of angina was 46% and that of a negative test 93%. By using multivariate logistic regression, a risk score could be calculated that increased the sensitivity to 74% at equal specificity. Almost 40% of the total population had a risk score of less than -2.9 (only 1 patient in this group had CAD). It is concluded that coronary arteriography can safely be omitted in many patients with severe aortic regurgitation if they have no symptoms of myocardial ischemia or risk factors known to increase its incidence.