The purpose of our study was to examine the ability of clinical and resting electrocardiographic variables to provide useful estimates of the probability of three-vessel or left-main coronary artery disease. The study group consisted of 680 patients with symptomatic coronary artery disease who underwent exercise equilibrium radionuclide angiography and coronary angiography within 6 months. Sixteen clinical and electrocardiographic variables were examined by logistic regression analysis. The independently predictive variables were then used to develop convenient graphic estimates of the probability of three-vessel or left-main disease and to classify patients into high-risk (>35%), intermediate-risk (15-35%), or low-risk (<15%) groups. Five variables were independently predictive of left-main or three-vessel disease: age, typical angina, diabetes, gender, and both history and electrocardiographic evidence of a prior myocardial infarction. A single graph was constructed that displayed the probability of severe coronary artery disease as a function of a five-point cardiac risk scale, which incorporated these variables. Two hundred sixty-two patients (39% of the study group) were classified as high risk; 127 of these patients (48%) had three-vessel or left-main disease. An additional 96 patients were classified as low risk; nine of these patients (9%) had three-vessel or left-main disease. Five clinical variables that were obtained on an initial patient assessment can provide useful estimates of the likelihood of severe coronary disease.
(Arch Intern Med. 1992;152:309-312)