Coronary arteriography was performed on 18 asymptomatic, apparently healthy adults with elevated coronary calcium scores. To extend the range of observation to subjects with low calcium scores, arteriograms from 18 patients with exertional dyspnea and/or valvular heart disease and low calcium scores were also analyzed; these 18 patients were considered asymptomatic from the point of view of coronary artery disease (CAD). For the comparison of symptomatic and asymptomatic persons, 3 age and sex-matched symptomatic patients were also selected for each of the original 18 asymptomatic subjects. Arteriograms were analyzed by computer-assisted quantitative coronary arteriography at a remote site without knowledge of the calcium score or any other patient characteristics. In the 18 asymptomatic subjects with elevated calcium scores, the mean calcium score was 573 +/- 504 (Agatston method) and the mean worst stenosis was 45% +/- 16%. For all 36 patients without symptoms of CAD, worst stenosis was closely correlated with the square root of the calcium score (r = 0.85, p <0.0001). Patients with symptomatic coronary disease and calcium scores < 1,000 had stenoses more severe than asymptomatic persons with similar calcium scores. Most asymptomatic adults with elevated calcium scores have nontrivial, nonobstructive CAD or preclinical obstructive CAD, and the relation between coronary calcium score and severity of stenosis is highly significant. These data indicate that electron beam tomography can be used to estimate the severity of CAD in asymptomatic persons.