To determine the predictive accuracy of fluoroscopically detected coronary artery calcification (CAC) and a positive submaximal exercise test, 129 asymptomatic men were screened; 13 had both coronary artery calcification and positive exercise test (greater than or equal to 1.0 mm ST-segment depression). These 13 men were studied at coronary arteriography. They had a mean age of 44 years (range 41-56 years); none had history or symptoms of heat disease and all had normal resting ECGs at entry. CAC was detected in one artery in 10 men, in two arteries in two men, and in three arteries in one man. Coronary artery disease (CAD) was considered clinically significant if any major coronary branch was narrowed > 50%. Coronary arteriography revealed 12 men with clinically significant CAD (one-vessel CAD in four, two-vessel CAD in five and three-vessel CAD in three men) and one man with minor one-vessel CAD. The predictive accuracy was 100% for minor CAD and 92% for clinically significant CAD. The location of CAC and CAD correlated, but the absence of CAC did not rule out the presence of CAD at coronary arteriography. Furthermore, CAC did not indicate the location of the highest stenotic (most occlusive) lesions seen at arteriography. Follow-up for the 13 patients was 36 months; three patients developed typical angina and one patient developed a transmural myocardial infarction. This study suggests that the predictive accuracy of CAC and a positive exercise test in the middle-aged non-hyperlipidemic asymptomatic male is very high (100% for CAD and 92% for clinically significant CAD) and that CAC and a positive exercise test predict an early appearance of angina or myocardial infarction in previously asymptomatic men.