Intimal-medial thickness (IMT) of the extracranial carotid arteries measured by B-mode ultrasonography has been used as a marker of systemic and coronary atherosclerosis. Previous studies have indicated that maximum and mean carotid IMT are significantly correlated with the extent and severity of coronary artery disease (CAD), but the clinical usefulness of these markers is limited because they are neither specific nor sensitive enough to identify patients with or without significant CAD. The correlation of a new IMT marker, variance of IMT, with coronary risk factors and coronary atherosclerosis was investigated in 200 patients who underwent carotid ultrasonography and coronary angiography. IMT was measured in 16 sites of the extracranial carotid arteries for the calculation of mean, maximum and variance of IMT. Univariate analysis showed that these three indexes were significantly correlated with age, serum lipoprotein (a) and hypertension. However, age was correlated weakly with variance of IMT. There were significant gender differences in the mean and maximum IMT but not in the variance. There were also significant correlations of mean IMT with smoking, and maximum and variance of IMT with high-density lipoprotein. Multiple logistic regression analysis in 100 age and sex matched patients indicated that the only significant predictor for CAD in this subgroup was variance of IMT (odds ratio = 1.6). These results indicated that each risk factor causes different morphologic manifestations in the carotid atherosclerotic lesion. Variance of IMT, which represents the irregularity of carotid IMT, was correlated well with CAD and appears to be useful for assessing systemic and coronary atherosclerosis.