We sought to determine the associations among thoracic aortic atherosclerosis, coronary atherosclerosis and the function of high density lipoprotein (HDL) in a case-control study. The function of HDL can be assessed by the fractional esterification rate of cholesterol in low density lipoprotein (LDL)- and very low density lipoprotein (VLDL)-depleted plasma (FER H D L ), which reflects a balance of cholesterol uptake by HDL and cholesterol ester (CE) transport in the reverse cholesterol transport (RCT) system in humans. Cases (n=51, age: 64.3+/-8.0 years) and controls (n=51, age: 58.7+/-13.1 years) were defined as subjects with/without angiographically proven coronary artery disease (CAD), respectively and examined for thoracic aortic atherosclerosis (TAA) by transesophageal echocardiography. The severity of TAA was determined by the ratio of average sclerotic areas (ASA) and average sclerotic lengths (ASL). The cases had significantly (P<0.05) higher values of ASA (0.22+/-0.18 vs. 0.10+/-0.11), ASL (0.82+/-0.56 vs. 0.48+/-0.45), ASA/ASL ratio (0.23+/-0.08 vs. 0.17+/-0.09) and FER H D L (10.3+/-3.8 vs. 8.3+/-3.5% per hour) and lower HDL-C and apolipoprotein A-I levels than the controls. A receiver operating characteristic (ROC) curve analysis showed that ASA/ASL and FER H D L had moderate discriminating ability for CAD and the diagnostic accuracy of ASA/ASL was better than that of FER H D L (area under ROC curve: 0.703 and 0.656, respectively). Multivariate logistic regression analysis indicated that ASA/ASL and FER H D L were independent indicators for CAD [odds ratio (95% CI): 7.5 (2.4-27), P<0.01 and 4.0 (1.2-15), P<0.05] after adjusting for age, gender and other conventional risk factors, and that a high FER H D L value greatly increased the relative risk of CAD associated with a high ASA/ASL. The function of HDL, as assessed by FER H D L , enhances the ability of TAA to predict CAD.