OBJECTIVE Previous studies reported that many patients are at high risk for cardiovascular disease (CVD) despite achieving recommended low-density lipoprotein cholesterol (LDL-C) levels. Therefore, we investigated whether the association between LDL-C and the risk for incident CVD differed according to high-density lipoprotein cholesterol (HDL-C) levels using coronary artery calcium score (CACS) progression as a surrogate marker for predicting CVD. METHODS We investigated 2132 Korean men in a health screening program, in which CACS was measured at baseline and after 4 years. Coronary artery calcification (CAC) progression was defined as a change in CACS ≥0 over 4 years. We divided the subjects into nine groups according to baseline HDL-C and LDL-C levels and compared their risks for CAC progression. RESULTS After 4 years, 475 subjects (22.3%) exhibited CAC progression. We identified a positive relationship between baseline LDL-C levels and the risk for incident CAC. However, this association was attenuated by high baseline HDL-C levels. Multivariate logistic regression analysis adjusted for age, body mass index, systolic blood pressure, fasting glucose, smoking, and exercise status revealed that the odds ratios for incident CAC in the lowest HDL-C tertile were 3.08 for LDL-C tertile 3 and 2.02 for LDL-C tertile 2 compared to LDL-C tertile 1. However, these differences disappeared in the highest HDL-C tertile (HDL-C ≥54.0 mg/dL). CONCLUSIONS In this longitudinal study, we found that the positive relationship between LDL-C and the relative risk for incident CAC was attenuated by higher HDL-C levels. Therefore, HDL-C levels should be considered when estimating CVD risk.