Lipoprotein subclass levels may improve the prediction of cardiovascular disease (CAD) in individuals beyond the risk assessment provided by conventional enzymatically determined lipid levels. The objective of this study was to evaluate the associations between nuclear magnetic resonance (NMR) spectroscopy-determined lipoprotein subclasses and coronary calcification in postmenopausal women, and to determine whether the associations were independent of conventional lipid measures. Coronary artery calcification (CAC) was measured by electron beam computed tomography, and lipoprotein subclasses were determined by NMR spectroscopy (Liposcience, Inc., Raleigh, NC), in 286 healthy women (mean age = 61.7), at 8 years postmenopause. CAC was analyzed as categories 0, 1 to 99, and > or =100. The mean CAC was 53 (range, 0 to 1,175), and 54% of the women had 0 scores. Large high-density lipoprotein (HDL) was inversely associated with CAC category, but small HDL was not. All very low-density lipoprotein (VLDL) subclasses-small, medium, and large-were positively associated with CAC (p <0.01). Small low-density lipoprotein (LDL) was positively associated with CAC (p <0.01), but medium and large LDL were not. Smaller LDL particle size (p <0.01) and higher levels of LDL particles (p <0.001) were associated with higher CAC category. In separate ordinal logistic regression models, small LDL, LDL particles, and large VLDL were each positively associated (p <0.05) with higher CAC after adjustment for age, systolic blood pressure (SBP), current smoking, and conventional measures of LDL cholesterol, HDL cholesterol, and triglycerides. These results suggest that the measurement of lipoprotein subclasses may improve the prediction of CAD in postmenopausal women beyond that provided by the conventional lipid panel and CAD risk factors.