Small and medium sized HDL particles are protectively associated with coronary calcification in a cross-sectional population-based sample

Background and aims: Failure of trials to observe benefits by elevating plasma high-density lipoprotein cholesterol (HDL-C) has raised serious doubts about HDL-C's atheroprotective properties. We aimed to identify protective HDL biomarkers by examining the association of nuclear magnetic resonance (NMR) measures of total HDL-particle (HDL-P), large HDL-particle, and small and medium-sized HDL-particle (MS-HDL-P) concentrations and average HDL-particle size with coronary artery calcification (CAC), which reflects the burden of coronary atherosclerosis, and compare with that of HDL-C. Methods: Using a cross-sectional design, 504 Jerusalem residents (274 Arabs and 230 Jews), recruited by population-based probability sampling, had HDL measured by NMR spectroscopy. CAC was determined by multidetector helical CT-scanning using Agatston scoring. Independent associations between the NMR measures and CAC (comparing scores ≥100 vs. <100) were assessed with multivariable binary logistic models. Results: Comparing tertile 3 vs. tertile 1, we observed protective associations of HDL-P (multivariable-adjusted OR 0.42, 95% CI 0.22-0.79, p_{linear trend} = 0.002) and MS-HDL-P (OR 0.36, 95% CI 0.19-0.69), p_{linear trend} = 0.006 with CAC, which persisted after further adjustment for HDL-C. HDL-C was not significantly associated with CAC (multivariable-adjusted OR 0.59, 95% CI 0.27-1.29 for tertiles 3 vs. 1, p_{linear trend} = 0.49). Large HDL-P and average particle size (which are highly correlated; r = 0.83) were not associated with CAC: large HDL-P (OR 0.77, 95% CI 0.33-1.83, p_{linear trend} = 0.29) and average HDL-P size (OR 0.72, 95% CI 0.35-1.48, p_{linear trend} = 0.58). Conclusions: MS-HDL-P represents a protective subpopulation of HDL particles. HDL-P and MS-HDL-P were more strongly associated with CAC than HDL-C. Based on the accumulating evidence, incorporation of MS-HDL-P or HDL-P into the routine prediction of CHD risk should be evaluated.