Background Although traditional lipid parameters and coronary imaging techniques are valuable for cardiovascular disease (CVD) risk prediction, better diagnostic tests are still needed.Methods In a prospective, observational study, 795 individuals had extensive cardiometabolic profiling, including emerging biomarkers, such as apolipoprotein E–containing HDL-cholesterol (ApoE-HDL-C). Coronary artery calcium (CAC) score was assessed in the entire cohort, and quantitative coronary computed tomography angiography (CCTA) characterization of total burden, noncalcified burden (NCB), and fibrous plaque burden (FB) was performed in a subcohort (n = 300) of patients stratified by concentration of ApoE-HDL-C. Total and HDL-containing apolipoprotein C-III (ApoC-III) were also measured.Results Most patients had a clinical diagnosis of coronary artery disease (CAD) (n = 80.4% of 795), with mean age of 59 years, a majority being male (57%), and about half on statin treatment. The low ApoE-HDL-C group had more severe stenosis (11% vs. 2%, overall P < 0.001), with higher CAC as compared with high ApoE-HDL-C. On quantitative CCTA, the high ApoE-HDL-C group had lower NCB (β = –0.24, P = 0.0001), which tended to be significant in a fully adjusted model (β = –0.32, P = 0.001) and altered by ApoC-III in HDL levels. Low ApoE-HDL-C was significantly associated with LDL particle number (β = 0.31; P = 0.0001). Finally, when stratified by FB, ApoC-III in HDL showed a more robust predictive value of CAD over ApoE-HDL-C (AUC: 0.705, P = 0.0001) in a fully adjusted model.Conclusion ApoE-containing HDL-C showed a significant association with early coronary plaque characteristics and is affected by the presence of ApoC-III, indicating that low ApoE-HDL-C and high ApoC-III may be important markers of CVD severity.Trial Registration ClinicalTrials.gov: NCT01621594.Funding This work was supported by the NHLBI at the NIH Intramural Research Program.
Alexander V. Sorokin, Nidhi Patel, Khaled M. Abdelrahman, Clarence Ling, Mart Reimund, Giorgio Graziano, Maureen Sampson, Martin P. Playford, Amit K. Dey, Aarthi Reddy, Heather L. Teague, Michael Stagliano, Marcelo Amar, Marcus Y. Chen, Nehal N. Mehta, Alan T. Remaley
Results comparing logistic regression models with AUC ROCs were reported.