Association of DNA Methylation-Derived C-Reactive Protein Predictors With All-Cause Mortality and Blood Lipids in Adults Aged ≥ 50 Years in the United States.

Abstract

Chronic low-grade inflammation contributes to aging, metabolic dysfunction, and age-related diseases. High-sensitivity C-reactive protein (HsCRP) is widely used but limited by short-term variability. DNA methylation-based CRPMort derived from GrimAge2 may provide a more stable measure of chronic inflammatory burden, yet its predictive value for mortality and association with lipid metabolism remain unclear. Associations of CRPMort and HsCRP with all-cause and cardiovascular mortality and lipid traits were evaluated using weighted cox regression and linear regression, Kaplan-Meier curves, restricted cubic splines (RCS), and time-dependent receiver operating characteristic curve (timeROC) analyses within the NHANES database, with subgroup, mediation, sensitivity, and Mendelian randomization analyses. Higher CRPMort was significantly associated with all-cause mortality (hazard ratio [95% confidence interval]: 1.65 [1.10, 2.46], p = 0.015) and showed near-significant positive associations with triglycerides (p = 0.051) and residual cholesterol (p = 0.056). RCS analyses demonstrated a linear positive relationship between CRPMort and both all-cause mortality and lipid traits, including triglycerides and residual cholesterol. HsCRP showed no significant associations with mortality or lipid traits (all p > 0.05). TimeROC curves revealed that compared to HsCRP, CRPMort had more superior long-term predictive performance and partially mediated the effect of chronological age on all-cause mortality. Importantly, concurrent elevations of CRPMort and HsCRP were associated with the highest risks of all-cause mortality and dyslipidemia. GrimAge2-derived CRPMort is a robust predictor of long-term all-cause mortality and may capture chronic inflammation linked to triglyceride-rich lipoproteins beyond HsCRP. Combined assessment of both inflammatory markers may enhance risk stratification and inform aging-related cardiometabolic research.