Acceleration of DNA methylation (epigenetic) clocks especially GrimAge2 can predict incident acute coronary syndrome in older adults

Background/Introduction Despite advances in atherosclerotic cardiovascular disease (ASCVD) risk assessments by standardised risk calculators for primary prevention, cardiovascular events and complications remain difficult to predict, particularly in old age and for secondary prevention. DNA methylation (DNAm) age calculated from epigenetic clocks strongly correlates with chronological age but whether epigenetic age acceleration (EAA) has relevance to acute cardiovascular events has been unexplored. Purpose To determine the capacity of EAA to predict incident acute cardiovascular events. Methods Community-living older adults aged >/=60y were profiled for blood DNAm using Infinium MethylationEPIC v1 arrays (Illumina). Five epigenetic clocks including Horvath, Hannum, PhenoAge, GrimAge and GrimAge2 were analysed. EAA was calculated from regression residuals where positive and negative values indicated acceleration and deceleration of biological age, respectively. ASCVD and comorbidities were documented at baseline. Prospective outcomes of mortality and hospitalisation were collected. 1H-nuclear magnetic resonance spectroscopy (Nightingale) was used to quantify blood concentrations of glycosylated acetyls (GlycA) as a composite marker of inflammation. p<0.05 was considered significant. The study was approved by institutional ethics committee and all participants provided written informed consent. Results DNAm at baseline was analysed in 535 older adults (median age 78y, IQR 69–85, range 60–99). DNAm age independently calculated from five epigenetic clocks were all positively correlated with chronological age and frailty (all p<0.0001), with no sex differentiated effects. Of five EAA estimates, only PhenoAge (PhenoAA), GrimAge (GrimAA) and GrimAge2 (Grim2AA) were associated with GlycA (p<0.01, p<0.0001 and p<0.0001, respectively). By logistic regression, only GrimAA and Grim2AA remained significantly associated with baseline ASCVD risk factors and comorbidities (diabetes mellitus, p<0.05 to p<0.01; angina pectoris, both p<0.05; history of ischaemic heart disease, both p<0.05; history of myocardial infarction, both p<0.05). During a median follow-up of 4.68y (IQR 3.91–6.44), 30 individuals were lost to follow-up, 55 (10.9%) died, while 178 (35.2%) and 101 (20.0%) experienced first all-cause and cardiovascular hospitalisations, respectively. In all analyses performed, Grim2AA was superior to GrimAA. Of note, the median Grim2AA in persons hospitalised for ASCVD predominantly secondary to ACS (not including stroke) was +0.73y (IQR -1.36 to +4.39, n=42). For every 1-year increase in Grim2AA, the hazard of future cardiovascular hospitalisation was increased by 5.7–6.1% (all p<0.05), and the hazard of ASCVD hospitalisation predominantly secondary to acute coronary syndrome (ACS) was increased by 11% (p<0.01). Conclusion(s) GrimAge2 can predict incident cardiovascular and ASCVD hospitalisation predominantly secondary to ACS.