Abstract P012: Three Biological Age Estimates May Capture Distinct Aspects of Aging

Background: Biological age ( BA ) may reflect an individual’s aging process better than chronological age ( CA ). The objective of our study was to construct BA measures from different types of biomarkers and test their associations with mortality and age-related disease in a community-based sample. Methods: We selected 6 clinical predictors that capture pulmonary, vascular, atherosclerosis, insulin sensitivity, inflammatory, and kidney domains of aging, and 9 inflammatory biomarkers measured in Framingham Heart Study Offspring cohort participants at exams 7 (1998-2001, N=3539, mean age 62±10) and 8 (2005-2008, N=3021, mean age 67±9). We used the Klemera-Doubal method to calculate a clinical variable BA and an inflammatory marker BA. We computed BA using DNA methylation (DNAm) data at exam 8 using Horvath’s method. For each of the measures we computed the difference ( ΔAge ) between BA and CA and modeled the effects of ΔAge after accounting for CA. We followed participants through 2014 for all-cause mortality (N=713), cardiovascular disease (CVD, N=412), coronary heart disease (CHD, N=223), stroke (N=129), and cancer (N=509). Results: Inflammatory and clinical ΔAge were correlated (=0.35, =0.33, for exams 7 and 8 respectively), and also across exams (inflammatory ΔAge exam 7 vs 8: =0.61; clinical ΔAge: =0.76). DNAm ΔAge was not significantly correlated with exam 8 inflammatory or clinical ΔAge. After adjusting for CA and sex, larger inflammatory and clinical ΔAge, corresponding to older BA than CA, was associated with significantly increased hazards of all-cause mortality, CVD, and CHD (Table). The clinical and inflammatory ΔAge were significant (p<0.05) in models containing both measures. DNAm ΔAge was associated with increased hazards of all-cause mortality, CVD, cancer, and stroke, and remained significant in a model for mortality that also included inflammatory ΔAge (p<0.05). Conclusions: Our findings suggest the three BA measures may be complementary in predicting risk for age-related disease.