Abstract P270: Prediction of Telomere Shortening Using a Genetic Risk Score: The Multi-ethnic Study of Atherosclerosis (mesa)

In epidemiological studies, telomere(definition) shortening has been associated with age-related diseases and mortality. We assessed whether a genetic risk score (GRS) combining seven telomere length (TL)-associated genetic variants identified in a European-American genome-wide association study (GWAS) predicts TL and change in TL over 10 years in a multi-ethnic sample. Research Hypotheses: 1. In cross-sectional analysis, the GRS will be inversely associated with TL. 2. In longitudinal analysis, the GRS will be associated with change in TL over a 10-year period. Methods: TL was assessed on a sample of 1198 white, African-American and Hispanic participants, who attended both, MESA Exam 1 and 5, 10 years apart. Participants were 45-84 years at Exam 1. We used MESA measured or imputed genotype data to create the TL GRS and constructed the GRS using weighted estimates from the prior GWAS. Our main exposure was TL GRS. For hypothesis 1, we used linear mixed-effects model (LMEM) including all valid observations from both exams and adjusted for an exam indicator, age, sex and specific genetic principal components (PC). For hypothesis 2, we used LMEM with time-varying covariates and adjusted for follow-up time since Exam 1, age, sex and PC. We stratified our analysis by race/ethnic group and used the top three race/ethnic-specific genetic PC to account for population stratification. Results: In cross-sectional analysis, we observed a significant inverse association between GRS and TL (estimates= -0.18 (S.E.=0.08), P =0.02 for white; -0.18 (0.07), P <0.01 for black; and -0.13 (0.05), P =0.01 for Hispanic) after full adjustment. In the longitudinal analysis, only among Hispanics, higher TL GRS predicted a reduction in TL over a 10-year period (-0.22 (0.06, P <0.001)), but after adjustment for PC, the association became not statistically significant. Conclusion: In the cross-sectional analysis, higher TL GRS was associated with shorter TL in the three race/ethnic groups studied. TL GRS did not predict telomere attrition in any of the race/ethnic groups.