0218 Sleep and Circadian Polygenic Risk Scores Predict Actigraphy-derived Traits and Cognition in Older Adults

Abstract Introduction Sleep disturbances are prevalent in older adults and have been associated with major pathophysiological characteristics of Alzheimer’s disease. Genome-wide association studies (GWAS) of sleep and circadian traits identified hundreds of genetic loci influencing variability in these phenotypes. GWAS summary statistics can be used to derive polygenic risks scores (PRS), a quantitative measure of genetic risk for a particular phenotype. In this study, we assessed whether sleep and circadian PRS can predict actigraphy-derived measures and cognition scores among older adults followed at the University of Kansas Alzheimer’s Disease Research Center. Methods We collected 7-days of actigraphy and cognitive assessments of verbal memory, attention, and executive function among participants of European ancestry with genome-wide genotyping data. Linkage disequilibrium reference panels and GWAS summary statistics for insomnia, chronotype, and sleep duration were used to calculate PRS using a Bayesian regression approach with continuous shrinkage priors (PRS-CS). We used Spearman correlations and general linear models adjusted for age, sex, and education to test for associations of PRS with actigraphy-derived traits (sleep duration, sleep efficiency, wake after sleep onset [WASO], number of awakenings, and sleep midpoint) and cognitive factors. Results The final dataset consisted of 61 participants (mean [standard deviation (SD)] age 74.1 [6.5] years; 39.3% female; mean 16 [3] years of education). No associations were observed between early chronotype PRS and early sleep midpoint, nor sleep duration PRS and objective sleep duration. However, early chronotype PRS was negatively correlated with WASO (ρ=-0.32; p=0.012) and positively correlated sleep efficiency (ρ=0.30; p=0.018). Adjusted analyses suggest that a one SD increase in early chronotype PRS was associated with lower WASO (β[95%CI]=-15.0[-29.3;-0.76] minutes; p=0.043) and higher sleep efficiency (2.9[0.6;5.1] %; p=0.014). Furthermore, one SD increase in sleep duration PRS was associated with better performance on attention (2.9 [0.6;5.1] SD; p=0.009). No significant associations with insomnia PRS were observed. Conclusion This study identified novel associations between early chronotype and sleep duration PRS and relevant actigraphy-derived traits and attention in a clinical cohort of older adults. Results may inform the potential role of genetically informed sleep traits towards understanding their impact on cognition. Support (if any) NIA P30AG035982; General Faculty Research Fund (University of Kansas)