Polygenic risk scores for Alzheimer's disease in relation to cognitive change: A representative sample from the general population followed over 16 years.

Polygenic risk scores for Alzheimer's disease (AD-PRSs) have been associated with cognition. However, few studies have examined the effect of AD-PRS beyond the APOE gene, and the influence of genetic variants related to level of cognitive ability (COG-PRS) on cognitive performance over time in the general older population. A population-based sample of 965 individuals born in 1930, with genetic and standardized cognitive data on six psychometric tests (Thurstone's picture memory, immediate recall of 10 words, Block design, word fluency, figure identification, delayed recall of 12 items), were examined at age 70, 75, 79, and 85 years. Non-APOE AD-PRSs and COG-PRSs (P < 5e−8, P < 1e−5, P < 1e−3, P < 1e−1) were generated from recent genome-wide association studies. Linear mixed effect models with random intercepts and slope were used to analyze the effect of APOE ε4 allele, AD-PRSs, and COG-PRSs, on cognitive performance and rate of change. Analyses were repeated in samples excluding dementia. APOE ε4 and AD-PRS predicted change in cognitive performance (APOE ε4*age: β = −0.03, P < 0.0001 and AD-PRS *age: β = −0.01, P = 0.02). The results remained similar in the sample excluding those with dementia. COG-PRS predicted level of cognitive performance, while APOE ε4 and AD-PRS did not. COG-PRSs did not predict change in cognitive performance. We found that genetic predisposition of AD predicted cognitive decline among 70-year-olds followed over 16 years, regardless of dementia status, while polygenic risk for general cognitive performance did not.