Abstract
BACKGROUND: Accumulating evidence suggests that the incidence of cancer and dementia are inversely associated. Bias does not appear to fully account for the relationship, but causal explanations have not been adequately investigated. We thus considered a possible inverse shared genetic basis.</p>
METHODS: We constructed polygenic risk scores for cancer (PRS cancer ) and Alzheimer disease (PRS AD ) in European ancestry UK Biobank (UKB) and Health and Retirement Study (HRS) participants aged 60 years or older. Linear mixed-effects models evaluated associations of PRS cancer with cognition, and logistic regression evaluated associations of PRS AD with cancer.</p>
RESULTS: In UKB, PRS cancer was nominally associated with improved fluid intelligence ( β : 0.12, 95% CI: 0.01-0.22). Twelve variants in PRS cancer , including 7 in the human leukocyte antigen (HLA) complex, were positively associated with fluid intelligence, and 7 were inversely associated ( P <5.8×10 -5 ). PRS cancer and its contributing variants were not associated with cognitive outcomes in HRS. PRS AD was not associated with cancer risk in either study cohort.</p>
DISCUSSION: Though not conclusive, the direction of the association between PRS cancer and fluid intelligence was consistent with our a priori hypothesis that cancer risk variants would decrease cognitive decline. The association pattern with HLA-related variants suggests potential relevance of immune surveillance for the inverse association between dementia and cancer.</p>