Medication Exposure and Neurodegenerative Disease Risk Across National Biobanks

AbstractAdvanced age, genetics, and environmental exposures are leading contributors to the development of neurodegenerative disorders (NDD). In this study, we used data from the UK Biobank (UKB) and the All of Us (AoU) initiative to determine if exposure to specific medications are associated with an increased or decreased risk of NDD, including Alzheimers (AD), Parkinsons disease (PD), and all-cause dementia (DEM). We investigated the associations between these diseases and prescription drug exposures through an unbiased analysis, assessing both lifetime risk and risk from exposures occurring more than ten years before diagnosis, while also accounting for comorbid conditions. MethodsCox proportional hazard models were used to evaluate both lifetime and ten-year lag-associated risks of developing a NDD following exposure to specific prescription medications. This analysis followed a two-stage design, incorporating separate discovery and replication cohorts sourced from national-scale biobanks. FindingsWe pulled data from over 700,000 health records from individuals of European ancestry to survey a total of 480 prescription medication exposures. After multiple test corrections, we found 241 significant associations between medication exposure and risk of NDDs in our discovery cohort, with 157 of these replicated in an independent dataset. After adjusting for potential comorbidities, 15 medication-NDD associations remained significant, some of which were attenuated after accounting for APOE-{varepsilon}4 status. Most of these significant pairings were associated with increased risk, however, two antibiotics, one proton pump inhibitor, and one statin had replicated effects inversely associated with disease risk. InterpretationWhile correlation does not imply causation and some associations may be driven by medications used to treat prodromal stages of disease, we have utilized large, unbiased datasets to identify and replicate associations between commonly prescribed medications and NDD risk. Additional longitudinal and mechanistic investigations are warranted.