Assessing Exposure-Response Relationships between Air Pollutants and Alzheimer’s Disease and Dementia Incidence in the U.S. Medicare Population

Abstract Introduction: emerging evidence suggests that air pollution is a modifiable risk factor for Alzheimer’s disease (AD) and related dementia (ADRD). However, uncertainty remains regarding the shape and magnitude of exposure-response relationships across pollutants, and how to control best for confounding. Methods: we examined exposure-response relationships between long-term exposure to PM2.5, NO2, and warm-season O3 and the incidence of AD and dementia using Medicare data (2000–2016). Outcomes were ascertained from Medicare claims, and exposure estimates from high-resolution spatiotemporal models were assigned based on residential ZIP code. We used a conventional decile-based categorical model via quasi-Poisson regression with adjustment for a number of confounders. This was followed by two alternative models, on both the relative and absolute scales: regression adjustment and inverse probability weighting (IPW), to both flexibly characterize exposure-response relationships and improve control for confounding through the use of pseudopopulations. Results: we observed differences in effect size and trend across three approaches, with the largest divergence seen for NO2 and O3. For PM2.5, all models indicated a monotonic exposure-response relationship with both AD and dementia, but the alternative models estimated stronger effects (steeper slopes) for AD than the conventional model, whereas the opposite was observed for dementia. For NO2, the conventional quasi-Poisson model suggested a stronger nonlinear association with AD, whereas the regression adjustment and IPW approaches yielded more attenuated estimates. For dementia, all NO2 modeling approaches suggested small positive associations. For O3, there was little evidence of exposure-response relationships across all methods for either AD or dementia. Conclusion: these findings reinforce the evidence that long-term PM2.5─and, to a lesser extent, NO2─are associated with increased ADRD risk, even at concentrations below current regulatory standards. By characterizing potentially nonlinear exposure–response relationships on both relative and absolute scales and incorporating counterfactual estimation approaches, this study provides a more comprehensive assessment of the potential impact of air pollution on AD and dementia incidence.