Hypertension and Stroke as Mediators of Air Pollution Exposure and Incident Dementia

Abstract

Importance: Fine particulate matter air pollution (PM2.5) has been consistently associated with cardiovascular disease, which, in turn, is associated with an increased risk of dementia. As such, vascular dysfunction might be a mechanism by which PM2.5 mediates dementia risk, yet few prior epidemiological studies have examined this potential mechanism.

Objective: To investigate whether hypertension and stroke serve as mediators and modifiers of the association of PM2.5 with incident dementia.

Design, setting, and participants: As part of the Environmental Predictors of Cognitive Health and Aging (EPOCH) Project, this cohort study used biennial survey data collected between 1998 and 2016 from respondents of the Health and Retirement Study (HRS), a nationally representative, population-based, cohort in the US. Eligible participants were those over 50 years of age who were free of dementia at baseline and had complete exposure, mediator, outcome, and demographic data from the HRS. Data analysis was conducted from August to November 2022.

Exposures: Exposure to PM2.5, calculated for the 10 years preceding each person's baseline examination according to residential histories and spatiotemporal models.

Main outcomes and measures: Incident dementia was identified using a validated algorithm based on cognitive testing and informant reports. The 4-way decomposition causal mediation analysis method was used to quantify the degree to which hypertension and stroke mediated or modified the association of PM2.5 with incident dementia after adjustment for individual-level and area-level covariates.

Results: Among 27 857 participants (mean [SD] age at baseline, 61 [10] years; 15 747 female participants [56.5%]; 19 249 non-Hispanic White participants [69.1%]), 4105 (14.7%) developed dementia during the follow-up period (mean [SD], 10.2 [5.6] years). Among participants with dementia, 2204 (53.7%) had a history of hypertension at baseline and 386 (9.4%) received a diagnosis of hypertension during the follow up. A total of 378 participants (9.2%) had a history of stroke at baseline and 673 (16.4%) developed stroke over the follow-up period. The IQR of baseline PM2.5 concentrations was 10.9 to 14.9 μg/m3. In fully adjusted models, higher levels of PM2.5 (per IQR) were not associated with increased risk of incident dementia (HR, 1.04; 95% CI, 0.98 to 1.11). Although there were positive associations of prevalent stroke (HR, 1.67; 95% CI, 1.48 to 1.88) and hypertension (HR, 1.15; 95% CI, 1.08 to 1.23) with incident dementia compared with those free of stroke and hypertension during follow-up, there was no statistically significant association of PM2.5 with stroke (odds ratio per IQR increment in PM2.5, 1.08; 95%CI, 0.91 to 1.29) and no evidence of an association of PM2.5 with hypertension (odds ratio per IQR increment in PM2.5, 0.99; 95%CI, 0.92 to 1.07). Concordantly, there was no evidence that hypertension or stroke acted as mediators or modifiers of the association of PM2.5 with incident dementia. Although the nonmediated interaction between PM2.5 and hypertension accounted for 39.2% of the total excess association (95% CI, -138.5% to 216.9%), the findings were not statistically significant.

Conclusions and relevance: These findings suggest that although hypertension may enhance the susceptibility of individuals to air pollution, hypertension and stroke do not significantly mediate or modify the association of PM2.5 with dementia, indicating the need to investigate other pathways and potential mediators of risk.

Figure 1.. Conceptual Model for the Analysis of Stroke or Hypertension Mediating the Association of PM_2.5 With Dementia Association by Vascular Conditions

The figure shows how hypertension or stroke (stroke used as the example) could serve as mediators of the association of PM_2.5 with dementia. C is the variable representing potential exposure-mediator, exposure-outcome, and mediator-outcome confounders. CDA indicates controlled direct association; INTref, reference interaction; INTmed, mediated interaction; and PIA, pure indirect association.

Figure 2.. Illustration of the 4-Way Decomposition of Total Association

The controlled direct association is due to neither mediation nor interaction. The reference interaction is only due to interaction. The mediated interaction is due to both mediation and interaction. The pure indirect association is only due to mediation.

Figure 3.. Illustration of the Association of PM_2.5 Exposure With Dementia and Potential Mediators

Odds ratios (ORs) and 95% CIs were used to analyze the association of PM_2.5 exposure with potential mediators (left side) using logistic regression, and hazard ratios (HRs) and 95% CIs were used to analyze the association of potential mediators with dementia (right side) using proportional hazard regression. The models were assessing the association of an IQR difference (10.9-14.9 μg/m³) in PM_2.5 with vascular conditions and were adjusted for age, interview date, sex, race and ethnicity, educational attainment, ownership of the primary residence, quartile of the total household wealth, urbanicity levels, neighborhood social economic status, and spatial basis functions (10 df), whereas the models assessing the association of vascular conditions with dementia were also adjusted for PM_2.5.