Mixtures of long-term exposure to ambient air pollution, built environment and temperature and stroke incidence across Europe

Abstract

Introduction: The complex interplay of multiple environmental factors and cardiovascular has scarcely been studied. Within the EXPANSE project, we evaluated the association between long-term exposure to multiple environmental indices and stroke incidence across Europe.

Methods: Participants from three traditional adult cohorts (Germany, Netherlands and Sweden) and four administrative cohorts (Catalonia [region Spain], Rome [city-wide], Greece and Sweden [nationwide]) were followed until incident stroke, death, migration, loss of follow-up or study end. We estimated exposures at residential addresses from different exposure domains: air pollution (nitrogen dioxide (NO₂), particulate matter < 2.5 μm (PM_2.5), black carbon (BC), ozone), built environment (green/blue spaces, impervious surfaces) and meteorology (seasonal mean and standard deviation of temperatures). Associations between environmental exposures and stroke were estimated in single and multiple-exposure Cox proportional hazard models, and Principal Component (PC) Analyses derived prototypes for specific exposures domains. We carried out random effects meta-analyses by cohort type.

Results: In over 15 million participants, increased levels of NO₂ and BC were associated with increased higher stroke incidence in both cohort types. Increased Normalized Difference Vegetation Index (NDVI) was associated with a lower stroke incidence in both cohort types, whereas an increase in impervious surface was associated with an increase in stroke incidence. The first PC of the air pollution domain (PM_2.5, NO₂ and BC) was associated with an increase in stroke incidence. For the built environment, higher levels of NDVI and lower levels of impervious surfaces were associated with a protective effect [%change in HR per 1 unit = -2.0 (95 %CI, -5.9;2.0) and -1.1(95 %CI, -2.0; -0.3) for traditional adult and administrative cohorts, respectively]. No clear patterns were observed for distance to blue spaces or temperature parameters.

Conclusions: We observed increased HRs for stroke with exposure to PM_2.5, NO₂ and BC, lower levels of greenness and higher impervious surface in single and combined exposure models.

Keywords: Air pollution; Green space; Meta-analysis; Multiple environmental exposures; Stroke; Urban exposome.