Comparison of air pollution exposure assessment methods and the association with children's respiratory health

Abstract

INTRODUCTION: Epidemiological studies of the associations of long-term exposure to outdoor air pollution with asthma onset and lung function in children have used different exposure assessment methods. Little is known about how these different methods affect the magnitude of the effect estimates. The aim of this study was to compare associations of long-term air pollution exposures, estimated with different exposure assessment methods, with asthma incidence and lung function. METHODS: Eight exposure assessment methods, differing in modelling (dispersion, empirical) and monitoring strategy (fixed site, mobile), were applied to estimate annual average air pollution levels at the residential addresses of 3,687 participants of the Dutch PIAMA birth cohort. Associations of air pollution exposure with asthma and lung function were assessed and compared between methods. Heterogeneity in the associations was assessed with meta-analyses. RESULTS: Estimated exposure levels and contrasts differed substantially between methods. Exposure estimates from the different methods were moderately to highly correlated, with Pearson correlations ranging from 0.5 to 0.9. Higher air pollution levels were consistently associated with higher asthma incidence and lower FEV1. However, the magnitude of the association differed between methods (e.g. the ORs (95 % CI) for asthma incidence ranged from 1.09 (0.99; 1.21) to 2.56 (1.50; 4.36) for BC per 1 µg/m3 increment). CONCLUSION: Different air pollution exposure assessment methods resulted in consistent conclusions about the presence and direction of associations with asthma incidence and lung function in children, but associations differed in magnitude. Differences in exposure assessment methods may partially drive heterogeneity in associations between different studies.