Comparison of Air Pollution-Mortality Associations Using Observed Particulate Matter Concentrations and Reanalysis...

Air pollution poses a health hazard in all countries. However, complete data on ambient particulate matter (PM) concentrations are not available in all world regions. Reanalysis data is already a valuable source of exposure data in epidemiological studies examining the relationship between temperature and health. Nevertheless, the performance of reanalysis data in assessing the short-term health effects of particulate air pollution remains unclear. We assessed the performance of CAMS reanalysis (EAC4) data from the European Centre for Medium-Range Weather Forecasts, compared with daily PM concentrations from field monitoring stations, to estimate short-term exposure to PM with an aerodynamic diameter less than 10 μm (PM₁₀) on daily mortality in 33 Spanish provincial capital cities using a two-stage time series regression design. The shape of the PM₁₀ distribution varied substantially between PM observations and CAMS global reanalysis of atmospheric composition (EAC4) reanalysis data, with correlation ranging from 0.21 to 0.58. The pooled mortality risk for a 10 μg/m³ increase in PM₁₀ showed similar estimates using PM concentrations {relative risks (RR) = 1.007, 95% confidence intervals (95% CI) = [1.002, 1.011]} and EAC4 reanalysis data (RR = 1.011, 95% CI = [1.006, 1.015]). However, the city-specific PM₁₀ beta coefficients estimated using PM concentrations and EAC4 reanalysis data showed a low correlation (r = 0.22). The use of reanalysis data should be approached with caution when assessing the association between particulate matter air pollution and health outcomes, particularly in cities with small populations.