Effect modification of pollen on the association between particulate matter air pollution and mortality: a multi-city analysis in Kyushu, Japan.

Pollen is a biogenic pollutant of growing concern due to rising temperatures under climate change. Previous studies suggest an exacerbating effect of pollen on allergenicity-related adverse health outcomes but remains inconclusive. This study aimed to examine whether pollen modifies the relationship between particulate matter air pollutant and mortality. We collected daily data on pollen, suspended particulate matter (SPM), and mortality from eight cities across Kyushu, Japan, during the pollen season (February to April) between 1989 and 2014. A two-stage time-series study was conducted, incorporating an interaction term between SPM and pollen levels to quantify effect modification by pollen in the first stage, followed by pooling the city-specific estimates in the second stage. A total of 262,288 all-cause deaths occurred during the study period. SPM was associated with respiratory mortality one day after exposure (lag 1) with a relative risk (RR) of 1.010 and corresponding 95% confidence interval (CI) of 1.000 to 1.019. Days were stratified into binary, lower or higher pollen days based on pollen concentration. A pattern of effect modification of SPM by pollen on respiratory mortality was observed. The risk of respiratory mortality associated with SPM at lag 1 during higher pollen days above the 75th percentile (RR: 1.021, 95% CI: 1.005-1.037) was larger than that during lower pollen days (RR: 1.003, 95% CI: 0.987-1.018). This pattern of effect modification remains regardless of variation in pollen day cutoffs. Targeted adaptive strategies should be developed according to specific pollen concentrations expected to elevate the health burden of air pollution during spring seasons in western Japan.