Health effects of fine particle sources and dimensional fractions: a 3-year time-series study in an urban area of Northern Italy

Abstract

Introduction Short-term exposure to fine particle mass has been associated with acute effects on health, but few studies examined source-specific exposure and role of size-segregated particle distribution. Within Supersite project we investigated the association between exposure to sources or fraction of particulate and health outcomes in the city of Bologna. Methods A daily series of chemical profile and fraction of particulate data has been collected for the period 2012-2014. Source apportionment methods were applied to identify specific sources of pollution. Particles number concentration (PNC) was converted into surface of particles, or collapsed into nucleation, Aitken and accumulation mode. The association of these daily data with mortality and morbidity was studied using Poisson regression models, adjusted by seasonality, meteorology, holidays, influenza epidemics and population summer decrease. Both single and mediated lags were examined. The effect estimates are expressed as percent changes in the number of cases per interquartile range increase of pollutant. Results 6 major sources were identified: traffic, industry, biomass burning, oil combustion, secondary nitrates and sulphates. Traffic and biomass burning showed significant associations with natural deaths at lags 3,4 (lag 4: 3.52%;95%CI:1.18;5.91) and lag 1,2,3 (lag 3: 3.50%;95%CI:0.82;6.25) respectively. Cardiovascular (CV) morbidity was associated only with nucleation at lag 3 (6.39%,95%CI:1.81;11.19). Respiratory (Resp) morbidity was variously associated with all sources, with stronger effects with industry (lag 0-1:4.81%,95%CI:1.50;8.23), and an association was observed also with accumulation at lag 6. Conclusions Particulate emission sources were consistently associated with respiratory diseases, while specific sources (traffic and biomass burning) were associated with natural mortality. PNC analyses suggested associations between smallest and biggest particles and CV and Resp outcomes respectively.