Air pollution and mortality in twenty-five Italian cities: results of the EpiAir2 Project

Objectives: this study aims at presenting the results from the Italian EpiaAir2 Project on the short-term effects of air pollution on adult population (35+ years old) in 25 Italian cities. Design: the short-term effects of air pollution on resident people died in their city were analysed adopting the time series approach. The association between increases in 10µg/m(3) in PM10, PM2.5, NO2 and O3 air concentration and natural, cardiac, cerebrovascular and respiratory mortality was studied. City-specific Poisson models were fitted to estimate the association of daily concentrations of pollutants with daily counts of deaths. The analysis took into account temporal and meteorological factors to control for potential confounding effect. Pooled estimates have been derived from random effects meta-analysis, evaluating the presence of heterogeneity in the city specific results. Setting and participants: it was analysed 422,723 deaths in the 25 cities of the project among people aged 35 years or more, resident in each city during the period 2006-2010. Main outcome measures: daily counts of natural, cardiac, cerebrovascular, and respiratory mortality, obtained from the registries of each city. Demographic information were obtained by record linkage procedure with the civil registry of each city. Results: mean number of deaths for natural causes ranged from 513 in Rovigo to 20,959 in Rome. About 25% of deaths are due to cardiac diseases, 10% to cerebrovascular diseases, and 7% to respiratory diseases. It was found an immediate effect of PM10 on natural mortality (0.51%; 95%CI 0.16-0.86; lag 0-1). More relevant and prolonged effects (lag 0-5) have been found for PM2.5 (0.78%; 95%CI 0.12-1.46) and NO2 (1.10%; 95%CI 0.63-1.58). Increases in cardiac mortality are associated with PM10 (0.93%; 95%CI 0.16-1.70) and PM2.5 (1.25%; 95%CI 0.17-2.34), while for respiratory mortality exposure to NO2 has an important role (1.67%; 95%CI 0.23-3.13; lag 2-5), as well as PM10 (1.41%; 95%CI - 0.23;+3.08). Results are strongly homogeneous among cities, except for respiratory mortality. No effect has been found for cerebrovascular mortality and weak evidence of association has been observed between ozone and mortality. Conclusions: a clear increase in mortality associated to air pollutants was observed. More important are the effects of NO2 (on natural mortality), mostly associated with traffic emissions, and of PM2.5 (on cardiac and respiratory mortality). Nitrogen dioxide shows an independent effect from the particulate matter, as observed in the bi-pollutant models.