OBJECTIVES: to estimate the short-term impact of air pollutants on adult population of 23 Italian cities on the calendar period 2006-2009 as part of the EpiAir2 project. DESIGN, MATERIALS AND METHODS: for each city, the short-term impact of air pollution on mortality was estimated. In particular, it was calculated the number of deaths attributable to particulate matter levels (PM10 and PM2.5) exceeding different thresholds deriving from the European Union legislation and the World Health Organization guidelines (PM10: 20 and 40 μg/m3, reduction of 20% up to 20 μg/m3 and more than 35 days per year with daily average above 50 μg/m3; PM2.5: 10, 18 and 25 μg/m3, reduction of 20% up to 18 μg/m3). The impact evaluation was done combining the effect estimates of air pollutant, the observed mortality level and the air pollution concentration measured by the ambient monitors. Regarding effects, the posterior city-specific distributions arising from a Bayesian metanalysis was employed. Uncertainty around the impact estimates was obtained by Monte Carlo methods. RESULTS: overall, in the 23 cities considered in the present study attributable deaths to short-term effects of PM10 concentrations above 20 μg/m3 as annual average and of PM 2.5 concentrations above 10 μg/m3 on the period 2006-2009 was respectively 0.9% (assuming independence among cities the 80% credibility interval is 0.4-1.4) and 0.8% (80%CrI 0.2-1.3) of natural mortality. The impact was larger for the cities of the river Po valley, the Florence area and the large conurbations of Rome, Naples and Palermo: considering PM 10 1.0% (80%CrI 0.4-1.5) vs. 0.4% (80%CrI 0.2-0.7) of natural deaths among the other cities in the study. Pollution reduction actions as prescribed by the European Union legislation, i.e. a 20% reduction up to 20 μg/m 3 for PM10 and up to 18 μg/m3 for PM 2.5, would have saved, over all the cities covered by this study, 42% and 51% of all attributable deaths, respectively. CONCLUSIONS: during the study period, air pollution exposure was responsible for a relevant impact on mortality in the enrolled cities. Policies based on percent reduction of PM 10 and PM2.5 as prescribed by the European Union could substantially reduce the short term impact on population mortality.

cited By 6