P I – 1–4 Source-related components of pm2.5 and long-term health effects: epidemiological findings of supersite project in italy

Background/aim Supersite project collected detailed measurements on chemicals, physical and toxicological parameters of atmospheric aerosols in Emilia-Romagna region (Northern Italy) in 2012–2014. Aim of the present study was the assessment of long-term effects of PM2.5 and its sources on natural, respiratory and cardiovascular mortality in an open residential cohort. Methods We built a cohort (age ≥29 years) in 47 municipalities for 2001–2010. We collected mortality data for 2009–2013. 3 year measurements of daily concentration and constituents of PM2.5 were made at 4 sites and were analysed through a Source Apportionment approach identifying 6 sources. For each source, the annual percentage contribution in each site was estimated, and these data were used to estimate exposure, according to both the site and to the yearly average value of PM2.5 at each address, derived from a LUR model. Yearly concentration map from regional dispersion models were also used to consider temporal variations. We used Cox models with time-dependent exposures. Bi-pollutant model was performed, adjusting for the remaining PM2.5 mass, to verify independent effect of each source. Results The cohort includes about 2 million inhabitants (6,011,667 person-years) with 82 624 deaths for natural causes. Identified sources were Traffic, Biomass Burning, Oil Combustion, Anthropogenic Mix; Secondary Nitrates and Secondary Sulfate. We observed an association between PM2.5 total mass and natural mortality (HR=1.007; 95% CI: 1.004 to 1.011 per 1 µg/m 3 PM2.5). 1-unit increment of BB and OC were associated with an increased risk of natural mortality, independent from the remain part of PM2.5 (HR=1.015; 95% CI: 1.005 to 1.025 and HR=1.035; 95% CI: 1.005 to 1.065, respectively). Respiratory mortality showed HR comparable to natural mortality (HR=1.008; 0.995–1.020 per 1 µg/m3 PM2.5), while association with cardiovascular mortality were equal to 1.003 (95% CI: 0.998 to 1.009 per 1 µg/m 3 PM2.5). Conclusion A large residential cohort was built to examine the long-term impact on mortality of exposure to Source-related components of PM2.5. This study suggests an effect of long-term exposures for some specific sources, both on natural and respiratory mortality indicating a relevant role of Biomass Burning and Oil Combustion. These findings can be useful to orientate the pollution reduction policies