Short-term effects of desert and non-desert PM10 on mortality in Sicily, Italy

Background: Increased PM10 concentrations are commonly observed during Saharan dust advections. Limited epidemiological evidence suggests that PM10 from anthropogenic and desert sources increase mortality. We aimed to evaluate the association between source-specific PM10 (non-desert and desert) and cause-specific mortality in Sicily during 2006-2012 period. Methods: Daily PM10 concentrations at 1-km(2) were estimated in Sicily using satellite-based data, fixed monitors and land use variables. We identified Saharan dust episodes using meteorological models, back-trajectories, aerosol maps, and satellite images. For each dust day, we estimated desert and non-desert PM10 concentrations. We applied a time-series approach on 390 municipalities of Sicily to estimate the association between PM10 (non-desert and desert) and daily cause-specific mortality. Results: 33% of all days were affected by Saharan dust advections. PM10 concentrations were 8 mu g/m(3) higher during dust days compared to other days. We found positive associations of both non-desert and desert PM10 with cause-specific mortality. We estimated percent increases of risk (IR%) of non-accidental mortality equal to 2.3% (95% Confidence Interval [CI]: 1.4, 3.1) and 3.8% (3.2, 4.4), per 10 mu g/m(3) increases in non-desert and desert PM10 at lag 0-5, respectively. We also observed significant associations with cardiovascular (2.4% [1.3, 3.4] and 4.5% [3.8, 5.3]) and respiratory mortality (8.1% [6.8, 9.5], and 6.3% [5.4, 7.2]). We estimated higher effects during April-September, with IR% = 4.4% (3.2, 5.7) and 6.3% (5.4, 7.2) for non-desert and desert PM10, respectively. Conclusions: Our results confirm previous evidence of harmful effects of desert PM10 on non-accidental and cardio-respiratory mortality, especially during the warm season.