Short-term effects of particulate air pollution on mortality in 25 cities in Italy and the role of the individual susceptibility–results of the EPIAIR2 project

Abstract

Background. A surveillance system (EPIAIR) is ongoing since 2001 in Italy to monitor short-term effects of air pollution. The relationship between PM10 and mortality has been already studied for the period 2001-2005, identifying specific subgroups of population most sensitive to air pollution. Aim. To estimate the association between exposure to PM10 and cause-specific mortality, and to evaluate individual characteristics as potential effect modifiers of this association in an extended period until 2010. Methods. Data on 422,723 natural deaths, age 35+, were collected in 25 Italian cities within the period 2003-2010. Information about cause of death, socio-demographic characteristics and chronic conditions from previous hospitalizations was gathered. Cardiac, cerebrovascular and respiratory diseases were investigated. Exposure data were collected from background monitoring stations. A case-crossover analysis was conducted within each city, followed by random-effects meta-analysis. Results. We found a 0.45% (95% confidence interval [CI]:0.07-0.83) increase in natural mortality for a 10µg/m3 increase of PM10 (lag 0-1). The effect for cardiac mortality lasted several days with an increase of 0.9% (95%CI:0.12-1.68) at lag 0-5. Higher effects of PM10 on natural mortality were found among people aged 85+ years (0.95%, 95%CI:0.39-1.51, p-value of the relative effect modification (p-REM)=0.080), and during summer season (1.51%,95%CI:0.18-2.85, p-REM=0.085). Subjects with previous hospitalizations for diabetes were more susceptible to the effect of PM10 (1.02%, 95%CI:0.30-1.74, p-REM=0.089); previous hospitalizations for anemia, diseases of pulmonary circulation, and heart failure were also suggestive of an effect modification. Conclusions. A strong evidence of a short-term effect of PM10 on natural and cardiac mortality up to the fifth day has been found. Old age, warm season, and diabetes turned out to be significant effect modifiers.