Characterising exposure to PM aerosols for an epidemiological study

Recent health studies evidence that epidemiological studies must be combined with accurate analyses of the physico-chemical properties of the particles, and vice versa, in order to determine the effects of atmospheric PM on human health. The present study focuses on the chemical characterisation of PM10 and PM2.5 aerosols, with the aim to characterise the exposure to this pollutant of the pregnant women in a cohort study in Valencia, Spain. PM sampling was carried out during five campaigns (I month by trimester) between 2004 and 2005, when the women in the study were pregnant. Four sites were selected, representing a gradient of urban, metropolitan, suburban and rural environments. The study includes the determination of PM levels, major and trace elements, and 15 polycyclic aromatic hydrocarbons (PAHs). The results evidenced a marked gradient regarding PM levels from urban to rural sites, which was also observed for anthropogenic PM components (OM + EC, NO3-, Fe, SO42-, NO3-, NH4+) but not for elements of mineral origin (Al2O3, Ca, K, Mg). PAH levels were similar or slightly higher that those reported for other Spanish sites but lower than in Europe. Findings show that the relationship between exposure to PM and adverse health effects cannot be determined only by the study of PM levels; other parameters such as chemical and physical properties of the aerosols must also be taken into account. Results also allow evaluating the adequacy of central monitor measurements as proxies for actual exposure to local emissions (e.g., vehicular emissions), which are known to vary considerably with distance to the source. The sampling strategy used in this study, devised jointly by PM monitoring and epidemiology specialists could be considered a guideline for the design of air quality monitoring studies aimed at epidemiological analysis. (C) 2007 Elsevier Ltd. All rights reserved.