Size Fractionate Particulate Matter, Vehicle Traffic ... - ACS Publications


Size Fractionate Particulate Matter, Vehicle Traffic...

1 downloads 67 Views 575KB Size

Environ. Sci. Technol. 2009, 43, 4707–4714

Size Fractionate Particulate Matter, Vehicle Traffic, and Case-Specific Daily Mortality in Barcelona, Spain ´ N,† L . P E R E Z , * ,† M . M E D I N A - R A M O ¨ N Z L I , †,‡ A . A L A S T U E Y , § J . P E Y , § N. KU ´ REZ,§ R. GARCIA,† A. TOBIAS,| N. PE X. QUEROL,§ AND J. SUNYER† Center for Research in Environmental Epidemiology (CREAL), Dr Aiguader 88, 08003 Barcelona, Spain, Catalan Institute for Research and Advanced Studies (ICREA), 08010 Barcelona, Spain, Institute of Environmental Analysis and Water Research, IDAEA, CSIC, C/ Jordi Girona, 18, 08034 Barcelona, Spain, and National School of Public Health, Institute of Health Carlos III, C/ Sinesio Delgado 8, 28029 Madrid, Spain

Received November 7, 2008. Revised manuscript received February 9, 2009. Accepted February 24, 2009.

Recent epidemiological research suggests that short-term effects of particle matter (PM) in urban areas may preferentially be driven by fine fractions. Questions remain concerning the adverse health effects of coarse particles generated by noncombustion, traffic-related processes and the mechanism of action of PM. Using a time-stratified case-crossover design, we investigated the association between three independent size fractions, coarse (PM10-2.5), intermodal (PM2.5-1), and very fine PMs (PM1), and three health outcomes, respiratory, cardiovascular, and cerebrovascular mortality in Barcelona, Spain, during the period of March 2003-December 2005. Using existing data, we examined the chemical composition of each fraction to explore the effects of PM from different sources and the mechanisms of action. We found that increased levels of PM1 and PM10-2.5 were associated with increased levels of cardiovascular and cerebrovascular mortality at lag 1 and lag 2. At lag 1, the odds ratio (OR) for a 10 µg/m3 increase in PM1 was 1.028 [95% confidence interval (CI), 1.000-1.058] for cardiovascular mortality and 1.063 (95% CI, 1.004-1.124) for cerebrovascular mortality. At lag 1, the odds ratio per a 10 µg/ m3 increase of PM10-2.5 was 1.059 (95% CI, 1.026-1.094) for cardiovascular mortality and 1.098 (95% CI, 1.030-1.171) for cerebrovascular mortality. Association with respiratory mortality was only detected for PM2.5-1 at lag 2 (OR, 1.206 per a 10 µg/ m3 increase; 95% CI, 1.028-1.416). Chemical composition data showed that PM in Barcelona was generated in a large proportion by vehicle traffic. Vehicle traffic PM, generated by combustion and noncombustion processes, should be considered in air pollution mitigation strategies in urban areas. * Corresponding author e-mail: [email protected]; fax ++ (34) 93 316 0635. † Center for Research in Environmental Epidemiology (CREAL) Municipal Institute of Medical Research (IMIM), Centro de Investigacio´n Biome´dica en Red de Epidemiologı´a y Salud Pu ´ blica (ClBERESP). ‡ Catalan Institute for Research and Advanced Studies (ICREA). § Institute of Environmental Analysis and Water Research, IDAEA, and CSIC. | Institute of Health Carlos III. 10.1021/es8031488 CCC: $40.75

Published on Web 03/26/2009

 2009 American Chemical Society

Introduction Past research demonstrates consistent associations among exposure to airborne particulate matter (PM) and mortality and morbidity and other markers of pathologic processes (1-3). Adverse short-term effects of PM are reported for PM with a diameter less than 10 µm (PM10). Recent research suggests that effects may be driven mainly by fine PM with a diameter less than 2.5 µm (PM2.5), generated by combustion processes (4). The effects of coarse PM with a diameter between 10 and 2.5 µm (PM10-2.5) are less clear. Coarse PM has typically been associated with terragenous sources (5). However, recent studies have shown that a large fraction of coarse PM in urban areas is formed by resuspended tire, brake, and pavement residue, and that some of these components are similar to components found in combustion PM (6-10). Several epidemiological studies have detected short-term effects of PM10-2.5 in urban areas (1, 2, 5, 11, 12), but these studies provide limited information about the relationship between PM sources and health effects. For example, sources of PM10-2.5 can vary considerably across and within areas; therefore, studies that pool data from different geographic areas (12) or studies conducted in areas without high vehicle traffic (5, 11, 13) are not sufficient. In addition, studies which examined the toxicological effects of road dust found contrasting results (14, 15). This study investigates the association between daily PM and health effects in Barcelona, Spain. Barcelona has one of the highest PM levels in western Europe and has intense vehicle traffic (16). We analyzed three independent size fractions, coarse (PM10-2.5), intermodal (PM2.5-1), and very fine (PM1), and three health outcomes, respiratory, cardiovascular, and cerebrovascular mortality. Using existing data, we examined the chemical composition of each fraction to explore the effects of PM from different sources and the mechanism of action of PM. This study builds on past research and provides new information for future research and public health and policy.

Materials and Methods Study Area. The study was conducted in the city of Barcelona, located on the northeast coast of the Mediterranean Sea. The city of Barcelona has a population of approximately 1.8 million inhabitants and is one of the most densely populated cities in Europe (16). The high population density results in the majority of people living and working in square block type areas adjacent to streets and main arteries with high road traffic. Air Pollution Data. Daily levels of particulate matter (PM) concentrations were obtained between March 27, 2003, and December 31, 2005, from real time measurements performed at a single monitoring site. The PM monitoring site is an urban background station and is exposed to road traffic emissions from Diagonal Avenue (approximately 150 m distance), which is one of the largest arteries in the city. An optical counter was used for real time PM10, PM2.5, and PM1 measurements. The data were continuously validated and corrected with gravimetric methods. This was done by simultaneously collecting PM10, PM2.5, and PM1 samples (three 24 h samples per week for each size fraction for the whole study period) with high-volume samplers (30 m3/h) equipped with Digitel cat off inlets and quartz microfiber filters (Schleicher & Schuell, QF20). PM filters were stabilized at 21 °C and 25% humidity for VOL. 43, NO. 13, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4707

48 h. PM10 levels were obtained using standard gravimetric methods, and a mean uncertainty of 1.2 µg/m3 was determined. Daily mass concentrations for coarse particles (PM10-2.5) and intermodal particles (PM2.5-1) were obtained by subtracting PM2.5 from PM10 and PM1 from PM2.5, respectively. All days for which PM10, PM2.5, and PM1 data were available were retained. Days when at least one fraction measurement was missing or presented improbable results (i.e., PM10-2.5 greater than PM2.5) were discarded. In addition to PM mass, the chemical composition of PM was also determined on samples collected on filters approximately once a week from October 1, 2005, to October 1, 2006. The speciation scheme was based on the analysis of major and trace elements by inductively coupled plasma atomic emission spectroscopy (ICP-AES) and inductively coupled plasma-mass spectrometry (ICP-MS). The determination of soluble anions were based on analysis by ion chromatography, and ammonium was determined using a selective electrode (these two techniques used water extractions on 1/4 of each filter). Lastly, total carbon was determined using an elemental LECO analyzer. Fractions of blank filters were analyzed in the batches of their respective filter samples, and the corresponding blank concentrations were subtracted from each sample. It is important to note that for major and trace element content HF:HNO3:HClO4 digestion (1/2 of each filter) was carried out in order to dissolve the mineral matter. A few milligrams of reference material NIST 1633b was added to a fraction of a blank filter to confirm the accuracy of the analysis of the acidic digestions. Uncertainties were around 10% for total carbon concentrations, between 4 and 10% for sulfate, nitrate, and ammonium,