Science of the Total Environment 409 (2011) 4934–4938

Contents lists available at SciVerse ScienceDirect

Science of the Total Environment

j ourna l homepage: www.e lsev ie r .com/ locate /sc i totenv

Coarse particles and mortality in three Chinese cities: The China Air Pollution andHealth Effects Study (CAPES)

Renjie Chen a,b,1, Yi Li c,1, Yanjun Ma d,1, Guowei Pan e, Guang Zeng f, Xiaohui Xu g,Bingheng Chen a,b, Haidong Kan a,b,⁎a School of Public Health, Key Lab of Public Health Safety of the Ministry of Education, Fudan University, Shanghai, Chinab G_RIoCE (Research Institute for the Changing Global Environment) and Fudan Tyndall Centre, Fudan University, Shanghai 200433, Chinac Chinese Academy of Meteorological Sciences, Beijing, Chinad Shenyang Institute of Atmospheric Environment, China Meteorological Administration, Shenyang, Chinae Liaoning Provincial Center for Disease Control and Prevention, Shenyang, Chinaf Department of Preventive Medicine, School of Management, Beijing University of Chinese Medicine, Beijing, Chinag Department of Epidemiology and Biostatistics, College of Public Health and Health Professions, University of Florida, Gainesville, FL, USA

⁎ Corresponding author at: Box 249, 130 Dong-An RTel./fax: +86 21 6404 6351.

E-mail address: haidongkan@gmail.com (H. Kan).1 These authors contributed equally to this work.

0048-9697/$ – see front matter © 2011 Elsevier B.V. Alldoi:10.1016/j.scitotenv.2011.08.058

a b s t r a c t

a r t i c l e i n f o

Article history:Received 4 July 2011Received in revised form 23 August 2011Accepted 25 August 2011Available online 16 September 2011

Keywords:Air pollutionCAPESCoarse particlesMortalityTime-series

Evidence concerning the health risks of coarse particles (PM10-2.5) is limited. There have been no multi-cityepidemiologic studies of PM10-2.5 in developing Asian countries. We examine the short-term association be-tween PM10-2.5 and daily mortality in three Chinese cities: Beijing, Shanghai, and Shenyang. PM10-2.5 concen-trations were estimated by subtracting PM2.5 from PM10 measurements. Data were analyzed using the over-dispersed generalized linear Poisson models. The average daily concentrations of PM10-2.5 were 101 μg/m3 forBeijing (2007–2008), 50 μg/m3 for Shanghai (2004–2008), and 49 μg/m3 for Shenyang (2006–2008). In thesingle-pollutant models, the three-city combined analysis showed significant associations between PM10-2.5

and daily mortality from both total non-accidental causes and from cardiopulmonary diseases. A 10-μg/m3 in-crease in 1-day lagged PM10-2.5 was associated with a 0.25% (95% CI: 0.08 to 0.42) increase in total mortality,0.25% (95% CI: 0.10 to 0.40) increase in cardiovascular mortality, and 0.48% (95% CI: 0.20 to 0.76) increase in respi-ratory mortality. However, these associations became statistically insignificant after adjustment for PM2.5. PM2.5

was significantly associated with mortality both before and after adjustment for PM10-2.5. In conclusion, therewere no statistically significant associations between PM10-2.5 and daily mortality after adjustment for PM2.5 inthe three Chinese cities.

oad, Shanghai 200032, China.

rights reserved.

© 2011 Elsevier B.V. All rights reserved.

1. Introduction

Ambient air pollution is a complex mixture composed of both solidparticles and gaseous pollutants. Among various air pollutants, particu-late matter (PM) often shows the strongest evidence for adverse healtheffects (Brunekreef and Holgate, 2002). PM can be characterized as dis-crete particles spanning several orders of magnitude in size: PM10 (de-fined as particulate matter less than 10 μm in aerodynamic diameter);PM2.5, also known as fine particles (defined as those particles lessthan 2.5 μm in aerodynamic diameter); and PM10-2.5, also known ascoarse particles (defined as those particles between 10 and 2.5 μm inaerodynamic diameter). Most prior studies have only used PM2.5 orPM10 as PM measurements, leaving the effects of other particle sizes –particularly PM10-2.5 – not well understood (Pope and Dockery, 2006).

Compared with PM2.5, PM10-2.5 have different sources, compositionand depositionmode in the human airway (Lippmann and Schlesinger,2000; Wilson and Suh, 1997). In addition, findings from existing epi-demiological studies of PM10-2.5 have been limited and inconclusive(Brunekreef and Forsberg, 2005; Chang et al., 2011; Mallone et al.,2011); some studies found significant health hazards of PM10-2.5,while others did not. The inadequate evidence of health effects ofPM10-2.5 had led the US Environmental Protection Agency (EPA) to re-ject a proposal to replace the existing daily PM10 standard with dailyPM10-2.5. Furthermore, most studies of PM10-2.5 were conducted in de-veloped countries, with only a small number of studies conducted inAsia. As the results, there remains a need for studies in cities of developingcountries, where characteristics of outdoor air pollution (e.g. air pollu-tion level, chemical composition and size of particles, and fate andtransport of pollutants), meteorological conditions, and socio-demographic status of local residents (e.g. disease pattern, age struc-ture, and social economic status), may be different from developedcountries.

As the largest developing country, China may have the worst PMpollution in the world (Kan et al., 2009). The current Chinese Air

Table 1Summary statistics (mean and SD) of daily death numbers, air pollution levels andweather conditions in Beijing, Shanghai and Shenyang.

Beijing Shanghai Shenyang

Population size (millions) 7.1 6.5 3.5Daily death numbers

All-natural causes 118 (22) 119 (22) 67 (10)Cardiovascular 54 (13) 46 (12) 31 (7)Respiratory 14 (6) 13 (5) 7 (3)

Air pollutants (24-h average, μg/m3)PM10 172 (93) 105 (54) 141 (66)PM2.5 82 (52) 55 (30) 94 (52)PM2.5–10 101 (67) 50(31) 49 (30)

Weather conditionsTemperature (°C) 15 (11) 19 (9) 9 (12)Humidity (%) 54 (21) 69 (12) 66 (15)

4935R. Chen et al. / Science of the Total Environment 409 (2011) 4934–4938

Quality Standard includes PM10 only, and PM2.5 and PM10-2.5 are stillnot criteria pollutants. In the literature, only one study in China hasinvestigated the health impact of PM10-2.5 and this is due to lack ofmon-itoring data (Kan et al., 2007). The objective of this paper is to examinethe short-term associations between PM10-2.5 and daily mortality inthree Chinese cities – Beijing, Shanghai, and Shenyang. This study is acomponent of the “China Air Pollution and Health Effects Study”(CAPES) initiated by the China Ministry of Environmental Protection(Chen et al., 2011).

2. Materials and methods

2.1. Data collection

We conducted our analysis of PM10-2.5 in Beijing, Shanghai andShenyang. To our knowledge, PM10 and PM2.5 are simultaneouslymeasured only in these three cities in China. Study areas were re-stricted to the urban areas of the three cities due to inadequate airpollution monitoring stations in the suburban areas. The study pe-riods were from January 1, 2007 to September 30, 2008 for Beijing;March 4, 2004 to December 31, 2008 for Shanghai; and August 9,2006 to December 31, 2008 for Shenyang.

PM2.5 and PM10-2.5 are not regularly monitored in China. Similar toprevious studies (Brunekreef and Forsberg, 2005), PM10-2.5 concen-trations were estimated by subtracting PM2.5 from PM10 measure-ments. The 24-hour average concentrations of PM10 and PM2.5 weremeasured at each city using the tapered element oscillatingmicrobalance(TEOM)method. According to the rules of Chinese government, the loca-tions of thesemonitoring stations aremandated not to be in the direct vi-cinity of traffic or industrial sources; not to be influenced by localpollution sources; and to avoid buildings, housing, and large emitterssuch as coal-, waste-, or oil-burning boilers, furnaces, and incinerators.

Mortality data of urban residents were obtained from China Centerfor Disease Control and Prevention. The causes of death were codedaccording to International Classification of Diseases, 10 (ICD-10). Themortality data were classified into deaths due to total non-accidentalcauses (ICD-10: A00-R99), cardiovascular disease (ICD-10: I00-I99),and respiratory disease (ICD-10: J00-J98).

To allow adjustment for the effect of weather conditions on mor-tality, meteorological data (daily mean temperature and relative hu-midity) were obtained at each city.

2.2. Statistical methods

The CAPES project follows the same analytical approach as thePublic Health and Air Pollution in Asia (PAPA) program (Wong etal., 2008). Daily counts of deaths and PM10-2.5 levels were linked bydate and were therefore analyzed with time-series methods (Bell etal., 2004).

To control for long-term and seasonal trends, generalized linearmodeling, with natural spline smoothers, was used to model dailymortality. According to the PAPA Protocol, we used the partial auto-correlation function (PACF) to guide the selection of model parame-ters. Specifically, we used 4–6 degrees of freedom (df) per year fortime trend for all mortality outcomes. When the absolute magnitudeof the PACF plot was less than 0.1 for the first two lag days, the basicmodel was regarded as adequate; if this criteria was not met, auto-regression (AR) terms for lag up to 7 days were introduced to im-prove the model (Kan et al., 2008). Day of the week (DOW) was in-cluded as a dummy variable in the basic models. Residuals of thebasic model were examined to check whether there were discernablepatterns and autocorrelation by means of residual plots and PACFplots.

After establishing the basic model, PM10-2.5 and covariates (in-cluding temperature, relative humidity, and PM2.5 concentrations)were introduced in the model. Based on the existing literature

(Dominici et al., 2006), 3 df (whole period of study) for temperatureand humidity could adequately control for their effects on mortalityand was therefore used in the model. To examine the temporal rela-tionship of PM10-2.5 with mortality, models with different lag struc-tures from lag 0 to lag 2 were fitted. A lag of day 0 (L0) correspondsto the current-day PM2.5, and a lag of day 1 (L1) refers to theprevious-day PM2.5.

We fitted both single-pollutant and two-pollutant models to as-sess the stability of PM's health effect. In the single-pollutant models,PM10-2.5 and PM2.5 were included alone in the model. In the two-pollutant models, PM10-2.5 and PM2.5 were included jointly at thesame lag. Given the difficulty of determining the optimal values ofdf for time trend, we conducted sensitivity analyses to test the impactof alternative df values on the estimated effects of PM10-2.5.

After the health effects of PM10-2.5 and PM2.5 at each city were es-timated, we calculated combined estimates of excess risk of mortalityand their standard errors using a random-effects model. Estimateswere weighted by the inverse of the sum of variance within andbetween-city.

All analyses were conducted in R 2.10.1 using the MGCV package.The results are presented as the percent change in daily mortality per10 μg/m3 increase of PM concentrations.

3. Results

Table 1 summarizes the mortality and PM data in the three cities.During the study periods, the mean daily death numbers for all non-accidental causes, cardiovascular causes and respiratory causes – inrespective order – were 118, 54 and 14 in Beijing; 119, 46, and 13in Shanghai; and 67, 31, and 7 in Shenyang (Table 1). Among allnon-accidental deaths, cardiorespiratory diseases accounted for 58%in Beijing, 50% in Shanghai, and 57% in Shenyang.

Generally, the PM10-2.5 levels in three Chinese cities were muchhigher than those reported in developed countries (Chang et al.,2011; Graff et al., 2009; Host et al., 2008; Lipsett et al., 2006; Maligand Ostro, 2009; Peng et al., 2008; Puett et al., 2009; Yeatts et al.,2007; Zanobetti and Schwartz, 2009) (Table 1). The average dailyconcentrations of PM10-2.5 were 101 μg/m3 for Beijing, 50 μg/m3 forShanghai, and 49 μg/m3 for Shenyang. The ratios between PM10-2.5

and PM10 were 0.59 in Beijing, 0.48 in Shanghai, and 0.35 in Shenyang.In general, PM10-2.5 was strongly correlated with PM10 (correlation co-efficients ranging from 0.74 to 0.86), and had a moderate correlationwith PM2.5 (correlation coefficients ranging from 0.28 to 0.53) (Table 2).

In the single-pollutant models, the city-specific associations ofPM10-2.5 with mortality varied by causes of death and lag structures(Table 3). The three-city combined analysis showed significant asso-ciations between PM10-2.5 and daily mortality from both total non-accidental causes (L0, L1 and L2) and from cardiopulmonary diseases(L1 only). Specifically, a 10-μg/m3 increase in PM10-2.5 (L1) was asso-ciated with a 0.25% (95% CI: 0.08 to 0.42) increase in total mortality,

Table 2Spearman correlation coefficients of PM2.5–10 with other metrics of PM and weatherconditions in three Chinese cities.

Beijing Shanghai Shenyang

PM10 0.82 0.86 0.74PM2.5 0.28 0.53 0.40Temperature −0.15 −0.27 −0.27Relative humidity −0.18 0.51 −0.36

a) Total mortality %

-0.2-0.1

00.10.20.30.40.50.60.7

L0 L1 L2 L0 L1 L2 L0 L1 L2 L0 L1 L2

Coarse PM Coarse PM adjusted for fine PM

Fine PM Fine PM adjustedfor coarse PM

b) Cardiovascular mortality %

-0.3-0.2-0.1

00.10.20.30.40.50.60.70.8

L0 L1 L2 L0 L1 L2 L0 L1 L2 L0 L1 L2

Coarse PM Coarse PM adjusted for fine PM

Fine PM Fine PM adjustedfor coarse PM

c) Respiratory mortality %

-0.6

-0.4

-0.2

0

0.2

0.4

0.6

0.8

1

L0 L1 L2 L0 L1 L2 L0 L1 L2 L0 L1 L2

Coarse PM Coarse PM adjusted for fine PM

Fine PM Fine PM adjustedfor coarse PM

Fig. 1. Percentage change in daily mortality per a 10-μg/m3 increase in particulate matter.a) Total mortality. b) Cardiovascular mortality. c) Respiratory mortality.

4936 R. Chen et al. / Science of the Total Environment 409 (2011) 4934–4938

0.25% (95% CI: 0.10 to 0.40) increase in cardiovascular mortality, and0.48% (95% CI: 0.20 to 0.76) increase in respiratory mortality. Similarly,PM2.5 showed statistically significant associations with daily mortalityin both city-specific and combined analysis (Table 3); a 10-μg/m3 in-crease in PM2.5 (L1) was associated with a 0.32% (95% CI: 0.22 to 0.42)increase in total mortality, 0.46% (95% CI: 0.30 to 0.62) increase in car-diovascular mortality and 0.50% (95% CI: 0.19 to 0.81) increase in respi-ratory mortality (Table 3).

Fig. 1 shows the combined average estimates and 95% CI of the per-centage increases in daily mortality associated with PM10-2.5 and PM2.5

in both single- and two-pollutant models. When adjusted for PM2.5,the effects of PM10-2.5 were no longer statistically significant for eachmortality outcome. Specifically, a 10-μg/m3 increase in PM10-2.5 (L1)was associated with a nonstatistically significant 0.14% (95% CI: −0.11to 0.39) increase in total mortality, 0.13% (95% CI: −0.03 to 0.29) in-crease in cardiovascularmortality, and 0.36% (95% CI:−0.05 to 0.77) in-crease in respiratory mortality when adjusted for PM2.5. However, theassociations of PM2.5 with daily mortality remained significant after ad-justment for PM10-2.5.

4. Discussion

In this multi-city time-series study, we did not find statisticallysignificant associations between PM10-2.5 and daily mortality whenadjusted for PM2.5. In contrast, the health effects of PM2.5 remainedsignificant after adjustment for PM10-2.5. To our knowledge, this isthe first multi-city analysis in an Asian developing country to examinethe health impact of coarse and fine particulate air pollutionsimultaneously.

Table 3Percent change of daily mortality associated per a 10-μg/m3 increase of particulate matter – effect estimates of individual cities and combined effects.

Pollutant City Cause of death Lag

L0 L1 L2

PM10-2.5 Beijing Total 0.12 (0.00,0.24)* 0.35 (0.24,0.47)* 0.06 (−0.06,0.18)Cardiovascular 0.04 (−0.13,0.22) 0.26 (0.09,0.44)* 0.03 (−0.15,0.21)Respiratory 0.12 (−0.22,0.46) 0.50 (0.17,0.83)* 0.04 (−0.31,0.38)

Shanghai Total 0.07 (−0.15,0.28) 0.21 (0.01,0.41)* 0.13 (−0.06,0.33)Cardiovascular 0.29 (−0.05,0.63) 0.27 (−0.05,0.59) 0.28 (−0.03,0.59)Respiratory 0.17 (−0.46,0.81) 0.63 (0.04,1.22)* 0.75 (0.17,1.32)*

Shenyang Total 0.24 (−0.08,0.57) 0.03 (−0.29,0.35) 0.35 (0.04,0.66)*Cardiovascular 0.44 (−0.02,0.90) 0.10 (−0.36,0.57) 0.69 (0.25,1.13)*Respiratory 0.04 (−0.99,1.08) −0.24 (−1.29,0.82) −0.44 (−1.46,0.58)

Combined Total 0.12(0.02,0.22)* 0.25(0.08,0.41)* 0.11(0.01,0.20)*Cardiovascular 0.13(−0.02,0.28) 0.25(0.10,0.39)* 0.29(−0.06,0.64)Respiratory 0.13(−0.16,0.41) 0.48(0.20,0.76)* 0.19(−0.40,0.77)

PM2.5 Beijing Total 0.53 (0.37,0.69)* 0.37 (0.22,0.51)* 0.09 (−0.06,0.23)Cardiovascular 0.58 (0.35,0.81)* 0.49 (0.28,0.70)* 0.05 (−0.16,0.26)Respiratory 0.66 (0.21,1.11)* 0.48 (0.07,0.89)* 0.47 (0.07,0.88)*

Shanghai Total 0.47 (0.22,0.72)* 0.18 (−0.06,0.43) 0.06 (−0.18,0.31)Cardiovascular 0.41 (0.00,0.81)* 0.26 (−0.14,0.66) 0.00 (−0.40,0.40)Respiratory 0.61 (−0.15,1.37) 0.71 (−0.05,1.47) 0.17 (−0.59,0.92)

Shenyang Total 0.35 (0.17,0.53)* 0.32 (0.14,0.51)* 0.09 (−0.08,0.27)Cardiovascular 0.46 (0.19,0.73)* 0.49 (0.22,0.75)* 0.23 (−0.03,0.49)Respiratory 0.29 (−0.29,0.88) 0.41 (−0.17,0.99) −0.26 (−0.83,0.31)

Combined Total 0.46(0.35,0.57)* 0.32(0.22,0.43)* 0.09(−0.02,0.19)Cardiovascular 0.51(0.35,0.67)* 0.46(0.30,0.61)* 0.10(−0.05,0.25)Respiratory 0.54(0.21,0.86)* 0.50(0.19,0.81)* 0.22(−0.08,0.52)

*pb0.05.

4937R. Chen et al. / Science of the Total Environment 409 (2011) 4934–4938

We estimated a 0.25% increase in daily death numbers per 10 μg/m3

increase in PM10-2.5 (lag 1) which although small could have publichealth significance. However, when adjusted for PM2.5, the associationswere no longer statistically significant, suggesting either the adverse ef-fects of PM10-2.5 are attributable to the co-existing hazard of PM2.5, orthat our study lacked adequate power to detect an independent healtheffect of PM10-2.5.

The null associations between PM10-2.5 and daily mortality in ourstudy are consistent with previous studies. For example, Brunekreefand Forsberg (2005) systemically reviewed the time-series studiesof the effects of PM10-2.5 and PM2.5 on mortality, and identified fourstudies that not only reported separate estimates for PM10-2.5 andPM2.5, but also the results of a two-pollutant analysis. Three of thefour studies found that the effects of PM10-2.5 were no longer statisti-cally significant after adjustment for PM2.5; whereas the PM2.5 effectsremained significant (Brunekreef and Forsberg, 2005). Only a studyfromMexico City found significant effects of PM10-2.5 after further ad-justment for PM2.5 (Castillejos et al., 2000).

Unsurprisingly, we found significant associations of PM2.5 after weadjusted for PM10-2.5. PM2.5 is composed of many organic and inor-ganic compounds, including sulfate, nitrate, organic and elementalcarbon, earthen dust and biological materials. In contrast, PM10-2.5 ispredominately composed of crustal-related materials such as calcium;aluminum; silicon; magnesium; iron; and primary organic materialssuch as pollen, spores, and plant and animal debris (Fang et al., 2000;Lin et al., 2005, 2007; Wang and Shooter, 2005). Generally, thesenature-generated particles (e.g. PM10-2.5) are less toxic to the cardiorespi-ratory system than combustion-related particles, such as oil fly ash par-ticles. For instance, Costa et al found that the in vivo pulmonary toxicityof urban particles varied with sizes, with the greatest toxicity from par-ticles b1.7 μm and the lowest toxicity from those N3.5 μm (Costa andDreher, 1997). In addition, compared with PM2.5, PM10-2.5 had lowerconcentrations of soluble transitionmetals,whichmight also contributeto the lower toxicity of PM10-2.5 (Dreher et al., 1996). These characteris-tics, combined with different pattern of deposition in the lung, supportthe hypothesis that PM2.5 may have greater toxic effect than PM10-2.5.

In our combined analysis, a 10 μg/m3 increase in PM2.5 concentra-tions (L1) corresponded to 0.32% (95%CI: 0.22 to 0.43) increase intotal mortality. Compared with prior estimates in developed coun-tries, our estimates in China were lower. For example, a multi-cityanalysis in 112 U.S. cities showed that a 10 μg/m3 increment ofPM2.5 was associated with 1.00% increase in total mortality (Zanobettiand Schwartz, 2009). Our effect estimates strengthened the assump-tion that lower exposure-response functions may exist in Chineseair pollution studies compared with those conducted in developedcountries (Aunan and Pan, 2004). This difference may be explainedby the different characteristics of the study contexts, such as localPM levels, population sensitivity to PM, age structure, and especiallyparticle composition and toxicity. At higher concentrations, the riskof death per unit increase of pollutant concentrations often tendedto be reduced (Pope et al., 2009).

Some limitations of our analysis should be noted. First, as in othertime-series studies, we used available outdoor monitoring data torepresent the population exposure to ambient PM. The resulting mea-surement error may have implications for interpreting the effect ofPM (Zeger et al., 2000). Although a study suggested that this mea-surement error would generally tend to bias estimates downward(Samet et al., 2000), we lack information on personal exposure to par-ticles to quantify this bias. Second, our analysis was restricted to threelarge Chinese cities. Future study of coarse PM should include notonly large cities, but also small cities and rural areas. Third, our as-sessment of ambient particles was derived entirely from one moni-toring station at each city. Compared with studies in North America,the data we collected were limited by the numbers of cities and limitedtime period, whichmay have limited our power to find significant asso-ciation between PM10-2.5 and mortality.

In summary, we did not find significant associations of PM10-2.5 withdaily mortality after adjustment for PM2.5 in the three Chinese cities.The health effect of PM2.5 remained after adjustment for PM10-2.5. Ourfindings can supplement the limited evidence of PM10-2.5 and PM2.5 inChina, and may have implications for China's environmental policy.

Acknowledgements

The study was supported by the National Basic Research Program(973 program) of China (2011CB503802), Gong-Yi Program of ChinaMinistry of Environmental Protection (200809109 and 201209008),National Natural Science Foundation of China (30800892), and Programfor New Century Excellent Talents in University (NCET-09-0314).

The authors declare they have no competing financial interests.

References

Aunan K, Pan XC. Exposure-response functions for health effects of ambient air pollutionapplicable for China – a meta-analysis. Sci Total Environ 2004;329:3-16.

Bell ML, Samet JM, Dominici F. Time-series studies of particulate matter. Annu Rev PublicHealth 2004;25:247–80.

Brunekreef B, Forsberg B. Epidemiological evidence of effects of coarse airborne particleson health. Eur Respir J 2005;26:309–18.

Brunekreef B, Holgate ST. Air pollution and health. Lancet 2002;360:1233–42.CastillejosM, Borja-Aburto VH, Dockery DW, Gold DR, Loomis D. Airborne coarse particles

and mortality. Inhal Toxicol 2000;12:61–72.Chang HH, Peng RD, Dominici F. Estimating the acute health effects of coarse particulate

matter accounting for exposure measurement error. Biostatistics 2011. doi:10.1093/biostatistics/kxr002.

Chen R, Pan G, Zhang Y, Xu Q, Zeng G, Xu X, et al. Ambient carbon monoxide and dailymortality in three Chinese cities: the China Air Pollution and Health Effects Study(CAPES). Sci Total Environ 2011;409:4923–8.

Costa DL, Dreher KL. Bioavailable transition metals in particulate matter mediatecardiopulmonary injury in healthy and compromised animal models. EnvironHealth Perspect 1997;105(Suppl 5):1053–60.

Dominici F, Peng RD, Bell ML, Pham L, McDermott A, Zeger SL, et al. Fine particulate airpollution and hospital admission for cardiovascular and respiratory diseases. JAMA2006;295:1127–34.

Dreher K, Jaskot R, Kodavanti U, Lehmann J, Winsett D, Costa D. Soluble transitionmetals mediate the acute pulmonary injury and airway hyperreactivity inducedby residual oil fly ash particles. Chest 1996;109:33S–4S.

Fang GC, Chang CN, Wu YS, Wang V, Fu PP, Yang DG, et al. The study of fine and coarseparticles, and metallic elements for the daytime and night-time in a suburban areaof central Taiwan, Taichung. Chemosphere 2000;41:639–44.

Graff DW, CascioWE, Rappold A, Zhou H, Huang YC, Devlin RB. Exposure to concentratedcoarse air pollution particles causes mild cardiopulmonary effects in healthy youngadults. Environ Health Perspect 2009;117:1089–94.

Host S, Larrieu S, Pascal L, Blanchard M, Declercq C, Fabre P, et al. Short-term associa-tions between fine and coarse particles and hospital admissions for cardiorespira-tory diseases in six French cities. Occup Environ Med 2008;65:544–51.

KanH, ChenB,Hong C.Health impact of outdoor air pollution in China: current knowledgeand future research needs. Environ Health Perspect 2009;117:A187.

Kan H, London SJ, Chen G, Zhang Y, Song G, Zhao N, et al. Differentiating the effects offine and coarse particles on daily mortality in Shanghai, China. Environ Int2007;33:376–84.

KanH, London SJ, ChenG, Zhang Y, SongG, ZhaoN, et al. Season, sex, age, and education asmodifiers of the effects of outdoor air pollution on daily mortality in Shanghai, China:the Public Health and Air Pollution in Asia (PAPA) Study. Environ Health Perspect2008;116:1183–8.

Lin CC, Chen SJ, Huang KL, Hwang WI, Chang-Chien GP, Lin WY. Characteristics ofmetals in nano/ultrafine/fine/coarse particles collected beside a heavily traffickedroad. Environ Sci Technol 2005;39:8113–22.

Lin CC, Chen SJ, Huang KL, Lee WJ, Lin WY, Liao CJ, et al. Water-soluble ions in nano/ultrafine/fine/coarse particles collected near a busy road and at a rural site. EnvironPollut 2007;145:562–70.

Lippmann M, Schlesinger RB. Toxicological bases for the setting of health-related airpollution standards. Annu Rev Public Health 2000;21:309–33.

Lipsett MJ, Tsai FC, Roger L, Woo M, Ostro BD. Coarse particles and heart rate variabilityamong older adults with coronary artery disease in the Coachella Valley, California.Environ Health Perspect 2006;114:1215–20.

Malig BJ, Ostro BD. Coarse particles and mortality: evidence from a multi-city study inCalifornia. Occup Environ Med 2009;66:832–9.

Mallone S, Stafoggia M, Faustini A, Gobbi GP, Marconi A, Forastiere F. Saharan dust andassociations between particulate matter and daily mortality in Rome, Italy. EnvironHealth Perspect 2011. doi:10.1289/ehp.1003026.

Peng RD, Chang HH, Bell ML, McDermott A, Zeger SL, Samet JM, et al. Coarse particulatematter air pollution and hospital admissions for cardiovascular and respiratory dis-eases among Medicare patients. JAMA 2008;299:2172–9.

Pope III CA, Burnett RT, Krewski D, Jerrett M, Shi Y, Calle EE, et al. Cardiovascularmortalityand exposure to airborne fine particulate matter and cigarette smoke: shape of theexposure-response relationship. Circulation 2009;120:941–8.

4938 R. Chen et al. / Science of the Total Environment 409 (2011) 4934–4938

Pope III CA, Dockery DW. Health effects of fine particulate air pollution: lines thatconnect. J Air Waste Manag Assoc 2006;56:709–42.

Puett RC, Hart JE, Yanosky JD, Paciorek C, Schwartz J, Suh H, et al. Chronic fine andcoarse particulate exposure, mortality, and coronary heart disease in the Nurses'Health Study. Environ Health Perspect 2009;117:1697–701.

Samet JM, Dominici F, Curriero FC, Coursac I, Zeger SL. Fine particulate air pollution andmortality in 20 U.S. cities, 1987–1994. N Engl J Med 2000;343:1742–9.

Wang H, Shooter D. Source apportionment of fine and coarse atmospheric particles inAuckland, New Zealand. Sci Total Environ 2005;340:189–98.

Wilson WE, Suh HH. Fine particles and coarse particles: concentration relationshipsrelevant to epidemiologic studies. J Air Waste Manag Assoc 1997;47:1238–49.

Wong CM, Vichit-Vadakan N, Kan H, Qian Z. Public Health and Air Pollution in Asia(PAPA): a multicity study of short-term effects of air pollution on mortality. EnvironHealth Perspect 2008;116:1195–202.

Yeatts K, Svendsen E, Creason J, Alexis N, Herbst M, Scott J, et al. Coarse particulate matter(PM2.5-10) affects heart rate variability, blood lipids, and circulating eosinophils inadults with asthma. Environ Health Perspect 2007;115:709–14.

Zanobetti A, Schwartz J. The effect of fine and coarse particulate air pollution onmortality: a national analysis. Environ Health Perspect 2009;117:898–903.

Zeger SL, Thomas D, Dominici F, Samet JM, Schwartz J, Dockery D, et al. Exposuremeasurement error in time-series studies of air pollution: concepts and conse-quences. Environ Health Perspect 2000;108:419–26.