- Email: [email protected]
A comparison of mortality related to urban air particles between periods with Asian dust days and without Asian dust days in Seoul, Korea, 2000–2004
ARTICLE IN PRESS
Environmental Research 105 (2007) 409–413 www.elsevier.com/locate/envres
A comparison of mortality related to urban air particles between periods with Asian dust days and without Asian dust days in Seoul, Korea, 2000–2004 Jong-Tae Lee, Ji-Young Son, Yong-Sung Cho Department of Public Health, Graduate School of Hanyang University, Hanyang University, Haengdang-dong 17, Seongdong-gu, Seoul 133-791, Republic of Korea Received 18 January 2007; received in revised form 28 May 2007; accepted 18 June 2007 Available online 20 July 2007
Abstract Recent papers have reported that Asian dust events have been associated with increased risks of all-cause mortality and rates of respiratory illness. The current study was designed to estimate the relative risk of mortality associated with Asian dust events. We used the daily counts of non-accidental deaths, air pollution and meteorological data in Seoul, Korea from 2000 to 2004. We divided all days during the study period into two groups according to the presence or absence of Asian dust events. For each group, we conducted timeseries analysis to estimate the relative risk of total non-accidental death when the concentration of each air pollutant increased by the inter-quartile range (IQR). The average concentrations of every air pollutant on the days without a dust event were lower than those on the days with such an event. We found that the effect sizes of air pollution on daily death rates in the model without Asian dust events were larger than those in the model with Asian dust events, and were statistically significant for all air pollutants (PM10, CO, NO2, and SO2) except for O3. Our results suggest that we are likely to underestimate the risk of urban air particles if we analyze the effect size of air pollution on daily mortality during Asian dust events. We hypothesize that the real health effect is much larger than previous results suggested. r 2007 Elsevier Inc. All rights reserved. Keywords: Asian dust event; Mortality; Urban air particles; Generalized additive model; Underestimation
1. Introduction Asian dust (AD) is the phenomenon whereby particles suspended in ascending air currents in the desert area of the northern part of China diffuse widely, are transported and descend slowly. It is an example of the long-distance transportation of air pollutants that affects the air quality of Korea. Large-scale movement of sand clouds, known as Whang-sa, has an influence over an extensive geographical area, and may result in health effects such as ocular and respiratory irritation (Monn et al., 1997). In Korea, dust-related events for the last two centuries can be found in various kinds of historical records (Wada, 1917). It is, however, only quite recently that AD has Corresponding author. Fax: +822 2298 0248.
E-mail address: [email protected] (J.T. Lee). 0013-9351/$ - see front matter r 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.envres.2007.06.004
drawn much attention due to its increasing frequency and strength (Korea Meteorological Administration (KMA), 2005). Papers over the last few years reported that AD events were associated with an increased risk of all-cause mortality (Kwon et al., 2002) and respiratory symptoms (Park et al., 2005). An animal study also showed that inflammatory markers in the lung and peripheral blood of rats increased with exposure levels of concentrated ambient particles collected during an AD storm in Taiwan (Lei et al., 2004). The occurrence of AD events in the Korean peninsula are closely related to meteorological conditions, including high surface wind speed and barometric instability (Chun et al., 2001). The occurrence of AD events in the Korean peninsula is, therefore, most frequent during spring and irregular during winter and fall. There have been no major dusts events recorded during summer, which is the period of the Asian monsoon with heavy rainfall.
ARTICLE IN PRESS 410
J.T. Lee et al. / Environmental Research 105 (2007) 409–413
The health effects of particulate matter are assumed to depend on both the size and the chemistry of the particles (Oberdo¨rster, 1996; Donaldson and MacNee, 1998). During AD events, highly increased mass concentrations of ambient particles are observed in comparison with those days without AD events. A recent study indicated that during AD events the concentration of larger particles (more than 0.82 mm in aerodynamic diameter) was distinctly increased, but the concentration of smaller particles was reduced. And measurements of optical depth also indicated that the atmosphere is more turbid, with larger particles (Chun et al., 2001) in April 1998 and January 1999. However, the potentially hazardous metal fractions of particles did not change or decreased during AD events compared with usual conditions (Kim et al., 2003). Hence, we estimated the relative risks (RRs) of excess mortality associated with ambient air particles, controlling for AD events during the period 2000–2004. From these results, we can estimate the RR of urban air particles on mortality during the days of AD events separately.
rate after controlling for potential confounding variables. The confounding variables considered in this study were long-term trends and seasonal patterns of air pollution, weather conditions and date. Also, to take the lag effect into consideration, we applied the individual pollutant concentrations of the same day 1–7 lagged days, or moving averages from 2 to 3 days. Akaike’s information criterion (AIC) was used to measure the model’s goodness of fit. All analyses were carried out with S-PLUS 2000 software (Statistical Sciences, Seattle, WA). We analyzed the data with a more stringent convergence criterion than the default settings of GAM to avoid biased estimates of the regression coefficient and standard errors (Dominici et al., 2002). To compare the magnitude of the effect of air pollutants on mortality, we calculated the RR of non-accidental death for inter-quartile changes of concentrations of PM10 and other air pollutants for each group (WAD, W/O AD).
2. Materials and methods
There were 63 AD event days in Seoul between 2000 and 2004 (Table 1). The occurrence of AD events was mostly in March and April. Table 2 shows summary statistics of the variables included in the analysis. We categorized all days during the study period into two groups according to the presence of AD events. The average concentrations of air pollutants between the days without AD events and those days with AD events were not significantly different except for PM10 (Table 2). The t-test results to assess statistical differences indicate that there was no difference in the meteorological variables and total death counts between the two groups (WAD, W/O AD). Fig. 1 shows the daily average PM10 concentration from 2000 to 2004. During
2.1. Data The daily counts of non-accidental deaths (ICD-10 codes, A00-R99) in Seoul between January 1, 2000 and December 31, 2004 were supplied by the National Statistical Office, Republic of Korea. Daily ambient air pollution levels in Seoul were obtained from 27 monitoring stations distributed evenly throughout Seoul and operated by the Department of Environment, Republic of Korea, during the same period. Routine monitoring for sulfur dioxide, nitrogen dioxide, ozone, carbon monoxide, and respirable particles was conducted using standardized reference methods (Lee et al., 1999b). We used 24-h averages as the basic measure in this analysis, except for ozone and carbon monoxide, for which we took the maximum daily 8-h moving average as the exposure index. The National Meteorological Administration of the Republic of Korea provided information on the occurrence of AD events each day, as well as daily weather information, including air temperature, relative humidity and pressure, for Seoul during the same study period.
3. Results
Table 1 Asian dust days in Seoul, Korea, 2000–2004 Year
Date
2000
7 March, 23–24 March, 27–28 March 7–8 April, 23 April, 26–27 April
2001
2 January 3–7 March, 20–25 March 7–12 April, 24–26 April 16–19 May 13–14 December
2002
12–13 January 17–19 March, 21–23 March 8–12 April, 16–17 April 11–12 November
2003
27 March 12–13 April
2004
25 February 10–11 March, 30–31 March 23 April
2.2. Statistical analysis All days during the study period were categorized into two groups according to the presence or absence of AD events. We, therefore, called each group ‘‘with AD days’’ (WAD) and ‘‘without AD days’’ (W/O AD). For each group, we conducted time-series analysis to estimate the RR of total death when the concentration of each air pollutant increased by the inter-quartile range (IQR). Time-series analysis using the generalized additive model (GAM) evaluated the effect of air pollution on mortality
ARTICLE IN PRESS J.T. Lee et al. / Environmental Research 105 (2007) 409–413
411
Table 2 Summary statistics of daily concentrations of air pollutants and total death counts on model with/without Asian dust days included in Seoul, Korea, 2000–2004 Variable
Mean (SD)a Model with Asian dust days included (N ¼ 1827 days)
70.00 (47.80) PM10 (mg/m3) CO (ppm) 0.92 (0.42) NO2 (ppb) 37.34 (12.23) SO2 (ppb) 5.21 (2.18) 26.82 (14.72) O3 (ppb) Temperature (1C) 12.90 (10.09) Humidity (%) 62.46 (14.42) Pressure (hPa) 1016.10 (8.09) Total death counts (per 93.26 (12.02) day) a
Model without Asian dust days included (N ¼ 1764 days)
Asian dust days only (N ¼ 63 days)
65.77 0.92 37.38 5.20 26.70 13.03 62.79 1016.19 93.18
188.49 1.00 36.41 5.62 30.07 9.15 54.27 1013.59 95.38
(33.60) (0.41) (12.22) (2.17) (14.78) (10.19) (14.28) (8.15) (12.08)
(142.85) (0.47) (12.58) (2.35) (12.56) (5.62) (14.90) (5.55) (10.07)
Standard deviation (SD).
AD events, the concentration of PM10 increased greatly compared with those days without AD events, especially during the spring of 2002. The time-series analysis showed that the effect size of PM10 on daily mortality in the model W/O AD was larger than that in the model WAD (Table 3). For PM10, the RR of total non-accidental death was indicated by a 1.0% (95% CI, 0.2–1.8%) increase in the model W/O AD and a 0.7% (95% CI, 0.2–1.3%) increase in the model WAD. In both analyses, the RRs on the best statistically lagged day (the day with minimum AIC) were statistically significant for all air pollutants (PM10, CO, NO2, and SO2), except for O3. For other pollutants, in the WAD models, each increase of IQR in CO, NO2, SO2, and O3 corresponded, respectively, to a 3.3% (95% CI, 2.5–4.1%), 2.4% (95% CI, 1.6–3.1%), 2.5% (95% CI, 1.7–3.3%), and 0.4% (95% CI, 0.6–1.4%) increase. The estimate of percentage increase in the W/O AD model was 3.3% (95% CI, 2.5–4.2%) for CO, 2.5% (95% CI, 1.7–3.3%) for NO2, 2.7% (95% CI, 1.8–3.5%) for SO2, and 0.5% (95% CI, 0.5–1.5%) for O3 (Table 3). 4. Discussion and conclusions This study was designed to estimate the RR of urban air particles on mortality during the days of AD events. We found that the degree of increase in RR in the model W/O AD was larger than that in the model WAD, and statistically significant for all air pollutants (PM10, CO, NO2, and SO2) except for O3. These results could be explained by two factors. The first is that the chemical composition of AD consists of relatively less toxic components for humans. In the case of air particulate matter, excess mortality due to PM2.5 is much higher than that due to PM10 (Schwartz, 1996; USEPA, 2001). Schwartz et al. (1999) reported that the coarse particles from windblown dust are not associated with mortality risk. During AD events, mass concentra-
Table 3 Estimated percentage increases in the risk of total death and 95% confidence intervals for total death by an inter-quartile range (IQR)a using a single-pollutant modelb Air pollutants [lag time]
PM10 (mg/m3) [lag1+2+3] CO (ppm) [lag1] NO2 (ppb) [lag1] SO2 (ppb) [lag1] O3 (ppb) [lag1+2]
Percentage increases Model with Asian dust days included (N ¼ 1827 days)
Model without Asian dust days included (N ¼ 1764 days)
0.7 3.3 2.4 2.5 0.4
1.0 3.3 2.5 2.7 0.5
(0.2, 1.3) (2.5, 4.1) (1.6, 3.1) (1.7, 3.3) (0.6, 1.4)
(0.2, 1.8) (2.5, 4.2) (1.7, 3.3) (1.8, 3.5) (0.5, 1.5)
a The IQR was 41.49 mg/m3 for PM10, 0.54 ppm for CO, 17.93 ppb for NO2, 3.06 ppb for SO2, 21.03 ppb for O3. b The model included the following variables: an intercept, indicator variables for days of the week, and smooth spline functions for date, temperature, humidity, and air pressure.
tions of particulate matter (hourly maximum concentration 606–839 mg/m3 in Seoul, Korea, 2001) are 7–14 times higher than the usual level of particulate matter (daily average 72 mg/m3 in Seoul, Korea, 2000) (Kim and Kim, 2003). However, the relative proportion of fine particulate matter (less than 2.5 mm in aerodynamic diameter) was decreased and the relative concentration of coarse particulate matter (more than 2.5 mm in aerodynamic diameter) was increased (Chun et al., 2001; Kim et al., 2002). This suggests that the ambient air particles during AD events were composed of particles of larger diameter. Also, a recent study reported that dust-borne metals are hazardous to health (Laden et al., 2000). The hazardous metal fractions of particles collected during AD event days did not change or decreased compared with those of particles collected during days without dust event (Kim et al., 2003). Kim and his colleagues analyzed the metallic components by comparing the concentrations in AD with those of a non-AD (NAD) period, and fine particle
ARTICLE IN PRESS 412
J.T. Lee et al. / Environmental Research 105 (2007) 409–413
Fig. 1. Mean concentrations of PM10 in Seoul, Korea, 2000–2004.
fractions with coarse ones. The authors reported that the magnitude of enrichment factor (EF) values of the most hazardous metals during the AD period was even smaller than the NAD counterpart. They also reported that the existence of low EF values might be ascribed to elevated input of crystal components such as aluminum, iron, etc. accompanying the AD event. Second, dust storms have a certain characteristic that differentiates them from the non-observable ambient urban air particles. The recent development of weather alarms using satellite remote sensing systems successfully predicts unusual dust events. People can easily recognize the occurrence of AD events from the mass media such as newspapers, TV, radio, etc. Therefore, people may tend to change their outdoor behavioral patterns to reduce or
avoid inhalation of particles, thereby reducing exposure to outdoor air pollution during AD days. The occurrence of AD events in the Korean peninsula is closely related to meteorological conditions, including high surface wind speed and barometric instability (Chun et al., 2001). In our results, however, there was no difference in meteorological variables (air temperature, relative humidity and air pressure) between the two groups of study period (WAD, W/O AD). Many studies have suggested that particulate air pollution has an effect on daily mortality rate. Recent literature on the effects of air pollution on daily death rate in Seoul, Korea, from the 1990s reported a 1.3–3.7% increase in the rate of all-cause mortality for inter-quartile increases in PM10 concentration (Lee et al., 1999a, b; Lee and
ARTICLE IN PRESS J.T. Lee et al. / Environmental Research 105 (2007) 409–413
Schwartz, 1999). By comparison, our results highlighted a contrast between increases in mortality of 0.7% per IQR (41.49 mg/m3) rise in PM10 in the WAD model versus increases in mortality of 1.0% in the W/O AD model. The differential effects on daily mortality for models with and without AD have not been previously identified. Our results show that the effect sizes of air particles on daily mortality in the model W/O AD were larger than those in the model WAD. Therefore, the actual size of the adverse health effect from air particles emitted from a city such as Seoul could be much larger than those estimated risks without controlling for days of AD events. Acknowledgment This study was supported by the Ministry of Environment, Republic of Korea (Eco-technopia 2004). References Chun, Y.S., Boo, K.O., Kim, J.Y., Park, S.U., Lee, M.H., 2001. Synopsis, transport, and physical characteristics of Asian dust in Korea. J. Geophys. Res. 106 (16), 18461–18469. Dominici, F., McDermott, A., Zeger, S.L., Samet, J.M., 2002. On the use of generalized additive models in time-series studies of air pollution and health. Am. J. Epidemiol. 156 (3), 193–203. Donaldson, K., MacNee, W., 1998. The mechanism of lung injury caused by PM10. In: Hester, R.E., Harrison, R.M. (Eds.), Issues in Environmental Science and Technology. The Royal Society of Chemistry, London. Kim, K.H., Kim, M.Y., Shin, J.Y., Choi, K.H., Kang, C.H., 2002. Insights into factors determining the aerosol distribution characteristics of the Asian dust on the basis of the concurrent analysis of PM2.5, PM10, and TSP during the spring season of 2001. J. Kor. Soc. Atmos. Environ. 18 (5), 419–426. Kim, K.H., Choi, G.H., Kang, C.H., Lee, J.H., Kim, J.Y., Youn, Y.H., Lee, S.R., 2003. The chemical composition of fine and coarse particles in relation with the Asian Dust events. Atmos. Environ. 37, 753–765.
413
Kim, K.H., Kim, M.Y., 2003. The effects of Asian Dust on particulate matter fractionation in Seoul, Korea during spring 2001. Chemosphere 51 (8), 707–721. Korea Meteorological Administration (KMA). 2005. Annual Climatological Report. Kwon, H.J., Cho, S.H., Chun, Y.S., Lagarde, F., Pershagen, G., 2002. Effects of the Asian Dust events on daily mortality in Seoul, Korea. Environ. Res. 90, 1–5. Laden, F., Neas, L.M., Dockery, D.W., Schwartz, J., 2000. Association of fine particulate matter from different sources with daily mortality in six US cities. Environ. Health Persp. 108 (10), 941–947. Lee, J.T., Schwartz, J., 1999. Reanalysis of the effects of air pollution on daily mortality in Seoul, Korea: a case-crossover design. Environ. Health Persp. 107 (8), 633–636. Lee, J.T., Dockery, D.W., Kim, C.B., Jee, S.H., Chung, Y., 1999a. A meta-analysis of ambient air pollution in relation to daily mortality in Seoul, 1991–1995. J. Prev. Med. Pub. Health 32 (2), 177–182. Lee, J.T., Shin, D.C., Chung, Y., 1999b. Air pollution and daily mortality in Seoul and Ulsan, Korea. Environ. Health Persp. 107 (2), 149–154. Lei, Y.C., Chan, C.C., Wang, P.Y., Lee, C.T., Cheng, T.J., 2004. Effects of Asian dust event particles on inflammation markers in peripheral blood and bronchoalveolar lavage in pulmonary hypertensive rats. Environ. Res. 95, 71–76. Monn, C., Carabias, V., Junker, M., Waeber, R., Karrer, M., Wanner, H.U., 1997. Small-scale spatial variability of particulate matter o10 mm (PM10) and nitrogen dioxide. Atmos. Environ. 31 (15), 2243–2247. Oberdo¨rster, G., 1996. Significance of particle parameters in the evaluation of exposure-dose-response relationships of inhaled particles. Particul. Sci. Technol. 14, 135–151. Park, J.W., Lim, Y.H., Kyung, S.Y., An, C.H., Lee, S.P., Jeong, S.H., Ju, Y.S., 2005. Effects of ambient particulate matter on peak expiratory flow rates and respiratory symptoms of asthmatics during Asian dust periods in Korea. Respirol 10, 470–476. Schwartz, J., 1996. Air pollution and hospital admissions for respiratory disease. Epidemiology 7, 20–28. Schwartz, J., Norris, G., Larson, T., Sheppard, L., Claiborne, C., Koenig, J., 1999. Episodes of high coarse particle concentrations are not associated with increased mortality. Environ. Health Persp. 107 (5), 339–342. USEPA (US Environmental Protection Agency). 2001. Air quality criteria for particulate matter. EPA 600/P-99/002bB. Office of Research and Development. Wada, Y., 1917. A Report of Historical Observation Record in Korea. Chosun-Chongdokbu, Incheon.
Environmental Research 105 (2007) 409–413 www.elsevier.com/locate/envres
A comparison of mortality related to urban air particles between periods with Asian dust days and without Asian dust days in Seoul, Korea, 2000–2004 Jong-Tae Lee, Ji-Young Son, Yong-Sung Cho Department of Public Health, Graduate School of Hanyang University, Hanyang University, Haengdang-dong 17, Seongdong-gu, Seoul 133-791, Republic of Korea Received 18 January 2007; received in revised form 28 May 2007; accepted 18 June 2007 Available online 20 July 2007
Abstract Recent papers have reported that Asian dust events have been associated with increased risks of all-cause mortality and rates of respiratory illness. The current study was designed to estimate the relative risk of mortality associated with Asian dust events. We used the daily counts of non-accidental deaths, air pollution and meteorological data in Seoul, Korea from 2000 to 2004. We divided all days during the study period into two groups according to the presence or absence of Asian dust events. For each group, we conducted timeseries analysis to estimate the relative risk of total non-accidental death when the concentration of each air pollutant increased by the inter-quartile range (IQR). The average concentrations of every air pollutant on the days without a dust event were lower than those on the days with such an event. We found that the effect sizes of air pollution on daily death rates in the model without Asian dust events were larger than those in the model with Asian dust events, and were statistically significant for all air pollutants (PM10, CO, NO2, and SO2) except for O3. Our results suggest that we are likely to underestimate the risk of urban air particles if we analyze the effect size of air pollution on daily mortality during Asian dust events. We hypothesize that the real health effect is much larger than previous results suggested. r 2007 Elsevier Inc. All rights reserved. Keywords: Asian dust event; Mortality; Urban air particles; Generalized additive model; Underestimation
1. Introduction Asian dust (AD) is the phenomenon whereby particles suspended in ascending air currents in the desert area of the northern part of China diffuse widely, are transported and descend slowly. It is an example of the long-distance transportation of air pollutants that affects the air quality of Korea. Large-scale movement of sand clouds, known as Whang-sa, has an influence over an extensive geographical area, and may result in health effects such as ocular and respiratory irritation (Monn et al., 1997). In Korea, dust-related events for the last two centuries can be found in various kinds of historical records (Wada, 1917). It is, however, only quite recently that AD has Corresponding author. Fax: +822 2298 0248.
E-mail address: [email protected] (J.T. Lee). 0013-9351/$ - see front matter r 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.envres.2007.06.004
drawn much attention due to its increasing frequency and strength (Korea Meteorological Administration (KMA), 2005). Papers over the last few years reported that AD events were associated with an increased risk of all-cause mortality (Kwon et al., 2002) and respiratory symptoms (Park et al., 2005). An animal study also showed that inflammatory markers in the lung and peripheral blood of rats increased with exposure levels of concentrated ambient particles collected during an AD storm in Taiwan (Lei et al., 2004). The occurrence of AD events in the Korean peninsula are closely related to meteorological conditions, including high surface wind speed and barometric instability (Chun et al., 2001). The occurrence of AD events in the Korean peninsula is, therefore, most frequent during spring and irregular during winter and fall. There have been no major dusts events recorded during summer, which is the period of the Asian monsoon with heavy rainfall.
ARTICLE IN PRESS 410
J.T. Lee et al. / Environmental Research 105 (2007) 409–413
The health effects of particulate matter are assumed to depend on both the size and the chemistry of the particles (Oberdo¨rster, 1996; Donaldson and MacNee, 1998). During AD events, highly increased mass concentrations of ambient particles are observed in comparison with those days without AD events. A recent study indicated that during AD events the concentration of larger particles (more than 0.82 mm in aerodynamic diameter) was distinctly increased, but the concentration of smaller particles was reduced. And measurements of optical depth also indicated that the atmosphere is more turbid, with larger particles (Chun et al., 2001) in April 1998 and January 1999. However, the potentially hazardous metal fractions of particles did not change or decreased during AD events compared with usual conditions (Kim et al., 2003). Hence, we estimated the relative risks (RRs) of excess mortality associated with ambient air particles, controlling for AD events during the period 2000–2004. From these results, we can estimate the RR of urban air particles on mortality during the days of AD events separately.
rate after controlling for potential confounding variables. The confounding variables considered in this study were long-term trends and seasonal patterns of air pollution, weather conditions and date. Also, to take the lag effect into consideration, we applied the individual pollutant concentrations of the same day 1–7 lagged days, or moving averages from 2 to 3 days. Akaike’s information criterion (AIC) was used to measure the model’s goodness of fit. All analyses were carried out with S-PLUS 2000 software (Statistical Sciences, Seattle, WA). We analyzed the data with a more stringent convergence criterion than the default settings of GAM to avoid biased estimates of the regression coefficient and standard errors (Dominici et al., 2002). To compare the magnitude of the effect of air pollutants on mortality, we calculated the RR of non-accidental death for inter-quartile changes of concentrations of PM10 and other air pollutants for each group (WAD, W/O AD).
2. Materials and methods
There were 63 AD event days in Seoul between 2000 and 2004 (Table 1). The occurrence of AD events was mostly in March and April. Table 2 shows summary statistics of the variables included in the analysis. We categorized all days during the study period into two groups according to the presence of AD events. The average concentrations of air pollutants between the days without AD events and those days with AD events were not significantly different except for PM10 (Table 2). The t-test results to assess statistical differences indicate that there was no difference in the meteorological variables and total death counts between the two groups (WAD, W/O AD). Fig. 1 shows the daily average PM10 concentration from 2000 to 2004. During
2.1. Data The daily counts of non-accidental deaths (ICD-10 codes, A00-R99) in Seoul between January 1, 2000 and December 31, 2004 were supplied by the National Statistical Office, Republic of Korea. Daily ambient air pollution levels in Seoul were obtained from 27 monitoring stations distributed evenly throughout Seoul and operated by the Department of Environment, Republic of Korea, during the same period. Routine monitoring for sulfur dioxide, nitrogen dioxide, ozone, carbon monoxide, and respirable particles was conducted using standardized reference methods (Lee et al., 1999b). We used 24-h averages as the basic measure in this analysis, except for ozone and carbon monoxide, for which we took the maximum daily 8-h moving average as the exposure index. The National Meteorological Administration of the Republic of Korea provided information on the occurrence of AD events each day, as well as daily weather information, including air temperature, relative humidity and pressure, for Seoul during the same study period.
3. Results
Table 1 Asian dust days in Seoul, Korea, 2000–2004 Year
Date
2000
7 March, 23–24 March, 27–28 March 7–8 April, 23 April, 26–27 April
2001
2 January 3–7 March, 20–25 March 7–12 April, 24–26 April 16–19 May 13–14 December
2002
12–13 January 17–19 March, 21–23 March 8–12 April, 16–17 April 11–12 November
2003
27 March 12–13 April
2004
25 February 10–11 March, 30–31 March 23 April
2.2. Statistical analysis All days during the study period were categorized into two groups according to the presence or absence of AD events. We, therefore, called each group ‘‘with AD days’’ (WAD) and ‘‘without AD days’’ (W/O AD). For each group, we conducted time-series analysis to estimate the RR of total death when the concentration of each air pollutant increased by the inter-quartile range (IQR). Time-series analysis using the generalized additive model (GAM) evaluated the effect of air pollution on mortality
ARTICLE IN PRESS J.T. Lee et al. / Environmental Research 105 (2007) 409–413
411
Table 2 Summary statistics of daily concentrations of air pollutants and total death counts on model with/without Asian dust days included in Seoul, Korea, 2000–2004 Variable
Mean (SD)a Model with Asian dust days included (N ¼ 1827 days)
70.00 (47.80) PM10 (mg/m3) CO (ppm) 0.92 (0.42) NO2 (ppb) 37.34 (12.23) SO2 (ppb) 5.21 (2.18) 26.82 (14.72) O3 (ppb) Temperature (1C) 12.90 (10.09) Humidity (%) 62.46 (14.42) Pressure (hPa) 1016.10 (8.09) Total death counts (per 93.26 (12.02) day) a
Model without Asian dust days included (N ¼ 1764 days)
Asian dust days only (N ¼ 63 days)
65.77 0.92 37.38 5.20 26.70 13.03 62.79 1016.19 93.18
188.49 1.00 36.41 5.62 30.07 9.15 54.27 1013.59 95.38
(33.60) (0.41) (12.22) (2.17) (14.78) (10.19) (14.28) (8.15) (12.08)
(142.85) (0.47) (12.58) (2.35) (12.56) (5.62) (14.90) (5.55) (10.07)
Standard deviation (SD).
AD events, the concentration of PM10 increased greatly compared with those days without AD events, especially during the spring of 2002. The time-series analysis showed that the effect size of PM10 on daily mortality in the model W/O AD was larger than that in the model WAD (Table 3). For PM10, the RR of total non-accidental death was indicated by a 1.0% (95% CI, 0.2–1.8%) increase in the model W/O AD and a 0.7% (95% CI, 0.2–1.3%) increase in the model WAD. In both analyses, the RRs on the best statistically lagged day (the day with minimum AIC) were statistically significant for all air pollutants (PM10, CO, NO2, and SO2), except for O3. For other pollutants, in the WAD models, each increase of IQR in CO, NO2, SO2, and O3 corresponded, respectively, to a 3.3% (95% CI, 2.5–4.1%), 2.4% (95% CI, 1.6–3.1%), 2.5% (95% CI, 1.7–3.3%), and 0.4% (95% CI, 0.6–1.4%) increase. The estimate of percentage increase in the W/O AD model was 3.3% (95% CI, 2.5–4.2%) for CO, 2.5% (95% CI, 1.7–3.3%) for NO2, 2.7% (95% CI, 1.8–3.5%) for SO2, and 0.5% (95% CI, 0.5–1.5%) for O3 (Table 3). 4. Discussion and conclusions This study was designed to estimate the RR of urban air particles on mortality during the days of AD events. We found that the degree of increase in RR in the model W/O AD was larger than that in the model WAD, and statistically significant for all air pollutants (PM10, CO, NO2, and SO2) except for O3. These results could be explained by two factors. The first is that the chemical composition of AD consists of relatively less toxic components for humans. In the case of air particulate matter, excess mortality due to PM2.5 is much higher than that due to PM10 (Schwartz, 1996; USEPA, 2001). Schwartz et al. (1999) reported that the coarse particles from windblown dust are not associated with mortality risk. During AD events, mass concentra-
Table 3 Estimated percentage increases in the risk of total death and 95% confidence intervals for total death by an inter-quartile range (IQR)a using a single-pollutant modelb Air pollutants [lag time]
PM10 (mg/m3) [lag1+2+3] CO (ppm) [lag1] NO2 (ppb) [lag1] SO2 (ppb) [lag1] O3 (ppb) [lag1+2]
Percentage increases Model with Asian dust days included (N ¼ 1827 days)
Model without Asian dust days included (N ¼ 1764 days)
0.7 3.3 2.4 2.5 0.4
1.0 3.3 2.5 2.7 0.5
(0.2, 1.3) (2.5, 4.1) (1.6, 3.1) (1.7, 3.3) (0.6, 1.4)
(0.2, 1.8) (2.5, 4.2) (1.7, 3.3) (1.8, 3.5) (0.5, 1.5)
a The IQR was 41.49 mg/m3 for PM10, 0.54 ppm for CO, 17.93 ppb for NO2, 3.06 ppb for SO2, 21.03 ppb for O3. b The model included the following variables: an intercept, indicator variables for days of the week, and smooth spline functions for date, temperature, humidity, and air pressure.
tions of particulate matter (hourly maximum concentration 606–839 mg/m3 in Seoul, Korea, 2001) are 7–14 times higher than the usual level of particulate matter (daily average 72 mg/m3 in Seoul, Korea, 2000) (Kim and Kim, 2003). However, the relative proportion of fine particulate matter (less than 2.5 mm in aerodynamic diameter) was decreased and the relative concentration of coarse particulate matter (more than 2.5 mm in aerodynamic diameter) was increased (Chun et al., 2001; Kim et al., 2002). This suggests that the ambient air particles during AD events were composed of particles of larger diameter. Also, a recent study reported that dust-borne metals are hazardous to health (Laden et al., 2000). The hazardous metal fractions of particles collected during AD event days did not change or decreased compared with those of particles collected during days without dust event (Kim et al., 2003). Kim and his colleagues analyzed the metallic components by comparing the concentrations in AD with those of a non-AD (NAD) period, and fine particle
ARTICLE IN PRESS 412
J.T. Lee et al. / Environmental Research 105 (2007) 409–413
Fig. 1. Mean concentrations of PM10 in Seoul, Korea, 2000–2004.
fractions with coarse ones. The authors reported that the magnitude of enrichment factor (EF) values of the most hazardous metals during the AD period was even smaller than the NAD counterpart. They also reported that the existence of low EF values might be ascribed to elevated input of crystal components such as aluminum, iron, etc. accompanying the AD event. Second, dust storms have a certain characteristic that differentiates them from the non-observable ambient urban air particles. The recent development of weather alarms using satellite remote sensing systems successfully predicts unusual dust events. People can easily recognize the occurrence of AD events from the mass media such as newspapers, TV, radio, etc. Therefore, people may tend to change their outdoor behavioral patterns to reduce or
avoid inhalation of particles, thereby reducing exposure to outdoor air pollution during AD days. The occurrence of AD events in the Korean peninsula is closely related to meteorological conditions, including high surface wind speed and barometric instability (Chun et al., 2001). In our results, however, there was no difference in meteorological variables (air temperature, relative humidity and air pressure) between the two groups of study period (WAD, W/O AD). Many studies have suggested that particulate air pollution has an effect on daily mortality rate. Recent literature on the effects of air pollution on daily death rate in Seoul, Korea, from the 1990s reported a 1.3–3.7% increase in the rate of all-cause mortality for inter-quartile increases in PM10 concentration (Lee et al., 1999a, b; Lee and
ARTICLE IN PRESS J.T. Lee et al. / Environmental Research 105 (2007) 409–413
Schwartz, 1999). By comparison, our results highlighted a contrast between increases in mortality of 0.7% per IQR (41.49 mg/m3) rise in PM10 in the WAD model versus increases in mortality of 1.0% in the W/O AD model. The differential effects on daily mortality for models with and without AD have not been previously identified. Our results show that the effect sizes of air particles on daily mortality in the model W/O AD were larger than those in the model WAD. Therefore, the actual size of the adverse health effect from air particles emitted from a city such as Seoul could be much larger than those estimated risks without controlling for days of AD events. Acknowledgment This study was supported by the Ministry of Environment, Republic of Korea (Eco-technopia 2004). References Chun, Y.S., Boo, K.O., Kim, J.Y., Park, S.U., Lee, M.H., 2001. Synopsis, transport, and physical characteristics of Asian dust in Korea. J. Geophys. Res. 106 (16), 18461–18469. Dominici, F., McDermott, A., Zeger, S.L., Samet, J.M., 2002. On the use of generalized additive models in time-series studies of air pollution and health. Am. J. Epidemiol. 156 (3), 193–203. Donaldson, K., MacNee, W., 1998. The mechanism of lung injury caused by PM10. In: Hester, R.E., Harrison, R.M. (Eds.), Issues in Environmental Science and Technology. The Royal Society of Chemistry, London. Kim, K.H., Kim, M.Y., Shin, J.Y., Choi, K.H., Kang, C.H., 2002. Insights into factors determining the aerosol distribution characteristics of the Asian dust on the basis of the concurrent analysis of PM2.5, PM10, and TSP during the spring season of 2001. J. Kor. Soc. Atmos. Environ. 18 (5), 419–426. Kim, K.H., Choi, G.H., Kang, C.H., Lee, J.H., Kim, J.Y., Youn, Y.H., Lee, S.R., 2003. The chemical composition of fine and coarse particles in relation with the Asian Dust events. Atmos. Environ. 37, 753–765.
413
Kim, K.H., Kim, M.Y., 2003. The effects of Asian Dust on particulate matter fractionation in Seoul, Korea during spring 2001. Chemosphere 51 (8), 707–721. Korea Meteorological Administration (KMA). 2005. Annual Climatological Report. Kwon, H.J., Cho, S.H., Chun, Y.S., Lagarde, F., Pershagen, G., 2002. Effects of the Asian Dust events on daily mortality in Seoul, Korea. Environ. Res. 90, 1–5. Laden, F., Neas, L.M., Dockery, D.W., Schwartz, J., 2000. Association of fine particulate matter from different sources with daily mortality in six US cities. Environ. Health Persp. 108 (10), 941–947. Lee, J.T., Schwartz, J., 1999. Reanalysis of the effects of air pollution on daily mortality in Seoul, Korea: a case-crossover design. Environ. Health Persp. 107 (8), 633–636. Lee, J.T., Dockery, D.W., Kim, C.B., Jee, S.H., Chung, Y., 1999a. A meta-analysis of ambient air pollution in relation to daily mortality in Seoul, 1991–1995. J. Prev. Med. Pub. Health 32 (2), 177–182. Lee, J.T., Shin, D.C., Chung, Y., 1999b. Air pollution and daily mortality in Seoul and Ulsan, Korea. Environ. Health Persp. 107 (2), 149–154. Lei, Y.C., Chan, C.C., Wang, P.Y., Lee, C.T., Cheng, T.J., 2004. Effects of Asian dust event particles on inflammation markers in peripheral blood and bronchoalveolar lavage in pulmonary hypertensive rats. Environ. Res. 95, 71–76. Monn, C., Carabias, V., Junker, M., Waeber, R., Karrer, M., Wanner, H.U., 1997. Small-scale spatial variability of particulate matter o10 mm (PM10) and nitrogen dioxide. Atmos. Environ. 31 (15), 2243–2247. Oberdo¨rster, G., 1996. Significance of particle parameters in the evaluation of exposure-dose-response relationships of inhaled particles. Particul. Sci. Technol. 14, 135–151. Park, J.W., Lim, Y.H., Kyung, S.Y., An, C.H., Lee, S.P., Jeong, S.H., Ju, Y.S., 2005. Effects of ambient particulate matter on peak expiratory flow rates and respiratory symptoms of asthmatics during Asian dust periods in Korea. Respirol 10, 470–476. Schwartz, J., 1996. Air pollution and hospital admissions for respiratory disease. Epidemiology 7, 20–28. Schwartz, J., Norris, G., Larson, T., Sheppard, L., Claiborne, C., Koenig, J., 1999. Episodes of high coarse particle concentrations are not associated with increased mortality. Environ. Health Persp. 107 (5), 339–342. USEPA (US Environmental Protection Agency). 2001. Air quality criteria for particulate matter. EPA 600/P-99/002bB. Office of Research and Development. Wada, Y., 1917. A Report of Historical Observation Record in Korea. Chosun-Chongdokbu, Incheon.