Short-term exposure to ambient particulate matter and emergency ambulance dispatch for acute illness in Japan

Science of the Total Environment 566–567 (2016) 528–535

Contents lists available at ScienceDirect

Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv

Short-term exposure to ambient particulate matter and emergency ambulance dispatch for acute illness in Japan Saira Tasmin a,⁎, Kayo Ueda a, Andrew Stickley b, Shinya Yasumoto c, Vera Ling Hui Phung a, Mizuki Oishi a, Shusuke Yasukouchi a, Yamato Uehara a, Takehiro Michikawa d, Hiroshi Nitta d a

Department of Environmental Engineering, Graduate School Engineering, Kyoto University, Kyoto, Japan Stockholm Center on Health and Social Change (Scohost), Södertörn University, 141 89 Huddinge, Sweden Kinugasa Research Organization, Ritsumeikan University, Japan d Center for Environmental Health Sciences, National Institute for Environmental Studies (NIES), Japan b c

H I G H L I G H T S • Examined the association between short-term exposure to ambient suspended particulate matter and emergency ambulance dispatches for acute illness, a relatively new indicator for evaluating the health effects of air pollution. • Using ambulance dispatch data, the present study demonstrated an increased risk of dispatches for acute illness associated with short-term exposure to suspended particulate matter. • There was the significant effect modification in this association by the type of medical condition, with the effects being stronger for less severe medical conditions.

a r t i c l e

i n f o

Article history: Received 27 January 2016 Received in revised form 2 May 2016 Accepted 8 May 2016 Available online 26 May 2016 Editor: D. Barcelo Keywords: Suspended particulate matter (SPM) Emergency ambulance dispatch (EAD)

G R A P H I C A L

A B S T R A C T

Figure shows area-specific and combined relative risk (RR) and 95% CIs of emergency ambulance dispatch for acute illness associated with a 10 μg/m3 increase in suspended particulate matter (SPM) at lag 0–1

a b s t r a c t Short-term exposure to air pollution may be linked to negative health outcomes that require an emergency medical response. However, few studies have been undertaken on this phenomenon to date. The aim of this study therefore was to examine the association between short-term exposure to ambient suspended particulate matter (SPM) and emergency ambulance dispatches (EADs) for acute illness in Japan. Daily EAD data, daily mean SPM and meteorological data were obtained for four prefectures in the Kanto region of Japan for the period from 2007 to 2011. The area-specific association between daily EAD for acute illness and SPM was explored using generalized linear models while controlling for ambient temperature, relative humidity, seasonality, long-term trends, day of the week and public holidays. Stratified analyses were conducted to evaluate the modifying effects of age, sex and medical conditions. Area-specific estimates were combined using metaanalyses.

Abbreviations: WHO, World Health Organization; PM, particulate matter; EAD, emergency ambulance dispatch; SPM, suspended particulate matter; PM10, particulate matter with diameter b10 μm; FDMA, Japanese Fire and Disaster Management Agency; Ox, photochemical oxidants; NO2, nitrogen dioxide; SO2, sulfur dioxide; GLM, generalized linear models; PACF, partial autocorrelation function; DOW, day of the week; df, degree of freedom; RR, relative risk; CI, confidence interval; PAF, population attributable fraction. ⁎ Corresponding author at: Room 363, Cluster C1-3, Katsura campus, Nishikyo-ku, Kyoto 615-8540, Japan. E-mail address: [email protected] (S. Tasmin).

http://dx.doi.org/10.1016/j.scitotenv.2016.05.054 0048-9697/© 2016 Elsevier B.V. All rights reserved.

S. Tasmin et al. / Science of the Total Environment 566–567 (2016) 528–535 Short-term exposure Japan

529

For the total study period the mean level of SPM was 23.7 μg/m3. In general, higher SPM was associated with a significant increase in EAD for acute illness [estimated pooled relative risk (RR): 1.008, 95% CI: 1.007 to 1.010 per 10 μg/m3 increase in SPM at lag 0–1]. The effects of SPM on EAD for acute illness were significantly greater for moderate/mild medical conditions (e.g. cases that resulted in b 3 weeks hospitalization or no hospitalization) when compared to severe medical conditions (e.g. critical cases, and cases that led to N 3 weeks hospitalization or which resulted in death). Using EAD data, this study has shown the adverse health effects of ambient air pollution. This highlights the importance of reducing the level of air pollution in order to maintain population health and well-being. © 2016 Elsevier B.V. All rights reserved.

1. Introduction According to the World Health Organization (WHO), air pollution has become the world's single biggest environmental health risk. An estimated 3.7 million deaths occurred worldwide due to ambient air pollution in 2012 (Bell et al., 2013). Among the air pollutants, particulate matter (PM) affects more people than any other pollutant. The adverse effects of short-term exposure to both coarse and fine PM are well documented in epidemiological studies and stretch across a range of health outcomes including mortality (Adar et al., 2014; Atkinson et al., 2014). The majority of the studies that have found an association between ambient PM and adverse health outcomes have been conducted in North America and Europe (Atkinson et al., 2014; Peters et al., 2000) with fewer studies occurring in Asia (Lee et al., 2015). This may be an important gap in the research as the chemical composition of PM can vary with regard to its major emission sources and atmospheric conditions (Arruti et al., 2011; Mues et al., 2013). Indeed, epidemiological studies have shown that the levels and composition of PM vary significantly across regions and that these differences in composition are directly linked to variations in the adverse effects on human health (Levy et al., 2012; Samoli et al., 2005), making regional assessment of this association essential. Studies showing the adverse health outcomes resulting from short-term exposure to ambient PM have mostly used mortality data (Peters et al., 2000; Samoli et al., 2013) or hospital-based data, such as hospital admissions or emergency room visits for non-fatal health outcomes (Qiu et al., 2014; Stafoggia et al., 2013; Zheng et al., 2015). Some studies have also made use of alternative data sources such as telephone calls for asthma (Laurent et al., 2008), doctors' house calls (Chardon et al., 2007), and information on doctor consultations (Hajat et al., 2002). In addition, more recently, emergency ambulance dispatch (EAD) data has begun to be used as a proxy for acute health outcomes. It has been suggested that these data may serve as a particularly useful resource to examine the health effects of environmental exposures (Alessandrini et al., 2011; Ueda et al., 2012). Moreover, very recent epidemiological studies have shown that short-term exposure to both fine and coarse PM is associated with an increase in EADs (Michikawa et al., 2015a; Michikawa et al., 2015b). However, as yet, research using EAD data is still comparatively limited. For example, although previous studies that have used EAD data have focused on different indices of PM exposure such as coarse PM (Michikawa et al., 2015b) and fine PM (Michikawa et al., 2015a) as well as PM with a diameter b 10 μm (PM10) (Sajani et al., 2014), not all indices of PM have been investigated. This is important as previous toxicological studies have shown that PM-induced toxic effects are often dependent upon the specific characteristics of PM, including its size and composition (Mirowsky et al., 2013). In Japan, for instance, although suspended particulate matter (SPM), i.e. particles with a diameter of less than approximately 7 μm, have been monitored under the Japanese Air Quality Standard since 1972 (Japanese Ministry of Environment, 2009), up until now, no studies have examined the association between short-term exposure to SPM and EAD.

Given this, the aim of the present study was to examine the effect of short-term exposure to SPM on EAD for acute illness in four prefectures in central Japan. We further investigated whether age, sex and severity of medical condition could modify the effect of short-term exposure to SPM on EAD for acute illness in this setting. 2. Methods 2.1. Setting The present study was conducted in 20 areas across four prefectures (Ibaraki, Saitama, Chiba and Kanagawa) in the Kanto region, located in the largest island of Japan, covering an area of approximately 17,466 km2 (Fig. 1). The capital city, Tokyo, is located in the Kanto region and this region is the most highly developed, urbanized, and industrialized part of Japan, while the prefectures included in this study are those that surround Tokyo. The weather in this region is generally mild and characterized by a humid subtropical climate with four distinct seasons. There were about 25.4 million residents in these four prefectures according to the 2010 census. Among the prefectures, Kanagawa has the largest population with 9 million people, followed by Saitama with 7.2 million people, Chiba with 6.2 million people, while Ibaraki has a population of approximately 3 million people. Each prefecture is divided into several areas by its prefectural office based on climate and various socioeconomic characteristics. In this study, in order to examine the relationship between SPM and EAD for acute illness with as much precision and accuracy as possible, area-level analyses were performed within each prefecture. 2.2. Outcome data The outcome variable in this study was the daily number of EADs for acute illness. EAD daily data for the period from 1 January 2007 to 31 December 2011 were obtained from the Japanese Fire and Disaster Management Agency (FDMA) for individual areas across the study region. Ambulance services for emergency purposes are provided free of charge by all local governmental fire defense headquarters throughout Japan and anyone can summon an ambulance by making an emergency telephone call (dial 119) (Tanigawa and Tanaka, 2006). Anonymous ambulance dispatch data for the four study prefectures were extracted from the FDMA dataset. For each record, information on the cause of the dispatch, date and time of the event, medical condition, primary diagnosis and basic information regarding the person requiring the dispatch, such as age and sex was available. In this study, only ambulance dispatches for acute illness were extracted from the total record. The medical condition of those transported to hospital was determined by an emergency medical doctor upon their arrival at the hospital. Five categories were used to describe the medical condition of the patients in the FDMA database: dead, critical (for patients with a condition where death is imminent), serious (for patients who were likely to require hospitalization for N3 weeks), moderate (for patients who would require hospitalization for b3 weeks) and mild (for patients who required no hospitalization) (Japanese Fire and Disaster Management Agency,

530

S. Tasmin et al. / Science of the Total Environment 566–567 (2016) 528–535

Fig. 1. Location of fire stations across four study prefectures in the Kanto region, Japan.

2003). Daily ambulance dispatch data were collected from all available fire stations within each area and aggregated. 2.3. Exposure data Air pollution data was obtained from the atmospheric environment database in the National Institute for Environmental Studies. Hourly measurements of SPM, photochemical oxidants (Ox), nitrogen dioxide (NO2), and sulfur dioxide (SO2) were collected from the existing background monitoring stations throughout all 20 areas in the four prefectures. In most countries, PM10 is defined as particles which have a 50% efficiency cut-off at a 10 μm aerodynamic diameter. However, under the Japanese Air Quality Standard, PM is collected with a 100% cutoff point at a 10 μm aerodynamic diameter and referred to as SPM (Sasaki and Sakamoto, 2006). That is why, when compared to PM10, the particle size of SPM is smaller and approximately equivalent to particles with an aerodynamic diameter of 7 μm when based on a 50% cut-off. Mixtures of ozone and other secondary oxidants generated by a photochemical reaction are referred to as Ox and were used in the current study as a proxy for Ozone. The daily 24-hour average concentration of SPM, NO2, and SO2 was calculated from the hourly measurements. An average of the 8-h maximum Ox concentration within the day was regarded as the daily Ox concentration. If any day had missing values for more than 4 h, data for that day were excluded. Air pollution data from all background monitoring stations were available across all the areas in the four prefectures. Each area had at least one background monitoring station and most of the areas had more than one background monitoring station. In the area-specific analysis, we used data from only one monitoring station in each area as the air pollutants were highly correlated within each area. Meteorological information including ambient temperature and relative humidity was obtained from the database of the Japan Meteorological Agency. Daily (24-hour) mean ambient temperature and relative humidity values were calculated using hourly measurements. Ethical approval for this study was provided by the Ethics Committee of the Graduate School of Engineering at Kyoto University.

2.4. Statistical analysis The statistical analysis was conducted in two phases. In the first part, area-specific time-series models were used for the analysis as EAD, air pollution and meteorological data are linked by date. Poisson regression generalized linear models (GLM) were used to fit the time-series data. A natural cubic spline function was used to control for long-term trends and seasonality, as well as the daily mean temperature and relative humidity. We used 7 degrees of freedom (df) per year for time trends (Dominici et al., 2000). For both temperature and humidity, we controlled for a 4-day moving average (lag 0–3) from the event day (lag 0) to 3 days prior (lag 3) with 3 df using previous literature as a guide (Zhang et al., 2015). Day of the week (DOW), influenza epidemics and public holidays were adjusted for by using dummy variables in the models as these variables may have a potential confounding effect on the association between SPM and EADs for acute illness. Influenza epidemic weeks were defined if the weekly number of influenza cases exceeded the 90th percentile of their distribution during the study period (Ng et al., 2014). The residuals from the basic models were plotted using residual plots and partial autocorrelation function (PACF) plots while undertaking regression diagnostics to check whether there were any discernible patterns or autocorrelation. We examined the association between SPM and EAD for acute illness at different lag structures, including both single day lags (from lag 0 to lag 3) and cumulative lags (the cumulative exposure over the current day and the previous day, lag 0–1). Similar effects were observed for the single day lag 0 and cumulative day lag 0–1. Therefore, we reported the estimated effects of the cumulative lag 0–1 in the results to compare our effect estimates with those from previous studies about EAD data and air pollution (Michikawa et al., 2015a; Sajani et al., 2014). We also performed a stratified analysis to investigate any modifying effect by age, sex, medical condition and diagnosis of specific diseases (i.e. respiratory and cardiovascular). For the medical conditions, we combined the dead, critical and serious categories together due to the small number of cases in these categories and defined this subgroup as severe. We applied a multiple comparison to test the statistical significance of the

S. Tasmin et al. / Science of the Total Environment 566–567 (2016) 528–535

difference between the effects of stratified subgroups (e.g., the estimated effects between children vs. adults) (Zeka et al., 2006). To assess the robustness of the association between SPM and daily EAD for acute illness, we fitted two-pollutant models. In the two-pollutant models, in addition to SPM, NO2, SO2 and Ox were included one at a time as control variables at the same lag. Finally, meta-analyses were performed that combined the mortality effect estimates across the areas while accounting for heterogeneity using the maximum likelihood method (Van Houwelingen et al., 2002). The results are expressed as the relative risk (RR) and 95% confidence interval (CI) for EAD for acute illness associated with a 10-μg/m3 increase in SPM at lag 0–1. All analyses were conducted using STATA 13.1 (Stata Corporation, Texas, USA). To quantify the public health burden, we calculated the population attributable fraction (PAF) using the classical equation (RR-1)/RR (Cardaba Arranz et al., 2014). Then, the annual reduction in EAD for acute illness attributable to a 10 μg/m3 decrease in SPM was estimated by multiplying the PAF by the number of EAD acute illness cases per year (Michikawa et al., 2015a). The number of EAD acute illness cases per year was obtained by averaging the EAD data for acute illness for the years 2007–2011. 3. Results From a total of 4,642,945 EAD records for all cases, 2,886,261 (62%) EAD acute illness cases were extracted for the 20 areas in the four prefectures during the period from 1 January 2007 to 31 December 2011 (Table 1). Table 2 summarizes the basic characteristics of the EAD for acute illness data across all the study areas during the study period. From all EADs for acute illness, 9% were for children (aged b 18 years), 40% were for adults (aged 18–64 years) and 51% were for the elderly (aged ≥ 65 years). Females accounted for just under half of the EAD acute illness records. The overwhelming majority (about 90%) of the EADs for acute illness were for moderate and mild medical conditions. Descriptive statistics for the EADs for acute illness, meteorological factors and daily mean concentrations of air pollutants aggregated by prefecture are shown in Table 3. The daily mean number of EADs for acute illness varied from 164 to 586 across the prefectures. As expected, daily mean EAD values for acute illness varied with population size, with higher values being observed for those prefectures which were larger. The concentration of air pollutants was very similar among the prefectures with mean values ranging from 22.2 μg/m3 to 24.5 μg/m3 for SPM. The mean daily average temperature ranged from 13.6 °C to 16.4 °C, while the mean relative humidity (RH) ranged from 65.3% to 74.4%, reflecting the humid subtropical climate of the Kanto area. Fig. 2 shows the area-specific and the pooled estimated relative risk (RR) associated with a 10 μg/m3 increment in SPM at lag 0–1. SPM was associated with a significant increase in acute illness EADs in most of the

531

Table 2 Basic characteristics of emergency ambulance dispatch for acute illness data in 20 areas across four study prefectures in the Kanto region, Japan (2007–2011). Variables

Frequency

Percent (%)

Age 1 Children (b18 years) 2 Adult (18–64 years) 3 Elderly (≥ 65 years)

247,816 1,156,075 1,482,370

9 40 51

Sex 1 Male 2 Female

1,406,839 1,290,557

52 48

Medical conditiona 1 Dead 2 Critical 3 Serious (hospitalization ≥3 weeks) 4 Moderate (hospitalization b3 weeks) 5 Mild (no hospitalization) 6 Others

57,273 20,706 224,329 1,111,176 1,398,801 1006

2 0.7 8 40 50 0.04

a

Determined by medical doctor upon hospital arrival.

areas. The combined relative risk of ambulance dispatch for acute illness was 1.008 (95% [CI]: 1.007 to 1.010) across all areas for a 10 μg/m3 increase in SPM at lag 0–1. Overall, there was no heterogeneity in the area-specific estimates (p N 0.05 for the Q-statistic). Therefore, only the pooled estimates of the results across all areas will be presented in the subsequent results. The PAF calculated from this combined RR was 0.008. Using this PAF, the reduction in annual EAD for acute illness cases attributable to a 10 μg/m3 decrease in SPM was estimated to be approximately 4581 (estimated range 4013–5715) for the study area. The relative risk (RR) associated with a 10 μg/m3 increment in SPM at lag 0–1 by age, sex, medical condition and disease diagnosis is shown in Table 4. The effect of SPM on acute illness EADs was significant for all age groups and for both males and females. There was no significant difference in the effects of SPM on acute illness EAD by age or sex strata. The effects of higher SPM were significantly larger for moderate and mild medical conditions when compared to severe medical conditions. Indeed, the estimated EAD risk for severe medical conditions was not significantly associated with a 10 μg/m3 increment in SPM (RR: 1.002, 95% CI: 0.997 to 1.007). The subgroup analysis by disease diagnosis revealed a significant association for all acute illness and respiratory diseases but not for cardiovascular diseases. In the two-pollutant models, the effect of SPM on EAD for acute illness was robust after the inclusion of the gaseous pollutants, i.e. NO2, SO2 and Ox (Fig. 3). Fig. 4 shows the results from a sensitivity analysis performed to examine the lag structures. The largest effects were observed at single day lag 0 (RR: 1.009, 95% CI: 1.008 to 1.011) and cumulative day lag 0–1 (RR: 1.008, 95% CI: 1.007 to 1.010).

Table 1 Categorization of emergency ambulance dispatch data across 20 areas in four study prefectures in the Kanto region, Japan (2007–2011). Incidence type Type 1 Type 2 Type 3 Type 4 Type 5 Type 6 Type 7 Type 8 Type 9 Type 10 Type 11 Type 12 Type 13 Type 14 Total n = 4.642,945

Fire Natural disaster Drowning Traffic Industrial accident Sports Injuries Violence Self-injury Acute illness Transferring Transferring with physician Transport of equipment Others (e.g. pregnancy/stillbirth)

Frequency

Percent (%)

6983 906 1467 568,045 46,523 35,356 638,191 40,447 55,083 2,886,261 347,292 19 1 16,371

0.15 0.02 0.03 12.23 1.00 0.76 13.75 0.87 1.19 62.16 7.48 b0.01 b0.01 0.35

532

S. Tasmin et al. / Science of the Total Environment 566–567 (2016) 528–535

Table 3 Summary statistics of daily emergency ambulance dispatches for acute illness and environmental variables across four study prefectures in the Kanto region, Japan (2007–2011). Mean ± SD

EAD for acute illness SPM (μg/m3) NO2 (ppb) SO2 (ppb) a Ox (ppb) Temperature (°C) Relative humidity (%) a

Ibaraki

Saitama

Chiba

Kanagawa

164.20 ± 26.70 22.20 ± 11.46 11.17 ± 4.72 2.00 ± 0.85 42.99 ± 14.70 14.25 ± 8.16 72.46 ± 11.74

436.42 ± 60.58 24.02 ± 11.79 16.97 ± 6.44 1.09 ± 0.62 49.91 ± 17.13 13.56 ± 8.65 74.37 ± 13.56

393.33 ± 62.75 24.42 ± 12.91 12.37 ± 6.08 1.65 ± 1.03 41.51 ± 17.98 16.44 ± 7.69 67.14 ± 14.57

586.69 ± 99.51 24.51 ± 10.45 15.33 ± 6.68 2.29 ± 0.94 50.12 ± 21.79 16.40 ± 7.59 65.34 ± 14.45

Daily maximum 8 h mean concentrations of photochemical oxidants (Ox).

4. Discussion The results of this time-series analysis showed that short-term exposure to ambient SPM was significantly associated with increased EADs for acute illness in 20 areas in four prefectures in the Kanto region of Japan during 2007–2011. The association remained robust after adjustment for additional gaseous pollutants, i.e. NO2, SO2 and Ox. An analysis stratified by the severity of the medical condition revealed that the significant effect of ambient SPM was observed only for cases with less severe (i.e. moderate and mild) medical conditions. An age- and sexstratified analysis revealed no significant effect modification of ambient SPM on EAD for acute illness by either age or sex. EAD is a relatively new indicator for evaluating the health effects of air pollution. Using this data source, we found that SPM was associated with a significant increase in EAD for acute illness at lag 0–1. This result is consistent with the findings from several earlier studies conducted in Japan that used ambulance dispatch data in relation to air pollution which showed that other indices of particulate air pollution, including desert dust, were associated with increased EADs for acute illness (Kashima et al., 2014; Michikawa et al., 2015a; Ueda et al., 2012). Specifically, one study found that a 10 μg/m3 unit increase in PM2.5 at lag 0–1 was associated with a 0.8% increase (OR: 1.008, 95% CI: 1.002 to 1.014) in EAD for acute illness in Fukuoka city, Japan (Michikawa et al., 2015a). The same research group also showed a statistically significant increase in EAD for acute illness associated with coarse PM in those aged 65 and above (OR:1.020, 95% CI: 1.005–1.035) at lag 0–1 in Fukuoka city (Michikawa et al., 2015b). The findings from the current study have confirmed that a similar association exists for SPM and EAD for acute illness but over a much larger geographical region: 20 areas in four

prefectures of the Kanto region, Japan, that cover 121 cities in total. Our results also accord with those from other studies outside Japan. For example, one study conducted in six towns in the Emilia-Romagna region in Italy found that a 10 µg/m3 increase in PM10 at lag 0–1 was associated with a significant increase in EADs for non-traumatic diseases (percentage change: 0.86%, 95% CI: 0.61–1.1%) (Sajani et al., 2014). The specific magnitude of the effect we found for SPM (a 0.8% increase in EAD for acute illness at 0–1 day lag) also tallies with the above result from the Italian study that used a different index of particulate matter. Thus, the results of the current study provide further evidence of the adverse health effects of particulate pollution using EAD as a health indicator. EAD data can serve as an important source of information in epidemiological studies. In particular, EAD data can have a much wider coverage of nonfatal acute health outcomes compared to hospital-based data (Alessandrini et al., 2011). Moreover, it is an especially useful source of data in a country like Japan where there are very few acute health outcome databases available for non-fatal acute health outcomes. Also, EAD data is available in real time in many countries and can be used for surveillance purposes in the future (Sajani et al., 2014). Another interesting finding was the significant effect modification in the association between EAD and ambient SPM by the severity of the medical condition. More specifically, our results showed that SPM was significantly associated with increased EAD only for patients with moderate and mild medical conditions, but not with severe (i.e. serious, critical and fatal) medical conditions. To the best of our knowledge, this study is the first to focus on how different levels of illness severity can modify this association. This is an important finding as previous epidemiological studies (Atkinson et al., 2014; Lee et al., 2015; Lu et al.,

Fig. 2. Area-specific and combined relative risk (RR) and 95% CIs of EAD for acute illness associated with a 10 μg/m3 increase in SPM at lag 0–1.

S. Tasmin et al. / Science of the Total Environment 566–567 (2016) 528–535

533

Table 4 Relative risks with 95% CIs for emergency ambulance dispatch for acute illness per 10 μg/ m3 increase in SPM at lag 0–1 by age, sex, medical condition and diagnosis: pooled estimates of four prefectures in the Kanto region, Japan (2007–2011). RR Age Children (b18 years)

95% CI

1.013

p-Valuea -

(1.010, 1.016) Adult (18–64 years)

1.010

0.49 (1.007, 1.012)

Elderly (≥ 65 years)

1.006

0.10 (1.004, 1.008)

Sex Male

1.010

Fig. 3. Combined relative risks (RR) and 95% CIs of EAD for acute illness for a 10 μg/m3 increase in SPM at lag 0–1 in single and two-pollutant models. -

(1.007, 1.012) Female

1.007

0.10 (1.005, 1.010)

Medical conditionb Severec

1.002

(0.997, 1.007)

Moderate (hospitalization b3 weeks)

1.008

0.03 (1.006, 1.010)

Mild (no hospitalization)

1.010

0.01 (1.008, 1.013)

Diagnosis All acute illness

1.008

(1.007, 1.010)

Respiratory

b0.01

1.018 (1.013, 1.023)

Cardiovascular

b0.01

1.000 (0.996, 1.005)

a Multiple comparison of estimates between the subgroups (for age, sex, medical condition and diagnosis the reference group is indicated by ‘-’). b Determined by medical doctor upon hospital arrival. c Severe medical condition include dead, critical and serious (hospitalization ≥3 weeks) cases.

2015) have highlighted the adverse effects of particulate pollution on both severe (i.e. mortality and hospital admission) and less severe (i.e. respiratory symptoms) outcomes. The reason for this difference is unclear. More evidence will be needed from future studies in order to better explain this phenomenon i.e. whether it is real or possibly an artifact of the way the data were recorded/coded etc. We did not find any significant difference in the effects of SPM on EAD acute illness by age or sex. Children and the elderly are considered to be at greater risk for air pollution-related health effects (Gouveia and Fletcher, 2000; Schwartz, 2004; Zhang et al., 2015). For example, Bell and coauthors found that for a 10 μg/m3 increase in PM10 the elderly (aged ≥ 64 years) had a statistically higher risk of mortality (relative risk [RR]: 0.64%, 95% CI: 0.50, 0.78) compared with younger people (aged b64 years) (RR: 0.34%, 95% CI: 0.25, 0.42) (Bell et al., 2013). However, making direct comparisons between the results obtained in our study and those obtained from other studies is complicated by the difference in health outcomes, exposure indices and statistical methods used. In addition, prior studies have also provided conflicting results about the sex-specific effect modification of short-term exposure to ambient particulate matter on mortality and emergency hospital admissions (Bell et al., 2013). However, in accordance with the results from the current study, earlier studies conducted in Fukuoka city, Japan, that evaluated the effect of both coarse and fine particulate matter on

EADs for acute illness (Michikawa et al., 2015a; Michikawa et al., 2015b) also did not find any sex-specific effect modification. We observed a significant association between SPM and EAD for respiratory diseases, but not for cardiovascular diseases. The significant association between EAD for respiratory diseases and SPM found in our study accords with the findings from previous epidemiological studies that used EAD data or other hospital-based outcomes in relation to PM indices (Delfino et al., 1997; Neuberger et al., 2013). In contrast, our result for cardiovascular diseases conflicts with the finding from an earlier study where a significant association was observed between cardiovascular morbidity and PM indices (Neuberger et al., 2013). However, similar to our findings, previous studies that used EAD data to evaluate the health effects of coarse and fine particles in Japan also found no association between PM indices and cardiovascular disease (Michikawa et al., 2015a; Michikawa et al., 2015b). The special characteristics of the cardiovascular disease pattern in Japan might in part, explain why no association was observed between cardiovascular diseases and SPM in our study. Specifically, unlike in western countries where coronary heart disease is the most common cardiovascular complication, in Japan the incidence of stroke is much higher (Ueshima et al., 2008). Our analysis using two-pollutant models showed that the addition of SO2 and Ox but not NO2 slightly attenuated the risk of SPM compared to in the single pollutant model. With the available data however, it is not possible to determine whether this difference between the oneand two-pollutant models reflects a real difference in toxicity. For example, our results may have been influenced by statistical co-linear effects among the pollutants (Sarnat et al., 2005), while a previous study in Japan suggested that the effects of PM on mortality can be distorted after adjustment for co-pollutants (i.e.NO2, SO2 and Ox) (Ueda et al., 2009). Having said this, studies by Michikawa et al. showed that the effects of coarse and fine PM on EAD in Fukuoka city, Japan were the same after the addition of co-pollutants such as NO2, SO2 and Ox (Michikawa et al., 2015a; Michikawa et al., 2015b). Although it is uncertain what underlies this variation, the current study was conducted in a

Fig. 4. Combined relative risks (RR) and 95% CIs of EAD for acute illness associated with a 10 μg/m3 increase in SPM at different lag days.

534

S. Tasmin et al. / Science of the Total Environment 566–567 (2016) 528–535

different region of Japan and it is possible that the regional difference in the air pollution mix might be important in this context. One of the main strengths of this study was its geographical and temporal coverage - as it comprised an area-specific analysis of multiple years for a large region in the most highly urbanized part of Japan, covering 17,466 km2 and including 25.4 million residents. Previous research has shown that multi-city studies produce more stable results and are less affected by bias compared to small studies (Dominici et al., 2000). As we made use of all of the acute illness EAD data, about 2.88 million cases, for the 20 areas in the four study prefectures, the possibility of any selection bias was minimal. Nevertheless, this study has several limitations that should also be acknowledged. First, we cannot rule out the possibility of exposure misclassification. Like most previous air pollution epidemiology time-series studies we used air pollution concentrations from fixed monitoring sites in each area which may have resulted in exposure misclassification because the air pollution level reading obtained from a fixed monitoring site may differ from the actual level of individual exposure (Holliday et al., 2014; Schwartz et al., 2007; Sheppard et al., 2012). However, in our study, the air pollutant concentrations obtained from the available monitoring stations within each area had a very high spatial correlation. Furthermore, there is some evidence that exposure misclassification may result in estimates being reduced (Goldman et al., 2011; Strickland et al., 2015). Second, the categorization of medical conditions was initially made by an emergency medical doctor, which in some cases, might have been subject to later change especially in relation to the duration of the hospitalization. It is thus possible that some of the cases in the analysis were misclassified. 5. Conclusions In conclusion, using ambulance dispatch data, the present study demonstrated an increased risk of EAD for acute illness associated with short-term exposure to SPM. Furthermore, the results suggest that the effects of SPM may only be evident for emergency ambulance dispatch for less severe medical conditions. These results not only highlight the importance of reducing the level of air pollution in order to maintain population health and well-being, but also suggest that service providers should be alerted to the increased demand for emergency medical services in the presence of higher levels of air pollution. Competing interests The authors declare no conflict of interest. Acknowledgments This research was supported by the Environment Research and Technology Development Fund (S-12) of the Ministry of the Environment of Japan. References Adar, S.D., Filigrana, P.A., Clements, N., Peel, J.L., 2014. Ambient coarse particulate matter and human health: a systematic review and meta-analysis. Curr Environ. Health. Rep. 1, 258–274. Alessandrini, E., Zauli Sajani, S., Scotto, F., Miglio, R., Marchesi, S., Lauriola, P., 2011. Emergency ambulance dispatches and apparent temperature: a time series analysis in Emilia-Romagna, Italy. Environ. Res. 111, 1192–1200. Arruti, A., Fernandez-Olmo, I., Irabien, A., 2011. Regional evaluation of particulate matter composition in an Atlantic coastal area (Cantabria region, northern Spain): spatial variations in different urban and rural environments. Atmos. Res. 101, 280–293. Atkinson, R.W., Kang, S., Anderson, H.R., Mills, I.C., Walton, H.A., 2014. Epidemiological time series studies of PM2.5 and daily mortality and hospital admissions: a systematic review and meta-analysis. Thorax 69, 660–665. Bell, M.L., Zanobetti, A., Dominici, F., 2013. Evidence on vulnerability and susceptibility to health risks associated with short-term exposure to particulate matter: a systematic review and meta-analysis. Am. J. Epidemiol. 178, 865–876. Cardaba Arranz, M., Munoz Moreno, M.F., Armentia Medina, A., Alonso Capitan, M., Carreras Vaquer, F., Almaraz, G.A., 2014. Health impact assessment of air pollution in Valladolid, Spain. BMJ Open. 4, e005999.

Chardon, B., Lefranc, A., Granados, D., Gremy, I., 2007. Air pollution and doctors' house calls for respiratory diseases in the greater Paris area (2000-3). Occup. Environ. Med. 64, 320–324. Delfino, R.J., Murphy-Moulton, A.M., Burnett, R.T., Brook, J.R., Becklake, M.R., 1997. Effects of air pollution on emergency room visits for respiratory illnesses in Montreal, Quebec. Am. J. Respir. Crit. Care Med. 155, 568–576. Dominici, F., Samet, J.M., Zeger, S.L., 2000. Combining evidence on air pollution and daily mortality from the 20 largest US cities: a hierarchical modelling strategy. J. Roy. Stat. Soc. a Sta. 163, 263–284. Goldman, G.T., Mulholland, J.A., Russell, A.G., Strickland, M.J., Klein, M., Waller, L.A., et al., 2011. Impact of exposure measurement error in air pollution epidemiology: effect of error type in time-series studies. Environ. Health 10, 61. Gouveia, N., Fletcher, T., 2000. Time series analysis of air pollution and mortality: effects by cause, age and socioeconomic status. J. Epidemiol. Community Health 54, 750–755. Hajat, S., Anderson, H.R., Atkinson, R.W., Haines, A., 2002. Effects of air pollution on general practitioner consultations for upper respiratory diseases in London. Occup. Environ. Med. 59, 294–299. Holliday, K.M., Avery, C.L., Poole, C., McGraw, K., Williams, R., Liao, D., et al., 2014. Estimating personal exposures from ambient air pollution measures: using meta-analysis to assess measurement error. Epidemiology 25, 35–43. Kashima, S., Yorifuji, T., Suzuki, E., 2014. Asian dust and daily emergency ambulance calls among elderly people in Japan: an analysis of its double role as a direct cause and as an effect modifier. J. Occup. Environ. Med. 56, 1277–1283. Laurent, O., Pedrono, G., Segala, C., Filleul, L., Havard, S., Deguen, S., et al., 2008. Air pollution, asthma attacks, and socioeconomic deprivation: a small-area case-crossover study. Am. J. Epidemiol. 168, 58–65. Lee, H., Honda, Y., Hashizume, M., Guo, Y.L., Wu, C.F., Kan, H., et al., 2015. Short-term exposure to fine and coarse particles and mortality: a multicity time-series study in East Asia. Environ. Pollut. 207, 43–51. Levy, J.I., Diez, D., Dou, Y., Barr, C.D., Dominici, F., 2012. A meta-analysis and multisite time-series analysis of the differential toxicity of major fine particulate matter constituents. Am. J. Epidemiol. 175, 1091–1099. Lu, F., Xu, D., Cheng, Y., Dong, S., Guo, C., Jiang, X., et al., 2015. Systematic review and metaanalysis of the adverse health effects of ambient PM2.5 and PM10 pollution in the Chinese population. Environ. Res. 136, 196–204. Michikawa, T., Ueda, K., Takeuchi, A., Kinoshita, M., Hayashi, H., Ichinose, T., et al., 2015a. Impact of short-term exposure to fine particulate matter on emergency ambulance dispatches in Japan. J. Epidemiol. Community Health 69, 86–91. Michikawa, T., Ueda, K., Takeuchi, A., Tamura, K., Kinoshita, M., Ichinose, T., et al., 2015b. Coarse particulate matter and emergency ambulance dispatches in Fukuoka, Japan: a time-stratified case-crossover study. Environ. Health Prev. Med. 20, 130–136. Mirowsky, J., Hickey, C., Horton, L., Blaustein, M., Galdanes, K., Peltier, R.E., et al., 2013. The effect of particle size, location and season on the toxicity of urban and rural particulate matter. Inhal. Toxicol. 25, 747–757. Mues, A., Manders, A., Schaap, M., van Ulft, L.H., van Meijgaard, E., Builtjes, P., 2013. Differences in particulate matter concentrations between urban and rural regions under current and changing climate conditions. Atmos. Environ. 80, 232–247. Neuberger, M., Moshammer, H., Rabczenko, D., 2013. Acute and subacute effects of urban air pollution on cardiopulmonary emergencies and mortality: time series studies in Austrian cities. Int. J. Environ. Res. Public Health 10, 4728–4751. Ng, C.F., Ueda, K., Takeuchi, A., Nitta, H., Konishi, S., Bagrowicz, R., et al., 2014. Sociogeographic variation in the effects of heat and cold on daily mortality in Japan. J. Epidemiol. 24, 15–24. Peters, A., Skorkovsky, J., Kotesovec, F., Brynda, J., Spix, C., Wichmann, H.E., et al., 2000. Associations between mortality and air pollution in Central Europe. Environ. Health Perspect. 108, 283–287. Qiu, H., Tian, L.W., Pun, V.C., Ho, K.F., Wong, T.W., Yu, I.T., 2014. Coarse particulate matter associated with increased risk of emergency hospital admissions for pneumonia in Hong Kong. Thorax 69, 1027–1033. Sajani, S.Z., Alessandrini, E., Marchesi, S., Lauriola, P., 2014. Are day-to-day variations of airborne particles associated with emergency ambulance dispatches? Int. J. Occup. Environ. Health 20, 71–76. Samoli, E., Analitis, A., Touloumi, G., Schwartz, J., Anderson, H.R., Sunyer, J., et al., 2005. Estimating the exposure-response relationships between particulate matter and mortality within the APHEA multicity project. Environ. Health Perspect. 113, 88–95. Samoli, E., Stafoggia, M., Rodopoulou, S., Ostro, B., Declercq, C., Alessandrini, E., et al., 2013. Associations between fine and coarse particles and mortality in Mediterranean cities: results from the MED-PARTICLES project. Environ. Health Perspect. 121, 932–938. Sarnat, J.A., Brown, K.W., Schwartz, J., Coull, B.A., Koutrakis, P., 2005. Ambient gas concentrations and personal particulate matter exposures: implications for studying the health effects of particles. Epidemiology 16, 385–395. Sasaki, K., Sakamoto, K., 2006. Diurnal characteristics of suspended particulate matter and PM2.5 in the urban and suburban atmosphere of the Kanto plain, Japan. Water Air Soil Pollut. 171, 29–47. Schwartz, J., 2004. Air pollution and children's health. Pediatrics 113, 1037–1043. Schwartz, J., Sarnat, J.A., Coull, B.A., Wilson, W.E., 2007. Effects of exposure measurement error on particle matter epidemiology: a simulation using data from a panel study in Baltimore. MD. J. Expo. Sci. Environ. Epidemiol. 17 (Suppl. 2), S2–10. Sheppard, L., Burnett, R.T., Szpiro, A.A., Kim, S.Y., Jerrett, M., Pope 3rd, C.A., et al., 2012. Confounding and exposure measurement error in air pollution epidemiology. Air Qual. Atmos. Health 5 (2), 203–216. Stafoggia, M., Samoli, E., Alessandrini, E., Cadum, E., Ostro, B., Berti, G., et al., 2013. Shortterm associations between fine and coarse particulate matter and hospitalizations in southern Europe: results from the MED-PARTICLES project. Environ. Health Perspect. 121, 1026–1033.

S. Tasmin et al. / Science of the Total Environment 566–567 (2016) 528–535 Strickland, M.J., Gass, K.M., Goldman, G.T., Mulholland, J.A., 2015. Effects of ambient air pollution measurement error on health effect estimates in time-series studies: a simulation-based analysis. J. Expo. Sci. Environ. Epidemiol. 25, 160–166. Tanigawa, K., Tanaka, K., 2006. Emergency medical service systems in Japan: past, present, and future. Resuscitation 69, 365–370. Ueda, K., Nitta, H., Ono, M., Takeuchi, A., 2009. Estimating mortality effects of fine particulate matter in Japan: a comparison of time-series and case-crossover analyses. J. Air Waste Manage. Assoc. 59, 1212–1218. Ueda, K., Shimizu, A., Nitta, H., Inoue, K., 2012. Long-range transported Asian dust and emergency ambulance dispatches. Inhal. Toxicol. 24, 858–867. Ueshima, H., Sekikawa, A., Miura, K., Turin, T.C., Takashima, N., Kita, Y., et al., 2008. Cardiovascular disease and risk factors in Asia: a selected review. Circulation 118, 2702–2709. Van Houwelingen, H.C., Arends, L.R., Stijnen, T., 2002. Advanced methods in meta-analysis: multivariate approach and meta-regression. Stat. Med. 21, 589–624. Zeka, A., Zanobetti, A., Schwartz, J., 2006. Individual-level modifiers of the effects of particulate matter on daily mortality. Am. J. Epidemiol. 163, 849–859.

535

Zhang, Y., Wang, S.G., Ma, Y.X., Shang, K.Z., Cheng, Y.F., Li, X., et al., 2015. Association between ambient air pollution and hospital emergency admissions for respiratory and cardiovascular diseases in Beijing: a time series study. Biomed. Environ. Sci. 28, 352–363. Zheng, X.Y., Ding, H., Jiang, L.N., Chen, S.W., Zheng, J.P., Qiu, M., et al., 2015. Association between air pollutants and asthma emergency room visits and hospital admissions in time series studies: a systematic review and meta-analysis. PLoS One 10, e0138146.

Web references Japanese Ministry of the Environment, 2009r. Environmental Quality Standards in JapanAir Quality. http://www.env.go.jp/en/air/aq/aq.html. Japanese Fire and Disaster Management Agency, 2003r. Working Group Report for Emergency Ambulance Service Statistics. http://www.fdma.go.jp/html/intro/form/pdf/ kyu-tokei/210217-1_2_6kyu-tokei.pdf.