Effects of air pollution on respiratory hospital admissions in İstanbul, Turkey, 2013 to 2015

Accepted Manuscript Effects of air pollution on respiratory hospital admissions in İstanbul, Turkey, 2013 to 2015 Özkan Çapraz, Ali Deniz, Nida Doğan PII:

S0045-6535(17)30640-9

DOI:

10.1016/j.chemosphere.2017.04.105

Reference:

CHEM 19167

To appear in:

ECSN

Received Date: 12 September 2016 Revised Date:

1 April 2017

Accepted Date: 22 April 2017

Please cite this article as: Çapraz, Ö., Deniz, A., Doğan, N., Effects of air pollution on respiratory hospital admissions in İstanbul, Turkey, 2013 to 2015, Chemosphere (2017), doi: 10.1016/ j.chemosphere.2017.04.105. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

PM 10

SC

RI PT

today’s dangerous: air pollution !!! PM 2. 5

PM 10

NO2

AC C

EP

TE D

M AN U

NO 2

PM 2. 5

ACCEPTED MANUSCRIPT 1

Effects of Air Pollution on Respiratory Hospital Admissions in İstanbul, Turkey, 2013

2

to 2015

3

Özkan Çapraz1, Ali Deniz2∗, Nida Doğan2 1

5

2

Marmara Clean Air Center, Şişli, İstanbul, Turkey, [email protected], İstanbul Technical University, Faculty of Aeronautics and Astronautics,

RI PT

4

6

Department of Meteorology, Maslak, İstanbul, Turkey [email protected],

7

[email protected]

Abstract

M AN U

9

SC

8

We examined the associations between the daily variations of air pollutants and

11

hospital admissions for respiratory diseases in İstanbul, the largest city of Turkey. A

12

time series analysis of counts of daily hospital admissions and outdoor air pollutants

13

was performed using single-pollutant Poisson generalized linear model (GLM) while

14

controlling for time trends and meteorological factors over a 3-year period (2013 -

15

2015) at different time lags (0-9 days). Effects of the pollutants (Excess Risk, ER) on

16

current-day (lag 0) hospital admissions to the first ten days (lag 9) were determined.

17

Data on hospital admissions, daily mean concentrations of air pollutants of PM10,

18

PM2.5 and NO2 and daily mean concentrations of temperature and humidity of

19

İstanbul were used in the study. The analysis was conducted among people of all

20

ages, but also focused on different sexes and different age groups including children

21

(0 – 14 years), adults (35 – 44 years) and elderly (≥65 years). We found significant

22

associations between air pollution and respiratory related hospital admissions in the

23

city. Our findings showed that the relative magnitude of risks for an association of the

24

pollutants with the total respiratory hospital admissions was in the order of: PM2.5,

25

NO2, and PM10. The highest association of each pollutant with total hospital

26

admission was observed with PM2.5 at lag 4 (ER = 1.50; 95% CI =1.09 − 1.99), NO2

27

at lag 4 (ER = 1.27; 95% CI = 1.02−1.53) and PM10 at lag 0 (ER = 0.61; 95% CI

28

=0.33 − 0.89) for an increase of 10 µg/m3 in concentrations of the pollutants. In

AC C

EP

TE D

10

∗ Corresponding Author: Ali DENİZ, İstanbul Technical University, Faculty of Aeronautics and Astronautics Engineering, Department of Meteorology, 34469, Maslak, İstanbul, Turkey, [email protected], Phone: +90 (212) 285 31 22, fax: +90 (212) 285 31 22

1

ACCEPTED MANUSCRIPT 29

conclusion, our study showed that short-term exposure to air pollution was positively

30

associated with increased respiratory hospital admissions in İstanbul during 2013-

31

2015. As the first air pollution hospital admission study using GLM in İstanbul, these

32

findings may have implications for local environmental and social policies.

33

Keywords: Air pollution, Respiratory diseases, Hospital admission, Generalized

35

linear model (GLM), İstanbul, Turkey

RI PT

34

36

1. Introduction

38

Air pollution has become one of the most serious environmental concerns for urban

39

areas throughout the world over the past several decades. Various epidemiological

40

studies in recent years reported associations between elevated air pollution levels and

41

increased death and hospitalization rates due to respiratory and cardiovascular

42

diseases (Samet et al., 2000, Atkinson et al., 2001, Katsouyanni et al., 2001, Chen et

43

al., 2004, Pope and Dockery, 2006, Dominici et al. 2006, Samoli et al., 2008,

44

Mahiyuddin et al., 2013, Li et al., 2016). Some epidemiological studies showed that

45

air pollution affects human health even air pollutant concentrations are below the air

46

quality standards (Peters et al., 2001, Pope III et al., 2002, Peng et al., 2006).

47

In Turkey, many studies have been published about air pollution and some of them

48

linked air pollution to adverse population health in cities based on data that were

49

collected in the 1990s and 2000s (Keles and Ilıcali, 1998; Savaş et al., 2002,

50

Hapçıoğlu et al., 2005, Tomaç et al., 2005, Tecer et al., 2008, Toros et al. 2013,

51

Ozcan and Cubukcu, 2015, Çapraz et al., 2016). For İzmir, it was reported that there

52

is a statistically significant relation between the number of asthma cases and the level

53

of air pollution urban air pollution in the six core districts in Izmir between the years

54

2007 and 2010 (Özcan and Çubukcu, 2015). The findings indicate that high levels of

AC C

EP

TE D

M AN U

SC

37

2

ACCEPTED MANUSCRIPT sulfur dioxide (SO2) and particulate matter (PM10) were significant causes of asthma.

56

In a study undertaken in İstanbul, significant associations between air pollution and

57

daily mortality from cardiovascular disease, respiratory diseases, and total non-

58

accidental causes were found over a 6-year period (2007 - 2012). An increase of 10

59

µg/m3 in concentrations of PM10, SO2 and NO2 over 10 days of lag corresponded to

60

RR = 1.0018 (95% CI = 0.9569−1.0489), RR = 1.2116 (95% CI = 0.9727−1.5091)

61

and RR = 1.0253 (95% CI = 0.9829−1.0694) increase of respiratory mortality,

62

respectively (Çapraz et al., 2016).

63

Air pollution is an important risk factor for respiratory health effects. Respiratory

64

diseases and the related mortality have been increasingly associated with exposure to

65

air pollutants. Sensitive and vulnerable groups such as pregnant women, children, the

66

elderly and those already suffering from respiratory and other serious illnesses or

67

from low income groups are especially affected from air pollution. Studies have

68

shown that the number of respiratory diseases in children and elderly increases due to

69

the higher air pollution concentrations (Braga et al., 1999, Braga et al., 2001, Viegi et

70

al., 2009). According to these studies, children are more susceptible because they

71

need twice the amount of air inhaled by adults and the elderly are more affected due

72

to their weak immune and respiratory systems and they have been exposed to a great

73

amount of air pollution throughout their lives.

74

In the current study which is the first air pollution and respiratory hospital admission

75

study using GLM in Turkey, we conducted a time-series study of the relationship

76

between daily mean levels of air pollutants (PM10, PM2.5 and NO2) and daily hospital

77

admissions for respiratory conditions in İstanbul, using Poisson regression in

78

generalized linear model (GLM) while controlling for time trends and meteorological

AC C

EP

TE D

M AN U

SC

RI PT

55

3

ACCEPTED MANUSCRIPT 79

factors. The analysis was conducted among people of all ages, but also focused on

80

different sexes and subjects in the range 0 – 14 years, 35 – 44 years and ≥65 years.

81

2. METHODOLOGY

83

2.1 Study area

84

İstanbul (29º N and 41º E) is located in the northwestern part of Turkey, separated as

85

Asian and European parts by Bosporus strait which is approximately 30 km in length.

86

It is the most populated and industrialized city in Turkey, with an area of 5 400 km2.

87

The city also forms one of the largest urban settlements in Europe with a population

88

of 14.1 million. The climate is moderate in İstanbul. The average temperature is

89

22.7°C in summer and 6.5°C in winter (Unal et al., 2011). The temperatures can drop

90

below zero in winter season.

91

Today, İstanbul has especially particulate matter and NO2 pollution depending on

92

different emission sources (İncecik and İm, 2012, Özdemir et al., 2014). Since the

93

mid-1990’s, SO2 concentrations are gradually decreased to very low levels in the city

94

due to the widespread use of natural gas in heating and industry (İncecik and İm,

95

2012). The major emission sources in the city are motor vehicles, industrial processes,

96

construction activities, residential heating and ship emissions. Traffic emissions in

97

İstanbul have been increased due to the rapid increase in car numbers since 1990s.

98

According to Turkish Statistical Institute, the number of registered vehicles was

99

3.000.000 in January 2016 in İstanbul (Turkish Statistical Institute, 2016). People are

100

exposed daily to continuous particulate matter and NO2 pollution due to the frequent

101

traffic jams in the city. PM10 concentrations show a seasonal pattern with higher

102

concentrations in winter and lower concentrations in summer. Shallow mixing height

103

and adverse weather conditions (high pressure systems and lower wind speeds) lead

AC C

EP

TE D

M AN U

SC

RI PT

82

4

ACCEPTED MANUSCRIPT to elevated air pollution levels in winter season (Unal et al., 2011). Long-range

105

transportation of air pollutants is another factor affecting the air quality of the city.

106

İstanbul is in a common route for air parcels, where air pollutants are carried over

107

European continent crossing over the city to the Asian and Mediterranean regions. Air

108

masses arriving to İstanbul are seasonally dependent and include air pollution

109

originating in European and Black Sea countries during winter, and desert dust loaded

110

air originating in Northern Mediterranean (Saharan) and Middle-Eastern countries

111

during spring (Karaca et al., 2009). Particulate matter concentrations occasionally can

112

reach very high values during the desert dust transportation events.

113

2.2 Data

114

The data of daily respiratory hospital admission counts of residents living in İstanbul

115

from March 1, 2013 to March 31, 2015 (761 days) was obtained from the database of

116

Republic of Turkey Ministry of Health. The selection was made according to the

117

International Classification of Disease, Tenth Revision (ICD-10) by the World Health

118

Organization. Daily hospital admission counts from respiratory diseases (ICD-10:

119

J00-J98) among residents of all ages were considered. In addition, hospital admissions

120

for different sexes and different age groups (children, adults and elderly) were also

121

separately analyzed.

122

Hourly air pollution data, including PM10, PM2.5 and NO2, were obtained from the

123

database of Ministry of Environment and Urbanization, the government agency in

124

charge of collection of air pollution data in Turkey. The daily concentrations for each

125

pollutant were averaged from the available monitoring results of 11 fixed-site stations

126

of Air Quality Monitoring Network in İstanbul (Başakşehir, Mecidiyeköy, Silivri,

127

Şirinevler, Ümraniye, Üsküdar, Kağıthane, Esenyurt, Sultanbeyli, Sultangazi and

128

Esenler) under Republic of Turkey Ministry of Environment and Urbanization (Figure

AC C

EP

TE D

M AN U

SC

RI PT

104

5

ACCEPTED MANUSCRIPT 1). To allow adjustment for the effect of weather on hospital admissions, hourly

130

meteorological data (temperature and relative humidity) were obtained from the Air

131

Quality Monitoring Stations where meteorological measurements are also made. We

132

have used daily means of the pollutants and weather variables calculated from the

133

hourly data to represent the daily reading for Istanbul. The percentage of data for each

134

pollutant was: PM10 (99.9%), PM2.5 (99.9%) and NO2 (99.4%).

135

Figure1.

136 137

2.3. Statistical analysis

138

In this study, Poisson regression in single-pollutant generalized linear model (GLM)

139

with Distributed Lag Model (DLM) was used to estimate the association between

140

pollutants and hospital admissions at different lags. The GLM with Poisson regression

141

consists in relating the response variable (hospital admissions), which can take on

142

only non-negative integers, with the explanatory variables (pollutants concentration).

143

Natural cubic splines were used with the model to adjust for long-term seasonality

144

patterns in respiratory hospital admissions and other time-varying covariates like

145

meteorological parameters that might confound the association between air pollution

146

and admissions. In our study, the time variable (day), daily mean temperature, daily

147

mean humidity, the day of the week, and the holiday indicator were the covariates.

148

Hospital admission data are Poisson distributed and Poisson regression provides an

149

estimation of the relative risk (RR) as RR=exp(β) where β

150

coefficient associated with a unit increment in an air pollutant. The results are

151

presented in terms of the excess risk (ER). ER, the percentage increase in hospital

152

admissions for standard exposure increments was derived from (RR−1)×100%, where

153

RR was the relative risk derived from exp(β). Positive ER values indicate the

AC C

EP

TE D

M AN U

SC

RI PT

129

6

is the regression

ACCEPTED MANUSCRIPT percentage increase (%) in respiratory hospital admissions for a 10 µg/m3 increase in

155

pollutant concentration. Negative ER values indicate that there is no relationship

156

between air pollution increases and hospital admissions.

157

In the studies of air pollution and health, it is common to find a relation between the

158

air pollutants concentration of one day to the health outcomes of some lag days and

159

then the lag days that best fit to data is chosen (Tadano, 2012). In the current study,

160

we have chosen 9 lag days for our model which was the one that best fits to our data.

161

We examined the effects of the pollutants (PM10, PM2.5 and NO2) on respiratory

162

hospital admissions for a 10 µg/m3 increase in pollutant concentrations with different

163

lag structures of single day lag from lag 0 to lag 9. Also, the overall cumulative

164

effects of a 10-unit (10 µg/m3) increase in PM10, PM2.5 and NO2 over 9 days of lag

165

(i.e. summing all the contributions up to the maximum lag) were observed for

166

prolonged effects. Air pollution and hospital admission relationship was conducted

167

for genders (male and female) and age groups (0 – 14 years, 35 – 44 years and ≥65

168

years). During the application of the GLM, Akaike Information Criterion (AIC) was

169

used in order to decide which number of degrees of freedom to use. In order to

170

minimize autocorrelation, we have examined the plot of the partial autocorrelation

171

function (PACF) of the residuals in the core model, which would bias the standard

172

errors. Residuals for the lag days were included in the model if the partial ACF plot

173

showed autocorrelations between one day and the previous days.

174

After adjusting the GLM with Poisson regression including all the time trends and

175

explanatory variables and choosing the degrees of freedom that better fits the data; the

176

fitted model was tested using z-test to assure that it is the right one to be applied to the

177

case-study. According to z-test, values of p < 0.05 were considered statistically

178

significant in our study. In order to generate GLM and make calculations, R version

AC C

EP

TE D

M AN U

SC

RI PT

154

7

ACCEPTED MANUSCRIPT 0.99.491 was used (R Development Core Team, 2016). Functions in dlnm and splines

180

package of R were used to build GLMs.

181

3. Results

182

3.1. Descriptive analysis

183

From March 1, 2013 to March 31, 2015 (761 days), a total of 12,884,628 hospital

184

admissions for respiratory conditions were recorded, with 4,652,912 in children (0-14

185

years), 1,279,611 in adults (35-44 years), and 1,929,661 in elderly (≥65 years) in

186

İstanbul. On average, there were approximately 17,160 hospital admissions per day

187

(Table 1).

188

Table 1.

189

Daily admission counts for men (n = 5,892,222) accounted for 45.7% and daily

190

admission counts for women (n = 6,992,406) accounted for 54.3% of the total hospital

191

admissions. During our study period, the yearly mean concentrations were 56.3µg/m3

192

for PM10, 41.8 µg/m3 for NO2 and 29.0 µg/m3 for PM2.5, respectively (Table 1).

193

Yearly mean values of PM10 and NO2 were below the Turkish air quality guidelines

194

except PM2.5 which doesn’t have threshold values (PM10 annual mean concentration

195

<60 µg/m3, NO2 annual mean concentration <60 µg/m3 by 2015). On the other hand

196

yearly mean values of PM10, PM2.5 and NO2 exceeded the threshold concentrations of

197

the European Union air quality directive (PM10 annual mean concentration <40

198

µg/m3, PM2.5 annual mean concentration <25 µg/m3, NO2 annual mean concentration

199

<40 µg/m3 by 2015). The comparison of the pollutants indicated that PM2.5 has the

200

highest coefficient of variation (CV = 0.52) over time, followed by PM10 (CV = 0.46)

AC C

EP

TE D

M AN U

SC

RI PT

179

8

ACCEPTED MANUSCRIPT and NO2 (CV = 0.36). The daily mean temperature and humidity were 14.5 °C and

202

79.6% in İstanbul (Table 1).

203

Table 2 shows the Pearson correlation coefficients between air pollutants

204

concentrations, temperature and humidity. PM10, PM2.5 and NO2 were closely

205

correlated with each other with the Pearson correlation coefficients ranging from 0.68

206

to 0.84. Air pollutant concentrations were negatively correlated with temperature and

207

humidity. The highest correlation noted was between PM10 and PM2.5 (r = 0.84),

208

followed by PM10 and NO2 (r = 0.70).

209

Table 2.

210

3.2. Analytical Results

211

3.2.1. Single-pollutant model for total and gender-specific hospital admissions

212

The excess risk estimate (ER) and 95% confidence intervals associated with an

213

increase in pollutant concentration of 10 µg/m3 was obtained for each pollutant (PM10,

214

PM2.5 and NO2). For all pollutants nine lags were evaluated (lags 0-9 days). Table 3

215

summarizes the results of the single-pollutant models for the estimates of respiratory

216

related hospital admissions for total and gender-specific admissions. Statistically

217

significant associations with respiratory hospital admissions at different lags were

218

found for all the pollutants.

219

Table 3.

220

Our findings showed that the relative magnitude of risks for an association of the

221

pollutants with the total hospital admissions was in the order of: PM2.5, NO2, and

222

PM10. The highest association of each pollutant with hospital admission was observed

223

with PM2.5 at lag 4 (ER = 1.50; 95% CI =1.09 − 1.99), NO2 at lag 4 (ER = 1.27; 95%

224

CI = 1.02−1.53) and PM10 at lag 0 (ER = 0.61; 95% CI =0.33 − 0.89). These findings

AC C

EP

TE D

M AN U

SC

RI PT

201

9

ACCEPTED MANUSCRIPT show that PM2.5 showed an important association with respiratory morbidity, with

226

1.5% of morbidity associated with changes in pollutant concentration. The ER

227

increased for PM2.5 and NO2 from lag 0 and reached a maximum value at lag 4, before

228

declining to the lowest value at lag 9 for PM2.5 and at lag 8 for NO2. The ER for PM10

229

reached a maximum value at lag 0 and then decreased, but increased again at lag 3

230

and lag 4.

231

According to our results, the relative magnitude of risks for men was in the same

232

order of total hospital admissions: PM2.5, NO2, and PM10. The highest association of

233

each pollutant was observed with PM2.5 at lag 4 (ER = 1.42; 95% CI =0.86- 1.98),

234

NO2 at lag 4 (ER = 0.96; 95% CI = 0.61- 1.31) and PM10 at lag 0 (ER = 0.68; 95% CI

235

=0.29−1.07). For women, the relative magnitude of risks was in the same order of:

236

PM2.5, NO2, and PM10. The highest association of each pollutant with hospital

237

admission for women was observed with PM2.5 at lag 3 (ER = 1.57; 95% CI =1.08-

238

2.06), NO2 at lag 4 (ER = 1.49; 95% CI = 1.12- 1.85) and PM10 at lag 4 (ER = 0.64;

239

95% CI =0.39 −0.90). Females are being more likely to be effected from outdoor air

240

pollution than males since excess risks of PM2.5, NO2 and PM10 for women were

241

generally higher than males (Table 3). The results are graphically presented in the

242

Supplementary appendix.

243

Table 5 shows the cumulative effect for gender-specific hospital admissions for an

244

increase of 10 µg/m3 in PM10, PM2.5 and NO2 over 9 days of lag (i.e. summing all the

245

effects up to maximum lag), together with its 95% confidence intervals (CI). An

246

increase of 10 µg/m3 in concentrations of PM10, PM2.5 and NO2 over 9 days of lag

247

corresponds to ER = 3.28 (95% CI = 2.53−4.04), ER = 7.22 (95% CI = 5.37−9.09)

248

and ER = 3.66 (95% CI = 2.48 − 4.85) increase of total respiratory hospitalizations,

249

respectively.

AC C

EP

TE D

M AN U

SC

RI PT

225

10

ACCEPTED MANUSCRIPT Table 5

251

3.2.2. Single-pollutant model for age-specific hospital admissions

252

Table 4 summarizes the results of the single-pollutant models for the estimates of

253

respiratory related hospital admissions for age-specific admissions. Statistically

254

significant associations were found for respiratory admissions for all-ages groups.

255

Age was clearly an effect modifier of respiratory hospital admissions, and people

256

were more likely to be effected from particulate matter (PM10 and PM2.5) if they were

257

older. The relative magnitude of risks for an association of the pollutants with the

258

respiratory hospital admissions for the children (aged 0-14) was in the order of: PM2.5,

259

NO2 and PM10. The highest association of each pollutant with hospital admission was

260

observed with PM2.5 at lag 8 (ER = 1.25; 95% CI =0.47- 2.03), NO2 at lag 9 (ER =

261

1.08; 95% CI = 0.54−1.63) and PM10 at lag 0 (ER = 0.71; 95% CI =0.33-1.09). The

262

relative magnitude of risks for the adults (aged 35-44) was in the same order of

263

children: PM2.5, NO2, and PM10. The highest association of each pollutant with

264

hospital admission for the adults (aged 35-44) was observed with PM2.5 at lag 4 (ER =

265

2.20; 95% CI =0.85- 3.56), NO2 at lag 0 (ER = 1.27; 95% CI = -1.39- 4.00) and PM10

266

at lag 4 (ER = 1.06; 95% CI =0.47−1.65). Finally, the relative magnitude of risks for

267

the elderly (aged ≥65) was in the order of: PM2.5, PM10, and NO2. The highest

268

association of each pollutant with hospital admission was observed with PM2.5 at lag 2

269

(ER = 3.83; 95% CI =1.58- 4.46), PM10 at lag 2 (ER = 1.26; 95% CI = 0.66 −1.86)

270

and NO2 at lag 4 (ER = 1.16; 95% CI =0.28- 2.05).

271

Table 4

272

According to our results, the excess risks of PM10, PM2.5 and NO2 increase with

273

advancing age (Table 4). Based on the association with hospital admissions for

274

respiratory illnesses of the elderly (≥ 65), the ER for PM2.5 showed a higher increase

AC C

EP

TE D

M AN U

SC

RI PT

250

11

ACCEPTED MANUSCRIPT from lag 0 to lag 2 compared with the other pollutants, which ERs were closer to one

276

another. The ER decreased for PM2.5 after lag 2 but increased again after lag 5. NO2

277

showed increasing patterns up to lag 4, and then declined slowly to the lowest value at

278

lag 9. After increasing up from lag 0 to lag 3, the RR of PM10 decreased gradually and

279

reached a minimum at lag 9. The results are graphically presented in the

280

Supplementary appendix.

281

Table 6 shows the cumulative effect for age-specific hospital admissions for an

282

increase of 10 µg/m3 in PM10, PM2.5 and NO2 over 9 days of lag (i.e. summing all the

283

effects up to maximum lag), together with its 95% confidence intervals (CI). An

284

increase of 10 µg/m3 in concentrations of PM10, PM2.5 and NO2 over 9 days of lag

285

corresponds to ER = 7.66 (95% CI = 5.12−10.26), ER = 17.9 (95% CI = 11.45−24.73)

286

and ER = 7.28 (95% CI = 3.49 − 11.22) increase of respiratory hospitalizations for

287

elderly (aged ≥65 age).

288

Table 6

289

4. Discussion and Conclusions

290

Our study showed that short-term exposure to outdoor air pollutants (PM10, PM2.5 and

291

NO2) was positively associated with respiratory hospital admissions in İstanbul.

292

According to our results, the risk of respiratory illness due to air pollution can occur

293

up to ten days (nine days lag) after the exposure. PM2.5 was the most significantly

294

associated pollutant with the respiratory hospital admissions in the city followed by

295

NO2 and PM10. According to our study, there is an apparent relationship between

296

aging and respiratory diseases. The elderly (≥65 years) are at higher risk of suffering

297

from respiratory diseases than other age groups due to the air pollution. Also, females

AC C

EP

TE D

M AN U

SC

RI PT

275

12

ACCEPTED MANUSCRIPT seem to be at higher risk of suffering from respiratory diseases than males since they

299

had generally higher ER values.

300

If we compare the results of our study with different studies, these estimates are

301

consistent with the existing literature. A significant association between acute

302

hospitalizations for COPD and PM10, NO2 and CO was found in Vancouver, Canada.

303

The magnitude of effects slightly increased with increasing days of exposure. (Yang

304

et al., 2005). In a study undertaken in the UK, there was a very strong positive

305

association between all respiratory admissions with SO2, weak positive associations

306

with PM10, PM2.5, and black smoke in the 0–14 age group (Anderson et al., 2001).

307

Dominici et al. found evidence of positive associations between day-to-day variation

308

in PM2.5 concentration and hospital admissions for all outcomes, except injuries, for at

309

least 1 exposure lag. For respiratory outcomes, the largest effects occurred at lags 0

310

and 1 for COPD and at lag 2 for respiratory tract infections (Dominici et al., 2006). In

311

Hong Kong, Ko et al. reported that, a 10 µg/m3 increase in PM10, PM2.5 and NO2

312

concentrations was associated with a 2.4%, 3.1% and 2.6% increase in COPD

313

admissions with a cumulative lag of 0–5, respectively (Ko et al., 2007). Larrieu

314

assessed the associations between daily levels of PM10, NO2, and O3 and medical

315

home visits in Bordeaux (France, 2000-2006). The risk of visits for upper and lower

316

respiratory diseases was significantly increased by 1.5% (95% confidence interval

317

(CI): 0.3, 2.7) and 2.5% (95% CI: 0.5, 4.4) during the 3 days following a 10 µg/m3

318

increase in PM10 levels, respectively. Similarly, an increased risk of visits for lower

319

respiratory diseases was observed for a 10 µg/m3 increment in NO2 (ER 2.6%; 95%

320

CI, 0.2-4.9). Estimates of ER were much higher in the elderly (≥65 years) for

321

respiratory diseases (upper and lower), regardless of the type of pollutant, suggesting

322

a higher effect in this subgroup. This difference was particularly evident for upper

AC C

EP

TE D

M AN U

SC

RI PT

298

13

ACCEPTED MANUSCRIPT respiratory diseases, with ERs of 12.3% (95% CI: 4.9, 19.7) and 8.3% (95% CI: 2.0,

324

14.7) being associated with 10 µg/m3 increases in NO2 and PM10 levels, respectively,

325

in this subgroup. In the other age groups, effects were close to each other (Larrieu et

326

al., 2009). According to a time-stratified case-crossover study which was carried out

327

in six Italian cities, there was an immediate increase in emergency hospitalizations for

328

respiratory diseases associated with PM10, and even more strongly associated with

329

NO2. According to the study, the NO2 effects on hospitalization for respiratory

330

diseases were almost twice as large as those due to PM10 (Faustini et al., 2013). In

331

another study, significant associations between particulate matter and respiratory

332

admissions were found in eight Mediterranean cities in Southern Europe. Lag 0–5

333

effect estimates were 1.36% (95% CI: 0.23, 2.49%) for a 10µg/m3 increase in PM2.5,

334

and 1.15% (95% CI: 0.21, 2.11%) for a 14.4µg/m3 increase in PM10 (Stafoggia et al.,

335

2013). In a recent study undertaken in Vietnam, PM10, NO2 and SO2 were positively

336

associated with daily hospital admission for respiratory diseases in Ho Chi Minh City.

337

At lag-0 day, the risk of respiratory admissions increased by 0.7% for a 10 µg/m3

338

increase in PM10; by 8% for a 10 µg/m3 increase in NO2, and by 2% for a 10 µg/m3

339

increase in SO2. Females were found to be more sensitive to exposure to air pollutants

340

than males (Phung et al., 2016). In summary, exposure to outdoor air pollution was

341

associated with respiratory hospital admissions around the world. As the first air

342

pollution hospital admission study using GLM in İstanbul, our findings may

343

supplement key scientific information on air pollution-related health effects for

344

İstanbul by providing local decision-makers with information needed to set priority of

345

air pollution control measures. Further research will be needed to find the effects of

346

the various pollutants and to gain insights into the modification of individual socio-

347

demographic characteristics and season on air pollution health effects.

AC C

EP

TE D

M AN U

SC

RI PT

323

14

ACCEPTED MANUSCRIPT Acknowledgements

349

Authors thank to Turkish State Ministry of Health and Ministry of Environment and

350

Urbanization for providing the related data.

351

Appendix A. Supplementary data

352

References

353 354 355 356

Anderson, H.R., Bremner, S.A., Atkinson, R.W., Harrison, R.M., Walters, S., 2001. Particulate matter and daily mortality and hospital admissions in the west Midlands conurbation of the United Kingdom: associations with fine and coarse particles, black smoke and sulphate. Occupational and Environmental Medicine 58, 504–510.

SC

RI PT

348

358 359 360 361 362

M AN U

357

Atkinson, R.W., Anderson, H.R., Sunyer, J., Ayres, J., Baccini, M., Vonk, J.M., Boumghar, A., Forastiere, F., Forsberg, B., Touloumi, G., Schwartz, J., Katsouyanni, K., 2001. Acute effects of particulate air pollution on respiratory admissions. Results from APHEA 2 Project. American Journal of Respiratory and Critical Care Medicine 164, 1860–1866.

369 370 371 372 373

374 375

EP

368

Braga, A.L.F., Conceição, G.M.S., Pereira, L. A. A., Kishi, H. S., Pereira, J. C. R., Andrade, M. F., Gonçalves, F.L.T., Saldiva, P.H.N., Latorre, M.R.D O., 1999. Air pollution and pediatric respiratory hospital admissions in Sao Paulo, Brazil. Journal of Occupational and Environmental Medicine 1, 95-102.

Braga, A.L.F., Saldiva, P.H.N., Pereira, L.A.A., Menezes, J.J.C., Conceição, G.M.S., Lin, C.A., Zanobetti, A., Schwartz, J. & Dockery, D.W., 2001. Health effects of air pollution exposure on children and adolescents in Sao Paulo, Brazil. Pediatric Pulmonology, Vol. 31, pp. 106-113.

AC C

364 365 366 367

TE D

363

Çapraz, Ö., Efe, B., Deniz, A., 2015. Study on the Association between Air Pollution and Mortality in İstanbul, 2007-2012. Atmospheric Pollution Research 7, 147-154.

376

15

ACCEPTED MANUSCRIPT 377 378

Chen, B., Hong, C., Kan, H., 2004. Exposures and health outcomes from outdoor air pollutants in China. Toxicology 198, 291–300.

379

Dominici F, Peng RD, Bell ML, Pham L, McDermott A, Zeger SL, et al. 2006. Fine particulate air pollution and hospital admission for cardiovascular and respiratory diseases. Journal of the American Statistical Association 295, 1127–1134

RI PT

380 381 382 383

Faustini, A., Stafoggia, M., Colais, P., Berti, G., Bisanti, L., Cadum, E., Cernigliaro, A., Mallone, S., Scarnato, C., Forastiere, F., EpiAir Collaborative Group, 2013. Air pollution and multiple acute respiratory outcomes. European Respiratory Journal 42, 304-313.

SC

384 385 386 387 388

Hapçıoglu, B., İşsever, H., Koçyigit, E., Dişçi, R., Vatansever, S., Özdilli, K., 2005. The effect of air pollution and meteorological parameters on chronic obstructive pulmonary disease at an İstanbul Hospital. Indoor and Built Environment 15, 149153.

M AN U

389 390 391 392 393

400 401 402 403

404 405 406

TE D

397 398 399

Katsouyanni K, Touloumi G, Samoli E, et al., 2001. Confounding and effect modification in the short-term effects of ambient particles on total mortality: results from 29 European cities within the APHEA2 Project. Epidemiology 12, 521–531.

EP

396

İncecik, S., İm, U., 2012. Air Pollution in Mega Cities: A Case Study of İstanbul. Air Pollution - Monitoring, Modelling and Health, Intech Open Access Publisher, 77-116

Keles, N., Ilıcalı, C., 1998. The impact of outdoor pollution on upper respiratory diseases. Rhinology 36, 24-27.

AC C

394 395

Karaca, F., Anil, I., Alagha O., 2009. Long-range Potential Source Contributions of Episodic Aerosol Events to PM10 Profile of a Megacity. Atmospheric Environment 43, 5713 – 5722.

407

408 409 410

Ko F.W., Tam W., Wong T.W., Chan, D.P.S., Tung, A.H., Lai, C.K.W., Hui, D.S.C., 2007. Temporal relationship between air pollutants and hospital admissions for chronic obstructive pulmonary disease in Hong Kong. Thorax 62, 780–785. 16

ACCEPTED MANUSCRIPT 411

412 413 414

Larrieu S., Lefranc A., Gault G., Edouard C., Franck C., Bernard J., Laurent F., 2009. Are the short-term effects of air pollution restricted to cardiorespiratory diseases? American Journal of Epidemiology 169, 1201-1208.

416 417 418

RI PT

415

Li, Y., Henze, D.K., Jack, D., Henderson, B.H., Kinney, P.L., 2016. Assessing public health burden associated with exposure to ambient black carbon in the United States. Science of The Total Environment 539, 515-525.

Mahiyuddin, W.R.W., Sahani, M., Aripin, R., Latif, M.T., Thach, T-Q., Wong, C-M., 2013. Short-term effects of daily air pollution on mortality. Atmospheric Environment 65, 69-79.

M AN U

420 421 422

SC

419

423

424 425

Özcan, S.N., Çubukçu, M.K., 2015. Evaluation of Air Pollution Effects on Asthma Disease: The case of İzmir. Procedia-Social and Behavioral Sciences 202, 448-455.

426

432 433 434 435 436 437

TE D

EP

431

Özdemir, H., Pozzoli, L., Kindap, T., Demir, G., Mertoglu, B., Mihalopoulos, N., Theodosi, C., Kanakidou, M., Im, U., Unal, A., 2014. Spatial and temporal analysis of black carbon aerosols in Istanbul megacity. Science of The Total Environment 473474, 451–458.

Peters, A., Dockery, D.W., Muller, J.E., Mittleman, M.A., 2001. Increased particulate air pollution and the triggering of myocardial infarction. Circulation 103, 2810- 2815.

AC C

427 428 429 430

Peng, R.D.; Dominici, F. & Louis, T.A., 2006. Model choice in time series studies of air pollution and mortality. Journal of the Royal Statistical Society Series 169, 179203.

438 439 440 441 442

Phung, D., Hien, T.T, Kindap, T., Linh, H.N, Luong, L.M.T., Morawska, L., Chu, C., Binh, N.D, Thai, P.K, 2016. Air pollution and risk of respiratory and cardiovascular hospitalizations in the most populous city in Vietnam. Science of The Total Environment 557-558, 322–330.

443 17

ACCEPTED MANUSCRIPT 444 445 446 447

Pope III, C.A., Burnett, R.T., Thun, M.J., Calle, E.E., Krewski, D., Ito, K., Thurston, G.D., 2002. Lung cancer, cardiopulmonary mortality, and long-term exposure to fine particulate air pollution. Journal of the American Medical Association 287, 11321142.

448

Pope III C.A, Dockery, DW., 2006. Health effects of fine particulate air pollution: lines that connect. Journal of Air & Waste Management Assocciation 56, 709–742.

RI PT

449 450 451

R Development Core Team. R: a language and environment for statistical computing, version 2.15.0. Vienna: R Foundation for Statistical Computing; 2016.

SC

452 453 454

Samet J.M., Zeger S.L, Dominici F., Curriero, F., Coursac, I., Dockery, D.W., Schwartz, J., Zanobetti, A., 2000. The National morbidity, mortality, and air pollution study. Part II: Morbidity and mortality from air pollution in the United States. Research report (Health Effects Institute) 94, 5–79.

M AN U

455 456 457 458 459

468

469 470 471 472 473 474

TE D

465 466 467

Savaş, B. Serdar et al. In Thomson, S. and Mossialos, E., eds., 2002. Health care systems in transition: Turkey. European Observatory on Health Care Systems, 4, Copenhagen, 108 pages.

EP

464

Samoli, E., Peng, R., Ramsay, T., Pipikou, M., Touloumi, G., Dominici, F., Burnett, R., Cohen, A., Krewski, D., Samet, J., Katsouyanni, K., 2008. Acute effects of ambient particulate matter on mortality in Europe and North America: results from the APHENA study. Environmental Health Perspectives 116, 1480-1486.

AC C

460 461 462 463

Stafoggia, M., Samoli, E., Alessandrini, E., Cadum, E., Ostro, B., Berti, G., Faustini, A., Jacquemin, B., Linares, C., Pascal, M., Randi, G., Ranzi, A., Stivanello, E., Forastiere, F., Med-Particles Study Group, 2013. Short-term Associations between Fine and Coarse Particulate Matter and Hospitalizations in Southern Europe: Results from the MED-PARTICLES Project. Environmental Health Perspective 121, 10261033.

475

476 477

Tadano, Y.S., Ugaya, C.M.L., Franco, A.T., 2012. Methodology to Assess Air Pollution Impact on Human Health Using the Generalized Linear Model with Poisson 18

ACCEPTED MANUSCRIPT 478 479

Regression, Air Pollution - Monitoring, Modelling and Health, 281-304, doi: 10.5772/33385

480 481 482 483

Tecer, L.H., Alagha, O., Karaca, F., Tuncel, G., Eldes, N., 2008. Particulate matter (PM2.5, PM10-2.5, and PM10) and Children’s Hospital Admissions for Asthma and Respiratory Diseases: a bidirectional case-crossover study. Journal of Toxicology and Environmental Health 71, 512-520.

485 486 487

RI PT

484

Tomaç, N., Demirel, F., Acun, C. ve Ayoğlu, F., 2005. Prevalence and risk factors for childhood asthma in Zonguldak, Turkey. Allergy and Asthma Proceedings, 26, 397402.

Toros, H., Erdun, H., Çapraz, Ö., Özer, B., Daylan, B. S., Öztürk, İ.A., 2013. Air Pollution and Quality Level in Metropolitan Turkey for Sustainable life. European Journal of Science and Technology 1, 12-18.

492

493 494

Turkish Statistical Institute http://www.tuik.gov.tr/

501 502 503 504 505

506 507 508

Transportation

statistics.

Unal, Y.S., Toros, H., Deniz, A., İncecik, S., 2011. Influence of meteorological factors and emission sources on spatial and temporal variations of PM10 concentrations in İstanbul metropolitan area. Atmospheric Environment 45, 55045513.

EP

500

2016.

AC C

496 497 498 499

(TÜİK),

TE D

495

M AN U

489 490 491

SC

488

Viegi, G., Maio, S., Simoni, M., Baldacci, S., Annesi-Maesano, I., 2009. The epidemiological link between ageing and respiratory diseases. European Respiratory Monograph: Respiratory Diseases in the Elderly. European Respiratory Society 43, 117.

Yang, Q., Chen, Y., Krewski, D., Richard, T.B., Yuanli, S., Kimberlyn, M.M., 2005. Effect of short-term exposure to low levels of gaseous pollutants on chronic obstructive pulmonary disease hospitalizations. Environmental Research 99, 99-105

509

19

ACCEPTED MANUSCRIPT Table 1 Summary statistics of daily hospital admission numbers, air pollutant concentrations and weather conditions in İstanbul, Turkey (20013–2015). Mean+SD

Coefficient of Variaton (CV)

Min P(25)

P(50)

P(75)

Max

54091

Total (n = 12 884 628)

17160 ± 10501.1

0.62

1909

5299

17963

24860

Man (n = 5 892 222)

7851 ± 4671.6

0.60

951

2666

8146

11309

24216

Woman (n = 6 992 406)

9310 ± 5833.0

0.63

922

2638

9795

13570

29875

0-14 yrs (n = 4 652 912)

6179 ± 3640.9

0.60

971

2829

5812

8824

20462

35-44 yrs (n = 1 279 611)

1682 ± 1061.0

0.63

141

472

1856

2430

5803

65+ yrs (n = 1 929 661)

2559 ± 1670.3

0.66

174

396

2971

3745

7187

311.0

Air pollution concentrations (µg/m3) PM10

56.3 ± 26.6

0.46

17.6

40.2

48.5

67.3

PM2.5

29.0 ± 15.5

0.52

9.6

19.2

24.0

35.5

89.3

NO2

41.8 ± 16.0

0.36

15.0

30.6

39.4

51.1

86.7

Temperature (°C)

14.5 ± 6.9

0.46

1.0

8.9

14.6

20.4

27.7

Humidity (%)

79.6 ± 11.4

0.14

47.3

73.8

80.1

86.9

100.0

SC

Weather

RI PT

Daily admission counts

Table 2 Pearson correlation coefficients between daily air pollutant concentrations and weather conditions in İstanbul (2013–2015). PM2.5

NO2

Temperature

PM10

1

PM2.5

0.84

NO2

0.70

0.68

Temperature

-0.12

-0.38

-0.11

1

Humidity

-0.25

0.02

-0.41

-0.40

1 1

Table 3

Humidity

M AN U

PM10

-0.39

ER(%) (mean and 95% CI) of daily hospital admission stratified by gender associated with 10g/m3 increase of pollutant concentrations in İstanbul, Turkey (2013–2015). NO2

0.61 (0.33, 0.89) 0.16 (-0.01, 0.32) 0.17 (1.00, 0.35) 0.35 (0.19, 0.52) 0.49 (0.32, 0.67) 0.49 (0.31, 0.67) 0.34 (0.18, 0.51) 0.14 (-0.04, 0.31) 0.06 (-0.10, 0.23) 0.42 (0.15, 0.69)

-1.76 (-2.62, -0.88) 0.10 (-0.32, 0.53) 1.10 (0.69, 1.51) 1.49 (1.14, 1.83) 1.50 (1.09, 1.90) 1.31 (0.90, 1.73) 1.08 (0.65, 1.50) 0.87 (0.29, 1.45) 0.73 (0.17, 1.29) 0.63 (-0.06, 1.32)

-0.68 (-1.47, 0.12) -0.19 (-0.47, 0.09) 0.51 (0.26, 0.76) 1.06 (0.84, 1.29) 1.27 (1.02, 1.53) 1.07 (0.81, 1.34) 0.56 (0.31, 0.80) -0.20 (-0.29, 0.25) -0.28 (-0.54, -0.02) 0.32 (0-0.08, 0.72)

0.68 (0.29, 1.07) 0.09 (-0.14, 0.32) 0.03 (-0.22, 0.28) 0.19 (-0.04, 0.42) 0.35 (0.11, 0.60) 0.40 (0.15, 0.65) 0.30 (0.07, 0.54) 0.13 (-0.12, 0.39) 0.04 (-0.19, 0.28) 0.30 (-0.07, 0.68)

-1.57 (-2.79, -0.35) 0.43 (-0.55, 0.64) 0.98 (0.41, 1.55) 1.38 (0.91, 1.86) 1.42 (0.86, 1.98) 1.21 (0.64, 1.78) 0.89 (0.30, 1.49) 0.58 (-0.23, 1.39) 0.34 (-0.44, 1.13) 0.27 (-0.70, 1.24)

-0.11 (-1.20, 1.00) -0.15 (-0.51, 0.21) 0.29 (-0.05, 0.63) 0.75 (0.44, 1.06) 0.96 (0.61, 1.31) 0.80 (0.44, 1.16) 0.34 (0.01, 0.66) -0.19 (-0.55, 0.17) -0.38 (-0.74, -0.02) 0.34 (-0.20, 0.89)

0.57 (0.18, 0.97) 0.20 (-0.04, 0.43) 0.27 (0.03, 0.52) 0.49 (0.26, 0.72) 0.64 (0.39, 0.90) 0.63 (0.37, 0.88) 0.45 (0.21, 0.68) 0.21 (-0.04, 0.47) 0.14 (-0.09, 0.38) 0.55 (0.17, 0.94)

-1.91 (-3.14, -0.66) 0.19 (-0.42, 0.80) 1.22 (0.64, 1.80) 1.57 (1.08, 2.06) 1.55 (0.97, 2.13) 1.39 (0.80, 1.98) 1.24 (0.63, 1.85) 1.16 (0.33, 1.99) 1.11 (0.31, 1.92) 0.99 (0.01, 1.99)

-0.94 (-2.06, 0.19) -0.20 (-0.59, 0.18) 0.64 (0.29, 1.00) 1.27 (0.94, 1.60) 1.49 (1.12, 1.85) 1.25 (0.87, 1.63) 0.69 (0.35, 1.02) 0.05 (-0.32, 0.42) -0.24 (-0.60, 0.13) 0.37 (-0.20, 0.94)

PM10

Total

L0 L1 L2 L3 L4 L5 L6 L7 L8 L9 L0 L1 L2

AC C

Men

EP

Lag

TE D

PM2.5

Gender

L3 L4 L5 L6 L7 L8 L9

Women

L0 L1 L2 L3 L4 L5 L6 L7 L8 L9

Note: Bold type indicates the highest association of each pollutant. Results are shown graphically in the Supplementary.

ACCEPTED MANUSCRIPT Table 4 ER(%) (mean and 95% CI) of daily hospital admission stratified by age group associated with 10g/m3 increase of pollutant concentrations in İstanbul, Turkey (2013–2015). PM2.5

0-14

L0 L1

0.71 (0.33, 1.09) 0.05 (-0.18, 0.28) -0.07 (-0.31, 0.18) 0.04 (-0.19, 0.26) 0.16 (-0.08, 0.41) 0.19 (-0.06, 0.43) 0.11 (-0.12, 0.34) 0.01 (-0.23, 0.26) 0.08 (-0.15, 0.31) 0.59 (-0.20, 0.99)

-2.63 (-3.8, -1.44) 0.01 (-0.57, 0.59) 1.02 (0.46, 1.58) 1.13 (0.67, 1.59) 0.92 (0.37, 1.48) 0.77 (0.21, 1.34) 0.83 (0.25, 1.42) 1.08 (0.28, 1.88) 1.25 (0.47, 2.03) 0.90 (-0.06, 1.87)

-0.98 (-2.07, 0.12) 0.01 (-0.37, 0.38) 0.67 (0.33, 1.02) 0.97 (0.65, 1.28) 0.91 (0.56, 1.26) 0.60 (0.24, 0.96) 0.20 (-0.13, 0.52) -0.07 (-0.42, 0.29) 0.10 (-0.25, 0.47) 1.08 (0.54, 1.63)

0.87 (-0.03, 1.77) 0.23 (-0.31, 0.78) 0.38 (-0.17, 0.94) 0.78 (0.26, 1.30) 1.06 (0.47, 1.65) 1.02 (0.42, 1.62) 0.66 (0.10, 1.21) 0.12 (-0.47, 0.72) -0.24 (-0.79, 0.31) 0.07 (-0.82, 0.96)

-0.65 (-3.51, 2.30) -0.35 (-1.75, 1.07) 0.65 (-0.69, 2.01) 1.65 (0.52, 2.80) 2.20 (0.85, 3.56) 2.08 (0.70, 3.47) 1.36 (-0.05, 2.80) 0.39 (-1.52, 2.35) -0.25 (-2.10, 1.64) 0.27(-2.03, 12.62)

1.27 (-1.39, 4.00) -0.56 (-1.45, 0.34) -0.46 (-.128, 0.37) 0.36 (-0.40, 1.12) 1.07 (0.23, 1.91) 1.20 (0.33, 2.08) 0.65 (-0.14, 1.44) -0.30 (-1.16, 0.56) -1.01 (-1.85, -0.15) -0.47 (-1.79, 0.87)

L2

0.66 (-0.27, 1.24) 1.51 (0.95, 1.74) 1.55 (1.00, 1.86)

L3 L4 L5 L6

1.23 (0.73, 1.57) 0.85 (0.25, 1.34) 0.58 (-0.07, 1.17) 0.48 (-0.16, 1.10)

L7 L8 L9

0.47 (-0.21, 1.19) 0.34 (-0.26, 1.04) -0.25 (-0.12, 0.80)

-2.23 (-5.14, 0.76) 3.17 (1.74, 4.63) 3.83 (2.46, 5.21) 2.56 (1.45, 3.67) 1.38 (0.02, 2.76) 1.29 (-0.57, 2.75) 2.31 (0.78, 3.87) 3.53 (1.47, 5.64) 3.01 (1.03, 5.04) -0.20 (-4.38, 0.39)

-0.32 (-4.53, 2.59) 0.69 (-1.14, 1.62) 1.15 (0.28, 2.03) 1.30 (0.49, 2.12) 1.33 (0.44, 2.23) 1.31 (0.38, 2.25) 1.23 (0.36, 2.10) 0.97 (0.04, 1.91) 0.35 (-0.55, 1.25) -0.93 (-2.40, 0.56)

L2 L3 L4 L5 L6 L7 L8 L9 35-44

L0 L1 L2 L3 L4 L5 L6 L7 L8 L9

65+

L0 L1

NO2

RI PT

PM10

SC

Lag

M AN U

Age

Table 5

TE D

Note: Bold type indicates the highest association of each pollutant. Results are shown graphically in the Supplementary.

Cumulative ER(%) of gender-specific daily hospital admission associated with 10 µg/m3 increase of pollutant concentrations over 9 days of lag in Istanbul (mean and 95% CI). Man

%ER

%ER

PM10

3.28 (2.53, 4.04)

2.67 (1.51, 3.60)

4.23 (3.13, 5.33)

PM25

7.22 (5.37, 9.09)

5.63 (3.10, 8.22)

8.79 (6.11, 11.5)

EP

Total

Woman %ER

Table 6

AC C

NO2 3.66 (2.48, 4.85) 2.10 (1.22, 4.13) 4.43 (2.81, 6.08) Note: Values marked with an asterisk (*) are statistically insignificant. Other values are significant at the 0.05 level (p < 0.05).

Cumulative ER(%) of age-specific daily hospital admission associated with 10 µg/m3 increase of pollutant concentrations over 9 days of lag in Istanbul (mean and 95% CI).

PM10 PM25 NO2

0-14

35-44

65+

%ER

%ER

%ER

1.88 (0.87, 2.90)

5.03 (2.50, 7.63)

7.66 (5.12, 10.26)*

5.35 (2.86, 7.90)

7.56 (1.48, 14.00)*

17.90 (11.45, 24.73)

3.52 (1.97, 5.09)*

1.73 (-1.68, 5.44)

7.28 (3.49, 11.22)*

Note: Values marked with an asterisk (*) are statistically insignificant. Other values are significant at the 0.05 level (p < 0.05).

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

Fig.1. Map of İstanbul showing air quality monitoring stations with red dots. Station names from right to left, respectively: (1) Sultanbeyli, (2) Ümraniye, (3) Üsküdar, (4) Mecidiyeköy, (5) Kağıthane, (6) Sultangazi, (7) Esenler, (8) Şirinevler, (9) Başakşehir, (10) Esenyurt, (11) Silivri.

ACCEPTED MANUSCRIPT

Highlights İstanbul has PM10, PM2.5 and NO2 pollution depending on different emission sources.

RI PT

Short-term exposure to PM10, PM2.5 and NO2 is positively associated with increased respiratory hospital admissions in İstanbul. PM2.5 has the strongest impact on respiratory hospital admissions.

Women and elderly people were more sensitive to respiratory risk of air pollution than other groups.

AC C

EP

TE D

M AN U

SC

The first air pollution and respiratory hospital admission study using GLM in İstanbul, Turkey.