Association Between Ambient Air Pollution and Hospitalization for Respiratory Diseases in Perth, Australia

Research article

Association Between Ambient Air Pollution and Hospitalization for Respiratory Diseases in Perth, Australia

Corresponding author: Dr Krassi Rumchev,School of Public Health, Curtin University, U1987 Perth WA 6845, Australia, Tel: + 61 8 9266 4342; + 61 8 92662958, Email: k.rumchev@curtin.edu.au

Abstract

Background: Numerous studies continue to demonstrate that air pollution is a significant contributor for respiratory hospital- izations among adults and children.

Methods: We examined the associations between daily changes in exposure levels of selected air pollutants including nitrogen dioxide (NO2), particulate matter with <10µm (PM10) and <2.5µm in aerodynamic diameter (PM2.5), and hospitalization for respi- ratory symptoms, respiratory infections and asthma in Perth, Western Australia for a 5-year period. A Poisson model allowing for autocorrelation and over-dispersion was applied in the analysis for each air pollutant and also carried out separately for different seasons, defined as “warm” (November to April) and “cold” (May to October) seasons.

Results: Daily hospital admissions due to respiratory infections were significantly related to daily changes in concentrations of NO2, PM10 and PM2.5. Significant lagged effects on hospitalization for asthma and respiratory symptoms were established for PM2.5. Respiratory symptoms hospitalization was also significantly affected by NO2 at longer lags.

Conclusions: Our findings have important public health implications as we established significant associations between air pol- lutants and respiratory hospitalization in a city where the average concentrations of NO2, PM10 and PM2.5 are below the national standards.

Keywords: Air Pollution; Hospital Admissions; Respiratory Symptoms; Australia

Introduction

Air quality is not a new concept, however, continues to be a significant health and environmental problem worldwide in three major fractions based on size which is defined to a 50% cut point at a specific aerodynamic diameter. Coarse particles are defined as particles smaller than 10μm in aerodynamic diameter (PM ), fine particles are smaller than 2.5μm (PM )

10 2.5

cluding in Australia. Studies from around the world continue

to demonstrate relationships between air pollution and ad- verse health effects among adults and children. Some of the air pollutants considered as a major health concern in urban ar- eas include particulate matter and gases such as ozone, oxides of nitrogen, carbon monoxide and sulphur dioxide [1].

Particulate matter (PM) consists of solid and liquid particles which vary in size, number, shape, surface area, chemical com- position, solubility and origin [2]. PM can be separated into

and ultrafine particles (PM1.0) smaller than 0.1μm in aerodynamic diameter. In urban areas, the coarse particulates typical- ly contain resuspended dust from roads and industrial activi- ties, and biological material such as pollen grains and bacterial fragments. Fine particles are largely formed from combustion processes related to traffic, wood burning, power generation and industrial processes including smelters, cement plants, and paper mills. According to the World Health Organisation (WHO) [3] fine particles are considered more dangerous than course particles because they consist of high proportion of various toxic metals and acids, and aerodynamically they can penetrate deeper into the respiratory tract when compared with PM10. Numerous epidemiological and clinical studies have linked particulate air pollution to long term and short term ef- fects of morbidity on respiratory, pulmonary and cardiovascu- lar diseases [4-6]. Three independent Australian studies con- ducted in Perth, Melbourne and Brisbane found a significant relationship between exposure to particulate matter and hos- pital admissions for respiratory symptoms in adults [7-9]. The results are consistent with those found in Europe [10-13] and USA [14,15]. The Australian National Environment Protection Council (NEPC) [16] estimated that exposure to higher con- centrations of PM2.5 in Sydney, Melbourne, Brisbane and Perth caused a total of 1,611 premature deaths annually. Another air pollutant, studied extensively and associated with respiratory illnesses, including bronchitis, pneumonia, decreased resis- tance to pulmonary infections and decreased lung function is nitrogen dioxide (NO2), an oxidised form of nitric oxide [17]. According to WHO [18], fuel combustion with emissions due mainly to motor vehicle exhausts, heating and power sources, is one of the main sources of NO2.

Air Quality in Australia is regulated by the Protection Measure for Ambient Air Quality developed through the Environment Protection and Heritage Council. The Protection Measure sets air quality standards that are legally binding on each level of the Government. The standards relate to six criteria air pollut- ants including particulates, nitrogen dioxide, carbon monox- ide, photochemical oxidants, sulphur dioxide, and lead.

The climate regime in Perth is a Mediterranean – style climate with hot, dry summers and cool, wet winters. The air quality in Perth is considered as acceptable, with the air pollutants’ concentrations within the national standardswith some occa- sional episodes of poor air quality [19]. The major sources of air pollution in Perth include motor vehicles, domestic sourc- es (principally wood heaters) and industries. Regional areas of Western Australia may also experience poor air quality at times, pollution being caused by bushfires and windblown dust, industrial facilities and hazard reduction burns.

This paper evaluates the relationship between hospital admis- sions for respiratory illnesses and daily changes in PM2.5, PM10 and NO2 during a five year period (2004-2008) in Perth, West- ern Australia.

Methodology

We conducted a time-series analysis to examine the impact of daily changes in PM2.5, PM10 and NO2 on daily hospitalization for respiratory symptoms, respiratory infections and asthma in Perth, Western Australia for all age groups. The study in- cluded a period of five years and according to the Australian Demographic Statistics (ABS) the total population in Perth during the study period was 2,296,411.

Morbidity data

Daily hospital admissions for respiratory illnesses to all hos- pitals in Perth metropolitan area for the period between 2004 and 2008 were obtained through the Health Department, Western Australia. The standard diagnostic tool, International Classification of Diseases (ICD), was applied in this study. The retrospective data for daily hospitalization for acute respirato- ry infections (ICD codes J00-J22), respiratory symptoms (ICD codes J23-J99, excluding J45) and asthma (ICD code J45) were provided as an overall daily estimate for all age groups. Re- spiratory health diagnoses were based on primary discharge information obtained from the patients’ charts. All transfers from other hospitals were subsequently excluded. Residents living outside the metropolitan area were also excluded from the analysis. Hospital admissions for diseases related to the di- gestive system (ICD codes K00-K93) were also obtained for the study period and used as a control confounding factor as no re- search demonstrated a relationship between gastrointestinal disease and air pollution.

Air Quality Data

Currently there are twelve air quality monitoring stations es- tablished across Perth metropolitan areas to assess the state of air quality against the National Environment Protection Measure (NEPM) standards. Due to missing data from some stations, three monitoring stations with the most completed data, including Duncraig, Caversham and South Lake located north, middle and south of Perth CBD (Central Business Dis- trict), were averaged to represent the air quality for metropol- itan Perth area.

The retrospective data for daily concentrations (24 hr average) of PM10, PM2.5, and NO2 during the study period between 2004 and 2008 were provided by the Department of Environment and Conservation (DEC), Western Australia, however, daily average concentrations of PM2.5 were available only from the meteorological station in Duncraig. All data were collected by direct measurements as Tapered Element Oscillating Micro- balance was used to obtain continuous readings of PM and the chemiluminescence method was applied for collecting the NO2 data.

Statistical Analysis

Descriptive statistics were generated for relevant variables by using the IBM SPSS Statistics for Windows, Version 22 (IBM Corp. Released 2010 Armonk, NY).

To explore the relationship between hospital admissions for asthma, respiratory symptoms, and respiratory infections col- lected over the 5-year period starting from 01 January 2004 to 31 December 2008 and daily concentrations (24 hr average) of PM10, PM2.5, and NO2, a Poisson model allowing for autocor- relation and over-dispersion among count data, implemented by Stata routine Arpois[20] , was applied for each air pollutant and successfully applied in other similar studies[21-23]. The autocorrelation among the time series data was controlled by including different order of autoregressive terms depending on different models. Ljung–Box portmanteau (Q) test was used to assess whether any autocorrelation remained in the regres- sion residuals. Seasonality was controlled by inclusion of the function for k = 0.5, 1, 2, 3, 4, 5, 6 and t = 1,2, …, 1827 (the number of study days). Long-term trends were controlled by the inclusion of the number of the day and its square. Linear or quadratic trend (if appropriate), indicator of years, week days, and public holidays, humidity, and temperature (lag of humidity or temperature were included only if the model was improved) and their squares (if appropriate), and the admis- sion number of digestive disorder were included in the model as potential confounding factors.

Since the health effects of PM10, PM2.5, and NO2 are in a time de- pendent fashion, hospital admissions for asthma, respiratory infection and respiratory symptoms might not be due to expo- sure to air pollution on the day of the admission but over the preceding days as well. In the Poisson modeling, we therefore examined the effect of legging exposure for 0,1,2, 3 and 4 days (lag0, lag1, lag2, lag3 and lag4 days, respectively). Consecutive cumulative lag effects were also considered in the modeling (lag0-1: average of lag0 and lag1; lag0-2: average of lag 0 to lag2; lag0-3: average of lag0 to lag3; lag0-4: average of lag0 to lag4).

Since there was no information available for asthma hospital admissions for the period between 2004 and 2006, the Pois- son modeling for asthma was conducted only for the period of two years, 2007 and 2008.

To examine if a season might have an impact on the relation- ship between air pollutants and hospital admissions, the Pois- son modeling was carried out separately for different seasons, defined as “warm” (November to April) and “cold” (May to Oc- tober) seasons.

The results are presented as the average percentage (%) change in hospital admissions and associated 95 percent confi- dence intervals (95% CI) for each interquartile range increase in the relevant air pollutant.

Results

Air Quality

Summary statistics for the concentrations of air pollutants, meteorological variables and daily respiratory hospitaliza- tions are presented in Table 1. It can be seen that the mean concentrations of all pollutants were well below the national standards. Significantly (p<0.01) lower concentrations of PM10 and PM2.5 were measured in 2008 compared to the previous years, with similar trends recorded for NO2 although

Year Statistics Concentrations of air pollutants Meteorological measures Number of hospital admissions
PM10

(µg/m3)

PM2.5 (µg/m3) NO2(ppb) T0C RH (%) Asthma Respiratory Symptoms Respiratory Infections
2004 Mean(SD) Median Min-Max 17.52 (6.12)

16.7

5.66-47.07

7.93 (3.07)

7.30

3-24.4

7.13 (3.17)

6.65

1.1 -18.4

18.20 (5.24)

17.55

8.6-33.0

62.68 (16.18)

64.10

20.8-97.7

No

sufficient data

17.01 (5.7)

16.5

5.0-37.0

34.27

(12.85)

32.0

8.0-72.0

2005 Mean(SD) Median

Min-Max

16.73 (7.49)

14.99

2.93-55.23

7.82 (3.35)

7.2

3-40.5

7.15 (3.34)

6.73

1.3 – 17.4

17.38 (4.78)

16.8

7.7-31.7

67.12 (14.02)

68.8

17.9-96.6

No

sufficient data

18.18 (6.32)

17.0

2.0-36.0

35.69 (14.9)

32.0

8.0-87.0

2006 Mean(SD) Median Min-Max 17.42 (6.35)

16.5

3.85-37.29

8.33 (3.04)

7.7

3.3-27.38

7.14 (3.45)

6.9

1 – 17.9

18.22 (4.88)

17.60

7.9-30.80

61.77 (14.37)

63.7

19.1-92.6

No

sufficient data

16.59 (5.42)

16.0

2.0-32.0

32.46

(11.59)

30.0

9.0-71.0

2007 Mean(SD) Median Min-Max 16.41 (6.40)

15.08

4.83-48.32

7.47 (2.25)

7. 10

2.91-19.10

6.70 (3.13)

6.33

1.1-16.2

18.05 (4.80)

17.05

10.3-33.5

63.05 (14.87)

65.9

15.2-95.7

7.85 (3.80)

8.00

1.0-24.0

43.60 (20.5)

46.0

7.0-94.0

29.62

(12.87)

26.0

5.0-71.0

2008 Mean(SD)

Median Min-Max

15.24 (6.16)

13.79

5. 94-46.16

7.50 (2.95)

6.74

3.48-36.61

6.75 (3.34)

6.18

1.1-14.7

18.02 (5.13)

17.3

8.1-32.0

62.05 (14.89)

63.7

24.0-91.8

7.55 (3.72)

7.00

0-20.00

46.50 (21.48)

48.5

6.0-96.0

30.94 (11.2)

30.0

9.0-61.0

Overall Mean(SD) Median Min-Max 16.69 (6.57)

15.41

2.93-55.23

7.81(2.96)

7.203

2.91-40.50

6.98a (3.28)

6.53

1.0-18.4

17.97 (4.97)

17.2

7.7-33.5

63.32 (14.99)

65.4

15.2-97.7

7.70 (3.76)

7

0-24

28.38 (19.57)

20

2-96

32.60

(12.93)

30

5-87

Table 1. Summary descriptive statistics of air pollutants concentrations, meteorological measures, and number of hospital admissions by year, Perth WA

the differences were not statistically significant. The overall average temperature and relative humidity were 17.90C and 63.4%, respectively and there were no significant changes in both meteorological parameters over the study period.

As mentioned earlier, the asthma data was incomplete for the study period but when compared with other hospital data recorded for 2007 and 2008, most people were hospitalized with respiratory symptoms followed by those with respirato- ry infections and asthma. With regards to seasonal differences in respiratory hospitalizations, more people were admitted to a hospital during the cold season when compared with the warmer periods (Table 2). Significantly higher (p<0.05) con- centrations of NO2 were recorded during the cold months which is in contrast with the PM10 and PM2.5 concentrations (Table 2).

Hospital admissions for respiratory infections were also sig- nificantly associated with exposures to PM10 and PM2.5, indicat- ing that for every 8.06 µg/m3 increase in the concentration of PM10 and 3.10 µg/m3 in PM2.5, the number of people admitted to hospitals with respiratory infections increased by 1.87% (95% CI: 0.46-3.31) and by 1.90% (95% CI: 0.91-2.91%), respective-

ly. In addition, exposures to NO2, significantly increased the re- spiratory infections hospitalization by 2.17% at leg1. Further- more, significant cumulative lag health impacts were found at longer lags for NO2 (at lag0-1, lag0-2, lag0-3, and lag0-4) and PM2.5 (at lag0-1 and lag0-2) but no lag effect was determined for PM10 (Table 3).

Association Between Air Pollution And Number of Hospital Admissions for Respiratory Symptoms, 2004-2008

Mean (SD) Median Min Max IQR
Cold period (n =920)
Concentration
NO2 (ppb) 8.16 (3.26) 8.23 1.1 17.9 4.8
PM10 (µg/m3) 14.00 (4.64) 13.37 5.01 38.98 5.75
PM2.5 (µg/m3) 7.50 (2.74) 6.95 2.91 27.38 3.10
Admission
Respiratory infections 41.23 (11.59) 40 12 87 16
Respiratory symptoms 31.56 (20.02) 23 5 94 27
Asthma (2007 ~ 2008) 9.62 (3.57) 9 2 24 5
Warm period (n =907)
Concentration
NO2 (ppb) 5.78 (2.84) 5.23 1.0 18.4 3.5
PM10 (µg/m3) 19.42 (7.09) 18.70 2.93 55.23 8.35
PM2.5 (µg/m3) 8.13 (3.14) 7.50 3.00 40.50 3.20
Admission
Respiratory infections 23.83 (6.92) 23 5 53 9
Respiratory symptoms 25.15 (18.56) 17 2 96 18
Asthma (2007 ~ 2008) 5.75 (2.82) 5 0 15 4

Table 2 . Seasonal differences in concentrations of air and daily respiratory hospitalizations (2004-2008), Perth WA

Significant (p<0.05) but weak correlations between air pollut- ants (24h exposure) and meteorological variables were estab- lished. For the entire study period, course and fine particulates were significantly correlated with each other (r=0.324) and NO2 was significantly correlated with PM2.5 (r = 0.24) but not with PM10. Particulate air pollution and NO2 were significantly (p<0.05) correlated with temperature and relative humidity.

Association Between Air Pollution and Number of Hospi- tal Admissions for Respiratory Infections, 2004-2008

All pollutants showed significant associations with respiratory infections hospitalization with the greatest effect seen for NO2. With every 5.0 ppb increase in NO2 the number of hospital ad- missions for respiratory infections increased by almost 4% on average at lag0 (95% CI: 2.31-5.71) (Table 3).

The Poisson models failed to establish significant associations between hospitalization for respiratory symptoms and daily changes in concentrations of NO2, PM2.5 and PM10 on the day of the admission (lag0), however, there were several significant lag effects. The number of patients admitted to a hospital with respiratory symptoms increased by 2.22% (95% CI: 0.11%,

4.37%) and 3.42% (95% CI: 1.27%, 5.61%) for every 0.5ppb

increase in NO2 exposure at lag1 and lag4, respectively. Signif- icant cumulative lag effects of NO2 were found for hospital ad- missions of respiratory symptoms at lag0-1, lag0-2, lag0-3, and lag0-4. (Table 3). The current study also showed significant lag effects of exposure to PM2.5 on respiratory symptoms hos- pitalization but failed to demonstrated such effects for PM10 (Table 3).

NO2 PM10 PM2.5
Respiratory infection
Lag0 3.99 (2.31, 5.71)* 1.87 (0.46, 3.31)* 1.90 (0.91, 2.91)*
Lag1 2.17 (0.50, 3.86)* -0.30 (-1.68, 1.11) 0.64 (-0.36, 1.65)
Lag2 0.18 (-1.46, 1.84) -0.89 (-2.28, 0.30) -0.25 (-1.15, 0.76)
Lag3 -0.43 (-2.06, 1.22) -1.11 (-2.51, 0.30) -0.52 (-1.53, 0.50)
Lag4 -1.03 (-2.22, 0.61) -1.39 (-2.78, 0.02) -0.86 (-1.87, 0.15)
Lag 0-1 4.01 (2.22, 5.84)* 1.07 (-0.47, 2.62) 1.71 (0.63, 2.80)*
Lag 0-2 3.51 (1.59, 5.47)* 0.38 (-1.22, 2.00) 1.26 (0.13, 2.42)*
Lag 0-3 3.03 (1.00, 5.10)* -0.21 (-1.86, 1.47) 0.91 (-0.27, 2.10)
Lag 0-4 2.48 (0.31, 4.69)* -0.66 (-2.37, 1.07) 0.48 (-0.74, 1.72)
Respiratory symptoms
Lag0 1.91 (-0.16, 4.03)# 1.32 (-0.42, 3.08) 0.82 (-0.48, 2.13)
Lag1 2.22 (0.11, 4.37)* 0.09 (-1.62, 1.82) 0.28 (-1.00, 1.57)
Lag2 1.17 (-0.91, 3.30) 0.44 (-1.29, 2.20) 0.69 (-0.59, 1.99)
Lag3 0.98 (-1.10, 3.10) 0.61 (-1.14, 2.39) 0.95 (-0.40, 2.31)
Lag4 3.42 (1.27, 5.61)* 0.67 (-1.07, 2.44) 1.52 (0.18, 2.88)*
Lag 0-1 2.73 (0.51, 4.99)* 0.90 (-0.99, 2.82) 0.72 (-0.67, 2.14)
Lag 0-2 2.92 (0.53, 5.36)* 0.90 (-1.07, 2.92) 0.96 (-0.50, 2.45)
Lag 0-3 3.08 (0.54, 5.68)* 1.02 (-1.04, 3.13) 1.29 (-0.25, 2.86)
Lag 0-4 4.37 (1.61, 7.21)* 1.18 (-0.97, 3.39) 1.84 (0.22, 3.49)*
Asthma (2007 ~ 2008)
Lag0 3.71 (-1.36, 9.04) -0.43 (-4.94, 4.30) 2.10 (-1.44, 5.75)
Lag1 2.42 (-2.61, 7.71) 1.26 (-3.30, 6.04) 3.91 (0.35, 7.59)*
Lag2 2.41 (-2.60, 7.67) 2.75 (-1.84, 7.56) 3.30 (-0.25, 6.96)
Lag3 -2.28 (-7.07, 2.76) 0.50 (-4.00, 5.22) -0.80 (-4.30, 2.83)
Lag4 0.19 (-4.71, 5.35) 2.00 (-2.48, 6.69) -0.25 (-3.64, 3.27)
Lag 0-1 3.78 (-1.61, 9.47) 1.02 (-4.05, 6.35) 4.05 (0.13, 8.12)*
Lag 0-2 4.68 (-1.11, 10.82) 2.25 (-3.13, 7.91) 4.95 (0.79, 9.28)*
Lag 0-3 2.91 (-3.20, 9.41) 2.17 (-3.44, 8.11) 3.89 (-0.43, 8.40)
Lag 0-4 2.62 (-4.01, 9.71) 2.92 (-3.01, 9.22) 3.55 (-0.97, 8.28)

*Statistically significant at 5% level

# Statistically significant at 10% level

Table 3. Percentage change (95% CI) in hospital admission (respiratory infection, respiratory symptoms, and asthma) per interquartile range increases in air pollutants at different lags for WA 2004-2008

Association Between Air Pollution and Number of Hospi- tal Admissions for Asthma, 2007-2008

Due to lack of complete asthma data, the Poisson modelling for asthma hospitalization was developed only for two year peri- od, between January 2007 and December 2008. The study did not establish significant associations between asthma hospi- talization and daily changes in concentrations of NO2 and PM10 (Table 3). However, we established lag effects of exposure to PM2.5 as the risk for asthma hospitalization increased by 3.91% (95% CI: 0.35%, 7.59%) for each unit increase in IQR of PM2.5. at lag 1.

Seasonal Difference in Respiratory Hospitalization

The Poison modeling was also applied to assess if seasonality had an impact on the relationship between respiratory hospi- tal admissions and air pollution. The results are presented in Table 4 and showed no significant association between expo

sure to air pollutants and number of people admitted to hos- pitals with respiratory symptoms, except at lag4 for exposure to NO2 during both seasons. Significant associations, however, were established for hospital admissions with respiratory in- fections and asthma which are discussed in detail below.

Cold Period (May to October)

The association between hospital admissions for respiratory infections and exposure to NO2 during the cold season fol- lowed similar pattern for all year around (Supp – Figure 1).

Exposure to NO2 increased asthma hospitalizations by 6.10% (95% CI: 0.02-12.56) (Table 4). Significant lag effects of NO2 exposures were observed for hospital admissions with asthma with the most significant effects at lag0-1 (6.77%) and lag0-2 (6.86%).

With regards to hospital admissions for respiratory

Lag0 Lag1 Lag2 Lag3 Lag4 Lag 0-1 lag 0-2 lag 0-3 Lag 0-4
Respiratory infections
NO2
Cold 4.15*

(2.22, 6.12)

2.38*

(0.45, 4.35)

0.31

(-1.60, 2.24)

0.09

(-1.83, 2.04)

0.32

(-1.62, 2.30)

4.44*

(2.30, 6.64)

3.62*

(1.54, 5.75)

3.32*

(1.18, 5.51)

3.16*

(0.93, 5.45)

Warm 2.92*

(0.49, 5.41)

1.92

(-0.44, 4.34)

-0.22

(-2.40, 2.00)

-0.63

(-2.86, 1.66)

-0.24

(-2.61, 2.19)

3.20*

(0.53, 5.93)

2.83#

(-0.07, 5.82)

2.08

(-0.83, 5.00)

1.64

(-1.41, 4.79)

PM10
Cold 0.88

(-0.71, 2.49)

-0.23

(-1.80, 1.36)

-0.58

(-2.14, 1.01)

-0.72

(-2.28, 0.85)

-0.91

(-2.45, 0.66)

0.51

(-1.12, 2.17)

0.11

(-1.58, 1.84)

-0.26

(-2.03, 1.54)

-0.69

(-2.49, 1.16)

Warm 2.41* (0.25, 4.44) -0.76

(-2.77, 1.29)

-0.89

(-3.03, 1.30)

-0.50

(-2.65, 1.70)

-0.74

(-2.77, 1.33)

1.06

(-1.40, 3.58)

0.26

(-2.07, 2.64)

0.31

(-2.10, 2.77)

0.33

(-2.10, 2.82)

PM2.5
Cold 1.62*

(0.24, 3.02)

1.08

(-0.28, 2.47)

0.11

(-1.24, 1.47)

0.07

(-1.30, 1.46)

-0.39

(-1.75, 0.99)

1.75*

(0.29, 3.22)

1.41#

(-0.01, 2.87)

1.28

(-0.19, 2.77)

0.94

(-0.52, 2.42)

Warm 2.48*

(0.83, 4.16)

-0.35

(-1.94, 1.27)

-0.90

(-2.60, 0.83)

-0.33

(-2.06, 1.44)

-0.38

(-2.10, 1.38)

1.47

(-0.32, 2.75)

0.72

(-1.26, 2.67)

0.52

(-1.59, 2.67)

0.30

(-1.85, 2.51)

Respiratory symptoms
NO2
Cold 0.96

(-1.42, 3.40)

2.07

(-0.40, 4.61)

0.62

(-1.79, 3.09)

0.48

(-1.96, 2.99)

2.77*

(0.26, 5.35)

1.97

(-0.70, 4.71)

1.83

(-0.79, 4.51)

1.77

(-0.93, 4.55)

2.65

(-0.21, 5.59)

Warm 1.94

(-1.07, 5.03)

0.44

(-2.42, 3.39)

0.86

(-1.99, 3.79)

1.79

(-1.21, 4.87)

2.93*

(0.08, 6.02)

1.57

(-1.65, 4.91)

1.81

(-1.67, 5.40)

2.46

(-1.12, 6.18)

4.04*

(0.08, 8.15)

PM10
Cold 1.40

(-0.69, 3.53)

0.19

(-1.89, 2.31)

-0.20

(-2.25, 1.89)

-0.31

(-2.34, 1.77)

0.39

(-1.65, 2.47)

0.94

(-1.21, 3.14)

0.65

(-1.59, 2.94)

0.42

(-1.92, 2.81)

0.56

(-1.86, 3.04)

Warm -0.18

(-2.81, 2.51)

-0.67

(-3.13, 1.85)

1.03

(-1.65, 3.19)

1.76

(-0.94, 4.53)

1.25

(-1.29, 3.86)

-0.35

(-3.35, 2.75)

0.22

(-2.72, 3.25)

0.92

(-2.16, 4.10)

1.30

(-1.79, 4.48)

PM2.5
Cold 1.48

(-0.43, 3.43)

0.55

(-1.32, 2.46)

0.04

(-1.78, 1.90)

0.21

(-1.68, 2.13)

1.72

(-0.18, 3.65)

1.29

(-0.72, 3.33)

0.98

(-0.97, 2.96)

0.93

(-1.08, 2.99)

1.43

(-0.60, 3.50)

Warm -0.03

(-2.09, 2.07)

-0.40

(-2.36, 1.61)

1.44

(-0.63, 3.55)

2.26*

(0.04, 4.52)

1.01

(-1.15, 3.22)

1.81

(-0.44, 4.21)

-0.34

(-2.50, 1.87)

1.66

(-1.08, 4.49)

2.27

(-0.55, 5.18)

Asthma (2007 ~2008)
NO2
Cold 6.10*

(0.02,12.56)

5.42

(-0.78,12.01)

3.40

(-2.74,9.94)

-3.13

(-8.76,2.86)

1.78

(-4.25,8.18)

6.77*

(0.04,13.95)

6.86*

(0.11,14.07)

4.68

(-2.31,12.17)

5.11

(-2.41,13.21)

Warm -1.58

(-8.79,6.20)

-2.95

(-10.13,4.80)

2.55

(-4.41,10.02)

3.68

(-3.32,11.20)

2.41

(-4.79,10.16)

-3.35

(-11.05,5.01)

-1.21

(-9.65,8.01)

0.93

(-7.62,10.27)

2.02

(-7.22,12.19)

PM10
Cold 0.54

(-4.84, 6.23)

2.51

(-2.99,8.32)

3.87

(-1.59,9.62)

-2.42

(-7.66,3.11)

-1.96

(-7.15,3.51)

1.78

(-3.94,7.85)

3.28

(-2.90,9.85)

1.86

(-4.69,8.86)

0.80

(-6.13,8.25)

Warm 0.26

(-6.27,7.25)

1.33

(-5.07,8.16)

1.48

(-4.85,8.23)

4.80

(-2.07,12.06)

7.69*

(1.07,14.74)

1.72

(-6.18,10.27)

2.53

(-5.15,10.83)

4.22

(-3.66,12.73)

6.25

(-1.65,14.78)

PM2.5
Cold 3.74

(-1.83,9.62)

4.98

(-0.59,10.85)

5.94*

(0.43,11.75)

-2.46

(-7.75,3.14)

0.16

(-5.27,5.90)

5.40

(-0.57,11.72)

6.78*

(0.85,13.07)

5.09

(-1.12,11.69)

4.78

(-1.61,11.58)

Warm 1.91

(-3.29,7.45)

3.89

(-1.30,9.36)

1.74

(-3.31,7.06)

1.86

(-3.15,7.14)

1.07

(-3.62,6.00)

4.64

(-1.33,10.97)

5.16

(-1.66,12.45)

5.72

(-1.43,13.38)

5.54

(-1.50,13.09)

*Statistically significant at 5% level

# Statistically significant at 10% level

Table 4. Percentage change (95%CI) in hospital admission for respiratory infections, respiratory symptoms, and asthma by cold (May to October) and warm (November to April) periods per interquartile increase in air pollutants at different lags, WA 2004 – 20088.

symptoms, a significant effect was found only for NO2 expo- sures indicating that for every 0.5 ppb increase in the con- centration of NO2 the respiratory symptoms hospitalization increased by 2.77% (95% CI: 0.26-5.35).

with the findings of another Australian study conducted ear- lier in Sydney by Morgan[27]. Inconsistent results, however, were reported by a study conducted in Brisbane, Australia[8]. Petroeschevsky and colleagues failed to establish association

Exposure to PM

2.5

increased respiratory infections hospital-

between NO2 and hospitalization for respiratory illnesses but showed a significant impact of particulate air pollution on re-

ization by 1.62% (95%CI: 0.24-3.02) but no effect was es-

tablished for exposure to PM10 during this period (Table 4).

spiratory hospitalizations. Similar results were reported by other Australian studies of Morgan[27] and Voigt[28]. In our

Significant lag effects were established for asthma hospitaliza-

tion and exposure to PM

2.5

(5.94% at lag2 and 6.78% at lag0-2)

study we demonstrated that even at PM2.5 .concentrations of

7.2 µg/m3, which is more than twice less than the reported

(Table 4; Supp-Figure 3).

Warm Period (November to April)

During the warm period, the exposure to air pollutants had a significant impact on hospital admissions for respiratory in- fections. For every unit increase in IQR of NO2 (3.5 ppb), PM10 (8.35 µg/m3) and of PM2.5 (3.2 µg/m3) hospital admissions for

concentration in the study of Hinwood [7] (18.4 µg/m3), fine particulate air pollution in Perth still have a significant impact on respiratory hospitalization. Significant impacts on hospital admissions for respiratory symptoms and asthma were ob- served for longer lagging exposures to fine air particulate pol- lution (PM2.5) but failed to show any significant effects of PM10.

respiratory infections increased by 2.92% (95% CI: 0.49-5.41),

2.41% (95% CI: 0.25- 4.44), and 2.48% (95% CI: 0.83-4.16),

respectively (Table 4, Supp-Figure 1). Significant increase in respiratory infections hospitalization was also established for exposure to NO2 at lag 0-1 (Table 4).

A significant 4-day lag effect of NO2 was found for respiratory symptoms hospitalization with an increase of 2.93% (95% CT: 0.08-6.02) (Table 4). Furthermore, a significant effect of cumu- lative lagging exposures to NO2 was established for respiratory symptoms hospitalization with 4.04 % increase at lag 0-4 but the study failed to demonstrate any significant NO2 impact on asthma hospitalization during the warm period (Supp-Figure 2).

Discussion

The study demonstrated that ambient levels of PM10, PM2.5, and NO2, which are primarily generated from combustion process- es, industrial activities and wood burning, were associated with the risk of hospital admissions for respiratory illnesses in Perth, Western Australia. Similar results were reported in a Perth study conducted earlier (1992-1998) by Hinwood and colleagues[7]. Hinwood found that daily variations in NO2 con- centrations were associated with respiratory disease hospital- ization among the 65 years and above age group. In a more re- cent study from Perth [6] and conducted between 2002-2008, emergency department admissions for asthma among chil- dren, aged 0-4 years, were also significantly associated with daily changes in NO2. The findings were not expected given the low concentrations of NO2 measured in Perth, compared with other cities including Hong Kong [24], Palermo in Italy [25] and Ontario in Canada [26]. The average NO2 concentration measured in our study was 6.98 ppb which is consistent with the concentration of 6.28 ppb reported by the study of Pereira and colleagues[6] but lower than the average concentration of 10.3 ppb reported in the study of Hinwood [25] conducted more than ten years ago. Despite the apparent low exposure levels, the results presented herein indicate that current NO2 levels in Perth still have a significant impact on daily hospi- talization for respiratory diseases. Our findings are consistent

Previous studies have also failed to show consistent effects of PM10 on asthma hospitalization. A study from Birmingham, England [29] showed no significant effect of PM10 on asthma hospitalization although the high concentrations measured for PM10 (130.9 µg/m3). In contrast, effects of PM10 on asthma hospitalization were found in a study in Seattle[30], in Utah Valley[31] and in California[32]. Similar findings were also re- ported in other countries including Canada, Europe and Hong Kong[24,26,33-36].

We can’t provide a clear explanation for such discrepancy re- garding the respiratory health effects of NO2 and air particulate matter but reasons may include population characteristics, cli- mate conditions, and also the complex mixture and different elemental composition of air particulate matter.

The present study has limitations which are similar to other studies as it unable to distinguish between mixtures of air pol- lutants and the confounding effect of co-pollutants. Moreover, due to the correlation between air pollutants and in particular between NO2 and PM2.5, it is difficult to determine the contribu- tion of each pollutant on health effects[37]. Another limitation is that exposure to air pollutants was assessed at aggregated levels by using data from ambient monitoring stations rather than at individual levels, however, any measurement error is most likely to be non-differential and produces conservative estimates of associations[38]. We acknowledge that age of study population may have an impact on the study outcomes, however, due to the limited access to data we couldn’t charac- terise the relationship between air pollution and respiratory hospital admissions for different age groups which is consid- ered as a study limitation. Future studies may be conducted if age groups are made available.

Despite the study limitations, there are number of strengths to this study which include the relatively long time series vari- ables available for up to 5 years, the controlling for long-term time trends, meteorological parameters (temperature, humidi- ty) and other confounding effects (years, seasons, day of week, and public holidays) considered in the Poisson modelling.

In summary, we provide supporting evidence that the ambient air pollution, in particular air particulate matter and nitrogen dioxide, has still significant impact on respiratory hospitaliza- tions in Perth, WA. The results have important public health implications as the established associations are for a city where the concentrations of PM2.5, PM10 and NO2 are relatively low and meet the National Air Quality standards.

Conclusion

This study, similar to previous studies, suggests that, urban air pollution may be responsible for the increase of respiratory hospitalization in Perth, Western Australia. Motor vehicles are one of the major sources of ambient NO2 and PM, and the find- ings of the study reinforce the need for public policy measures to better control air pollution.

Acknowledgment

We wish to acknowledge the contribution of Shima Jadav for assisting with the data collection. We thank the Health Depart- ment and the Department of Environment and Conservation in Western Australia for providing the health and air quality data, respectively.

References
  1. Stieb DM, Judek S, Burnett RT. Meta-analysis of time-series studies of air pollution and mortality: Effects of gases and par- ticles and the influence of cause of death, age and season. Jour- nal of Air and Waste Management Association . 2002, 52 (4): 47-84.
  2. Pope C, Dockery W. Health effects of fine particulate air pollution: lines that connect. J Air Waste Manag Assoc. 2006, 56(6): 709-742.
  3. WHO (World Health Organisation (WHO). Air quality guide- lines for particulate matter, ozone, nitrogen dioxide and sul- fur dioxide. Global update 2005. Summary of risk assessment. WHO Press, 2006. Geneva, Switzerland.
  4. Leitte AM, Petrescu C, Franck U, Richter M, Suciu O et al. Re- spiratory health, effects of ambient air pollution and its mod- ification by air humidity in Drobeta-Turnu Severin, Romania. Science of The Total Environment. 2009, 407(13): 4004-4011.
  5. Simpson R, Williams G, Petroeschevsky A, Best T, Morgan G et al. The short-term effects of air pollution on daily mortality in four Australian cities. The Australian and New Zealand Jour- nal of Public Health . 2005, 29 (3): 205-212.
  6. Pereira G, Cook A, De Vos A, Holman D. A case-crossover anal- ysis of traffic-related air pollution and emergency department presentations for asthma in Perth, Western Australia. Medical Journal of Australia.2010, 193(09): 511-514.
  7. Hinwood A, de Klerk N, Rodriguez C, Jacoby P, Runnion T et al. The relationship between changes in daily air pollution and hospitalizations in Perth, Australia 1992 – 1998: A case-cross- over study. International Journal of Environmental Health Re- search . 2006, 16 (1): 27-46.
  8. Petroeschevsky A, Simpson R, Thalib L, Rutherford S. Asso- ciations between outdoor air pollution and hospital admis- sions in Brisbane, Australia. Archives of Environmental Health. 2001, 56(1): 37-52.
  9. Denison L, Simpson R, Petroeschevsky A, Thalib L, WilliamsG. Air pollution and daily hospital admissions in Melbourne 1994-1997. Melbourne (AUST): EPA Victoria, November 2001, Report N0 789.
  10. Sunyer J, Spix C, Quénel P, Ponce-de-León A, Pönka A et.al. Urban air pollution and emergency room admissions for asth- ma in four European cities: the APHEA Project. Thorax. 1997, 52(9): 760-765.
  11. Atkinson RW, Anderson HR, Sunyer J, Ayres J, Baccini M et al. Acute effects of particulate air pollution on respiratory admissions—Results from APHEA 2 project. Am J Respir Crit Care. 2001, 164(10 Pt 1): 1860-1866.
  12. Dear K, Ranmuthugala G, Kjellstrom T, Skinner C, Hanigan I. Effects of temperature and ozone on daily mortality during the August 2003 heat wave in France. Archives of Environmental and Occupational Health .2005, 60(4): 205-12.
  13. Filleul L, Cassadou S, Médina S, Fabres P, Lefranc A et al. The relation between temperature, ozone, and mortality in nine French cities during the heat wave of 2003. Environmen- tal Health Perspectives. 2006, 114(9): 1344-1347.
  14. Samet M, Zeger L, Dominici F, Curriero F, Coursac I et al. The National Morbidity, Mortality, and Air Pollution Study. Part II: Morbidity and mortality from air pollution in the United States. Res Rep Health Eff Inst. 2000, 94(2): 5-70.
  15. Metzger KB, Tolbert PE, Klein M, Peel JL, Flanders WD et al. Ambient air pollution and cardiovascular emergency depart- ment visits. Epidemiology. 2004, 15(1): 46-56.
  16. National Environmental Protection Council (NEPC). Na- tional Environmental Protection Measure Review Report, May 2011.
  17. Suwanwaiphatthana W, Ruangdej K, Turner-Henson A. Out- door Air Pollution and Children’s Health. Pediatric Nursing. 2010, 36(1): 8-33.
  18. WHO (World Health Organisation) WHO Air Quality Guide- lines for Europe, 2nd Edition, 2000, WHO Regional office for Europe.
  19. Department of Environment and Conservation (DEC). Western Australia Air Monitoring Report 2011, Government of Western Australia.
  20. Tobías A, Sáez M, Galán I, Campbell MJ. Sensitivity analysis of common statistical models used to study the short-term ef- fects of air pollution on health. Int J Biometeorol 2003, 47(4): 227-229.
  21. Marks G, Colquhoun JR, Girgis ST, Hjelmroos KM, Treloar AB et al. Thunderstorm outflows preceding epidemics of asth- ma during spring and summer. Thorax. 2001, 56: 468-471.
  22. Schwartz J, Spix C, Touloumi G, Bacharova L, Barumam- dzadeh T et.al. Methodological issues in studies of air pollution and daily counts of deaths or hospital admissions. Journal of Epidemiology and Community Health . 1996, 50(Suppl 1): S3- S11.
  23. Katsouyanni K, Zmirou D, Spix C, Sunyer J, Schouten JP et.al. Short-term effects of air pollution on health: a European approach using epidemiological time-series data. The APHEA project: background, objectives, design. Eur Respir J 1996, 8: 1030-1038.
  24. Qiu H, Yu I, Tian L, Wang X, Tse LA, et al.Effects of coarse particulate matter on emergency hospital admissions for re- spiratory diseases: a time-series analysis in Hong Kong. Envi- ron Health Perspect. 2012, 120(4): 572-576.
  25. Tramuto F, Cusimano R, Cerame G, Vultaggio M, Calamusa G et al. Urban air pollution and emergency room admissions for respiratory symptoms: a case-crossover study in Palermo, Italy. Environ Health . 2011, 10: 31.
  26. Lin M, Chen Y, Burnett RT, Villeneuve PJ, Krewski D. Effect of short-term exposure to gaseous pollution on asthma hospi- talisation in children: a bi-directional case-crossover analysis. J Epidemiol Community Health . 2003, 57(1): 50-55.
  27. Morgan G, Corbett S, Wlodarczyk J. Air pollution and hos- pital admissions in Sydney, Australia, 1990 to 1994. Am Public Health Assoc 1998, 88(12): 1761- 1766.
  28. Voigt T, Bailey M, Abramson M. Air pollution in the Latrobe Valley and its impact upon Respiratory morbidity. Aus NZ J Public Health . 1998, 22(5): 556-561.
  29. Wordley J, Walters S, Ayres JG. Short term variations in hos- pital admissions and mortality and particulate air pollution. Occup Environ Med. 1997, 54(2): 108-116.
  30. Mittleman M, Maclure M, Robins JM. Control sampling strategies for case-crossover studies: an assessment of relative efficiency. Am J Epidemiol. 1995, 142(1): 91-98.
  31. Pope C. Respiratory hospital admissions associated with PM10 pollution in Utah, Salt Lake, and Cache valleys. Arch Envi- ron Health. 1991, 46(2): 90-97.
  32. Lipsett M, Hurley S, Ostro B. Air pollution and emergency room visits for asthma in Santa Clara County, California. Envi- ron Health Perspect. 1997, 105(2): 216-222.
  33. Lin Mei, Chen Y, Burnett RT, Villeneuve PJ, Krewski D. The Influence of Ambient Coarse Particulate Matter on Asthma Hospitalization in Children: Case-Crossover and Time-Series Analyses. Environ Health Perspect. 2002, 110(6): 575-581.
  34. Host S, Larrieu S, Pascal L, Blanchard M, Declercq C et.al. Short-term associations between fine and coarse particles and hospital admissions for cardiorespiratory diseases in six French cities. Occup Environ Med . 2008, 65(8): 544-551.
  35. Tecer L, Alagha O, Karaca F, Tuncel G, Eldes N. Particulate matter (PM(2.5), PM(10-2.5), and PM(10)) and children’s hos- pital admissions for asthma and respiratory diseases: a bidi- rectional case-crossover study. J Toxicol Environ Health . 2008, 71(8): 512-520.
  36. Wong T, Lau TS, Yu TS, Neller A, Wong SL et al. Air pollu- tion and hospital admissions for respiratory and cardiovascu- lar diseases in Hong Kong. Occup Environ Med. 1999, 56(10): 679-683.
  37. Luginaah IN, Fung KY, Gorey KM, Webster G, Wills C. Associ- ation of ambient air pollution with respiratory hospitalization in a government-designated area of concer: the case of Wind- sor, Ontario. Environ Health Perspect. 2005, 113(3): 290-296.
  38. Jalaludin B, Khalaj B, Sheppeard V, Morgan G. Air pollution and ED visits for asthma in Australian children: a case-cross- over analysis. Int Arch Occup Environ Health . 2008,81(8): 967-974.

Be the first to comment on "Association Between Ambient Air Pollution and Hospitalization for Respiratory Diseases in Perth, Australia"

Leave a comment

Your email address will not be published.


*