Investigating the relationship between environmental factors and respiratory health outcomes in school children using the forced oscillation technique

Accepted Manuscript Title: Investigating the relationship between environmental factors and respiratory health outcomes in school children using the Forced Oscillation Technique Authors: Jonathon Boeyen, Anna C. Callan, David Blake, Amanda J. Wheeler, Peter Franklin, Graham L. Hall, Claire Shackleton, Peter D. Sly, Andrea Hinwood PII: DOI: Reference:

S1438-4639(16)30271-1 http://dx.doi.org/doi:10.1016/j.ijheh.2017.01.014 IJHEH 13045

To appear in: Received date: Revised date: Accepted date:

27-8-2016 28-12-2016 3-1-2017

Please cite this article as: Boeyen, Jonathon, Callan, Anna C., Blake, David, Wheeler, Amanda J., Franklin, Peter, Hall, Graham L., Shackleton, Claire, Sly, Peter D., Hinwood, Andrea, Investigating the relationship between environmental factors and respiratory health outcomes in school children using the Forced Oscillation Technique.International Journal of Hygiene and Environmental Health http://dx.doi.org/10.1016/j.ijheh.2017.01.014 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.

Investigating the relationship between environmental factors and respiratory health outcomes in school children using the Forced Oscillation Technique. Jonathon Boeyen a, Anna C. Callan b, David Blake a, Amanda J. Wheeler a, Peter Franklin c, Graham L. Hall d, Claire Shackleton e, Peter D. Sly e. Andrea Hinwood a* a

Centre for Ecosystem Management, Edith Cowan University, 270 Joondalup Drive, Joondalup, WA 6027, Australia; e-mail: [email protected]. [email protected]. [email protected]. [email protected]. b School of Medical and Health Sciences, Edith Cowan University, 270 Joondalup Drive, Joondalup, WA 6027, Australia; e-mail: [email protected]. c School of Population Health, The University of Western Australia, 35 Stirling Highway, Crawley, WA 6009, Australia; e-mail: [email protected]. d Children’s Lung Health Telethon Kids Institute, School of Physiotherapy and Exercise Science Curtin University, Centre for Child Health Research, University of Western Australia, PO Box 855, West Perth WA 6872 Australia e-mail: [email protected] e Child Health Research Centre, The University of Queensland, 62 Graham St, South Brisbane Qld 4101 Australia; e-mail: [email protected]. *Corresponding author: Andrea Hinwood Centre for Ecosystem Management Edith Cowan University 270 Joondalup Drive, Joondalup, WA 6027, Australia Telephone: +61 8 6304 5372 E-mail: [email protected]

1

Abstract The environmental factors which may affect children’s respiratory health are complex, and the influence and significance of factors such as traffic, industry and presence of vegetation is still being determined. We undertook a cross-sectional study of 360 school children aged 5 to 12 years who lived on the outskirts of a heavy industrial area in Western Australia to investigate the effect of a range of environmental factors on respiratory health using the forced oscillation technique (FOT), a non-invasive method that allows for the assessment of the resistive and reactive properties of the respiratory system. Based on home address, proximity calculations were used to estimate children’s exposure to air pollution from traffic and industry and to characterise surrounding green space. Indoor factors were determined using a housing questionnaire. Of the outdoor measures, the length of major roads within a 50m buffer was associated with increased airway resistance (Rrs8). There were no associations between distance to industry and FOT measures. For the indoor environment the presence of wood heating and gas heating in the first year of life was associated with better lung function. The significance of both indoor and outdoor sources of air pollution and effect modifiers such as green space and heating require further investigation.

Keywords: children, FOT, green space, industry, traffic, respiratory health

2

Introduction It is well established that children are more vulnerable to the effects of air pollution as they spend more time participating in outdoor physical activities, display elevated breathing rates, inhale and retain more air than adults per unit of body weight, and their narrow bronchioles are more prone to constriction upon exposure to environmental irritants (Hansen et al., 2003; Committee on Environmental Health, 2004). Environmental factors have been shown to have an influence on the respiratory health of children and both indoor and outdoor sources of air pollution have been implicated (Brauer et al. 2007; Markandya & Wilkinson, 2007; FuentesLeonate et al. 2009; Gillespie-Bennett et al. 2011).

Exposure to traffic-related air pollution has been associated with the prevalence and exacerbation of asthma and respiratory symptoms in children (Brauer et al. 2007; Gehring et al. 2010; Gordian et al. 2006; Kim et al. 2008; McConnell et al. 2006; Nordling et al. 2008). However adverse effects to respiratory health are not observed in all studies (Fuertes et al. 2013; Rosenlund et al. 2009). Exposure to industry-related air pollution may also result in adverse respiratory health effects in children (Cara et al. 2007; Rusconi et al. 2011), however, the type and volume of pollutants emitted from industrial facilities varies enormously, and hence the potential for health impacts is also variable. Industrial activities which may influence children’s respiratory health include fossil fuel electricity generation (Markandya & Wilkinson, 2007), metal works (Cara et al. 2007) and the use of heavy vehicles and their subsequent diesel emissions (Riedl & Diaz-Sanchez, 2005). Indoor sources of air pollution include gas and wood burning for heating and cooking which have been associated with respiratory symptoms and effects in children (Triche et al. 2002; Fuentes-Leonarte et al. 2009; Gillespie-Bennett et al. 2011). Contrary findings have also been reported (Bennett et al. 2010).

Few studies have investigated the influence of effect modifiers such as urban vegetation on air quality and human exposure to air pollution. Nowak et al. (2006) found that urban trees absorb a variety of air pollutants, improving air quality and potentially reducing exposures and hence reducing the effects of exposure to air pollutants. A study of pregnant women by Dadvand et al. (2012) associated living in greener areas with reduced levels of personallymonitored air pollution exposure to fine particulate matter (PM) and nitrogen oxides (NOx). Studies investigating the relationship between vegetation density and asthma and allergic 3

sensitisation have generated conflicting results with some showing improvements in health outcomes with increasing greenness and others the reverse (Fuertes et al. 2014; Lovasi et al. 2008; 2013). Given these findings, further work is required to clarify the role of these potential effect modifiers and sources of pollution on children’s respiratory health. This is imperative as exposure to pollutants during critical stages of children’s growth may affect both lung structure and function (Kajekar, 2007).

The forced oscillation technique (FOT) is a non-invasive lung function method which allows the assessment of respiratory system resistance (Rrs) and reactance (Xrs) related to respiratory system compliance (Oostveen et al., 2003). The FOT is an attractive lung function assessment technique for children due to its reduced requirement for participant cooperation and high success rates when compared with other established lung function tests such as spirometry (Navajas & Farre, 2001). The FOT has been used to evaluate alterations in respiratory mechanics in infants, children, young people and adults with common respiratory diseases, including recurrent wheeze, asthma, bronchial hyperresponsiveness, cystic fibrosis and broncopulmonary dysplasia (Oostven et al. 2003; Rosenfeld et al. 2013). Additionally, the FOT has been used to examine the effects of atopy and environmental tobacco exposure among other variables (Calogero et al. 2013; Gray et al. 2015), suggesting that it may be useful when investigating the effects of environmental factors on lung function.

This cross-sectional study was conducted to examine the influence of environmental factors on the lung function (using FOT) of primary school-aged children in the Kwinana region of Western Australia using questionnaire-based information and spatial analyses. The Kwinana Industrial Area contains a range of light and heavy industries and is Western Australia’s primary area for industrial development. Industries include oil and alumina refining, nickel processing, cement works, chemical and pesticides manufacturing as well as shipping and metal works-type activities (www.npi.gov.au).

Materials and Methods Study Population and Recruitment Children, aged between 5 and 12 years, were recruited from ten primary schools in the Kwinana region of Western Australia (Figure 1). The number of children recruited for the study was 591 of a total of 2466 who were invited, a response fraction of 24.0%. Parents completed a respiratory health questionnaire and a questionnaire on the child’s current and 4

past home environments. Children with parental consent and personal assent completed lung function testing using the FOT and atopy testing (skin prick test). All testing was conducted at the schools during normal school hours. Testing was done in June 2009 (winter). All children who volunteered to participate in the study and were at school on the day of testing were included in the study. The presence of evidence that children presented with an upper respiratory tract infection (‘cold’) at the time of testing was noted.

Data Collection Questionnaire Information on each child’s demographic, lifestyle and health characteristics was obtained as reported by parents via questionnaire. The questionnaire included questions found in the literature to impact upon lung health. These included: family history of asthma, premature birth, maternal smoking, smoking in the home, use of unflued gas heaters in the home, presence of an internal garage, exposure to pets and parental education. Duration of residency in Kwinana was also examined, as children born and raised in locations with different air pollution levels can have different respiratory health outcomes (McConnell et al. 2006). Information was also provided on respiratory symptoms/disease in the last year and medication use, both historical and present.

Lung function Respiratory system impedance, a measure of respiratory function, was attempted in all children consented into the study using the FOT according to international guidelines (Oostveen 2003, Beydon et al. 2007) and as reported previously by Hall et al. (2007) and Calogero et al. (2013). Measurements were obtained from a commercially available FOT device (12M, Chess Medical, Belgium) which uses a pseudo-random noise forcing signal between 4 and 48 Hz. The device was verified daily using a known resistance and measured impedances corrected for the impedance of the device and mouthpiece.

Briefly children performed the tests sitting upright with their head in a neutral position. Measurements were obtained with the child breathing via a mouthpiece incorporating a bacterial filter (Suregard, Bird Healthcare, Australia). Children were instructed to breathe normally while wearing a nose clip with the cheeks and floor of the mouth supported by an investigator. Measurements were excluded if leak, mouth or tongue movement, swallowing, glottal closure, talking or audible noise occurred during the test. Three to five acceptable 5

measurements of respiratory system impedance (Zrs) were obtained for each child. Rrs and Xrs were calculated at each frequency and the mean of the total acceptable measurements was reported. The frequency dependence of the Rrs between 4 and 24 Hz and the resonant frequency (the point at which Xrs is zero) were determined from the average of Zrs of all acceptable measurements. The area under the reactance curve (AX) was calculated between 6 Hz up to the resonant frequency (being the frequency at which the Xrs curve equals zero). We aimed to obtain a within-test variability of Rrs of less than 10%. However, individual measurements and the subsequent averaged Zrs data were not excluded if this criterion was not met. The primary FOT outcomes reported in this study were Rrs and Xrs at a frequency of 8 Hz (Rrs8 and Xrs8), Fres and AX, expressed as standardised z-scores controlling for each individual’s age, sex and height based on reference equations from Calogero et al. (2013) and Fdep (the frequency dependence). Higher scores for Rrs8, Fres and AX are associated with poorer respiratory function (Oostveen et al. 2003), whereas higher scores for Xrs8 and Fdep correspond to better respiratory function (Calogero et al. 2013).

Atopy Atopy testing was undertaken using skin prick tests for the following locally prevalent allergens: peanut, cat, dog, grass mix, house dust mite and mould mix, together with negative (saline) and positive (histamine) controls. A positive response was taken as a weal of 3 mm or greater in size than the saline control. Children were marked as atopic by the investigators if a positive response to any of the six allergens was shown.

Ethical approval This study received approval from the Princess Margaret Hospital Human Research Ethics Committee in 2009. Subsequently, approval was obtained from both Princess Margaret Hospital and Edith Cowan University (ECU) Human Research Ethics Committee (Project Number 11329) specifically for the analysis described in this paper. No personal identifiers were sent to ECU researchers and all addresses were geocoded.

Exposure assessment Exposure assessment was conducted using children’s homes as proxies for their location, with linear and buffer calculations used to estimate children’s exposure to traffic and industry and to determine their surrounding greenness based on the ESCAPE protocols (Beelen et al. 6

2014). Linear distance from each child’s residence to the geographic centre of the Kwinana Industrial Area was measured. Nearest distance of a child’s residence to a weighted centre point for NOx emissions (derived from annual pollution inventory data for heavy industry in the Kwinana region) was used as an estimate of exposure to industry. NOx has previously been shown as a strong marker for many hazardous air pollutants including PM and other oxides of nitrogen (Beckerman et al. 2008). All spatial analyses were conducted using ArcGIS (ESRI, released 2013, ArcGIS Version 10.2., Redlands CA: ESRI).

Nearest distance to the closest major road and road density within a buffer were used as surrogates of traffic exposure as applied in previous respiratory health studies (Brauer et al. 2007). For road density calculations, buffers of 50m, 100m, and 200m, 300 and 500m in radius were applied following the ESCAPE protocols (Beelen et al. 2014), with smaller buffer sizes excluded due to the lower number of subjects participating in this study, and larger buffer sizes not considered due to the size of the study area. Road-based traffic calculations were conducted using Main Roads Western Australia’s road hierarchy data set for 2012, with major roads classified as such if they were a ‘Primary’, ‘District’ or ‘Regional’ distributor under the related criteria (Main Roads Western Australia, 2011a; 2011b). Surrounding greenness at each child’s location was determined using the Normalized Difference Vegetation Index (NDVI) derived from Landsat 7 ETM SCL-off data at a resolution of 30 m x 30 m. The Landsat 7 ETM imagery obtained was cloud-free, captured during the daytime and dated within the month of the original KCRHS sampling period. NDVI approximates green vegetation density based on the visible (red) and near-infrared light reflected by the land surface (Wade & Sommers, 2006). NDVI values fall within -1 to 1, where values of ≤0 indicate areas with no green vegetation content and values approaching 1 indicate areas with high green vegetation density (Wade & Sommers, 2006). Buffer calculations were used to approximate vegetation around children’s locations in a similar manner to the techniques applied by Lovasi et al. (2013) and Markevych et al. (2014) in related health studies. As limited information on how proximity to vegetation influences children’s lung function exists, the mean NDVI values within a 100 m, 200 m, 300 m and 500 m radius of children’s locations were considered. Surface water features were masked out prior to calculation of the mean to avoid influencing the NDVI results.

Statistical Analyses 7

The independent effect of children’s individual characteristics (personal, health, housing and lifestyle) and environmental variables on their respiratory function was investigated using SPSS (IBM Corp., released 2013, IBM SPSS Statistics for Windows Version 22.0, Armonk, NY: IBM Corp.). Fdep and standardised z-scores for Rrs8, Xrs8, Fres, and AX were used as indicators of the children’s respiratory function. Spearman correlations and Mann Whitney U tests were performed to identify individually significant relationships (p-value <0.05), with all variables with a p-value <0.05 included in subsequent multiple regression analyses.

Multiple linear regressions for each standardised FOT measure were performed. Each FOT variable was considered separately. All variables shown to have an effect on each FOT variable in descriptive analyses and for which literature indicated it may be significant were included in the analysis. All included variables added to each model via forced entry, using a p-value of >0.05 for removal. Where included variables were thought to overlap (e.g. length of roads within a 50 m buffer and length of roads within a 100 m buffer) the variable with the smallest beta value was excluded from the regression model. These models allowed the effects of exposure to traffic, industry and surrounding green space on children’s respiratory health to be assessed, as well as other factors known to affect children’s respiratory health such as current cold. Only those variables that were significant in regression models were included in the final analysis.

Results Characteristics of the study population Of the 591 children who participated in the study only 360 had satisfactory FOT measures (60.9% of participating children and 14.6% of the total invited population). The personal, health, housing and lifestyle characteristics of these children are presented in Table 1.

Several characteristics known to influence respiratory health were prevalent in the cohort, including maternal smoking during pregnancy and maternal smoking during the child’s first year of life (approximately 35%). Approximately 13% of children were reported as currently having asthma with over 40% reported to have used asthma medication in their lifetime. Cough and wheeze symptoms in the past year were reported in relatively few participants. Nearly 40% of the cohort currently had a cat in the home. The highest average level of 8

education of parents of children in this study was secondary high school, with approximately 30% of parents having a university education. Most children were reported as having lived in the Kwinana region for more than 5 years with a mean (range) residency time of 6.5 years (0.21 to 12.3 years).

Table 1. Personal characteristics of the study population with valid FOT measures. Measurement (N = 360) n Mean Range Age (years) 360 9.1 5.0 - 12.4 Height (cm) 360 134 108 - 164 Weight (kg) 359 34.7 17.6 - 87.6 BMI 359 18.6 12.9 - 33.0 Characteristic Currently has asthma Ever taken medication for asthma Wheeze in the past year Cough lasting two weeks in the past year Ever had bronchitis and had cough during the night in the past year Premature birth (gestation <37 weeks) Parental history of ever having asthma, wheezing, eczema or allergic rhinitis Mother smoked during pregnancy Mother smoked during the child’s first year of life Mother smokes now

n 47 137 58 27 23

n valid 356 331 354 353 350

% prevalence 13.2 41.4 16.4 7.6 6.6

25 260

356 351

7.0 74.1

115 110

321 320

35.8 34.4

95

318

29.9

331

Smoking frequency inside the home Never 296 Occasionally 28 Frequently 7 Pet in home at present Cat in the home at present Cat in the home during the first year of life Father’s highest level of education Primary school Secondary school Higher than secondary school Mother’s highest level of education Primary school Secondary school Higher than secondary school

279 119 100

89.4 8.5 2.1 321 304 286

86.9 39.1 35.0

213 11 126 76

5.2 59.2 35.7 223

2 120 101

0.9 53.8 45.3

9

The housing characteristics of the cohort are shown in Table 2. Most homes were older (>5 years) with a third having a garage with internal access to the home. Use of gas in homes was common with over three-quarters of participants using gas for cooking, and gas heating currently being used in nearly 50% of the children’s homes. Approximately 14% of children lived in a home with no reported heating and there were several types of heating used in the same home for approximately 25% of participants. Less than 10% of children reported living within 50 m of a major road.

Table 2. Housing characteristics of the study population with valid FOT measures. Characteristic (n = 360) Age of home <5 years old Home has an attached vehicle garage Home has vehicle garage with internal access at present

n 76 205 98

n valid 331 308 306

% prevalence 23.0 66.6 32.0

360

Types of heating used at present No heating 51 One type 217 Two or more types 92

14.2 60.3 25.5

Types of heating used in the first year of life No heating 94 One type 177 Two or more types 89

360

Gas used for cooking in home at present Any wood burning for heat at present Any wood burning for heat used during first year of life If gas heating used vented to outside at present If gas heating used vented to outside during first year of life.

253 54 88

321 360 360

78.8 15.0 24.4

115 67

301 301

38.2 22.3

20 30 66 96

346 346 346 346

5.8 8.7 19.1 27.7

26.1 49.2 24.7

Questionnaire-reported nearest major road within

50m 100m 150m 200m

Environmental variables such as proximity to traffic and surrounding greenness were estimated for children using their home address. For 19 children, environmental variables 10

were not calculated due to concerns about the positional accuracy of their geocodes. The environmental variables calculated for children with valid FOT measures are summarised in Table 3. The mean NDVI values around children’s homes were low (means <0.16) indicative of low vegetation content being present in Kwinana’s residential area. No children were located within 3.7 km of the weighted centre point for NOx emissions for heavy industry in the Kwinana region (Table 3).

Table 3. Descriptive analysis of GIS-calculated environmental variables for traffic, greenness and industry based on the location of children’s homes. Environmental variable (n = 360) Mean Range SD Length of roads (m) within a buffer size of 50 m 103 0.00 - 241 49.8 100 m 418 0.00 - 945 154 200 m 1,550 0.00 - 2890 517 300 m 3,290 59.9 - 5,620 1,070 500 m 8,210 588 - 15,300 2,560 Length of major roads (m) within a buffer size of 50 m 5.44 0.00 - 161 24.9 100 m 29.3 0.00 - 397 91.8 200 m 143 0.00 - 1,330 283 300 m 348 0.00 - 2,080 490 500 m 997 0.00 - 3,520 970 Length of Primary Distributor roads (m) within a buffer size of 50 m 2.4 0.00 - 138 12.5 100 m 15.4 0.00 - 375 60.5 200 m 82.8 0.00 - 1030 206 300 m 211 0.00 - 1594 388 500 m 613 0.00 - 2715 810 NDVI within a buffer size of 100 m 0.12 -0.04 - 0.35 0.07 200 m 0.13 -0.01 - 0.32 0.07 300 m 0.14 0.03 - 0.31 0.06 500 m 0.15 0.05 - 0.32 0.05 Linear distance to industrial area centre point weighted by annual NOx emissions (m)

7,930

3,790 - 25,200

2,920

Surface water area (m2) within a buffer size of 500 m 1000 m

9,112 84,500

0 - 292,000 0 - 1,530,000

28,464 160,000

Baseline FOT z-score outcomes and Fdep for the children are shown in Table 4. Children also completed Skin Prick Tests (SPT) and of the 343 children who successfully completed them, 40.8% were atopic. 11

The effect of questionnaire-reported and geographical information system (GIS) generated variables on FOT outcomes was assessed and shown in Supplementary Tables S1-3. Atopy, characterised by skin prick tests had no influence on any of the FOT measures. Increased resistance (Rrs8) was weakly correlated with increased length of roads within 200m (rs = 0.112, p = 0.038) and 500m (rs = 0.120, p = 0.026) and the length of primary roads within 100m (rs = 0.123, p = 0.023); reactance was weakly correlated with length of roads within 50m (rs = -0.113, p = 0.035) (Spearman correlation coefficients for all environmental variables shown in Supplementary Table S1).

Table 4. FOT values of the study population. Measure* n Mean SD Range Resistance (z-score) Rrs8 360 0.33 0.93 -2.04 - 3.03 Reactance (z-score) Xrs8 360 -0.23 0.98 -3.94 - 1.63 Area under reactance curve (z-score) AX 360 0.20 1.03 -2.16 - 3.38 Resonant frequency (z-score) Fres 357 0.19 1.01 -2.51 - 4.30 Frequency dependence of Rrs Fdep 360 -0.10 0.07 -0.33 - 0.11 *Frequency 8 Hz; Rrs8, Xrs8, AX and Fres presented as z-scores standardised using reference equations from Calogero et al. (2013).

Green space variables were not associated with FOT outcomes (Supplementary Table S1). For linear distance to industry using a cut point of 5km, living within 5km was shown to have a negative effect on Fres (z = 2.243, p = 0.025) and Fdep (z = -2.598, p = 0.009). Children who lived in homes <5 years had poorer lung function than those who lived in homes >5 years: Xrs8 (z = -2.525, p = 0.012); AX (z = 2.419, p = 0.016) and Fdep (z = -2.125, p = 0.034) (Supplementary Table S2).

Premature birth was associated with poorer lung function as measured by Fres (z = 2.910, p = 0.004) and children who had ever taken medication for asthma had poorer lung function: Rrs8 (z = 2.347, p = 0.019). There was a significant difference between those who presented with a current cold (n = 37) and those without for AX (z = 2.318, p = 0.020); Fres (z = 2.341, p = 0.019) and Fdep (z = -2.011, p = 0.044) with children presenting with a cold having poorer lung function. Of the information on symptoms, only the questionnaire-reported diagnosis of bronchitis (n = 33) was associated with poorer lung function as measured by Rrs8 (z = 2.630, p = 0.009).

12

When the presence of a heater at the present time was investigated there were significant differences in Xrs8 (z = -2.603, p = 0.009); AX (z = 2.422, p = 0.015) and Fdep (z = -4.414, p = <0.001) showing better lung function associated with heating use. The same pattern was shown for any heating and no heating at the first year of life: Xrs8 (z = -3.501; p = <0.001); AX (z = 3.051, p = 0.002); Fdep (z = -3.641, p = <0.001) (Supplementary Table S3). When home heating type was analysed, both the nomination of gas and wood heating types were associated with better lung function. Wood burning for heat at present compared with no wood burning heater at the present time resulted in no significant differences in lung function except for Fdep (z = -1.981, p = 0.048). When wood heating use in the first year of life was examined, lung function measures were better: Xrs8 (z = -2.090, p = 0.037); AX (z = 2.057, p=0.040) in those using wood heaters. Similar results were observed for gas heating at present and in the first year of life as well as central heating for the first year of life (Supplementary Table S3). The combined effect of emission heaters versus non-emission heater use was assessed with those using emission-based heating having better lung function (Supplementary Table S3). The effect of venting gas from a gas heater to the outside was also assessed and for participants reporting venting their gas heater to the outside at the present time, a higher Rrs8 was observed indicating poorer lung function. The same result was seen when data for the first year of life was examined (Supplementary Table S3).

Few variables, considered significant in univariate analyses, remained significant predictors of lung function in subsequent multivariate regression analyses. Very small percentages of the variation in FOT variables were explained by environmental factors as shown by the very small R2 values (between 1.8% and 6.4% of the variation in the respective FOT variables (Table 5). Lung function as measured by FOT was influenced by asthma medication use (Rrs8), being born preterm (Fres) and had recorded a current cold at the time of testing (AX and Fres). After adjusting for these factors, FOT outcomes demonstrated small but significant changes associated with environmental factors. Home heating in the first year of life was associated with a small but significant improvement in Rrs8, Xrs8, AX and Fdep. Increased length of roads within a 50m buffer was associated with poorer lung function as measured by Rrs8 but did not influence other FOT outcomes (Table 5). Age of home <5 years was linked to worse AX and Fdep. Resonant frequency (Fres) was not associated with environmental factors.

13

Table 5. Linear regression analyses of children’s environmental variables and their influence on standardised FOT measures (z-scores). Outcome

Rrs8 n = 318

Variables in Model

Constant Any heating during the first year of life Length of major roads within 50m Ever taken medication for asthma

Xrs8 n = 358

Constant Any heating during the first year of life

AX n = 329

Constant Any heating during the first year of life Age of home <5 years Current cold

Fres n = 351

Constant Born prematurely Current cold

Fdep n = 330

Constant Any heating during the first year of life Age of home <5 years

Standardised coefficients* Beta p-value

95% Confidence Interval for Beta Lower Upper bound bound 0.174 0.534 -0.254 -0.061

R2 0.073

Adjusted R2** 0.064

-0.260 0.271

0.030

0.028

-0.029 -0.231

0.327 0.001

0.048

0.039

0.012 0.051

0.073 0.093

0.596 0.811

0.077 0.001 0.022

-0.011 0.272 0.058

0.210 1.088 0.728

0.050

0.044

0.680 0.393

<0.001 0.059

-0.109 <0.001

-0.085 0.015

0.024

0.018

0.103 -0.113

0.039

-0.037

-0.001

-0.174

0.002 0.001

0.159

0.004

0.002

0.009

0.132

0.016

0.046

0.438

0.174

<0.001 0.001

-0.540 0.070

-0.115

0.101 0.051

0.335 0.452

* Standardised Coefficient Beta: estimate from the linear regression modelling if model were fitted to standardised data. ** Adjusted R2 is the explanatory ability of the model to predict the variability in the responses where there are a variety of predictor variables.

Discussion This study suggests the impact of environmental factors on children’s lung function is complex with subtle variations in effect. Using the forced oscillation technique to quantify respiratory function, we found no effect of green space and a small and non-significant effect 14

from proximity to industry. In contrast proximity to traffic and newer homes were associated with poorer respiratory function outcomes and home heating in the first year of life was associated with improved lung function.

Increased length of road within a 50 m buffer was a predictor of poorer lung health (as measured by Rrs8), while several other traffic variables were significant in univariate analyses. This finding was consistent with the results of studies reported by Nordling et al. (2008) and Rosenlund et al. (2009). Dales et al. (2009), also reported a modest increase in the odds of children’s asthma using road density within a 200 m buffer around homes as a surrogate for exposure to traffic-related air pollution, and Kim et al. (2004) and McConnell et al. (2006) identified higher risks of asthma living within 25 m of a major road and within 5075 m of a major road. This finding also supports the suggestion by Hoek et al. (2008) that variability in traffic-related air pollution is limited at distances greater than approximately 100 m away from a major road. While this study did not find surrounding greenness to be a factor that influenced children’s lung function, there are mixed results in the literature. A negative association between street tree density and the prevalence of asthma was reported by Lovasi et al. (2008) and later, Lovasi et al. (2013) identified a positive association between tree canopy coverage and the prevalence of asthma and allergic sensitisation. Conflicting results were also reported by Fuertes et al. (2014) in a study of two populations of children residing in different areas where vegetation density was positively associated with allergic rhinitis and eyes and nose symptoms in one population, and in the second, vegetation density was negatively associated with allergic rhinitis, eyes and nose symptoms and aeroallergen sensitisation (Fuertes et al., 2014).

Univariate analyses indicated an effect of distance from the centre of the Kwinana Industrial Area using a cut point of 5km, with better lung function found for children living greater than 5km away, however this was not significant in regression analyses. Two stationary air monitors were active within the study area as part of the Department of Environment and Conservation’s Background Air Quality Study (BAQS) during the sampling period and air pollution levels in the Kwinana airshed were well below ambient standards for both NO2 and PM2.5 and comparable to those found elsewhere in the Perth metropolitan region, 15

(Department of Environment and Conservation, 2011). It is possible that the amount of air pollution emitted from industry in the Kwinana Industrial Area is low, or that the residential area is located too far from the source for emitted air pollutants to have any influence on children’s respiratory health, with most children living >5km away from the centre of the industrial zone. The locational methods used in this study may also have been too simplistic to address the complexity of the pollutants emitted to the airshed and subsequent dispersion patterns of emissions.

Indoor sources of air pollution have previously been shown to have adverse effects on lung function and the housing questionnaire sought to identify whether housing was an important predictor. The finding that unflued gas heating and wood heating when being used at the present time and in the first year of life resulted in better lung function was not expected although the effect sizes were very small. Most of the studies that have investigated indoor emitting heaters (those producing pollutants such as nitrogen dioxide (NO2), PM and volatile organic compounds (VOCs) such as unflued gas heaters and wood heaters report increased respiratory symptoms (Phoa et al., 2002; Nitschke et al., 2006; Marks et al., 2010; Bennett et al., 2010) and reduced lung function (Nitschke et al., 2006). In this study better lung function was observed for any heating compared with no heating, and therefore the unexpected findings may be due to temperature rather than emissions. Pierse et al. (2013) showed an improvement in lung function in asthmatic children with a 1°C increase in temperature in homes and Howard-Chapman et al. (2008) reported no change in lung function in children in homes where an improved heating system (less emissions) was installed, however they did observe a significant reduction in respiratory symptoms in children. Marks et al. (2010) did not observe any effect of unflued gas heaters on lung function in their study of school children using a double blind cluster randomised cross over study in New South Wales, although they did see effects on respiratory symptoms not observed in this study. Further, a wood heater study of children and adults in Australia did not find an increase in adverse respiratory symptoms in homes with higher wood heater use and smoke (Bennett et al. 2010). Li et al. (2016) examined the effect of cold temperatures on lung function found a statistically significant decrease in forced vital capacity (FVC) with decreasing ambient temperature. We did not measure temperature or indoor air pollution in this study so cannot determine if either of these were associated with lung function. It is possible that colder temperatures during 16

infancy and childhood may increase the risk of adverse respiratory outcomes and hence any heating may improve respiratory health as the data suggests in this study. There are two other possible explanations for these findings. Firstly the lack of heating may be related to socioeconomic factors (Hegewald & Crapo, 2007) not assessed in this study but which may influence lung function development. Of the asthmatic children in this study, 14% reported no heating and 29% of non-asthmatics reported no heating used at present or in the first year of life. It is possible that parents with asthmatic children may be more attentive to their comfort and that the results simply reflect this via the provision of heating. Secondly, it may be a chance finding. Larger longitudinal studies would be required to investigate these issues.

Age of home, use of asthma medication and being born prematurely were also identified as significant predictors of children’s respiratory health in the linear regression analyses, although again the models explained little of the variance in the FOT variables. Being born prematurely and neonatal chronic lung disease have been consistently demonstrated to lead to decrements in lung function and increased respiratory symptoms (Vergheggen et al. 2016) and the findings of this study confirm earlier studies (Harju et al. 2014). Age of home <5 years was a significant predictor, however the results must be treated with caution due to the small effect size, nevertheless it is possible that there are pollutants associated with newer homes which may contribute to poorer lung function (Jaakola et al. 2004; Mendell, 2007). There were a number of limitations with this study, most importantly the cross-sectional nature of the design of the study and hence temporal variations could not be taken into account when measuring children’s lung function. Other temporal variations that were not accounted for included wind speed and direction, atmospheric stability, mixing height and topography, all factors which may influence air pollution exposure (Hagler et al. 2009; Hagler et al. 2010; Hitchins et al. 2000; Hu et al. 2009).

Traffic intensity is often a significant indicator of exposure to traffic-related air pollution (Hoek et al., 2008) and relies on count data which were limited to a few available counters in this study. Given the small area, discernment of different types of traffic and in particular heavy traffic was challenging and these factors may have been important as they have been shown to be important for discerning exposures in other studies (Kanaroglou & Buliung, 2008). Similarly, it was only possible to obtain annual emissions data for a limited proportion 17

of industry within the Kwinana region. This was not ideal and led to the use of singular points (the weighted centre point for annual NOx emissions or geographic centre point of the industrial area) in estimating children’s exposure to industry-related air pollution via distance calculations, which lacked specificity.

In this investigation, NDVI was used to characterise the vegetation content of the Kwinana region. NDVI is not able to discern between different types of vegetation and is sensitive to certain soil types, clouds and atmospheric effects (Weier & Herring, 2000). Also, as Landsat 7 imagery has a resolution of 30 m, the influence of vegetation in smaller scales on air pollution exposure may not have been captured. However, NDVI has been used successfully in exposure studies using a resolution of 30 m (Markevych et al. 2014; Villeneuve et al. 2012).

Only 14.6% of the invited school population participated in this study. When the results of the prevalence of asthma and other factors which influence respiratory disease were compared with other Australian data (University of Queensland, 2012), fewer (-7.1%) of the Kwinana children were found to be atopic following skin prick testing. However many more children had mothers who smoked during pregnancy (+13.1%) or during their first year of life (+13.5%) compared with those from the Australian Child Health and Air Pollution Study (ACHAPS) (University of Queensland, 2012). The prevalence of other individual characteristics between this group of children and ACHAPS cohorts were similar, and therefore the recruited children can be considered largely representative of the wider population of Australian primary school children (University of Queensland, 2012).

The authors acknowledge that parents with children with asthma or other adverse respiratory symptoms may have been more inclined to accept the invitation to participate in this study while they were not targeted specifically and this may have influenced the findings. It is also possible that acute exposure to air pollutants leading up to children’s lung function measurements impacted their measurement on the day of testing. This could not be accounted for in calculations as no information was collected on children’s activities immediately prior their testing (e.g. method of transport to school).

18

Had extensive air quality monitoring been conducted during the sampling period at various locations within the Kwinana region (e.g. primary schools) interpolation methods could have increased the ability of the GIS data model to estimate children’s exposures (Briggs, 2005). Additionally, as participation in the study was not mandatory, the study population may have been biased (e.g. skewed towards children with parents who are more educated or have a greater awareness of health risks, who could have introduced measures to reduce exposures) and hence not representative of the wider population. Associations have been found between socioeconomic status (SES) and reduced lung function (Hegewald & Crapo, 2007), and given the Kwinana region is below-average in SES versus other regions in Australia (Australian Bureau of Statistics, 2011) it is possible that SES factors not investigated in this study may have explained some of the variation in children’s lung function.

Despite these limitations, this study has enabled the assessment of a number of air pollution sources and effect mediators at the individual level to be examined using the forced oscillation technique.

Conclusions This study aimed to investigate the relationship between environmental factors and children’s respiratory health using the forced oscillation technique. The results provide some evidence for the effect of traffic proximity to adversely affect lung function as measured by respiratory resistance. The results also suggest home heating may be a more important factor in children’s respiratory health in their early years than other environmental exposures, however larger studies with indoor and outdoor measures of exposure, including temperature, would be required to confirm this. There are few studies of effect modifiers such as temperature and green space and further effort is required to establish the relative importance of sources of pollution, housing location and design and green space to better understand how to manage and prevent impacts on the developing respiratory system.

Acknowledgements The authors would like to thank the Kwinana parents and children for their participation, the KCRHS investigators for the use of the data, and the Centre for Ecosystem Management and Department of Health for funding the study.

References 19

Australian Bureau of Statistics (ABS). 2011. 2033.0.55.001 – Census of population and housing: socio-economic indexes for areas (SEIFA), Australia, 2011: local government area (LGA) index of relative socio-economic disadvantage, 2011. Retrieved from http://www.abs.gov.au/ausstats/[email protected]/mf/2033.0.55.001 Beckerman, B., Jerrett, M., Brook, J. R., Verma, D. K., Arain, M. A. & Finkelstein, M. M. 2008. Correlation of nitrogen dioxide with other traffic pollutants near a major expressway. Atmospheric Environment. 42(2), 275-290. Beelen, R., Raaschou-Nielsen, O., Stafoggia, M., Andersen, Z. J., Weinmayr, G., Hoffmann, B. et al. 2014. Effects of long-term exposure to air pollution on natural-cause mortality: an analysis of 22 European cohorts within the multicentre ESCAPE project. Lancet. 383(9919), 785-795. Bennett, C. M., Dharmage, S.C., Matheson, M., Gras, J.L., Markos, J., Mesaros, D. et al. 2010. Ambient woodsmoke exposure and respiratory symptoms in Tasmania, Australia. Sci Total Environ. 409(2), 294-299. Beydon N., Davis S. D., Lombardi E., Allen J. L., Arets H. G., Aurora P. et al. 2007. American Thoracic Society/European Respiratory Society Working Group on Infant and Young Children Pulmonary Function Testing. An official American Thoracic Society/European Respiratory Society Statement: pulmonary function testing in preschool children. Am J Respir Crit Care Med. 175, 1304-1345. Brauer, M., Hoek, G., Smit, H. A., de Jongste, J. C., Gerritsen, J., Postma, D. S. 2007. Air pollution and development of asthma, allergy and infections in a birth cohort. Eur Respir J. 29(5), 879-888. Briggs, D. 2005. The role of GIS: coping with space (and time) in air pollution exposure assessment. J Toxicol Environ Health A. 68(13-14), 1243-1261. Calogero, C., Simpson, S. J., Lombardi, E., Parri, N., Cuomo, B., Palumbo, M. et al. 2013. Respiratory impedance and bronchodilator responsiveness in health children aged 2-13 years. Pediatric Pulmonology. 48(7), 707-715. Cara, A., Buntinx, F., Akker, M., Dinana, G. & Manolovici, C. 2007. Industrial air pollution and children’s respiratory health: a natural experiment in Calarasi. European Journal of General Practice. 13(3), 135-143. Committee on Environmental Health. 2004. Ambient air pollution: health hazards to children. Pediatrics. 114(6), 1699-1707. Dadvand, P., de Nazelle, A., Triguero-Mas, M., Schembari, A., Cirach, M., Amoly, E. et al. 2012. Surrounding greenness and exposure to air pollution during pregnancy: an 20

analysis of personal monitoring data. Environmental Health Perspectives. 120(9), 12861290. Dales, R., Wheeler, A. J., Mahmud, M., Frescura, A. M., Liu, L. 2009. The influence of neighborhood roadways on respiratory symptoms among elementary schoolchildren. J Occup Environ Med. 51(6), 654-660. Department of Environment and Conservation (DEC). 2011 Background Air Quality Monitoring in Kwinana 2005-2010. Technical Report. Prepared by the Department of Environment and Conservation Air Quality Management Branch. Fuentes-Leonarte, V., Ballester, F., Ten, J, M. 2009. Sources of indoor air pollution and respiratory health in preschool children. Journal of Environmental and Public Health. doi: 10.1155/2009/727516. Fuertes, E., Markevych, I., von Berg, A., Bauer, C., Berdel, D., Koletzko, S. et al. 2014. Greenness and allergies: evidence of differential associations in two areas in Germany. Journal of Epidemiology and Community Health. 68(8), 787-790. Fuertes, E., Standl, M., Cyrys, J., Berdel, D., von Berg, A., Bauer, C. et al. 2013. A longitudinal analysis of associations between traffic-related air pollution with asthma, allergies and sensitization in the GINIplus and LISAplus birth cohorts. PeerJ. 1, e193. doi: 10.7717/peerj.193 Gehring, U., Wijga, A. H., Brauer, M., Fischer, P., de Jongste, J. C., Kerkhof, M. et al. 2010. Traffic-related air pollution and the development of asthma and allergies during the first 8 years of life. Am J Respir Crit Care Med. 181(6), 596-603. Gillespie-Bennett, J., Pierse, N., Wickens, K., Crane, J., Howden-Chapman, P. and the Housing Heating and Health Study Research Team. 2011. The respiratory health effects of nitrogen dioxide in children with asthma. European Respiratory Journal. 38(2), 303309. Gordian, M. E., Haneuse, S., Wakefield, J. 2006. An investigation of the association between traffic exposure and the diagnosis of asthma in children. J Expo Sci Environ Epidemiol. 16(1), 49-55. Gray, D., Czovek, D., Smith, E., Willemse, L., Alberts, A., Gingle, Z., et al. 2015. Respiratory impedance in healthy unsedated South-African infants: effects of maternal smoking. Respirology. 20(3), 467-473. Hagler, G. S. W., Baldauf, R. W., Thoma, E. D., Long, T. R., Snow, R. F., Kinsey, J. S. et al. 2009. Ultrafine particles near a major roadway in Raleigh, North Carolina: downwind

21

attenuation and correlation with traffic-related pollutants. Atmospheric Environment. 43(6), 1229-1234. Hagler, G. S. W., Thoma, E. D. Baldauf, R. W. 2010. High-resolution mobile monitoring of carbon monoxide and ultrafine particle concentrations in a near-road environment. Journal of the Air & Waste Management Association. 60(3), 328-336. Hansen, C., Neller, A., O’Farrell, T., van der Pols, J., Simpson, R. Williams, G. 2003. Air Quality and Child Health – Current Evidence and Priorities for Australia. Brisbane, Queensland: Department of the Environment and Heritage. Harju, M., Keski-Nisula, L., Georgiadis, L., Raisanen, S., Gissler, M., Heinonen, S. 2014. The burden of childhood asthma and late preterm and early term births. The Journal of Pediatrics. 164(2), 295-299. Hegewald, M. J. & Crapo, R. O. 2007. Socioeconomic status and lung function. Chest. 135(2), 1608-1614. Hitchins, J., Morawska, L., Wolff, R. Gilbert, D. 2000. Concentrations of submicrometre particles from vehicle emissions near a major road. Atmospheric Environment. 34(1), 51-59. Hoek, G., Beelen, R., de Hoogh, K., Vienneau, D., Gulliver, J., Fischer, P., Briggs, D. 2008. A review of land-use regression models to assess spatial variation of outdoor air pollution. Atmospheric Environment. 42(33), 7561-7578. Howden-Chapman, P., Pierse, N., Nicholls, S., Gillespie-Bennett, J., Viggers, H., Cunningham, M. et al. 2008. Effects of improved home heating on asthma in community dwelling children: randomised controlled trial. BMJ. 337 doi: 10.1136/bmj.a.1411 Hu, S. S., Fruin, S., Kozawa, K., Mara, S., Paulson, S. E. & Winter, A. M. 2009. A wide area of pollutant impact downwind of a freeway during pre-sunrise hours. Atmospheric Environment. 43(16), 2541-2549. Jaakkola, J. J., Parise, H., Kislitsin, V., Lebedeva, N.I. and Spengler, J.D. 2004. Asthma, wheezing, and allergies in Russian schoolchildren in relation to new surface materials in the home. Am. J. Public Health. 94, 560–562. Kajekar, R. 2007. Environmental factors and developmental outcomes in the lung. Pharmacology & Therapeutics. 114(2), 129-145. Kanaroglou, P. S. & Buliung, R. N. 2008. Estimating the contribution of commercial vehicle movement to mobile emissions in urban areas. Transportation Research Part E. 44(2), 260-276. 22

Kim, J., Smorodinsky, S., Lipsett, M., Singer, B. C., Hodgson, A.T., Ostro, B. 2004. Trafficrelated air pollution near busy roads. American Journal of Respiratory and Critical Care Medicine. 170(5), 520-526. doi: 10.1164/rccm.200403-281OC Kim, J. J., Huen, K., Adams, S., Smorodinsky, S., Hoats, A., Malig, B. et al. 2008. Residential traffic and children’s respiratory health. Environmental Health Perspectives. 116(9), 1274-1279. Li, Q., Wang, H. J., Song, Y., Ma, J., Song, J. Y., Gou, Y., 2016. Association between children’s forced vital capacity and long-term exposure to local ambient temperature in China: A National cross sectional survey. Sci Total Environ. 1, 557-558: 880-887. doi: 10.1016/j.scitotenv.1016.02.135.Epub 2016 April 2017 Lovasi, G. S., O’Neil-Dunne, J., Lu, J., Sheehan, D., Perzanowski, M. S., MacFaden, W. et al. 2013. Urban tree canopy and asthma, wheeze, rhinitis, and allergic sensitization to tree pollen in a New York City birth cohort. Environmental Health Perspectives. 121(4), 494-500. Lovasi, G. S., Quinn, J. W., Neckerman, K. M., Perzanowski, M. S. Rundle, A. 2008. Children living in areas with more street trees have lower prevalence of asthma. Journal of Epidemiology and Community Health, 62(7), 647-649. Main Roads Western Australia. 2011a. Road hierarchy for Western Australia: road types and criteria. Retrieved from https://www.mainroads.wa.gov.au/Documents/Road%20Hierarchy%20Criteria%20%20April%202011.u_3158367r_1n_D11%5E2392185.PDF Main Roads Western Australia. 2011b. Statewide Traffic Digest: 2003/04 – 2008/09. Retrieved from https://www.mainroads.wa.gov.au/Documents/Traffic%20Digest.u_3169030r_1n_D11 %5E2398970.PDF Markandya, A. & Wilkinson, P. 2007. Electricity generation and health. Lancet. 370(9591), 979-990. Markevych, I., Thiering, E., Fuertes, E., Sugiri, D., Berdel, D., Koletzko, S., et al. 2014. A cross-sectional analysis of the effects of residential greenness on blood pressure in 10year old children: results from the GINIplus and LISAplus studies. BMC Public Health. 14(1), 477. Marks, G. B., Ezz, W., Aust, N., Toelle, B.G., Xuan, W., Belousova, E. et al. 2010. Respiratory health effects of exposure to low-NOx unflued gas heaters in the

23

classroom: a double blind, cluster-randomised crossover study, Environ Health Perspect. 118(10), 1476-1482. McConnell, R., Berhane, K., Yao, L., Jerrett, M., Lurmann, F., Gilliland, et al. 2006. Traffic, susceptibility, and childhood asthma. Environmental Health Perspectives. 114(5), 766-772. Mendell, M. J. 2007. Indoor residential chemical emissions as risk factors for respiratory and allergic effects in children: a review. Indoor Air. 17(4), 259-277. Navajas, D., Farré R. 2001. Forced oscillation technique: from theory to clinical applications. Monaldi Arch Ches Dis. 56, 555-62. Nordling, E., Berglind, N., Melen, E., Emenius, G., Hallberg, J., Nyberg, F et al. 2008. Traffic-related air pollution and childhood respiratory symptoms, function and allergies. Epidemiology. 19(3), 401-408. Nowak, D. J., Crane, D. E., Stevens, J. C. 2006. Air pollution removal by urban trees and shrubs in the United States. Urban Forestry and Urban Greening. 4(3-4), 115-123. Oostven, E., MacLeod, D., Lorino, H., Farré, R., Hantos, Z., Desager, K., et al. 2003. The Forced Oscillation Technique in clinical practice: methodology, recommendations and future developments. Eur Resp J.22, 1026–41. 10.1183/09031936.03.00089403 Pierse, N., Arnold, R., Keall, M., Howden-Chapman, P., Crane, J., Cunningham, M., Heating Housing and Health Study Group. 2013. Modelling the effects of low indoor temperatures on the lung function of children with asthma. J Epidemiol Community Health. 1;67(11), 918-25. doi: 10.1136/jech-2013-202632. Phoa, L. L., Toelle, B. G., Ng, K., & Marks, G. B. 2004. Effects of gas and other fume emitting heaters on the development of asthma during childhood. Thorax. 59(9), 741745. Riedl, M. & Diaz-Sanchez, D. 2005. Biology of diesel exhaust effects on respiratory function. Journal of Allergy and Clinical Immunology. 115(2), 221-228. Rosenfeld, M., Allen, J., Arets, H., Aurora, P., Beydon, N., Calogero, C. et al. on behalf of the ATS ad hoc committee. 2013. An Official ATS Workshop Report: Optimial Lung Function Tests for Monitoring Cystic Fibrosis, Bronchopulmonary Dysplasia and Recurrent Wheezing in in Children < 6 Years of Age. Ann Am Thorac Soc. 10(2). S111. Rosenlund, M., Forastiere, F., Porta, D., De Sario, M., Badaloni, C. Perucci, C. A. 2009. Traffic-related air pollution in relation to respiratory symptoms, allergic sensitisation and lung function in schoolchildren. Thorax. 64(7), 573-580. 24

Rusconi, F., Catelan, D., Accetta, G., Peluso, M., Pistelli, R., Barbone, F., et al. 2011. Asthma symptoms, lung function, and markers of oxidative stress and inflammation in children exposed to oil refinery pollution. J. Asthma. 48(1), 84-90 Triche, E.W., Belanger, K., Beckett, W., Bracken, M.B., Holford, T.R., Gent, J. et al. 2002. Infant respiratory symptoms associated with indoor heating sources. Am J Respir Crit Care Med. 166(8), 1105-1011. University of Queensland. 2012. Australian Child Health and Air Pollution Study: Final Report. Brisbane, Queensland: University of Queensland. Verheggen, M., Wilson, A.C., Pillow, J.J., Stick, S.M., Hall, G.L. 2016. Respiratory function and symptoms in young preterm children in the contemporary era. Pediatr Pulmonol. 51(12), 1347-1355. Villeneuve, P. J., Jerrett, M., Su, J. G., Burnett, R. T., Chen, H., Wheeler, A. J., & Goldberg, M. S. 2012. A cohort study relating urban green space with mortality in Ontario, Canada. Environ Res. 115, 51-58. Wade, T. & Sommer, S. 2006. A to Z GIS: An Illustrated Dictionary of Geographic Information Systems. Redlands, California: Esri Press. Weier, J. & Herring, D. 2000. Measuring vegetation (NDVI & EVI). Retrieved from http://earthobservatory.nasa.gov/Features/MeasuringVegetation/measuring_vegetation_2.p hp

25

Figure 1. Map of the study area including the location of the 10 primary schools from which participants were recruited.

26