Long term impact of coal mine fire smoke on lung mechanics in exposed adults

In 2014, a six-week long fire at the Hazelwood open cut coal mine exposed residents in the adjacent town of Morwell to high concentrations of fine particulate matter with an aerodynamic diameter <2.5µm (PM 2.5). The long-term health consequences are being evaluated as part of the Hazelwood Health Study (HHS). Approximately 3.5 to 4 years after the mine fire, adults from Morwell (n=346) and the comparison town Sale (n=173) participated in the longitudinal Respiratory Stream of the HHS. Individual fire-related PM 2.5 exposure was retrospectively modelled. Lung mechanics were assessed using the forced oscillation technique (FOT), which utilises pressure waves to measure respiratory system resistance (Rrs) and reactance (Xrs). Multivariate linear regression was used to evaluate associations between PM 2.5 and transformed Rrs5, area under the reactance curve (AX5) and Xrs5 controlling for key confounding factors. There were clear dose-response relationships between increasing mine fire PM 2.5 and worsening lung mechanics, including a reduction in post-bronchodilator Xrs5 and an increase in AX5. A 10 µg/m3 increase in mine fire related PM2.5 was associated with a 0.015 (95%CI: 0.004, 0.027) reduction in exponential(Xrs5) post bronchodilator, which was comparable to 4.7 years of aging. Similarly, the effect of exposure was associated with a 0.072 (0.005, 0.138) increase in natural log(AX5) post-bronchodilator, equivalent to 3.9 years of aging. This is the first study using FOT in adults evaluating long term respiratory outcomes after a medium-term ambient PM 2.5 exposure to coal mine fire smoke. These results should inform public health policies and planning for future events. . CC-BY-NC 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 20, 2020. ; https://doi.org/10.1101/2020.10.14.20213009doi: medRxiv preprint 3

Ambient air particulate matter (PM) exposure, from sources including vehicle exhaust, industry, biomass fuels and wildfires, collectively account for an estimated 7.5% of all deaths globally in 2016.[1] In particular, fine PM with an aerodynamic diameter < 2.5µm (PM 2.5) infiltrates deep into the peripheral lung. Short-term (days) exposure to PM 2.5 has been shown to be associated with cardiovascular and respiratory morbidity and mortality.[2,3] Guo et al[4] showed in a large cohort study that long-term ambient PM 2.5 exposure was consistently associated with reduced lung function, accelerated annual lung function decline and an increased risk of developing chronic obstructive pulmonary disease (COPD) in adults. Similarly in a review, Li et al[5] showed an association with long-term exposure to ambient air pollution levels and increased incidence of respiratory symptoms in children. The long-term sequelae of fine particle exposures on lung function, particularly from medium exposure episodes (weeks to months) such as landscape fires, have not been well characterised. Studies of wildfires predominantly use secondary data such as hospitalization and emergency presentations to identify respiratory associations.[6] Though long-term exposure to indoor coal burning has been found to be associated with worsening respiratory symptoms, reduced lung function and chronic obstructive pulmonary disease in adults - much remains unknown regarding the impact of coal mine fires on human lung health.[6] Addressing the gaps in the current available evidence is critical given the increasing incidence of catastrophic wildfires globally attributable to climate change.[7] The forced oscillation technique (FOT) is a methodology used to measure lung mechanics. FOT may be able to detect early changes in peripheral airway function that spirometry cannot.[8] To our knowledge, no study has assessed the long-term impact of PM 2.5 from exposure to coal mine fires, wildfires, or biomass fuel smoke in adults using FOT. In February 2014, embers from nearby bush fires started a fire in the Hazelwood open-cut brown coal mine, located in the Latrobe Valley, south-eastern Australia. It was an unprecedented event that generated significant air pollution from coal mine fire smoke over six weeks, particularly affecting residents in the adjacent town of Morwell. The most exposed population at the time numbered approximately 14,000.[9] This resulted in considerable community concerns about the potential long-term health effects of smoke exposure. The Hazelwood Health Study (www.hazel woodhealthstudy.org.au ) was established to investigate potential health effects in people who were exposed to smoke from the mine fire. . CC-BY-NC 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 20, 2020. ; https://doi.org/10.1101/2020.10.14.20213009doi: medRxiv preprint 4 The Hazelinks stream of the Hazelwood Health Study utilised hospital emergency presentations and admissions data to show that hospitalisation for respiratory conditions increased during the first month of the mine fire.[10] The Adult Survey stream of the Hazelwood Health Study[11] compared self-reported health outcomes between the most exposed community and an unexposed sample more than two years after the event. The Survey found increasing risks of respiratory symptoms, particularly cough, phlegm and wheeze, related to the mine fire exposure.[12] This analysis aimed to further investigate the association between exposure to mine fire smoke and long-term lung function as assessed by FOT, 3.5 to four years after the event.

Study design and setting The Respiratory Stream of the Hazelwood Health Study is a longitudinal follow-up study of selected participants from the Adult Survey.[11,13]. The study was conducted between August and December 2017 in Morwell (exposed), and between January and March 2018 in the nearby town of Sale (unexposed). Study data were collected and managed using REDCap (Research Electronic Data Capture)[14] electronic data capture tools hosted at Monash University, Australia. Participant eligibility and recruitment Participants were eligible for the Respiratory Stream of the Hazelwood Health Study if they had completed the Adult Survey, were at least 18 years of age on 9 February 2014 and had lived in the study area at the time of the mine fire. Adult Survey participants were excluded from the Respiratory Stream if they had specified no further contact, were of unknown age or sex, or were aged over 90 years. Participants were further excluded where a contraindication to spirometry was identified – including recent surgery, myocardial infarction, pneumothorax, pulmonary embolism, open pulmonary tuberculosis or known aneurysms.[15] A target sample size of 339 from Morwell and 170 from Sale was derived based on the ability to detect a 5ml/year or greater FEV 1 decline in exposed compared with non-exposed participants using a two-sample t-test with a two-sided p-value of 0.05 and 80% power. A weighted random sample (to correct for lower response rate in some subgroups of participants, such as young people) of 1,346 Adult Survey participants was invited for assessment of their respiratory function. Participants reporting an asthma attack or current asthma medication use in the Adult Survey were oversampled (40%) to provide ability for . CC-BY-NC 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 20, 2020. ; https://doi.org/10.1101/2020.10.14.20213009doi: medRxiv preprint 5 further evaluation in an asthmatic sample. Invitation to participate was by mail, email and/or SMS, and recruitment continued until the target sample size was achieved (see Figure 1). Participant characteristics Participant characteristics such as age, sex, ethnicity, employment status and smoking history were collected via questionnaires. Participants were classified as non-smokers (100 cigarettes in their lifetimes) or current smokers.[16] Height and weight were measured by trained personnel during the study visit. Education level and occupational exposures (employment in dusty or polluted environments for at least six months) were obtained from the Adult Survey.[11] Self-reported asthma status was captured via a modified European Community Respiratory Health Survey questionnaire.[17] Participants were identified as having spirometry consistent with COPD if post bronchodilator (BD) FEV 1/FVC < lower limit of normal (5 th percentile) using Global Lung Initiative spirometry reference values.[18] Exposure assessment Retrospective modelling of the spatial and temporal distribution of mine fire-related PM 2.5 concentrations by the Australian Commonwealth Scientific and Industrial Research Organisation (CSIRO) Oceans & Atmosphere[19,20] was used due to the absence of ground- level air pollution monitoring at the beginning of the mine fire. Individual level mean daily PM 2.5 exposures over the mine fire period (51 days, between 9 February to 31 March, 2014) were estimated through linking time-location diary data (reported in the Adult Survey) with the modelled fire-related PM2.5 exposure data as described by Johnson et al.[12] Clinical outcome measures Respiratory testing was performed by the same trained respiratory scientists at both sites using standard operating procedures in line with current respiratory measurement standards where available. Spirometry was measured using the EasyOne Pro Lab Respiratory Analysis System (ndd Medical Technologies AG, Zürich, Switzerland) in line with international standards.[21] Forced Oscillation Technique (FOT) parameters were measured using the Tremoflo C-100device (Thorasys, Montreal, Canada) in line with standards current at time of testing.[22] Parameters reported for FOT included respiratory system resistance and reactance at a frequency of 5Hz (Rrs5 and Xrs5 respectively), the area under the reactance curve (AX5) and resonant frequency (Fres). Data were imputed[23] where acceptability criteria were not met or coherence <0.80 for 5Hz or <0.90 at 11 or 19Hz.[22] Tests were . CC-BY-NC 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 20, 2020. ; https://doi.org/10.1101/2020.10.14.20213009doi: medRxiv preprint 6 performed before and 10 minutes after administration of a short acting bronchodilator (300µg salbutamol). Bronchodilator use in the previous 24 hours was recorded, as bronchodilators were unable to be withheld prior to assessment due to ethical reasons. Statistical methods Statistical weighting was developed and applied to all analyses to correct for over-sampling of asthmatics as well as possible attrition bias from the Adult Survey to clinical follow-up, see details in the online supplement. Descriptive statistics were used to compare patient characteristics and clinical outcomes between non-exposed Sale participants as well as the tertiles of PM 2.5 exposure level in Morwell (low, medium or high exposure). Crude statistical significance was assessed using Pearson chi-squared tests for categorical measures and t-tests for continuous measures. Multivariate linear regression models were fitted to analyse the association between mean PM2.5 exposure and outcomes, controlling for key confounders including age, height, weight, sex, smoking status, self-reported asthma and/or COPD, employment, education level and occupational exposure. Standardised z-scores and %predicted[24] for FOT outcome variables were not used in the analysis due to poor regression model fit and high proportions of participants outside of reference prediction range (mostly due to older age and heavier weight). Therefore, possible outcome transformation methods and nonlinear associations were explored using both Box-Cox transformation and fractional polynomial regression models. The best outcome transformation methods were identified as logarithmic transformations for Rrs5, AX5 and Fres and exponential transformation for Xrs5. Additional non-linearity was not observed between transformed outcomes and predicators such as age, weight and height. Missing data were addressed using multiple imputation using chained equations. Due to the lack of a low or no exposure sample in Morwell, as well as possible differences between Morwell and Sale participants, two sets of regression models were carried out for each outcome variable: one model including a binary variable indicating township of participant (Morwell or Sale), and the other model excluding this variable. Sensitivity analyses were performed with unweighted and complete case models. Statistical analyses were performed using Stata version 15 (Stata Corporation, College Station, Texas 2015). . CC-BY-NC 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 20, 2020. ; https://doi.org/10.1101/2020.10.14.20213009doi: medRxiv preprint 7 Ethical considerations The Monash University Human Research Ethics Committee (MUHREC) approved the Hazelwood Health Study: Cardiovascular and Respiratory Streams (approval number 1078). All participants provided written informed consent.

Participant characteristics and PM2.5 exposure This cross-sectional analysis included all participants in the first round of Respiratory Stream data collection, which comprised a total of 519 participants (346 from Morwell, and 173 from Sale). Refer to Figure 1 for flow of participants. Table 1 shows the participant characteristics by exposure level to mine fire smoke. The mean (standard deviation; SD) PM2.5 exposure levels for non-exposed (Sale) and for Morwell (low, medium and high exposure groups) were 0.1 (0.4), 5.9 (1.8), 11.5 (1.5) and 27.8 (10.3) µg/m3, respectively. There were differences between exposure groups for gender distribution and weight, with those in the high exposure group having a higher proportion of males and heavier weight. Other participant characteristics were comparable between exposure groups. Table 1: Participant characteristics by exposure group. Characteristic Sale Morwell low exposure Morwell medium exposure Morwell high exposure p-value N=173 N=109 N=113 N=124 Age group / years, n (weighted %) 18-44 44 (22%) 36 (26%) 36 (30%) 35 (25%) 0.74 45-64 74 (42%) 43 (44%) 43 (37%) 50 (37%) 65+ 55 (36%) 30 (29%) 34 (34%) 39 (38%) Gender, n (weighted %) Male 62 (36%) 43 (46%) 46 (43%) 62 (56%) 0.02 Caucasian/White, n (weighted %) 171 (99%) 108 (99%) 112 (100%) 123 (99%) 0.92 Employed, n (weighted %) 89 (47%) 44 (38%) 46 (40%) 50 (35%) 0.34 Higher education*, n (weighted %) 107 (63%) 56 (57%) 54 (54%) 74 (64%) 0.39 . CC-BY-NC 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 20, 2020. ; https://doi.org/10.1101/2020.10.14.20213009doi: medRxiv preprint 8 BMI kg/m2, n (weighted %) Underweight/Normal (BMI<25) 40 (24%) 23 (20%) 21 (18%) 15 (11%) 0.06 Overweight (25≤ BMI<30) 66 (38%) 35 (33%) 31 (29%) 35 (30%) Obese (BMI≥ 30) 67 (38%) 51 (47%) 61 (53%) 74 (59%) Smoking status, n (weighted %) Non-smoker 82 (49%) 58 (52%) 60 (54%) 49 (36%) 0.10 Ex-smoker 66 (39%) 35 (33%) 34 (33%) 51 (47%) Current smoker 25 (12%) 16 (15%) 19 (13%) 24 (17%) Asthma and/or COPD †, n (weighted %) 73 (37%) 58 (41%) 57 (40%) 63 (39%) 0.96 Historical occupational exposure, n (weighted %) 64 (37%) 43 (44%) 44 (39%) 54 (46%) 0.47 Age / years 57.3 (20.0) 54.7 (14.3) 54.5 (15.3) 56.7 (14.7) 0.50 Height / cm 166.5 (11.2) 166.0 (9.6) 166.4 (8.8) 167.1 (7.9) 0.86 Weight / kg 81.2 (24.4) 86.7 (22.7) 86.8 (20.2) 88.8 (17.0) 0.009 Mean (SD) PM 2.5 exposure / µg/m3 0.1 (0.4) 5.9 (1.8) 11.5 (1.5) 27.8 (10.3) * Certificate, University or other Tertiary Institute degree † Spirometric COPD and/or self-reported asthma attack in the last 12 months PM2.5 exposure and lung function Forced Oscillation Technique variables are dependent on sex, age, height and weight – hence unadjusted results lack meaning and were not included in the analysis. As shown in Figures 2A and 2B and Table S1 , all outcome variables were skewed and displayed slightly larger variation in baseline compared to post bronchodilator outcomes. A clear dose response pattern was observed between exposure level and FOT outcomes. Results from multivariate linear regression analysis (Table 2) revealed a negative association between increasing mine fire related PM 2.5 exposure and post bronchodilator reactance at 5Hz, with Morwell included or excluded as a predictor. With Morwell excluded as a predictor, a 10 µg/m 3 increase in mine fire related PM 2.5 was associated with 0.015 reduction in post bronchodilator exponential transformed Xrs5. This was equivalent to 4.7 years of aging estimated in the regression model (see Table S2). When Morwell was excluded as a predictor, regression . CC-BY-NC 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 20, 2020. ; https://doi.org/10.1101/2020.10.14.20213009doi: medRxiv preprint 9 analysis suggested that increased exposure to mine fire related PM 2.5 was associated with increased area under the post bronchodilator reactance curve (AX5). The effect of exposure was associated with a 0.072 increase in ln(AX5) post-bronchodilator; being equivalent to 3.9 years of aging (see Table S3). More detailed regression results for post bronchodilator Xrs5 and AX5 are shown in supplementary Tables S2 to S3 . Sensitivity analyses (results not shown) suggest that un-weighted and complete case results were consistent with main findings. Table 2. Summary table for multivariate linear regressions of FOT parameters – regression coefficients ( β ) and 95% confidence intervals. Mean exposure model (10 µg/m³) Including Morwell as predictor Mean exposure model (10 µg/m³) Excluding Morwell as predictor β -Coef (95% CI) p-value β -Coef (95% CI) p-value Baseline* Baseline ln(Rrs5) -0.003 (-0.035, 0.029) 0.87 -0.001 (-0.028, 0.026) 0.95 Baseline exp(Xrs5) -0.009 (-0.024, 0.006) 0.23 -0.008 (-0.020, 0.005) 0.23 Baseline ln(AX5) 0.030 (-0.056, 0.116) 0.50 0.038 (-0.034, 0.109) 0.31 Baseline ln(Fres) 0.001 (-0.030, 0.032) 0.97 0.006 (-0.019, 0.032) 0.62 Post BD† Post BD ln(Rrs5) 0.011 (-0.018, 0.041) 0.45 0.012 (-0.013, 0.036) 0.34 Post BD exp(Xrs5) -0.018 (-0.032, -0.003) 0.015 -0.015 (-0.027, -0.004) 0.011 Post BD ln(AX5) 0.063 (-0.017, 0.144) 0.12 0.072 (0.005, 0.138) 0.034 post BD ln(Fres) 0.017 (-0.010, 0.045) 0.22 0.021 (-0.001, 0.044) 0.07 * Regression models adjusted for age, gender, height, weight, employment, education, smoking status, asthma and COPD status and work exposure and whether participants had bronchodilator prior to the test. Missing data, including 44 records for baseline Rrs5, baseline Xrs5 and baseline AX5; 48 records for baseline Fres; 6 records for education level, were imputed using multiple imputation with chained equations. † Regression models adjusted for age, gender, height, weight, employment, education, smoking status, asthma and COPD status and work exposure. Missing data, including 41 records for post BD Rrs5, post BD Xrs5 and post BD AX5; 42 records for post BD Fres; 6 records for education level, were imputed using multiple imputation with chained equations. Note: exp- exponential; ln- natural log . CC-BY-NC 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 20, 2020. ; https://doi.org/10.1101/2020.10.14.20213009doi: medRxiv preprint 10

Assessment of participants nearly four years after the Hazelwood coal mine fire revealed an association between medium term mine fire related PM 2.5 exposure and more negative respiratory system reactance (Xrs5), specifically measured after administration of bronchodilator. To the best of our knowledge, this represents the first study using FOT analysis in adults to evaluate longer term respiratory function after a medium term PM 2.5 exposure related to coal mine fire smoke. The mechanism for the more negative reactance (a marker of the compliance of the respiratory system) is unclear. Previous studies of long-term exposure to air pollution and PM 2.5 have shown associations with increased respiratory morbidity and airflow obstruction.[4,6,25-27] Separately, it has been shown that measurements of reactance at 5- 6Hz via FOT are sensitive to airway closure[28-31] and expiratory flow limitation[32-34] in subjects with obstruction. A possible mechanism for the association seen between medium term exposure PM 2.5 and Xrs5 in this study may be early peripheral airway changes that occur with airflow limitation or accelerated lung aging. Interestingly, the association between PM 2.5 and Xrs5 was only observed in the post- bronchodilator data. A possible explanation for this finding is that participants were recruited from a general population with varying states of lung health and by assessing participants post-bronchodilator, variability of bronchomotor tone was minimised across participants[35- 37] allowing assessment of fixed pulmonary abnormalities. That is, the assessment of the relationship between PM 2.5 and Xrs5 could be undertaken without the confounding effects of bronchomotor tone. Importantly, these findings in adults are similar to the findings in children within the Hazelwood Health Study Early Life Follow up (ELF) stream. Shao and colleagues[38] demonstrated that infant or in utero exposures to coal mine fire emissions were associated with long-term impairment of lung reactance, with increased average PM 2.5 being significantly associated with worsening area under the reactance curve - a complementary parameter in the evaluation of reactance.[24] The study has several strengths. Unlike observational studies that have used only secondary data (such as hospitalization) to assess respiratory health, this research has built upon . CC-BY-NC 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 20, 2020. ; https://doi.org/10.1101/2020.10.14.20213009doi: medRxiv preprint 11 previously collected hospitalisation[10] and self-reported symptom data[12] with the inclusion of objective measures of lung mechanics. A further strength of this study was the inclusion of individual estimates of PM 2.5 exposure utilising a combination of detailed time- location diaries and spatially and temporally resolved modelling of PM 2.5 concentrations based upon coal combustion and weather conditions. However, the study also has some limitations. The study endeavoured to account for all relevant potential confounding factors in our analysis, such as age, gender, weight, BMI, education status, tobacco and occupational exposures. However, it is feasible that some of the observed results occurred by chance or were influenced by unknown confounding factors. Furthermore, at this stage in the study we only have cross-sectional data on lung mechanics. Future followup of the Hazelwood Health Study Respiratory Stream participants will better inform an investigation of the long-term implications of medium-duration coal mine fire- related smoke exposure on respiratory mechanics and lung health. In conclusion, a clear dose response association was observed between medium-duration PM 2.5 exposure levels from ambient coal mine fire smoke and a more negative respiratory system reactance in this cohort. This study adds new findings to the literature on the lung health effects of medium term PM 2.5 exposure. These inform public health policy and planning for future coal mine fires or similar medium duration PM 2.5 generating pollution events such as the recent megafires in Australia and the United States. Longitudinal data are required to confirm the findings of this study and to better understand the association of coal mine fire smoke and altered respiratory system reactance and potential accelerated lung aging in exposed populations. Acknowledgments We wish to thank Susan Denny, Kylie Sawyer, Shantelle Allgood and Kristina Thomas from the Monash University School of Rural Health, who assisted with the study. Most of all, the study team would like to acknowledge the contribution of all community members who have participated in the study to date. Financial support The Hazelwood Health Study is funded by the Victorian Department of Health & Human Services (Australia). However, this paper presents the views of the authors and does not represent the views of the Department. . CC-BY-NC 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 20, 2020. ; https://doi.org/10.1101/2020.10.14.20213009doi: medRxiv preprint 12

1. GB D 2016 Ri s k Fac to r C oll abor ato r s . G loba l, r egion al, a nd na tional compa rativ e ris k a s s e s sme nt o f 84 behav ioural, envi r onm en tal a nd oc cup ation al, and metaboli c r i sk s or clust er s of risk s , 1990 –2016: a s y s tem atic ana ly si s for the Global Bur d e n of Disea s e S tudy 201 6. La ncet 2017 ; 390: 1345–4 22. 2. Atkin son RW, Kang S , An der s on HR, e t al. Ep i demiolog ica l t i me s e rie s studi e s of PM2.5 a nd daily mortali ty and ho spital a dmissio ns : a sys t ematic review and me ta -analy s i s. Th or a x 2014 3. L iu C, Che n R, Ser a F , et al . Ambi en t P art i cula t e A i r Po llu tion and Dai ly Mor tali ty in 652 Citie s. Th e New Engla n d jou rnal o f m e dic ine 2 019;381(8 ):7 05-15 . doi: 10 .1056 /N E J M oa 1817364 [publi she d O nli n e Fir st : 2019 /08/23 ] 4. Guo C, Zhang Z, L au AKH, e t al . E ff e ct of l ong- term e xpo sur e t o fine par ticu l ate mat ter on lung func t i o n decl ine and ri sk o f c hr o nic ob s t r uc tive p ulmona ry di sea se i n Taiw an : a l on gitudinal , co hor t s tudy. L a nce t P l ane t He alth 201 8; 2(3): e114- e2 5. d oi: 10. 1016 /S2542 - 5 196(18)300 28 - 7 5. Li S , W il liams G , Jala ludin B , et a l . Pa nel s tudie s o f air pollu tion on c hildren' s lung function a nd re spira tor y s y mptoms: a lite rat ure rev iew. J A s t h m a 2012;49(9 ):895 -910. do i : 10.3 109/0277090 3.2012.7 24129 6. Mel ody S , John ston F . C oal mine fir e s and huma n heal th: W ha t do we k now? I n t e r na t io na l J o ur n a l of Co a l Geol ogy 2015;1 52 doi: 10.1016 /j. c oal.20 15.11.00 1 7. Balme s JR . Whe re T her e's Wild fir e, The r e 's Smoke . N Engl J M ed 20 18;3 78(10 ) :8 81 -83. do i : 10.1 056/N E JMp171 6846 8. Lundbl ad L KA, S iddiqui S, B ossé Y , et a l. Appl ica tion s o f os c illom e t ry in c linic al r e sea r c h a n d prac t i ce . Cana d ia n Journ al of Re s p ira t o ry, C r itic al Care, and S leep Med icine 2019:1-15. do i : 10.1 080/2474533 2.2019.1 649607 9. Aus tr a li an B urea u of S tati s tics (ABS ). Ca nber a, ACT : Da ta by Reg ion - M orwell (SA2 ). 201 8 [upda ted 2 018 Febr ua r y 05; c ited 2 019 Janua ry 21 ]. A v aila ble f rom : ww w.sta t.ab s .gov.a u / it t/ r .j sp?d a tabyreg i on#/ . 10. Guo Y, Ga o CX, D enne kamp M , et al. The a ssoc iati on of c o al m ine fi re smo ke with h os pi ta l emerge ncy pre se nta tion s and admi s sion s : Ti me seri es an al ysi s o f H a z e lwood H e alth Study . C h em os phe r e 2020 ;253:1266 67 . doi: h t t ps: //doi. org/10 .1 016/j.chem o s p her e.2 02 0.1266 67 11. Ikin J, Ca rroll M, Wal ker J, et al. C ohort P ro file: The Haze lwood H e al th S t u dy Adul t Coh ort . Inter na tion a l J ourn al o f E pide mi olo gy 2020;dy aa083 doi: 10. 1093 /ije/dya a 083 12. John so n AL, G ao CX, D enneka mp M, e t a l. A s s oc i ation s b etwe en Re s pi ra tor y Health Out c ome s and Coa l Mine Fir e PM 2.5 S moke E xposur e : A Cro s s-S ec tional Stu dy. I n te r n at i on al J ou r na l of Env iron men tal R es earch a nd P u blic Healt h 2 019;16 (21): 42 62. do i : doi:1 0.3390 /ijerph16 214262 13. T aylor S, Borg B, Gao CX, e t a l . The i mpa ct of th e H a zel wood co al mine fire smoke e xpo sure on as thma. Jour nal o f A s t hma 2020 ;in pr e s s 14. Harr i s PA, T aylor R, T hielk e R, et al. Rese arch elec t ronic d at a c aptur e (RE D Ca p) - - a metad a ta - drive n me t ho d ology an d w orkflow pr oce ss for pr ovidi ng t ran sl ation a l r e s e a r c h i nfor m a tic s suppo rt. J B i om e d Inf o r m 2009;4 2 (2) :377-81. d oi : 10.10 16/ j .jbi.200 8 .08.010 . CC-BY-NC 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 20, 2020. ; https://doi.org/10.1101/2020.10.14.20213009doi: medRxiv preprint 13 15. Co op er B G . A n upda t e on con traindi cation s f or lung fu nc t i on t e s ting . Thorax 201 1;66(8) :714-23 . doi: 10.1136/ thx.2 010.13988 1 [ publi she d O nli ne Fir s t : 2010/07 /31] 16. Wor l d He alth Orga niz a t io n. Guidel i nes for c on t roll ing a nd mon i t o r in g th e t o bacc o ep idemic . Genev a : World Hea lth Organizati on; 1 998. Ava ilable : http s : / /app s . who .int /iri s/handl e/ 10 665 / 42049 . 17. Europe an C ommunity Re spi rat or y Health Survey II SC . The Euro pean Com munity Re spira tor y Health S u r v ey I I. Eur Re s pir J 200 2;20(5 ): 1071 - 9 . doi: 1 0 .1183/09031 936.02 .0 004 6802 18. Q uan je r PH, St anoje vic S, Col e TJ, et al. Mul ti -eth nic re fe renc e va lue s for s pir ometry for th e 3 -95 - yr ag e range : th e g lobal l un g func tion 2 0 12 equati on s. Eur Res pir J 2012 ;40 (6 ):1 324-43. do i : 10.1 183/0903193 6.00080 312 19. L uhar A.K, Emmer son K.M , Rei sen F, Will iamson G . J, Co pe M . E . Modellin g smok e distributi on i n the v icin it y of a large and prolo nged fi r e f r om an ope n-cu t c oal mi ne. Atm os Envi r on . 2020 ;229:117 471. . 20. Rei se n F, Gill e tt R, Choi J, F i s h e r G, T or re P . Ch arac teri s tics of an op en -cut c o al mi ne fire poll utio n ev ent. Atmo sph eric E nvi r onm ent . 20 17; 1 5 1:140-151 . http s : / /doi. org/10 .1016/j.a tmo senv .201 6.12 .015 . 21. Mill e r MR , Hankin son J, Br u s asc o V, e t al . St anda rdi satio n o f s p iro me t ry . Eu r R e s p i r J 2005 ;26(2) :319-38 . doi: 10. 1183 /090319 36.05.00 034805 22. Oo s tve e n E , M ac Leod D , L orino H, et a l . T he forc ed o sci llati on t echniq ue i n c linic al pr a ctic e : methodol ogy, rec ommenda tion s an d f u ture de vel opme nts . Eur Re spir J 2003 ;22( 6):1026-41 . doi: 10.1183/090 31936 .03 .0008940 3 23. S te r ne J A , Whi te IR, Ca r l in J B, e t al . Multiple i mputa tion fo r m issing da ta in epide miologic al a nd cl inica l re sea rch: po te nt i a l and pit fall s . BM J 2009;3 38:b2393. d oi: 10. 1136 /bmj.b 2393 24. O o s tveen E, Bod a K, van der G rin ten CP , e t al . R e spira tory imp e danc e i n heal thy s ub ject s: bas elin e va lue s and broncho dila tor r esp on s e . Eur Re s p ir J 20 13;42(6) :15 13 -23. doi : 10.1 183/0903193 6.00126 212 [p ubli shed O nli ne Fir st: 2013/04 /20] 25. Sood A. In do or fuel expo sur e a nd the lu ng in bo th d ev eloping and d ev el oped coun tri es : a n upda t e . Clin Ch e st Med 2012;3 3 (4) :649-6 5. do i: 10.10 16 / j .ccm .2012.08.003 26. Bal can B , Akan S , Ugurlu AO , et a l. Eff ect s o f bioma s s smoke o n pulmon a r y fun ctio ns : a c a se co nt rol study. I nt J Chr o n Ob struct Pul m on Di s 2016;11:1 61 5-22. doi : 10.2 147/C O P D. S1 09056 27. P op e C A, 3r d, Xu X . Pa s s iv e ci gar ette s mok e, c o al he a t i ng, and r es pira tory symptom s o f nonsmoki ng w omen i n Ch ina . E nv iron Hea lth Per s p ect 1 993;101(4 ):3 14-6. d oi : 10.1 289/ehp .9310 1314 28. Dell ac a RL , Ande rss on Ole rud M, Za nnin E , et al. Lu ng rec ruitmen t a s s e s sed by total r es pira tor y sys t e m inpu t reac tan ce . In te n s i v e Care Med 2 009;35(12 ):2 164 -72. doi : 10.1007 /s00134-0 09 - 1673 -3 [publi sh ed Onlin e Fir s t : 2 009/10 /0 1] 2 9 . M i l n e S , J e t m a l a n i K , C h a p m a n D G , e t a l . R e s p i r a t o r y s y s t e m r e a c t a n c e r e f l e c t s c o m m u n i c a t i n g lung v olume in c hronic obstr u ctive p ul monary dise a se. J our nal o f Applie d P hysiolo gy 2019 ;126(5) :1223 - 3 1. d oi: 1 0.1152 /jappl physiol.005 03 .2018 30. N il sen K , G ov e K, T hien F , et al. Com pa r i s on o f tw o me thod s o f de te rmining lung de - r ec rui tment , using the forced o scill a tion techni qu e . Eur J Appl Phys i ol 2018;118 (10) :22 13-24. doi : 10.1 007/ s 004 21-018-3949 -1 [publi sh ed Online Fir st: 2 018 /08/ 0 1 ] . CC-BY-NC 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 20, 2020. ; https://doi.org/10.1101/2020.10.14.20213009doi: medRxiv preprint 14 31. N il sen K , Thien F, Th a mr in C, et al . Ea rly onse t o f ai r w ay de re crui tment as se s se d usi ng the force d oscill a tion t ec hniqu e i n s u bject s with a sthma. Jo urnal of A pp lie d P hysiol ogy 2019 ;126(5) :1399 - 4 08. d oi: 10.1152 /jap p lphysiol .00 534.2018 3 2 . D e l l a c a R L , D u f f y N , P o m p i l i o P P , e t a l . E x p i r a t o r y f l o w l i m i t a t i o n d e t e c t e d b y f o r c e d o s c i l l a t i o n and ne ga tive e xpir ato ry pre ssur e. Eur Re spir J 200 7;29 (2):36 3-74. d o i : 10.1 183/0903193 6.00038 006 [p ubli shed O nli ne Fir st: 2006/11 /03] 33. D ellac a RL, Rotge r M, Alive rti A , e t al. N o ninva sive de tec tion o f ex pir a to ry flow li mita tion i n CO PD p atie nt s dur ing na sal CP AP. Eu r R es p i r J 2006;27(5 ):9 83-91. do i : 10.1 183/0903193 6.06.000 80005 [pub li she d Onlin e F irst: 2006 / 02/01 ] 3 4 . D e l l a c a R L , S a n t u s P , A l i v e r t i A , e t a l . D e t e c t i o n o f e x p i r a t o r y f l o w l i m i t a t i o n i n C O P D u s i n g t h e force d o scill ati on t e ch nique . Eu r R esp i r J 200 4;23(2 ):232 -40 . [publ i s h e d O nline Fi rst : 2004 / 02/26 ] 35. Kelly VJ, Brown NJ, Ki ng GG, e t a l . The bronchodi la tor re s pon se o f in vi vo spec ific ai rway co mplianc e i n adul ts w ith ast hma . Ann Bio m ed E ng 20 11;39(3 ):112 5 -35. doi : 10.1 007/ s 104 39-010 -0206 -0 36. Herp el L B, Kanne r RE, Le e SM , e t a l . Var iabili t y of spi rome try in ch roni c ob struc t iv e pulmona r y dise a se: re sul t s f rom two clin ical t ri a ls. A m J R es p i r C r it C a r e M e d 20 06 ;173(10 ):1 1 06-13. do i : 10.1 164/r c cm.2 00506 -975 OC [publ i s h e d O nl ine Fi rs t: 2006/02 /25 ] 37. N isa r M, Ea r is JE, P e ar son MG, e t a l. Acute b r onc h odila tor trial s in c hronic ob s t ructiv e pulmona r y dise a se. Am Re v Re spir Di s 1992; 146(3 ):5 55-59. doi : 10. 1164/ajrcc m/ 1 46 .3.555 38. S hao J, Zo s k y GR, H a ll GL, e t a l. Early life ex po sur e t o coal mine fir e smoke e m ission s and a lte r e d lung function in youn g chi ldre n. Re spir ol ogy 2020; 25(2):198-205 . d oi: 10 .1111 /r e s p.1 3617 . CC-BY-NC 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 20, 2020. ; https://doi.org/10.1101/2020.10.14.20213009doi: medRxiv preprint . CC-BY-NC 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 20, 2020. ; https://doi.org/10.1101/2020.10.14.20213009doi: medRxiv preprint . CC-BY-NC 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 20, 2020. ; https://doi.org/10.1101/2020.10.14.20213009doi: medRxiv preprint . CC-BY-NC 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 20, 2020. ; https://doi.org/10.1101/2020.10.14.20213009doi: medRxiv preprint