Age-dependent Effect of Ambient Ozone on Emergency Asthma Hospitalizations in Hong Kong.
Yuanrong Dai,Hong Qiu,Shengzhi Sun,Yang,Hualiang Lin,Linwei Tian
DOI: https://doi.org/10.1016/j.jaci.2018.01.006
2018-01-01
Abstract:There is evidence that short-term air pollution exposure may increase the risk of asthma attacks. However, the association of ambient ozone exposure with the risk of asthma hospitalization has been inconsistent for childhood asthma.1Hajat S. Haines A. Goubet S.A. Atkinson R.W. Anderson H.R. Association of air pollution with daily GP consultations for asthma and other lower respiratory conditions in London.Thorax. 1999; 54: 597-605Crossref PubMed Scopus (148) Google Scholar, 2Silverman R.A. Ito K. Age-related association of fine particles and ozone with severe acute asthma in New York City.J Allergy Clin Immunol. 2010; 125: 367-373Abstract Full Text Full Text PDF PubMed Scopus (142) Google Scholar Ambient ozone may evoke airway inflammation as well as antivirus responses.3Wolcott J.A. Zee Y.C. Osebold J.W. Exposure to ozone reduces influenza disease severity and alters distribution of influenza viral antigens in murine lungs.Appl Environ Microbiol. 1982; 44: 723-731PubMed Google Scholar We hypothesized that the association between ambient ozone exposure and asthma hospitalization is age-dependent because of the differential vulnerability to respiratory virus infections, which are biological triggers of asthma attacks. We conducted a time-series study in Hong Kong to examine the short-term effect of ambient ozone on emergency hospital admissions for asthma in children, adults, and elders, respectively. We collected air pollution data from Hong Kong Environmental Protection Department, emergency hospital admission data from Hospital Authority, and meteorological data from Hong Kong Observatory for the period 2005 to 2014. Daily emergency hospital admissions for asthma (International Classification of Diseases, Ninth Revision code 493) in young children (≤6 years old), children (<18 years old), adults (18-64 years old), and elders (≥65 years old), respectively, were regressed over daily 24-hour mean concentrations of ozone, other air pollutants, and meteorological measurements including temperature and relative humidity. We used generalized additive distributed lag model to examine the association between ozone and asthma.4Zanobetti A. Schwartz J. Mortality displacement in the association of ozone with mortality: an analysis of 48 cities in the United States.Am J Respir Crit Care Med. 2008; 177: 184-189Crossref PubMed Scopus (127) Google Scholar We used the smoothing spline, s(.), with a priori model specifications and degree of freedom (df) to filter out seasonal patterns and long-term trend in asthma hospitalizations. Time-varying covariates controlled for included daily mean temperature, relative humidity, day of week, and official holidays. We examined the distributed lag effects and cumulative effects of ozone within a period of lag0 to lag20, from the current day to previous 20 days. Sensitivity analyses were conducted by applying different ozone exposure metrics (daytime 8-hour mean and maximum) and adjustment for coexisting pollutants. Stratified analyses by age group, sex, and season were performed. The results were expressed in terms of percentage changes (excess risk) in daily emergency asthma hospitalizations per interquartile range increment in O3, and 95% CIs. All analyses were conducted using "mgcv" and "dlnm" packages in the statistical environment R 3.3.3 (R Development Core Team, 2016; http://www.r-project.org). During the study period, there were a total of 62,367 emergency hospital admissions for asthma in Hong Kong, among which 35.1% were children, 32.6% adults, and 32.3% elders. The asthma hospitalization risk associated with ambient ozone varied by age group. An interquartile range increment in O3 (31.6 μg/m3) in a previous week (lag0-6) was associated with a 6.6% (95% CI 2.3%-10.7%) decrease in childhood asthma admissions, but 8.5% (95% CI 4.0%-13.2%) and 13.2% (8.4%-18.2%) increases in asthma for adults and elders, respectively (Table I). The prolonged effects of O3 over lag 7-20 days were also different across age groups. Plots of distributed lag models of O3 effect on asthma along lag 0-20 days (Fig 1) illustrated the apparently protective effect in children and adverse effect in adults and the elderly. The risk estimates for ozone were robust to the adjustment of copollutants. After adjustment for ambient ozone, the 2 pollutants PM2.5 and NO2 are associated with higher risks of asthma hospitalizations in children. Sensitivity analyses with different metrics of ozone did not change the risk estimates significantly; the associations did not vary by sex or season; results with the young children group (≤6 years old) were similar to those with children younger than 18 years (see this article's Online Repository at www.jacionline.org). The current study disclosed age-dependent patterns of ozone and asthma relationships in an extended time window of lag0-20 days, which may help explain inconsistent findings of ozone and asthma in the earlier literature.Table IAge-stratified percent changes and 95% CIs in asthma emergency hospitalizations with an IQR increase in ozone in different time windows of the lag 0-20 period, based on 2005-2014 data in Hong Kong∗Interquartile range (IQR) for ozone is 31.6 μg/m3, using 24-h mean concentration as the metric. Risk estimates were derived from Poisson generalized additive distributed lag models, constrained with a second-degree (quadratic) polynomial, while adjusting for time trend and seasonality, weather conditions, day of week, and public holidays.Distributed lagsAsthma in young children (age ≤ 6 y)Asthma in children (age < 18 y)Asthma in adults (age 18-64 y)Asthma in elders (age ≥ 65 y)Lags 0-6−8.6 (−13.4 to −3.4)−6.6 (−10.7 to −2.3)8.5 (4.0 to 13.2)13.2 (8.4 to 18.2) Lags 0-2−2.8 (−0.9 to 1.7)−2.0 (−4.8 to 0.9)5.3 (2.5 to 8.2)7.6 (4.7 to 10.6) Lags 3-6−6.0 (−8.2 to −3.7)−4.7 (−6.6 to −2.8)3.1 (1.2 to 5.0)5.1 (3.2 to 7.1)Lags 7-20−21.1 (−26.3 to −15.5)−18.7 (−23.3 to −14.0)−7.0 (−11.7 to −2.0)−3.0 (−8.0 to 2.3)Lags 0-20−27.9 (−34.3 to −20.8)−24.1 (−29.8 to −18.0)0.9 (−6.1 to 8.4)9.7 (2.0 to 18.1)Statistically significant risk estimates are in boldface.∗ Interquartile range (IQR) for ozone is 31.6 μg/m3, using 24-h mean concentration as the metric. Risk estimates were derived from Poisson generalized additive distributed lag models, constrained with a second-degree (quadratic) polynomial, while adjusting for time trend and seasonality, weather conditions, day of week, and public holidays. Open table in a new tab Statistically significant risk estimates are in boldface. As a highly reactive gaseous pollutant, ozone exerts both inflammatory and antiviral effects in the respiratory system. The oxygen radicals produced from this exogenous oxidant evoke oxidative stress and airway inflammation, which would directly cause asthma exacerbation. Short-term exposure to ozone may also exacerbate asthma by increasing bronchial allergen responsiveness. On the other hand, ozone was shown to be protective against influenza by reducing influenza disease severity and mitigating acute virus-induced lung injury,3Wolcott J.A. Zee Y.C. Osebold J.W. Exposure to ozone reduces influenza disease severity and alters distribution of influenza viral antigens in murine lungs.Appl Environ Microbiol. 1982; 44: 723-731PubMed Google Scholar which may relate to the ozone-primed immunity via the alarmin IL-335Mathews J.A. Krishnamoorthy N. Kasahara D.I. Cho Y. Wurmbrand A.P. Ribeiro L. et al.IL-33 drives augmented responses to ozone in obese mice.Environ Health Perspect. 2017; 125: 246-253Crossref PubMed Scopus (45) Google Scholar, 6Yang Q. Ge M.Q. Kokalari B. Redai I.G. Kemeny D.M. Bhandoola A. et al.Group 2 innate lymphoid cells mediate ozone-induced airway inflammation and hyperresponsiveness in mice.J Allergy Clin Immunol. 2017; 137: 571-578Abstract Full Text Full Text PDF Scopus (77) Google Scholar against influenza virus infection.7Monticelli L.A. Artis D. Innate lymphoid cells promote lung tissue homeostasis following acute influenza virus infection.Nat Immunol. 2012; 12: 1045-1054Crossref Scopus (1110) Google Scholar The inflammatory and antivirus effects of ozone may exert opposite effects on asthma hospitalizations, explaining the differential effects of ozone on asthma in different subpopulations. Children are more vulnerable to respiratory infections because of their relatively weak immune system and more social gathering on school days. There are 80% to 85% asthma exacerbations in school-age children in the community associated with viral infection.8Johnston S.L. Pattemore P. Sanderson G. Smith S. Campbell M. Josephs L. et al.The relationship between upper respiratory infections and hospital admissions for asthma: a time trend analysis.Am J Respir Crit Care Med. 1996; 154: 654-660Crossref PubMed Scopus (497) Google Scholar The inverse correlation between ozone and childhood asthma alludes to the antivirus effect of ozone and viral infections as a mediator for asthma hospitalization. The contribution of viral infection to asthma hospitalization may be larger in children than in adults and the elderly. In the children group, the protective effect of ozone was not observed in the immediate lag 0-2 days (Table I and Fig 1), which may suggest that ozone's inflammatory effect was more immediate whereas its antivirus effects was more delayed. There is a possibility that the real modifier of the ozone and asthma relationship is asthma phenotype instead of age itself because of the association between clinical phenotypes and age.9Wenzel S.E. Asthma phenotypes: the evolution from clinical to molecular approaches.Nat Med. 2012; 18: 716-725Crossref PubMed Scopus (1628) Google Scholar Furthermore, this study was conducted in Hong Kong, an Asian city with high density of population and higher pollution levels compared with Western countries. Replication of the current study is warranted in other geographical locations. In conclusion, the reduced risk of asthma hospitalization in children stands in apparent contrast with the elevated risk in adults and the elderly associated with ambient ozone. Age or age-related asthma phenotype may be a modifier of the relationship between ozone and asthma. The differential effects of ozone on asthma reflect the complex roles of ozone in inducing both inflammatory and antiviral responses in different subpopulations. Fig E2Plots of distributed lag estimates of age- and season-stratified percent changes and 95% CIs in asthma emergency hospitalizations with an IQR increase in ozone along the lag 0-20 days, based on 2005-2014 data in Hong Kong. The general patterns are similar between seasons. IQR, Interquartile range.View Large Image Figure ViewerDownload Hi-res image Download (PPT)Fig E3Plots of distributed lag estimates of age- and sex-stratified percent changes and 95% CIs in asthma emergency hospitalizations with an IQR increase in ozone along the lag 0-20 days, based on 2005-2014 data in Hong Kong. The associations did not vary by sex. IQR, Interquartile range.View Large Image Figure ViewerDownload Hi-res image Download (PPT)Fig E4Plots of distributed lag estimates of age-stratified percent changes and 95% CIs in asthma emergency hospitalizations with an IQR increase in air pollutant levels along the lag 0-20 days, derived from the single- and 2-pollutant regression models of 2005-2014 data in Hong Kong. The results with young children 6 years and younger are similar to those with children younger than 18 years. IQR, Interquartile range.View Large Image Figure ViewerDownload Hi-res image Download (PPT)Table E1Distribution of emergency hospital admissions for asthma, air pollution concentrations, and meteorological factors in Hong Kong, 2005-2014VariableMean ± SDMin25th50th75thMaxEmergency asthma admissions Age (y)17.1 ± 5.8313172143Young child (≤6 y)3.9 ± 2.7023522Child (age < 18 y)6.0 ± 3.7035829Adult (age 18-64 y)5.6 ± 2.8045719Elder (age ≥65 y)5.5 ± 2.7045718 SexFemale9.0 ± 3.70691122Male8.1 ± 3.60681026 SeasonWarm15.5 ± 5.4312151940Cool18.7 ± 5.8414182243Air pollution concentrations (μg/m3) O3 (24-h mean)36.7 ± 21.74.718.532.150.1134.4 O3 (8-h mean)47.7 ± 30.05.224.040.565.8194.3 PM2.5 (24-h mean)34.5 ± 21.04.917.730.046.4174.5 NO2 (24-h mean)55.7 ± 19.21341.752.966.7150.5Weather conditions Mean temperature (°C)23.4 ± 5.28.419.124.627.931.8 Relative humidity (%)78.4 ± 10.629.073.079.086.099.025th, 25th percentiles; Max., maximum; Min., minimum. Open table in a new tab 25th, 25th percentiles; Max., maximum; Min., minimum.