Published on

Desgarradoras experiencias cercanas a Termoeléctricas a Carbón a nivel mundial.

Published in: Health & Medicine
  • Be the first to comment

  • Be the first to like this

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide


  1. 1. ARTICLE IN PRESS Health & Place 16 (2010) 399–408 Contents lists available at ScienceDirect Health & Place journal homepage: www.elsevier.com/locate/healthplace Who is affected more by air pollution—Sick or healthy? Some evidence from a health survey of schoolchildren living in the vicinity of a coal-fired power plant in Northern Israel Tamar Yogev-Baggio a, Haim Bibi b, Jonathan Dubnov c,d, Keren Or-Hen e, Rafael Carel d, Boris A. Portnov a,n a Department of Natural Resources & Environmental Management, University of Haifa, Israel b Carmel Medical Center, Haifa, Israel c Haifa District Health Office, Ministry of Health, Israel d School of Public Health, University of Haifa, Israel e Graduate School of Education, University of Haifa, Israel a r t i c l e in fo abstract Article history: Objective: To evaluate the effects of exposure to air pollution by NOx and SO2 on the development of Received 17 July 2009 pulmonary function of children, characterized by different health status. Received in revised form Methods: A cohort of 1181 schoolchildren from the 2nd to 5th grades, residing near a major coal-fired 18 November 2009 power plant in the Hadera district of Israel, were subdivided into three health status groups, according Accepted 22 November 2009 to the diagnosis given by a physician at the beginning of the study period in 1996: (a) healthy children; (b) children experiencing chest symptoms, and (c) children with asthma or spastic bronchitis. Keywords: Pulmonary Function Tests (PFTs) were performed twice (in 1996 and 1999) and analyzed in conjunction Children with air pollution estimates at the children’s places of residence and several potential confounders Air pollution —height, age, gender, parental education, passive smoking, housing density, length of residence in the Pulmonary function test study area and proximity to the main road. Results: A significant negative association was found between changes in PFT results and individual exposure estimates to air pollution, controlled for socio-demographic characteristics of children and their living conditions. A sensitivity analysis revealed a decrease in the Forced Expiratory Volume during the First Second (FEV1) of about 19.6% for children with chest symptoms, 11.8% for healthy children, and approximately 7.9% for children diagnosed with asthma. Results of a sensitivity test for the Forced Vital Capacity (FVC) were found to be similar. Conclusion: Exposure to air pollution appeared to have had the greatest effect on children with chest symptoms. This phenomenon may be explained by the fact that this untreated symptomatic group might experience the most severe insult on their respiratory system as a result of exposure to ambient air pollution, which is reflected by a considerable reduction in their FEV1 and FVC. Since asthmatic children have lower baseline and slower growth rates, their PFT change may be affected less by exposure to air pollution, reflecting a well known relationship between pulmonary function change and height growth, according to which age-specific height is very similar for preadolescent children, but shifts upward with age during the growth spurt. & 2009 Elsevier Ltd. All rights reserved. 1. Introduction 2007; Schwartz, 2004). Several studies indicated that gaseous pollutants and particulate matters tended to increase the risk of According to several epidemiological studies, children are chronic chest symptoms and might have lifelong effects on more susceptible to air pollution than adults due to their children’s health (Bateson and Schwartz, 2008; Gauderman et al., increased respiratory rate and immature host defense mechan- 2004; Kajekar, 2007; Schwartz, 2004). Genetic factors, which isms, which lead to the increased absorption of air pollutants regulate antioxidant defenses, may act together with environ- (Bateson and Schwartz, 2008; Gauderman et al., 2000, 2002, 2004, mental factors further elevating a potential risk of respiratory diseases in children (Gilliland et al., 1999, 2002; Kajekar, 2007; Sly and Flack, 2008; Romieu et al., 2002; Schwartz, 2009). n Corresponding author. Fax: + 972 4 8249971. However, the question whether sensitivity to air pollution E-mail address: Portnov@nrem.haifa.ac.il (B.A. Portnov). differs among children characterized by different health status 1353-8292/$ - see front matter & 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.healthplace.2009.11.013
  2. 2. ARTICLE IN PRESS 400 T. Yogev-Baggio et al. / Health & Place 16 (2010) 399–408 (e.g., healthy children vs. asthmatic children) remains largely unanswered. Thus, for instance, Gauderman et al. (2004) did not find any significant effect of air pollution on asthmatic children, concluding that it may be influenced by a small sample size. Only a handful of studies looked into these differences, indicating that the long-term effects of air pollutants on pulmonary function (PF) growth may depend on children’s initial health status (see inter alia, Avol et al., 2001; Goren and Hellmann, 1997; Jedrychowski et al., 2002). Several earlier studies demonstrated that the effects of air pollution on children’s PFTs differed between healthy children and asthmatics. In particular, such differences were detected for NOx, SO2, O3 and PM exposures (see inter alia, Gauderman et al., 2004; Gilliland, 2009). Healthy preadolescents residing in neigh- borhoods with elevated levels of SO2 and suspended particulate matters (SPM) were found to exhibit a delay in their pulmonary function growth (Jedrychowski et al., 2002), while pollutant initiated inflammation was found to worsen the pulmonary function (PF) performance among asthmatic children (Lee et al., 2002; Pikhart et al., 2000; Raizenne et al., 1996). Exposure to air pollution was also found to increase the number of asthmatic children residing in urban areas (Delfino, 2002; Islam et al., 2007; Jerrett et al., 2008), thus supporting a causality relationship between air pollution and asthma. A major uncertainty of epidemiological studies is often associated with the estimation of individual exposure levels. A commonly used approach assumes that each person in a concerned region has the same exposure level, since the pollution often originates from the same pollution source (Elliott and Watenberg, 2000, 2004; Nuckols et al., 2004; Thurston et al., 2009). Exposure level assigned to each individual are often Fig. 1. Map of the study area. obtained from a few air quality monitors and reflect the mean concentrations in the entire urban area or community (Nuckols 2. Methods et al., 2004; Portnov et al., 2007). This approach may under- estimate the individual exposure to air pollutants in the most 2.1. Study area polluted areas and overestimate it in the least polluted neighbor- hoods, thus leading to a measurement error known as exposure misclassification (Elliott and Savitz, 2008; Nuckols et al., 2004; The Hadera district of Israel, which forms our study area, is Portnov et al., 2007). Furthermore, the use of mean air pollution situated on the Mediterranean shore, some 50 km north-east of levels ‘‘smoothes’’ air pollution ‘‘spikes’’ (or so-called ‘‘air Tel Aviv. Its total population amounts to approx. 350,000 pollution events’’), and may thus result in unreliable estimates residents (ICBS (Central Bureau of Statistics, Jerusalem, Israel), of health effects, which may be especially harmful for children. 2008). The study area hosts two townships (Hadera and Pardes- Several epidemiological surveys carried out in Israel compared Hana-Karkur), two small suburban neighborhoods (Beit Eliezer the prevalence of respiratory symptoms, related to asthma and and Givat Olga), and a major coal-fired power plant with bronchitis, among subgroups of children residing in highly production capacity of about 2590 MW (see Fig. 1). Sources of polluted areas with the prevalence of these symptoms among ambient air pollution in the area originate mainly from the power children residing in less polluted areas. Most of these studies plant’s smokestacks, as well as from several small industrial detected the attenuation of PF development and an increase in the plants scattered throughout the study area, and from motor prevalence of respiratory morbidity in polluted areas (Goren et al., vehicles. 1990, 1991; Goren and Hellmann, 1997; Peled et al., 2005). However, these studies were rather inconclusive in regard to the 2.2. Air pollution estimates differential effect of air pollution on children characterized by different health status. Air quality data for our study were obtained from the network Following our previous studies (Dubnov et al., 2007; Portnov of 12 air quality monitoring stations (AQMSs), maintained by the et al., 2007), in the present analysis, we attempt to determine Hadera Association of Towns for the Environmental Protection whether the previously detected effect of air pollution on children’s (ATEP) and providing continuous measurements of air pollution PF development vary across subgroups of children characterized by by SO2 and NOx. In the present study, we adopted the ‘‘event’’ different health status, i.e., healthy children, asthmatics, and children approach, used in several previous studies and found to be an with chest symptoms. effective and sensitive measure for local air pollution levels The study is based on a cohort of 1181 schoolchildren residing (Dubnov et al., 2007; Portnov et al., 2007). According to this in the Hadera district of Israel (Fig. 1), and subdivided into three approach, half-an-hour concentrations of NOx and SO2 that health status groups—(a) healthy children (children without simultaneously exceeded some predefined air pollution levels known respiratory diseases and symptoms); (b) children who (i.e., 0.125 ppm for NOx and 0.070 ppm for SO2) were multiplied experienced chest symptoms during the year preceding the by the average concentrations during the ‘‘air pollution events’’ pulmonary function test; and (c) children diagnosed by a and then summed up over the entire study period, 1996–1999 physician as suffering from either asthma or spastic bronchitis. (for more detail on this approach, see Dubnov et al., 2007). The
  3. 3. ARTICLE IN PRESS T. Yogev-Baggio et al. / Health & Place 16 (2010) 399–408 401 aforementioned concentration levels are quarters of the local air 2007) being similar to those used elsewhere (Leonardi et al., pollution standards for NOx and SO2, respectively, and are 2002). In particular, the following three ‘‘health status’’ categories recommended by ATEP for air pollution event analysis (ATEP (subgroups) were distinguished: (Association of Towns for Environmental Protection), 2005; Goren et al., 1995). 1. Healthy subgroup: The child was never diagnosed by a In order to insure a time lag between the air pollution data and physician with either spastic bronchitis or asthma and had PFT results, air pollution data were taken for calendar years (i.e., not had any chest symptoms during the last 12 months prior to January 1st through December 31st), while PF tests were carried PFT. out a few months later after the start of a calendar year, i.e., 2. Chest symptoms subgroup: The child had at least one of the between April and June (see the subsection on Study Population). following symptoms during the 12 months preceding the PFT Over the entire study period (1996–1999), air pollution events test: wheezing with or without upper respiratory tract occurred approximately 0.3% of the time in the least polluted part infection (URTI); coughing with or without URTI; coughing of the study area and about 1.1% of the time in its most polluted with phlegm or wheezing with dyspnea. part. Whereas the overall annual numbers of ‘‘air pollution 3. Asthma/spastic bronchitis subgroup: The child was diagnosed by events’’ appeared to be stable in 1996–1999, the integrated a physician with either asthma or spastic bronchitis and concentration values of the air pollutants exhibited a general exhibited at least one of the above respiratory symptoms, as tendency to rise (see Appendix 1). idem in (2). The NOx  SO2 ‘‘event’’ concentrations, observed at the location of AQMSs were then interpolated by the kriging interpolation In accordance with the ‘‘intent-to-treat principle’’ for a fixed method (Beers and Kleijnen, 2003; Dubnov et al., 2007; Kim et al., cohort analysis (Rothman et al., 2008), the classification was 2009), which enabled us to obtain a continuous surface of based on the results of the 1996 survey, that is, on the health NOx  SO2 air pollution estimates covering the entire study area status of the children at the beginning of the study period. Only (see Fig. 1). the children with a known health status in 1996, who did not To illustrate differences in the children’s PF performances change their home address during the entire length of the study across communities with different levels of air pollution, we also period (that is, 1181 children; see Section 2.3), were included in estimated the average NOx  SO2 ‘‘event’’ concentrations for 20 the analysis. Small Census Areas (SCAs), into which the study area is divided, and then grouped these SCAs into three air pollution zones, 2.5. Individual exposure estimates with distinctively different air pollution levels: low pollution (NOx  SO2 o312 ppm); medium pollution (312oNOx  SO2 o2640 ppm), and high pollution (NOx  SO2 42640 ppm) (see In the present study, we used an indirect approach to Fig. 2). The classification was performed using the Jenks’ ‘‘natural estimating the individual air pollution exposure (Dubnov et al., breaks’’ method in the ArcGIS9.xTM software. This method 2007; Jerrett et al., 2008; Kim et al., 2009). According to this determines the best arrangement of values into classes by approach, the homes of the children participating in the survey comparing the sum of squared differences of values from the were positioned on a map and superimposed upon NOx  SO2 air means of their classes and thus identifies ‘‘natural break points’’ in pollution levels at corresponding locations (see Section 2.2 and the data distribution by picking the class breaks that best group Fig. 1). Next, the levels of NOx and SO2 air pollution, to which each similar values and maximize differences between the classes child was exposed, were calculated. The task was performed in the (Minami, 2000). ArcGIS 9TM software, using its ‘‘spatial join’’ tool, which makes it possible to match between different geographic layers (maps) based the spatial relationship between them (Minami, 2000). 2.3. Study population 2.6. Statistical analysis The health, demographic and PFT data used in the present analysis are described in detail in Dubnov et al. (2007). In brief, The analysis was carried out in four phases. During the first the same cohort of 1492 schoolchildren performed pulmonary phase, changes in the proportional shares of different health status function tests (FVC and FEV1) in 1996 and, again, in 1999. In total groups in the total population between 1996 and 1999 was 1181 out of 1492 children resided at the same address and had examined separately for the sub-regions with low, medium and information about their health status at the beginning of the high air pollution (see Section 3.1 and Fig. 2). During the second study period. This group of 1181 children formed the study cohort phase of the analysis, the interaction between the ‘‘health status’’ covered by the present analysis. Changes in the results of groups (see Section 3.1) and air pollution levels (Section 3.2) was pulmonary functions tests (DFVC and DFEV1) between the years investigated by the analysis of variance (ANOVA), which was used were calculated as the percent difference between the observed to determine whether the mean DPFT values, observed in and expected lung function volumes (Hankinson et al., 1999). For different health status groups of children, differ by air pollution each child in the sample, health and socioeconomic data were zones. During the third phase of the analysis, the multiple obtained from individual questionnaires (Dubnov et al., 2007; regression analysis was run to indentify and measure the effect Feris, 1978), filled out by the children’s parents. The question- of air pollution on 1996–1999 changes in the children’s PF test naires included questions on housing density; exposure to (both DFEV1 and DFVC), adjusted for several potential confoun- tobacco smoking; parental education and duration of living in ders, such as the duration of residence in the study area, passive the study area. The same questionnaire was used in both 1996 smoking in the family, housing density, father’s education, and and 1999. proximity to the main road. The analysis was performed separately for each health status subgroup covered by the study 2.4. Health status classification (see Section 3.3). During the analysis, the normality and the heterogeneity of variance assumptions were verified and the The classification of children into health status subgroups was results were found to be satisfactory. Lastly, during the fourth based on a health questionnaire described above (Dubnov et al., phase of the analysis, an estimation test was run to determine the
  4. 4. ARTICLE IN PRESS 402 T. Yogev-Baggio et al. / Health & Place 16 (2010) 399–408 Fig. 2. Subdivision of the study area into air pollution regions. magnitude of changes in the children’s DFEV1 and DFVC, with known health status in both 1996 and 1999.] As the table attributed to plausible changes in ambient air pollution levels. shows, the percent of children, diagnosed as healthy, decreased The general characteristics of the entire cohort of school- between 1996 and 1999 in all air pollution regions, but most children surveyed are reported in Table 1, while Table 2 reports substantially in the ‘‘high air pollution’’ zone. For example, the the characteristics of children in different health status groups. number of healthy children in the least polluted (‘‘low air pollution’’) zone dropped from 46.7% children diagnosed as healthy in 1996 to 40.4% such children in 1999. Concurrently, 3. Results the corresponding decrease in the number of healthy children in the ‘‘high air pollution’’ zone was much higher, from 45.9% of 3.1. Changes in the health status groups’ sizes children diagnosed as healthy in 1996 to 34.5% of such children in 1999. Table 3 illustrates changes in the numbers of children in each The opposite trend is found in the ‘‘chest symptoms’’ group. health status group in different air pollution regions, which The number of children in this health status group increased in all occurred between 1996 and 1999. [The table reports only children air pollution zones, but, again, most substantially, in the ‘‘high air
  5. 5. ARTICLE IN PRESS T. Yogev-Baggio et al. / Health & Place 16 (2010) 399–408 403 pollution’’ region. In particular, the number of children exhibiting Characteristically, the direction of changes appear to be less chest symptoms in this air pollution zone increased between consistent in respect to the ‘‘pulmonary diseases’’ subgroup, 1996 and 1999 by nearly twofold, from 20 children diagnosed which may be explained by a relatively small numbers of children with chest symptoms in 1996 to 37 such children in 1999, in this group overall, which is generally insufficient for general- or by 85%. ization. Table 1 General characteristics of the schoolchildren surveyed in 1996 and 1999. Variable 1996 1999 Mean S.D. Mean S.D. Total number of children surveyed 1181 1181 Observed FVC (l) 2.0 0.5 2.9 0.7 Predicted FVC (%) 94.3 12.4 93.9 11.0 Observed FEV1 (l) 1.8 0.4 2.7 0.6 Predicted FEV1 (%) 100.7 13.7 97.6 11.6 Female gender (%) 54 56 Age (years) 9.2 1.5 12.2 1.5 Height (cm) 136.3 10.6 153.2 10.6 Father’s education (years of schooling) 12.7 2.7 12.9 2.8 Passive smoking in the family %a 53 48 Duration of residence in the study area (years) 8.0 2.7 11.0 2.7 Proximity to the main road (%)b 36.0 Housing density (persons per room) 1.3 0.7 1.2 0.5 a At least one person smokes in the child’s home. b Living within 50 m from a main road (the first row of buildings). Table 2 Characteristics of children in different health status groups (according to the 1996 survey). Variable Healthy (N = 537) Chest symptoms (N = 357) Pulmonary diseases (N= 287) Prob. 1996 1996 1996 Mean S.D. Mean S.D. Mean S.D. Age (years) 9.3 1.6 9.1 1.5 9.2 1.5 0.310 Father’s education (years of schooling) 12.8 2.6 12.5 2.9 12.8 2.7 0.285 Female gender (%) 56.4 55.7 47.7 0.044 Height (cm) 136.6 10.7 135.9 10.8 136.2 10.3 0.598 Housing density (person/room) 1.3 0.7 1.4 0.6 1.3 0.7 0.400 Length of residence in the study area (years) 8.4 2.5 7.7 2.8 7.8 2.8 o0.001 Observed FEV1 (l) 1.9 0.4 1.8 0.4 1.8 0.4 0.078 Observed FVC (l) 2.0 0.5 2.0 0.5 2.0 0.5 0.094 Passive smoking in the family %a 47.1 56.7 58.5 0.002 Predicted FEV1 (%) 101.6 13.8 100.1 13.6 99.6 13.6 0.092 Predicted FVC (%) 94.9 12.0 93.5 12.9 94.5 12.4 0.249 Proximity to the main road (%)b 39.1 32.5 33.1 0.076 a At least one person smokes in the child’s home. b Percent of children living within 50-m from a main road. Table 3 Changes in the number of children in different health status groups by air pollution regions in 1996–1999.a Air pollution region Health group Number of children in 1996 Number of children in 1999 % Change Prob. Low Healthy 268 232 13.4 o0.001 Chest symptoms 170 208 + 22.4 Pulmonary diseases 136 134 –1.5 Total 574 574 Medium Healthy 97 74 –23.7 o0.001 Chest symptoms 77 84 + 9.1 Pulmonary diseases 40 56 + 40.0 Total 214 214 High Healthy 40 30 –25.0 o0.01 Chest symptoms 20 37 + 85.0 Pulmonary diseases 27 20 –25.9 Total 87 87 a Includes only children with reported health status in both 1996 and 1999 (875 children out of 1181 children in the sample).
  6. 6. ARTICLE IN PRESS 404 T. Yogev-Baggio et al. / Health & Place 16 (2010) 399–408 Table 4 Interaction of health status groups with air pollution levels (method: analysis of variance (ANOVA); dependent variables—DFEV1)1. Health status group Air pollution zone Fa(df.) Low Medium High Mean (S.D.) N Mean (S.D.) N Mean (S.D.) N Healthy À 2.12 (12.77) 360 À 4.44 (13.22) 136 À 11.47 (11.98) 40 10.16 (2,535) P o0.001 Chest symptoms À 2.75 (12.95) 230 À 1.40 (11.18) 104 À 16.07 (17.59) 20 11.33 (2,356) P o0.001 Pulmonary diseases À 1.82 (11.96) 194 À 3.23 (12.87) 66 À 5.92 (11.84) 27 1.48 (2,286) P =0.229 Note: Total number of children—1181. a F-statistic and its degree of freedom (df), post-hoc differences. Table 5 Factors affecting the pulmonary function change by children subgroups (dependent variable—DFEV1; method—multiple regression analysis). Variable Model 1 (all children) Model 2 (healthy) Ba Sig.b 95% CIc Ba Sig.b 95% CIc Lower bound Upper bound Lower bound Upper bound Age 2.564 o0.001 1.889 3.240 2.728 o0.001 1.601 3.855 Father’s education 0.139 0.221 À 0.084 0.361 0.393 0.030 0.038 0.749 Gender À 0.084 0.889 À 1.265 1.097 0.325 0.725 À 1.489 2.140 Height 0.396 o0.001 0.302 0.490 0.341 o0.001 0.187 0.495 Housing density 0.264 0.594 À 0.705 1.233 0.199 0.800 À 1.344 1.741 NOx  SO2 À 3.81E À 03 o0.001 À 0.005 À 0.003 À 3.02E À 03 o0.001 À 0.005 À 0.001 Passive smoking 1.098 0.072 À 0.097 2.293 1.868 0.044 0.053 3.683 Road proximity À 1.062 0.092 À 2.298 0.175 À 0.147 0.877 À 2.007 1.713 Duration of residence 0.124 0.349 À 0.135 0.383 0.268 0.252 À 0.191 0.728 (Constant) À 82.449 o0.001 À 92.102 À 72.796 À 82.952 o0.001 À 98.307 À 67.596 R2 0.41 0.40 No. of children 1181 537 Variable Model 3 (chest symptoms) Model 4 (pulmonary diseases) Ba Sig.b 95% CIc Ba Sig.b 95% CIc Lower bound Upper bound Lower bound Upper bound Age 1.996 0.001 0.813 3.179 2.808 o 0.001 1.597 4.019 Father’s education À 0.168 0.394 À 0.557 0.220 0.146 0.500 À 0.279 0.570 Gender 0.053 0.963 À 2.163 2.268 À 0.228 0.843 À 2.491 2.034 Height 0.515 o 0.001 0.352 0.679 0.366 o 0.001 0.189 0.542 Housing density À 0.608 0.540 À 2.557 1.341 0.966 0.242 À 0.656 2.588 NOx  SO2 À 7.58E–03 o 0.001 À 0.010 À 0.005 À 2.04E À 03 0.069 À 0.004 0.000 Passive smoking À 0.864 0.448 À 3.102 1.374 1.776 0.135 À 0.558 4.109 Road proximity À 1.908 0.115 À 4.282 0.466 À 1.111 0.364 À 3.516 1.294 Duration of residence À 0.198 0.399 À 0.659 0.263 0.281 0.205 À 0.154 0.716 (Constant) À 83.299 o 0.001 À 100.841 À 65.758 À 83.032 o 0.001 À 101.156 À 64.908 R2 0.45 0.44 No. of children 357 287 a Unstandardized regression coefficient. b Two-tailed significance level. c 95% confidence interval. 3.2. Air pollution—health status groups’ interaction 3.3. Multivariate regression analysis Table 4 reports the results of the ANOVA analysis and Table 5 reports the results of the multivariate regression illustrates the significance of differences in the average DFEV1 analysis of the factors affecting the FEV1 change across different values observed in different health status groups and in different health status groups of children between 1996 and 1999 (The air pollution zones. (The results of the analysis for DFVC are results for the FVC change were found to be similar and can be essentially similar and are not reported here for brevity’s sake.) obtained from the authors upon request.) In the analysis, As Table 4 shows, the effect of ‘‘air pollution levels—the individual air pollution estimates (see Section 2.5) were children’s health status’’ interaction on DPFT appears to be controlled for several potential confounders: age, height, gender, statistically significant (Roy’s=2.37, Po0.05), thus indicating that duration of residence in the study area, passive smoking in the the effect of air pollution on a child’s PFT change does appear to family, housing density, father’s education, and proximity to the depend on the child’s health status at the beginning of the study closest main road (see Appendix 1). The models appear to provide period (that is, in 1996). a reasonably good fit (R2 =0.40–0.45).
  7. 7. ARTICLE IN PRESS T. Yogev-Baggio et al. / Health & Place 16 (2010) 399–408 405 Table 6 Changes in the DFEV1 values as a response to plausible changes in NOx  SO2, air pollution levels. Air pollution levels (ppm) Healthy Chest symptoms Pulmonary diseases Low pollution (NOx  SO2 = 106) À 4.005 0.050 À 2.578 Medium pollution NOx  SO2 =1507) À 8.235 À 10.566 À 5.437 High pollution (NOx  SO2 = 2700) À 11.837 À 19.606 À 7.872 Note: Based on the models in Table 5; input constants are set to the average values as follows: age= 9.2 (average for the sample); proximity to the main road = 0 ( 450 m); passive smoke= 0 (no); gender= male (1); duration of living in the study area= 8 years (average for the sample); housing density= 1.3 persons per room (average for the sample); father’s education =12.71 years (average for the sample); height = 136.28 cm (average for the sample). As Table 5 shows, DFEV1 is significantly and positively using the regression models reported in Table 5 and holding all correlated with the age and height of a child (Po0.01). The effect other variables (accept for air pollution) constant. of the NOx  SO2 interaction term on the children’s PFT perfor- As Table 6 shows, the largest DFEV1 deficit associated with mance appears to be negative and highly significant in most of the living in the high air pollution area is observed in the ‘‘chest models (Po0.01), implying that increasing air pollution levels symptoms’’ subgroup ( À19.6%), followed by healthy children appear to have, ceteris paribus, a significant and negative effect on ( À 11.8%) and children with chronic respiratory diseases ( À 7.9%), the children’s PF growth. Characteristically, the strongest thus suggesting that healthy children and children with chest ‘‘NOx  SO2—DFEV1’’ association is detected in the ‘‘chest symp- symptoms appear to be most susceptible to chronic respiratory toms’’ subgroup and among the healthy (Po0.001; see Table 5; effects associated with their exposure to air pollution. Models 2–3). However, this association is much weaker for the Possible explanations of this trend may be attributed to differences ‘‘PF diseases’’ subgroup (P 40.05; see Table 5; Model 4), implying in PFT change among children in different health status groups. The that PF growth in children with chronic pulmonary diseases relation between pulmonary function and height is well known: the appears to be least affected by air pollution, as compared with age-specific height is very similar for the preadolescent children, but other groups under study. A possible explanation for this shifts upward with age during the growth spurt (Wang et al., 1993). phenomenon will be suggested in the following section. Some studies have shown that children with asthma exhibited impaired growth (Balfour-Lynn, 1986; Littlewood et al., 1988; Murray et al., 1976; Wales et al., 1991; Wolthers and Pedersen, 1991). This 4. Discussion slow growth may lessen the effect of air pollution on PF changes and differences in measurements between ‘‘initial performance values’’ Since individuals differ in their metabolic rate and immune (FEV1_1996) and ‘‘final performance values’’ (FEV1_1999). As Wang system functioning, some are likely to be more susceptible to et al. (1993) reported, FEV1 of asthmatic children was, on the average, ambient air pollution than others (Arshad et al., 2005; D’Amato 1.5–3.0% lower than that of children who had no asthma or wheezing. et al., 2005; Kim, 2004; Paden, 2005). In the present study, a In particular, asthmatic children tend to exhibit airway hyper- cohort of 1181 schoolchildren divided into three subgroups based responsiveness and obstruction, which is reflected by low values on their health status was tested twice for their PFT performance of polluted air they inhale (GINA (Global Strategy for Asthma in 1996 and again in 1999. Management and Prevention), 2008; Sears et al., 2003). As the analysis revealed, the number of healthy children in the Thus, Xuan et al. (2000) carried an 8-year long survey in which study cohort decreased over the study period overall, but most five follow-up studies were conducted at 2-year intervals. Height substantially in the most air polluted areas. In particular, the has been established as the best predictor of lung function number of healthy children in the ‘‘low pollution’’ zone dropped growth, in both cross-sectional and longitudinal comparisons. The during the study period by some 13.4%, from 268 children relationship between height and age in both sexes was found to diagnosed as healthy in 1996 to 232 such children in 1999. be linear for 8–12 year old, but had a curved relation for older Concurrently, the corresponding decrease in the number of healthy children. In our study, the PFT values were measured twice during children in the ‘‘high air pollution’’ zone was 25%. Since the analysis a 3-year period and the results were found to be close to those of covered only children who resided at the same address during the Xuan et al. (2000) with data collected during a longer survey. Even entire study period, these changes in the size of the health status if the children were equally exposed to air pollution before the groups appear to reflect a clinical situation in which air pollution in study period, the changes we observed during the 3 years, covered the study area insults the children’s respiratory system. the study, are likely to reflect the impact of recent air pollution The analysis of variance also confirmed that the interaction exposure. This assumption is based mainly on the fact that the between air pollution levels and the children’s health status was growth spurt in our children (the 8–12 age group) occurred associated significantly with the children’s PFT change during the study period and the relationship between their height (P40.001), thus indicating that the PF development of children growth and DPFT was linear. in different health status groups (that is, healthy children, As Xuan et al. (2000) also found, subjects who had wheeze children with chest symptoms and children with pulmonary and/or asthma during the course of the study were likely to have diseases) appeared to respond differently to air pollution, bronchoconstriction, and hence lower levels of FEV1. This would estimated for the places of the children’s residence. cause such children to have an overall lower level of lung function Regression models, adjusted for the children’s health and (FEV1) but would not explain the observed reduced rate of growth socio-demographic characteristics, also indicated that air pollu- in PFT since there is no reason to suspect that the degree of tion estimates (measured in this study by the NOx  SO2 bronchoconstriction would increase over time. Hence, we believe interaction term) appeared to have a stronger negative effect on that these background data provide an explanation why asthma the PFT change in healthy children and children with chest may be associated with a reduced rate of growth in FEV1. symptoms, than in children with already diagnosed chronic In our study the most affected population were children not respiratory diseases (i.e., spastic bronchitis and asthma). diagnosed as asthmatics but experienced chest symptoms. In fact, A test of DFEV1 response to plausible changes in NOx  SO2 air- those children may belong to an ‘‘undiagnosed’’ and, therefore, pollution levels is reported in Table 6. The test was performed by ‘‘untreated-for-asthma’’ group. This group of children may
  8. 8. ARTICLE IN PRESS 406 T. Yogev-Baggio et al. / Health & Place 16 (2010) 399–408 experience the most severe insult on their respiratory system as An additional sort of bias, attributed to exposure misclassifica- reflected by a greater reduction of their PFT from the baseline.1 tion, can also arise from an assumption that average exposure Our results are generally consistent with the results of levels estimated for townships and SCAs are used to approximate previous studies carried out in the same region (Dubnov et al., the air pollution exposure of individuals (Jerrett et al., 2008; 2007; Portnov et al., 2007), which reported a decline of children’s Portnov et al., 2007; Kim et al., 2009). Since, in the present study, PFT performance in line with increasing air pollution levels. As in we used individual exposure estimates, this source of bias was these previous studies, we used the individual exposure estimate also minimized. approach, based on geographic information systems (GIS) tech- While children spend most of their time at home, a significant nologies, which are more sensitive than zonal approaches, proportion of their ‘‘active’’ time is spent at school which traditionally used in the epidemiological research (Jerrett et al., contributes to the total exposure. Since according to the Israeli 2008; Sahsuvaroglu et al., 2009; Kim et al., 2009). Earlier municipal law, schoolchildren of the elementary and secondary evidence, accumulated during the 1990s, also indicated a rise in schools (1st through 8th grade) attend schools near their homes, the prevalence of acute pulmonary symptoms (upper respiratory most of them stay in the same air pollution zone, which tract infection (URTI), and lower respiratory tract infection (LRTI) minimizes the chance of exposure misclassification. among asthmatics and non-asthmatics children (Brauer et al., Lastly, the effect of air pollution on asthmatic children, who 2002; Goren and Hellmann, 1997). Although preadolescent (6–10 receive medical treatment, and the effect of air pollution on year old) children exhibited a more acute prevalence of these children suffering from chest symptoms, but not taking medica- symptoms, no parallel decline in their PFTs was reported. These tions, may be different (Delfino et al., 2008; Liu et al., 2009). adverse effects air pollution on lung development responses were However, we had no robust information on such ‘‘undiagnosed’’ also detected by studies carried out elsewhere (Gauderman et al., asthmatic children and had to rely on information obtained from 2002, 2004; Horak et al., 2002). questionnaires filled out by the children’s parents. Although there is no shortage of studies investigating the association between air pollution and children pulmonary func- tion performance (Gauderman et al., 2000, 2002, 2004, 2007; 6. Conclusion Horak et al., 2002; Jedrychowski et al., 1999, 2002; Lee et al., 2002; Peters et al., 1999a, 1999b), we are unaware of other As the present study indicates, the long-term exposure to studies that evaluated the effect of air pollution on children ambient air pollution appears to have more detrimental effects on subgroups characterized by different initial health status and used pulmonary function growth among children with chest symptoms individual exposure estimates approach. and healthy than among asthmatics, with the estimated PFT deficit ranging from À 7.9% for asthmatics to À19.6% for children exhibiting chest symptoms. A likely explanation of this trend is 5. Study limitations that different effects of exposure to ambient air pollution on the airways in these groups when lungs rapidly develop. In asthmatic Population-based epidemiologic studies are often susceptible children, the height growth is often stunned. As a result, the to a selection bias arising from a loss of information and prediction pulmonary function values can be less accurate than misclassification of exposure (Elliott and Watenberg, 2000; for children whose growth track is normal. Since asthmatic Greenland and Morgenstern, 1989). Differences in personal children had lower baseline of PF and lower growth rates, the characteristics (physiological, genetic, diseases, lifestyle, smoking, reduction of the PFT values may be less affected. Therefore, the alcohol, drugs, sexual orientation, etc.), and differences in effect of air pollution on this children subgroup may be less pollution sources (industries and power plants, motor vehicles, pronounced. However, further studies are needed to understand naturally occurring dust storms), as well as differences in better this phenomenon. Nevertheless, the data already available topography, and meteorology could lead to these biases (Elliott provide evidence that children, particularly those with increased et al., 1992; Morgenstern and Thomas, 1993; Lash et al., 2009). susceptibility, such as untreated children with chest symptoms, Since the same cohort of children was tested twice (in 1996 and may benefit substantially from a reduction in the current levels of again in 1999), and their health status and residential conditions air pollution, especially in densely populated urban areas. were estimated by individual questionnaires, these sources of bias are less likely to occur in the present study. As the analysis revealed, the number of children with PFT deficiency was highest in air polluted areas which may be Appendix 1 influenced by the concentration of low income families in air polluted neighborhoods. To rule out such a possibility, we See (Table A1). analyzed the income levels of families by small census areas, into which the entire study area is divided, and did not find significant differences in the average monthly incomes (P40.3). Table A1 Therefore, social–economic status differences are unlikely to Overall number of ‘‘air pollution events’’ in the study area and average integrated provide explanation for the observed PF discrepancies in the concentration values (ICV) of SO2 and NOx in 1996–1999. Source: After Dubnov et al. (2007). study cohort. To minimize a potential selection bias, we used in our analysis Year Overall number of ‘‘air Integrated concentration values the ‘‘intent-to-treat’’ approach, commonly used in clinical trials pollution events’’ (ICV) [ppm]n and based on the fixed health status definition of the research SO2 NOx subjects, as recorded at the beginning of the study period (Rothman et al., 2008). 1996 159 5.0 5.6 1997 150 6.1 7.6 1998 157 7.3 8.0 1 Medical surveillance and medicine intake were analyzed in our study as 1999 162 8.0 9.3 potential confounders, but were not found significant. These results are not Total 628 26.4 30.5 reported here for brevit sake and can be obtained from the authors upon request.
  9. 9. ARTICLE IN PRESS T. Yogev-Baggio et al. / Health & Place 16 (2010) 399–408 407 References Goren, A.I., Hellmann, S., Glaser, E.D., 1995. Use of outpatient clinics as a health indicator for communities around a coal-fired power plant. Environ. Health Perspect. 103 (12), 1110–1115. Arshad, S.H., Kurukulaaratchy, R.J., Fenn, M., Matthew, S., 2005. Early life risk Greenland, S., Morgenstern, H., 1989. Ecological bias, confounding, and effect factors for current wheeze, asthma, and bronchial hyperresponsiveness at 10 modification. Int. J. Epidemiol. 18 (1), 269–274. years of age. Chest 127 (2), 502–508. Hankinson, J., Odencrantz, J., Fedan, K., 1999. Spirometric reference values from a ATEP (Association of Towns for Environmental Protection), 2005. Annual Report sample of the general U.S. population. Am. J. Respir. Crit. Care Med. 159, for 2004. Hadera, Israel (Available at: /http://www.igudhadera.co.il/S) (in 179–187. Hebrew). Horak, F., Studnicka, M., Gartner, C., Spengler, J.D., Tauber, E., Urbanek, R., Veiter, A., Avol, E.L., Gauderman, W.J., Tan, S.M., London, S.J., Peters, J.M., 2001. Respiratory Frischer, T., 2002. Particulate matter and lung function growth in children: a effects of relocating to areas of differing air pollution levels. Am. J. Respir. Crit. three-year follow up study in Austrian schoolchildren. Eur. Respir. J. 19, Care Med. 164, 2067–2072. 838–845. Balfour-Lynn, L., 1986. Growth and childhood asthma. Arch. Dis. Child. 61, Islam, T., Gauderman, W.J., Berhane, K., McConnell, R., Avol, E., Peters, J.M., 1049–1055. Gilliland, F.D., 2007. Relationship between air pollution, lung function and Bateson, T.F., Schwartz, J., 2008. Children’s response to air pollutants. J. Toxicol. asthma in adolescents. Thorax 62 (11), 957–963. Environ. Health A 71 (3), 238–243. Jedrychowski, W., Flak, E., Mroz, E., 1999. The adverse effect of low levels of Beers, W.V., Kleijnen, J.P.C., 2003. Kriging for interpolation in random simulation. J. ambient air pollutants on lung function growth in preadolescent children. Oper. Res. Soc. 54, 255–262. Environ. Health Perspect. 107, 669–674. Brauer, M., Hoek, G., Van Vliet, P., Meliefste, K., Fischer, P.H., Wijga, A., et al., 2002. Jedrychowski, W., Maugeri, U., Jedrychowska-Bianchi, I., 2002. Body growth rate in Air pollution from traffic and the development of respiratory infections and preadolescent children and outdoor air quality. Environ. Res. 90 (1), 12–20. asthmatic and allergic symptoms in children. Am. J. Respir. Crit. Care Med. 166 Jerrett, M., Shankardass, K., Berhane, K., Gauderman, J., Kuenzli, N., Avol, E., (8), 1092–1098. Gilliland, F., Lurmann, F., Molitor, J.N., Molitor, J.T., Thomas, D.C., Peters, J., ICBS, Central Bureau of Statistics, Jerusalem, Israel, 2008. Statistical Abstract of McConnell, R., 2008. Traffic-related air pollution and asthma onset in children: Israel. Available at: /http://www.cbs.gov.il/shnaton59/st02_09x.pdfS. a prospective cohort study with individual exposure measurement. Environ. D’Amato, G., Liccardi, G., D’Amato, M., Holgatew, S., 2005. Environmental risk Health Perspect. 116 (10), 1433–1438. factors and allergic bronchial asthma. Clin. Exp. Allergy 35 (9), 1113–1124. Kajekar, R., 2007. Environmental factors and developmental outcomes in the lung. Delfino, R.J., 2002. Epidemiologic evidence for asthma and exposure to air toxics: Pharmacol. Ther. 114, 129–145. linkages between occupational, indoor, and community air pollution research. Kim, J.J., 2004. Ambient air pollution: health hazards to children. Pediatrics 114, Environ. Health Perspect. 110 (4), 573–589. 1699–1707. Delfino, R.J., Staimer, N., Tjoa, T., Gillen, D., Kleinman, M.T., Sioutas, C., Cooper, D., Kim, S.Y., Sheppard, L., Kim, H., 2009. Health effects of long-term air pollution: 2008. Personal and ambient air pollution exposures and lung function influence of exposure prediction methods. Epidemiology 20 (3), 442–450. decrements in children with asthma. Environ. Health Perspect. 116 (4), Lash, T.L., Fox, M.P., Fink, A.K., 2009. Applying Quantitative Bias Analysis to 550–558. Epidemiologic Data Series: Statistics for Biology and Health, vol. XII. Springer Dubnov, J., Barchana, M., Rishpon, S., Levental, A., Segal, I., Carel, R., Portnov, B.A., 194pp. 2007. Estimating the effect of air pollution from a coal-fired power station on Lee, J.T., Kim, H., Song, H., Hong, Y.C., Cho, Y.S., Shin, S.Y., Hyun, Y., Kim, Y.S., 2002. the development of children’s pulmonary function. Environ. Res. 103 (1), Air pollution and asthma among children in Seoul, South Korea. Epidemiology 87–98. 13 (4), 481–484. Elliott, P., Cuzick, J., English, D., Stern, R. (Eds.), 1992. Methods for Small Area Leonardi, G.S., Houthuijs, D., Nikiforov, B., Volf, J., Rudnai, P., Zejda, J., Gurzau, E., Studies. Oxford University Press, Oxford (1996 reprint).. Fabianova, E., Fletcher, T., Brunekreef, B., 2002. Respiratory symptoms, Elliott, P., Watenberg, D., 2000. Bias and confounding in spatial epidemiology. In: bronchitis and asthma in children of Central and Eastern Europe. Eur. Respir. Elliott, P., Wakefield, J., Best, N., Briggs, D.J. (Eds.), Spatial Epidemiology: J. 20 (4), 890–898. Methods and Applications. Oxford University Press, Oxford, pp. 8–84. Littlewood, J.M., Johnson, A.W., Edwards, P.A., Littlewood, A.E., 1988. Growth Elliott, P., Watenberg, D., 2004. Spatial epidemiology: current approaches and retardation in asthmatic children treated with inhaled beclomethasone future challenges. Environ. Health Perspect. 12 (9), 998–1006. dipropionate. Lancet 1, 115–116. Elliott, P., Savitz, D.A., 2008. Design issues in small-area studies of environment Liu, L., Poon, R., Chen, L., Frescura, A.M., Montuschi, P., Ciabattoni, G., Wheeler, A., and health. Environ. Health Perspect. 116 (8), 1098–1104. Dales, R., 2009. Acute effects of air pollution on pulmonary function, airway Feris, B.G., 1978. Recommended respiratory disease questionnaire for use with inflammation, and oxidative stress in asthmatic children. Environ. Health adults and children in epidemiological research. Epidemiology Standardization Perspect. 117 (4), 668–674. Project. Am. Rev. Respir. Dis. 118, 1–120. Minami, M., ESRI, 2000. Using GIS ArcMap. Redlands, California. Gauderman, W.J., Gilliland, F.D., Avol, E., Vora, H., Thomas, D.T., Rappaport, E., Morgenstern, H., Thomas, D., 1993. Principles of study design in environmental Berhane, K., Lurmann, F., McConnell, R., Stram, D., Peters, J.M., 2002. epidemiology. Environ. Health Perspect. 101 (4), 23–38. Association between air pollution and lung function growth in Southern Murray, A.B., Fraser, B., Hardwick, D.F., Pirie, G., 1976. Chronic asthma and growth California children: results from a Second Cohort. Am. J. Respir. Crit. Care Med. failure in children. Lancet, 197–198. 166 (1), 74–84. Nuckols, J.R., Ward, M.H., Jarup, L., 2004. Using geographic information systems for Gauderman, W.J., Vora, H., McConnell, R., Berhane, K., Gilliland, F., Thomas, D., exposure assessment in environmental epidemiology studies. Environ. Health Lurmann, F., Avol, E., Kunzli, N., Jerrett, M., Peters, J., 2007. Effect of exposure to Perspect. 112 (9), 1007–1015. traffic on lung development from 10 to 18 years of age: a cohort study. Lancet Paden, D.B., 2005. The epidemiology and genetics of asthma risk associated with 369, 535–537. air pollution. J. Allergy Clin. Immunol. 115 (2), 213–219. Gauderman, W.J., Avol, E., Gilliland, F., Vora, H., Thomas, D., Berhane, K., McConnell, Peled, R., Friger, M., Bolutin, A., Bibi, H., Epstein, L., Pilpel, D., Scharf, S., 2005. Fine R., Kunzli, N., Lurmann, F., Rappaport, E., Margolis, H., Bates, D., Peters, J., 2004. particles and meteorological conditions are associated with lung functions in The effect of air pollution on lung development from 10 to 18 years of age. N. children with asthma living near two power plants. Public Health 119 (5), Engl. J. Med. 351 (11), 1057–1067. 418–425. Gauderman, W.J., McConnell, R., Gilliland, F., London, S., Thomas, D., Avol, E.L., Peters, J.M., Avol, E.L., Gauderman, W.J., Linn, W.S., Navidi, W., London, S.J., Vora, H., Berhane, K., Rappaport, E., Lurmann, F., Margolis, H., Peters, J., 2000. Margolis, H., Rappaport, E., Vora, H., Gong Jr, H., Thomas, D.C., 1999a. A study of Association between air pollution and lung function growth in Southern twelve southern California communities with differing levels and types of air California children. Am. J. Respir. Crit. Care Med. 162 (4), 1383–1390. pollution: II. Effects on pulmonary function. Am. J. Respir. Crit. Care Med. Gilliland, F.D., Rappaport, E.B., Berhane, K., Islam, T., Dubeau, L., Gauderman, W.J., 159 (3), 768–775. McConnell, R., 2002. Effects of glutathione S-transferase P1, M1, and T1 on Peters, J.M., Avol, E.L., Navidi, W., London, S.J., Gauderman, W.J., Lurmann, F., Linn, acute respiratory illness in school children. Am. J. Respir. Crit. Care Med. 166 W.S., Margolis, H., Rappaport, E., Gong Jr., H., Thomas, D., 1999b. A study of (3), 346–351. twelve Southern California communities with differing levels and types of air Gilliland, F.D., McConnell, R., Peters, J., Gong Jr., H., 1999. A theoretical basis for pollution: I. Prevalence of respiratory morbidity. Am. J. Respir. Crit. Care Med. investigating ambient air pollution and children’s respiratory health. Environ. 159 (3), 760–767. Health Perspect. 107 (3), 403–407. Pikhart, H., Bobak, M., Kriz, B., Danova, J., Celko, M.A., Prikazsky, V., Pryl, K., Briggs, Gilliland, F.D., 2009. Outdoor air pollution, genetic susceptibility, and asthma D., Elliott, P., 2000. Outdoors air concentrations of nitrogen dioxide and sulfur management: opportunities for intervention to reduce the burden of asthma. dioxide and prevalence of wheezing in schoolchildren. Epidemiology 11 (2), Pediatrics 123, S168–S173. 153–160. GINA, Global Strategy for Asthma Management and Prevention, 2008. Reports Portnov, B.A., Dubnov, J., Barchana, M., 2007. On Ecological fallacy, assessment update. Available at: /http://www.ginasthma.orgS. errors stemming from misguided variable selection, and the effect of Goren, A.I., Goldsmith, J.R., Brenner, S., Egoz, N., Rishpon, S., 1990. Prevalence of aggregation on the outcome of epidemiological study. J. Expo. Sci. Environ. respiratory conditions among schoolchildren exposed to different levels of air Epidemiol. 17, 106–121. pollutants in the Haifa Bay area, Israel. Environ. Health Perspect. 89, 225–231. Raizenne, M., Neas, L.M., Damokosh, A.I., Dockery, D.W., Spengler, J.D., Koutrakis, Goren, A.I., Goldsmith, J.R., Hellmann, S., Brenner, S., 1991. Follow-up of school- P., Ware, J.H., Speizer, F.E., 1996. Health effects of acid aerosols on North children in the vicinity of a coal-fired power plant in Israel. Environ. Health American children’s pulmonary function. Environ. Health Perspect. 104, Perspect. 94, 101–105. 506–514. Goren, A.I., Hellmann, S., 1997. Changing prevalence of asthma among school- Rothman, K.J., Greenland, S., Lash, T.L. (Eds.), 2008. Modern Epidemiology. children in Israel. Eur. Respir. J. 10 (10), 2279–2284. Lippincott Williams & Wilkins.
  10. 10. ARTICLE IN PRESS 408 T. Yogev-Baggio et al. / Health & Place 16 (2010) 399–408 ´ Romieu, I., Sienra-Monge, J.J., Ramirez-Aguilar, M., Tellez-Rojo, M.M., Moreno- Sly, P.D., Flack, F., 2008. Susceptibility of children to environmental pollutants. ´ ´ Macıas, H., Reyes-Ruiz, N.I., del Rıo-Navarro, B.E., Ruiz-Navarro, M.X., Hatch, G., Ann. N. Y. Acad. 1140, 163–183. ´ Slade, R., Hernandez-Avila, M., 2002. Antioxidant supplementation and lung Thurston, G.D., Bekkedal, M.Y.V., Roberts, E.M., Ito, K., Pope, C.A., Glenn, B.S., functions among children with asthma exposed to high levels of air pollutants. ¨ Ozkaynak, H., Utell, M.J., 2009. Use of health information in air pollution health Am. J. Respir. Crit. Care Med. 166 (5), 703–709. research: past successes and emerging needs. J. Expo. Sci. Environ. Epidemiol. Sahsuvaroglu, T., Jerrett, M., Sears, M.R., McConnell, R., Finkelstein, N., Arain, A., 19, 45–58. Newbold, B., Burnett, R., 2009. Spatial analysis of air pollution and childhood Wales, J., Barnes, N., Swift, P., 1991. Growth retardation in children on steroids for asthma in Hamilton, Canada: comparing exposure methods in sensitive asthma. Lancet 338, 1535. subgroups. Environ. Health 8, 14. Wang, X., Dockrey, D.W., Wypij, D., Fay, M.E., Ferris Jr., B.J., 1993. Pulmonary Schwartz, J., 2004. Air pollution and children health. Pediatrics 113 (4), function between 6 and 18 years of age. Pediatr. Pulmonol. 15, 75–88. 1037–1043. Wolthers, O.D., Pedersen, S., 1991. Growth of asthmatic children during treatment Schwartz, D.A., 2009. Gene–environment interactions and airway disease in with budesonide: a double blind trial. BMJ 303, 163–165. children. Pediatrics 123 (3), S151–S159. Xuan, W., Peat, J.K., Toelle, B.G., Marks, G.B., Berry, G., Woolcock, A.J., 2000. Lung Sears, M.R., Greene, J.M., Willan, A.R., Weicek, E.M., Taylor, D.R., Flannery, E.M., function growth and its relation to airway hyper responsiveness and recent et al., 2003. A longitudinal, population based cohort of childhood asthma wheeze. Results from a longitudinal population study. Am. J. Respir. Crit. Care followed to adulthood. N. Engl. J. Med. 349, 1414–1422. Med. 161 (6), 1820–1824.