Published ahead of print on April 19, 2007, doi:10.1164/rccm.200510-1678OC
© 2007 American Thoracic Society doi: 10.1164/rccm.200510-1678OC
Lung Function Growth in Children with Long-Term Exposure to Air Pollutants in Mexico City1 Instituto Nacional de Salud Publica, Cuernavaca, Mexico; 2 Instituto Nacional de Enfermedades Respiratorias, Mexico City, Mexico; 3 Universidad Autónoma Metropolitana, Mexico City, Mexico; 4 Medical School, UNAM, Mexico City, Mexico; and 5 School of Public Health, University of North Carolina, Chapel Hill, North Carolina Correspondence and requests for reprints should be addressed to Isabelle Romieu, M.D., M.P.H., Sc.D., Instituto Nacional de Salud Publica, 655 Avenida Universidad, Col. Santa Maria Ahuacatitlán, Cuernavaca, Morelos 62508, México. E-mail: iromieu{at}correo.insp.mx
Rationale: Although short-term exposure to air pollution has been associated with acute, reversible lung function decrements, the impact of long-term exposure has not been well established. Objectives: To evaluate the association between long-term exposure to ozone (O3), particulate matter less than 10 µm in diameter (PM10), and nitrogen dioxide (NO2) and lung function growth in Mexico City schoolchildren. Methods: A dynamic cohort of 3,170 children aged 8 years at baseline was followed from April 23, 1996, through May 19, 1999. The children attended 39 randomly selected elementary schools located near 10 air quality monitoring stations and were visited every 6 months. Statistical analyses were performed using general linear mixed models. Measurements and Main Results: After adjusting for acute exposure and other potential confounding factors, deficits in FVC and FEV1 growth over the 3-year follow-up period were significantly associated with exposure to O3, PM10, and NO2. In multipollutant models, an interquartile range (IQR) increase in mean O3 concentration (IQR, 11.3 ppb) was associated with an annual deficit in FEV1 of 12 ml in girls and 4 ml in boys, an IQR range (IQR, 36.4 µg/m3) increase in PM10 with an annual deficit in FEV1 of 11 ml in girls and 15 ml in boys, and an IQR range (IQR, 12.0 ppb) increase in NO2 with an annual deficit in FEV1 of 30 ml in girls and 25 ml in boys. Conclusions: We conclude that long-term exposure to O3, PM10, and NO2 is associated with a deficit in FVC and FEV1 growth among schoolchildren living in Mexico City.
Key Words: lung function growth • air pollution • children
Epidemiologic studies have shown that acute exposure to ambient air pollution is associated with a range of respiratory events in children (1–3). Although there is growing evidence that air pollution exposure is likely to affect lung growth (4–10), there is controversy on which pollutant is most harmful to health. Long-term exposure to ozone (O3) has been associated with significantly decreased lung function in retrospective and prospective cohorts of children (6, 11) and young adults (4, 5), and the Children's Health Study (CHS) (7–9) has reported that nitrogen dioxide (NO2), acid vapor, and elemental carbon had the strongest effect. The metropolitan area of Mexico City experiences significant air pollution problems. O3 levels are high, and the average 1-hour daily maximum frequently exceeds 110 ppb (the Mexican standard) (12). Studies conducted among children with asthma who live in Mexico City have shown a decrement in lung functions and an increase in respiratory symptoms (13–16), suggesting that children with lifelong exposure to a heavily polluted environment, mainly to ozone pollution, have detectable abnormalities that could be indicative of small airway disease or decreased total lung capacity. We conducted a prospective dynamic cohort study to evaluate the long-term effect of ambient air pollution on the lung function growth of Mexico City schoolchildren. Some of the results of this study have been previously reported in the form of abstracts (17, 18).
Study Design We selected 10 fixed-site air monitoring stations in Mexico City and randomly selected 39 elementary schools from among those located within 2 km of the stations. The study cohort consisted of students of the selected schools who were 8 years of age at the beginning of the study, who had not been diagnosed as having asthma, and whose parents had signed a consent letter. A substantial number of children entered or left the cohort during the course of the study. At baseline, a questionnaire was completed by the parents of 1,819 children, and a spirometric test was administered to each child (phase 1). An additional 1,351 participants of the same age (±1 mo) as the previously enrolled children were added to the cohort in subsequent phases. The cohort was followed every 6 months (spring and fall) for 3 years with spirometric tests and two questionnaires, one completed by the parents and the other by the children and their teachers. Lung function testing was conducted by trained technicians following American Thoracic Society standards (19), using computerized dry rolling-seal spirometers (model 922; SensorMedics, Yorba Linda, CA). The spirometry quality control program has been previously reported (20).
Air Pollution Monitoring Long-term exposure for each day of the study period was estimated as the averages over the previous 6 months of the daily O3 8-hour mean, PM10 24-hour means, and NO2 24-hour means averaged over the previous 6 months. These averages vary depending on the station assigned to each school. Only low concentrations of SO2 and CO were registered, so their effects were not analyzed.
Statistical Analysis
Table 1 and Table E1 in the online supplement present the characteristics of the study population (n = 3,170) by study phase and sex. Anthropometric measurements and lung function variables increased over time for both sexes. Figure 1 presents the location of the monitoring stations used in the study. Air pollutant concentrations over the 3-year study period were higher in spring than in fall (Table E2). Over the study period, 8-hour mean O3 concentrations ranged from 60 ppb (SD, 25) in the northeast area to 90 ppb (SD, 34) in the southwest, and 24-hour mean PM10 concentrations ranged from 53 µg/m3 (SD, 32) in the southwest to 97 µg/m3 (SD, 49) in the northeast (Table 2). O3 was negatively associated with PM10 (r = –0.23; p < 0.001) and positively associated with NO2 (r = 0.166; p < 0.001).
Table 3 presents O3, PM10, and NO2 means and percentiles of 6-month mean concentrations and interquartile range (IQR) during the study period. The widest IQR was observed in 6-month mean PM10 concentrations.
Table 4 presents the results by sex of our final mixed models for FVC, FEV1, and FEF25–75% adjusted for age, body mass index, height, height by age, weekday time spent in outdoor activities, environmental tobacco smoke exposure, previous-day mean air pollutant concentration, and time since first test. One-pollutant models showed an association between ambient air pollutants and deficits in lung growth. In girls, a 11.3-ppb increase (IQR) in O3 was associated with an annual deficit of –35 ml (95% confidence interval [CI], –41 to –29) in FVC, –24 ml (95% CI, –30 to –19) in FEV1, and –20 ml/s (95% CI, –32 to –8) in FEF25–75%. The annual deficits for boys were –25 ml (95% CI, –31 to –19) in FVC, –16 ml (95% CI, –21 to –11) in FEV1, and –8 ml/s (95% CI, –19 to 4) in FEF25–75%. Ambient PM10 and NO2 concentrations were similarly negatively associated with lung growth. In girls, a 36.4 µg/m3 increase (IQR) in PM10 was associated with an annual deficit of –39 ml (95% CI, –47 to –31) in FVC and –29 ml (95% CI, –36 to –21) in FEV1. The corresponding deficits for boys were –33 ml (95% CI, –41 to –25) in FVC and –27 ml (95% CI, –34 to –19) in FEV1. Slightly larger coefficients were observed for the effect of NO2. For a 12.0-ppb increase (IQR) in NO2, the annual deficits were –48 ml (95% CI, –55 to –41) for FVC and –32 ml (95% CI, –39 to –26) for FEV1 in girls and –45 ml (95% CI, –53 to –37) for FVC and –26 ml (95% CI –33 to –19) for FEV1 in boys. No significant effect of PM10 and NO2 was observed on FEF25–75%. Estimates from two-pollutant models were not substantially different. In multipollutant models, the negative association of O3, PM10, and NO2 with lung function growth persisted, but the effect was slightly stronger for O3 in girls than boys. Because the observed impact was greater on FVC than on FEV1, the FEV1/FVC ratio for both sexes tended to increase with higher pollutant concentrations in all models (Table 4). When the percentage annual changes in predicted values were calculated on the basis of the reference equations for Mexican children (21), the results were similar (Table E4).
Figure 2 presents the estimated growth in FVC (Figure 2A), FEV1 (Figure 2B), and FEF25–75% (Figure 2C) for the 25, 50, and 75 percentiles of O3, PM10, and NO2 concentrations by sex, obtained from multipollutant models. At the beginning of the study and at each phase of follow-up, children exposed to lower O3 and PM10 concentrations had better lung function values than children exposed to higher concentrations.
Because FEF25–75%/FVC is a marker of low volume in small airways and because this might modify the effect of O3 on FEF25–75%, we stratified by tertiles of FEF25–75%/FVC. O3 was significantly related to a deficit in lung growth (–5.1% per 10 ppb; 95% CI –8.7 to –1.5) among girls with the lowest FEF25–75%/FVC (lower two tertiles). No effect was observed among girls in the highest tertile or among boys.
Our study revealed significant deficits in lung function growth in children with long-term exposure to air pollutants. In one-pollutant models, O3, PM10, and NO2 were associated with a significant deficit in FVC and FEV1 growth in girls and boys. The FEV1/FVC ratio increased because exposure had a greater impact on FVC than on FEV1. An association between O3 and a deficit in FEF25–75% growth was observed only among girls with a low FEF25–75%/FVC ratio. The effect on FEV1 during the 3 years of follow-up was slightly greater than that reported for exposure to maternal smoking among children in the United States (22, 23). A deficit in FVC and FEV1 growth was observed for O3, PM10, and NO2 after adjusting for the acute effect of these pollutants (previous-day concentrations) and for confounding factors. In multipollutant models, O3 and NO2 had the strongest effect among girls. These results are in part consistent with previous results from the CHS, which revealed a deficit in growth mostly with exposure to PM2.5, NO2, and inorganic acid vapor (7, 9, 15). In the CHS, O3 exposure was associated with a growth deficit in peak expiratory flow rate in fourth graders followed for only 4 of the 8 years of follow-up (8, 22). Lower lung functions in children exposed to higher O3 concentrations have been reported in cross-sectional studies (24), including one based on CHS data that observed lower peak expiratory flow rate and maximal midexpiratory flow particularly in girls spending more time outdoors (25) and in retrospective (11) and prospective (6) cohort studies. Deficit in lung growth was associated with a set of pollutants including O3, PM10, and NO2. Their main source, particularly that of NO2 and O3, is traffic related: The former is directly emitted from tailpipes, and the latter is a photochemical reaction to exhaust gases (26). Because the pollutants are correlated, independent effects could not be accurately estimated. The mechanism of action by which long-term exposure to air pollution produces changes in lung development has not been established. Human and animal studies have demonstrated several changes in lung morphology related to O3 exposure (27–29), particularly with a predominantly restrictive pattern (30). Calderon and colleagues (16) have suggested that chronic and sustained inhalation of a complex mixture of air pollutants, including O3 and PM, might be associated with small airway disease. Recently, oxidative stress resulting from increased exposure to oxidant compounds (O3, NO2, and particulate components) has been identified as a major feature underlying the toxic effects of air pollutants (31–33). The resulting increased expression of proinflammatory cytokines leads to an enhanced inflammatory response (32) and potential chronic lung damage. It is not clear whether this might result in permanent loss or whether the pattern of exposure (repeated peak exposure versus average exposure) is relevant. Because of the shortness of the 3-year follow-up period and the nonlinear pattern of childhood lung function growth (34), we were unable to estimate the impact on lung function attained in early adulthood. In our study, exposure to pollutants was associated with a higher FEV1/FVC ratio, suggesting a restrictive pattern similar to that already described in animals (30) and humans (35). However, the effect of air pollution on the resulting functional pattern has been inconsistent (36, 37), and inflammatory changes in small airways have been observed (11). Several factors need to be considered when interpreting our results. The exposure values used in studies on the long-term effects of air pollution on lung function growth are usually fixed-site monitoring station data that have been averaged over communities. To reduce exposure misclassification, our study was based on schools located within 2 km of the monitoring stations. In addition, we conducted microenvironmental and personal exposure assessments in a randomly selected subsample of 60 children, using passive O3 samplers and personal PM10 monitors. PM10 and O3 concentrations from personal, indoor, and outdoor monitors were significantly correlated with the measurements obtained from the fixed-site air monitoring stations (38). Our results were not substantially modified when we adjusted our models for potential confounding factors. However, we did not have information on variables such as smoking during pregnancy, birth weight, and atopy, which have been associated with reduced lung function growth (39, 40). Because socioeconomic status might have a differential distribution across monitoring stations and pollution concentrations and might be a determinant of lung function in our population, we tested for interactions between maternal and paternal education levels and air pollutant effects. None of these interactions was significant, suggesting that socioeconomic status could not explain our results. Lung function testing was conducted by trained technicians, and all spirometry test results were reviewed by a pulmonologist blinded to the location of the school attended by the child. The reproducibility of the tests was good and has been previously reported (20). In the event of missing lung function data, if the data were missing in one phase but not in the next the rate of change could be imputed from the mixed model because the coefficients obtained from the mixed model analysis estimated mean effects.
Conclusions
The authors thank Steve Marshall, Kiros Berhane, James Gauderman, Nino Kuenzli, and Rob McConnell for their important input on this study; the Mexico City monitoring network (RAMA) and the field team for providing high-quality data; and the school principals, teachers, students and parents for their participation. They also thank Garth Evans for reviewing the English manuscript.
Supported by the Mexican Sciences and Technology Council (CONACYT), SALUD-2005-01-13956 and by the National Center for Environmental Health–Centers for Disease Control and Prevention, Atlanta, GA. This article has an online supplement, which is accessible from this issue's table of contents at www.atsjournals.org Originally Published in Press as DOI: 10.1164/rccm.200510-1678OC on April 19, 2007 Conflict of Interest Statement: R.R.-M. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. R.P.-P. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. G.O.-F. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. L.M.-A. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. H.M-M. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. T.F. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. W.M. received $21,000 from API in 2006 as a contract to develop and publish ozone exposure-response models. D.L. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. I.R. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. Received in original form October 27, 2005; accepted in final form April 19, 2007
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
AT A GLANCE COMMENTARY
