Abstract
Rationale: The role of air pollution exposure in the development of asthma, allergies, and related symptoms remains unclear, due in part to the limited number of prospective cohort studies with sufficiently long follow-ups addressing this problem.
Objectives: We studied the association between traffic-related air pollution and the development of asthma, allergy, and related symptoms in a prospective birth cohort study with a unique 8-year follow-up.
Methods: Annual questionnaire reports of asthma, hay fever, and related symptoms during the first 8 years of life were analyzed for 3,863 children. At age 8, measurements of allergic sensitization and bronchial hyperresponsiveness were performed for subpopulations (n = 1,700 and 936, respectively). Individual exposures to nitrogen dioxide (NO2), particulate matter (PM2.5), and soot at the birth address were estimated by land-use regression models. Associations between exposure to traffic-related air pollution and repeated measures of health outcomes were assessed by repeated-measures logistic regression analysis. Effects are presented for an interquartile range increase in exposure after adjusting for covariates.
Measurements and Main Results: Annual prevalence was 3 to 6% for asthma and 12 to 23% for asthma symptoms. Annual incidence of asthma was 6% at age 1, and 1 to 2% at later ages. PM2.5 levels were associated with a significant increase in incidence of asthma (odds ratio [OR], 1.28; 95% confidence interval [CI], 1.10–1.49), prevalence of asthma (OR, 1.26; 95% CI, 1.04–1.51), and prevalence of asthma symptoms (OR, 1.15; 95% CI, 1.02–1.28). Findings were similar for NO2 and soot. Associations were stronger for children who had not moved since birth. Positive associations with hay fever were found in nonmovers only. No associations were found with atopic eczema, allergic sensitization, and bronchial hyperresponsiveness.
Conclusions: Exposure to traffic-related air pollution may cause asthma in children.
Exposure to ambient air pollution can exacerbate existing asthma. The role of exposure to ambient air pollution in the development of childhood asthma, allergy, and related symptoms, however, remains less clear due in part to the limited number of prospective birth cohort studies.
In this prospective birth cohort study, we found positive associations between traffic-related air pollution levels outside subjects' homes, and the incidence and prevalence of asthma during the first 8 years of life. Our results provide evidence that air pollution exposure may contribute to the pathogenesis of asthma in children.
The prevalence of respiratory symptoms changes rapidly during childhood, and only a minority of children with reports of symptoms in early childhood develop asthma at school age (14, 15). It is therefore important to study the incidence of asthma over a sufficiently long period of time in relation to air pollution exposure, and this has not been done so far in the context of birth cohort studies.
The 8-year follow-up of the Dutch Prevention and Incidence of Asthma and Mite Allergy (PIAMA) Study (16) has been completed. The follow-up is unique because of its length and the availability of annual standardized questionnaires on the respiratory health of the children and its determinants. Moreover, at the age of 8 years, a more reliable diagnosis of asthma and allergies is possible. This gives us the unique opportunity to study the association between traffic-related air pollution and the development of asthma, allergy, and related symptoms longitudinally over an 8-year period. In addition, objective measures such as allergen-specific IgE and bronchial hyperresponsiveness at age 8 are available for a subset of the population. Some of the results of this study have been previously published in the form of an abstract (17).
The PIAMA Study is a prospective birth cohort study (16). Women were recruited in 1996–1997 during their second trimester of pregnancy from a series of communities in the north, west, and center of The Netherlands. Nonallergic pregnant women were invited to participate in a “natural history” study arm. Pregnant women identified as allergic through the screening questionnaire were allocated primarily to an intervention arm with a random subset allocated to the natural history arm. The intervention involved the use of mite-impermeable mattress and pillow covers. The institutional review boards of the participating institutes approved the study protocol, and written informed consent was obtained from all participants.
Air pollution concentrations at the birth address of each participant were estimated by land-use regression models described elsewhere (18). Briefly, four 2-week measurements of NO2, fine particles (PM2.5, particulate matter less than 2.5 μm in diameter), and “soot” (determined as the reflectance of the PM2.5 filters) were performed at each of 40 sites within 1 year and then adjusted for temporal trends to calculate long-term average concentrations. Geographic information system (GIS) data were also collected regarding traffic, road, and population density in the vicinity of each monitoring location. Regression models were developed to relate the annual average concentrations measured at the 40 monitoring sites with the GIS variables. The models explained 73, 81, and 85% of the variability in the annual average concentrations for PM2.5, soot, and NO2, respectively. These models were then used to obtain a unique ambient air pollution concentration estimate at the birth address of each participant. More information about the regression models is provided in the online supplement.
Information about respiratory health and atopic eczema was obtained in yearly questionnaires that were completed by the parents approximately at the time of the child's birthday until the age of 8 years. From these questionnaires, the same health outcomes as in previous analyses (6, 19) were selected to describe asthma, allergies, and related symptoms. The exact definitions of the health outcomes studied are given in the online supplement.
At age 8 years, a blood sample was taken from all children who gave consent. Specific IgE levels for house dust mite, cat, dog, cocksfoot (Dactylis glomerata) and birch pollen, Alternaria alternata, egg, and milk were measured in blood samples by a radioallergosorbent test-like method used at the Sanquin Laboratories (Amsterdam, The Netherlands). Sensitization was defined as a positive reaction (IgE level ≥0.35 IU/ml) to one of the allergens tested.
At age 8 years, all children of allergic mothers and a random sample of children of nonallergic mothers were invited for a medical examination including determination of bronchial hyperresponsiveness (BHR). BHR was defined as a decrease in FEV1 of 20% or more at a cumulative dose of methacholine bromide not exceeding 0.61 mg. More details are provided in the online supplement.
A series of potential confounding variables including sex, study arm (intervention or natural history), use of mite-impermeable mattress covers, allergies of mother and father, maternal and paternal education, maternal smoking during pregnancy, breastfeeding, presence of a gas stove in the child's home, presence of older siblings, smoking, signs of dampness and pets in the child's home, day-care attendance, and Dutch nationality were selected as in previous analyses (6, 19). High parental education was defined as having at least completed high school. Smoking was defined as any smoking at the child's home assessed by validated questionnaires completed by the parents. Confounders were defined for the first year of life to coincide as much as possible with air pollution exposure, which was estimated for addresses at birth.
Longitudinal associations between air pollution levels at the birth address and yearly questionnaire reports of binary health outcomes at ages 1 to 8 years were analyzed by generalized estimation equations (GEE) with a logit link (20). A seven-dependent correlation matrix was chosen to account for correlations between repeated observations in the same individual at the eight different points in time. For outcomes measured at six points in time, a five-dependent matrix was chosen. Associations between air pollution and categorical outcomes that were assessed at age 8 years only were analyzed by standard logistic regression (outcomes with two categories) and polytomous logistic regression (21) (outcomes with more than two categories), respectively. Air pollution levels were entered as continuous variables without transformation in all models. All associations were calculated with and without adjustment for confounding variables. All odds ratios are presented for an interquartile range increase in estimated exposures (10.4 μg · m−3 for NO2, 3.2 μg · m−3 for PM2.5, and 0.57 × 10−5 m−1 for soot).
Moving houses during the follow-up might have resulted in measurement error. We therefore included interaction terms between air pollution levels and moving status in the models to estimate associations between air pollution and outcomes for children who did and did not move separately.
The study started with 3,963 newborns. For a total of 3,874 (97.8%) of these children, the parents returned at least one of the yearly questionnaires at ages 1 to 8 years. Response was high at all ages (Figure E1 in the online supplement). After exclusion of children with missing information about air pollution levels (n = 9) and all health outcomes studied (n = 2), 3,863 children remained for the final analysis. Measurements of allergen-specific IgE and BHR were available for 1,700 and 936 children, respectively. General characteristics of all children are presented in Table 1; general characteristics of the subsets of children with measurements of specific IgE and BHR are presented in Table E1 in the online supplement. By design, participants in the intervention study were overrepresented among the children with measurements of BHR and consequently the percentage of children with allergic mothers was higher in this subpopulation. In the subset of children with measurements of specific IgE, because of selective participation the proportion of children with allergic mothers was also somewhat higher than in the larger population. In addition, in both subsets the percentage of children with highly educated parents, the percentage of children who were breastfed at 3 months, and the percentage of Dutch children were somewhat higher than in the large study population; maternal smoking during pregnancy, smoking at the child's home, and pet ownership were less prevalent than in the large population.
Variable | n/N | % | ||
|---|---|---|---|---|
| General Characteristics | ||||
| Female sex | 1,864/3,863 | 48.3 | ||
| Study region | ||||
| North | 1,208/3,863 | 31.3 | ||
| Middle | 1,561/3,863 | 40.4 | ||
| West | 1,094/3,863 | 28.3 | ||
| Study arm | ||||
| Natural history study | 3,128/3,863 | 81.0 | ||
| Intervention, mite impermeable mattress cover | 382/3,863 | 9.9 | ||
| Intervention, placebo mattress cover | 353/3,863 | 9.1 | ||
| Allergic mother | 1,183/3,865 | 30.6 | ||
| Allergic father | 1,175/3,853 | 30.5 | ||
| Mother's education, % high | 1,328/3,796 | 35.0 | ||
| Father's education, % high | 1,491/3,752 | 39.7 | ||
| Mother smoking during pregnancy | 664/3,809 | 17.4 | ||
| Breastfeeding at 3 mo | 1,857/3,807 | 48.8 | ||
| Gas stove* | 3,158/3,821 | 82.6 | ||
| Unvented gas water heater | 152/3,509 | 4.3 | ||
| Any other siblings (at birth) | 1,942/3,861 | 50.3 | ||
| Any smoking at home* | 1,026/3,725 | 27.5 | ||
| Signs of dampness in living room/child's bedroom* | 310/3,688 | 8.4 | ||
| Any pets in home* | 1,886/3,793 | 49.7 | ||
| Day-care attendance ≥ 4 h/wk* | 896/3,725 | 24.1 | ||
| Dutch nationality | 3,479/3,691 | 94.3 | ||
| Did not move house during first 8 yr of life | 1,851/3,857 | 48.0 | ||
| Health Outcomes at Age 8 years | ||||
| Wheeze phenotype | ||||
| Never | 1,758/3,016 | 58.3 | ||
| Early transient | 811/3,016 | 26.9 | ||
| Late onset | 122/3,016 | 4.0 | ||
| Persistent | 325/3,016 | 10.8 | ||
| Bronchial hyperresponsiveness | 400/936 | 42.7 | ||
| Allergic sensitization | ||||
| Any allergen positive | 693/1,700 | 40.8 | ||
| Inhalant allergens only | 406/1,698 | 23.9 | ||
| Food allergens only | 144/1,698 | 8.5 | ||
| Inhalant and food allergens | 141/1,698 | 8.3 | ||
| Atopic asthma | 54/1,591 | 3.4 | ||
| Nonatopic asthma | 13/1,591 | 0.8 | ||
The percentage of children with questionnaire-based health outcomes that have been assessed annually are presented in Figure 1. Prevalence of health outcomes that have been assessed at age 8 years only (wheezing phenotypes, BHR, allergic sensitization, atopic and nonatopic asthma) are presented in Table 1. The annual prevalence of asthma was between 3 and 6%. The annual incidence of asthma was highest in the first year of life (6.1%) and relatively constant (1.4–2.4%) at the other ages. Prevalence of asthma-related symptoms decreased with age. The percentage of children who ever had hay fever increased from 1.2% at age 3 years to 5.1% at age 8 years. The prevalence of atopic eczema was relatively constant at all ages (14.4–18.0%).
Figure 1. Percentage of children with asthma, hay fever, atopic eczema, and related symptoms at ages 1 to 8 years. NA = not available.
[More] [Minimize]The distributions of annual average NO2, PM2.5, and soot levels at the participants' birth addresses are presented in Table 2. The estimated exposures for the various pollutants were highly correlated (r = 0.93, 0.96, and 0.97 for the correlation between NO2 and PM2.5, NO2 and soot, and PM2.5 and soot, respectively).
NO2 | PM2.5 | Soot | |
|---|---|---|---|
| (μg/m3) | (μg/m3) | (10−5 m−1) | |
| Minimum | 12.6 | 13.5 | 0.77 |
| 10th percentile | 14.7 | 14.0 | 1.15 |
| 25th percentile | 18.5 | 14.9 | 1.35 |
| 50th percentile | 26.0 | 17.3 | 1.78 |
| Mean | 25.4 | 16.9 | 1.72 |
| 75th percentile | 28.9 | 18.1 | 1.92 |
| 90th percentile | 34.7 | 19.1 | 2.17 |
| Maximum | 58.4 | 25.2 | 3.68 |
The overall associations with air pollution levels were statistically significant (P < 0.05) for prevalence and incidence of asthma, and prevalence of asthma symptoms, wheeze, and sneezing, runny/blocked nose during the past 12 months before and after confounder adjustment (Figure E2 in the online supplement and Figure 2, respectively). Age-specific associations from models with air pollution–age interaction terms revealed little variation of the effects of air pollution on asthma and related symptoms with age. All odds ratios (except for incident asthma at age 2 yr) were greater than one and differences in air pollution effects between ages were small with the exception of asthma incidence, for which associations became somewhat stronger at ages 6–8 years. The odds ratios for sneezing, runny/blocked nose increased with age and were greater than one from age 5 years onward. The odds ratios for atopic eczema were all close to unity. Overall, changes in odds ratio due to confounder adjustment were small (less than 10% with few exceptions). Odds ratios were similar for NO2, PM2.5, and soot because of the high correlation between pollutants.
Figure 2. Adjusted overall and age-specific association between annual average levels of air pollution at the birth address and asthma, hay fever, atopic eczema, and related symptoms during the first 8 years of life. Results are presented as adjusted odds ratios (ORs) with 95% confidence intervals. We adjusted for all characteristics from Table 1 except study region. ORs were calculated for an interquartile range increase in air pollution levels (10.4 μg · m−3 for NO2, 3.2 μg · m−3 for PM2.5, and 0.57 × 10−5 m−1 for soot).
[More] [Minimize]Crude and adjusted associations between air pollution levels at the birth address and health outcomes, which have been assessed at age 8 years only, are presented in Table E1 and Table 3, respectively. Associations with air pollution levels were positive for all wheezing phenotypes, which is consistent with the positive associations between air pollution levels and wheeze at all ages. No association was found between air pollution levels and BHR. Air pollution levels were associated with a significantly increased risk of allergic sensitization in unadjusted analyses but the associations were somewhat reduced and no longer statistically significant after adjustment for potential confounders. When we distinguished between atopic and nonatopic asthma at age 8 years, positive associations with air pollution levels seemed to be limited to nonatopic asthma in crude analyses (crude odds ratios [95% confidence intervals] for PM2.5, NO2, and soot were 1.85 [0.92–3.73], 2.98 [1.21–7.37], and 2.06 [0.99–4.30], respectively, for nonatopic asthma; and 0.95 [0.64–1.40], 1.00 [0.63–1.58], and 0.97 [0.64–1.46]), respectively, for atopic asthma). Adjusted models did not converge for nonatopic asthma because of the small number of children with nonatopic asthma.
NO2 | PM2.5 | Soot | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| OR | 95% CI | OR | 95% CI | OR | 95% CI | |||||||
| Wheeze phenotype (N = 2,668†) | ||||||||||||
| Early transient | 1.17 | 0.97–1.41 | 1.29 | 1.04–1.62 | 1.22 | 1.00–1.48 | ||||||
| Late onset | 1.13 | 0.99–1.29 | 1.18 | 1.01–1.37 | 1.13 | 0.99–1.30 | ||||||
| Persistent | 1.30 | 0.99–1.72 | 1.37 | 0.98–1.91 | 1.30 | 0.98–1.74 | ||||||
| Bronchial hyperresponsiveness (N = 818†) | 1.04 | 0.85–1.28 | 0.98 | 0.76–1.24 | 1.04 | 0.84–1.29 | ||||||
| Allergic sensitization (N = 1,499†,‡) | ||||||||||||
| Any allergen | 1.11 | 0.95–1.30 | 1.16 | 0.96–1.39 | 1.12 | 0.95–1.32 | ||||||
| Inhalant allergens only | 1.07 | 0.89–1.29 | 1.14 | 0.92–1.42 | 1.09 | 0.90–1.32 | ||||||
| Food allergens only | 1.12 | 0.85–1.46 | 1.21 | 0.88–1.67 | 1.16 | 0.87–1.53 | ||||||
| Inhalant and food allergens | 1.17 | 0.89–1.56 | 1.09 | 0.78–1.52 | 1.10 | 0.82–1.48 | ||||||
We explored the possibility of regional confounding by additional adjustment for study region. Results for PM2.5 are presented in Table 4; results for NO2 and soot are presented in Table E3 of the online supplement. Results for the models without adjustment for study region (which are presented in Figure 2 and Table 3) are included as well in these tables to facilitate comparison. The positive associations between air pollution levels, prevalent and incident asthma, and symptoms of asthma remained stable. The association between air pollution levels and wheeze decreased to unity. Because study region is an important determinant of air pollution levels in the land-use regression models that were used to estimate exposures, the adjustment for region may be an overadjustment.
Without Adjustment for Study Region | With Adjustment for Study Region | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| n† | OR | 95% CI | OR | 95% CI | ||||||
| During the first 8 years of life | ||||||||||
| Prevalent asthma | 3,184‡ | 1.26 | 1.04–1.51 | 1.36 | 0.99–1.88 | |||||
| Incident asthma | 3,143‡ | 1.28 | 1.10–1.49 | 1.26 | 0.97–1.63 | |||||
| Asthma symptoms | 3,156‡ | 1.15 | 1.02–1.28 | 1.22 | 0.98–1.51 | |||||
| Wheeze | 3,184‡ | 1.20 | 1.08–1.33 | 1.04 | 0.85–1.28 | |||||
| Sneezing, runny/blocked nose | 3,156‡ | 1.12 | 1.01–1.24 | 1.14 | 0.95–1.37 | |||||
| Hay fever | 3,156‡ | 1.05 | 0.83–1.32 | 0.73 | 0.45–1.20 | |||||
| Atopic eczema | 3,184‡ | 1.00 | 0.90–1.11 | 1.11 | 0.91–1.35 | |||||
| At age 8 years | ||||||||||
| Bronchial hyperresponsiveness | 818 | 0.98 | 0.76–1.24 | 1.04 | 0.66–1.66 | |||||
| Allergic sensitization | 1,499 | 1.16 | 0.96–1.39 | 1.06 | 0.76–1.49 | |||||
Moreover, we explored the impact of the study design (i.e., the division into an intervention arm and a natural-history study arm) by excluding the participants of the intervention study, which represent approximately one fifth of the total study population. The results remained essentially unchanged (data not shown).
Less than half the population (48.0%) still lived at their birth address at age 8 years. Children who had moved house at least once during the follow-up, on average had lived at their birth address for 3.2 (SD 2.1) years before moving; 14.0% of the movers (7.3% of the total study population) moved during the first year of life. To check whether effects were different between children who did and who did not move house, we included terms for the interaction between air pollution and moving status in the models (for the repeated measures in the models for the overall effects, i.e., the models without air pollution–age interaction). Results for PM2.5 are shown in Table 5. Results for NO2 and soot were similar (Table E4 in the online supplement). Effects were larger and significant only for those who did not move house. There was no association between air pollution levels and hay fever in the entire study population, but there was a statistically significant positive association for children who did not move house.
Nonmovers | Movers | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| n† | OR | 95% CI | n† | OR | 95% CI | |||||||
| During the first 8 years of life | ||||||||||||
| Prevalent asthma | 1,575‡ | 1.34 | 1.03–1.75 | 1,605‡ | 1.19 | 0.92–1.53 | ||||||
| Incident asthma | 1,551‡ | 1.36 | 1.09–1.69 | 1,588‡ | 1.20 | 0.98–1.48 | ||||||
| Asthma symptoms | 1,556‡ | 1.27 | 1.07–1.49 | 1,596‡ | 1.05 | 0.90–1.23 | ||||||
| Wheeze | 1,575‡ | 1.25 | 1.08–1.45 | 1,605‡ | 1.15 | 1.00–1.32 | ||||||
| Sneezing, runny/blocked nose | 1,556‡ | 1.14 | 0.98–1.34 | 1,596‡ | 1.09 | 0.95–1.25 | ||||||
| Hay fever | 1,556‡ | 1.43 | 1.01–2.04 | 1,596‡ | 0.81 | 0.58–1.12 | ||||||
| Atopic eczema | 1,575‡ | 1.03 | 0.89–1.21 | 1,605‡ | 0.97 | 0.84–1.12 | ||||||
| At age 8 years | ||||||||||||
| Bronchial hyperresponsiveness | 464 | 0.99 | 0.69–1.41 | 351 | 0.96 | 0.69–1.34 | ||||||
| Allergic sensitization | 617 | 1.25 | 0.96–1.61 | 881 | 1.07 | 0.83–1.39 | ||||||
We demonstrated positive associations between levels of traffic-related air pollution at the birth address and the incidence and prevalence of asthma, and the prevalence of asthma-related symptoms in children who were monitored from birth until 8 years of age. In addition, traffic-related air pollution was found to be associated with symptoms of rhinitis during the first 8 years of life. Positive associations with hay fever were found in nonmovers only.
The present work extends results from previous analyses within the same birth cohort study at ages 2 and 4 years, respectively (6, 19, 22), in several ways: first, the follow-up was extended to the age of 8 years; and second, the data were analyzed longitudinally rather than cross-sectionally, making full use of the cohort design. We could show that the positive associations with wheeze and doctor-diagnosed asthma at ages 2 and 4 years persisted over time. At age 4 years, air pollution levels were associated with allergic sensitization to food allergens, but not to inhalant allergens. At age 8 years, associations with sensitization to food allergens and inhalant allergens were similar but both became nonsignificant after confounder adjustment (6). We found positive associations of air pollution levels with symptoms of rhinitis (sneezing, runny/blocked nose). Positive associations between air pollution levels and hay fever were limited to the participants who did not move house since birth. In line with our previous findings, we did not find an association between air pollution levels and atopic eczema. The relatively stable prevalence of atopic eczema in the present study is similar to the prevalence of physician-diagnosed eczema during the first 6 years of life reported from a comparable German birth cohort study (23).
One of the strengths of the present study is its prospective design with yearly questionnaire follow-ups over a period of 8 years, starting at birth. This enables us to fill the gap between those cohort studies reporting associations between traffic-related air pollution and asthma and related symptoms in early life (6, 8, 11, 19, 24) and those reporting associations with asthma prevalence and incidence at school age (9, 12). Response rates were high during the entire follow-up with 82% of the original population (3,269 of 3,963) returning the questionnaire at age 8 years. In addition, for a large subgroup of the population objective measures such as allergen-specific IgE and bronchial hyperresponsiveness at age 8 years were available. Associations with air pollution were found primarily for questionnaire-reported health outcomes. Associations with allergic sensitization became insignificant after confounder adjustment. The most important confounders were nationality, parental education, and day-care attendance: non-Dutch children, children of highly educated parents, and children attending day-care were on average exposed to slightly higher levels of air pollution and had higher rates of sensitization. Changes in the association between air pollution levels and allergic sensitization due to adjustment for single confounders were generally small (<4%), but when all confounding variables were entered into the model the associations became nonsignificant. At age 8 years, we were able to distinguish between atopic and nonatopic asthma. There was some indication that the effect might be limited to nonatopic asthma. However, because the number of children with atopic asthma, and in particular those with nonatopic asthma, were small, results must be interpreted with caution.
A potential limitation of our study may be that asthma was defined by parental reporting of physician-diagnosed asthma, rather than by direct physician diagnosis using standardized criteria such as lung function measurements and bronchial hyperresponsiveness. Bronchial hyperresponsiveness was measured in a subset of 936 children at the age of 8 years, but no association with air pollution was found. The reason for this is not clear. The percentage of children with bronchial hyperresponsiveness in our study was relatively high among children with and without asthma (75.0 and 41.1%, respectively). This may be due to the oversampling of children with allergic mothers in the study design. The high prevalence of BHR could also imply that the cutoff at a cumulative dose not exceeding 0.61 mg of methacholine bromide was not strict enough for our population.
A potential limitation of the exposure assessment might be that exposure was defined as ambient air pollution at the participants' residential address and that exposure at nonresidential addresses such as day-care centers or schools, where participants regularly spend part of their time, was not included. Validation studies comparing personal exposures with exposures estimated from land-use regression models are scarce. A study among 62 pregnant women in Vancouver, Canada found a moderate correlation (r = 0.49) between short-term (48-h) personal exposure measurements of NO and NO2 exposures estimated by land-use regression (25). The correlation was stronger in the subset of women who spent more time at home. In our birth cohort of approximately 4,000 children, personal exposure assessment was not feasible and no information was available about time–activity pattern and exposure at places other than home. However, given the age of our cohort, we believe that participants spent sufficient amounts of their time at home to yield valid exposure estimates from the land-use regression models. Apart from the sensitivity analyses on children who did/did not move house, we were not able to account for changes in residential address. This might have resulted in some measurement error, leading to attenuation of effect estimates, as indicated by the higher effects in children who did not move to another home compared with children who did move. However, in a birth cohort study from Cincinnati exposure to diesel exhaust particles was estimated by a land-use regression model as a time-weighted average of the estimated exposure at all addresses where the child spent more than 8 hours per week, that is, taking into account day-care addresses and changes in residential addresses due to moving (26). Overall, the differences between the estimated exposure at the birth record address and the time-weighted average exposure during the first 3 years of life were found to be small.
We presently know little about the relevance of the timing of the exposure in addition to the level of exposure, and it is unclear whether early life exposure at a time when the lungs are considered to be more susceptible is more important than later exposure. The estimated air pollution level at the birth address can be considered a good estimate of early life exposure for basically all participants of this study (only 7.3% of the total population moved during the first year of life). However, for children who moved house, the estimated air pollution level at the birth address is probably a poor estimate of their current exposure. The fact that effects were stronger in movers compared with nonmovers may indicate that the effects of air pollution are not due solely to early life exposure, but result at least partly from current exposure and that the relevance of early versus current exposure is not that clear. More research on this topic is needed. Moreover, the present results illustrate the impact of measurement error when relying on birth addresses alone and underline the importance of accounting for moving in this type of study.
For children who lived at their birth address during the entire follow-up, exposure might also have changed due to overall (long-term) temporal changes in air pollution levels. To investigate the long-term validity of the Traffic-related Air Pollution on Childhood Asthma (TRAPCA) model, in 2007, we went back to the original TRAPCA sites and performed four 1-week measurements (one measurement per season) of NO2. Results indicated that the original TRAPCA model was highly predictive (R2 = 0.80) of NO2 concentrations measured at the same sites almost 10 years later. A detailed description of these measurements will be published elsewhere. Further support for the validity of the estimated air pollution levels for children who did not move house comes from the National Air Quality Monitoring Network: annual average NO2 and PM10 levels were relatively stable between 2000 and 2007 (27).
An important issue regarding the present findings concerns the issue of pollutant-specific effects, that is, which (set of) pollutant(s) is responsible for the observed effects. As in our previous analyses (6, 19) it was not possible to disentangle the effects of the specific pollutants that were measured because of the high correlations among the different pollutants. In outdoor air, NO2 often is highly correlated with other combustion products, in particular fine particulate matter (28). An expert panel reviewed the biological plausibility of epidemiological findings on criteria pollutants (29). Taking into account findings from epidemiological, human clinical, and toxicological studies, the experts concluded that NO2 may be acting as a surrogate for a mixture of pollutants.
We conclude from our results that exposure to traffic-related air pollution may cause asthma in children monitored from birth to 8 years of age.
The authors thank all the children and their parents for their cooperation. The authors also thank all the field workers and laboratory personnel involved for their efforts, and Ada Vos for data management.
| 1. | Trasande L, Thurston GD. The role of air pollution in asthma and other pediatric morbidities. J Allergy Clin Immunol 2005;115:689–699. |
| 2. | Sarnat JA, Holguin F. Asthma and air quality. Curr Opin Pulm Med 2007;13:63–66. |
| 3. | Brunekreef B, Sunyer J. Asthma, rhinitis and air pollution: is traffic to blame? Eur Respir J 2003;21:913–915. |
| 4. | Heinrich J, Wichmann HE. Traffic related pollutants in Europe and their effect on allergic disease. Curr Opin Allergy Clin Immunol 2004;4:341–348. |
| 5. | Gilmour MI, Jaakkola MS, London SJ, Nel AE, Rogers CA. How exposure to environmental tobacco smoke, outdoor air pollutants, and increased pollen burdens influences the incidence of asthma. Environ Health Perspect 2006;114:627–633. |
| 6. | Brauer M, Hoek G, Smit HA, de Jongste JC, Gerritsen J, Postma DS, Kerkhof M, Brunekreef B. Air pollution and development of asthma, allergy and infections in a birth cohort. Eur Respir J 2007;29:879–888. |
| 7. | Morgenstern V, Zutavern A, Cyrys J, Brockow I, Koletzko S, Kramer U, Behrendt H, Herbarth O, von Berg A, Bauer CP, et al. Atopic diseases, allergic sensitization, and exposure to traffic-related air pollution in children. Am J Respir Crit Care Med 2008;177:1331–1337. |
| 8. | Nordling E, Berglind N, Melen E, Emenius G, Hallberg J, Nyberg F, Pershagen G, Svartengren M, Wickman M, Bellander T. Traffic-related air pollution and childhood respiratory symptoms, function and allergies. Epidemiology 2008;19:401–408. |
| 9. | McConnell R, Berhane K, Gilliland F, London SJ, Islam T, Gauderman WJ, Avol E, Margolis HG, Peters JM. Asthma in exercising children exposed to ozone: a cohort study. Lancet 2002;359:386–391. |
| 10. | Ryan PH, LeMasters GK, Biswas P, Levin L, Hu S, Lindsey M, Bernstein DI, Lockey J, Villareal M, Khurana Hershey GK, et al. A comparison of proximity and land use regression traffic exposure models and wheezing in infants. Environ Health Perspect 2007;115:278–284. |
| 11. | Andersen ZJ, Loft S, Ketzel M, Stage M, Scheike T, Hermansen MN, Bisgaard H. Ambient air pollution triggers wheezing symptoms in infants. Thorax 2008;63:710–716. |
| 12. | Shima M, Nitta Y, Ando M, Adachi M. Effects of air pollution on the prevalence and incidence of asthma in children. Arch Environ Health 2002;57:529–535. |
| 13. | Gauderman WJ, Avol E, Lurmann F, Kuenzli N, Gilliland F, Peters J, McConnell R. Childhood asthma and exposure to traffic and nitrogen dioxide. Epidemiology 2005;16:737–743. |
| 14. | Martinez FD, Wright AL, Taussig LM, Holberg CJ, Halonen M, Morgan WJ; Group Health Medical Associates. Asthma and wheezing in the first six years of life. N Engl J Med 1995;332:133–138. |
| 15. | Kurukulaaratchy RJ, Matthews S, Arshad SH. Defining childhood atopic phenotypes to investigate the association of atopic sensitization with allergic disease. Allergy 2005;60:1280–1286. |
| 16. | Brunekreef B, Smit J, de Jongste J, Neijens H, Gerritsen J, Postma D, Aalberse R, Koopman L, Kerkhof M, Wijga A, et al. The Prevention and Incidence of Asthma and Mite Allergy (PIAMA) birth cohort study: design and first results. Pediatr Allergy Immunol 2002;13:55–60. |
| 17. | Gehring U, Brauer M, Fischer P, de Jongste JC, Kerkhof M, Oldenwening M, Wijga A, Brunekreef B. Traffic-related air pollution and the development of asthma during the first 8 years of life: the PIAMA Study [abstract]. Epidemiology 2009;20:S34. |
| 18. | Brauer M, Hoek G, van Vliet P, Meliefste K, Fischer P, Gehring U, Heinrich J, Cyrys J, Bellander T, Lewne M, et al. Estimating long-term average particulate air pollution concentrations: application of traffic indicators and geographic information systems. Epidemiology 2003;14:228–239. |
| 19. | Brauer M, Hoek G, van Vliet P, Meliefste K, Fischer PH, Wijga A, Koopman LP, Neijens HJ, Gerritsen J, Kerkhof M, et al. Air pollution from traffic and the development of respiratory infections and asthmatic and allergic symptoms in children. Am J Respir Crit Care Med 2002;166:1092–1098. |
| 20. | Diggle P, Liang K, Zeger SL. Analysis of longitudinal data. Oxford: Clarendon Press; 1994. |
| 21. | Hosmer DW, Lemeshow S. Applied logistic regression. New York: Wiley; 1989. |
| 22. | Brauer M, Gehring U, Brunekreef B, de Jongste J, Gerritsen J, Rovers M, Wichmann HE, Wijga A, Heinrich J. Traffic-related air pollution and otitis media. Environ Health Perspect 2006;114:1414–1418. |
| 23. | Chen CM, Sausenthaler S, Bischof W, Herbarth O, Borte M, Behrendt H, Kramer U, Williams HC, Wichmann HE, Heinrich J. Perinatal exposure to endotoxin and the development of eczema during the first 6 years of life. Clin Exp Dermatol (In press). |
| 24. | Gehring U, Cyrys J, Sedlmeir G, Brunekreef B, Bellander T, Fischer P, Bauer CP, Reinhardt D, Wichmann HE, Heinrich J. Traffic-related air pollution and respiratory health during the first 2 yrs of life. Eur Respir J 2002;19:690–698. |
| 25. | Nethery E, Leckie SE, Teschke K, Brauer M. From measures to models: an evaluation of air pollution exposure assessment for epidemiological studies of pregnant women. Occup Environ Med 2008;65:579–586. |
| 26. | Ryan PH, LeMasters GK, Levin L, Burkle J, Biswas P, Hu S, Grinshpun S, Reponen T. A land-use regression model for estimating microenvironmental diesel exposure given multiple addresses from birth through childhood. Sci Total Environ 2008;404:139–147. |
| 27. | Beijk R, Mooibroek D, Hoogerbrugge R. Air quality in The Netherlands 2007. RIVM report 680704005. 2008. National Institute for Public Health and the Environment, Bilthoven, The Netherlands. |
| 28. | Sarnat JA, Schwartz J, Catalano PJ, Suh HH. Gaseous pollutants in particulate matter epidemiology: confounders or surrogates? Environ Health Perspect 2001;109:1053–1061. |
| 29. | Brown JS, Graham JA, Chen LC, Postlethwait EM, Ghio AJ, Foster WM, Gordon T. Panel discussion review: session four—assessing biological plausibility of epidemiological findings in air pollution research. J Expo Sci Environ Epidemiol 2007;17:S97–S105. |
