Although adverse respiratory health outcomes from exposure to ambient air pollution are biologically plausible, research linking exposure to asthma has been inconclusive 12. Recent research has emphasized the growing contribution and heightened toxic potential of traffic-related air pollution (TAP) near major vehicular corridors 3, as well as significant associations between exposure to TAP and onset of asthma 4. Other studies have found positive, significant associations between exposure to TAP and adverse respiratory outcomes 5678910111213, while others have reported null associations 141516.

The inconsistencies in linking TAP and asthma may be due to exposure measurement error in some studies, which arise partly from the way exposures to traffic pollution are estimated and derived. These exposure estimates include: self-reported traffic density at residence 1112; number of cars passing by per 24 hours on the nearest street to a home or school 71718; distance between the nearest street and home 8916171920; identification of the street with highest traffic density relative to a child's school or home 1021; perception of residential nuisances related to traffic pollution 22; indices which combine traffic and distance 142324; cumulative exposure indices 2526; and estimation of pollution exposure at the home using geographic information systems (GIS) and land use regression models 27.

Susceptibility factors have also been suggested to contribute to these observed inconsistencies, including early life exposure 28, duration of residence, parental asthma history, and gender 2029. Additionally, Douwes et al30 suggest there to be a growing importance of investigating specific subtypes of asthma, namely non-atopic or non-allergic asthma. Nystad et al31 show increasing rates of children with non-atopy related asthma, defined as asthma without atopic diseases (specifically hayfever or eczema), compared to atopy-related asthma. Ronmark et al. 32 report different patterns of risk factors for atopic and non-atopic asthma, defined as asthma with or without at least one positive skin test for type-1 allergy. The separation of atopic and non-atopic asthma is often not considered in the air pollution and asthma literature, and as non-atopic asthma is thought to contribute much of the increased incidence of asthma, it may be accentuating observed inconsistencies in reported results. Further to these complexities, differences exist between the sexes, with girls being more susceptible to asthma than boys 2033; despite boys having higher prevalence rates. Age also seems to influence the progression toward onset 34.

In this article, we examine the relationship between within-city or 'intraurban' contrasts in air pollution exposure and childhood asthma in Hamilton, Canada. Further, we test these associations within asthmatic subgroups stratified by the presence or absence of other atopic diseases, gender and age to determine whether these susceptibility factors influence the relationship between air pollution and asthma.

The prevalence rates of all asthma and asthma without hayfever are shown in Table 1. With an overall prevalence of 18%, boys had higher prevalence for asthma (21%) than girls (15%), consistent with other Canadian studies on the life course of asthma 66. The difference between the two sexes was more apparent in the younger age group, with boys having higher rates of asthma without hayfever than girls. Prevalence rates for asthma ever were similar to the larger study from which our population was drawn, but our sub-population had slightly lower rates of wheezing ever for the younger children (data not shown).

In total, sixteen exposure surfaces were created. These included the surfaces derived using the Theissen polygon method for O₃, NO_x, and SO₂, and a splined surface developed for PM₁₀. Figures 2, 3 and 4 show the surfaces for PM₁₀Spline, O₃Theissen and NO₂LUR. All surfaces, except O₃, showed a pollution gradient within the city that followed the expected trend of higher intensities in the northeast near the industrial core, and decreasing pollution levels towards the outskirts of the city. The O₃ surface followed the opposite pattern, with lower levels in the downtown area and higher levels at the edges of the city.

The different pollution metrics reflect both different sources and different approaches to modelling exposure. The SO₂Theissen polygon surface indicates the presence of point-source industries in the northeast end of downtown Hamilton. The Theissen polygon surfaces created discrete categories of pollution levels equal to the measurements obtained at the fixed MOE sites. As these are not smooth continuous surfaces, they may not accurately reflect the real variation of pollution – an inherent unavoidable characteristic of creating such polygons around the monitoring locations 67. The O₃Theissen surface had a similar categorizing effect. The PM₁₀Spline, on the other hand, may have over-smoothed the true variation of pollution, again, due to the nature of this interpolation technique. Both the kriged and LUR NO₂ surfaces were based on a denser network of monitors within the city. With the large variation in concentrations measured by the monitors, the kriging methodology was not able to capture the full spatial variation without incorporating some unavoidable errors included in the estimation. The highest errors, however, tended to be outside the area encompassing the children's residence locations. Visually, the NO₂LUR surface appeared most heterogeneous, with the highest variation occurring around roads and densely populated areas of the lower city.

The concentrations of estimated pollutants were then assigned to the postal code of each child's residential address. Pollution exposures in each group were quite similar for most pollutants (Table 2) with the exception of smaller ranges of PM₁₀ exposure for girls and NO₂LUR exposures for boys. A correlation matrix was constructed for the independent variables (see Additional file 1). The pollutants had low correlations, except for O₃ with NO_x, which were inversely related due likely to the scavenging effect of ozone by local sources of NO 68. As expected, the two measures of NO₂ were correlated. The DI had a weak positive correlation with the pollutants, except with O₃ where there was a weak negative correlation. Dwelling value and average income were highly correlated (r = 0.73) and followed very similar patterns in their correlations with pollutants. The rate of repair and percent old houses were also highly correlated (r = 0.72). To avoid introducing multicollinearity, only one of the two variables in each correlated set was retained for the multivariate analysis. DI and smoking did not have strong associations with the other variables, and thus were kept for further testing in the multiple regression models.

Bivariate logistic regression revealed positive, but insignificant, associations between pollution exposures and asthma outcomes when the whole population was tested. A detailed table is available in the online appendix (see Additional file 2). The odds ratio (OR) for asthma with NO₂LUR, for example, was 1.02 per ppb (95% CI, 1.00–1.04) but the association was insignificant. Samples were stratified based on literature suggesting that differences exist between the sexes 69, that rates differ by age, and that asthma and asthma without hayfever have different risk factors 31. When testing the predictive potential of atopy related status for asthma, a positive association was found; namely that children with hayfever symptoms were more likely to have asthma than those without hayfever symptoms (OR = 3.03; 95%CI, 2.20–4.17). After testing interactions between the pollutants, atopy and subgroups, we found effects suggestive of an interaction between hayfever and pollutants in all girls for NO₂LUR (p = 0.156). The power to test for interactions in epidemiological studies is often poor, resulting in researchers missing important interactions due to lower power 70. As noted by Selvin 71, relaxing the type 1 error p value from the traditional 5% to 20% is a common approach in epidemiological studies, one that can allow for interaction tests in studies that are not powered for effect modification. In this instance, we had substantive reasons to test for interaction, and the sub-group analysis indicates that girls are more susceptible than boys. Given this, the literature of subgroup interaction and the empirical evidence in our data, we subsequently stratified the sample into subgroups by age, sex, and by age and sex. It is important to note that all subgroups were investigated; however, due to the large number of tested associations between the subgroups, and the voluminous resulting tables, only the significant results for the susceptible groups are included in this paper to avoid detracting from the main findings by listing tables of insignificant effects.

Tables 3, 4, 5 and 6 show the associations for the stratified analysis conducted for the non-atopy related asthma population within the subgroups. Asthma without hayfever was associated with NO₂LUR for all girls and older girls. We also ran trivariate logistic regressions on the significant associations identified in the bivariate tests for asthma without hayfever (see Table 7). The effects of pollutants remained robust. NO₂LUR retained significance with asthma without hayfever in all girls for each confounding variable. For the subpopulation of older girls, the odds ratios were generally stronger after adjustment for potential compositional and contextual variables. We used these regressions to test for the effect of the compositional and contextual variables on the percent change in the regression coefficients of the models. Testing the most robust group of associations (NO₂LUR and all girls) identified the deprivation index (DI) and rate of repair as the variables that reduced the regression coefficient by more than 10%. These were retained as the confounding variables for the multiple variable logistic regressions.

We also tested the effect of co-pollutants on our models (see Table 8). For the populations of all girls and older girls, the effect of NO₂LUR was larger after adjusting for SO₂, PM₁₀ and O₃. Calculated for a 1-unit increase in NO₂, the odds ratio for asthma without hayfever among all girls was 1.46 times (after controlling for PM₁₀, SO₂, O₃, DI and rate of repair), and 2.71 times greater among older girls.

Sensitivity analyses were conducted with a generalized linear model (GLM) with a natural spline smoother 65. The coefficient changed slightly from 0.128 to 0.130 when the smoother was applied to 10 degrees of freedom (df), and to 0.129 for 20 df (p < 0.05). Using the natural spline smoother, increasing spans indicated a more localized analysis. This sensitivity analysis lends further support to the notion that confounding probably does not bias the coefficients as the effects were robust.

For further sensitivity analysis, we tested the alternative indicators of atopic conditions (eczema and runny nose not associated with a cold) to assess whether the selection of indicator made a difference in the air pollution and asthma relationship. The results were sensitive to the selection of other indicators of atopic conditions, as they were positive for the most part, but were no longer significant.

We also assessed the sensitivity of the association of wheeze ever and current wheeze with the same symptom indicators of atopic conditions, and the pattern of effects was similar to that observed for asthma. After controlling for confounders and copollutants, NO₂LUR remained significant with wheeze ever without hayfever (OR = 1.13, 95%CI, 1.01–1.23) and current wheeze without hayfever (OR = 1.28, 95%CI, 1.06–1.55) for all girls (OR = 1.15, 95%CI, 1.00–1.31) and older girls (OR = 1.35, 95%CI, 1.10–1.66).

There were no significant associations between any of the exposure estimates and asthma in the whole population, but large effects were detected the subgroup of children without hayfever (predominately in girls). More specifically, after controlling for confounders we observed significant associations between NO₂LUR and non-atopy related asthma in all girls and older girls. The NO₂LUR surface provided the only robust associations with all girls and older girls after running the co-pollutant models and GLM sensitivity analyses. Sensitivity analyses for the definition of respiratory symptoms without hayfever showed that using wheeze instead of asthma as a respiratory health effect still revealed similar findings; however, if alternative indicators of atopic disease (eczema or runny nose without a cold) were used, the effects were not robust.

The sizes of the effects found in the asthma without hayfever models are significant. For older girls, the OR is 3.46 (95% CI, 1.19–10.07) for the NO₂LUR (calculated for mean-minimum pollution). Taking the exposure coefficients from the co-pollutant models brings the estimates closer to OR = 2.98 (95% CI, 0.98–9.02) for older girls and to OR = 1.85 (95% CI, 0.92–3.73) for all girls using the same exposure contrast. The high collinearity between the pollutants probably contributed to the wider confidence intervals in the co-pollutant models. An analysis among a sub-sample with home-based exposure measurements from the Southern California Children's Health Study (CHS) revealed an association between asthma prevalence and home and outdoor NO₂ of OR = 1.83 (95% CI, 1.04–3.22) per increase of one IQR (5.7 ppb) in exposure 72. Thus, while both studies observed similarly sized effects, our effects were only within the asthma subgroup without hayfever.

Other researchers have also commented on the relevance and importance of non-atopy related respiratory symptoms. Heinrich and colleagues 73 evaluated TAP exposure using self-administered subjective questionnaires assessing traffic intensity in a population of 6896 adults. High traffic intensity increased the risk for non-allergic asthma, but not for atopic symptoms including allergic sensitization. Non-allergic asthma in this study was identified as having current asthma but a negative screening assay for specific sensitizations to mite, cat, dog, pollen and fungal allergens. Douwes et al. 30 highlighted the importance of investigating non-allergic asthma, as allergic asthma contributes to approximately 50% of asthma cases; thus, much of the increase in asthma incidence may been attributed to non-allergic asthma. Romanet-Manent and colleagues' 74 estimation of the contribution of non-allergic asthma is slightly lower, at 10–30% of asthmatics. Clinical differences between atopic and non-atopic asthma were initially identified by Rackeman et al. 75 to include key factors such as gender, increasing age, increased severity of lung function, and chronic rhinosinusitis. Other recent studies support this earlier differentiation of asthma, placing increasing emphasis on the genetic investigations of each subtype 76.

Douwes et al. 30 also hypothesize that exposures to bacterial endotoxins and air pollutants may induce non-allergic or non-atopic asthma. Nystad et al. 31 suggest that increases in exposure and adjuvant factors may be contributing to the asthmatic response without triggering atopic symptoms, such as hayfever and eczema. Non-allergic asthma is generally associated with increased neutrophil and interleukin-8 levels 77, but the biological mechanisms in which NO₂ causes the non-atopic effects to occur are not as clear. Particulate pollution has shown to induce neutrophilic air inflammation 78. Seaton and Dennecamp 79 propose that strong correlations between NO₂ and ultrafine particles (UFP = particulate matter less than 0.1 micrometre in diameter) may explain why NO₂ would appear significant in health studies.

In the most stringent analysis controlling for confounders and co-pollutants, effects were observed in all girls and older girls and only for the NO₂LUR model, a result consistent with recent findings from the CHS cohorts in Southern California 20. Female sex has shown to increase the risk of a non-allergic type of asthma in an adult population 74 although no mechanism for this difference was suggested. Gold et al. 80 have suggested that gender differences in asthma rates might be due to differences inherent in the mechanical properties of the lung and inflammatory responses. Alternatively, Venn et al. 81 proposed that hormonal changes occurring in early puberty may affect prevalence rates, as well as differential exposures to triggers for wheeze or asthma, such as smoking. Berhane et al. 82 have found that duration and age of onset of asthma differs between the sexes, thus having differential impacts on lung function. There may also be additional factors influencing exposure times to pollution levels that we were not able to account for in this study.

There are limitations in this study that warrant comment. We had few individual-level covariate data for the children; thus, we created proxies for true individual factors by the use of information available at the census tract level. While contextual level data is important 83, lack of control for the individual factors, such as smoking at home, which has shown to impact the respiratory health of children, may have biased our results1684. The sensitivity analysis with the GLM indicated that some confounding might have remained, possibly due to individual level factors with spatial attributes not captured within our proxy variables. The use of the GLM model has, however, reduced this possibility as the removal of spatial structures in the data reduces residual confounding.

Non-atopic asthma is widely-accepted to be identifiable from the absence of positive skin prick tests for IgE mediated sensitizations to common allergens which include dust mites, pet and particular foods 30. In the absence of this objective evidence relative to atopic indicators, we had to rely on associated symptoms. Upton et al. 85 suggest that clinicians and epidemiologists often use other allergic manifestations such as hayfever to identify atopy when the skin test data is unavailable; indeed they use that same definition to identify 20 year intergenerational trends in prevalence of asthma and hayfever. They do, however, emphasize that the difference between atopic and non-atopic asthma cannot be validated solely on the presence or absence of hayfever. Ronmark et al. 32 defined atopic asthma as asthma with at least one positive skin prick allergy test, while Romanet-Manent et al. 74 defined this as allergic asthma. To avoid adding to these differences and complications, we chose to report our outcome as asthma without hayfever, rather than non-atopic or non-allergic asthma, and acknowledge the fact that we cannot be certain of the atopic/non-atopic difference with the available data.

Since the ISAAC study took place in 1994 and 1995, and the monitoring for the kriging and LUR modelling was conducted in 2002, temporal discontinuity exists between the health outcomes and both of our NO₂ surfaces. In spite of this potential limitation, research shows that there has been little change in pollution levels within this time period, and, more importantly, that the spatial variation of pollution has not changed substantially from 1995 to 2002 42. Hamilton's distinct topography (especially the Niagara Escarpment), downtown industrial core, and prevailing wind direction contribute to a consistent spatial pattern of pollution that has long been known to decrease from the industrial core outwards to the rest of Hamilton 86. Similar to European findings 87, although pollution levels may fluctuate, the spatial pattern of NO₂ appears consistent over time.

We found significant associations between exposure to modeled NO₂ and asthma without hayfever outcomes in children living in Hamilton. Girls with asthma without hayfever, and particularly older girls, were most susceptible to the effects of NO₂ or a closely associated co-pollutant. The effects were sensitive to the method of exposure estimation, and more refined exposure models produced the most robust associations. Given the potential role of non-atopic asthma, further studies should include objective determination of atopic status using skin allergy tests or serum IgE levels. Since NO₂ may proxy for other pollutants that can cause adverse health outcomes such as ultrafine particles 79, further research is needed on the identification and measurement of which pollutants may be associated with the observed effects.

CHS: Southern California Children's Health Study; CI: confidence interval; df: degrees of freedom; DI: deprivation index; DMAP: Distance Mapping and Analysis Program; DMTI: Desktop Mapping Technologies Inc; GLM; generalized linear model; IDW: inverse distance weighted; IgE: serum immunoglobulin E; IQR: interquartile; ISAAC: International Study of Asthma and Allergies in Childhood; LUR: land use regression; MOE: Ministry of the Environment; O3: ozone; ON: Ontario; OR: odds ratio; PCA: principal components analysis; ppb: parts per billion; PM10: particulates of 10 micrometers or less; TAP: traffic-related air pollution; UFP: ultrafine particles; USA: United States of America.

The authors declare that they have no competing interests.

TS participated in study design, conducted the analysis and drafted the paper. MRS provided the ISAAC data, and with RC contributed to expertise and interpretation of data. NF participated in pollution surface creation. AA and BN were both involved in data interpretation and initial study design. RB provided statistical guidance and expertise. MJ participated in study conception and design, data analysis and interpretation, and in writing the paper. All authors have contributed to the overall direction of the research, have made editorial comments, and have read and approved the final manuscript.