There is a strong body of evidence that traffic-related air pollutants aggravate asthmatic symptoms. Short-term exposure to air pollutants has been associated with e.g. peak flow values 1, daily symptoms 2, anti-asthmatic medication 3, emergency visits 4 and hospital admissions 5. However, exposure to traffic exhaust may also induce onset of asthma and asthma-like symptoms in healthy individuals. Oxidative stress has been suggested as the major, underlying mechanism behind many of the toxic reactions induced by air pollutants 6. Childhood asthma is associated with decreased levels of various components in the antioxidant defences 7. Reactive oxygen species (ROS) and free radicals are generated by traffic related pollutants such as nitrogen dioxide (NO₂) and particulates. Reduced glutathione, ascorbic acid, uric acid, tocopherol and other antioxidants in the airways counteract the effects of ROS. Oxidative stress arises when the neutralising antioxidant defence is overwhelmed. An inflammatory response involves an endogenous production of additional ROS and further inflammation. High levels of ROS cause a depletion of local antioxidants and lead to widespread inflammatory reactions also outside the target tissues.

There is increasing evidence mainly from experimental studies that exposure to diesel exhaust particles also has immunologic effects influencing the development of allergic sensitization. Several animal and human in vivo and in vitro studies strongly suggest that diesel exhaust particles act as adjuvants and augment the allergic reaction 8. Combined exposure to diesel exhaust particles and antigen could switch the immune response towards IgE 9. A recent animal study indicated that exposure to diesel exhaust particles induces epigenetic changes 10. Concomitant exposure to ambient diesel exhaust particles and antigen in mice altered the methylation of T helper genes. Changes in the methylation affect the differentiation of the T helper cells and thereby possibly also the risk of allergic sensitization and asthma.

It has been argued that children are more susceptible to adverse effects from traffic exhaust 11. Exposure to air pollutants early in life may have long-lasting effects 12. The newborn child has an immature immune system and the respiratory epithelium in the developing lung has an increased permeability. A tenfold rise in the number of alveoli occurs during the first four years of life. The development of the lungs is not finished until adulthood 13. The ratio of lung surface area to body weight is much larger in children than in adults with a huge difference in air intake per minute in relation to body weight and alveolar surface area between children and adults 14. In addition, children tend to be more active than adults and spend more time outdoors.

Misclassification of exposure to pollutants may contribute to weak associations and inconsistent findings in population studies. In previous epidemiological research focusing on the associations between traffic pollution and asthma, allergies or sensitisation, investigators have used measured indicator pollutants as NO₂ from stations representing the surrounding area or various surrogates to assess exposure. In the category of surrogates we find self-reported or measured traffic density in buffer zones around the home address, and self-reported or measured distance of the home to the nearest street.

More recently, exposure at the home address has also been estimated using dispersion models or land use regression models. While atmospheric dispersion models require meteorological information and emission data (i.e. emission factors for vehicle fleet), the land use regression models explain existing contrasts in monitored concentrations using, for example, road type and population density. In many cases the concentration values used come from monitoring campaigns designed to meet the needs of the modelling. A high temporal resolution is not important, but the monitoring sites must be sufficient to reflect the full range of conditions in the study population.

Also for traffic pollutants particle concentration is usually studied as particulate matter (PM) without any chemical specification. PM₁₀ reflects the mass concentration of particles less than 10 microns and PM_2.5 of particles less than 2.5 microns. Other common indicators of traffic exhaust are NO₂ and nitrogen oxides (NO_x), where the fraction emitted as NO₂ tends to increase due to new diesel engines with oxidative catalysts.

Misclassification of the outcome is often a concern in questionnaire-based studies of respiratory disease, particularly in young children with less distinctive symptoms. Lower respiratory illness and wheezing symptoms are frequent in early childhood but very few of these children develop asthma. Moreover, childhood asthma is not a homogeneous disorder and different phenotypes of wheeze and asthma have been identified based on prognosis, atopic status and lung function 15. Sensitization is assessed by objective measurements and misclassification is therefore a minor problem. Sensitization to inhalant allergens predominates after infancy 1617. Allergic sensitization is common in asthma and early sensitization predicts persisting symptoms 18.

A comprehensive review from WHO was presented in 2005. Rather substantial evidence suggested that the inflammatory processes associated with exposure to traffic related pollutants contributed to an increased risk of non-allergic respiratory symptoms. Antioxidants and possibly surfactant components could be important determinants of individual susceptibility 1920. However, it was less clear whether exposure to traffic exhaust may induce allergic sensitization. The evidence from population based studies has been weak and the findings have not been consistent as summarized in the WHO review 1920.

Several reports based on well-designed cohort studies using different forms of exposure modelling have been published after the WHO report. The aim of this review was to assess the evidence from recent prospective studies in children to support a contribution of long-term traffic pollution to the development of asthma-like symptoms and allergic sensitization.

We have reviewed original articles in English published since 2002 and found in PubMed in Oct 2008. We have only included cohort studies assessing potential effects of long-term exposure to traffic exhaust on respiratory health and allergic sensitization. In birth cohort studies prevalence has been seen as an equally important outcome variable as the incidence. Case-control studies and cross-sectional studies and studies on ozone have been excluded. We have also excluded studies on effects of short-term exposures.

The following search terms have been used in PubMed and Medline:

Traffic, air pollution, diesel particulate, asthma, sensitization, cohort, child

We have evaluated 15 papers based on 10 different cohort studies from Germany, the Netherlands, Sweden, Norway, the US and Japan, in all six birth cohorts and four prospective studies where the children were enrolled at or after entry at school. Most of these papers are too recent to have influenced the WHO review. The birth cohorts from Germany (Munich), the Netherlands and Sweden (Stockholm) participated in TRAPCA, an international collaboration on the impact of Traffic Related Air Pollution on Childhood Asthma 21. Several different types of exposure variables are used: contrasts in community level of pollutants 222324, measurements outside homes 25, distance to a large road or traffic flow 92627, concentrations outside homes according to atmospheric dispersion models with emission data 282930 and concentrations estimated with statistical models (land use regression models) including monitoring data 262731323334. Also for a specific type of exposure variable there are differences in how exposure is described, for example the averaging time used or concentration classes presented.

Data from all evaluated studies are summarized in Table 1 and 2. Studies are first by study country.

All epidemiological studies have certain limitations. We have in this review only included prospective studies implying that exposure to air pollutants was assessed before the development of symptoms. In all, 13 papers based on data from 9 cohorts have evaluated the relationship between traffic exposure and respiratory health. An increased risk of respiratory symptoms was demonstrated in all studies. The outcome varied, however, by the age of the child. With the limited number of studies with both the same exposure measure and the same definition of the health outcome, we found it too early to now perform a formal meta-analysis. Different types of exposure assessment have been used in these studies: traffic proximity (GIS), dispersion modelling, pollution measurements and regression models. Some studies seem to have a very high spatial resolution in exposure data, but it is difficult to compare these different approaches. However, internal validations show for example that the dispersion model in the Norwegian study 30 seems to perform less well than the model in the similar Swedish study 2829. Such differences may contribute to inconsistent findings.

The association between traffic-related air pollutants and respiratory health during infancy was assessed in three birth cohorts. In two of the large European birth cohorts, exposure to traffic exhaust was related to slightly increased odds ratios for coughing symptoms 33, asthmatic bronchitis 26, upper respiratory infections and doctor-diagnosed asthma 31 In the birth cohort form Cincinnati, distance to stop-and-go traffic 9 and increasing levels of traffic soot had a dose-response relationship with cough without wheeze during the first year of life 34. Wheeze and cough induced by respiratory infections are common in early childhood. Wheezing is not specific for asthma and many of the wheezers are symptom-free when they start at school 36.

Collaborative centres in the TRAPCA study assessed potential effects of traffic exhaust also at four 2935 or six years 27. The findings were fairly similar. There were some inconsistencies, however. In the Dutch cohort, exposure to traffic related pollutants was associated with an increased risk of doctor-diagnosed asthma and wheeze 35. In Munich, exposure to pollutants and distance to main road were associated with doctor-diagnosed asthma or asthmatic bronchitis whereas the association between traffic exposure and reported wheeze was less consistent 27. In the Swedish cohort, traffic exposure had an association with persistent wheeze but no association with doctor-diagnosed asthma 2829. Inconsistencies in wheezing and doctor diagnosed asthma between countries could possibly be related to discrepancies in the interpretation of the term "wheeze" and variability in diagnostic routines and labelling of symptoms.

Five studies assessed the outcome in schoolchildren. In the two Japanese cohorts, urban schoolchildren exposed to traffic related pollutants and/or living close to main road had an increased prevalence of asthma in comparison with rural children 2324. It could not be excluded, however, that a low prevalence of asthma was attributed to protective factors related to rural living. Two cohorts were recruited from the CHS in southern California. Lung function and exposure to ambient air pollutants appeared to have interactive effects on the development of asthma in one of the studies. Exposure assessment was less sophisticated when compared with the European studies and levels of pollutants were only defined as high or low 22. Residential exposure to NO₂ was assessed in the other study from California but this study had limitations due to the study size and included only 26 new cases of asthma over a period of eight years. In the cohort study from East Boston, the children were followed annually from birth up to 18 years. Long-term exposure to traffic exhaust and exposure to violence appeared to have synergistic effects on the risk of asthma 32. However, reporting of symptoms could be more likely in children exposed to violence. Control of confounding was limited also in this study due to the fairly small sample size.

Asthma is a heterogeneous disease and there are several phenotypes of childhood wheezing 37. Potential effects of traffic related air pollutants on different wheezing phenotypes were only assessed in one of the birth cohorts 2829. Adjustment for relevant covariates, e.g. socioeconomic status and parental education had certain shortcomings in three of the cohorts 232432. Unmeasured residual confounding could affect the outcome in any of the studies. Nevertheless, the consistency in the results indicates that traffic exhaust contributes to the development of respiratory symptoms in healthy children. This is supported by recent genetic association studies strongly suggesting that variability in the susceptibility to air pollutants is influenced by polymorphism in genes involved in airway inflammation and oxidative stress 3839. Type and timing of exposures regulate the expression of the genes 40. Glutathione-S-transferase is a catalyst of the conjugation of reduced glutathione and an important component in the protection against reactive oxygen species 41. An interaction has been demonstrated between glutathione-S-transferase P1 (GSTP1) gene polymorphism and outdoor air pollution on asthma among schoolchildren in Taiwan 42. Microsomal epoxide hydrolase (EPHX1) plays a role in the detoxification of polyaromatic hydrocarbons (PAHs). EPHX1 and GSTP1 variants affect the risk of asthma in children exposed to traffic exhaust 43. Transforming growth factor (TGF)-β1 is involved in airway inflammation and the TGF-β1 gene is upregulated by oxidant stress. Living close to highways in California was associated with an increased risk of lifetime asthma in children with certain variants in the TGF-β1 gene 43.

Potential effects of traffic exhaust on the development of allergic sensitization were only assessed in the four European birth cohorts. Long-term exposure to outdoor air pollutants had no association with sensitization in ten-year-old schoolchildren in Norway 30. In contrast, German, Dutch and Swedish preschool children had an increased risk of sensitization related to traffic exhaust despite fairly similar levels of outdoor air pollution as in Norway 272935. A strong exposure-response association was demonstrated in Germany and the Netherlands but not in Sweden, where a significant difference in sensitization was presented only for the fifth to ninety-fifth percentile difference in air pollution exposure within the cohort. Traffic related pollutants may induce the release of allergenic granules from grass pollen 44 and affect the morphology and possibly also the allergenicity of the pollen grains 45. Exposure to traffic-related pollutants was associated with sensitization to outdoor allergens in Germany and in Sweden. In the Netherlands, however, traffic exhaust was associated with sensitization only to food allergen. There is some evidence from experimental studies in mice that exposure to diesel exhaust could counteract the development of oral tolerance against food allergens 46.

In the Norwegian study, sensitization to single allergens in subpopulations suggested that confounding and selection bias could influence the outcome. In the Netherlands, sensitisation was assessed in a subgroup of children comprising 20% of the original cohort. This sample differed in many respects from the original cohort. Vulnerability to adverse effects from environmental factors could be related to genetic polymorphism and differ between individuals or subgroups within a population. Assessment of gene-gene and gene-environment interactions on sensitization has so far only been carried out in a subgroup of the Swedish birth cohort. Exposure to traffic-related air pollutants was associated with an increased risk of sensitization at four years in children with specific variants of GSTP1 and this association was modified by polymorphism in genes encoding TNF. Such findings strongly indicate that traffic-related effects on sensitization could be restricted to individuals with a specific genetic polymorphism. Harmful effects of environmental factors are diluted and difficult to detect in a large population where only some individuals are genetically susceptible 3847.

The findings from recent cohort studies indicate that traffic exhaust contributes to the development of respiratory illness in childhood. A growing body of evidence also suggests that traffic related air pollutants may induce sensitization. So far, however, very few cohort studies have assessed effects on sensitization and the findings are not consistent. Effects on sensitization could be difficult to detect if genetic susceptibility is uncommon in the study population. Ongoing cohorts should reassess effects of traffic exhaust on respiratory health and sensitization in older children. Further genetic association studies are required and may identify individuals vulnerable to adverse effects from traffic related pollutants. Future studies should also evaluate effects of traffic exhaust on the development and long term outcome of different phenotypes of asthma and wheezing symptoms.

CCAAPS: Cincinnati Childhood Allergy and Air Pollution Study; CHS: Southern California Children's Health Study; ECAT: Elemental carbon attributable to traffic sources; EPHX1: Microsomal epoxide hydrolase; GINI: German Infant Intervention Programme; GSTP1: Glutathione S-transferase P1; LISA: Influences of Lifestyle Related Factors on the Immune System and Development of Allergies in Children; NO_x: Nitrogen oxides; PIAMA: Prevention of Asthma and Mite Allergy; PM: Particulate matter; ROS: Reactive oxygen species; TNF: Tumour necrosis factor; TRAPCA: Traffic Related Air Pollution on Childhood Asthma.

The authors declare that they have no competing interests.

LB and BF have designed the study, reviewed the included papers and prepared the final manuscript together.