Ambient nitrogen dioxide is a widely available measure of traffic-related air pollution and is inconsistently associated with the prevalence of asthma symptoms in children. The use of this relationship to evaluate the health impact of policies affecting traffic management and traffic emissions is limited by the lack of a concentration-response function based on systematic review and meta-analysis of relevant studies. Using systematic methods, we identified papers containing quantitative estimates for nitrogen dioxide and the 12 month period prevalence of asthma symptoms in children in which the exposure contrast was within-community and dominated by traffic pollution. One estimate was selected from each study according to an a priori algorithm. Odds ratios were standardised to 10 μg/m3 and summary estimates were obtained using random- and fixed-effects estimates. Eighteen studies were identified. Concentrations of nitrogen dioxide were estimated for the home address (12) and/or school (8) using a range of methods; land use regression (6), study monitors (6), dispersion modelling (4) and interpolation (2). Fourteen studies showed positive associations but only two associations were statistically significant at the 5 % level. There was moderate heterogeneity (I2 = 32.8 %) and the random-effects estimate for the odds ratio was 1.06 (95 % CI 1.00 to 1.11). There was no evidence of small study bias. Individual studies tended to have only weak positive associations between nitrogen dioxide and asthma prevalence but the summary estimate bordered on statistical significance at the 5 % level. Although small, the potential impact on asthma prevalence could be considerable because of the high level of baseline prevalence in many cities. Whether the association is causal or indicates the effects of a correlated pollutant or other confounders, the estimate obtained by the meta-analysis would be appropriate for estimating impacts of traffic pollution on asthma prevalence.