BACKGROUND: In a recent population-based case-control study using 2,400 cases of childhood cancer, we found a statistically significant association between residential radon and acute lymphoblastic leukemia risk. HYPOTHESIS: Traffic exhaust in the air enhances the risk association between radon and childhood leukemia. METHODS: We included 985 cases of childhood leukemia and 1,969 control children. We used validated models to calculate residential radon and street NO(x) concentrations for each home. Conditional logistic regression analyses were used to analyze the effect of radon on childhood leukemia risk within different strata of air pollution and traffic density. RESULTS: The relative risk for childhood leukemia in association with a 10(3) Bq/m(3)-years increase in radon was 1.77 (1.11, 2.82) among those exposed to high levels of NO(x) and 1.23 (0.79, 1.91) for those exposed to low levels of NO(x) (p(interaction,) 0.17). Analyses for different morphological subtypes of leukemia and within different strata of traffic density showed a non-significant pattern of stronger associations between radon and childhood leukemia within strata of higher traffic density at the street address. INTERPRETATION: Air pollution from traffic may enhance the effect of radon on the risk of childhood leukemia. The observed tendency may also be attributed to chance.