Background and aim: Acute leukemia (AL) affect approximately 500 children under 15 each year in France. Living near major roads and being exposed to traffic-related air pollutants are increasingly suggested as risk factors for childhood AL. The present GIS-based registry-based case-control study investigates the hypothesis that prenatal exposures to traffic-related pollutants increase the risk of AL, using exposure at the residence at birth as a proxy of prenatal exposure. Methods: the study included all the AL cases born and diagnosed in France during 2010-2015 (n=728) and 11,908 controls representative of the French births during the same period. Addresses of residence at birth were geocoded. Exposure indicators were the length of major roads within 150 m of the address, and the concentrations of NO2, fine particulate matter (PM2.5) and black carbon (BC) estimated by land use regression models. Odds-ratios were estimated for several exposure categories and under the hypothesis of a log-linear increase in childhood AL risk with increasing level of exposure, using unconditionnal logistic regression models (all AL) and polytomic logistic models (AL subtypes). The analyses were also stratified according to the degree of urbanization of the municipality. Results: childhood AL was not associated with major road length. Overall, PM2.5 and BC concentrations were associated with an increased risk of childhood AL (OR=1.3 [1.2-1.4] and OR=1.1 [1.0-1.2] per inter-quartile range increase, respectively). However, no positive association was observed for BC when the effect of BC was isolated from that of total PM2.5. Stratifying by urbanization status did not change the results for PM2.5 but the associations were weaker for BC especially in most urban areas. Non-significant positive associations were observed for NO2. Conclusions: our findings suggest a role of prenatal exposure to traffic-related air pollution in childhood AL. Keywords: Childhood leukemia; road traffic; air pollution; prenatal exposure