This paper describes an ecological study investigating whether there is an excess incidence of acute leukaemia among children aged 0-14 years living in the vicinity of the nuclear sites in Belgium. Poisson regression modelling was carried out for proximity areas of varying sizes. In addition, the hypothesis of a gradient in leukaemia incidence with increasing levels of surrogate exposures was explored by means of focused hypothesis tests and generalized additive models. For the surrogate exposures, three proxies were used, that is, residential proximity to the nuclear site, prevailing winds and simulated radioactive discharges, on the basis of mathematical dispersion modelling. No excess incidence of acute leukaemia was observed around the nuclear power plants of Doel or Tihange nor around the nuclear site of Fleurus, which is a major manufacturer of radioactive isotopes in Europe. Around the site of Mol-Dessel, however, two- to three-fold increased leukaemia incidence rates were found in children aged 0-14 years living in the 0-5, 0-10 and the 0-15 km proximity areas. For this site, there was evidence for a gradient in leukaemia incidence with increased proximity, prevailing winds and simulated radioactive discharges, suggesting a potential link with the site that needs further investigation. An increased incidence of acute leukaemia in children aged 0-14 years was observed around one nuclear site that hosted reprocessing activities in the past and where nuclear research activities and radioactive waste treatment are ongoing.