A statistical association between childhood leukaemia and an abdominal X-ray examination of the pregnant mother was first reported in 1956 from a case-control study of childhood cancer mortality conducted in Great Britain. This study, later called the Oxford Survey of Childhood Cancers (OSCC), was continued and eventually showed a highly statistically significant approximately 50% proportional increase in the risk of childhood leukaemia associated with antenatal diagnostic radiography. The association has been confirmed by many case-control studies carried out around the world, the appropriately combined results of which show a highly statistically significant increase in risk that is compatible with the OSCC finding. There is no doubt about the reality of the statistical association, but a causal interpretation has been questioned. On balance, however, the evidence points to low-level irradiation of the fetus increasing the risk of leukaemia in childhood, with an excess relative risk coefficient of around 50 Gy(-1) (equivalent to an excess absolute risk coefficient of about 3% Gy(-1)), although the uncertainty associated with these coefficients is considerable and they are likely to be overestimates. In contrast to exposure in utero, the evidence from case-control studies for an association between childhood leukaemia and postnatal exposure to medical diagnostic irradiation is equivocal and sometimes conflicting. Since standard radiation risk models predict that low-level exposure in the early years of life should produce an increased risk of childhood leukaemia that is roughly similar to that arising from fetal exposure, this absence of persuasive evidence is likely to be due to various problems with the studies. This is unfortunate given the rise in relatively high dose diagnostic procedures (e.g. paediatric CT scans) that would be predicted to materially increase the relative risk of childhood leukaemia.