Childhood leukaemia and low-level radiation - are we underestimating the risk?

The Seascale childhood leukaemia 'cluster' can be interpreted as indicating that the risk of childhood leukaemia arising from low-level exposure to ionising radiation has been underestimated. Indeed, several variants of such an interpretation have been advanced. These include exposure to particular radionuclides, an underestimation of the radiation risk coefficient for childhood leukaemia, and the existence of a previously unrecognized risk of childhood leukaemia from the preconceptional irradiation of fathers. However, the scientific assessment of epidemiological associations is a complex matter, and such associations must be interpreted with caution. It would now seem most likely that the Seascale 'cluster' does not represent an unanticipated effect of the exposure to ionising radiation, but rather the effect of unusual population mixing generated by the Sellafield site which has produced an increase in the infection-based risk of childhood leukaemia. This episode in the history of epidemiological research provides a timely reminder of the need for great care in the interpretation-of novel statistical associations. (author)