CASE-CONTROL STUDY OF LEUKEMIA AND NON-HODGKINS-LYMPHOMA IN CHILDREN IN CAITHNESS NEAR THE DOUNREAY NUCLEAR INSTALLATION

Objective-To examine whether the observed excess of childhood leukaemia and non-Hodgkin's lymphoma in the area around the Dounreay nuclear installation is associated with established risk factors, or with factors related to the plant, or with parental occupation in the nuclear industry. Design-Case-control study. Setting-Caithness local government district. Subjects-14 cases of leukaemia and non-Hodgkin's lymphoma occurring in children aged under 15 years diagnosed in the area between 1970 and 1986 and 55 controls matched for sex, date of birth, and area of residence within Caithness at time of birth. Main outcome measures-Antenatal abdominal x ray examination; drugs taken and viral infections during pregnancy; father's occupation; father's employment at Dounreay and radiation dose; distance of usual residence from the path of microwave beams, preconceptional exposure to non-ionising radiation in the father; and other lifestyle factors. Results-No raised relative risks were found for prenatal exposure to x rays, social class of parents, employment at Dounreay before conception or diagnosis, father's dose of ionising radiation before conception, or child's residence within 50 m of the path of microwave transmission beams. Results also proved negative for all lifestyle factors except an apparent association with use of beaches within 25 km of Dounreay. However, this result was based on small numbers, arose in the context of multiple hypothesis testing, and is certainly vulnerable to possible systematic bias. Conclusion-The raised incidence of childhood leukaemia and non-Hodgkin's lymphoma around Dounreay cannot be explained by paternal occupation at Dounreay or by paternal exposure to external ionising radiation before conception. The observation of an apparent association between the use of beaches around Dounreay and the development of childhood leukaemia and non-Hodgkin's lymphoma might be an artefact of multiple testing and influenced by recall bias.