A comparison of the risks of leukaemia and non-Hodgkin's lymphoma in the first generation offspring (F 1 ) of the Danish Thorotrast patients with those observed in other studies of parental pre-conception irradiation

The high doses of alpha-particle radiation which Thorotrast delivers to the testes makes the study of offspring of patients, who were administered this diagnostic contrast medium, particularly relevant to the controversy regarding childhood leukaemia in the offspring of the male Sellafield workforce, as it has been suggested that the alpha-particle dose from actinides may, in part, account for the observed association between paternal pre-conception external irradiation and leukaemia. The risks of childhood leukaemia and non-Hodgkin's lymphoma found among the offspring of the male Danish cohort of neurosurgical patients injected with Thorotrast for cerebral arteriography are compared with those observed in the offspring of the Sellafield (western Cumbria), Ontario and Scottish radiation workers and in the offspring of the Japanese atomic bomb survivors. Risks are compared using linear and exponential forms of a relative risk model. The preconception exposure risks of leukaemia, or of leukaemia and non-Hodgkin's lymphoma, in the Danish study are statistically incompatible with the corresponding risks in the children of the Sellafield workforce (P < 0.01), which is entirely due to incompatibility with the risks in the children of the Sellafield workforce born in Seascale (P < 0.01). Indeed no cases arose in the Danish study. The statistical incompatibility between the offspring of the Danish patients and the Seascale-born offspring of the Sellafield workforce is independent of the models used and is insensitive to uncertainties in the gonadal dose estimates in the respective datasets or to uncertainties in the estimates of the background cancer rates in the Seascale and Danish populations. The risks of leukaemia and non-Hodgkin's lymphoma in the Danish dataset are statistically compatible with the risks observed in the offspring of the Sellafield workforce born elsewhere in western Cumbria and with those seen among the offspring of the Ontario or Scottish radiation workforces as well as with those in the offspring of the Japanese atomic bomb survivors. The statistical compatibilities of the leukaemia preconception exposure risks in these various groups are also independent of the models used. It is most unlikely that the Seascale childhood leukaemia cases are caused by paternal exposures to alpha-particle emitters such as plutonium. In each non-Seascale dataset the risks are consistent with there being no excess hazard of leukaemia associated with paternal pre-conception exposure to ionising radiation, and it would seem that the association seen for Seascale-born children should be attributed to chance.