Mortality of workers at the Sellafield plant of British Nuclear Fuels

Abstract

The mortality of all 14 327 people who were known to have been employed at the Sellafield plant of British Nuclear Fuels at any time between the opening of the site in 1947 and 31 December 1975 was studied up to the end of 1983. The vital state of 96% of the workers was traced satisfactorily and 2277 were found to have died, 572 (25%) from cancer. On average the workers suffered a mortality from all causes that was 2% less than that of the general population of England and Wales and 9% less than that of the population of Cumberland (the area in which the plant is sited). Their mortality from cancers of all kinds was 5% less than that of England and Wales and 3% less than that of Cumberland. In the five years after their first employment Sellafield workers had an overall mortality that was 70% of that of England and Wales, probably due to healthier members of the population being selected for employment. Raised death rates from cancers of several specific sites were found, but only for those of ill defined and secondary sites was the excess statistically significant (30 observed, 19·7 expected). For cancers of the liver and gall bladder there was a significant deficit of deaths (four observed, 10·5 expected). Workers in areas of the plant where radiation exposure was likely were issued with dosimeters to measure their external exposure to ionising radiations. Personal dose records were maintained for workers who entered such areas other than infrequently. Workers with personal dose records (“radiation” workers) had lower death rates from all causes combined than other workers but the death rates from cancer in the two groups were similar. Compared with the general population radiation workers had statistically significant deficits of liver and gall bladder cancer, lung cancer, and Hodgkin's disease. There were excesses of deaths from myeloma (seven observed, 4·2 expected) and prostatic cancer (19 observed, 15·8 expected) but these were not significant and there was no evidence of an excess of leukaemia (10 deaths observed, 12·2 expected) or cancer of the pancreas (15 observed, 17·8 expected). Non-radiation workers had a significant deficit of leukaemia (one death observed, 5·1 expected) and a significant excess of cancers of ill defined and secondary sites (13 deaths observed, 5·8 expected). For no type of cancer was the ratio of observed to expected deaths significantly different between radiation and non-radiation workers. For non-neoplastic conditions radiation workers in general had lower death rates than other workers, and for none of the causes of death examined was the mortality significantly higher among radiation workers. Also there was no significant evidence of a relation between cancers of specific sites and duration of employment at Sellafield as a radiation worker. Comparisons were made of the mortality of radiation workers who had accumulated different exposures to radiation as measured by dosimeter records. For all causes of death combined there was a significant negative association between death rates and accumulated doses which was no longer present when exposures in the 15 years preceding death were discounted. There were positive associations between accumulated radiation dose and death rates from bladder cancer, multiple myeloma, leukaemia, and all lymphatic and haematopoietic neoplasms. These were not statistically significant when exposure up to the time of death or up to two years before death was considered. Nevertheless, when exposures recorded in the 15 years before death were ignored these associations, with the exception of that for leukaemia, became significant. The association of radiation dose with bladder cancer has not been reported in other studies of radiation workers but the findings for myeloma were consistent with the excess reported for radiation workers in the United States. Excess death rates from all cancers were estimated to be reduced by six deaths per million person years at risk per 10 mSv (1 rem) for radiation accumulated up to the time of death (95% confidence limits -22 and 13) and increased by 17 deaths per million person years at risk per 10 mSv for radiation exposure accumulated 15 or more years before death (95% confidence limits —30 and 70). The findings were compatible with the estimates of the carcinogenic risk of low doses of ionising radiations made by the International Commission on Radiological Protection (about five deaths per million person years at risk per 10 mSv), but on these data alone it was impossible to exclude with confidence a risk 10-fold higher or the absence of any risk at all. © 1986, British Medical Journal Publishing Group. All rights reserved.