The objective of this study was to answer the following question, a crucial one for establishing radiation protection standards for the public: is the risk of cancer the same for a given low dose of radiation whether the dose is received over a short or extended period?
The study covered nearly 96 000 employees of seven nuclear facilities in Canada, the United States and the United Kingdom. Direct individual estimation of doses is systematically carried out among nuclear industry workers. The large number of workers included in the study allowed the most precise direct estimation so far of carcinogenic risk due to protracted low-dose exposures.
The data available previously to estimate the health effects of ionizing radiation were derived from studies of the mortality of atomic bomb survivors in Hiroshima and Nagasaki (acute, high doses) and of patients irradiated for therapeutic purposes (fractionated, high-dose exposures), as well as from many animal studies.
Environmental and occupational exposures are generally (except in the case of accidents) at low doses and are received over long periods of time. Hence, up to now, protection standards for these types of exposures have been set using extrapolation models, i.e. mathematical models to predict:
The validity of these extrapolations has been questioned. Some scientists claim that risks are in fact much higher than predicted, while others contend that a low level of exposure over a long time is in fact protective against cancer.
The IARC therefore set up a study, in collaboration with scientists from the three participating countries, to estimate directly the effect of low-dose protracted exposures on the risk of leukaemia (excluding chronic lymphocytic leukaemia, CLL, which is not thought to be caused by ionizing radiation) and of all other cancers.
The excess relative risk for death from leukaemia (excluding CLL) obtained is 2.2 per Sievert (Sv). Thus, a person having accumulated a dose of 0.1 Sv over a lifetime or over his or her employment has a 1.22 times greater risk of dying from leukaemia than a nonexposed person (Less than 8% of the workers included in our study had received such doses). No increased risk of mortality from all cancers except leukaemia was observed.
The risk estimates described here are somewhat lower than the linear extrapolations based on the Hiroshima and Nagasaki data, but they are compatible with risks ranging from almost zero to twice these extrapolations.
The risks which form the basis of current radiation protection recommendations (see below) are not invalidated by the results of this study.