The evidence about cancer risks from radioactive particles is of interest to the committee because of suggestions that these risks are higher than those from the equivalent quantities of more uniform radiation (see summary of meeting 5). A review of the literature carried out for the committee showed that much of the relevant evidence is about 20 years old and is related to ‘hot particles’, ie relatively large and active particles. The committee’s interest is primarily in the small and less active particles that are most likely to be found in the environment.

The evidence from the 1980s is mainly about particles on the skin and in the lung. Most of the in vitro and animal data indicate that the risks from ‘hot particles’ are lower than those from uniform radiation. There are very few epidemiological studies on exposure of the public to particles in the environment but there are a number of studies on exposure of workers to plutonium particles, and some studies on radioactive particles used in diagnostic medicine. In general, the worker and medicine studies indicate that risks from particles are similar to or lower than risks from uniform irradiation.

The committee noted that a recent study of Sellafield plutonium workers shows no excesses of lung, liver or bone cancers. Recent studies of workers at Mayak do show excesses of these cancers but the doses to the lungs, liver and bone of these people were much higher than at Sellafield. The risk factors estimated from the Mayak data using standard dosimetric assumptions seem to be consistent with the factors derived from studies of populations exposed to external radiation. There are a number of uncertainties in extrapolating the Sellafield and Mayak results to the public. In particular, the particles inhaled by the workers were larger and more highly radioactive than those found routinely in the environment.


‘Hormesis’ is the term used for the idea that low doses of radiation have beneficial, rather than detrimental, effects on human health. ‘Thresholds’ is the term used here to refer to the idea that there are levels of radiation dose to the whole body or to particular body organs below which no detrimental effects occur. There are relatively few advocates of hormesis or thresholds or both in the UK, but there are larger numbers in other countries, especially the US. The committee had contacted a US organisation called ‘Radiation, Science and Health’ [insert link to web site] to request references to the studies that this organisation feels provide the strongest evidence for hormesis and thresholds. The committee discussed these papers and others supplied by members.

There are some human cancers (eg soft tissue sarcomas and some bone cancers) that appear not to occur unless doses to the key cells are high enough. There are also cancers in mice that show similar threshold-like responses. Some committee members are of the view that it is not appropriate to generalise from these particular situations. There is also a view that the results of some studies that appear to support hormesis and thresholds can be explained in terms of a biphasic dose response relationship (see summary of meeting 7). In vitro evidence for mammalian cells suggests that low doses of radiation do not increase activity in the principal genes that control repair of DNA double strand breaks. Nor is there conclusive evidence from animal studies that this sort of response occurs in vivo. Some members pointed out that, even if low radiation doses do boost the immune system in the short term, the end result could be accelerated aging or a shortened lifespan.

There are a number of epidemiological studies that show lower than expected numbers of cancer deaths or an apparent dose threshold. These include studies of workers who have ingested or inhaled radium or plutonium. There are also studies on exposure to radon in homes that fail to show the expected relationship between radon levels and numbers of deaths from lung cancer. A difficulty with these radon studies is that they are for geographical areas, not individual people or households, so they cannot take proper account of the large lung cancer risks from smoking. Some committee members are of the view that all the radon epidemiological data together (ie the data for uranium miners and other workers and the data for the public) support the linear, no threshold hypothesis.


The committee intends to carry out its own epidemiological study of infant leukaemia in Great Britain following the Chernobyl accident. For this purpose, it has requested data on leukaemia incidence from the Childhood Cancer Research Group. The committee will compare the numbers of leukaemias that occurred in children born well before the accident with those in children born in the seven months after the accident, in the subsequent three years and in years after that. The comparison will be carried out for the areas where the radioactivity levels were highest (north Wales, Cumbria and south-west Scotland) and for the areas where the levels were lowest (England south-east of a line from the Wash to Bristol). All the epidemiologists on the committee will receive the data and will do the analyses.


The Oxford Survey of Childhood Cancers (OSCC) is the main source of information on childhood cancer risks following in utero irradiation. For this reason it is of interest to the committee, even though it is about external irradiation only. The source of the exposure in the survey was abdominal x-rays to mothers during pregnancy, which were carried out much more frequently in the past than they are today.

The OSCC was a national case-control study of childhood cancer mortality, set up by the late Professor Alice Stewart. The survey was based on all children in England, Scotland and Wales aged under 16 who died from leukaemia or other cancers during 1953-81, and live controls matched individually by age, gender and geographical location at death. Information was obtained from death certificates, interviews with the mothers, and by examining records of family doctors, antenatal clinics and radiologists. The most recent analyses of cancer risks were based on data for 14,759 traced cases (among over 22,000 notified deaths), who died aged 0-15 years during 1953-79, and their matched controls.

The OSCC results show that, overall, children exposed to radiation in utero had about a 40% greater risk of contracting cancer than unexposed children. This risk was highest for children born during the 1940s and decreased for children born during the 1950s and 1960s as radiation doses were reduced. Smaller studies in the northeast US, Sweden and elsewhere confirmed the association between prenatal x-rays and childhood cancer.

In the OSCC there are similar relative risks of leukaemia and solid cancers from in utero irradiation. This is in contrast to other studies that show a higher relative risk of leukaemia from radiation exposure during childhood. The OSCC does not show that in utero irradiation is especially likely to lead to infant leukaemia, but the study is not very precise in this area because the lowest age group considered is 0 to 2 years. (The 0-1 year old group would need to be considered to examine infant leukaemia risks in detail.) Also, most of the x-rays were carried out in the third trimester of pregnancy, so a higher risk in the first or second trimesters would not necessarily be detected. The lowest dose at which there is a statistically significant risk in the OSCC is about 10 milligray.

The committee also discussed a US study of a small number of women who, for medical reasons, were given the radioactive isotope iron-55 during pregnancy. (This occurred in the 1940s and would not be permitted today.) The study shows a significant excess of childhood cancers but none of infant leukaemia. The information on the quantities of iron-55 given to the mothers is poor and there are large uncertainties about the radiation doses to the foetuses, so it is not possible to obtain quantitative risk estimates from the study.

Next (Meeting 7)