Panel Report

Many participants expressed concern about the acceptability of the risk estimates made by the International Commission on Radiological Protection (ICRP). Some felt that history has shown continuing increases in the assessed risk per unit dose (risk factor) and that this process must be expected to continue. Others felt the present risk estimates were unduly conservative and would soon be relaxed.

Participants also expressed concerns about the composition of the ICRP, which has been described as a "self-perpetuating oligarchy." This body was established by the International Congress of Radiology in 1928 and originally comprised experienced radiologists who advised their less experienced colleagues. New members are appointed by the existing membership, as needs dictate, based on relevant research experience and scientific publications. Today, the Commission incorporates radiologists, radiation physicists and radiobiologists with extensive research experience in the biological effects of radiation. This membership enables the ICRP to provide effective guidance on scientific questions relating to the health hazards arising from radiation exposures. However, it also means that the Commission is somewhat insulated from the concerns of many members of the general public, who have no direct representation on it.

The Panel has recognized its responsibility to study all the information brought before it and to state clearly whether it believes the risk estimates published by the ICRP in 1991 (ICRP Publication 60) provide an appropriate basis for assessing radiological risks. This appendix reviews the information that the Panel considered and the conclusions that it reached.

Scientific knowledge of the hazards arising from exposure to ionizing radiation has been gradually accumulating throughout this century, often at the personal cost of the individuals exposed. The Panel appreciates that this is an ongoing process and that many unanswered questions remain.

Wilhelm Roentgen discovered X-rays in 1897, and hospitals quickly began to use them. Tissue injuries due to overexposure were soon reported. By the end of the last century, it became evident that, in severe cases, these were likely to develop into malignancies. Most early radiologists lost thumbs or fingers as a result of holding film holders in position during exposures. Many eventually died from the resulting malignancies.

A similar but smaller-scale story followed the discovery of radium by the Curies. Its use for medical implants to treat malignancies developed quickly, but the need to restrict direct handling of such sources was not fully appreciated until later.

Other concerns developed when laboratory experiments showed that exposures to ionizing radiation could lead to mutagenic effects in fruit flies and, possibly, other animals. Before World War I, the radiological societies of several countries generated detailed recommendations for establishing approved radiation safety practices. Later, the ICRP was established and issued more formal recommendations, which were continuously updated until the start of World War II.

There was, therefore, a substantial and well-developed understanding of radiation risks when the development of atomic weapons began in the U.S. and Canada. Many hospital physicists with knowledge of radiation hazards, particularly those experienced in handling radium, were recruited by the Manhattan Project. They developed techniques for controlling and chemically processing the large quantities of radioactive substances generated by the reactors in which plutonium was being produced. These scientists founded the health physics profession, and established safe procedures for working with the raw products on a very large scale. It was many years before comprehensive data on the risks associated with ingesting or inhaling all these new artificial radioactive isotopes could be generated. Nevertheless, the working practices developed by the scientists responsible for the Project minimized detrimental health effects among the staff working on it, to such an extent that little useful data for establishing a numerical relationship between exposures and risk could be obtained.

When two atomic bombs were dropped on Japan in 1945, thousands of individuals received very large or fatal exposures to radiation. It was quickly recognized that the study of future illnesses among the irradiated survivors would provide a unique source of data on the direct risks associated with radiation exposures. Their offspring could provide similar data about genetic risks to the population. The U.S. and Japanese governments set up the joint Radiation Effects Research Foundation (RERF) to conduct these studies and evaluate the results.

The main problem lay in accurately estimating the radia-tion doses received by the survivors. Since they obviously had not been wearing radiation dosimeters, some very sophisticated procedures were developed to evaluate individual doses. However, there was a limit to the precision these procedures could achieve, particularly at the lower exposure levels. Therefore, the United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR) has collected additional international information on human radiation exposures on an ongoing basis. UNSCEAR has reported periodically to the United Nations General Assembly since 1958. Virtually all significant data have been collated and reviewed in RERF and UNSCEAR reports.

The RERF studies showed that elevated levels of leukemia among the atomic bomb survivors became evident after about three years, peaked after seven years and then declined to normal levels after 11 or 12 years. It was expected that an excess incidence of solid tumour cancers would appear much later. Scientists anticipated that this would show a similar curve, which would rise more slowly and peak after a much longer interval of perhaps 30 or 35 years. However, the expected peak in observed excess cancers has not appeared; although more than 50 years have now elapsed, cancer incidence among the highly irradiated survivors has remained about two per cent higher than among a control group of non-irradiated Japanese throughout the last two decades. This led to the adoption of what is known as the "relative risk model." This model predicts that the excess risk of solid tumours arising from radiation exposure is not an absolute quantity (as it appears to be with leukemia), but is instead proportional to their natural level of incidence. Since cancer incidence rises with advancing age, this model predicts a larger total number of radiation-induced cancers, although most of these would arise very late in life.

Following these studies, the ICRP lowered the recommended occupational dose limits to avoid the total lifetime risk of death from a radiation-induced cancer being greater than that predicted by the old radiobiological model (regardless of the fact that most of these cancers are now expected to occur later in life, and to lead to a much smaller loss of life expectancy than formerly predicted). The new recommended dose limits, based on average doses over a five-year period, have not yet been incorporated into the Atomic Energy Control Board (AECB) regulations-a fact that was stressed repeatedly during the panel hearings.

The atomic bomb survivors received their radiation dose more than 50 years ago. Estimates of the excess lifetime cancer risk they have experienced will be updated regularly until they have all died. As a result, radiation risk estimates based on these survivors will be revised frequently. However, since most survivors are now very elderly, it is unlikely that this risk estimate will change significantly.

Many participants who do not fully understand why such changes have taken place suggested to the Panel that future large increases in the risk estimates should be expected. The Panel does not accept these suggestions. It also notes that a significantly larger proportion of the bomb survivors studied has remained alive over this 50-year period than would be the case for a similar control group of non-irradiated Japanese with the same age distribution. This means that the bomb survivors actually have a greater life expectancy than their contemporaries. Although cancer deaths among these survivors are, on average, about two per cent higher than for the control population, deaths from all other causes are lower by an even greater percentage.

Many scientists regard this as evidence that moderate doses of radiation can stimulate the immune system and thereby increase resistance to a wide range of other diseases-a view supported by thousands of animal studies reported in the scientific literature-but many other scientists believe such a conclusion is premature. Any increased resistance of the study population to disease may result from better medical care provided to the bomb survivors as a group, or from the early death of those bomb survivors who were constitutionally less strong during the immediate postwar period, before the detailed study of the remaining survivors began.

Since the excess cancer risk remains small even for the most highly irradiated individuals, it is impossible to determine whether the origin of any given cancer in any irradiated individual is related to radiation. Radiation-induced cancer risks are therefore referred to as "stochastic" risks, which can only be expressed as a probability-even if the individuals concerned have actually developed cancers.

There are two difficulties in estimating this stochastic risk from data relating to the atom bomb survivors. Inaccurate data, including uncertainty about the actual radiation doses the survivors received, may make the risk estimate incorrect; and the applicability of data derived from these survivors to workers experiencing occupational radiation exposures may be limited. Whereas the survivors received a single, almost instantaneous, acute radiation dose when the bomb exploded, most occupationally exposed workers slowly build up a significant radiation exposure as a result of chronic low-level exposure over many years. Risk estimates derived from other irradiated populations are needed to confirm the applicability of data derived from the bomb survivors. Many such studies have been undertaken of other exposed populations, including groups living in areas of high background radiation and groups exposed to unusually large medical or occupational radiation doses. Although such studies have been fairly compatible with those of the atomic bomb survivors, either the numbers of exposed individuals or the doses involved have been such that the statistical reliability of the derived radiation risk estimate is lower than for the bomb survivor studies.

To fill this gap, the International Agency for Research on Cancer has sponsored large-scale studies of workers in major nuclear establishments who have received significant occupational exposures over periods of many years. The results from the agency's first study were presented to the Panel during the public hearings. [Bliss Tracey, in Nuclear Fuel Waste Environmental Assessment Panel Public Hearings Transcripts, March 13, 1996, pp. 65-66.] The study combined the data from such workers in three different countries (the U.S., Canada and the U.K.). All the workers involved had been using approved dosimetry devices, so their actual radiation exposures were known with relative accuracy. There was a very small and not statistically significant excess of leukemias among the most highly exposed group, but no excess of solid tumours. Although the numbers involved in this collective international study were sufficiently large to provide reasonable statistical confidence in the results, an even larger study, involving a longer follow-up period and workers from more countries, is now in progress.

In addition, a great deal of data relating to the lung cancer risk arising from high radon levels in homes has now been collected. In general, no excess cancers have been demonstrated. Health Canada conducted a large-scale study in Winnipeg which, if the risks calculated from studies of uranium miners applied, would have conclusively identified an association between lung cancer incidence and radon levels in homes. No such association could be identified. Professor Cohen of the University of Pittsburgh has collected a large amount of data, covering most of the U.S., which strongly suggests a negative relationship between radon levels in homes and lung cancer incidence.

On the basis of these results and other data, many health physicists have come to believe that the linear relationship between radiation exposure and risk adopted by the ICRP in 1949 is not correct and that, in practice, there is a threshold level below which radiation exposures are not harmful but may even be beneficial. In the opinion of the Panel, this belief is not sufficiently supported to justify abandoning the current conservative approach to radiation control. However, it has been tentatively adopted by the Health Physics Society, the largest single organization of professional health physicists in the world.

The theoretical basis for a linear-no-threshold relationship was the demonstration, from radiobiological experiments, that the number of chromosome breaks associated with radiation exposure was directly proportional to the dose. From this it was assumed that, since cancerous cells have experienced chromosomal changes, there would be a risk of cancer induction that would also be proportional to the dose.

It is now recognized that numerous chromosomal breaks occur in every living person all the time, that most of these are repaired, and that damaged cells that cannot be repaired are eliminated by the immune system. The efficiency with which the immune system does this is the primary factor governing the cancer risk experienced by any individual. This risk increases with age, primarily because the effectiveness of the immune system de-creases with increasing age.

Many scientists believe it can be demonstrated in the laboratory that, while large radiation doses depress the immune system, chronic low-level exposures stimulate it. This debate is ongoing among many groups of physicians and scientists who estimate other stochastic risks, such as those associated with environmental chemicals and other factors. These groups have come together in the BELLE (Biological Effects of Low-level Exposures) organization. BELLE publishes a regular newsletter to maintain contact between scientists interested in this problem but working in different fields. It also maintains an Internet Web site that enables those interested to follow some of the ongoing scientific studies.

Most of the scientific information about the carcinogenic effects of radiation on humans has been obtained from studies of people irradiated at levels between 10 millisievert (mSv) and 30,000 mSv. Around the world, the typical annual background radiation dose is about 3 mSv (the EIS quoted an average figure of 2.6 mSv for Canada [Atomic Energy of Canada Limited, Environmental Impact Statement, p. 35.] ). The ICRP-recommended annual dose limit for additional non-medical exposures of members of the public is 1 mSv (ICRP Publication 60). Diagnostic X-ray procedures may involve local tissue doses of anywhere between about 1 mSv and 30,000 mSv.

The atomic bomb survivor studies showed no statistically significant increase in cancer incidence for doses below 20 mSv, but for doses above this the cancer risk appears to vary linearly with the received dose. The present controversy relates to whether this linear response extends down as far as the lowest doses for which experimental data can be generated. Adequate scientific data to answer this question definitively are not yet available.

AECB regulatory documents require that the annual risk to any individual following the closure of a high-level waste repository be below one in a million fatal cancers and serious genetic effects. On the basis of the linear hypothesis, this corresponds to a radiation dose of below 0.05 mSv per year, which means that any resulting health effects would be so small that they would fall on a part of the response graph far below that for which any experimental data will ever be available. Unless and until non-linearity in the response has been conclusively demonstrated for all types of radiation exposure, this leaves little alternative except to use the linear response hypothesis to calculate a ceiling value for the risk associated with such very small exposures, while recognizing that this is probably a conservative approach. The Panel has therefore accepted that the ICRP Publication 60 radiation risk factor currently provides the most appropriate basis for evaluating the radiological health implications associated with developing a high-level waste repository.