> Statement to the Subcommittee on Energy and Environment  of the Committee on Science, United States House of Representatives,  July 18, 2000, by  Steve Wing, Associate Professor, Department of Epidemiology,
School of Public Health, University of North Carolina

> Living Dangerously, by Rob Edwards, New Scientist, 28 February 1998

This first report of the European Committee on Radiation Risk is intended for regulators and those who have to make decisions about the health effects of radioactive releases. It presents a rational model for calculating the health risks of exposure to ionizing radiation. Unlike the existing framework of modelling radiation risk, the ECRR model uses evidence from the most recent research, from new discoveries in radiation biology and from human epidemiology to create a system of calculation which gives results which are in agreement both with the mechanism of radiation action at the level of the living cell and observation of disease in exposed populations.

This follows concerns about the conventional risk models advised by the International Commission on Radiological Protection, a body which has been widely criticised for lack of balance and for being self appointed and too close to the nuclear industry. The ICRP model entirely fails to explain ill health in populations exposed to internal radioactivity. The ECRR cites massive amounts of evidence; examples are effects following Chernobyl, the persistent 10-fold excess of childhood leukaemia near Sellafield, lymphoma in veterans exposed to depleted Uranium dust during the Gulf War and the Balkans, and breast cancer in the cohort of women who were adolescent during 1957 - '63 when nuclear weapons-testing was at its height. The UK government is sufficiently worried about the inability of the ICRP model to explain or predict such clear evidence of harm from internal radioactive exposures that in 2001 it set up its own Committee Examining Radiation Risk from Internal Emitters (CERRIE). Dr Chris Busby who is Scientific Secretary to the ECRR is a founder member of CERRIE and also sits on the UK Ministry of Defence Depleted Uranium Oversight Board (DUOB). In this volume, the committee explains how the present risk model came to be universally used, and points out its scientific shortcomings. It also addresses the ethical basis of releasing radioactive materials to the environment.

The volume is essential reading for anyone involved in legislation in this area and should also be of interest to members of the public who need to estimate the effects of nuclear discharges.

Summary of Contents

The report outlines the committee’s findings regarding the effects on human health of exposure to ionising radiation and presents a new model for assessing these risks. It is intended for decision-makers and others who are interested in this area and aims to provide a concise description of the model developed by the committee and the evidence on which it depends. The development of the model begins with an analysis of the present risk model of the International Commission on Radiological Protection (ICRP) which is the basis of and dominates all present radiation risk legislation. The committee regards this ICRP model as essentially flawed as regards its application to exposure to internal radioisotopes but for pragmatic reasons to do with the existence of historical exposure data has agreed to adjust for the errors in the ICRP model by defining isotope and exposure specific weighting factors for internal exposures so that the calculation of effective dose (in Sieverts) remains. Thus, with the new system, the overall risk factors for fatal cancer published by ICRP and other risk agencies may be used largely unchanged and legislation based upon these may also be used unchanged. It is the calculation of the dose which is altered by the committee's model.

1. The European Committee on Radiation Risk arose out of criticisms of the risk models of the ICRP which were explicitly identified at the European Parliament STOA workshop in February 1998; subsequently it was agreed that an alternative view should be sought regarding the health effects of low level radiation. The committee consists of scientists and risk specialists from within Europe but takes evidence and advice from scientists and experts based in other countries.

2. The report begins by identifying the existence of a dissonance between the risk models of the ICRP and epidemiological evidence of increased risk of illness, particularly cancer and leukaemia, in populations exposed to internal radioactive isotopes from anthropogenic sources. The committee addresses the basis in scientific philosophy of the ICRP risk model as applied to such risks and concludes that ICRP models have not arisen out of accepted scientific method. Specifically, ICRP has applied the results of external acute radiation exposure to internal chronic exposures from point sources and has relied mainly on physical models for radiation action to support this. However, these are averaging models and cannot apply to the probabilistic exposures which occur at the cell level. A cell is either hit or not hit; minimum impact is that of a hit and impact increases in multiples of this minimum impact, spread over time. Thus the committee concludes that the epidemiological evidence of internal exposures must take precedence over mechanistic theory-based models in assessing radiation risk from internal sources.

3. The committee examines the ethical basis of principles implicit in the ICRP models and hence in legislation based on them. The committee concludes that the ICRP justifications are based on outmoded philosophical reasoning, specifically the averaging cost-benefit calculations of utilitarianism. Utilitarianism has long been discarded as a foundation for ethical justification of practice owing to its inability to distinguish between just and unjust societies and conditions. It may, for example, be used to underpin a slave society, since it is only overall benefit which is calculated, and not individual benefit. The committee suggests that rights-based philosophies such as Rawls's Theory of Justice or considerations based on the UN Declaration of Human Rights should be applied to the question of avoidable radiation exposures to members of the public resulting from practice. The committee concludes that releases of radioactivity without consent can not be justified ethically since the smallest dose has a finite, if small, probability of fatal harm. In the event that such exposures are permitted, the committee emphasises that the calculation of ‘collective dose’ should be employed for all practices and time scales of interest so that overall harm may be integrated over the populations.

4. The committee believes that it is not possible accurately to determine ‘radiation dose to populations’ owing to the problems of averaging over exposure types, cells and individuals and that each exposure should be addressed in terms of its effects at the cell or molecular level. However, in practice this is not possible and so the committee has developed a model which extends that of the ICRP by the inclusion of two new weighting factors in the calculation of effective dose. These are biological and biophysical weighting factors and they address the problem of ionisation density or fractionation in time and space at the cell level arising from internal point sources. In effect, they are extensions of the ICRP’s radiation weighting factors employed to adjust for differences in ionisation density resulting from different quality radiations (e.g. alpha-, beta and gamma).

5. The committee reviews sources of radiation exposure and recommends caution in attempting to gauge the effects of novel exposures by comparison with exposures to natural radiation. Novel exposures include internal exposures to artificial isotopes like Strontium-90 and Plutonium-239 which bind specifically to DNA but also include micrometer range aggregates of isotopes (hot particles) which may consist of entirely man-made isotopes (e.g. Plutonium) or altered forms of natural isotopes (e.g. depleted Uranium). Such comparisons are presently made on the basis of the ICRP concept of ‘absorbed dose’ which does not accurately assess the consequence for harm at the cell level. Comparisons between external and internal radiation exposures may also result in underestimates of risk since the effects at the cell level may be quantitatively very different.

6. The committee argues that recent discoveries in biology, genetics and cancer research suggest that the ICRP target model of cellular DNA is not a good basis for the analysis of risk and that such physical models of radiation action cannot take precedence over epidemiological studies of exposed populations. Recent results suggest that very little is known about the mechanisms leading from cell impact to clinical disease. The committee reviews the basis of epidemiological studies of exposure and points out that many examples of clear evidence of harm following exposure have been discounted by ICRP on the basis of invalid physical models of radiation action. The committee reinstates such studies as a basis for its estimates of radiation risk.

7. The committee reviews the models of radiation action at the cell level and conclude that the ‘linear no threshold’ model of the ICRP is unlikely to represent the response of the organism to increasing exposure except for external irradiation and for certain end points in the moderately high dose region. Extrapolations from the Hiroshima lifespan studies can only reflect risk for similar exposures i.e. high dose acute exposures. For low dose exposures the committee concludes, from a review of published work, that health effects relative to the radiation dose are proportionately higher at low doses and that there may be a biphasic dose response from many of these exposures owing to inducible cell repair and the existence of high-sensitivity phase (replicating) cells. Such dose-response relationships may confound the assessment of epidemiological data and the committee points out that the lack of a linear response in the results of epidemiological studies should not be used as an argument against causation.

8. In further considering mechanisms of harm, the committee concludes that the ICRP model of radiation risk and its averaging methods exclude effects which result from anisotropy of dose both in space and in time. Thus the ICRP model ignores both high doses to local tissue caused by internal hot particles, and sequential hits to cells causing replication induction and interception (second event), and merely averages all these high risk situations over large tissue mass. For these reasons, the committee concludes that the unadjusted ‘absorbed dose’ used by ICRP as a basis of risk calculations is flawed, and has replaced it with an adjusted ‘absorbed dose’ which uses enhancement weightings based on the biophysical and biological aspects of the specific exposure. In addition, the committee draws attention to risks from transmutation from certain elements, notably Carbon-14 and Tritium, and has weighted such exposures accordingly. Weightings are also given to radioactive versions of elements which have a particular biochemical affinity for DNA e.g. Strontium and Barium and certain Auger emitters.

9. The committee reviews the evidence which links radiation exposure to illness on the basis that similar exposures define the risks of such exposures. Thus the committee considers all the reports of associations between exposure and ill health, from the A-bomb studies to weapons fallout exposures, through nuclear site downwinders, nuclear workers, reprocessing plants, natural background studies and nuclear accidents. The committee draw particular attention to two recent sets of exposure studies which show unequivocal evidence of harm from internal irradiation at low dose. These are the studies of infant leukemia following Chernobyl, and the observation of increased minisatellite DNA mutations following Chernobyl. Both of these sets of studies falsify the ICRP risk models by factors of between 100 and 1000. The committee uses evidence of risk from exposures to internal and external radiation to set the weightings for the calculation of dose in a model which may be applied across all exposure types to estimate health outcomes. Unlike the ICRP the committee extends the analysis from fatal cancer to infant mortality and other causes of ill health including non-specific general health detriment.

10. The committee concludes that the present cancer epidemic is a consequence of exposures to global atmospheric weapons fallout which peaked in the period 1959-63 and that more recent releases of radioisotopes to the environment from the operation of the nuclear fuel cycle will result in significant increases in cancer and other types of ill health.

11. Using both the ECRR's new model and that of the ICRP the committee calculates the total number of deaths resulting from the nuclear project since 1945. The ICRP calculation, based on figures for doses to populations up to 1989 given by the United Nations, results in 1,173,600 deaths from cancer. The ECRR model predicts 61,600,000 deaths from cancer, 1,600,000 infant deaths and 1,900,000 foetal deaths. In addition, the ECRR predicts a 10% loss of life quality integrated over all diseases and conditions in those who were exposed over the period of global weapons fallout.

12. The committee lists its recommendations. The total maximum permissible dose to members of the public arising from all human practices should not be more than 0.1mSv, with a value of 5mSv for nuclear workers. This would severely curtail the operation of nuclear power stations and reprocessing plants, and this reflects the committee’s belief that nuclear power is a costly way of producing energy when human health deficits are included in the overall assessment. All new practices must be justified in such a way that the rights of all individuals are considered. Radiation exposures must be kept as low as reasonably achievable using best available technology. Finally, the environmental consequences of radioactive discharges must be assessed in relation to the total environment, including both direct and indirect effects on all living systems.

Mr. Chairman and Members of the Committee, thank you for inviting me to testify about health effects of low level radiation. I am an epidemiologist on the faculty at the University of North Carolina where I have studied radiation health effects among workers at Oak Ridge, Los Alamos, Hanford and Savannah River under funding from the Departments of Energy and Health and Human Services. Epidemiology, the study of disease in human populations, is especially important in risk estimation and standard setting because animal and laboratory studies necessitate extrapolation from high to low doses, from molecules and cells to organisms, and from other species to humans (1-3).

We know that ionizing radiation can cause cancer and inherited mutations by damaging DNA. Although epidemiologists have studied populations exposed to both high and low levels of radiation, extrapolation of risks from high to low doses has led to a debate over whether a straight line extrapolation, the linear no-threshold model, is appropriate. My testimony will make three points: current cancer risk estimates are too low by a factor of ten or more; current standards do not adequately protect workers and the public;  and, a large and growing body of scientific evidence shows that there is no basis for further relaxation of radiation protection standards.

Extrapolation from high dose studies.

High dose studies examine special populations including patients receiving radiation treatments.  By far the most influential are studies of survivors of the bombings of Hiroshima and Nagasaki that are currently the primary basis for cancer risk estimates. However, the A-bomb studies are flawed due to selective survival, poor dose measurement and confounding exposures (4-7).

The atomic bombings produced massive immediate casualties as well as delayed deaths due to lingering effects of radiation, infectious epidemics, and the destruction of food, housing, and medical services (8).  Only the healthiest survived these conditions, especially among those who are most vulnerable, the young and the old.  By 1950, when a list of survivors was assembled for long-term study, persons most susceptible to radiation had already died. The healthy survivor effect leads to underestimation of risks, particularly for exposures in utero, during childhood, and at older adult ages (6).

Detection of radiation risks depends upon the ability of an epidemiological study to classify persons according to their exposure levels. A-bomb survivors were not wearing radiation badges, therefore their exposures had to be estimated by asking survivors about their locations and shielding at the time of detonation.  In addition to the typical types of recall bias that occur in surveys, stigmatization of survivors made some reluctant to admit their proximity (9). Acute radiation injuries such as hair loss and burns among survivors who reported they were at great distances from the blasts (10, 11) suggests the magnitude of these errors, which would lead to under estimation of radiation risks.

Another bias occurs because of the higher exposures of distant survivors to residual radiation.  Fallout affected distant survivors in both cities (8, 12). In addition, survivors who were shielded or exposed at greater distances were strong enough to enter the areas near the hypocenters of the blasts within hours of detonation, exposing themselves to residual radiation created by the atomic weapons (8, 12-14). Residual radiation exposures of lower dose survivors leads to an underestimate of radiation risks.

Direct observation from low dose studies.

In 1956 Dr. Alice Stewart and colleagues reported in The Lancet that fetal exposures during obstetric x-ray examinations are associated with elevated childhood cancer rates (15). The fetus is especially sensitive to radiation due to rapid cell division. Stewart's findings have been replicated in numerous other low dose studies (6, 16-18), and standards for medical practice now dictate that small doses of radiation associated with a single x-ray should be avoided during pregnancy.

Long-term studies of cancer among nuclear workers began to appear in the 1970s when Mancuso, Stewart and Kneale reported that small doses of radiation received at older ages raised cancer rates among workers at the plutonium production facility in Hanford, Washington (19). Manhattan Project scientists realized in the early 1940s that workers in the weapons plants faced special hazards, and they created a unique resource for health studies at some facilities by issuing each employee a radiation monitor that was incorporated into the security badge required at work. Although dose records are poor for many workers and veterans, long-term studies of well-monitored workers have now been reported from nuclear facilities in the U.S., the United Kingdom and Canada. Despite the fact that workers are generally healthy adults, many of these studies have demonstrated relationships between low level radiation and cancer death, particularly among older workers. The greater sensitivity of older adults to ionizing radiation was not recognized in A-bomb studies due to selective survival, however this observation is consistent with studies that show reductions in immune function and efficiency of DNA repair with aging (6, 20). Risk estimates from many occupational studies are approximately 10 times higher than estimates based on follow-up of A-bomb survivors (21-33), showing that current protection standards are too lax.  In our recent study of multiple myeloma among Oak Ridge, Hanford, Los Alamos and Savannah River workers, doses between 5 and 10 rems were associated with a threefold elevated risk, and doses over 10 rems were associated with a fivefold elevated risk (33).  None of the multiple myeloma cases had recorded doses over the current U.S. occupational limit of five rems per year.

From the United Kingdom comes evidence that paternal preconception exposures are associated with risk of childhood cancer, stillbirth and an excess of male compared to female births (34-36). The ability of radiation to induce heritable genetic mutations in experimental animals has been recognized since the 1920s (37).  This recent evidence suggests that small doses of radiation delivered in the period prior to conception can lead to genetic effects in human offspring. Evidence on genomic instability following exposure to alpha radiation raises concerns for both carcinogenic and inherited genetic effects (38-40).

The belief that radiation risks at low doses could be extrapolated from high dose studies led some to predict that cancer risks of radiation could not be detected among nuclear workers. Although this has turned out to be false, some researchers have pooled data from different worker populations in order to increase sample size, believing that this would increase power to detect radiation risks (41-43). Unfortunately, pooling populations with different types of radiation, exposure conditions, measurement qualities and worker selection factors, achieves statistical precision at the cost of accuracy, diluting radiation effects (43).

Diseases and genetic mutations caused by radiation do not carry a marker showing their origins, therefore epidemiologists look for excess rates of disease in populations with higher radiation exposures. However, it is easy to design an epidemiological study of environmental or occupational radiation exposure that is unable to detect low level effects. Only in special circumstances, such as the cases of well-monitored workers and certain medical exposures (44), is it possible to quantify low doses and subsequent risk. The sensitivity of epidemiological studies is compromised because people generally cannot be traced between the time they are exposed and the time disease develops, and because medical information (other than
cause of death) is not routinely available for populations without universal medical care. It is incorrect to conclude that low level radiation is safe on the basis of studies that lack careful radiation measurements and follow-up of medical outcomes.  Unfortunately such conclusions have been made based on studies of geographic variation in average background radiation (45).

Furthermore, some scientists have mistakenly claimed that there is no evidence of radiation health effects below some arbitrary level. Not only do such statements ignore an extensive medical literature on in utero and occupational radiation; they reflect a basic misunderstanding of how epidemiology works.  In order to detect the risks from a hazardous agent, epidemiologists study a range of exposure levels. For example, we compare lung cancer rates of never-smokers to rates among people who smoke less than a pack a day, one pack a day, two packs a day, and three or more packs a day. It would be incorrect to separate the data for people who smoke one cigarette a day and declare that low levels of smoking are safe. Conclusions about health effects of agents such as radiation and cigarettes should be derived from data on a range of exposures.

The current state of knowledge.

As knowledge about ionizing radiation has grown, health effects have been recognized from activities that until recently were thought to be safe. Despite past assurances about the safety of nuclear weapons tests, the National Cancer Institute's recent study indicates that tens of thousands of Americans can expect to get thyroid cancer from just one of the radionuclides released by atmospheric testing (46). The fact that radiation protection standards have been reduced as scientific study of low doses increases is another measure of concern (7). Although the International Commission on Radiation Protection recommended in 1990 that the 5 rem per year limit for nuclear workers be reduced to 2, the U.S. continues to permit workers to be exposed to more than twice the radiation dose allowed by countries that adopted the international standard, including Canada and the European Union.

The nuclear age is little more than a half-century old. Although much has been learned about radiation during this time, there is much more that remains to be understood about human health effects. It is increasingly clear that there is great variability in the sensitivity of humans to low level radiation due to factors such as age, genetic susceptibility and exposures to chemical agents, infection or nutritional factors.  Decisions about exposure standards should take account of the special risks faced by the young, the old and the genetically susceptible. Public health and moral principles demand that we protect the most vulnerable. As amply documented by the Secretarial Panel for the Evaluation of Epidemiologic Research appointed by Admiral Watkins (47), President Clinton's Advisory Committee on Human Radiation Experiments (48), a taskforce of the Physicians for Social Responsibility (49), and numerous publications in the scientific literature (50-54), the body of scientific knowledge about the health effects of ionizing radiation has been compromised by concerns about secrecy and public relations. In its 1995
report, the President's Advisory Committee on Human Radiation noted that, "By the mid-1960s the possibility that data gathering could only get the AEC (Atomic Energy Commission) into more trouble became an incentive to 'not study at all'" (48). These attitudes have continued to affect research in recent decades (51, 52). In the case of regulatory standards that are intended to protect the health of workers and the public, policy makers should consider scientific evidence and testimony with the understanding that scientists have been restrained from fully investigating the effects of low level ionizing radiation.
Current radiation standards already fail to adequately protect workers and the public, even if flawed risk estimates from A-bomb studies are used. The 1994 GAO report on Nuclear Health and Safety notes that exposures permitted by current Nuclear Regulatory Commission and Department of Energy guidelines, according to those agencies, would lead to 1 in 300 premature cancer deaths in the general public and 1 in 8 among workers (55).  No other carcinogens are permitted such lax standards. I strongly urge members of Congress and the regulatory agencies to exercise precaution and prudence in order to protect the health and lives of the public and of future generations who will be affected by decisions on production and
disposition of nuclear materials.
 
References
1. National Research Council, Committee on the Biological Effects of Ionizing Radiation. Health effects of exposure to low levels of Ionizing Radiation (BEIR V). Washington, DC:National Academy Press, 1990.
2. International Commission on Radiological Protection. 1990 Recommendations of the International Commission on Radiological Protection, vol 21. Oxford:Pergamon Press, 1991.
3. United Nations Scientific Committee on the Effects of Atomic Radiation. Sources and effects of ionizing radiation. New York:United Nations, 1993.
4. Stewart A. Detection of late effects of ionizing radiation: why deaths of A-bomb survivors are so misleading. Int J Epidemiol. 14:52-56(1985).
5. Stewart A. A-bomb data:  detection of bias in the life span study cohort. Environmental Health Perspectives 105:1519-21(1997).
6. Stewart A. The role of epidemiology in the detection of harmful effects of radiation. Environmental Health Perspectives 108:93-96(2000).
7. Wing S, Richardson D, Stewart A. The relevance of occupational epidemiology to radiation protection standards. New Solutions 9:133-151(1999).
8. Committee for the Compilation of Materials on Damage Caused by the Atomic Bombs in Hiroshima and Nagasaki. Hirosima and Nagasaki:  The physical, medical, and social effects of the atomic bombings. Tokyo:Iwanami Shoten, Publishers, 1981.
9. Lindee MS. Suffering made real: American science and the survivors at Hiroshima. Chicago, Ill.:University of Chicago Press, 1994.
10. Neriishi K, Stram DO, Vaeth M, Mizuno S, Akiba S. The observed relationship between the occurrence of acute radiation effects and leukemia mortality among A-bomb survivors. Radiat Res. 125:206-213(1991).
11. Neriishi K, Wong FL, Nakashima E, Otake M, Kodama K, Choshi K. Relationship between cataracts and epilation in atomic bomb survivors. Radiation Research 144:107-13(1995).
12. Rabbitt Roff S. Hotspots: The Legacy of Hiroshima and Nagasaki. London:Cassell, 1995.
13. Rabbitt Roff S. Residual radiation in Hiroshima and Nagasaki. Lancet 348(1997).
14. Yamahata Y. Nagasaki Journey:  The photographs of Yosuke Yamahata. Rohnert Park, CA:Pomegranate, 1995.
15. Stewart AM, Webb J, Giles D, Hewitt D. Malignant diseases in childhood and diagnostic irradiation in utero. Lancet 2:447(1956).
16. Doll R, Wakeford R. Risk of childhood cancer from fetal irradiation [see comments]. British Journal of Radiology 70:130-9(1997).
17. McMahon B. Prenatal X-ray exposure and childhood cancer. Journal of the National Cancer Institute 28:1173(1962).
18. Bithell JF, Stewart AM. Pre-natal irradiation and childhood malignancy: a review of British data from the Oxford Survey. British Journal of Cancer 31:271-87(1975).
19. Mancuso TF, Stewart A, Kneale G. Radiation exposures of Hanford workers dying from cancer and other causes. Health Physics 33:369-385(1977).
20. Moriwaki S-I, Ray S, Tarone R, Kraemer K, Grossman L. The effect of donor age on the processing of UV-damaged DNA by cultured human cells:  Reduced DNA repair capacity and increased DNA mutability. Mutation Research 364:117-123(1996).
21. Beral V, Fraser P, Carpenter L, Booth M, Brown A, Rose G. Mortality of employees of the Atomic Weapons Establishment, 1951-82. British Medical Journal 297:757-70(1988).
22. Kneale GW, Mancuso TF, Stewart AM. Hanford Radiation Study III:  A Cohort Study of the Cancer Risks from Radiation to Workers at Hanford (1944-77 deaths) by the Method of Regression Models in Life-tables. BJIM 38:156-166(1981).
23. Kneale GW, Stewart AM. Reanalysis of Hanford data: 1944-1986 deaths. American Journal of Industrial Medicine 23:371-89(1993).
24. Kneale GW, Stewart AM. Factors affecting recognition of cancer risks of nuclear workers. Occupational & Environmental Medicine 52:515-23(1995).
25. Richardson DB, Wing S. Greater sensitivity to radiation exposures at older ages among workers at Oak Ridge National Laboratory: Follow-up through 1990. International Journal of Epidemiology 28:428-436(1999).
26. Richardson D, Wing S. Radiation and mortality among workers at Oak Ridge National Laboratory:  Positive associations for doses received at older ages. Environmental Health Perspectives 107:649-656(1999).
27. Ritz B, Morganstern H, Moncau J. Age at exposure modifies the effects of low-level ionizing radiation on cancer mortality in an occupational cohort. Epidemiology 10:135-140(1999).
28. Ritz B, Morganstern H, Froines J, Young B. Effects of exposure to external ionizing radiation on cancer mortality in nuclear workers monitored for radiation at Rocketdyne/Atomics International. American Journal of Industrial Medicine 5:21-31(1999).
29. Ritz B. Radiation exposure and cancer mortality in uranium processing workers. Epidemiology 10:531-538(1999).
30. Wing S, Shy CM, Wood JL, Wolf S, Cragle DL, Frome EL. Mortality among workers at Oak Ridge National Laboratory: Evidence of radiation effects in follow-up through 1984. JAMA 265:1397-402(1991).
31. Wing S, Shy CM, Wood JL, Wolf S, Cragle DL, Tankersley W, Frome EL. Job factors, radiation and cancer mortality at Oak Ridge National Laboratory: follow-up through 1984. American Journal of Industrial Medicine 23:265-79(1993).
32. Wing S. A Review of Recent Findings on Radiation and Mortality at Oak Ridge National Laboratory. In: Neue Bewertung des Strahlenrisikos (Lengfelder E, Wendhauser H, eds). Munich:Medizin Verlag, 1993;217-228.
33. Wing S, Richardson D, Wolf S, Mihlan G, Crawford-Brown D, Wood J. A case control study of multiple myeloma at four nuclear facilities. Annals of Epidemiology 10:144-153(2000).
34. Gardner M, et al. Results of case-control study of leukaemia and lymphoma among young people near Sellafield nuclear plant in West Cumbria. British Medical Journal 300:423-429(1990).
35. Parker L, Pearce M, Dickenson H, Aitkin M, Craft A. Stillbirths among offspring of male radiation workers at Sellafield nuclear reprocessing plant. The Lancet 354:1407-14(1999).
36. Dickinson HO, Parker L, Binks K, Wakeford R, Smith J. The sex ratio of children in relation to paternal preconceptional radiation dose: a study in Cumbria, northern England. Journal of Epidemiology & Community Health 50:645-52(1996).
37. Muller HJ. Artificial transmutation of the gene. Science 66:84-87(1927).
38. Khadim MA, MacDonald DA, Goodhead DT, Lorimore SA. Transmission of chromosome instability after plutonium alpha particle irradiation. Nature 355(1992).
39. Morgan WF, Day JP, Kaplan MI, McGhee EM, Limoli CL. Genomic instability induced by ionizing radiation. Radiation Research 146:247-258(1996).
40. Watson GE, Lorimore SA, G. WE. Long-term in-vivo transmission of alpha particle-induced chromosomal instability in murine haemopoietic cells. Int J Radiat Biol 69:175-182(1996).
41. Cardis E, Gilbert ES, Carpenter L, Howe G, Kato I, Armstrong BK, Beral V, Cowper G, Douglas A, Fix J, Fry SA, Kaldor J, Lave C, Salmon L, Smith PG, Voelz GL, Wiggs LD. Effects of low doses and low dose rates of external ionizing radiation: cancer mortality among nuclear industry workers in three countries [see comments]. Radiation Research 142:117-32(1995).
42. Gilbert ES, Cragle DL, Wiggs LD. Updated Analyses of Combined Mortality Data for Workers at the Hanford Site, Oak Ridge National Laboratory, and Rocky Flats Weapons Plant. Radiation Research 136:408-421(1993).
43. Frome EL, Cragle DL, Watkins JP, Wing S, Shy CM, Tankersley WG, West CM. A mortality study of employees of the nuclear industry in Oak Ridge, Tennessee. Radiation Research 148:64-80(1997).
44. Preston-Martin S, Thomas DC, White SC, Cohen D. Prior exposure to medical and dental x-rays related to tumors of the parotid gland. Journal of the National Cancer Institute 80:943-9(1988).
45. Cohen B. Test of the linear no-threshold theory of radiation carcinogenesis for inhaled radon decay products. Health Physics 68:157-174(1995).
46. National Cancer Institute. Estimated Exposures and Thyroid Doses Received by the American People from Iodine-131 in Fallout Following Nevada Atmospheric Nuclear Bomb Tests NIH Publication No. 97-4264. Washington, DC: US Department of Health and Human Services, Natinoal Institutes of Health, National Cancer Institute, 1997.
47. Secretarial Panel for the Evaluation of Epidemiologic Research Activities. Report to the Secretary by the Secretarial Panel for the Evaluation of Epidemiologic Research Activities for the U.S. Department of Energy. Washington, DC: US Department of Energy, 1990.
48. Advisory Committee on Human Radiation Experiments. Final Report, Advisory Committee on Human Radiation Experiments. Washington, DC:US Government Printing Office, 1995.
49. Geiger HJ, Rush D, Michaels D. Dead Reckoning: A critical review of the Department of Energy's Epidemiologic Research. Washington, DC: Physicians for Social Responsibility, 1992.
50. Morgan KZ. Health physics: its development, successes, failures, and eccentricities. American Journal of Industrial Medicine 22:125-33(1992).
51. Lyon J. Nuclear weapons testing and research efforts to evaluate health effects on exposed populations in the United States. Epidemiology 10:557-560(1999).
52. Wilkinson G. Seven years in search of alpha:  The best of times, the worst of times. Epidemiology 10(1999).
53. Sterling TD. The health effects of low-dose radiation on atomic workers: a case study of employer-directed research. International Journal of Health Services 10:37-46(1980).
54. Greenberg M. The evolution of attitudes to the human hazards of ionizing radiation and to its investigators. American Journal of Industrial Medicine 20:717-721(1991).
55. Government Accounting Office. Nuclear Health and Safety: Consensus on acceptable radiation risk to the public is lacking GAO/RCED-94-190. Washington, DC: United States General Accounting Office, 1994.

YOUNG children and old people around the world could be exposed to damaging doses of radiation from nuclear plants and other sources because the database that is used to set safe limits is flawed.

A new analysis by a leading British epidemiologist suggests that the young and old are more sensitive to radiation damage than was previously thought. The international system of radiation safety limits is mostly based on epidemiological studies of 76 000 people from Hiroshima and Nagasaki who were still alive five years after their cities were obliterated by American atom bombs in 1945. The rates at which they have contracted cancers compared with people from other Japanese cities are used by regulatory agencies to estimate the risks of exposing people to radiation from nuclear plants, bomb tests and fallout from accidents such as that at Chernobyl in 1986.

But Alice Stewart, famous for her work in the 1950s revealing the dangers of X-raying pregnant women, argues that the atom bomb survivors are not a normal, homogeneous population. She says her analysis shows that children and old people are more vulnerable to radiation, and that a high proportion of them died between 1945 and 1950 before studies of the Hiroshima and Nagasaki residents began. The young and old are therefore under-represented among the survivors. "The atom bomb data are no good as a basis for radiation safety regulations," she says.

Aged 91, Stewart is an honorary professor at the University of Birmingham School of Medicine. Her study, which is due to be published by the Scientific and Technological Options Assessment unit of the European Parliament in the next few weeks, compares 2601 survivors who suffered from acute radiation injuries with 63 072 survivors who did not. Stewart found that of those with acute injuries, children who were under 10 when the bombs exploded were a thousand times as likely to die of cancer as people aged between 10 and 55. People over 55 at the time of the explosions were twice as likely to die of cancer as those aged between 10 and 55, the study shows. Among those who did not suffer acute injuries, children under 10 at the time of the explosions were three times as likely to die of cancer as other age groups.

Stewart says her results show that the very young and very old are particularly sensitive to radiation. She suggests that the immune systems of the young and old are more easily harmed because they are either still developing or are breaking down. A damaged immune system makes people more vulnerable to cancers and infections, she says.

The National Radiological Protection Board, which advises the British govemment, accepts that the Japanese database is not perfect because it lacks information on the first five years of exposure. But the board's spokesman, Mike Clark, points out that safety limits are also derived from other databases-including Stewart's earlier work on X-rays.