British Nuclear Energy Society. Third International Conference on Health Effects of Low Dose Radiation: Challenges for the 21st Century, Stratford-upon-Avon, UK, 11 - 14 May 1997

Abstract

The BNES has held two previous meetings concerned with the topic of health effects of low dose radiation, one in London in 1987, the other in Bournemouth in 1992. Both previous meetings were well attended and covered some of the more controversial recent issues in radiobiology and epidemiology. At the moment there does not seem to be any burning issue engrossing the minds of the scientific community; nevertheless I think that this third meeting was just as successful as its predecessors, being attended by 130 delegates from more than 20 countries (although, as Chairman, I may be prejudiced). When I said there were no current controversial issues, perhaps I was slightly inaccurate. I think most readers will be aware that the linear no-threshold hypothesis which has formed the basis of radiological protection for so long has come under renewed criticism from certain quarters, especially within North America and France. A number of UK delegates at Stratford took the opportunity to defend the hypothesis, although representatives of the opposing point of view were either absent, or kept rather quiet. In his opening address, Sir Richard Doll reviewed the history of radiation epidemiology and the limitations of this branch of science in telling us about the risks of exposure to low doses of radiation. He concluded that he personally would continue to postulate a linear dose-response relationship at low doses, though with a slope that might be less steep at lower doses than at moderate doses. Support for the view that there is no threshold for the carcinogenic effects of radiation, or, if there is that it must be quite low, was expressed by other speakers from the UK. Richard Wakeford reviewed the evidence from in utero exposure to radiation, concluding that there is strong evidence against a threshold of dose at any level which is meaningful to radiological protection; Colin Muirhead reviewed the evidence from the atomic bomb survivors for the shape of the dose response curve at low doses and the existence of a threshold. Among solid cancers, the evidence for a threshold is not convincing, apart from non-melanoma skin cancer. (The evidence for a threshold is stronger for leukaemia.) In addition, Roger Cox, in his fascinating review of some recent developments in animal models, concluded that we can be reasonably confident of the monoclonal origin of cancer. As I mentioned earlier, proponents of the threshold hypothesis were not much in evidence, with the exception of Philippe Duport (Canada), who argued that the linear non-threshold hypothesis predicted numbers of cancers in excess of those observed in a number of experiments on animals exposed to alpha emitters and who argued that there were indications in certain populations of miners that there might be thresholds determined by a combination of dose and dose rate. David Brenner made a strong plea for radiobiology to play a much bigger role in the design and interpretation of epidemiological studies, which on their own could not provide information about environmentally relevant exposures. He presented some examples of situations where information from one discipline had been used to augment specific areas of uncertainty in the other, mentioning, among other things, the inverse doserate effect for radon observed among uranium miners and independently in laboratory experiments in vivo and in vitro, and once again, evidence from radiobiology for the shape of the dose-response curve at low doses. Roger Cox addressed interaction between scientific disciplines too. Large and expensive animal studies which address only the quantitative aspects of tumorigenesis are now fading from the scene, to be replaced by studies where the emphasis is mechanistic and usually based on molecular genetic approaches. The new animal models, said Dr Cox, could serve as bridges between in vitro studies on DNA damage repair, genomic instability etc, and human epidemiological studies on organ specific carcinogenesis, age-dependence, dose and dose rate effects and so on. He reviewed the acute myeloid leukaemia model in CBA and other mouse strains, intestinal tumorigenesis in the min (multiple intestinal neoplasia) mouse, and studies of the role of the p53 gene in tumorigenesis in mice genetically manipulated to be heterozygously deficient in the gene. The epidemiological studies identified by John Boice as possibly yielding most evidence on the risks of low doses into the next century were identified as follows: Combined studies of nuclear workers; Studies of certain populations exposed after Chernobyl, such as the clean-up workers; Second cancers among survivors of childhood cancer; Studies of domestic radon exposure. The major uncertainties remaining, the challenges facing epidemiology in the next century, concern the expression of lifetime risk after exposure in childhood, the effects of exposure to 131I in childhood, the interaction of radiation with other environmental agents and with genetic factors, individual susceptibility and the possibility of molecular signatures which could uniquely implicate radiation as a cause of specific cancers. Some of these uncertainties are likely to be resolved more quickly than others. Several talks presented at the conference describe work which will contribute to resolution of these issues. Keith Binks presented the results of studies of mortality in workers at BNFL's Springfields and Capenhurst plants, and outlined plans for some quite exciting studies on the BNFL workforce as a whole, notably of mortality and morbidity in relation to both external dose and radionuclide specific internal doses. A number of case-control studies are also in progress or in the planning stage, notably of leukaemia, multiple myeloma and prostate cancer. Sarah Darby addressed problems with estimation of the magnitude of the risk from domestic radon exposure, notably those associated with uncertainties in radon concentrations, and illustrated these with an example drawn from her current case-control study of lung cancer in Cornwall. The distressing increase in the incidence of thyroid cancer in children exposed to fallout from the Chernobyl accident continues and was comprehensively reviewed by Sir Dillwyn Williams. The incidence of childhood thyroid cancers is about one per million per year in most countries, so that in Belarus, which has a population of about 2.3 million children, about 16 cases would have been expected in the period between 1990 and 1996. 457 cases have actually been diagnosed during that time. The incidence of thyroid cancer in exposed children in the Gomel oblast of Belarus is currently nearly 200 times the UK incidence whilst that in children aged 0 - 4 at the time of the accident is considerably greater. There seems to be no question but that the increased incidence is due to radioactive isotopes of iodine released during the accident. Sir Dillwyn stressed the much increased sensitivity of very young children, even when allowance is made for the fact that thyroid doses may have been greater. Another paper from a joint Belarus/Swiss/Italian team provided authoritative estimates of the geographical distribution of thyroid cancer incidence in Belarus and the distribution of radioiodine pollution. These distributions took account of population migration effects and the effect of temporal changes in wind direction during the relevant period. The map showing the two distributions showed a high degree of visual correlation and a calculated correlation coefficient of 0.8. It is still rather early to expect to see any increase in adult cancers following the accident. A number of populations are being closely followed up, however, and a very interesting paper concerned mutations in the p53 gene and other molecular changes in Chernobyl clean-up workers from Estonia. The incidence of cancer to date in these workers is much as expected, but the incidence of p53 mutations appears to be low in the sample studied. The role of mutations in the p53 gene in induction of radiogenic cancer was a topic addressed by several speakers. Andersson and co-workers, for example, had looked at p53 status in thorotrast-related liver cancers. The question was raised, however, of whether searching for mutations in the p53 gene was fruitful in the case of radiation-induced cancers, since radiation causes gross chromosome damage rather than point mutations. Some findings were reported that suggest we should remain aware of possibilities for future surprises. Yuri Dubrova and co-workers reported an apparent high-frequency increase in minisatellite re-arrangements in children born in the Mogilev district of Belarus, suggesting germline mutation in their parents as a result of Chernobyl. Also, several laboratories have now observed that radiation can induce an on-going genomic instability in somatic cells, as a high frequency event, including in bone marrow stem cells as described by Eric Wright. At present it is not known whether or not either of these two phenomena can lead to long-term health consequences. Speakers were cautious in their conclusions concerning the adaptive response and whether or not it could lead to an actual net benefit from exposure at low doses. Hilton Smith reviewed the evidence from in vitro studies, from animal experiments and from epidemiology but thought that the evidence for cellular adaptation expressed as a reduction in the expected incidence of cancer following low doses remained equivocal. It was important to judge the balance between stimulated cellular repair and residual damage. Jerome Pushkin was similarly cautious and listed some outstanding questions which needed to be addressed before any final conclusions could be drawn. Other papers in the adaptive response session addressed the existence, magnitude and possible mechanisms of the adaptive response in a number of in vivo and in vitro systems. Much of the discussion centred around the persistence, usually quite short, of the protective effect after delivery of a conditioning dose. I think the conference overall showed that there are still many exciting questions waiting to be answered concerning the interaction of low doses of radiation with biological systems. Not the least useful aspect of the meeting was the opportunity it provided to epidemiologists and radiobiologists to exchange ideas. As a number of speakers indicated, multidisciplinary approaches may be fruitful in the future. It remains to be seen (perhaps at the fourth BNES conference in 2002) whether this turns out to be the case.