tions. One is the possibility of internal contamina- tion by a radionuclide (as adumbrated in the leading article by Dr Valerie Beral).3. There are several reasons ...
CORRESPONDENCE * All letters must be typed with double spacing and signed by all authors. * No letter should be more than 400 words. * For letters on scientific subjects we normally reserve our correspondence columns for those relating to issues discussed recently (within six weeks) in the BMJ.
* We do not routinely acknowledge letters. Please send a stamped'addressed envelope ifyou would like an acknowledgment. * Because we receive many more letters than we can publish we may shorten those we do print, particularly when we receive several on the same subject.
Leukaemia and lymphoma among young people near Sellafield SIR,-In a careful and detailed case-control study Professor Martin J Gardner and colleagues have produced strong evidence of a causal link between occupational radiation exposure of fathers at Sellafield and the subsequent occurrence of acute leukaemia in their children.' The authors suggest that this is most probably due to a germ cell mutation, but the alternative possibility is that fathers contaminate the home environment and this leads to irradiation of the fetus in utero or the infant in the early months of life. I have recently published a theoretical model of childhood leukaemogenesis based on spontaneous mutation in dividing cells that is relevant to this problem.2 I derived an equation that enables the calculation of the probability of a malignant clone arising as a result of a number of specific independent mutations (n) occurring in a population of dividing cells with a mutation rate per gene per cell generation of m. The epidemiological and kinetic data on childhood acute lymphoblastic leukaemia are best explained if m'- 1 x 10 6 n=4. Furthermore, it can be shown that if the mutation rate is constant throughout all cell generations during intrauterine and early extrauterine life then the incidence of acute lymphoblastic leukaemia is proportional to m4. Thus a 10-fold increase in the incidence, as seen in Seascale, would occur if the mutation rate was increased by a multiple of 1 8. Interestingly, if the increased mutation rate is confined to extrauterine life-that is, the last few cell generations-then the increase in incidence would be proportional to m. In other words a 10-fold increase in the mutation rate would be required to increase the incidence 10 times. The model can also be used to calculate the effects of a germ cell mutation. If the fetus inherits one defective gene then only three subsequent mutations are required. The chance of acute lymphoblastic leukaemia arising in these circumstances is between 0 1 and 0-9 per affected child compared with the chance in the general population of 3x 10 by age 15 years. A 10-fold increase in incidence due to germ cell mutation would require a mean frequency of mutation in spermatozoa of at least 3 x 10 ' if all affected infants develop acute lymphoblastic leukaemia and 3x 10 2 if only 10% develop it. In fact these figures are underestimates because they ignore the possibility of selection against mutant spermatozoa. The most commonly quoted estimate of the mutant frequency of genes, which is based on in vitro measurement and the appearance of new mutants, is 2x 10 5.' Thus a 10-fold increase in the incidence of acute lymphoblastic leukaemia due to germ line mutation would require a 100-fold or 1000-fold increase in
676
the mutation rate. It is difficult to believe that Sellafield workers could sustain this without a. measurable effect on general' health and a pronounced increase in the incidence of other cancers, particularly considering that smoking increases the mutant frequency by only 50%.' Thus a 10-fold increase in the incidence of acute lymphoblastic leukaemria as a result of a radiation induced increase in the mutation rate would require a 100-fold or 1000-fold increase if acting on the germ cell; a 10-fold increase if acting on lymphocytes during early extrauterine life; but only a 1 8-fold increase if acting throughout intrauterine life, which seems the most plausible mechanism even though the pathways of exposure are unclear. J A MORRIS Department of Pathology, Lancaster Moor Hospital, Lancaster LA I 3JR
non-radioactive derivatives might be expected to have a profound effect. If radionuclides were localised in spermatogonia any mutations induced would bc amplified during spermatogenesis and could lead to) a pcrceptiblv higher incidencc of genetic lesions in the offspring. If such a mechanism were to be ani important factor in the pathogenesis of childhood leukaemia found in the study of Professor Gardner and colleagues there might be a case for assessing the nutritional zinc state of workers in nuclear installations to ensure that those in contact with radionuclides are not deficient in this essential trace element. Zinc is known to alleviate the toxic effects of metals such as cadmium, either by the induction of increased concentrations of metallothionein or by binding to and masking vital sulphur containing moieties, and the absorption of metals such as iron, copper, and zinc is interdependent.' PATRICK A RILEY
1 Gardner MJ, Snee MP, Hall AJ, Powell CA, Downes S, Terrell JD. Results of a case-control study of leukaemia and lymphoma among young people near Sellafield nuclear plant in West Cumbria. BrMedJ7 1990;300:423-9. (17 February.) 2 Morris JA. A mutational theory of leukaemogenesis. J Clin Pathol 1989;42:337-40. 3 Cole J, Green MHL, James SE, Henderson L, Cole H. A further assessment of factors influencing measurements of thioguanine-resistant mutant frequency in circulating T lymphocytes. Mutat Res 1988;204:493-507.
SIR,-The results of Professor Martin J Gardner and colleagues,' which showed the high relative risk of leukaemia in children whose fathers had been exposed to external radiation in the period before conception, when contrasted with the absence of any similar tendency in the offspring of Japanese men who survived the atomic bomb explosions2 suggest several mechanistic interpretations. One is the possibility of internal contamination by a radionuclide (as adumbrated in the leading article by Dr Valerie Beral).3 There are several reasons for serious consideration to be given to the possibility of such internal contamination, particularly in the light of thetissue distribution of metallic elements such as zinc.4 Relatively high concentrations of zinc are found in semen' and there is a close association between zinc and DNA in the so called "zinc finger" domains of DNA binding proteins.6 Jonas et alf proposed that substitution of zinc at these vulnerable sites by other cations, such as transition metals (for example, copper and iron), could lead to damage of DNA by the generation of reactive oxygen species of a type known to be produced by radiation,' and potentially mutagenic alterations in DNA bases by hydroxyl radical attack are known to occur.9 Clearly, a similar substitution by relatively short lived radionuclides that are able to bind to thiols at zinc sites adjacent to the genome could be a source of local radiation induced metagenesis, and even the presence of unnatural
Department ot Chemical Pathology, University College and Middlesex School of Medicine, London W 11P 6DtB ROBIN L WILLSON
Department of Biology and Biochemistry, Bruncl University, Uxbridge 1 Gardner MJ, Snee MP, Hall AJ, Powell CA, D)ownes'S, Terrell JD. Results of case-control study of leukaemia and lymphoma among young people near Sellafield nuclear plant in West Cumbria. BrMedj 1990;300:423-9. (17 February.) 2 Ishimaru T, Ishimaru M, Mikami M. Leukaemia incidence among individuals exposed in utero, children of atomic bomb survitvors and their controls, Hiroshima and Nagasaki, 1945-79. Hiroshima: Radiation Effects Research Foundation, 1981. (RERF technical report No 11-81.) 3 Beral V. Leukaemia and nuclear installations. Br Med J 1990;300:411-2. (17 February.) 4 Willson RL. Iron, zinc, free radicals and oxygen in tissue disorders and cancer control. Ciba Found Symp 1977;51: 333-54. 5 Underwood EJ, Somers M. Studies of zinc nutrition in sheep: the relation of zinc to growth, testicular development and spermatogenesis in young rams. Australian Journal of Agricultural Research 1969;20:889-97. 6 Miller J, McLachlan AD, Klug A. Repetitive zinc-binding dpmains in the protein transcription factor IIIA from Xenopus occytes. EMBOJ 1985;4:1609-14. 7 Jonas SK, Riley PA, Willson RL. Hydrogen peroxide cytotoxicity. BiochemJ 1989;264:651-5. 8 Willson RL. Zinc and iron in free radical pathology and cellular control. In: Mills CF, ed. Zinc in human biology. Berlin: Springer Verlag, 1989:147-73. 9 Arouma OI, Halliwell B, Dizdaroglu M. Iron ion-dependent modification of bases in DNA by the supcroxide radicalgenerating system hypoxanthine/xanthine oxidase. j Biol Chem 1989;264:13024-8. 10 Sandstrom B, Lonnerdal B. Promoters and antagonists of zinc absorption. In: Mills CF, ed. Zinc in human biology. Berlin: Springer Verlag, 1989:57-78.
SIR,-Clusters of cases of childhood leukaemia have been shown to occur around the Sellafield and Dounreay nuclear reprocessing facilities in the United Kingdom.' Radiation exposure to the general population in the vicinity of these establishments cannot account for such excesses on current
BMJ VOLUME 300
10 MARChI 1990
Number of cases of childhood cancer in Telford Leukaemia anid lymphoma* Observed Expected 1961-7(0 1971-85
9 26
Solid tumourst Observed Expectcd 16 28
10.9 17-5
12-3 16-4
tAll cases of leukaemia and non-Hodgkin's lNmphoma. ,AII other cases excluding those of leukaemia and non-Hodgkin's Iymphoma. tUor all results (observed v expected).
All tumours Observed Expected 25 54
23-2 33-9
p Valuet
NS 0-05
bigger population. Moreover, the notion that a worker should be employed not for his skill or labour but for his body's ability to absorb radiation would break new ground in the ethics of occupational health. JOHN COULTER 'I'he Senate, IParliament House, Canberra, Australia 260() 1 Gardner MJ, Snee MP, Hall AJ, Powell CA, Downes S, rerrell
dose-response relations. Research by Professor Martin J Gardner and colleagues on the cases occurring around the Sellafield plant in particular shows raised relative risks in cases in which the father was employed at the plant and whose employment entailed increased exposure to radiation.2 The effect of industrial and chemical processes, however, cannot be ignored. Furthermore, radiation exposure cannot explain the presence of other clusters of cases of leukaemia that have been shown to occur in areas with no nuclear reprocessing plants.' Kinlen proposed an alternative explanation in that when new centres of population are developed in isolated communities cases of leukaemia may occur as an abnormal response to infection or to specific leukaemogenic viruses introduced by the resulting population mixing.4 We therefore investigated the incidence of all types of childhood cancer in the vicinity of Telford, Shropshire. This is a designated new town where a large influx of residents occurred in the early 1970s, analogous in many ways to that which occurred in the areas surrounding Sellafield and Dounreay, but the town does not have a similar source of radiation. Shropshire is one of five counties covered by the West Midlands Regional Children's Tumour Registry, which collects detailed information on all cases of childhood cancer back to 1957. It has therefore been possible to investigate Kinlen's hypothesis by comparing the incidence of childhood cancer in the 10 years (1961-70) before the major population expansion with that occurring in the 15 years (1971-85) after population influx and consequent mixing. Observed cases diagnosed during these two periods were compared with the calculated expected numbers on the basis of the local childhood population and the average regional incidence of cancer for the same periods. The results (table) showed an increase over expected numbers of cases of leukaemia and lymphoma and "solid tumours" during 1971-85. The results for cases of leukaemia and lymphoma in isolation seem to support Kinlen's population mixing hypothesis, but the observation of an increase in the number of cases of solid tumours would not be predicted on the basis of his hypothesis, which implicated leukaemogenic viruses. Our observations suggest that alternative explanations may be more appropriate, such as the environmental aspects of the increasing urbanisation of the area or the less likely possibility of improvements in cancer registration for the Telford area, in particular during 1971-85. The results also highlight the relatively small number of cases in such studies. K R MUIR R GRAHAM
S E PARKES I R MANN
West Midlands Regional Children's Tumour Registry, Children's Hospital, Birmingham B16 8ET
I Committee on Medical Aspects of Radiatiott in the Environment. Second report. London: HMSO, 1988. 2 Gardner MJ, Snee MP, Hall AJ, Powell CA, Downes S, Terrell JD. Restllts ot' case-conitrol study of letukaemia and lympltoma among yotung people ttear Sellafield tauclear plant in West C>,umbria. BrAcdJ7 1990;300:423-9. (17 Fehruar%. 3 Openshaw S, Charlton M, Craf't AW, Birch J. Investigation of leukaemia clulsters by use ott a geographical analysis machine.
Lancet 1988;i:272-3. 4 Kinlen L. Evidence for an infective cause of childhood letikaemia: comparison ott a Scottish new town with ttuclear reprocessing sites in Britaitt. Iancet 1988;ii: 1323-7.
BMJ
VOLUME 300
10 MARCH 1990
JD. Restults of case-control study of leukaemia and lymphoma SIR,-The study by Professor Martin J Gardner among young people ncar Sellafield niuclear plant in West and colleagues of the incidence of leukaemia in the Cumbria. Br Medj 1990;300:423-9. (17 February.) children of male workers at Sellafield' suggests 2 Beral V. Leukaemia and nuclear installations. Br Med Jf 1990; 300:411-2. (17 February.) extremely useful directions for further scientific A. In: An epidemiologist takes a look at radiattion risks. study as well as having important ethical con- 3 Stewart Washington: United States Department of Health, Education, sequences for exposure of workers in radiation and Welfare, 1973:43-70. environments. Not all of these points were made in Dr Valerie Beral's editorial.2 Firstly, how do you decide whether the leu- SIR,-Professor Martin J Gardner and colleagues kaemia is due to mutation of sperm stem cells have provided grounds for suspecting that paternal caused by exposure to radiation or to the father irradiation in the period before conception may be carrying radiation home by some as yet undeter- the cause of the excesses of cases of childhood mined route and thus exposing the child to direct leukaemia and non-Hodgkin's lymphoma near radiation? Stewart's massive study in Oxford of the Sellafield.' This cannot have been one of the main consequences of radiation exposure during fetal hypotheses of their study, as in that case the life showed an increased incidence of several controls would have been matched to the cases on cancers in those exposed in utero.' Each cancer the basis of length of residence in the parish or area showed a limited age at onset. All had occurred before conception.' The residential pattern in by the age of 10 with the incidence of myeloid Seascale has been unusual, with houses alle'cated leukaemia and lymphomas peaking in the ninth and often changed on the basis of grade and and 10th years of life and lymphatic leukaemia in occupation and with much movement into and the third and fourth years. If the Sellafield cancers within the parish by people from all over Britain. were caused by mutations in the sperm you would Many people rented houses there for only a short expect the latency to be no longer than these period, often in the early part of their residence intervals. If the mutations resulted from postfetal in West Cumbria. For such workers there is radiation exposure the latency may well have been limited opportunity for radiation exposure while in longer. Seascale. It is therefore relevant to ask about the Secondly, ifthese cancers resulted from exposure numbers of fathers of control children who had of the sperm stem cells to radiation causing lived as long in Seascale before the conception of mutations that were passed to the children (whether the children as had the corresponding fathers from external radiation or absorbed radionuclides of children with leukaemia or non-Hodgkin's lodging near or in the testes is immaterial), then the lymphoma: matching on this variable might alter cancers should be looked on as only one index of the estimates of relative risk. genetic damage. Mutation results when DNA Given the high proportion of controls lost damage is followed by faulty repair. During to the radiation analysis, it would be unwise the process of repair an incorrect sequence is not to consider the possibility of other biases, so established, which then passes to all descendent potentially important are the findings. If failure to cells. There is no a priori reason to expect- that return a questionnaire often meant that radiation radiation would specifically target or damage, or details could not be traced (because of lack of a date both, the sequence that potentially codes for of birth) the findings could reflect less cooperation leukaemia. It is far more likely that the damage and among appreciably irradiated fathers than among subsequent faulty repair randomly distributed fathers of other controls in Seascale, where so through the DNA of the sperm. Some of these much publicity has been given to a possible link mutations will be lethal, and you might therefore between radiation and leukaemia. expect to find an increased rate of spontaneous If the four cases in Seascale that were excluded abortion among women who are pregnant by these from the study for technical reasons were, like fathers. There should also be an increase in a most of the seven cases studied, also associated variety of genetically determined consequences in with paternal radiation before conception then we the surviving children who do not have leukaemia. could have an adequate explanation of the excesses IThirdly, those children with leukaemia will have in terms of this claimed new risk factor. These four the mutated gene in every cell in their body, cases, which have contributed to concern as much including their germ line cells. If they are treated as the others, are cases 1, 2, and 16 listed in and survive to reproductive years their offspring the Black report' and also the case discovered by will be similarly affected. Gardner et al in which death occurred shortly after Lastly, the application of the results of Professor leaving Seascale. If, say, none (or only one) of the Gardner's and colleagues' study to setting radiation fathers of these four children had been exposed to standards raises many unusual ethical possibilities. radiation (in any nuclear establishment) before the If maximum allowable levels of exposure are conception of their children the combined picture retained should men who have not completed their of all 11 cases would look rather different. This families be excluded from work that entails would then not justify dismissing an infective exposure to radiation? If men in their reproductive basis'" from having a role, as Professor Gardner years are to work at these jobs the allowable levels and colleaguesimply. It is tantalising not to know may need to be reduced to no more than 1 mSv per about the overall picture. year. Is this achievable? Should those already I. J KINLEN exposed be advised not to have children? The Cancer Rcsearch Campaign Epidemiology Unit, largest uranium mine in the world has recently Uttiversity of Edinburgh, begun operation in south Australia. Its under- Edinburgh ElH- 9JZ ground miners are exposed to an average of 15 mSv of radiation per year, and this is with high I Gardner MIJ, Snee Ml', Hall AJ, IPowell CA, Downes S. 'Ferrell JD. Restilts of case-control study of leukacmia and lymphoma standards of hygiene. Increasing the turnover of among young people near Sellafield nuclear plant in West miners will not reduce the total number of cases of Cumbria. BrMedj 1990;300:423-9. (17 February.) leukaemia but will merely spread the risk over a 2 Gardner MJ, Hall AJ, Snee MI", I)owttes S, P'owell CA, Terrell
677
