Search

Background: The study cohort of the survivors of the A-bombs in Japan, used as the basis of the internationally accepted estimates of cancer radiation risk, was collected more than 5 years after the bombing and did not include those who died of bomb-related injuries before that date. This paper tests whether the people who survived, in spite of bomb-related injuries, are homogeneous in respect of variation of cancer risk with age with survivors without such injuries., Methods: Appropriate statistical models are derived and fitted to survivor data by maximum likelihood and the resultant statistics used to test the homogeneity assumption., Results: Significant differences were found between those with no injuries and those with multiple injuries and shown to be largely due to exposures before 10 or after 55 years of age having exceptionally high risks of late effects of radiation for survivors showing early effects, i.e. bomb-related injuries., Conclusions: Certain accepted dogmas about the biology of radiation risks in humans, such as 'cancer is the only late effect of radiation' and 'leukaemia is uniquely radiogenic amongst cancers', may be significantly in error. These are discussed.

Newly released data from the Radiation Effects Research Foundation on the survivors of the Hiroshima and Nagasaki A-bombing allow a reassessment of radiation hazards. It appears that deaths from marrow damage (such as aplastic anaemia) continued after 1950. The Life Span Study cohort appears biased in favour of persons with high immunological competence, the result of infants and the elderly being more likely to die before 1950 than young adults. A study of survivors of in utero exposures suggests that embryos are more sensitive to the lethal effects of radiation than more mature foetuses. Current estimates of cancer risks from radiation may only apply to young adults with high immunological competence; young children and the elderly may be at greater risk.

Objectives: To test the "Gardner hypothesis" that childhood leukaemia and non-Hodgkin lymphoma can be caused by fathers' exposure to ionising radiation before the conception of the child, and, more generally, to investigate whether such radiation exposure of either parent is a cause of childhood cancer., Design: Case-control study., Setting: Great Britain., Subjects: 35,949 children diagnosed as having cancer, together with matched controls., Main Outcome Measures: Parental employment as radiation worker as defined by inclusion in the National Registry for Radiation Workers and being monitored for external radiation before conception of child; cumulative dose of external ionising radiation for various periods of employment before conception; dose during pregnancy., Results: After cases studied by Gardner and colleagues were excluded, fathers of children with leukaemia or non-Hodgkin lymphoma were significantly more likely than fathers of controls to have been radiation workers (relative risk 1.77, 95% confidence interval 1.05 to 3.03) but there was no dose-response relation for any of the exposure periods studied; indeed, the association was greatest for those with doses below the level of detection. No increased risk was found for fathers with a lifetime preconception dose of 100 mSv or more, or with a dose in the 6 months before conception of 10 mSv or more. There was no increased risk for the group of other childhood cancers. Mothers' radiation work was associated with a significant increase of childhood cancer (relative risk 5.00, 1.42 to 26.94; based on 15 cases and 3 controls). Only four of the case mothers and no controls were radiation workers during pregnancy., Conclusions: These results do not support the hypothesis that paternal preconception irradiation is a cause of childhood leukaemia and non-Hodgkin lymphoma; the observed associations may be chance findings or results from exposure to infective or other agents. If there is any increased risk for the children of fathers who are radiation workers, it is small in absolute terms: in Britain the average risk by age 15 years is 6.5 per 10,000; our best estimate, using all available data, is that the increase is 5.4 per 10,000. For mothers, the numbers are too small for reliable estimates of the risk, if any, to be made.

Objectives: To discover how the age when a given dose of ionising radiation is received (exposure age) affects the subsequent cancer risk, and whether the types of cancer caused by repeated exposure to small doses during adult life differ from naturally occurring cancers at that age., Method: A nested case-control design with all possible controls in a cohort of nuclear workers, and a Mantel-Haenszel test (requiring only one degree of freedom) to discover whether there was any level of exposure age where the null hypothesis of no effects of radiation was rejected. This analysis was followed by inspection of how different types of cancers were related to the cancer risk., Results: For radiation received at least 15 years before a cancer death (to allow for cancer latency) evidence of a dose related risk was found which was largely the result of exposures during the last 10 years of working life (between 55 and 65 years of age). The relative frequency of site specific cancers showed no signs of being different for radiogenic and idiopathic cancers, and there was no evidence of the exceptionally strong association between radiation and leukaemia found in atomic bomb data and other high dose situations., Conclusions: Sensitivity to carcinogenic effects of radiation increases progressively with age during adult life and, provided the dose is too small to produce many cell deaths, the ratio of leukaemias to solid tumours is no different for radiogenic and idiopathic cancers.

Objectives: To discover whether direct estimates of the risks of cancer for nuclear workers agree with indirect estimates based on survivors of the atomic bomb; whether relations between age at exposure and risk of cancer are the same for workers and survivors, and whether dosimetry standards are sufficiently uniform to allow pooling of data from different nuclear industrial sites., Method: Data from five nuclear sites in the United States were included in a cohort analysis that as well as controlling for all the usual factors also allowed for possible effects of three cancer modulating factors (exposure age, cancer latency, and year of exposure). This analysis was first applied to three distinct cohorts, and then to two sets of pooled data., Results: From each study cohort there was evidence of a risk of cancer related to dose, and evidence that the extra radiogenic cancers had the same overall histological manifestations as naturally occurring cancers and were largely the result of exposures after 50 years of age causing deaths after 70 years. There were, however, significant differences between the five sets of risk estimates., Conclusions: Although the risks of cancer in nuclear workers were appreciably higher than estimates based on the cancer experiences of survivors of the atomic bomb, some uncertainties remained as there were non-uniform standards of dosimetry in the nuclear sites. The differences between nuclear workers and survivors of the atomic bomb were largely the result of relations between age at exposure and risk of cancer being totally different for workers and survivors and, in the occupational data, there were no signs of the special risks of leukaemia found in atomic bomb data and other studies of effects of high doses.

Reanalysis of A-bomb survivor data has shown the following: a) in the high-dose (> 1 Gy) subgroups of the life span study cohort of 5-y survivors, there is a significant deficit of individuals who were < 10 y or > 50 y at the time of the bomb; and b) in the cohort on in utero children, there is a significant deficit of individuals who were < 8 wk of fetal age when exposed. This paper discusses how this selection bias has affected the perception of three effects of A-bomb radiation: marrow damage, carcinogenesis, and second-generation effects.

Reanalysis of Hanford data by a method, which is new only in the sense that it makes new uses of standard epidemiological procedures, has produced evidence of a cancer risk at low dose levels. By a conservative estimate, about three per cent of the pre-1987 cancer deaths of Hanford workers had occupational exposures to external radiation as the critical (induction) event. These radiogenic cancers were evenly distributed between five diagnostic groups, but as a result of there being much greater sensitivity to "cancer induction by radiation" after, rather than before, 50 years of age, they were concentrated among the cancers which proved fatal after 70 years of age. The reanalysis provides no support for the idea that radiation is more likely to cause leukemia than solid tumors, or the idea that there is reduced cancer effectiveness of radiation at low dose levels (dose rate effectiveness factor or DREF hypothesis), but the estimated proportion of radiogenic cancers was much higher for the 175 nonfatal cancers (which had other certified causes of death) than for the 1,732 fatal cases. Finally, according to the latest publication of the US Committee on Biological Effects of Ionizing Radiation (BEIR V), dose rate is more important than exposure age, and even a single exposure to 10 rem would only increase the normal cancer risk by four percent. Nevertheless, for all recorded exposures of Hanford workers, the estimated doubling dose was close to 26 rem; for exposures after 58 years, it was close to 5 rem, and for exposures after 62 years, it was less than 1 rem.

In 1964, the Atomic Energy Commission agreed to sponsor "a study of the lifetime health and mortality experiences of all employees of AEC contractors." The commission put in charge of this study a physician (Thomas Mancuso) who had recently shown how the U.S. Social Security system could be used to identify the dates and causes of death of all insured workers. As director of the AEC project, Mancuso was at liberty to include any or all the postwar offshoots of the Manhattan Project. His master plan included workers from Oak Ridge, Los Alamos, and Hanford, but it soon became apparent that his attempts to link radiation exposures to subsequent events were proving more successful at Hanford than elsewhere. The authors of this paper, who participated in the study, review the controversy surrounding its eventual publication.

Cancer risk coefficients for ionizing radiation are currently based on the assumption that, after the bombing of Hiroshima and Nagasaki, there were no late effects of early selection (survival of the fittest) or acute marrow damage. These negative findings were the result of applying a linear model of relative risk to the deaths of 5-y survivors. By applying a linear-quadratic model to these deaths (i.e., a model with more than one degree of freedom), we have obtained evidence of longstanding competition between selection effects of the early deaths and other radiation effects, and also evidence that late effects of radiation include marrow damage as well as cancer. Consequently, the present method of risk estimation--by linear extrapolation of high dose effects--should no longer be used for estimating the cancer effects of occupational exposures or background radiation.

A modified Mantel-Haenszel analysis of data from the Oxford Survey of Childhood Cancers has shown that cases associated with foetal irradiation (X-rayed cases) accounted for a higher proportion of deaths between 5 and 10 years than of earlier or later deaths. This finding is compatible with somewhat later origins for the cancers actually caused by the radiation exposures (radiogenic cases) than for other (idiopathic) cases which proved fatal before 10 years of age. Therefore the usual time for incurring congenital anomalies (or the first trimester of foetal life) could be the commonest time for initiating childhood cancers. The theoretical implications of this and other findings of the Oxford Survey are discussed within the framework of a theory which assumes that all mutant cells have cancer potentialities and that defects in the immune surveillance mechanism favour multiplication of these cells (or endogenous sources of self-replicating foreign proteins) as well as live pathogens (or exogenous sources of self-replicating foreign proteins).

In a case/control study of 8059 matched pairs, the effect of maternal exposure to drugs and illnesses during pregnancy on the relative risk (RR) of cancer in the child was investigated using conditional logistic regression techniques. Acute respiratory infections, particularly viral infections such as influenza, were associated with a significantly increased RR of all childhood cancers and of neoplasms of the reticulo-endothelial system (RES) in particular, (RR = 1.69 all cancers, RR = 1.81 RES neoplasms, RR = 1.59 solid cancers). An analysis of illnesses according to their physiological effects yielded a significant association between childhood leukaemia and febrile illnesses (RR = 1.27 RES neoplasms). A significant increase in RR was associated with maternal history of epilepsy (RR = 1.31 all cancers) rather than with exposure to anticonvulsant drugs. Vaccines showed a pattern of RR similar to that of acute viral infections. Consumption of antipyretics and analgesics significantly increased the RR of childhood cancer (RR = 1.36 all cancers). An analysis of drugs according to their metabolic reactions yielded a significant association between those undergoing amino acid conjugation (predominantly antipyretics and analgesics) and childhood cancer risk (RR = 1.76 solid cancers).

An analysis of data collected during the course of the Oxford Survey of Childhood Cancer has shown that it is possible to recognize different facets of memory bias without systematic checking of individuals' records, and to make use of the biased data. The position of foetal irradiation in the aetiology of childhood cancers has been re-affirmed, but there is no support for the idea that exposure of parental gonads to diagnostic X-rays is conducive to cancer in the next generation.

Evidence of early loss of immunological competence in cases of neoplasms occurring in juveniles was found in an analysis of OSCC data (Oxford Survey of Childhood Cancers). The effects observed included heightened sensitivity to infection from birth onwards for all types of childhood cancer, higher levels of sensitivity for leukaemia than for lymphomas, and higher levels for lymphomas than for other solid tumours. The findings as a whole are consistent with in utero loss of immunological competence, which is an essential promoter of cancers of foetal origin and thus allows the outcome of an in utero cancer induction to be influenced both by maternal levels of immunological competence and postnatal infection.

Though most of the production work at Hanford is done by manual workers, 46% of the most dangerous jobs are performed by people who have professional or technical qualifications. For these privileged workers occupational mortality risks are positively correlated with radiation doses but for manual workers, who have relatively high death rates, there is an inverse relation with dose. The high ratio of professional to manual workers is clearly the reason for the industry having fewer observed than expected deaths and the inverse relation with dose for less privileged workers is probably a sign that there has been selective recruitment of the most highly paid manual workers--that is, skilled craftsmen into the more dangerous occupations. Evidence of this selective recruitment was obtained by equating danger levels with levels of monitoring for internal radiation. Therefore, there should be some control for these levels in any analysis of cancer effects of the measured dose of radiation.

Data of 10,556 case-control pairs from the Oxford Survey of Childhood Cancers and related sources have shown that when cancers originate in the reticuloendothelial system (RES neoplasms) they are liable to cause loss of immunological competence before they are clinically recognizable. Since these early effects may have lethal consequences, the true prevalence of RES neoplasms is difficult to identify, especially in infection-sensitive age groups and populations with high death rates from infection. An inevitable consequence of a nuclear holocaust is a high infection death rate. Therefore, a population of A-bomb survivors is a totally unsuitable one for studying the precise nature of the association between ionizing radiation and human cancers.

A Mantel-Haenszel analysis of fetal irradiation subfactors indicated that most of the "extra" X-rayed cases in the Oxford Survey of Childhood Cancers were radiation induced. First trimester exposures were rare but probably ten times more dangerous than later exposures. Ratios of observed: expected numbers of cancer deaths were lower for children with abnormal X-rays than for other X-rayed children, and lower for recent than remote exposures. The first of these differences was probably due to several antenatal conditions having positive associations with obstetric radiography and several causes of early (noncancer) deaths; the second one was probably due to a progressive lowering of film doses between 1940 and the present time. A rare cause of fetal irradiation (hydramnios), whose associations with congenital defects are well documented, led to the discovery that two faults in the International Classification of Diseases and Causes of Death have contributed to mistaken ideas about the etiology of childhood cancers: Neoplasms were not listed among the official causes of stillbirths, and cystic tumors of the kidneys and lungs of infants were not listed as neoplasms.

This paper reports on results from the study initiated by Mancuso into the health risks from low-level radiation in workers engaged in plutonium manufacture at Hanford Works, Washington State, USA, and attempts to answer criticisms of previous reports by an in-depth study. Previous reports have aroused much controversy because the reported risk per unit radiation dose for cancers of radiosensitive tissues was much greater than the risk generally accepted on the basis of other studies and widely used in setting safety levels for exposure to low-level radiation. The method of regression models in life-tables isolates the effect of radiation after statistically controlling for a wide range of possible interfering factors. Like the risk of lung cancer for uranium miners the dose-response relation showed a significant downward curve at about 10 rem. There may, therefore, be better agreement with other studies, conduct at higher doses, than is widely assumed. The findings on cancer latency (of about 25 years) and the effect of exposure age (increasing age increases the risk) are in general agreement with other studies. An unexplained finding is a significantly higher dose for all workers who developed cancers in tissues that are supposed to have low sensitivity to cancer induction by radiation.

Data from the Oxford Survey of Childhood Cancers were subjected to the Mantel-Haenszel procedures to recognize independent effects of associated factors in retrospective data. Our results showed that several birth factors, including fetal irradiation, social class, maternal age, and sibship position, exerted separate effects on childhood cancers in general, and reticuloendothelial system neoplasms in particular.

A study based upon an unusually large series of childhood cancers and matched controls found a significant deficit of case/control pairs in which the cancer case had fewer immunizations against infectious diseases than the matched control. All types of immunizations and cancers were affected but the case/control differences were more pronounced for older cases with late immunizations than for younger cases with early immunizations, and more pronounced for solid tumours than leukaemia. Therefore there may be immune system responses to immunizations (or simulated infections) which make it difficult for small clones of cancer cells to enlarge and are more successful in preventing localised tumours in adolescents than childhood leukaemias.

This paper continues the series by Mancuso, Stewart, and Kneale (MSK) on studies of cancer risks for radiation workers at Hanford. It concentrates on the statistical problems posed by the need to estimate and control for job related mortality risks when there are several changes of occupation and no certainty about how different occupations are related to two socioeconomic factors which have strong health associations--namely, education and income. The final conclusion is that for tissues which are sensitive to cancer induced by radiation there is a risk of cancer for Hanford exposures whose dose response is curvilinear with long latency and increasing effect with increasing exposure age.

A slight rearrangement of the data included in a recent report from the Radiation Effects Research Foundation (RERF) has shown differences between cardiovascular and other non-malignant diseases of A-bomb survivors which probably result from two factors: selection effects of early infection deaths and residual effects of marrow damage. Both effects were dose related but neither was obvious becasue one reduced the risk of later infection deaths and the other increased the risk. Allowance for these factors is bound to alter present RERF estimates for cancer effects of radiation and the change will probably be in an upward direction, thus bringing these estimates closer to ones based on radiation workers.