- Email: [email protected]
Late radiation responses in man: Current evaluation from results from Hiroshima and Nagasaki
A~. S~zc~ ~ Vol.3,Nu.8,pp.231—239,l983 Printed in Great Britain. All rights reserved.
0273—1177/83 $0.00 + .50 Cop~’rjght ©COSPAR
LATE RADIATION RESPONSES IN MAN: CURRENT EVALUATION FROM RESULTS FROM HIROSHIMA AND NAGASAKI William J. Schull Center for Demographic and Population Genetics, Graduate School of Biomedical Sciences, University of Texas Health Science Center, P.O. Box 20334, Houston, TX 77025, U.S.A.
ABSTRACT Among the late effects of exposure to the atomic bombings of Hiroshima and Nagasaki, none looms larger than radiation related malignancies. Indeed, the late effects of A—bomb radiation on mortality appear to be limited to an increase in malignant tumors. At present, it can be shown that cancers of the breast, colon, esophagus, lungs, stomach, thyroid, and urinary tract as well as leukemia and multiple myeloma increase in frequency with an increase in exposure. No significant relationship to radiation can as yet be established for malignant lymphoma, nor cancers of the rectum, pancreas or uterus. Radiation induced malignancies other than leukemia seem to develop proportionally to the natural cancer rate for the attained age. For specific age—at—death intervals, both relative and absolute risks tend to be higher for those of younger age at the time of bombing. Other late effects include radiation—related lenticular opacities, disturbances of growth among those survivors still growing at the time of exposure, and mental retardation and small head sizes among the in utero exposed. Chromosoinal abnormalities too are more frequently encountered in the peripheral leucocytes of survivors, and this increase is functionally related to their exposure. Some uncertainty continues to surround both the quantity and quality of the radiation released by these two nuclear devices, particularly the Hiroshima bomb. A recent reassessment suggests that the gamma radiation estimates which have been used in the past may be too low at some distances and the neutron radiation estimates too high at all distances; moreover, the energies of the neutrons released now appear “softer’ than previously conjectured. These uncertainties are not sufficiently large, however, to compromise the reality of the increased frequency of malignancy, but make estimates of the dose response, particularly in terms of gamma and neutron exposures, tentative. INTRODUCTION Thirty—five years of studies of the survivors of the atomic bombings of Hiroshima and Nagasaki have established a number of late effects of radiation exposure: some unequivocally some less c~rtainly. Among the former are the occurrence of radiation—related lenticular opacities; an increase in leukemia and a variety of malignant solid tumors such as those of the breast, lung, stomach and thyroid; disturbances of growth among survivors who had not achieved their adult dimensions at the time of exposure; an increase in mental retardation and small head sizes among those exposed in utero; and an increase in the frequency of chromosomal aberrations in peripheral lymphocytes. Less definite is an increase in certain aspects of humoral and cell mediated immunity, or in the frequencies of cancers of the colon, esophagus, salivary glands and urinary tract. These studies have also sought but failed to find a variety of other conceivable elfects, such as an impingement on fertility or a nonspecific life—shortening. Either no effect has occurred in these instances or the nature or extent of the effect lies outside present measurement capabilities. We now turn to a more measured account of the data on which these assertions rest. THE FOLLOW—UP STUDIES Throughout most of the first decade of the existence of the Atomic Bomb Casualty Commission (AECC) — the agency of the National Academy of Sciences—National Research Council charged with the responsibility for the study of the late effects of exposure to atomic radiation, most studies of the survivors were opportunistic both in the research directions they pursued and in the samples they employed (see [1] for a thorough account of the history of this agency). A notable exception was the Genetics Study [2,31. It was not until 1956
231
W.J. Schull
that the first steps toward a unified program were taken. The latter centered on the investigation of a fixed population of survivors and controls, based largely upon the nationwide enumeration of survivors in 1950 as a part of Japan’s first post—war decennial census [4]. The program entails (1) a mortality surveillance of approximately 110000 individuals (82000 of whom were exposed), the so—called Life Span Study (LSS); (2) a series of now biennial, but initially annual, physical examinations of a subset of approximately 20000 members of the Life Span Study, the Adult Health Study; and finally (3) an autopsy program aimed at the postmortem examination of as many members as practicable of the LSS sample. For a variety of reasons this latter aspect of the unified program has been deemphasized recently, but not before some 10000 autopsies had occurred [5]. In both cities tumor and tissue registries, managed by local medical associations, are available to support this program of inquiry. There is also available a number of ad hoc samples, such as those on which studies of those individuals exposed in utero rest. ESTIMATES OF EXPOSURE It was not until 1957 that the first estimates of individual exposures became available. Prior studies had attempted to relate possible effects either to the distances of survivors from the hypocenter or to the presence of symptoms associated with acute radiation sickness or both. These preliminary estimates were of limited utility, however, for they were largely available only on those individuals exposed in the open, a minority of the survivors within 2000 meters of the blast hypocenter in both cities. This limitation spurred a more comprehensive program of dose assessment, one culminating in the so—called T65 dose estimates [6,7]. The latter have been the basis of most appraisals of dose—response relationships until quite recently. Under the T65 dosimetry, survivors within 1600 meters of the hypocenter in Hiroshima (2000 in Nagasaki) were assigned separate gamma and neutron exposures based upon their individual distances from the hypocenter and the attenuation of the “free—in—air” doses at those distances attributable to the shielding they may have experienced. Shielding factors were estimated experimentally [61. Individuals exposed beyond 1600 or 2000 meters were assigned the gamma and neutron exposures which prevailed at their specific distances uncorrected for the effects of shielding; it should be noted that at 1600 and 2000 meters, the total exposure in terms of tissue kerma (kinetic energy released in tissue) was then thought to be about 19 and 18 rads for Hiroshima and Nagasaki, respectively. Recent reassessments based on further data have called into question the T65 estimates [8,9]. These newer appraisals suggest (1) that the gamma exposures at some distances may be greater than previously thought but the neutron exposures are less at all distances, and (2) the energies associated with the neutrons released was much lower than previously conjectured. At present it is not possible to estimate individual exposures under the newly proposed dosimetry save for those individuals who were exposed in the open, a minority of the proximally exposed, for new building and body shielding factors are not as yet available. It appears clear, nonetheless, that these factors will be different from those associated with the T65 dosimetry. Marcum [10] has suggested that the average gamma shielding factors for Hiroshima and Nagasaki may be more nearly 0.55 and 0.50 rather than the 0.90 and 0.81 given for these two cities by Milton and Shohoji [7]. It should be noted that to the average, than those previously or a malignant solid tumor, or kerma or absorbed dose will be which follow.
extent these reassessments lead to lower exposures, on the assumed to be true, the absolute risk of death from leukemia the occurrence of a lenticular opacity and the like per unit increased over those specific values cited in the paragraphs
RADIATION RELATED MALIGNANCIES Leukemia In 1952, Folley, Borges and Yamawaki accumulated persuasive evidence that leukemia was increased among the survivors of the atomic bombings of Hiroshima and Nagasaki [Ii]. The next several years were devoted to the study of the early hematologic and preclinical phases, and the effects of age at the time of bombing and calendar time on the incidences and types of radiation—related cases of leukemia [12—141. It was soon apparent that the peak annual incidence occurred in the early l9SOs and has declined steadily therafter [13,151. These studies were hampered, however, by the failure to define a manageable cohort for scrutiny, the absence of a systematic method of case detection, and the inability to assign individuals estimates of their exposures. The occurrence of leukemia could be functionally related only to the distances from the blast hypocenter of the exposed individuals and the presence or absence of a history of severe radiation complaints, that is, epilation, oropharyngeal lesions, or purpura [161. Subsequently w~iththe establishment in 1955 of the Unified Study Program with its emphasis on fixed samples, the development of a Leukemia Registry (1958), and the evolution of means to assign to individuals doses of gamma and neutron radiation, our understanding of the leukemogenic effect of A—bomb exposure improved materially. It is now clear that acute granulocytic, acute lymphocytic, acute leukemias of other types (such
Late Radiation Responses in Man
233
as monocytic, stem—cell type, and erythroleukemia) and chronic granulocytic leukemia are all radiation related. As yet, however, chronic lymphocytic leukemia does not appear to be radiation related, but this is an infrequent malignancy among the Japanese (only two cases have been seen in Hiroshima and Nagasaki combined, both in Nagasaki, and neither exposed) and thus our ability to distinguish modest increases is still limited. The most recent appraisal [17] of the incidence of leukemia in atomic bomb survivors and their controls in the Life Span Study cohort, an investigation from October 1950 through December 1978, notes the following: (a) The risk of all types of leukemia increases with dose in both cities except among those survivors who received less than 100 rad in total kerma in Nagasaki. (b) The shape two major should be leukemias
of the dose response curve is different between the two cities and between the types of leukemia (acute leukemia and chronic granulocytic leukemia). It noted, however, that this difference between the cities, but not the type of is clouded by the reassessment to which we have previously alluded.
(c) The excess risk among those survivors who received 100 rad or more (total kerma) has gradually declined with years after exposure in both cities. The excess risk had disappeared among Nagasaki survivors by 1970, but the risk at that time was still high among exposed survivors in Hiroshima who were 30 years of age or older at the time of the bombing. At the time of their analysis, the only age ATB (at time of bombing) group for which there continued to be an excess risk was 45 years or older. These authors conclude that their reanalysis continues to support previous observations that the leukemogenic effect of radiation on those individuals exposed at younger ages was greatest in the early post—bomb period and has declined more rapidly in subsequent years whereas the effect on those exposed at older ages appeared later and has persisted longer. Given the concern for the effects of low levels of exposure (only a few rad or so), it has not been possible to show that the crude annual incidence rate for all leukemias is significantly higher in those survivors who received exposures in the range of 1—49 rad as contrasted with survivors who received less than one rad. Thus, in sensu strictu, these data do not establish a low dose effect. However, given the nature of the increase in the annual incidence of these various forms of leukemia among those individuals who received 50 rad or more, prudence dictates a cautious attitude at these lower levels of exposure. Insofar as other radiation—related malignancies of the lymphatic and hematopoietic tissue are concerned, Nishiyama and his colleagues [18] were the first to report an increased prevalence of malignant lymphoma and multiple mnyeloma in survivors of the atomic bomb in Hiroshima exposed to 100 rads or more (a similar relationship was not seen in Nagasaki). Most of the effect they saw was, however, attributable to malignant lymphoma rather than myeloma. They identified only six cases of the latter. Earlier reports had suggested an increased frequency of myelofibrosis with myeloid metaplasia in proximally exposed survivors in Hiroshima [19,20]. Ichimaru and his colleagues [21] have recently reviewed all of the cases of multiple myeloma seen from October 1950 through December 1976 in the Life Span Study sample. They find the risk of myeloma to increase consistently and significantly with radiation exposure. Indeed, the standardized relative risk is about five times greater among individuals exposed to 100 rads or more than in the control group. They further note no differences between the cities nor in the average period from exposure to onset of this malignancy, which was 20.3 years in the high dose group (100—600 rad), 20.6 years in the low dose group (1—99 rad) and 22.7 years in the control group (less than one rad), respectively. Only 22 confirmed cases of multiple myeloma were seen among the survivors in the years 1950—1976 and the data are as yet too sparse to examine with assurance the possible effects of age at the time of exposure upon subsequent risk. Malignant Neoplasms of the Lip, Oral Cavity, and Pharynx A recent review of all malignancies under this rubric seen in the years 1957 to 1976 among members of the Life Span Study sample identified 85 possible cases, exclusive of salivary gland tumors. Sixty—three of these 85 were considered definite, that is, had been histologically confirmed [22]. Save for cancers of the tongue, the number of malignancies for each specific site (lip; gum; mouth other than lip, gum or tongue; tonsils; nasopharynx; hypopharynx and pharynx) is invariably small, and no one of these tumors is significantly associated with dose. Patently, however, given the infrequency of most of these malignancies, our ability to detect any but the most pronounced radiation effects is extremely small. Malignancies of the Digestive Organs and Peritoneum At least three such cancers seem radiation related, namely, those of the esophagus, stomach and large bowel excluding the rectum. Associations have been sought for the liver [23,24] and the gall bladder, bile ducts and Vater’s ampulla [251 but none have been seen. Some 90 cases of primary carcinoma of the pancreas were diagnosed from 1956—1970, in the ABCC
234
W.J. Schull
autopsy series, but no clear dependence on exposure emerged [25]. To return to the three significantly associated malignancies, the effect of ionizing radiation on the frequency of death attributed to an esophageal malignancy appears, retrospectively, to have been first apparent in the years 1954—1958 [27,281. At present, the estimated excess mortality attributable to this tumor per million person years/rad (PYR) is 0.15. A conspicuous elevation of the frequency of this tumor is only seen among survivors who were 50 years of age or more ATB, and while the effects from the two cities cannot be shown to be significantly different, only the one in Hiroshima is significant by conventional statistical standards. The Situation with respect to malignancies of the stomach remains enigmatic. Nakamura [29,301, using death certificates for 1950—1973, reported a consistent increase in mortality with increasing radiation dose in Hiroshima, the rate being highest among those survivors who had received 400 or more rads, whereas the evidence of a radiation effect in Nagasaki was very weak. An excess was found only at doses above 500 rad. Beebe et al. [27] reported that the evidence of an association of tumors of this organ with radiation derived from the steady accumulation of excess deaths rather than some recent change, based on an examination of successive four—year time increments beginning in 1950 and terminating in 1974. They observed no single four years in which the excess mortality was of more than borderline significance. Kato and Schull [28], however, in the most recent analysis find the estimated mortality PYR to be significantly elevated in 1975—1978. The absolute risk of death from a malignancy of the stomach in these years per million PYR was 1.62 (90% confidence interval: 0.11, 3.14). Again, although the cities do not differ significantly one from another as judged in terms of estimated mortality per million PYR, only the effect in Hiroshima is a statistically significant one. Malignant neoplasms of the large intestine, excluding the rectum, have in the past not appeared to be radiation related. Kato and Schull [28], however, now find the estimated mortality per million PYR for this tumor when cities, sexes, and all ages ATB are combined to be significantly elevated. It is 0.30 (90% confidence interval: 0.16, 0.43) or approximately twice the risk associated with esophageal tumors. This change in events is attributable mainly to a marked increase in the number of deaths ascribed to this malignancy in the years 1975—1978. It is seen in both cities, more strikingly so in Hiroshima than in Nagasaki, however, and more conspicuously affects two age groups (10—19 or 35—49 years ATB). Malignancies of the Respiratory and Intrathoracic Organs Harada and Ishida [311,in an analysis of the Hiroshima City Tumor Registry, were the first to suggest that the incidence of lung cancer was significantly higher among those who were exposed within 1500 m of ground zero. Subsequent studies [see, for example, 32, 33 and 341 confirmed and extended their observations. The continuing mortality surveillance on the Life Span Study sample reveals the absolute risk of death from this malignancy per million PYR to be 0.61 (90% confidence interval: 0.37, 0.86). Both relative and absolute risks are higher in Hiroshima than in Nagasaki although the values for the two cities do not differ significantly. Only those for Hiroshima meet the conventional criteria for statistical significance. However, it should be noted that the absolute risk increased markedly in both cities in the years 1975—1978. For the cities collectively, for example, the absolute risk in any four year period beginning in 1955 which had previously varied within the range 0.30— 0.73 rose dramatically to 2.59 (95% confidence interval: 1.54, 3.64). The carcinogenic effect which previously was seen primarily in the older age groups is now readily seen in those individuals who were 20—34 and even 10—19 years of age ATB. Efforts to find evidence of an influence of either smoking or occupational exposure on this apparent radiation effect have failed thus far [28,35]. Malignancies of Bone,
Connective Tissue and Breast
Yamamoto and Wakabayashi [36], in a study of benign and malignant bone tumors found at autopsy or in surgical specimens in Hiroshima and Nagasaki in the years 1950—1965, were unable to demonstrate an association of these tumors and exposure distances. With the exception of osteosarcomas, however, the number of malignant bone tumors of a specified kind was too small to permit consideration according to the histological type of the tumor. Death certificate data have not been particularly informative here, for the number of deaths ascribed to malignant tumors of the bone are so few. Under this rubric, the malignancy most clearly radiation related is cancer of the breast. Wanebo et al. [37] were the first to establish this. Their study was limited, however, to women who were members of the Adult Health Study group. Subsequent investigations have focused on members of the Life Span Study sample, some 63000 females in all. The most recent of these [38] finds that (1) the distribution of histological types of mammary cancers does not vary significantly with radiation dose; (2) among all women who received at least ten rad, those irradiated before age 20 will experience the highest rates of breast cancer throughout their lifetimes; and (3) the incidence of malignant mammary tumors appears to be linearly related to dose in both cities. Nakamura and his colleagues [39] have examined the relationship between exposure
Late Radiation Responses in Man
235
and other epidemiological risk factors in the development of breast cancer among the survivors. Specifically, they examine the association of such risk factors as family history of malignant neoplasms, education, menstrual, marital and reproductive histories, and history of breast feeding. They conclude that the risk increases in proportion to the number of other risk factors involved. Based on mortality findings alone, Kato and Schull [28] report the absolute risk of death from cancer of the breast to be 0.50 per million PYR (females only) (95% confidence interval: 0.29, 0.72). For some malignancies, the cities do not differ significantly, albeit at face value the risk is less in Nagasaki than in Hiroshima, where the effect is statistically significant. Malignancies of the Genito—Urinary System Efforts to relate malignancies of this system to exposure to the A—bombs have involved both incidence studies and the continued mortality surveillance of the Life Span Study sample. In the former studies, Bean and his colleagues [40] were unable to demonstrate an association with exposure to an atomic bomb of the occurrence of prostatic adenocarcinoma in 1357 male members of the Life Span Study sample who were 50 years of age or older at death and were autopsied at ABCC—RERF between 1961 and 1969. Nor could they demonstrate an effect of exposure on histologic type or biologic activity. Sawada [41], in a much earlier study, attempted to evaluate the effect of A—bomb radiation on the occurrence of gynecologic tumors as seen in a series of routine examinations of 1785 exposed women and 1802 control subjects, but unfortunately, almost half of the women of this series would not accept an examination of their pelves. Carcinoma of the cervix was the most frequently encountered tumor; however, its occurrence could not be shown to depend on exposure nor was the age of onset significantly different between control and irradiated individuals. The most recent examination of deaths in the Life Span Study sample [28] finds a marginally significant increase in deaths attributed to malignancies of the urinary tract; the absolute risk (excess deaths! million PYR) was 0.15 (95% confidence interval: 0.04, 0.26). There was, however, no significant increase in deaths attributable to cancer of the uterus. Other Malignancies Malignancies of the eye, brain, “other” and unspecified parts of the nervous system; the thyroid gland and other endocrine glands and their related structures fall within this rubric. Included too are malignant neoplasms of “other and ill—defined” sites, malignancies without specification of site, and secondary malignant neoplasms of the lymph nodes, respiratory and digestive systems, as well as other sites. We are concerned only with primary tumors and of those within this rubric, only malignant neoplasms of the thyroid have been significantly associated with exposure to A—bomb radiation. Hollingsworth et al. [42] first noted, in a study of thyroid disease in Hiroshima, that carcinoma of the thyroid constituted some 7% of the total number of individuals with thyroid disorders and that a greater number of these malignancies were seen among the more heavily exposed individuals. They cautiously noted, however, that differences among distance groups were not statistically significant. Subsequent studies have removed any doubt of the association of this tumor with exposure. For example, Parker and his associates [43] summarized the results of the continuing, periodic examinations of the Adult Health Study group in the years 1958— 1971. In all, 74 cases of thyroid cancer in the Adult Health Study were seen and histologically confirmed in these years. They found thyroid carcinoma to be commoner in women and significantly more prevalent in persons exposed to 50 or more rads of atomic radiation. They also found carcinoma of the thyroid to be diagnosed during life more often in persons less than 20 years ATB although they failed to confirm the previously reported radiosensitivity in childhood. An increased risk was seen even among persons who were 50 years of age or older at the time of radiation exposure. It warrants note here that this is one of the few malignancies where a significant relationship with ionizing radiation has been demonstrable largely on clinical grounds. No effect has emerged through the continuing mortality surveillance of the Life Span Study sample, but few of the individuals with thyroid malignancies, confirmed either at autopsy or on histological section, have had their deaths ascribed to these malignancies. A recent analysis, however, of the incidence of cancer within the Life Span Study sample in Nagasaki, based on Tumor Registry data, revealed a significant excess of cancers of this particular kind; the absolute risk per million PYR rad was 1.32 (95% confidence interval: 0.88, 0.76). Before turning to other radiation related effects, it is important to note two general features associated with radiation related malignancies. First, the time from exposure to death is shortened for leukemia depending upon the dose received by the survivor but not for other cancers, and radiation—induced cancers other than leukemia seem to develop proportionally to the natural cancer rates for the attained age. Second, for specific age—at— death intervals, both relative and absolute risks tend to be higher for the young at the time of bombing. Simply stated, the young, particularly those in the first decade of life, seem especially vulnerable to radiation induced damage.
236
W.J. Schull
OTHER RADIATION—RELATED EFFECTS As has been implied,
an increased probability of malignancy is the greatest hazard confronting survivors, but there are other effects as well. The latter include an increased risk of radiation—related lenticular opacities, disturbances of growth among those survivors who had not achieved their adult dimensions at the time of exposure, an increase in mental retardation and small head sizes among the in utero exposed and an increase in the frequency of chromosomal aberrations in peripheral lymphocytes.
Cataracts The formation of radiation induced cataracts was the first late effect recognized unequivocally in individuals exposed to A—bomb radiation in Hiroshima and Nagasaki. Since the
original description of such cataracts by Cogan, Martin and Kimura (44], numerous other clinical, histopathological and statistical studies have appeared (see [45] for further references). It is now apparent that a radiation induced cataract is in its early stages, at least, a highly characteristic lesion. It is a central, posterior subcapsular opacity, easily visible with a alit lamp or an ophthalmoscope. Unfortunately, the word “cataract”
connotes to many a defect which impairs vision although it is commonly used to describe any detectable change in translucency in the lens. The vast majority (over 80%) of these lesions are nonprogressive, do not impinge on visual acuity and commonly the individual
involved is unaware of their presence. It is generally believed that the radiation damaged cells of the germinative epithelium are unable to differentiate into normal lens fibers, and that their remnants are gradually pushed to the posterior pole under the pressure of the remaining undamaged cells which continue to differentiate normally. The opacification itself arises from the denaturation of the normally transparent protein in the damaged cells, as the latter breaks down. Customarily the effects of ionizing radiation are viewed as either “stochastic,” if the
probability
of their occurrence is a direct function of dose, or “nonstochastic,”
if it is
the severity of the effect which is exposure—dependent (almost always in this instance there is a threshold). It is not always clear whether a particular radiation related effect should be viewed as a stochastic or a nonstochastic one, for often the sequence of biological events which culminate in the effect are unknown. Radiation carcinogenesis, for example, as well as mutation is generally assumed to be a stochastic effect; whereas lenti— cular opacities and mental retardation have commonly been viewed as nonstochastic. Otake and Schull’s [45] recent reanalysis of the Hiroshima and Nagasaki data-on lenticular opacities seems to bear out the classification of this lesion as nonstochastic. They find, among the numerous linear and quadratic models which they fitted to the frequency of occurrence of lenticular opacities, the “best” fit, that is, the one with the smallest chi—square and largest tail probability is a model which postulates a linear relationship with exposure and a threshold. Indeed, the threshold which they estimate from the Japanese experience is one consistent with the threshold which has been commonly conjectured, based on clinical experience.
Fetal Development It has lor.g been known that irradiation fetus.
has a damaging effect upon the development of the
This knowledge has prompted a continued clinical and mortality surveillance of those
individuals who were exposed in utero to the atomic bombs in Hiroshima and Nagasaki. Indeed, we have known for thirty years that among these children there is an increased prevalence of
mental retardation [46—50]. Several developments have occurred recently which impinge on these earlier studies. First, the appropriateness of the T65 exposure estimates have been questioned [8,9]. Second, Hashizumi et al. [51] have estimated the in utero fetal exposures to radiation from the atomic bombs; and Kerr [52] has given the estimates of absorbed dose to a first trimester fetus in terms of tissue kerma in air. While these latter estimates are predicated on radiant energies of the kind postulated in the T65 dosimetry, they provide the first and to date only basis for assessing fetal absorbed doses. Finally, but no less importantly, if the vulnerability of a developing organ to radiation damage is related to rate of growth, Dobbing and Sands’ [53] studies of the quantitative growth of the human brain provide a compelling basis for expecting those children exposed between 10 and 18 weeks of gestation, as measured from the last menstrual period, to be at the highest risk of mental retardation of all children exposed in utero. Two principal findings emerge from this reanalysis. First, although the developing forebrain is undoubtedly vulnerable throughout gestation, the highest risk of damage occurs at the time when the most rapid proliferation of neuronal elements occurs, namely, 10—17 weeks after conception. Overall, the risk is four times greater in these weeks than in subsequent ones. Second, in this period of rapid neuronal growth, damage — expressed as subsequent mental retardation — appears linearly related to the dose the fetus receives. Indeed, the variation in frequency of mental retardation with dose not explained by a linear model is consistently not significant. Radiation—related mental retardation in the latter stages of gestation (after the 17th week) is not fully explicable by a linear model and the observed values suggest that at these ages a threshold may exist.
Late Radiation Responses in Man
23]
It should be noted that the cohort of individuals exposed in utero have not as yet been shown to have an increased risk of premature mortality from all causes nor has it been possible to demonstrate that their risk of leukemia or any of the other malignancies known to be radiation related in those survivors not exposed in utero is increased. Thus, to date, the most significant finding within this cohort is the increased risk of mental retardation and diminished head size [49]. Chromosomal Aberrations No less than 15 years ago [54], radiation—induced chromosomal aberrations were shown to persist in cultured peripheral blood lymphocytes derived from Hiroshima and Nagasaki A—bomb survivors long after their radiation exposure. Recently, Awa and his colleagues [55] have reconfirmed the increased frequency of cells with chromosome aberrations with increasing dose in both cities. However, in every dose group, the frequency of aberrant cells was consistently higher in Hiroshima than in Nagasaki. They suggested that a higher neutron dose in Hiroshima than in Nagasaki may have been a major component contributing to this difference. This seems less likely now, given the dose reassessment. However, even under the reassessed values, there continues to be a difference between the two cities. Among the types of chromosomal aberrations which have been identified thus far, reciprocal translocations were observed to predominate, and they played an important role in determining the dose—aberration relationship. It warrants note too that this is one of the few effects which seems statistically discernible even within the 1—9 rad dose range. The health significance of these findings continues to be problematic. Some years ago an attempt was made to correlate the occurrence of chromosomal abnormalities in A—bomb survivors with the findings on their periodic examinations within the Adult Health Study program [56]. No consistent correlation between the clinical parameters evaluated and the degree of cytogenetic abnormalities emerged. The thought, however, that these chromosomal aberrations may be prodromal continues to titillate investigators but a persuasive answer to their role seems unlikely to emerge until more deaths have occurred among those individuals who have been studied karyotypically over the past decade and a half. Growth Disturbances in Children Finally, we have adverted briefly to growth disturbances among those survivors who had not reached skeletal maturity at the time of their exposure. Nehemias [57] noted, “Differences with respect to size which are statistically very significant are found among the subgroups defined on the basis of the radiation exposure index. The actual physical differences are small, however. The differences have been shown to be in the direction of decrease in size with increase in degree of radiation exposure.” More recently, Belsky and Blot [58] have examined the stature of adults exposed in childhood to the atomic bombs of Hiroshima and Nagasaki. They found adult height to be significantly lowered among Hiroshima males and females exposed to high doses (100+ rad) of irradiation when they were ages 0—5 years ATB; however, the effect of dose declined with increase in age ATB. For Nagasaki, no statistical differences in means for height were noted by A—bomb dose group although for females who were 0—5 years ATB the highest dose group had the smallest mean height. Belsky and Blot conclude that a growth retarding effect of radiation can be demonstrated for the youngest heavily exposed children. If their findings are taken in concert with those of Nehemiss, a somatic effect- on growth seems highly probable. The mechanism(s) is, of course, unknown. SUMMARY To summarize,
cancers of the breast, colon, esophagus,
lungs, stomach, thyroid and urinary
tract as well as leukemia and multiple myeloma increase in frequency with an increase of exposure. Furthermore, it has been demonstrated that radiation—related lenticular opacities, disturbances of growth among those survivors who had not achieved their adult dimensions at the time of exposure, and mental retardation and small head sizes among the in utero exposed are increased. Chromosomal aberrations in peripheral lymphocytes are more frequent too. The health implications of some of these effects are uncertain, for example the chromosomal aberrations, but most unquestionably compromise both the duration of life and its quality. Indeed, it can be shown that among the atomic bomb survivors there is a significant foreshortening of life, explicable in terms of their increased probabilities of death from a malignant tumor. Among these effects, however, if they are examined on a dose level—by dose level basis, significant effects emerge generally only among those individuals who have received exposures of 50 rad (tissue kerma) or higher. Only chromosomal aberrations can be shown to be unequivocally elevated at lesser exposures. Under these circumstances, the hazards one foresees at these lower dose levels is functionally related to the statistical model one presumes best describes the fundamental dose—response relationship. Prudence counsels that this relationship should be construed as linear until such time as proof to the contrary exists.
238
W.J. Schull
REFERENCES 1. 2.
3. 4. 5. 6.
7. 8.
9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45.
G.W. Beebe, Epidemiologic Reviews 1, 184 (1979) J.V. Neel and W.J. Schull, The Effect of Exposure to the Atomic Bombs on Pregnancy Terminations in Hiroshima and Nagasaki, National Academy of Sciences—National Research Council, Washington, D.C. (1956) W.J. Schull, H. Otake and J.V. Heel, Science 213, 1220 (1981) G.W. Beebe and M. Usagawa, Atomic Bomb Casualty Commission Technical Report 12—68 (1968) T. Yamamoto, I.M. Moriyama, M. Asano and L. Guralnick, Radiation Effects Research Foundation Technical Report 18—78 (1978) J.A. Auxier, Ichiban: Radiation Dosimetry for the Survivors of the Bombings of Hiroshima and Nagasaki, Technical Information Center, Energy Research and Development Administration (1977) R.C. Milton and T. Shohoji, Atomic Bomb Casualty Commission Technical Report 1—68 (1968) G.D. Kerr, in: Proceedings of the Fourth Symposium on Neutron Dosimetry, Gesellschaft fur Strahlen—und Umweltforschung, Munich—Neuherberg, June 1—5, 1981 Volume 1, 501 (1981) W.E. Loewe and E. Mendelsohn, Lawrence Livermore Laboratory Publication UCRL—86595 (1981) J. Marcum, an unpublished manuscript entitled House Attenuation Factors for Radiation at Hiroshima and Nagasaki prepared for Director, Defense Nuclear Agency J.H. Folley, W. Borges and T. Yamawaki, American Journal of Medicine 13, 11 (1952) R.D. Lange, W.C. Moloney and T. Yamawaki, Blood 9, 574 (1954) W.C. Moloney, New England Journal of Medicine 253, 88 (1955) W.C. Mooney and R.D. Lange, Blood 9, 663 (1954) M. Tomonaga, New Zealand Medical Journal of Hematology 65 (Supplement), 863 (1966) W.C. Moloney and M.A. Kastenbaum, Science 121, 308 (1955) H. Ichimaru, T. Ishimaru, M. Mikami, Y. Yamada and T. Ohkita, Journal of Radiation Research (in press) H. Nishiyama, R.E. Anderson, T. Ishimaru, K. Ishida, Y. Ii and N. Okabe, Cancer 32, 1301 (1973) R.E. Anderson, T. Hoshino and T. Yamamoto, Annals of Internal Medicine 60, 1 (1964) R.E. Anderson and T. Yamamoto, in: Myeloproliferative Disorders of Animals and Man, Hanford Biological Symposium, Richland, Washington (1968) N. Ichimaru, T. Ishimaru, M. Mikami and M. Matsunaga, Radiation Effects Research Foundation Technical Report 9—79 (1979) J.A. Pinkston, T. Wakabayashi, T. Yamamoto, M. Asano, Y. Hirada, H. Kumagami and M. Takeichi, Radiation Effects Research Foundation Technical Report X—80 (1980) W.M. Schreiber, H. Kato and J.D. Robertson, Cancer 26, 69 (1970) M. Asano, S. Seyama, H. Itakura, T. Hamada, S. lijima and H. Kato, Transactions of the Society of Pathology, Japan 67, 143 (1978) J.D. Robertson, H. Kato and W.M. Schreiber, Atomic Bomb Casualty Commission Technical Report 7—70 (1970) R.W. Cihak, T. Kawashima and A. Steer, Cancer 29, 1133 (1972) G.W. Beebe, H. Kato and C.E. Land, Radiation Research 25, 138 (1978) H. Kato and W.J. Schull, Radiation Research 90 (1982) K. Nskamura, Radiation Effects Research Foundation Technical Report 8—77 (1977) K. Nakamura, Lancet ii, 86 (1977) T. Hirada and M. Ishida, Journal of the National Cancer Institute 25, 1253 (1960) A. Ciocco, Atomic Bomb Casualty Commission Technical Report 18—65 (1965) G.W. Beebe, T. Yamamoto, Y.S. Matsumoto and S.E. Gould, Atomic Bomb Casualty Commission Technical Report 8—67 (1967) C.K. Wanebo, K.G. Johnson, K. Sato and T.W. Thorslund, American Reviews of Respiratory Diseases 98, 778 (1968) T. Ishinaru, T.W. Cihak, C.W. Land, A. Steer and A. Yamada, Cancer 38, 1723 (1975) T. Yaniamoto and T. Wakabayashi, Acts Pathologica Japonica 19, 201 (1969) C.K. Wanebo, K.G. Johnson, K. Sato and T.W. Thorslund, Atomic Bomb Casualty Commission Technical Report 13—67 (1967) M. Tokunaga, J.E. Norman, Jr., M. Asano, S. Tokuoka, H. Ezaki, I. Nishimori and Y. Tsuji, Journal of the National Cancer Institute 62, 1347 (1979) K. Nakamura, D.H. McGregor, H. Kato and T. Wakabayashi, Radiation Effects Research Foundation Technical Report 9—77 (1977) M.A. Bean, R. Yatani, P.1. Liu, K. Fukazawa, F.W. Ashley and S. Fujita, Cancer 32, 496 (1973) H. Sawada, Atomic Bomb Casualty Commission Technical Report 6—59 (1959) D.R. Hollingsworth, H.B. Hamilton, H. Tamagaki and G.W. Beebe, Medicine 42, 47 (1976) L.N. Parker, J.L. Belsky, T. Ysmamoto, S. Kawamoto and R.J. Keehn, Annals of Internal Medicine 80, 600 (1974) D.C. Cogan, S.F. Martin and S.J. Kimura, Science 110, 654 (1949) M. Otake and W.J. Schull, Radiation Research 91 (1982)
Late Radiation Responses in Man
46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58.
C. Plummer, Pediatrics 10, 687 (1952) R.W. Miller, Pediatrics 18, 1 (1956) R.W. Miller and W.J. Blot, Lancet ii, 784 (1972) W.J. Blot and R.W. Miller, Radiology 106, 617 (1973) M. Otake and W.J. Schull, Lancet (in press) (1982) T. Hashizume, T. Maruyama, K. Nishizawa and A. Nishimura, Journal of Radiation Research 1.4, 346 (1973) G.D. Kerr, Health Physics 37, 487 (1979) J. Dobbing and J. Sands, Archives of Diseases of Childhood 48, 757 (1973) A.D. Bloom, S. Neriishi, A.A. Awa, T. Honda and P.G. Archer, Lancet ii, 802 (1967) A.A. Awa, T. Sofuni, T. Honda,M. Itoh, S. Neriishi and M. Otake, Radiation Effects Research Foundation Technical t 12—77 (1977) R.A. King, J.L. Belsky, M. Otake, A.A. Awa and T. Matsui, Atomic Bomb Casualty Commission Technical Report 15—72 (1972) J.V. Nehemias, Atomic Bomb Casualty Commission Technical Report 22—61 (1961) J.L. Belsky and W.J. Blot, Atomic Bomb Casualty Commission Technical Report 35—71 (1971)
239
0273—1177/83 $0.00 + .50 Cop~’rjght ©COSPAR
LATE RADIATION RESPONSES IN MAN: CURRENT EVALUATION FROM RESULTS FROM HIROSHIMA AND NAGASAKI William J. Schull Center for Demographic and Population Genetics, Graduate School of Biomedical Sciences, University of Texas Health Science Center, P.O. Box 20334, Houston, TX 77025, U.S.A.
ABSTRACT Among the late effects of exposure to the atomic bombings of Hiroshima and Nagasaki, none looms larger than radiation related malignancies. Indeed, the late effects of A—bomb radiation on mortality appear to be limited to an increase in malignant tumors. At present, it can be shown that cancers of the breast, colon, esophagus, lungs, stomach, thyroid, and urinary tract as well as leukemia and multiple myeloma increase in frequency with an increase in exposure. No significant relationship to radiation can as yet be established for malignant lymphoma, nor cancers of the rectum, pancreas or uterus. Radiation induced malignancies other than leukemia seem to develop proportionally to the natural cancer rate for the attained age. For specific age—at—death intervals, both relative and absolute risks tend to be higher for those of younger age at the time of bombing. Other late effects include radiation—related lenticular opacities, disturbances of growth among those survivors still growing at the time of exposure, and mental retardation and small head sizes among the in utero exposed. Chromosoinal abnormalities too are more frequently encountered in the peripheral leucocytes of survivors, and this increase is functionally related to their exposure. Some uncertainty continues to surround both the quantity and quality of the radiation released by these two nuclear devices, particularly the Hiroshima bomb. A recent reassessment suggests that the gamma radiation estimates which have been used in the past may be too low at some distances and the neutron radiation estimates too high at all distances; moreover, the energies of the neutrons released now appear “softer’ than previously conjectured. These uncertainties are not sufficiently large, however, to compromise the reality of the increased frequency of malignancy, but make estimates of the dose response, particularly in terms of gamma and neutron exposures, tentative. INTRODUCTION Thirty—five years of studies of the survivors of the atomic bombings of Hiroshima and Nagasaki have established a number of late effects of radiation exposure: some unequivocally some less c~rtainly. Among the former are the occurrence of radiation—related lenticular opacities; an increase in leukemia and a variety of malignant solid tumors such as those of the breast, lung, stomach and thyroid; disturbances of growth among survivors who had not achieved their adult dimensions at the time of exposure; an increase in mental retardation and small head sizes among those exposed in utero; and an increase in the frequency of chromosomal aberrations in peripheral lymphocytes. Less definite is an increase in certain aspects of humoral and cell mediated immunity, or in the frequencies of cancers of the colon, esophagus, salivary glands and urinary tract. These studies have also sought but failed to find a variety of other conceivable elfects, such as an impingement on fertility or a nonspecific life—shortening. Either no effect has occurred in these instances or the nature or extent of the effect lies outside present measurement capabilities. We now turn to a more measured account of the data on which these assertions rest. THE FOLLOW—UP STUDIES Throughout most of the first decade of the existence of the Atomic Bomb Casualty Commission (AECC) — the agency of the National Academy of Sciences—National Research Council charged with the responsibility for the study of the late effects of exposure to atomic radiation, most studies of the survivors were opportunistic both in the research directions they pursued and in the samples they employed (see [1] for a thorough account of the history of this agency). A notable exception was the Genetics Study [2,31. It was not until 1956
231
W.J. Schull
that the first steps toward a unified program were taken. The latter centered on the investigation of a fixed population of survivors and controls, based largely upon the nationwide enumeration of survivors in 1950 as a part of Japan’s first post—war decennial census [4]. The program entails (1) a mortality surveillance of approximately 110000 individuals (82000 of whom were exposed), the so—called Life Span Study (LSS); (2) a series of now biennial, but initially annual, physical examinations of a subset of approximately 20000 members of the Life Span Study, the Adult Health Study; and finally (3) an autopsy program aimed at the postmortem examination of as many members as practicable of the LSS sample. For a variety of reasons this latter aspect of the unified program has been deemphasized recently, but not before some 10000 autopsies had occurred [5]. In both cities tumor and tissue registries, managed by local medical associations, are available to support this program of inquiry. There is also available a number of ad hoc samples, such as those on which studies of those individuals exposed in utero rest. ESTIMATES OF EXPOSURE It was not until 1957 that the first estimates of individual exposures became available. Prior studies had attempted to relate possible effects either to the distances of survivors from the hypocenter or to the presence of symptoms associated with acute radiation sickness or both. These preliminary estimates were of limited utility, however, for they were largely available only on those individuals exposed in the open, a minority of the survivors within 2000 meters of the blast hypocenter in both cities. This limitation spurred a more comprehensive program of dose assessment, one culminating in the so—called T65 dose estimates [6,7]. The latter have been the basis of most appraisals of dose—response relationships until quite recently. Under the T65 dosimetry, survivors within 1600 meters of the hypocenter in Hiroshima (2000 in Nagasaki) were assigned separate gamma and neutron exposures based upon their individual distances from the hypocenter and the attenuation of the “free—in—air” doses at those distances attributable to the shielding they may have experienced. Shielding factors were estimated experimentally [61. Individuals exposed beyond 1600 or 2000 meters were assigned the gamma and neutron exposures which prevailed at their specific distances uncorrected for the effects of shielding; it should be noted that at 1600 and 2000 meters, the total exposure in terms of tissue kerma (kinetic energy released in tissue) was then thought to be about 19 and 18 rads for Hiroshima and Nagasaki, respectively. Recent reassessments based on further data have called into question the T65 estimates [8,9]. These newer appraisals suggest (1) that the gamma exposures at some distances may be greater than previously thought but the neutron exposures are less at all distances, and (2) the energies associated with the neutrons released was much lower than previously conjectured. At present it is not possible to estimate individual exposures under the newly proposed dosimetry save for those individuals who were exposed in the open, a minority of the proximally exposed, for new building and body shielding factors are not as yet available. It appears clear, nonetheless, that these factors will be different from those associated with the T65 dosimetry. Marcum [10] has suggested that the average gamma shielding factors for Hiroshima and Nagasaki may be more nearly 0.55 and 0.50 rather than the 0.90 and 0.81 given for these two cities by Milton and Shohoji [7]. It should be noted that to the average, than those previously or a malignant solid tumor, or kerma or absorbed dose will be which follow.
extent these reassessments lead to lower exposures, on the assumed to be true, the absolute risk of death from leukemia the occurrence of a lenticular opacity and the like per unit increased over those specific values cited in the paragraphs
RADIATION RELATED MALIGNANCIES Leukemia In 1952, Folley, Borges and Yamawaki accumulated persuasive evidence that leukemia was increased among the survivors of the atomic bombings of Hiroshima and Nagasaki [Ii]. The next several years were devoted to the study of the early hematologic and preclinical phases, and the effects of age at the time of bombing and calendar time on the incidences and types of radiation—related cases of leukemia [12—141. It was soon apparent that the peak annual incidence occurred in the early l9SOs and has declined steadily therafter [13,151. These studies were hampered, however, by the failure to define a manageable cohort for scrutiny, the absence of a systematic method of case detection, and the inability to assign individuals estimates of their exposures. The occurrence of leukemia could be functionally related only to the distances from the blast hypocenter of the exposed individuals and the presence or absence of a history of severe radiation complaints, that is, epilation, oropharyngeal lesions, or purpura [161. Subsequently w~iththe establishment in 1955 of the Unified Study Program with its emphasis on fixed samples, the development of a Leukemia Registry (1958), and the evolution of means to assign to individuals doses of gamma and neutron radiation, our understanding of the leukemogenic effect of A—bomb exposure improved materially. It is now clear that acute granulocytic, acute lymphocytic, acute leukemias of other types (such
Late Radiation Responses in Man
233
as monocytic, stem—cell type, and erythroleukemia) and chronic granulocytic leukemia are all radiation related. As yet, however, chronic lymphocytic leukemia does not appear to be radiation related, but this is an infrequent malignancy among the Japanese (only two cases have been seen in Hiroshima and Nagasaki combined, both in Nagasaki, and neither exposed) and thus our ability to distinguish modest increases is still limited. The most recent appraisal [17] of the incidence of leukemia in atomic bomb survivors and their controls in the Life Span Study cohort, an investigation from October 1950 through December 1978, notes the following: (a) The risk of all types of leukemia increases with dose in both cities except among those survivors who received less than 100 rad in total kerma in Nagasaki. (b) The shape two major should be leukemias
of the dose response curve is different between the two cities and between the types of leukemia (acute leukemia and chronic granulocytic leukemia). It noted, however, that this difference between the cities, but not the type of is clouded by the reassessment to which we have previously alluded.
(c) The excess risk among those survivors who received 100 rad or more (total kerma) has gradually declined with years after exposure in both cities. The excess risk had disappeared among Nagasaki survivors by 1970, but the risk at that time was still high among exposed survivors in Hiroshima who were 30 years of age or older at the time of the bombing. At the time of their analysis, the only age ATB (at time of bombing) group for which there continued to be an excess risk was 45 years or older. These authors conclude that their reanalysis continues to support previous observations that the leukemogenic effect of radiation on those individuals exposed at younger ages was greatest in the early post—bomb period and has declined more rapidly in subsequent years whereas the effect on those exposed at older ages appeared later and has persisted longer. Given the concern for the effects of low levels of exposure (only a few rad or so), it has not been possible to show that the crude annual incidence rate for all leukemias is significantly higher in those survivors who received exposures in the range of 1—49 rad as contrasted with survivors who received less than one rad. Thus, in sensu strictu, these data do not establish a low dose effect. However, given the nature of the increase in the annual incidence of these various forms of leukemia among those individuals who received 50 rad or more, prudence dictates a cautious attitude at these lower levels of exposure. Insofar as other radiation—related malignancies of the lymphatic and hematopoietic tissue are concerned, Nishiyama and his colleagues [18] were the first to report an increased prevalence of malignant lymphoma and multiple mnyeloma in survivors of the atomic bomb in Hiroshima exposed to 100 rads or more (a similar relationship was not seen in Nagasaki). Most of the effect they saw was, however, attributable to malignant lymphoma rather than myeloma. They identified only six cases of the latter. Earlier reports had suggested an increased frequency of myelofibrosis with myeloid metaplasia in proximally exposed survivors in Hiroshima [19,20]. Ichimaru and his colleagues [21] have recently reviewed all of the cases of multiple myeloma seen from October 1950 through December 1976 in the Life Span Study sample. They find the risk of myeloma to increase consistently and significantly with radiation exposure. Indeed, the standardized relative risk is about five times greater among individuals exposed to 100 rads or more than in the control group. They further note no differences between the cities nor in the average period from exposure to onset of this malignancy, which was 20.3 years in the high dose group (100—600 rad), 20.6 years in the low dose group (1—99 rad) and 22.7 years in the control group (less than one rad), respectively. Only 22 confirmed cases of multiple myeloma were seen among the survivors in the years 1950—1976 and the data are as yet too sparse to examine with assurance the possible effects of age at the time of exposure upon subsequent risk. Malignant Neoplasms of the Lip, Oral Cavity, and Pharynx A recent review of all malignancies under this rubric seen in the years 1957 to 1976 among members of the Life Span Study sample identified 85 possible cases, exclusive of salivary gland tumors. Sixty—three of these 85 were considered definite, that is, had been histologically confirmed [22]. Save for cancers of the tongue, the number of malignancies for each specific site (lip; gum; mouth other than lip, gum or tongue; tonsils; nasopharynx; hypopharynx and pharynx) is invariably small, and no one of these tumors is significantly associated with dose. Patently, however, given the infrequency of most of these malignancies, our ability to detect any but the most pronounced radiation effects is extremely small. Malignancies of the Digestive Organs and Peritoneum At least three such cancers seem radiation related, namely, those of the esophagus, stomach and large bowel excluding the rectum. Associations have been sought for the liver [23,24] and the gall bladder, bile ducts and Vater’s ampulla [251 but none have been seen. Some 90 cases of primary carcinoma of the pancreas were diagnosed from 1956—1970, in the ABCC
234
W.J. Schull
autopsy series, but no clear dependence on exposure emerged [25]. To return to the three significantly associated malignancies, the effect of ionizing radiation on the frequency of death attributed to an esophageal malignancy appears, retrospectively, to have been first apparent in the years 1954—1958 [27,281. At present, the estimated excess mortality attributable to this tumor per million person years/rad (PYR) is 0.15. A conspicuous elevation of the frequency of this tumor is only seen among survivors who were 50 years of age or more ATB, and while the effects from the two cities cannot be shown to be significantly different, only the one in Hiroshima is significant by conventional statistical standards. The Situation with respect to malignancies of the stomach remains enigmatic. Nakamura [29,301, using death certificates for 1950—1973, reported a consistent increase in mortality with increasing radiation dose in Hiroshima, the rate being highest among those survivors who had received 400 or more rads, whereas the evidence of a radiation effect in Nagasaki was very weak. An excess was found only at doses above 500 rad. Beebe et al. [27] reported that the evidence of an association of tumors of this organ with radiation derived from the steady accumulation of excess deaths rather than some recent change, based on an examination of successive four—year time increments beginning in 1950 and terminating in 1974. They observed no single four years in which the excess mortality was of more than borderline significance. Kato and Schull [28], however, in the most recent analysis find the estimated mortality PYR to be significantly elevated in 1975—1978. The absolute risk of death from a malignancy of the stomach in these years per million PYR was 1.62 (90% confidence interval: 0.11, 3.14). Again, although the cities do not differ significantly one from another as judged in terms of estimated mortality per million PYR, only the effect in Hiroshima is a statistically significant one. Malignant neoplasms of the large intestine, excluding the rectum, have in the past not appeared to be radiation related. Kato and Schull [28], however, now find the estimated mortality per million PYR for this tumor when cities, sexes, and all ages ATB are combined to be significantly elevated. It is 0.30 (90% confidence interval: 0.16, 0.43) or approximately twice the risk associated with esophageal tumors. This change in events is attributable mainly to a marked increase in the number of deaths ascribed to this malignancy in the years 1975—1978. It is seen in both cities, more strikingly so in Hiroshima than in Nagasaki, however, and more conspicuously affects two age groups (10—19 or 35—49 years ATB). Malignancies of the Respiratory and Intrathoracic Organs Harada and Ishida [311,in an analysis of the Hiroshima City Tumor Registry, were the first to suggest that the incidence of lung cancer was significantly higher among those who were exposed within 1500 m of ground zero. Subsequent studies [see, for example, 32, 33 and 341 confirmed and extended their observations. The continuing mortality surveillance on the Life Span Study sample reveals the absolute risk of death from this malignancy per million PYR to be 0.61 (90% confidence interval: 0.37, 0.86). Both relative and absolute risks are higher in Hiroshima than in Nagasaki although the values for the two cities do not differ significantly. Only those for Hiroshima meet the conventional criteria for statistical significance. However, it should be noted that the absolute risk increased markedly in both cities in the years 1975—1978. For the cities collectively, for example, the absolute risk in any four year period beginning in 1955 which had previously varied within the range 0.30— 0.73 rose dramatically to 2.59 (95% confidence interval: 1.54, 3.64). The carcinogenic effect which previously was seen primarily in the older age groups is now readily seen in those individuals who were 20—34 and even 10—19 years of age ATB. Efforts to find evidence of an influence of either smoking or occupational exposure on this apparent radiation effect have failed thus far [28,35]. Malignancies of Bone,
Connective Tissue and Breast
Yamamoto and Wakabayashi [36], in a study of benign and malignant bone tumors found at autopsy or in surgical specimens in Hiroshima and Nagasaki in the years 1950—1965, were unable to demonstrate an association of these tumors and exposure distances. With the exception of osteosarcomas, however, the number of malignant bone tumors of a specified kind was too small to permit consideration according to the histological type of the tumor. Death certificate data have not been particularly informative here, for the number of deaths ascribed to malignant tumors of the bone are so few. Under this rubric, the malignancy most clearly radiation related is cancer of the breast. Wanebo et al. [37] were the first to establish this. Their study was limited, however, to women who were members of the Adult Health Study group. Subsequent investigations have focused on members of the Life Span Study sample, some 63000 females in all. The most recent of these [38] finds that (1) the distribution of histological types of mammary cancers does not vary significantly with radiation dose; (2) among all women who received at least ten rad, those irradiated before age 20 will experience the highest rates of breast cancer throughout their lifetimes; and (3) the incidence of malignant mammary tumors appears to be linearly related to dose in both cities. Nakamura and his colleagues [39] have examined the relationship between exposure
Late Radiation Responses in Man
235
and other epidemiological risk factors in the development of breast cancer among the survivors. Specifically, they examine the association of such risk factors as family history of malignant neoplasms, education, menstrual, marital and reproductive histories, and history of breast feeding. They conclude that the risk increases in proportion to the number of other risk factors involved. Based on mortality findings alone, Kato and Schull [28] report the absolute risk of death from cancer of the breast to be 0.50 per million PYR (females only) (95% confidence interval: 0.29, 0.72). For some malignancies, the cities do not differ significantly, albeit at face value the risk is less in Nagasaki than in Hiroshima, where the effect is statistically significant. Malignancies of the Genito—Urinary System Efforts to relate malignancies of this system to exposure to the A—bombs have involved both incidence studies and the continued mortality surveillance of the Life Span Study sample. In the former studies, Bean and his colleagues [40] were unable to demonstrate an association with exposure to an atomic bomb of the occurrence of prostatic adenocarcinoma in 1357 male members of the Life Span Study sample who were 50 years of age or older at death and were autopsied at ABCC—RERF between 1961 and 1969. Nor could they demonstrate an effect of exposure on histologic type or biologic activity. Sawada [41], in a much earlier study, attempted to evaluate the effect of A—bomb radiation on the occurrence of gynecologic tumors as seen in a series of routine examinations of 1785 exposed women and 1802 control subjects, but unfortunately, almost half of the women of this series would not accept an examination of their pelves. Carcinoma of the cervix was the most frequently encountered tumor; however, its occurrence could not be shown to depend on exposure nor was the age of onset significantly different between control and irradiated individuals. The most recent examination of deaths in the Life Span Study sample [28] finds a marginally significant increase in deaths attributed to malignancies of the urinary tract; the absolute risk (excess deaths! million PYR) was 0.15 (95% confidence interval: 0.04, 0.26). There was, however, no significant increase in deaths attributable to cancer of the uterus. Other Malignancies Malignancies of the eye, brain, “other” and unspecified parts of the nervous system; the thyroid gland and other endocrine glands and their related structures fall within this rubric. Included too are malignant neoplasms of “other and ill—defined” sites, malignancies without specification of site, and secondary malignant neoplasms of the lymph nodes, respiratory and digestive systems, as well as other sites. We are concerned only with primary tumors and of those within this rubric, only malignant neoplasms of the thyroid have been significantly associated with exposure to A—bomb radiation. Hollingsworth et al. [42] first noted, in a study of thyroid disease in Hiroshima, that carcinoma of the thyroid constituted some 7% of the total number of individuals with thyroid disorders and that a greater number of these malignancies were seen among the more heavily exposed individuals. They cautiously noted, however, that differences among distance groups were not statistically significant. Subsequent studies have removed any doubt of the association of this tumor with exposure. For example, Parker and his associates [43] summarized the results of the continuing, periodic examinations of the Adult Health Study group in the years 1958— 1971. In all, 74 cases of thyroid cancer in the Adult Health Study were seen and histologically confirmed in these years. They found thyroid carcinoma to be commoner in women and significantly more prevalent in persons exposed to 50 or more rads of atomic radiation. They also found carcinoma of the thyroid to be diagnosed during life more often in persons less than 20 years ATB although they failed to confirm the previously reported radiosensitivity in childhood. An increased risk was seen even among persons who were 50 years of age or older at the time of radiation exposure. It warrants note here that this is one of the few malignancies where a significant relationship with ionizing radiation has been demonstrable largely on clinical grounds. No effect has emerged through the continuing mortality surveillance of the Life Span Study sample, but few of the individuals with thyroid malignancies, confirmed either at autopsy or on histological section, have had their deaths ascribed to these malignancies. A recent analysis, however, of the incidence of cancer within the Life Span Study sample in Nagasaki, based on Tumor Registry data, revealed a significant excess of cancers of this particular kind; the absolute risk per million PYR rad was 1.32 (95% confidence interval: 0.88, 0.76). Before turning to other radiation related effects, it is important to note two general features associated with radiation related malignancies. First, the time from exposure to death is shortened for leukemia depending upon the dose received by the survivor but not for other cancers, and radiation—induced cancers other than leukemia seem to develop proportionally to the natural cancer rates for the attained age. Second, for specific age—at— death intervals, both relative and absolute risks tend to be higher for the young at the time of bombing. Simply stated, the young, particularly those in the first decade of life, seem especially vulnerable to radiation induced damage.
236
W.J. Schull
OTHER RADIATION—RELATED EFFECTS As has been implied,
an increased probability of malignancy is the greatest hazard confronting survivors, but there are other effects as well. The latter include an increased risk of radiation—related lenticular opacities, disturbances of growth among those survivors who had not achieved their adult dimensions at the time of exposure, an increase in mental retardation and small head sizes among the in utero exposed and an increase in the frequency of chromosomal aberrations in peripheral lymphocytes.
Cataracts The formation of radiation induced cataracts was the first late effect recognized unequivocally in individuals exposed to A—bomb radiation in Hiroshima and Nagasaki. Since the
original description of such cataracts by Cogan, Martin and Kimura (44], numerous other clinical, histopathological and statistical studies have appeared (see [45] for further references). It is now apparent that a radiation induced cataract is in its early stages, at least, a highly characteristic lesion. It is a central, posterior subcapsular opacity, easily visible with a alit lamp or an ophthalmoscope. Unfortunately, the word “cataract”
connotes to many a defect which impairs vision although it is commonly used to describe any detectable change in translucency in the lens. The vast majority (over 80%) of these lesions are nonprogressive, do not impinge on visual acuity and commonly the individual
involved is unaware of their presence. It is generally believed that the radiation damaged cells of the germinative epithelium are unable to differentiate into normal lens fibers, and that their remnants are gradually pushed to the posterior pole under the pressure of the remaining undamaged cells which continue to differentiate normally. The opacification itself arises from the denaturation of the normally transparent protein in the damaged cells, as the latter breaks down. Customarily the effects of ionizing radiation are viewed as either “stochastic,” if the
probability
of their occurrence is a direct function of dose, or “nonstochastic,”
if it is
the severity of the effect which is exposure—dependent (almost always in this instance there is a threshold). It is not always clear whether a particular radiation related effect should be viewed as a stochastic or a nonstochastic one, for often the sequence of biological events which culminate in the effect are unknown. Radiation carcinogenesis, for example, as well as mutation is generally assumed to be a stochastic effect; whereas lenti— cular opacities and mental retardation have commonly been viewed as nonstochastic. Otake and Schull’s [45] recent reanalysis of the Hiroshima and Nagasaki data-on lenticular opacities seems to bear out the classification of this lesion as nonstochastic. They find, among the numerous linear and quadratic models which they fitted to the frequency of occurrence of lenticular opacities, the “best” fit, that is, the one with the smallest chi—square and largest tail probability is a model which postulates a linear relationship with exposure and a threshold. Indeed, the threshold which they estimate from the Japanese experience is one consistent with the threshold which has been commonly conjectured, based on clinical experience.
Fetal Development It has lor.g been known that irradiation fetus.
has a damaging effect upon the development of the
This knowledge has prompted a continued clinical and mortality surveillance of those
individuals who were exposed in utero to the atomic bombs in Hiroshima and Nagasaki. Indeed, we have known for thirty years that among these children there is an increased prevalence of
mental retardation [46—50]. Several developments have occurred recently which impinge on these earlier studies. First, the appropriateness of the T65 exposure estimates have been questioned [8,9]. Second, Hashizumi et al. [51] have estimated the in utero fetal exposures to radiation from the atomic bombs; and Kerr [52] has given the estimates of absorbed dose to a first trimester fetus in terms of tissue kerma in air. While these latter estimates are predicated on radiant energies of the kind postulated in the T65 dosimetry, they provide the first and to date only basis for assessing fetal absorbed doses. Finally, but no less importantly, if the vulnerability of a developing organ to radiation damage is related to rate of growth, Dobbing and Sands’ [53] studies of the quantitative growth of the human brain provide a compelling basis for expecting those children exposed between 10 and 18 weeks of gestation, as measured from the last menstrual period, to be at the highest risk of mental retardation of all children exposed in utero. Two principal findings emerge from this reanalysis. First, although the developing forebrain is undoubtedly vulnerable throughout gestation, the highest risk of damage occurs at the time when the most rapid proliferation of neuronal elements occurs, namely, 10—17 weeks after conception. Overall, the risk is four times greater in these weeks than in subsequent ones. Second, in this period of rapid neuronal growth, damage — expressed as subsequent mental retardation — appears linearly related to the dose the fetus receives. Indeed, the variation in frequency of mental retardation with dose not explained by a linear model is consistently not significant. Radiation—related mental retardation in the latter stages of gestation (after the 17th week) is not fully explicable by a linear model and the observed values suggest that at these ages a threshold may exist.
Late Radiation Responses in Man
23]
It should be noted that the cohort of individuals exposed in utero have not as yet been shown to have an increased risk of premature mortality from all causes nor has it been possible to demonstrate that their risk of leukemia or any of the other malignancies known to be radiation related in those survivors not exposed in utero is increased. Thus, to date, the most significant finding within this cohort is the increased risk of mental retardation and diminished head size [49]. Chromosomal Aberrations No less than 15 years ago [54], radiation—induced chromosomal aberrations were shown to persist in cultured peripheral blood lymphocytes derived from Hiroshima and Nagasaki A—bomb survivors long after their radiation exposure. Recently, Awa and his colleagues [55] have reconfirmed the increased frequency of cells with chromosome aberrations with increasing dose in both cities. However, in every dose group, the frequency of aberrant cells was consistently higher in Hiroshima than in Nagasaki. They suggested that a higher neutron dose in Hiroshima than in Nagasaki may have been a major component contributing to this difference. This seems less likely now, given the dose reassessment. However, even under the reassessed values, there continues to be a difference between the two cities. Among the types of chromosomal aberrations which have been identified thus far, reciprocal translocations were observed to predominate, and they played an important role in determining the dose—aberration relationship. It warrants note too that this is one of the few effects which seems statistically discernible even within the 1—9 rad dose range. The health significance of these findings continues to be problematic. Some years ago an attempt was made to correlate the occurrence of chromosomal abnormalities in A—bomb survivors with the findings on their periodic examinations within the Adult Health Study program [56]. No consistent correlation between the clinical parameters evaluated and the degree of cytogenetic abnormalities emerged. The thought, however, that these chromosomal aberrations may be prodromal continues to titillate investigators but a persuasive answer to their role seems unlikely to emerge until more deaths have occurred among those individuals who have been studied karyotypically over the past decade and a half. Growth Disturbances in Children Finally, we have adverted briefly to growth disturbances among those survivors who had not reached skeletal maturity at the time of their exposure. Nehemias [57] noted, “Differences with respect to size which are statistically very significant are found among the subgroups defined on the basis of the radiation exposure index. The actual physical differences are small, however. The differences have been shown to be in the direction of decrease in size with increase in degree of radiation exposure.” More recently, Belsky and Blot [58] have examined the stature of adults exposed in childhood to the atomic bombs of Hiroshima and Nagasaki. They found adult height to be significantly lowered among Hiroshima males and females exposed to high doses (100+ rad) of irradiation when they were ages 0—5 years ATB; however, the effect of dose declined with increase in age ATB. For Nagasaki, no statistical differences in means for height were noted by A—bomb dose group although for females who were 0—5 years ATB the highest dose group had the smallest mean height. Belsky and Blot conclude that a growth retarding effect of radiation can be demonstrated for the youngest heavily exposed children. If their findings are taken in concert with those of Nehemiss, a somatic effect- on growth seems highly probable. The mechanism(s) is, of course, unknown. SUMMARY To summarize,
cancers of the breast, colon, esophagus,
lungs, stomach, thyroid and urinary
tract as well as leukemia and multiple myeloma increase in frequency with an increase of exposure. Furthermore, it has been demonstrated that radiation—related lenticular opacities, disturbances of growth among those survivors who had not achieved their adult dimensions at the time of exposure, and mental retardation and small head sizes among the in utero exposed are increased. Chromosomal aberrations in peripheral lymphocytes are more frequent too. The health implications of some of these effects are uncertain, for example the chromosomal aberrations, but most unquestionably compromise both the duration of life and its quality. Indeed, it can be shown that among the atomic bomb survivors there is a significant foreshortening of life, explicable in terms of their increased probabilities of death from a malignant tumor. Among these effects, however, if they are examined on a dose level—by dose level basis, significant effects emerge generally only among those individuals who have received exposures of 50 rad (tissue kerma) or higher. Only chromosomal aberrations can be shown to be unequivocally elevated at lesser exposures. Under these circumstances, the hazards one foresees at these lower dose levels is functionally related to the statistical model one presumes best describes the fundamental dose—response relationship. Prudence counsels that this relationship should be construed as linear until such time as proof to the contrary exists.
238
W.J. Schull
REFERENCES 1. 2.
3. 4. 5. 6.
7. 8.
9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45.
G.W. Beebe, Epidemiologic Reviews 1, 184 (1979) J.V. Neel and W.J. Schull, The Effect of Exposure to the Atomic Bombs on Pregnancy Terminations in Hiroshima and Nagasaki, National Academy of Sciences—National Research Council, Washington, D.C. (1956) W.J. Schull, H. Otake and J.V. Heel, Science 213, 1220 (1981) G.W. Beebe and M. Usagawa, Atomic Bomb Casualty Commission Technical Report 12—68 (1968) T. Yamamoto, I.M. Moriyama, M. Asano and L. Guralnick, Radiation Effects Research Foundation Technical Report 18—78 (1978) J.A. Auxier, Ichiban: Radiation Dosimetry for the Survivors of the Bombings of Hiroshima and Nagasaki, Technical Information Center, Energy Research and Development Administration (1977) R.C. Milton and T. Shohoji, Atomic Bomb Casualty Commission Technical Report 1—68 (1968) G.D. Kerr, in: Proceedings of the Fourth Symposium on Neutron Dosimetry, Gesellschaft fur Strahlen—und Umweltforschung, Munich—Neuherberg, June 1—5, 1981 Volume 1, 501 (1981) W.E. Loewe and E. Mendelsohn, Lawrence Livermore Laboratory Publication UCRL—86595 (1981) J. Marcum, an unpublished manuscript entitled House Attenuation Factors for Radiation at Hiroshima and Nagasaki prepared for Director, Defense Nuclear Agency J.H. Folley, W. Borges and T. Yamawaki, American Journal of Medicine 13, 11 (1952) R.D. Lange, W.C. Moloney and T. Yamawaki, Blood 9, 574 (1954) W.C. Moloney, New England Journal of Medicine 253, 88 (1955) W.C. Mooney and R.D. Lange, Blood 9, 663 (1954) M. Tomonaga, New Zealand Medical Journal of Hematology 65 (Supplement), 863 (1966) W.C. Moloney and M.A. Kastenbaum, Science 121, 308 (1955) H. Ichimaru, T. Ishimaru, M. Mikami, Y. Yamada and T. Ohkita, Journal of Radiation Research (in press) H. Nishiyama, R.E. Anderson, T. Ishimaru, K. Ishida, Y. Ii and N. Okabe, Cancer 32, 1301 (1973) R.E. Anderson, T. Hoshino and T. Yamamoto, Annals of Internal Medicine 60, 1 (1964) R.E. Anderson and T. Yamamoto, in: Myeloproliferative Disorders of Animals and Man, Hanford Biological Symposium, Richland, Washington (1968) N. Ichimaru, T. Ishimaru, M. Mikami and M. Matsunaga, Radiation Effects Research Foundation Technical Report 9—79 (1979) J.A. Pinkston, T. Wakabayashi, T. Yamamoto, M. Asano, Y. Hirada, H. Kumagami and M. Takeichi, Radiation Effects Research Foundation Technical Report X—80 (1980) W.M. Schreiber, H. Kato and J.D. Robertson, Cancer 26, 69 (1970) M. Asano, S. Seyama, H. Itakura, T. Hamada, S. lijima and H. Kato, Transactions of the Society of Pathology, Japan 67, 143 (1978) J.D. Robertson, H. Kato and W.M. Schreiber, Atomic Bomb Casualty Commission Technical Report 7—70 (1970) R.W. Cihak, T. Kawashima and A. Steer, Cancer 29, 1133 (1972) G.W. Beebe, H. Kato and C.E. Land, Radiation Research 25, 138 (1978) H. Kato and W.J. Schull, Radiation Research 90 (1982) K. Nskamura, Radiation Effects Research Foundation Technical Report 8—77 (1977) K. Nakamura, Lancet ii, 86 (1977) T. Hirada and M. Ishida, Journal of the National Cancer Institute 25, 1253 (1960) A. Ciocco, Atomic Bomb Casualty Commission Technical Report 18—65 (1965) G.W. Beebe, T. Yamamoto, Y.S. Matsumoto and S.E. Gould, Atomic Bomb Casualty Commission Technical Report 8—67 (1967) C.K. Wanebo, K.G. Johnson, K. Sato and T.W. Thorslund, American Reviews of Respiratory Diseases 98, 778 (1968) T. Ishinaru, T.W. Cihak, C.W. Land, A. Steer and A. Yamada, Cancer 38, 1723 (1975) T. Yaniamoto and T. Wakabayashi, Acts Pathologica Japonica 19, 201 (1969) C.K. Wanebo, K.G. Johnson, K. Sato and T.W. Thorslund, Atomic Bomb Casualty Commission Technical Report 13—67 (1967) M. Tokunaga, J.E. Norman, Jr., M. Asano, S. Tokuoka, H. Ezaki, I. Nishimori and Y. Tsuji, Journal of the National Cancer Institute 62, 1347 (1979) K. Nakamura, D.H. McGregor, H. Kato and T. Wakabayashi, Radiation Effects Research Foundation Technical Report 9—77 (1977) M.A. Bean, R. Yatani, P.1. Liu, K. Fukazawa, F.W. Ashley and S. Fujita, Cancer 32, 496 (1973) H. Sawada, Atomic Bomb Casualty Commission Technical Report 6—59 (1959) D.R. Hollingsworth, H.B. Hamilton, H. Tamagaki and G.W. Beebe, Medicine 42, 47 (1976) L.N. Parker, J.L. Belsky, T. Ysmamoto, S. Kawamoto and R.J. Keehn, Annals of Internal Medicine 80, 600 (1974) D.C. Cogan, S.F. Martin and S.J. Kimura, Science 110, 654 (1949) M. Otake and W.J. Schull, Radiation Research 91 (1982)
Late Radiation Responses in Man
46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58.
C. Plummer, Pediatrics 10, 687 (1952) R.W. Miller, Pediatrics 18, 1 (1956) R.W. Miller and W.J. Blot, Lancet ii, 784 (1972) W.J. Blot and R.W. Miller, Radiology 106, 617 (1973) M. Otake and W.J. Schull, Lancet (in press) (1982) T. Hashizume, T. Maruyama, K. Nishizawa and A. Nishimura, Journal of Radiation Research 1.4, 346 (1973) G.D. Kerr, Health Physics 37, 487 (1979) J. Dobbing and J. Sands, Archives of Diseases of Childhood 48, 757 (1973) A.D. Bloom, S. Neriishi, A.A. Awa, T. Honda and P.G. Archer, Lancet ii, 802 (1967) A.A. Awa, T. Sofuni, T. Honda,M. Itoh, S. Neriishi and M. Otake, Radiation Effects Research Foundation Technical t 12—77 (1977) R.A. King, J.L. Belsky, M. Otake, A.A. Awa and T. Matsui, Atomic Bomb Casualty Commission Technical Report 15—72 (1972) J.V. Nehemias, Atomic Bomb Casualty Commission Technical Report 22—61 (1961) J.L. Belsky and W.J. Blot, Atomic Bomb Casualty Commission Technical Report 35—71 (1971)
239