Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 377
THE CHILDREN OF ATOMIC BOMB SURVIVORS: A Genetic Study Congenital Malformations, Stillbirths, and Early Mortality among the Children of Atomic Bomb Survivors: A Reanalysis MASANORI OTAKE,* WILLIAM J.SCHULL,†,1AND JAMES V.NEEL‡ * Department of Statistics, Radiation Effects Research Foundation (RERF), Hiroshima, Japan; †RERF, and The Genetics Centers, Graduate School of Biomedical Sciences, University of Texas Health Science Center, Houston, Texas; and ‡Department of Human Genetics, University of Michigan, Ann Arbor, Michigan OTAKE, M., SCHULL, W.J., AND NEEL, J.V.Congenital Malformations, Stillbirths, and Early Mortality among the Children of Atomic Bomb Survivors: A Reanalysis. Radiat. Res. 122, 1– 11 (1990). Of all the data sets pertinent to the estimation of the genetic risks to humans following exposure to ionizing radiation, potentially the most informative is that composed of the cohort of children born to atomic bomb survivors. We present here an analysis of the relationship between parental exposure history and untoward pregnancy outcomes within this cohort, using to the fullest extent possible the recently revised estimates of the doses received by their parents, the so-called DS86 doses. Available for study are 70,073 terminations, but DS86 doses have not been or presently cannot be computed on the parents of 14,770. The frequency of untoward pregnancy outcomes, defined as a pregnancy terminating in a child with a major congenital malformation, and/or stillborn, and/or dying in the first 14 days of life, increases with combined (summed) parental dose, albeit not significantly so. Under a standard linear model, when the sample of observations is restricted to those children whose parents have been assigned the newly established DS86 doses (n =55,303), ignoring concomitant sources of variation and assuming a neutron RBE of 20, the estimated increase per sievert in the predicted frequency of untoward outcomes is 0.00354 (±0.00343). After adjustment for concomitant sources of variation, the estimated increase per sievert in the proportion of such births is 0.00422 (±0.00342) if the neutron RBE is assumed to be 20. A “one-hit” model with appropriate adjustments for extraneous sources of variation results in an almost identical value, namely, 0.00412 (±0.00364). When the sample is extended to include parents lacking the full array of dose parameters necessary to calculate the DS86 dose, but sufficient for an empirical conversion of the previously employed T65DR dose system to its DS86 equivalent, we find under the linear model that the estimated increase per sievert in untoward pregnancy outcomes is some 31% higher than that published previously, 0.00264 (±0.00277), assuming an RBE of 20, after adjustment for extraneous sources of variation. (Since a dose could not be calculated in 367 of the 70,073 outcomes, the n=69,706.) The corresponding value with the one-hit model is 0.00262 (±0.00294). INTRODUCTION Estimation of the genetic risks to humans following exposure to ionizing radiation has been and continues to be one of the most important and difficult tasks that confront radiobiologists. A variety of kinds of experiences and data are available on which to base estimates. These include (a) exposure to diagnostic and therapeutic doses of X rays and radioactive materials such as radium or cobalt-60; (b) occupationally incurred exposures, for example, in uranium mining or the maintenance of nuclear reactors; (c) geographic areas with “high” natural or manmade background radiation levels, and (d) the atomic bombings of Hiroshima and Nagasaki. Exposure from these sources varies substantially in both radiation quality and dose. The exposure may be acute or chronic, of a single or more qualities, whole-body or partial, and possibly confounded by the health of the exposed person. Indeed, so disparate are the exposures and the methods of case ascertainment that have been used that it is impossible to examine these data in any collective manner. Of all the possible data sets, the cohort of children born to atomic bomb survivors is potentially the most informative. Recognition of this fact led to the initiation of a joint United States-Japan program of investigation shortly after the war, conducted initially by the Atomic Bomb Casualty Commission (ABCC) and now by the Radiation Effects Research Foundation (RERF). Those studies have provided more data than all other similar studies combined. It needs to be noted, however, that although the sample sizes available are large by the standards of conventional epidemiological studies, the average gonadal doses are well below those employed experimentally. On the other hand, the study deals with a population of different socio-economic states and occupations, a variety of ages, and both sexes, and can be termed a “natural population.” These data do not confound medical indications for the use of ionizing radiation with the effects of ionizing radiation. Finally, the new system of dosimetry, the DS86, provides better estimates of individual doses than the earlier T65DR Reproduced, with permission, from Radiation Research, vol. 122, © 1990 by Academic Press. 1 Permanent director, RERF.
OCR for page 378
THE CHILDREN OF ATOMIC BOMB SURVIVORS: A Genetic Study dosimetry for the survivors. These dose estimates are unquestionably more reliable than those known for the majority of other populations. Most of the methods of detection of mutations that are feasible in the study of human populations have been employed in the study of the offspring of A-bomb survivors. These include a search for changes in the frequency of (a) certain population characteristics, for example, the occurrence of major congenital defects or premature death (1–4); (b) sentinel phenotypes (5,6); (c) chromosomal abnormalities (7,8); and (d) biochemical variants of a structural or kinetic nature (9,10). Although these changes are diverse, the aim of identifying them is the same—to estimate the probability of mutation per unit exposure to ionizing radiation and to ascertain the public health implications of an increase in the number of mutations. The purpose of this paper is to reexamine one facet of these studies of the offspring of the survivors of the atomic bombing of Hiroshima and Nagasaki: the clinical data on adverse pregnancy outcomes obtained in the years 1948– 53. This reanalysis has been prompted by the revised dose estimates. The results of this reanalysis, of interest in its own right, will in a subsequent publication be combined with the results of similar analyses of all of the other data now available on these children. The aim is to generate an estimate of the lowest amount of radiation (95% probability level) that will produce a mutational impact equal to that arising spontaneously each generation, i.e., a minimal doubling dose, and also an estimate of the actual doubling dose suggested by the analysis.2 MATERIALS AND METHODS The Study Sample We have described the nature of the birth registration process involved in the ascertainment between 1948 and 1953 of the 70,073 pregnancies lasting at least five lunar months which resulted in the potential study group, the completeness of this process, and the information that was collected (1–4). Briefly, a continuous genetic surveillance of the children born in Hiroshima and Nagasaki subsequent to the atomic bombing began in 1948. The initial study, the one of concern here, utilized the postwar Japanese rationing system as a case-finding mechanism. A provision of this system entitled pregnant women who registered their pregnancies after the fifth lunar month to supplementary foodstuffs. As a consequence of the stringency of the postwar economic circumstances, most women registered their pregnancies, and it was possible through this means to identify more than 95% of all pregnancies in these cities persisting for at least 20 weeks of gestation and arrange for the examination of the progeny by a physician. These observations were generally made in the home and were supported by an infant autopsy program and a second examination of some 30% of surviving infants at the clinical facilities of ABCC 8–10 months after their birth [for a fuller description of this study see (1)]. The examination of newly born infants, in a program such as that just described, can provide the following information: sex, birth weight, viability at birth, presence of major malformation, and occurrence of death during the first weeks of life, all of which are potentially confounded by a variety of extraneous factors. Of the extraneous factors for which data exist, the most important are year of birth, city, sex, birth rank, and the ages of the parents at the birth of the child. In this analysis our interest is focused on three aberrant outcomes of pregnancy: the birth of a child with a major congenital anomaly, and/or who was stillborn, and/or who succumbed within 2 weeks following birth. [Note: A major congenital anomaly in this context is defined as one that is incompatible with survival, is life-threatening, or seriously compromises the individual's capacity to function normally in society. A detailed listing of the major congenital malformations encountered in this study has been presented previously (1,11).] We designate a pregnancy which terminated in an infant exhibiting one or more of these potential indicators of radiation damage as an untoward pregnancy outcome. Elsewhere we will estimate what fraction of these untoward outcomes can be considered directly related to mutation in the preceding generation.2 Dosimetry To avoid ambiguity, we define the four terms that we shall use to describe dose and exposure: the free-in-air (FIA) kerma (kinetic energy released in material), kerma in shielded areas (termed here “shielded kerma”), organabsorbed dose, and organ dose equivalent. The first describes the kerma in tissue at a point in air over bare ground (i.e., not in or near a building); the second, the kerma of the individual with allowance for structural shielding; the third, the radiation (?, neutron) absorbed by the organ or tissue under consideration; and the last, the sum of the products of the various absorbed doses times their quality factors, or relative biological effectiveness (RBE). The gonad neutron doses were low, and we assume an RBE of 20 for the genetic effect of neutrons (12) and an RBE of 1 for ? rays. Kerma and organ-absorbed doses are expressed in grays and the organ dose equivalent in sieverts. Our previous analyses of radiation-related risks among the offspring of A-bomb survivors have been based on the estimated T65DR doses of their parents (13). This system of dosimetry estimated a kerma in air with allowance for shielding, where appropriate, for most individual survivors (98%) within the RERF Life Span Study (LSS) sample, which included most of the proximally exposed parents (14). Organ-absorbed doses were assigned using fixed coefficients to describe the attenuation of radiation through organs and tissues before it reached the specific organ or tissue of interest (15). Shielded kerma and organ-absorbed doses were estimated separately for neutrons and ? rays. The work of a number of investigators (16–18) led to a reassessment of the A-bomb radiation dosimetry sponsored by the United States and Japan (19). This reappraisal has resulted in the Dosimetry System 1986 (DS86) (20). Patently, evaluation of the genetic effects of the A-bombs requires precise knowledge regarding the radiation exposure of the pertinent individuals, in this instance, the parents of the children in the study sample. Accordingly we describe in some detail exactly how these doses were estimated. We begin by distinguishing between two sets of doses, one the DS86 system and the other an ad hoc dose system we have developed. DS86 in this context implies a dose calculated by one of the methods installed at the RERF by the Science Applications International Corp. The bases and manner of computing doses in these instances are described either in Roesch (20) or in Preston and Pierce (21). Recently, as an addition to the descriptions in (20) and (21), the system has been extended to include direct computation of doses for those survivors in Nagasaki who were terrain-shielded or exposed in factories. This involves approximately 1000 survivors in the LSS 2 J.V.Neel, W.J.Schull, A.A.Awa, C.Satoh, H.Kato, M.Otake, and Y.Yoshimoto. The children of parents exposed to atomic bombs: Thoughts on the genetic doubling dose of radiation for humans. Manuscript in preparation.
OCR for page 379
THE CHILDREN OF ATOMIC BOMB SURVIVORS: A Genetic Study TABLE I A Comparison of the Shielding of Parents with Reference to Whether DS86 Doses Have Been Assigned DS86 dose known DS86 dose unknown Shielding category Mothers (17,010) Fathers (17,010) Mothers (12,207) Fathers (12,207) In open 3.4 5.9 4.1 6.1 Japanese building 86.4 81.7 71.2 65.2 Other 10.2 12.4 24.7 28.7 sample. Rules have also been formulated for the indirect estimation of doses for a larger number of survivors exposed beyond 2000 m. Specifically, this includes individuals exposed in the open or for whom no shielding information is available. While this recent development has further reduced the number of survivors without DS86 doses, there are still groups of individuals for whom doses are not available. Many of these individuals were exposed within 2000 m and can therefore be presumed to have received significant doses. It is this latter group, which involves one or both of the parents for approximately 14,000 pregnancy terminations, that is the primary focus of the ad hoc doses. The exposure data on these parents were collected between 1948 and 1954 (cf. (1)), a decade before the present detailed exposure history was codified, and while a T65DR-type dose can usually be calculated, the data are simply not sufficiently detailed for the DS86 system. The 14,000 children described above constitute a 26% addition to the sample for this study, and including them in the study will enable us to provide information applicable to other samples where a similar problem exists. We have therefore returned to the original sample tape for this study. Through matching parental master file numbers on the tape with the roster of individuals, we have divided the parents into two groups: those for whom a DS86 dose has been calculated, and those for whom it has not. Within each of the two groups, we have tabulated the categorical data on parental shielding recorded in the course of the study. The results are shown in Table I. The tabulated values are percentages of individuals reporting this particular type of shielding among 12,207 parents in the DS86 unknown group and 17,010 among the DS86 known. It should be noted that the most common shielding by far in each group is a Japanese wooden TABLE II A Comparison of the T65DR Doses (G4) Assigned when a DS86 Dose Can Be Assigned and when it Cannot, further Subdivided by Distance and by Presence or Absence of Symptoms T65DR Dose Distance Symptoms 0.01–6.0 0.01–0.09 0.1–0.19 0.2–0.49 0.5–0.99 1.0+ 0 a. T65DR known, DS86 known <2.0 km Symptoms No. 2471 217 240 374 319 1321 15 Dose 165.1 5.8 14.7 31.1 75.0 278.3 0.0 No symptoms No. 10235 2034 2555 2448 1612 1586 6276 Dose 55.7 5.7 14.6 31.5 70.2 208.7 0.0 Total No. 12706 2251 2795 2822 1931 2907 6291 Dose 77.0 5.7 14.6 31.5 71.0 240.3 0.0 ≥2.0 km Symptoms No. 225 168 57 295 Dose 5.3 2.1 14.7 0.1 No symptoms No. 4180 3632 547 1 19297 Dose 3.3 1.6 14.3 19.6 0.0 Total No. 4405 3800 604 1 19592 Dose 3.4 1.6 14.3 19.6 0.0 b. T65DR known, DS86 unknown <2.0 km Symptoms No. 470 19 28 37 68 318 22 Dose 204.2 4.5 15.7 34.7 74.7 280.2 0.0 No symptoms No. 1408 213 229 297 260 409 95 Dose 90.2 5.4 15.5 32.1 70.8 230.6 0.0 Total No. 1878 232 257 334 328 727 117 Dose 118.7 5.3 15.5 32.4 71.6 252.3 0.0 ≥2.0 km Symptoms No. 15 9 6 2 Dose 7.3 3.9 12.5 0.2 No symptoms No. 269 206 63 65 Dose 5.4 2.7 13.9 0.2 Total No. 284 215 69 67 Dose 5.5 2.8 13.8 0.2 Note. Further explanation in text.
OCR for page 380
THE CHILDREN OF ATOMIC BOMB SURVIVORS: A Genetic Study TABLE III The Distribution of Joint Parental DS86 Gonadal Dose Equivalents, Based on a Neutron RBE of 20, by City and Dose Category, for the Cohort of Children Examined for Untoward Pregnancy Outcome Dose category (Sv) City Total <0.01 0.01–0.09 0.10–0.49 0.50–0.99 1.00–2.49 2.50+ 0.01+ Hiroshima Subjects 31283 24449 3761 1958 586 416 113 6834 N mean (Sv)a 0.01 0.00 0.00 0.02 0.08 0.23 0.87 0.04 G mean (Gy) 0.06 0.00 0.04 0.22 0.61 1.29 2.84 0.26 T mean (Sv) 0.07 0.00 0.04 0.24 0.69 1.52 3.71 0.30 Nagasaki Subjects 24020 20760 1724 711 488 286 51 3260 N mean (Sv) 0.00 0.00 0.00 0.01 0.03 0.09 0.32 0.02 G mean (Gy) 0.05 0.00 0.03 0.25 0.69 1.36 3.34 0.34 T mean (Sv) 0.05 0.00 0.03 0.25 0.73 1.46 3.66 0.36 Both cities Subjects 55303 45209 5485 2669 1074 702 164 10094 N mean (Sv) 0.01 0.00 0.00 0.01 0.06 0.18 0.70 0.03 G mean (Gy) 0.05 0.00 0.03 0.23 0.65 1.32 3.00 0.29 T mean (Sv) 0.06 0.00 0.03 0.24 0.70 1.49 3.70 0.32 a N=20×neutron dose, G=? dose, T=total (?+20×neutron) dose. building. Further, although more parents are in the “other” category when the DS86 is unknown, this category does not necessarily imply shielding in a concrete structure or air raid shelter, but includes parents stated to be behind fences or trees, under the eaves of buildings, and the like. Presumably a DS86 dose could have been computed on most of the DS86 unknown group in wooden structures had detailed shielding information been available, and possibly on some of the “others” as well. To proceed further, we have divided the DS86 known and DS86 unknown groups into two mutally exclusive subgroups for those parents for whom ? and neutron exposures have been calculated by the T65DR dose system: survivors exposed at 2000 m or less and those exposed at more than 2000 m. These two groups were further sorted into the asymptomatic and the symptomatic, i.e., those survivors who reported one or more of three cardinal signs or symptoms of acute radiation sickness: epilation, subcutaneous bleeding, or oropharyngeal lesions. Individuals in the eight groups that resulted were then distributed over five dose intervals, and within each interval the mean T65DR dose was computed. The results are shown in Table II. It will be noted that the mean T65DR is virtually the same for the DS86 unknown and the DS86 known groups within each dose interval, and this is true with respect to persons exposed at 2000 m or less, more than 2000 m, or classified as symptomatic. We also note that within a T65DR dose group the percentage of symptomatic individuals is approximately the same in the DS86 known and unknown groups. These facts suggest that for these individuals the distribution of T65DR doses is independent of factors that determine whether or not a DS86 dose can be computed. If this be so, then the distribution of DS86 doses within the subgroups where both doses are known can be used to provide an ad hoc TABLE IV The Distribution of Joint Parental Gonadal Absorbed Dose Equivalent by City and Dose Category in the Cohort of Births Seen in 1948–1953, Extended (DS86+ad hoc Dose) Cohorta Dose category (Sv) City Total <0.01 0.01–0.09 0.10–0.49 0.50–0.99 1.00–2.49 2.50+ 0.01+ Hiroshima Subjects 34113 26300 4229 2230 667 507 180 7813 N mean (Sv)b 0.01 0.00 0.00 0.02 0.08 0.24 0.93 0.05 G mean (Gy) 0.07 0.00 0.04 0.22 0.61 1.29 2.89 0.28 T mean (Sv) 0.08 0.00 0.04 0.23 0.69 1.53 3.82 0.33 Nagasaki Subjects 35593 30997 2257 1088 737 409 105 4596 N mean (Sv) 0.00 0.00 0.00 0.01 0.03 0.09 0.29 0.02 G mean (Gy) 0.05 0.00 0.03 0.26 0.68 1.35 3.07 0.37 T mean (Sv) 0.05 0.00 0.03 0.27 0.71 1.44 3.36 0.40 Both cities Subjects 69706 57297 6486 3318 1404 916 285 12409 N mean (Sv) 0.01 0.00 0.00 0.01 0.06 0.18 0.69 0.04 G mean (Gy) 0.06 0.00 0.03 0.23 0.65 1.31 2.96 0.32 T mean (Sv) 0.06 0.00 0.03 0.24 0.70 1.49 3.65 0.36 a A neutron RBE of 20 is assumed. b N=20×neutron dose, G=? dose, T=total (?+20×neutron) dose.
OCR for page 381
THE CHILDREN OF ATOMIC BOMB SURVIVORS: A Genetic Study TABLE V Increments or Decrements of Change in the Individual Frequencies of Congenital Malformation, Stillbirths, and Neonatal Deaths in the Original Cohort of Births per Sievert of Joint Parental Gonadal Dose Equivalent Based Upon an Assumed Neutron RBE of 20, DS86 Cohort Only Variable Regression coefficient Standard error Regression model: Pi=Constant+ (Background) +bD Dosei Malformation Joint parental exposure 0.00099 0.00184 Birth order of child 0.00087* 0.00042 Year of birth 0.00120** 0.00032 Stillbirths Joint parental exposure 0.00151 0.00199 Birth order of child –0.00054 0.00042 Year of birth –0.00048 0.00035 Neonatal deaths Joint parental exposure 0.00237 0.00233 Birth order of child 0.00011 0.00049 Year of birth 0.00094** 0.00041 Regression model: Pi=1-exp-(Constant+ (Background)+bD Dosei) Malformation Joint parental exposure 0.00106 0.00233 Birth order of child 0.00091*** 0.00054 Year of birth 0.00116** 0.00042 Stillbirths Joint parental exposure 0.00138 0.00244 Birth order of child –0.00064 0.00054 Year of birth –0.00050 0.00044 Neonatal deaths Joint parental exposure 0.00233 0.00272 Birth order of child 0.00028 0.00059 Year of birth 0.00101 0.00049 Note. Significance levels: *(P<0.05), **(P<0.01), ***(P<0.10). estimate of the total kerma for the subgroups where the DS86 dose is unknown. Various methods of assigning an ad hoc dose come to mind. We proceeded on the principle that within the six dose intervals shown in Table II (0.01–0.19,...,?1.0 Gy), the T65DR dose on an individual could be multiplied by the ratio of the DS86 to the T65DR mean total kerma to obtain an ad hoc kerma dose. Accordingly, within the DS86 known group, the mean total kerma and its ? and neutron components, for the six dose intervals, were computed for the two cities separately (data not shown). From these data, we have calculated the ratio of the means for the total DS86 and total T65DR kerma. These ratios we term the dose conversion factors. These factors are strikingly different between the two cities, as was anticipated because of the difference in the neutron component attributed to the two explosions in the T65DR doses. Furthermore, whereas the dose conversion factors are functionally dependent upon the dose interval in Hiroshima, there is a much lesser dependence, if one at all, in Nagasaki. Again, this is in keeping with the changes in kerma associated with the DS86 system [see (20)]. To obtain the ad hoc estimates of total kerma, we have multipled each individual's T65 estimate of kerma within a dose interval in the unknown DS86 group (for the six intervals) by the dose conversion factor within that interval in the city where the parent was exposed. To partition the estimated total kerma into neutron and ? components, the ad hoc total kerma dose was multipled by the proportion of the known DS86 dose within the interval ascribed to neutrons. Finally, the ad hoc total kerma was divided into neutron and ? doses using these proportions. Where the T65DR dose was stated to be zero, we have presumed the DS86 to be the same and assigned this value as the ad hoc estimate. To obtain the putative gonadal doses, the ad hoc kerma estimates were multipled by the DS86 average ovary and testis transmission factors for ? and neutron emanations based on the LSS cohort. These values are 0.74 and 0.16 for ? rays and neutrons, respectively, for the ovary and 0.78 and 0.32 for the testes. We have disregarded the neutron-gamma capture factors, for it seemed too tenuous to attempt to specify the appropriate fraction of the total kerma assignable to neutron-capture gamma rays. In keeping with the higher frequency of one or more symptoms of radiation exposure encountered in the DS86 unknown dose group within 2000 m at the time of bombing (ATB), which was 25% compared to 19% for the DS86 known group, we find that the system described above assigns to them an average kerma dose of 0.75 Gy in contrast to the dose of 0.53 Gy computed for the DS86 known group within 2000 m. The approximate nature of this procedure is clear. Possibly the most important assumption on which it rests is the tacit one that the distribution of types of shielding is the same, or approximately so, in the DS86 known and unknown groups since the DS86 doses used to obtain the crude adjustment factors are estimates of shielded kerma. As shown earlier, this assumption seems reasonably well met by the data on shielding as categorically defined. In this connection, we are impressed at how well the biological dosimetry (frequency of epilation, chromosomal damage) reinforces the physical. Table III, using only those pregnancy terminations for which there are DS86 doses, or one or both parents are known not to have been in Hiroshima or Nagasaki ATB, summarizes the number of pregnancies ascertained, the mean combined (summed) parental absorbed gonadal dose, and the dose equivalent, assuming an RBE for neutrons of 20, by city for six dose categories: 0, 0.01–0.09, 0.10–0.49, 0.50–0.99, 1.00–2.49, and ?2.50 Gy. Table IV gives the comparable values when the parents for whom ad hoc doses were calculated are included. The lowest combined dose group, that is, those individuals exposed to less than 0.01 Gy, includes not only those persons present in the city and exposed to less than the stated dose but also any spouses who were not present in the city at the time of the bombing [the so-called not-in-city group (NIC)]. As will be noted from a comparison of the average gonadal absorbed doses with the average dose equivalents given in Table III (see also Table IV), other values for the neutron RBE would not materially alter the results to be described subsequently, for under the DS86 system of dosimetry the neutron contribution is small, and the RBE assigned to neutrons is of minor importance in determining the slope of the dose-response relationship. Statistical Methods In the analysis of untoward pregnancy outcomes, the effects of radiation have been estimated using two overlapping samples of observations. The
OCR for page 382
THE CHILDREN OF ATOMIC BOMB SURVIVORS: A Genetic Study TABLE VI The Distribution of Untoward Pregnancy Outcomes (UPO) and Total Subjects by Parental Gonadal Dose Equivalents in Sieverts (DS86) Based on an Assumed Neutron RBE of 20, Sexes and Cities Combined, DS86 Cohort Only Father's gonadal dose equivalent (Sv) Mother's gonadal dose equivalent (Sv) <0.01 0.01–0.09 0.10–0.49 0.50–0.99 1.0+ Subj UPO Subj UPO Subj UPO Subj UPO Subj UPO Mean dose 1.00+ 388 19 21 2 9 0 9 1 15 1 1.66 0.50–0.99 651 44 19 0 24 1 47 4 17 1 0.70 0.10–0.49 1655 81 124 9 209 13 41 2 20 0 0.24 0.01–0.09 3790 179 700 27 138 5 27 2 45 2 0.04 <0.01 45234 2257 1104 60 510 21 238 12 268 17 0.00 Mean dose 0.00 0.03 0.25 0.71 2.08 first sample (n=55,303) is restricted to those pregnancy terminations for which there are DS86 doses for both parents or it is known they were not in the city ATB. The second sample (n=69,706) consists of the first sample augmented by those pregnancy terminations where an ad hoc dose could be computed on the basis of a previous T65DR dose estimate. Data were collected on 76,626 infants, but after exclusions because the pregnancy had not been previously registered or the biomedical or exposure data were deficient in some respect [see (1)], the number available for analysis was 69,706. In the original analysis (1), the offspring of consanguineous parents were excluded because of the possible heterogeneity they would introduce to the sample, but they are included in the present analysis since a subsequent analysis (22) did not indicate that they responded differently to radiation. Information regarding the diagnosis of a major congenital defect, the occurrence of a stillbirth, or death in the first 14 days following birth is available on all these infants. The decision to restrict this treatment of death to the first 14 days rests on the following considerations: In the years from 1948–1954 over 90% of all pregnancy terminations occurred in the home rather than a hospital or clinic, and most mothers were reluctant to take their newly born child out until it was several weeks old. Of necessity, therefore, the initial examination of the infant occurred in the home, generally within the first 2 weeks following delivery. Thus the information available to us on mortality from this study can be presumed to be complete for only about the first 14 days of life. (Subsequent deaths became known through another program, the findings of which will be described elsewhere.) Although a variety of dose-response models have been fitted to the data, the results of only two will be described in detail. The first involves fitting a linear dose-response model to the occurrence of the various indicators of radiation-related damage treated as binary response variables. Data were available on six background or concomitant variables, in addition to summed parental dose, which might influence the occurrence of an event of interest: city (Hiroshima, Nagasaki; Hiroshima was assigned the value 1 and Nagasaki 0), sex (male, female; males were assigned the value 1, females 0), maternal age (years), paternal age (years), year of birth following the bombing, and birth rank. Specifically, we have fitted a model of the following form: where Pi is the expected frequency of the event of interest in the ith individual (i=1, 2, ... n; i.e., the total number of subjects) having background characteristics xij (j=1, 2,..., 6), and dose i. The constants, bj, and bD are the parameters to be estimated. The effects associated with the background characteristics are useful in providing standards of comparison with the risk seen with exposure. When these latter variables are treated as risk factors to be estimated rather than as part of the background, the model indicated above takes a slightly different form, indicated in Table V. The second dose-response model, employed in an earlier analysis of the data (2), involves fitting an exponential curve of the kind described as a “one-hit” model, of the following form: where Pi, xij, and Dosei are as previously defined. It is important to note in this connection that in the earlier analysis employing this model (2) 25 dose categories were used (five maternal and five paternal groupings), and that the background variables were treated somewhat differently. Also, the simple linear model described above is a first approximation to the exponential, and the results of fitting the two models would not be expected to differ greatly. We offer the results of the exponential model only for continuity with earlier analyses which have used this model. In both instances, the parameters of the model have been estimated by the method described elsewhere (2). RESULTS Analysis employing DS86 doses only. Of the 55,303 pregnancy terminations analyzed, 2760 culminated in an untoward pregnancy outcome (Table VI). Among these latter terminations were 770 infants with a major congenital defect, 894 who were stillborn, and 1230 who, although live-born, died before the fifteenth day after birth. These individually designated categories are overlapping, and 134 of the 2760 events occurred in conjunction with another. As shown in Table VI, which presents the relationship between parental dose category and untoward pregnancy outcome, the background or control rate was 4.99%. Table VII sets forth the results of the analysis of these data when the concomitant sources of variation which can be identified in these data are taken into account. We note no
OCR for page 383
THE CHILDREN OF ATOMIC BOMB SURVIVORS: A Genetic Study TABLE VII Increments or Decrements of Change in the Frequency of Untoward Pregnancy Outcomes in the Original Cohort of Births per Sievert of Joint Parental Gonadal Dose Equivalent Based upon an Assumed Neutron RBE of 20, DS86 Cohort Only Variable Regression coefficient Standard error Regression model: Pi=Constant+ (Background)+bD Dosei Background effects Constant 0.04282 0.00662 City 0.00110 0.00189 Sex 0.00262 0.00184 Mean age of father –0.00039* 0.00022 Mean age of mother 0.00034 0.00031 Birth order of child 0.00039 0.00075 Year of birth 0.00173** 0.00061 Excess risk Joint parental exposure 0.00422 0.00343 Cov(Constant, Dose)=-0.2980×10-9 Corr(Constant, Dose)=-0.0131 Regression model: Pi=1-exp-(Constant+ (Background)+bD Dosei) Background effects Constant 0.04324 0.00714 City 0.00110 0.00203 Sex 0.00262 0.00198 Mean age of father –0.00040* 0.00024 Mean age of mother 0.00034 0.00034 Birth order of child 0.00046 0.00080 Year of birth 0.00172** 0.00066 Excess risk Joint parental exposure 0.00412 0.00364 Cov(Constant, Dose)=–0.3515×10-9 Corr(Constant, Dose)=–0.0135 Note. Significance levels: *(P<0.10), **(P<0.01). significant effects of city, sex, or ages of mothers or fathers. A somewhat larger number of untoward pregnancies occurred in Hiroshima than in Nagasaki, and among the progeny of irradiated males. Adverse outcomes decreased slightly with paternal age and increased with maternal age. These findings are consistent with our previous studies of this cohort (1,2,12,23). The regression on parental radiation exposure is positive as expected on the hypothesis that radiation produces mutations but well below the level of statistical significance. The results obtained by fitting the one-hit model differ negligibly from those just described (see Table VII). In Table V we present separate analyses for malformations, stillbirths, and neonatal deaths using both statistical approaches, to determine whether there might be a stronger suggestion of a radiation effect upon some one of the three indicators “hidden” within the gross regression coefficient. In fact, within the statistical limits imposed by the data, all three indicators behave in a similar manner. The role of possible sources of extraneous variability is potentially troublesome. We have examined, albeit crudely, their effects on the estimate of the genetic risk by ignoring all of the concomitant variables and fitting a simple linear model to the data given in Table VII. We assigned as the dose in each of the 25 cells the sum of the individual mean parental doses given in the marginal entries (data not shown). A further purpose of this analysis was to determine the effect of grouping of doses on the estimates of intercept and slope. The estimates of the latter two parameters in the restricted sample (with an RBE of 20) obtained by the method of maximum likelihood are 0.0497 and 0.00354 (±0.00343), respectively. As can be seen from a comparison of these estimates with those given in Table VII, grouping and ignoring the extraneous sources of variation does not appear to have a profound effect on either estimate. The estimates change by about 16% when compared with the linear model including the background factors, suggesting that the extraneous sources of variation are more or less randomly distributed with respect to dose. Analysis employing DS86 and ad hoc doses. The use of ad hoc doses results in the addition of 14,403 to the sample. Of the 69,706 pregnancy terminations included in this extended sample, 3498 culminated in an untoward pregnancy outcome (Table VIII). Among these untoward outcomes were 950 infants with major congenital defect, 1148 who were stillborn, and 1565 liveborn infants who died within 14 days of birth. When abnormal terminations were scored in this manner, 165 of the 3613 events occurred in conjunction with one other, resulting in the 3498 abnormal infants. As shown in Table VIII, which presents the distribution of untoward pregnancy outcomes with respect to parental dose, the background rate was 5.02%. Table IX presents the results of the analysis of the extended cohort. All of the findings observed in the more restricted subset of data are repeated. Again the regression of untoward pregnancy outcome upon parental radiation exposure is positive, which would be expected if the exposure had induced deleterious dominant mutations, but the regression is well below statistical significance, and insignificantly lower than that obtained with the DS86 only cohort. Furthermore, none of the components of the untoward pregnancy outcomes gives any hint of responding differentially to parental radiation (Table X). There is no suggestion of significant heterogeneity between the two data sets. As
OCR for page 384
THE CHILDREN OF ATOMIC BOMB SURVIVORS: A Genetic Study TABLE VIII The Distribution of Untoward Pregnancy Outcomes (UPO), and Total Subjects by Parental Gonadal Dose Equivalent in Sieverts (DS86) Based on an Assumed Neutron RBE of 20, Sexes and Cities Combined, Extended (DS86+ad hoc Dose) Cohort Father's gonadal dose equivalent (Sv) Mother's gonadal dose equivalent (Sv) <0.01 0.01–0.09 0.10–0.49 0.50–0.99 1.0+ Subj UPO Subj UPO Subj UPO Subj UPO Subj UPO Mean dose 1.00+ 528 24 36 3 13 0 12 1 19 1 1.78 0.50–0.99 834 58 25 0 27 2 53 4 24 1 0.69 0.10–0.49 1984 95 151 10 235 15 53 2 30 1 0.23 0.01–0.09 4428 207 755 28 179 7 54 2 71 3 0.03 <0.01 57322 2880 1428 79 726 36 339 16 380 23 0.00 Mean dose 0.00 0.03 0.25 0.70 2.12 expected, the standard errors of the various regression terms are appropriately decreased in the larger data set. Comparison with previous analysis. Our previous analysis of these data with the one-hit model, employing the T65DR doses and a somewhat different computational technique, yielded a regression on summed parental doses per sievert of 0.001824±0.003232 (2). The regression most nearly comparable to this in the present analysis (one-hit model, full data set, RBE of 20) is 0.00264±0.00277. This increase in the regression under the current analysis reflects the reduction in the gonad dose estimated with the DS86 schedule and the use of slightly different statistical procedures. The reduction in the estimated gonadal dose is complex, depending, inter alia, on a major reduction in the neutron component attributed to the Hiroshima bomb which entailed a change in the neutron RBE from 5 to 20 and a decrease in the transmission of the radiation by Japanese-style buildings, but an increase in the estimate of tissue transmission of the radiation, especially for ? radiation. The exact reduction in gonadal dose varies with distance from the hypocenter, but for all parents receiving doses ?0.01 Sv it is about 30%. As noted, none of the models we have used reveals a statistically significant effect of combined parental exposure, although all show that the risk for an untoward outcome of pregnancy increases with increasing dose. Previous analyses have also failed to disclose a significant dose effect (1,2). We note once again, however, that in associating errors with these regressions we do not imply a test of the null hypothesis. Radiation has resulted in an increased frequency of mutation in every well-studied organism, and it is inconceivable that humans are an exception. We will discuss below the appropriate uses of these data. UNCERTAINTIES As in most if not all epidemiological studies of similar scope, a number of uncertainties attend this one and the analyses we have presented. Two of these warrant particular comment, namely, errors in the estimation of the organ absorbed doses, and the completeness of the ascertainment of death and defect. Errors in the Estimation of the Gonadal Absorbed Doses All estimates of the doses to survivors of the A-bombing are subject to at least three sources of error that stem from the FIA dose curves themselves, the estimation of the attenuation of energy through tissues, materials, and the like, and the assertions of the survivors as to their locations. There is no particular reason, however, to believe that these errors are any larger or smaller in the present study than in any other study based on the survivors of the atomic bombing of these cities, where it has been estimated that individual doses could be in error by as much as 30% (24). Be this as it may, these errors can affect inferences on the overall shape of the dose-response curve as well as parameter values defining that shape (24–26). Completeness of the Ascertainment of Death and Defect As we have previously stated, ascertainment of the pregnancies whose outcomes we report occurred at or shortly after the fifth lunar month of gestation, and generally well before the pregnancy terminated. Thus the design was inherently prospective, and from a variety of lines of evidence, including examination of the births officially recorded in these cities, we estimate that more than 95% of qualifying pregnancies were identified. Many of those pregnancies which were not registered did subsequently come to our attention through the attending physician or midwife. Most involved children either conceived out of wedlock or born to women who had applied for rations elsewhere or were unaware of the provisions of the rationing system. There is no evidence that pregnancies persisting for 20 weeks or more and terminating untowardly went unregis-
OCR for page 385
THE CHILDREN OF ATOMIC BOMB SURVIVORS: A Genetic Study TABLE IX Increments or Decrements of Change in the Frequency of Untoward Pregnancy Outcomes in the Original Cohort of Births per Sievert of Joint Parental Gonadal Dose Equivalent Based upon an Assumed Neutron RBE of 20, Extended (DS86+ad hoc) Cohort Variable Regression coefficient Standard error Regression model: Pi=Constant+ (Background)+bD Dosei Background effects Constant 0.03856 0.00582 City 0.00100 0.00167 Sex 0.00238 0.00165 Mean age of father –0.00023 0.00020 Mean age of mother 0.00034 0.00028 Birth order of child 0.00019 0.00066 Year of birth 0.00179** 0.00055 Excess risk Joint parental exposure 0.00264 0.00277 Cov(Constant, Dose)=-0.2827×10-9 Corr(Constant, Dose)=–0.0175 Regression model: Pi=1-exp-(Constant+ (Background) +bD Dosei) Background effects Constant 0.03868 0.00627 City 0.00101 0.00180 Sex 0.00237 0.00177 Mean age of father –0.00024 0.00022 Mean age of mother 0.00035 0.00030 Birth order of child 0.00022 0.00071 Year of birth 0.00179** 0.00059 Excess risk Joint parental exposure 0.00262 0.00294 Cov(Constant, Dose)=-0.3299×10-9 Corr (Constant, Dose)=-0.0179 Significance levels: **(P<0.01). tered more frequently if one or both parents were exposed to the bombing. DISCUSSION In principle, an increase in untoward outcomes in the pregnancies of women exposed to the atomic bombs could reflect an altered maternal physiology and/or depressed socioeconomic status and/or an increased mutation rate in the woman and/or her husband. The average child entering into this study was conceived about 5 years following the bombings, by which time any acute effects of the bombings had certainly disappeared. The intensive studies sponsored by RERF have revealed an increase in a variety of malignant tumors as a late sequela of the bombings [review in (27)], but aside from leukemia, these had not yet made their appearance at the time the present observations were made. We will therefore argue that an altered maternal physiology is unlikely to have affected the indicator, but would only inflate any observed effect. With respect to socioeconomic TABLE X Increments or Decrements of Change in the Individual Frequencies of Congenital Malformation, Stillbirths, and Neonatal Deaths in the Original Cohort of Births per Sievert of Joint Parental Gonadal Dose Equivalent Based Upon an Assumed Neutron RBE of 20, Extended (DS86+ad hoc) Cohort Variable Regression coefficient Standard error Regression model: Pi=Constant+ (Background)+bD Dosei Malformation Joint parental exposure 0.00101 0.00154 Birth order of child 0.00064* 0.00036 Year of birth 0.00131** 0.00028 Stillbirths Joint parental exposure 0.00092 0.00163 Birth order of child –0.00059 0.00038 Year of birth –0.00028 0.00032 Neonatal deaths Joint parental exposure 0.00128 0.00185 Birth order of child 0.00026 0.00043 Year of birth 0.00068* 0.00037 Regression model: Pi=1-exp-(Constant+ (Background)+bD Dosei) Malformation Joint parental exposure 0.00120 0.00194 Birth order of child 0.00064 0.00047 Year of birth 0.00130** 0.00038 Stillbirths Joint parental exposure 0.00091 0.00200 Birth order of child –0.00075 0.00048 Year of birth –0.00029 0.00040 Neonatal deaths Joint parental exposure 0.00128 0.00218 Birth order of child 0.00047 0.00052 Year of birth 0.00074* 0.00044 Significance levels: *(P<0.10), **(P<0.01).
OCR for page 386
THE CHILDREN OF ATOMIC BOMB SURVIVORS: A Genetic Study status, we have reported that the parents who were not exposed because they came to Hiroshima and Nagasaki following the bombings as released service men, repatriates, spouses, or immigrants were slightly younger and had a little more education and somewhat higher occupational ratings than the exposed parents (1–4). In principle, this might result in a higher indicator frequency in the children of exposed parents and also lead to an overestimate of radiation effects. It is important to note that the estimate of genetic damage following exposure to ionizing radiation derived here rests on mutational events manifesting themselves as an untoward pregnancy outcome recognizable between 20 weeks after fertilization and 14 days following birth. Clearly this does not represent all of the mutational damage that could be expressed throughout life or even all of that from fertilization to 14 days postpartum. Pregnancies terminating prior to the 20th week were not ascertained in this study, and to the extent that such occurrences were dose-related, reflecting mutational damage, our data would underestimate the risk. The magnitude of this possible underestimation is uncertain, indeed impossible to estimate, since pregnancies terminating within 4 weeks of fertilization, before a menses is missed or other symptoms of pregnancy become apparent, are often unrecognized by the prospective mother. Although the proportion of pregnancies terminating this early in gestation is not precisely known, it appears relatively large. However, from the societal standpoint, these early losses are much less traumatic than the untoward pregnancy outcomes of this study. Since radiation has caused genetic damage in every species properly studied in an experimental setting, we assume that some genetic damage resulted from the Hiroshima-Nagasaki experience. We must then also accept that the data from these children, despite their limitations, provide the best available basis for estimating the confidence limits to be placed on the computed genetic risk involved in exposure to ionizing radiation. Briefly, under a linear dose-response function, the excess relative risk for genetic effects can be defined as the ratio of the slope to the intercept. When concomitant sources of variability are taken into account (Tables VII and IX), this ratio at an RBE of 20 is 0.0986, based on the restricted sample. The 95% lower confidence limit on this doubling dose estimate is approximately 0.15 Sv. The comparable values for the extended sample are 0.0685, and 0.19 Sv, respectively. To utilize these estimates to derive a doubling dose applicable to the genetic component, however, it is necessary to postulate what fraction of the indicator may be attributed to spontaneous mutation in the preceding generation. We shall provide an estimate of this fraction based upon current information and combine the findings of this study with other genetic endpoints that have been measured in the children of survivors to derive estimates of both the minimal and most likely doubling dose in a future paper. 2 ACKNOWLEDGMENTS This research was conducted at the Radiation Effects Research Foundation, Hiroshima, Japan. The Radiation Effects Research Foundation (formerly ABCC) was established in April 1975 as a private non-profit-making Japanese foundation, supported equally by the Government of Japan through the Ministry of Health and Welfare, and the Government of the United States through the National Academy of Sciences under contract with the Department of Energy. RECEIVED: June 7, 1989; ACCEPTED: October 30, 1989 REFERENCES 1. J. V. NEEL and W. J. SCHULL, The Effect of Exposure to the AtomicBombs on Pregnancy Termination in Hiroshima and Nagasaki. Publ. No. 461, National Academy of Sciences-National Research Council, Washington, DC, 1956. 2. W. J. SCHULL, M. OTAKE, and J. V. NEEL, Genetic effects of the atomic bombs: A reappraisal. Science 213, 1220–1227 ( 1981). 3. J. V. NEEL, H. KATO, and W. J. SCHULL, Mortality in the children of atomic bomb survivors and controls. Genetics 76, 311–326 ( 1974). 4. H. KATO, W. J. SCHULL, and J. V. NEEL, A cohort-type study of survival in the children of parents exposed to atomic bombings. Am. J. Hum. Genet. 18, 339–373 ( 1966). 5. W. J. SCHULL and J. V. NEEL, Maternal radiation and mongolism. Letter to the Editor. Lancet 1, 537–538 ( 1962). 6. W. J. SCHULL, J. V. NEEL, M. OTAKE, A. AWA, C. SATOH, and H. B. HAMILTON, Hiroshima and Nagasaki: Three and a half decades of genetic screening . In Environmental Mutagens and Carcinogens (T. Sugimura, S. Kondo, and S. Takebe, Eds. ), pp. 687–700. University of Tokyo Press, Tokyo, 1982. 7. A. A. AWA, A. D. BLOOM, M. C. YOSHIDA, S. NERIISHI, and P. ARCHER, A cytogenetic survey of the offspring of atomic bomb survivors. Nature 218, 367–368 ( 1968). 8. A. A. AWA, T. HONDA, S. NERIISHI, T. SOFUNI, H. SHIMBA, K. OHTAKI, M. NAKANO, Y. KODAMA, M. ITO, and H. B. HAMILTON, Cytogenetic studies of the offspring of atomic bomb survivors. In Cytogenetics: Basic and Applied Aspects (B. Obe and A. Basler, Eds. ), pp. 166–183. Springer-Verlag, Berlin, 1987. 9. C. SATOH, A. A. AWA, J. V. NEEL, W. J. SCHULL, H. KATO, H. B. HAMILTON, M. OTAKE, and K. GORIKI, Genetic effects of atomic bombs. In Human Genetics Part A: The Unfolding Genome (B. Bonne-Tamir, Ed. ), pp. 267–276. A. R. Liss, New York, 1982. 10. J. V. NEEL, C. SATOH, K. GORIK, J. ASAKAWA, M. FUJITA, N. TAKAHASHI, T. KAGEOKA, and R. HAZAMA, Search for mutations altering protein charge and/or function in children of atomic bomb survivors: Final report. Am. J. Hum. Genet. 42, 663–676 ( 1988). 11. J. V. NEEL, A study of major congenital defects in Japanese infants. Am. J. Hum. Genet. 10, 398–445 ( 1958). 12. D. GRAHN, C. H. LEE, and B. F. FARRINGTON, Interpretation of cytogenetic damage induced in the germ line of male mice exposed for over 1 year to 239Pu alpha particles, fission neutrons, or 60Co gamma rays. Radiat. Res. 95, 566–583, ( 1983). [See also D. GRAHN, B. A. FARRINGTON, and C. H. LEE, Genetic injury in hybrid male mice exposed to low doses of 60Co gamma-rays of fission neutrons. I. Response to single doses. Mutat. Res. 129, 215–229 ( 1984). ]
OCR for page 387
THE CHILDREN OF ATOMIC BOMB SURVIVORS: A Genetic Study 13 . J. A. AUXIER, Ichiban—Radiation Dosimetry for the Survivors of theBombings of Hiroshima and Nagasaki . REDA Critical Review Series, TID-27080, NTIS 1977. [See also J. T. CHEKA, F. W. SANDERS, et al . Distribution of weapons radiation in Japanese residential structures . USAEC Report CEX-62, 11, 1965. ] 14 . R. C. MILTON and T. SHOHOJI, Tentative 1965 Radiation Dose Estimation for Atomic Bomb Survivors, Hiroshima and Nagasaki . Technical Report 1–68, Atomic Bomb Casualty Commission, Hiroshima, 1968. 15. G. D. KERR, Organ dose estimates for Japanese atomic bomb survivors. Health Phys. 37, 487–508 ( 1979). 16 . P. P. WHALEN, Status of Los Alamos efforts related to Hiroshima and Nagasaki dose estimates. In Reevaluation of Dosimetric Factors:Hiroshima and Nagasaki (V. P. Bond and J. W. Thiessen, Eds. ). Technical Information Center, U. S. Dept. of Commerce, Springfield, VA, 1982. 17 . W. E. LOEWE and E. MENDELSOHN, Revised dose estimates at Hiroshima and Nagasaki. Health Phys. 41, 663–665 ( 1981). [See also W. E. LOEWE and E. MENDELSOHN, Neutron and gamma doses at Hiroshima and Nagasaki. Nucl. Sci. Eng. 81, ( 1982). ] 18 . G. D. KERR, Findings of a recent Oak Ridge National Laboratory review of dosimetry for Japanese Atomic Bomb survivors. In Reevaluation of Dosimetric Factors: Hiroshima and Nagasaki (V. P. Bond and J. W. Thiessen, Eds. ). Technical Information Center, U. S. Dept. of Commerce, Springfield, VA, 1982. 19 . G. D. KERR, T. HASHIZUME, and C. W. EDINGTON, Historical Review. In Final Report of US-Japan Reassessment of Atomic BombDosimetry in Hiroshima and Nagasaki (W. C. Roesch, Ed. ), pp. 1– 13. Radiation Effects Research Foundation, Hiroshima, 1987. 20 . W. C. ROESCH, Ed. , Final Report of US-Japan Reassessment ofAtomic Bomb Radiation Dosimetry in Hiroshima and Nagasaki . Radiation Effects Research Foundation, Hiroshima, 1987. 21 . D. L. PRESTON and D. A. PIERCE, The Effect of Changes in Dosimetry on Cancer Mortality Risks in the Atomic Bomb Survivors . Technical Report 9–87, Radiation Effects Research Foundation, Hiroshima, 1987 [See also Radiat. Res. 114,437–466 ( 1988). ] 22 . W. J. SCHULL and J. V. NEEL, Atomic bomb exposure and the pregnancies of biologically related parents. Am. J. Publ. Health 49, 1621– 1629 ( 1959). 23 . W. J. SCHULL, Empirical risks in consanguineous marriages: Sex ratio, malformation and viability. Am. J. Hum. Genet. 10, 294–343( 1958). 24 . S. JABLON, Atomic Bomb Radiation Dose Estimation at ABCC . Technical Report 23–71, Atomic Bomb Casualty Commission, Hiroshima, 1971. 25 . E. S. GILBERT, Some Effects of Random Dose Measurement Errorson the Analysis of Atomic Bomb Survivor Data . Technical Report 12– 82, Radiation Effects Research Foundation, Hiroshima, 1982. 26 . E. S. GILBERT and J. L. OHARA, Analysis of atomic bomb radiation dose estimation at RERF using data on acute radiation symptoms. Radiat. Res. 100, 124–138 ( 1984). 27 . I. SHIGEMATSU and A. KAGAN, Cancer in Atomic Bomb Survivors,pp. 1–196. Plenum Press, New York/London, 1986.
OCR for page 388
THE CHILDREN OF ATOMIC BOMB SURVIVORS: A Genetic Study This page in the original is blank.
Representative terms from entire chapter: