Technical Report No. 16-87
Severe mental retardation among the prenatally exposed survivors of the atomic bombing of Hiroshima and Nagasaki: A comparison of T65DR and DS86 dosimetry systems
Otake M, Yoshimaru H, Schull WJEditor’s note: Publications based on this report were published in Congenital Anomalies 29:309-20, 1989, and J Radiat Res (Tokyo) 32S:249-64, 1991.
Summary
In March 1986, as a result of a comprehensive reevaluation of the exposure of the survivors of the A-bombing of Hiroshima and Nagasaki, a new method for the estimation of individual doses was introduced, termed the Dosimetry System 1986 (DS86). Important differences obtain between the estimation of organ-absorbed doses in this system and the one previously employed at RERF, the T65DR. The fetal absorbed doses associated with the latter are merely the estimates of maternal shielded kerma multiplied by average correction factors; whereas in the new system they are computed individually without the use generally of explicit, average correction factors and thus allow better for the scattering of radiant energy that occurs within tissues. Actual fetal absorbed doses, as such, are not yet available, and therefore, the comparisons described here rest on the computed dose to the mother’s uterus. The DS86 sample itself consists of 1,544 individuals (96.6%) of the 1,598 belonging to the clinical sample on whom T65DR doses are available, including all of the 30 individuals diagnosed to be severely mentally retarded. A variety of models with and without a threshold have been fitted to the individual as well as grouped dose data to ascertain the most suitable dose-response relationship.
Briefly the findings of this comparison are as follows: The risk of severe mental retardation due to radiation exposure changes little from one dosimetric system to the other. The highest risk of radiation damage to the embryonic and fetal brain occurs 8-15 weeks after fertilization under both the T65DR and DS86 systems. Although other dose-response models will fit the data, damage to the 8-15 week old fetus expressed as the frequency of severe mental retardation appears adequately described by a simple linear model without a threshold. The risk at 1 Gy is about 46% with the T65DR system and 43% with the DS86 under a simple linear model, and 56% and 48%, for the T65DR and DS86 dosimetries, respectively, under an exponential linear model.
Somewhat more evidence exists under the DS86 system of a threshold to the dose-response relationship in the 8-15 week interval than existed with the T65DR doses. However, the location and reality of the threshold are difficult to assess. Without exclusion of five cases of probable nonradiation-related etiologies, and using a linear model, the grouped dose data suggest a threshold at about 0.20 Gy; whereas the individual dose data yield a threshold in the neighborhood of 0.40 Gy. The estimate of the threshold varies substantially with the model fitted and whether the five cases of mental retardation with probable nonradiation-related etiologies are or are not included.
Damage to the fetus 16-25 weeks after fertilization seems linear-quadratically or quadratically related to dose, especially in the DS86 sample, and suggests a threshold in the neighborhood of 0.70 Gy (DS86 dose), under a linear model using the individual dose data, with a lower 95% confidence bound of 0.21 Gy. Grouped dose data give the same lower bound, but an estimate of the threshold of 0.64 Gy.
In March 1986, as a result of a comprehensive reevaluation of the exposure of the survivors of the A-bombing of Hiroshima and Nagasaki, a new method for the estimation of individual doses was introduced, termed the Dosimetry System 1986 (DS86). Important differences obtain between the estimation of organ-absorbed doses in this system and the one previously employed at RERF, the T65DR. The fetal absorbed doses associated with the latter are merely the estimates of maternal shielded kerma multiplied by average correction factors; whereas in the new system they are computed individually without the use generally of explicit, average correction factors and thus allow better for the scattering of radiant energy that occurs within tissues. Actual fetal absorbed doses, as such, are not yet available, and therefore, the comparisons described here rest on the computed dose to the mother’s uterus. The DS86 sample itself consists of 1,544 individuals (96.6%) of the 1,598 belonging to the clinical sample on whom T65DR doses are available, including all of the 30 individuals diagnosed to be severely mentally retarded. A variety of models with and without a threshold have been fitted to the individual as well as grouped dose data to ascertain the most suitable dose-response relationship.
Briefly the findings of this comparison are as follows: The risk of severe mental retardation due to radiation exposure changes little from one dosimetric system to the other. The highest risk of radiation damage to the embryonic and fetal brain occurs 8-15 weeks after fertilization under both the T65DR and DS86 systems. Although other dose-response models will fit the data, damage to the 8-15 week old fetus expressed as the frequency of severe mental retardation appears adequately described by a simple linear model without a threshold. The risk at 1 Gy is about 46% with the T65DR system and 43% with the DS86 under a simple linear model, and 56% and 48%, for the T65DR and DS86 dosimetries, respectively, under an exponential linear model.
Somewhat more evidence exists under the DS86 system of a threshold to the dose-response relationship in the 8-15 week interval than existed with the T65DR doses. However, the location and reality of the threshold are difficult to assess. Without exclusion of five cases of probable nonradiation-related etiologies, and using a linear model, the grouped dose data suggest a threshold at about 0.20 Gy; whereas the individual dose data yield a threshold in the neighborhood of 0.40 Gy. The estimate of the threshold varies substantially with the model fitted and whether the five cases of mental retardation with probable nonradiation-related etiologies are or are not included.
Damage to the fetus 16-25 weeks after fertilization seems linear-quadratically or quadratically related to dose, especially in the DS86 sample, and suggests a threshold in the neighborhood of 0.70 Gy (DS86 dose), under a linear model using the individual dose data, with a lower 95% confidence bound of 0.21 Gy. Grouped dose data give the same lower bound, but an estimate of the threshold of 0.64 Gy.