Ionizing radiation and genetic risks IV. Current methods, estimates of risk of Mendelian disease, human data and lessons from biochemical and molecular studies of mutations

Abstract

This paper is aimed at a synthesis of conclusions and concepts from the first three papers of this series and an inquiry of their relevance to the estimation of the risk of autosomal dominant and X-linked diseases in man, due to exposure to ionizing radiation. For a population under conditions of continuous irradiation, the doubling-dose method (DD method) enables the prediction of the excess risk of dominant and X-linked diseases at equilibrium. Per unit dose, this quantity is the product of the natural prevalence of these diseases (assumed to be 10,000/10 6 livebirths) and the reciprocal of the DD. The DD currently used is 1 Gy and is based primarily on data on the induction of recessive specific-locus mutations in male mice. The estimate of risk to the first generation is derived from that at equilibrium; the figure is about 15% of the equilibrium value (i.e., 15 cases/10 6 livebirths/cGy). with the direct method, the first-generation risk of dominant disease is estimated using data on the induction of dominant skeletal and cataract mutations in male mice and a number of correction factors. The estimates are about 10–20 cases and 0–9 cases, respectively, for irradiation of males and females, per 10 6 livebirths/cGy. In the Japanese studies, no significant adverse genetic effects, attributable to exposure of the parents to the atomic bombs, could be demonstrated with respect to any of the endpoints used. Most of the latter are clinically and socially relevant but mutationally insensitive. On the basis of these data, Neel and colleagues have estimated that the gametic DD for genetic effects of radiation in man is at least about 4–5 times the 1 Gy value thus far used. The concepts, assumptions, and the data-base used with the DD method have been re-examined. Arguments are advanced to support the thesis that ionizing radiation is probably not very efficient in inducing the very specific molecular changes that are known to underlie spontaneous mutations which cause naturally occurring dominant genetic diseases. It is suggested that (i) the DD estimate of 1 Gy that is used to estimate risk for autosomal dominant and X-linked diseases is conservative and (ii) the 1% prevalence figure for these diseases that is used for this purpose may be too high. If these suggestions are correct, then the estimate of risk for the dominant and X-linked diseases may need to be revised downwards. The question of the total risk of genetic disease (i.e., an estimate which includes the very large class of multifactorial diseases) is not addressed here and is not answerable at present using the DD method, one important reason being that the mutation-responsive component of these diseases cannot be reliably estimated. The choice of an overall DD used for genetic risk estimation depends on what indicators are perceived to be relevant in the human context and is largely judgemental. Since recessive predominate among radiation-induced mutations, adverse genetic effects of radiation exposure are primarily those associated with induced recessive mutations in the heterozygous condition. Some suggestions are made for further work to improve genetic risk estimates.