Models and assumptions underlying genetic risk assessment.

Various methods employed for estimating the genetic risks of radiation are reviewed. With the doubling-dose method, genetic damage is expressed as an increase in cases of known genetic disease. The actual doubling dose is based on figures obtained with the mouse. There have been no recent data on induced mutation frequencies. Recent results suggest that the prevalence figure for multifactorial disease may be at least one order of magnitude higher than before. Various assumptions underlying the doubling-dose concept are discussed in the light of recent findings on: (1) spontaneous mutations resulting from insertion elements, and (2) the comparability between spontaneous and induced mutations. The so-called direct method makes use of figures for induction of dominant mutations affecting the skeleton and the lens of the eye in the mouse, and of translocation induction in monkeys. Induction rates are converted to overall rates of induced dominant effects in man by applying certain assumptions. The proportionality between dose and effect is the basis for all genetic risk assessments. The possible significance of data on human lymphocytes indicating a threshold below 4 rad and the induction of repair enzymes by low radiation doses is discussed. The parallelogram approach is based on the principle that estimates can be obtained on the amount of genetic damage that cannot always be assessed directly. Thus mutations in mouse germ cells can be predicted by using mutation frequencies in cultured mammalian cells and O6-ethylguanine adducts. Measurement of haemoglobin mutations in human and mouse erythrocytes, and of HPRT-deficient mutations in lymphocytes of man and mouse should make more precise estimates of mutation frequencies in human germ cells possible. The development of a database on mutations in somatic cells of the mouse, their induction frequencies and molecular nature are considered an important priority. Used in combination with mouse germ-cell mutation frequencies, they should enable more precise risk estimates on the basis of mutations in somatic cells of man.