The Use of Short-Term Genotoxicity Tests in Risk Assessment

Abstract

The discipline of risk assessment is currently under pressure from two directions. On the one hand, more and more decisions must be made concerning how resources are to be allocated to reduce human exposure to harmful chemicals, and the assessment of risk from those exposures is often essential in making such decisions. On the other hand, many have emphasized the gaps in scientific knowledge and the reliance on assumptions in risk assessments, and the resulting regulations have been criticized as too lenient by some (Epstein and Swartz 1988; Karstadt 1988; Perera 1988; Bailar et al. 1988) and as too stringent by others (Wilson 1988; Freedman and Zeisel 1988). Risk assessments will always have some component of uncertainty, because we often care about risk at levels well below what is detectable [for example, many environmental regulations strive for a lifetime increased cancer risk of no more than 10“5or 10”6 (Anderson 1988), but the lowest detectable risk in humans is about 10^-3, the risk of diethylstilbestrol exposure, which causes a very rare form of tumor (IARC 1987)], and so some extrapolation will always be necessary. Also, because generating information requires the expenditure of money, the ideal risk assessment with perfect information can be approached only for a very limited number of chemicals. Thus, progress in risk assessment will come not only from identifying areas of uncertainty where more research is needed, but also from developing rational procedures for making decisions in the presence of uncertainty.