Quantitative cross-species extrapolation in noncancer risk assessment

A procedure has been developed to extrapolate dose-effect findings across species and quantify some of the associated uncertainty via traditional statistical methods. The relationship between dose-effect curves which are known in two species can be described by a k-parameter dose equivalence equation (DEE). When a DEE is determined for as many agents of a given "family" as are. known, a h-dimensional distribution of DEE parameters would obtain. When an actual extrapolation is to be made for a new agent of the same family which has not been (or cannot be) tested in the species to which results are to be extrapolated, the best estimate of the parameters of the new DEE would be some measure of central tendency and the best estimate of the uncertainty would be the variance/covariance of the h-dimensional distribution of DEE parameters. The method of extrapolation seeks to compliment the mechanistic knowledge or be a substitute for mechanistic methods in the large majority of cases where such understanding is lacking.