Linear interpolation algorithm for low dose risk assessment of toxic substances.

Abstract

In order to detect potential toxic effects of substances, relatively high doses generally are administered to relatively small numbers of laboratory animals. It is impossible to estimate low levels of disease incidence with precision at low environmental dose levels even with large numbers of laboratory animals. However, upper limits on risk can be obtained for convex dose response curves by linear interpolation between the lowest experimental dose level and zero. A simple mathematical algorithm is provided for low dose risk assessment from dose response data and the performance of this procedure is evaluated for a variety of toxicological data, including but not limited to carcinogenesis. The low dose confidence limits resulting from linear interpolation are similar to those obtained from the Armitage-Doll multistage model.