Gaylor D W, Kodell R L
J Environ Pathol Toxicol. 1980 Nov;4(5-6):305-12.
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.
为了检测物质的潜在毒性作用,通常会给相对少量的实验动物施用相对高的剂量。即使使用大量实验动物,也不可能在低环境剂量水平下精确估计低发病率。然而,对于凸形剂量反应曲线,可以通过在最低实验剂量水平和零之间进行线性插值来获得风险上限。本文提供了一种从剂量反应数据进行低剂量风险评估的简单数学算法,并针对包括但不限于致癌作用的各种毒理学数据评估了该程序的性能。线性插值得出的低剂量置信限与从阿米蒂奇 - 多尔多阶段模型获得的置信限相似。