Choice of Dose Measure for Extrapolating Carcinogenic Risk from Animals to Humans

Publisher Website
Google Scholar
Abstract
When estimating carcinogenic risk in humans from animal data, animal to human extrapolation is often accomplished by assuming that humans and animals are equally sensitive when dose is measured appropriately in common units in both species. This paper reports on an empirical investigation of several dose units, based on data from 23 chemicals for which both animal and human data were available. The dose units were evaluated using a measure called “bias,” which represents roughly the factor by which the TD25 calculated from animal data must be multiplied to correspond, on average, to the TD25 calculated from human data. (A TD25 is the average human dose rate per unit of body weight in mg kg-1 d-1, administered between ages 20 and 65, that would cause an extra lifetime cancer risk of 25%.) Biases of greater than one mean that estimates of human risk from animal data exceeded estimates made from human data. Five dose measures are evaluated in this paper; they are based on several risk assessment methods. The resulting ranges of biases are: dose rate per surface area in mg m−2 d−1, 2.1 to 12; dose rate per unit of body weight in mg kg−1 d−1, 0.36 to 1.6; parts per million in diet, 1.1 to 6.2; cumulative dose per unit of body weight in mg kg−1 per lifetime, 13 to 84; parts per million in air, 1.1 to 4.5. These findings suggest that all of these measures of dose except dose rate per unit of body weight tend to result in overestimation of human risk.

This publication has 0 references indexed in Scilit: