WORK PLANS - Laserfiche WebLink

Home Browse Search

aw epidemiological evidence. Use of animal data for human health assessment may under- or overestimate potential human health risk. In order to adjust for this uncertainty, EPA generally applies safety factors ranging from 10 to 1,000 for the results of studies in sensitive animal populations to develop toxicity values to be used in human health evaluations. This conservative approach is likely to overestimate the potential human health risks. There is uncertainty whether animal carcinogens are also carcinogenic in humans. While many chemical substances are carcinogenic in one or more animal species, only a small number of chemical substances are known to be human carcinogens. The fact that some chemicals are carcinogenic in some animals but not in others raises the possibility that not all animal carcinogens are human carcinogens. EPA assumes that humans are as �.., sensitive to all carcinogens as the most sensitive animal test species. This policy decision is designed to prevent underestimating risk, increasing the likelihood that the potential carcinogenic risk will be overestimated. Dose Extrapolation Toxicity testing is generally performed using high doses over a relatively short period of time, while environmental exposures are usually low doses over a relatively long period of time. Extrapolating from high doses to low doses has considerable uncertainty. For noncarcinogenic chemicals, little or no toxic effect may occur below a threshold value, due to the body's natural detoxifying and compensatory mechanisms. However, because of uncertainty associated with dose extrapolation, EPA often bases the RfDs for noncarcinogenic effects on the most sensitive animal species at the lowest dose. This risk assessment has incorporated EPA RfD values which were derived in this manner. The use of these conservatively derived RfD values is likely to overestimate the potential health impact of noncarcinogenic chemicals at the site. The model used by EPA to determine slope factors for carcinogens is a linearized multistage model, which provides a conservative estimate of cancer risk at low doses and `-� is likely to overestimate the actual slope factor. EPA acknowledges that the actual slope factor is likely to be between zero and this estimate. Inadequate knowledge of the validity and accuracy of this model, however, increases uncertainty and the tendency to overestimate cancer risks. S:\LDC\YELLO.RPr May 4, 1995 7-6 i.+