Preliminary estimates of the virtually safe dose for tumors obtained from the maximum tolerated dose.

The purpose of this paper was to examine the correlation between the maximum tolerated dose (MTD) and the low-dose estimate of the virtually safe dose (VSD) for animal carcinogens. Chronic bioassay results from the National Cancer Institute/National Toxicology Program carcinogenesis screening program were used. Estimates of the VSD were obtained by linear low-dose extrapolation for which an adequate dose-response relationship existed at the same tumor site in the same sex for both rats and mice. Estimates of the VSD were compared with the MTD for 69 tumor sites from 38 chemicals for rats and mice. The MTDs ranged from high to low toxicity (1 ppb to 4.4% in the diet). The overall geometric mean of the ratio of the MTD to the VSD corresponding to a maximum estimated risk of 10(-6) was 3.8 x 10(5). Of the 138 cases, only 3 cases were more than a factor of 10 from the mean ratio. This suggested that a quick estimate of the VSD could be obtained by dividing the MTD, obtained from a subchronic study, by 400,000. Further, if the human exposure is less than 10(-7) X MTD, the estimated risk is likely to be negligible even if the chemical is a carcinogen. It may not be worthwhile to conduct a chronic bioassay for the purpose of demonstrating a negligible risk, if the chemical is likely to be carcinogenic, unless the human exposure is extremely low.(ABSTRACT TRUNCATED AT 250 WORDS)