QUICK ESTIMATE OF THE REGULATORY VIRTUALLY SAFE DOSE BASED ON THE MAXIMUM TOLERATED DOSE FOR RODENT BIOASSAYS

With a limited subset of National Cancer Institute/National Toxicology Program (NCI/NTP) bioassays, Gaylor (Regul. Toxicol. Pharmacol. 9, 10 1-108, 1989) showed that the regulatory virtually safe dose (VSD), cor responding to an estimated lifetime cancer risk of less than 10(-6), c ould be estimated within a factor of 10 simply by dividing the maximum tolerated dose (MTD), estimated from the results of a 90-day study, b y 380,000. The purpose of this current study was to extend the analysi s to all carcinogens in the Carcinogenic Potency Database (CPDB) utili zing the TD50 (average daily dose rate in mg/kg body wt/day that was e stimated to halve the probability of remaining tumor-free at a specifi ed tissue site throughout a a-year study). Using the relationship betw een the upper bound on the low-dose slope (q(1)) and the TD50 reporte d by Krewski et al. (Risk. Anal. 13, 383-398, 1993) and the ratio of t he maximum dose tested (Max-D)/TD50 obtained in our present analysis, an estimate of the regulatory VSD was given by the MTD/740,000, for NC I/NTP rodent carcinogens. This was about a factor of two lower than th e limited analysis conducted by Gaylor. There was little difference wh en the chemicals were divided into mutagens and nonmutagens. Ninety-si x percent (134 of the 139 NCI/NTP rodent carcinogens) of the regulator y VSDs calculated from the individual TD(50)s obtained from the 2-year bioassays were within a factor of 10 of the MTD/740,000. Gold et al. (Environ. Health Perspect. 79, 259-272, 1989) investigated the distrib ution of the TD50 among ''near-replicate'' experiments (where the same chemical was tested more than once and was positive in the same strai n, sex, and species by the same route). The distribution of TD(50)s fr om near-replicate experiments is similar to the distribution of the Ma x-D/TD50. Hence, the estimate of the regulatory VSD based on the Max-D /740,000, for NCI/NTP bioassays, is about as precise as the estimate o btained from a 2-year bioassay. This questions the advisability of con ducting a 2-year bioassay for purposes of regulatory risk estimation. Since resources are available to test only a small fraction of chemica ls to which humans are exposed, a preliminary estimate of the regulato ry VSD can be useful in setting testing priorities. If the expected hu man exposure level is below the regulatory VSD estimated from the MTD, a chemical may be assigned low priority for testing in a 2-year bioas say. Based on previous work the FDA proposed a ''threshold of regulati on'' of a dietary concentration of 0.5 ppb for all substances used in food-contact articles. The results of the present analysis could be us ed to make the procedure more chemical specific based upon an availabl e MTD. The high correlation between the MTD and estimate of cancer pot ency (TD50) can be exploited to provide a preliminary, hypothetical up per bound estimate of cancer risk for exposure to a chemical without c onducting a 2-year animal bioassay. Thus, the expected level of human exposure relative to the MTD can be used to determine the priority for further research on a chemical, such as mechanistic studies. (C) 1995 Academic Press, Inc.