Estimating "safe" levels, a hazardous undertaking

Abstract

Various problems beset the question of identifying chemical carcinogens in the environment or setting permissible levels for potential carcinogens. Issues arising are cost-benefit questions, existence of thresholds, appropriate experimental designs, how to extrapolate to man, results from tests on laboratory animals, etc. Certain approaches implicitly involve use of a double standard, with much more stringent measures taken when clearer evidence of carcinogenicity is found. Such double standards may discourage careful testing of carcinogens as this could more probably lead to imposition of the stricter measure. Even-handed application of devices like that recommended by Mantel and Bryan for setting "safe" levels could avoid this difficulty and would encourage more adequate testing. Why laboratory testing should be at high or moderately high levels is explained and the futility of "mega-mouse" experiments at very low dose levels is indicated. A surface-area rule for extrapolating dose levels from laboratory animal to man is suggested, but this is indicated to lead approximately to direct equivalence when dose levels are expressed as dietary concentrations.