A methodology for assessing mutagenic hazards of chemicals

A comprehensive framework for identifying substances which represent a potential threat to public health due to mutagenicity and for relative ranking of their hazards has been developed. The methodology is designed to evaluate a range of genotoxic endpoints of potential significance to humans and is not merely a substitute for, or an adjunct to, the carcinogenicity assessment. A range of endpoints in both somatic and germ cells is considered. The biological test systems utilized here include humans and other mammals, bacteria, Drosophila, yeasts, molds, and plants. Bioassays conducted in vivo and in vitro, with and without metabolic activation, are included. Seventy-five different assays are considered, of which 73 are currently included in the Gene-Tox database. The tests are grouped into three categories on the basis of significance to humans, as well as sensitivity, specificity, validity, and reliability. Seven in vivo mammalian tests comprise Group I. The remaining 68 tests are divided between Groups II and III, in a decreasing order of significance and confidence in the tests. Tests with virtually identical endpoints and organisms are listed together. The system accomplishes two tasks: 1) organizes the data from short-term tests on chemicals, and 2) classifies chemicals into one of five hazard categories, designated by letters A to E, on the basis of that information. Classification into hazard categories depends on the overall strength of evidence that an agent may cause mutations in humans. Generally, score for each chemical is a function of the number and combination of results in each of the three groups of tests. Specifically, it is derived by weighing several variables: the number and type of endpoints measured, the number and type of species represented, the significance of positive and non-positive results, the relevance of specific tests for predicting effects in humans, the group classification (I, II or III) of each test result and the overall pattern presented. Assessment of data for one hundred chemicals shows a good representation of scores from A to E, with category E most commonly represented, followed by D and C.