The prospects for a simplified and internationally harmonized approach to the detection of possible human carcinogens and mutagens.

It is proposed that the many sets of Regulatory Guidelines for the assessment of chemical carcinogenicity and mutagenicity should be simplified and harmonized in light of current experimental data. Data are discussed which illustrate that an absolute distinction would be drawn between assays conducted in vitro from those in vivo, and that the genotoxicity of a chemical can be adequately defined using a combination of the Salmonella mutation assay and one for the assessment of chromosome aberrations in vitro. It is specifically recommended that once a chemical has shown a clear positive response in vitro, further short-term assays should be conducted in vivo; this avoids considering the 'weight of evidence' of in vitro data, the dangers of which are illustrated. It has now been unequivocally established that not all in vitro genotoxins prove carcinogenic to mammals. It is therefore recommended that all new in vitro genotoxins should be assessed in vivo using the mouse bone marrow micronucleus assay, and if a negative response is observed, a liver genotoxicity test. At present an assay for the induction of unscheduled DNA synthesis (UDS) in the liver is the most well developed for this purpose. Current data indicate that an in vitro genotoxin found to be inactive in these two in vivo assays will be neither carcinogenic nor mutagenic to the germ cells of mammals. Equally, genotoxicity produced in mammals indicates a carcinogenic and mutagenic potential which can usually only be countered by appropriate chronic bioassays. The use of short-term in vivo assays in this critical role requires attention to the selection of appropriate dose-levels and routes of exposure - these issues are discussed. The above testing strategy will not detect certain animal carcinogens, some of which are specifically discussed. These carcinogens have been variously referred to in the literature as epigenetic/non-genotoxic/hormonal/toxic/ambiguous or ambivalent carcinogens. It is suggested that they present a minor potential hazard to man when compared with that of genotoxic carcinogens and that their short-term detection can only be achieved by the development of new whole mammal assays employing non-genetic endpoints. This is in contrast to the present tendency to employ additional genotoxicity assays for their detection in the unjustified belief that they possess an exquisite specificity of genotoxic action. This article represents a personal view, but the testing strategy proposed is based to a large extent on the original three-tier approach of Bridges.(ABSTRACT TRUNCATED AT 400 WORDS)