Long-term chemical carcinogenesis bioassays predict human cancer hazards. Issues, controversies, and uncertainties.

Long-term carcinogenesis bioassays are the most valued and predictive means for identifying potential carcinogenic hazards of various agents to humans. Agents may be chemicals, chemical mixtures, multiple chemicals, combinations of chemicals, residues and contaminants, commercial products and formulations, and various exposure circumstances. Life-styles, dietary factors, and occupational exposure circumstances are very difficult, but not totally impossible, to evaluate experimentally. Historically, the first chemical bioassay took place in the early part of this century: Yamagiwa and Ichikawa in 1915, showed that coal tar applied experimentally to rabbit ears caused skin carcinomas. Since then, nearly 1500-2000 bioassays of one sort or another have been carried out. Importantly, however, some of these bioassays must be considered inadequate for judging the absence of carcinogenicity, since there were various limitations on the way they were performed: too few animals, too short a duration, too low exposure concentrations, too limited pathology, as examples. Thus, each bioassay must be critically evaluated, especially those reported to be negative, because "false negatives" are certainly more hazardous to human health than are "false positives". Likewise, one must be careful not to discount bioassay results simply because a target organ in rodents may not have a direct counterpart in humans (e.g., Zymbal glands), or because an organ site in rodents may not be a major site of cancers in humans (e.g., mouse liver). The design and conduct of a bioassay is not simple, however, and one must be fully aware of possible pitfalls as well as viable and often necessary alternatives. Similarly, evaluating results and interpreting findings must be approached with the utmost objectivity and consistency. These and other select issues, controversies, and uncertainties possibly encountered in long-term bioassays are covered in this paper. One fact remains abundantly clear: for every known human carcinogen that has been tested adequately in laboratory animals, the findings of carcinogenicity are concordant.