Thresholds of carcinogenicity in the ED01 study.

The results of the articles on the carcinogenicity of 2-acetylaminofluorene (J. H. Farmer et al., 1980, J. Environ. Pathol. Toxicol. 3, 55-68; N. A. Littlefield et al., 1980, J. Environ. Pathol. Toxicol. 3, 17-34) in approximately 25,000 female mice were reanalyzed by the procedure proposed earlier (W. J. Waddell, 2002, TOXICOL: Sci. 68, 275-279) using the Rozman scale (K. K. Rozman et al., 1996, Drug Metab. Rev. 28, 29-52). In contrast to some conclusions of the lack of a threshold for carcinogenesis that have been made in the past from this study, this reanalysis showed a clear and consistent threshold for bladder neoplasms at about 10 19.5 molecules/kg/day and for liver neoplasms at about 10 19.1 molecules/kg/day. The slopes of the dose-response curves for bladder neoplasms from 17 months through 33 months were consistently very steep, while those for liver neoplasms increased from a shallow slope at 18 months to a steep slope at 33 months. This is interpreted to indicate that the mechanism of carcinogenesis may be different in the two organs. A linear response for percentage tumors plotted against dose on a logarithmic scale is confirmed by this analysis, which is based on the fundamental principle that chemical potential effects a linear response. Furthermore, this application continues to show a sharp threshold for carcinogenesis. The implications of these observations should be important in the extrapolation of results from animal experiments to human risk assessment.