Reproducibility of low-dose extrapolation procedure: comparison of estimates obtained using different rodent species...

Carcinogenic risk assessment based on low-dose extrapolation of dose-response relationships is characterized by a significant level of uncertainty. The aim of this study was to investigate the reproducibility of risk estimates carried out using different animal species and strains. The Weibull and multistage mathematical modes, as well as the "linearized multistage" model (EPA "conservative" procedure), have been used for dose-response relationship fitting. Whenever necessary, the Michaelis-Menten function was introduced in the models to account for a possible metabolic process. The analysis was directed to dietary sodium saccharin and vinyl chloride inhalation carcinogenicity (nine total dose-response relationships). Data from different strains, species, and researchers were available for these substances. The general pattern of two sodium saccharin curves was an upward curve in both cases with a high consistency among parameters and risk estimates (the latter ranging within a factor 3). For vinyl chloride the trend was clearly downward for six carcinogenic dose-response relationships from different species; low-dose risk estimates, based on the upper confidence limits of the linear component, range within a factor 3. The study has indicated the reproducibility of the low-dose extrapolation process using different strains and species.