Comparison of Diesel Exhaust (DE) Unit Risk Estimates Derived from Animal Bioassays.* P. A. Valberg, A. Y. Watson and S. M. McCarthy, Gradient Corporation, 44 Brattle Street, Cambridge, MA 02138
Several animal studies have reported increased lung tumors in rats exposed to inhaled diesel engine exhaust (DE); epidemiologic studies suggest that DE may be a potential human carcinogen. We examined DE unit risks predicted for human lung cancers, as calculated by Smith and Stayner (1991), Hattis and Silver (1992), and Pepelko and Chen (1993). All three sets of authors used identical rat bioassay data linking chronic diesel exhaust inhalation to the development of lung tumors in rats (Mauderly et at., 1987). Although different in detail, the dose-response models were uniformly linear and non-threshold. However, each set of authors chose a different approach in relating the calculated rat DE unit risk to human DE unit risk. The dose metric used varied among the studies (e.g., air concentrations, mass of retained lung burden, mass retained per unit lung surface area). Species scaling of dose levels from animals to humans was done in different ways, and the phenomenon of lung overload (decreased lung clearance at high lung burdens) was not always considered. The unit risks derived differ by an eighty-fold factor between the highest, at 8 x 10-4 (Hattis and Silver), and the lowest at 0.1 x 10-4 (Pepelko and Chen); the unit risk is the predicted human cancer risk for continuous lifetime exposure to 1 mg/m3 DE particle concentrations. We found that the majority of this difference could be attributed to choice of dose metric and how each group treated particle overload in rat and human lungs. Despite these quantitative results, a number of unanswered questions and uncertainties frustrate DE risk assessment: (1) Cause of species-specific biological susceptibility (e.g., rats are susceptible to DE carcinogenicity, hamsters are not, and mice are equivocal), (2) Occurrence of an apparent threshold in the rat tumor response (at about 2 to 4 mg/m3), (3) Inclusion of squamous cysts as lung tumors (which was done in the above calculations, but was discounted in a recent rat study), (4) Differences in emission products and exposure scenarios between laboratory studies and human exposure, (5) Disagreement on the appropriate lung-dose metric and its low-dose extrapolation, (6) The role of lung overload both within and among species and (7) The significance of inhaled carbon-black carcinogenicity in rats. Additional data and analyses will be necessary before the potential carcinogenicity of diesel exhaust in humans can be quantitatively predicted.
*Work supported by the American Mining Congress, Washington, DC.