Analysis of diesel-exhaust unit-risk estimates derived from animal bioassays

Chronic inhalation of diesel exhaust (DE) causes lung tumors in rats; epidemiologic studies suggest that DE may be a potential human carcinogen. We compared the DE unit risks for human lung cancer as derived by Smith and Stayner (1991), Hattis and Silver (1992), Pepelko and Chen (1993), California Environmental Protection Agency, Office of Environmental Health Hazard Assessment (1994), and the U.S. Environmental Protection Agency (USEPA) (1994). All five sets of authors used identical rat bioassay data (Mauderly et al., 1987). Although different in detail, the dose-response models were uniformly linear and nonthreshold. However, each set of authors chose a different approach in relating the rat data to DE unit risk in humans. The predicted unit risks for continuous lifetime exposure to 1 microgram/m3 DE differed by an 80-fold factor between the highest [8 x 10(-4), Hattis and Silver] and the lowest [0.1 x 10(-4), Pepelko and Chen]. The choice of dose-input parameters and how each group treated particle overload were the major factors affecting the different risk estimates. Several unanswered questions undermine the current use of the rat bioassay for DE risk assessment: (1) Differences in emission products and exposure scenarios between laboratory studies and human exposure, (2) occurrence of an apparent threshold in the rat-lung-tumor response, (3) uncertainty as to the appropriate lung-dose metric and its low-dose extrapolation, (4) the role of lung overload, and (5) the cause of species-specific biological susceptibility.