Derivation of a Drinking Water Action Level. M. Whittaker, ENVIRON Corporation, Arlington, VA; and L. Bestervelt and G. Ball, NSF International, Ann Arbor, MI
U.S. EPA’s proposed cancer risk assessment guidelines were used to derive a maximum drinking water level (MDWL) for benzaldehyde. In the absence of biologically-based or case-specific models, these guidelines identify linear and nonlinear approaches to extrapolate from empirically-derived data. Both approaches were employed to contrast differences in methodology. Benzaldehyde doses associated with mouse forestomach squamous-cell papillomas were scaled in proportion to (body weight) 0.75 to derive human equivalent doses. For modes of action requiring linear extrapolation, Weibull and quantal polynomial models were used to derive 95% lower bound estimates (LBs) of benzaldehyde concentrations associated with 10% increases in forestomach squamous-cell papillomas (LED10). Slopes from these extrapolations were used to calculate approximately equivalent MDWLs associated with a 1 in 100,000 cancer risk. For modes of action allowing nonlinear methods, Weibull and quantal polynomial models were used to identify 95% lower bound estimates (LBs) corresponding to 10% increases in forestomach squamous-cell papillomas (LED10). These concentrations were modified by uncertainty factors and used to derive approximately equal MDWLs. Toxicology data on benzaldehyde and other benzyl group chemicals suggest that benzaldehyde has a nonlinear tumorigenic mode of action. Because of this, the MDWL derived using nonlinear extrapolation was selected over the MDWL using linear extrapolation.
Go to . . .