Abstract:
Published estimates of water-skin permeability coefficients of organic contaminants found in potable water are highly variable due to differences both in experimental conditions and mathematical approaches. Human in vivo experiments are potentially the most credible sources of information about dermal absorption, but require relatively sophisticated mathematical models (e.g., physiologically based pharacokinetic models) to interpret resulting biomonitoring data. Within those models alternative mathematical representations of skin can yield substantially different estimates of the permeability coefficient from the same data. Gordon et al. (1998) conducted experiments at different water temperatures in which breath chloroform data were collected from bathing human volunteers. Their data have been reevaluated using three PBPK models that differ only in the manner in which skin is represented. Values of the water-skin permeability coefficient (Kp) were estimated by Markov chain Monte Carlo (MCMC) ditting to observed breath concentrations from Gordon et al.'s trials. Statistically significant differences in estimates of permeability coefficients were found across temperatures and across models. Lessons learned from this exercise are applicable to the larger question of the relative importance of ingestion and dermal exposures to water contaminants. A protocol given by the U.S. EPA Office of Solid Waste and Emergency Response leads to the conclusion that the dermal contribution is small compared to ingestion for chloroform and other trihalomethanes (THMs). However, an assessment published by the U.S. EPA Office of Research and Development comes to a different conclusion. reporting that oral and dermal exposures to THMs are roughly equivalent. The latter conclusion is supported by present analysis. Disparate findings in these documents and prior publications are attributable to different modeling approaches and parameter assumptions. Factors that lead to underestimation of the contribution of dermal exposure include assumption of steady state absorption when unwarranted, inconsistent model parameters, and failure to consider temperature effects on skin permeability and water quality.