Reconstructing week-long exposures to volatile organic compounds using physiologically based pharmacokinetic models.

Reconstruction of human exposure to toxic chemicals using physiologically based pharmacokinetic (PBPK) models and biomarkers is an attractive prospect, because biomarker measurements generally provide the most direct evidence of dose. Previously it has been shown that it is possible to reconstruct short-term (30 minute) exposure to chloroform, and that it is possible in some cases to resolve the total dose between two routes of uptake (Georgopoulos et al., 1994). In this paper it is shown that it is mathematically feasible to reconstruct longer term exposures to volatile organic compounds (VOCs), using benzene as a paradigm for other VOCs, and exhaled breath concentration as a biomarker of exposure. First, it is shown that exhaled breath concentration is an appropriate biomarker for long-term exposure to benzene, since benzene accumulates in fat and is eliminated in exhaled breath. Application of a benzene PBPK model (Travis et al., 1990) showed that benzene continues to accumulate in the fat compartment for over 10 days, and consequently fat acts as an integrator of dose during this period. Second, the benzene PBPK model is used to reconstruct exposure using the maximum likelihood approach. Since no data were available for long-term exposures of this duration, "data" with a normally distributed random error and 30% coefficient of variation were generated by the PBPK model for a variety of daily exposures. It was shown that in most cases it is possible to estimate cumulative exposure within 40% of the actual values, even when the exposure concentration-time profile is unknown. The estimated exposure is found to always be an underestimate of the true exposure when the exposure concentration is assumed to be constant.