Estimation of interindividual pharmacokinetic variability factor for inhaled volatile organic chemicals using a probability-bounds approach

The derivation of reference concentrations (RfCs) for systemically acting volatile organic chemicals (VOCs) uses a default factor of 10 to account for the interindividual variability in pharmacokinetics (PK) and pharmacodynamics (PD). The magnitude of the PK component of the interindividual variability factor (IVF; also referred to as human kinetic adjustment factor (HKAF)) has previously been estimated using Monte Carlo approaches and physiologically based pharmacokinetic (PBPK) models. Since the RfC derivation considers continuous lifetime human exposure to VOCs in the environment, algorithms to compute steady-state internal dose (SS-ID), such as steady-state arterial blood concentration (Ca) and the steady-state rate of amount metabolized (RAM), can be used to derive IVF-PKs. In this context, probability-bounds (P-bounds) approach is potentially useful for computing an interval of probability distribution of SS-ID from knowledge of population distribution of input parameters. The objective of this study was therefore to compute IVF-PK using the P-bounds approach along with an algorithm for SS-ID in an adult population exposed to VOCs. The existing steady-state algorithms, derived from PBPK models, were rewritten such that SS-ID could be related, without any interdependence, to the following input parameters: alveolar ventilation (Qp), hepatic blood flow (Ql), intrinsic clearance (CLint) and blood:air partition coefficient (Pb). The IVF-PK was calculated from the P-bounds of SS-ID corresponding to the 50th and 95th percentiles. Following either specification of probability distribution-free bounds (characterized by minimal, maximal, and mean values) or distribution-defined values (mean, standard deviation and shape of probability distribution where: Qp = lognormal, Ql = lognormal, CLint = lognormal and Pb = normal) in RAMAS Risk Calc® software version 3.0 (Applied Biomathematics, Setauket, NY), the P-bound estimates of SS-ID for benzene, carbon tetrachloride, chloroform and methyl chloroform were obtained for low level exposures (1 ppm). Using probability distribution-defined inputs, the IVF-PK for benzene, carbon tetrachloride, chloroform and methyl chloroform were, respectively, 1.18, 1.28, 1.24, and 1.18 (based on P-bounds for Ca), and 1.31, 1.58, 1.30, and 1.24 (based on P-bounds for RAM). A validation of the P-bounds computation was performed by comparing the results with those obtained using Monte Carlo simulation of the steady-state algorithms. In data-poor situations, when the statistical distributions for all input parameters were not known or available, the P-bounds approach allowed the estimation of IVF-PK. The use of P-bounds method along with steady-state algorithms, as done in this study for the first time, is a practical and scientifically sound way of computing IVF-PKs for systemically acting VOCs.