Molecular structure-based prediction of the steady-state blood concentrations of inhaled organics in rats

Chronic exposure to volatile organic chemicals (VOCs) in the environment leads to steady-state conditions. The establishment of quantitative relationships between steady-state blood concentrations and molecular structures of VOCs can be potentially useful. The objective of this study was therefore to investigate the relationship between the steady-state arterial blood concentration (Ca(ss)) in the rat and the molecular structures of 19 VOCs belonging to multiple chemical families (alkanes, haloalkanes, haloalkenes, and aromatics). The overall approach consisted of developing quantitative relationships between molecular fragments (CH(3), CH(2), CH, C, C horizontal lineC, H, Cl, benzene ring, and H in benzene ring structure) in alkanes, haloalkanes, haloethylenes, and aromatic hydrocarbons, as well as their Ca(ss) (associated with 1 mu mu mol/L inhalation exposure) according to an additive fragment model. This modeling approach implies that each fragment in the structure of a chemical has an additive and constant contribution to its Ca(ss). A multilinear regression was performed using a commercially available statistical software package, and the results obtained were essentially the contributions associated with each of the nine structural fragments toward Ca(ss) in the rat continuously exposed to 1 mu mu mol/L VOC in the air. The resulting model estimated adequately the Ca(ss) of VOCs initially used in the calibration (estimated/experimental ratio: 1.04 +/- 0.30, mean +/- standard deviation [SD]). This molecular structure vs. Ca(ss) relationship was then evaluated using an external dataset on Ca(ss) for three aliphatic hydrocarbons (octane, 2-methyl octane, and 1-nonene; 100 ppm exposures). The ratio of predicted to experimental Ca(ss) for these chemicals ranged from 0.6 to 1.2. The results of this study suggest that steady-state blood concentrations of inhaled VOCs can be predicted using structure-activity type models.