Physiologically based toxicokinetic modelling as a tool to assess target organ toxicity in route-to-route extrapolation--the case of coumarin.

Coumarin (1,2-benzopyrone) is occurring in food, and is also used in cosmetics. In order to perform a risk assessment for both oral and dermal exposure, we applied a physiologically based approach to model kinetics in humans by simulating both routes of exposure. The concentration-time profile in liver revealed a higher peak concentration (C(max-hep)) for the oral when compared to the dermal route. The area under the concentration-time curve in the liver (AUC(hep)) was found the same for both routes if the same extent of absorption is assumed. Dose response information from published rat studies were used to identify the metric relevant for liver toxicity. Liver exposure levels resulting from doses and durations as outlined in the studies were simulated in a rat model. We obtained 31 data pairs of C(max-hep) and AUC(hep). Liver toxicity was observed at doses which resulted in simulated C(max-hep) values exceeding a certain liver concentration whereas we could not identify a clear cut off value of AUC(hep). Our findings support the notion that liver toxicity of coumarin in rats is related to C(max-hep) rather than to AUC(hep). If these findings can be transferred to the situation in humans, the result demonstrates that route specific differences in organ peak concentrations have to be considered when performing route-to-route extrapolation.