Modelling of patient EMS exposure: translating pharmacokinetics of EMS in vitro and in animals into patients.

In order to support the toxicological risk assessment for the ethyl methanesulfonate (EMS) exposure of patients ingesting contaminated Viracept tablets (Müller and Singer, 2009), there was a need to correlate the effects observed in in vivo genotoxicity studies with mice to EMS exposure and to estimate human exposure to EMS at the level of contamination of Viracept tablets. The species differences in volume of distribution of EMS, a key factor for determination of its C(max), were small in the species investigated (mouse, rat, monkey), the species differences in clearance, the key factor involved in AUC assessment, were large (Lavé et al., 2009). Because of this uncertainty in extrapolation of clearance across species we used a conservative approach for human exposure predictions in terms of AUC where clearance was assumed to solely reflect the chemical stability of EMS neglecting additional clearance pathways such as metabolism and exhalation. This approach was compared to the estimates obtained from allometric scaling based on rat clearance, the species leading to the lowest clearance predicted in man. We found that both approaches led to nearly identical predictions of the human AUC. Thus, we predict a human AUC of 13 microM h for patients ingesting the most contaminated Viracept tablets, corresponding to a maximal daily intake of 0.055 mg/kg of EMS. The C(max) of EMS in these patients is predicted to be 0.85 microM. In order to provide a basis for toxicological risk assessment, these maximal human AUC and C(max) values are to be compared to the AUC and C(max) values in mice at the EMS dose of 25mg/kg which was found to be the threshold dose for induction of mutagenic effects, i.e. the dose at which no mutagenic effects were observed (Gocke et al., 2009-a). We calculate AUC and C(max) in mice at the threshold dose to be 350 microM h and 315 microM, respectively. Thus we conclude that a large safety factor can be deduced, whatever the basis of comparison, as is discussed in detail by Müller et al. (2009).