Update of time extrapolation factors for risk assessment: The benefit of combined databases and probabilistic analyses

In regulatory risk assessment, extrapolation factors (EFs) are used for filling data gaps, thus reducing the amount of animal testing. Extrapolation is common practise to consider i.e. time-, interspecies-, intraspecies extrapolation. Time extrapolation is performed in cases where a short time study is available, but safety assessment for chronic exposure conditions is required. The NO(A)EL (no observed (adverse) effect level) of the long term study is then estimated by applying the corresponding extrapolation factor. In the present project, time extrapolation factors for oral and inhalation administration routes are determined by a probabilistic approach. Paired studies with oral or inhalation exposure for the same chemical but with different study durations were extracted from literature, the ECHA CHEM, ToxRef DB,IMI Etox DB, the ELINCS DB and the RepDose DB. Datasets were compiled for the extrapolations of subacute to subchronic and of subchronic to chronic. Probabilistic analyses were carried out with the resulting EF-datasets oral: subchronic to chronic; oral: subacute to subchronic; inhalation: subchronic to chronic. In the case of the inhalation route, local and systemic EFs were analysed separately. In the largest dataset, oral: subchronic to chronic EF for groups of compounds sharing structural properties were analysed. These included aliphatic esters, haloalkanes, phenols or phthalates. The EF derived for specific groups did not differ significantly from the EF of the entire dataset. The analysis of subgroups with different toxicological potency, however, revealed a consistent trend in all three datasets. Compounds with relatively low NOELs in the short term study showed on average significantly lower time EFs com-pared to those being less toxic. Cut-off values were set to define two potency groups per dataset. The resulting EFs per potency group are shown and discussed.