Percutaneous penetration of drugs applied in transdermal delivery systems: an in vivo based approach for evaluating computer generated penetration models.

Human skin is a viable pathway for administration of therapeutics. Transdermal delivery systems (TDS) have been approved by the US-FDA since 1981. To enable the risk assessment of dermal exposure, predictive mathematical models are used. In this work the accuracy of predicted flux of the models is compared to experimental human in vivo data of drugs applied in US-FDA approved TDS. A database of pharmacokinetic data of drugs applied in TDS was used and updated. Three mathematical models (QSAR) were used to calculate predicted fluxes, and compared to the human in vivo data. For more than half of the drugs applied in TDS, the predicted flux by the mathematical models was an underestimation compared to the flux calculated with the experimental in vivo data. The flux was over- or underestimated by a factor 10-100. All mathematical models were significantly correlated with the experimental in vivo data. The process of percutaneous penetration has several influencing factors, TDS minimize some of these factors. Limitations are discussed. Further research is needed, with a focus on validation and standardization of the permeability coefficient. This work offers observations that should give a stimulus for refinement on the appropriate usage and limitations of mathematical models.