Relating transdermal delivery plasma pharmacokinetics with in vitro permeation test (IVPT) findings using diffusion and compartment-in-series models.

Increasing emphasis is being placed on using in vitro permeation test (IVPT) results for topical products as a surrogate for their in vivo behaviour. This study sought to relate in vivo plasma concentration - time pharmacokinetic (PK) profiles after topical application of drug products to IVPT findings with mechanistic diffusion and compartment models that are now widely used to describe permeation of solutes across the main skin transport barrier, the stratum corneum. Novel in vivo forms of the diffusion and compartment-in-series models were developed by combining their IVPT model forms with appropriate in vivo disposition functions. Available in vivo and IVPT data were then used with the models in data analyses, including the estimation of prediction intervals for in vivo plasma concentrations derived from IVPT data. The resulting predicted in vivo plasma concentration - time profiles for the full models corresponded closely with the observed results for both nitroglycerin and rivastigmine at all times. In contrast, reduced forms of these in vivo models led to discrepancies between model predictions and observed results at early times. A two-stage deconvolution procedure was also used to estimate the in vivo cumulative amount absorbed and shown to be linearly related to that from IVPT, with an acceptable prediction error. External predictability was also shown using a separate set of in vitro and in vivo data for different nitroglycerin patches. This work suggests that mechanistic and physiologically based pharmacokinetic models can be used to predict in vivo behaviour from IVPT data for topical products.