Lukas Kovar (1), Dominik Selzer (1), Andreas Weber (1) and Thorsten Lehr (1)
(1) Clinical Pharmacy, Saarland University, Saarbruecken, Germany
Introduction: The importance of transdermal administration of drugs has been increasing steadily over the past decades [1]. One major reason for that is the growing interest in the transdermal administration as an alternative to more traditional methods of drug delivery allowing both convenient and painless drug delivery, a sustained release profile without plasma fluctuations and the avoidance of first pass hepatic metabolism [1,2]. However, delivery efficiency for many compounds is limited due to the low permeability of the skin, especially the stratum corneum. Hence, the possibility of predicting plasma concentration-time profiles of administered drugs based on their physicochemical properties and patch characteristics is of great interest for the development of new transdermal patches and dosing regimens.
Objectives:
- Development and evaluation of a semi-physiological transdermal model (SPTM) and a physiological transdermal model (PTM) within a whole-body physiologically based pharmacokinetic (PBPK) framework
- Prediction of fentanyl and nicotine plasma concentration-time profiles with the SPTM and PTM after the administration of transdermal drug patches
- Comparison of predictive performance between the SPTM and PTM
Methods: Two parent-metabolite PBPK models of fentanyl and nicotine, respectively, were developed in PK-Sim® and MoBi® (version 8.0) [3]. Physicochemical model parameters as well as plasma profiles and fractions of drug excreted unchanged to urine were obtained from the literature. For model development, 102 plasma profiles from 51 clinical studies were split into an internal training (n=50) and an external test (n=52) dataset. Initially, fentanyl and nicotine PBPK models were established by fitting unknown model parameters to intravenous and oral plasma profile training data. Subsequently, the models were complemented by the SPTM and the PTM. The SPTM consists of a vehicle and two skin compartments, which has shown to be an adequate option to model skin kinetics within a transdermal route of application [4]. The PTM consists of the vehicle, stratum corneum, epidermis and dermis, each divided into 40 layers. Model evaluation was performed by comparing plasma profiles predicted with observed as well as predicted to observed area under the plasma concentration-time curve (AUC) and maximum plasma concentration (Cmax) values. Moreover, as quantitative measures, the geometric mean fold errors (GMFE) of predicted and observed AUC and Cmax as well as the mean relative deviation (MRD) values of all predicted compared to observed plasma concentrations were calculated.
Results: The fentanyl model includes metabolism to norfentanyl through CYP3A4, an unspecific hepatic clearance, glomerular filtration and a renal excretion via the p-glycoprotein. Nicotine elimination was implemented through CYP2A6 and CYP2B6 metabolism to cotinine, glomerular filtration and an unspecific hepatic clearance mechanism. Absorption kinetics of the SPTM were estimated by fixing the obtained pharmacokinetic parameters from the training dataset and estimating unknown skin model parameters. On the contrary, most model input parameters for the PTM were estimated via Quantitative Structure Activity Relationships (QSARs). The models for both compounds show promising results, e.g. with GMFE of predicted versus observed AUC for nicotine of 1.07 for both the SPTM and the PTM, GMFE of predicted versus observed Cmax of 1.12 (SPTM) and 1.09 (PTM) and overall MRD values of 1.23 (SPTM) and 1.27 (PTM), respectively. Moreover, the individual AUC and Cmax ratios of all plasma profiles were within the common acceptance criterion of a twofold deviation.
Conclusions: Both the SPTM and the PTM were capable to successfully describe and predict all transdermal plasma profiles. Parameters of the PTM can be interpreted physiologically and estimated via QSARs. The results indicate that extrapolation from parameter sets from one dose to higher and lower doses as well as from single to multiple dosing regimens is possible for both models. The mechanistic model could be used to de novo predict nicotine and fentanyl plasma concentrations based on patch characteristics as well as in vitro dissolution data.
Funding: Supported by the German Federal Ministry of Education and Research (BMBF), 031L0153, “Alternativmethoden zum Tierversuch” and 03XP0196, ”NanoCare4.0–Anwendungssichere Materialinnovationen”.
References:
[1] Selzer D et al.: Finite and infinite dosing: Difficulties in measurements, evaluations and predictions. Adv Drug Deliv Rev (2012) 65(2):278-94.
[2] Lee H et al.: Device-assisted transdermal drug delivery. Adv Drug Deliv Rev (2018) 127:35-45.
[3] Eissing T et al.: A computational systems biology software platform for multiscale modeling and simulation: integrating whole-body physiology, disease biology, and molecular reaction networks. Front Physiol (2011) 2:4.
[4] Davies M et al.: Determining Epidermal Disposition Kinetics for Use in an Integrated Nonanimal Approach to Skin Sensitization Risk Assessment. Tox Sci (2011) 119(2):308-318.
Reference: PAGE () Abstr 9358 [www.page-meeting.org/?abstract=9358]
Poster: Drug/Disease Modelling - Absorption & PBPK