Predicting transdermal fentanyl delivery using mechanistic simulations for tailored therapy

Abstract

Transdermal drug delivery is a key technology for administering drugs. However, most devices are "one-sizefits-all", even though drug diffusion through the skin varies significantly from person-to-person. For nextgeneration devices, personalization for optimal drug release would benefit from an augmented insight into the drug release and percutaneous uptake kinetics. Our objective was to quantify the changes in transdermal fentanyl uptake with regards to the patient's age and the anatomical location where the patch was placed. We also explored to which extent the drug flux from the patch could be altered by miniaturizing the contact surface area of the patch reservoir with the skin. To this end, we used validated mechanistic modeling of fentanyl diffusion, storage, and partitioning in the epidermis to quantify drug release from the patch and the uptake within the skin. A superior spatiotemporal resolution compared to experimental methods enabled in-silico identification of peak concentrations and fluxes, and the amount of stored drug and bioavailability. The patients' drug uptake showed a 36% difference between different anatomical locations after 72 h, but there was a strong interpatient variability. With aging, the drug uptake from the transdermal patch became slower and less potent. A 70-year-old patient received 26% less drug over the 72-h application period, compared to an 18-year-old patient. Additionally, a novel concept of using micron-sized drug reservoirs was explored in silico. These reservoirs induced a much higher local flux (µg cm -2 h -1 ) than conventional patches. Up to a 200-fold increase in the drug flux was obtained from these small reservoirs. This effect was mainly caused by transverse diffusion in the stratum corneum, which is not relevant for much larger conventional patches. These micronsized drug reservoirs open new ways to individualize reservoir design and thus transdermal therapy. Such computer-aided engineering tools also have great potential for in-silico design and precise control of drug delivery systems. Here, the validated mechanistic models can serve as a key building block for developing digital twins for transdermal drug delivery systems.

Predicting Transdermal Fentanyl Delivery Using Mechanistic Simulations for Tailored Therapy

“…The funder was not involved in the study design, collection, analysis, interpretation of data, the writing of this article or the decision to submit it for publication. This manuscript has been released as a pre-print at bioRxiv (Defraeye et al, 2020).…”

Inverse mechanistic modeling of transdermal drug delivery for fast identification of optimal model parameters

“…Also, first-generation TDDS are analyzed in-silico to increase our insight into how these systems interact with the patient's skin [10]. Here, drug uptake differences between different substances [11] or age categories of patients [12] were targeted. Also, the impact of the composition of the stratum corneum and viable epidermis on the drug uptake was quantified [13], [14].…”

“…A key advantage of such first-principles-based modeling is the high spatial and temporal resolution they offer. Transient changes in concentration profiles can be identified within different skin layers during the uptake process [12], [15]. This spatiotemporal resolution enables us to virtually probe the drug transport inside and out of the skin in time and space.…”

“…Another advantage is that such in-silico tests can be performed swiftly and noninvasively, without any side effects for the patient. Physics-based modeling opens new ways of tailoring TDD therapy for patients or patient groups [12]. This deterministic approach is also suffering from statistical variability in the data due to biological variability or measurement uncertainty.…”

“…Simulating, however, also consumes a lot of time and resources, making it is usually not more efficient than measuring the parameters separately. As a third way, modelers just source these parameters from literature data during model setup instead ( [12], [24]- [26]). Due to a lack of appropriate data, often, data of other drugs with similar molecular weight and lipophilicity are taken [21].…”