(M1030-05-30) Modelling and Simulation of Fentanyl Matrix Patches and Risk Assessment from Manufacturing Errors

Purpose: Modelling and simulation (M&S) can provide insights into formulation impact on drug delivery when clinical data is not available. In this case study a recently reported manufacturing error^[1], where multiple dermal fentanyl patches were stacked in one dose packet, was simulated to investigate the potential impact of these defective patches could have on the plasma fentanyl concentration. This demonstrates the potential of M&S to bridge the gap in clinical information and assist in the risk assessment of product variance.

Methods: A physiologically based pharmacokinetic (PBPK) model was developed using the Simcyp V23 simulator software. An in vivo model was developed using literature IV fentanyl dosing studies for parameterisation of distribution and elimination. The dermal absorption model was initially developed using the IVPT module within Simcyp’s MPML MechDermA model, for the reference licenced branded fentanyl matrix patch (Duragesic, Janssen-Cilag). The model used was the mechanistic patch “drug in adhesive” which was parameterised using literature data for in vitro release (IVRT) and in vitro permeation testing (IVPT). In vitro-in vivo extrapolation (IVIVE) coupled the IVPT parameterised dermal absorption to the IV developed distribution and elimination model. This in vivo model was validated using literature data for patch exposure. The validated PBPK model was subsequently used to simulate the plasma profiles for scenarios arising from the reported manufacturing error (Figure 1). Scenario 1 involved three patches perfectly aligned atop each other; scenario 2 involved a mismatch to the overlays and increased that area of patch exposure to skin by 10%; scenario 3 was application of three patches adjacent to each other as a reference worst case scenario.

Results: A PBPK model was developed using literature IV dosing studies. Dermal absorption was developed using both IVRT and IVPT data, which resulted in well-captured IVPT simulations. IVIVE to the in vivo patch use captured the reported clinical data for Duragesic patch application (Figure 2). Simulations of the manufacturing error (Figure 3) predicted plasma concentrations would be elevated. In scenario 1, with a perfect overlap, the mean C_max and AUC_0-144 increased by 22% and 19% respectively, with the T_max delayed and shifted to the patch removal time, increasing the time that a patient may be exposed to an elevated plasma profile. Increasing the surface area by 10% due to non-aligned overlap (scenario 2) led to a mean C_max and AUC_0-144 increase of 32% and 30% respectively. While no degree of increased plasma concentration is linked to adverse events, it has been suggested that an increase of 25% may be significant^[2]. Scenario 3 was simulated (Figure 3) as a worst-case scenario to contrast an overlapping patch system to individually placed patches. This simulation resulted in a profile that was 3-fold increase over that of the single patch and highlights the importance of the area of patch coverage.

Conclusion: Simulations of these defective patches in virtual patients resulted in higher plasma concentrations, 22-32%, potentially in a range that may result in adverse events. However, due to the alignment of the patches on top of each other, the effect is not directly proportional to dose. These simulations reaffirm the FDA actions to recall the affected brand and highlights how mechanistic formulation modelling can be leveraged to examine the potential risks from manufacturing defects.

References: [1] USFDA, “Alvogen Issues Voluntary Nationwide Recall for One Lot of Fentanyl Transdermal System 25 mcg/h Due to a Defective Delivery System,” can be found under https://www.fda.gov/safety/recalls-market-withdrawals-safety-alerts/alvogen-issues-voluntary-nationwide-recall-one-lot-fentanyl-transdermal-system-25-mcgh-due-defective, 2025.
[2] E. J. M. Kuip, M. L. Zandvliet, S. L. W. Koolen, R. H. J. Mathijssen, C. C. D. van der Rijt, Br J Clin Pharmacol 2017, 83, 294–313.


Figure 1: Proposed scenarios for erroneous patch manufacturing application on individuals, with perfect overlay of three patches (Scenario 1), slight offset to (scenario 2) and worst-case of no overlap (scenario 3).


Figure 2: Validation simulations for in vivo application of Duragesic patches using literature clinical data. Solid line is mean prediction, grey lines are 5th and 95th percentiles. Solid points are observed data


Figure 3: Simulations of potential plasma profiles after applications of patches suffering from manufacturing errors. Red lines are erroneous formulations with mean (dark) and 5th and 95th percentiles (light), blue lines are standard patch simulation mean (dark) and 5th and 95th percentiles (light).