(T1230-04-23) PBPK Modeling to Predict the Effect of Gastric pH on Bioequivalence of Dabigatran Etexilate Capsules

Purpose: In generic drug development, physiologically based pharmacokinetic (PBPK) modeling and simulation has played an important role in the design of clinical bioequivalence studies and provided risk assessment to support the assessment of Abbreviated New Drug Applications (ANDAs). A noteworthy application of PBPK modeling is to predict the impact of gastric pH changes after co-administration of acid-reducing agents (ARAs) such as proton pump inhibitors (PPIs) on the bioequivalence of weak base drug products. Dabigatran etexilate (DABE) is a weak base drug which demonstrates pH-dependent solubility and DABE products also shows pH-dependent dissolution characteristics that may be formulation-dependent¹. Therefore, there is a potential of having different pH impact caused by the co-administration of an ARA on the absorption between a generic drug product and its reference listed drug (RLD) with formulation differences. The purpose of this study was to use a PBPK absorption model of DABE capsules and virtual bioequivalence (VBE) simulations to assess the impact of gastric pH change on the BE of a generic DABE capsule product. This work will also support the future implementation of ICH M13A to help assess the potential impact of ARA on BE outcome and determine if an additional in vivo BE study is needed for weak base drugs in subjects with elevated gastric pH.

Methods: Simcyp^TM simulator (Version 22) was used to develop the PBPK model for DABE capsules (Figure-1). The disposition model of DABE capsules was modified based on previously developed DABE PBPK models^2-5. The absorption model of DABE capsules was developed using the Advanced Dissolution, Absorption, and Metabolism (ADAM) module. The dissolution data of DABE capsules (RLD) in quality control (QC) media at pH 2 and plasma profiles following 150 mg oral dose under fasted condition were used to estimate the absorption parameters. The developed PBPK model was validated using pharmacokinetic (PK) data from 75 mg, 200 mg, and 600 mg dose of DABE (RLD). Administration of a PPI is known to increase the gastric pH up-to 5.0⁶. Therefore, in case of PPI co-administration under fasted condition, available dissolution profiles of RLD in QC media with elevated pHs (4.5 and 6.8) were used as model inputs. Since all the available QC dissolution profiles (at pH 4.5 and 6.8) of RLD resulted in underprediction of the PK profile, a hypothetical bio-predictive dissolution profile was generated to achieve a better model prediction of the PK data. To generate the hypothetical dissolution profile, QC dissolution profile of RLD at pH 4.5 was used and scaled up at each time point to get overall higher dissolution values. When applying this model to predict the PK profiles of RLD and generic products in Indian population, model parameters (e.g., clearance and regional permeability values) were adjusted based on published report of population PK differences of DABE between Indian and Caucasian populations^7-9. This modified model was validated using data from multiple sources. For conducting VBE simulations, hypothetical biopredictive dissolution profiles were generated for both generic and RLD products in a similar way and used as model inputs.

Results: Development and validation of DABE PBPK model in Caucasian population showed good predictive capability (PE ~ 20%) under fasted condition, where dissolution data generated under quality control (QC) condition (pH 2) were used as model input (Figure-2a). In case of PPI coadministration, the available dissolution profiles in QC media with elevated pHs (4.5 and 6.8) failed to provide satisfactory prediction (PE > 50% and PE > 200%, at pHs 4.5 and 6.8 respectively) when used as model input. Using a hypothetical bio-predictive dissolution profile as model input, the PBPK model demonstrated good predictive performance (PE < 20%) (Figure-2b). Furthermore, this PBPK model also demonstrated good predictive capability (PE < 20%) for Indian population after adjusting the clearance and regional absorption parameters based on the published reports. Lastly, using the hypothetical bio-predictive dissolution profiles as model inputs, VBE of DABE 150 mg capsule was conducted under fasted conditions with PPI coadministration in Indian healthy subjects. Those VBE simulations demonstrated BE between the generic and RLD products (Figure-3).

Conclusion: In this research work, a PBPK model for DABE was developed and validated. Using the hypothetically generated bio-predictive dissolution profiles (to mimic elevated pH conditions) as model inputs, the model was able to predict the PK profiles when the DABE was co-administered with PPI. Furthermore, virtual BE simulations predicted BE between a generic and RLD when both products were co-administered with PPI, suggesting low risk of non-BE for this specific generic product, thus an in vivo BE study with PPI coadministration may not be needed. Our work showed the utility of PBPK modeling as an alternative for BE assessment. Since a hypothetically generated bio-predictive dissolution data was used in the model, further experimental bio-predictive dissolution data might be needed to refine and validate this model for predicting the effect of ARAs on BE.

References: 1. Harada A, Ikushima I, Haranaka M, Yanagihara A, Nakayama D. Am J Cardiovasc Drugs. 2020;20(3):249-58.
2. Farhan N, Cristofoletti R, Basu S, Kim S, Lingineni K, Jiang S, et al. CPT Pharmacometrics Syst Pharmacol. 2021;10(3):199-210.
3. Laizure SC, Parker RB, Herring VL, Hu ZY. Drug Metab Dispos. 2014;42(2):201-6.
4. Blech S, Ebner T, Ludwig-Schwellinger E, Stangier J, Roth W. Drug Metab Dispos. 2008;36(2):386-99.
5. Ebner T, Wagner K, Wienen W. Drug Metab Dispos. 2010;38(9):1567-75.
6. Miehlke S, Madisch A, Kirsch C, Lindner F, Kuhlisch E, Laass M, et al. Aliment Pharmacol Ther. 2005;21(8):963-7.
7. Liesenfeld KH, Lehr T, Dansirikul C, Reilly PA, Connolly SJ, Ezekowitz MD, et al. J Thromb Haemost. 2011;9(11):2168-75.
8. Oeser SG, Bingham JP, Collier AC. Pharmaceutics. 2018;10(1).
9. Balram C, Sharma A, Sivathasan C, Lee EJ. Br J Clin Pharmacol. 2003;56(1):78-83.

Acknowledgements: Disclainer: The views expressed in this abstract are from the authors only and do not necessarily reflect the official policies of the Food and Drug Administration or Department of Health and Human Services, nor does any mention of trade names, commercial practices, or organization imply endorsement by the United States Government.


Figure-1: General workflow for development, validation, and application of the PBPK model for DABE.


Figure-2: Predicted profile of total Dabigatran following 150 mg capsule under a) Fasted condition without PPI and b) Fasted condition with co-administration of PPI (pantoprazole).


Figure-3: Virtual bioequivalence study between reference and test DABE 150 mg capsule under fasted condition and PPI (pantoprazole) coadministration; a) free dabigatran simulation and b) total dabigatran (free DAB + DAB-G conjugate) simulation.