(W1130-01-03) Simulation of Variable Fasted Gastric Motility and pH as a Key to Understand Contrasting Properties of Immediate Release Pellet-Filled Capsules and Bare Pellets

Purpose: Recent studies showed that migrating motor complex (MMC) influences drug product disintegration and absorption [1]. However, simulation of gastric pressure events remains a challenging aspect of the in vitro characterization of orally-administered medicines. Not only do they vary with timing of occurrence and pressure magnitude, but also pharmacopoeial methods do not include them. Therefore, the goal of this study was to develop a workflow for the generation of individual dissolution profiles, representative for a random time of intake during the MMC cycle [2]. The method served to elucidate contrasting performance of Pradaxa pellet-filled capsules and bare pellets in vivo: administration of capsules was characterized by longer and more variable t_max, as well as lower systemic exposure in comparison with pellets [3].

Methods: Dissolution experiments were performed in the novel PhysioCell apparatus (Physiolution Polska), which is capable of generating physiologically-relevant pressure events of up to 500 mbar [4]. Also, the device simulates temperature and flow rate gradients, and allows for the online pH measurement and adjustment. In this study it was used to perform tests that varied with dissolution medium pH, and timing of intragastric stress (IST) and gastric emptying (GET) events (Table 1). A three-dimensional matrix of bio-relevant experiments was created with Central Composite Design (CCD). The dissolution medium, supplying the PhysioCell, consisted of HCl of pH adequate for each test defined in the experimental matrix. The gastric emptying sequence included a switch of the medium to 1 mM phosphate buffer (PB), accompanied with three 300 mbar pressure waves. Obtained dissolution profiles served as a training dataset for machine learning (ML) algorithms: Neural Networks and Random Forest. Models of the dissolution as a function of dissolution medium pH, IST and GET were then verified and used to generate multiple individual dissolution profiles under customized test parameters.

Results: Figure 1 presents a comparison of the two dosage forms under Test 1 conditions: IS and GE simulated at 19 and 30 min, respectively and gastric fluid pH of 2.6. In this test pellets started dissolving immediately, and the drug was completely dissolved and evacuated from the flow-through cell upon transition to the intestinal conditions. On the contrary, the disintegration of capsules was triggered with the IS. The drug only partially dissolved before the gastric emptying, but the dissolution under intestinal conditions was partially facilitated by the tartaric acid included in the formulation. Once the buffering effect was over, pH exceeded 4.0 (pKa of dabigatran etexilate) and the drug precipitated immediately. Similar trends were observed under all test conditions, as they were discriminative for the two dosage forms. Further on, two models of dissolution profiles as a function of gastric pH, IST and GET were constructed with ML algorithms and verified with independent dissolution experiments. Correlation coefficients of the model predicted verification profiles to the actual results equaled 0.9778 and 0.9799 for capsules and pellets, respectively. Models allowed to generate multiple dissolution profiles under customized test conditions (Figure 2). Obtained profiles reflected capsules’ susceptibility to disintegrate under pressure events and immediately release the drug, as seen in vitro. They also mirrored increased dissolution rate under low values of pH. On the other hand, profiles generated for pellets were uniform, despite various simulated conditions. Their simulated performance was robust in contrast to variable dissolution profiles of capsules in reaction to physiological conditions. Therefore, both dissolution results and model-generated profiles showed faster and less variable dissolution of pellets, which can explain n-vivo observed differences.

Conclusion: Characterization of IR dosage forms based on bio-relevant dissolution apparatus, physiological test space and ML simulations provides an unique insight into its performance under variable fasted gastric conditions. Such a workflow might be prospectively used to generate individual dissolution profiles, representative for a random intake time during MMC cycle, for the further pharmacokinetic simulations.

References: 1. Hens B, Tsume Y, Bermejo M, Paixao P, Koenigsknecht MJ, Baker JR, et al. Low Buffer Capacity and Alternating Motility along the Human Gastrointestinal Tract: Implications for in Vivo Dissolution and Absorption of Ionizable Drugs. Mol Pharm. 2017;14:4281–94.
2. Staniszewska M, Myslitska D, Romański M, Polak S, Garbacz G, Dobosz J, et al. In Vitro Simulation of the Fasted Gastric Conditions and Their Variability to Elucidate Contrasting Properties of the Marketed Dabigatran Etexilate Pellet-Filled Capsules and Loose Pellets. Mol Pharm. 2024;acs.molpharmaceut.4c00025.
3. Boehringer Ingelheim. Relative Bioavailability of Dabigatran After Administration of Different Application Forms of a Single Oral Dose of 150 mg Dabigatran Etexilate (Capsule, Powder for Reconstitution Into Solution, Pellets on Food) in Healthy Male and Female Volunteers (an Open-label, Randomised, Three-way Crossover, Clinical Phase I Study) [Internet]. clinicaltrials.gov; 2017 Apr. Report No.: NCT02171611.
4. Staniszewska M, Romański M, Dobosz J, Kołodziej B, Lipski U, Garbacz G, et al. PhysioCell; — a Novel, Bio-relevant Dissolution Apparatus: Hydrodynamic Conditions and Factors Influencing the Dissolution Dynamics. AAPS PharmSciTech. 2023;24:65.

Acknowledgements: The work has been partially financed by the Polish National Centre for Research and Development (POIR.01.02.00-00-0011/17).

Table 1. Matrix of bio-relevant dissolution experiments generated with CCD.

Figure 1. Comparison of the dissolution of capsules and bare pellets under exemplary test conditions.


Figure 2. ML-model generated dissolution profiles for capsules (a - c) and pellets (d - f).