(458d) Screening of Parameters Influencing Performance of Tangential/Crossflow Filtration Systems in Continuous API Manufacturing Processes with a Mechanistic CFD Model

The latest revision of Annex 1 has had many impacts on the pharmaceutical industry since its release, but one of the most significant changes is increased scrutiny upon sterilized filtration systems with the Pre-Use Post Sterilization Integrity Testing (PUPSIT) requirement [1]. PUPSIT requires pharmaceutical manufacturers to test sterilized filter assemblies prior to use by performing non-destructive integrity testing such as bubble point, diffusive flow, water intrusion or pressure hold testing, which (while not uncommon) was not previously standard practice across the industry [2]. As a result, this revision will impose significant difficulties upon some pharmaceutical manufacturers and require adjustments to their current manufacturing processes to remain in compliance with Good Manufacturing Practice (GMP) [2].

However, in cases where PUPSIT is not possible due to process constraints, alternative approaches may be acceptable, provided that a thorough risk assessment is performed, and compliance achieved through implementation of appropriate controls to mitigate risk in non-integral filtration systems. Per Annex 1, this risk assessment should include “in depth knowledge and control of the filter sterilization process to ensure the potential damage to the filter is minimized” [1].

The latest revision of Annex 1 was finalized in August 2022 and will be coming into effect in August 2023, one year after publication (excluding lyophilization). As the deadline grows closer, it is an especially critical time to investigate and better understand the multitude of factors influencing filtration systems in order to perform effective risk analysis and develop proper mitigation strategies.

One form of filtration showing considerable interest in recent years is cross flow filtration, also known as tangential flow filtration (TFF), which is in part due to its compatibility with continuous manufacturing processes and flow chemistry [3]. A major benefit of TFF systems relative to standard dead-end filtration is reduced buildup of particulate on the membrane’s surface (the “filter cake”) which is detrimental to effectiveness and can block a filter. In a TFF setup, the flow tangential to the membrane effectively “washes away” the particulate buildup and extends filter lifetime, which is ideal when employing continuous manufacturing of APIs.

In this work, a mechanistic CFD model of a TFF system is developed, calibrated and verified with experimental data. The model verification involved comparing the model prediction with the collected experimental data (residence time distribution, saturation rates, hold times, etc.) at various conditions which encompassed the standard operational range of said filtration unit. This model was then used to perform screening experiments with various material and operational parameters that impact the performance of a filter. Some system parameters used in screening include membrane thickness, permeability, fluid viscosity, and diffusion coefficients.

In a traditional lab setting, such an investigation would require many unique experimental setups, various materials, and likely utilize a Design of Experiments approach (DoE) with a limited number of runs as a fractional factorial to minimize time and cost. These fractional factorial approaches, while efficient, can limit the usefulness of generated statistical models. Without sufficient datapoints spanning the design space, interactions between parameters cannot be fully investigated and may lead to important interactions being missed [4].

However, with a mechanistic CFD model, the restrictions on the number of runs which accompany physical experiments are reduced and a full factorial experiment is possible. Upon completion of the screening runs, analysis was performed using statistical software to create various empirical linear and nonlinear models which are useful for understanding the complexities and interactions of the various parameters.

The primary objective of this study is to enhance our understanding of TFF/CFF filtration systems and demonstrate a method that can be widely applied across the pharmaceutical industry as the latest revision of Annex 1 comes into effect. This approach can be performed on other types of filtration systems and geometries within pharmaceutical manufacturing and other industries. The approach is simple and straightforward, which allows it to be easily adapted for screening experiments with more parameters under investigation.

References

[1] The Rules Governing Medicinal Products in the European Union Volume 4 EU Guidelines for Good Manufacturing Practice for Medicinal Products for Human and Veterinary Use, Annex 1: Manufacture of Sterile Medicinal Products. Brussels 22.8.22 C(2022) 5938 final https://health.ec.europa.eu/system/files/2022-08/20220825_gmp-an1_en_0.pdf

[2] 2022 Annex 1 Workshop (Palm Springs), Oct 20-21 2022, hosted by the Parenteral Drug Association.

[3] Huter, M; Strube, J. (2019). Model-based design and process optimization of continuous single pass tangential flow filtration focusing on continuous bioprocessing. Processes, 7(6), 317. https://doi.org/10.3390/pr7060317

[4] Hessing, T. (2023, March 25). Partial/fractional factorial design. Six Sigma Study Guide. Retrieved March 31, 2023, from https://sixsigmastudyguide.com/partialfractional-factorial-design/#:~:te....

Acknowledgements

This work is supported by the US Food and Drug Administration (FDA) under contract number 75F40122C00122.