Find out more about the drug product filtration system
Sterile filtration of a drug product is the final process step to ensure the product is sterile prior to final container filling. It is critical both from a regulatory and patient safety perspective. The recent update to the EU GMP Annex 1 requires pre-use post-sterilization integrity testing (PUPSIT) of sterilizing-grade filters as part of the contamination control strategy to mitigate risk of contamination.
The drug product filtration system is automated and optimized to reproducibly complete PUPSIT and post-use filter integrity testing. It also provides in situ (point-of-use) leak testing of flow kits to mitigate the potential risk of using damaged flow kits due to transportation, storage, or manual installation before committing to the final sterile filtration step of the drug product. The tests assess the performance of the barriers to contamination to ensure a closed sterile flow path prior to filtration.
In this study, we installed flow kits with filter sizes from 1 inch (KA1) to 10 inch (NP6) positioned in redundant and single filtration configurations. We then performed pre-use leak test, and pre-use (PUPSIT) and post-use integrity test (forward flow and bubble point test) of the filters using automated methodologies. 148 pre-use (PUPSIT) and post-use filter integrity tests and 65 pre-use leak tests were performed on the system flow path of which all passed, proving the drug product filtration system barriers to contamination are reliable and robust in mitigating the risk to both the product and patient.
Introduction
Final sterile filtration of drug product prior to filling is an essential step in drug product manufacturing. This step ensures the product is sterile and safe prior to filling into the final vials/syringes to be given to a patient. Therefore, it is imperative this filtration is completed in a robust manner ensuring all barriers to contamination are performing as required.
The drug product filtration system facilitates fully automated in situ leak testing, filter wetting, pre-use post-sterilization integrity testing (PUPSIT), and post-use integrity testing. The filters can be configured in two arrangements, either single or redundant, depending on the process needs. Sterile gas filters are positioned above each liquid filter to ensure the flow path is fully closed and allows natural venting of air during filter priming and sterile air to be used for leak and filter integrity testing. In situ leak testing has been incorporated into the automated method, directly after flow kit installation to ensure the full flow path is leak free prior to product commitment. Fully automated filter priming, flushing, and pressurized soaking has been integrated into the method prior to pre-use and post-use integrity testing to mitigate the risk of false failures due to incomplete filter wetting.
In situ leak test and filter integrity test performance data generated from typical application tests with the system show the simplicity and reliability of these automated methods using the full range of filter sizes (KA1 to NP6) and filter arrangements (single and redundant).
Material and methods
The equipment and materials used for this study are detailed in Tables 1 to 3. Two sterilizing grade liquid filters were used in different filter formats: Fluorodyne™ II DFL, a 0.2 µm rated sterilizing grade liquid filter comprising of two symmetric membrane layers of hydrophilic modified polyvinylidene fluoride (PVDF), and Supor™ EKV, a 0.2 µm rated sterilizing-grade liquid filter comprising of one asymmetric prefilter membrane layer (upstream) and one symmetric membrane layer (downstream) made of hydrophilic polyethersulfone (PES).
Table 1. Equipment used for this study
| Description |
| 20 L flush bag trolley with weight scale |
| 200 L square container stainless steel with load cells |
| LevMixer™ drive gen IV |
Table 2. Single-use manifolds used for this study
| Description |
| Drug product filtration system sampling |
| Drug product filtration system flush bag F1 20 L |
| Drug product filtration system flush bag F2 20 L |
| Drug product filtration system feed single filter |
| Drug product filtration system feed single filter w/o flow/temp |
| Drug product filtration system filter 1 KA1DFL |
| Drug product filtration system filter 1 KA2DFL |
| Drug product filtration system filter 1 KA3DFL |
| Drug product filtration system filter 1 KA4DFL |
| Drug product filtration system filter 1 NP5LEKV |
| Drug product filtration system filter 1 NP6DFL |
| Drug product filtration system filter 1 NP6EKV |
| Drug product filtration system filter 2 KA1DFL |
| Drug product filtration system filter 2 KA2DFL |
| Drug product filtration system filter 2 KA3DFL |
| Drug product filtration system filter 2 KA4DFL |
| Drug product filtration system filter 2 NP5LEKV |
| Drug product filtration system filter 2 NP6DFL |
| 200 L LevMixer™ biocontainer bag |
Table 3. Product materials used in this study
| Description |
| Reverse osmosis (RO) water |
| Glycerol |
The drug product filtration system can be set-up for two different filtration configurations, either single or redundant filtration. Additionally, a flush trolley with integrated weight scales can be positioned next to the system to hold either two 20 L or two 50 L flush bags. For this study both filtration configurations and the 20 L flush bag trolley were used (Fig 1). The hardware and flow kits were installed and operated as per the operating instructions. Additionally, the process was performed using an automated recipe which was proactively defined by the user prior to study start.
Fig 1. Typical process set-up for drug product filtration system. The picture shows the system in a redundant filtration set-up (two Kleenpak™ Nova NP6 filters) equipped with a 50 L stainless steel container with a LevMixer™ drive gen IV unit (left) and a 20 L flush bag trolley (right).
Flow kit leak testing
Flow kit leak testing was performed prior to using a new flow kit. To test flow kits for leakage, two types of tests can be performed using the Palltronic Flowstar IV LGR. The leak test checks for potential leaks in flow kit tubing, while the single-use system (SUS) leak test checks the flow kit tubing including the 2D biocontainer(s) for potential leaks. To check the entire installed flow kit on the system for potential leaks, four main leak tests were performed covering different flow paths:
- Leak test 1: Covers the flow path from the inlet pump to filter position 1. This is a flow kit leak test which checks potential leakage of the flow kit tubing.
- Leak test 2: Covers the flow path from filter position 2 to the recovery bag and the filter 2 flush bag. This is a SUS leak test which checks potential leakage of the connected 2D bags and flow kit tubing.
- Leak test 3: Covers the flow path from filter position 1 to filter position 2. This is a flow kit leak test which checks the potential leakage of the flow kit tubing.
- Leak test 4: Covers the flow path from filter position 1 to the filter 1 flush bag. This is a SUS leak test which checks potential leakage of the connected 2D flush bag.
The full range of flow kit combinations were leak tested as part of the system validation, see Figure 2 showing each leak test flow path on the system. The full flow kit list and input parameters for leak testing can be found in Table 2 and Table 5, respectively. The maximum acceptable diffusive gas flow limit for each flow path was based on the volume and type of components used within each flow path studied. These results are presented in Table 8. Leak tests were performed at ambient temperature (15°C to 25°C). During the test, the temperature was controlled within ± 1°C from the temperature at the start and end of the test. After completion of the SUS leak test the 2D 20 L biocontainers were fully evacuated automatically using the air evacuation device (silicon cover).
Fig 2. Leak testing flow paths available on the system. The picture shows the redundant filtration configuration of the drug product filtration system with a 20 L flush bag, however, the general flow path defined for each leak test is the same regardless of configuration or size.
Filter integrity testing
The system can accommodate in-line filters sized from 1 inch (KA1) to 10 inch (NP6) (Fig 3). This study encompassed the full range of filter sizes available positioned in both the filter 1 and filter 2 position for single or redundant filtration. The PUPSIT and the post-use integrity test of the filters was fully automated by the system. Filter integrity tests were performed at ambient temperature (15°C to 25°C). During the test, the temperature was controlled to be within ± 1°C from the temperature at the start and end of the test.
Fig 3. Filter size range with indicative batch volume ranges, that were used in both single and redundant filtration configurations.
For the PUPSIT , the filters were primed, flushed, and integrity tested in situ using an automated method. Filter priming was performed for all filter capsules at 0.10 L/min regardless of filter media and size, which allowed a slow and reliable filling of the filter capsule and natural air venting during filter priming (approximate fill rate = 1” of filter cartridge per min). Each filter was flushed according to recommended filter flushing procedures for the specific filter media and size to minimize filter extractables/leachables (Table 4). Forward flow and bubble point integrity tests were performed according to the parameters shown in Table 5 and integrity test results including acceptance criteria are shown in Table 7.
RO water and 67% glycerol:water solutions were used as the product for the application runs. When water was used as the product no additional flushing was completed prior to post-use integrity testing. When 67% glycerol:water solution was used as the product, prior to post-use integrity testing, each filter was then flushed until the filter hold-up was fully replaced with water. This was determined by measuring the viscosity of a sample taken from the filter effluent after the full flush was completed. Once the viscosity of the sample measured was equal to that of the reference reverse osmosis water sample (0.998 cP) the filter was considered flushed. The integrity testing parameters are presented in Table 6. For the full list of filters studied and the associated acceptable air flows, see Table 7.
Table 4. Filter flushing flow rate and time per capsule size
| Capsule size and media | Flow rate (L/min) | Flush time (min) |
| KA1 DFL | 0.10 | 5 |
| KA2 DFL | 0.15 | 5 |
| KA3 DFL | 0.27 | 5 |
| KA4 DFL | 0.60 | 5 |
| KA5 DFL | 0.49 | 5 |
| KA6 DFL | 1.00 | 5 |
Table 5. Leak testing parameters used to perform leak tests on flow kits
| Process parameter | Value |
| Filter integrity test instrument | Palltronic Flowstar IV LGR |
| Test gas | Air |
| Leak test pressure (flow kit tubing only) | 1200 mbar |
| SUS leak test pressure (flow kit including biocontainer) | 50 mbar |
| Leak test time | 600 s |
Table 6. Filter integrity test parameters used for the forward flow (FF) and bubble point (BP) tests
| Process parameter | Value |
| Filter integrity test instrument | Palltronic Flowstar IV LGR |
| Filter wetting fluid | RO water |
| Test gas | Air |
| Forward flow filter integrity test time | 600 s |
| Bubble point module factor | KA1 = 0.1 KA2 = 0.2 KA3 = 0.4 KA4 = 0.5 NP5 = 0.5 NP6 = 1 |
Results and discussion
Automated filter wetting and filter integrity testing
The recent update to the EU GMP Annex 1 requires pre-use post-sterilization integrity testing (PUPSIT), in addition to, post-use filter integrity testing, as filter integrity is critical for the final sterilizing grade filtration step. The drug product filtration system combines both automated filter wetting and automated in situ filter integrity testing to minimize risk due to operator error and mitigate the likelihood of obtaining false failures of filter integrity test due to incomplete wetting. For this study, we performed both forward flow and bubble point integrity tests with water-wetted filters to assess effective wetting and reliable filter integrity testing.
The filters integrated into the single-use flow kits passed all 148 pre-use (PUPSIT) and post-use forward flow and bubble point integrity tests (Table 7). An example of the forward flow and bubble point integrity test execution by the Flowstar IV LGR is shown in Figure 4 and Figure 5, respectively. The integrity testing results detailed were completed with filters arranged in the two available configurations (single or redundant), in a range of sizes (KA1 to NP6) with two different Cytiva sterilizing-grade filters (Fluorodyne™ II DFL and Supor™ EKV) confirming the system reliability and integrity over a full range of processing applications.
Table 7. Summary of the 148 filter integrity tests (forward flow and bubble point) performed by the drug product filtration system based on filter type and size
| Filter type and size | Integrity test type | Sample number | Maximum acceptable flow (FF)/minimum bubble point test pressure (BP) | Average measured flow (FF)/bubble point test pressure (BP) ± standard deviation (σ) | Result |
|
Fluorodyne™ II KA1DFL |
Forward flow |
Pre-use = 17 Post-use = 17 |
≤ 1.0 mL/min |
Pre-use = 0.7 ± 0.1 mL/min Post-use = 0.7 ± 0.1 mL/min |
PASS |
|
Fluorodyne™ II KA1DFL |
Bubble point |
Pre-use = 17 Post-use = 17 |
≥ 3180 mbar |
Pre-use = 4010 ± 60 mbar Post-use = 4040 ± 64 mbar |
PASS |
|
Fluorodyne™ II KA2DFL |
Forward flow |
Pre-use = 4 Post-use = 0 |
≤ 2.0 mL/min |
Pre-use = 1.2 ± 0.3 mL/min Post-use = N/A |
PASS |
|
Fluorodyne™ II KA2DFL |
Bubble point |
Pre-use = 4 Post-use = 0 |
≥ 3180 mbar |
Pre-use = 4110 ± 75 mbar Post-use = N/A |
PASS |
|
Fluorodyne™ II KA3DFL |
Forward flow |
Pre-use = 2 Post-use = 2 |
≤ 3.4 mL/min |
Pre-use = 2.5 ± 0.1 mL/min Post-use = 2.3 ± 0.1 mL/min |
PASS |
|
Fluorodyne™ II KA3DFL |
Bubble point |
Pre-use = 2 Post-use = 2 |
≥ 3180 mbar |
Pre-use = 4080 ± 35 mbar Post-use = 4050 ± 71 mbar |
PASS |
|
Fluorodyne™ II KA4DFL |
Forward flow |
Pre-use = 2 Post-use = 0 |
≤ 7.4 mL/min |
Pre-use = 5.2 ± 0.2 mL/min Post-use = N/A |
PASS |
|
Fluorodyne™ II KA4DFL |
Bubble point |
Pre-use = 2 Post-use = 0 |
≥ 3180 mbar |
Pre-use = 3980 ± 35 mbar Post-use = N/A |
PASS |
|
Supor™ NP5LEKV |
Forward flow |
Pre-use = 2 Post-use = 0 |
≤ 7.5 mL/min |
Pre-use = 5.6 ± 0.7 mL/min Post-use = N/A |
PASS |
|
Supor™ NP5LEKV |
Bubble point |
Pre-use = 2 Post-use = 0 |
≥ 3320 mbar |
Pre-use = 3910 ± 14 mbar Post-use = N/A |
PASS |
|
Fluorodyne™ II NP6DFL |
Forward flow |
Pre-use = 14 Post-use = 14 |
≤ 12.0 mL/min |
Pre-use = 7.8 ± 0.6 mL/min Post-use = 7.2 ± 0.5 mL/min |
PASS |
|
Fluorodyne™ II NP6DFL |
Bubble point |
Pre-use = 14 Post-use = 14 |
≥ 3180 mbar |
Pre-use = 3860 ± 76 mbar Post-use= 3900 ± 73 mbar |
PASS |
Fig 4. Example of a forward flow integrity test sequence completed for a NP6DFLP6G filter wetted with water. Test gas was air, test pressure = 2760 mbar, maximum acceptable flow limit = 12.0 mL/min, test time = 600 s. Each phase of the integrity test (pressurization, stabilization, and measurement) is labelled with vertical reference lines showing the transition point into each phase. (*pressurization phase, †vent phase after test completion)
The forward flow integrity test sequence using the Flowstar IV LGR integrity test instrument starts with a pressurization phase to the entered and validated test pressure (e.g., 2760 mbar). Next the stabilization phase starts where the stabilization time is automatically controlled by the integrity test instrument and finishes as soon as the flow has stabilized within defined limits. Once the stabilization phase is completed, the measurement phase commences where the diffusive gas flow through the wetted filter membrane is measured over the set test time.
If the measured diffusive gas flow is less than or equal to the specified maximum acceptable flow limit (mL/min) the test is considered a pass. If the measured diffusive gas flow is greater than the specified maximum acceptable flow limit (mL/min) the test is a fail. The measured flow for the tested NP6DFLP6G filter (Fig 4), with a specified maximum acceptable flow limit (mL/min) of 12.0 mL/min, was 8.1 mL/min providing a pass result.
Fig 5. Example of a bubble point integrity test sequence completed for a water wetted NP6DFLP6G filter. Test gas was air and the minimum bubble point pressure was 3180 mbar. Each phase of the integrity test (pressurization, leak test, and measurement) is labelled with vertical reference lines showing the transition point into each phase. (*pressurization phase, †vent phase after test completion)
The bubble point integrity test sequence using the Flowstar IV LGR integrity test instrument starts with a pressurization phase to 80% of the entered minimum bubble point pressure. After this and prior to the actual bubble point measurement phase a leak test phase is performed to establish that there is no gross leak/defect in the filter system under test. During the leak test phase, the instrument measures the gas flow occurring at this gas pressure and compares this against a limit value which is determined by setting a module factor. If the gas flow measured is below the limit value, the bubble point test proceeds and enters the measurement phase. The applied module factors for the bubble point testing is listed in Table 6. In the bubble point measurement phase, the pressure in the upstream volume of the filter system is increased in increments of 50 mbar. Between each pressure increment the inlet supply pressure is closed and pressure decay is measured over a short interval.
If the measured bubble point pressure is equal to or greater than the membrane specified minimum bubble point pressure, the test is considered a pass. If the measured bubble point pressure is less than the membrane specified minimum bubble point pressure, the test is considered a fail. The measured bubble point pressure for the tested NP6DFLP6G filter, with a specified minimum bubble point pressure of 3180 mbar, was 3950 mbar providing a pass result (Fig 5).
In situ flow kit leak testing
Sterile filtration of the drug product is the last critical process step to ensure the product is fully sterile prior to final filling. Automated in situ (point-of-use) leak testing of flow kits helps to mitigate the potential risk of using damaged flow kits due to transportation, storage, or manual installation.
The flow kits passed all 65 pre-use leak tests performed (Table 8). The leak test sequence completed by the integrated Palltronic Flowstar IV LGR integrity test instrument is shown in Figure 6, demonstrating the precision control by the instrument. The 65 leak tests encompassed the full range of filtration flow kits available on the system. The absence of leak detection in the drug product filtration system filtration manifolds confirms that the system’s flow path was closed prior to the commitment of product, minimizing the risk of contamination.
Table 8. Summary of the 65 leak tests performed using the drug product filtration system for flow kits (leak test) and flow kits including biocontainers (SUS leak test) (Results reported based on leak test type 1 through 4, filter type, and flow path. The test parameters are shown in Table 4.)
| Leak test | Sample number | Maximum acceptable flow (mL/min) | Average actual flow (mL/min) ± standard deviation (σ) | Result |
| Leak test 1 Filter: KA1DFL |
5 | ≤ 1.8 | 0.7 ± 0.2 | Pass |
| Leak test 1 Filter: KA2DFL |
1 | ≤ 1.8 | 0.9 | Pass |
| Leak test 1 Filter: KA3DFL |
1 | ≤ 1.9 | 0.8 | Pass |
| Leak test 1 Filter: KA4DFL |
1 | ≤ 2.0 | 0.6 | Pass |
| Leak test 1 Filter: NP6DFL |
7 | ≤ 2.4 | 0.6 ± 0.2 | Pass |
| Leak test 1 Flow kit connection for single filtration configuration |
3 | ≤ 1.8 | 0.7 ± 0.5 | Pass |
| Leak test 2 Filter: KA1DFL Flush bag: 20 L |
8 | ≤ 48.0 | 6.2 ± 4.2 | Pass |
| Leak test 2 Filter: KA2DFL Flush bag: 20 L |
1 | ≤ 48.0 | 7.9 | Pass |
| Leak test 2 Filter: KA3DFL Flush bag: 20 L |
1 | ≤ 48.0 | 2.7 | Pass |
| Leak test 2 Filter: KA4DFL Flush bag: 20 L |
1 | ≤ 48.0 | 4.9 | Pass |
| Leak test 2 Filter: NP5LEKV Flush bag: 20 L |
1 | ≤ 48.0 | 9.4 | Pass |
| Leak test 2 Filter: NP6DFL Flush bag: 20 L |
7 | ≤ 48.0 | 5.2 ± 2.5 | Pass |
| Leak test 3 Filter: KA1DFL redundant |
5 | ≤ 1.7 | 0.8 ± 0.2 | Pass |
| Leak test 3 Fillter: KA2DFL redundant |
1 | ≤ 1.8 | 0.8 | Pass |
| Leak test 3 Filter: KA3DFL redundant |
1 | ≤ 1.8 | 0.9 | Pass |
| Leak test 3 Filter: KA4DFL redundant |
1 | ≤ 2.2 | 0.7 | Pass |
| Leak test 3 Filter: NP5LEKV redundant |
1 | ≤ 2.1 | 1 | Pass |
| Leak test 3 Filter: NP6DFL redundant |
7 | ≤ 2.9 | 0.5 ± 0.3 | Pass |
| Leak test 4 Filter: KA1DFL Flush bag: 20 L |
1 | ≤ 10.0 | 0.5 | Pass |
| Leak test 4 Filter: KA2DFL Flush bag: 20 L |
1 | ≤ 10.0 | 0 | Pass |
| Leak test 4 Filter: KA3DFL Flush bag: 20 L |
1 | ≤ 10.0 | 3.8 | Pass |
| Leak test 4 Filter: KA4DFL Flush bag: 20 L |
1 | ≤ 10.0 | 0 | Pass |
| Leak test 4 Filter: NP5LEKV Flush bag: 20 L |
1 | ≤ 10.0 | 6.6 | Pass |
| Leak test 4 Filter: NP6DFL Flush bag: 20 L |
7 | ≤ 10.0 | 6.2 ± 2.5 | Pass |
Fig 6. Example of a flow kit leak test sequence completed for leak test 3 (flow path from filter NP6DFL position 1 to filter NP6DFL position 2) The test gas was air with a leak test pressure of 1200 mbar, a maximum acceptable flow of 2.9 mL/min, and a leak test time of 600 s. Each phase of the leak test is labelled with vertical reference lines showing the transition point between each phase. (*pressurization phase, †vent phase after test completion)
The leak test sequence using the Flowstar IV LGR integrity test instrument starts with a pressurization phase to the entered test pressure (e.g., 1200 mbar). After this the stabilization phase starts and is automatically controlled by the integrity test instrument and finishes as soon as the gas flow has stabilized within defined limits. Once the stabilization phase is completed, the measurement phase commences where the gas flow is measured over a set test time. During the first 150 s the integrity test instrument uses values measured during the stabilization phase to calculate the average gas flow. The flow reported on the graph is the average gas flow value for the previous 150 s. If the measured gas flow is less than or equal to the specified maximum acceptable flow limit (mL/min) the test is a considered a pass. If the measured gas flow is greater than the specified maximum acceptable flow limit (mL/min) the test is considered a fail. The measured gas flow for the tested flow kit, with a specified maximum acceptable flow limit (mL/min) of 2.9 mL/min, was 0.4 mL/min so the result was a pass (Fig 6).
Conclusion
Sterile filtration of a drug product is the final process step to ensure a drug product is sterile prior to final container filling. It is critical both from a regulatory and patient safety perspective. The recent update to the EU GMP Annex 1 requires PUPSIT of sterilizing-grade filters as part of the contamination control strategy to mitigate the risk of contamination. The drug product filtration system is optimized for automated PUPSIT, post-use filter integrity testing, and in situ (point-of-use) leak testing of the entire single-use flow kit path prior to drug product commitment, reducing the risk of product contamination for large scale GMP applications. Additionally, the ability to completely automate in situ priming, wetting, and flushing of filters and the option to perform pressurized static soaks on filters to further improve filter wetting, provides a reliable and robust process to perform filter integrity testing on filters sized from KA1 to NP6 and in either single or redundant arrangements. This further mitigates the risk of false failures.
Further reading
CY49222-25APR25-AN