Sterile Filtration and PUPSIT - Questions and Answers at a Glance (Part 2)
Sterile filtration is a key step in the aseptic manufacture of pharmaceutical products. It serves to remove microbiological contamination and ensure the sterility of the end product. A closely related topic is PUPSIT (Pre-Use Post-Sterilization Integrity Testing), which checks the integrity of the filter after sterilization but before the actual filtration. PUPSIT is not an easy system to implement. Many companies face technical, spatial, and organizational challenges. Existing systems in particular often cannot be easily adapted for PUPSIT implementation. While some manufacturers have already successfully implemented PUPSIT, others are still experiencing difficulties or need to conduct a thorough risk analysis to justify an alternative approach or the complete avoidance of PUPSIT.
To understand the relevance and complexity of this topic, it is essential to consider the regulatory background. The revised Annex 1, which came into force in 2023, introduced significantly more detailed and stringent requirements for the manufacture of sterile medicinal products. Despite its importance from a regulatory perspective and in terms of patient safety, implementing PUPSIT in practice can be complex. Many systems in aseptic production environments were not developed in accordance with the new Annex 1 requirements. Significant redesigns, investments, and operational adjustments are often necessary to adapt these systems to enable reliable and validated PUPSIT procedures. In addition, space constraints in cleanrooms, the need for additional equipment, and changes to SOPs and training programs add to the burden.
We therefore collected questions about sterile filtration from seminar participants, discussed them with an expert, and compiled them. Read the second of three sets of questions below.
1. Where must the sterile filter be located?
- In the A area on the filling line
- Under an LF in the B area or an LF in the C area
- In the B area or C area
Recommendation
Vienna, Austria3-6 November 2026
Contamination Control & Workshop Risk Assessment in Contamination Control
From my point of view there is no clear rule, but it must be justified. What makes it clear is the process in which you apply the filter. For example, if your filter operates in a closed system and is sterilized in the closed system and maintained sterile (assured by integrity of the piping downstream of the filter outside of grade A) in the closed system, the filter can even be located in zone D (new in C). In order to assure low bioburden prior to sterile filtration (for example open valves upstream for degassing the filter) it can also make sense to locate the filter in zone C or B. If you have an aseptic assembly in an open system, the filter must be located and assembled in zone A, validated by aseptic process simulation. I think the rationale for location should be assessed in the Contamination Control Strategy.
2. How can the installation for the filter integrity test be technically implemented?
- Fixed piping vs Hose connections
Fixed piping has the advantage to make the testing more robust and stable. Hose connections are not so pressure stable and very often you need to operate above 3500 mbar during FIT which the connections sometimes cannot withstand. - Where is the test device located?
The outer surfaces of new generation devices are able to withstand H2O2 decontamination. So, you could bring it into the cleanroom. However, the inner surfaces according to my experiences should not be decontaminated and therefore remain a "Black box". Therefore, I would avoid to transfer them into a cleanroom. On the other hand they should be located as close as possible to the filter since for most devices the upstream volume plays a role in the test accuracy. If the connection is too long between the device and filter, you may face test fluctuations or no stable results. - Where can the PUPSIT rack be located?
- In the A area on the filling line
That depends on your process. If you are using single-use systems and you don't want to be worried about integrity of the system downstream the sterilizing filter, I would put the whole assembly in grade A with the disadvantage that you need a connection with your test device in grade A. In that case you also need to protect your area from non-sterile gas coming from the test device. So, it is also not optimal and you need a lot of place and aseptic manipulations in grade A. - In the B area or C area
If you are operating your filter in a closed system and sterilize it there and you protect your product filter during integrity test with an additional gas filter to assure low bioburden during testing, I think C should be sufficient. Since you are venting the filter as well, an additional protection by LF could be helpful, although venting happens under pressure. Therefore, any contamination risk due to venting can be considered as low.
3. What must a risk assessment include if no PUPSIT is carried out?
A clear description is listed in Annex 1, chapter 8.87. In general you have to assess that a masking effect which might alter the post use integrity test result is excluded. One possibility is to demonstrate that there is no fouling effect caused by your product to be filtered. Also, before you make this risk assessment you should assess why you cannot perform PUPSIT.
4. What is your opinion on the requirement in Annex 1 regarding the use of 2 sterile filters?
Using a redundant filter gives you additional safety margin, especially from a business perspective to assure in case of integrity test failure of one filter to save the batch. On the other hand it comes with additional losses due to adsorption etc. So, depending on the process robustness and batch costs, an additional filter could make sense. However, with a proper prefilter and a controlled closed system with smooth filtration conditions, I believe 1 filter is perfectly sufficient.
5. In the case of sterile filtration prior to sterilization in the final container by autoclaving, is there also a requirement for renewed filtration (after 24 h storage time)? Or can renewed bacteria-reducing filtration be dispensed with here?
Since the sterilization is in the final container, we are not applying sterile filtration but low bioburden filtration. So, in theory the paragraph 8.94 of Annex 1 is not applicable. Also, you would not need to validate this filter for bacterial retention. What you need to assure is a low bioburden (10 cfu/100 ml) prior to final sterilization. Hence, your filtration needs to guarantee this. Nevertheless, using a filter more than a working day you still need to consider other process constraints such as endotoxins which might come from bacteria being held back in the filter, filtration performance may decrease, particles could increase (compatibility) etc. In other words, you would also need to validate this usage time from a quality attributes point of view, which includes low bioburden as well.
Recommendation
Vienna, Austria3-5 November 2026
Contamination Control - Requirements, Measures and Strategies
6. What options do you see for approval if the forward flow test was passed at the start of production but not after the end of production, but sterilization by autoclaving takes place in the final container? Particles etc. are inconspicuous?
Since the sterilization is done in the autoclave, bioburden prior to sterilization, the success of autoclaving and sterility test results are major contributors to batch release from a microbiological perspective. The filter prior to autoclaving is from my point of view a pre-filter which needs to assure low bioburden. Hence, sterile filtration is not applicable and by definition there is no requirement for performing an integrity test on a pre-filter. Hence, the failed result post-use would trigger a closer investigation on bioburden. If the function of the filter shall also reduce particles, then particles need to be verified as well. Finally, the investigation is determined by the function of the filter.
7. What is your opinion on the theoretical possibility that e.g.: a protein-containing, viscous product could close an initial leak in the filter during backfilling?
The PDA has published a study showing the masking effect on bubble point tests. In summary, there must be a huge fouling happening which has a significant influence on filtration performance before masking effects have been seen in these studies. More details can be seen in the PDA study.
In summary
Sterile filtration and PUPSIT are integral components of a robust aseptic production strategy. While the new EU GMP Annex 1 defines clear requirements, there is still room for scientifically sound, risk-based decisions - especially for clinical batches or small volumes. It is important that every decision is documented in a comprehensible manner, justified in relation to the process, and embedded in the contamination control strategy (CCS). This is the only way to ensure long-term regulatory acceptance and product safety. Read the other sets of questions in part 1 and 3.
About the Author
Matthias Schaar has been working for Novartis in Switzerland since 2007. He began building up his knowledge in microbiological quality assurance and quality control. Currently, his main focus is on supporting the validation team and routine manufacturing in the context of sterile filter validation and its application.
