Filter validation: Product-specific studies that need to be performed by the filter user
Next to the information that can be obtained by the filter manufacturer, there are studies that need to be performed with the product and – if possible – under process conditions. The filter user, meaning the pharmaceutical company, is responsible for these studies. Apart from the studies that have been implemented in-house during development, a cooperation with the filter manufacturer is suitable for that. A quick overview can be found in the following paragraphs.
A viability test provides the information whether the product in question has a bactericidal effect on the bacteria used for the bacteria retention test (usually Brevundimonas diminuta). This test is necessary to construct the following bacteria retention test in an expedient way. For the test, a defined number of B. diminuta is incubated in product solution for the duration of the entire filtration process. At the end, the number of bacteria shouldn’t be lower than before. If the product is supposed to have a bactericidal effect, there are several possibilities to modify the bacteria retention test. For example, it’s possible to shorten the incubation time with bacteria for as long as it’s tolerable for them, in combination with a previously performed pre-conditioning of the filter (without bacteria) with the product for the rest of the time, to cover the entire duration of the filtration process.
The bacteria challenge test (BCT / BCA) according to ASTM F838-15a is supposed to prove that the filter is able to provide a sterile filtrate for the entire duration of the filtration and isn’t affected in its functionality by the product solution. According to the recommendations by the FDA worst case production conditions have to be used for the test. This includes among others the maximum filter use time and pressure (please check also our corresponding blog article). Additionally, instead of using Brevundimonas diminuta for ASTM F838-15a, a bioburden isolate can also be used if smaller or better than B. diminuta. In case of unimpaired viability direct inoculation of the bacteria into the product solution is the preferred method. If this is not possible, alternative options are presented. It is not necessary to test exactly the same filter as the one used in production, but the specific type of filter membrane. The integrity of the filters to be tested must be checked and the results have to be part of the report.
A product-specific bubble point test or a diffusion test as filter integrity test is used to prove that the filter wasn’t damaged in any way during production that could have led to an impairment of bacteria retention. It’s expedient to assess the integrity before as well as after filtration in the case of a sterile filtration for final filling. A determination before-hand gives a go-ahead for filling or allows a filter change, a downstream determination is proof that the process was carried out correctly and is required by the FDA. For verification purposes it makes sense to list a defined number of integrity values that have been identified under process conditions.
It doesn’t make sense to study the release of particles from the filter by only rinsing with water, as performed by the filter supplier as the contact of product and membrane might lead to particle generation. This aspect has to be taken into account accordingly. Such a test can be performed for example at laboratory scale by circulating filtration using the same filter membrane. The corresponding compendial requirements are considered to be the acceptance criteria (USP <787> / <788>, Ph. Eur. 2.9.19 and JP 6.07). The results of the process validation can be listed additionally.
The possibility that components of the product composition can adsorp to the filter membrane has been described repeatedly (amongst others, Zhou 2008, Mahler 2010, Lei 2013, Pillai 2016). This applies particularly to the adsorption of the protein itself and surfactants like polysorbate 20 or 80. Since changing the product composition can have, amongst others, dramatic effects on the potency and stability of the drug, it’s necessary to analyze the product’s adsorption behavior by appropriate tests. According to PDA report 26, they are generally conducted during process development as small scale studies and confirmed later at-scale. This can be done using the data from the process validation, for example.
Multiple / re-filtrations generally don’t take place in standard processes. Correspondingly, testing them isn’t a requirement by the PDA, yet seems to be expedient considering the possibility of production interruptions taking place. For example, a sterile filter failing the filter integrity test directly before final filling would make a return the product solution necessary. After a temporarily storage and exchange of the defective sterile filter, the same product solution has to be filtered once more. In this case it’s useful to have information whether the increased shearing stress caused by the repeated filtration has or hasn’t had a negative impact on the product quality. The protein content for example can be studied accordingly, using a defined number of circulating filtrations, and can then be compared with the release specification. Single use is also recommended by the FDA, but if justified, repeated use is also allowed. In this case, the maximum number of refiltrations must be validated.
Pharmaceutical companies are responsible for guaranteeing the safety of their drugs. This applies for possibly extractable substances like extractables and leachables as well. This is the reason why they have to conduct their own extractables / leachables studies if the data provided by the filter manufacturer isn’t sufficient for a risk assessment. There’s no industry standard for the implementation of such studies at the moment. There are recommendations for conducting them in the PDA report 26, but it seems to allow great variability. For example, it names multiple methods for analysis (ranging from RP-HPLC to GC-MS to TOC), but doesn’t specify preferences. The mentioned methods are vastly different, considering that quantification as well as identification should be performed. The extractables team of the BioPhorum Operations Group (BPOG) has started an initiative to change this. Accordingly, they have published a paper in which they give recommendations for conducting extractions, for the applicable analysis methods (LC-MS, GC-MS and IPC-MS) and for information to be put in the concluding test report. A wide support of the pharmaceutical industry can be assumed, because 18 BPOG member companies from more than 26 locations have contributed to this paper. Organizations setting standards in the industry (like ASTM and USP) are supposed to gain access to the developed testing protocol, so that they can develop a new standard by using it.