In a previous blog post, we ended with a definition of a high-performing particle filter as one that may be capable of removing 99.98% of particles greater than a defined size. We did add the caveat that this may only hold true for non-deformable particles, and that the filtration matrix itself needs to be unaffected by the process itself. Such filters are often described as being “absolute,” as their performance can be rigorously tested using standard testing criteria.
Removing bacteria and viruses—filter performance in critical processes
This level of performance is, however, far from adequate for critical filtration processes tasked to remove undesirable species such as bacterial and viruses. These are common, potential contaminants within a biological process. Instinctively it would be reasonable to just look for a more appropriate definition and measure of performance. This is largely true, but we will challenge this shortly.
The classical definition of the performance of a sterilizing-grade filter is one that is capable of producing sterile effluent when challenged with 107 0.2 micron organisms per cm2 of filtration media. This differs from particulate-rated filters in two ways. Firstly, zero breakthrough when challenged at 107 equates to > 99.99999% retention. This is many orders of magnitude higher than the best particulate-rated filters. Secondly, this absolute measure of performance is per unit area and therefore increases as the size of the filter increases. Testing criteria therefore need to be specific to the filter being used. This level of performance is proportionate with the criticality of the typical applications. Drugs that may be contaminated with bacteria are clearly unsafe and sterile filtration is just one in a long list of controls necessary to assure safety.
Just as for particulate filters this performance also assumes two things. Firstly, that the bacteria used in the challenge adequately represents those that could reasonably be expected in the process, both in terms of size and physical characteristics, such as deformability. Secondly, it assumes that the filter will perform in exactly the same way under the real process conditions as for those used in the generic test. Both of these assumptions have driven the pharmaceutical industry to carefully perform process-specific validation to better control any residual risk.
FDA definition of a sterilizing-grade filter
This is also reflected in the Food and Drug Administration’s (FDA) definition of a sterilizing-grade filter. Prior to 1987 this mirrored our definition above and defined a sterilizing-grade filter as one that retains 107 0.2 micron organisms per cm2 of filtration media. Often current definitions simply define such a filter as one that is capable of removing all process organisms under the process conditions. This is still commonly simulated by the same conditions, but there are exceptions. For cultures, notably these most often refer to the organism used. Some common contaminants such as R.picketii and Mycoplasma species are known to penetrate traditional filters designed to function as sterilizing-grade filters. When these are known contaminants not controlled by other means, filters such as those that carry a 0.1 micron label are often used to ensure retention.
Finally, some risk factors for a reduction in performance (penetration) come from the process fluid itself and can be exacerbated by some process conditions. Attributes such as high viscosity and the presence of detergents or lipids have been seen to be problematic for some filter constructions. Optimal filter selection and careful selection of the process parameters ultimately control these risks, with the actual performance then being rigorously documented with process-specific validation data.
When it comes to the necessary performance of a filter, if there is any doubt, it is always wise to speak to an expert in filtration.