Adjuvants are becoming more common in vaccine and other drug formulations to increase therapeutic response. Some of these substances, however, are close enough in size to bacteria that they are unable to pass through sterilizing-grade filters. Others have low surface tension that can reduce a filter's bacterial retention. As a result, adjuvants can cause premature plugging of filter membranes and reduce filter capacity. Pharmaceutical Technology spoke to several industry experts to gain insight on resolving these technical challenges.

IMAGE: COURTESY SARTORIUS STEDIM

Reducing blockage

Bromm (Sartorius Stedim): One possibility for filter manufactures to deal with these challenges is to develop sterilizing-grade filters that specifically address these needs. According to our experience at Sartorius Stedim, highly asymmetric membranes, such as polyethersulfone (PES) membranes provide higher flow rate and capacity for such type of formulations compared with symmetric membranes. According to practical experiences, the use of a heterogeneous double-layer membrane construction provides total throughput advantages compared with single layer membrane filters. The prefilter (i.e., upstream layer) protects the final membrane (i.e., downstream layer) from premature plugging. Of high importance is to find the optimal graduation between two membranes. Studies with model solutions and test results with actual formulations in field tests have demonstrated that the combination of a finer prefilter membrane with the final 0.2 µm membrane achieves better results compared with combinations with a coarser prefilter membrane for adjuvants applications.

Pharmaceutical manufacturers should carry out filtration studies to compare the performance of different membrane materials and construction principals of filters to find out the optimal solution for their specific formulation. Furthermore, the use of prefilters should be considered in such studies to protect final sterilizing-grade filters effectively and to reduce costs and filtration time. These studies can be used to determine the optimal parameters for the filtration process, such as differential pressure or temperature. Increasing the temperature can enhance filterability depending on the stability of the solution at higher temperatures. The same filter-selection process may be applied for other protein therapeutics or vaccines.

Martin (Pall): Pharmaceutical manufacturers can reduce filter plugging by optimizing formulation and process conditions for desired filter life, along with selection of appropriate filters with suitable capacity. Filter manufacturers can provide technical support for this process by conducting feasibility (filterability) trials, selecting appropriate filter-media grades, sizing of filter cartridges or capsules, as well as ultimately applying that knowledge to the development of new filters capable of providing greater capacity.

Process parameters such as pressure, temperature, and flux (i.e., flow per unit area) can have a large impact on filter throughput and capacity. For example, with complex plugging biological fluids, performing the filtration in a constant flow mode, increasing pressure differential to maintain flux rather than operating under a constant pressure mode can often have a positive impact on filtration throughput (capacity). Process temperature can also have an impact but is product-dependent and needs feasibility (filterability) tests to determine whether an improvement can be achieved through modification. Optimizing these performance variables is an acceptable (and recommended) technique to reduce the risk of premature blockage for vaccines or protein therapeutics.

Koklitis (3M): The plugging of membrane filter systems by adjuvants is particularly undesirable when the process step has been validated to provide sterility assurance. The risk of filter plugging can be reduced by careful control of the filtration operating conditions, such as inlet pressure and optimum flux. The lifetime of the sterilizing-grade filter membrane will be greatly determined by the particle load in the process feedstream and the capacity may be extended with a prefiltration stage. A prefilter rated at 0.45 µm will remove larger emulsion micelles or liposomes which might ordinarily plug a sterilizing 0.2 µm membrane. Another option is to consider a 0.2 µm-rated bioburden reduction membrane as a prefilter. This can be of the same material as the final sterilizing membrane to simplify validation and may be effective for removing larger particle sizes from the process stream as a result of its pore size distribution. The prefiltration system selected should be sized appropriately to meet the demands of the process stream to minimize the expense associated with the final sterilizing membrane stage. When emulsions are used, the pharmaceutical manufacturer could investigate an adjuvant formulation with a sufficiently small particle size to make it filter-sterilizable.

Some studies with oil-in-water emulsions have shown that increasing the pressure drop across the membrane can increase filter capacity. The coating of bacteria on the membrane with emulsion has been considered to contribute to bacterial penetration. In such instances, higher bacterial retention may be achieved by increasing the temperature if cold conditions are currently used. However, the reasons for adopting cold filtration (e.g., to maintain protein stability) may present an obstacle to implementing a change.

Powell (Asahi): This is rather hard to answer because the blocking can occur due to a wide range of issues related to product use and conditions such as pH, conductivity, protein concentration, viscosity, temperature, membrane incompatibility with what is in the adjuvant, and so forth. The best solution would be to better characterize the adjuvant, the product, and the combination to find the most stable and best filter condition possible, where material is not precipitating, too viscous, too high a concentration, and/or at the early stage or "edge" of aggregation and the filter type where the adjuvant's oil, if present, does not bind to or change/damage the membrane itself.

There are really two choices: the brute force method, where one throws more membrane at the problem, or the better method, which would be to choose the right adjuvant for the job and choose conditions that fit into a high stability window of operation for the API. Another more sophisticated solution to these kinds of clogging problems is to use a cascade of filters that end in the final desired porosity. The upstream filter(s) can act as prefilters to increase final filter capacity.