A sterilizing-grade filter is one that is capable of retaining a minimum 1 × 107 cfu of B. diminuta (ATCC 19146) per square centimeter of effective filter area when tested in accordance with ASTM F838-05 (2013), Standard Test Method for Determining Bacterial Retention of Membrane Filters Utilized for Liquid Filtration (2).

The designation “sterilizing grade” implies a sterilizing action only if other conditions are met, including the integrity test specification established by the filter manufacturer and validated by the user (see Validation, below).

Sterilizing-grade filters typically are microporous membranes that have nominal pore-size ratings of about 0.2 µm. These membranes are fabricated with various materials, have relatively narrow pore-size distributions, and can be integrity tested. The integrity test results can be correlated with microbial retention. Membrane filters that have a nominal pore size of 0.45 µm can be validated to produce sterile filtrates under some conditions; for example, some liquids require high differential pressures to achieve useful flow rates, and these pressures are not suitable for use with 0.2-µm–rated filters. When manufacturers use 0.45-µm filters, they should ensure particularly stringent control of presterilization bioburden.

Microporous membranes remove microorganisms from the liquid by two primary mechanisms: sieve retention, which relies on physical blockage of particles that are larger than the pores they encounter, and adsorption, which is a charge-related phenomenon whereby particles are bound to the membrane surfaces. Both mechanisms should be considered during the development, qualification, and validation of sterilizing filtration processes.

Sieve retention, for the most part, is independent of filtration conditions and the microbial challenge level, but the composition of the liquid, temperature, and filtration conditions can affect sieve retention under certain conditions.

Adsorption is a charge-related phenomenon that is influenced by the composition of the membrane, the properties of the filtered liquid, the filtration conditions, and the number and type of microorganisms present in the liquid.

The number and type of microorganisms and other particles present in the material to be filtered can affect retention. Sieve retention implies that every microorganism is larger than the largest pore in the filter. Microorganisms and filter pores are not uniform in size, and some microorganisms may be smaller than the largest pores. Even if a microorganism is smaller than the pore it encounters, it may be retained if a site is available and the filtration conditions are conducive to adsorption. Retention probability is related to the number of microorganisms in the upstream bioburden (4,5).

Microporous membranes consist of a polymer matrix that is penetrated by interstices commonly referred to as pores. Compared with those in depth filters, the pores in microporous membranes have a relatively narrow size distribution. The size, number, and shape of the pores determine the filter’s retention capabilities. With the exception of the pores in track-etched filter membranes, pores are not cylindrical; they are made up of a series of pseudopolyhedral structures with varying internal diameters (6).

The pore size, pore-size distribution, and membrane porosity are a function of the manufacturing process. Careful design and control of that process are necessary to ensure that the resulting membrane has the desired integrity test value, microbial retention capability, and uniformity.

Microorganisms have a variety of shapes and sizes. If the microorganism encounters a membrane pore that is smaller than its smallest diameter, the microorganism likely will be captured by sieve retention. If, however, the microorganism is smaller than the pore it encounters, it may be retained by adsorption if the residence time, electrostatic charge, pH, fluid chemistry, and membrane material are conducive to adsorption.

Some genera of microorganisms are deformable, so that at high-pressure differentials or flow rates a microorganism may be forced through a filter pore that is only slightly smaller than the organism (7). Mollicute bacteria lack cell walls and thus are small and pliable enough to pass through filter membranes under certain conditions.

Grow-through (i.e., passage of microorganisms through a filter as a function of time) may result from one or more scenarios. The microorganism may penetrate as it multiplies: a parent cell divides into two smaller daughter cells that negotiate pore passageways. Penetration can occur with time, because the increasing number of microorganisms overwhelms the few larger pores that are encountered. However, because of the limited time periods typically involved, the controls on bioburden, and the often limited availability of nutrients, grow-through is considered a rare phenomenon in pharmaceutical processes (see Monitoring of Bioburden 1229.3) (8).

Several factors related to the filter matrix can affect microbial retention. These include the material from which the filter is made, the pore size and pore-size distribution, whether the membrane is isotropic or anisotropic (i.e., whether membrane pore structure is uniform from face to face or “tapers” from one face to the other), membrane thickness, and whether the filter consists of single or multiple layers.

The membrane material is especially important if adsorption is a significant mechanism in a particular filtration scheme. For example, polyamide exhibits stronger microbial adsorption than does cellulose ester (3).

Microporous membranes that have a relatively wide pore-size distribution are less likely to retain microorganisms, particularly at high challenge levels, than membranes of comparable pore-size ratings with narrower pore-size distributions. This relates to the probability of a microorganism’s encountering a pore larger than itself.

Thicker membranes generally are more retentive than thinner membranes of the same type and pore-size rating, owing to the higher probability of entrapment or adsorption within the pore structure because of the increased distance that bacteria must travel in thicker membranes. This distance favors entrapment, as does increased residence time within the pore, which favors adsorptive retention (3).

Multi-layered membrane filters exhibit a higher probability of retention than do single-layered membrane filters of the same thickness because the small number of largest pores is the factor affecting retention and the probability of a large pore in one layer being congruent to a large pore in the adjoining layer of a double-layer filter is negligible (3).

The composition of the filtered solution can adversely affect the membrane material if an incompatibility exists, causing damage to the membrane and affecting both retention and physical integrity, unless detected before the filter is selected for use. In addition, if adsorption is a significant retention mechanism, then solution properties such as pH and the presence of surfactants become important.

Surface charge and ionic strength are important variables. Bacterial and membrane surfaces in aqueous media are negatively charged, resulting in repulsion. The repulsive force is balanced by attractive forces, which include hydrophobic surface energy minimizing forces, operative only over short distances, and hydrogen bonding. High ionic strength allows the surfaces to close—because of discharge through the electrolyte—to the point where hydrophobic adsorption can occur (3). Also, high ionic strength can draw water out of the cell, reducing its size, which may lower the probability of retention, depending upon the composition of the prefiltration bioburden. Surface tension and the presence of surfactants influence retention. Adsorption of surfactant by the filter and the microorganism creates repulsion, leading to a decreased probability of retention. Surfactants in concentrations as low as 0.05% have been shown to inhibit adsorption and decrease the retention of latex spheres (9).

Differential pressure, flow rate, and temperature are among the factors that can affect microbial retention. Hydraulic shock should be avoided, not only because it affects pressure differential and flow rate but also because it can damage the filter. Flow rate is proportional to differential pressure, and higher flow rates reduce adsorption because the contact time is reduced (10).

Microbial retention may be reduced at higher temperatures when these result in higher flow rates due to decreased solution viscosity. Temperature effects are not significant in filters where sieve retention is the primary removal mechanism.

Filtration efficacy can be defined in terms of a log reduction value, which is the logarithm of the quotient produced by dividing the upstream challenge population by the recovered downstream population.

The log reduction value is influenced by the number and size of the challenge microorganisms, the filter design and membrane polymer, the filtration process parameters, and the properties of the solution.

To determine the specific log reduction value of a filter, the challenge test should permit some passage of the test microorganism through the filter in order to produce a denominator. Sterilizing-grade membrane filters should not permit passage of the specified challenge microorganisms for that filter rating. For this reason the log reduction value of sterilizing-grade filters is described as equal to or greater than the log of the challenge population.

As noted, microbial retention in sterilizing filtration relies on a combination of sieve retention and adsorption. Validation of sterilizing filtration therefore requires determination of the effect of the liquid on the filter, determination of the effect of the filter on the liquid, and demonstration that the filter can consistently yield sterile solutions under the intended conditions of use. The liquid to be filtered can affect the pore structure of the membrane, can have electrostatic properties different from the standard challenge suspension used to establish integrity test specifications, and can change the size and shape of the challenge microorganisms. Factors that should be considered when developing a sterilizing-filtration validation protocol include the surface tension, pH, temperature, and osmolarity of the liquid to be filtered; the compatibility of the material or solution components with the filter itself; the pressures, flow rates, and hydraulic shock likely to be encountered; and the maximum filtration time and volume to be filtered. The effect of sterilization (steam, radiation, or gas) on the filter’s retention capability also should be considered.

B. diminuta (ATCC 19146) is used as the challenge organism unless it is not viable in the liquid to be filtered. Viability studies should be used to confirm that the liquid has no adverse effects on the challenge organism. If the challenge organism is viable in the liquid to be filtered, the liquid should be inoculated to achieve a challenge level of 1 × 107 cfu/cm2, and the filtrate should be evaluated for the presence of the challenge organism. If B. diminuta is not viable in the liquid, several options are available, and analysts can (1) modify the liquid to ensure the viability of the challenge organism (e.g., adjust pH or remove the bactericidal component), (2) reduce the exposure time to ensure that the challenge organism remains viable, or (3) change the challenge organism from B. diminuta to one that has been isolated from the liquid to be filtered. These studies should employ production process pressure differentials or process flux values as appropriate.

If possible, the liquid to be filtered should be used because in some instances the challenge organism has penetrated a filter in contact with the liquid but has been retained by the same filter when inoculated into a surrogate fluid (11).

Three different lots of filter membranes should be used for the microbial retention studies. The membranes should have preuse integrity test values that are near the filter manufacturer's specification in order to minimize the possibility that production filters will fail to meet the integrity test value established during the validation exercise. Successful microbial challenge studies result in no microorganisms detectable in the filtrate. The sterilization process for the filter, its housing, and associated equipment should be validated. The filter should be sterilized within its housing rather than relying on aseptic assembly following sterilization of the filtration system components. The assembled filtration apparatus can be steam sterilized in an autoclave, using a vacuum cycle, with particular attention to the orientation and wrapping of the housing and any associated tubing to allow condensate drainage and steam penetration. Validated steam sterilization-in-place cycles also can be used.

Sterilization methods and cycles should be carefully chosen and designed to preclude damage to the filter. Pre- and poststerilization integrity testing can be used to confirm that the sterilization procedure does not change the integrity test value and to demonstrate that the sterilization process does not damage the filter.

Finally, the scale of the validation approach should be considered. Two approaches are common. One uses a small section of membrane material, typically a 47-mm disk, to represent the filter. This approach validates the microbial retention capability of the membrane. A second approach uses the intended filter configuration, typically a 10-inch cartridge in its housing. This approach validates the filter system performance in addition to the retention capability of the membrane material. Analysts should consider the volume of the test liquid and operational considerations when they choose which approach to use.

Integrity testing can be used to show that a filter has the correct pore-size rating, is installed properly in its housing, and has not been damaged by the process used to sterilize it. These integrity test methods generally rely on detecting gas flow caused by pressure differential across a wetted membrane. Figure 1 shows the relationship between gas flow and pressure differential for high-surface-area and low-surface-area membranes (a 47-mm membrane disk can be considered low surface area, and a multicartridge array of 10-inch cartridge filters can be considered high surface area).

A filter that successfully passes an integrity test based on specifications developed during an effective sterilizing filtration validation study is capable of producing a sterile filtrate.

Integrity test methods for microporous membrane filters include bubble point, diffusive flow, and pressure hold. Membrane type and size and the particular filtration process influence the choice of an appropriate integrity test method. The sensitivity of the bubble point test decreases with increasing filter area because of increasing diffusive flow. Bubble point, diffusive flow, and pressure-hold integrity tests are used for hydrophilic microporous membrane filters. As shown in Figure 1, diffusive flow occurs at pressures below the bubble point, and bulk flow takes place above the bubble point. The bubble point marks the transition between diffusive flow and bulk flow. The exact bubble point is difficult to detect in high-surface-area filters, because the high membrane surface area generates significant diffusion. Integrity testing may be performed manually or by automated equipment designed specifically for that purpose.

The decision not to perform preuse, presterilization integrity testing should be based on a formal risk assessment.

Preuse, poststerilization integrity testing may create an unnecessary risk for microbial contamination of the filter and associated downstream tubing and equipment. Preuse, poststerilization integrity testing is unnecessary if effective validation studies have demonstrated that the process for sterilizing the filter does not affect the integrity test value of the filter.

Postfiltration integrity testing should be conducted to ensure that the filter was not damaged during the filtration process.

The filter manufacturer is responsible for ensuring that the filter production process has been validated and is well controlled and that the sterilizing-grade filters meet the requirements of ASTM F838-05 (2013). The filter manufacturer determines the integrity test specification for the filters, usually adding a safety factor to ensure that each filter will meet that specification. The filter manufacturer conducts extractable and leachable studies to ensure that the filter does not release objectionable levels of these materials into the solvent systems typically employed in pharmaceutical manufacturing. The filter manufacturer conducts cleanliness tests to assure that the filter does not adversely affect the USP particulate requirements of the product. The filter manufacturer provides technical support and troubleshooting advice if the filter user encounters a problem.

The filter user is responsible for establishing microbial retention at the validated integrity test value, establishing microbial retention in the liquid to be filtered, and validating the use and sterilization of the filter and housing. The filter user is responsible for determining that the filter is not additive or extractive to the extent that the filtered liquid is adversely affected.