Sterility Testing: Membrane Filtration or Direct Inoculation?

The choice between membrane filtration and direct inoculation in sterility testing depends on the product, the applicable compendial reference and the method’s ability to reveal contamination under the actual test conditions. Filtration is the route to assess whenever the nature of the product permits it; direct inoculation is not a shortcut around interference studies. Both approaches require demonstrated suitability for the matrix and an interpretable result. [1–3]

This article helps build the rationale for selection: what to check, which evidence to retain and when unresolved questions prevent a decision. The decision matrix and two cases are original GuideGxP material. They are not an analytical procedure and do not replace the approved method or the current compendial text applicable to the product.

1. What the test demonstrates—and what it cannot

A satisfactory result indicates that contaminating microorganisms were not detected in the sample examined under the test conditions. It does not, by itself, demonstrate the absence of contamination in every unit of a batch. EU GMP Annex 1 treats the finished-product test as the last of a series of controls: it cannot compensate for failures in product and process design, procedural or validation parameters. [1,3]

This distinction changes the starting question. It is not enough to ask which technique most easily produces a negative result. Ask which route preserves detection capability, meets the applicable requirements and leaves the test conditions reconstructable. A clear sample does not prove the absence of inhibition; a closed system does not automatically establish that handling was correct.

Sampling remains another essential decision, to be defined using applicable references and process knowledge. This article does not propose universal numbers of containers, quantities per medium, times or temperatures. If the laboratory has not established which method and sampling plan are authorised, equipment selection is premature.

2. Establish the reference before comparing techniques

The selection file should identify the product, formulation, presentation, market, registered or approved method and version of the text used. This is a GuideGxP organisational recommendation to prevent the word “compendial” from concealing different references. An established internal procedure does not guarantee that it still reflects the current formulation and conditions.

ICH Q4B Annex 8 addresses interchangeability of Ph. Eur. 2.6.1, USP <71> and JP 4.06 within specified conditions and regional implementation. It does not permit unrestricted substitution or remove the need to demonstrate product suitability. Testing conditions for medical devices are outside the scope of the cited ICH recommendation. [2]

The technical principles presented here draw on the WHO International Pharmacopoeia chapter, thirteenth edition, 2025, which was read in full. Current licensed full texts of USP and Ph. Eur. were not consulted; the laboratory must check them where applicable. WHO supports understanding of the reasoning and is not declared automatically equivalent to the authorised method for every product.

3. Filtration: separate the sample while preserving detection

In filtration, the system retains microorganisms present in the examined portion on the membrane while the liquid passes through the filter. The membrane is then placed under the specified culture conditions, through aseptic transfer or introduction of medium into a compatible system. The WHO chapter requires a membrane with established retention capability and equipment that permits aseptic operations. [1]

A potential advantage is separating retained material from matrix components that would hinder growth, using treatment specified and demonstrated to be suitable. This does not mean that every residue is eliminated. The membrane, diluent, any neutralising substances and washing conditions belong to the method being assessed; they are not a list of freely interchangeable accessories.

Assess filterability in the actual preparation intended, compatibility with materials and solvents, the quantity to be examined and the feasibility of the route. Poor passage cannot automatically be resolved by increasing pressure, dilution or washing. Compendial constraints must be respected and the change shown not to compromise the test. WHO also sets limits on washing: “wash more” is not a universal solution to inhibition. [1]

The commercial category of a formulation does not decide the issue. WHO also covers oily preparations and those soluble in appropriate solvents, under defined conditions. “Oily product means direct inoculation” is therefore premature. The file should distinguish a genuine incompatibility from a lack of data on the configuration considered.

4. Direct inoculation: control what remains in the medium

In direct inoculation, the specified portion of product is introduced into the culture medium. The matrix therefore remains part of the test system and can affect both growth and its observation. The technique may be appropriate when filtration is impracticable or according to the applicable method, but those effects require explicit assessment. [1]

Evaluate antimicrobial activity, product–medium interaction, permitted dilution or neutralisation and the mixture’s behaviour during observation. A positive control without product demonstrates a property of the medium and control conditions; alone, it does not establish that the same growth can be detected in the presence of the sample.

Formulation-related turbidity may prevent a change attributable to growth from being distinguished. An uninterpretable reading must not be converted into a negative result. The method needs to specify the applicable handling of that situation and the evidence to record. Improvised steps should not be added at the end simply to obtain a convenient reading. [1]

The organisational comparison should also consider containers, space, manipulations and incubation capacity under the required conditions. These factors make a solution practicable, but do not take precedence over method suitability. A route that is easier for the operator is not necessarily more informative about the product.

5. Method suitability: demonstrate the absence of relevant interference

The suitability study challenges the method’s ability to reveal reference microorganisms under the intended conditions, including the presence of the product or its treatment. Comparison with the relevant control without product allows assessment of whether antimicrobial activity is absent or has been adequately eliminated. A generic statement about the filter material is insufficient. [1,2]

Define which matrix properties could interfere with the test and which parts of the study address each concern. Record the composition or variant represented, preparation, system, materials, media, controls, observations and any changes. GuideGxP recommends maintaining a readable link between the hypothesised risk, the test performed and the resulting evidence.

If the expected growth is not demonstrated under the required conditions, a lack of growth in an ordinary product sample does not make the method suitable. The interference must be addressed and suitability reassessed according to the applicable reference. WHO provides for this assessment for a new product and when experimental conditions change. Do not extend a conclusion to different matrices without adequate justification. [1]

Also distinguish medium growth promotion, method suitability and the product test result. They are connected evidence, but answer different questions. Calling all of them a “positive control” makes it difficult to understand what was actually verified.

6. An original matrix for the selection rationale

This table is a GuideGxP tool, not a regulatory ranking. For each row, note the applicable reference, available evidence and remaining question. If a critical condition is unresolved, the choice remains a proposal requiring verification.

A useful decision goes beyond “membrane preferred”. It states the product, configuration and method to which it applies, the alternatives considered, who assessed the data and the conditions requiring a new review. Cells without evidence must remain visible rather than being filled with an unsupported compliance judgement.

7. Two simulated matrices, two lines of reasoning

Case A: a clear aqueous solution with a potentially inhibitory component. The laboratory considers filtration because the preparation appears compatible with that route. Clarity does not, however, demonstrate an absence of antimicrobial activity. The plan checks system and materials, permitted treatment and suitability under final conditions. A favourable decision remains conditional on these results, rather than filtration ease alone.

If residual inhibition emerges during the study, the team documents the problem and evaluates permitted changes with a new demonstration of suitability. It does not keep increasing washing without limit or switch to direct inoculation on the assumption that this removes the same problem. The decision matrix shows which condition remains unsatisfied.

Case B: a viscous, opaque suspension. In the second simulated case, the laboratory gathers evidence that the intended preparation cannot be filtered in the assessed configuration. It therefore considers direct inoculation within the applicable method. Two separate questions remain: does the matrix inhibit growth, and does its opacity prevent reliable interpretation?

The plan addresses both through suitability checks and the observation route allowed by the reference. Availability of larger incubators does not close these questions. If the reading remains uninterpretable, the proposal is not ready for routine use. The cases do not demonstrate experimental performance or prescribe one technique for all solutions or suspensions.

8. Make the test reconstructable

Before routine use, link the approved method, suitability evidence, staff competence and equipment status. Check identity and status of media, diluents and materials, not just their presence on the bench. WHO good practices for microbiology laboratories support this coordinated management of resources, activities and records. [4]

Records should allow reconstruction of which sample was examined, by whom, with which materials and conditions, which controls were performed and which observations led to the result. For electronic systems, retain original data and the necessary context; a final transcription saying “complies” does not describe what happened during the test.

Annex 1 requires aseptic conditions and medium quality control according to the relevant reference. It also requires decontamination of sample external surfaces not to compromise method sensitivity or sample reliability. A treatment that protects the environment must therefore not be assumed harmless to the test without assessment. [3]

As an original operational check, ask the reviewer to reconstruct a test from retained records: identity, method version, materials, conditions, controls, observations, anomalies and decision. If any connection depends solely on the analyst’s memory, identify the documentation gap before it becomes an investigation problem.

9. Uncertain or positive results: preserve evidence before concluding

An uncertain observation requires recording and management under the approved procedure. Preserve relevant evidence, inform responsible functions and distinguish the observed data, hypotheses and conclusion. Do not automatically turn turbidity, reading difficulties or a control anomaly into “laboratory error”.

The WHO chapter treats growth as non-compliance with the test unless invalidity for causes unrelated to the product is clearly demonstrated under the specified conditions. A negative retest does not, by itself, cancel the initial result. The investigation should assess test and control data, environment, procedure and relevant microorganism information without selecting only favourable elements. [1]

Decisions on invalidation, further testing and product impact belong to the applicable quality process. This article provides no alternative release criteria. The initial selection rationale also helps here: a complete file shows whether a difficulty was anticipated, which evidence addressed it and what changed.

Operational conclusion

Select the technique from the product and applicable reference, then demonstrate that the configuration preserves detection and interpretability. Document interference, controls and the boundaries of the conclusion. If critical evidence is missing, make the uncertainty explicit and resolve it before routine use: having a system available is not yet a rationale for the method.

Continue in the Microbiology and Sterility hub and the article on organising a pharmaceutical microbiology laboratory to connect selection with actual workflows and resources.

Sources and applicability

Sources checked on 30 September 2026. The WHO chapter in the thirteenth edition, 2025, was read in full; ICH Q4B Annex 8 Step 5 and relevant Annex 1 sections were read. Only the public USP preview was consulted; the laboratory must check the current full USP/Ph. Eur. text. The matrix, cases and review questions are original GuideGxP recommendations, not additional compendial requirements.

1. WHO — The International Pharmacopoeia, 3.2 Test for sterility. Thirteenth edition, 2025.
2. EMA / ICH — Q4B Annex 8, Sterility Test. Step 5, EMA/CHMP/ICH/645592/2008; 2017 document, implemented from December 2009.
3. European Commission — EU GMP Annex 1. 2022 revision, particularly 10.5–10.9.
4. WHO — Good practices for pharmaceutical microbiology laboratories. TRS 961, Annex 2, 2011.
5. USP — General Chapter <71> Sterility Tests. Public preview, 2017 citation; not a full consultation of the current edition.