Bacterial endotoxin testing: method selection and interference control

Dilution can reduce interference, but it can also conceal contamination relative to the limit being assessed. Selecting a bacterial endotoxins test, BET, therefore starts with the product decision and reaches the reagent through suitability evidence. This guide provides a selection matrix and a numerical example, not universal release limits.

1. Define what the test must detect

BET detects endotoxins associated with Gram-negative bacteria. It neither counts viable microorganisms nor demonstrates the absence of every pyrogen. A sample can have a low bioburden count and contain endotoxins. Likewise, a compliant BET result does not replace sterility testing or an assessment of non-endotoxin pyrogens.

Identify the product or material, process point, matrix, route of use and intended decision. If the risk includes other pyrogens, assess an appropriate strategy, including approaches such as the monocyte activation test where relevant and acceptable. Choose according to risk and applicable requirements, not merely the equipment available.

2. Compare the technique and its compendial status

Limulus amebocyte lysate, LAL, techniques can detect gel formation, turbidity or colour development. Recombinant options include recombinant Factor C, rFC, and recombinant cascade reagent, rCR. Define the principle, response and controls for the actual procedure.

Status checked on 30 September 2026. USP 〈85〉 covers BET; 〈86〉 includes rFC and rCR without making them automatically applicable to every monograph. FDA’s 2026 update emphasises suitability for the intended purpose. Renaming the reagent is insufficient.

In Ph. Eur., rFC becomes method G in 2.6.14 and 2.6.32 is suppressed in Issue 13.1. The implementation date is 1 January 2027: publication and implementation are different. Check the current text, authorisation and market before transitioning.

3. Connect limit, sensitivity and dilution

The product limit depends on the applicable specification, dose, administration route and units. Do not confuse concentration with dose or transfer limits between products. The maximum valid dilution, MVD, expresses how far dilution can proceed while retaining the ability to assess that limit with the method’s sensitivity.

Teaching example with entirely hypothetical inputs: an already justified limit L = 2 EU/mL of product; applicable sensitivity λ = 0.05 EU/mL in the tested solution. EU means endotoxin units. MVD = L/λ = 2/0.05 = 40, a dimensionless factor. If the limit were in EU/mg, first convert it using the concentration in mg/mL. Never apply this formula without checking units.

At a 20-fold dilution, the analytical threshold expressed for the original product is 0.05 × 20 = 1 EU/mL; at 100-fold it becomes 5 EU/mL, above L. No detected signal at 100-fold would not demonstrate compliance with 2 EU/mL. Account for every preparation step and the definition of λ relevant to the technique.

4. Resolve interference without losing useful sensitivity

Inhibition and enhancement can arise from pH, ionic composition, colour, turbidity or interactions between the matrix and reaction system. Do not assign a standard dilution to a whole product family. Study the actual preparation, including any treatment and its effect on endotoxin detectability.

Simulated case: using the inputs above, the positive product control fails its predefined criterion at 10-fold dilution; at 20-fold, recovery and other controls are acceptable. The latter dilution can support a conclusion only within a suitable method and an approved procedure. A result below 0.05 EU/mL in the tested solution corresponds to less than 1 EU/mL in the product. If the control worked only at 100-fold, the problem would remain unresolved: a different approach and further data would be needed.

5. Separate system controls from product suitability

Document reference or control standards, potency traceability, lots, preparation, storage and expiry. Apply negative controls, sensitivity confirmation or a standard curve, and positive product controls as required by the technique. Recovery of a known addition, or spike, assesses interference in the tested preparation; it does not by itself certify every earlier step.

Define criteria before testing and retain raw readings, plate maps, dilutions, calculations and review records. Supplier documentation supports generic system performance; evidence on your material demonstrates local applicability. Reader qualification, software control and method suitability answer complementary questions.

6. Manage holding time, masking and anomalous results

A control added immediately before analysis cannot reconstruct endotoxin behaviour during storage. Justify the container, mixing, temperature and holding time with relevant data. Distinguish immediate interference from loss of recovery or masking over time, often discussed as low endotoxin recovery, LER.

The 2019 study by Reich and colleagues shows that endotoxin source and conditions influence masking. This is experimental evidence, not a universal storage recipe.

If controls fail, investigate test validity; if a valid result exceeds the specification, initiate the relevant OOS process. Retain initial results and repeat tests, with their rationale. Repetition provided for by the method does not authorise searching for a dilution that delivers a preferred value.

7. Control the method change

Within change control, connect purpose, covered products, comparator, interference, detection capability and expected variability. Establish the verification or validation needed and the handling of discordant results. Assess market-specific regulatory obligations in advance; a compendial chapter does not mean automatic acceptance of a change.

After implementation, monitor invalid tests, recovery, reagent lots and formulation or software changes. The useful benefit is a reliable result within the laboratory workflow. For other QC decisions, return to the microbiology laboratory hub.

Sources and access limitations

• FDA, Pyrogen and Endotoxins Testing, Edition 2, March 2026: nonbinding guidance; questions 2–7 address anomalies, samples and transition.
• FDA’s 2026 clarification on recombinant reagents.
• USP 〈85〉, 〈86〉 and 〈1085〉: public previews, not complete licensed texts. The cited EDQM FAQs clarify the transition but do not replace the chapters.
• Reich et al., 2019, masking study: indexed abstract consulted; no operating parameters derived.