Bacterial endotoxin test (revision 23)

The bacterial endotoxin test detects lipopolysaccharide, a component of the outer membrane of Gram-negative bacteria that provokes a febrile response when introduced parenterally. The test exploits a clotting cascade in amoebocyte lysate derived from horseshoe crab blood, which is activated by picogram quantities of endotoxin.^[1]

Endotoxin is not a living organism and is not removed by sterilisation. A preparation can be sterile and pyrogenic at the same time: autoclaving kills bacteria but leaves their lipopolysaccharide intact and biologically active. Endotoxin is the commonest but not the only pyrogen, and a separate rabbit pyrogen test exists for the wider class.^[2] This is why endotoxin and sterility are separate determinations answering separate questions.^[1]

Results are expressed in endotoxin units, normalised to mass for a drug substance or to volume for a solution. Compendial limits for parenteral products are calculated from the maximum dose, and a research-chemical certificate reporting an endotoxin figure without the basis of normalisation is difficult to interpret.^[3]

Three variants are recognised. The gel-clot method observes whether a firm clot forms at a given dilution and gives a limit result rather than a number. The turbidimetric method follows the increase in turbidity as the cascade proceeds and is quantitative. The chromogenic method measures release of a coloured product from a synthetic substrate and is also quantitative.^[1]

Recombinant factor C assays, which use a recombinant version of the initiating enzyme rather than crab-derived lysate, have been introduced and are increasingly accepted. They avoid the animal-derived reagent and are more specific, since they respond only to endotoxin and not to β-glucans.

Interference is the routine difficulty. Sample matrices can enhance or inhibit the cascade, so a validated method includes a positive product control demonstrating recovery of a known spike. A result reported without evidence that interference was assessed is not fully interpretable.^[1]

Endotoxin contamination arises from the water and the equipment, not usually from the synthesis. Water is the principal route: bacteria grow readily in stagnant water systems, and even after they are killed their lipopolysaccharide persists. Glassware carries endotoxin unless depyrogenated by dry heat, which is a more aggressive process than sterilisation.^[1] Water-system control is accordingly treated as a manufacturing rather than a testing problem.^[4]

Lyophilised research peptides are not manufactured as parenteral products and are not required to meet a parenteral endotoxin limit. Some certificates report endotoxin nonetheless; many do not. Absence of the determination is not evidence of a problem, and its presence is not evidence that the material is suitable for any particular purpose.

For comparison, the compendial limit for most parenteral products corresponds to 5 EU per kilogram of body weight per hour, translated into a per-milligram limit by the maximum dose. A figure quoted in EU/mg without the dose basis cannot be compared against that limit by a reader.^[3]

A low endotoxin figure establishes that the sample tested contained little detectable lipopolysaccharide. It does not establish sterility — see Sterility testing — and does not address non-endotoxin pyrogens, which the test does not detect at all.^[1]

Nor does it transfer between vials. Endotoxin contamination is frequently heterogeneous, arising from a particular piece of equipment or a particular fill, so a determination on one unit constrains other units only through the assumption of lot uniformity that the whole apparatus of lot traceability exists to support.

• USP General Chapter 85
• Sterility testing
• USP General Chapter 71
• Limit of detection
• Certificate of analysis