Endotoxin and sterility

Endotoxin and sterility are the two microbiological questions that can appear on a peptide report. Each is a separate question with its own method, output and limit, and neither says anything about identity, purity or amount. The LAL endotoxin method page compares the assay techniques. This lesson is about reading the numbers: detection floors, valid dilutions, spike recoveries and what a sterility result certifies.

Why are endotoxin and sterility different questions?

The incubation and media details come from the harmonised sterility text, which uses fluid thioglycollate medium, "primarily intended for the culture of anaerobic bacteria" but which also detects aerobic bacteria, and soya-bean casein digest medium for fungi and aerobic bacteria (harmonised sterility test text, EMA). The method pages for LAL endotoxin and sterility testing cover the procedures.

What is an endotoxin unit actually counting?

Bacterial endotoxins are lipopolysaccharides (LPS) from Gram-negative bacteria, and the horseshoe crab blood-cell lysate behind the LAL test starts its clotting reaction in response to picogram to nanogram amounts of them (Iwanaga, Proceedings of the Japan Academy, 2007). The result is not a weight. The harmonised text has laboratories prepare their standards from a reference standard endotoxin and states that "Endotoxin is expressed in Endotoxin Units (EU). One EU is equal to one International Unit (IU) of endotoxin" (harmonised Bacterial Endotoxins Test text, EMA). An EU is a unit of activity measured against that standard, and EU per mg expresses it per milligram of the sample, which is why that is the form to look for on a peptide report.

Why are endotoxin results so often written as "less than"?

Because the assay has a floor. Each LAL reagent has a sensitivity, written λ (lambda), the lowest endotoxin concentration it reliably detects. Samples are usually diluted before testing, partly to stop the sample interfering with the reaction. If nothing is detected in the diluted solution, the most the analyst can say is that the original sample held less than λ multiplied by the dilution.

Illustrative calculation, with invented numbers: a peptide is dissolved at 1 mg per mL, then diluted 1 in 50 for testing, using a reagent with λ of 0.005 EU per mL. Nothing is detected.

Had the analyst diluted 1 in 200, the same clean sample would have been reported as "< 1 EU/mg". Neither figure is the endotoxin content; both are floors set partly by the analyst's choices. Comparing two "less than" results compares floors, not measurements.

How far can a sample be diluted? The maximum valid dilution

Dilution cannot go on forever, because at some point the floor rises above the limit you are testing against and a "less than" result could hide a failure. The harmonised Bacterial Endotoxins Test text defines the maximum valid dilution (MVD) as "the maximum allowable dilution of a sample solution at which the endotoxin limit can be determined" and gives the formula: MVD = endotoxin limit x concentration of sample solution / λ (harmonised Bacterial Endotoxins Test text, EMA).

Continue the illustrative example with an invented limit of 5 EU per mg. MVD = 5 x 1 / 0.005 = 1,000. Any dilution up to 1 in 1,000 can still show whether the sample is under 5 EU per mg. The FDA's endotoxin guidance adds that if a test fails at less than the MVD, it should be repeated at a greater dilution not exceeding the MVD, and if an out-of-specification result at the MVD cannot be attributed to testing error, the lot should be rejected (US FDA, Pyrogen and Endotoxins Testing Q&A).

What does a spike recovery tell you?

It tells you whether the sample was hiding endotoxin from the reagent. Some samples damp the reaction (inhibition) or push it along (enhancement). To check, the laboratory adds a known amount of standard endotoxin to the sample and measures how much it gets back. In the photometric methods, the harmonised text treats the sample as free of interfering factors when "the recovery of the endotoxin added ... is within 50% to 200%". The same text requires the standard curve's correlation coefficient to be at least 0.980 (harmonised Bacterial Endotoxins Test text, EMA).

An illustrative report line, "< 0.25 EU/mg, kinetic chromogenic, spike recovery 92%", tells you the method, the floor, and that a known amount of endotoxin added to this sample was found. A line that reads "< 0.25 EU/mg" with no recovery is not wrong, but you cannot tell from it whether the sample suppressed the signal.

What does a sterility result actually certify?

Much less than the word suggests. The harmonised text is explicit: "a satisfactory result only indicates that no contaminating micro-organism has been found in the sample examined in the conditions of the test" (harmonised sterility test text, EMA). It is a finding about the units tested, not a guarantee about the rest.

That is why the method sets a minimum number of containers. For parenteral preparations, the harmonised table calls for, per medium, 10 per cent or 4 containers, whichever is greater, in batches of up to 100 containers, and 2 per cent or 20 containers (10 for large-volume parenterals), whichever is less, in batches of more than 500 (harmonised sterility test text, EMA). ICH Q4B Annex 8(R1) recommends that Ph. Eur. 2.6.1, JP 4.06 and USP chapter 71 can be used as interchangeable in the ICH regions, subject to stated conditions (ICH Q4B Annex 8(R1)%20Guideline.pdf)). A single vial sent in by a buyer cannot meet that design. It can show that one vial grew something, which is meaningful, but a clean single vial says very little about the batch.

Where do these tests sit in a full specification?

The European Medicines Agency's guideline on synthetic peptides lists bacterial endotoxins and microbiological purity among the tests a peptide substance specification may include, alongside identification, purity, assay, counter-ion, water and solvents (EMA, synthetic peptides guideline, 2025). A buyer's test orders part of that list. The which test tool helps you choose what to order for a given concern, and the tirzepatide page lists what a standard test reports for that compound. The course continues at /learn.

Check your understanding

1. Two reports say "< 0.5 EU/mg" and "< 2 EU/mg". Does the first sample have less endotoxin? Not necessarily. Both are floors set by dilution and reagent sensitivity. Neither is a measured content.
2. A spike recovery of 30% is reported. What does that suggest? The sample inhibited the reaction, which is outside the 50% to 200% range in the harmonised text, so a low result may be unreliable.
3. One vial is sterility tested and complies. Is the batch sterile? Unknown. The harmonised method tests many containers per batch, and even then a pass shows only that no organism was found in the units examined.