Key takeaways
- Endotoxin is a molecule from dead or living Gram-negative bacteria; sterility is about living organisms. They need different tests.
- LAL and recombinant factor C tests report endotoxin in EU per mg; the pharmacopoeial chapters are USP <85> and <86>, Ph. Eur. 2.6.14 and 2.6.32.
- The pharmacopoeial sterility test (USP <71>, Ph. Eur. 2.6.1) incubates samples for 14 days; a microbial count is a different, weaker test.
- FDA's July 2026 review of BPC-157 certificates found purity results but no endotoxin or bioburden results.
- Low endotoxin levels can activate sensitive immune cells in the lab, so an untested reagent can create a result that belongs to the contaminant.
01Two tests, two different questions
A peptide can be the right molecule at high purity and still carry bacterial debris or living microbes. HPLC and mass spectrometry are not built to see either.
| Test | Question it answers | What it detects | How the result is written | Pharmacopoeial chapter |
|---|---|---|---|---|
| Bacterial endotoxins (BET) | How much bacterial endotoxin is in the sample? | Lipopolysaccharide from Gram-negative bacteria, alive or dead | A number in EU per mg (or per mL), against a stated limit | USP <85> and <86> · Ph. Eur. 2.6.14 and 2.6.32 |
| Sterility | Does anything alive grow from the sample? | Living bacteria and fungi | Growth or no growth after 14 days, with the number of containers tested | USP <71> · Ph. Eur. 2.6.1 |
| Microbial count (bioburden) | How many organisms are present in a non-sterile material? | Living bacteria, yeasts and moulds that grow on the test media | Counts, often written TAMC and TYMC | USP <61> · Ph. Eur. 2.6.12 |
The three pharmacopoeias (US, European, Japanese) have harmonised these chapters, so a method run to one is accepted as interchangeable with the others under conditions set out by ICH.
02What endotoxin is, and why sterilising does not remove it
Endotoxins are lipopolysaccharides (LPS) found in the outer membrane of Gram-negative bacteria. FDA's inspection guide describes their release from bacteria after the cell dies and breaks open. That is why endotoxin can be present when no living bacteria are.
Endotoxin is also hard to get rid of. The same FDA guide notes that it is difficult to remove from products once present, and notes that some texts recommend dry heat at 250 °C for 45 minutes to make glassware pyrogen-free. Filtering or sterilising a solution can kill or remove bacteria and still leave the endotoxin behind.
Results are reported in endotoxin units (EU). The US, European and Japanese reference standards are interchangeable because each is calibrated against the WHO International Standard for Endotoxin. For a peptide, the result is usually written per mg of material.
03How the tests work: LAL and recombinant factor C
LAL. Limulus amebocyte lysate is made from horseshoe crab blood cells. Endotoxin sets off a clotting cascade in the lysate, and the test reads that reaction in one of three ways: a gel clot forms or not (gel-clot), the mixture turns cloudy (turbidimetric), or a colour develops (chromogenic). FDA has accepted LAL in place of the older rabbit pyrogen test for decades; a November 1977 Federal Register notice first set out conditions for using it as a finished-product test. When techniques disagree, the gel-clot limit test makes the final call.
Recombinant factor C (rFC). Factor C is the first enzyme in the crab's cascade. A recombinant version, read by fluorescence, removes the need for crab blood. The European Pharmacopoeia added chapter 2.6.32 for rFC in July 2020, in force from 1 January 2021. USP chapter <86>, covering rFC and recombinant cascade reagents, was published for early adoption in November 2024 and became official in May 2025, as an alternative to <85>.
Interference. The sample itself can block or boost the reaction. A valid result depends on a test for interfering factors: endotoxin is spiked into the sample and must be recovered. Labs dilute the sample to overcome interference, but only up to a maximum valid dilution.
LAL and rFC detect endotoxin only. Other fever-causing contaminants need a pyrogen test, such as the monocyte activation test that FDA lists as an alternative assay.
04Sterility testing is a different test
The pharmacopoeial sterility test (USP <71>, Ph. Eur. 2.6.1, JP 4.06) asks a yes-or-no question about living organisms. The sample is either filtered through a membrane or added directly to two growth media, fluid thioglycollate medium and a soybean-casein digest broth. The media are incubated for 14 days and checked for cloudiness that signals growth.
Two consequences follow. The test consumes the containers it uses, and it reports only on those containers. And it takes two weeks, so a sterility line on a certificate dated a few days after manufacture deserves a question.
A microbial count is not the same thing. The enumeration tests in USP <61> and Ph. Eur. 2.6.12 are written for non-sterile products and report how many organisms grew, often as a total aerobic microbial count (TAMC) and a total yeast and mould count (TYMC). A low count is useful information. It is not a sterility result.
05Why many peptide certificates skip both
FDA put this on the record in its briefing for the July 2026 Pharmacy Compounding Advisory Committee. The certificates supplied for BPC-157 acetate reported appearance, identification, purity, solubility, related substances and acetic acid content. FDA found no results for specified impurities, aggregates, bioburden or bacterial endotoxin levels. Its literature search found that most BPC-157 certificates contain only a purity result.
There are practical reasons the gap is common. Endotoxin needs its own method and an interference check for each peptide. Sterility needs two weeks and uses up vials. A lab that only runs HPLC cannot answer either question.
- Endotoxin: method named (gel-clot, kinetic chromogenic, rFC), a number in EU/mg, and the limit it was judged against.
- Interference check passed, or the dilution used.
- Sterility: the chapter followed, how many containers were tested, the incubation period, and the result.
- The batch number, the date and the lab, as on any certificate of analysis.
- "Endotoxin: pass" with no number and no limit.
- "Sterile" based on a microbial count.
- A sterility result dated within days of the test start.
06What it means for research models
In cell work, endotoxin is a confounder. In a 2014 study, researchers at the University of Salzburg measured endotoxin in commercially available recombinant proteins and found levels in the range stated on data sheets, and some higher. They then showed that human CD1c+ dendritic cells could be activated by lipopolysaccharide at amounts equivalent to the contamination found in some of those products. Their conclusion: with LPS-sensitive cells, even low endotoxin contamination may generate erroneous data.
That study tested recombinant proteins, not synthetic peptides. The principle carries over to any reagent added to an immune or inflammation readout: if endotoxin was never measured, an observed effect could belong to the contaminant rather than the molecule.
In animal models, endotoxins are pyrogens, substances that cause fever, which is why the oldest test for them used rabbits (USP <151>). A study design that cannot rule them out has a gap in its controls.
07What these tests can't tell you
- Whether a vial stayed clean after it was opened or reconstituted. The result describes the sample at the time of testing.
- Whether other batches are the same. Each lot needs its own result.
- Identity, purity or quantity. Those come from HPLC, mass spectrometry and net peptide content.
- Whether a substance is safe or effective in people. Quality testing and evidence of effect are separate questions; see evidence grades.
Plain-English summary of pharmacopoeial and FDA sources, checked 17 November 2026. Not legal or medical advice.