Endotoxin, Sterility and Heavy Metals: What a Purity Number Leaves Out

An HPLC purity number leaves out endotoxin, sterility and heavy metals entirely, because the instrument that produces it cannot detect any of the three. Each needs its own test, run by a different method, and a certificate of analysis that does not report a result for one of them has said nothing about it.

That is not a flaw in HPLC. It is a description of what the method was built to do, and of the three questions it was never asked.

Why a chromatogram cannot see them

Reversed-phase HPLC separates the dissolved, peptide-like components of a sample and measures how much ultraviolet light each one absorbs. Purity is the main peak's share of everything the detector registered. Our explainer on what an HPLC trace shows covers the mechanics.

Three kinds of contaminant fall outside that picture:

Endotoxin is a fragment of bacterial cell wall that matters at trace levels, far below what a purity percentage resolves.
Living microorganisms are not a dissolved chemical at all. A handful of cells in a vial does not produce a peak.
Metals such as lead or arsenic are not detected by a method tuned to the bonds in a peptide backbone.

A sample can be 99 percent pure by HPLC and fail all three. The purity figure would not move.

Endotoxin: USP <85>

What it is. Bacterial endotoxins are lipopolysaccharides found in the outer membrane of gram-negative bacteria, according to the FDA's inspection guide on the subject. They are released when bacterial cells die and break apart. The effect the guide emphasizes is fever, and substances that cause it are called pyrogens.

Why it is separate from sterility. Endotoxin is debris, not a living organism, so killing or filtering out bacteria does not get rid of it. The FDA guide states that endotoxins are not removed by sterilizing or microbiological filters and are difficult to remove once present. It lists dry heat at 250 degrees Celsius for 45 minutes among the conditions used to destroy them. A product can be sterile and still carry endotoxin.

How it is tested. The compendial method is the Bacterial Endotoxins Test in United States Pharmacopeia chapter <85>, often called the LAL test after its reagent, Limulus amebocyte lysate, which reacts to endotoxin. The FDA has accepted LAL testing in place of the older rabbit pyrogen test for decades, and its current guidance describes gel clot, photometric and kinetic versions of the method. A companion chapter, <86>, covering recombinant reagents not derived from animals, was set to become official in May 2025.

What a result looks like. Endotoxin is reported in endotoxin units, abbreviated EU, against a stated limit. A certificate line reading "bacterial endotoxins" with a number, a unit, a limit and a method is a result. A line reading "pass" with none of those is a claim.

Regulators treat this as a basic characterization question for peptides. When FDA staff reviewed semax for its July 2026 advisory meeting, a lack of endotoxin data was one of the specific gaps they cited in concluding the substance was not well characterized.

Sterility: USP <71>

What it is. A sterility test looks for living bacteria and fungi. Under the procedure laid out in USP <71> and described in the FDA's Pharmaceutical Microbiology Manual, samples are either passed through a membrane filter or added directly to growth media, then incubated for not less than 14 days in two media, fluid thioglycollate medium and soybean-casein digest medium, chosen to support different kinds of organisms. If nothing grows, the sample passes.

What that means for a certificate. A genuine sterility result cannot be produced overnight. A certificate dated the day after a lot was filled either did not include a compendial sterility test or reported it before the incubation was finished.

Its built-in limit. Sterility testing is destructive, so only a sample of a batch can be tested. The FDA's guidance on aseptic processing is direct about the consequence: sterility tests are limited in their ability to detect contamination because of the small sample size typically used. Its own illustration is a lot of 10,000 units with a 0.1 percent contamination rate, tested using 20 units. That lot has a 98 percent chance of passing.

This is why pharmaceutical manufacturing treats sterility as something a validated process delivers and a test spot-checks. A passing sterility result is evidence about the units tested and the process behind them. It is not proof about every vial.

Heavy metals: USP <232> and <233>

What it is. The modern term is elemental impurities. USP chapter <232> sets the limits and chapter <233> sets the analytical procedures. Together they replaced the old Heavy Metals chapter, <231>, and became applicable to drug products with USP monographs on January 1, 2018, according to FDA guidance. USP had said the older test was no longer sensitive enough to detect a number of impurities at levels known to be toxic.

Which elements. The chapters are aligned with an international guideline, ICH Q3D, which sorts elements into classes by toxicity and by how likely they are to turn up. Class 1 is arsenic, cadmium, mercury and lead, described in the guideline as human toxicants with limited or no use in pharmaceutical manufacturing, typically arriving through commonly used materials. Other classes cover metals such as nickel, cobalt, palladium and platinum, some of which are used as catalysts.

What a result looks like. A real elemental impurities result names the elements tested and gives a measured value for each, against a limit. A single line reading "heavy metals: conforms" with no elements, no numbers and no method tells you much less.

How to read a certificate for these three

Treat each as its own question and look for the same four things each time:

The test is named. Bacterial endotoxins, sterility, elemental impurities. If a test is not listed, it was not reported.
A method is cited. A USP chapter number or an equivalent pharmacopeial method.
A result and a limit appear. A number with units for endotoxin and metals. For sterility, a statement of no growth after the incubation period.
The lot matches. The result has to belong to the batch in front of you. Our guide to reading a certificate of analysis covers lot matching in detail.

One further caution applies to sterility and endotoxin in particular. Both depend on how a product was filled and handled, not only on the raw material. A result for a bulk powder does not describe a finished vial filled from it later.

As our explainer on independent testing notes, these analyses are ordered separately from purity and identity, and a certificate that reports only those two has said nothing about the rest. An absent test is not a failed test. It is an unanswered question, and it should be read that way.

What these tests do not establish

Passing all three says a sample met defined limits for three kinds of contamination. It says nothing about whether a compound is safe or effective in a person, which is a question for clinical trials. Research material is sold for laboratory use, and none of these results changes that.

Frequently asked questions

Does a 99 percent purity result mean a peptide is sterile?

No. HPLC purity measures the relative amount of the main peptide component. It does not detect microorganisms. Sterility is a separate test with a 14-day incubation.

What is an endotoxin?

A lipopolysaccharide from the outer membrane of gram-negative bacteria, released when the cells die. It can remain in a product after the bacteria themselves are gone.

What is the LAL test?

The Bacterial Endotoxins Test described in USP <85>. It uses Limulus amebocyte lysate, a reagent that reacts to endotoxin, and reports a result in endotoxin units.

Can a product be sterile and still fail endotoxin?

Yes. Sterilizing filters remove living bacteria but, according to the FDA, do not remove endotoxins.

Which heavy metals should a certificate list?

At minimum the four Class 1 elements in ICH Q3D: arsenic, cadmium, mercury and lead, each with its own measured result.

Sources:

The Assay is published by Merit Sciences, which sells research compounds and publishes its own certificates of analysis. Research use only.

For research use only · Not for human or veterinary use · No research compound discussed here is FDA-approved for any indication