How to Read a Certificate of Analysis (COA)

A Certificate of Analysis, or COA, is a lab report. It tells you what a specific batch of a product actually contains. For peptides sold as “research chemicals,” a COA is pretty much the only honest look at quality you get. But it’s only as good as the lab that made it and the batch it describes. This guide walks you through what the tests mean, and how to tell a real COA from one that means nothing.

What a COA is (and is not)

A COA reports test results for one specific batch (also called a “lot”) of one product. A good one names the compound, the lot number, the date it was made or tested, the testing methods used, and the lab that did the work.

Here’s what a COA is not: it is not proof the product is safe, it is not a promise that every vial is good, and it is not a sign that the product is legal or okay to use in people. A COA only tells you what was in the sample the lab tested. That isn’t always the same as what’s in the vial in your hand.

The core tests: purity and identity

Two tests do most of the heavy lifting, and they answer two different questions.

HPLC (high-performance liquid chromatography) measures purity — how much of the sample is the thing you actually want. The machine separates everything in the sample, and each ingredient shows up as a “peak” on a graph. Purity is usually figured out by taking the size of the main peak and dividing it by the size of all the peaks added together (this is called “area under the curve”), then turning that into a percentage. So a “98%” result means the peptide you’re after makes up 98% of what the machine saw. The other 2% is other peptide-related stuff.

One important catch: HPLC purity is only measured against what this test can detect. It does not count water, salts, or certain other particles, and on its own it cannot prove the main peak really is the right molecule.

Mass spectrometry (MS) answers a different question — identity, meaning is this actually the right compound? It measures molecular weight, so you can compare the observed mass (what the machine measured) against the theoretical mass (the weight the compound should have based on its known sequence). A close match — usually within about half a dalton (a tiny unit of molecular weight) for these methods — is a good sign the product is what the label says.

Purity without identity is empty. A sample can be 99% pure and still be 99% of the wrong thing. A trustworthy COA shows both tests.

The safety-oriented tests

For anything that’s going to be mixed with liquid and injected, contamination matters even more than purity. These tests follow standards set by the United States Pharmacopeia (USP), a group that sets quality rules for medicines:

Sterility (USP <71>) — whether the material grew any detectable bacteria or fungi.
Bacterial endotoxin (USP <85>) — endotoxins are lipopolysaccharides, which are pieces of the outer shell of certain bacteria (Gram-negative bacteria). They handle heat well, survive ordinary sterilizing, and can cause fever and inflammation. They’re checked using the LAL (Limulus amebocyte lysate) method, or with newer lab-made reagents such as recombinant Factor C (USP <86>, official May 2025). Results come in endotoxin units (for example, EU/mg). For injectable drugs, the threshold human pyrogenic dose — the level that can trigger a fever — is 5.0 EU/kg per hour (and far lower, 0.2 EU/kg per hour, for the intrathecal route, meaning injection near the spinal cord).
Heavy metals — things like lead, arsenic, and mercury, which can come from how the product was made or from its raw ingredients.

Most research-grade peptide COAs only report purity and identity and skip these safety tests completely. That missing part tells you something too.

Third-party vs vendor-supplied

A third-party COA is made by an independent lab that has nothing to gain from the sale — ideally a lab with ISO/IEC 17025 accreditation. (ISO/IEC 17025 is the international standard for whether a testing lab is actually competent at its job.)

A vendor-supplied (in-house) COA is made by the seller themselves. That doesn’t make it useless, but the seller clearly has a reason to make the product look good, and a report like this is easy to fake. Lean toward COAs that name a real, independent lab, that sit on the lab’s own letterhead, and that include contact details you can actually check.

Spotting fake, stale, or meaningless COAs

Watch for these warning signs:

No lot number, or a lot that doesn’t match your vial. A COA for a different batch tells you nothing about your product.
Old or reused. One COA copied across every batch for years; no date; or a “made on” date from long before you bought it.
No methods listed, or no lab named. Something like “Purity: 99%” with no graph, no instrument, and no analyst behind it.
Missing the graphs. A number without the actual chromatogram or spectrum image behind it can’t be checked.
Too perfect to be real. “100.0%” purity, or the exact same numbers on every product.
Details that don’t add up — the wrong molecular formula, sequence, or theoretical mass for the compound it names.
Easily edited files with mismatched fonts, crooked text, or messy formatting.

When a lab is named, you can sometimes reach out to them directly to confirm the report is real.

Bottom line

A strong COA covers one specific batch, is recent, comes from a named independent lab, and shows both HPLC purity (with the graph) and MS identity (observed mass vs theoretical mass) — plus sterility and endotoxin results for anything injectable. If you see a number with no graph, no lot, no date, or no lab behind it, treat it as marketing, not proof.

And keep the bigger picture in mind: as of 2026, peptides sold this way are not FDA-approved for use in people, and a clean COA doesn’t change that. (Even when a peptide might be allowed for pharmacy compounding, that is not the same thing as FDA approval.) This is reference information, not medical advice.

Sources

New profiles and evidence updates only. No spam, and we never sell or share your email.