2026-09-07

What a COA Can Prove

A COA can support a narrow claim about a tested sample under a stated method. It can help connect a lot to an identity result, a purity-related result, or another listed specification. It cannot turn a research material into an approved medicine or prove that a material is sterile, stable, safe, effective, or suitable for people.

That is the whole game: match the claim to the evidence. A polished PDF does not widen the test’s scope.

Start with the sentence you want to prove

Write the claim before opening the certificate:

• “A laboratory found a signal consistent with the named analyte in this lot.”
• “The sample met this method’s stated purity criterion.”
• “The material stayed within specification through shipping or storage.”

The first two may be COA questions. The third needs stability evidence, defined conditions, and a time point. “Safe for humans” needs a clinical, manufacturing, and regulatory record far beyond a COA.

This split prevents a common error: using a strong analytical result as a shortcut for a weak clinical or safety claim.

The minimum chain of custody

A useful certificate identifies the material, lot or batch, sample, laboratory, method, date, result, units, and acceptance criteria. Compare the lot on the listing or record with the lot on the certificate. Compare the named analyte with the claim. Check whether the sample date and report date make sense together.

A report can be genuine and still be irrelevant to the item in front of you. A certificate for lot A does not establish the properties of lot B. A result from a bulk sample does not automatically describe a filled container. A report with no method or units leaves the conclusion underdefined.

For regulated manufacturing, FDA’s Q7 guidance on APIs says certificates should include the material name, batch number, release date, tests, acceptance limits, and numerical results when applicable. The same guidance treats supplier reliability as something to verify over time, not something to assume from a logo.

What the methods can support

HPLC can separate components under particular conditions and produce a chromatographic profile. An area percentage may support a method-specific composition claim. Mass spectrometry can show a measured mass-to-charge signal consistent with an expected molecular mass. Those are useful observations with defined boundaries.

A result is stronger when the report names the reference standard, method, detector, sample identifier, observed value, expected value, and acceptance criterion. A line that says “passed” without the test details is a conclusion asking for trust.

The FDA data-integrity guidance is a useful reminder that reliable data are complete, consistent, accurate, attributable, legible, contemporaneous, and original or a true copy. A screenshot of one peak is a fragment of a record, not the record itself.

What remains open

A COA usually cannot settle:

• sterility or endotoxin status;
• identity of every possible related substance;
• stability after an unverified storage history;
• accuracy of every label statement;
• biological activity in a person;
• clinical benefit or risk;
• authorization for a particular human use.

Those questions need other evidence and different accountable institutions. A COA is not defective because it cannot answer them. The mistake is asking it to.

A five-minute worksheet

Use the sample review page to capture the exact lot and document date. In your notes, make six columns: claim, source, sample, method, date, remaining uncertainty. Link the result to the Explore library only after you have separated a documented fact from an interpretation. The Field Manual has a printable version of the same matrix.

Classify each conclusion as documented, partly documented, unclear, or contradicted. “Unclear” is a valid result. It is more honest than a confidence score built from missing fields.

Limitations

This is educational material, not medical, legal, or laboratory advice. Methods, specifications, and regulatory duties vary by product and jurisdiction. Research-market documentation does not establish human-use suitability. The article does not provide dosing, injection, reconstitution, administration, vendor, or purchase guidance.

Sources

• FDA: Q7 Good Manufacturing Practice Guidance for Active Pharmaceutical Ingredients
• FDA: Data Integrity and Compliance With Drug CGMP
• ICH Q2(R2): Validation of Analytical Procedures