A certificate of analysis, or COA, can exist in more than one version. A corrected report may fix an administrative detail, clarify information or replace an earlier document. The important question is not simply which PDF looks newest. It is how the versions connect, what changed and whether the document still describes the same tested sample.
This guide is for readers in Ireland and elsewhere in Europe who want to interpret laboratory documents, not select or use a cannabinoid product. Report control is one part of evidence literacy; it is not a safety guarantee or a legal assessment.
Why the report number matters
Begin with the laboratory name, report identifier, sample identifier, issue date and any revision label. Keep these separate from the date a file was uploaded to a website. An upload date tells you when that copy appeared, not when the laboratory performed the work.
NIST’s reporting-results crosswalk highlights unique document identification, identification of the tested item, relevant dates and authorisation. It is an explanatory laboratory-quality resource, not an Irish or EU product rule. These details help a reader connect a result with its documentary context.
Is a revised report evidence of a new test?
Not necessarily. An amendment and a fresh analysis answer different questions. A document can change without the laboratory measuring the sample again. Conversely, new testing can involve a different sample or method, so its result should not be silently substituted for the original.
As a practical reading check, ask whether the change concerns identification, wording, units, reported measurements or interpretation. If additional analysis is mentioned, ask which sample was examined and when. Our explanation of laboratory methods and sample matrices helps explain why a shared substance name is insufficient to establish comparability.
How should an amendment connect to the original?
European co-operation for Accreditation’s FAQ on amendments to test reports explains that changed information should be identifiable and reasons included where appropriate. It also stresses that changes must remain connected to the actual item tested, calibrated or sampled. A customer request does not justify inserting information unrelated to that item.
A complete replacement should have its own unique identification and refer to the original it replaces. EA’s further clarification on replacement reports discusses differing interpretations of whether every change must be marked within a complete new report. Its discussion emphasises clear information and context. For a reader, the useful response is to request an explanation when the relationship between versions is unclear, rather than assume that an unmarked difference proves misconduct.
A fictional example: two PDFs, one sample
Imagine report A-104, issued on 3 March, and A-104-R1, issued on 8 March. Both identify sample S-22. The second says it replaces the first and corrects a typographical error in the sample description. This tells you about a documented correction; it does not itself show that another sample was collected or that testing was repeated.
Now imagine that the second PDF instead lists sample S-23, with no explanation. The documents may concern separate work. Do not combine a measurement from one with a conclusion from the other. Seek clarification of the identifiers and the relationship between the reports before treating them as a continuous record. These identifiers are invented solely to illustrate the reading process.
What should you check before sharing a result?
- Obtain the complete issued report rather than a cropped table or screenshot.
- Record the report number, revision, issue date and sample identifier together.
- Read amendment notes and references to superseded documents.
- Check whether the underlying analysis changed or only the reporting information.
- When copies conflict, contact the laboratory using independently verified contact details and ask which document is current.
- Do not edit a PDF yourself to reconcile discrepancies or combine pages from different versions.
This is a practical checklist, not a substitute for the laboratory’s procedures. If confirmation is unavailable, describe the version as unverified. That is more accurate than choosing the copy with the most reassuring number.
What a clean version history cannot establish
Document control cannot make a sample representative of an entire batch. That separate issue is covered in our guide to batch sampling and measurement uncertainty. Nor does an identifiable report demonstrate that every possible contaminant was tested.
Likewise, a laboratory’s accreditation has defined boundaries; see what accreditation scope means. The measured properties, relevant methods and scope still need to be read on their own terms. A well-organised report history makes evidence easier to follow, but does not turn limited evidence into a universal assurance.
Educational information. Sources reviewed 22 October 2026. No individual report or laboratory is assessed here.