Checking a certificate of analysis means confirming, in order: the document belongs to the delivered batch; the standard it claims is the one that applies; every expected test is present with a limit at least as strict as that standard's; quantitative results are numbers within those limits; methods, dates and signatures are stated; and the document agrees with itself.
Why a procedure beats a glance
Most certificates are looked at rather than checked: the batch number is glanced against the label, the word "complies" is noted, and the document is filed. The faults that matter rarely announce themselves at a glance. A printed limit quietly looser than the monograph, a quantitative test answered with "complies", a retest date that contradicts the manufacturing date: each survives a glance and each has consequences.
A fixed order also protects the reviewer. Checking results before establishing which specification governs them produces confident nonsense, because a result can pass one standard and fail another. The sequence below is arranged so that each step gives the next one its meaning, and it is the same sequence the automated analysis follows.
The ten steps
Work through these in order. Steps 1 to 3 establish what you are looking at; 4 to 7 do the comparisons; 8 to 10 test whether the document can be trusted as a document.
- 1. Tie the certificate to the goods. The material name, grade and batch number on the certificate must match the container label and the order. A certificate that cannot be tied to the delivered batch proves nothing about it (ICH Q7 11.40 to 11.41).
- 2. Identify the issuer. Original manufacturer, or a repacker, agent or broker? A re-issued certificate should name the laboratory that did the testing, refer to the original manufacturer, and attach the original certificate (ICH Q7 11.43 to 11.44).
- 3. Establish the governing standard. What does the certificate claim: a pharmacopoeial monograph, the supplier's specification, your agreed specification? Everything after this step is a comparison against that claim, so pin it down first.
- 4. Check the test list is complete against the claimed standard. A claimed monograph with monograph tests missing is a gap in the claim, before any result is read (ICH Q7 11.42).
- 5. Compare printed limits against the claimed standard's limits. A looser printed limit means every "pass" on that row was judged against a weaker promise. This is a separate question from whether the result is good.
- 6. Read the results against the printed limits. Quantitative tests need numbers with units; "complies" belongs only to qualitative tests such as appearance or identification (ICH Q7 11.42).
- 7. Check the arithmetic and internal consistency. Individual impurities against their total, assay against its neighbours, values that appear twice and must agree.
- 8. Check methods. Each test should name its method; a result produced by an unstated method cannot be compared with a monograph limit that presumes a specific procedure.
- 9. Check the dates. Manufacturing date, release date, retest or expiry date: present, unambiguous, and consistent with each other and with today's date if the material is about to be used.
- 10. Check authorisation. A dated signature or approved electronic equivalent from the issuer's quality unit, with the issuer's name and contact details in the text, not only in a logo (ICH Q7 11.43).
What failure looks like at each step
The same faults recur across supplier certificates. This table is the working map from step to the finding it usually produces.
| Step | Typical failure | Why it matters |
|---|---|---|
| 1. Batch tie | Certificate covers a different batch, or two batches share one certificate | Every result on the page is about other goods |
| 2. Issuer | Trader's certificate with no testing laboratory named and no original attached | The chain of evidence stops at a letterhead |
| 3. Standard | "Complies with USP" with no edition and in-house limits printed | The claim and the contents describe different standards |
| 4. Test list | Claimed monograph, impurity tests absent | Silence exactly where scrutiny belongs |
| 5. Limits | Printed assay limit wider than the monograph's | A pass against a weaker promise |
| 6. Results | "Complies" against assay or loss on drying | Cannot be compared, trended or relied on |
| 7. Consistency | Impurity rows exceed the printed total | The document contradicts itself |
| 8. Methods | No method beside water content | Different methods give different numbers for the same material |
| 9. Dates | Retest date before the manufacturing date | The document's own timeline cannot be true |
| 10. Authorisation | Titles without names, or no date at the signature | No accountable person released the batch |
When checking the document is not enough
A sound certificate is necessary, not sufficient. Every major framework keeps at least one specific identity test on each batch with the recipient regardless of the paperwork: 21 CFR 211.84(d)(1) and (d)(2) in the United States, EU GMP Chapter 5 section 5.35 in Europe, ICH Q7 7.30 for API manufacturers. Reliance on a supplier's certificate for the remaining tests must itself be earned: a documented supplier evaluation, full analyses at intervals compared against the certificates, and audits where risk warrants them (21 CFR 211.84(d)(2), EU 5.36, ICH Q7 7.31).
And a document check cannot tell you whether the document is genuine or whether the laboratory that signed it did the work. Where the stakes or the doubts are high enough, the answer is a laboratory, not a better reading of the paper. What a careful check does is tell you which certificates deserve that escalation, with the reasons written down.
Definitions
- Certificate of analysis (CoA)
- A document issued for a batch of material listing the tests performed, the acceptance limits and the results, authorised by the issuer's quality unit.
- Acceptance criteria
- The numerical limits, ranges or other measures a test result must meet for the material to conform, set by the governing specification.
- Monograph
- The section of a pharmacopoeia setting the quality standard for one material: its tests, methods and acceptance criteria.
Sources
- ICH Q7, Good Manufacturing Practice Guide for Active Pharmaceutical Ingredients (Step 4, 10 November 2000), International Council for Harmonisation (sections 7.30 to 7.32 and 11.40 to 11.44)
- 21 CFR 211.84, Testing and approval or rejection of components, drug product containers, and closures, U.S. Code of Federal Regulations (eCFR) (paragraphs (d)(1) to (d)(3))
- EudraLex Volume 4, EU Guidelines for GMP, Part I, Chapter 5: Production (revision in operation from 1 March 2015), European Commission (sections 5.35 and 5.36)
Primary sources are cited in preference to summaries. Where a source is licensed (for example a pharmacopoeial monograph) the reference is given and the text is not reproduced.
Dr. Majid Qaria
Dual-PhD biomedical scientist and pharmaceutical entrepreneur working across emerging-market supply chains.
Why we built this →Spotted an error or a newer figure? We correct within a week: info@pharmatrust.tech.