What a Certificate of Analysis Actually Proves

“99.4% PURE” is a beautiful number.

It is specific enough to feel objective and close enough to perfection to end the conversation. Place it beside a chromatogram with one large peak, add the name of a laboratory, and a buyer may feel that every important question has been answered.

Usually, one question has been answered.

A certificate of analysis is not a passport declaring a material safe, effective, sterile, legal, correctly filled, or appropriate for a particular use. It is a record of tests performed on a sample under stated conditions.

Its value depends on four things:

• Was the correct sample tested?
• Was the right method used for the question?
• Is the result interpreted accurately?
• Does the conclusion remain inside the limits of the test?

Reading a COA well begins by refusing to ask one number to do six jobs.

First: does the report belong to the product?

Before interpreting chemistry, establish identity in the ordinary documentary sense.

Look for:

• Product name
• Exact sequence or unambiguous chemical description
• Modification and salt form when relevant
• Lot or batch number
• Sample or report number
• Date received
• Date tested
• Laboratory name and contact information
• Methods performed
• Results and acceptance criteria
• Reviewer or authorized signature

Then compare the lot number on the report with the lot number on the material.

A report for “TB-500” or “BPC-157” without a batch identifier may show that a sample with that label was tested at some point. It does not establish that the current inventory came from the same production lot.

A QR code is not inherently verification. It may lead to a permanent marketing page displaying one favorable report for every batch.

Traceability comes before interpretation.

Product name is not always chemical identity

Commercial names can be imprecise.

A peptide may exist in different salt or counterion forms. It may be modified, conjugated, truncated, amidated, acetylated, or otherwise altered. A familiar nickname can refer to related but non-equivalent materials.

A report should make clear what the laboratory understood the sample to be.

This is especially important when theoretical molecular mass is compared with measured mass. The expected result depends on the actual chemical form.

If the identity on the vial, the report, and the underlying specification do not align, the purity percentage cannot repair the ambiguity.

What HPLC tells you

High-performance liquid chromatography separates components according to how they interact with a column and the mobile phase moving through it.

As components leave the column, a detector records signals. The resulting chromatogram contains peaks.

In a simplified area-normalization approach:

Area of the designated main peak
÷
Area of all integrated peaks
× 100
= reported chromatographic purity

If the principal peak accounts for 99.4 percent of the integrated signal, the report may state 99.4 percent purity by that method.

That can be meaningful. Under the stated conditions, one detected chromatographic component dominated the sample.

But it does not automatically mean:

• 99.4 percent of the total vial weight is target peptide
• 99.4 percent of the labeled milligrams are present
• The largest peak has been conclusively identified
• Every impurity was detected
• Every impurity separated into its own peak
• The sample is sterile
• Endotoxin is absent
• The material will remain unchanged during storage
• The delivered batch is the one tested

HPLC is not one universal method. Results depend on the column, mobile phase, gradient, detector, wavelength, run time, sample preparation, reference standard, system suitability, and integration choices.

A number without a method is an orphan.

The chromatogram deserves to be seen

A complete report should allow a technically qualified reader to examine more than the summary box.

Useful information may include:

• Full chromatogram
• Retention time
• Integration boundaries
• Secondary peaks
• Baseline
• Detector conditions
• Column and mobile phases
• Gradient
• Run time
• Reference comparison
• System-suitability results
• Calculation method

A cropped image showing one tall peak and no surrounding baseline is persuasive design, not complete analytical context.

One complication is co-elution: more than one component may emerge together closely enough to appear as a single peak under a particular method. Higher-resolution chromatography, different methods, or mass-spectrometric investigation may be needed to understand what the main peak contains.

Purity is partly a property of the material and partly a property of the method’s ability to separate what is there.

What mass spectrometry adds

Mass spectrometry measures ions according to their mass-to-charge ratio.

For peptide identity, the laboratory can compare observed mass with the theoretical mass expected from the proposed sequence and chemical form. Multiple charge states may appear, and more advanced tandem mass spectrometry can provide additional structural or sequence evidence through fragmentation.

A close match supports the conclusion that the sample contains a molecule consistent with the expected peptide.

It still does not answer every question.

Mass spectrometry can support identity without quantifying how much target peptide is present in the vial. HPLC can report relative chromatographic purity without proving the largest peak is the intended molecule.

The methods are complementary:

• Chromatography asks: what components are separated and in what relative proportions?
• Mass spectrometry asks: is the observed mass consistent with the proposed identity?
• Assay asks: how much target material is present?
• Other tests ask: what additional chemical, physical, or microbiological attributes matter?

Strong documentation often uses orthogonal methods—different techniques that approach identity or quality from different directions.

Purity is not content

Suppose a vial is labeled 10 milligrams and the HPLC result is 99 percent.

It is tempting to multiply the two and conclude the vial contains 9.9 milligrams of target peptide.

That calculation may be invalid.

The dry material can include:

• Water
• Counterions
• Residual solvents
• Salts
• Excipients or bulking agents
• Process-related residues
• Peptide-related impurities

Chromatographic area percentage is not automatically a measurement of total target mass per vial.

Establishing content may require a quantitative assay calibrated against an appropriate standard, amino-acid analysis, water determination, counterion analysis, fill controls, or a combination of methods appropriate to the material.

The correct question is not merely:

What percentage does the chromatogram show?

It is also:

What evidence supports the labeled quantity of target peptide?

The remaining fraction has an identity too

If a report shows 99 percent target-related peak area, what makes up the remaining 1 percent?

Potential peptide-related and process-related impurities can include:

• Truncated or deletion sequences
• Incomplete coupling products
• Oxidized material
• Deamidated products
• Isomers
• Residual protecting groups
• Residual reagents and solvents
• Aggregates
• Degradation products

One unknown impurity is not equivalent to another. Meaning depends on identity, concentration, biological activity, and the research context.

A total impurity percentage can conceal whether the remainder is composed of many tiny known peaks, one larger unknown peak, or substances not well detected by the method.

A mature quality discussion asks about the impurity profile, not only the headline purity.

Chemical purity does not establish sterility

HPLC and mass spectrometry are chemical analytical methods.

They do not, by themselves, establish the absence of:

• Viable bacteria
• Fungi
• Bacterial endotoxins
• Particulate contamination
• Contamination introduced after the tested sample was taken

Those questions require separate, appropriate methods.

Likewise, a claim of “sterile filtered” describes a process step. It is not automatically the same as a passing validated sterility test on the finished material.

The attributes required for a laboratory experiment depend on the experiment. The point is not that every material requires every possible test. It is that an unperformed test cannot be inferred from an unrelated result.

A COA is a photograph, not a time machine

The report describes the sample when it was tested.

It does not automatically describe the product after:

• Months of storage
• Heat exposure
• Repeated temperature cycles
• Moisture ingress
• Damaged packaging
• Shipment through uncontrolled conditions
• Preparation or reconstitution
• Extended in-use storage

Stability is a program, not a date printed beside a purity result.

A credible stability assessment considers storage conditions, time points, stability-indicating methods, packaging, and the attributes that may change. A release test establishes that a sample met specifications at release. It does not prove indefinite preservation.

Who selected the sample?

The analytical result can be genuine while the sampling process remains weak.

Ask:

• Did the laboratory receive a manufacturer-selected sample?
• Was a finished product purchased independently?
• Was the vial sealed?
• Was chain of custody documented?
• Could the tested sample differ from ordinary inventory?
• Is the laboratory competent for the method?
• Can the report be verified directly?
• Are reports current and batch-specific?

Third-party testing means an outside laboratory performed a test. It does not necessarily mean the third party independently selected the sample, designed the specification, audited manufacturing, or verified the supply chain.

The phrase should be unpacked.

The practical COA review

A serious review can be conducted in sequence.

1. Match the batch

No match, no reliable conclusion about the delivered lot.

2. Clarify the chemical identity

Sequence, modification, salt form, and naming should agree.

3. Identify the methods

Which questions were actually tested?

4. Inspect the raw presentation

Look beyond the final percentage at chromatograms, spectra, conditions, and calculations.

5. Separate purity from assay

Do not infer milligrams from peak-area percentage without appropriate quantitative evidence.

6. Examine impurities

Are important peaks identified or merely summed?

7. Ask about untested attributes

Sterility, endotoxin, content, water, solvents, aggregation, stability, and particulates are separate questions.

8. Verify the laboratory and report

Confirm traceability when the decision requires it.

9. Match the evidence to the research use

The required characterization depends on what the laboratory intends to investigate.

What the number can honestly say

A careful interpretation of a 99.4 percent HPLC result might read:

Under the stated chromatographic conditions, the designated principal peak represented approximately 99.4 percent of the integrated detector response in the tested sample. Separate evidence is required to establish definitive identity, absolute content, relevant impurity characterization, microbiological attributes, stability, correspondence to the delivered lot, and suitability for the intended research.

That sentence is not good packaging copy.

It is good science.

A certificate of analysis can be a meaningful piece of evidence. It becomes misleading only when it is asked to certify conclusions the laboratory never tested.

The strongest buyers do not ask whether a company “has COAs.”

They ask whether the documentation answers the question that actually matters.

---

Selected references

• International Council for Harmonisation: Q2(R2) analytical procedure validation
• International Council for Harmonisation: Q1A(R2) stability testing
• United States Pharmacopeia: Reference standards for synthetic peptide therapeutics
• U.S. Food and Drug Administration: Analytical approaches to peptide impurities

Continue the Journal

• Peptides Without the Hype
• Evidence Before Hype: How to Read a Longevity Claim
• Measure What Matters: Biomarkers Without Obsession