Factory COA vs an outside lab: why the numbers differ, and when the gap matters

Manufacturer COA vs third-party test: the short answer

A manufacturer’s COA and a third-party test can both be correct and still disagree. They often measure different things, on different samples, at different times, against different standards. Five causes explain most gaps:

Cause	Which way it usually pushes	How to check
Net peptide or total powder	Total powder reads higher	The basis stated on each report
Reference standard	Either way, by a fixed proportion	The standard’s source and assigned content
Salt and water	Their weight reads as peptide on a powder basis	Counter-ion and water results
HPLC method and integration	Purity moves by points	Wavelength, chromatogram, peak table
Sample and age	The later test usually reads lower	Dates, lot, storage, number of vials

Read both reports with this table beside you. For a general guide to every part of a report, see how to read a COA.

1. Net peptide or total powder

Lyophilised peptide powder is not all peptide. It also holds counter-ion salt and water. Bachem’s worked example: 5 mg of powder with 85% net peptide content and 98.7% purity holds 4.195 mg of the target peptide.

A report can express mg on either basis. The EMA guideline on synthetic peptides expresses assay on a counter-ion-free, anhydrous basis and defines strength as the mass of peptide base. Some reports instead give the mass of powder in the vial, or that mass multiplied by HPLC purity. Those numbers include salt and water.

Peptigrity, which compares measured mg with label claims, found that about 40% of the certificates it tracked in September 2026 read over 110% of the label. It treats high readings with caution because gross-mass reporting inflates them. Net peptide content in research-grade peptides is typically about 70% to 90% of the powder. A factory COA on a powder basis and an outside lab on a net peptide basis can therefore differ by 10% to 30% with no fault on either side. Content per vial is reported as free base, against a reference standard. For more, see net peptide content.

2. The reference standard

Content is measured by comparing the sample with a reference standard of known content. If the standard’s assigned content is off, every result from that standard is off by the same proportion. Finnrick reported in May 2026 that one lab issued five COAs that were all off by the same proportion, a pattern that pointed to a reference standard calibration error.

Our research found no pharmacopoeial monograph for newer peptides such as retatrutide, so labs use in-house methods with standards they buy or make. Labs assign those standards in different ways. Janoshik, for example, told Krysia’s Substack (August 2026) that it standardises its own reference standards by CHNS analysis and NMR and reports net peptide. Check whether the factory COA says where its standard came from and how its content was assigned.

3. Salt and water

Counter-ion results show how much of the powder is TFA or acetate salt. Water content by Karl Fischer shows how much is water. Both add weight without adding peptide.

The USP authors show how the parts add up. In their worked mass balance, 99% chromatographic purity multiplied by the 94% of powder left after acetate, TFA, solvents and residue gives 0.93 mg of peptide per mg of powder, on a dry basis. They opened vials under dry nitrogen for water testing, because peptides take up water from air. A powder that sat open, or travelled in humid conditions, can weigh more and hold proportionally less peptide.

4. The HPLC method

ICH Q6B, written for biotech proteins, calls purity “method-dependent”, and the same holds for synthetic peptides. HPLC purity is the main peak’s share of UV peak area at one wavelength, and it changes with the column, the gradient, the wavelength and how peaks are integrated.

Integration alone can move a result. Finnrick’s corrections log records a 92.83% purity in March 2026 that came from noise integrated as peaks; the lab reissued the COA. In a published comparison, two labs reported 99.985% and 97.779% purity for the same tesamorelin batch (Krysia’s Substack, August 2026). Without a chromatogram and a peak table, you cannot tell which choices sit behind a factory’s purity figure. See what purity means.

5. Sample and age

A factory COA describes a sample the factory took, usually at release. Your result describes the vials you sent, after filling, shipping and storage. Two effects follow:

• Fill differs between vials. One vial tells you about one vial. Test several vials from the batch to see the spread.
• Peptides degrade. Heat, moisture and time lower purity and content. A COA dated months before your test can be accurate for its day and still read high now. See peptide degradation.

Buyers face the same gap one step down the chain, between a seller’s COA and their own vial: see vendor COA vs testing your own vial.

What published same-batch comparisons show

Outside labs also disagree with each other, which sets a baseline for what to expect against a factory COA. Three published single cases:

Published case	Product	Spread
graymarket, September 2025	Retatrutide, one batch, three labs	One lab read about 15% lower than the other two
Krysia’s Substack, August 2026	Tirzepatide labelled 60 mg, three labs	62.3 mg to 75.44 mg
Peptide Critic, multi-lab comparison	Retatrutide, one batch, five labs plus the seller’s report	10.18 mg to 12.80 mg, about 26%

These cases do not show that any one lab is wrong. They show that basis, standards and methods move results, so a single number is never the whole story. For the wider picture, see why labs disagree.

When the gap is a real problem

Some differences no method choice can explain. Treat the batch as failed, or test it again with more vials, if:

• The mass is wrong. An identity result that does not match the expected mass means a different molecule.
• The peptide is missing. No main peak, or a peak at the wrong mass.
• Content is far below the label on every basis. Even after salt, water and method are allowed for.
• Large unknown peaks appear that the factory chromatogram does not show.
• The lot does not match between the COA and your vials.

For the checks to run on the factory document itself, see checking a Chinese manufacturer’s COA.

How HK Analytical reports make the comparison possible

Our lab is in Hong Kong, and samples from mainland China reach it by cross-border courier. Our reports give content per vial as free base against a reference standard, purity with the full chromatogram, and identity as observed against expected mass, and every number comes with its raw data. Counter-ion and water content can be added to the same order, so salt and water are measured, not assumed. The example report shows the layout. To compare a factory COA with your own vials, request a quote.