“Tested to USP standards” sounds like one settled thing. It isn’t. USP testing standards are a set of separate tests, and each one answers a different question about what’s in the vial: is it the right molecule, how much of it is there, what else came along for the ride, and is it clean enough to formulate with?
This post walks through the tests you’ll most often see on an API certificate of analysis, what each one is really telling you, and how to read the paperwork without simply taking anyone’s word for it. It’s written for pharmacists and procurement teams, not analytical chemists, so the jargon stays short.
What USP is, and what it isn’t
USP, the United States Pharmacopeia, writes public quality standards for drug substances and drug products, and FDA can enforce them. In practice that means two kinds of documents. A monograph is the specification for one named substance: which tests to run, how to run them, and what results count as acceptable. A general chapter covers a method or topic that many monographs share, like chromatography or endotoxin testing, so each monograph doesn’t have to spell it out again.
Here’s the part that trips people up. Plenty of APIs don’t have a monograph at all, and that can be true of newer peptides in particular. For synthetic peptides, USP has an informational chapter, 〈1503〉, on the quality attributes of peptide drug substances. A proposed companion, 〈1504〉, covers starting materials, and USP’s 2023 list still shows it as in development. Informational means guidance. It describes attributes worth controlling and isn’t a pass/fail specification for any one product.
So “USP grade” or “meets USP” only means something when it arrives with a named monograph and real numbers for the lot in front of you. Without those, it’s a label, not a result.
The tests you’ll see on a COA, and what each one is for
None of this is exotic. It’s just rarely explained in one place. Here are the rows that show up most often on an API certificate of analysis, with the question each one is really asking.
| Test | The question it answers | What to look for |
|---|---|---|
| Appearance | Does the material look the way the spec says it should? | A real description of color and form, not just “conforms.” |
| Identification | Is it the right molecule? | HPLC retention time against a reference standard, plus mass spectrometry. For peptides, two independent methods is the norm, because one alone rarely settles it. Mass spec can’t separate some isomers, for example. |
| Assay | How much of the active is actually in there? | A number with units and a stated basis. HPLC against a quantitative standard is preferred. Amino acid analysis (see USP 〈1052〉) and nitrogen content are also used for peptides. |
| Purity and related substances | What else came through synthesis and purification? | HPLC, with the general method rules in USP 〈621〉. Look for individual impurities and their limits, not one headline percentage. |
| Water content | How much of the powder’s weight is water? | Usually Karl Fischer titration (USP 〈921〉). Peptides pick up moisture readily, so this number matters. |
| Counterion content | How much of the weight is the salt partner, such as acetate or TFA? | Peptides are often supplied as salts, and the counterion has weight. USP has procedural chapters for acetic acid (〈503〉) and TFA (〈503.1〉) in peptides, and its peptide expert panel calls counterion and water the most important specific tests. |
| Residual solvents | Are leftover process solvents under the safety limits? | Gas chromatography against the limits in USP 〈467〉. USP’s general notices apply residual-solvent control even when a monograph lists no test. |
| Elemental impurities | Any metals from catalysts, equipment, or raw materials? | USP 〈232〉 sets limits and 〈233〉 gives the procedures. They replaced the old heavy-metals color test. |
| Endotoxin and bioburden | Is it suitable for injectable use? | Endotoxin by USP 〈85〉, microbial counts by USP 〈61〉 and 〈62〉. The peptide panel says APIs headed for injectables need both. If neither is listed, ask why. |
Not every test gets run on every lot. For synthetic peptide APIs, the USP panel’s position is that routine lot release must at least establish identity, purity, and assay. The rest are chosen based on the product, the process, and where the material is going. So a short list of tests isn’t automatically a red flag. A short list with no explanation is.
How to read a COA like a skeptic
The habit that catches the most problems is simple: treat the COA as a set of numbers that have to agree with each other, not a row of green checkmarks. Start with the basics.
- The lot number matches your vial and your invoice. A COA for “the product” in general isn’t a COA for your lot.
- The manufacturing site is named. If it only names a brand or a trading company, you can’t tell where the material came from.
- Results are numbers sitting next to limits. “Conforms” says very little. A result like 98.6% against a limit of not less than 95.0% says a lot. (Those figures are only an example.)
- Each test names its method. A USP chapter, a monograph, or the manufacturer’s validated procedure.
- The dates are there. Date tested, date manufactured, and a retest or expiry date.
- Someone signed it. An authorized quality representative, with a way to reach them.
Then do the one check most people skip. HPLC purity is usually reported as a peak-area percentage. It tells you how much of the detected peptide material is the right peptide. It says nothing about the water and salt weighing down the powder, so a lot can read 99% pure and still be well under 99% peptide by weight.
USP’s peptide expert panel recommends a mass balance to close that gap. Add the assay (peptide content), the counterion content, and the water content. The total should land near 100%. If a COA gives you purity but none of the other three, you can’t run the check, and that’s worth a question.
A few things should slow you down:
- “USP grade” or “meets USP” with no monograph named
- Purity listed on its own, with no assay, water, or counterion result
- A COA you can’t trace back to the manufacturing site, especially one issued by a reseller
- A missing lot number, test date, or signature
- Identical results across different lots, down to the last decimal
Any one of these can have an innocent explanation. Two or three together is a reason to pause the order and ask.
When there’s no monograph
No monograph doesn’t mean no standard. It means the standard is the manufacturer’s own specification, and that’s where the useful questions are.
USP’s peptide guidance gives a menu of attributes worth controlling: identity, assay, purity, counterion, water, and the rest of the table above. The manufacturer picks the tests, sets the limits, and has to show the methods are fit for purpose. In the expert panel’s guidance, the HPLC purity method should be validated (USP 〈1225〉 is the reference), at least one method should be stability-indicating, and when more than one method is used they should separate by different principles.
The weak spot is impurity limits. For peptide-related impurities, the panel noted that no generally accepted guidance exists, so the levels get set case by case, based on safety data for that product. Two COAs for the same peptide can therefore carry different impurity limits and both be legitimate. What you want is a manufacturer that can tell you what its limits are and why it chose them.
With a non-monograph peptide, ask for three things: the specification (tests and limits), the method behind each test, and how the limits were justified. A serious manufacturer has those answers on file. The panel wrote its guidance in 2014 and USP has issued more since, so check the current USP-NF text before you rely on any specific point.
What this means if you’re buying for compounding
If you compound under section 503A of the FD&C Act, testing is only part of the picture, and FDA’s rules put eligibility and paperwork first. State-licensed pharmacists and physicians may compound only with bulk drug substances that meet one of three conditions. They comply with an applicable USP or NF monograph if one exists. Or they’re a component of an FDA-approved drug when no monograph exists. Or they appear on FDA’s 503A bulks list when neither of those applies.
On top of that, FDA says the substance must come with a valid certificate of analysis and must have been made by an establishment registered with FDA. It also tells compounders to know their bulks suppliers.
Two practical takeaways. A spotless COA doesn’t make a substance eligible, because eligibility is a separate question that gets answered before testing comes into it. And “valid COA” and “registered manufacturer” are things you can check rather than assume: read the COA the way the last section describes, and confirm the manufacturer’s FDA registration yourself.
If you’re buying for research rather than compounding, the 503A rules above are specific to compounding. The documentation habits still carry over. Your own QA or procurement policy usually sets the bar, and a lot-specific COA with real numbers is the common denominator.
This is general information, not legal or regulatory advice. Your counsel and your state board of pharmacy have the final word on what your pharmacy needs.
Frequently asked questions
What does “USP grade” actually mean?
It should mean the lot was tested against a specific USP monograph and met it. If the document doesn’t name the monograph and show the lot’s results, treat “USP grade” as a label rather than a result.
Does every API have a USP monograph?
No. Plenty don’t, and that can be true of newer peptides in particular. USP has an informational chapter, 〈1503〉, on quality attributes of synthetic peptide drug substances, but that’s guidance rather than a specification for any one product.
What’s the difference between HPLC purity and peptide content?
Purity is usually a peak-area percentage: the share of the detected peptide material that is the right peptide. Peptide content, or assay, is how much of the powder’s weight is the peptide. Water and counterions account for the difference, which is why the mass balance check is worth doing.
Is a certificate of analysis enough?
It’s the starting point, and for compounding under section 503A, FDA requires a valid one. But a COA doesn’t establish that a substance is eligible for compounding, and it’s only as trustworthy as its source. Check that it’s lot-specific, shows real numbers and methods, and traces back to a registered manufacturing site.
The short version
For any lot, ask three things: what each test was, what method ran it, and whether the numbers add up. If the paperwork can answer those, the rest of the conversation gets easier.
You can see how we approach documentation on our Quality & Compliance page, browse the API catalog, or request a quote and tell us which documents you need to see before you order.
Sources
- FDA, Bulk Drug Substances Used in Compounding Under Section 503A of the FD&C Act (content current as of May 14, 2026)
- USP, Peptide-related USP-NF Documentary Standards and USP Reference Standards (effective December 22, 2023)
- USP Therapeutic Peptides Expert Panel, Control Strategies for Synthetic Therapeutic Peptide APIs, Part I: Analytical Consideration, Pharmaceutical Technology, March 2014