The Creatine Purity Problem: Why COA Nitrogen Values Don't Tell the Whole Story

Creatine monohydrate has over 1,000 peer-reviewed studies behind it, a safety profile that even skeptical researchers tend to respect, and a global market estimated at over $700 million annually. It is, by most measures, one of the most straightforward sports nutrition ingredients in the category.

So it’s worth asking: why do third-party market surveys still find creatine products with actual purity levels ranging from 85% to 99.9%, depending on the manufacturer and origin?

The answer isn’t usually outright counterfeiting. It’s something quieter — a persistent mismatch between what standard certificates of analysis actually measure and what they need to measure to confirm creatine purity. We see this at Qalitex with enough regularity that it warrants a thorough explanation. Because brands that understand the gap can close it. Brands that don’t keep accepting supplier COAs that look fine right up until they don’t.

The Nitrogen Testing Trap

Most basic creatine COAs report purity via total Kjeldahl nitrogen (TKN) analysis. The underlying logic is sound on its face: creatine monohydrate (C₄H₉N₃O₃) contains 26.4% nitrogen by mass, so a sample with the correct nitrogen percentage should, in theory, be predominantly creatine.

For raw material and ingredient-level verification, Ayah Labs specializes in contract testing and supplier qualification.

The problem is that several structurally related compounds have similar — or significantly higher — nitrogen content.

Creatinine, the cyclic dehydration metabolite that forms when creatine loses a water molecule, contains 31.1% nitrogen by mass. Glycocyamine (guanidinoacetic acid), a direct biosynthetic precursor to creatine, has a comparable nitrogen profile. Dicyandiamide (DCD), a byproduct that accumulates when creatine synthesis isn’t tightly controlled, runs around 66.6% nitrogen — the same value as melamine, which is exactly why melamine became the adulterant of choice during the 2008 Chinese protein scandal.

A lot blended with even 6–8% creatinine can produce a Kjeldahl nitrogen reading that falls squarely within the expected range for pure creatine monohydrate. The number looks right. The supplier COA gets signed off. The product ships. But the actual creatine content is materially lower than what’s on the label, and the creatinine content — a metabolic waste product with no ergogenic value — is elevated before the consumer opens the bottle.

This isn’t theoretical. Independent testing programs, including ConsumerLab’s ongoing supplement reviews, have repeatedly identified commercial creatine products with creatinine impurity levels that range from trace quantities up to values representing a material label inaccuracy under 21 CFR § 101.36. The variation doesn’t track neatly with price point or country of origin. It tracks with whether the manufacturer was using HPLC or nitrogen analysis at the time of release.

What HPLC Purity Testing Actually Finds

High-performance liquid chromatography separates compounds by their differential interactions with a stationary phase, then quantifies each one independently. Instead of summing nitrogen and inferring purity, you get a chromatographic map of what’s actually in the sample — peak by peak.

A properly executed HPLC purity assay for creatine monohydrate will quantify the following simultaneously:

• Creatine monohydrate (the labeled compound)
• Creatinine (dehydration metabolite — the most common impurity)
• Glycocyamine / guanidinoacetic acid (synthesis precursor)
• Dicyandiamide (synthesis byproduct, abbreviated DCD)

The USP creatine monohydrate monograph specifies HPLC as the identification and assay method and sets clear acceptance criteria: creatinine NMT 0.2% (w/w), and assay content between 98.0% and 102.0% of labeled creatine monohydrate. Those limits exist because USP’s expert committees looked at the chemistry and identified exactly where this ingredient fails.

A product “tested” by nitrogen analysis alone cannot demonstrate compliance with the USP creatinine limit. It literally cannot — because the method doesn’t measure creatinine. The purity value on that COA is a nitrogen-derived estimate, not a compound-specific assay.

It’s also worth being direct about dicyandiamide. DCD accumulates during the industrial synthesis of creatine when reaction conditions — temperature, pH, reaction time — aren’t tightly controlled. At elevated concentrations, DCD has been flagged for potential renal and thyroid effects in animal feeding studies, which is precisely why NSF International includes DCD screening in the NSF Certified for Sport program’s testing requirements. If your supplier’s COA doesn’t report DCD, it’s not because DCD isn’t present. It’s because the test wasn’t run.

What 21 CFR Part 111 Actually Requires

Under the FDA’s dietary supplement current Good Manufacturing Practice regulations at 21 CFR Part 111, every manufacturer and brand owner has explicit identity testing obligations for incoming ingredients — obligations that don’t disappear because a supplier provides paperwork.

Subpart E of Part 111 requires that each incoming component be tested to establish its identity before it’s used in manufacturing. The FDA permits brands to rely on a supplier’s COA to support release decisions, but only after establishing the reliability of that supplier through their own testing. The standard language is worth understanding precisely: you must conduct at least one test to verify the identity of the component yourself, using a scientifically valid method. And you must periodically reconfirm supplier performance.

In practice, many brands accept a supplier’s nitrogen-based COA as a substitute for independent identity testing. That’s not what the regulation allows, and it’s not what an FDA investigator will accept during an inspection. A Form 483 observation for “failure to test incoming components for identity” is among the most common findings in supplement facility inspections — and it’s entirely avoidable.

The evidentiary standard that holds up best in an enforcement context is testing conducted by an ISO/IEC 17025-accredited third-party laboratory. ISO 17025 accreditation means the lab’s methods have been independently validated, its calibration is traceable to national standards, and its data chain-of-custody meets an auditable standard. A supplier’s in-house COA — even a good one — doesn’t carry the same weight, and brands that treat the two as equivalent are taking on risk that isn’t visible until a recall or enforcement action makes it visible.

Reading a Creatine COA Like a Lab Professional

Most brands review a COA and focus on three things: the purity percentage, the heavy metals values, and the microbial counts. For many ingredients, that’s a reasonable shortlist. For creatine monohydrate, it systematically misses the most predictable failure mode.

Here’s what to look for specifically:

Test method for assay. The method should identify HPLC explicitly. “Total nitrogen,” “Kjeldahl,” or “by calculation” in the method field means the purity number is not creatine-specific. Require an HPLC-based assay before accepting the lot.

Creatinine — explicitly reported. The COA should show a separate line for creatinine with a quantified value, not just a pass/fail. Per USP criteria, that value should be below 0.2% (w/w). A COA that says “purity: 99.9%” with no creatinine breakdown is not telling you what you need to know.

Dicyandiamide. Not universally listed on standard COAs, but worth requiring — especially for raw material sourced from Chinese manufacturers, which represents the large majority of global creatine production. If your product is seeking NSF Certified for Sport or Informed Sport certification, DCD testing will be required anyway. Getting it done on incoming lots is the more efficient path.

Moisture / water of crystallization. Creatine monohydrate contains one molecule of crystalline water, which accounts for approximately 11.9% of its molecular weight. Product that has partially dehydrated, or that was mislabeled to begin with, delivers less creatine per gram than stated. Karl Fischer titration or thermogravimetric analysis (TGA) for moisture content is a low-cost addition to the protocol and catches this scenario cleanly.

Issuing laboratory. Who signed the COA matters. An ISO 17025-accredited external lab and a manufacturer’s internal QC department are not equivalent, even if they ran the same methods. Know which one you’re looking at.

At Qalitex, our standard creatine monohydrate testing protocol covers HPLC identity and purity per USP methodology, explicit creatinine quantification, DCD screening, Karl Fischer moisture analysis, and a four-element heavy metals panel via ICP-MS (lead, arsenic, cadmium, mercury). For clients supporting third-party certification programs — NSF, Informed Sport, or Banned Substances Control Group — we can align the test scope with each program’s specific requirements from the outset.

The Certification Layer: What It Actually Adds

NSF Certified for Sport and Informed Sport are the two most widely recognized third-party programs in the sports nutrition category. Both require HPLC-based identity and purity testing, LC-MS/MS banned substance screening across hundreds of compounds, and regular facility audits. A creatine product certified under either program has been through more rigorous scrutiny than anything a standard supplier COA provides.

What certification adds beyond methodology is accountability. Both programs maintain publicly searchable product databases. If a product fails post-market testing, ends up in an FDA MedWatch adverse event report, or gets flagged by a professional sports organization, certification at an ISO 17025-accredited lab is the clearest available evidence that the brand exercised reasonable diligence. It doesn’t prevent every problem, but it changes the conversation considerably.

That accountability matters because dietary supplements operate on a self-certification model. There’s no pre-market approval pathway. The brand is legally responsible for ensuring safety and label accuracy before the first unit ships — and that responsibility sits with the brand even when the manufacturing, the raw material supply chain, and the testing were all outsourced.

Creatine purity testing done properly — HPLC assay, creatinine quantification, DCD screening, moisture analysis — runs 3 to 4 business days at a competent ISO 17025 lab. The cost is a fraction of a percent of what a single product recall, a deactivated Amazon listing, or a state AG enforcement action would cost. The only reason to skip it is not knowing the risk is there.

And now you do.