Peptide Impurity Thresholds: What a COA Purity Number Does Not Tell You

Peptide impurity thresholds for synthetic peptides are not the ICH Q3A figures most suppliers quote. ICH Q3A(R2) explicitly excludes peptides from its scope, and the framework that actually governs them is the European Pharmacopoeia general monograph, which requires peptide-related impurities to be reported above 0.1 percent, identified above 0.5 percent, and qualified above 1.0 percent. That difference matters, because it means a headline purity figure can legally sit alongside a half-percent impurity that nobody has ever structurally characterised.

Most discussion of research peptide quality stops at a single number. A certificate of analysis reports 98.4 percent or 99.1 percent by reversed-phase HPLC, the number clears some internal bar, and the conversation ends. The number is real, but it is the least informative part of the document. What determines whether a compound will behave reproducibly in a receptor binding assay is the composition of the remaining one or two percent, and whether anyone bothered to find out what is in it.

Why ICH Q3A Does Not Apply to Synthetic Peptides

ICH Q3A(R2), the harmonised guideline on impurities in new drug substances adopted at Step 4 on 25 October 2006, is the document most commonly cited when impurity limits come up. Its threshold table is widely reproduced: for substances in the daily intake tier at or below two grams per day, impurities are reported above 0.05 percent, identified above 0.10 percent, and qualified above 0.15 percent.

Those numbers are correct, and they are irrelevant here. The preamble of Q3A(R2) states plainly that the guideline covers new drug substances produced by chemical synthesis, then lists what falls outside it: biological and biotechnological products, peptide, oligonucleotide, radiopharmaceutical, fermentation products and semi-synthetic products derived from them, herbal products, and crude products of animal or plant origin. Peptides are named in the exclusion list. A supplier quoting a 0.10 percent identification threshold for a synthetic peptide is applying a framework that was written to exclude the very class of molecule in question.

The exclusion is not an oversight. Small molecule impurity control assumes a manufacturing process that produces a modest number of structurally distinct byproducts. Solid-phase peptide synthesis does the opposite. Every coupling and every deprotection step is an opportunity to generate a species that differs from the target by a single residue, a single stereocentre, or a protecting group that failed to leave. The resulting impurity population is large, closely related to the parent, and chromatographically difficult to separate from it.

The Thresholds That Actually Apply

The European Medicines Agency resolved the ambiguity with the Guideline on the Development and Manufacture of Synthetic Peptides, reference EMA/CHMP/CVMP/QWP/367182/2025. It was adopted by the CHMP on 1 December 2025 and by the CVMP on 4 December 2025, and it came into effect on 1 June 2026. The guideline states directly that synthetic peptides are excluded from the scope of ICH Q3A and that consequently the limits laid down in that guideline are not applicable.

In their place, the guideline points to the European Pharmacopoeia general monograph Substances for Pharmaceutical Use. Under that monograph, peptide-related impurities should be reported above 0.1 percent, identified above 0.5 percent, and qualified above 1.0 percent. The same thresholds carry through to finished products.

Reporting, Identification, and Qualification Are Three Different Claims

These three words are routinely used interchangeably in supplier marketing, and they are not synonyms. ICH Q3A(R2) defines each precisely, and the definitions transfer cleanly to the pharmacopoeial framework.

A reporting threshold is the level above which an impurity must appear in the analytical record at all. Below it, the peak is simply not carried forward. An identification threshold is the level above which the structure of that impurity must be determined, meaning someone has to establish what the molecule actually is rather than describing it by retention time. A qualification threshold is the level above which biological safety data are required to support the amount present.

The practical consequence for synthetic peptides is a wide identification gap. Under the pharmacopoeial framework, an impurity present at 0.4 percent sits above the reporting threshold but below the identification threshold. It is a real, measurable component of the material, it appears on a properly constructed impurity table, and no regulatory expectation requires anyone to know what it is. Compare that to the 0.10 percent identification threshold under ICH Q3A and the peptide framework is roughly five times more permissive at exactly the point where structural knowledge begins.

What a 98 Percent Purity Figure Leaves Undefined

Run the arithmetic against those thresholds. A peptide reported at 98.0 percent purity carries 2.0 percent of material that is not the target compound. That 2.0 percent could be a single impurity, in which case it sits at twice the pharmacopoeial qualification threshold and well above the identification threshold. It could equally be twenty separate species at 0.1 percent each, none of which individually triggers identification. Both scenarios produce the same headline number and represent entirely different materials.

This is why the impurity table, not the purity percentage, is the informative part of a certificate. A document that reports a single assay value with no accompanying breakdown of individual and total impurities has not told you whether the non-target fraction was characterised, estimated, or simply subtracted. Our approach to third-party COA documentation treats the impurity profile and the chromatogram as the substance of the report rather than an appendix to it.

Area Percent Assumes Every Impurity Absorbs Light Equally

There is a second assumption buried in almost every reported purity figure. Purity by RP-HPLC is normally calculated by area normalisation, where each peak area is expressed as a percentage of total integrated area. That calculation is only accurate if the target peptide and every impurity produce the same detector response per unit mass.

ICH Q3A(R2) acknowledges this explicitly in its section on analytical procedures. It permits acceptance criteria based on analytical assumptions such as equivalent detector response, while noting that where the response factors of the drug substance and the relevant impurity are not close, a correction factor should be applied or the impurities should in fact be overestimated. The assumption is allowed, but it is supposed to be examined and stated.

For peptides the assumption is imperfect in a specific and predictable direction. Detection at 214 nm responds primarily to the amide bond, so response scales roughly with peptide length, and a deletion sequence missing one residue generates slightly less signal than the parent. Detection at 280 nm depends on aromatic residues, so an impurity that has lost a tryptophan or tyrosine can absorb dramatically less than the target while being present in substantial quantity. In both cases the error runs the same way. The impurity is understated and the reported purity is flattered. This is one reason orthogonal techniques matter, and why quantitative NMR and mass-based methods are used to check what area normalisation reports.

Co-Elution and the Limits of a Single Method

The EMA guideline addresses a failure mode that a single chromatogram cannot reveal. It states that where one analytical method is not appropriate to separate all impurities, additional independent methods may be needed, and that when co-eluting impurities are observed as one peak, the 1.0 percent qualification threshold applies to that combined peak unless otherwise justified.

That instruction exists because peptide impurities are structurally close to the parent. A single-residue deletion, an epimer at one stereocentre, or a deamidation product can differ from the target by a fraction of a minute in retention time, or not at all under a given gradient. A method that has not been challenged with forced degradation samples may be reporting a clean 99 percent peak that contains three species. The analytical framework for demonstrating that a method can actually separate what it claims to separate is set out in ICH Q2, and the specifics of that exercise are covered in our discussion of peptide analytical method validation.

A Documented Example from Tirzepatide Synthesis

The impurity classes described above are not hypothetical. Wang and colleagues published a mechanistic study of diketopiperazine formation during solid-phase peptide synthesis of tirzepatide in ACS Omega in 2022, volume 7, issue 50, pages 46809 to 46824, with authors from Eli Lilly and Company.

The study traced the formation of diketopiperazine byproducts and the associated double-amino-acid deletion impurities to specific points in the synthesis. Diketopiperazine formation occurred principally during Fmoc deprotection and during post-coupling aging of the unstable Fmoc-Pro-Pro-Ser-resin intermediate, and the authors observed comparable behaviour at other intermediates containing a penultimate proline. They also reported that Fmoc deprotection could proceed in dimethylformamide, dimethyl sulfoxide, N-methyl-2-pyrrolidone, and acetonitrile without added piperidine, which explains why aging an intermediate in solvent alone can generate the impurity. Substituting Bsmoc-protected amino acids eliminated the diketopiperazine byproducts.

The relevance is direct. A double-amino-acid deletion impurity in a compound the size of tirzepatide differs from the target by a small fraction of total molecular mass and is a prime candidate for co-elution. It is precisely the species that an unresolved chromatogram will fold into the main peak, and precisely the species that a purity figure alone will never disclose. The broader catalogue of peptide-related impurity classes, including truncation, deletion, insertion, and racemisation products, was reviewed comprehensively by D’Hondt and colleagues in the Journal of Pharmaceutical and Biomedical Analysis in 2014, volume 101, pages 2 to 30.

One further distinction is worth drawing. The pharmacopoeial thresholds above govern peptide-related impurities, meaning species structurally derived from the synthesis of the target sequence. Process-related contaminants sit under separate frameworks with their own acceptance criteria. Organic solvents carried through from coupling and cleavage are controlled under ICH Q3C and measured by gas chromatography headspace analysis. Catalytic and equipment-derived metals fall under ICH Q3D and are quantified by ICP-MS. The trifluoroacetate counterion introduced during reversed-phase purification is a third category again, reported as residual TFA content. A certificate that reports chromatographic purity alone has addressed one of these four categories.

Key Research Findings

• ICH Q3A(R2), adopted at Step 4 on 25 October 2006, names peptides in its scope exclusion list alongside biological, oligonucleotide, radiopharmaceutical, and fermentation products.
• ICH Q3A(R2) thresholds for the daily intake tier at or below two grams per day are 0.05 percent reporting, 0.10 percent identification, and 0.15 percent qualification. These do not apply to synthetic peptides.
• EMA/CHMP/CVMP/QWP/367182/2025 was adopted by the CHMP on 1 December 2025 and came into effect on 1 June 2026, confirming that ICH Q3A limits are not applicable to synthetic peptides.
• European Pharmacopoeia thresholds for peptide-related impurities: report above 0.1 percent, identify above 0.5 percent, qualify above 1.0 percent.
• Co-eluting impurities observed as a single peak are subject to the 1.0 percent qualification threshold unless otherwise justified, per EMA/CHMP/CVMP/QWP/367182/2025.
• Wang et al., ACS Omega, 2022, 7(50), 46809 to 46824, traced tirzepatide diketopiperazine and double-deletion impurity formation to Fmoc deprotection and post-coupling aging at Fmoc-Pro-Pro-Ser-resin, and eliminated it using Bsmoc protection.
• D’Hondt et al., Journal of Pharmaceutical and Biomedical Analysis, 2014, 101, 2 to 30, reviewed peptide-related impurity classes and the analytical techniques used to profile them.

Applying This to Certificate Evaluation

Three questions follow from the framework above. First, does the certificate report individual and total impurities, or only an assay value? A single number cannot distinguish one 2.0 percent impurity from twenty at 0.1 percent. Second, is a chromatogram included, and does it show a baseline that returns cleanly between peaks? Integration decisions are invisible in a summary table. Third, was more than one analytical principle applied, so that co-elution under a single gradient does not pass unnoticed?

None of this requires a reader to distrust chromatography. Reversed-phase HPLC is the correct primary tool, and a well-run method with a validated separation is highly informative. The point is narrower. The purity percentage is a summary statistic that rests on a threshold framework most people misattribute, an equal-response assumption that peptides violate in a known direction, and a separation that may or may not have been challenged. Reading the impurity table, and knowing which thresholds govern it, is what turns that summary statistic back into information. Our guidance on reading a certificate of analysis walks through the document section by section.

For research purposes only. Not for human consumption. Not for diagnostic or therapeutic use.