Skip to content
Peptide Science Journal
Regulatory & Standards

Impurity thresholds for synthetic peptides: the number underneath the purity figure

A certificate of analysis reporting 99.2% purity makes a claim about 99.2% of the material. The remaining 0.8% is described only by subtraction. Whether that residue is a single well-characterized by-product or two dozen unresolved species is not something the purity figure can distinguish, and the difference matters more than the headline number does.

Regulatory frameworks address this by defining thresholds: concentration levels above which a specific obligation attaches. Below the threshold, an impurity may be reported as an anonymous fraction of the total. Above it, someone must name it, determine its structure, or demonstrate that it is safe. Thresholds are the mechanism that converts a purity percentage into an actual description of a material. For synthetic peptides, the threshold framework was revised at the end of July 2026, and the revision is a useful occasion to examine what these numbers are doing.

Three thresholds, and what each one demands

The general pharmaceutical framework comes from ICH Q3A, which governs impurities in new drug substances. It defines three tiers, each attached to a different obligation.

The reporting threshold is the level above which an impurity must appear in documentation at all. The identification threshold is the level above which its structure must be determined. The qualification threshold is the level above which its biological safety must be established, typically through toxicological data. An impurity present below the reporting threshold is, for regulatory purposes, part of an undifferentiated remainder. One above the qualification threshold requires an evidentiary case of its own.

The values scale inversely with how much drug substance is administered, on the straightforward logic that a fixed percentage of a large mass is a larger absolute exposure. For a drug substance with a maximum daily dose of 2 grams or less, ICH Q3A sets the reporting threshold at 0.05%, the identification threshold at 0.10%, and the qualification threshold at 0.15% — with the identification and qualification tiers also capped in absolute terms, so that whichever constraint binds first applies. Higher administered masses pull the percentages down.

The structural point is that these numbers are not measurements. They are decisions about where regulatory attention stops, calibrated to a toxicological model in which risk scales with quantity. That model fits small organic molecules well. It fits peptides considerably less well.

Why peptides required different logic

The dominant safety concern for a peptide impurity is usually not toxicity. It is immunogenicity — the possibility that a structurally related variant provokes an immune response, and that the response cross-reacts with the intended sequence or with an endogenous counterpart. This risk does not scale cleanly with mass. A sequence variant present at a low level can carry a T-cell epitope that the parent sequence does not, and a quantity-based threshold has no way to see that.

FDA addressed this in a 2021 guidance covering abbreviated applications for synthetic versions of five peptides previously approved as recombinant products: glucagon, liraglutide, nesiritide, teriparatide, and teduglutide. Rather than applying an absolute toxicological threshold, the guidance built a comparative framework. A proposed synthetic product had to show that every peptide-related impurity it shared with the reference product was present at the same level or lower. For any new impurity — one absent from the reference product entirely — the level had to be no more than 0.5% of the drug substance, and each such impurity had to be individually characterized and justified. Above 0.5%, the guidance treated the immunogenicity question as one that abbreviated applications could not resolve.

Beneath that sat a second number. Applicants were directed to identify each peptide-related impurity present at 0.10% of the drug substance or greater, with “identify” defined explicitly as characterizing the structure. The guidance noted that identification in the 0.10%–0.5% band exists specifically to support assessment of immunogenicity risk, and that applicants might be asked to look below 0.10% depending on the product.

Two features of this framework are worth isolating. First, the reference point is another product rather than an absolute limit, so the same impurity at the same concentration can be acceptable in one context and disqualifying in another. Second, the justification standard is not merely quantitative: applicants had to show that new impurities did not contain sequences with increased affinity for major histocompatibility complex, did not raise aggregation propensity, and did not produce distinct innate immune stimulation relative to the reference. Those are structural and biological questions that a percentage cannot answer.

What changed in July 2026

On 28 July 2026, FDA withdrew the 2021 guidance, stating that it no longer reflected the agency’s current scientific thinking, and indicating in the CDER guidance agenda an intention to revise it. The same day, the agency published seventeen revised draft product-specific guidances covering peptide products, with comments accepted through 28 September 2026.

The scope of the revised set is the most visible change. The 2021 document addressed five peptides. The July 2026 guidances cover calcitonin salmon, dasiglucagon, glucagon, liraglutide, pegcetacoplan, semaglutide, teriparatide, tirzepatide, and vosoritide across their respective listed products — a list that now includes several of the acylated and multi-receptor sequences that have dominated peptide research activity in recent years, and which the 2021 framework never contemplated.

FDA characterized the revised recommendations as spanning five areas: whether recombinantly, synthetically, or semi-synthetically produced peptides may be submitted through the abbreviated pathway; innate immune response testing; impurity thresholds; higher order structure assessment; and biological activity assessment. Three of those five are the same categories the 2021 guidance had folded into its impurity justification requirement. Their promotion into named assessment areas suggests a framework in which structure and biological behavior are evaluated directly rather than inferred from an impurity percentage.

The practical status is worth stating precisely: the 2021 guidance is withdrawn, the replacements are drafts open for comment, and the general revision is pending. Anyone citing the 0.5% and 0.10% figures should treat them as the prior framework’s numbers rather than current recommendations, and the product-specific documents as the operative reference until the general guidance is reissued.

What a threshold demands analytically

A threshold is only meaningful if the analytical method can see the level it names. Identifying an impurity at 0.10% of drug substance requires detecting it reliably well below that, resolving it from the main peak and from neighboring impurities, and generating enough structural information to determine a sequence. FDA’s 2021 recommendation of ultra-high-performance liquid chromatography with high-resolution mass spectrometry reflects those requirements: chromatographic resolution to separate, high-resolution mass measurement to assign composition, and fragmentation to locate modifications along the chain.

Reversed-phase separation alone is not sufficient for this, and the reason is a recurring theme in peptide analysis. The impurity classes that matter most for immunogenicity are frequently the ones reversed-phase chromatography handles worst. Aspartimide-derived isomers, D-epimers, and deamidation products are mass-identical or nearly so to the target and often co-elute with it. An impurity that co-elutes with the main peak is not merely unidentified — it is invisible, and it inflates the purity figure by being counted as product. This is why threshold frameworks pair with orthogonal method requirements, and why a single reversed-phase number, however high, does not establish that nothing sits above a threshold.

The same constraint governs the reporting threshold from the other direction. A method’s limit of quantitation sets a floor below which the “remainder” genuinely cannot be characterized. Whether that floor sits at 0.02% or 0.2% determines how much of a material’s composition the purity figure is summarizing rather than describing, and that parameter is a property of the method rather than of the peptide.

Why none of this appears on a research-grade certificate

The frameworks described here attach to marketing applications. Material supplied for research use is outside their scope entirely, and the practical consequence is that no threshold structure applies to a research-grade certificate of analysis at all.

This produces a specific interpretive problem. A research certificate reporting 99% purity generally does not state a reporting threshold, so there is no way to know what fraction of the remaining 1% was below the method’s quantitation limit and simply absent from the calculation. It rarely states an identification threshold, so individual impurities are typically unnamed regardless of level. It essentially never addresses qualification, because no toxicological or immunogenicity assessment has been performed on anything. The number is a chromatographic area ratio produced by one method under one set of conditions, and the regulatory apparatus that would otherwise convert such a number into a description of composition is not present.

That does not make the figure useless, but it does bound what can be concluded from it. Two certificates reporting the same purity from methods with different quantitation limits, different gradients, and different detection wavelengths are not reporting the same property. The threshold frameworks exist precisely because a percentage on its own underdetermines the composition of a material, and the absence of those frameworks does not make the underdetermination go away.

The July 2026 revision is a reminder that even where thresholds do apply, they are provisional. The 0.5% figure stood for five years, was grounded in reasoning about batch-to-batch variability and immunogenicity risk, and has now been withdrawn as no longer reflecting current scientific thinking — with its replacement still in draft. Thresholds encode a judgment about how much of a material needs to be known, and that judgment moves as analytical capability and understanding of immunogenicity move. The number under a purity figure is not a constant of nature. It is a record of what someone concluded was worth looking for, at a particular time, with particular instruments.

Further reading

Research use only. This post is for educational and reference purposes on peptide synthetic and analytical chemistry. It does not constitute medical, veterinary, or dosing guidance.