Is 98% Purity Good Enough? What the Number Does and Doesn’t Tell You
Short answer. A purity percentage on its own cannot answer that. The questions are which impurities make up the other 2%, whether each one at 0.10% or more has been identified, and whether any single unknown exceeds 0.5%, the thresholds published for synthetic peptide drug substances. A 98% lot with characterized impurities is better documented than a 99.5% lot with unknowns.
What this means for your clinic. Stop asking suppliers for a number and start asking for the impurity table underneath it: every peak at 0.10% or more, identified, with the largest unknown stated. Then write your own acceptance criterion into the supplier file, so that “good enough” is a decision your clinic made in advance rather than one made lot by lot.
Why does the headline number mislead?
Purity by HPLC is the main peak’s share of everything the detector saw. Two lots can carry the same figure and be different products underneath it, because the figure says nothing about what the remainder is.
A 98% lot might carry 2% of impurities that the laboratory has separated, identified by mass and named, none of them above 0.5%. A 99.5% lot might carry a single 0.5% peak that nobody has identified. The first file tells you what the lot contains. The second tells you only that something is there. That is the argument of this post, and it is why the clinic’s guide to verifying peptides treats the impurity table, not the headline, as the thing to read on a certificate of analysis.
There is a second reason the number misleads. Purity is a proportion of detector response, and the method decides it: column, gradient and wavelength can move the same lot by a fraction of a percent in either direction. So 98% from one laboratory set against 99.5% from another is not always a comparison of two lots. It may be a comparison of two methods.
Which thresholds have actually been published, and for what?
The only published thresholds specific to synthetic peptides come from FDA’s May 2021 guidance on abbreviated new drug applications for certain highly purified synthetic peptide drug products. It asks an applicant to identify each peptide-related impurity present at 0.10% or more, and it says a proposed product should not contain any new peptide-related impurity above 0.5%. A new impurity between those lines needs an assessment of its potential to provoke an immune response.
The scope matters as much as the numbers. That guidance covers five named peptide products in the generic pathway and nothing else. This series cites it because it is the only published benchmark for synthetic peptide impurities, not because it is a rule for any other product, supplier or clinic.
The general impurity guideline for small molecules, ICH Q3A(R2), sets reporting, identification and qualification thresholds for new drug substances, and it expressly excludes peptides from its scope. A supplier who quotes ICH Q3A limits for a peptide is quoting a document that says it does not apply.
The reason the peptide benchmark cares so much about identification is immunogenicity. The 2021 guidance ties its impurity limits to the risk that a new impurity provokes an immune response. Wu and colleagues (2017), writing from FDA’s Office of Pharmaceutical Quality, review how impurities and aggregates arising in synthesis and storage bear on the safety of peptide products, immunogenicity included. That is the guidance’s rationale: an impurity that sits beside the parent on a chromatogram may still be a different molecule to the immune system, which is why each one at 0.10% or more should be known by name.
What does an impurity profile show that a purity number can’t?
An impurity profile is the table under the chromatogram: each peak beyond the main one, its retention time, its area percent and, where the laboratory has done the work, its identity by mass. D’Hondt and colleagues (2014) catalog what those peaks usually are in a synthetic peptide: deletion and insertion sequences from incomplete coupling or deprotection, diastereomers from racemization, oxidized residues, adducts left by protecting groups, dimers, and degradation products such as pyroglutamate and succinimide forms. A profile shows which of those a lot carries. A number does not.
Two illustrative lots make the point.
| Lot A (illustrative) | Lot B (illustrative) | |
|---|---|---|
| Purity by HPLC | 98.0% | 99.5% |
| Peaks at 0.10% or more | Four, each identified by mass and named | One |
| Largest single unknown | None | 0.5%, unidentified |
| Impurities identified | 2.0% of 2.0% | 0% of 0.5% |
| What the lot file can say | What the lot contains | Only that something is there |
On EV-01 the main peak is 99.80%, which leaves 0.20% spread across the small peaks in the integration table. The question to put to that table is not whether 99.80% is high. It is whether any single remaining peak reaches 0.10%, and if one does, whether the report identifies it. On a certificate that stops at the purity line, that question cannot be asked.
How does a clinic write its own acceptance criterion?
An acceptance criterion is a written statement of what your clinic will accept before a lot leaves quarantine. It belongs in the supplier file, dated and signed by whoever approved it, so nobody has to decide under pressure when a shipment arrives.
Name the benchmark you are drawing on and write beside it that it is a benchmark for synthetic peptide drug substances in one regulatory pathway, not a rule that binds the clinic.
Set three lines: a minimum purity by a stated method, identification of every impurity at 0.10% or more, and a cap on the largest single unknown.
Require the impurity table and the chromatogram with every certificate, so the three lines can be checked rather than taken on trust.
Write down what happens when a lot misses a line: the lot stays in quarantine, the supplier is asked a fixed set of questions, and the outcome is recorded as a deviation.
When a lot misses, the questions to the supplier are specific. Which impurity, at what level, identified as what. Whether earlier lots carried it. Whether the manufacturer has stability data showing it grows in storage. The post on What Are Peptide-Related Impurities, and Which Ones Should Concern a Clinic? covers what the answers mean, and the post on What Should a Clinic Do If a Supplier Won’t Provide a Batch-Specific COA? covers a supplier who will not answer.
Whether a lot that misses the criterion is used is a decision for the licensed provider, with the manufacturer’s data in front of them. The clinic’s part is the record showing the question was asked and answered. The figures your clinic adopts are yours, chosen for your products and written into your file.
How Helios does it
Under the Helios verification standard, step 2 requires a lot-specific certificate before shipment, the manufacturer’s certificate plus an independent laboratory report for the exact lot, and the record it produces is the certificate set. Helios Consultants LLC applies the standard lot by lot: the set must carry the panel the standard specifies, which includes the purity chromatogram and its integration table rather than a purity figure alone. Under step 6, every lot is held in quarantine until steps 2 to 5 pass, and any failure produces a deviation record and a written disposition. The sequence is set out in how Helios verifies every lot. If you want help setting this up in your clinic, book a consultation.
Related questions
What Does HPLC Purity Actually Measure on a Peptide COA?
What Are Peptide-Related Impurities, and Which Ones Should Concern a Clinic?
What Can a Peptide COA Not Tell You?
Which Tests Should a Complete Peptide COA Include?
Sources
FDA, ANDAs for Certain Highly Purified Synthetic Peptide Drug Products That Refer to Listed Drugs of rDNA Origin, Guidance for Industry (May 2021). https://www.fda.gov/regulatory-information/search-fda-guidance-documents/andas-certain-highly-purified-synthetic-peptide-drug-products-refer-listed-drugs-rdna-origin
ICH Q3A(R2), Impurities in New Drug Substances. https://database.ich.org/sites/default/files/Q3A%28R2%29%20Guideline.pdf
Wu LC, et al. Int J Pharm. 2017. PMID 28027918. https://doi.org/10.1016/j.ijpharm.2016.12.051
D’Hondt M, et al. J Pharm Biomed Anal. 2014. PMID 25044089. https://doi.org/10.1016/j.jpba.2014.06.012
Educational content for licensed professionals. Not legal or medical advice. No product is offered for sale on this page.