Analytical Method Validation: What ICH Q2(R2) Asks of a Purity Number
The TB-500 10mg batch certificate on this site reports 98.62% purity by HPLC-UV at 214 nm, with a content mass of 9.84 mg against an expected 10.00 mg. Every part of that is a real measurement made by a third-party laboratory on the batch being shipped, and publishing it openly is more than most of this market does.
Here is what it does not say. It does not say the analytical procedure that produced 98.62% was validated. It does not say over what concentration range the response was demonstrated, what the procedure’s specificity was tested against, how repeatable the number is, or what its uncertainty is. Those are separate claims from the number itself, and the difference between a measured value and a validated one is the subject of a specific regulatory document that most people buying research peptides in the United States have never had reason to open.
Validation, stated plainly
Validating an analytical procedure means demonstrating, with data, that the procedure is fit for the purpose it is being used for. Not that the instrument works. Not that the analyst is competent. That this procedure, applied to this kind of sample, over this range, produces results that are specific to the analyte, accurate, and reproducible enough for the decision being made with them.
The governing text is ICH Q2(R2), Validation of Analytical Procedures, whose final version was adopted on 1 November 2023 with error corrections on 30 November. In the United States it is not a foreign document: FDA issued Q2(R2) as a guidance for industry in March 2024. The guidance is explicitly nonbinding — it “does not establish any rights for any person and is not binding on FDA or the public,” and applicants may propose alternative approaches — but it is the reference frame US regulators and US contract laboratories actually work in.
Its stated scope is analytical procedures for release and stability testing of commercial drug substances and drug products, with the principles extending to other control-strategy procedures on a risk-based basis and to clinical development in a phase-appropriate manner. Research-grade material sold as a laboratory reagent sits outside that scope. That is precisely why the framework is useful as a reading tool rather than as a compliance stick: it tells you what a fully characterized purity claim would look like, so you can see how far any given certificate is from one.
What the guideline actually asks for
Q2(R2) does not ask for the same evidence for every kind of test. It sorts performance characteristics by what the procedure is for:
- Identity tests need specificity and nothing else.
- Limit tests for impurities need specificity and a validated lower range limit.
- Quantitative impurity tests need specificity, range and response, a lower range limit, accuracy, and precision.
- Assay and potency tests need specificity, range and response, accuracy, and precision.
A peptide purity figure computed as area percent is doing the work of a quantitative impurity determination and an assay at the same time — it is asserting both how much main peak there is and how little of everything else. Under this framework that is the most demanding column in the table, not the least.
The definitions are worth having in the reader’s hands. Specificity is the extent to which other substances interfere with determination of the analyte; the guideline distinguishes it from selectivity, which is relative rather than absolute. Accuracy is the closeness of agreement between a measured value and an accepted reference value. Precision is the degree of scatter across multiple samplings of the same homogeneous sample, and it comes in three tiers: repeatability under the same conditions over a short interval, intermediate precision across days, analysts, environments and equipment within one laboratory, and reproducibility between laboratories. Robustness is the capacity to meet performance criteria during normal use, established by deliberately varying procedure parameters.
The 2023 change that matters most here
The headline structural revision in Q2(R2) is that linearity was replaced by range and response. The older framing invited a laboratory to report a correlation coefficient and move on. The revised framing asks whether the calibration model — linear, non-linear, or multivariate — appropriately relates signal to the quality attribute across the reportable range, and asks separately for verification of the lower range limit. Multivariate procedures get explicit provisions, with distinct calibration, internal testing and validation phases.
For anyone reading certificates, that change has a practical consequence: an r² near 1.000 is not, on its own, evidence of anything much. It is a statement about the tightness of points around a fitted line, not about whether the fit is the right model, and not about whether the procedure behaves at the low end where impurity peaks actually live.
The numbers behind the numbers
Q2(R2) is more concrete than its reputation suggests. Where a linear response is being demonstrated, a minimum of five concentrations distributed across the range is recommended. Accuracy is assessed over an appropriate number of determinations and concentrations covering the reportable range, with three concentrations at three replicates each given as the example. Repeatability is either a minimum of nine determinations covering the reportable range, or a minimum of six determinations at 100% of the test concentration. Detection limit is generally accepted at a signal-to-noise ratio of 3:1 and quantitation limit at 10:1.
Those are small numbers. The work behind a validated purity method is not exotic; it is a couple of weeks of careful laboratory time. Which makes its absence from a certificate a choice rather than an impossibility.
Where the target comes from: ICH Q14
Q2(R2) was adopted alongside ICH Q14, Analytical Procedure Development, on the same date. Q14 covers how a procedure is designed before anyone validates it, and it introduces the analytical target profile — a prospective summary of the performance characteristics describing the intended purpose and the anticipated performance criteria. Q14 states directly that the analytical target profile serves as the foundation from which validation criteria under Q2 are derived.
The order matters and is routinely inverted in commercial practice. Validation does not decide whether a number is good enough. It demonstrates that the procedure meets criteria someone set in advance, on purpose, for a stated use. A validation report with no target profile behind it is a set of experiments in search of an acceptance criterion.
Q14 also distinguishes a minimal approach — identify the attribute, select the technology, evaluate performance, document the control strategy — from an enhanced approach using risk assessment, design of experiments, and defined proven acceptable ranges. It notes that robustness is normally established during development and need not be repeated at validation if it has already been done.
A worked example, and what it cost to find out
Bavand Savadkouhi and colleagues published a reversed-phase HPLC method for eptifibatide acetate — a cyclic heptapeptide — in the Iranian Journal of Pharmaceutical Research in 2017 (16(2):490–497, PMID 28979304). It is an ordinary, well-executed validation, and two of its findings are exactly the kind of thing a certificate never surfaces.
First, the sensitivity. Using the conventional signal-to-noise approach, the authors calculated a detection limit of 15 µg/mL and a quantitation limit of 45 µg/mL. Then they tested recovery at that calculated quantitation limit and got 70%, which failed their acceptance criterion. They set the working quantitation limit at 0.15 mg/mL instead — more than three times higher than the calculated value. A signal-to-noise calculation told them the method could quantify at 45 µg/mL. An accuracy experiment told them it could not.
Second, the wavelength. Eptifibatide has a UV maximum near 219 nm, and the authors did not use it. They ran detection at 275 nm because the trifluoroacetic acid in the mobile phase interfered at the lower wavelength, and 275 nm gave a flat baseline and a symmetrical peak. That is a specificity decision, documented, with a reason.
Hold that against a research peptide certificate. Peptides from solid-phase synthesis are typically isolated as trifluoroacetate salts, and the near-universal detection wavelength on these certificates — ours included — is 214 nm, chosen because that is where the peptide bond absorbs. Trifluoroacetate also absorbs in that region. The point is not that 214 nm is wrong; it is standard practice and it works. The point is that whether it was the right choice for a given peptide and a given mobile phase is a specificity question with a documented answer in a validation report, and a certificate that reports the wavelength without the validation is reporting the setting without the justification.
The stability-indicating counterpart is Bisht, Rupenthal, Sreebhavan and Jaiswal, who published a method for a connexin43 mimetic peptide in the Journal of Pharmaceutical Analysis in 2017 (7(6):365–373). Its title carries the load: a stability-indicating method is one qualified by forced degradation, meaning the authors deliberately stressed the peptide and showed the method still separated intact material from what the stress produced. A method that has never been shown degradation cannot be assumed to see it.
What to ask a US domestic supplier
The US research-peptide market has consolidated around a small set of trust signals: a batch certificate, a named third-party laboratory, a purity figure, and increasingly a domestic ship-from address. Ours ship same-day out of Santa Barbara under a Wyoming entity, and the certificate is downloadable without an account. Those are real differentiators and they are not the same thing as method validation.
Reasonable questions, in ascending order of how few suppliers can answer them:
- Which laboratory ran it, and is the report batch-specific?
- What method, what detection wavelength, and what column and gradient?
- Was the procedure validated, and against what performance criteria?
- What is the reportable range, and where does the lower range limit sit relative to the impurity peaks being integrated?
- What is the uncertainty on the reported figure?
A supplier who answers the first two honestly and says “no” to the third is being straight with you. A two-decimal purity figure quoted with no uncertainty and no validation statement behind it should be read as what it is: a laboratory’s result from a procedure whose fitness has been asserted rather than demonstrated.
Our certificates carry a sample-identification block and an HPLC-UV purity-and-content section. They do not carry a validation statement, a reportable range, or an uncertainty. We would rather write that sentence than let a reader infer otherwise from a number with two decimal places.
All compounds referenced are supplied for research use only. They are laboratory reagents for in vitro and preclinical investigation by qualified researchers, are not for human use, are not foods or drugs, and nothing here describes or endorses administration to a person.
References
- ICH Q2(R2) — validation of analytical procedures. Final version adopted 1 November 2023, with error corrections 30 November 2023. Performance characteristics by test type, definitions of specificity, accuracy, precision, detection and quantitation limits and robustness, and the replacement of linearity by range and response.
- US Food and Drug Administration — Q2(R2) validation of analytical procedures, guidance for industry, March 2024. Nonbinding recommendations adopting the ICH text.
- ICH Q14 — analytical procedure development. Adopted 1 November 2023. Minimal and enhanced development approaches, the analytical target profile, and its role as the foundation for validation criteria under Q2.
- Bavand Savadkouhi M, Vahidi H, Ayatollahi AM, Hooshfar S, Kobarfard F — reversed-phase HPLC method development and validation for eptifibatide acetate in bulk drug substance and dosage forms. Iranian Journal of Pharmaceutical Research, 2017; 16(2):490–497. PMID 28979304; PMCID PMC5603858.
- Bisht R, Rupenthal ID, Sreebhavan S, Jaiswal JK — a stability-indicating reversed-phase HPLC method for quantification of a connexin43 mimetic peptide and its degradation kinetics in biological fluids. Journal of Pharmaceutical Analysis, 2017; 7(6):365–373.