Analytical Method Validation: Why a Purity Number Depends on Its Method

Why a purity number needs analytical method validation

An HPLC purity is calculated, not observed. The harmonized pharmacopoeial chromatography chapter describes that calculation, normalization: each peak’s area as a percentage of the total area of all peaks, excluding solvent and mobile-phase peaks “and those at or below the disregard limit or reporting threshold”. Every input to that sum depends on the method: whether impurities separate from the main peak, whether small peaks are detected, whether the detector responds in proportion. Change the column, gradient or wavelength and the number can change with it.

Validation is the evidence that those inputs behave, which is how ICH Q2(R2) frames its objective: showing that a procedure is fit for its intended purpose. Eurachem’s guide for testing laboratories makes the same point from the other side: a result fit for its intended use “must be sufficiently reliable so that any decision based on it can be taken with confidence”. ICH adopted Q2(R2) on 1 November 2023, and it took effect at the European Medicines Agency on 14 June 2024 alongside its companion, ICH Q14 on analytical procedure development. Both are guidelines for medicines; we quote them because they set out publicly, and in detail, what a validated method should demonstrate.

The validation characteristics in plain English

ICH Q2(R2) groups the method validation parameters into specificity or selectivity, range, accuracy, precision and robustness. The table adds the two checks that sit around them and says what each protects a purity figure against.

Specificity: can the method see the impurities that matter?

For a purity method, specificity comes first, because an impurity that co-elutes with the main peak is simply counted as product, and no amount of precision or linearity data will reveal it. ICH Q2(R2) separates the two terms. Specificity describes “the ultimate state, measuring unequivocally a desired analyte”; selectivity is a relative term for how far analytes in a mixture “can be measured without interferences from other components with similar behaviour”.

For purity and impurity tests, the guideline says discrimination can be established “by stressing or spiking product” to achieve appropriate impurity levels and, where impurities cannot be prepared or isolated, by comparison with an orthogonal procedure. Peptides make this demanding. The US FDA describes peptide-related impurities as sequences “related to, but different from” the active ingredient through insertion, deletion or other modification, and for certain generic synthetic peptide drugs it recommends “sensitive and high resolution analytical procedures (e.g., UHPLC-HRMS)” to detect them. Close relatives of a sequence can elute near the main peak, which is why reading the chromatogram and LC-MS identity confirmation are complementary checks rather than alternatives.

Accuracy, precision and range

The definitions keep two often-blurred ideas apart. In ICH Q2(R2), accuracy expresses the closeness of agreement between a measured value and one “accepted either as a conventional true value or as an accepted reference value”, while precision expresses the “closeness of agreement (degree of scatter)” among repeated measurements of the same homogeneous sample. A method can be precise and wrong, repeating the same biased number.

Q2(R2) considers precision at three levels. Repeatability covers the same conditions over a short interval, and the guideline suggests at least nine determinations across the reportable range, or six at 100% of the test concentration. Intermediate precision adds “different days, different environmental conditions, different analysts and different equipment”; reproducibility compares laboratories. Accuracy can be shown with, for example, three concentrations in triplicate.

Range sets the boundaries. The guideline’s example reportable range for an assay runs from 80% to 120% of declared content, and for an impurity from the reporting threshold to 120% of its acceptance criterion. A purity figure leans on both: the main peak’s range and the impurities’. Where peptide content rather than purity is the question, an independent technique such as amino acid analysis gives a cross-check from a different principle.

Detection and quantitation limits

The detection limit is “the lowest amount of an analyte in a sample which can be detected but not necessarily quantitated as an exact value”; the quantitation limit is the lowest amount that “can be quantitatively determined with suitable precision and accuracy”. Laboratories often abbreviate them LOD and LOQ. Q2(R2) offers several ways to establish them, including signal-to-noise ratios of 3:1 for detection and at least 10:1 for quantitation, or calculation from the standard deviation of the response and the calibration slope (3.3σ/S and 10σ/S).

For a purity figure, the quantitation limit matters more than it looks. Q2(R2) says that for impurity tests it “should be equal to or below” the reporting threshold, and the harmonized chromatography chapter writes the same requirement into system suitability. Because normalization leaves out peaks at or below the reporting threshold, a method that cannot see small impurities does not report them as impurities: it reports a higher purity. For related-substances tests in pharmacopoeial monographs, the chapter notes, the reporting threshold “is generally 0.05%”.

System suitability: checking the system on the day

Validation happens before a method enters routine use; system suitability is checked every time it runs. Q2(R2) calls the system suitability test “an integral part of analytical procedures”, and its glossary, in wording taken from ICH Q14, says such tests verify that the measurement system and associated operations “are fit for the intended purpose and increase the detectability of unacceptable performance”.

For chromatography, the chapter harmonized across the European, Japanese and United States pharmacopoeias (Ph. Eur. 2.2.46, JP 2.00, USP <621>) puts numbers on this. System sensitivity is defined by signal-to-noise; the quantitation peak’s symmetry factor must be 0.8 to 1.8 unless a monograph says otherwise; resolution or peak-to-valley ratio may be specified; and for assays, unless a monograph sets its own requirement, the relative standard deviation of replicate sample injections of a reference solution has a ceiling tied to the upper content limit. For an upper limit of 102.0% and six replicates, that ceiling is 0.85%. The revised chapter became official in the USP on 1 December 2022 and took effect in the Ph. Eur. on 1 January 2023, and the EDQM notes that compliance with the system suitability test is always required when chromatographic conditions are adjusted.

Measurement uncertainty and pass/fail decisions

The international metrology vocabulary defines measurement uncertainty as a “non-negative parameter characterizing the dispersion of the quantity values being attributed to a measurand”. It is usually reported as an expanded uncertainty, the standard uncertainty multiplied by a coverage factor k; NIST notes that when the normal distribution applies, k = 2 “defines an interval having a level of confidence of approximately 95 %”. A purity of 98.6% ± 0.5% (k = 2), to take a hypothetical figure, says far more than a bare 98.6%.

Standards Council of Canada guidance says a robust estimation of measurement uncertainty “must be performed as part of the validation”, and Eurachem observes that most of the information needed can be obtained during validation. Reporting it is conditional: ISO/IEC 17025:2017 calls for uncertainty in a report when it is relevant to the validity or application of the result, when a customer requires it, or when it “affects conformity to a specification limit”. Under simple acceptance, a value exactly at the limit carries, in UKAS’s words, “an associated risk of false decision of 50 %”; a guard band, derived from the uncertainty, exists “with the purpose of limiting the risk of false acceptance”.

Questions to ask about a certificate’s method

None of this asks a buyer to repeat the validation; it lists questions a competent laboratory, or a supplier relaying its answers, should be able to handle.

1. Which procedure produced the number? Technique, column, gradient and detection wavelength, or a method reference.
2. Was it validated, or verified, for this peptide? SCC guidance applies validation to non-standard, laboratory-developed and modified methods, and verification to standard methods adopted without significant change.
3. How was specificity shown? Which impurities were spiked or generated by stress, and was an orthogonal technique such as LC-MS used?
4. What were the quantitation limit and reporting threshold? Peaks below them are not in the purity figure.
5. Did the run meet system suitability, and against which criteria?
6. What is the uncertainty, and which decision rule stands behind any “pass”?
7. Does the report tie the tested sample to your batch? Our guide to what a certificate should show covers that check.

A certificate need not answer all seven on its face. What matters is whether the answers exist when asked.

What’s established and what isn’t

Applying this when comparing suppliers in Canada

The same vocabulary now sits in Canadian drug guidance: Health Canada’s table of implemented ICH guidelines lists Q2(R2) from 17 October 2025 and Q14 from 12 January 2026. That is pharmaceutical guidance, and we make no claim that it applies to research reagents; it is useful here as a current, shared definition of what “validated” should mean.

For testing laboratories, the Standards Council of Canada publishes requirements and guidance on method verification and validation, describing validation as ensuring a method has “the capabilities and performance characteristics consistent with specified requirements” and is fit for its intended use. If a report names an accreditation, it is reasonable to ask which methods that accreditation covers.

In practice, compare suppliers on whether these questions get answered, not on the size of the headline number. We are a supplier too, so put the same questions to us: the certificates we publish are on our Lab Testing page, and method questions can go through the contact button below.

Read the certificate behind the number

Janoshik test reports for the research peptides below are listed on our Lab Testing page, so you can read each purity result in context rather than as a bare number on a label.

BPC-15715-amino-acid research peptideView product GHK-CuCopper tripeptideView product TB-500Thymosin beta-4 research peptideView product KPVLys-Pro-Val tripeptideView product

Frequently asked questions