HPLC method validation under ICH Q2(R2) is the documented demonstration that a chromatographic procedure is fit for its intended purpose: that it measures the right analyte (specificity), gives a proportional response over a defined reportable range, and returns results that are accurate and precise enough to support a release, stability or impurity decision. ICH Q2(R2) was adopted at Step 4 on 1 November 2023, alongside the new ICH Q14 guideline on analytical procedure development, and replaces Q2(R1). This guide sets out what changed, which performance characteristics apply to which type of HPLC method, the acceptance criteria laboratories commonly use, and how to structure a validation protocol that holds up to regulatory review.
What changed from Q2(R1) to Q2(R2)
Q2(R1) served the industry for almost three decades, and the core science has not been thrown away. Specificity, linearity, accuracy, precision and detection/quantitation limits all still matter. What Q2(R2) does is reorganise and modernise the framework:
- Terminology. “Validation characteristics” are now called performance characteristics. “Range” is described as the reportable range, supported by the response (linear or non-linear) and a lower range limit (the QL for quantitative impurity tests, the DL for limit tests).
- Robustness moves upstream. The guideline states that robustness is typically evaluated during development, before the validation study, and points to ICH Q14 for details. In practice, your robustness data now belongs in the development report rather than being an afterthought in the validation protocol.
- Non-linear and multivariate procedures are explicitly covered, which matters for techniques such as NIR or some detector responses, although most HPLC-UV methods remain linear.
- Combined accuracy and precision. Laboratories may assess the total impact of bias and variability against a combined performance criterion (for example, using a prediction or tolerance interval) instead of judging each separately.
- Platform procedures. Where an established platform procedure is applied to a new purpose, validation testing can be abbreviated if scientifically justified.
Regionally, the European Medicines Agency brought Q2(R2) into effect on 14 June 2024, and the US FDA issued both Q2(R2) and Q14 as final guidance in March 2024.
ICH Q2(R2) tells you which performance characteristics to demonstrate and gives minimum designs (for example, at least five concentrations for linearity and at least six or nine determinations for repeatability), but it does not set numerical acceptance criteria. Those must be pre-defined in your protocol and justified by the method’s intended purpose and the specification it supports.
Performance characteristics by HPLC method type
The first decision in any validation is what the method is for. An HPLC assay, a quantitative related-substances method and a limit test each need a different set of experiments. The table below summarises the expectations commonly applied under Q2(R2) for typical HPLC procedures.
| Performance characteristic | Identification | Impurities (quantitative) | Impurities (limit test) | Assay / content |
|---|---|---|---|---|
| Specificity / selectivity | Yes | Yes | Yes | Yes |
| Response (linearity) and reportable range | No | Yes | No | Yes |
| Lower range limit | No | QL | DL | No |
| Accuracy | No | Yes | No | Yes |
| Repeatability | No | Yes | No | Yes |
| Intermediate precision | No | Yes | No | Yes |
| Robustness | Evaluated during development (ICH Q14), summarised in the validation package | |||
Retention time alone is rarely sufficient for identification in a regulated context; HPLC identity tests are usually paired with a diode-array spectrum match or an orthogonal technique such as LC-MS.
Parameters, experimental design and typical acceptance criteria
Specificity
Specificity shows that the analyte response is free from interference by the matrix, known impurities, degradation products and the blank. For an HPLC related-substances method, inject the diluent blank, placebo, the API, each available impurity individually and a spiked mixture. Where impurity standards are not available, stressed samples (acid, base, oxidative, thermal, photolytic) demonstrate that degradants are resolved from the main peak. This is why validation of a stability-indicating method is closely tied to forced degradation studies. Peak purity by diode-array detection, or mass confirmation, strengthens the case.
Response and reportable range
Q2(R2) recommends a minimum of five concentrations appropriately distributed across the range. Report the regression equation, slope, intercept, correlation coefficient and a plot of residuals. A visually random residual pattern is more informative than a high r value on its own. The guideline’s recommended reportable ranges include:
- Assay of a drug substance or product: 80% to 120% of the declared content or of the specification acceptance criterion.
- Content uniformity: 70% to 130% of the declared content.
- Impurities: from the reporting threshold to 120% of the specification acceptance criterion.
- Purity by area %: from 80% of the lower specification acceptance criterion to the upper criterion or 100%.
Accuracy
Accuracy is normally shown by recovery of a known amount of analyte spiked into placebo (drug product) or by comparison with a reference standard of established purity (drug substance). Q2(R2) gives the example of three concentrations with three replicates each across the reportable range. For impurity methods, spike the impurities into the sample matrix, including a level at or near the QL. The quality of your reference material directly limits the accuracy you can demonstrate, so characterised standards with a complete certificate of analysis are essential; see our guide to impurity reference standards.
Precision
Repeatability requires either a minimum of nine determinations covering the reportable range (for example three concentrations, three replicates each) or a minimum of six determinations at 100% of the test concentration. Intermediate precision examines within-laboratory variation: different days, analysts and equipment, ideally with a different column lot. Reproducibility (between laboratories) is relevant when a method is transferred or standardised across sites.
Detection and quantitation limits
Q2(R2) accepts several approaches. With the signal-to-noise approach, a ratio of 3:1 is generally acceptable for estimating the DL and at least 10:1 for the QL. With the standard deviation and slope approach, DL = 3.3σ/S and QL = 10σ/S, where σ is the standard deviation of the response and S the slope of the calibration curve. Whichever approach you use, confirm the QL experimentally by showing acceptable accuracy and precision at that level.
Typical acceptance criteria
Because ICH does not fix numbers, most laboratories work from internally justified criteria. The values below are common starting points for small-molecule HPLC-UV methods, not regulatory limits; tighten or relax them according to the specification and the method’s purpose.
| Parameter | Assay method (typical) | Related substances method (typical) |
|---|---|---|
| Specificity | No interference at analyte retention time; peak purity passes | All specified impurities resolved (resolution commonly ≥ 1.5); no blank or placebo interference |
| Linearity | r ≥ 0.999; random residuals | r ≥ 0.99 from QL to 120% of limit |
| Accuracy (mean recovery) | 98.0–102.0% | Often 80–120% near QL, 90–110% at specification level |
| Repeatability (%RSD) | ≤ 2.0% | Often ≤ 10% at QL, ≤ 5% at specification level |
| Intermediate precision | Overall %RSD ≤ 2.0%; no significant day/analyst bias | Comparable to repeatability criteria |
| Solution stability | Change within a pre-defined limit over the hold time | No new peaks; impurity levels stable within limit |
Worked example: setting the range for an impurity method
Suppose an API has a specified impurity with an acceptance criterion of NMT 0.15%, a reporting threshold of 0.05%, and the test solution is prepared at 1.0 mg/mL.
- Reportable range. From the reporting threshold (0.05%) to 120% of the specification (0.18%). At 1.0 mg/mL this corresponds to 0.5 µg/mL to 1.8 µg/mL of the impurity.
- Lower range limit. The QL must be at or below the reporting threshold, so target a QL of about 0.03% (0.3 µg/mL) to leave margin, and confirm S/N ≥ 10 with acceptable precision at that level.
- Linearity levels. Five or more levels, for example QL, 50%, 80%, 100% and 120% of the specification limit.
- Accuracy. Spike at three levels (for example QL, 100% and 120% of the limit), three preparations each, into the API or placebo matrix.
- Relative response factor. If the impurity will be quantified against the API peak, determine the RRF from the slopes of the impurity and API calibration lines and fix it in the method.
The same logic for an assay at 0.5 mg/mL gives a reportable range of 0.4 to 0.6 mg/mL (80–120%), with linearity typically run at 80, 90, 100, 110 and 120%.
Writing the validation protocol and report
A validation protocol should be approved before any experiment is run. A practical structure is:
- Objective, scope and intended purpose (assay, related substances, dissolution, cleaning, etc.), with reference to the analytical target profile or specification.
- Method summary: column, mobile phase, gradient, flow rate, temperature, detector wavelength, injection volume, system suitability requirements.
- Materials: reference standards and working standards with lot numbers, potency and expiry; samples and placebo.
- Instruments and software, with qualification status and electronic records controls compliant with 21 CFR Part 11 where applicable.
- Each performance characteristic with its experimental design, calculations and pre-defined acceptance criteria.
- Handling of deviations, out-of-specification results and failed acceptance criteria.
The report should present raw data references, chromatograms, statistical outputs and a clear conclusion against every criterion. Any deviation must be explained with its impact assessed, not quietly dropped.
Common pitfalls in HPLC validation
- Validating an under-developed method. If robustness was never studied, small shifts in pH, column temperature or organic proportion will surface as failures during validation or transfer. Q2(R2) and Q14 now make this expectation explicit.
- Specificity without stressed samples. A method that resolves known impurities but has never seen degraded samples cannot be called stability-indicating.
- QL by calculation only. A 10σ/S estimate that is never verified with real injections is a frequent review comment.
- Range that does not match the specification. If the impurity limit is later widened, the validated range may no longer cover 120% of it.
- Inadequate solution stability. Standards and samples held in the autosampler for long sequences need demonstrated stability for that duration.
- Weak reference material. Uncharacterised or poorly assigned standards undermine accuracy and RRF determination.
- System suitability set too loosely. Criteria should be drawn from validation and robustness data so that they catch genuine performance loss; see the applicable pharmacopoeial chapter (USP <621> or Ph. Eur. 2.2.46) for definitions.
How SDS Labs can help
SDS Labs provides analytical method development, validation and routine testing for raw materials, intermediates, APIs and finished formulations, using HPLC with UV, fluorescence and refractive index detection. Work is carried out under an ISO/IEC 17025:2017 quality system, with data generated in line with 21 CFR Part 11 and cGMP requirements, and we hold impurity and working standards that can support specificity and accuracy studies.
If you are planning a new validation or bridging an existing Q2(R1) method to Q2(R2) expectations, share your method and specification through our contact page, call +91 90295 20144 or email info@sdslabs.in.
Frequently asked questions
What is the difference between ICH Q2(R1) and Q2(R2)?
Q2(R2) keeps the core concepts of specificity, linearity, accuracy, precision and detection or quantitation limits, but renames them performance characteristics, introduces the reportable range and lower range limit, covers non-linear and multivariate procedures, allows combined accuracy and precision criteria, and moves robustness into development under the companion guideline ICH Q14. It was adopted at Step 4 on 1 November 2023.
What are the validation parameters for an HPLC method?
For a quantitative HPLC assay or impurity method, the parameters are specificity, response (linearity) over the reportable range, accuracy, repeatability and intermediate precision, plus the quantitation limit for impurity methods. Limit tests need specificity and the detection limit. Robustness is studied during development and summarised in the validation package, and solution stability is normally included as well.
How many concentrations are needed for linearity under ICH Q2(R2)?
ICH Q2(R2) recommends a minimum of five concentrations appropriately distributed across the reportable range. For an assay this typically means 80, 90, 100, 110 and 120% of the test concentration. For an impurity method, levels usually run from the quantitation limit up to 120% of the specification acceptance criterion, with residuals examined alongside the correlation coefficient.
How are LOD and LOQ determined in HPLC?
ICH Q2(R2) accepts a signal-to-noise approach, where about 3:1 is generally acceptable for the detection limit and at least 10:1 for the quantitation limit, or a calculation from the standard deviation of the response and the calibration slope (3.3σ/S and 10σ/S). Whichever approach is used, the quantitation limit should be confirmed experimentally with acceptable accuracy and precision.
Does ICH Q2(R2) specify acceptance criteria such as 98–102% recovery?
No. ICH Q2(R2) describes which characteristics to evaluate and minimum experimental designs, but does not fix numerical acceptance criteria. Values such as 98.0–102.0% mean recovery and not more than 2.0% RSD for an assay are common industry starting points. Criteria must be pre-defined in the protocol and justified against the specification and intended use of the method.