Analytical method validation (ICH Q2(R2))

Every result a QC lab reports is used to make a decision: release a batch, reject it, or set a shelf life. That decision is only as good as the method behind the number. Method validation is the documented evidence that a procedure is fit for its intended purpose — that it measures the right thing, accurately and reliably, across the range you need.

GMP rules require it. In the US, 21 CFR 211 expects laboratory methods to be accurate, sensitive, specific and reproducible. ICH Q2(R2), adopted in 2023, describes the characteristics to study and how to present the data. USP <1225> covers validation of compendial-style procedures and largely follows the ICH approach.

Validation proves the method is suitable. System suitability, run each day, proves the system is working on that day. You need both.

These three words are often mixed up. They answer different questions, and the amount of work differs a lot.

• Compendial methods are already validated by the pharmacopoeia — you do not repeat full validation, but you must show they work for your sample matrix.
• A change in the drug product composition (for example a new excipient) can invalidate specificity, so check it with a change control assessment.

ICH Q2(R2) lists the performance characteristics to consider. Which ones you study depends on the method's purpose: an identification test needs specificity only, while a quantitative impurity method needs almost everything.

• Specificity / selectivity — the method measures the analyte without interference from impurities, degradants, excipients or the blank. For stability-indicating methods this is shown with forced degradation samples and peak purity.
• Response (linearity) — the relationship between concentration and signal. Q2(R2) widens the old term 'linearity' to cover non-linear responses, which some techniques (for example some immunoassays) show.
• Range — the interval between the lowest and highest levels where accuracy, precision and response have been shown to be acceptable.
• Accuracy — how close the result is to the true value, usually shown by recovery of known amounts spiked into placebo or sample.
• Precision — how close repeated results are to each other. Repeatability is same analyst, same day, same equipment; intermediate precision adds different days, analysts or instruments within one lab. Reproducibility (between labs) matters mainly for transfer or standardisation.
• Detection limit (LOD) and quantitation limit (LOQ) — the lowest amount that can be detected, and the lowest that can be measured with acceptable accuracy and precision. Needed for impurity and limit tests.
• Robustness — how well the method copes with small deliberate changes, such as ±0.2 pH unit in the buffer or ±2 °C in column temperature. Under ICH Q14 this is studied during development rather than as a late validation tick-box.

Q2(R2) also allows accuracy and precision to be assessed together (a combined approach, for example using a total error or tolerance interval), if you justify it.

ICH does not set numerical limits — the criteria must be justified for your method and product. The values below are typical starting points seen in many pharma QC labs for HPLC methods. Your validation protocol and SOP define the actual limits.

Labelled typical for a reason: a biological assay, a dissolution method or a trace impurity at ppm levels will justify quite different numbers.

Validation is a planned, documented study — not a set of experiments written up afterwards. The approved protocol fixes the experiments and the acceptance criteria before any data are generated, so the results cannot be 'fitted' to the limits.

1. 1Confirm the method is developed and stable: system suitability criteria set, instrument qualified and calibrated, reference standards in date.
2. 2Write the validation protocol: purpose, method reference, characteristics to study, experimental design (levels, replicates), acceptance criteria, and responsibilities. QA approves it.
3. 3Execute the protocol. Record everything contemporaneously in the notebook or LIMS; keep chromatograms and audit trails. Follow ALCOA+ data integrity principles.
4. 4Handle deviations. Any failure or departure from the protocol is documented and investigated — never quietly repeated.
5. 5Calculate and review: recovery, RSD, regression statistics. Have a second person check calculations and raw data.
6. 6Write the validation report: results against each criterion, deviations and their impact, and a clear conclusion that the method is (or is not) suitable. QA approves it.
7. 7Release the method for routine use through the test procedure / SOP, with training records for the analysts.

Take an HPLC assay for paracetamol tablets. For specificity you inject the diluent blank, a placebo blend and stressed samples (acid, base, peroxide, heat, light) and show the paracetamol peak is pure and free from interference. For linearity you prepare five standards from 80% to 120% of the working concentration and plot area against concentration.

Accuracy uses placebo spiked with paracetamol at 80%, 100% and 120%, three preparations each, reporting percentage recovery. Repeatability is six sample preparations from one homogeneous composite; intermediate precision repeats this on another day with another analyst and column. A related substances method on the same product would add LOD/LOQ for the impurities (for example 4-aminophenol) and accuracy at the specification limit.

• Use the standard-curve calculator to check slope, intercept and correlation from your linearity data.
• Use the %RSD calculator for repeatability and system precision.

ICH Q14 (Analytical Procedure Development) was adopted alongside Q2(R2). It treats a method as having a lifecycle: development, validation, routine use and change. It starts with an Analytical Target Profile (ATP) — a statement of what the method must measure and how well — and encourages risk assessment and multivariate experiments to understand which parameters matter.

The 'minimal' approach (traditional development, then validation) is still acceptable. The 'enhanced' approach builds more knowledge up front, which can make later changes easier to justify. USP <1220> describes a similar lifecycle idea for compendial users.

• Ongoing monitoring — trending system suitability and control sample results — is part of the lifecycle, not an extra.
• Changes go through change control; the risk assessment decides whether partial revalidation is needed.

When you adopt a pharmacopoeial method — for example a USP or Ph. Eur. monograph assay — you do not revalidate it, but you must show it works in your hands, for your product. USP <1226> describes this verification. The extent depends on how complex the method is and how different your sample is from what the monograph assumed.

Simple general tests, such as loss on drying or pH, usually need no verification beyond normal qualification and training. A chromatographic assay or impurity method for a new formulation typically needs at least specificity (no excipient interference), repeatability and, for impurities, an LOQ check against the reporting threshold.

Pharmacopoeias are revised. When a monograph method changes, assess whether your verification still covers it — check the current edition rather than an old printout.