Robustness and System Suitability: Why a Method That Worked on Tuesday Fails on Thursday

Every validation characteristic described so far is retrospective. Specificity, accuracy, precision, range — all of them were demonstrated at a particular time, by particular people, on particular equipment. None of them says anything about the run you are doing today. Two things do.

Quick answer

What is robustness? ICH Q14 defines it as a measure of an analytical procedure's capacity to meet the expected performance criteria during normal use. It is tested by deliberate variations of analytical procedure parameters, and — the clause that matters most in peptide work — it should consider the duration of the analysis, including stability of sample preparations and reagents.

What is a system suitability test? ICH Q14 defines SSTs as tests developed and used to verify that the measurement system and the analytical operations associated with the procedure are fit for the intended purpose, and to increase the detectability of unacceptable performance. Robustness is designed in; system suitability is checked every time.

The clause about duration

Most discussions of robustness concentrate on instrument parameters: flow rate, column temperature, mobile phase composition, pH, wavelength. Those matter. But the guideline explicitly extends robustness to cover how long the analysis takes and whether the prepared samples survive it.

For peptides that is not a footnote. A sequence of thirty injections can run overnight, which means the last vial in the tray was prepared many hours before it was measured. If the peptide is degrading or adsorbing in that interval, the result is a downward drift across the run that looks exactly like a real concentration difference between samples.

How quickly that happens depends on the solution: the stability question is covered in how long a reconstituted peptide is stable, and the surface losses in adsorption losses. An autosampler tray at ambient temperature is a worse place to keep a dilute peptide solution than almost anywhere else in the laboratory.

Where robustness sits in the lifecycle

ICH Q2(R2) states that robustness is typically evaluated as part of development, prior to the execution of the validation study. ICH Q14 adds that for most procedures robustness evaluation is conducted during development, and if it was already conducted there, it does not need to be repeated during validation.

This explains something that otherwise looks like an omission: robustness data frequently does not appear in a validation report at all, because it legitimately belongs to the development file. ICH Q2(R2) notes the evaluation can be submitted as part of development data on a case-by-case basis or should be made available upon request.

Q14 also notes that data from validation studies — intermediate precision in particular — can complement robustness evaluation, which makes sense: varying days, analysts and equipment is already a partial robustness experiment. That relationship is described in repeatability, intermediate precision and reproducibility.

Which parameters get varied

Q14 says prior knowledge and risk assessment inform the selection, and that those parameters likely to influence procedure performance over the intended period of use should be studied. It also notes that for procedures with inherent high parameter variability — the guideline names those requiring biological reagents — wider parameter ranges may need to be investigated.

A reversed-phase peptide separation has a familiar set of candidates:

Univariate examination of a single parameter can establish a proven acceptable range; Q14 encourages multivariate experiments where parameter interactions matter, since parameters rarely drift one at a time.

System suitability: the check that runs every time

An SST is part of the analytical procedure itself rather than part of its validation. Q14 places it inside the analytical procedure control strategy — the set of controls that ensures the procedure stays fit for purpose during routine use across its lifecycle — and says the control strategy should be defined before validation and confirmed after it.

The SST is typically conducted with one or more predefined materials, including the use of positive and/or negative controls, and is designed to verify selected procedure attributes, with acceptance criteria based on the procedure's performance criteria.

The phrase worth holding onto is from the definition: SSTs increase the detectability of unacceptable performance. The purpose is not to confirm that things are fine. It is to make a bad day visible before the results from it enter a record.

A worked example from the peptide literature

A 2022 validation of a method for the therapeutic peptide teduglutide evaluated robustness and system suitability alongside linearity, limits, accuracy and specificity — and did something instructive: it ran the validation on both the UV and the mass spectrometry signals from the same separation, then compared the validation performance parameters obtained from each to select the most suitable.

Two detectors, one chromatographic run, different validation performance. The method is not one thing; it is a chain, and the detector is a link with its own characteristics. The same study used forced degradation — thermal stress at 40 °C and 60 °C, mechanical shaking and accelerated light exposure — explicitly framed as the impact of medicine mishandling under hospital conditions, which is robustness thinking applied to the sample rather than the instrument.

What a research buyer can take from this

• Re-inject a standard at the end of every run. It is one injection and it tests the whole time axis.
• Prepare dilute solutions close to the time of measurement, not the night before.
• Define your own system suitability check — a known sample, a criterion, a decision rule — and apply it before accepting results, not after a result looks surprising.
• Record the conditions that varied, not just the result. A column change or a new mobile phase batch is the first thing to look at when numbers shift.
• Treat a drifting series as a method question first. Before concluding that lots differ, confirm the run itself was stable — the reasoning behind lot-to-lot variability and bridging.

Key takeaways

• ICH Q14 defines robustness as the capacity to meet expected performance criteria during normal use, tested by deliberate parameter variations.
• Robustness explicitly covers the duration of the analysis, including the stability of sample preparations and reagents.
• It is normally evaluated during development and need not be repeated during validation, which is why it is often absent from validation reports.
• Intermediate precision data can complement robustness evaluation.
• Procedures using biological reagents may need wider parameter ranges investigated.
• A system suitability test is part of the analytical procedure and exists to increase the detectability of unacceptable performance.
• In one published peptide method, UV and mass spectrometry signals from the same run gave different validation performance and had to be compared to select the better one.

Frequently asked questions

Why is robustness not in the validation report?

Because the guidelines place it in development. ICH Q2(R2) says it may be submitted as development data case by case or made available on request. Its absence is normal rather than a gap.

What is the difference between robustness and intermediate precision?

Intermediate precision measures how much results scatter when ordinary things vary by themselves. Robustness deliberately pushes parameters to see whether performance survives. One observes; the other provokes.

Can I do a robustness test without special equipment?

The sample-side half, yes. Measure a prepared solution immediately and again after the length of your typical run. That tests the duration clause directly and needs nothing extra.

What makes a good system suitability criterion?

One tied to the attribute that matters and that fails before results do. For a separation, resolution of the critical pair is the usual choice, for the reason set out in specificity.

Does any of this apply if I only read certificates?

Indirectly but usefully. It tells you that a certificate reflects a procedure that was fit for purpose on a particular day, verified by checks you cannot see. That is a reasonable thing to rely on, and a different thing from a guarantee about your own vial.

Sources

• International Council for Harmonisation. ICH Harmonised Guideline Q14: Analytical Procedure Development. Final version adopted 1 November 2023. ICH guideline
• International Council for Harmonisation. ICH Harmonised Guideline Q2(R2): Validation of Analytical Procedures. Final version adopted 1 November 2023. ICH guideline
• Pérez-Robles R, Salmerón-García A, Clemente-Bautista S, Jiménez-Lozano I, Cabañas-Poy MJ, Cabeza J, Navas N. Method for identification and quantification of intact teduglutide peptide using (RP)UHPLC-UV-(HESI/ORBITRAP)MS. Anal Methods. 2022;14(43):4359–4369. DOI

Journal citation retrieved via PubMed. Guideline text is quoted from the adopted ICH Q14 and Q2(R2) documents.

This article describes analytical methodology only. It contains no usage instructions, no dosing information and no protocols of any kind.

For laboratory and in-vitro research use only. Not for human or veterinary use.