The difference between cleaning validation vs cleaning verification looks like a matter of terminology, but in an audit it decides which evidence the inspector expects to see: a formal validation proving that the cleaning process is reproducible, or an analytical check repeated after every batch. Confusing the two approaches — or using verification as a permanent shortcut where validation is required — is a classic finding generator. In this article we compare the two concepts under EU GMP Annex 15, look at when it is legitimate to operate in verification mode and how to set defensible acceptance limits.
Cleaning validation vs cleaning verification: the Annex 15 definitions
The glossary of EudraLex Volume 4 Annex 15 defines the two terms sharply. Cleaning validation is the documented evidence that an approved cleaning procedure reproducibly removes residues of the previous product and of cleaning agents below the scientifically set maximum allowable carryover level. Cleaning verification is instead the gathering of evidence, through chemical analysis after each batch or campaign, that residues have actually been reduced below that level.
The distinction is therefore between a prospective demonstration of process capability (validation, which once completed allows equipment release on the basis of the validated procedure) and a case-by-case check (verification, which proves nothing about reproducibility: it only says that this time the equipment is clean).
When verification is acceptable and when validation is required
Annex 15 (§10.3) recognises that a cleaning validation programme can take time to complete and that for some products it may be necessary to operate under validation with verification after each batch. The typical cases where cleaning verification is the correct or transitional approach are:
- Investigational medicinal products (IMPs): formulations and strengths change constantly and there is no repetitive process to validate; Annex 15 cites this case explicitly.
- Occasional or one-off manufacture: products made rarely, pilot batches, ongoing technology transfer.
- The transitional phase of a validation: until the number of runs defined in the protocol is completed, equipment release relies on the analytical results of each cleaning.
- After significant changes (cleaning agent, parameters, equipment), pending revalidation.
For recurring commercial manufacture, on the other hand, cleaning validation is the regulatory expectation: permanent verification in place of validation is sustainable only with a robust documented justification, and in any case requires sufficient data to conclude that the equipment is clean and available for further use. The same principle applies to APIs: ICH Q7 (section 12.7) asks that cleaning validation be directed at the situations with the greatest risk of contaminating the next product.
Topics like this one — definitions that look like subtleties and turn into audit findings — are the bread and butter of The Pragmatic GMP, GuideGxP's free weekly newsletter: every week one GxP topic broken down into operational, defensible decisions. Subscribe here.
Operational comparison table
| Aspect | Cleaning validation | Cleaning verification |
|---|---|---|
| What it demonstrates | The cleaning procedure removes residues reproducibly | Residues are below the limit in that specific cleaning |
| When it is performed | Once (number of runs defined by risk assessment), then maintained over the lifecycle | After every batch or campaign |
| Typical scope | Recurring commercial manufacture | IMPs, occasional manufacture, transitional pre-validation phase, post-change |
| Basis of the limits | HBEL/PDE from toxicological evaluation, converted into MACO | The same science-based limits: verification does not allow "more convenient" limits |
| Equipment release | Based on the validated procedure and routine controls | Only after a compliant analytical result for that cleaning |
| Audit evidence | Protocol, report, worst case, hold times, maintained validated state | Batch-by-batch analytical results and trending |
Acceptance limits: HBEL, MACO and visual inspection
Whatever the regime, the limits do not change in nature. Annex 15 (§10.6) requires carryover limits to be based on a toxicological evaluation: this is where HBELs come from (health based exposure limits, typically expressed as PDE) under the dedicated EMA guideline, to be converted into MACO and into per-surface or rinse limits. The EMA Q&A on HBELs and prevention of cross-contamination clarifies that HBELs must be determined for all products and that the traditional criteria — 1/1000 of the minimum therapeutic dose and 10 ppm — may continue to be used only where they are more stringent than the HBEL-derived limit, not as a convenient alternative.
Finally, beware of visual inspection: Annex 15 (§10.2) considers it an important part of the acceptance criteria, but it is not generally acceptable as the sole criterion. "Visually clean" on its own holds up only in scenarios well justified by risk, and it replaces the analytical rationale neither in validation nor in verification. Repeated cleaning and retesting until an acceptable result is obtained is not an acceptable approach either.
A solid cleaning validation protocol also covers the factors that verification, by its nature, does not explore: identification of the worst case across products and equipment, the hardest-to-clean sampling locations, the maximum times between end of production and cleaning (dirty hold time) and between cleaning and reuse (clean hold time), and the maximum campaign length on which the validation is based. These are exactly the questions an inspector asks when all they find is batch-by-batch verification data: without these answers, the equivalence of the two approaches cannot be demonstrated.
The most common audit findings
- Perpetual verification with no justification: sampling after every batch for years, with no document explaining why validation was never completed.
- Different limits between the two regimes: verification run against limits less strict than those calculated from HBELs, as if batch-by-batch testing "compensated" for it.
- No documented worst case: the validation does not identify the variable factors (worst-case product, hardest-to-clean locations, dirty/clean hold times, maximum campaign length).
- Visually clean as the sole criterion without supporting studies or a risk rationale.
- Validated state not maintained: after the initial report there is no periodic monitoring, trending or reassessment following changes.
GuideGxP recommendation
Treat cleaning verification for what it is: a control tool, not a substitute for validation. Define in an SOP or in the Validation Master Plan when you operate in verification mode (IMPs, occasional manufacture, post-change transitions), with which limits and for how long, and set an explicit exit criterion towards validation for recurring products. Calculate HBELs for every product on site, convert them into MACO with a traceable rationale and use the trending of verification data as capability evidence when the time comes to validate.
To set defensible limits, worst cases, recovery studies and complete protocols, the GuideGxP Operational Guide to Cleaning Validation in GMP (Second Edition + Toolkit) covers HBEL, MACO and audit-ready documentation step by step, with ready-to-use Excel and Word templates.
