Cleanroom Qualification: ISO 14644 Tests, HEPA Integrity, Recovery and Room Performance

The qualification protocol for a new suite comes up for review with a test table headed "testing to ISO 14644-3". Every row states a method, but none states the occupancy state, the source of the acceptance criterion or the URS requirement it verifies. Recovery appears among the mandatory items of requalification; filter integrity is declared covered "by the manufacturer's certificate"; classification is planned at rest only, because "in operation is covered by environmental monitoring".

This is not an execution problem but an architecture problem: what is demonstrated, under which condition, against which criterion. A protocol that does not answer those three questions produces valid data and indefensible conclusions. The wider picture is in the Cleanrooms & HVAC Systems hub.

Why the qualification file matters beyond start-up

At tender stage, a specification that does not say who performs which tests and who issues the report makes bids incomparable. At start-up, every badly written criterion becomes a deviation with the suite idle. In operation, qualification data are the baseline for set points and for alert and action levels, and requalification and change control are measured against them.

Regulatory framework: what actually constrains

Classification, qualification, routine monitoring

Annex 1 §4.24 requires qualification, classification included, to be clearly distinguished from operational environmental monitoring [REQUIREMENT]. Classification measures total particles only, by the ISO 14644-1 method: it is a subset of qualification. Qualification is the documented GMP process of Annex 15 applied to the whole system, with the §4.25 list and criteria from the URS and Annex 1. Routine monitoring is a continuous, risk-based programme with alert and action levels from site data; for GMP, section 9.

Two consequences follow. ISO 14644-2 reclassification does not replace requalification under §4.32, and the reverse is equally true; Annex 1 in fact carries no explicit reference to ISO 14644-2. And a rich monitoring programme does not compensate for a classification missing in one state: it is required at rest and in operation for all four grades [REQUIREMENT]. Zoning defines the perimeter.

The CQV sequence and the commissioning boundary

Qualification does not start in the field but with a verifiable URS: that is where the criteria no standard fixes come from, namely temperature, humidity, airflow tolerances and alarm delays. DQ verifies that the design answers the URS; IQ that the installation matches the approved design; OQ that the system runs within the stated ranges, alarms and interlocks included; PQ that performance holds.

Commissioning remains an engineering process. Within EudraLex, reuse of verification performed before formal qualification is governed as FAT/SAT by Annex 15 §3.4–3.7, provided documentation is adequate and transport and installation are shown not to have altered performance [REQUIREMENT]. The phrase commissioning leverage belongs to ISPE and ASTM E2500-25 [STANDARD], not to Annex 15; the package enabling that reuse is covered in construction and turnover.

The minimum test list, test by test

Installed filter system integrity

This is the in situ test of the mounted system, not the factory classification of the filter. In Europe E/H/U classification remains EN 1822-1:2019, while the test methods are now EN ISO 29463-4, which describes the scan leak test of elements, and EN ISO 29463-5; EN 1822-2/-3/-4/-5 have been withdrawn [STANDARD]. The certificate covers the element before installation; the in situ test challenges medium, frame, gasket, housing and clamping, with aerosol upstream and scanning downstream. The ≥99.97 % retention of particles >0.3 µm quoted by FDA [GUIDANCE] is not the same quantity as the European classes, based on MPPS; and a leak test is never a replacement criterion.

Airflow volume and velocity

Airflow volume shows that supply air matches the design and supports classification and clean up; the tolerance is not regulatory and belongs in the URS. Velocity applies to unidirectional systems: §4.30 requires it to be justified in the protocol, measurement position included, with a guidance range of 0.36–0.54 m/s at the working position unless otherwise justified in the CCS [GUIDANCE]. The FDA figure of 0.45 m/s ± 20 % is a typical value in a footnote [GUIDANCE].

Pressure difference, airflow direction, visualisation

Pressure difference shows that the cascade holds under all intended operating conditions: §4.14 gives a minimum of 10 Pa between adjacent rooms of different grade, explicitly qualified as a guidance value [GUIDANCE], and §4.16 requires continuous recording of the differentials identified as critical, with alarms [REQUIREMENT]. §4.15 then requires airflow patterns to be visualised, to show that there is no ingress from lower to higher grade and that air does not travel from less clean areas or over operators towards higher grades [REQUIREMENT]: a qualitative test with reading criteria agreed beforehand, covered in airflow visualisation studies, which CFD does not replace.

Particle classification

Particles ≥0.5 µm and ≥5 µm are measured against the Table 1 limits for each grade and state [REQUIREMENT]. One figure explains why classification is not enough: the ≥0.5 µm at rest limit is 3,520 particles/m³ for grade A and grade B alike [REQUIREMENT], while the two environments differ in airflow, gowning and microbial limits, so particle counts alone do not qualify a grade. Sampling points follow ISO 14644-1 (§4.28). At rest means HVAC running, equipment installed but not operating, no personnel; in operation means equipment running in the defined mode with the maximum number of people performing or simulating routine work.

Microbial, temperature and humidity

Airborne and surface microbial contamination is determined during qualification; the number of locations comes from a documented risk assessment and from classification and visualisation results (§4.31), against the Table 2 limits [REQUIREMENT]. The European reference for biocontamination control is EN 17141:2020; ISO 14698-1 and -2 were withdrawn on 16 March 2026 [STANDARD]. Neither Annex 1 nor 21 CFR states temperature or humidity values: ranges come from the URS, the process and gowning.

Recovery: a qualification parameter, not a requalification item

Recovery appears neither in the §4.25 list nor in the §4.32 list. It lives in §4.29(iii): at rest limits are to be reached after a clean up period, guidance value less than 20 minutes, determined in qualification and applied in procedure to restore the qualified state after a disruption [GUIDANCE]. The site value is measured and becomes procedural data; ISO 14644-3 supplies the method, not the criterion [STANDARD]. Listing it among the mandatory items of §4.32 misreads the text, though repeating it after work affecting airflow volume or distribution remains defensible on a risk basis [QRM].

Qualification test matrix

Calibration, alarms and interlocks

A test is worth no more than the measurement chain behind it: instruments identified by serial number, certificates valid on the day, calibration verified afterwards as well. Alarms are tested as functions: threshold, delay, recording, notification. §4.13 prohibits airlock and pass-box doors from opening together, with interlocking for grades A and B [REQUIREMENT], to be tested in failure modes too. Where data originate in a BMS or EMS, Annex 11 applies [REQUIREMENT].

Deviations, acceptance and release

Release is a Quality act based on the qualification report, not on the sum of the test sheets: the report states what was tested, under which conditions, with which departures and with which conclusion on fitness for intended use. Non-blocking residual items are released with a condition of use and an owner; an internal criterion stricter than the regulatory requirement is changed through QRM, while Annex 1 requirements allow no waiver. Periodic review and requalification start from there.

What QRM contributes

ICH Q9(R1), Step 4 on 18 January 2023, contains no HVAC parameters, but Annex II.4 applies QRM to facilities, equipment and utilities, including the scope of qualification. Formality as a continuum (§5.1) scales test depth to the impact of the area; the treatment of subjectivity and bias (§5.3) is why criteria are approved before execution, by people who do not execute them. What remains to be justified is the number of microbial locations, sample extent and which tests are repeated after a correction [QRM].

Worked example: Site Vega

Site Vega is a realistic but fictional example: a suite of grade B and C rooms with interlocked airlocks. Three issues emerge during qualification. In operation classification of a grade C room exceeds the limit at one repeatable location; the in situ test finds a leak at a filter housing although factory certificates were compliant; the measured clean up period exceeds the design assumption.

The answers were different. The out-of-limit location was traced to the position of a return grille relative to the simulated activity: correction and repeat of the coupled tests. The housing was resealed and rescanned, with a note in the report on the distinction between certificate and in situ test. The measured clean up became site data written into procedure, with the Annex 1 guidance value as a benchmark, not a contractual criterion.

Levels of prescriptiveness

Operational checklist

1. Build the URS requirement – test – criterion matrix before writing protocols.
2. State the occupancy state and configuration for every test.
3. Name the source of every criterion: URS, Annex 1 or site data.
4. Keep the filter factory certificate separate from the in situ test.
5. Check that the specification cites EN 1822-1:2019 and EN ISO 29463.
6. Justify the velocity measurement position in the protocol.
7. Plan classification at rest and in operation, with microbial locations from a written risk assessment.
8. Determine the clean up period in qualification and write it into procedure.
9. Test alarms and interlocks in failure modes as well.
10. Keep traceable raw data available for change control.

Recurring mistakes and red flags

• Presenting classification as completed qualification, or treating ISO 14644-2 reclassification as Annex 1 requalification.
• Accepting the filter factory certificate instead of the in situ test.
• Stating that the recovery test is mandatory in requalification under §4.32.
• Attributing acceptance criteria, or the 0.36–0.54 m/s range, to ISO 14644-3.
• Repeating a failed test without change control on the protocol.

To turn these principles into protocols that survive inspection, The Pragmatic GMP is the GuideGxP newsletter: short analyses on qualification, HVAC and contamination control, from its dedicated page.

Key takeaways

• Classification, qualification and routine monitoring are distinct: §4.24 requires it.
• The §4.25 list defines the perimeter; criteria come from the URS, Annex 1 or site data.
• Installed filter system integrity is an in situ test, different from the factory certificate.
• ISO 14644-3 supplies test methods and no GMP acceptance criteria.
• The clean up period is determined in qualification and is not a periodic requalification item.

References

• EudraLex Volume 4
• Annex 1 (PDF)
• Annex 15 (PDF)
• ISO 14644-1
• ISO 14644-3
• ISO 29463-1
• FDA Aseptic Processing Guidance
• PIC/S Publications