1What a smoke study must prove.
A smoke study, or airflow visualisation study, releases a visible tracer into the air of a cleanroom, RABS or isolator and films where it goes. EU GMP Annex 1 asks for airflow patterns to be visualised “to demonstrate that there is no ingress from lower grade to higher grade areas and that air does not travel from less clean areas (such as the floor) or over operators or equipment that may transfer contamination to the higher grade areas”.¹ FDA’s aseptic guidance says: “In situ air pattern analysis should be conducted at the critical area to demonstrate unidirectional airflow and sweeping action over and away from the product under dynamic conditions.”²
| Point | EU GMP Annex 1 | FDA aseptic guidance |
|---|---|---|
| Status | EU GMP guideline, in operation since 25 August 2023 | Nonbinding guidance of September 2004; warning letters cite 21 CFR 211.113(b) and 211.42(c)(10) |
| States | At rest and in operation, simulating operator interventions (4.15) | Dynamic conditions, including interventions and equipment design (IV.A) |
| Where | Cleanrooms and zones, small egress points, open isolator interfaces, RABS during interventions and door openings (4.15, 4.20); blow-fill-seal, by visualisation or equivalent studies (8.109) | The critical area (IV.A) and blow-fill-seal (Appendix 2); open isolator ports need overpressure “supported by qualification studies” (Appendix 1) |
| Record | Video retained, outcome documented (4.15) | “well documented with written conclusions”; video useful (IV.A) |
| Air speed | 0.36 to 0.54 m/s at the working position (guidance value), correlated with the video (4.30) | 0.45 m/s ±20% (0.36 to 0.54 m/s) “has generally been established”; later checks hold the study’s range (IV.A note 5, IV.D) |
| Feeds into | Monitoring locations (4.15, 4.31, 9.4) and training (7.18) | Evaluating later equipment configuration changes (IV.A) |
The PIC/S GMP Guide carries the same Annex 1, in force since 25 August 2023 with minor differences, and MHRA points UK manufacturers to EudraLex Volume 4.⁵⁶ PIC/S PI 058-1, in force since 28 October 2026, is the reading of Annex 1 that PIC/S, the EU and WHO agreed. Where a barrier system’s air speed sits outside 0.36 to 0.54 m/s, it says the visualisation should “demonstrate protection of the critical process operations (first air protection)”.⁷ First air is filtered air “that has not been interrupted prior to contacting exposed product and product contact surfaces”.¹ For test methods, Annex 1 points to the ISO 14644 series.¹ For the wider Annex, see what EU GMP Annex 1 requires.
US 503B outsourcing facilities have FDA’s draft CGMP guidance of January 2020. It asks for airflow studies, such as an in-situ smoke study, under simulated operational conditions at the ISO 5 critical area, repeated “when any changes are made to the critical area that might affect airflow”.⁸
2Acceptance criteria, location by location.
Write each criterion before the study, as something a reviewer can see on the video. Annex 15 asks protocols to fix acceptance criteria in advance, and a result that misses one is recorded as a deviation and investigated.⁹
| Location | Passes when the video shows | Source |
|---|---|---|
| Grade A, at rest | Smoke leaves the filter, falls and sweeps over and away from exposed product and product-contact parts, with no stagnant or reversing flow | 4.15, 4.19;¹ IV.A² |
| Grade A, interventions | First air reaches open product before touching a glove, sleeve, tool or machine part, through each listed intervention | 4.15, 8.16, glossary¹ |
| RABS doors | No grade B air drawn in during interventions or door openings | 4.20(ii)¹ |
| Open isolator openings | Air leaves at every interface and exit port; none is drawn in | 4.20(i)(c);¹ Appendix 1² |
| Egress points | No air from the lower grade room enters grade B where filled, closed product leaves | 4.15¹ |
| Room doors | Air moves from the cleaner room to the less clean one, with the door open as well as closed | 4.14, 4.15;¹ IV.C² |
| Closed isolators | Any turbulence adds no risk to exposed product; a processing line inside still needs first air | 4.19(i)(b);¹ Appendix 1² |
| Lyophiliser loading | Transport devices and venting do not adversely affect airflow patterns | 8.126(iii)¹ |
Turbulence away from exposed product is not a failure in itself: Annex 1 expects turbulent air to “flush the cleanroom via mixed flow dilution”. Where the video shows air movement is a contamination risk, it asks for corrective action, such as design improvement.¹
3Smoke study protocol template.
Copy these sections into your protocol. Each is tied to the clause that asks for it.
| Section | What it holds | Requirement |
|---|---|---|
| Scope | Room, line, barrier and every configuration covered: formats, fill volumes, line speeds | 4.15;¹ Baxter letter⁴ |
| Prerequisites | All equipment installed as it will run, since it can change airflow; HVAC at qualified settings, air speed measured at the working position | 4.30;¹ PI 058-1 4.29i, 4.30⁷ |
| Locations | Point of fill, stopper bowl, open components, RABS doors, isolator interfaces, transfer ports, egress points, room doors | 4.15, 4.20, 4.28¹ |
| States | At rest: HVAC on, equipment installed but idle, nobody present. In operation: equipment running, the maximum number of people | 4.15, 4.29¹ |
| Interventions | Every inherent and corrective intervention on the authorised list, done as the procedure describes. For each, in each line configuration: the clip that shows it and the media fill that challenged it | 8.16, 9.34;¹ Fareva letter¹⁰ |
| Criteria | A pass statement for each location and state | Annex 15 2.4⁹ |
| Filming | Smoke release point and camera position for each sequence; each clip named by location, intervention and date | 4.15;¹ K.C. letter¹¹ |
| Failures | Deviation, impact on batches already made, corrective action, repeat study | Annex 15 2.8;⁹ 4.15¹ |
| Report | Result against each criterion; proposed monitoring locations; use in training | Annex 15 2.9;⁹ 4.15, 4.31, 7.18¹ |
| Contractor work | Site staff confirm a vendor’s protocol fits site procedures before approving it | Annex 15 2.6⁹ |
The footage is a GMP record. EU GMP Chapter 4 counts electronic and photographic media as documentation, and says “Secure controls must be in place to ensure the integrity of the record throughout the retention period”.¹² Keep the original, unedited files with the protocol and report, retrievable by clip name; a training cut is a copy.
4Running the study on the day.
The guidance says what the video must show, not how to shoot it. FDA’s warning letters name what inspectors found wrong in the footage.
- Choose the tracer. ASTM E3379, a consensus standard guide to these studies, asks for a generator that produces “constant, neutrally buoyant, and non-toxic tracer particles/smoke”.¹³ If you are choosing between water or WFI fog and a glycol-type smoke, weigh what each leaves on product-contact surfaces, and plan the clean-down before you start.
- Rehearse each sequence. Dry-run every intervention with the operator, the person holding the smoke source and the camera operator before the filmed take.
- Put the smoke where the question is. At K.C. Pharmaceuticals, “the smoke source was not always positioned at the HEPA filter face to confirm unidirectional airflow and uniform velocity.” At Jubilant HollisterStier, one video showed “inadequate smoke at the site of the critical activity”.¹¹¹⁴ Release smoke at the filter face for the at-rest pass, then at each critical site during each intervention, following the hand or tool as it moves.
- Film the air, not the operator’s back. FDA faulted K.C. for interventions in which “the camera is placed behind the operator and the impact of the intervention on the airflow cannot be visualized.”¹¹ Fix each camera position in the protocol, and name each sequence on film by location and intervention.
- Use the people who do the job. In operation means “the maximum number of personnel present performing or simulating routine operational work”.¹ Reach decides whether a body crosses first air: at Jubilant, an operator “extended their upper body over the conveyor”, and FDA faulted the response for not addressing hazards that included “intervention ergonomics”.¹⁴ Where reach changes the path, film the tallest and the shortest operators who perform that intervention.
- Record changes on the day. A significant change to the approved protocol during execution is documented as a deviation and justified.⁹
- Judge the film before the line relies on it. Jubilant rated airflow “Excellent” despite the inadequate visualisation.¹⁴ At K.C. in 2023, the quality unit had not finished reviewing a repeat study, and it was not made available to the investigators.¹⁵ ASTM E3379 asks for a subject matter expert to have “direct oversight of airflow studies”.¹³ Have someone who did not film the study review every clip, and record pass or fail with the reason, clip by clip.
5Example: a smoke study on a RABS filling line.
This example is invented. Study AV-26-07 covers Line 2, a vial filling line with grade A inside an active RABS in a grade B room. Before filming, engineering measured air speed at nine working-position points: 0.42, 0.45, 0.47, 0.39, 0.44, 0.46, 0.41, 0.50 and 0.42 m/s, a mean of 0.44 m/s, all inside Annex 1’s 0.36 to 0.54 m/s guidance range.¹ Operators performed each intervention as the procedure describes, and engineering and microbiology filmed on 13 October 2026.
| Location or intervention | State | What the video showed | Result |
|---|---|---|---|
| Point of fill | At rest | Smoke falls over needles and open vials, then out at the RABS base | Pass |
| Stopper bowl | At rest | Smoke sweeps down over the bowl and track | Pass |
| Egress to capping | At rest | At the opening to the grade C capping room, smoke flows out of grade B | Pass |
| Egress to capping | In operation | Stoppered vials pass through the opening; smoke still flows out to grade C | Pass |
| Needle change | In operation | Gloved hands stay beside and below the needles; first air reaches open vials | Pass |
| Fallen vial removal | In operation | Smoke over the tongs is swept down and away from open vials | Pass |
| RABS door, jam clearing | In operation | Air flows out through the open door; no grade B air enters | Pass |
| Stopper bowl top-up | In operation | Smoke passing over the glove sleeve falls into the open bowl | Fail |
| Room door to airlock | In operation | Air leaves the filling room as the door opens | Pass |
What followed:
- QA raised deviation DEV-26-0412 the same day, because the top-up missed a pre-set criterion.⁹
- QA assessed the batches filled since the line’s last study, using monitoring results at the bowl and the last two media fills, which had included the top-up.
- Change CC-26-0198 added a covered stopper chute, so the sleeve stays below the bowl rim. Until the new top-up had passed a repeat study and a media fill, top-ups were made only with filling stopped and open vials cleared.
- A repeat study on 3 November 2026 passed. Microbiology added an active air sample at the bowl rim, training added the clip, and the next media fill includes the new top-up, as Annex 1 asks for interventions to be evaluated through risk management and process simulation.¹
6When to run a smoke study, and when to repeat it.
Annex 1 lists “airflow direction test and visualisation” among the tests of cleanroom qualification, but not among the minimum tests for requalification, which is due at least every 6 months for grades A and B.¹ Neither Annex 1 nor FDA’s aseptic guidance sets a fixed interval for repeating a study.¹² The trigger is change, and the contamination control strategy should say which changes count.
| Situation | Smoke study? | Basis |
|---|---|---|
| New room, line or barrier | Yes, at rest and in operation | 4.15, 4.25¹ |
| Routine requalification | Not in the minimum list; repeat if your risk assessment says so | 4.32¹ |
| HVAC, filter or layout | Assess in change control; repeat where airflow could change | 4.32;¹ IV.A² |
| New intervention | Assess its airflow impact, usually on film, and evaluate it by risk management and APS before it joins the list | 8.16¹ |
| New configuration | Film it, or show the studied one is the worst case | Baxter letter⁴ |
| Air speed outside range | Show first air protection at the working position | 4.30;¹ PI 058-1 4.30⁷ |
| Contamination signal | Review the videos against current practice and equipment | 9.4;¹ IV.A² |
FDA’s reason for the last row: “even successfully qualified systems can be compromised by poor operational, maintenance, or personnel practices.”²
7What inspectors cite.
FDA’s inspection data for fiscal year 2025 list “Inadequate smoke studies” six times on drug Form 483s. The cite is section 501(a)(2)(A) of the FD&C Act, the insanitary-
“Furthermore, your firm has not conducted dynamic airflow studies during filling and stoppering operations.” Fareva Amboise, FDA warning letter, 10 April 2026¹⁰
FDA asked Fareva for new dynamic smoke studies “with thorough and complete evaluations of aseptic interventions and operator positioning within the critical filling areas”, and a video of them.¹⁰
“Your dynamic smoke studies did not demonstrate unidirectional airflow protection of the production line while barrier (b)(4) were (b)(4). The study showed air moving at a sharp angle toward (b)(4) vents, leaving critical areas of your aseptic processing line potentially unprotected.” Par Health USA and Endo USA, FDA warning letter 320-26-68, 15 April 2026³
“Turbulent airflow may be appropriate for a (b)(4) that does not house aseptic processing equipment. In contrast, your firm uses (b)(4) that include processing line equipment.” Baxter International, FDA warning letter 320-26-132, 29 September 2026⁴
8Where teams go wrong.
- No study in either state. FDA told one firm it “failed to conduct smoke studies under static and dynamic conditions”, and that “dynamic smoke studies are necessary to show that (b)(4) manipulations do not interfere with laminar airflow”.¹⁸
- Mitigating a design flaw instead of fixing it. At Par, equipment design flaws compromised unidirectional airflow: “Personnel blocked first pass air to (b)(4) with their forearms” during an installation step. FDA said the firm was “attempting to partially mitigate significant aseptic processing line hazards” rather than make more extensive changes.³ Annex 1 asks for engineering solutions whenever possible to minimise operator incursion.¹
- Leaving the video in the validation file. Annex 1 wants the outcome used to set monitoring locations and considered in operator training.¹
9How it runs in Seal.
In Seal, a smoke study runs as a qualification test. Commissioning and qualification instantiates an approved test template against the actual asset and configuration, and a failed acceptance becomes an exception tied to the requirement, step, observation and evidence. A retest adds new evidence without overwriting the original failure. In the contamination control strategy, each control links to the qualification and monitoring evidence behind it, and a change cannot close until the affected claims are reviewed.
Environmental monitoring keeps each excursion’s product-impact assessment on timing, airflow, organism, adjacent locations, interventions, cleaning and batch evidence, and aseptic process simulation records which interventions each media fill actually challenged.
References
- 1EudraLex Volume 4, Annex 1: Manufacture of Sterile Medicinal Products, C(2022) 5938 final, paragraphs 4.14 to 4.32, 7.18, 8.16, 8.109, 8.126, 9.4, 9.34 and glossary, European Commission, 22 August 2022, in operation since 25 August 2023. https:/
/ health. ec. europa. eu/ system/ files/ 2022- 08/ 20220825_ gmp- an1_ en_ 0. pdf (read 8 November 2026) - 2Guidance for Industry: Sterile Drug Products Produced by Aseptic Processing, Current Good Manufacturing Practice, sections IV.A (with footnote 5), IV.C and IV.D and Appendices 1 and 2, FDA, September 2004. https:/
/ www. fda. gov/ media/ 71026/ download (read 8 November 2026) - 3Par Health USA, LLC and Endo USA, Inc., warning letter 320-26-68, MARCS-CMS 722121, FDA, 15 April 2026. https:/
/ www. fda. gov/ inspections- compliance- enforcement- and- criminal- investigations/ warning- letters/ par- health- usa- llc- endo- usa- inc- 722121- 04152026 (read 8 November 2026) - 4Baxter International Inc., warning letter 320-26-132, MARCS-CMS 733231, FDA, 29 September 2026. https:/
/ www. fda. gov/ inspections- compliance- enforcement- and- criminal- investigations/ warning- letters/ baxter- international- inc- 733231- 09292026 (read 8 November 2026) - 5Entry into force of revised Annex 1, PIC/S GMP Guide (PE 009-17), PIC/S news, August 2023. https:/
/ picscheme. org/ en/ news/ entry- into- force- of- revised- annex- i- pics- gmp- guide- pe- 009- 17 (read 8 November 2026) - 6Good manufacturing practice and good distribution practice, MHRA, GOV.UK. https:/
/ www. gov. uk/ guidance/ good- manufacturing- practice- and- good- distribution- practice (read 8 November 2026) - 7PI 058-1, Recommendation: Interpretation of the Most Important Changes in the Annex 1 on the Manufacture of Sterile Products (2022), questions 4.29i and 4.30, PIC/S, in force 28 October 2026. https:/
/ picscheme. org/ docview/ 11495 (read 8 November 2026) - 8Current Good Manufacturing Practice: Guidance for Human Drug Compounding Outsourcing Facilities Under Section 503B of the FD&C Act, revised draft guidance, revision 2, section III.B, FDA, January 2020. https:/
/ www. fda. gov/ media/ 88905/ download (read 8 November 2026) - 9EudraLex Volume 4, Annex 15: Qualification and Validation, paragraphs 2.4 to 2.9, European Commission, in operation since 1 October 2015. https:/
/ health. ec. europa. eu/ system/ files/ 2016- 11/ 2015- 10_ annex15_ 0. pdf (read 8 November 2026) - 10Fareva Amboise, warning letter, MARCS-CMS 723502, FDA Center for Veterinary Medicine, 10 April 2026. https:/
/ www. fda. gov/ inspections- compliance- enforcement- and- criminal- investigations/ warning- letters/ fareva- amboise- 723502- 04102026 (read 8 November 2026) - 11K.C. Pharmaceuticals, Inc., warning letter 320-26-112, MARCS-CMS 729870, FDA, 12 August 2026. https:/
/ www. fda. gov/ inspections- compliance- enforcement- and- criminal- investigations/ warning- letters/ kc- pharmaceuticals- inc- 729870- 08122026 (read 8 November 2026) - 12EudraLex Volume 4, Part I, Chapter 4: Documentation, principle and paragraph 4.10, European Commission, in operation since 30 June 2011. https:/
/ health. ec. europa. eu/ system/ files/ 2016- 11/ chapter4_ 01- 2011_ en_ 0. pdf (read 8 November 2026) - 13ASTM E3379-25a, Standard Guide for Critical Airflow Visualization, scope 1.4.1 and section 4.1, ASTM International, 2025 (public summary page). https:/
/ store. astm. org/ e3379- 25a. html (read 8 November 2026) - 14Jubilant HollisterStier General Partnership, warning letter 320-26-81, MARCS-CMS 723537, FDA, 28 May 2026. https:/
/ www. fda. gov/ inspections- compliance- enforcement- and- criminal- investigations/ warning- letters/ jubilant- hollisterstier- general- partnership- 723537- 05282026 (read 8 November 2026) - 15K.C. Pharmaceuticals Inc., warning letter, MARCS-CMS 654986, FDA, 3 August 2023. https:/
/ www. fda. gov/ inspections- compliance- enforcement- and- criminal- investigations/ warning- letters/ kc- pharmaceuticals- inc- 654986- 08032023 (read 8 November 2026) - 16Inspection Observations, fiscal year 2025 data set, Drugs sheet, cite 22766, FDA. https:/
/ www. fda. gov/ inspections- compliance- enforcement- and- criminal- investigations/ inspection- references/ inspection- observations (read 8 November 2026) - 1721 U.S.C. 351(a)(2)(A), Adulterated drugs and devices (FD&C Act section 501), U.S. Government Publishing Office. https:/
/ www. govinfo. gov/ content/ pkg/ USCODE- 2024- title21/ html/ USCODE- 2024- title21- chap9- subchapV- partA- sec351. htm (read 8 November 2026) - 18Sciarra Aeromed, Inc., warning letter, MARCS-CMS 728248, FDA Center for Veterinary Medicine, 23 September 2026. https:/
/ www. fda. gov/ inspections- compliance- enforcement- and- criminal- investigations/ warning- letters/ sciarra- aeromed- inc- 728248- 09232026 (read 8 November 2026)
AQuestions and answers
Can a smoke study be done at rest only?
No. Annex 1 asks for airflow patterns to be studied both at rest and in operation, simulating operator interventions, and FDA expects the analysis under dynamic conditions. An at-rest film shows the design works in an empty room; only the in-operation film shows whether people, tools and moving equipment break first air.
How long should smoke study videos be kept?
Annex 1 says the videos should be retained but sets no period. EU GMP Chapter 4 ties retention to the activity a record supports: validation raw data that supports the marketing authorisation is kept while the authorisation is in force, and process validation raw data at least as long as the records of every batch released on the strength of it. A safe rule is to keep each video as long as the batch records of every batch filled in the configuration it qualified.
Does a smoke study replace air velocity measurement?
No. Annex 1 lists airflow velocity and airflow visualisation as separate qualification tests and asks for the two to correlate. Velocity shows the air is fast enough at the working position; the video shows where it goes. A study filmed at a different air speed from the one running today proves little.
Do closed isolators need a smoke study?
Yes. Annex 1 asks for airflow patterns within cleanrooms and zones to be visualised, and in a closed isolator the study shows that any turbulence does not increase the risk to exposed product. Where a closed isolator houses a processing line, Annex 1 asks for first air and unidirectional flow over exposed product; FDA’s guidance says turbulent flow can be acceptable in closed isolators that are compact and house no processing line. Open isolators also need studies at their interfaces to show no air comes in.
Can a contractor run a smoke study?
Yes, but the site stays responsible for it. EU GMP Annex 15 asks site staff to confirm that a third party’s protocol is suitable and complies with internal procedures before approving it. The interventions on film must be done the way your operators do them in production, and QA judges the videos against criteria written before filming.