“We acknowledge your retraining” — the sentence FDA writes before rejecting your CAPA
483 response filed 9 June 2025 · Citations 21 CFR 211.22(d), 211.28(a), 211.42(c)(10), 211.113(b), 211.188
This firm did more than most. It halted sterile production on the first day of the inspection, engaged an air-certification vendor, finalised draft SOPs, and retrained its staff on gowning, cleanroom conduct, material transfer, visual inspection, and aseptic technique. FDA rejected the training-based corrective action every time it appeared — six times in one letter.
Issue 01 ended on a line it did not develop: training may appear in a response, but never as the root cause answer. This issue is about the other half. Retraining is not a corrective action. It is the delivery mechanism for one, and FDA evaluates it as such.
What makes this letter unusually useful is that FDA rejected the same move six times and gave its reasons each time. The reasons are the same four, in rotation.
What FDA found
The firm is a 503B outsourcing facility producing sterile injectables. Investigators recorded ten insanitary conditions in aseptic processing, most of them about first air: an operator placing gloved hands directly over open sterile containers; a second row of exposed vials set behind the first in the ISO 5 hood, obstructing airflow to critical surfaces; an operator resting their arms on the hood work surface; hands reaching outside the ISO 5 area for supplies and returning unsanitised; exposed skin in the aseptic processing area.
Alongside these, FDA cited visibly rusty equipment, no disinfection at transitions from lower to higher air quality, smoke studies that did not demonstrate unidirectional airflow under dynamic conditions, and CGMP violations spanning the quality control unit, gowning, cleaning and disinfection, and batch records.
- 12 May 2025
- Inspection opens; the firm ceases sterile production the same day.
- 15 May 2025
- Form FDA 483 issued.
- 9 Jun 2025
- Firm responds. Retraining is the corrective action for six findings.
- 1 Aug 2025
- Sterile production resumes.
- 1 Dec 2025
- Warning letter. Every training-based corrective action is rejected.
What the firm said
The response was not defensive and did not blame anyone. For each observation the firm named the procedure at issue and trained its people to it — SOP 0003.00 on gowning “with emphasis on mirror self inspection,” SOP 0004 on cleanroom conduct, SOP 2001.00 on material transfer, SOP 2004 on visual inspection, plus internal training on aseptic operations. FDA acknowledged all of it, in identical form:
“We acknowledge your retraining to SOP 0003.00 with emphasis on mirror self inspection. However, your response is inadequate…” FDA, Warning Letter 717972
FDA also allowed that some corrective actions “appear adequate.” This is not a firm that ignored its 483. It is a firm that answered every observation with the most common corrective action in the industry.
Why FDA rejected it
The procedure was never changed
Staff were retrained on gowning with emphasis on mirror checks — but SOP 0003.00 “was not updated to reflect the correct gowning sequence or to establish procedures that enforce mirror checks and supervisory oversight.” The training taught something the document did not require. The same gap appears on disinfection protocols and on visual inspection, where defect classification and AQL criteria were never added.
The training did not cover the observed behaviours
FDA noted that aseptic technique retraining “did not address specific behaviors that compromise first air (e.g., blocking critical surfaces, resting forearms on ISO 5 surfaces, leaning into hoods and exposing skin)” — the exact conduct the investigators had documented. Nor did it reach the gowning deficiencies actually seen, including safety glasses in place of cleanroom goggles and crossing the demarcation line on the unclassified side.
Attendance is not competency
Across findings, FDA asked for the same missing evidence: “training curriculum, attendance, competency assessments, completion dates,” post-training assessment, and supervisory oversight of the retrained personnel. A signed training record establishes that a session occurred. It does not establish that behaviour changed, and it says nothing about the week after the trainer leaves.
Nothing was said about product already made
Retraining is prospective by construction, so it is silent on everything released under the old behaviour. FDA asked for what the response omitted: a retrospective review of batches inspected under improper techniques, an impact assessment for distributed lots AP-202505 through AP-202513, and a risk assessment for batches released without validated airflow studies.
The pattern worth carrying forward
Retraining is attractive because it is fast, cheap, documentable, and closes a CAPA record inside the 15-day window. It also carries a hidden premise: that the person knew less than they needed to. That premise is usually false. An operator who rests their forearms on an ISO 5 surface generally knows they shouldn’t. They do it because the hood is arranged so that it is the natural way to reach the work.
Training addresses knowledge. Almost none of these findings were knowledge problems — they were layout, workflow, supervision, and procedure problems, and training leaves every one of them exactly as it was.
- Did the procedure change? If the SOP still says what it said before the observation, you have trained people to work differently from your own document. The document wins on the next inspection.
- How will you know it worked, and who is watching? Name the competency assessment, the observer, and how often observation continues once the training closes. Attendance records answer neither question.
- What about the batches made before? Training is prospective and says nothing about released product. If your response has no retrospective review, FDA will ask for one — and will read its absence as a scope failure.
A defensible response would have paired each training commitment with the three things it cannot supply on its own: the revised procedure that makes the correct behaviour the documented one, the verification and ongoing observation that show it held, and the retrospective review that bounds what the old behaviour already released. Training would still appear — as the last step, not the answer.
Retraining tells FDA you believe the person was the problem. Everything FDA asks for next is a test of whether that was true.
Veritas Quality Consultants works with pharmaceutical, medical device, and biologics manufacturers on Form 483 responses, root cause analysis, and CAPA development — built to FDA expectations inside the 15-day window.
Request a consultation