The FDA's August 2025 warning letter to ABBE Laboratories describes a 2009 cleaning validation still standing while a harder-to-clean product was introduced onto the same shared kettle — and no change-control step that caught it.
What happened
On 6 August 2025 the US FDA issued Warning Letter 320-25-98 to ABBE Laboratories, Inc. (FEI 3004433842), an OTC drug contract manufacturer and repackager in Farmingdale, New York. Signed by Francis Godwin, Director of CDER's Office of Manufacturing Quality, it follows an inspection from 24 to 27 February 2025 and calls the firm's 11 March 2025 response inadequate. It cites four CGMP violations; process validation and annual product review were repeat findings from 2018.
The second violation, under 21 CFR 211.67(b), is the one that belongs in a cleaning-validation discussion. FDA states the firm has not demonstrated that its cleaning practices adequately remove contaminants from the shared equipment used for its OTC drug products. Two observations follow. The 2009 cleaning validation study for the Kettle (b)(4) did not include a determination of the worst-case drug product for that equipment and lacked the identity of the drug product cleaned during the study. And the firm failed to perform a new cleaning validation study when it began manufacturing a new drug product, End Zit Acne Drying Lotion, on the same shared equipment — a product the letter says the firm itself determined to be the most difficult to clean, because of its titanium dioxide active ingredient. The study also lacked analysis for API residues.
The letter adds a documentation lapse: procedure ABBE-010 requires all cleaning and sanitisation information in logbooks, but the firm stopped documenting cleaning activities and equipment status around 2020.
FDA asked for a CAPA plan based on a retrospective assessment of the cleaning programme; improvements to the cleaning validation programme emphasising worst case — drugs of lower solubility in their cleaning solvents, drugs difficult to clean, the hardest-to-clean swabbing locations, and maximum dirty and clean hold times; and, in the same list, a description of the steps that must be taken in the change management system before introduction of new manufacturing equipment or a new product.
Why it matters
This is not a letter about a company that never validated its cleaning. ABBE Laboratories had a study on file. The problem is that it stayed where it was written while the portfolio moved on: a 2009 study, a documentation stop around 2020, and a harder-to-clean product run on the same kettle without a corresponding study. That failure mode is more uncomfortable than an absent validation, because it survives reviews that check for the existence of a document rather than its reach.
FDA's remedy is equally instructive: it links the cleaning validation programme and the change management system in one remediation bullet. The corrective action for a validation that did not track product changes is a described change-control step, not merely a new study.
What is particularly interesting
The firm appears to have done the scientific thinking and then not acted on it in the validation record. Per the letter, ABBE determined that End Zit Acne Drying Lotion was the most difficult to clean because of titanium dioxide — precisely the worst-case reasoning a cleaning programme is supposed to produce. The gap was not analytical judgement; it was treating that judgement as an observation rather than a trigger to repeat the study. Worst-case knowledge that does not change the validated state is operationally no different from not having it.
Two details sharpen the point. The 2009 study did not identify the drug product cleaned, so it could not later be shown to cover what followed. And the documentation stop means that for roughly five years there was no contemporaneous evidence that the defined cleaning steps had been performed — a failure no retrospective analysis repairs.
How such a gap can realistically develop
None of this requires negligence to be plausible. A 2009 study is written against the products and equipment of 2009, usually by people who have since left. A contract manufacturer's portfolio changes with its customers, and a new formulation arrives as a commercial decision with a dossier attached — not as a validation project. The kettle is already qualified, so no equipment change is raised, and the cleaning procedure still describes the same steps, so the SOP is not obviously wrong. Only one question is missing: does the validated state still describe what we now make on this equipment? If that question is nobody's specific job, the answer defaults to yes. Logbook documentation decays the same way — dropped in a busy period, never missed because nobody looks for the absence.
Questions worth asking in our own programme
- Which products does each cleaning validation actually cover, and can records show that every product made on that equipment since is inside that set?
- When a new product is introduced, what in our change-control procedure requires a written decision on whether it changes the worst case, and who signs it?
- If our own assessment concluded that a product is the most difficult to clean, what documented action follows, and where does that conclusion sit in the validation file?
- Does our cleaning validation include analysis for API residues, supported by recovery data for the sampling method used?
- Can contemporaneous records demonstrate that every defined cleaning step is performed as written?
- Are maximum dirty and clean hold times defined, validated and consistent with shop-floor practice?
Wider context
FDA has used comparable worst-case lists in several cleaning-related warning letters in recent years. The ABBE letter is a site-specific action under 21 CFR parts 210 and 211 and creates no new global requirement. What it illustrates is the lifecycle expectation behind 211.67: cleaning validation describes a defined state of control, and one never revisited against a changing portfolio is a document, not a control.
Sources
- US FDA, Warning Letter 320-25-98 to ABBE Laboratories, Inc., 6 August 2025 (MARCS-CMS 708309). https://www.fda.gov/inspections-compliance-enforcement-and-criminal-investigations/warning-letters/abbe-laboratories-inc-708309-08062025
- US FDA, Warning Letters database entry, ABBE Laboratories, Inc. (issue 08/06/2025; posted 08/12/2025; CDER).
- 21 CFR 211.67(b); 211.84(d)(1)–(2); 211.100(a); 211.22(d).
Validation takeaway
A cleaning validation describes a state of control at a point in time; every new product on shared equipment is a question about whether that state still holds. The FDA letter shows the failure mode in its mildest plausible form: the firm had identified the most difficult product to clean and still did not repeat the study — which is why the requested corrective action names the change-management step, not just the missing validation.
Questions worth asking in our own programme
- For each shared equipment train, which products does the current cleaning validation actually cover, and can we show from records that every product made on it since is inside that set?
- When a new product is introduced, what in our change-control procedure requires a written decision on whether it changes the worst case, and who signs it?
- If our own technical assessment concluded that a product is the most difficult to clean, what documented action follows, and where does that conclusion appear in the validation file?
- Does our cleaning validation include analysis for API residues on the shared equipment, supported by recovery data for the sampling method used?
- Can we demonstrate from contemporaneous records that every defined cleaning step is performed as written — today, not in principle?
- Are maximum dirty and clean hold times defined, validated and consistent with what actually happens on the shop floor?
