The FDA rejected a retrospective MACO argument in a February 2025 warning letter — not because the arithmetic was wrong, but because the cleaning it assumed had not been performed. What cleaning-validation teams can take from it.
What happened
On 26 February 2025 the US FDA issued Warning Letter 320-25-48 to Granules India Limited (FEI 3004097901), signed by Francis Godwin, Director of CDER's Office of Manufacturing Quality. It follows an inspection of the firm's finished-dose site at Gagillapur Village, Telangana, from 26 August to 6 September 2024. FDA reviewed the firm's 27 September 2024 Form 483 response and found it inadequate. The letter cites three CGMP violations and a separate data-integrity concern; the first — 21 CFR 211.67(b), written procedures for cleaning and maintenance of equipment — is the one that matters here.
The Warning Letter identifies significant contamination in multiple ducts of non-dedicated equipment used in preparing finished drug products. Filters had been installed to prevent contamination, but FDA concluded that inadequate cleaning and maintenance rendered them ineffective. Swab samples from the ducts — specifically downstream of high-efficiency particulate air (HEPA) filters — detected residues from multiple previously manufactured drug products, plus microbial contamination reported as too numerous to count (TNTC). FDA also states the firm lacked documented cleaning procedures for the duct sections between the HEPA filter and the cleaning point. Product and equipment details are redacted as "(b)(4)".
In response, the firm supplied an impact assessment, limited reserve-sample testing and an assurance of low microbiological risk, judging the risk low based on retrospective maximum allowable carry-over (MACO) calculations. FDA rejected that reasoning:
> "Cross-contamination is not uniform, and testing of limited reserve samples alone cannot ensure products are contaminant-free. While MACO is helpful in validated cleaning scenarios, it cannot apply where cleaning has not been performed, or exposure is inconsistent."
FDA added that the batch-size denominators in the firm's MACO calculations were inconsistent with the size of the relevant vessels, and that the limits were not adequately justified.
Why it matters
This is not a story about arithmetic. It is a story about the boundary conditions of a calculation.
MACO answers one bounded question: given that a defined cleaning procedure has been executed on a defined surface, how much residue may remain and still be acceptable in the next product?
Where cleaning has not been performed, or where exposure is not uniform — residue accumulating in ductwork and redistributed by airflow — the calculation loses its foundation. There is no defined surface, no defined residue level and no executed cleaning step to evaluate. A carry-over figure cannot substitute for the missing procedure it presumes; it is persuasive only where that regime exists, is documented, and is verified in execution.
What is particularly interesting
What FDA asked for instead. Not a corrected calculation. It asked for a comprehensive, independent retrospective assessment of cleaning effectiveness: the identity of the residues, which other shared equipment may have been improperly cleaned, and whether cross-contaminated product was released.
The worst-case list. FDA enumerated the conditions it expects a programme to identify and evaluate: drugs of higher toxicity, higher potency, lower solubility in the cleaning solvent, and difficult-to-clean characteristics; the most difficult swabbing locations; and maximum hold times before cleaning.
Recovery, in one clause. The letter directs the firm to "ensure use of appropriate limits that take into account recovery study results" — a reminder that the limit and the sampling method demonstrating it are one argument, not two.
The neighbouring findings connect. Bird droppings and feathers were observed on air purification units and ducting (211.58), and the HEPA filter integrity procedure did not define the quality unit's role in verifying critical air-handling conditions (211.68(a)). The duct contamination sits downstream of maintenance and oversight, not independent of it.
How such a gap can realistically develop
Air-handling ductwork is rarely anyone's cleaning-validation scope. It is installed by engineering, maintained under a preventive-maintenance programme, and treated as infrastructure rather than product-contact equipment. Cleaning validation teams map surfaces that contact product; duct interiors sit outside that map, and no one owns the question until an inspector asks.
The MACO argument has a similar origin. When reserve samples confirm a contaminant is present, MACO is the tool a team reaches for — quantitative, already validated for process equipment, reassuring. What gets skipped is the qualifying step: asking whether the calculation's premise — a defined cleaning procedure, executed, with uniform exposure — holds here.
In a mature, busy, multi-product site these are less failures of competence than failures of scope: a boundary drawn once, correctly enough at the time, and never revisited as equipment and products changed.
Questions worth asking in our own programme
- Which document defines cleaning and maintenance of the internal surfaces of our air-handling ductwork and similar non-product-contact infrastructure, and who owns it? If none exists, what is the documented basis for that conclusion?
- Wherever a MACO value appears, is the cleaning procedure that the calculation presumes documented, executed at the assumed frequency, and verified in execution?
- Would our worst-case rationale actually select a low-solubility, high-potency compound? If not, what does the rationale say about solubility and potency as selection drivers?
- Are our acceptance limits stated as recovery-adjusted values, traceable to a recovery study on the same material and method as routine sampling?
- Do our hold-time definitions state the events behind the timestamps — including the maximum time before cleaning begins — rather than only the intervals?
- Is there any shared equipment item in routine use without a current, approved cleaning procedure?
Wider context
A warning letter is a site-specific regulatory action, not a globally binding requirement, and should not be framed as universal law. Its citations sit under 21 CFR 211.67(b), 211.58 and 211.68(a), and its force is directed at one facility.
What travels beyond the site is the reasoning: a risk calculation is only as valid as the process it models. That is consistent with health-based exposure limit thinking, with lifecycle cleaning validation, and with expectations around worst-case selection and sampling recovery across FDA, EMA and PIC/S references. The firm's own reserve samples confirmed cross-contamination; the disagreement was about interpretation, not detection.
Sources
- FDA, Warning Letter 320-25-48 (MARCS-CMS 697115), Granules India Limited, 26 February 2025.
- US FDA, Data Integrity and Compliance With Drug CGMP (referenced in the letter).
- 21 CFR 211.58, 211.67(b), 211.68(a).
Validation takeaway
A MACO calculation is only valid inside the scenario it models. If the cleaning procedure it presumes is not documented, not executed, or the exposure is not uniform, the number carries no weight — and the corrective action FDA asks for is an assessment of cleaning effectiveness, not a better calculation.
Questions worth asking in our own programme
- Which document defines cleaning and maintenance of the internal surfaces of our air-handling ductwork and similar non-product-contact infrastructure, and who owns it? If none exists, what is the documented basis for concluding it is not needed?
- Wherever a MACO value appears in our records, is the cleaning procedure that the calculation presumes documented, executed at the assumed frequency, and verified in execution?
- Would our worst-case rationale actually select a low-solubility, high-potency compound? If not, what does the rationale say about solubility and potency as selection drivers?
- Are our acceptance limits stated as recovery-adjusted values, traceable to a recovery study performed on the same material and method as routine sampling?
- Do our hold-time definitions state the events behind the timestamps — including the maximum time before cleaning begins — rather than only the intervals?
- Is there any shared equipment item in routine use without a current, approved cleaning procedure?
