A 2026 FDA warning letter to Dabur India rejects a shared-equipment cleaning programme that relied exclusively on visual inspection. The objection is narrower than it appears — and it turns on missing quantitative limits, not on visual inspection itself.
What happened
FDA inspected Dabur India Limited's drug manufacturing facility at Village Saily, Silvassa (Dadra and Nagar Haveli and Daman and Diu, India), FEI 3001413302, from 12 to 16 January 2026. Warning Letter 320-26-105, dated 24 July 2026, was addressed to Chief Executive Officer Mr. Mohit Malhotra. It cites significant violations of CGMP for finished pharmaceuticals (21 CFR parts 210 and 211); the firm's drug products are adulterated under section 501(a)(2)(B) of the FD&C Act. FDA found the firm's 4 February 2026 Form FDA 483 response inadequate.
Cleaning is the letter's third observation, cited under 21 CFR 211.67(b). FDA's wording is specific: "your cleaning validation program for the non-dedicated Unit (b)(4) Line (b)(4) relied exclusively on the visual inspection of (b)(4) to determine adequate cleaning. Your cleaning validation program lacked scientifically determined maximum allowable carryover limits, quantitative residue testing, or direct surface sampling of difficult-to-clean areas." No cleaning validation at all had been performed for that line, despite at least (b)(4) OTC products being manufactured on it.
The firm committed to validated analytical methods, carryover limits, direct surface swab sampling and cleaning revalidation, and stated one line is now dedicated. FDA found this inadequate on three grounds: no interim cross-contamination controls during remediation; testing only (b)(4) samples across (b)(4) batches is "statistically insufficient"; and no retrospective review of all U.S.-market products within expiry affected by non-validated cleaning on non-dedicated equipment.
FDA also cites data integrity and process validation failures, and placed all site drugs on Import Alert 66-40 on 5 June 2026.
Why it matters
Read carefully, this is not an argument that visual inspection is unsound. FDA objects to exclusivity and to the absence of a quantitative anchor. A visual check not linked to a demonstrated visual detection limit answers a different question than a cleaning validation programme must: not "does the surface look clean," but "is the residue below a limit derived from toxicological and dose-based reasoning?"
A visual result is qualitative; a carryover limit is a number. Where the programme contains no number — no MACO, no residue method, no defined sampling points — there is nothing against which the observation can be checked, no process capability to establish, and no way to distinguish a comfortable pass from a marginal one. The gap is not visual inspection itself but the absence of the framework that would let it mean something.
What is particularly interesting
The requested remediation is a worst-case framework, not a scolding. FDA asks for identification of drugs with higher toxicities, higher potencies, lower solubility in the cleaning solvent, difficult-to-clean characteristics, swabbing locations for areas most difficult to clean, and maximum hold times before cleaning. That is the familiar anatomy of a defensible programme; the agency wants the framework to exist, not particular cleaning chemistry.
Sampling location justification is called out twice — as the "direct surface sampling of difficult-to-clean areas" that was absent, and again in the worst-case list. Where a programme is visual-only, the locations hardest to inspect and clean are exactly where the assumption is weakest.
Statistical sufficiency matters in retrospectives. A retrospective cross-contamination review is itself a sampling exercise needing a defensible size — and the interval between finding a gap and closing it is not risk-free.
How such a gap can realistically develop
Visual-only strategies rarely arrive as a decision; they accumulate. A programme may be written when a line runs a narrow product family, where visual examination was accepted as the practical criterion because the assumption of similarity held. Over years the product range broadens through routine business decisions, and the cleaning rationale is never revisited because nothing triggered a review. Change management covers the equipment and the new product, but not the question "does this invalidate our cleaning rationale?"
In parallel, visual inspection itself often remains unqualified: no study establishing its detection limit, no inspector qualification, no defined viewing conditions. It appears to work for years — until an inspector asks for the basis.
Questions worth asking in our own programme
- For every shared equipment train, is there a documented numeric acceptance limit, or does the rationale rest on visual examination alone?
- Where visual inspection is used, has a visual detection limit been established, and is it demonstrably below the health-based or dose-based residue limit?
- Are inspectors qualified, viewing conditions defined and the technique described in a procedure reviewed against ASTM E3263?
- When a product is introduced onto existing equipment, does change management formally ask whether the cleaning validation rationale still holds?
- Are sampling locations justified on cleanability, including the areas hardest to clean?
- If a cleaning validation gap were found today, what interim controls would protect product, and how large would a credible retrospective review have to be?
Wider context
The regulatory position is consistent and not new. EU GMP Annex 15 states that a visual check for cleanliness is an important part of the acceptance criteria but that "it is not generally acceptable for this criterion alone to be used." FDA's 1998 Guide to Inspections of Validation of Cleaning Processes notes that although visual examination should be considered, "relying only on visual examination would not be scientifically sound."
ASTM has meanwhile formalised the legitimate use of the technique. ASTM E3263 provides statistically valid procedures for determining the visual detection limit and qualifying inspectors; its scope is deliberately narrow, applying where risk management has justified the hazard analysis, the cleaning process is repeatable and validated, and visual inspection can be relied upon at the residue limit justified by the HBEL.
The fair reading is a qualified element within a quantitative framework, not a replacement for one.
Sources
- U.S. FDA, Warning Letter 320-26-105, Dabur India Limited, 24 July 2026 (MARCS-CMS 728828), https://www.fda.gov/inspections-compliance-enforcement-and-criminal-investigations/warning-letters/dabur-india-limited-728828-07242026
- ASTM E3263-22ε1, Standard Practice for Qualification of Visual Inspection of Pharmaceutical Manufacturing Equipment and Medical Devices for Residues.
- FDA, Guide to Inspections of Validation of Cleaning Processes, 1998.
- EudraLex Volume 4, EU GMP Annex 15: Qualification and Validation.
Validation takeaway
Visual inspection is a legitimate and often valuable element of a cleaning validation programme — but it is an element, not a substitute for one. FDA's objection to Dabur India's programme was that visual examination carried the entire acceptance decision for non-dedicated equipment, with no scientifically determined maximum allowable carryover limit, no quantitative residue testing and no direct surface sampling of difficult-to-clean areas. The practical lesson is one of linkage: a visual result only means something when anchored to a qualified visual detection limit and a numerically justified residue limit, and when sampling locations deliberately include the surfaces hardest to clean and hardest to inspect. Where a programme can demonstrate that linkage and a comfortable margin, visual inspection can be relied upon; where it cannot, the acceptance criterion is an assumption rather than a control.
Questions worth asking in our own programme
- For every shared equipment train, is there a documented numeric acceptance limit, or does the rationale rest on visual examination alone?
- Where visual inspection is used, has a visual detection limit been established, and is it demonstrably below the health-based or dose-based residue limit — with what margin?
- Are inspectors qualified, viewing conditions defined, and the technique described in a procedure reviewed against ASTM E3263 principles?
- When a new product is introduced onto existing equipment, does change management formally ask whether the cleaning validation rationale still holds?
- Are sampling locations chosen and justified on cleanability and inspectability, including the areas hardest to clean?
- If a cleaning validation gap were found today, what interim controls would protect product during remediation, and how large would a credible retrospective review have to be?
