Cleaning Validation: The Number That Sets Your Limit

Almost every active substance in commerce is made on equipment that made something else the week before. The industry's answer to that fact is a calculation: the maximum amount of the previous product that may carry into the next one and remain acceptable. The calculation has a name, a basis, and a direct effect on the impurity floor of the material a buyer receives.

It is also a subject that almost never comes up at purchase, which is odd, because it is one of the few quality controls that physically touches every batch.

How the number is built

Two approaches coexist. The dose-based one starts from the previous product's pharmacological potency: a fraction, conventionally one thousandth of its smallest single dose, becomes the allowed daily exposure for the next product's user, and that exposure is converted into a mass allowance by dividing by the next product's maximum daily intake. The concentration-based one is blunter: the previous product may not exceed ten parts per million in any subsequent batch.

The stricter of the applicable criteria governs. For a potent previous product the dose-based limit is far tighter than 10 ppm; for a benign one the reverse. The result is a mass, which is then distributed across the equipment surfaces, verified by swabs with a recovery study, and confirmed by rinse samples where the geometry allows.

Element	What it decides
Safety factor of the previous product	How small the allowance starts
Maximum daily intake of the next product	How the allowance converts to a concentration
Worst-case pairing	Which previous product sets the binding limit
Hardest-to-clean location	Where the swab must be taken
Swab recovery study	Whether the measured value can be trusted

Why the pairing matters more than the formula

The formula is generic. The pairing is specific, and for a bile-acid product the worst case is usually another bile acid. Structurally close molecules are chromatographically close, which means carryover can hide inside the related-substances profile rather than appearing as a discrete unknown peak. It is also pharmacologically related, which drives the safety factor down and the allowance with it.

A site that runs a single product family has a simpler problem and a sharper question: which member of the family sets the limit, and does the swab method distinguish it from the product being made? A site that runs a mixed portfolio has the more common problem, and the answer is a matrix of pairings with the worst case identified.

Visual cleanliness and its ceiling

Visual inspection is the cheapest control and the most misunderstood. Under good lighting, a trained eye detects residue at around a milligram per square metre, sometimes better on a smooth surface. Calculated allowances are routinely a hundred to a thousand times lower.

Visual cleanliness is therefore a precondition: a visibly soiled surface has failed before any swab is taken. It is not a demonstration that the allowance was met, and a cleaning record that consists of a signature and the word "visually clean" is describing a preliminary step as though it were the control.

The swab is the control, and the recovery study is what makes the swab meaningful. A swab that recovers half of what is present reports half the residue, and the allowance is met arithmetically while the surface carries twice the limit.

What this means for a bile-acid buyer

The carryover allowance sets a floor under the impurity profile: any related substance observed has to be distinguishable from what the previous product could contribute. For most of the family that is manageable, because the profile is characterised. The two places it bites are the mass balance and the inorganic residue, where an unexplained contribution has to be attributed to something.

It also matters when the previous product is not a bile acid at all but a processing aid, a dye from a non-pharma campaign or a material with an element that the elemental impurity assessment did not anticipate. Campaign scheduling is a quality decision for exactly this reason, and it is a legitimate question at audit.

What a buyer can reasonably ask

Three facts, none of which compromises a supplier's position. The basis on which the MACO applying to their product was calculated. The worst-case pairing that was validated for the line in question. And whether the swab method's recovery has been studied for that pairing. The full protocol stays with the site; the three facts are what make the limit interpretable rather than decorative.

The regulatory framing is public: the ICH Q7 guidance sets the expectations for cleaning and validation, the EudraLex volume 4 annexes describe the European interpretation, and the EDQM publishes the certification conditions that assume cross-contamination is controlled.

One question tests the whole system: what ran on this line immediately before our batch, and what limit governed the changeover? An answer with a number and a calculation is a control. An answer with a cleaning log reference is a filing system.

Where this leaves a supply-chain partner

We can state the basis of the carryover limit that applied to a given batch, the worst-case pairing validated for the line, and what the changeover verification showed. Where a customer's sensitivity to a specific carryover is higher than the standard allowance, we discuss campaign position and dedicated equipment rather than assuming the general limit is enough.

That conversation happens rarely and it is the one that prevents the deviation nobody can explain: the unattributable peak on an otherwise clean certificate.

The swab, the recovery study and the geometry

The swab is where the calculated allowance meets the physical surface. Its result is a mass recovered from a defined area, extrapolated to the whole equipment train, and the extrapolation assumes the residue distributes the way the validation's worst case did. Three practical elements decide whether the assumption holds.

Location. The swab must be taken at the hardest-to-clean position, which is rarely the centre of a vessel: it is a valve body, a gasket groove, a spray ball shadow. A protocol that swabs convenient locations is sampling the easy geometry and calling it the equipment.

Recovery. The swab method is itself validated by spiking a known mass onto a coupon of the same material and measuring what comes back. A recovery of seventy per cent means the true residue is a fraction higher than the measurement, and the limit comparison has to be made after that correction. Skipping the recovery study does not make the limit stricter; it makes the compliance claim uninterpretable.

Rinse sampling. Where the geometry defeats swabbing, a final rinse is analysed instead, and it measures the solvent's ability to carry residue away rather than the residue itself. The two are related but not identical, and a rinse result alone is a weaker demonstration than a swab at the worst location.

What a buyer does with this

The point of asking is not to audit the supplier's protocol. It is to interpret the impurity line on the certificate with the right baseline. A related substance at 0.06 per cent means something different on a line whose validated carryover could contribute 0.02 per cent of a structurally similar compound than on a dedicated line, and the difference is invisible without the three facts.

The regulatory framing is public and consistent: the ICH Q7 guidance treats cleaning as a validated process with defined acceptance criteria, the European GMP text sets out the same expectation with annex detail, and the EDQM certification framework assumes cross-contamination control for every certificate it supports.

One question maps the whole subject: what was the worst-case pairing validated for the line that made this batch, and what allowance did it set? An answer with two numbers is a control. An answer describing the cleaning procedure in general terms is a procedure, which is a different document with a different job.

Campaign scheduling as a control

The carryover allowance is a number, and the campaign order is the decision that gives it meaning. A line that runs a potent product and then a sensitive one has a hard case; the reverse order is easier. Sites that understand this sequence their campaigns deliberately, and sites that do not sequence them by customer deadline.

Two practices reduce the exposure without new equipment. A campaign-end cleaning with a longer verification, used when the next product is the sensitive one. And a bracketing approach where the swab result after the changeover is compared against the specific pairing rather than against the line's general limit.

Dedicated equipment is the third option and the most expensive. It is justified where the carried material's limit is so low that cleaning verification cannot demonstrate compliance at a reasonable cost, and the decision belongs in the supplier's quality system rather than in a commercial negotiation.

What the numbers on a swab report mean

A swab result is a mass per area, corrected for recovery, extrapolated across the equipment train. Three conversions sit between that number and the customer's risk: area to total residue, total residue to carryover into the next batch, and carryover into the next batch to a concentration in a dose. Each conversion carries an assumption, and the assumptions are what the MACO calculation exists to make explicit.

For a buyer, the useful question is not the swab number. It is which assumption the number depends on, and whether that assumption was set for a product like theirs. That is a two-sentence answer for a site that has done the calculation, and a long silence for one that has inherited a limit.

Frequently asked questions

What is a maximum allowable carryover?

The largest quantity of the previous product that may remain on shared equipment and still be acceptable in the next batch. It is calculated from safety or dose-based criteria, distributed over the next batch's size, and verified by swab or rinse results after cleaning.

Why is visually clean not a limit?

Because the eye detects residue down to roughly a milligram per square metre under good lighting, which is orders of magnitude coarser than the calculated allowance. Visual cleanliness is a precondition of a valid swab, not a substitute for one.

How is the carryover calculated?

The dose-based approach takes one thousandth of the previous product's smallest single dose, or a safety-factor variant, and divides by the maximum daily intake of the next product. The 10 ppm approach caps the previous product at ten parts per million in any next batch. The stricter of the applicable criteria governs.

Does carryover matter more for a bile acid than for other products?

It matters whenever the previous product shares an analytical blind spot with the next one. Related bile acids are structurally close, chromatographically similar and pharmacologically related, so the worst-case pairing for a bile-acid line is another bile acid.

What should a buyer be able to ask for?

The basis of the MACO calculation as it applies to their product, the recovery studies behind the swab method, and the worst-case pairing that was validated. Not the full protocol, which belongs to the site, but the three facts that make the limit interpretable.

How does cleaning validation interact with the certificate?

Indirectly but materially: an unidentified impurity in the next batch may be the previous product. That is why a certificate with an unexplained peak deserves the question about what ran on that line before, and why the question is reasonable rather than accusatory.