WorkflowQuality Operations & CAPA

Cleaning Validation (MACO)

Cleaning validation kept current on shared trains: when a product enters, agents recompute each carryover limit, re-rank the worst case and check every swab site, then the lead and QA sign.

Follow one new product onto a shared train ↓
Watch with transcript →
The problem

A new product's small batch quietly lowers the other limits on the train

In a contract plant, many client products share one equipment train. Each residue limit on it depends on the next product's batch and daily dose, the previous product's exposure limit, the train's whole contact area and each swab site's recovery. When technology transfer adds a product, the lead types one new row into the matrix spreadsheet and redoes the governing pair.

The client's exposure limit is still to be provided and arrives later as a PDF in an email thread. Whether the old study covers the new limits is judged by eye. Results that no longer meet them surface at the first clean after the new product, or in front of an inspector.

Built forCleaning validation leadsValidation engineersQC analystsToxicologists

The Problem Today

A new product on a shared train changes the carryover limit of each product before it, and old swab results go unread against the new limits.

10 ppm
an example residue limit in FDA's cleaning validation guide published

Redone by hand for each product

  • Batch sizes
  • Daily doses
  • PDE reports
  • Swab data
  • Old results against new limitsA site that passed the last study can fail the new sum
  • Exposure limits chased by emailThe client's report arrives as a PDF in a thread, often late

The Solution

  1. The train re-read, site by siteItems, materials, swab and rinse sites and results on file, each cited
  2. Every product pair recomputedHealth-based and parts-per-million limits for each next product, the lowest taken
  3. A challenger reads it as an inspectorIt caught a stainless-steel recovery copied onto a silicone gasket
  4. The study signed twiceThe lead approves, QA countersigns, then each client is told

Built-in controls

  • Each limit traced to the record it came from
  • No limit from an exposure limit nobody accepted
  • No agent edits the cleaning procedure

What You Get

Each new client product enters a shared train with its residue limits recomputed and its study planned before its first batch.

  • More client products per shared trainA new product joins an existing line instead of waiting for its own
  • The sums outlive their authorA new engineer reads the same written-out sum and challenger
  • Clients see which limits movedEach client whose limit changed is told, with the sum behind it
Features

Cleaning validation re-checked pair by pair when a new product joins

A client's new tablet enters a shared granulation, compression and coating train, and validation, toxicology and QA each act on the re-check.

demoThe plant, its client products, the toxicologists and every limit shown are sample data.

  1. 01Shared train

    Cleaning validation matrix drawn on the equipment, each swab site pinned

    • Each swab and rinse site on its item. The lead sees where the residue evidence sits on the equipment itself, with its material, limit and status a click away.
    • The matrix in force, as approved. Before any agent reads, the drawing shows the approved matrix as it stands, so nothing new appears on screen unread.
    • A banner for the waiting product. The change control that adds a product opens a re-check with its due date, so the impact is assessed before the first batch.
  2. 02Re-check

    Shared equipment re-check that shows which swab sites lost their margin

    • Six agents, each with its source. Each agent shows the passage it read and its confidence, so the validation engineer checks the work behind each figure.
    • Sites turn as their limit falls. A site whose old result no longer meets its new reading limit turns red on the drawing and is marked as needing a new study.
    • The sum redone as a draft. The carryover sum under the drawing is labeled a draft revision, and the banner says nothing is approved yet.
  3. 03Product matrix

    MACO calculation for each product pair, with the lowest limit taken

    • Health-based and 10 ppm, pair by pair. Each next product's own batch and largest daily dose set the limit for each residue, and the lower of the two figures governs.
    • The worst case product re-scored. Solubility, cleanability and toxicity are scored again, so the lead sees whether the validated marker still covers the new product.
    • A missing PDE marked missing. No health-based limit is computed for a residue whose exposure limit has not arrived, and the ranking says which toxicity score it assumed.
  4. 04Challenger

    Inspector-style challenge of the draft that catches the agents' own error

    • A steel recovery on a silicone gasket. The challenger found that the matrix in force and the agents' draft both used the stainless-steel recovery for a silicone gasket.
    • Each point answered from its clause. Questions on the area counted, the pairing and the lower limit are answered with the clause that settles them.
    • What no record settles goes to a person. The clean hold time and the exposure report each go to the cleaning validation lead or the toxicologist to decide.
  5. 05Swab site

    Swab limit per site, before and after, with each past run's share

    • One site's sum, written out. Limit per area, times swab area, times the material's recovery gives the reading limit, the way the validation report writes it.
    • Each past run against both limits. The validation engineer sees the share each of the three runs took of the old limit and of the draft, side by side.
    • What the new study does at this site. A site that needs a new study is swabbed in duplicate in each run, and after each clean until the report is signed.
  6. 06PDE request

    PDE for cleaning validation requested from the client in fixed fields

    • Fields for the PDE and its basis. Point of departure, adjustment factors, critical effect and the signed report come back as data the agents can check.
    • Drafted by agents, sent by the lead. The cleaning validation lead sends the request, and its due date follows the client's quality agreement.
  7. 07Client portal

    The client's toxicologist answers in a portal and attaches the report

    • Only the client's own requests. The client's toxicologist sees the requests addressed to them and none of the CDMO's other records.
    • A comment changes nothing by itself. Free text is set aside for a person to read, so only the fields and the attached report can reach the matrix.
  8. 08PDE acceptance

    HBEL accepted by the CDMO's toxicologist only after opening the report

    • Fields checked against the report. The agents check that the answer matches the signed report and that its factors reproduce the stated PDE before anyone may accept it.
    • Accepted by the toxicologist alone. The CDMO's toxicologist opens the report, picks a reason and accepts; only then are the new product's health-based limits computed.
  9. 09Study plan

    Cleaning validation protocol approved by the lead, countersigned by QA

    • Three consecutive full cleans. The drafted study names its runs and sites, the recovery study and swab method due before run one, and that a failed run opens a deviation.
    • Two signatures, two people. The cleaning validation lead approves and QA validation countersigns; only then does the new matrix come into force.
    • Clients told what moved. QA validation sends each client whose limit changed a notification, as its quality agreement asks.

The lead owns the matrix; a toxicologist owns every exposure limit in it

Validation runs the re-check and the study, QC owns what a swab can measure, and toxicology decides which exposure limits the sums use.

Cleaning validation leads

Each train's product matrix, its re-checks and the study plans they sign

  • The impact of a new product seen on the equipment
  • Exposure limits back as fields, not email
  • A plan that lists what still blocks it
Validation engineers

Start re-checks, answer challenger points, approve verification results

  • Each sum written out to check by hand
  • Rehearsals of a batch, dose or PDE change
  • Results to approve that someone else entered
QC analysts

Swab and rinse methods, recovery studies, quantitation limits and results

  • Limits below the method's reach flagged early
  • Each site's recovery tied to its material
  • Results entered against the limits in force
Toxicologists

Read each client exposure report before a limit is computed from it

  • Reports whose factors are already checked
  • Acceptance or send-back with a reason code
  • No limit set on a number they have not read
More screens

The gasket's limit fell twice, once for silicone and once for the new product

The site the challenger corrected, the question only a person could answer, a rehearsal, the week's verifications, every train, and the signatures.

The gasket at its true recoveryThe matrix carried the gasket at steel's 85%; at silicone's 62% its old limit was 38.3 µg, and against the draft limit its three runs reach 94% to 108%.
A clean hold question only the lead answersThe old study held cleaned equipment for 48 hours; the lead picks a 72-hour time point for run three and a reason code before closing the point.
A limit the swab method could not proveRehearsing a far lower exposure limit for the residue drops the train's limit to 25 mg and puts most sites below what the method can measure; nothing is saved.
Cleaning verification after each cleanQC enters swab results against the limits in force, a validation engineer who did not enter them approves, and a failed clean opens a deviation instead of a re-clean.
All shared trains in four plantsEach train with its products, sites, lowest limit and worst case, open re-checks with their due dates, and annual reviews confirmed by opening a sample first.
Clean hold, approval and countersignature, each with a reasonThe clean hold decision, the lead's approval and QA's countersignature on the study, each with its reason code, its meaning and the session.

demoSample data, from the working solution.

The toxicologist signs before any limit is computed, and QA signs last

Toxicologist at the CDMO
Sees
The client's signed report, and whether its factors reproduce the stated PDE
Signs
Acceptance of the exposure limit, or a send-back with a reason code
Cleaning validation lead
Sees
The plan, the sums and each challenger point, opened before signing
Signs
Approval of the matrix and the study as written
QA validation
Sees
The plan read to its end against the limits procedure
Signs
The countersignature that puts the new matrix in force, or a send-back
Validation engineer
Sees
Swab and rinse results a QC analyst entered
Signs
Approval of a cleaning verification they did not enter

A failed swab opens a deviation instead of another clean and retest

Re-cleaning to passEU GMP's annex on qualification and validation rules out repeated cleaning and retesting until acceptable, so a failed verification opens a deviation.
Editing the cleaning procedureThe agents draft studies and verification plans; the procedure itself changes only through change control.
A limit from an unread exposure reportHealth-based limits wait until the CDMO's toxicologist has opened the signed report and accepted it.
Releasing equipment for the next batchA validation engineer who did not enter the results approves them; until then the train stays cleaned, awaiting results.

Limits rebuilt from the client's exposure reports and the plant's swab files

Product registerEach product's batch size and largest daily dose, per train.
Client exposure-limit reportsSigned PDE reports from each client's toxicologist, read only once accepted.
Equipment registerItems, product-contact areas and materials of each shared train.
Cleaning validation reports and recovery studiesThe sites, runs and results on file, and each material's recovery.
Swab and rinse methodsEach method's quantitation limit, so no limit is set below what it can measure.
Cleaning and use logs and QC sample logsThe verification results after each clean, by train and site.

Questions validation teams ask about cleaning validation

What is cleaning validation?

Cleaning validation is documented proof that a cleaning procedure removes each product's residue to a limit that protects the next product made on the same equipment. Here agents keep that proof current on shared trains: when a product enters, they recompute each carryover limit, re-rank the worst case and compare the results on file with the new limits.

How is MACO calculated for a shared train?

For each pair, the residue's exposure limit times the next product's batch size, divided by that product's largest daily dose, gives the health-based limit. A parts-per-million limit on the next batch is computed beside it, and the lower one is taken. The lowest pair sets the train's limit, which is spread over the whole product-contact area and then over each swab.

How is the worst case product chosen?

On solubility, cleanability and toxicity, as EU GMP's annex on qualification and validation asks for shared equipment. The agents re-score each product when one enters and say whether the validated marker still covers the newcomer; a hard-to-clean product can become the worst case before its exposure limit is known.

Why does swab recovery depend on the material?

Recovery is measured per material in the approved recovery study, and a factor copied from stainless steel onto a gasket makes that site's limit look looser than it is. The challenger checks each site's material against the study before the plan can be signed.

What happens when the client has not supplied a PDE?

No health-based limit is computed from it. The agents draft a structured request to the client's toxicologist, the lead sends it, and the CDMO's toxicologist accepts the answer only after opening the signed report and seeing the factors reproduce it.

Does cleaning verification continue while the new study runs?

Yes. Until the study report is signed, each clean before and after the new product is swabbed at every site, and the next batch cannot be charged until QC's results are approved by someone who did not enter them.

Pick the shared train
you trust least.

Send one train's product register, validation report and recovery study, and we will re-check its limits pair by pair.