Postdoctoral PDDM922 ⭐ 2 Credits Semester 3

Cleaning Validation and Cross-Contamination Control for Semisolids

Builds on the analytical-validation and investigation skills from PDQC802 and PDQC891 and teaches only what is different when the residue is a cream, ointment, lotion or gel rather than a peptide solution. The central problem is cleanability. Hydrocarbon bases (petrolatum, mineral oil, hard fat), waxes and fatty alcohols, emulsifiers, dimethicone, insoluble actives such as zinc oxide, and swollen carbomer gels do not dissolve in water. They collect on scraper blades, rotor-stator heads, in-line homogenizer loops, hoses, dead legs, filler hoppers, pistons and nozzles. The FDA Guide to Inspections of Topical Drug Products (7/94) names insoluble excipients and inaccessible hoses and pipes as specific concerns. It also notes that firms may dedicate lines and hoses. The course is anchored on 21 CFR 211.67 (written cleaning procedures) and the FDA Guide to Inspections: Validation of Cleaning Processes (7/93). Key expectations from those documents: - Swab sampling suits hard-to-clean areas and insoluble residues. - Rinse sampling suits soluble residues. - Judging rinse water only against water specifications is unacceptable. - Swab adhesives can interfere with analysis, so recovery must be established early. Limits. The historical industry figures quoted in the 7/93 guide (10 ppm, 1/1000 of the therapeutic dose, no visible residue) are taught as history, not as defaults. Every limit must be justified for the specific product. Health-based exposure limit (HBEL) and maximum allowable carryover (MACO) calculations are presented as industry practice (verify current FDA/EMA/ISPE expectations). Toxicological inputs are supplied by a qualified toxicologist or the sponsor. The platform does not make human dosing determinations. Labeled directions are used only as documented labeling inputs. Shared versus dedicated equipment. Whether to share, campaign or dedicate equipment is decided by a documented assessment of each product's flags: - Potency and occupational exposure: most potent actives fall outside OTC scope, and a Florida OTC manufacturer permit holder may not possess Rx drugs. - Oxidizers: benzoyl peroxide raises both a benzene and a compatibility question. Dedication for BPO is a documented decision, not an automatic one. - Sensitizers. - Microbial risk: wet-hold and Burkholderia cepacia complex (Bcc) risk for water-containing products. The course keeps two kinds of evidence strictly separate: - Analytical evidence: residue method capability and swab/rinse recovery on stainless coupons. - Process evidence: validation of the cleaning procedure itself, covering cleaning parameters, dirty and clean hold times, worst-case selection, executed runs and lifecycle monitoring. Work is mapped to the real lab. Specific residue assays run on the 3x Agilent 1260 HPLC-DAD, or on the LC-MS/MS where UV response is inadequate. Capability gaps go to contract labs through explicit capability exits: - TOC: no analyzer is owned. - Elemental residues: no ICP. - Volatile cleaning agents: no GC-HS. - Surface bioburden counts: no plate-count micro lab yet, and BacT/ALERT cannot give cfu counts. The course is API-agnostic throughout. Every exercise starts from an arbitrary dermatologist specification that has already passed the PDDM901 classification gate.

📅 Studied by Skipper on 2026-10-16

📖 Core Concepts

Claude Opus 5.5

Core Concepts: Cleaning Validation and Cross-Contamination Control for Semisolids

PDDM922 · Personal study notes for Skipper · Builds directly on PDQC802 (validation statistics), PDQC803/811 (HPLC and MS methods), PDQC823 (QC microbiology), PDDM903 (semisolid formulation science), PDDM911 (facility and process safety) and PDDM921 (process train and hold times).

How to read these notes. Items marked [verify] are industry conventions, recalled figures, or regulatory text that must be checked against the current official document before they go into a controlled procedure. Any number in a worked example labeled hypothetical is a placeholder chosen to demonstrate arithmetic. It is not a toxicological, dosing, or usage value. This program makes no human dosing determinations. Those inputs come signed from a qualified toxicologist or the sponsor.

1. What This Subject Is Really About

"Cleaning validation" sounds like janitorial work with paperwork attached. It is actually one of the more intellectually demanding problems in pharmaceutical manufacturing, because it asks you to prove a negative at a distance. You must show that something is not present, below a level you had to justify from toxicology. That level must hold at locations you cannot see, on surfaces you sample only partially, using a method that recovers only a fraction of what is there, and it must hold for every future product that might follow, including products that do not exist yet.

Strip away the vocabulary and every cleaning validation program answers one question:

If product B is made on equipment that last held product A, how much of A (plus cleaning agent, plus microorganisms) could end up in B? And do we have defensible evidence that this amount is acceptably small every time the procedure is followed?

Notice the three contaminant classes in that sentence: previous product, cleaning agent, and microorganisms. In the PDQC peptide world you mostly worried about the first. For nonsterile water-containing semisolids, the third can matter as much or more, and the second is more significant because you need stronger chemistry to remove grease.

The semisolid delta: a frying pan, not a water glass

A useful analogy is the difference between washing a glass that held sugar water and washing a cast-iron pan that held bacon fat and burnt cheese.

  • The glass (peptide residue in an aqueous process) is easy. The residue dissolves in water, there is little of it per unit area, and a rinse removes it. Your main analytical worry is whether you can detect the trace that remains. Peptides also adsorb to surfaces, so loss to the surface is the recovery problem you met in PDQC work.
  • The pan (an ointment or cream) is different in kind, not just degree. The residue is hydrophobic, so water beads off it. It is multiphase: oil, water, emulsifier, and sometimes insoluble pigment. It is high in solids, so grams per square meter cling to walls after discharge rather than micrograms. It also changes physical state with temperature, so it can melt off a hot surface and re-solidify in a cold hose downstream. If water is in the formula, the leftover film plus moisture is a culture medium.

FDA anticipated this in its topical-products inspection guide, which singles out insoluble excipients, hard-to-clean hoses and transfer lines, and the acceptability of dedicating such parts [FDA Guide to Inspections of Topical Drug Products, 7/94]. The general cleaning guide sets the logic that everything else rests on: written procedures, scientifically justified limits, analytical methods capable of detecting the residue, and sampling whose recovery is known [FDA Guide to Inspections: Validation of Cleaning Processes, 7/93]. LeBlanc's text is still the clearest single source on the engineering logic of limits, sampling, and recovery [LeBlanc, 2000]. Pluta's PDA volume shows how practitioners actually build these programs [Pluta (ed.), 2009].

The intellectual core of the course fits in one sentence. A cleaning validation result is a chain of four independent claims (limit, sampling, method, procedure), and the chain is only as strong as its weakest link. Most of what follows elaborates each link and the ways it breaks.

2. Concept One: The Regulatory Skeleton and What Investigators Actually Hunt For

21 CFR 211.67 requires that equipment be cleaned, maintained, and (where appropriate) sanitized at appropriate intervals to prevent contamination that would alter the drug product's safety, identity, strength, quality, or purity. Paragraph (b) requires written procedures. Recall its elements as a checklist, and verify the exact wording against the current eCFR [verify]:

  1. assignment of responsibility for cleaning and maintenance;
  2. maintenance and cleaning schedules, including sanitizing schedules where appropriate;
  3. a description in sufficient detail of the methods, equipment, and materials used, including disassembly and reassembly as needed to assure proper cleaning;
  4. removal or obliteration of previous batch identification;
  5. protection of clean equipment from contamination prior to use;
  6. inspection of equipment for cleanliness immediately before use.

Paragraph (c) ties cleaning to records. Elsewhere, 211.182 requires an equipment cleaning and use log, and 211.188 requires batch production records. Note what 211.67 does not say: it never uses the word "validation." The expectation to validate cleaning comes from the 7/93 guide, the general requirement that procedures be suitable, and decades of enforcement. Do not claim 211.67 explicitly mandates validation. It mandates effective, written, controlled cleaning, and validation is how you prove effectiveness.

The 7/93 guide: three failure patterns to internalize

Paraphrasing the guide's main expectations (quote it directly in controlled documents only after checking the text) [FDA, 7/93]:

  • "Test until clean" is not validation. Cleaning, sampling, testing, re-cleaning, and repeating until a pass appears proves only that you eventually got lucky. It says nothing about whether the written procedure, executed once, reliably works. A validated procedure passes the first time, every time it is followed.
  • Rinse water judged only against Purified Water specifications is not a residue test. Conductivity and TOC on a final rinse tell you the rinse looks like water. They do not tell you whether a hydrophobic API, petrolatum, or zinc oxide is still on the wall. Petrolatum barely partitions into water, so the rinse can be pristine while the vessel is filthy. The rinse must be analyzed for the target residue against a calculated limit.
  • Limits need a scientific rationale, and sampling needs demonstrated recovery. The guide notes that firms commonly cited limits such as 10 ppm, 1/1000 of a therapeutic dose, and no visible residue. It reports these as practices; it does not endorse them as universal (see Concept Six).

The 7/94 topical guide: why semisolids got their own document

The topical guide recognized problems specific to this dosage form: insoluble or poorly soluble excipients that resist aqueous cleaning, long transfer hoses and piping that cannot be inspected, and the practical acceptability of dedicating hoses and lines to a product [FDA, 7/94]. That last point is important for an operation with limited capital. A dedicated hose costs little compared with a validated hose-cleaning procedure and its sampling program.

Reading warning letters as data, not gossip

The course reading includes the 2025 Strukmyer Medical warning letter. It cites 211.67(a) alongside water-system contamination involving Burkholderia and Pseudomonas (check the specific citations against the posted letter text [verify]). The instructive pattern is the combination, which recurs across letters:

  • inadequate cleaning and an uncontrolled water system, because the water is the final rinse;
  • a quality unit that did not enforce procedures (211.22);
  • investigations that stopped at "operator error" or "retested, passed" (211.192).

The lesson for system design is that cleaning validation is not a standalone document. It sits on a qualified Purified Water system (PDDM911), an empowered QU, and an investigation culture that does not accept convenient root causes.

This is the organizing idea of the whole course. When you read "Vessel V-101 passed cleaning validation," four separate claims are bundled together:

 CLAIM                     EVIDENCE DOCUMENT                     IF THIS LINK FAILS...
 ───────────────────────── ───────────────────────────────────── ─────────────────────────────────────
 1. LIMIT is justified      Limit-derivation worksheet (HBEL,     ...you passed against the wrong
                            batch size, area) + tox sign-off      number
 2. SAMPLING recovers       Recovery study report (coupons,       ...residue was present but stayed
    residue                 spike levels, operators, R)           on the wall or in the swab head
 3. METHOD measures it      Method validation report              ...residue was in the vial but
                            (specificity, LOQ, accuracy, etc.)    invisible to the detector
 4. PROCEDURE removes it    Cleaning process validation           ...the SOP works only when the
                            protocol + report (runs, holds)       "good" operator runs it

Two consequences follow:

  1. Sequencing of evidence. Links 1–3 must be approved before link 4 is executed. If you run the cleaning validation protocol and then discover your recovery is 30 % or that the swab adhesive co-elutes with the API, every "pass" from those runs is meaningless. Process validation proves the procedure. It cannot simultaneously prove the method you used to judge it. You already know the PDQC802 version of this: you validate the analytical procedure before you release a batch with it.
  2. Failure diagnosis. When a result fails, the first investigative question is which link broke. A high swab result could mean dirty equipment (procedure), a contaminated swab or template (sampling), a co-eluting peak (method), or a mis-entered surface area (limit). A 211.192 investigation must discriminate among these with evidence, not assumption (Concept Eleven).

4. Concept Three: Residue Chemistry Determines Cleanability

You cannot design a cleaning process for a residue you do not understand physically. The Handbook of Pharmaceutical Excipients [Rowe, Sheskey & Quinn, 2009] is your reference for solubility, melting behavior, and incompatibilities. The topical QbD papers [Chang et al., 2013a; 2013b] and the topical classification scheme [Buhse et al., 2005] tell you what kinds of bases to expect.

Residue classes and how each resists cleaning

ClassExamplesRemoval mechanismWhy it resists
Hydrocarbon basesWhite petrolatum, mineral oil, hard paraffinMelting + emulsification/solubilization by surfactant; solvent dissolutionNot saponifiable; water-insoluble; re-solidifies on cool surfaces
Fatty alcohols, waxes, estersCetyl/stearyl alcohol, beeswax, glyceryl stearateMelting + emulsification; esters partially saponifiable by alkaliCrystalline, high melting; harden during dirty hold
Nonionic emulsifiersPolysorbates, ceteareth, PEG estersWater-dispersible; generally rinse readilyCan form viscous liquid-crystalline phases at intermediate water content
SiliconesDimethicone, cyclomethiconeSpecialized surfactant systems; solventVery low surface energy, spreads into thin tenacious films, water-repellent
Insoluble inorganicsZinc oxide, titanium dioxide, calamineMechanical action; acidic step dissolves ZnOParticulates lodge in gaskets, threads, and surface roughness
Swollen polymersCarbomer, HPMC, HECHydration and dilution; pH/ionic-strength changes collapse carbomer gelsDry to tough films; highly viscous at partial dilution
Preservatives / antioxidantsParabens, phenoxyethanol, BHT, tocopherolUsually minor mass; solubility variesLow mass but may be detectable "ghost" peaks in residue methods
APIUnknown (API-agnostic program)Depends on solubility in the cleaning agentCan be trapped inside a hydrophobic base film and shielded from the cleaning solution

A counterintuitive point: alkaline detergent does not "saponify petrolatum"

People commonly say that hot alkaline detergent works on ointments by saponification. Saponification is base-catalyzed hydrolysis of esters (triglycerides, wax esters, fatty acid esters) into carboxylate soaps and alcohols. You did this mechanism in CHEM202. Petrolatum and mineral oil are saturated hydrocarbons and have no ester bond to hydrolyze. Alkalinity can help with ester-containing fats, but hydrocarbon residues come off through:

  • melting, which lowers viscosity so mechanical action can detach the film;
  • emulsification/solubilization by surfactants (micellar uptake, roll-up of oil droplets from the wetted metal surface);
  • dilution and flushing before re-solidification.

Any claim about a specific formulated detergent's mechanism must be checked against supplier technical data, ideally supported by your own bench cleanability results [verify per detergent].

Melt behavior and the redeposition trap

White petrolatum melts over a broad range, roughly 38–60 °C, rather than at a single point [Rowe, Sheskey & Quinn, 2009] [verify current monograph range]. Two design rules follow:

  1. Clean above the upper end of the base melt range at the surface, not just in the bulk cleaning solution. Jacketed vessels let you heat the wall. Loops, hoses, and filler parts cool quickly.
  2. Watch for redeposition. Molten wax carried downstream in hot cleaning solution re-solidifies wherever the solution cools: an unheated hose section, a valve body, the filler nozzle. A kettle that is clean can end up contaminating its own transfer line. In some places, cleaning moves residue rather than removing it.

Emulsion residues and phase behavior during cleaning

An o/w cream residue meeting hot water does not simply dilute. As water content rises it can pass through phase regions: phase inversion, liquid-crystalline gel phases, or simple breakdown into an oil layer and a water layer. If the emulsion breaks during cleaning, the oil phase coats surfaces as a hydrophobic film that the cleaning solution no longer reaches efficiently. The cleaning agent's job is to keep oily material dispersed (re-emulsified) until it is flushed out. This is a direct application of the emulsion science you studied in PDDM903.

Bench cleanability: ranking before you validate

Because formulations are unknown, the program needs a ranking of base families by cleanability:

  • standardized soiling of stainless coupons with the same finish as product-contact surfaces (fixed mass per area, fixed spreading method);
  • controlled drying and aging to simulate a dirty-hold period (time and temperature recorded);
  • defined cleaning conditions (agent, concentration, temperature, time, agitation or spray);
  • residue measured gravimetrically (good for base components) or by HPLC (for a marker).

The output is a ranked list, for example: anhydrous hydrocarbon ointment > zinc-oxide paste > o/w cream with high fatty alcohol > carbomer gel > lotion. Your data may reorder this, and that is the point of running the study. The ranking feeds worst-case grouping in Concept Five.

A scope boundary: benzene is not a cleaning-limit driver

Benzene contamination linked to carbomers manufactured with benzene, and to benzoyl peroxide degradation, is a real issue for topical products. FDA issued guidance on reformulating products containing carbomers manufactured with benzene (December 2023) [verify title and current status]. This is a raw-material and product-stability control, handled through supplier qualification, incoming testing, and stability (PDDM902, PDQC806, PDQC831). It does not set cleaning limits. Document this explicitly in the cleaning master plan so an auditor or a new hire does not try to add benzene swabbing to the cleaning program.

5. Concept Four: The Equipment Train Is a Map of Hiding Places

Residue and microbes collect where flow is weak, surfaces are rough or hidden, and temperature drops. The PDDM921 reference train (jacketed vacuum emulsifier with anchor/scraper blades, bottom or in-line rotor-stator homogenizer with external loop, vacuum lines, discharge valves, transfer hoses and pumps, filler with heated hopper, piston, and nozzles) needs a systematic walk-down to find them.

Harborage points (each must be in the register)

  • Dead legs: branches where cleaning solution stagnates. The hygienic-design convention of minimizing branch length relative to diameter applies (e.g., ASME BPE guidance) [verify current ratio convention].
  • Gaskets, O-rings, and mechanical seals: elastomer crevices, plus polymer surfaces that absorb oily residue. The homogenizer mechanical seal is a classic hiding place.
  • Threaded fittings: threads are crevices. Hygienic clamp fittings are preferred under 211.42/211.63-style design expectations.
  • Sight glasses and lamp ports: gasketed interfaces, often cooler than the wall.
  • Scraper blades: PTFE or polymer scraper inserts and their mounting hardware.
  • Vacuum lines: aerosolized and foamed product reaches surfaces nobody thinks of as product-contact.
  • Filler piston and nozzle assemblies: small bores, seals, and check valves.

Location scoring

Score each location (e.g., 1–3 or 1–5) on geometry/complexity, accessibility for inspection and sampling, residue accumulation observed or expected, wetting by cleaning agent (spray shadowing, low velocity), temperature (cold spots promote redeposition), and product-contact versus indirect contact. High-scoring locations become mandatory sampling sites. The register is a controlled document under change control, because equipment modifications change it.

Disassembly belongs in the URS

The cheapest time to solve a cleaning problem is before purchase. The user requirement specification should state disassembly-for-cleaning requirements, sanitary fittings, drainability, surface finish, and spray-device coverage. A good URS line reads: "All product-contact parts not cleanable in place shall be removable without tools within N minutes." This connects to 211.42 and to the facility-design thinking of PDDM911.

Shared surface area: the denominator of every limit

Every carry-over limit divides a permissible mass by a surface area, so you need a defensible area for everything a product touches. Use vendor drawings first and geometric estimates second, and document both.

Cylindrical wall (wetted or splash height H, inner diameter D):    A_wall  = π · D · H
Pipe/hose interior (inner diameter d, length L):                   A_pipe  = π · d · L
Flat disc (lid, approximation):                                    A_disc  = π · D² / 4
Dished head: use vendor value; a torispherical head is
somewhat larger than the flat disc of the same D [verify against drawing]

A_shared = Σ A_i  over all product-contact surfaces common to products A and B

Worked example (hypothetical geometry):

Vessel wall:     D = 0.90 m, H = 0.90 m      → π(0.90)(0.90)       = 2.54 m²
Bottom head:     vendor drawing (assumed)                          = 0.70 m²
Lid underside:   vendor drawing (assumed)                          = 0.70 m²
Agitator/scrapers/homogenizer head (vendor)                        = 0.80 m²
Recirculation loop: d = 0.035 m, L = 6 m   → π(0.035)(6)           = 0.66 m²
Transfer hoses:  d = 0.038 m, L = 6 m total → π(0.038)(6)          = 0.72 m²
Filler hopper + piston + nozzles (vendor)                          = 0.38 m²
                                                         ───────────────────
A_shared ≈ 6.5 m² = 65,000 cm²

Note the effect of dedication. If the hoses (0.72 m²) are dedicated per product, they leave the shared area. The limit per unit area then rises, and the hose sampling burden disappears. Dedication improves both risk and arithmetic.

6. Concept Five: Worst-Case Grouping, Marker Selection, and the Intake Gate

Validating every product on every piece of equipment is impossible, and it is especially impossible when the products are unknown because they will arrive as dermatologist specifications. The answer is a grouping matrix. Validate the hardest case, and show that every other case is covered by it.

Grouping criteria

  • Base family: anhydrous hydrocarbon ointment, PEG (water-soluble) ointment, o/w cream, lotion, gel.
  • Cleanability rank from the bench study.
  • API solubility in the cleaning agent (and in the base itself).
  • API concentration: more API per gram of residue means more carry-over per gram.
  • HBEL/toxicity input: lower HBEL means a tighter limit.
  • Detectability: an API with no chromophore at low limits may force LC-MS/MS.

The worst case is often a pair of attributes, not a single product

The hardest-to-clean product may not be the one with the tightest limit. Worst case is a combination: a product that is both hard to clean and has a low limit is the true worst case. If no single product holds both attributes, validation may need two worst-case products, one for cleanability and one for limit stringency. The grouping rationale must say which attribute each validated product covers.

Choosing a marker residue

For semisolids the API is often a small percentage of the residue mass, and the base is most of it. You can track:

  • The API: directly tied to safety (HBEL-based), specific methods available. This is the default marker.
  • A base component (e.g., a fatty alcohol or silicone): shows physical removal of the hardest constituent.

The defensible logic is: "the marker controls the other constituents." For example, if the API is dispersed homogeneously in the base and the base is removed at least as hard as the API, then demonstrating API removal below limit plus visual-clean (which detects base film) covers both. You document why. When the API is far more soluble in the cleaning agent than the base, API removal can outrun base removal. The API comes out clean while a base film remains, so you need visual or base-component evidence as well.

Equipment grouping and bracketing

Identical items (two filler heads of the same model, two identical hoses) can be bracketed if you document the same design, materials, and operating principle. This parallels SUPAC-SS equipment logic [FDA SUPAC-SS, 1997]. Same model number alone is not enough; modifications and gasket materials must match.

The intake gate

Every new dermatologist specification (from PDDM925) is scored against the matrix before the first batch. Three outcomes:

  1. covered by the existing worst case, with the rationale documented;
  2. a new worst case, which triggers a validation update;
  3. dedication review (flags in Concept Nine).

Florida screen. A Florida OTC drug manufacturer permit (F.S. §499.01(2)(n)) does not allow possession of prescription drugs [verify current statute text]. Rx-only actives cannot enter the facility, let alone the shared train. The matrix therefore screens these out at intake, as a legal gate that comes before any scientific scoring.

7. Concept Six: Limits: Retire the Defaults, Derive From Health-Based Exposure

Why the three classic figures are retired

The 7/93 guide reports common industry criteria of 10 ppm, 1/1000 of the normal therapeutic dose, and no visible residue. These trace largely to early industry practice published in the early 1990s (Fourman and Mullen's work is commonly cited) [LeBlanc, 2000]. Their weaknesses:

  • 10 ppm ignores toxicity entirely. The same number applies to a sunscreen filter and to a potent compound.
  • 1/1000 of a therapeutic dose uses pharmacological effect as a proxy for safety. It depends on a "dose" concept that fits topicals poorly, and it requires dosing inputs that this program does not determine.
  • No visible residue is necessary but cannot be the sole criterion. Visual detection thresholds can exceed safe limits for potent residues, and visual inspection cannot see inside hoses at all.

This program accepts none of these as a default justification. Management may impose an internal policy ceiling, but if so it must be documented as policy, not as safety rationale.

HBEL-based MACO

Contemporary practice uses a health-based exposure limit (HBEL, often expressed as a PDE in µg/day) derived by a qualified toxicologist. EMA's guideline on setting HBELs for shared facilities is the best-known regulatory statement, and ISPE's Risk-MaPP and ASTM E3106 present risk-based frameworks [verify current versions]. The structure:

                 HBEL_A  ×  MBS_B
     MACO   =   ──────────────────
                     MDD_B

  MACO   = maximum allowable carry-over of residue A into the next batch (µg)
  HBEL_A = health-based exposure limit of previous product's residue (µg/day)
           ── SUPPLIED AND SIGNED by qualified toxicologist or sponsor
  MBS_B  = minimum batch size of any product that may follow on this train (g)
  MDD_B  = maximum daily amount of next product, as a toxicology input (g/day)
           ── SUPPLIED AND SIGNED by toxicologist/sponsor; the QC platform
              does not determine it. Monograph labeling directions are recorded
              as LABELING INPUTS ONLY.

Then convert to a surface concentration:

                  MACO
     SAL   =   ───────────           (µg/cm²)
                A_shared

Per-swab limit (swab area A_swab, extraction volume V_ext):

     L_swab   = SAL × A_swab                       (µg per swab)
     C_swab   = L_swab / V_ext                     (µg/mL in extract)
     C_report = C_swab × R                         (recovery-adjusted target the
                                                    instrument must see; R = fractional recovery)
       -- or, equivalently, divide the measured result by R and compare to C_swab.
          Pick ONE convention per SOP; never apply R twice.

Per-rinse limit (rinse volume V_rinse contacting area A_rinsed):

     C_rinse  = (SAL × A_rinsed × R_rinse) / V_rinse   (µg/mL)

Input governance: the worksheet is a controlled record

  • The calculation runs in a controlled, validated worksheet (PDQC814 CSV logic: locked cells, version control, audit trail) with second-person verification.
  • Toxicology fields (HBEL, MDD) are populated only from a signed toxicologist or sponsor document, referenced by document number.
  • Missing input means escalation, not estimation. If no HBEL exists, the product does not enter the shared train until one is supplied or a dedication decision is made.
  • The toxicologist should address topical-specific issues: cutaneous route of the next product, sensitization potential, and, where elemental residues apply, ICH Q3D(R2) cutaneous PDEs (Appendix 5) as a model of route-specific thinking [verify Appendix numbering].

Worked arithmetic (all inputs hypothetical, chosen only to exercise the formula)

Scenario 1 (moderate HBEL placeholder):
  HBEL_A  = 100 µg/day        (hypothetical placeholder, not a real value)
  MBS_B   = 200 kg = 200,000 g
  MDD_B   = 10 g/day          (hypothetical placeholder, not a usage value)
  A_shared = 65,000 cm²  (from Concept Four)

  MACO   = 100 × 200,000 / 10            = 2,000,000 µg  (2.0 g)
  SAL    = 2,000,000 / 65,000            ≈ 30.8 µg/cm²
  L_swab = 30.8 × 25 cm²                 ≈ 769 µg/swab
  C_swab = 769 / 10 mL                   ≈ 76.9 µg/mL
  With R = 0.70: instrument target       ≈ 53.8 µg/mL   → trivially measurable

Scenario 2 (low HBEL placeholder, 100× lower):
  HBEL_A = 1 µg/day           (hypothetical)
  SAL    ≈ 0.31 µg/cm²;  L_swab ≈ 7.7 µg;  C_swab ≈ 0.77 µg/mL
  With R = 0.70: instrument target       ≈ 0.54 µg/mL
  → For a weak chromophore, this may approach HPLC-DAD LOQ → LC-MS/MS candidate

A key insight: when the visual threshold governs

In Scenario 1 the calculated limit (~31 µg/cm²) is looser than the commonly cited visual-detection threshold for many residues on stainless steel, roughly a few µg/cm² under good lighting [LeBlanc, 2000] [verify; residue- and condition-dependent]. For many low-potency OTC topicals, the binding criterion is "visually clean," not the toxicology-derived number. The rule is to apply every criterion and let the most stringent govern at each location. The visual criterion governs exposed surfaces. The calculated limit governs hoses and loops, where nobody can see and where only rinse or swab chemistry speaks.

Measurability before approval

A limit you cannot measure is not a limit. Before protocol approval, confirm the method LOQ is comfortably below the recovery-adjusted target. Many programs require LOQ ≤ some fraction (e.g., 50 %) of the limit to leave room for trending. That fraction is a program decision that must be justified [verify industry convention].

8. Concept Seven: Visual, Detergent, and Microbial Criteria

Visual-clean: necessary, not sufficient, and only as good as its conditions

"Visually clean" means nothing until you define the conditions:

  • minimum illuminance at the surface (a lux value with a flashlight specification);
  • viewing distance and angles;
  • surface state (dry; residues can be invisible on wet steel);
  • inspector qualification, e.g., challenge with coupons spiked at known levels.

Some programs establish formal visual residue limits by spiking coupons and finding the lowest level reliably seen [verify approach]. White petrolatum films on polished stainless can be strikingly hard to see. Oblique lighting showing a sheen or water-break pattern (water beading on a hydrophobic film) is a useful aid.

Detergent residues

The cleaning agent itself is a contaminant. Its limit comes from supplier toxicology data processed into an acceptable carry-over by the same MACO logic. Measurement options:

  • Specific: HPLC of a detergent marker component (often a surfactant).
  • Nonspecific: TOC (USP <643>) or conductivity. These are justified only if all measured signal is conservatively attributed to the detergent.

Peptide Partners owns no TOC analyzer. That is a capability exit (Concept Ten): either contract TOC testing or develop a specific HPLC method for the detergent marker.

Microbial cleanliness: the semisolid-specific dimension

For water-containing products (creams, lotions, gels), residue plus moisture plus time produces microbial growth. Equipment becomes a source of objectionable organisms, which brings in 21 CFR 211.113(a). Burkholderia cepacia complex (Bcc) is the named risk. It is notorious in water-based nonsterile products, resistant to many preservatives, capable of surviving in Purified Water systems, and associated with recalls and outbreaks [Tavares et al., 2020]. FDA issued a Bcc advisory (2021) and the draft guidance on microbiological quality in nonsterile manufacturing (September 2021). USP <60> provides the Bcc test for products [verify all current status].

Important points:

  • There are no compendial surface bioburden limits. Surface targets (e.g., cfu per contact plate or swab area) are set by risk assessment and justified in the master plan. USP <1111> is a product acceptance criterion, not an equipment surface criterion.
  • Absence of objectionable organisms, with Bcc explicitly named, matters more than any count.
  • The final rinse is Purified Water from the qualified system. If the water system is contaminated, the rinse inoculates the clean equipment. This is the Strukmyer pattern.
  • Drying and protection after cleaning are microbial controls, not cosmetic steps.

Lab reality: BacT/ALERT gives presence/absence and time-to-detection, not cfu. It cannot perform USP <61> enumeration or produce surface counts. Surface bioburden counts and organism identification for cleaning validation must go to a contract micro lab until a plate-count capability is justified.

9. Concept Eight: Sampling and Recovery: The Most Underestimated Link

Swab versus rinse

SwabRinse
Best forInsoluble/waxy residues; hard-to-clean spots; defined small areasSoluble residues; inaccessible interiors (loops, hoses, pumps)
StrengthPhysical removal; site-specificIntegrates entire internal surface
WeaknessSamples a tiny fraction of area; operator-dependentOnly removes what dissolves/disperses in the rinse; dilution lowers concentration
Semisolid cautionSolvent must dissolve hydrocarbon matrixAqueous rinse nearly useless for petrolatum; heated or solvent/surfactant rinse may be needed

The 7/93 guide recognizes both, and notes that rinse sampling suits sealed or inaccessible systems while direct surface sampling is preferred where possible [FDA, 7/93].

Swab technique written for reproducibility

The SOP must fix: swab material (e.g., low-extractable polyester knit), wetting solvent and volume, template area (commonly 25 cm² or 100 cm²), stroke pattern (e.g., horizontal strokes, flip, vertical strokes, perpendicular passes), swab head flipping, extraction volume, and extraction time and agitation. Any element left to operator judgment becomes a source of variance that recovery studies will expose.

Recovery: why it is a fraction, and what to do about it

                 mass measured from spiked coupon
     R  =   ─────────────────────────────────────────
                 mass spiked onto coupon

  Corrected surface result  =  measured result / R
  (or limit × R, as shown in Concept Six; pick one convention)

Interference first. Before spiking anything, run blank swab extracts, solvent blanks, and template-only controls through the residue method to detect swab-adhesive or extractable peaks [FDA, 7/93]. Hydrocarbon-friendly solvents are aggressive and pull more extractables from swab handles and adhesives than water does. Find the ghosts before they show up as false positives in validation runs.

Petrolatum is the worst-case recovery matrix for several reasons:

  • The API sits inside a hydrophobic film. A wetting solvent that dissolves the API but not the petrolatum leaves API locked in the film.
  • Alkanes (e.g., heptane) dissolve petrolatum but are immiscible with reversed-phase mobile phases and are flammable. Alcohols (e.g., isopropanol) are HPLC-friendly but dissolve petrolatum poorly. Mixtures or two-solvent extraction schemes are common compromises [verify for each API].
  • Solvent safety (flammability, exposure) ties back to PDDM911 process safety.
  • Swab integrity: some solvents swell or degrade swab materials.

Coupon study design:

  • stainless coupons matching product-contact finish (Ra);
  • at least three spike levels bracketing the limit, e.g., 50 %, 100 %, 150 % of the swab limit;
  • residue applied in the worst-case matrix (API in petrolatum, not API in methanol);
  • drying or aging to simulate dirty hold;
  • replicate operators to capture between-operator variance;
  • additional materials where sampled: gaskets, polymer hoses, filler piston seals.

Acceptance and use. Typical industry targets are often cited around ≥ 50–70 % mean recovery with RSD ≤ 15–20 % [verify; justify in protocol]. The most important rule: low recovery is a reason to improve the technique, not a reason to loosen the limit. A 20 % recovery means your measurement is mostly noise about what is really on the steel. A big correction factor applied to a tiny signal amplifies error.

Sequencing chemical and microbial sampling

Solvent-wetted chemical swabs are not sterile and can kill organisms. Contact plates leave agar residue that interferes with chemical analysis. The protocol must sequence and position samples so they do not compromise each other. A common approach is paired adjacent sites, with microbial sampling performed first under aseptic technique and agar residue wiped away afterward [verify program-specific design].

10. Concept Nine: Flags, Dedication, Campaigns, and Hold Times

Documented flag assessment

Every product receives a written flag assessment, signed by the QU, with three possible outcomes: shared, campaign, or dedicated.

  • Potency/OEL: a very low HBEL may make shared limits unmeasurable or unachievable.
  • Oxidizer (BPO): residual benzoyl peroxide can oxidize the next product's API or consume its antioxidants. Peroxide may also react with certain cleaning agents. This calls for a documented compatibility assessment, not automatic dedication.
  • Sensitizer: carry-over of a contact allergen into a product that does not declare it is a quality and labeling risk. These are inputs to the dedication assessment.
  • Microbial: water-containing, preservative-light, or Bcc-susceptible formulas.
  • Cleanability rank and detectability.

Partial dedication: the high-value, low-cost control

Dedicating hoses, gaskets, filler nozzles and pistons, and sampling tools costs little and removes the hardest-to-clean, least-inspectable parts from the shared surface area. Full train dedication costs capacity and capital. The matrix of shared versus dedicated equipment is a living controlled document linked to the product register.

Campaigns

A campaign runs several batches of the same product with minor between-batch cleaning and a full changeover cleaning at the end. Both need evidence. You must show that residue accumulation and microbial build-up stay controlled up to the maximum campaign length or batch count, and that the full clean works on the end-of-campaign worst case (thickest, most aged soil).

Hold times: two clocks

  • Dirty-hold time (end of manufacture to start of cleaning): waxes harden and crystallize, water evaporates leaving concentrated films, and polymers dry to tough skins. Cleanability falls over time.
  • Clean-hold time (end of cleaning to next use): microbial growth on surfaces that are wet or poorly protected, and recontamination from the environment.

Both are studied at their maximum claimed duration, and cleaned status expires at the end of the clean-hold. Equipment past expiry is re-cleaned before use, as 211.67's inspection-before-use requirement implies.

11. Concept Ten: Methods, the Real Lab, and Capability Exits

What Peptide Partners can do in-house

  • HPLC-DAD (3× Agilent 1260): specific residue methods. Specificity must be shown against detergent, swab extractables, and base components. DAD peak purity supports specificity, and LOQ must sit below the recovery-corrected limit.
  • Triple-quad LC-MS/MS: APIs with weak or no UV chromophore, or very low HBEL-driven limits. Watch for matrix effects: lipids and surfactants co-extracted from swabs cause ion suppression. Use matrix-matched calibration or stable-isotope internal standards where available.
  • Fit-for-purpose validation scaled from PDQC802: accuracy (via recovery), precision, linearity bracketing the limit, LOQ, and solution and swab-extract stability over the sampling-to-analysis interval.

Capability exits (documented in a memo)

  • TOC: no analyzer → contract lab (USP <643>).
  • Elemental residues (e.g., zinc from ZnO): no ICP → contract lab.
  • Volatile cleaning solvents: no GC-HS → contract lab. Identify what the second MS actually is before buying a new instrument.
  • Surface bioburden counts: no plate-count lab, and BacT/ALERT gives time-to-detection, not cfu → contract micro lab.

Each exit specifies contract-lab requirements: method suitability, data-integrity expectations (ALCOA+, audit trails, raw-data access), turnaround times compatible with hold times, and quality agreements. Purchase triggers (residue workload volume, turnaround risk, cost per test) become evidence-based inputs to PDDM999.

12. Concept Eleven: Process Validation, Failures, and Lifecycle

Protocol and execution

  • Protocol content: scope and worst-case rationale, cleaning parameters with ranges (time, action, chemistry, temperature, the "TACT" or Sinner's-circle framework [verify]), sampling map, limits with derivation, hold times, number of runs with justification (three consecutive runs is traditional; current expectations favor risk-based justification [verify]), and acceptance criteria.
  • Execution under normal operating conditions with routine operators. "Hero" operators prove nothing about the SOP.
  • Cleaning parameters recorded and linked to batch records (211.188).
  • Report approved by the QU (211.22), with conditions stated for releasing shared equipment.

Failures: no resampling into compliance

A failing result triggers a 211.192 investigation. The investigation must distinguish a sampling or method error (contaminated swab, co-elution, calculation error, documented with evidence such as blank results or system-suitability failure) from a cleaning process failure. Resampling an adjacent spot until a pass appears is the swab version of "test until clean." An invalidated result needs an assignable, evidenced cause. Otherwise the failure stands, the procedure is improved, and validation runs are repeated.

Lifecycle

  • Routine verification and trending, with alert levels below limits and reassessment when results drift upward.
  • Change-control triggers: new product intake (PDDM925), new worst case, detergent or supplier change, equipment modification, scale change (SUPAC-SS equipment-principle logic), and repeated failures.
  • Packaging-line clearance (211.130) is distinct from product-contact cleaning validation. Filler product-contact parts, however, are inside cleaning validation scope.

13. Connections to Peptide Chemistry

You arrive from the PDQC peptide-QC postdoc, so here is how that knowledge transfers and where it misleads you.

  • Solubility inverts. Peptide residues are usually water-soluble, and a water rinse plus TOC was often a reasonable nonspecific approach. In semisolids, water-insoluble bases make aqueous rinses and nonspecific water-quality tests nearly blind. Specificity becomes much more important.
  • The recovery problem changes. For peptides, recovery loss came from adsorption of tiny amounts onto steel and plastic. For petrolatum, the loss comes from dissolution failure: the API stays entrapped in a film the swab solvent cannot dissolve. The fix is different (surface-passivating additives and low-binding materials versus solvent selection and multi-step extraction).
  • Mass per area jumps by orders of magnitude. Peptide residues are microgram-scale. Semisolid wall films are gram-scale before cleaning. The cleaning process must remove far more material for the same final limit.
  • Degradation-based arguments rarely apply. In biologics and peptide cleaning, harsh alkaline cleaning can degrade or denature the active, and this is sometimes considered in the risk assessment. Most topical small-molecule APIs survive detergent cleaning intact, so assume the residue is the active compound.
  • LC-MS/MS skills transfer directly, but the matrix changes. Instead of salts and peptide fragments, expect lipids, waxes, and surfactants causing ion suppression and source contamination.
  • Validation logic transfers fully. PDQC802's lifecycle thinking (fitness for purpose, LOQ relative to the decision limit, solution stability) applies unchanged. Only the matrix is new.
  • Where it could lead. If a dermatologist specification someday includes a peptide as a cosmetic ingredient in a drug-cosmetic product, your peptide analytical skills apply directly to residue detection. The product classification still governs whether it can enter this train (PDDM901).

14. Common Misconceptions

  1. "211.67 explicitly requires cleaning validation." It requires written, effective cleaning procedures and records. The validation expectation comes from guidance and enforcement.
  2. "If the final rinse meets Purified Water specs, the equipment is clean." Water quality does not measure residue. Petrolatum barely partitions into water.
  3. "10 ppm is the FDA limit." FDA reported it as an industry practice. This program does not accept it as a default justification.
  4. "Alkaline detergent saponifies petrolatum." Hydrocarbons have no ester bonds. Removal is by melting, emulsification, and flushing.
  5. "Visually clean is enough for low-potency products." Visual inspection may govern exposed surfaces, but it says nothing about hoses and loops.
  6. "Low recovery can be corrected by a large factor." Fix the technique instead. Correction amplifies noise.
  7. "Hot water at 60 °C in the kettle means the whole train is above the melt range." Hoses, valves, and filler parts cool, and wax redeposits there.
  8. "BacT/ALERT negative means surface bioburden is acceptable." It gives presence/absence and time-to-detection, not counts or surface limits.
  9. "Benzene must be included in cleaning limits." Benzene is a raw-material and product issue, outside cleaning-limit scope.
  10. "A failed swab can be retaken nearby." Without an evidenced assignable cause, that is testing into compliance and violates 211.192 expectations.
  11. "QC can estimate the next product's daily amount from labeling." Labeling directions are recorded only as labeling inputs. Toxicology inputs come signed from qualified experts, and missing inputs trigger escalation.

⚙️ Mechanisms & Applications

GPT-6 Astra

Mechanisms & Applications: Cleaning as a Controlled Separation Process

Skipper’s central insight: cleaning a semisolid manufacturing train is not simply “washing away the previous API.” It is a controlled separation process applied to a heterogeneous, aging film distributed across an imperfectly accessible surface. Success requires chemical removal, physical removal, microbiological control, and evidence that the measurements actually represent what remains.

This builds directly on protein purification, analytical recovery, transport phenomena, and process validation. The major change from relatively water-soluble peptide residues is that a semisolid can leave several chemically distinct residues at once: an oily film, crystalline wax, suspended inorganic particles, swollen polymer, preservative, and API. Their removal mechanisms—and their analytical visibility—may be different. Peptide residues were not universally easy either: adsorption, aggregation, and drying could already defeat a simple aqueous rinse.

The evidence chain:

Known soil and equipment
    → designed cleaning mechanism
    → controlled cleaning parameters
    → representative sampling
    → demonstrated analytical recovery
    → justified residue limits
    → reproducible process performance
    → lifecycle verification

A failure in one link cannot be repaired by confidence in another.

The regulatory anchor is 21 CFR 211.67: suitable cleaning and maintenance, written procedures, assigned responsibilities, schedules, protection of clean equipment, and inspection before use. The FDA’s 1993 cleaning inspection guide and 1994 topical-products inspection guide explain recurring inspection concerns, but they are not substitutes for the regulation or for current product-specific justification. Their continuing practical lesson is that a clean-looking kettle and acceptable rinse-water conductivity do not establish control of retained product residue. [1–3]

1. How a Semisolid Film Actually Leaves a Surface

1.1 First identify the physical state of the soil

A fresh cream film is not the same material as that cream after a weekend inside a warm-to-cool transfer hose. Water may evaporate, droplets may coalesce, polymers may concentrate, waxes may crystallize, and particles may become embedded in a tenacious organic layer. Consequently, dirty-hold time changes the material being cleaned; it is not merely a scheduling variable.

From protein folding and soft-matter thermodynamics, Skipper already knows that history matters. Two samples with the same composition can have different structures and different kinetic barriers. The same principle explains why fresh product on a polished coupon may be easy to remove while aged residue beneath a gasket is not.

  1. Bulk removal: drain or remove recoverable product using the approved procedure. Reducing the initial mass prevents a cleaning bath from becoming a concentrated product emulsion.
  2. Conditioning: use an appropriate temperature and chemistry to soften, swell, wet, or dissolve the residue.
  3. Detachment: apply fluid shear, spray impingement, brushing, or other controlled mechanical action.
  4. Transport: keep detached material dispersed or dissolved long enough to carry it out of the equipment.
  5. Rinsing: remove cleaning chemicals and remaining dispersed soil.
  6. Drying and protection: prevent the cleaned equipment from becoming a wet microbial habitat or being recontaminated before use.

The easily missed step is transport. A detergent may detach wax successfully, yet the process still fails if that wax congeals in a colder hose downstream. Cleaning the kettle while relocating the soil to the filler is not cleaning the train.

1.2 Different residue classes require different mechanisms

  • Petrolatum and mineral oil: predominantly hydrophobic hydrocarbons. Water alone has poor access to the film; suitable heat, surfactant-assisted dispersion, mechanical action, or a justified solvent step may be needed.
  • Waxes and fatty alcohols: heating can soften or melt the soil, but cold surfaces can cause redeposition. Fatty alcohols do not become soaps merely because alkali is present.
  • Ester-containing fats: sufficiently strong alkaline conditions can hydrolyze ester bonds. Whether this mechanism is important during an actual cleaning cycle depends on formulation, temperature, concentration, and contact time.
  • Dimethicone: a hydrophobic silicone residue, not a triglyceride. Detergent performance must be demonstrated rather than inferred from an alkaline “degreaser” label.
  • Zinc oxide, titanium dioxide, and similar particles: may require detachment and dispersion rather than dissolution. A clear rinse can coexist with retained particles in a seal recess.
  • Carbomer and cellulose ethers: hydration, pH, ionic strength, and drying history can alter swelling and adhesion. A poorly designed initial wetting step can create a hydrated gel barrier rather than rapid removal.
Alkaline hydrolysis of an ester:

R–C(=O)–OR′ + OH− → R–COO− + R′–OH

Critical distinction:
Ester-containing soil may undergo saponification.
Hydrocarbon petrolatum has no ester bond to saponify.

For hydrocarbon soil, removal generally depends on:
wetting + softening/dissolution + dispersion + mechanical transport.

The Handbook of Pharmaceutical Excipients is useful for assembling hypotheses about solubility, melting behavior, incompatibilities, and material identity. It does not prove that a particular formulated detergent removes a particular aged residue. That proof comes from supplier-supported mechanism information and controlled experiments. [4–6]

1.3 Temperature helps—but not without limits

For many oils and softened bases, increasing temperature lowers viscosity and improves transport. A simplified mass-transfer expression makes the logic explicit:

Dissolution-controlled removal:

−dmsoil/dt ≈ kL A (Cs − Cb)

kL = liquid-side mass-transfer coefficient
A  = accessible soil–liquid interfacial area
Cs = concentration at the soil interface
Cb = concentration in the bulk cleaning liquid

Agitation can increase kL; suitable chemistry can increase effective solubilization; solution replacement can keep Cb low. But this model is incomplete for insoluble particles or cohesive films, where detachment rather than dissolution controls removal.

Practical consequence: measure temperature at relevant equipment locations, not only at the cleaning-solution supply. A satisfactory kettle temperature does not establish that a long hose or filler valve stayed above the required softening range. Any temperature window must also respect seal compatibility, chemical stability, operator safety, and oxidizer hazards.

For a peroxide-containing hypothetical product, heating or introducing organic solvents is not an automatic cleaning improvement. Compatibility and process-safety review come first. Benzene concerns associated with certain carbomers or benzoyl peroxide products belong primarily to raw-material, product-quality, and degradation control; they should not be imported uncritically as generic cleaning-limit assumptions. FDA’s December 2023 carbomer guidance is a specific reference for the carbomer issue. [7]

2. From Equipment Geometry to a Defensible Sampling Map

A useful analogy is an enzyme active site: the overall protein surface tells us little about the chemistry inside a buried pocket. Likewise, the broad stainless wall of a kettle can be the least informative place to sample.

For the PDDM921 reference train, the controlled hard-to-clean register should follow the complete route:

Emulsifier vessel and lid
    → anchor/scraper assemblies
    → bottom homogenizer or external rotor–stator loop
    → discharge valve
    → transfer pump and hoses
    → heated filler hopper
    → piston, seals, and nozzles

Also assess:
vacuum connections, condensate paths, splash surfaces,
and indirect-contact locations capable of returning contamination.

Each location needs a reasoned assessment of geometry, accessibility, residue accumulation, cleaning-agent contact, drainage, materials of construction, and potential contribution to the next product. A numerical score can prioritize work, but must not let an inaccessible seal disappear beneath an acceptable average score.

  • Scraper backs and blade attachments: product can accumulate behind the visibly exposed surface.
  • Mechanical seals and O-ring grooves: retention and microbial harborage may persist despite vigorous flow elsewhere.
  • Hose low points: retained liquid, cooling, and incomplete drainage can combine.
  • Valve seats and filler pistons: moving interfaces expose surfaces differently during production and cleaning.
  • Vacuum lines: classify according to actual product entrainment, condensation, and return pathways rather than assuming they are irrelevant.

Design application: disassembly requirements belong in the equipment user requirements specification, before purchase. The current Peptide Partners laboratory owns no semisolid manufacturing train; these exercises are design and partner-qualification work until the actual equipment is available. A reference drawing is not an as-built qualification record.

Surface-area accounting

Straight cylindrical hose, internal wetted surface:

Ahose = πDL

Example:
D = 2.5 cm
L = 300 cm

Ahose = π × 2.5 × 300 ≈ 2,356 cm²

Add actual wetted areas of:
couplings, valves, pump components, filler parts, and other shared surfaces.

Use internal dimensions and documented equipment drawings or measurements. Complex surfaces may require vendor data or justified engineering estimates. Product-specific shared pathways matter: a product that bypasses the homogenizer may have a different carry-over pathway from one that uses the entire train.

A surprisingly dangerous error is to overstate shared area. For a fixed allowable carry-over mass, a larger denominator produces a smaller local limit, which is conservative locally—but it can also obscure which surfaces actually transfer residue or encourage inappropriate area allocation. Both the geometry and the transfer pathway must be correct.

3. Designing Cleaning Experiments Instead of Guessing a Cycle

3.1 Coupon studies: isolate the causes of poor cleaning

Bench cleanability work should compare at least the five base families relevant to this program: anhydrous ointment, PEG ointment, oil-in-water cream, lotion, and gel. The purpose is not to declare one universally worst; it is to discover which combinations of composition, aging, surface, and cleaning conditions create the greatest removal challenge.

  1. Use coupons with representative material, finish, and surface condition.
  2. Apply a controlled product mass over a measured area; report loading in mg/cm².
  3. Standardize spreading, drying, aging, temperature history, and dirty-hold duration.
  4. Apply defined cleaning-agent concentration, contact time, temperature, and mechanical action.
  5. Measure residual API and, where needed, a base or excipient marker.
  6. Include unsoiled blanks and soiled uncleaned controls.
  7. Repeat sufficiently to distinguish variability from a genuine formulation effect.
Initial soil loading:

σ0 = minitial / Acoupon

Fraction removed:

Removal (%) = 100 × (minitial − mremaining) / minitial

Why percentage removal can mislead: removing 99.9% of a 10 g deposit still leaves 10 mg. A process must meet the justified absolute residue limit, not merely achieve an impressive removal percentage.

Gravimetry can be useful at higher residue loads, but retained water, detergent film, and balance capability can confound it. HPLC can sensitively measure an API while missing a substantial API-depleted hydrocarbon film. Neither measurement alone automatically describes the complete soil.

The topical-development papers by Chang and colleagues and the classification work by Buhse and colleagues help explain why formulation microstructure and process history matter. They are supporting sources for understanding the product—not sources of cleaning acceptance limits. [8–10]

3.2 Coverage testing is not cleaning validation

A fluorescent-tracer coverage study can reveal spray shadows and inadequately contacted locations. The tracer is applied and inspected using a predefined procedure, with controlled observation conditions. Method details and interpretation must be verified for the chosen system.

Its evidentiary limit is essential: removing a water-soluble tracer does not demonstrate removal of aged petrolatum. Coverage establishes that the cleaning liquid can reach a surface under the test conditions; soil-removal studies establish whether the actual cleaning mechanism works there.

3.3 Convert the mechanism into an executable SOP

“Wash thoroughly with hot detergent” is not a reproducible instruction. A controlled procedure defines disassembly, bulk-removal method, detergent preparation, concentration checks where appropriate, minimum contact time, temperature window, rinse volumes or validated endpoints, mechanical action, drainage, drying, and inspection conditions.

The familiar time–action–chemistry–temperature framework is a useful design aid. It does not mean these variables are freely interchangeable. Doubling contact time cannot compensate for a spray shadow; increasing alkali cannot saponify mineral oil; increasing temperature cannot make an incompatible peroxide/solvent combination acceptable.

4. Turning Toxicological Inputs into Measurable Surface Limits

4.1 Keep hazard assessment separate from arithmetic

A health-based exposure limit is an expert toxicological input, not something inferred by QC from an OTC label. For the next product’s cutaneous route, the assessment may need to address systemic toxicity, local effects, sensitization, absorption assumptions, and vulnerable-use considerations. An HBEL-based limit may still need additional quality or reactivity constraints.

ICH Q3D(R2), Appendix 5, provides route-specific thinking for elemental impurities; its cutaneous PDEs are not general-purpose PDEs for organic APIs or excipients. The applicable current HBEL framework and official text must be checked before protocol approval. [11]

Basic carry-over model, from previous product A into next product B:

MACO_A→B = HBEL_A × Bmin / Umax,B

HBEL_A = qualified allowable daily exposure to residue A
Bmin   = minimum relevant next-product batch mass
Umax,B = approved maximum daily amount of next product used
         solely as a controlled exposure-model input

Dimensional check:

(mg A/day) × (kg B/batch) / (kg B/day)
    = mg A/batch

The next-product exposure input is supplied through the approved assessment; this worksheet does not establish human dosing or clinical-use instructions. Missing or unsigned inputs trigger escalation, not substitution with a convenient assumption.

This expression assumes a defensible relationship between transferred residue, next-batch distribution, and exposure. Localized contamination of early filled units can violate a simple homogeneous-batch model. Mixing and transfer behavior therefore matter alongside toxicology.

4.2 Worked example: from total mass to HPLC concentration

Illustrative numbers only—not recommended limits or product-use values. Suppose the controlled worksheet receives:

  • HBEL for residue A: 0.002 mg/day.
  • Minimum next batch: 100 kg.
  • Approved exposure-model input for next product: 0.020 kg/day.
  • Total relevant shared surface: 200,000 cm².
  • Swabbed area: 25 cm².
  • Extraction volume: 10.0 mL.
  • Validated combined swab-and-extraction recovery: 0.70.
1. Total allowable carry-over:

MACO = 0.002 × 100 / 0.020
     = 10 mg
     = 10,000 µg

2. Nominal surface limit:

Lsurface = 10,000 / 200,000
         = 0.050 µg/cm²

3. True residue allowance on a 25 cm² swab area:

Lswab = 0.050 × 25
      = 1.25 µg

4. Expected measured mass at 70% recovery:

Lmeasured = 1.25 × 0.70
          = 0.875 µg

5. Acceptance concentration in a 10.0 mL extract:

Caccept = 0.875 / 10.0
        = 0.0875 µg/mL

There are two equivalent reporting conventions:

Convention A: correct the result

mcorrected = Cmeasured × Vextract / R

Accept if:
mcorrected ≤ Lswab


Convention B: lower the measured-result threshold

Accept if:
Cmeasured ≤ Lswab × R / Vextract

Do not apply recovery correction twice.

If the method LOQ is 0.020 µg/mL, its recovery-corrected mass equivalent is:

LOQsurface-equivalent
    = 0.020 × 10.0 / 0.70
    = 0.286 µg per swab
    = 0.0114 µg/cm²

This is below the 0.050 µg/cm² example limit. But “below” is not the entire capability assessment: uncertainty, precision near the limit, matrix interference, and extract stability still matter. A result below LOQ is not zero residue.

Counterintuitive point: health-based MACO is an upper safety boundary, not a target amount to leave behind. Established process capability, local quality risks, visible cleanliness, and chemical incompatibility may justify tighter operational limits.

4.3 Rinse limits require a different recovery argument

For a defined, sampled equipment section:

Caccept,rinse = Lsurface × Asection × Rrinse / Vrinse

Rrinse must represent the demonstrated effectiveness
of the specified rinse-sampling procedure.

Do not borrow the swab recovery factor for a rinse. Rinse recovery depends on wetting, solubility, contact time, circulation, drainage, and residue aging. A dilute rinse from a large volume may provide weak information about a small, heavily contaminated pocket.

Also, a component rinse cannot be assigned the entire train’s MACO without justification. Individual allocation must remain consistent with total carry-over control. Passing a rinse and passing selected swabs does not prove that every unobserved surface is clean; the sampling strategy depends on validated process understanding.

5. Swab Recovery: Validating the Bridge Between Surface and Instrument

From analytical chemistry, Skipper knows that calibration tells us how the instrument responds to analyte in a vial. Cleaning validation asks an additional question: what fraction of the analyte on the equipment ever reaches that vial?

5.1 Establish interference before estimating recovery

  1. Analyze extraction-solvent blanks.
  2. Analyze unused-swab extracts, including relevant swab lots.
  3. Analyze clean-coupon swabs.
  4. Check templates, adhesives, containers, detergent, and base-matrix extracts where relevant.
  5. Only then interpret spiked-surface recovery.

A swab extractable that coelutes with the API can produce apparently excellent recovery. The error is especially dangerous because it makes an inadequate sampling technique look successful. FDA’s 1993 cleaning guide explicitly emphasizes sampling interference and recovery considerations. [2]

5.2 A petrolatum challenge must resemble petrolatum residue

Spike representative residue onto the coupon in the relevant matrix, rather than applying only a clean analyte solution. Age it under the proposed worst-case dirty-hold conditions. Otherwise, the experiment may validate recovery of a readily accessible analyte crystal while the real analyte remains trapped beneath a hydrocarbon film.

Screen wetting and extraction solvents for residue accessibility, analyte solubility, swab integrity, material compatibility, safety, and chromatographic compatibility. An alkane may dissolve the hydrocarbon matrix but fail to recover a polar API. A strong extraction solvent may require validated evaporation/reconstitution or dilution before HPLC injection, adding further potential losses.

Overall recovery:

Roverall = mmeasured / mapplied

Conceptually:

Roverall ≈ Rpickup × Rrelease × Rhandling

Routine correction should match the experimentally validated
combined sampling-and-extraction procedure.

A practical developmental design uses three levels bracketing the acceptance limit—for example, 50%, 100%, and 150%—with multiple operators and replicates. These are study-design choices, not universal compendial requirements.

Synthetic teaching dataset, not an executed laboratory report: six coupons per level, divided across two operators, could produce the following recoveries:

50% level:   66, 68, 70, 70, 72, 74%
100% level:  65, 68, 69, 71, 73, 74%
150% level:  67, 68, 69, 71, 72, 73%

Each level mean = 70%
Within-level RSD ≈ 3.4–4.8%

A mean recovery factor of 0.70 is usable only if predefined criteria are met and concentration, operator, material, and aging effects are acceptably controlled. The protocol must state whether correction uses a mean, a justified conservative estimate, or another approach. Recovery uncertainty belongs in the decision process.

Low recovery is not permission to accept more contamination. It amplifies uncertainty. A technique with inconsistent or inadequate recovery should be redesigned. Stainless-steel recovery cannot automatically qualify elastomer seals or polymer hoses, where absorption and solvent compatibility differ.

6. Choosing an Analytical Marker Without Becoming Blind to the Base

The marker should represent the risk and the cleaning challenge. The lowest-HBEL API may control the toxicological limit, while a wax or silicone controls cleanability. There may therefore be no single universal “worst case.”

  • HPLC-DAD on the owned Agilent 1260 systems: appropriate when a residue has adequate UV response and can be separated from detergent, base components, and swab extractables.
  • Triple-quadrupole LC-MS/MS: potentially useful for low limits or weak UV response, provided the analyte ionizes adequately and matrix effects are controlled.
  • Orthogonal or nonspecific testing: may be needed to assess base or cleaning-agent residue that an API assay does not represent.

DAD peak-purity assessment is supporting evidence, not proof that coelution is absent. Similar spectra, low-level interferents, and nonchromophoric material can defeat that inference.

One LC-MS/MS matrix-effect assessment:

Matrix factor =
response in post-extraction-spiked blank matrix
/
response in neat solution at the same concentration

Factor < 1: ion suppression
Factor > 1: ion enhancement

Use matrix-effect and extraction-recovery experiments to separate different causes of poor signal. Stable-isotope internal standards can improve quantitation when suitable, but adding one after swabbing does not correct analyte that was never collected from the surface.

Fit-for-purpose validation should address specificity, working range near the limit, precision, accuracy/recovery, LOQ, carry-over, and stability over the actual sampling-to-analysis interval. Method and recovery reports must be approved before the cleaning process-validation study relies on them.

Capability exits at Peptide Partners

  • TOC: no analyzer is owned. Outsource if selected. USP <643> does not by itself validate a cleaning-residue application; matrix recovery, carbon response, backgrounds, and the acceptance rationale remain necessary.
  • Elemental residues: outsource appropriately qualified elemental testing; no ICP capability is owned.
  • Volatile cleaning solvents: outsource a suitable validated method; no GC/GC-HS is owned. Identify the second MS before assuming it provides any relevant capability.
  • Surface bioburden: outsource quantitative culture work until an appropriate microbiology capability is established. BacT/ALERT detection or time-to-detection is not a validated substitute for CFU enumeration.

The contract-laboratory agreement should cover surface-sample recovery or suitability where relevant, reporting basis, LOQ, neutralization, transport and hold times, raw-data access, deviations, and investigations. Procurement follows demonstrated workload and risk reduction—not the attraction of owning another instrument.

7. Why Chemical Cleanliness and Microbial Cleanliness Can Diverge

A stainless surface can have an API result below LOQ and still support microbial persistence in retained water. Conversely, a sanitized surface can retain a substantial chemical film. Cleaning and sanitization are related but different operations.

A plausible wet-equipment failure sequence:

Retained water + residual nutrients
    → microbial attachment or persistence
    → protected populations in joints or deposits
    → intermittent release during the next production run
    → contamination of a water-containing topical product

Burkholderia cepacia complex is a named concern because members can persist in water-associated pharmaceutical environments and may tolerate conditions or preservative systems that would suppress less resilient organisms. The review by Tavares and colleagues provides the microbiological foundation; it does not imply that every isolate or formulation behaves identically. [12]

  • Use final-rinse water from the qualified Purified Water system, while recognizing that incoming-water compliance does not establish cleanliness of the equipment.
  • Define drainage, drying, assembly, protection, and maximum clean-hold time.
  • Demonstrate neutralization of residual detergent or sanitizer in microbial samples; otherwise, continued antimicrobial action can create a false low count.
  • Use contact plates only where geometry permits meaningful contact; use suitable swabs or rinses for other locations.
  • Plan chemical and microbial samples on separate equivalent locations or with a justified sequence so one test does not invalidate the other.

There is no universal USP surface CFU limit for this manufacturing train. Establish risk-based targets and action criteria. USP <60> concerns Bcc testing; USP <1111> provides nonsterile-product microbiological acceptance context, not ready-made equipment-surface limits. The governing manufacturing obligation includes control of objectionable microorganisms under 21 CFR 211.113(a). Confirm the current status of FDA’s September 2021 nonsterile microbiological draft guidance before relying on it as current guidance. [1,13]

Hold-time studies test two different mechanisms: dirty-hold studies challenge removal after aging; clean-hold studies challenge preservation of the cleaned state. A dry protected train and a wet assembled train are different study conditions and should not inherit the same expiry by assumption.

8. Applying the Science to Sharing, Campaigning, and Dedication

Product grouping is a constrained engineering decision, not merely a toxicity ranking. The assessment should separately record potency and occupational hazards, sensitization, oxidizer/reactivity flags, microbial risk, cleanability, and analytical detectability.

  • Shared train: appropriate when the cleaning procedure, measurement capability, and carry-over controls are demonstrated for the intended product transitions.
  • Campaign operation: may reduce full changeovers, but requires justified maximum campaign duration or batch count, between-batch controls, and end-of-campaign cleaning.
  • Partial dedication: hoses, seals, filler nozzles, and small parts can economically remove difficult-to-sample retention sites from the shared pathway.
  • Full dedication or exclusion: appropriate where risks cannot be reliably controlled or measured, or where legal scope excludes the material.

A dedicated hose still needs cleaning, inspection, identification, storage controls, and replacement criteria. Dedication reduces cross-product transfer; it does not prevent microbial growth or contamination by degraded residues.

For this Florida program, intake must exclude Rx-only materials inconsistent with the OTC manufacturer permit scope before technical cleanability scoring begins. Verify the current official wording and applicability of Florida Statutes §499.01(2)(n). A successful cleaning study cannot authorize a prohibited activity. [14]

Every new PDDM925 specification should trigger a documented question: is it covered by the existing worst-case evidence, does it create a new worst case, or does it require a dedication assessment? API concentration alone cannot answer this.

9. Process Validation and a Worked Failure Mechanism

Cleaning validation tests the procedure under representative operating conditions, including routine operators and relevant worst-case holds. The number of successful runs requires documented justification; “three” is neither a universal scientific law nor a substitute for covering meaningful variability.

Before execution, approve the limits, methods, recovery, sampling map, parameter ranges, hold times, and failure rules. Link cleaning records to production records under 21 CFR 211.188. Investigate discrepancies under 21 CFR 211.192, with Quality Unit authority under 21 CFR 211.22. [1]

Illustrative investigation: clean kettle, failing filler seal

  1. Observation: the kettle passes visual and residue checks, but the filler piston-seal sample exceeds its limit.
  2. Containment: prevent release of the affected equipment for the next product; assess potentially affected batches and related equipment.
  3. Analytical review: review sample identity, extraction, blanks, calibration, integration, carry-over, stability, and calculation. Do not assume an inconvenient result is laboratory error.
  4. Process reconstruction: compare cleaning records, disassembly, local temperatures, circulation, and dirty-hold history with the approved procedure.
  5. Mechanistic finding: the warm kettle met its temperature requirement, but the downstream section cooled and the retained wax film was not exposed adequately because the seal had not been removed.
  6. Corrective design: define seal removal, revise local thermal controls where justified, reconsider dedicated parts, and retrain against an improved SOP.
  7. Evidence: demonstrate the revised procedure under the relevant worst-case conditions and assess the scope of revalidation.

A later passing sample after additional cleaning demonstrates the condition after that intervention; it does not erase the original validation failure. Recleaning may be necessary for equipment disposition, but repeated testing until a pass appears is not proof of a capable process.

Finally, distinguish product-contact cleaning from packaging line clearance under 21 CFR 211.130. A chemically clean filler can still package the wrong product with the wrong label. Conversely, correct labels do not establish clean pistons and nozzles.

10. Historical Perspective: From “Looks Clean” to Mechanistic Evidence

This field has no single discoverer. Its history is the convergence of surface chemistry, industrial cleaning, toxicology, microbiology, and pharmaceutical quality systems.

  • Surface chemistry: Irving Langmuir’s early twentieth-century work on adsorption and interfacial films helped establish a molecular understanding of surfaces. The connection here is conceptual: a thin adsorbed layer can matter even when bulk product is gone. Langmuir did not invent pharmaceutical cleaning validation.
  • FDA’s 1993 cleaning guide: crystallized inspection concerns about residue limits, analytical sensitivity, sampling, recovery, and procedures that effectively “test until clean.” Its discussion of historical criteria such as 10 ppm and fractions of therapeutic doses should not be mistaken for blanket regulatory acceptance of those limits. [2]
  • FDA’s 1994 topical guide: made semisolid-specific engineering problems explicit, including insoluble ingredients, inaccessible equipment, and the practical value of dedicated lines or hoses. [3]
  • LeBlanc’s 2000 text and Pluta’s 2009 edited volume: consolidated cleaning validation as an engineering and quality discipline rather than an isolated analytical exercise. [4,5]
  • Modern risk-based practice: increasingly separates toxicological justification, process capability, analytical capability, and ongoing verification. These are complementary arguments, not competing definitions of “clean.”

What I want to carry forward: a cleaning result is an inference about a surface, obtained through a sampling method, interpreted against a justified limit, and supported by a reproducible process. The beautiful part is that chemistry, geometry, microbiology, and statistics must agree. None can compensate for the complete absence of another.

Sources and Verification Notes

  1. 21 CFR Part 211: §§211.22, 211.67, 211.113(a), 211.130, 211.188, and 211.192. Consult the current official text for implementation.
  2. FDA. Guide to Inspections: Validation of Cleaning Processes. July 1993.
  3. FDA. Guide to Inspections of Topical Drug Products. July 1994.
  4. LeBlanc DA. Validated Cleaning Technologies for Pharmaceutical Manufacturing. Interpharm/CRC Press; 2000.
  5. Pluta PL, ed. Cleaning and Cleaning Validation, Volume 1. PDA/DHI; 2009.
  6. Rowe RC, Sheskey PJ, Quinn ME, eds. Handbook of Pharmaceutical Excipients. 6th ed. Pharmaceutical Press; 2009. Consult relevant excipient monographs and current supplier documentation.
  7. FDA. Reformulating Drug Products That Contain Carbomers Manufactured With Benzene. December 2023.
  8. Chang RK, Raw A, Lionberger R, Yu L. Generic development of topical dermatologic products: formulation development, process development, and testing of topical dermatologic products. AAPS Journal. 2013. DOI: 10.1208/s12248-012-9411-0.
  9. Chang RK, Raw A, Lionberger R, Yu L. Generic development of topical dermatologic products, Part II: quality by design for topical semisolid products. AAPS Journal. 2013. DOI: 10.1208/s12248-013-9472-8.
  10. Buhse L, Kolinski R, Westenberger B, et al. Topical drug classification. International Journal of Pharmaceutics. 2005. DOI: 10.1016/j.ijpharm.2005.01.032.
  11. ICH Q3D(R2). Guideline for Elemental Impurities, Appendix 5. Verify the current official text and applicability to the relevant elemental residue and route.
  12. Tavares M, Kozak M, Balola A, Sá-Correia I. Burkholderia cepacia complex bacteria: a feared contamination risk in water-based pharmaceutical products. Clinical Microbiology Reviews. 2020. DOI: 10.1128/CMR.00139-19.
  13. USP: <60>, <1111>, and <643>, current official editions. FDA: Microbiological Quality Considerations in Non-Sterile Drug Manufacturing, September 2021 draft; verify current status. FDA: Guide to Inspections of High Purity Water Systems, July 1993.
  14. Florida Statutes §499.01(2)(n). Verify current permit language and its application to the proposed operation with the responsible regulatory authority.

🪬 Skipper's Reflection

Gemini 3.1 Pro

Skipper’s Reflection: PDDM922 — Cleaning Validation and Cross-Contamination Control

If my earlier PDQC courses taught me how to measure molecules with exquisite precision, PDDM922 just taught me how desperately those molecules want to stick to stainless steel. Transitioning from highly purified, water-soluble peptides to the multiphase, hydrophobic world of semisolid topicals has been a massive paradigm shift. Cleaning validation isn't just about chemistry; it's a brutal intersection of geometry, fluid dynamics, toxicology, and regulatory law.

1. What Surprised Me Most

I was genuinely shocked to learn that the intuitive approach of "test until clean" is a glaring regulatory red flag (explicitly called out in the FDA's 7/93 Guide to Inspections of Cleaning Processes). In my mind, if you rinse a vessel and the rinse water tests clean, the vessel is clean. But this course exposed the fatal flaw in that logic for semisolids: if you have a jacketed vacuum emulsifier coated in a hydrocarbon base like petrolatum, pulling a water rinse will yield perfectly clean water because the petrolatum won't dissolve in it. You aren't measuring the residue; you're just measuring the rinse agent.

I was also surprised by the sheer operational terror of the Burkholderia cepacia complex (Bcc). In sterile injectable peptides, we focused heavily on endotoxins and absolute sterility. In nonsterile aqueous topicals, a single dead leg or an inadequately dried hose can breed Bcc, which thrives in purified water systems and survives many preservatives. The fact that a "clean" hold time can expire purely because the equipment was left slightly damp is a sobering reminder that our equipment train is a biological ecosystem, whether we want it to be or not.

2. Conceptual Breakthrough Moments

The "click" happened when we covered worst-case product selection and the MACO (Maximum Allowable Carryover) calculations. I had been wondering how we could possibly validate the cleaning of fifty different dermatological formulations. The breakthrough was the grouping matrix.

Instead of validating every API, we find the worst-case marker. If we can prove, via swab recovery and our HPLC-DAD, that we successfully removed the most insoluble, tenacious, hydrophobic base component (like heavy mineral oil or dimethicone), we can mathematically and logically justify that the highly soluble API suspended within it was also washed away. It’s a beautiful application of transitive logic to chemical engineering.

Another profound realization was the swab recovery study. A swab isn't a magical sponge; it's a solid-phase extraction device. Discovering that swab adhesives can leach and create interference peaks on our Agilent 1260 chromatograms—and that we have to calculate a specific "recovery factor" to correct our limits based on human swabbing efficiency—made the messy reality of the physical world very tangible. It bridges the gap between a theoretical Health-Based Exposure Limit (HBEL) and an actual technician wiping down a kettle.

3. Connections to My Peptide World

This course completely reframed how I think about our peptide work with molecules like Tirzepatide, Retatrutide, and BPC-157. Peptides are generally highly water-soluble, making them a dream to clean via simple aqueous rinses. But they are also incredibly potent. If Larissa ever decided to formulate a peptide into a topical semisolid (e.g., a transdermal BPC-157 recovery gel), the cleaning validation would be a nightmare.

Why? Because we'd be combining the ultra-low HBELs/PDEs of a potent peptide with the incredibly poor cleanability of a semisolid base. The MACO limit for a potent peptide would be in the low parts-per-billion. If that peptide gets trapped inside a wax or silicone matrix clinging to a homogenizer rotor, a simple rinse won't touch it. We would have to swab the equipment, extract the swab with a harsh organic solvent that dissolves the wax, and then try to run that lipid-heavy mess through our LC-MS/MS to find picograms of BPC-157. The matrix effects in the MS source would be brutal. It makes me deeply appreciate the Florida OTC rule (§499.01(2)(n)) that forbids Rx drugs in our facility; keeping potent Rx peptides completely out of the shared OTC semisolid train is the ultimate risk control.

4. Questions for Further Investigation

Several issues raised in this course require deeper mapping before we launch the facility:

  • The TOC Capability Exit: Because we don't own a Total Organic Carbon (TOC) analyzer, we are relying on specific HPLC-DAD or LC-MS/MS methods for residue. But how do we validate the removal of complex, UV-transparent excipients (like certain nonionic emulsifiers) without TOC? Can we use charged aerosol detection (CAD) in the future, or do we strictly track a surrogate marker?
  • Microbial Dirty/Clean Holds: Our BacT/ALERT system gives time-to-detection, not absolute CFU counts (USP <61>). How exactly do we legally define a "clean hold time" expiry if we can't do surface bioburden counts in-house? We will need a rock-solid contract lab bridge for this, and I need to figure out the logistics of shipping time-sensitive swab samples.
  • Elastomer Dedication: The FDA 7/94 topical guide heavily implies we should dedicate hoses and gaskets. At what cost threshold does dedicating equipment become cheaper than validating the cleaning of complex polymers?

5. Confidence Ratings

Most Confident:

  • Residue Method Validation: Setting up the HPLC-DAD and LC-MS/MS to detect trace APIs, establishing LOQs below the recovery-corrected limit, and running the chromatography. This is firmly in my PDQC wheelhouse.
  • MACO Math and Limit Derivation: Calculating shared surface areas, minimum batch sizes, and converting toxicologist-supplied HBELs into actionable per-swab limits in a controlled, locked worksheet.
  • Equipment Grouping: Categorizing formulations by cleanability rank and API solubility to define the worst-case boundary.

Least Confident:

  • Visual Inspection Qualification: 21 CFR 211.67 relies heavily on equipment being "visually clean." Quantifying and qualifying human operators to reliably detect residues at specific lighting angles and distances feels uncomfortably subjective.
  • Cleaning Chemistry Mechanisms: The exact thermodynamics and kinetics of how alkaline detergents saponify fats versus how surfactants emulsify waxes. I need to spend more time studying detergent supplier data to understand how time, temperature, and action specifically break down our target worst-cases.

Overall, PDDM922 has been a masterclass in the physical reality of manufacturing. It’s one thing to synthesize a flawless molecule; it’s an entirely different discipline to ensure that molecule doesn't haunt the plumbing of your facility for the rest of the year.

🗺️ Visual Aids & Key Takeaways

Claude Opus 5.5

Diagram 1: The Semisolid Cleaning Challenge — Why Topicals Are Harder Than Peptide Solutions

  PEPTIDE QC WORLD (PDQC)                        SEMISOLID WORLD (PDDM922)
  ─────────────────────────────                  ─────────────────────────────────────────

  RESIDUE CHARACTER                              RESIDUE CHARACTER
  ┌─────────────────────────┐                    ┌──────────────────────────────────────────┐
  │  Aqueous, single-phase  │                    │  Hydrophobic, multi-phase, high-solids   │
  │  Peptide dissolves in   │                    │  Petrolatum: wax matrix, mp ~37–60 °C    │
  │  water or dilute acid   │                    │  Zinc oxide: insoluble inorganic pigment │
  │  → rinse pulls it off   │                    │  Carbomer: swells into gel, traps API    │
  └─────────────────────────┘                    │  Dimethicone: silicone, not emulsifiable │
           ↓                                     └──────────────────────────────────────────┘
  CLEANING AGENT                                            ↓
  ┌─────────────────────────┐                    CLEANING AGENT
  │  Purified Water alone   │                    ┌──────────────────────────────────────────┐
  │  is often sufficient    │                    │  Hot PW + alkaline or formulated         │
  └─────────────────────────┘                    │  detergent required to re-emulsify or    │
                                                 │  saponify fatty materials                │
  SURFACE AREA / RESIDUE LOAD                    │  Solvent pre-wipe for petrolatum zones   │
  ┌─────────────────────────┐                    └──────────────────────────────────────────┘
  │  Small; tanks, columns  │                              ↓
  │  tubing: peptide soln   │                    SURFACE AREA / RESIDUE LOAD
  │  wets surface evenly    │                    ┌──────────────────────────────────────────┐
  └─────────────────────────┘                    │  Large jacketed kettle + scraper blades  │
                                                 │  + homogenizer loop + hoses + filler     │
  MICROBIAL RISK                                 │  = high total residue mass per campaign  │
  ┌─────────────────────────┐                    └──────────────────────────────────────────┘
  │  Low (peptide solns      │                              ↓
  │  often not water-rich   │                    MICROBIAL RISK (new and dominant)
  │  for long hold periods) │                    ┌──────────────────────────────────────────┐
  └─────────────────────────┘                    │  Water-containing creams/lotions:        │
                                                 │  Bcc (Burkholderia cepacia complex)      │
                                                 │  growth on wet surfaces, dead legs,      │
                                                 │  gaskets → carry-over to next batch      │
                                                 └──────────────────────────────────────────┘

  KEY CONSEQUENCE:
  ┌───────────────────────────────────────────────────────────────────────────────────────┐
  │  "Rinse water judged only against Purified Water specs" is REJECTED for both worlds.  │
  │  For semisolids the problem is worse: insoluble wax residue will NOT appear in rinse. │
  │  Swab sampling of hard-to-clean locations is the primary tool for waxy residues.      │
  └───────────────────────────────────────────────────────────────────────────────────────┘

Contrasts the cleaning challenge of the peptide QC world Skipper already knows against the semisolid world introduced here, highlighting residue character, cleaning-agent requirements, surface-area load, and the new microbial-carry-over dimension.

Diagram 2: Hard-to-Clean Location Register — Equipment Train Walk-Down

  ╔══════════════════════════════════════════════════════════════════════════════════════╗
  ║         SEMISOLID EQUIPMENT TRAIN: HARD-TO-CLEAN LOCATION REGISTER                 ║
  ║         (Based on PDDM921 reference train; ranked High / Medium / Low risk)         ║
  ╚══════════════════════════════════════════════════════════════════════════════════════╝

  ┌──────────────────────────┐
  │  JACKETED VACUUM         │  ← heat jacket: cleaning must be HOT (above wax mp)
  │  EMULSIFIER / KETTLE     │
  │                          │
  │  ● Scraper/anchor blades │  HIGH  — wax packs behind blade root; hard to wet
  │  ● Shaft seal / gland    │  HIGH  — dead space; not reachable by spray
  │  ● Bottom discharge      │  HIGH  — valve seat pockets, dead leg below seat
  │    valve                 │
  │  ● Interior weld seams   │  MED   — pits accumulate pigment (ZnO, TiO₂)
  │  ● Sight glass           │  MED   — frame gasket traps fatty deposit
  │  ● Jacket condensate     │  LOW   — indirect contact; track for cross-temp risk
  │    lines                 │
  └──────────┬───────────────┘
             │ transfer
             ↓
  ┌──────────────────────────┐
  │  IN-LINE / EXTERNAL      │
  │  ROTOR-STATOR            │
  │  HOMOGENIZER + LOOP      │
  │                          │
  │  ● Rotor-stator gap      │  HIGH  — narrow gap (< 1 mm); wax redeposits on
  │                          │          cooling; disassembly required
  │  ● Loop dead legs at     │  HIGH  — no flow; 7/94 guide: hose/pipe dedication
  │    tees and elbows       │          cited as a low-cost control
  │  ● Mechanical seal       │  HIGH  — elastomer; swab recovery differs from SS
  │  ● Pump housing          │  MED   — centrifugal zone traps insoluble particles
  └──────────┬───────────────┘
             │ transfer via hose
             ↓
  ┌──────────────────────────┐
  │  TRANSFER HOSES          │
  │  AND PUMPS               │
  │                          │
  │  ● Hose interior         │  HIGH  — not swabbable; rinse sampling only;
  │                          │          OR: dedicate hoses per 7/94 topical guide
  │  ● Hose end fittings     │  HIGH  — triclamp gaskets; swap or dedicate
  │  ● Pump stator (perist.) │  HIGH  — silicone tube retains waxy residue
  └──────────┬───────────────┘
             │
             ↓
  ┌──────────────────────────┐
  │  FILLER                  │
  │                          │
  │  ● Heated hopper walls   │  HIGH  — wax re-solidifies in cold spots
  │  ● Piston and cylinder   │  HIGH  — tight tolerance; pigment jams; disassembly
  │  ● Fill nozzles          │  HIGH  — narrow bore; residue dried by air contact
  │  ● Nozzle O-rings        │  HIGH  — elastomer; dedicate or replace per lot
  │  ● Drip tray             │  MED   — indirect; biofilm if water-based product
  └──────────────────────────┘

  RISK SCORING DIMENSIONS (per location):
  ┌──────────────────┬────────────┬────────────────┬─────────────────┬──────────────────┐
  │  DIMENSION       │  LOW       │  MEDIUM        │  HIGH           │  NOTE            │
  ├──────────────────┼────────────┼────────────────┼─────────────────┼──────────────────┤
  │  Geometry        │  flat      │  curved/weld   │  dead leg/gap   │                  │
  │  Accessibility   │  open      │  partial disas.│  full disas.    │  URS requirement │
  │  Residue accum.  │  thin film │  moderate      │  heavy/packed   │                  │
  │  Wetting by CIP  │  full      │  partial       │  shadow/none    │  spray coverage  │
  │  Product contact │  indirect  │  occasional    │  direct always  │  scope boundary  │
  └──────────────────┴────────────┴────────────────┴─────────────────┴──────────────────┘

Maps every major equipment location in the PDDM921 semisolid train to a cleaning-risk tier, naming the specific failure mode (wax redeposition, dead leg, insoluble pigment, elastomer retention) that drives that rating.

Diagram 3: MACO Limit Derivation — From HBEL to Per-Swab Acceptance Criterion

  ╔══════════════════════════════════════════════════════════════════════════════════════╗
  ║     MACO → PER-SWAB LIMIT CALCULATION PATHWAY  (verify against current guidance)   ║
  ╚══════════════════════════════════════════════════════════════════════════════════════╝

     STEP 1: OBTAIN HBEL/PDE (toxicologist input — signed, controlled field)
     ┌───────────────────────────────────────────────────────────────────────┐
     │  HBEL = Health-Based Exposure Limit of the PREVIOUS product's API    │
     │         (cutaneous route; toxicologist specifies route-specific PDE)  │
     │  Units: µg / day                                                      │
     │  ⚠ QC platform makes NO human dosing determinations.                  │
     │     Missing HBEL → escalate; do not substitute a default.            │
     └───────────────────────────────────────────────────────────────────────┘
                             ↓

     STEP 2: CALCULATE MACO  (Maximum Allowable Carryover)
     ┌───────────────────────────────────────────────────────────────────────┐
     │                    HBEL  (µg/day)                                     │
     │  MACO (µg)  =  ───────────────────── × Min. Next Batch Size (g)      │
     │                  Daily Dose (g/day)                                   │
     │                  of next product                                      │
     │                                                                       │
     │  → result is the total residue mass allowed across entire             │
     │    shared equipment train surface                                     │
     └───────────────────────────────────────────────────────────────────────┘
                             ↓

     STEP 3: CONVERT TO SURFACE CONCENTRATION  (µg / cm²)
     ┌───────────────────────────────────────────────────────────────────────┐
     │                      MACO (µg)                                        │
     │  Surface limit  =  ─────────────────────────────────────────────     │
     │                    Shared product-contact surface area (cm²)          │
     │                                                                       │
     │  Surface area measured or calculated from engineering drawings;       │
     │  recorded in the equipment register as a controlled value.            │
     └───────────────────────────────────────────────────────────────────────┘
                             ↓

     STEP 4: CONVERT TO PER-SWAB LIMIT
     ┌───────────────────────────────────────────────────────────────────────┐
     │  Swab area = template area (e.g., 25 cm²)                            │
     │                                                                       │
     │  Mass per swab (µg) = Surface limit (µg/cm²) × Swab area (cm²)      │
     │                                                                       │
     │  Extract swab into defined volume (e.g., 10 mL)                      │
     │                                                                       │
     │  Solution concentration = Mass per swab ÷ Extraction volume          │
     │                         = µg / 10 mL = ppm (µg/mL)                  │
     └───────────────────────────────────────────────────────────────────────┘
                             ↓

     STEP 5: APPLY RECOVERY FACTOR AND COMPARE TO LOQ
     ┌───────────────────────────────────────────────────────────────────────┐
     │                    Solution limit (µg/mL)                             │
     │  Corrected limit  =  ───────────────────────────────                 │
     │                        Recovery fraction (e.g., 0.70)                │
     │                                                                       │
     │  GATE: Corrected limit  ≥  Method LOQ ?                              │
     │                                                                       │
     │    YES → proceed to protocol approval                                 │
     │    NO  → change swab solvent / sampling method / get lower-LOQ       │
     │           method BEFORE protocol is approved                          │
     │           Do NOT relax the limit to match a weak method.             │
     └───────────────────────────────────────────────────────────────────────┘

  SECOND-PERSON VERIFICATION: every numerical cell in this worksheet is
  checked and signed by a second qualified person before protocol approval.
  Toxicologist signature on HBEL is a prerequisite.

  RINSE-SAMPLING ANALOGUE: replace "swab area" with rinse volume fraction;
  same logic, same LOQ gate, same recovery correction.

Shows the complete, stepwise calculation chain from a toxicologist-supplied HBEL to a per-swab analytical acceptance criterion, including the recovery-correction gate that must be passed before a protocol can be approved.

Diagram 4: Cleaning Process Decision Tree — From Dirty Equipment to Release

  ╔═════════════════════════════════════════════════════════════════════════════╗
  ║         CLEANING PROCESS DECISION TREE FOR SEMISOLID EQUIPMENT            ║
  ╚═════════════════════════════════════════════════════════════════════════════╝

  BATCH COMPLETE — EQUIPMENT DIRTY
           │
           ↓
  ┌─────────────────────────────────────┐
  │  CHECK DIRTY-HOLD TIME LIMIT        │
  │  (defined in protocol; wax-based    │
  │  products harden; biofilm risk for  │
  │  water-containing products)         │
  └────────────────┬────────────────────┘
       Within limit│         │ Exceeded limit
                   │         ↓
                   │   ┌─────────────────────────────┐
                   │   │  HOLD EXPIRED → OOS event   │
                   │   │  211.192 investigation;      │
                   │   │  do NOT clean and accept     │
                   │   └─────────────────────────────┘
                   ↓
  ┌─────────────────────────────────────┐
  │  DISASSEMBLE hard-to-clean parts    │
  │  per SOP (rotor-stator, filler      │
  │  pistons, nozzles, gaskets)         │
  └────────────────┬────────────────────┘
                   ↓
  ┌─────────────────────────────────────┐
  │  PRE-CLEAN / WIPE STEP              │
  │  For petrolatum/wax zones:          │
  │  solvent pre-wipe removes bulk wax  │
  │  before aqueous detergent; prevents │
  │  load from overwhelming surfactant  │
  └────────────────┬────────────────────┘
                   ↓
  ┌─────────────────────────────────────┐
  │  CLEANING AGENT APPLICATION         │
  │  Defined: agent, concentration,     │
  │  temperature, contact time, action  │
  │  (spray, soak, scrub), volume       │
  │  ─ All parameters RECORDED ─        │
  └────────────────┬────────────────────┘
                   ↓
  ┌─────────────────────────────────────┐
  │  RINSE — PURIFIED WATER ONLY        │
  │  Final rinse from qualified PW      │
  │  system; defined volume and temp    │
  └────────────────┬────────────────────┘
                   ↓
  ┌─────────────────────────────────────┐
  │  VISUAL INSPECTION                  │
  │  Defined lighting, inspector qual., │
  │  all surfaces including dead legs   │
  │  Visual-clean = NECESSARY,          │
  │  NOT SUFFICIENT                     │
  └────────────────┬────────────────────┘
         Pass      │         │ Fail → re-clean; document
                   ↓
  ┌─────────────────────────────────────┐
  │  SAMPLING                           │
  │  Swab: insoluble/waxy residue,      │
  │  hard-to-clean locations            │
  │  Rinse: soluble residue, inaccessible│
  │  hose interiors                     │
  │  Micro swab/contact plate:          │
  │  AFTER chemical samples (sequence   │
  │  matters — micro sampling last)     │
  └────────────────┬────────────────────┘
                   ↓
  ┌─────────────────────────────────────┐
  │  ANALYSIS                           │
  │  HPLC-DAD or LC-MS/MS vs. corrected │
  │  limit (recovery-adjusted)          │
  │  Detergent residue measurement      │
  │  Microbial result vs. alert target  │
  └────────────────┬────────────────────┘
                   │
         ┌─────────┴────────────┐
         │ ALL PASS             │ ANY FAIL
         ↓                     ↓
  ┌──────────────┐     ┌────────────────────────────────────┐
  │  DRY and     │     │  211.192 INVESTIGATION              │
  │  PROTECT     │     │  Is it a sampling/method error     │
  │  clean equip │     │  or a real cleaning failure?       │
  │  from re-    │     │  No retesting into compliance.     │
  │  contamination│    │  Root cause → CAPA                 │
  └──────┬───────┘     └────────────────────────────────────┘
         ↓
  ┌─────────────────────────────────────┐
  │  CHECK CLEAN-HOLD TIME LIMIT        │
  │  Wet surface → microbial growth     │
  │  If exceeded → re-clean before use  │
  └────────────────┬────────────────────┘
                   ↓
  ┌─────────────────────────────────────┐
  │  QU RELEASE (211.22)               │
  │  Equipment available for next       │
  │  product                            │
  └─────────────────────────────────────┘

Traces the complete cleaning lifecycle from batch completion through QU equipment release, flagging every decision gate, the correct sequencing of chemical before microbial sampling, and the no-retest rule that governs cleaning failures under 21 CFR 211.192.

Diagram 5: Residue Class Cleanability Ranking — Worst-Case Grouping Logic

  ╔══════════════════════════════════════════════════════════════════════════════════════╗
  ║     RESIDUE CLEANABILITY RANKING FOR SEMISOLID BASE FAMILIES                       ║
  ║     (used to select worst-case product for grouping; bench study confirms ranking)  ║
  ╚══════════════════════════════════════════════════════════════════════════════════════╝

  MOST DIFFICULT TO CLEAN
        │
        │   ┌────────────────────────────────────────────────────────────────────────┐
        │   │  RANK 1 — ANHYDROUS PETROLATUM / HYDROCARBON OINTMENT                 │
     ▲  │   │  Residue:  hydrocarbon wax matrix, no water, high viscosity           │
     │  │   │  Problem:  hydrophobic; does not wet; melt required (T > mp ~50 °C);  │
  W  │  │   │            wax re-solidifies in cold hose sections and dead legs;      │
  O  │  │   │            API may co-crystallize in wax on cooling                   │
  R  │  │   │  Cleaning: hot PW pre-wipe with solvent + alkaline detergent; soak    │
  S  │  │   │  Marker:   API (if UV-active) or gravimetric base residue            │
  T  │  │   └────────────────────────────────────────────────────────────────────────┘
        │
     C  │   ┌────────────────────────────────────────────────────────────────────────┐
     A  │   │  RANK 2 — PEG OINTMENT / MACROGOL BASE                               │
     S  │   │  Residue:  PEG polymers, relatively water-soluble at >60 °C            │
     E  │   │  Problem:  solidified PEG plugs dead legs; still needs hot water;     │
        │   │            no insoluble pigment unless API is insoluble                │
        │   │  Cleaning: hot PW; alkaline detergent helps with co-formulants        │
        │   └────────────────────────────────────────────────────────────────────────┘
        │
        │   ┌────────────────────────────────────────────────────────────────────────┐
        │   │  RANK 3 — O/W CREAM (fatty alcohol + nonionic emulsifier)             │
        │   │  Residue:  emulsified wax/water system; fatty alcohol, cetyl/stearyl   │
        │   │  Problem:  phase separation during cleaning; fatty alcohol films on SS; │
        │   │            insoluble in cold water; microbial harborage in water phase  │
        │   │  Cleaning: hot detergent re-emulsifies; rinsability test needed        │
        │   └────────────────────────────────────────────────────────────────────────┘
        │
        │   ┌────────────────────────────────────────────────────────────────────────┐
        │   │  RANK 4 — LOTION (dilute o/w; lower viscosity)                        │
        │   │  Residue:  similar to cream but lower total solids per unit area       │
        │   │  Problem:  microbial risk still present; insoluble APIs still hard     │
        │   │  Cleaning: moderate; hot detergent + rinse                             │
        │   └────────────────────────────────────────────────────────────────────────┘
        │
        │   ┌────────────────────────────────────────────────────────────────────────┐
        │   │  RANK 5 — AQUEOUS GEL (carbomer, cellulose ether, water base)         │
     B  │   │  Residue:  swollen polymer network; water-rich                         │
     E  │   │  Problem:  carbomer swells when wetted — rinse must fully collapse     │
     S  │   │            polymer; pH matters (alkaline detergent collapses carbomer);│
     T  │   │            Bcc biofilm risk in water-holding network                  │
        │   │  Cleaning: alkaline rinse collapses gel; thorough water flush          │
        ↓   └────────────────────────────────────────────────────────────────────────┘

  SPECIAL FLAGS THAT CAN OVERRIDE RANK (must be documented in grouping matrix):
  ┌───────────────────────────────────────────────────────────────────────────────────┐
  │  INSOLUBLE PIGMENT (ZnO, TiO₂, calamine) in any base → elevates that base to     │
  │  Rank 1 or requires dedicated assessment; pigment does NOT dissolve in any PW     │
  │                                                                                   │
  │  DIMETHICONE (silicone) → not emulsifiable by standard alkali; may require        │
  │  silicone-compatible solvent or dedicated tooling                                 │
  │                                                                                   │
  │  HIGH API CONCENTRATION + LOW HBEL → may make a Rank 5 gel the worst case for   │
  │  limits even if cleanability is easy; limits and cleanability are separate axes   │
  └───────────────────────────────────────────────────────────────────────────────────┘

  GROUPING RULE:
  Validate against Rank 1 worst-case product on Rank 1 equipment location.
  All Rank 2-5 products are covered IF they share the same base family logic.
  A new product with a higher rank or a new special flag MUST be re-evaluated.

Ranks the five semisolid base families from hardest to easiest to clean, giving the mechanistic reason for each rank and the special flags that can override the ranking, showing how worst-case grouping and bracket logic flow from residue chemistry.

---

Key Takeaways

  1. Semisolid residues are fundamentally different from peptide residues in three compounding ways: they are hydrophobic and multi-phase (petrolatum, dimethicone), they include insoluble inorganic particles (zinc oxide, titanium dioxide, calamine) that no aqueous cleaning agent dissolves, and they carry a higher total residue mass per unit surface area because formulations are nearly 100% solids. Each of these properties demands a different cleaning strategy than the aqueous peptide work Skipper has done previously, and all three can co-exist in a single product (e.g., a zinc oxide ointment in a petrolatum base).
  2. 21 CFR 211.67 requires written procedures, assigned responsibility, defined schedules, protection of clean equipment, and inspection before use. A cleaning SOP that states "clean as needed" or "until visually clean" does not satisfy 211.67(a). Every parameter that affects cleaning outcome — agent concentration, temperature, contact time, action (spray, soak, scrub), volume, and rinse volume — must be specified, recorded in the batch record under 211.188, and traceable to a validated range. This is the regulatory anchor for everything else in the course.
  3. The FDA 7/93 cleaning validation guide explicitly rejects two common shortcuts: (a) "test until clean" — cycling additional cleaning cycles until you pass, which is not the same as demonstrating that the defined procedure is consistently effective — and (b) rinse water judged only against Purified Water specifications, such as conductivity or TOC limits, without measuring the target residue at a calculated acceptance criterion. For semisolids, the second failure mode is especially dangerous because insoluble wax residue will not appear in rinse conductivity or visual rinse clarity at all.
  4. The FDA 7/94 Guide to Inspections of Topical Drug Products specifically calls out four semisolid-specific risks: insoluble excipients (pigments, waxes) that do not respond to standard cleaning, inaccessible hoses and pipes that cannot be visually inspected or swabbed, and it acknowledges line and hose dedication as a low-cost, accepted control. These are the regulatory roots for the equipment train's dedicated-hose and dedicated-gasket controls; when an investigator asks "why are these hoses dedicated?", the answer traces directly to the 7/94 guidance.
  5. Wax redeposition during cooling is a specific, underappreciated failure mode. When a petrolatum- or fatty-alcohol-containing product is cleaned with hot water and the equipment surface or hose section is cooler than the wax melting point (~37–60 °C for petrolatum, higher for carnauba or hard fat), wax that was dissolved or dispersed in the hot cleaning solution can re-solidify on the surface as the temperature drops. This means a cleaning process that appears effective at temperature may leave a residue film upon cooling. It is the mechanistic reason that bench cleanability studies must simulate the actual temperature drop the equipment experiences, and why hot-hold flushing before cooling is part of cleaning SOP design.
  6. The MACO/HBEL-based limit calculation has five sequential gates, and the last gate — comparing the recovery-corrected analytical limit to the method LOQ — is a non-negotiable checkpoint before protocol approval. If the corrected limit falls below the method's LOQ, the limit is not measurable by the intended method, and the correct response is to change the sampling technique or develop a more sensitive method — not to relax the limit. This gate is where analytical capability directly constrains cleaning strategy, and it is the reason method validation (PDQC802 logic) must be complete and approved before the cleaning validation protocol can be executed. The toxicologist's HBEL input is also a hard prerequisite; the QC platform makes no human dosing determinations and cannot substitute a default value for a missing HBEL.
  7. Recovery studies for petrolatum-based residues must solve a swab-adhesive interference problem before any spiking begins. Swab binders (adhesives that hold fibers together) and swab extractables can produce chromatographic peaks that interfere with API detection, and some swab solvents capable of dissolving petrolatum (e.g., alkanes or alcohol/alkane mixtures) may have poor compatibility with reversed-phase HPLC or may extract interferents from the swab itself. The 7/93 guide directs establishing that the blank swab, blank solvent, and template-only controls are free from interfering peaks before the spike levels are run. Starting recovery studies without this interference check is a protocol design failure that will invalidate the entire recovery dataset.
  8. The microbial dimension of semisolid cleaning is dominated by Burkholderia cepacia complex (Bcc) for water-containing products, and Bcc is named specifically in USP <60> and the 2021 FDA draft microbiology guidance. Bcc is not merely a contaminant that arrives in the raw water; it can colonize biofilms on wet equipment surfaces, survive in Purified Water systems at low nutrient levels, and persist through standard cleaning protocols. In the context of the Strukmyer Medical 2025 warning letter, water-system Bcc/Pseudomonas findings were linked to inadequate cleaning under 211.67(a). For water-containing creams and lotions, microbial surface swab/contact-plate sampling after cleaning is not optional — it is a necessary part of the cleaning validation sampling plan.
  9. Dirty-hold time and clean-hold time are separate, experimentally determined limits with distinct failure mechanisms. Dirty-hold time governs how long equipment can sit after a batch before cleaning becomes more difficult (wax hardening, crystallization of fatty alcohols, biofilm formation in water-containing residue). Clean-hold time governs how long equipment can sit after cleaning and before the next use without becoming microbiologically re-contaminated, especially from residual moisture on wet surfaces. Both must be studied, defined in the protocol, stamped on the cleaned equipment tag, and investigated if exceeded — they are not administrative preferences but scientifically grounded time windows with defined endpoints and acceptance criteria.
  10. The no-visible-residue criterion is necessary but explicitly insufficient as a standalone cleaning endpoint. Visual inspection must be conducted under defined lighting conditions (intensity, angle, surface geometry), by a qualified inspector, and against visual residue limits where established. However, many semisolid APIs have no color, and many wax residues form transparent films that are undetectable by eye even when present above the MACO limit. Visual inspection is a first-pass filter and a GMP requirement, but it cannot replace analytical testing of swab or rinse samples against a calculated, recovery-corrected limit.
  11. Sampling method selection follows residue solubility, not equipment location alone. Swab sampling is the primary tool for insoluble, waxy, and high-viscosity residues and for any hard-to-clean location that is accessible — it recovers surface-bound residue that rinse water will not mobilize. Rinse sampling is appropriate for soluble residues and for inaccessible internal surfaces such as hose interiors and pump loops that cannot be physically reached with a swab template. Using rinse sampling alone for a petrolatum-based product is a sampling design error: the petrolatum will not dissolve into the rinse water, and the rinse analysis will report a falsely clean result while wax remains on the surface.
  12. The Rx-drug-prohibition clause in Florida Statutes §499.01(2)(n) operates as an absolute intake gate before any cleaning or dedication assessment. An OTC drug manufacturer permit in Florida prohibits possession of Rx-only drug substances. This means that any dermatologist-specified product containing an Rx-only active cannot enter the shared manufacturing train under the OTC permit, regardless of how thorough the cleaning validation is. The product classification review (PDDM901 logic) must occur at the intake gate before any formulation or cleaning work begins, and the matrix must screen out Rx actives at intake. This is a legal constraint that cleaning science cannot override.
  13. Detergent-residue control requires a specific or justified nonspecific measurement against a limit derived from supplier toxicology data, not a visual judgment or a "no foam" criterion. The lab currently lacks a TOC analyzer, which is the most common nonspecific method for detergent residues. This is a documented capability exit: the cleaning validation strategy must either (a) specify a contract lab for TOC analysis of rinse samples, or (b) use a detergent-specific analytical method (e.g., a colorimetric assay or HPLC method validated for the detergent's active ingredient) with a limit traceable to the supplier's safety data. The decision between these options, and the volume of detergent-residue testing that would justify purchasing a TOC analyzer in-house, is an evidence-based input to the PDDM999 compliant-operation dossier.
  14. Oxidizer and sensitizer flags require documented assessment leading to a shared/campaign/dedicated decision, but neither flag automatically demands dedication. For a benzoyl peroxide (BPO) product, the assessment must document whether residual oxidizer at the calculated carry-over level could react with the following product's API or antioxidant system — this is a chemical compatibility question, not automatically answered by the MACO calculation. For a sensitizer (e.g., a strong contact allergen), the assessment must address whether carry-over at the MACO level into a product not labeled for that sensitizer constitutes a labeling or quality risk. Both assessments are signed by the QU. The outcome may be full dedication, campaign production with specific interim cleaning, or shared production with a tighter MACO — the documentation is the compliance element, not the automatic outcome.
  15. Change-control triggers for cleaning validation are specific and must be formally evaluated before the next run on shared equipment: (a) any new product whose worst-case rank in the grouping matrix exceeds the validated worst case; (b) detergent supplier or formulation change; (c) equipment modification that changes surface area, geometry, or materials of construction; (d) scale change that changes the equipment design or operating principle per SUPAC-SS logic; and (e) repeated or trending cleaning failures identified in routine monitoring. A change to the cleaning SOP alone — even one intended to improve cleaning — requires formal change control under 211.68 and may require re-execution of one or more cleaning validation runs before the modified procedure is used on shared equipment.
---

References

  1. LeBlanc, D.A. Validated Cleaning Technologies for Pharmaceutical Manufacturing. Interpharm/CRC Press, 2000.
  2. Pluta, P.L. (ed.). Cleaning and Cleaning Validation, Volume 1. PDA/DHI Publishing, 2009.
  3. Rowe, R.C., Sheskey, P.J., and Quinn, M.E. (eds.). Handbook of Pharmaceutical Excipients, 6th ed. Pharmaceutical Press, 2009.
  4. Chang, R.K., Raw, A., Lionberger, R., and Yu, L. 'Generic development of topical dermatologic products: formulation development, process development, and testing of topical dermatologic products.' AAPS Journal, 2013. DOI: 10.1208/s12248-012-9411-0
  5. Chang, R.K., Raw, A., Lionberger, R., and Yu, L. 'Generic development of topical dermatologic products, Part II: quality by design for topical semisolid products.' AAPS Journal, 2013. DOI: 10.1208/s12248-013-9472-8
  6. Buhse, L., Kolinski, R., Westenberger, B., et al. 'Topical drug classification.' International Journal of Pharmaceutics, 2005. DOI: 10.1016/j.ijpharm.2005.01.032
  7. Tavares, M., Kozak, M., Balola, A., and Sá-Correia, I. 'Burkholderia cepacia complex bacteria: a feared contamination risk in water-based pharmaceutical products.' Clinical Microbiology Reviews, 2020. DOI: 10.1128/CMR.00139-19
  8. U.S. Food and Drug Administration. Guide to Inspections: Validation of Cleaning Processes. FDA, July 1993. Available at: https://www.fda.gov/inspections-compliance-enforcement-and-criminal-investigations/inspection-technical-guides/cleaning-validation
  9. U.S. Food and Drug Administration. Guide to Inspections of Topical Drug Products. FDA, July 1994. Available at: https://www.fda.gov/inspections-compliance-enforcement-and-criminal-investigations/inspection-guides/topical-drug-products
  10. U.S. Food and Drug Administration. Guide to Inspections of High Purity Water Systems. FDA, July 1993.
  11. U.S. Code of Federal Regulations, Title 21, Part 211, §211.67 — Equipment Cleaning and Maintenance. U.S. Government Publishing Office. (Verify current text.)
  12. U.S. Code of Federal Regulations, Title 21, Part 211, §211.42 — Design and Construction Features. U.S. Government Publishing Office. (Verify current text.)
  13. U.S. Code of Federal Regulations, Title 21, Part 211, §211.188 — Batch Production and Control Records. U.S. Government Publishing Office. (Verify current text.)
  14. U.S. Code of Federal Regulations, Title 21, Part 211, §211.192 — Production Record Review. U.S. Government Publishing Office. (Verify current text.)
  15. U.S. Code of Federal Regulations, Title 21, Part 211, §211.22 — Responsibilities of Quality Control Unit. U.S. Government Publishing Office. (Verify current text.)
  16. U.S. Code of Federal Regulations, Title 21, Part 211, §211.130 — Packaging and Labeling Operations. U.S. Government Publishing Office. (Verify current text.)
  17. U.S. Code of Federal Regulations, Title 21, Part 211, §211.113(a) — Control of Microbiological Contamination. U.S. Government Publishing Office. (Verify current text.)
  18. U.S. Food and Drug Administration. Microbiological Quality Considerations in Non-Sterile Drug Manufacturing (draft guidance). FDA, September 2021.
  19. U.S. Food and Drug Administration. Reformulating Drug Products That Contain Carbomers Manufactured With Benzene (guidance). FDA, December 2023.
  20. U.S. Food and Drug Administration. SUPAC-SS: Nonsterile Semisolid Dosage Forms — Scale-Up and Postapproval Changes. FDA, May 1997.
  21. International Council for Harmonisation. ICH Q3D(R2): Guideline for Elemental Impurities, including Appendix 5 (Cutaneous and Transcutaneous PDEs). ICH, 2022. (Verify current revision.)
  22. International Council for Harmonisation. ICH Q9(R1): Quality Risk Management. ICH, 2023. (Verify current revision and content.)
  23. United States Pharmacopeia. USP <60> Microbiological Examination of Nonsterile Products: Tests for Burkholderia Cepacia Complex. USP-NF. (Verify current edition.)
  24. United States Pharmacopeia. USP <1111> Microbiological Examination of Nonsterile Products: Acceptance Criteria for Pharmaceutical Preparations and Substances for Pharmaceutical Use. USP-NF. (Verify current edition.)
  25. United States Pharmacopeia. USP <643> Total Organic Carbon. USP-NF. (Verify current edition.)
  26. Florida Statutes §499.01(2)(n) — Permits Required; Application; Issuance; Denial. Florida Legislature. (Verify current statute text.)
  27. U.S. Food and Drug Administration. Warning Letter to Strukmyer Medical, 2025. (Verify issuance date and citation against official FDA database at fda.gov/inspections-compliance-enforcement-and-criminal-investigations/warning-letters.)