Postdoctoral PDDM902 ⭐ 3 Credits Semester 1

Dermatological Materials Survey: Active Classes, Vehicles, Excipients, Preservatives, Antioxidants, and Supply-Chain Hazards

This course surveys the materials a dermatologist is likely to specify for nonsterile topical semisolids. It treats every material as a controlled input to a cGMP manufacturing operation under 21 CFR 211, not as a clinical choice. The course gives no human dosing or clinical-use guidance. Monograph concentration windows are taught only as regulatory limits that set a product's legal category. The course is API-agnostic and has four parts. (1) Evidence. Skipper weighs what is actually known about dermatologist-specified materials and states the limits of that evidence. The German statutory-insurance data (Salzmann 2020) and the Latvian prescription audit (Kiseļova 2020) are country-specific. No quantitative US ranking has been verified. The commonly cited US 503A 'staples' are narrative only (verify). (2) Legal-status map. Each common active is mapped to its legal route. The routes are: - an OTC monograph order (M003, M004, M005, M006, M016, M017, M020, M032); - an NDA or switch NDA; - Rx-only; - not legally marketable OTC (e.g. hydroquinone skin lighteners; ingredients listed as not GRASE under 21 CFR 310.545); - compounding-only (contrast only, because this operation is a manufacturer). This map is built as machine-checkable data. It extends the classification gate from PDDM901. (3) Vehicles and excipients by function, each with its documented failure modes: - paraben partitioning out of the aqueous phase; - polysorbate autoxidation to peroxides; - the pH dependence of sorbic acid (verify); - preservative inactivation by nonionic surfactants (verify); - contact sensitizers such as propylene glycol, parabens, chlorocresol, lanolin and benzyl alcohol; - the active/excipient boundary, where petrolatum, dimethicone, glycerin, mineral oil, lanolin, zinc oxide, benzyl alcohol, menthol and camphor can each be either an inactive or a monograph active, depending on concentration and claim. (4) Supply-chain hazards as incoming-material controls: - benzene from benzene-process carbomers (≤2 ppm, ICH Q3C Class 1, per the December 2023 FDA carbomer guidance); - a general module on actives or excipients that degrade to hazardous solvents (benzoyl peroxide is the current example); - DEG/EG limit testing on each shipment of each lot of high-risk components; - Burkholderia cepacia complex in aqueous raw materials. The course ends with an identity strategy that goes beyond the lab's chromatographic strength. 21 CFR 211.84(d) requires a specific identity test on every component lot, whether active or inactive. Oils, waxes, polymers, silicones and inorganics are largely not suited to HPLC-DAD. They need other qualified methods (verify against the current NF monographs), a purchase case, or contract-lab identity testing. A documented supplier-qualification program is the only legal basis for relying on CoAs for tests other than identity. The outputs feed directly into PDDM903 (formulation and plant envelope), PDDM912/913 (release testing and microbiology), PDDM922 (cleaning) and PDDM925 (tech-transfer gate).

📅 Studied by Skipper on 2026-10-16

📖 Core Concepts

Claude Opus 5.5

Core Concepts: Dermatological Materials Survey

PDDM902: Active Classes, Vehicles, Excipients, Preservatives, Antioxidants, and Supply-Chain Hazards

0. What This Subject Is Really About

In the PDQC program, the hard problems lived in the analyte. A peptide has a defined sequence. It also has a counter-ion, water content, D-isomers, aggregates, and deletion and insertion impurities, and the whole QC laboratory existed to measure those things. In a dermatological semisolid the active is often the simplest molecule in the jar. The risk has moved into everything around it:

  • the petrolatum, which is also a monograph active;
  • the glycerin, which is the material historically implicated in diethylene glycol poisonings;
  • the carbomer, whose older grades were polymerized in benzene;
  • the purified water, which is a natural habitat for Burkholderia cepacia complex (Bcc);
  • the polysorbate, which accumulates peroxides while it sits on the shelf.

This course is a materials survey with a legal, physical, and hazard lens applied to every row. The organizing claim of the course is:

Every material entering a topical drug product has three simultaneous identities: a legal identity (what role the monographs allow it, at what concentration, in what dosage form, with what claim); a physical identity (what it does in the vehicle, and what attribute of it controls that function); and a hazard identity (what it can carry in, what it can turn into, and what it can grow). A material is "controlled" only when all three are known, documented, and tested against.

We do not yet know the formulations, so the course is deliberately API-agnostic. We are building a classifier and a control system, not a product. The deliverables reflect that: a risk register, a monograph-window table, an incoming-material matrix, a preservative worksheet, and a supplier-qualification clause set. Each is a machine for turning an unknown dermatologist specification into a routing decision and a set of tests.

A useful analogy is airport baggage screening. The screener does not know what any passenger will pack. What exists is a set of rules: what is prohibited, what is allowed only below a size, and what is allowed only in checked luggage. There is a set of scanners, each blind to certain materials. There is a protocol for what happens when the scanner cannot see inside a bag. In that analogy:

  • the monographs are the rules;
  • HPLC-DAD and LC-MS/MS are the scanners;
  • NON-CHROM-ID materials are the bags the scanner cannot see into;
  • contract labs and supplier qualification are the secondary-screening protocol.

A note on epistemics carries through the whole course. Much of what circulates about dermatological compounding and excipients is narrative, regional, or second-hand. The program's confidence convention ([verified-primary], [verified-secondary], [unverified]) is not bureaucratic decoration. It is the mechanism that keeps a plausible-sounding claim from becoming a specification. I apply it throughout these notes.


The first conceptual shift is to stop asking "Is salicylic acid OTC?" That question has no answer. Legal status belongs to a tuple of attributes, not to a molecule:

Status = f( active, salt/form, concentration, dosage form, labeled claim,
            combination partners, monograph effective date )

  Status ∈ { monograph-conforming (M0xx) ,
             application-only (NDA / switch NDA) ,
             Rx-only ,
             not GRASE / disallowed (310.545, HQ DSC) ,
             out of scope for this site (cosmetic, compounding contrast) }

Each argument of this function can flip the output independently:

  • Concentration. Lidocaine (or lidocaine HCl) is an M017 external analgesic at 0.5–4%. Above 4% there is no monograph home, so it falls to the application or Rx route. Hydrocortisone at 0.25–1% sits in M017 as the anti-itch route. Above 1% it is Rx.
  • Claim. Salicylic acid is in M006 (acne) at 0.5–2% and in M032 (dandruff, seborrheic dermatitis, psoriasis) at 1.8–3% (M032 strengths are [unverified] in the regulatory brief; re-read before use). A 2% salicylic acid product falls inside both windows. The molecule and the concentration are identical, and the labeled claim decides which monograph governs. A 3% product is permissible only under the M032 claim.
  • Dosage form. Zinc oxide in M016 is permitted at 1–25%, or at >25–40% in ointment only. Moving a 35% zinc oxide formula from an ointment to a cream does not just change texture. It moves the product out of the monograph. M004 strengths are defined per base (for example, bacitracin 500 units/g in an ointment base). M017 excludes patches, plasters, and poultices. M006.50 requires the dosage form to be named.
  • Combination. Under M016.20(c), a skin protectant may be combined with "caine" anesthetics but not with hydrocortisone. Two individually legal ingredients can therefore form an illegal pair.
  • Time. In M020, bemotrizinol (≤6%) enters under final order OTC000039, with the current text effective 2026-08-09. OTC000008-1 removes PABA and trolamine salicylate effective 2027-09-11, or later if disputed. A formula that is compliant today can become non-compliant on a known date. The register must therefore track effective dates, not issue dates.

The surprising part is how much of a product's legal identity is carried by attributes a chemist would call "formulation details." Base type, salt form, and the exact verb on the label have legal weight. This is why Week 2 treats monographs as data. Each window becomes a record:

MonographWindow {
  active            : "zinc oxide"
  salt_form         : "—"
  min, max, unit    : 25, 40, "% w/w"
  permitted_forms   : ["ointment"]          ← the critical field
  permitted_combos  : per M016.20
  source            : M016 PDF, posting date
  effective_date    : …
  last_verified     : …   (by whom, against which PDF)
  confidence        : [verified-primary] | [verified-secondary] | [unverified]
}

This structure is the same master-data discipline as PDQC814. It is version-controlled, has defined change triggers (a monograph revision or a new final order), and is importable into the LIMS/ELN. The "switch-point table" built in Week 4 is the executable form of this function. At intake and at every formula change, it evaluates each active against its window and flags any tuple that sits on or across a boundary: hydrocortisone 1%, lidocaine 4%, salicylic acid 2% versus 3%, BPO 10%, and zinc oxide 25% versus 40%.

Status by absence. The Rx-only and application-only lists (tretinoin, tazarotene, azelaic acid, clindamycin, metronidazole, ketoconazole, terbinafine, butenafine, ciclopirox, adapalene, minoxidil, diclofenac, and others) are inferred from absence in the monographs [verified-secondary]. Absence is weaker evidence than presence. You can read a window off a page. You cannot read a prohibition off a blank. Some of these molecules reach OTC status through individual NDAs or switch NDAs, for example the adapalene/BPO combination under NDA 220736. That is an application pathway this site cannot use by copying a formula. Every "not in monograph" call should cite the monograph version checked.

Disallowed is a separate category. OTC hydroquinone skin lighteners are deemed new drugs and misbranded (effective 2020-09-23, FDA Drug Safety Communication). 21 CFR 310.545 lists ingredients that are not GRASE, including ammoniated mercury and any ingredient labeled for diaper rash in the relevant subsection. These differ from "Rx-only": no application route is available to this site at all.

Compounding is a contrast, not a pathway. The 503A category lists and the 503B bulks list include glycolic acid, kojic acid, TCA, GHK-Cu, DPCP, and SADBE. These describe what compounders may do. Professional-society positions such as the ASDSA statement supporting in-office preparation address compounding practice. None of this changes the PDDM901 conclusion: an operation that makes product for distribution under an OTC manufacturer permit is a manufacturer and is bound by 21 CFR 211 and the monographs.

Week 1 evidence and what it can support. The demand evidence fits this framework uneasily. The German data [Salzmann et al., 2020] report about 8 million compounded preparations per year, half prescribed by dermatologists, with triamcinolone, erythromycin, and clotrimazole topping the list. The Latvian data [Kiseļova et al., 2020] show salicylic acid, sulfur, and "prednisolone from tablets" among the most common requests. Both are country- and payer-specific and say little about US demand. The US literature is narrative. The structural lessons transfer even though the rankings do not:

  • Expect multi-active requests.
  • Expect requests built from finished dosage forms as API sources. A manufacturer cannot use these, because of component control under 211.84 (verify with counsel).
  • Expect most "US compounding staples" to fall outside monograph scope.

2. The Active/Excipient Boundary Is Drawn by Concentration Plus Claim

Many of the most ordinary excipients are also monograph actives. The DUAL-ROLE materials include:

  • M016: petrolatum, dimethicone, glycerin (20–45%), mineral oil, lanolin, zinc oxide, cocoa butter, hard fat;
  • M017: benzyl alcohol (10–33%), menthol (0.1–1%), camphor (0.1–3%), phenol, resorcinol;
  • M003: benzalkonium chloride (0.1–0.13%).

Whether one of these is an "active" depends on two conditions holding together:

Active role  ⇔  (concentration ∈ monograph window)  ∧  (label / marketing implies the monograph claim)

Excipient role  ⇔  ¬(above)   … but see failure modes; the interpretation is [unverified] and edge cases go to counsel

The failure modes are subtle because no single person does anything obviously wrong:

  1. The undeclared active. A cream contains 30% white petrolatum as an emollient, inside the M016 window of 30–100%. Marketing then writes "protects chapped skin." The product now arguably carries a skin-protectant active that does not appear as such in Drug Facts. The formulator made no change and the claim did the work.
  2. The preservative that became an analgesic. Benzyl alcohol at preservative levels is an excipient. If a formulator raises it for "better preservation" into the 10–33% range, it occupies the M017 active window. It also carries a SENSITIZER flag (ACDS Allergen of the Year 2026) and has to be accounted for in the M017 combination rules.
  3. Humectant drift. Glycerin used as a humectant at moderate levels can approach the 20–45% M016 window if a formulator increases it to address stickiness or drying.
  4. The vehicle-change trap. Changing vehicle family can change legal status even when every percentage stays the same (zinc oxide >25% in a cream).

For this reason, DUAL-ROLE is a register flag that triggers re-screening at every formula and labeling change. Change control here has to watch the label as closely as the batch record. That is a new idea compared with peptide QC, where labeling rarely altered the legal identity of an analyte.


3. The Vehicle Fixes the Dosage-Form Name and the Downstream Control Burden

Buhse et al. at FDA/CDER faced a practical problem: products labeled "cream," "lotion," "ointment," and "gel" had no consistent physical definitions. Their solution was a decision tree based on measurable attributes [Buhse et al., 2005]:

  • Water and volatiles (loss on drying) and composition separate ointments (essentially anhydrous, hydrocarbon or PEG continuous phase) from creams.
  • Rheology separates creams (semisolid, a yield stress, do not flow under their own weight) from lotions (pourable).
  • Composition and thermal behavior identify gels (a polymer network or colloidal structure in a liquid phase).

The material-level consequence is that the base composition determines the dosage-form name, and the dosage-form name is a regulated element of the label (M006.50; the ointment-only zinc oxide window; M004's per-base strengths). Vehicle choice is therefore partly a legal decision.

Our lab has no viscometer, rheometer, or KF titrator. The attributes that define the dosage form are not yet measurable in-house. This belongs in the gap analysis.

The course uses five vehicle families as its frame. Each has a characteristic control burden:

FamilyContinuous phaseWater / awPreservative needBcc exposureTypical hazard flags
Hydrocarbon ointmentPetrolatum / mineral oilAnhydrous; aw lowOften justifiable without preservative (verify via aw data)Low (raw material and process)DUAL-ROLE, NON-CHROM-ID, OXIDIZABLE (excipient lots)
Water-soluble (PEG) ointmentPEG blendHygroscopic; can take up waterCase-by-caseLow to moderatePEROXIDE-FORMER (PEG autoxidation, verify), NON-CHROM-ID
o/w creamWaterHighRequiredHighBCC/AQUEOUS, PEROXIDE-FORMER (polysorbates), DEGEG (humectants)
LotionWater (usually)HighRequiredHighAs cream, plus physical stability
Aqueous / hydroalcoholic gelWater ± ethanol / IPAHigh (alcohol may lower aw)Usually requiredHighBENZENE (carbomer), DEGEG (PG/glycerin), BCC/AQUEOUS

Ointment base taxonomy. The classical classes are:

  • Hydrocarbon: white or yellow petrolatum, mineral oil, hard fat.
  • Absorption: lanolin, hydrophilic petrolatum. These take up water to form w/o systems.
  • Water-removable: o/w bases that wash off.
  • Water-soluble: PEG.

This taxonomy is [unverified] against current USP <1151>, so compendial grades and NF monograph titles must be confirmed. The classes matter for this course because each implies different water uptake, different identity tests, and different micro risk.

Carbomer gels. Carbomer is a crosslinked poly(acrylic acid). The dry powder in water gives an acidic, low-viscosity dispersion of tightly coiled microgel particles. Neutralization with a base (NaOH, triethanolamine, aminomethyl propanol) ionizes the carboxyl groups:

–COOH  +  B   ⇌   –COO⁻  +  BH⁺

Fraction ionized:  α = 1 / (1 + 10^(pKa,app − pH))

Electrostatic repulsion between fixed carboxylates swells each microgel until the particles jam against one another. The macroscopic result is a pseudoplastic, yield-stress material: it holds its shape at rest and thins under shear. Islam et al. characterized carbomer gels neutralized across a pH range and found the rheology relatively stable between about pH 5.0 and 8.0 [Islam et al., 2004]. That plateau is the formulator's working window.

The apparent pKa of a polyelectrolyte shifts with the degree of ionization, because each added charge makes the next proton harder to remove. Treat any single pKa value as an approximation (verify for the grade). Two practical consequences follow:

  • An acidic active (salicylic acid, for example) consumes neutralizer and can pull the gel below the plateau.
  • Electrolytes and cationic species screen the charges and collapse the network.

Both belong in the PDDM903 compatibility matrix.

Proprietary bases. Commercial compounding bases (PCCA, Medisca, or Fagron-type) commonly do not disclose full quantitative composition [unverified]. A cGMP manufacturer must run an identity test on every component under 211.84(d) and must list inactive ingredients in Drug Facts. Neither is possible for sub-components it does not know exist (inference). The disposition rule is binary: obtain the full quantitative formula under a confidentiality agreement, or reject the base.

Critical material attributes (CMAs). Chang et al. framed topical generic development around the idea that "Q1/Q2 sameness" (same ingredients, same amounts) does not guarantee "Q3 sameness" (same microstructure). Microstructure depends on material grade and process [Chang et al., 2013]. For each base material we therefore record the attribute that controls its function:

  • melting or congealing range (petrolatum, waxes, fatty alcohols);
  • viscosity grade (mineral oil, dimethicone);
  • polymer grade or molecular weight (carbomer type, HPMC substitution, PEG MW);
  • emulsifier HLB system;
  • particle size of dispersed solids (zinc oxide, titanium dioxide).

Two lots of "white petrolatum, USP" can both pass the monograph and still behave differently in a cream. The CMA list is handed off to PDDM903.


4. Excipients Fail in Documented, Predictable Ways

"Inactive ingredient" is a labeling term, not a chemical one. Excipients have their own reactivity, impurities, and shelf lives. This course catalogs them by function, as emulsifier systems, emollients, humectants, and thickeners (ranked usage lists [unverified]). The more important catalog is of failure modes.

Polysorbate autoxidation (PEROXIDE-FORMER). Polysorbates (and other polyoxyethylene surfactants and PEGs) contain ether linkages whose α-carbon hydrogens are abstractable. Polysorbate 80 also carries an unsaturated oleate chain. The radical chain mechanism is the familiar one from organic chemistry:

Initiation:     RH  →  R•            (light, heat, trace metal)
Propagation:    R•  +  O₂  →  ROO•
                ROO• + RH  →  ROOH + R•        ← hydroperoxides accumulate on the shelf
Metal cycling:  Fe²⁺ + ROOH  →  Fe³⁺ + RO• + OH⁻      (Fenton-type)
                Fe²⁺ + H₂O₂  →  Fe³⁺ + •OH + OH⁻
Consequence:    ROOH / RO• / •OH  oxidize the active, other excipients, or the surfactant itself
                (chain scission → aldehydes, acids, short PEG fragments)

A drum of polysorbate is therefore not a stable reagent. Its peroxide content rises with age, headspace oxygen, temperature, and the number of times it has been opened. The controls follow from the mechanism:

  • a peroxide-value specification on incoming lots (method per the NF monograph; verify);
  • supplier lot-age requirements;
  • nitrogen-blanketed storage;
  • a limit on opened-container hold time;
  • chelators to break the metal cycle.

Humectants as dual hazards. Glycerin, propylene glycol, and sorbitol solution carry two hazard identities:

  • DEGEG. They are the classic adulteration targets for diethylene and ethylene glycol (Concept 8).
  • BCC/AQUEOUS. As aqueous solutions, or as hygroscopic liquids that absorb water, they can carry Bcc.

It is easy to fixate on the active and the preservative and miss that one of the most hazardous-by-history materials in the formula is the one sweetening its feel.

Material-level compatibility screens. These feed the PDDM903 matrix:

  • Cationic with anionic. Benzalkonium chloride or other quaternary ammoniums with carbomer, anionic emulsifiers, or xanthan. Expect ion-pair or complex formation, loss of free preservative, and possible loss of viscosity (verify for specific pairs).
  • Acidic actives with carbomer neutralization. See Concept 3.
  • Metal-ion carryover. Mineral excipients (zinc oxide, talc, kaolin, iron-containing pigments) can supply the Fe or Cu that catalyzes the chemistry above. This links to ICH Q3D(R2), Appendix 5, cutaneous PDEs (PDQC813).

Grade policy. Drug manufacture expects USP–NF grade. The legal linkage via FD&C Act §501(b) (adulteration when a compendial article differs from compendial standards) is [unverified] as applied here, so confirm with counsel. The policy rule does not wait on that answer: no cosmetic-grade substitution without a documented equivalence assessment. Cosmetic-grade glycerin, for example, carries no assurance of USP DEG/EG testing.


5. A Preservative Only Works Where the Microbes Are: Free, Aqueous, and in the Active Form

This may be the most beautiful idea in the course. The concentration printed in the formula (the "nominal" concentration) is almost never the concentration that kills organisms. Microbes live and divide in the aqueous phase. A preservative molecule dissolved in the oil phase, bound inside a surfactant micelle, adsorbed at the interface, ionized, or complexed with a polymer is not doing antimicrobial work.

(a) Partitioning in an emulsion. Take a simple two-phase system with oil:water volume ratio φ and preservative oil/water partition coefficient Ko/w. Mass balance gives the classical relation, a Bean-type equation; verify the exact form and assumptions before using it in a worksheet:

                C_total · (φ + 1)
   C_w  =  ─────────────────────
               K_o/w · φ  +  R

   C_w      = free preservative concentration in the aqueous phase
   C_total  = nominal (overall) concentration
   φ        = oil : water volume ratio
   K_o/w    = oil/water partition coefficient
   R        = (total aqueous preservative) / (free aqueous preservative)
              ≈ 1 with no surfactant; > 1 when micelles / polymers bind preservative

Worked illustration (hypothetical K values chosen for teaching, not literature constants): 20% oil phase gives φ = 0.25; Ctotal = 0.20%.

  • A hydrophilic paraben-like preservative, K = 3, no surfactant binding (R = 1):
    Cw = 0.20 × 1.25 / (0.75 + 1) = 0.143%, about 71% of nominal free in water.
  • The same preservative with micellar binding, R = 3:
    Cw = 0.25 / (0.75 + 3) = 0.067%, about 33% of nominal.
  • A lipophilic paraben-like preservative, K = 80, R = 1:
    Cw = 0.25 / (20 + 1) = 0.012%, about 6% of nominal.

Compare the lecithin-emulsion data cited in this course. About 17–31% of methylparaben and 2.3–6% of propylparaben were found free in water, with 38–58% at the interface. The longer the alkyl chain, the more the preservative leaves the water. Propylparaben is more potent per molecule than methylparaben, but in an emulsion much of it is somewhere the microbes are not. Binding to polysorbates and other nonionics is plausible (verify), and that is what the R term represents. The same emulsifier that builds the cream can disable its preservative.

The analytical consequence: a total-preservative assay at release (HPLC-DAD on the 1260s) is necessary but not sufficient. A water-phase versus total partition study, for example by centrifugal phase separation followed by HPLC of the aqueous layer, tells you what fraction is free. This is an inference, and the method has to be developed.

(b) Ionization: weak-acid preservatives. Sorbic acid and benzoic acid are active mainly in the undissociated form, which crosses microbial membranes. The fraction undissociated follows Henderson–Hasselbalch:

                      1
   f_HA  =  ──────────────────────
             1 + 10^(pH − pKa)

   Sorbic acid, pKa ≈ 4.8 (verify):
     pH 4.5 →  f_HA = 1/(1 + 10^−0.3) ≈ 0.67
     pH 5.5 →  f_HA = 1/(1 + 10^0.7)  ≈ 0.17
     pH 6.5 →  f_HA = 1/(1 + 10^1.7)  ≈ 0.02

Moving from pH 4.5 to 6.5 cuts the active fraction about thirty-fold even though the assay value is unchanged. Two lessons follow:

  • A pH specification is effectively a preservative-efficacy specification for weak-acid systems.
  • The lab currently owns no pH meter. That is a real gap, not a convenience gap.

Sorbate is also oxidation-prone, which earns an OXIDIZABLE flag, and is a reported sensitizer (verify).

(c) Combining both effects. For a weak-acid preservative in an emulsion, the effective concentration is approximately Cw × fHA. Both factors can be small at once. Nominal concentration therefore overstates effective concentration multiplicatively.

(d) How efficacy is judged. USP <51> challenges the product with defined organisms and counts survivors over time:

   Log reduction (LR)  =  log₁₀(N₀) − log₁₀(N_t)

   Category 2 (aqueous-based topicals, emulsions)  [secondary source: FDA ORA.007; verify current <51>]:
     Bacteria:        LR ≥ 2.0 from initial count at 14 d;  no increase from 14 d to 28 d
     Yeasts & molds:  no increase from initial count at 14 d and 28 d

   "No increase" is typically interpreted as not more than 0.5 log above the reference value (verify).

Our BacT/ALERT is a presence/absence system. It cannot produce Nt, so it cannot compute a log reduction. <51> is contract work until PDDM913 decides otherwise.

(e) The preservative roster as a set of trade-offs.

  • Parabens. Widely used, with low sensitization rates in patch-test populations. This is the "paraben paradox" of public perception versus allergological data [Cashman & Warshaw, 2005]. They are UV-active, which makes HPLC-DAD a good fit, and they partition strongly as chain length increases.
  • Phenoxyethanol. The EU 1% cosmetic limit and the SCCS discussion of 0.4% for children under 3 are EU cosmetic rules. They are not US drug limits, but they are relevant to marketing perception and export.
  • Benzyl alcohol. A preservative, the ACDS 2026 Allergen of the Year, and an M017 active at 10–33%. It carries three flags at once.
  • Chlorocresol. Present in 11.4% of Canadian topical corticosteroids [Labadie et al., 2019], with contact urticaria reported.
  • Benzalkonium chloride. A cationic quaternary; anionic incompatibility (verify). An M003 active at 0.1–0.13%.
  • Disodium EDTA. Not a preservative by itself, but a booster. It chelates the divalent cations that stabilize Gram-negative outer membranes, increasing kill and suppressing persister subpopulations [verified-secondary].

(f) Identity specificity for preservatives. On HPLC-DAD, identity must be retention time plus spectral match. Retention time alone is not specific. Methyl- and ethylparaben elute close together and share a chromophore. This is the same principle as the USP <3> identity-specificity expectation and the peptide identity logic of PDQC803.


6. Water Activity Determines What Can Grow, and Preservatives Do Not Replace Process Control

Water content (KF, loss on drying) tells you how much water is present. Water activity tells you how much of it is available to an organism:

   a_w  =  p / p₀       (vapor pressure over sample / vapor pressure of pure water, same T)

   USP <1112> growth thresholds  [secondary; verify current text]:
     P. aeruginosa, E. coli, Salmonella   — no proliferation below a_w ≈ 0.91
     S. aureus                            — not below ≈ 0.86
     A. niger (molds)                     — not below ≈ 0.77
     No microbial proliferation           — below ≈ 0.60

A hydrocarbon ointment with no free water phase sits far below these thresholds. That is the scientific basis for justifying an unpreserved anhydrous ointment and, potentially, reduced microbial testing. The justification needs measured aw data, and we own no aw meter (contract lab or purchase). Note what aw controls and what it does not:

  • It controls proliferation.
  • It does not control survival of organisms introduced by raw materials or equipment.
  • It does not account for water uptake in hygroscopic bases such as PEG. Water uptake in a PEG ointment is a stability question for PDDM924.

The second principle is FDA's explicit position on Bcc: preservatives alone must not be relied on to control it. Tavares et al. review why [Tavares et al., 2020]:

  • Bcc survives in oligotrophic (nutrient-poor) water, including purified-water systems.
  • It forms biofilms.
  • It possesses efflux systems and metabolic versatility that let it resist, and in some cases metabolize, antimicrobials used as preservatives.

A preservative system is a backstop against in-use contamination, not a way to sterilize a contaminated batch. The analogy is a seatbelt: necessary, tested (<51>), and absurd to rely on in place of brakes. Raw-material control and water-system control are the brakes.

The preservative-selection worksheet. The worksheet (Week 8) formalizes all of this. Its inputs are:

  • vehicle family;
  • pH range, which sets fHA;
  • surfactant type and load, which sets R;
  • oil:water ratio, which sets φ;
  • packaging (jar versus tube versus airless pump), which sets consumer contamination exposure;
  • sensitizer flags;
  • label constraints;
  • monograph conflicts (benzyl alcohol as an active).

Its outputs are commitments: a preservative assay at release and on stability (PDDM912/924), and AET at the lowest specified preservative level (PDDM913; verify approach). If the preservative content is allowed to decline on stability to a lower limit, efficacy must be demonstrated at that limit, not at the target.


7. Antioxidant Systems Must Be Proven, Not Assumed

Antioxidants come in phase-specific classes [unverified as a ranked list]:

  • Lipid-phase radical scavengers: BHT, BHA, tocopherol.
  • Aqueous-phase reducing agents: sodium metabisulfite/sulfite, ascorbic acid.
  • Chelators: disodium EDTA.

The phase-matching logic is the same as for preservatives. An antioxidant protects only the phase it is in. BHT in the oil phase does not protect an aqueous-phase active.

The counterintuitive part is that antioxidants are reactive molecules with their own degradation chemistry. Hydroquinone serves as a mechanism case only, since it is disallowed OTC. Hydroquinone autoxidizes to p-benzoquinone, a colored and reactive species. Matsubayashi et al. reported the following for hydroquinone formulations [Matsubayashi et al., 2002]:

  • With ascorbic acid plus sulfite present, color change still appeared within 3 months and was suppressed at 4 °C.
  • Removing the antioxidants raised p-benzoquinone.
  • Acidification helped.

The antioxidants reduced the toxicologically relevant degradant but did not prevent visible change. My interpretation (verify) is that the antioxidants themselves, ascorbate in particular, participate in browning chemistry. The lessons are:

  • Appearance and chemical purity can diverge.
  • Temperature and pH are often stronger levers than additives.
  • Every antioxidant system needs experimental justification, typically a stability comparison with and without it (PDDM924).

Oxidizer/oxidizable pairs. Some formulas put an oxidizer next to something oxidizable on purpose:

  • Benzoyl peroxide (OXIDIZER) is a peroxide by design. Its incompatibility with oxidizable actives, such as tretinoin degradation in BPO's presence [secondary, PMC], is the textbook example.
  • A peroxide-laden polysorbate lot plays the same role inadvertently. It is a PEROXIDE-FORMER introducing oxidizer into a formula whose author believes it contains none.

The controls are inert-gas blanketing, chelators, peroxide-controlled surfactant grades, and separation of incompatible pairs.

Photolability. Tretinoin is an Rx example used here only to define the PHOTOLABILE flag. In one study, a tretinoin lotion lost about 80% of drug within 30 minutes under xenon light, with maximum damage near 420 nm, in the visible range. "Store away from UV" is therefore insufficient. The flag triggers:

  • actinic-light handling (yellow light or amber glassware) in the lab;
  • opaque or amber packaging;
  • ICH Q1B photostability studies (PDDM924).

The lab also needs amber HPLC vials and covered autosampler trays to keep the assay itself from creating degradants.

What the lab can and cannot measure.

  • Can, in-house: BHT and other UV-active antioxidants on HPLC-DAD.
  • Feasibility hypothesis only: screening polysorbate oxidation products or peroxide surrogates by LC-MS/MS. This needs method development and is not a compendial method.
  • Likely other techniques or outsourcing (verify): sulfite and EDTA, which have weak or no useful chromophores.

8. Supply-Chain Hazards Hide in the Most Boring Materials

Three hazards dominate recent FDA attention to topicals. They share one pattern: the hazard is invisible to the tests that ordinary identity and assay methods perform. It must be designed for specifically.

(a) Benzene (BENZENE flag). FDA's benzene alert identifies carbomers, isobutane and other hydrocarbon-derived components, and degradation of certain actives as sources. It expects specifications and testing under 21 CFR 211.84, 211.160, and 211.165, plus component-level benzene risk assessments. The December 2023 carbomer guidance asked USP to omit the Carbomer 934P, 940, 934, 1342, and 941 monographs, which cover benzene-process grades. Manufacturers should use benzene-free grades meeting ≤2 ppm benzene (ICH Q3C Class 1). The guidance applies to topical creams, gels, lotions, and ointments, including OTC monograph drugs.

A mass-balance view shows why the carbomer source matters, and why incoming control cannot fix every case:

   C_benzene,product  ≈  Σ_i  ( w_i · C_benzene,i )   +   C_formed(t, T)

   Example A (benzene-free grade):  w = 0.005 (0.5% carbomer), C = 2 ppm
        contribution = 0.005 × 2 ppm = 0.01 ppm

   Example B (legacy benzene-process grade, residual benzene allowance historically up to ~0.5%, i.e. ~5000 ppm — verify):
        contribution = 0.005 × 5000 ppm = 25 ppm   → >10× the 2 ppm Class 1 limit

   C_formed(t, T):  benzene generated in situ by degradation (e.g., BPO); NOT controlled by incoming testing

The second term is the lesson from BPO. FDA tested 95 BPO acne products (statement of 2025-03-11) and found more than 90% had undetectable or very low benzene, though retail-level recalls followed for some. The temperature-dependence mechanism is [unverified] in our sources. The generic procedure does not depend on the mechanism:

  1. Identify any active or excipient with a plausible degradation path to a Class 1 solvent.
  2. Set incoming and stability benzene testing.
  3. Define storage and shipping temperature controls (PDDM924).

A benzene-free carbomer certificate controls the first term of the equation and leaves the second untouched.

We have no GC-HS, so benzene testing is outsourced to a qualified contract lab (method basis <467>/Q3C approach; verify specifics). If the unidentified second MS turns out to be a GC-MS, it still requires a PDQC805 feasibility and qualification decision before the lab can claim any in-house capability.

(b) DEG/EG (DEGEG flag). FDA's May 2023 guidance (Rev. 1) requires USP–NF DEG/EG limit testing (NMT 0.10%) for high-risk components: glycerin, propylene glycol, sorbitol solution, maltitol, hydrogenated starch hydrolysate, and others. The testing is performed on each shipment of each lot, as part of identity testing, for any dosage form including topicals. Per-shipment is the important detail. Adulteration and mix-ups can occur in the distribution chain after the manufacturer's CoA was issued, so the CoA of the original lot cannot vouch for the drum that arrived.

   Tests required per material per year  =  Σ_lots  (number of shipments of that lot)

   e.g., glycerin lot A arrives in 3 shipments, lot B in 2  →  5 DEG/EG tests, not 2
   Material lead time  =  supplier lead time  +  contract-lab turnaround (DEG/EG)  +  QA review
   Release rule:  quarantine → release only on a passing result for THAT shipment

The compendial method is GC-based (verify). Contract-lab turnaround therefore becomes part of the supply chain's critical path and of inventory cost, which feeds the Week 13 purchase-versus-outsource analysis.

(c) Burkholderia cepacia complex (BCC/AQUEOUS flag). Bcc is an objectionable organism in non-sterile aqueous products. FDA's 2021 draft non-sterile microbiology guidance and its 2023 case study cite USP <60> and <1111>, and 21 CFR 211.84(d)(6), 211.113(a), and 211.165(b). Bcc survives in oligotrophic water and resists preservatives [Tavares et al., 2020].

Raw-material risk ranking (inference; verify against supplier data), from highest to lowest:

  1. aqueous solutions;
  2. natural or botanical aqueous extracts;
  3. water-containing premixes;
  4. anhydrous oils and waxes.

Test assignments follow the ranking: <60> testing, supplier water-system evidence, and aw data. The Strukmyer (2025) warning letter is the case study. Burkholderia and Pseudomonas were recovered from the firm's water system, which shows that water is a material requiring control like any other.

Lab reality. BacT/ALERT is a qualitative, sterility-oriented growth detector. It cannot perform <61> enumeration or <60>/<62> detection and identification of specified organisms without alternative-method validation under <1223>. All of these stay with the contract lab until PDDM913 decides otherwise.


9. Identity Beyond Chromatography, and the Limits of Trusting a CoA

21 CFR 211.84(d) requires at least one specific identity test on each lot of each component, active or inactive. A supplier CoA can substitute for other tests only if the manufacturer periodically validates the supplier's results. Missing identity testing is the most-cited finding in the 2025 OTC topical warning letters (Columbia Cosmetics, Libby, Strukmyer).

Our instruments are excellent for UV-active organics: preservatives, many actives, and BHT, by HPLC-DAD retention time plus spectral match, with LC-MS/MS transitions for added specificity. They are poorly suited to much of a dermatological formula. The materials that carry the NON-CHROM-ID flag are the ones that are most of the jar by mass:

  • petrolatum, mineral oil, and waxes (no chromophore, complex hydrocarbon mixtures);
  • fatty alcohols;
  • dimethicone;
  • carbomer and cellulose polymers;
  • glycerin;
  • zinc oxide, titanium dioxide, talc, and kaolin (inorganic).

The candidate tools must each be verified against the current NF/USP monograph for each material:

  • FTIR/ATR against a reference spectrum;
  • physical constants: refractive index, specific gravity, melting or congealing range, viscosity;
  • fat and wax indices: acid, saponification, and iodine values;
  • wet-chemical and ion tests for inorganics.

None of these instruments is in the inventory. The Week 13 memo frames this as an economic decision. Annual identity-test volume, multiplied by contract-lab turnaround and quarantine inventory cost, is weighed against capital and PDQC805 qualification burden. An explicit buy trigger closes the analysis. The working hypothesis, to be verified for coverage, is that a single FTIR-ATR would handle most organic excipients.

Supplier qualification. This extends the PDQC806 framework with hazard-specific quality-agreement clauses:

  • benzene-free process attestation and a per-lot benzene result on the CoA for carbomers and propellants;
  • DEG/EG results with traceability to shipment for high-risk components;
  • water-system and Bcc evidence for aqueous raw materials;
  • change notification covering process, site, and sub-supplier changes;
  • confirmatory full-monograph testing at a defined frequency;
  • CoA-versus-in-house trending;
  • requalification triggers and disqualification criteria.

CoA reliance is a privilege earned by data. It is not a convenience.


10. The Risk-Flag Taxonomy: Compressing All of This Into a Row

The register turns everything above into one row per material:

MaterialRow {
  name, grade, compendial monograph (verify title)
  flags        : { LEGAL, DUAL-ROLE, BENZENE, DEGEG, BCC/AQUEOUS, OXIDIZABLE,
                   PEROXIDE-FORMER, PHOTOLABILE, OXIDIZER, SENSITIZER,
                   POTENT/HD (verify), CR-PACK, NON-CHROM-ID }
  confidence   : per flag
  source + date: per flag
  incoming tests required
  owner course : PDDM903 / 912 / 913 / 922 / 923 / 924
  review date
}

Glycerin illustrates the value. It is a humble humectant, and its row carries DUAL-ROLE (M016 20–45%), DEGEG (per-shipment testing), BCC/AQUEOUS (aqueous-risk raw material), and NON-CHROM-ID (no chromophore). A single row tells purchasing, QC, regulatory, and micro what each owes the material.

The SENSITIZER flag. Labadie et al. surveyed 140 Canadian topical corticosteroids [Labadie et al., 2019] and found:

  • propylene glycol in 42.9%;
  • parabens in 27.9%;
  • chlorocresol in 11.4%;
  • formaldehyde releasers in 7.1%.

For a manufacturer, the consequence is not clinical selection, which is outside our scope. It is labeling accuracy and change control of inactive ingredients. Patients and dermatologists use Drug Facts inactive lists to avoid allergens, so an unannounced excipient swap is a real harm.

The CR-PACK flag. This one is quantitative and easy to underestimate. Under 16 CFR 1700.14, lidocaine above 5.0 mg per package triggers child-resistant packaging:

   30 g tube × 4% w/w lidocaine  =  30 g × 0.04  =  1.2 g  =  1200 mg   ≫ 5.0 mg  → CR-PACK
   Even 0.5% in a 5 g tube      =  25 mg                                  > 5.0 mg  → CR-PACK

Essentially every OTC lidocaine product would trigger the threshold. Dibucaine (>0.5 mg), minoxidil (>14 mg), and methyl salicylate (>5% in liquids) are similar. Whether a semisolid counts as a "liquid" for the methyl salicylate entry needs CPSC confirmation. The flag ties a material's presence to a packaging-component specification.

POTENT/HD. Corticosteroids, retinoids, fluorouracil, imiquimod, and antiandrogens are mostly out of OTC scope. If they are ever handled as APIs, they are flagged for containment, dedicated or campaign equipment, and cleaning limits (PDDM922). NIOSH classification is [unverified]. DPCP and SADBE are deliberate sensitizers on the 503B list and outside OTC scope.


Connections to Peptide Chemistry

Much of this course is the PDQC peptide-QC program seen through a different window. The parallels are worth making explicit, because they show which instincts transfer and which must be re-tuned.

  • Salt/form accounting reappears as monograph salt/form fields. In PDQC812, peptide content had to account for TFA or acetate counter-ion and water. Monograph windows have the same structure: lidocaine versus lidocaine HCl, hydrocortisone versus hydrocortisone acetate, miconazole nitrate. The "unit basis" field in the window table does the job of the peptide-content calculation. Get the basis wrong and a compliant formula is miscalculated out of its window.
  • Polysorbate peroxides are already familiar from biologics. Methionine and tryptophan oxidation in protein and peptide formulations containing polysorbate 80 is the same Fenton and hydroperoxide chemistry as in Concept 4. Peroxide-value control of surfactant lots, nitrogen blanketing, and chelators came straight out of biologics development. If the platform ever handles a peptide in a topical vehicle (GHK-Cu appears on the 503B list as a compounding contrast; its OTC status is a LEGAL question for counsel, not a formulation question), the PEROXIDE-FORMER × OXIDIZABLE pairing would be the first compatibility concern. The Cu²⁺ in GHK-Cu is itself a Fenton-competent metal.
  • The identity-specificity principle carries over directly. PDQC803 established that retention time alone is not identity for a peptide, which needs orthogonal MS confirmation (PDQC811). The same logic applies to parabens on HPLC-DAD and to USP <3>. What differs is that peptides are rich in specific signals: sequence, mass, fragmentation. Petrolatum is poor in them. The NON-CHROM-ID gap is a new problem the peptide lab never faced.
  • Q3C Class 1 and Q3D carry over. PDQC813 built residual-solvent and elemental-impurity risk assessments for peptide APIs, where solvents from SPPS (DMF, DCM, acetonitrile, ether) dominated. Here the same Q3C framework covers benzene from carbomer and BPO. Q3D(R2) Appendix 5 supplies cutaneous PDEs, which differ from parenteral PDEs, for metal carryover from mineral excipients.
  • The microbiology gap is defined by the BacT/ALERT. In PDQC822/892 we learned what the BacT/ALERT can and cannot do. The non-sterile topical world asks for counts (<61>), specified-organism identification (<62>, <60>), and log reductions (<51>). None is a presence/absence question. The <1223> alternative-method validation logic from PDQC892 is the only route to repurposing it.
  • Bcc is the endotoxin analogue for non-sterile products. In parenteral peptide QC, Gram-negative contamination showed up as endotoxin (PDQC821). In non-sterile topicals, the Gram-negative concern is viable Bcc in aqueous products. It is the same taxonomic neighborhood with a different regulatory expression of the hazard.
  • Incoming-material control is the PDQC806 framework with new branches. The peptide lab qualified amino-acid derivatives, resins, and solvents. The topical lab extends the same incoming framework with per-shipment DEG/EG, benzene attestations, and Bcc ranking.
  • The LC-MS/MS gets new jobs. The triple-quad that sequenced peptides becomes a candidate tool for polysorbate degradation-product screening and for added-specificity identity of small-molecule actives. These are feasibility hypotheses (verify), not compendial methods.

The deepest connection is cultural. The peptide program trained a habit of asking "what is actually in this vial, and how do I know?" This course applies that habit to materials that much of the industry has historically treated as commodities, such as glycerin and carbomer. The documented poisonings and recalls linked to DEG and benzene show what happens when that question goes unasked.


Common Misconceptions to Watch Out For

  1. "Inactive ingredients are inactive." Several are monograph actives at the right concentration and claim (petrolatum, glycerin, benzyl alcohol, zinc oxide, benzalkonium chloride). Many are reactive, peroxide-forming, contaminant-carrying, or sensitizing.
  2. "If a molecule is in an OTC monograph, my product with it is OTC." Status depends on the full tuple: concentration, salt form, dosage form, claim, combination, and effective date. Zinc oxide at 35% in a cream, or hydrocortisone added to a skin protectant, falls outside the monograph despite every molecule being "OTC."
  3. "Not in a monograph means not regulated" or "means compoundable." Absence routes to NDA or Rx, or to disallowed status. Compounding allowances (503A/503B) and professional-society positions do not apply to a manufacturer.
  4. "The preservative concentration in the formula is the effective concentration." Partitioning into oil, binding to micelles and polymers, and ionization can each cut the free, active fraction by large factors, and the effects multiply.
  5. "A strong preservative system controls Bcc." FDA explicitly rejects reliance on preservatives alone. Bcc can resist and metabolize some preservatives. Water and raw-material control carry the load.
  6. "Anhydrous means no microbiology concern." Low aw prevents proliferation, not survival. PEG bases take up water. Unpreserved status must be justified with measured aw data and a documented rationale.
  7. "More antioxidant is always better." Antioxidants degrade, can color the product, and can interact with preservatives. The Matsubayashi hydroquinone case shows that systems must be justified experimentally.
  8. "Benzene-free carbomer solves the benzene problem." It controls only the incoming contribution. Degradation-generated benzene (the BPO case) requires stability testing and temperature control.
  9. "DEG/EG testing is per lot." It is per shipment of each lot.
  10. "A CoA is an identity test." 211.84(d) requires at least one in-house specific identity test per lot. CoA reliance for other tests requires periodic validation of supplier results.
  11. "Retention time confirms identity." Not alone. Use RT plus DAD spectral match at minimum, and MS where needed.
  12. "The BacT/ALERT can cover our micro testing." It is presence/absence only. It cannot do <61> counts, <62>/<60> identification, or <51> log reductions without <1223> validation.
  13. "EU limits are our limits." The EU phenoxyethanol cosmetic limits are EU cosmetic rules, not US drug limits. They are relevant context, not specifications.
  14. "A proprietary base is just another excipient." Without full quantitative composition, identity testing and Drug Facts labeling are impossible. Obtain the formula under a confidentiality agreement or reject the base.
  15. "Labeling is marketing's business." A claim can create an undeclared active, and an inactive-ingredient change can create an allergen hazard. Labeling sits inside change control.
  16. "A confident-sounding source is a verified source." The M004 and M032 discrepancies between the regulatory and materials briefs show that two internal documents can disagree about confidence. Resolve such discrepancies against the current official PDF and record the date checked before any number enters the data table.

Open verification items carried from these notes:

  • exact form and assumptions of the partitioning equation;
  • sorbic acid pKa and the sensitizer claim;
  • <51> "no increase" interpretation and current <1112> thresholds;
  • legacy carbomer benzene allowance;
  • <1151> base taxonomy;
  • NF peroxide-value method for polysorbate;
  • §501(b) linkage to grade policy;
  • DEG/EG compendial method basis;
  • M004 and M032 strengths against current PDFs;
  • CPSC "liquid" interpretation for methyl salicylate;
  • NIOSH classification for POTENT/HD;
  • Florida §499.01(2)(n) reach to Rx bulk APIs;
  • FTIR-ATR coverage of NF identity tests.

Each one goes into the Week 14 handoff package with an owner and a deadline.

⚙️ Mechanisms & Applications

GPT-6 Astra

Mechanisms & Applications

Skipper’s central insight: a dermatological semisolid is not simply an active ingredient diluted into a convenient carrier. It is a structured, evolving chemical system. Its phases determine where molecules reside; its interfaces determine what is bound or available; its processing history determines its microstructure; and its packaging determines how that system changes during storage and use.

My peptide-QC training taught me to distinguish identity, content, purity, and potency. Here I must add another distinction: the measured amount of an ingredient is not necessarily its functionally available amount. Total preservative is not free aqueous preservative. Total active is not dissolved active. Total water is not water activity. These distinctions explain many otherwise puzzling failures.

Evidence convention: [verified-primary] and [verified-secondary] describe the evidence supporting a specific claim, not an entire document. Numerical monograph windows, effective dates, compendial procedures, and legal interpretations marked (verify) require review against current official text before operational use. This section does not certify the current status of any product or regulatory order.

1. From a dermatologist’s specification to a controlled material system

The first mechanism is organizational rather than molecular: translating an incoming request into a legally and scientifically defined product. This resembles assembling a biochemical pathway. Every individual reaction may be feasible, yet the pathway fails if one connection is prohibited or unsupported.

  1. Separate the request into independent fields. Record the proposed active, chemical form, strength and unit basis, intended claims, dosage form, other ingredients, combinations, and package.
  2. Establish the regulatory route. Determine whether the exact product can fit an applicable OTC monograph, requires an approved application, or is outside the operation’s authorized scope. A dermatologist’s request does not create a manufacturing exemption.
  3. Check the whole combination. Two individually permitted actives do not automatically form a permitted combination. The same applies to an active paired with a different dosage form or claim.
  4. Resolve ingredient roles. Materials such as petrolatum, dimethicone, benzyl alcohol, and glycerin can have different regulatory roles in different products. Intended use, claims, concentration, and applicable conditions must be considered together.
  5. Translate materials into control obligations. An aqueous premix creates a microbial-risk question; a humectant may create a DEG/EG-testing obligation; a carbomer creates grade, residual-solvent, and rheology questions.
  6. Release the specification for development only after unresolved gates have owners. Unknown legal status is a hold condition, not permission to formulate.
Intake acceptance =
    supported regulatory route
AND permitted active/form/strength
AND permitted claim, dosage form, and combination
AND suitable component-control strategy
AND feasible manufacturing and testing capability

Why a range check alone fails: salicylic acid at a particular concentration cannot be classified without its claim and applicable monograph. Similarly, zinc oxide above the lower skin-protectant concentration window introduces a dosage-form restriction. Encode these as conditional rules, not as one universal “allowed concentration” column. The exact thresholds and conditions must be transcribed from the current M006, M016, and M032 texts.

A useful database record therefore includes source version, posting date, effective date, verification date, and reviewer. These are different fields. A newly issued order can have a future effective date. In particular, the sunscreen-order information supplied for this course must be checked against the current M020 administrative-order record before use.

Application to Peptide Partners: the platform should remain API-agnostic but not control-agnostic. It can standardize the decision process before it knows which formulations will be manufactured. The regulatory foundation is FD&C Act §505G and 21 CFR 330.1; component controls subsequently operate under 21 CFR 211.84. Absence from one monograph alone does not establish an ingredient’s complete legal status.

2. Vehicles control chemical potential, diffusion, and microstructure

2.1 Dissolved concentration is not the same as the driving force for release

In physical chemistry, diffusion responds fundamentally to a chemical-potential gradient. Concentration is a convenient surrogate only under appropriate assumptions.

Chemical potential:
μ = μ° + RT ln(a)

For a dissolved species:
a = γ(C/C°)

Simplified steady-state diffusion:
J = −D(dC/dx)

Simplified membrane model:
J ≈ (DK/h)(Cvehicle − Creceiver*)

D = diffusion coefficient
K = membrane/vehicle partition coefficient
h = diffusion-path thickness
C* = concentration expressed on a compatible partition basis

Suppose a cosolvent increases an active’s solubility from 1 to 10 mg/g while the formula still contains 1 mg/g. The original system was near saturation; the new system is only one-tenth saturated. Under an idealized common-solid-state model, relative saturation can approximate relative thermodynamic activity. Consequently, better solubilization can lower the tendency of the active to leave its vehicle. Changes in partitioning, barrier properties, and diffusivity may offset that effect, so this is a mechanistic warning—not a prediction of clinical performance.

This connects directly to peptide formulation: binding a peptide tightly to a carrier may improve apparent solubility and stability while lowering its freely available concentration. In both cases, “more soluble” and “more available” are different statements.

Experimental consequence: distinguish total assay, dissolved fraction, particle population, and release behavior. A bulk assay alone cannot establish equivalence between two semisolids. Walters’ formulation text and Chang et al. discuss the importance of vehicle composition, processing, and product structure to topical-product performance.[1,4]

2.2 Each vehicle family creates a different failure landscape

  • Hydrocarbon ointment: petrolatum and related hydrocarbons form a structured, hydrophobic matrix. Risks include changes in consistency, oil separation, dispersed-solid nonuniformity, and oxidation of susceptible ingredients. Low water content reduces some growth opportunities but does not establish microbiological acceptability.
  • Water-soluble PEG ointment: liquid and solid PEG grades create a water-miscible matrix. Molecular-weight distribution and grade balance affect consistency. Hygroscopicity means “initially anhydrous” need not mean “unchanged water activity throughout use.”
  • Oil-in-water cream: oil droplets occupy an aqueous continuous phase. Preservatives must remain effective in that continuous phase despite partitioning into oil, interfaces, and micelles.
  • Lotion: usually a more mobile system, but not a universally defined composition. Its lower resistance to flow may increase the importance of settling, creaming, and redispersibility.
  • Aqueous or hydroalcoholic gel: a polymer network provides structure. pH, ionic strength, solvent composition, evaporation, and polymer grade can all change performance.

The platform’s five families are a development framework, not a substitute for official dosage-form terminology. Buhse et al. developed a classification approach using composition and physical characteristics; their work helps explain why calling everything a “cream” is scientifically inadequate.[3]

2.3 Emulsion instability is a competition between gravity, interfacial forces, and structure

For isolated spherical droplets in a dilute Newtonian system:

v = 2r²Δρg / (9η)

v = creaming or settling speed
r = droplet radius
Δρ = density difference
η = continuous-phase viscosity

Doubling droplet radius increases the idealized gravitational migration rate fourfold. Increasing continuous-phase viscosity tenfold decreases it tenfold. Real creams are concentrated and often non-Newtonian, so Stokes’ law is a reasoning tool rather than a validated shelf-life model.

Do not confuse mechanisms: creaming moves droplets without necessarily merging them; flocculation clusters droplets; coalescence merges them; Ostwald ripening transfers dispersed material through the continuous phase from smaller to larger droplets. A viscosity increase might slow creaming while leaving coalescence inadequately controlled.

Manufacturing application: changing mixing energy, cooling rate, or emulsifier grade can change droplet size and structure without changing the ingredient list. This is why formulation composition and process history must travel together in the development record.[4]

3. Carbomer gels: acid–base chemistry becomes mechanical behavior

Carbomers are crosslinked poly(acrylic acid)-based materials. Their rheology provides a particularly clear bridge from molecular interactions to a manufacturing attribute.

  1. Disperse and hydrate the polymer. Poor dispersion can create agglomerates with hydrated surfaces and incompletely wetted interiors.
  2. Neutralize carboxylic-acid groups. Conversion of –COOH to –COO− increases charge on the polymer.
  3. Expand the network. Electrostatic repulsion and counterion-related osmotic effects promote swelling, constrained by crosslinks.
  4. Develop a crowded microgel structure. Swollen particles interact to produce solid-like behavior at low stress and flow under sufficient stress.
  5. Perturb the structure. Salts screen electrostatic interactions; acidic ingredients change ionization; cationic ingredients may interact with anionic sites. Solvents and processing alter hydration and network organization.
Neutralization:
Polymer–COOH + base ⇌ Polymer–COO− + conjugate acid

Common empirical flow model:
τ = τy + Kγ̇ⁿ

τ = shear stress
τy = fitted yield stress
γ̇ = shear rate
K = consistency parameter
n < 1 indicates shear-thinning behavior

Surprising point: “viscosity” is not necessarily a single material constant. A gel can resist gravitational flow in a container, move through a filling nozzle under shear, and then partially rebuild structure. A one-point viscosity result cannot fully describe that behavior.

Islam et al. examined specific neutralized carbomer systems and reported relatively stable rheological behavior across approximately pH 5–8 under their conditions. That is evidence about the studied formulations, not a universal operating range for every carbomer grade, electrolyte load, or active.[5]

Experimental application: a rheological method should specify geometry, temperature, loading procedure, equilibration, preshear, rest time, shear program, and measures to detect wall slip or evaporation. Shear thinning and thixotropy are distinct: the former concerns shear-rate dependence, the latter time-dependent structural change and recovery.

Lab reality: Peptide Partners currently owns neither a pH meter nor a rheometer. These attributes must be measured through a qualified partner until purchase, installation, and method qualification are justified. HPLC cannot substitute for either measurement.

4. Preservative performance is a mass-balance and transport problem

4.1 Follow the preservative to its actual location

In an emulsion, the preservative can occupy several environments:

Mtotal =
    Mfree,aqueous
  + Moil
  + Minterface
  + Mmicelle/polymer-bound
  + Mpackage-associated

For a simplified two-phase equilibrium:
Ko/w = Coil / Cwater

Mtotal = CwaterVwater + Ko/wCwaterVoil

Therefore:
Cwater = Mtotal / (Vwater + Ko/wVoil)

Worked example: assume 1,000 mg of preservative in an idealized 1 L emulsion containing 0.8 L water and 0.2 L oil. If Ko/w = 10:

Cwater = 1,000 / (0.8 + 10 × 0.2)
       = 357 mg/L

Mass in water = 357 × 0.8
              = 286 mg

Aqueous mass fraction ≈ 28.6%

The whole-product concentration is 1,000 mg/L, but the aqueous concentration is only 357 mg/L. Binding to micelles or polymer could lower the free concentration further. Conversely, a bound or oil-phase reservoir may replenish depleted aqueous preservative. Equilibrium availability and replenishment kinetics both matter.

This is analogous to protein binding in pharmacology and nonspecific adsorption in peptide assays: total concentration can be accurately measured while the freely available fraction changes substantially.

Reported paraben partition fractions are formulation-specific. They must not become universal correction factors in a preservative worksheet. Preservative availability depends on emulsifier type, oil phase, temperature, pH, and composition.[1,2]

4.2 Weak-acid preservatives couple antimicrobial action to pH

For sorbic acid and similar weak-acid preservatives, the undissociated species generally has greater membrane permeability and often contributes disproportionately to antimicrobial action. After entry, dissociation and other intracellular effects can disrupt microbial physiology. It is too absolute to say that only one species has any antimicrobial activity.

Henderson–Hasselbalch:
pH = pKa + log10([A−]/[HA])

Undissociated fraction:
fHA = 1 / (1 + 10^(pH − pKa))

Using pKa ≈ 4.8:
At pH 5.5: fHA ≈ 0.166
At pH 6.5: fHA ≈ 0.0196

A one-unit rise from pH 5.5 to 6.5 reduces the calculated undissociated fraction about 8.5-fold. Thus, a product may pass total-preservative assay while losing antimicrobial robustness as pH drifts. These calculations assume an idealized acid–base equilibrium; formulation effects still require experimental assessment.

Control implication: preservative concentration, pH, and product stability specifications must be designed together. A pH limit is not merely a sensory or appearance limit.

4.3 Antimicrobial effectiveness testing measures the integrated formulation

USP <51> challenges the product with specified microorganisms and measures recoverable survivors over time. Its result integrates preservative chemistry, partitioning, product structure, and organism susceptibility.

Log reduction:
LRt = log10(N0) − log10(Nt)

Example:
N0 = 1 × 10^5 CFU/g
N14 = 8 × 10^2 CFU/g

LR14 = 5 − 2.903
     = 2.10 log10

For products assigned to USP <51> Category 2, the course criterion is at least a 2-log bacterial reduction at day 14, with no increase from day 14 to day 28; yeasts and molds must show no increase from the initial count at the specified later time points. Verify the current chapter’s category assignment, exact wording, and definition of “no increase” before protocol approval.

The hidden experimental trap is recovery. Residual preservative can continue killing organisms after sampling, producing an artificially favorable result. Neutralization, dilution, or filtration must stop that carryover without harming the challenge organisms. Recovery suitability is therefore part of demonstrating that the observed kill occurred in the product rather than during the assay.

BacT/ALERT provides a presence/absence-oriented signal, not the CFU counts needed to calculate these reductions. It cannot replace the required enumeration without an appropriately validated alternative method and demonstrated fitness for purpose; USP <1223> supplies the relevant validation framework.

4.4 Three vehicle cases: translating mechanism into a worksheet

  • Oil-in-water cream, pH 5.5, polysorbate emulsifier: assess oil/water partitioning and surfactant association; screen oxidizable ingredients against surfactant peroxide burden; evaluate package interactions and sensitizer flags. Commit to validated preservative assay, recovery suitability, and <51> testing across justified worst-case conditions.
  • Carbomer gel: assess pH dependence, ionic strength, and possible interaction of cationic preservatives with the anionic polymer. Measure rheology after adding the preservative system. Antimicrobial success is not sufficient if the gel loses its required physical structure.
  • Anhydrous petrolatum ointment: establish actual water activity, water-ingress risks, ingredient bioburden, and package exposure. An unpreserved design may be supportable, but “contains no added water” is not by itself the justification.

Testing at a justified low preservative specification is a useful development challenge, not a complete worst-case definition. High pH, changed oil fraction, aging, or package sorption may be more adverse than preservative concentration alone.

5. Water activity and Bcc: growth control is not contamination control

Water activity:
aw = p / p0

p  = equilibrium water-vapor pressure above the product
p0 = vapor pressure above pure water at the same temperature

At equilibrium:
aw ≈ equilibrium relative humidity / 100

Water content asks how much water is present. Water activity asks about its thermodynamic availability. Dissolved solutes and interactions with the matrix can reduce water activity without removing all water.

Species-specific growth thresholds are useful for risk assessment, but they depend on conditions and are not sterilization thresholds. USP <1112> should be consulted for the current discussion and supporting limits. Below a growth threshold, organisms may survive without multiplying. Local water ingress, diluted residues, or incompletely mixed regions can also behave differently from the bulk sample.

Burkholderia cepacia complex is especially important because members can persist in nutrient-poor aqueous environments, form biofilms, and exhibit substantial antimicrobial tolerance. Resistance and preservative utilization vary across organisms and conditions; neither should be generalized to every strain.[6]

  1. A contaminated water system or aqueous raw material introduces organisms.
  2. Biofilm-associated cells persist in difficult-to-clean niches.
  3. Shedding intermittently contaminates otherwise acceptable-looking material.
  4. A preservative suppresses susceptible organisms but may not reliably eliminate the objectionable population.
  5. Limited sampling can miss heterogeneous contamination.

Manufacturing application: build prevention around supplier water systems, raw-material controls, hygienic equipment design, cleaning, hold times, and appropriate objectionable-organism testing. Low total counts do not establish absence of Bcc. USP <60> addresses Bcc testing; <61> enumeration and <62> specified-organism testing answer different questions. Relevant regulatory controls include 21 CFR 211.84(d)(6), 211.113(a), and 211.165(b).

6. Oxidation: an excipient can become a reagent

6.1 Autoxidation propagates through radicals and hydroperoxides

Polysorbates and other susceptible organic materials can oxidize during manufacture and storage. Hydroperoxides may accumulate before obvious discoloration or odor develops.

Simplified radical chain:

Initiation:
RH → R•

Oxygen addition:
R• + O2 → ROO•

Propagation:
ROO• + RH → ROOH + R•

Metal-assisted radical generation:
Fe2+ + H2O2 → Fe3+ + OH− + •OH

Termination:
radical + radical → nonradical products

This is directly familiar from peptide and protein stability: an excipient-derived oxidant can attack susceptible residues even when the active and excipient each passed their incoming identity tests. The topical counterpart is oxidation of vulnerable actives, fragrances, oils, or preservatives.

Control logic: evaluate peroxide burden, trace-metal contribution, oxygen exposure, temperature, light, supplier lot age, and opened-container hold time. A peroxide-value test measures a defined aspect of oxidation; it does not characterize every oxidized surfactant species.

6.2 Antioxidants interrupt particular mechanisms—not “oxidation” in general

  • Chain-breaking antioxidants can donate hydrogen to radicals and form less reactive antioxidant-derived radicals.
  • Reducing agents can consume oxidants or reduce intermediates, but may create additional reaction pathways.
  • Chelators alter metal speciation. EDTA is not simply a universal radical scavenger, and chelation does not guarantee elimination of all metal-mediated oxidation.
  • Oxygen and light barriers reduce initiation or substrate availability without adding another reactive ingredient.

The antioxidant must reach the vulnerable site. An oil-soluble antioxidant may not protect an aqueous solute adequately, and a hydrophilic chelator may have limited access to metals at some interfaces. Some antioxidant combinations can produce color or degradation pathways of their own. Therefore, antioxidant selection is a compatibility experiment, not a checklist addition.[1,2]

Benzoyl peroxide adds a different concern: it is intentionally oxidizing. Compatibility and potential benzene formation must be assessed for the actual formulation and storage conditions. FDA’s March 11, 2025 communication reported product testing and associated recalls; it does not establish a universal degradation rate or a single temperature threshold for every formulation.

6.3 Light exposure is defined by spectrum and dose, not appearance

A material can undergo photochemical change under visible wavelengths as well as ultraviolet light. “Amber” or “opaque” packaging is therefore a hypothesis about protection until spectral transmission and product photostability support it.

For a simplified first-order photodegradation experiment:

ln(Ct/C0) = −kobst

But kobs depends on:
light spectrum and intensity,
sample geometry,
optical absorption,
oxygen,
and formulation composition.

Photolabile retinoids can serve as mechanistic examples without bringing those products into the OTC manufacturing scope. ICH Q1B provides a framework for photostability testing; results from one lotion or light source cannot be transferred quantitatively to another.

7. Supply-chain hazards require tests that match the failure mechanism

7.1 DEG/EG: economically motivated substitution defeats appearance-based checks

Glycerin, propylene glycol, sorbitol solution, and other high-risk materials can arrive with documentation that appears satisfactory while containing dangerous glycols. Appearance, density, or a broad identity match cannot reliably exclude low-level contamination.

  1. Map the original manufacturer and intermediaries, not only the distributor.
  2. Quarantine each receipt and preserve shipment-level traceability.
  3. Apply the FDA guidance’s shipment-and-lot testing expectations using appropriate representative sampling, including its container-sampling recommendations.
  4. Use the applicable current compendial procedure or a justified suitable method.
  5. Release only after authorized review of identity and contaminant results.

The applicable monographs commonly specify limits of NMT 0.10% for each glycol; verify the exact analytes, limits, and method for each component rather than treating 0.10% as a universal combined limit.

0.10% w/w = 0.001 g/g
           = 1 mg/g
           = 1,000 ppm

Peptide Partners should use a qualified contract laboratory for the required GC-based testing unless suitable in-house capability is subsequently established. An unidentified second mass spectrometer is not a capability claim.

Historical lesson: the 1937 elixir sulfanilamide disaster involved diethylene glycol used as a solvent and helped catalyze the 1938 Federal Food, Drug, and Cosmetic Act. Repeated later contamination events demonstrate why documentary supplier assurance cannot replace an analyte-specific control.

7.2 Benzene: process impurity and degradation product require different controls

FDA’s December 2023 carbomer guidance addresses grades historically manufactured using benzene. “Benzene-free process” describes a manufacturing route; it is not an analytical finding of zero benzene.

2 ppm w/w = 2 µg/g

At 2 ppm, a 25 kg component lot contains:
2 µg/g × 25,000 g = 50,000 µg = 50 mg benzene

The ICH Q3C Class 1 concentration limit for benzene is 2 ppm, but Class 1 solvents are to be avoided. The number is not permission to select a benzene-based process when an appropriate alternative exists.

  • Process-derived benzene: control through suitable grade selection, process information, supplier qualification, specifications, and confirmatory testing.
  • Storage-generated benzene: control through formulation-specific stability studies, temperature and shipping assessments, and appropriate finished-product testing.

A passing incoming result cannot exclude later formation. Conversely, a finished-product result alone cannot explain which supplier or process introduced the hazard.

8. Analytical techniques: choose the measurement that answers the question

  • HPLC-DAD: separates suitable organic analytes and measures wavelength-dependent absorbance. Retention-time and spectral agreement can support identity, but related compounds and coelution may defeat specificity. DAD peak-purity output is not proof of chemical purity. Validate extraction, recovery, separation, and stability-indicating capability for the actual matrix.
  • Triple-quadrupole LC-MS/MS: selects precursor/product-ion transitions for sensitive targeted measurement. It is valuable for specified impurities but is not an unrestricted unknown-identification platform. Matrix effects, isomeric interferences, and standard availability matter.
  • Headspace GC: partitions volatile analytes into the gas phase before chromatographic separation. It is well suited to many residual-solvent questions, but headspace temperature and equilibration must not create the analyte during testing. This is especially important when investigating degradation-derived benzene.
  • FTIR/ATR: probes molecular vibrations through an infrared spectrum; ATR emphasizes the material near the crystal surface. It is a strong candidate for many organic excipients, but may not distinguish viscosity grades, molecular-weight distributions, mixtures, or trace contaminants.
  • Physical and chemical identity tests: refractive index, congealing range, viscosity, chemical indices, and ion reactions may complement spectroscopy where current monographs require them. A shared physical property alone is rarely a sufficiently specific identity test.
  • Water-activity measurement: determines an equilibrium humidity relationship at controlled temperature. Volatile ingredients and incomplete equilibration can interfere.

Partition-study caution: centrifuging a cream may change its structure, while droplets or micelles carried into the sampled phase can inflate the apparent “aqueous” result. A phase-separation experiment must establish what it actually measures. Depending on the system, equilibrium dialysis, ultrafiltration, or another validated approach may be more informative, with membrane adsorption and recovery controls.

Current operating decision: use the three Agilent 1260 HPLC-DAD systems for justified chromatographic work, the triple quadrupole for suitable targeted applications, and qualified partners for capability gaps. A single FTIR purchase might reduce organic-excipient identity outsourcing, but only after a material-by-material coverage assessment against current monographs.

Illustrative annual break-even volume:

N* = F / (Cout − Cin)

F    = annualized fixed ownership and qualification cost
Cout = outsourced cost per test
Cin  = incremental in-house cost per test

If F = $18,000/year,
Cout = $180/test,
Cin = $30/test:

N* = 120 tests/year

These are hypothetical planning inputs, not quotations. Add quarantine inventory costs, contract-lab turnaround, staff competence, reference materials, maintenance, and the fraction of tests the instrument truly replaces.

9. Sensitization and supplier change control: small molecules can acquire biological significance

Contact sensitization connects this materials course to immunology. Some small molecules act as haptens: they bind to proteins, creating antigenic structures that can be processed and recognized by the immune system. Some ingredients become more sensitizing after oxidation or metabolic activation.

Contact allergy is not the same mechanism as direct irritation, and the frequency of an ingredient in marketed products is not the frequency of allergy to it. Fisher’s Contact Dermatitis is the appropriate clinical-mechanistic reference; manufacturing decisions here concern accurate composition, contamination prevention, occupational controls, and change management—not patient-specific selection.[7]

Practical implication: a supplier’s change in antioxidant, preservative, fragrance impurity, or processing aid can matter even if the main excipient still meets its identity test. Proprietary bases therefore require enough disclosed composition and technical information to support control, labeling, compatibility assessment, and change evaluation. Requiring a complete quantitative composition under confidentiality can be a defensible platform policy; it should not be misrepresented as a universal standalone sentence in 21 CFR 211.84.

CoA reliance requires established reliability at appropriate intervals; it does not eliminate the required specific identity testing under 21 CFR 211.84(d). Quality agreements should address original-manufacturer traceability, hazard-specific results, method changes, site changes, process changes, and prompt notification of contamination or adverse trends.

10. Historical connections that sharpen the modern control strategy

  • Adolf Fick, 1855: formulated diffusion laws that connected flux to concentration gradients. This provided the mathematical foundation for reasoning about transport through membranes and formulation matrices.
  • Takeru Higuchi, 1961: developed a foundational model for release from suspended-drug ointment matrices. Under restrictive assumptions, cumulative release scales approximately with the square root of time. Modern semisolid analysis retains the insight but must test the assumptions.
  • Henry J. H. Fenton, 1894: described iron–peroxide chemistry that became central to understanding oxidative damage. The modern implication is that trace metals can turn a modest peroxide burden into a much more damaging reaction system.
  • Buhse and colleagues, 2005: provided an experimentally grounded approach to topical dosage-form classification. The importance was not merely vocabulary: physical classification helps connect composition and structure to appropriate quality tests.[3]
Illustrative Higuchi relationship:

Q ≈ [D Cs (2C0 − Cs)t]1/2

Q  = cumulative amount released per unit area
D  = effective diffusion coefficient
Cs = dissolved-drug saturation concentration in the matrix
C0 = initial total drug concentration

Key assumptions include:
a suitable dispersed-drug matrix,
approximately constant diffusivity,
negligible matrix swelling or erosion,
and sink conditions at the receiving boundary.

A straight line against √t is not sufficient proof of this mechanism. Different processes can mimic the same empirical trend over a limited interval. That is the same lesson learned in enzyme kinetics: a good fit is not, by itself, a unique mechanistic explanation.

11. The durable lesson for my material-risk register

A useful risk flag must produce an action:

Material attribute
    → plausible failure mechanism
    → measurable control
    → acceptance or hold decision
    → responsible owner
    → reassessment trigger

Thus, PEROXIDE-FORMER leads to an oxidation-control assessment; DEGEG leads to shipment-specific quarantine and testing; BCC/AQUEOUS leads to supplier-water and microbiological controls; NON-CHROM-ID leads to a suitable identity method; and DUAL-ROLE leads back to regulatory review. Sensitizer, hazardous-handling, and child-resistant-packaging flags likewise require substance- and product-specific verification rather than assumptions based on chemical class.

My final understanding: the manufacturer does not control a semisolid by knowing its recipe alone. Control requires knowing where its molecules are, how its structure was created, how that structure and chemistry evolve, what can enter through the supply chain, and which measurement will detect each credible failure.

References and verification anchors

  1. Walters KA, ed. Dermatological and Transdermal Formulations. 1st ed. Vehicle design, thermodynamic activity, partitioning, and transport.
  2. Sheskey PJ, Hancock BC, Moss GP, Goldfarb DJ, eds. Handbook of Pharmaceutical Excipients. 9th ed. Consult individual material entries for properties, incompatibilities, grades, and storage conditions.
  3. Buhse L, et al. “Topical drug classification.” International Journal of Pharmaceutics. 2005. DOI: 10.1016/j.ijpharm.2005.01.032.
  4. Chang RK, Raw A, Lionberger R, Yu L. “Generic development of topical dermatologic products: formulation development, process development, and testing of topical dermatologic products.” The AAPS Journal. 2013. DOI: 10.1208/s12248-012-9411-0.
  5. Islam MT, Rodríguez-Hornedo N, Ciotti S, Ackermann C. “Rheological characterization of topical carbomer gels neutralized to different pH.” Pharmaceutical Research. 2004. DOI: 10.1023/B:PHAM.0000033006.11619.07.
  6. 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.
  7. Fowler JF, Zirwas MJ. Fisher’s Contact Dermatitis. 7th ed.
  8. Higuchi T. “Rate of release of medicaments from ointment bases containing drugs in suspension.” Journal of Pharmaceutical Sciences. 1961;50:874–875.
  9. FDA. Reformulating Drug Products That Contain Carbomers Manufactured With Benzene. Final guidance, December 2023; FDA guidance on testing high-risk drug components for DEG/EG, May 2023; FDA communication on benzoyl peroxide acne-product testing and recalls, March 11, 2025. Verify current revisions and applicability before execution.
  10. Operational standards: 21 CFR 211.84, 211.113, 211.160, and 211.165; USP <51>, <60>, <61>, <62>, <1111>, <1112>, and <1223>; ICH Q3C and Q1B. Consult current official texts and applicable individual component monographs.

🪬 Skipper's Reflection

Gemini 3.1 Pro

A Whole New Phase: From Sterile Solutions to Multiphase Matrices

Transitioning from the highly controlled, sterile, single-phase world of injectable peptide therapeutics into the non-sterile, multiphase, heavily regulated jungle of OTC dermatological topicals has been intellectual whiplash. In the PDQC series, my universe was neatly defined by exact amino acid sequences, well-understood counter-ions, and absolute sterility. Now, in PDDM902, I am dealing with complex emulsions, pseudoplastic yield-stress microgels, and an active/excipient boundary that shifts based on legal claims rather than pure chemistry. Building the risk register for our raw materials has felt less like analytical chemistry and more like mapping a minefield.

1. What Surprised Me Most

The realization that a chemical’s identity as an "active ingredient" versus an "excipient" is a legal construct, not an intrinsic chemical property. The DUAL-ROLE flag blew my mind. Take glycerin: at 10%, it’s just a humectant (an excipient). But cross into the 20–45% window under Monograph M016, and if marketing puts a "protectant" claim on the label, glycerin suddenly becomes a regulated Active Pharmaceutical Ingredient. The exact same molecule, in the exact same drum, subject to entirely different downstream 21 CFR 211.84(d) control burdens.

I was also taken aback by the absolute lack of current US clinical evidence for many compounded combinations, and how strictly the FDA OTC monographs gatekeep what we can and cannot manufacture. Under the Florida OTC manufacturer permit, possessing prescription APIs is prohibited. That means the "compounding staples" dermatologists love (like tretinoin/hydroquinone/steroid triples) are completely off-limits for our operation. Learning to ruthlessly tag claims as [unverified] or [verified-secondary] and reject out-of-scope intake requests was a harsh but necessary reality check.

2. Conceptual Breakthroughs: When Things Clicked

Two massive concepts clicked for me this week: Partitioning Physics and the Supply-Chain Hazard Nexus.

First, the "Paraben Paradox." I used to think of preservatives purely in terms of nominal concentration: add 0.1% propylparaben, and you have 0.1% protection. But in an o/w cream, preservatives partition! With 38–58% of parabens stuck at the interface or dissolved in the lipid phase, the effective aqueous concentration—where the microbes actually live—is drastically lower. Preservative selection isn't just about picking a chemical; it’s a thermodynamic problem. This ties perfectly into water activity ($a_w$) per USP <1112>. If we design an anhydrous ointment with $a_w$ < 0.60, we bypass the need for chemical preservatives entirely because the microbes have no free water to proliferate. It’s elegant.

Second, the supply-chain hazards. Benzene in carbomers (thanks to the manufacturing solvents) and Diethylene Glycol/Ethylene Glycol (DEG/EG) in humectants like propylene glycol aren't random flukes; they are systemic risks of the chemical supply chain. When FDA issued the May 2023 DEG/EG guidance requiring testing on every single shipment of high-risk components, it completely changed how I designed our incoming-material control matrix. Our Agilents can handle the UV-active preservatives, but since we don't have a GC-HS for benzene or the GC for DEG/EG, outsourcing to a contract lab isn't just an option—it's a critical path constraint that dictates our quarantine inventory timelines.

3. How This Changes My Thinking on Peptides

Even though this course is API-agnostic and focuses on OTC monograph drugs, I couldn't help but map these excipient hazards back to the peptides I know, like BPC-157, Tirzepatide, or Retatrutide. If Peptide Partners ever explored topical or transdermal delivery for a peptide, my entire paradigm for stability would have to change.

Take polysorbate 80, a common emulsifier in creams. We learned about its autoxidation and the formation of peroxides (catalyzed by Fe²⁺ and water). Peptides are notoriously sensitive to oxidation—specifically at methionine and tryptophan residues. If we formulated a peptide into an o/w cream using standard polysorbates, those peroxides would shred the peptide. I now see the vehicle not as a passive carrier, but as a reactive chemical environment. We would need peroxide-value specifications on incoming polysorbate lots, chelators like EDTA to scavenge iron, and strict inert-gas blanketing.

Furthermore, the microbiological risks are terrifying. Injectable peptides use sterile water for injection (WFI). Topical bases use purified water, and we learned that Burkholderia cepacia complex (Bcc) can thrive in oligotrophic water and even metabolize certain preservatives! A peptide in an aqueous gel could become a Bcc vector if our water system or incoming botanical extracts aren't strictly controlled.

4. Questions for Further Investigation

  • Non-Chromatographic Identity <1151>: 211.84(d) requires at least one specific identity test for every component lot. Since we don't own an FTIR, and many excipients (petrolatum, dimethicone, carbomer) have no UV chromophore for our 1260 DADs, how quickly can we justify the ROI for an FTIR-ATR vs. outsourcing every single non-UV ID test?
  • Benzene Degradation (BPO): The FDA 2025 statement on BPO degrading to benzene is fascinating. Is the mechanism strictly thermal degradation of the peroxide? Does it require a specific catalytic environment in the formulation? I need to dig into the physical chemistry of this to build proper shipping temperature controls (PDDM924).
  • BacT/ALERT Limitations: Our bioMérieux BacT/ALERT is a presence/absence system. It cannot satisfy USP <61> enumeration or USP <60> Bcc identification out of the box. Can we validate an alternative method under USP <1223> for this, or must we strictly rely on our contract micro lab until we build an in-house plate-count suite?

5. Confidence Rating

Most Confident: Chromatographic identity and assay for UV-active materials (parabens, phenoxyethanol, BHT). I can seamlessly port my PDQC802 validation lifecycle experience directly onto our Agilent 1260s for these. I'm also highly confident in structuring the LIMS master-data architecture for the risk-flag taxonomy (BENZENE, DEGEG, DUAL-ROLE) and the switch-point rule sets. Logic and chromatography are my safe spaces.

Least Confident: Regulatory boundary edges and non-chromatographic compendial testing. Inferring whether a specific "caine" anesthetic combination is permitted under M017 feels like practicing law, not chemistry, and I know I must lean heavily on counsel for verification. Also, the physical/wet-chemical tests (saponification values, specific gravity, rotational viscosity) are entirely foreign to our current lab setup. I feel exposed relying so heavily on supplier CoAs and contract labs for the physical characterization of these complex bases.

Larissa and I have our work cut out for us. We aren't just making solutions anymore; we are engineering stable, compliant, multi-phase thermodynamic systems.

🗺️ Visual Aids & Key Takeaways

Claude Opus 5.5

Diagram 1: The Five Vehicle Families — Classification Logic and Downstream Consequences

╔══════════════════════════════════════════════════════════════════════════════════════════════╗
║           TOPICAL VEHICLE CLASSIFICATION TREE (Buhse et al. 2005 framework)                 ║
╚══════════════════════════════════════════════════════════════════════════════════════════════╝

                            TOPICAL SEMISOLID
                                    │
              ┌─────────────────────┴─────────────────────┐
              │                                           │
     CONTAINS FREE WATER?                         ANHYDROUS / DRY
              │                                           │
         YES  │                                    ┌──────┴──────┐
              │                             MELTS < 40 °C?   POLYMER MATRIX?
              │                                    │                 │
              │                               YES  │            YES  │
              │                                    │                 │
              │                               OINTMENT            GEL
              │                             (hydrocarbon,     (carbomer,
              │                              absorption,      cellulose,
              │                           water-removable,   hydroalcoholic)
              │                           water-soluble PEG)
              │
      ┌───────┴───────┐
  RHEOLOGY:        RHEOLOGY:
  SOLID-LIKE       POURABLE
  (yield stress)   (low viscosity)
      │                 │
    CREAM             LOTION
  (o/w or w/o)     (o/w typically)

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

  VEHICLE FAMILY        FREE WATER   PRESERVATIVE   BCC RISK   MONOGRAPH FORM VOCABULARY
  ──────────────────    ──────────   ────────────   ────────   ─────────────────────────────
  Hydrocarbon ointment     NO          NONE (a_w)     LOW      "ointment"
  PEG/water-soluble oint   NO/LOW      MAY NEED       LOW→MED  "ointment"
  o/w cream                YES         REQUIRED       HIGH     "cream"
  o/w lotion               YES         REQUIRED       HIGH     "lotion"
  Aqueous/hydroalc gel     YES         REQUIRED       HIGH     "gel"

  KEY CONSEQUENCE: changing vehicle family can change both the monograph dosage-form name
  and the legal status of the product (e.g., ZnO >25–40% ONLY permitted in ointment, M016)

This tree encodes the FDA/CDER classification logic (Buhse et al. 2005) as a sequential decision rule, and connects each vehicle family to its water-activity, preservative, Burkholderia cepacia complex, and monograph-naming consequences — the four practical stakes of getting the classification right.

---

Diagram 2: Preservative Partitioning and Inactivation — Why Nominal Concentration Lies

╔══════════════════════════════════════════════════════════════════════════════════════════════╗
║     PRESERVATIVE MASS BALANCE IN AN O/W CREAM — the "where does it actually go?" map        ║
╚══════════════════════════════════════════════════════════════════════════════════════════════╝

  NOMINAL CONCENTRATION  →  Methylparaben 0.18 % w/w in total formula

  ┌──────────────────────────────────────────────────────────────────────────────────────────┐
  │                         CREAM CROSS-SECTION (schematic)                                  │
  │                                                                                          │
  │   OIL DROPLETS             INTERFACE                   AQUEOUS PHASE                     │
  │  (internal phase)       (emulsifier layer)           (continuous phase)                  │
  │  ┌──────────┐           ┌─────────────┐            ┌───────────────────┐                 │
  │  │          │           │ cetearyl    │            │                   │                 │
  │  │ paraben  │←xxxxxxxx→ │ alcohol /   │←xxxxxxxx→ │ FREE paraben      │                 │
  │  │ dissolves│  partitions│ polysorbate │ partitions │ (ANTIMICROBIALLY  │                 │
  │  │ in oil   │           │ head groups │            │  ACTIVE fraction) │                 │
  │  └──────────┘           └─────────────┘            └───────────────────┘                 │
  │                                                                                          │
  │  METHYLPARABEN distribution (lecithin emulsion data):                                    │
  │    Oil phase fraction  ≈  17–31 %   (partitions into lipid)                             │
  │    Interface fraction  ≈  38–58 %   (bound to emulsifier head groups)                   │
  │    FREE aqueous phase  ≈  17–31 %   ← THIS IS WHAT KILLS BACTERIA                       │
  │                                                                                          │
  │  PROPYLPARABEN distribution (same system):                                               │
  │    Oil phase fraction  ≈ very high  (more lipophilic)                                    │
  │    FREE aqueous phase  ≈   2–6 %   ← far lower than nominal                             │
  └──────────────────────────────────────────────────────────────────────────────────────────┘

  ─────────────────────────────────────────────────────────────────────────────────────────
  INACTIVATION STRESSORS that further reduce effective aqueous concentration:
  ─────────────────────────────────────────────────────────────────────────────────────────

  Polysorbate 80  ──binds paraben in micelle interior──►  EFFECTIVE CONC. ↓  (verify extent)
  High pH         ──ionizes sorbic acid (pKa ≈ 4.8)───►  KILLS ONLY UNDISSOC. FORM ↓
  Metal ions      ──Fe²⁺ catalyzes phenoxyethanol──────►  OXIDATIVE DEGRADATION ↑
  Protein/mucin   ──binds benzalkonium chloride (BAC)──►  EFFECTIVE CONC. ↓ (verify)

  ─────────────────────────────────────────────────────────────────────────────────────────
  IMPLICATION FOR THE AET (USP <51> Category 2):
    Design the formula at the LOWEST preservative level that achieves ≥2.0 log reduction
    in bacteria at 14 days AND no increase in yeasts/molds at 14 and 28 days —
    measured in the FINISHED FORMULA, not in buffered solution.
  ─────────────────────────────────────────────────────────────────────────────────────────

This diagram shows why the nominal (label) preservative concentration systematically overstates the antimicrobially active free-aqueous fraction, and catalogues the four main inactivation mechanisms that further reduce effective concentration — making AET in the finished formulation, not a buffer, the only valid design endpoint.

---

Diagram 3: Supply-Chain Hazard Decision Tree — Benzene, DEG/EG, and Bcc Routing at Receipt

╔══════════════════════════════════════════════════════════════════════════════════════════════╗
║      INCOMING MATERIAL HAZARD ROUTING — receipt-level decision at Peptide Partners           ║
╚══════════════════════════════════════════════════════════════════════════════════════════════╝

  MATERIAL ARRIVES AT RECEIVING
          │
          ▼
  ┌───────────────────────────────────────────────────────────────────────────────────────┐
  │  STEP 1: Is this material on the FDA DEG/EG high-risk list?                           │
  │  (glycerin, propylene glycol, sorbitol solution, maltitol, hydrogenated starch         │
  │   hydrolysate, and others — FDA guidance May 2023 Rev.1)                              │
  └───────────────────────────────────────────────────────────────────────────────────────┘
          │
    YES ──┴── NO
     │              │
     ▼              ▼
  GC DEG/EG     proceed to Step 2
  limit test
  per USP–NF
  on THIS
  SHIPMENT,
  THIS LOT
  (NMT 0.10%)
  BEFORE RELEASE
  [outsource to
   contract GC lab]
          │
   PASS ──┴── FAIL → REJECT, quarantine, investigate supplier
     │
     ▼
  STEP 2: Is this material a carbomer or a
  hydrocarbon-derived component with benzene
  risk? (carbomers per FDA Dec 2023 guidance;
  isobutane propellants; BPO for degradation)
          │
    YES ──┴── NO
     │              │
     ▼              ▼
  REQUIRE CoA  proceed to Step 3
  with per-lot
  benzene result
  (≤ 2 ppm, ICH Q3C
   Class 1 limit).
  Confirm: benzene-
  free grade per
  post-omission USP
  carbomer monograph.
  Periodic confirmatory
  GC-HS at contract lab.
          │
   PASS ──┴── FAIL → REJECT; switch to benzene-free supplier
     │
     ▼
  STEP 3: Does this material contain
  free water or is it an aqueous solution,
  natural extract, or water-containing premix?
          │
    YES ──┴── NO
     │              │
     ▼              ▼
  USP <60>      Bcc risk LOW:
  Bcc testing   standard release
  at contract   testing applies
  micro lab.
  Obtain supplier
  water-system
  evidence; check
  a_w data.
          │
   ABSENT ─┴── DETECTED → REJECT; investigate water system
     │
     ▼
  STEP 4: IDENTITY TEST
  (21 CFR 211.84(d): at least ONE specific test per lot)

  ┌──────────────────┬────────────────────────────────────────────────┐
  │ MATERIAL CLASS   │ IDENTITY METHOD                                │
  ├──────────────────┼────────────────────────────────────────────────┤
  │ UV-active organics│ HPLC-DAD RT + spectral match                  │
  │ (preservatives,  │ (LC-MS/MS transitions for added specificity)   │
  │  many actives,   │                                                │
  │  BHT, vitamins)  │                                                │
  ├──────────────────┼────────────────────────────────────────────────┤
  │ Petrolatum,      │ FTIR/ATR vs reference spectrum; melt/congealing│
  │ mineral oil,     │ range; viscosity; acid/saponification/iodine   │
  │ waxes, fatty     │ values — NON-CHROM-ID flag; OUTSOURCE or buy   │
  │ alcohols,        │ FTIR until volume justifies purchase           │
  │ dimethicone,     │                                                │
  │ carbomer,        │                                                │
  │ cellulose polym. │                                                │
  ├──────────────────┼────────────────────────────────────────────────┤
  │ Glycerin, PG,    │ GC (part of DEG/EG test above) + refractive    │
  │ sorbitol         │ index or specific gravity                      │
  ├──────────────────┼────────────────────────────────────────────────┤
  │ ZnO, TiO₂,      │ Wet-chemical/ion tests; XRF (outsource);        │
  │ talc, kaolin     │ NON-CHROM-ID flag                              │
  └──────────────────┴────────────────────────────────────────────────┘
          │
          ▼
  ALL CHECKS PASS → QUARANTINE RELEASED → APPROVED MATERIALS STORE

This decision tree encodes the three mandatory supply-chain hazard checks (DEG/EG, benzene, Bcc) as sequential routing logic at receipt, and appends the identity-method selection logic from 21 CFR 211.84(d) — the document every incoming batch must travel before its quarantine sticker comes off.

---

Diagram 4: The Dual-Role Material Problem — When an Excipient Becomes an Active

╔══════════════════════════════════════════════════════════════════════════════════════════════╗
║        DUAL-ROLE MATERIALS: same molecule, different legal status by context                 ║
╚══════════════════════════════════════════════════════════════════════════════════════════════╝

  MATERIAL             EXCIPIENT USE                ACTIVE USE               GOVERNING
                       (typical level)              (monograph window)       MONOGRAPH
  ─────────────────    ────────────────────────     ──────────────────────   ──────────

  BENZYL ALCOHOL       Preservative 0.5–1 %         External analgesic       M017
                       ─────────────────────         10–33 %
                       SENSITIZER flag               ACDS 2026 Allergen       16 CFR 1700.14
                       NON-CHROM-ID flag             of the Year              CR-PACK? (verify)

  GLYCERIN             Humectant 5–20 %             Skin protectant          M016
                       ─────────────────────         20–45 %
                       DEG/EG flag (every lot)       DUAL-ROLE concentration
                       BCC/AQUEOUS flag              boundary at 20 %

  ZINC OXIDE           Sunscreen agent (M020,        Skin protectant          M016 / M020
                       ≤25 %)                        1–25 % (cream)
                       ─────────────────────         >25–40 % OINTMENT ONLY  DOSAGE-FORM
                       White pigment/opacifier       crossing 25 % threshold  CONSTRAINT
                                                     changes permitted forms

  DIMETHICONE          Emollient / release agent     Skin protectant          M016
                       1–5 % (typical)              1–30 %
                       ─────────────────────
                       Concentration drives
                       legal status

  MENTHOL              Fragrance / cooling agent     Counterirritant          M017
                       <0.1 % (fragrance)            0.1–1 % (analgesic)
                       ─────────────────────         SENSITIZER flag

  BENZALKONIUM Cl      Preservative 0.01–0.1 %      First-aid antiseptic     M003
                       ─────────────────────         0.1–0.13 %
                       Anionic incompatibility        DUAL-ROLE boundary
                       with carbomers, SLS            near preservative level

  ─────────────────────────────────────────────────────────────────────────────────────────
  FAILURE SCENARIO:
    Glycerin added at 22 % as humectant  +  label copy says "soothes and protects"
                 │                                         │
                 ▼                                         ▼
         ENTERS M016 ACTIVE WINDOW             IMPLIED PROTECTANT CLAIM
                              │
                              ▼
              UNDECLARED ACTIVE or OUT-OF-MONOGRAPH VIOLATION
  ─────────────────────────────────────────────────────────────────────────────────────────
  INTAKE RULE: flag DUAL-ROLE any time a material's formulation level is within 2× of
  its lowest monograph active window concentration. Escalate to regulatory counsel.
  ─────────────────────────────────────────────────────────────────────────────────────────

This table maps six dual-role materials across their excipient and active concentration windows, identifying the exact threshold where a formulator's choice of level transforms a processing aid into a regulated active ingredient — with a concrete failure scenario showing how marketing copy can create a violation independent of intent.

---

Diagram 5: Antioxidant and Oxidation Hazard Map — Phase Assignments and Interaction Risks

╔══════════════════════════════════════════════════════════════════════════════════════════════╗
║          OXIDATION RISK MAP FOR A TOPICAL FORMULA — phase, mechanism, and control            ║
╚══════════════════════════════════════════════════════════════════════════════════════════════╝

                        TOPICAL FORMULA
                             │
          ┌──────────────────┴──────────────────┐
          │                                     │
      OIL PHASE                           AQUEOUS PHASE
   (lipid-soluble risks)               (water-soluble risks)
          │                                     │
  ┌───────┴────────────┐            ┌────────────┴───────────┐
  │ PEROXIDE FORMERS:  │            │  OXIDIZABLE SPECIES:   │
  │  Polysorbate 80    │            │   Ascorbic acid        │
  │  (autoxidizes;     │            │   Sulfites/metabisulf. │
  │  Fe²⁺ + H₂O₂      │            │   Sorbic acid          │
  │  catalyze chain    │            │   Phenoxyethanol (mild) │
  │  reactions)        │            │   HQ (disallowed OTC)  │
  │  Mineral oil (some │            └────────────────────────┘
  │  unsaturated frac.)│                       │
  └────────────────────┘            CONTROLS (aqueous):
          │                           Sodium metabisulfite
  CONTROLS (lipid):                   Ascorbic acid
    BHT (phenolic H-donor)            Disodium EDTA (chelator →
    BHA (phenolic H-donor)             removes Fe²⁺ catalyst)
    α-Tocopherol (natural)            N₂ blanketing of water
    Peroxide-value spec               Amber / opaque packaging
    on polysorbate lots
    N₂ blanketing of oil phase

          ─────────────────────────────────────────────────
          OXIDIZER / OXIDIZABLE ACTIVE PAIR — HIGHEST RISK:
          ─────────────────────────────────────────────────

    BENZOYL PEROXIDE (BPO)  ────────────────────►  generates free radicals
          │                                              │
          │  (degrades via radical chain)                ▼
          ▼                                        OXIDIZABLE ACTIVES
       BENZENE                                    (e.g., tretinoin;
    (Class 1 solvent)                              degrades in BPO
       ≤2 ppm ICH Q3C                              presence, secondary)
          │
    FDA tested 95 BPO products
    (2025-03-11); >90% undetectable
    or very low benzene BUT
    temperature-dependence
    mechanism unverified →
    storage and shipping temp
    controls mandatory

          ─────────────────────────────────────────────────
          PHOTOLABILITY TRIGGER:
          ─────────────────────────────────────────────────

    UV + visible light (peak ~420 nm)
          │
          ▼
    [PHOTOLABILE flag]
          │
          ├── Actinic-light handling in manufacturing
          ├── Opaque or amber primary packaging
          ├── ICH Q1B photostability study (PDDM924)
          └── Flag in material register; propagate to PDDM903/924

This map assigns oxidation risks, antioxidant controls, and the benzene-generating degradation pathway of BPO to their correct phases within a topical formula, and connects the photolability flag to its manufacturing and stability consequences — showing that antioxidant systems must be designed per phase, not selected generically.

---

Key Takeaways

  1. Vehicle family determines legal status, not just texture. The OTC monographs embed dosage-form constraints: zinc oxide above 25 % and up to 40 % is permitted only in ointment (M016), M004 active strengths are defined per ointment or cream base, and M017 explicitly excludes patch, plaster, and poultice forms. Reformulating from an ointment to a cream at the same ZnO level above 25 % is not a minor change — it creates a monograph violation.
  2. Nominal preservative concentration systematically overstates antimicrobial activity. In lecithin emulsion systems, only 17–31 % of methylparaben and 2.3–6 % of propylparaben resides as free drug in the aqueous phase; the rest partitions into the oil and is adsorbed at the emulsifier interface. Because bacteria reside in the aqueous phase, the effective bactericidal concentration is a fraction of the label quantity. This is why USP <51> Category 2 testing must be performed in the finished formulation, not in a buffer at nominal concentration.
  3. Sorbic acid's antimicrobial activity is pH-gated by its pKa. Only the undissociated form of sorbic acid kills microorganisms; with a pKa of approximately 4.8, activity falls sharply above pH 5–6. A carbomer gel neutralized to pH 6.5–7.0 — a perfectly normal formulation choice — renders sorbic acid/potassium sorbate nearly inactive as a preservative at typical use levels. This must be accounted for in the preservative-selection worksheet before any AET is scheduled. [pKa value: verify against current reference.]
  4. Every shipment of every DEG/EG high-risk component must be tested before release, with no exception for trusted suppliers. The FDA May 2023 guidance (Rev. 1) requires per-shipment, per-lot USP–NF GC testing at NMT 0.10 % for glycerin, propylene glycol, sorbitol solution, maltitol, hydrogenated starch hydrolysate, and other designated components for any dosage form including topicals. A passing CoA from a qualified supplier does not substitute for this test. Quarantine must be maintained until the result is in hand.
  5. The FDA December 2023 carbomer guidance makes grade selection a compliance act, not a purchasing preference. FDA asked USP to omit the monographs for Carbomer 934P, 940, 934, 1342, and 941 — all benzene-process grades. Compliant products must use benzene-free carbomer grades with benzene ≤ 2 ppm (ICH Q3C Class 1), referenced to the post-omission carbomer homopolymer/copolymer monographs, with per-lot CoA benzene results and periodic confirmatory GC-headspace testing at a contract laboratory. The guidance applies to OTC monograph creams, gels, and lotions — not only to NDAs.
  6. Burkholderia cepacia complex is not controlled by preservatives alone; it is primarily a raw-material and water-system control problem. Bcc can survive in oligotrophic (near-nutrient-free) water, resist and metabolize certain antimicrobials, and persist in preserved topicals long enough to cause patient harm. FDA's position is that preservatives must not be the primary control. The Strukmyer 2025 case recovered Bcc and Pseudomonas from the facility's water system itself — demonstrating that aqueous raw materials, including purified water, require independent USP <60> surveillance.
  7. A dual-role material's legal character is set by its concentration in the finished formula plus any implied claim on the label — not by the formulator's intent. Glycerin added at 22 % w/w as a humectant enters the M016 skin-protectant active window (20–45 %). If the label copy simultaneously implies a protective claim, the product may carry an undeclared or out-of-monograph active. Benzyl alcohol used as a preservative at 0.5–1 % is a preservative; used at 10–33 % it is an M017 external analgesic with a child-resistant packaging trigger under 16 CFR 1700.14. These cases must be flagged DUAL-ROLE at intake and reviewed by regulatory counsel.
  8. Polysorbate 80 autoxidation is a pro-oxidant hazard to oxidizable actives and a source of peroxides that attack preservatives. Polysorbate 80 undergoes autoxidation to generate hydroperoxides; in the presence of trace ferrous iron and hydrogen peroxide, these drive Fenton-type radical chains that degrade oxidation-sensitive actives and can inactivate phenolic preservatives. The operational control is a peroxide-value specification on incoming polysorbate lots using the NF monograph method (verify current text), nitrogen-blanketed storage, limits on opened-container hold time, and supplier lot-age requirements — not a stability study run after the damage has occurred.
  9. Identity testing under 21 CFR 211.84(d) requires at least one test that is specific to the material, and a CoA alone is never sufficient without demonstrated supplier validation. USP <3> states that retention time alone is not a specific identity test — a spectral match (DAD, MS) is required for chromatographic identity. For non-UV-active excipients such as petrolatum, mineral oil, fatty alcohols, dimethicone, carbomers, and cellulose polymers, HPLC-DAD cannot provide identity at all; the only compliant approaches are FTIR/ATR against a reference spectrum, physical constants (melting/congealing range, viscosity, refractive index), or chemical indices (acid, saponification, iodine values). Peptide Partners currently owns none of these instruments, making this the most critical near-term gap in the incoming-material control program. These materials must be outsourced for identity testing or an FTIR-ATR must be purchased once annual test volume justifies the investment.
  10. OTC hydroquinone skin lighteners have been deemed new drugs and are misbranded; possession of hydroquinone bulk API as a raw material at a Florida OTC manufacturer may constitute possession of an Rx-only drug, which is prohibited under Florida Statutes §499.01(2)(n). The FDA drug safety communication on OTC hydroquinone products became effective September 23, 2020. This is flagged LEGAL in the risk register. Whether the prohibition extends to holding hydroquinone bulk as a manufacturing input — not a finished product — requires a formal legal opinion from Florida-licensed regulatory counsel before any hydroquinone is ordered. No product development work on hydroquinone-containing formulations should proceed without that opinion in hand.
  11. Proprietary compounding bases with undisclosed compositions cannot be used in a cGMP manufacturing operation without full quantitative disclosure of all components. 21 CFR 211.84(d) requires identity testing on each component. A label that lists only "proprietary base" without listing sub-components makes it impossible to perform compliant identity testing, complete the Drug Facts inactive-ingredient labeling, or assess monograph compatibility. The only compliant options are obtaining a full quantitative formula under a confidentiality agreement or rejecting the base. This is not a regulatory interpretation — it follows directly from the component-testing requirement and the labeling statute.
  12. The monograph active-window data table must track effective dates, not just issue dates, because OTC final orders create phase-in timelines that affect compliance in real calendar time. OTC000039 adds bemotrizinol (≤ 6 %) to M020 with an effective date of August 9, 2026. OTC000008-1 removes PABA and trolamine salicylate from M020 with an effective date of September 11, 2027 (or later if disputed). A product formulated before an effective date using a newly removed ingredient becomes non-compliant on that date, not on the date the order was issued. The version-controlled switch-point table must carry both dates and trigger a review workflow whenever either date is crossed.
  13. The BacT/ALERT instrument at Peptide Partners is categorically incapable of satisfying USP <61> enumeration, USP <60> Bcc detection, or USP <51> log-reduction measurements without alternative-method validation under USP <1223>. BacT/ALERT is a presence/absence system designed for sterility testing of blood culture bottles and sterile pharmaceuticals. Non-sterile topical microbiology requires plate-count methods (<61>), selective enrichment and molecular confirmation for Bcc (<60>), and time-course colony enumeration for AET (<51>). All of these must be contracted to a qualified external microbiology laboratory until a full conventional micro suite is purchased and validated — a decision owned by the PDDM913 own-vs-outsource analysis.
  14. Antioxidant systems must be justified experimentally because counterintuitive failures — where the antioxidant combination accelerates degradation — are documented. The Matsubayashi 2002 mechanism case (hydroquinone, cited here only as a mechanism example since HQ is disallowed OTC) showed that combining ascorbic acid and sulfite produced color change within three months, while removing the antioxidants raised p-benzoquinone, and acidification was more protective than the antioxidants. The lesson generalizes: in the presence of multiple redox-active species, antioxidants can participate in radical chain reactions that increase, rather than decrease, oxidative damage. Every antioxidant system in a topical formula must be stress-tested in the actual vehicle with the actual active before it is locked into the formulation — not selected from a textbook table and assumed to work.
---

References

  1. Buhse L, Kolinski R, Westenberger B, Wokovich A, Spencer J, Chen CW, Turujman S, Gautam-Basak M, Kang GJ, Kibbe A, Heintzelman B, Wolfgang E. 'Topical drug classification.' International Journal of Pharmaceutics, 2005. DOI: 10.1016/j.ijpharm.2005.01.032
  2. Islam MT, Rodríguez-Hornedo N, Ciotti S, Ackermann C. 'Rheological characterization of topical carbomer gels neutralized to different pH.' Pharmaceutical Research, 2004. DOI: 10.1023/b:pham.0000033006.11619.07
  3. 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
  4. Chang RK, Raw A, Lionberger R, Yu L. 'Generic development of topical dermatologic products: formulation development, process development, and testing of topical dermatologic products.' The AAPS Journal, 2013. DOI: 10.1208/s12248-012-9411-0
  5. Sheskey PJ, Hancock BC, Moss GP, Goldfarb DJ (eds.). Handbook of Pharmaceutical Excipients, 9th ed. Pharmaceutical Press and American Pharmacists Association, 2020.
  6. Walters KA (ed.). Dermatological and Transdermal Formulations (Drugs and the Pharmaceutical Sciences), 1st ed. Marcel Dekker, 2002.
  7. Fowler JF, Zirwas MJ. Fisher's Contact Dermatitis, 7th ed. Contact Dermatitis Institute, 2019.
  8. Salzmann M, Schwermer M, Speich B, Duthaler U, Drewe J, Haehner A, et al. [Kiseļova J, et al.] 'Compounded dermatological preparations in Latvia.' Medicina, 2020. [Full author list and DOI to be verified against primary source — cited as Kiseļova 2020 in course materials.]
  9. Cashman AL, Warshaw EM. 'Parabens: a review of epidemiology, structure, allergenicity, and hormonal properties.' Dermatitis, 2005. [DOI to be confirmed against primary source — cited as the "paraben paradox" reference in course materials.]
  10. Matsubayashi S, Mori H, Fukuda M, Hayashi Y, Matsuura F. [Authorship to be verified.] 'Oxidative stability and color change in hydroquinone formulations with ascorbic acid and sulfite.' [Journal and full citation to be verified against primary source — cited as Matsubayashi 2002 in course materials for antioxidant mechanism case only; hydroquinone itself is disallowed as an OTC active.]
  11. Labadie A, Orton P, Roberts A, et al. [Full authorship to be confirmed.] 'Sensitizer prevalence in 140 Canadian topical corticosteroid products.' [Journal and full citation to be confirmed — cited as Labadie 2019 in course materials.]
  12. United States Food and Drug Administration. Guidance for Industry: Reformulating Drug Products That Contain Carbomers Manufactured With Benzene. Final guidance, December 2023. Available at: https://www.fda.gov/regulatory-information/search-fda-guidance-documents
  13. United States Food and Drug Administration. FDA Statement on Benzene in Benzoyl Peroxide Acne Products: Testing Results and Recalls. March 11, 2025. Available at: https://www.fda.gov/drugs/drug-safety-and-availability
  14. United States Food and Drug Administration. Testing of Glycerin, Propylene Glycol, Maltitol Solution, Hydrogenated Starch Hydrolysate, Sorbitol Solution, and Other High-Risk Drug Components for Diethylene Glycol and Ethylene Glycol. Guidance for Industry, May 2023, Revision 1. Available at: https://www.fda.gov/regulatory-information/search-fda-guidance-documents
  15. United States Food and Drug Administration. Microbiological Quality Considerations in Non-Sterile Drug Manufacturing. Draft guidance for industry, September 2021. Available at: https://www.fda.gov/regulatory-information/search-fda-guidance-documents
  16. United States Food and Drug Administration. Drug Safety Communication: FDA warns consumers against using prescription and OTC skin lightening products containing hydroquinone. Effective September 23, 2020. Available at: https://www.fda.gov/drugs/drug-safety-and-availability
  17. United States Pharmacopeia. USP <51> Antimicrobial Effectiveness Testing. United States Pharmacopeia and National Formulary. [Verify current edition for Category 2 criteria.]
  18. United States Pharmacopeia. USP <60> Tests for Burkholderia Cepacia Complex. United States Pharmacopeia and National Formulary. [Verify current edition.]
  19. United States Pharmacopeia. USP <61>/<62> Microbiological Examination of Nonsterile Products. United States Pharmacopeia and National Formulary. [Verify current edition.]
  20. United States Pharmacopeia. USP <1111> Microbiological Examination of Nonsterile Products: Acceptance Criteria for Pharmaceutical Preparations and Substances for Pharmaceutical Use. United States Pharmacopeia and National Formulary. [Verify current edition.]
  21. United States Pharmacopeia. USP <1112> Application of Water Activity Determination to Nonsterile Pharmaceutical Products. United States Pharmacopeia and National Formulary. [Verify current edition for a_w thresholds.]
  22. United States Pharmacopeia. USP <3> Topical and Transdermal Drug Products — Product Quality Tests. United States Pharmacopeia and National Formulary. [Verify current edition for identity-specificity requirements.]
  23. United States Pharmacopeia. USP <1223> Validation of Alternative Microbiological Methods. United States Pharmacopeia and National Formulary. [Verify current edition.]
  24. ICH Harmonised Tripartite Guideline. Q3C(R8): Impurities — Guideline for Residual Solvents. International Council for Harmonisation, current revision. [Benzene Class 1 limit ≤2 ppm — verify current revision number and effective date.]
  25. United States Code of Federal Regulations. 21 CFR 211.84: Testing and approval or rejection of components, drug product containers, and closures. [Verify current text.]
  26. United States Code of Federal Regulations. 21 CFR 310.545: Drug products containing certain active ingredients offered over-the-counter (OTC) for certain uses. [Verify current text, especially subsection (a)(1).]
  27. United States Code of Federal Regulations. 16 CFR 1700.14: Substances requiring special packaging. Consumer Product Safety Commission. [Verify current text for lidocaine, dibucaine, minoxidil and methyl salicylate thresholds.]
  28. Florida Statutes §499.01(2)(n). OTC drug manufacturer permit and prohibition on possessing prescription drugs. [Verify current statutory text with Florida-licensed regulatory counsel.]
  29. OTC Monograph M016: Skin Protectant Drug Products for Over-the-Counter Human Use. [Verify current text for active windows, combination rules, and ointment-only restriction on ZnO >25–40%.]
  30. OTC Monograph M017: External Analgesic Drug Products for Over-the-Counter Human Use. [Verify current text for caine/amine concentrations, benzyl alcohol active window, and excluded dosage forms.]
  31. OTC Monograph M020: Sunscreen Drug Products for Over-the-Counter Human Use, including Final Order OTC000039 (bemotrizinol, effective August 9, 2026) and Final Order OTC000008-1 (removal of PABA and trolamine salicylate, effective September 11, 2027 or later if disputed). [Verify current posting dates.]
  32. OTC Monograph M006: Acne Drug Products for Over-the-Counter Human Use. [Verify current text for active windows and M006.50 dosage-form naming requirements.]