The evidence table

Every row, every source, every date

This is the whole input to the verdicts. A row is never “niacinamide is good”; it is “niacinamide, for hyperpigmentation”, with the document that gives it standing, the countable facts the evidence dots are derived from, the dose range the studies used, and every source with the date it was opened. A row with a thin trial base says so. An ingredient and concern not on this page is in the research queue.

231rows · table ev-2026.10-r231
91ingredients across 25 concerns
120rows carry at least one trial or systematic review
543sources, each opened on the date shown
193rows checked against their sources with a date on record

Standing answers “is it recognised for this?” and must rest on a document. The dots answer “how much has it been tested?” and are derived by a fixed rule from the trial count, replication, consistency and guideline endorsement recorded on the row — four dots need a guideline endorsement or a high-certainty replicated meta-analysis, three need a trial or a systematic review. Nobody is named on a row. Jump to: Acne · Acne Scarring · Aging · Blackheads · Compromised Skin Barrier · Dandruff · Dark Circles · Dark Spots · Dehydration · Dryness · Dullness · Eczema · Excess Oiliness · Fungal Acne · Hyperpigmentation · Keratosis Pilaris · Large Pores · Melasma · Nail Fungus · Post-Procedure · Psoriasis · Rosacea · Rough Texture · Sensitivity · Sun Damage

Acne 15 rows

Adapalene Established · regulator recognises it, trials Strong Well-established
Standing document
DIFFERIN- adapalene gel (US Food and Drug Administration OTC Drug Facts labeling, via DailyMed (National Library of Medicine)), opened Jul 27, 2026
Counted on the row
3 randomised trials, 900 participants pooled; a systematic review or meta-analysis; independently replicated; findings consistent; guideline endorsement strong; best source tier 1.
Studied dose
0.1%, leave-on, judge after ~12 wk
Dose note
Inflammatory-acne endpoint (papule/pustule reduction). Adapalene is a true topical retinoid, not a cosmetic 'retinol' — the OTC 0.1% gel is the same molecule/strength studied in the pivotal trials.
Vehicle
The OTC Differin 0.1% gel is the identical concentration and molecule used in the registration trials, so — unusually for this table — there is essentially no OTC-vs-Rx indirectness for the 0.1% presentation.
Checked
against its sources on Jul 27, 2026; next review Jul 17, 2027
Sources
Azelaic Acid Established · regulator recognises it, trials Moderate Promising
Standing document
AZELEX® (azelaic acid cream) 20% (US Food and Drug Administration prescribing information, via DailyMed (National Library of Medicine)), opened Jul 27, 2026
Counted on the row
3 randomised trials; independently replicated; findings consistent; guideline endorsement conditional; best source tier 1; GRADE certainty moderate.
Studied dose
10%, leave-on, judge after ~12 wk
Dose note
The acne indication sits with the 20% cream; 15% gel/foam is the rosacea presentation.
Vehicle
Gel delivers more efficiently than cream at equal nominal %, so 15% and 20% are not a clean dose ladder. OTC 10% has not been shown equivalent to Rx 20%.
Checked
against its sources on Jul 27, 2026; next review Jul 17, 2027
Sources
Benzoyl Peroxide Established · regulator recognises it, trials Strong Well-established
Standing document
21 CFR § 333.310 - Acne active ingredients. (US Code of Federal Regulations (Legal Information Institute, Cornell Law School)), opened Jul 27, 2026
Counted on the row
120 randomised trials; a systematic review or meta-analysis; independently replicated; findings consistent; guideline endorsement strong; best source tier 1; GRADE certainty moderate.
Studied dose
2.5–10%, judge after ~6 wk
Dose note
DOSE PLATEAU VERIFIED: 2.5% is equivalent to 5% and 10% on inflammatory lesion reduction, with less irritation than 10% (Mills 1986, n=153, three double-blind studies). Higher concentration buys irritation, not efficacy. 2.5% is the rational default.
Vehicle
Vehicle and particle size modulate irritation at equal nominal %. Bleaches fabric — a real adherence factor. Oxidizes tretinoin on contact; stable beside adapalene.
Checked
against its sources on Jul 27, 2026; next review Jul 17, 2027
Sources
Clascoterone Established · regulator recognises it, trials Moderate Well-established
Standing document
WINLEVI (clascoterone) cream, for topical use — INDICATIONS AND USAGE (U.S. FDA-approved prescription drug label (Sun Pharmaceutical Industries), via DailyMed / U.S. National Library of Medicine), opened Jul 27, 2026
Counted on the row
2 randomised trials, 1,440 participants pooled; findings consistent; guideline endorsement conditional; best source tier 1; GRADE certainty high.
Studied dose
1%, leave-on, judge after ~12 wk
Dose note
Prescription topical androgen-receptor inhibitor — the first new topical acne mechanism in decades.
Vehicle
Both pivotal trials were vehicle-controlled; the vehicle arm is the correct comparator and the drug separated from it significantly on all co-primary endpoints.
Checked
against its sources on Jul 27, 2026; next review Jul 17, 2027
Hectorite Thin · trials Limited Limited
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 3.
Dose note
Rinse-off contact time is short; clay's plausible mechanism is transient sebum/oil absorption, which is a mattifying/sebum property, NOT inflammatory-lesion reduction.
Vehicle
Clay is a rinse-off absorbent; any acne-relevant action is oil/sebum uptake during a brief application, distinct from a sustained anti-inflammatory or comedolytic effect on papule/pustule counts.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Glycolic Acid Thin · trials Moderate Limited
Counted on the row
1 randomised trial, 25 participants pooled; findings insufficient; no guideline endorsement recorded; best source tier 2.
Studied dose
10%, leave-on, judge after ~8 wk
Dose note
Glycolic acid has no prescription form. The ONLY placebo-controlled evidence on the inflammatory-lesion endpoint uses a CLINICIAN-ADMINISTERED 40% peel at pH 2.0, applied 5× at 2-week intervals. Cochrane 2020 describes the studied peel range as 20–70%. This is a PROCEDURE, not a consumer leave-on dose, and the gap is the single biggest caveat on this row.
Vehicle
pH is decisive and must not be collapsed: free-acid fraction (the active species) depends on the finished product's pH, so concentration alone does not describe the dose. The positive placebo-controlled peel was 40% at pH 2.0; the leave-on trial was 10% at pH 4 — NOT interchangeable interventions. Kaminaka's placebo was hydrochloric acid in a polyethylene glycol vehicle — a true low-pH SHAM, an unusually rigorous control that isolates the glycolic acid rather than the acidity. Consumer glycolic products are also frequently co-formulated with other actives.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Mandelic Acid Thin · trials Moderate Limited
Counted on the row
1 randomised trial, 50 participants pooled; findings insufficient; no guideline endorsement recorded; best source tier 3.
Studied dose
5–10%, leave-on
Dose note
The trial evidence is for a high-strength professional peel, not for the low-strength leave-on mandelic products that dominate the catalog. Peel-dose efficacy does not automatically transfer to a 5-10% leave-on.
Vehicle
Mandelic acid's larger molecular size gives slower, more even penetration than glycolic acid, which is the usual tolerability argument for it; but the controlled acne data are peel-format and single-arm-active, not vehicle-controlled leave-on.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Niacinamide Thin · trials Moderate Limited
Counted on the row
4 randomised trials, 374 participants pooled; findings insufficient; no guideline endorsement recorded; best source tier 1.
Studied dose
4–5%, leave-on, judge after ~8 wk
Dose note
The trials tested 4% in a GEL. A 10% niacinamide serum has no acne trial behind it at any level, and nothing in the record establishes dose-response for lesion counts — the concentration ladder is a marketing artifact here, not an evidence gradient.
Vehicle
All the comparative data is on gels, tested against antibiotic gels. Niacinamide is water-soluble and delivery is not the contested variable; the contested variable is whether the trials could detect an effect at all (see notes).
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Retinoids Established · regulator recognises it, trials Strong Well-established
Standing document
RETIN-A- tretinoin cream / RETIN-A- tretinoin gel (US Food and Drug Administration prescribing information, via DailyMed (National Library of Medicine)), opened Jul 27, 2026
Counted on the row
trial count not established; independently replicated; findings consistent; guideline endorsement strong; best source tier 1.
Studied dose
0.1%, leave-on, judge after ~12 wk
Dose note
This row is the retinoid CLASS. Adapalene 0.1% OTC carries genuine acne-indication evidence; do not read this well-established grade onto cosmetic retinol, which is graded separately.
Vehicle
Topical retinoids normalize follicular keratinization and are anti-inflammatory; an initial retinization/purge period is expected and is not treatment failure.
Checked
against its sources on Jul 27, 2026; next review Jul 17, 2027
Sources
Retinol Thin · trials Moderate Limited
Counted on the row
1 randomised trial; findings insufficient; no guideline endorsement recorded; best source tier 3.
Studied dose
0.2%, leave-on, judge after ~12 wk
Dose note
Only one small trial isolates OTC retinol at a stated concentration for inflammatory acne.
Vehicle
The isolating trial is a comparative (retinol vs sodium ascorbyl phosphate vs combination) design, not vehicle/placebo-controlled, so the effect vs no-treatment is not cleanly sized.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Salicylic Acid Established · regulator recognises it, trials Moderate Emerging
Standing document
§ 333.310 Acne active ingredients. (U.S. Food and Drug Administration — 21 CFR Part 333 Subpart D (Topical Acne Drug Products), Code of Federal Regulations via GovInfo), opened Jul 27, 2026
Counted on the row
1 randomised trial; findings insufficient; guideline endorsement conditional; best source tier 1; GRADE certainty low.
Studied dose
0.5–2%, judge after ~6 wk
Dose note
UNVERIFIED — dose range is background knowledge; the FDA OTC monograph was NOT opened this session.
Vehicle
Contact time is the dominant uncontrolled variable. A 2% leave-on and a 2% cleanser are not interchangeable despite the identical nominal %. Any tier for undifferentiated 'salicylic acid' overstates the rinse-off case.
Checked
against its sources on Jul 27, 2026; next review Jul 17, 2027
Sources
Sulfur 10% Established · regulator recognises it, trials Limited Limited
Standing document
§ 333.310 Acne active ingredients. (U.S. Food and Drug Administration — 21 CFR Part 333 Subpart D (Topical Acne Drug Products), Code of Federal Regulations via GovInfo), opened Jul 27, 2026
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 1.
Studied dose
3–10%
Dose note
Doses are regulatory/formulation conventions, not derived from a controlled inflammatory-acne trial.
Vehicle
Sulfur is malodorous and drying and is usually delivered in wash-off or drying-lotion vehicles or combined with sodium sulfacetamide, so any clinical signal is confounded by the co-active and the vehicle.
Checked
against its sources on Jul 27, 2026; next review Jul 17, 2027
Sources
Melaleuca Alternifolia Emerging · trials Moderate Emerging
Counted on the row
2 randomised trials, 184 participants pooled; independently replicated; findings consistent; no guideline endorsement recorded; best source tier 3.
Studied dose
5%, leave-on
Dose note
Endpoint is inflammatory lesion count reduction; both trials reported inflamed and non-inflamed lesion improvement.
Vehicle
Tea tree oil is a volatile essential oil and a known contact sensitizer (oxidized terpinolene/ascaridole); efficacy and irritation both depend on concentration and oxidation state of the oil in the finished product.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Tretinoin 0.025% Established · regulator recognises it, trials Strong Well-established
Standing document
RETIN-A- tretinoin cream / RETIN-A- tretinoin gel (US Food and Drug Administration prescribing information, via DailyMed (National Library of Medicine)), opened Jul 27, 2026
Counted on the row
trial count not established; a systematic review or meta-analysis; independently replicated; findings consistent; guideline endorsement strong; best source tier 1.
Studied dose
0.025–0.1%, leave-on, judge after ~12 wk
Dose note
Prescription-only topical retinoid; the archetypal member of the guideline-recommended topical-retinoid class.
Vehicle
The inflammatory-acne endpoint (papule/pustule reduction) is met; tretinoin is also comedolytic, so it spans both the acne and blackheads claims (graded separately).
Checked
against its sources on Jul 27, 2026; next review Jul 17, 2027
Sources
Zinc Gluconate Thin · trials Moderate Limited
Counted on the row
1 randomised trial, 30 participants pooled; findings insufficient; no guideline endorsement recorded; best source tier 1.
Studied dose
0.1–2%, leave-on
Dose note
The evidence base that people cite for 'zinc treats acne' is almost entirely ORAL zinc (Yee 2020 meta-analysis pooled oral and topical but the effect signal is oral) or ERYTHROMYCIN+zinc combination lotions (Zineryt), where the antibiotic is the active. Neither is a topical cosmetic zinc product. The one clean topical-zinc-monotherapy RCT (Cochran 1985, 2% zinc sulfate) found NO benefit over placebo and MORE irritation.
Vehicle
Topical zinc's cosmetic forms (zinc PCA, zinc gluconate) are studied mostly for sebum output, not inflammatory lesion count; where lesion-count benefit appears it is in oral supplementation or antibiotic-zinc combinations, not leave-on cosmetic zinc alone.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources

Acne Scarring 10 rows

Azelaic Acid Thin No evidence No evidence
Counted on the row
trial count not established; findings insufficient; no guideline endorsement recorded; best source tier 1.
Studied dose
up to 10%, leave-on, judge after ~12 wk
Dose note
No dose of azelaic acid has been studied against a textural/atrophic acne-scar endpoint.
Vehicle
Azelaic acid acts on comedogenesis, inflammation and melanogenesis (tyrosinase). It has no established collagen-remodeling mechanism, which is what atrophic scar improvement would require.
Checked
no verification date on record
Sources
Centella Asiatica Extract Thin · trials Limited Limited
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 3.
Dose note
The acne-scarring endpoint is measured change in ATROPHIC/hypertrophic scar depth/count or a validated scar-severity scale — NOT post-inflammatory pigment, and specifically ACNE scars (predominantly atrophic).
Vehicle
Centella's anti-fibrotic activity is studied on RAISED (hypertrophic/surgical) scars, which are the opposite morphology to the atrophic scars that dominate acne scarring — a delivery/indication mismatch, not a dose question.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Glycolic Acid Thin · trials Moderate Limited
Counted on the row
1 randomised trial, 58 participants pooled; findings insufficient; no guideline endorsement recorded; best source tier 2.
Studied dose
5–10%, leave-on, judge after ~8 wk
Dose note
The catalog products in this claim are OTC leave-on glycolic; the robust scar evidence is procedural 70% peels.
Vehicle
Atrophic-scar improvement is documented for repeated 70% in-office peels (intervention mismatch for a cosmetic leave-on). The only daily-topical arm I confirmed (15% cream, Erbağci 2000) showed weak, non-validated 'some useful effects on scars' and was clearly inferior to the peels in the same study.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Lactic Acid Thin No evidence No evidence
Counted on the row
trial count not established; findings insufficient; no guideline endorsement recorded; best source tier 3.
Studied dose
5–10%, leave-on, judge after ~4 wk
Dose note
The acne-scarring endpoint is measured atrophic scar depth/count or a validated scar scale — NOT texture-smoothing or pigment. Do not confuse lactic acid (an AHA) with POLY-L-lactic acid, an injectable dermal filler.
Vehicle
Superficial AHA chemistry can modestly improve surface texture, but atrophic acne scars are a dermal-volume defect; a superficial leave-on acid does not address scar depth.
Checked
no verification date on record
Sources
Niacinamide Thin No evidence No evidence
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded.
Studied dose
2–10%, leave-on, judge after ~8 wk
Dose note
Concentration is moot — no trial measures a scar-morphology endpoint at any dose.
Vehicle
N/A — no finished-product trial exists for this endpoint.
Checked
no verification date on record
Sources
Retinoids Thin · trials Limited Limited
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 3.
Dose note
The acne-scarring endpoint is measured ATROPHIC scar depth/count or a validated scar scale — not active-lesion reduction and not pigment.
Vehicle
Retinoids' strongest scar role is PREVENTION — treating active acne early prevents scar formation — rather than remodeling established atrophic scars via a leave-on cream.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Retinol Thin · trials Limited Limited
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 3.
Studied dose
0.1–1%, leave-on, judge after ~12 wk
Dose note
No cosmetic-retinol trial with an atrophic-scar endpoint was located.
Vehicle
The atrophic-scar evidence is entirely prescription 3rd/4th-generation retinoids measured against scar-count/SGA endpoints (trifarotene, adapalene). Cosmetic retinol has not been tested against a validated atrophic-scar measure in any trial I could confirm.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Tranexamic Acid Thin No evidence No evidence
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded.
Studied dose
2–5%, leave-on, judge after ~12 wk
Dose note
TXA is also delivered intradermally/via microneedling in the post-acne studies — but those target pigment or erythema, not scar texture.
Vehicle
Tranexamic acid is a plasmin inhibitor whose dermatologic mechanism is on the melanocyte/pigment and coagulation axes, not dermal collagen remodeling; nothing about its mechanism predicts an effect on atrophic/hypertrophic scar architecture.
Checked
no verification date on record
Sources
none recorded
Tretinoin 0.025% Thin · trials Limited Limited
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 3.
Studied dose
0.05–0.1%, leave-on, judge after ~12 wk
Dose note
No trial establishes topical tretinoin monotherapy for atrophic acne scar depth on a validated scar-severity endpoint.
Vehicle
CRITICAL claim separation: tretinoin's strong photoaging-WRINKLE evidence (sun-damage claim) and its PIH pigment evidence (hyperpigmentation claim) must NOT be read as evidence for atrophic/textural SCAR depth. Acne scarring is a measured change in scar depth/count or a validated scar-severity scale — a different endpoint entirely.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Ascorbic Acid Thin · trials Limited Limited
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 3.
Dose note
No vitamin-C-monotherapy dose is defined for an atrophic-scar endpoint.
Vehicle
Every located study either co-formulates vitamin C with tocopherol/ferulic acid or applies it AS AN ADJUNCT to microneedling / ablative CO2 laser — the procedure, not the vitamin C, is the scar-remodelling agent.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources

Aging 16 rows

Adenosine Thin · trials Moderate Limited
Counted on the row
1 randomised trial, 126 participants pooled; findings insufficient; no guideline endorsement recorded; best source tier 3.
Studied dose
1%, leave-on
Dose note
CRITICAL DOSE MISMATCH: the Korean MFDS functional-cosmetic threshold for adenosine is 0.04%, and NO published trial at 0.04% could be located. The only verifiable efficacy dose is 1% — 25× the regulatory floor. Products formulated at the MFDS minimum are NOT supported at their labeled dose by any literature that could be verified.
Vehicle
Abella 2006 gave each subject two of three products (adenosine cream, adenosine dissolvable film, or placebo) periorbitally, twice daily, 8 weeks, FOITS optical profilometry (Ra/Rz) at 0/3/8 weeks. The abstract does not state whether the placebo was a film, a cream, or both — so vehicle matching across arms is unconfirmed. This matters: a dissolvable film is occlusive and hydrating, and occlusion alone transiently reduces measured periorbital roughness, so an unmatched placebo would confound the film arm.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Bakuchiol Thin · trials Moderate Limited
Counted on the row
1 randomised trial, 44 participants pooled; findings insufficient; no guideline endorsement recorded; best source tier 1.
Studied dose
0.5%, leave-on
Dose note
Dose is poorly characterized: 4 of 15 human trials in the JDD 2024 review did not report concentration AT ALL. 0.5% is the only dose with any controlled evidence, from a single trial.
Vehicle
Both arms of the only controlled RCT were creams with NO vehicle-only arm, so the measured wrinkle improvement includes the base's moisturization effect.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Ubiquinone Thin · trials Limited Limited
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 3.
Studied dose
0.3–1%, leave-on
Dose note
Endpoint = INSTRUMENTED wrinkle depth/count or elasticity. In-vitro fibroblast/MMP data is a surrogate and does not count toward this endpoint.
Vehicle
CoQ10 is lipophilic and poorly penetrating; topical delivery is formulation-dependent, so an oral-supplement effect cannot be assumed to transfer to a leave-on cream.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Acacia Seyal Gum Extract Thin Contradicted No evidence
Counted on the row
trial count not established; findings consistent; no guideline endorsement recorded; best source tier 4.
Vehicle
Note the claim-specificity: topical collagen is a perfectly legitimate surface humectant. This row is the DERMAL-REPAIR claim only. A `collagen → dehydration` row would be an ordinary, defensible row.
Checked
no verification date on record
Sources
none recorded
Ferulic Acid Thin · trials Limited Limited
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 3.
Studied dose
0.5–1%, leave-on
Dose note
Endpoint = INSTRUMENTED wrinkle depth/count or elasticity. In-vitro antioxidant assays do not count.
Vehicle
Ferulic acid's established role in these products is as an antioxidant STABILIZER that lowers the pH window and doubles the photoprotection of the vitamin C + E base — a formulation function, not an independent anti-wrinkle actor. Every human aging dataset is the combination, so the vehicle/co-actives are inseparable from any measured effect.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Glycolic Acid Thin · trials Moderate Limited
Counted on the row
2 randomised trials, 116 participants pooled; a systematic review or meta-analysis; findings mixed; no guideline endorsement recorded; best source tier 2.
Studied dose
5–10%, leave-on, judge after ~8 wk
Dose note
The wrinkle-endpoint evidence splits by delivery. The only positive fine-wrinkle result comes from a 50% in-office PEEL, not a leave-on product. The one vehicle-controlled leave-on RCT (5%) found only a non-significant wrinkle trend.
Vehicle
pH and free-acid concentration govern AHA activity; a 50% peel and a 5% neutralized cream are not points on one dose ladder. Reading peel efficacy onto a leave-on product is indirectness_intervention + indirectness_dose.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
SH-Polypeptide-121 Thin · trials Moderate Limited
Counted on the row
9 randomised trials; findings mixed; no guideline endorsement recorded; best source tier 3.
Vehicle
'Growth factors' is not one intervention but a heterogeneous class delivered in wildly different preparations; a large fraction of the trials pair the topical with microneedling or fractional laser, so the topical GF effect cannot be isolated from the procedure.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Lactic Acid Thin · trials Limited Limited
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 1.
Studied dose
5–10%, leave-on, judge after ~4 wk
Dose note
Smith 1996 found dermal thickening and wrinkle-appearance improvement at 12% but explicitly NO dermal changes at 5% — the effect is concentration-dependent and the visible-wrinkle benefit sits at the top of the OTC range.
Vehicle
The wrinkle/dermal-thickness signal is a 12% finding; most catalog lactic-acid leave-ons are below that, so indirectness_dose applies to the lower-strength products.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Niacinamide Thin · trials Moderate Limited
Counted on the row
4 randomised trials, 106 participants pooled; independently replicated; findings consistent; no guideline endorsement recorded; best source tier 3.
Studied dose
4–5%, leave-on, judge after ~8 wk
Dose note
Bissett used 5%; Chiu and Kawada used 4%. No dose-ranging trial exists — 4% and 5% were never compared head-to-head, so dose_response cannot be claimed.
Vehicle
All confirmed trials deliver niacinamide in a moisturizer/cosmetic base, vehicle- or placebo-controlled except Fu 2010 (comparator 0.02% tretinoin, NO vehicle arm). Bissett/Chiu/Kawada are split-face — good control of subject-level confounding, but permits contralateral transfer of a small water-soluble molecule, which no abstract addresses. Kawada measured Ra from skin replicas; Chiu used unspecified 'noninvasive biometrological instruments' (wrinkle counts, not depth).
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Palmitoyl Dipeptide-7 Thin · trials Moderate Limited · contested
Counted on the row
2 randomised trials; a systematic review or meta-analysis; findings mixed; no guideline endorsement recorded; best source tier 2.
Studied dose
0.0003%, leave-on
Dose note
Dose reporting across the topical peptide literature is poor — two of the four human trials verified did not state the concentration of the active peptide at all, making any dose-response claim unsupportable.
Vehicle
Skin penetration is an UNRESOLVED problem for the most-marketed peptides: acetyl hexapeptide-8 showed only 0.22% penetration with the highest concentration remaining in the outer stratum corneum; GHK has a clogP of -2.24, outside the range considered appropriate for skin permeation. For these molecules the vehicle may not be a nuisance variable but the dominant one.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Retinaldehyde Emerging · trials Moderate Emerging
Counted on the row
2 randomised trials, 95 participants pooled; findings consistent; no guideline endorsement recorded; best source tier 3.
Studied dose
0.05–0.1%, leave-on
Dose note
Retinaldehyde is one metabolic step from retinoic acid (retinal to retinoic acid), so it is more potent than retinol but is a cosmetic, not an Rx retinoid — graded on its own trials, not the tretinoin FDA data.
Vehicle
The two RCTs are active-comparator (dose-vs-dose and vs glycolic peel), not vehicle-controlled, and appear manufacturer-linked; retinaldehyde is well tolerated relative to peels and acids.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Retinoids Established · regulator recognises it, trials Strong Well-established
Standing document
RENOVA® (tretinoin cream) 0.02% (US Food and Drug Administration prescribing information, via DailyMed (National Library of Medicine)), opened Jul 27, 2026
Counted on the row
trial count not established; independently replicated; findings consistent; guideline endorsement strong; best source tier 1.
Studied dose
0.02–0.05%, leave-on, judge after ~12 wk
Dose note
The FDA indication is explicitly NARROW: FINE wrinkles only. The label states tretinoin does NOT reverse photoaging and has NOT shown effect on coarse wrinkling, laxity, or lentigines.
Vehicle
Well-established for the fine-wrinkle/texture endpoint under sun-avoidance; NOT established for skin laxity or coarse/deep wrinkling, which the FDA label specifically excludes.
Checked
against its sources on Jul 27, 2026; next review Jul 17, 2027
Sources
Retinol Emerging · trials Moderate Emerging
Counted on the row
9 randomised trials; a systematic review or meta-analysis; independently replicated; findings mixed; no guideline endorsement recorded; best source tier 1.
Studied dose
0.075–0.5%, leave-on, judge after ~12 wk
Dose note
NO verified dose-response gradient — concentrations come from separate trials by different sponsors with different vehicles, endpoints and durations. Higher % cannot be said to outperform lower % from this evidence.
Vehicle
Retinol is chemically unstable and vehicle/stabilization is a FIRST-ORDER efficacy variable — the largest positive dataset is specifically a 'stabilized bioactive' formulation, so results do not transfer across arbitrary retinol products at the same nominal %. Requires two-step conversion (retinol -> retinaldehyde -> retinoic acid).
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Tretinoin 0.025% Established · regulator recognises it, trials Strong Well-established
Standing document
REFISSA- tretinoin cream (Tretinoin Cream, USP (Emollient) 0.05%) — INDICATIONS AND USAGE (US FDA prescribing information via DailyMed (NLM)), opened Jul 27, 2026
Counted on the row
8 randomised trials, 1,361 participants pooled; a systematic review or meta-analysis; independently replicated; findings consistent; guideline endorsement strong; best source tier 1.
Studied dose
0.02–0.05%, leave-on, judge after ~12 wk
Dose note
No verified head-to-head gradient establishing monotonic dose-response on wrinkle endpoints — dose_response NOT claimed as an upgrade.
Vehicle
Tretinoin IS retinoic acid and needs no intracellular conversion — the mechanistic reason its evidence is both stronger than retinol's and NON-TRANSFERABLE to it.
Checked
against its sources on Jul 27, 2026; next review Jul 17, 2027
Ascorbic Acid Thin · trials Moderate Limited
Counted on the row
2 randomised trials; findings consistent; no guideline endorsement recorded; best source tier 2.
Dose note
The randomized on-endpoint trials used 5% (Humbert) and an unstated Cellex-C concentration (Traikovich). No verified RCT tested the commercially dominant 15-20% L-AA + vitamin E + ferulic acid serum against vehicle on a wrinkle endpoint.
Vehicle
Same pH<3.5 / 20%-ceiling pharmacology as the pigment row. Formulation dependence is unusually consequential: the JDD 2023 review found EVERY included study used vitamin C combined with other actives, so vehicle and co-actives cannot be separated from the vitamin C effect ANYWHERE in this base. ADDED Oct 2026: the two counted RCTs are vehicle-controlled within-subject designs (Traikovich hemiface; Humbert side vs side). Humbert tested vitamin C as the sole active against its excipient, so the 'every study is a combination' finding of the JDD 2023 review describes that review's included set, not these two trials.
Checked
no verification date on record
Sources
Tocopheryl Acetate Thin · trials Limited Limited
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 3.
Studied dose
1%, leave-on
Dose note
At the concentrations vitamin E usually appears at in the catalog it is functioning as a FORMULATION antioxidant — protecting the product's oil phase from rancidity — not as an active dosed at a skin-efficacy level. A front-label 'Vitamin E' aging claim and a tocopherol entry near the end of an INCI list are usually two different facts about the same jar.
Vehicle
The common cosmetic form is TOCOPHERYL ACETATE, an ester, which is not itself the antioxidant — it requires enzymatic hydrolysis in skin to liberate free α-tocopherol. Flagging as a caveat, NOT as an established finding: I did not verify the conversion efficiency against a primary source this session (see research_todo). The distinction matters because the mechanism this row anchors on belongs to free α-tocopherol, while the ingredient in most products is the ester.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources

Blackheads 10 rows

Adapalene Established · trials Moderate Well-established
Counted on the row
3 randomised trials, 900 participants pooled; a systematic review or meta-analysis; independently replicated; findings consistent; no guideline endorsement recorded; best source tier 1.
Studied dose
0.1%, leave-on, judge after ~12 wk
Dose note
Comedonal endpoint = reduction in open/closed comedone (non-inflammatory lesion) count. This is a co-primary endpoint in every acne registration trial and the DEFINING action of a retinoid.
Vehicle
OTC 0.1% gel equals the studied strength; minimal OTC-vs-Rx indirectness for the comedone endpoint.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Charcoal Powder Thin No evidence No evidence
Counted on the row
trial count not established; findings insufficient; no guideline endorsement recorded; best source tier 3.
Dose note
Endpoint = reduction in open/closed comedone count or measured follicular-plug clearance. No study measures this for charcoal.
Vehicle
Charcoal appears almost entirely in rinse-off vehicles (masks, cleansers), where contact time is seconds to minutes — the format least able to affect a follicular plug even if a mechanism existed.
Checked
no verification date on record
Sources
Hectorite Thin · trials Limited Limited
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 3.
Dose note
The blackheads endpoint is reduction in open/closed COMEDONE count or measured follicular-plug clearance — not total acne lesion count, and not the immediate 'pore-strip' surface effect.
Vehicle
The idea that clay 'draws out' follicular plugs is the pore-strip myth: rinse-off masks remove surface debris and sebum casts transiently but do not clear the deep follicular plug that defines a comedone.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Glycolic Acid Thin · trials Moderate Limited
Counted on the row
2 randomised trials, 55 participants pooled; independently replicated; findings insufficient; no guideline endorsement recorded; best source tier 2.
Studied dose
5–10%, leave-on, judge after ~8 wk
Dose note
EVERY trial with a comedonal endpoint used an in-office PEEL (Kaminaka 40% at pH 2.0, applied 5 times at 2-week intervals; Ilknur 20-70%). There is no controlled comedone-count evidence for a 7% leave-on toner. This is the central gap for the 177 catalog products this row touches.
Vehicle
pH and contact time are decisive for an acid and are the whole difference between the evidence and the product. A 40% pH-2.0 peel neutralised after minutes by a clinician is a procedure; a 7% pH-3.8 toner left on the skin is a cosmetic. Reading the peel literature onto the toner is a dose/vehicle extrapolation, not a finding.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Lactic Acid Thin · trials Limited Limited
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 3.
Studied dose
5–10%, leave-on, judge after ~4 wk
Dose note
Comedone/blackhead reduction (open/closed comedone count) is the endpoint here — distinct from inflammatory acne and from the physicochemical 'exfoliation' job.
Vehicle
AHA keratolysis is pH- and concentration-dependent; the controlled comedone signal comes from a carbonated glycolic+lactic combination, so lactic acid's independent contribution is not isolable.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Mandelic Acid Thin · trials Moderate Limited
Counted on the row
1 randomised trial, 50 participants pooled; findings insufficient; no guideline endorsement recorded; best source tier 3.
Studied dose
45%, leave-on
Dose note
The trial dose is a serial in-office peel, materially different from the leave-on mandelic products in the catalog.
Vehicle
Active-controlled (45% mandelic peel vs 30% salicylic peel), no vehicle arm, and delivered as a professional peel rather than a leave-on.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Retinoids Established · trials Strong Well-established
Counted on the row
trial count not established; independently replicated; findings consistent; guideline endorsement strong; best source tier 1.
Studied dose
0.1%, leave-on, judge after ~12 wk
Dose note
Blackheads = open/closed comedone count. Unlike most ingredients (where acne and comedone evidence must be graded apart), retinoids are the one class whose comedolytic action is directly and abundantly measured.
Vehicle
Retinoids act on the microcomedo — the precursor lesion — which is precisely why their comedone-count reduction is well documented rather than inferred from inflammatory-lesion data.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Retinol Thin · trials Limited Limited
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 3.
Studied dose
0.1–1%, leave-on, judge after ~12 wk
Dose note
Retinol is a cosmetic precursor that must convert (retinol → retinaldehyde → retinoic acid) at a small fraction of tretinoin's potency. No cosmetic-retinol comedone-count trial was located.
Vehicle
The comedone-clearance literature is entirely prescription tretinoin/adapalene/tazarotene in gel/cream/lotion vehicles. Whether OTC retinol at cosmetic concentrations produces a measurable comedone-count reduction has not been tested in a controlled trial I could confirm.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Salicylic Acid Established · regulator recognises it, trials Moderate Emerging
Standing document
21 CFR § 333.350 - Labeling of acne drug products. (US Code of Federal Regulations (Legal Information Institute, Cornell Law School)), opened Jul 27, 2026
Counted on the row
1 randomised trial, 500 participants pooled; findings insufficient; no guideline endorsement recorded; best source tier 1.
Studied dose
0.5–2%, judge after ~6 wk
Dose note
The OTC monograph range. Below 0.5% is under-dosed; the pixie-dusting proxy applies when undisclosed.
Vehicle
Requires a functional pH (≤ ~4) to act as an exfoliant; neutralized formulations are inert for this claim (see inci-flag-rules → antagonism_pairs).
Checked
against its sources on Jul 27, 2026; next review Jul 17, 2027
Sources
Tretinoin 0.025% Established · trials Strong Well-established
Counted on the row
trial count not established; independently replicated; findings consistent; guideline endorsement strong; best source tier 1.
Studied dose
0.025–0.1%, leave-on, judge after ~12 wk
Dose note
This row is Rx TRETINOIN specifically. OTC retinol is a separate ingredient with separate (weaker) evidence — do not merge the two.
Vehicle
Comedolysis is tretinoin's defining mechanism: it normalizes follicular keratinocyte differentiation and cohesion, dissolving the microcomedo that seeds both open and closed comedones. This is the single best-established action of topical retinoids.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources

Compromised Skin Barrier 4 rows

Ceramides NP, AP, EOP Established · mechanism established Limited Established mechanism
Counted on the row
trial count not established; no guideline endorsement recorded; best source tier 2.
Studied dose
0.5–1%, leave-on, judge after ~4 wk
Vehicle
Only one trial tests a near-isolated ceramide (synthetic ceramide-2 at 0.5%/1% vs placebo emulsion). The others are multi-lipid finished formulations.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
  • Elias PM, Skin Barrier / Coderch et al., ceramides and skin barrier function — ceramides ~50% of SC intercellular lipid mass, essential to the permeability barrier — reference work
Dimethicone Established · mechanism established Limited Established mechanism
Counted on the row
trial count not established; no guideline endorsement recorded; best source tier 2.
Studied dose
1–15%, leave-on
Dose note
FDA-recognised skin protectant.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
  • International Cosmetic Ingredient Dictionary and Handbook — declared function — Personal Care Products Council
  • United States Pharmacopeia / National Formulary — excipient monograph — USP-NF
Niacinamide Established · mechanism established Limited Established mechanism
Counted on the row
trial count not established; no guideline endorsement recorded; best source tier 1.
Studied dose
2%, leave-on, judge after ~8 wk
Dose note
The TEWL evidence base sits at ~2%. NO trial with a TEWL endpoint exists at the 4-10% concentrations typical of retail niacinamide serums. Do not assume the 2% barrier finding scales upward.
Vehicle
SEVERE CONFOUNDING: Christman 2012 tested niacinamide/GLYCERIN moisturizers against OTHER moisturizers — not against their own vehicle — so glycerin and the lipid base are inseparable from niacinamide. Only Soma 2005 (2% vs white petrolatum) isolates it reasonably, and its comparator is an occlusive rather than a matched vehicle.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Petrolatum Established · mechanism established Limited Established mechanism
Counted on the row
trial count not established; no guideline endorsement recorded; best source tier 2.
Studied dose
1–100%, leave-on, judge after ~1 wk
Dose note
TEWL reduction by occlusion is a measured physical property.
Vehicle
This row is occlusion (sealing water in). It is NOT a claim to rebuild barrier lipids — that is clinical and contested.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
  • International Cosmetic Ingredient Dictionary and Handbook — declared function — Personal Care Products Council
  • United States Pharmacopeia / National Formulary — excipient monograph — USP-NF
  • Harry's Cosmeticology / Baran & Maibach Textbook of Cosmetic Dermatology — humectant, emollient and occlusive classification — standard reference text

Dandruff 10 rows

Ciclopirox Established · regulator recognises it, trials Moderate Promising
Standing document
CICLOPIROX SHAMPOO [SUN PHARMACEUTICAL INDUSTRIES, INC.] — Ciclopirox Shampoo 1% (DailyMed (US National Library of Medicine) — FDA prescription drug label), opened Aug 11, 2026
Counted on the row
11 randomised trials, 3,029 participants pooled; a systematic review or meta-analysis; findings consistent; guideline endorsement conditional; best source tier 1; GRADE certainty moderate.
Studied dose
1–1.5%, judge after ~4 wk
Dose note
Trials used timed rinse-off contact (typically 2-3 minutes, twice weekly) — contact time is part of the dosing, per the trial designs.
Vehicle
Rinse-off shampoo IS the studied vehicle for this claim — dandruff shampoo RCTs build timed contact into the design, so rinse-off does not void the claim (per claims-vocabulary note). Leave-on ciclopirox products for scalp are NOT covered by this evidence.
Checked
against its sources on Aug 11, 2026
Sources
Climbazole Thin · trials Moderate Limited
Counted on the row
1 randomised trial, 620 participants pooled; findings insufficient; no guideline endorsement recorded; best source tier 3.
Studied dose
0.5–2%, leave-on
Dose note
No dose-response data in humans. In-vitro Malassezia susceptibility work exists but establishes mechanism plausibility, not clinical efficacy.
Vehicle
Rinse-off shampoo is the intended vehicle and dandruff trials legitimately test rinse-off with timed contact — but climbazole has no vehicle-controlled trial in ANY vehicle, so vehicle questions are moot until a controlled monotherapy trial exists.
Checked
no verification date on record
Coal Tar 0.5% Established · regulator recognises it, trials Limited Limited
Standing document
21 CFR § 358.710 - Active ingredients for the control of dandruff, seborrheic dermatitis, or psoriasis (US Code of Federal Regulations — FDA OTC drug monograph (Cornell LII rendering of the current CFR)), opened Aug 11, 2026
Counted on the row
trial count not established; findings mixed; guideline endorsement conditional; best source tier 1.
Studied dose
0.5–5%, judge after ~4 wk
Dose note
Verified trials sit at both ends of the monograph range (0.5% and 4%) with different comparative outcomes — dose-effect within the range is not established.
Vehicle
Shampoo with timed contact is the studied and monographed vehicle. Coal tar's mechanism (keratoplastic/antiproliferative rather than primarily antifungal) means it also carries the scalp-psoriasis indication — but this row records only the dandruff/seb-derm claim.
Checked
against its sources on Aug 11, 2026
Sources
Ketoconazole 1% Established · regulator recognises it, trials Strong Well-established
Standing document
1% Ketoconazole Anti Dandruff Shampoo (DailyMed (US National Library of Medicine) — FDA OTC human drug label), opened Aug 11, 2026
Counted on the row
8 randomised trials, 2,520 participants pooled; a systematic review or meta-analysis; findings consistent; guideline endorsement strong; best source tier 1; GRADE certainty low.
Studied dose
1–2%, judge after ~4 wk
Dose note
The Cochrane-pooled and guideline-recommended strength is 2%. Trials build timed rinse-off contact (3-5 minutes) into dosing.
Vehicle
Rinse-off shampoo with timed contact is the studied vehicle — per the claims vocabulary, rinse-off does not void the dandruff claim. Ketoconazole in leave-on cosmetic formats or non-scalp presentations is outside this evidence.
Checked
against its sources on Aug 11, 2026
Sources
Piroctone Olamine Thin · trials Moderate Limited
Counted on the row
2 randomised trials; findings consistent; no guideline endorsement recorded; best source tier 3.
Studied dose
0.5–1%
Dose note
EU-standard antidandruff active. NOT in the US FDA OTC dandruff monograph — verified against the 21 CFR 358.710 active-ingredient list this session, which names only coal tar, pyrithione zinc, salicylic acid, selenium sulfide, and sulfur.
Vehicle
Rinse-off shampoo is the trial vehicle and, per the claims vocabulary, does not void the dandruff claim — timed rinse-off exposure is part of the trial design for this class.
Checked
no verification date on record
Salicylic Acid Established · regulator recognises it, trials Strong Promising
Standing document
Active ingredients for the control of dandruff, seborrheic dermatitis, or psoriasis. (21 CFR § 358.710, Code of Federal Regulations (Cornell Law School LII)), opened Aug 11, 2026
Counted on the row
2 randomised trials, 202 participants pooled; findings consistent; guideline endorsement strong; best source tier 1.
Studied dose
1.8–3%, judge after ~6 wk
Dose note
Keratolytic, not antifungal — it detaches flakes rather than suppressing Malassezia. A salicylic-acid-only product treats the scale, not the driver.
Vehicle
Rinse-off shampoo is the monograph and trial presentation; contact time is built into the trial design for this claim, so rinse-off does not void it.
Checked
against its sources on Aug 11, 2026
Selenium Sulfide 1% Established · regulator recognises it, trials Strong Well-established
Standing document
Active ingredients for the control of dandruff, seborrheic dermatitis, or psoriasis. (21 CFR § 358.710, Code of Federal Regulations (Cornell Law School LII)), opened Aug 11, 2026
Counted on the row
1 randomised trial, 246 participants pooled; independently replicated; findings insufficient; guideline endorsement strong; best source tier 1.
Studied dose
0.6–1%, judge after ~4 wk
Dose note
The confirmed tier-1 RCT tested the 2.5% Rx strength; OTC products are 1% or 0.6% micronized, hence indirectness_dose for the shelf presentation.
Vehicle
Rinse-off shampoo with timed contact is the trial and monograph presentation; rinse-off does not void this claim.
Checked
against its sources on Aug 11, 2026
Sources
Sulfur 10% Established · regulator recognises it, trials Strong Promising
Standing document
Active ingredients for the control of dandruff, seborrheic dermatitis, or psoriasis. (21 CFR § 358.710, Code of Federal Regulations (Cornell Law School LII)), opened Aug 11, 2026
Counted on the row
1 randomised trial, 48 participants pooled; findings insufficient; guideline endorsement strong; best source tier 1.
Studied dose
2–5%
Dose note
The monograph distinction matters for product copy: sulfur may claim dandruff control, not seb-derm control. Combination with salicylic acid is explicitly permitted within each ingredient's range.
Vehicle
Rinse-off shampoo; timed contact is part of the trial design for this claim.
Checked
against its sources on Aug 11, 2026
Tea Tree Oil Thin · trials Moderate Limited
Counted on the row
1 randomised trial, 126 participants pooled; findings insufficient; no guideline endorsement recorded; best source tier 1.
Studied dose
5%
Dose note
Not an FDA monograph dandruff active — verified against the 21 CFR 358.710 list this session. Products marketing tea tree for flakes without a monograph active grade on this row's evidence, not the antifungal class's.
Vehicle
Rinse-off shampoo with daily use in the trial; rinse-off does not void the dandruff claim per the vocabulary.
Checked
no verification date on record
Sources
Zinc Pyrithione Established · regulator recognises it, trials Strong Well-established
Standing document
§ 358.710 Active ingredients for the control of dandruff, seborrheic dermatitis, or psoriasis. (US Code of Federal Regulations (Legal Information Institute, Cornell Law School)), opened Aug 11, 2026
Counted on the row
2 randomised trials, 951 participants pooled; independently replicated; findings consistent; guideline endorsement strong; best source tier 1.
Studied dose
0.3–2%, judge after ~4 wk
Dose note
Concentration bands are format-specific in the monograph itself — a leave-on at rinse-off percentages is outside the GRASE determination.
Vehicle
Rinse-off contact time is built into the trial designs and the monograph ("applied and then washed off after brief exposure"), so rinse-off format does NOT void this claim — the vocabulary's explicit exception for dandruff. Deposition matters: ZPT is a particulate active, and formulation work (e.g. P&G's 'potentiated ZPT') shows delivery to the follicular infundibulum varies materially between shampoos at the same nominal percentage. Separately: the EU banned zinc pyrithione in cosmetics from March 2022 (CMR reclassification) — US OTC drug-monograph shampoos are unaffected, but EU-market reformulations mean catalog SKUs may differ by region (context note; not entered as evidence).
Checked
against its sources on Aug 11, 2026
Sources

Dark Circles 7 rows

Caffeine Thin · trials Limited Limited
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 2.
Studied dose
1–3%, leave-on
Dose note
No caffeine-only concentration has ever been tested against a control for dark circles. Any '3% caffeine' claim traces to a pad that also contained vitamin K.
Vehicle
Emu oil base with vitamin K in the one controlled study — an emollient vehicle plus a second putative vascular active. Any measured change in periorbital appearance could be emolliency, vitamin K, or caffeine, and the design cannot separate them.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Niacinamide Thin · trials Limited Limited
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 2.
Studied dose
5%, leave-on, judge after ~8 wk
Dose note
Quoting '5% niacinamide' on a dark-circle row would import a full-face spot dose onto a periorbital claim. That is citation drift in miniature — the same move that produced the 'niacinamide shrinks pores' myth from a pigmentation paper.
Vehicle
Not assessable for this claim — no niacinamide-attributable periorbital trial exists. The periorbital region is thinner and more permeable than the cheek, so even a well-established full-face result would not transfer without testing.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Palmitoyl Dipeptide-7 Thin · trials Limited Limited
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 3.
Dose note
NO peptide concentration is stated in ANY human study bearing this claim. All three uncontrolled studies describe only 'bioavailable peptides', 'TriHex peptides' or 'OS-01' without a percentage. The specific doses circulating in commerce (Eyeseryl/acetyl tetrapeptide-5 at 0.01%, Haloxyl at 3%) originate from SUPPLIER MARKETING SHEETS (Lipotec, Sederma), not from any peer-reviewed trial — tier 4, inadmissible.
Vehicle
NO vehicle-controlled trial of a topical peptide for dark circles exists. All three human studies with a dark-circle endpoint are single-arm, open-label, against baseline only. Because periorbital appearance responds to hydration and simple emollient plumping of thin infraorbital skin, a baseline-controlled design CANNOT separate a peptide effect from a vehicle effect. Every peptide is delivered in a multi-active matrix (vitamin C, caffeine, chrysin, botanicals), so even a positive result could not be attributed to the peptide.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Retinoids Thin · trials Limited Limited
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 3.
Dose note
Dark-circles endpoint requires distinguishing vascular vs pigmented subtype and measuring periorbital pigment or a validated scale — retinoids have neither an isolated nor subtype-matched trial here.
Vehicle
Periorbital skin is thin and retinoid-sensitive; retinoic acid on the eyelids is a documented cause of periorbital irritant/allergic contact dermatitis, so tolerability, not just efficacy, is a concern.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Retinol Thin · trials Limited Limited
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 3.
Dose note
No isolated-retinol dark-circle dose is defined; the one study combines retinol with a botanical.
Vehicle
The single located study is UNCONTROLLED (single-arm before/after, no vehicle or comparator), so no effect can be attributed to retinol or even to the product over time/placebo.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Ascorbic Acid Thin · trials Moderate Limited
Counted on the row
1 randomised trial, 14 participants pooled; findings insufficient; no guideline endorsement recorded; best source tier 3.
Studied dose
10%, leave-on, judge after ~12 wk
Dose note
Endpoint requires distinguishing the vascular vs pigmented dark-circle subtype — evidence must match subtype or it is indirect.
Vehicle
Form matters decisively: in the one trial, 10% sodium ascorbate acted on the VASCULAR subtype by thickening the dermis (masking bluish discoloration), while ascorbic acid glucoside did nothing — so results do not generalize across vitamin C derivatives.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Phytonadione Thin · trials Limited Limited
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 3.
Studied dose
1%, leave-on
Dose note
Endpoint requires distinguishing vascular vs pigmented dark-circle subtype; the vitamin K rationale (aiding clearance of extravasated blood) targets the vascular subtype but is not demonstrated in isolation.
Vehicle
Every human dark-circle study of vitamin K combines it with caffeine, retinol, or vitamins C/E, so its independent contribution is unmeasured.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027

Dark Spots 11 rows

Alpha-Arbutin Thin · trials Limited Limited
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 3.
Studied dose
1–5%, leave-on, judge after ~12 wk
Dose note
No controlled trial isolates alpha-arbutin's effect on discrete solar lentigines with a validated spot/colorimetry endpoint.
Vehicle
The one study that names 'dark spots' (PMID 39943675) is a NON-comparative, uncontrolled interventional study of a trihydroxybenzoic-acid-glucoside + alpha-arbutin formulation applied with sunscreen, lumping 'facial melasma OR dark spots' - no control arm, no ingredient isolation, and the sunscreen alone confounds any lentigines benefit.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Azelaic Acid Thin · trials Moderate Limited
Counted on the row
1 randomised trial, 50 participants pooled; findings insufficient; no guideline endorsement recorded; best source tier 2.
Studied dose
10%, leave-on, judge after ~12 wk
Dose note
Solar lentigines are a structural/keratinocyte-driven pigment lesion; azelaic acid's tyrosinase-modulating mechanism addresses melanocyte activity, which the trial's authors concluded does not resolve the structural component of lentigines.
Vehicle
The negative result is for discrete solar lentigines specifically. It must NOT be read as evidence against azelaic acid for melasma or post-inflammatory hyperpigmentation, which are distinct claims with their own (separate) evidence bases.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Glycolic Acid Thin · trials Limited Limited
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 3.
Studied dose
5–10%, leave-on, judge after ~8 wk
Dose note
Endpoint = measured lightening of DISCRETE lentigines / age spots. Most GA pigment data is in melasma or diffuse photoaging (different claims) or as professional peels (different intervention).
Vehicle
The only lentigo-specific positive signal is a 4-active combination serum, not a glycolic-acid leave-on; the vehicle and co-actives, not GA alone, may carry any effect.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Hydroquinone Emerging · trials Moderate Emerging
Counted on the row
1 randomised trial, 216 participants pooled; findings insufficient; no guideline endorsement recorded; best source tier 3.
Studied dose
2%, leave-on, judge after ~12 wk
Dose note
For discrete solar lentigines the better-studied topical is mequinol 2%/tretinoin 0.01% (Solage); HQ 3% is the active comparator, not the lead agent.
Vehicle
Solar lentigines are a discrete, well-demarcated pigment lesion distinct from diffuse melasma or post-inflammatory pigment; evidence must be measured on lentigines specifically to count here.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Kojic Acid Thin · trials Limited Limited
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 3.
Studied dose
1–4%, leave-on, judge after ~12 wk
Dose note
No trial defines a kojic-acid dose against a discrete solar-lentigo (colorimetry) endpoint.
Vehicle
Kojic acid is unstable and readily oxidizes; most positive data comes from combination formulas (kojic + hydroquinone), so the vehicle/co-active confounds any single-ingredient attribution.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Glycyrrhetinic Acid Thin No evidence No evidence
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 2.
Dose note
COULD NOT ESTABLISH any dose range for this claim, because no trial of licorice on discrete lentigines exists to have used one. Note also that 'licorice root extract' is not a defined dose of anything — glabridin content varies by extract.
Vehicle
Not assessable — no clinical trial on this endpoint exists.
Checked
no verification date on record
Sources
Niacinamide Thin · trials Limited Limited
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 2.
Dose note
No dose can be reported because NO clinical trial of niacinamide on discrete lentigines exists at any concentration. The doses circulating in the niacinamide pigment literature (5% half-face for facial hyperpigmentation; 5% for photoaged 'hyperpigmented spots'; 2% in a sunscreen for UV-induced tanning) all belong to OTHER endpoints and must not be carried over.
Vehicle
Not establishable — there is no on-endpoint trial from which to describe a vehicle, concentration, pH or duration. For the reviewer: niacinamide's proposed mechanism (blocking melanosome transfer, suppressing keratinocyte IL-6/ET-1) acts on pigment TRANSFER in actively pigmenting skin. Solar lentigines are architecturally distinct lesions with elongated rete ridges and a resident hyperactive melanocyte population — a plausible biological reason the mechanism may not transfer to this endpoint. That reasoning is hypothesis, not evidence; it is offered only to show the endpoint separation is not a technicality.
Checked
no verification date on record
Sources
Retinoids Thin · trials Moderate Limited
Counted on the row
1 randomised trial, 58 participants pooled; independently replicated; findings insufficient; no guideline endorsement recorded; best source tier 1.
Studied dose
0.1%, leave-on, judge after ~12 wk
Dose note
FAMILY ROW. 'Retinoids' spans Rx tretinoin/tazarotene (real lentigines evidence) and OTC retinol (none). The endpoint-matching data belongs to the specific Rx member, not the class.
Vehicle
The one lentigines RCT used Rx tretinoin; transferring it to the OTC members of the family is exactly the Rx-vs-OTC indirectness the system guards against, hence the indirectness_intervention downgrade.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Retinol Thin · trials Limited Limited
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 3.
Studied dose
0.1–0.5%, leave-on, judge after ~12 wk
Dose note
No OTC retinol monotherapy trial with a discrete-lentigines endpoint exists at any dose.
Vehicle
N/A — no OTC-retinol finished-product trial with a discrete solar-lentigines endpoint exists.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Tretinoin 0.025% Emerging · trials Moderate Emerging
Counted on the row
1 randomised trial, 58 participants pooled; independently replicated; findings insufficient; no guideline endorsement recorded; best source tier 1.
Studied dose
0.1%, leave-on, judge after ~12 wk
Dose note
Prescription-only; do not transfer to OTC retinol.
Vehicle
Endpoint-matching evidence is for discrete solar lentigines ('liver spots'), the exact dark-spots endpoint. Distinct from tretinoin's photoaging-wrinkle evidence (sun-damage claim) and its PIH evidence (hyperpigmentation claim).
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Ascorbic Acid Thin · trials Moderate Limited
Counted on the row
2 randomised trials; findings mixed; no guideline endorsement recorded; best source tier 2.
Dose note
No dose-response can be derived; every on-endpoint trial used a different vitamin C entity at a single concentration. CORRECTION to the prior note: derivatives HAVE been tested alone vs vehicle on lentigines (Miyamoto 2002, 3% MAP, n=120; Ishikawa 2019, 6% ascorbate-2-phosphate), but neither states randomized allocation. What has never been tested alone vs vehicle on lentigines is free high-percentage L-ascorbic acid, the dominant commercial form. Miyamoto also found spot area returned to pre-treatment levels 6 months after stopping.
Vehicle
All leave-on lotions/serums, twice daily, intra-individual split-body or half-face. Three of four use pH-independent esterified/glycosylated derivatives (ascorbate-2-phosphate, ascorbyl glucoside, ascorbyl glucosamine) rather than free L-AA; these require enzymatic cleavage in skin, so Pinnell 2001's pH<3.5 / ~20% absorption pharmacokinetics for free L-AA does NOT transfer to them. Khemis (free L-AA 10%) recorded 23 adverse events across 15 of 30 subjects, six possibly product-related and mostly halo depigmentation — a formulation-specific safety signal absent from the derivative lotions.
Checked
no verification date on record
Sources

Dehydration 9 rows

Aloe Barbadensis Leaf Extract Established · mechanism established Limited Measured
Counted on the row
trial count not established; no guideline endorsement recorded; best source tier 2.
Dose note
Predominantly water + polysaccharides; hydration is real but the extract is not standardised, so the mechanism is not definitional in the way glycerin's is.
Vehicle
Deliberately 'measured' not 'established-mechanism': botanical extracts vary batch to batch and the INCI name does not pin the composition.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
  • International Cosmetic Ingredient Dictionary and Handbook — declared function — Personal Care Products Council
  • Cosmetic Ingredient Review — safety assessment and reported function — CIR Expert Panel
Coco-Betaine Established · mechanism established Limited Established mechanism
Counted on the row
trial count not established; no guideline endorsement recorded; best source tier 2.
Studied dose
0.5–5%, leave-on
Dose note
Osmolyte humectant.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
  • International Cosmetic Ingredient Dictionary and Handbook — declared function — Personal Care Products Council
Butylene Glycol Established · mechanism established Limited Established mechanism
Counted on the row
trial count not established; no guideline endorsement recorded; best source tier 2.
Studied dose
1–10%, leave-on
Dose note
Humectant/solvent.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
  • International Cosmetic Ingredient Dictionary and Handbook — declared function — Personal Care Products Council
Ethylhexylglycerin Established · mechanism established Limited Established mechanism · contested
Counted on the row
trial count not established; no guideline endorsement recorded; best source tier 2.
Studied dose
20%, leave-on, judge after ~1 wk
Vehicle
Glycerin IS the vehicle in many formulations, which makes vehicle-controlled designs structurally awkward.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
  • USP glycerin monograph / Fluhr et al., glycerol and the skin (review) — glycerin is the reference humectant, drawing and holding water in the stratum corneum — reference work
Sodium Hyaluronate Established · mechanism established Limited Established mechanism
Counted on the row
trial count not established; no guideline endorsement recorded; best source tier 2.
Studied dose
0.1%, leave-on, judge after ~1 wk
Dose note
Molecular weight is a live variable and is almost never disclosed on a label.
Vehicle
Any HA effect is plausibly downstream of the humectant vehicle; no trial cleanly separates the polymer from its base.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
  • Draelos ZD, Cosmetic Dermatology: Products and Procedures — hyaluronic acid classified as a humectant water-binding agent holding many times its weight in water — reference work
Panthenol Established · mechanism established Limited Established mechanism
Counted on the row
trial count not established; no guideline endorsement recorded; best source tier 2.
Studied dose
2.5%, leave-on
Dose note
The 2.5% figure comes from the persisted output — re-verify.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
  • Cosmetic Ingredient Review assessment of panthenol / dexpanthenol — humectant and moisturizer that raises stratum-corneum water content — reference work
Dipropylene Glycol Established · mechanism established Limited Established mechanism
Counted on the row
trial count not established; no guideline endorsement recorded; best source tier 2.
Studied dose
1–10%, leave-on
Dose note
Humectant/solvent.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
  • International Cosmetic Ingredient Dictionary and Handbook — declared function — Personal Care Products Council
  • United States Pharmacopeia / National Formulary — excipient monograph — USP-NF
Glycine, Serine, Alanine, Sodium PCA Established · mechanism established Limited Established mechanism
Counted on the row
trial count not established; no guideline endorsement recorded; best source tier 2.
Studied dose
0.2–4%, leave-on
Dose note
A component of skin's own natural moisturizing factor.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
  • International Cosmetic Ingredient Dictionary and Handbook — declared function — Personal Care Products Council
  • Harry's Cosmeticology / Baran & Maibach Textbook of Cosmetic Dermatology — humectant, emollient and occlusive classification — standard reference text
Hydroxyethyl Urea Established · mechanism established Limited Established mechanism
Counted on the row
trial count not established; no guideline endorsement recorded; best source tier 2.
Studied dose
2–10%, leave-on, judge after ~2 wk
Dose note
Dose changes the claim, not just the strength: low-dose urea hydrates, high-dose exfoliates.
Vehicle
The dose boundary here is a claim boundary. A 10%+ urea product should be graded on rough-texture, not only dehydration.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
  • International Cosmetic Ingredient Dictionary and Handbook — declared function — Personal Care Products Council
  • United States Pharmacopeia / National Formulary — excipient monograph — USP-NF
  • Harry's Cosmeticology / Baran & Maibach Textbook of Cosmetic Dermatology — humectant, emollient and occlusive classification — standard reference text

Dryness 12 rows

Caprylic/Capric Triglyceride Established · mechanism established Limited Established mechanism
Counted on the row
trial count not established; no guideline endorsement recorded; best source tier 2.
Studied dose
1–30%, leave-on
Dose note
Emollient ester.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
  • International Cosmetic Ingredient Dictionary and Handbook — declared function — Personal Care Products Council
Ceramides NP, AP, EOP Established · mechanism established Limited Established mechanism
Counted on the row
trial count not established; no guideline endorsement recorded; best source tier 2.
Dose note
Concentrations are almost never disclosed in this literature — the trials test finished ceramide-containing products, not isolated ceramides at a stated %.
Vehicle
MOST ceramide evidence is on multi-ingredient FINISHED PRODUCTS, not ceramides in isolation. For a barrier/moisturization claim the indirectness is arguably LOW (the mechanism under test IS lipid-mixture replacement), but the row cannot attribute effect to ceramides specifically.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
  • Elias PM, Feingold KR (eds), Skin Barrier — stratum corneum ceramides are the principal intercellular lamellar lipid; their depletion is the biochemical signature of dry/xerotic skin — reference work
Cetearyl Alcohol & Stearic Acid Established · mechanism established Limited Established mechanism
Counted on the row
trial count not established; no guideline endorsement recorded; best source tier 2.
Studied dose
1–10%, leave-on
Dose note
FATTY alcohol — an emollient, NOT a drying alcohol. See inci-flag-rules -> drying_alcohol.exclude.
Vehicle
The confusion this row prevents is the commonest error in amateur INCI analysis.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
  • International Cosmetic Ingredient Dictionary and Handbook — declared function — Personal Care Products Council
  • Harry's Cosmeticology / Baran & Maibach Textbook of Cosmetic Dermatology — humectant, emollient and occlusive classification — standard reference text
Dimethicone Established · regulator recognises it, mechanism established Limited Established mechanism
Standing document
§ 347.50 Labeling of skin protectant drug products (indications for § 347.10 actives, incl. dimethicone) (U.S. Food and Drug Administration — 21 CFR Part 347 Subpart B, Code of Federal Regulations via GovInfo), opened Jul 27, 2026
Counted on the row
trial count not established; no guideline endorsement recorded; best source tier 2.
Studied dose
1–15%, leave-on
Dose note
Breathable occlusive film; FDA-recognised skin protectant.
Checked
against its sources on Jul 27, 2026; next review Jul 17, 2027
Sources
  • International Cosmetic Ingredient Dictionary and Handbook — declared function — Personal Care Products Council
  • United States Pharmacopeia / National Formulary — excipient monograph — USP-NF
  • Harry's Cosmeticology / Baran & Maibach Textbook of Cosmetic Dermatology — humectant, emollient and occlusive classification — standard reference text
Ethylhexylglycerin Established · mechanism established Limited Established mechanism
Counted on the row
trial count not established; no guideline endorsement recorded; best source tier 2.
Studied dose
2–20%, leave-on, judge after ~1 wk
Dose note
Humectancy contributes to dryness relief, though occlusion is the dominant mechanism for lipid-deficient skin.
Vehicle
Glycerin holds rows for BOTH dehydration and dryness — the mechanisms differ (binding vs sealing) and both are definitional.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
  • International Cosmetic Ingredient Dictionary and Handbook — declared function — Personal Care Products Council
  • Harry's Cosmeticology / Baran & Maibach Textbook of Cosmetic Dermatology — humectant, emollient and occlusive classification — standard reference text
Sodium Hyaluronate Established · mechanism established Limited Measured
Counted on the row
trial count not established; no guideline endorsement recorded; best source tier 3.
Dose note
Only LOW-molecular-weight HA beat vehicle on capacitance. Most catalog products do not declare MW, so the measured effect cannot be assigned to a given product.
Vehicle
MECHANISM IS NOT DEFINITIONAL FOR THIS CLAIM, which is why the row is `measured` and not `established-mechanism`. `dryness` in the vocabulary is LIPID-DEFICIENT skin, corrected by emollients and occlusives. HA is a water-binding humectant: that is definitional for `dehydration` (its sibling row), not for lipid deficiency. HA adds no lipid and is not a meaningful occlusive. What the row CAN stand on is instrumented human data on the claim's own endpoint: in elderly xerosis, LMW-HA lotion raised skin capacitance (corneometry) at 4 weeks vs vehicle (56.37 vs 49.01 AU, p<0.001) and vs HMW-HA (52.37 AU, p=0.004). The SECOND arm of the endpoint went the other way: the clinical symptom sum score (SRRC — the xerosis-scale arm) and TEWL did NOT differ between any groups. So the measurement is real but partial — more water in the stratum corneum, no demonstrated change in clinical xerosis signs or barrier water loss over 4 weeks. HMW-HA vs vehicle (52.37 vs 49.01) is not reported as significant in the abstract.
Checked
no verification date on record
Sources
Jojoba Esters Established · mechanism established Limited Established mechanism
Counted on the row
trial count not established; no guideline endorsement recorded; best source tier 2.
Studied dose
1–100%, leave-on
Dose note
Wax ester emollient.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
  • International Cosmetic Ingredient Dictionary and Handbook — declared function — Personal Care Products Council
Lanolin Established · regulator recognises it, mechanism established Limited Established mechanism
Standing document
§ 347.50 Labeling of skin protectant drug products (indications for § 347.10 actives, incl. lanolin) (U.S. Food and Drug Administration — 21 CFR Part 347 Subpart B, Code of Federal Regulations via GovInfo), opened Jul 27, 2026
Counted on the row
trial count not established; no guideline endorsement recorded; best source tier 2.
Studied dose
1–30%, leave-on
Dose note
Occlusive/emollient; a known sensitiser — priced in Safety, not here.
Checked
against its sources on Jul 27, 2026; next review Jul 17, 2027
Sources
  • International Cosmetic Ingredient Dictionary and Handbook — declared function — Personal Care Products Council
  • United States Pharmacopeia / National Formulary — excipient monograph — USP-NF
Paraffinum Liquidum + Cera Microcristallina Established · regulator recognises it, mechanism established Limited Established mechanism
Standing document
§ 347.50 Labeling of skin protectant drug products (indications for § 347.10 actives, incl. mineral oil) (U.S. Food and Drug Administration — 21 CFR Part 347 Subpart B, Code of Federal Regulations via GovInfo), opened Jul 27, 2026
Counted on the row
trial count not established; no guideline endorsement recorded; best source tier 2.
Studied dose
1–100%, leave-on
Dose note
Occlusive; second only to petrolatum.
Checked
against its sources on Jul 27, 2026; next review Jul 17, 2027
Sources
  • International Cosmetic Ingredient Dictionary and Handbook — declared function — Personal Care Products Council
  • United States Pharmacopeia / National Formulary — excipient monograph — USP-NF
Petrolatum Established · regulator recognises it, mechanism established Limited Established mechanism
Standing document
§ 347.50 Labeling of skin protectant drug products (indications for § 347.10 actives, incl. petrolatum) (U.S. Food and Drug Administration — 21 CFR Part 347 Subpart B, Code of Federal Regulations via GovInfo), opened Jul 27, 2026
Counted on the row
trial count not established; no guideline endorsement recorded; best source tier 2.
Studied dose
1–100%, leave-on, judge after ~1 wk
Dose note
The reference occlusive — reduces transepidermal water loss by up to ~99%, the benchmark every other occlusive is measured against.
Vehicle
Occlusion is physics. Petrolatum forms a hydrophobic film; water does not pass through it. This is the single least contested claim in all of skincare.
Checked
against its sources on Jul 27, 2026; next review Jul 17, 2027
Sources
  • International Cosmetic Ingredient Dictionary and Handbook — declared function — Personal Care Products Council
  • United States Pharmacopeia / National Formulary — excipient monograph — USP-NF
  • Harry's Cosmeticology / Baran & Maibach Textbook of Cosmetic Dermatology — humectant, emollient and occlusive classification — standard reference text
Shea Butter Established · mechanism established Limited Established mechanism
Counted on the row
trial count not established; no guideline endorsement recorded; best source tier 2.
Studied dose
1–100%, leave-on
Dose note
Emollient triglycerides + unsaponifiables.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
  • International Cosmetic Ingredient Dictionary and Handbook — declared function — Personal Care Products Council
  • Harry's Cosmeticology / Baran & Maibach Textbook of Cosmetic Dermatology — humectant, emollient and occlusive classification — standard reference text
Squalane Established · mechanism established Limited Established mechanism
Counted on the row
trial count not established; no guideline endorsement recorded; best source tier 2.
Studied dose
1–100%, leave-on
Dose note
Emollient; a saturated analogue of skin's own squalene.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
  • International Cosmetic Ingredient Dictionary and Handbook — declared function — Personal Care Products Council
  • Harry's Cosmeticology / Baran & Maibach Textbook of Cosmetic Dermatology — humectant, emollient and occlusive classification — standard reference text

Dullness 3 rows

Ethylhexylglycerin Thin · trials Limited Plausible mechanism
Counted on the row
trial count not established; no guideline endorsement recorded; best source tier 2.
Studied dose
2–20%, leave-on, judge after ~1 wk
Dose note
Hydrated corneocytes scatter light more evenly. Plausible, not definitional.
Vehicle
The honest floor for the catalog's most-claimed condition: a real mechanism, argued rather than established. This is what 5,357 product-claims actually rest on.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
  • Harry's Cosmeticology / Baran & Maibach Textbook of Cosmetic Dermatology — humectant, emollient and occlusive classification — standard reference text
Glycolic Acid Established · mechanism established Limited Measured
Counted on the row
trial count not established; no guideline endorsement recorded; best source tier 2.
Studied dose
4–10%, leave-on, judge after ~8 wk
Dose note
Removing dulling corneocytes changes surface light scattering — a real optical mechanism, but 'measured' rather than definitional because the OPTICAL outcome (not the exfoliation) is the claim.
Vehicle
Deliberately NOT established-mechanism. Exfoliation is definitional; the resulting luminance change is an inference from it. The subjective 'radiance' claim remains unscoreable.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
  • Harry's Cosmeticology / Baran & Maibach Textbook of Cosmetic Dermatology — humectant, emollient and occlusive classification — standard reference text
Sodium Hyaluronate Thin · trials Limited Plausible mechanism
Counted on the row
trial count not established; no guideline endorsement recorded; best source tier 2.
Dose note
Not applicable — no admissible trial exists for this endpoint. Reporting a dose range would imply an evidence base that is not there.
Vehicle
HA is a humectant; any short-term optical change is plausibly downstream of hydration and surface plumping — exactly what the vehicle of any HA serum also does. No trial separates the polymer's contribution from the base's.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027

Eczema 12 rows

Centella Asiatica Extract Thin · trials Limited Limited
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 2.
Dose note
The eczema endpoint is change in a validated severity index (EASI/SCORAD/POEM) in diagnosed atopic dermatitis — distinct from generic 'soothing' or barrier/dryness claims.
Vehicle
Centella's anti-inflammatory eczema signal comes from cell-culture (HaCaT) and mouse models; no human leave-on formulation has been tested against a validated eczema index.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Ceramides NP, AP, EOP Thin · trials Moderate Limited
Counted on the row
3 randomised trials, 213 participants pooled; independently replicated; findings consistent; no guideline endorsement recorded; best source tier 2.
Dose note
Ceramide content is essentially never disclosed on a cosmetic INCI, so a dose-response cannot be constructed even in principle from catalog data. The trials test named finished products, not a concentration.
Vehicle
For barrier lipids the RATIO is the claimed mechanism, not the presence of ceramides — the 3:1:1 ceramide:cholesterol:fatty-acid stoichiometry is the whole rationale for EpiCeram. An OTC cream containing 'ceramide NP' somewhere on the INCI is not a low-dose EpiCeram; it is a different object making a similar-sounding claim.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Colloidal Oatmeal Established · regulator recognises it, trials Moderate Limited
Standing document
§ 347.50 Labeling of skin protectant drug products (US FDA — 21 CFR 347.50 (via Cornell Legal Information Institute)), opened Jul 27, 2026
Counted on the row
2 randomised trials, 139 participants pooled; a systematic review or meta-analysis; findings mixed; no guideline endorsement recorded; best source tier 3.
Studied dose
1%, judge after ~2 wk
Dose note
1% is the OTC monograph skin-protectant level; trials tested the finished cream, not oatmeal in isolation.
Vehicle
Decisive limitation: both pivotal RCTs are ACTIVE-controlled (oatmeal cream vs a ceramide/prescription BARRIER cream), with no vehicle/placebo arm. They can show non-inferiority-to-another-moisturizer at best; they cannot isolate colloidal oatmeal's effect versus no treatment.
Checked
against its sources on Jul 27, 2026; next review Jul 17, 2027
Sources
Dimethicone Thin · trials Limited Limited
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 1.
Studied dose
1–5%, leave-on
Dose note
No trial defines a dimethicone dose isolated against an eczema severity endpoint.
Vehicle
Dimethicone is a silicone occlusive whose real, measurable action is reducing transepidermal water loss by forming a surface film — a barrier/occlusion mechanism. In AD trials it is never the sole functional ingredient; the finished moisturizer's benefit cannot be attributed to the silicone specifically.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Ethylhexylglycerin Established · trials Moderate Well-established
Counted on the row
3 randomised trials, 385 participants pooled; a systematic review or meta-analysis; findings consistent; no guideline endorsement recorded; best source tier 1; GRADE certainty high.
Studied dose
15–20%, leave-on, judge after ~1 wk
Dose note
COULD NOT ESTABLISH a dose-response relationship. Cochrane pooled by 'glycerol-containing', not by concentration, so the catalog's wide glycerin range (a few percent to ~20%+) is NOT resolved by this evidence. A product containing glycerin at 2% is not what was trialled.
Vehicle
This is the row where 'vehicle' is nearly the whole story: the comparator was the same moisturiser base WITHOUT glycerol, so the effect measured is glycerol's increment over an already-emollient vehicle. That is a fair test of the ingredient, and it is also why the increment is small — the base itself works.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Glycyrrhetinic Acid Thin · trials Moderate Limited
Counted on the row
4 randomised trials, 187 participants pooled; findings consistent; no guideline endorsement recorded; best source tier 3.
Dose note
Every controlled trial delivered the licorice actinide inside a multi-ingredient emollient, so the base cannot be separated from the licorice.
Vehicle
All four controlled trials tested licochalcone A WITHIN a moisturizer (glycerol, fatty acids, ceramides), never as a standalone. The comparator was a low-potency corticosteroid (1% hydrocortisone / 0.02% triamcinolone), not vehicle, so the design shows non-inferiority to a weak steroid but cannot attribute the SCORAD drop to licorice rather than the emollient base.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Niacinamide Thin · trials Moderate Limited
Counted on the row
1 randomised trial, 122 participants pooled; findings insufficient; no guideline endorsement recorded; best source tier 3.
Dose note
The intervention is a finished niacinamide-CONTAINING emollient regimen, not isolated niacinamide.
Vehicle
The active comparator problem is decisive: both treatment arms received a niacinamide-containing EMOLLIENT, and emollients are themselves first-line AD therapy. Without a niacinamide-free emollient arm, niacinamide's independent contribution to the SCORAD improvement cannot be separated from the vehicle.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Panthenol Thin · trials Moderate Limited
Counted on the row
2 randomised trials, 146 participants pooled; findings insufficient; no guideline endorsement recorded; best source tier 3.
Studied dose
5%, leave-on
Dose note
5% dexpanthenol is the tested level, but always co-formulated with other actives — never isolated.
Vehicle
Both trials test a COMBINATION emollient (panthenol + ceramide / linoleic acid / other), compared against an active reference emollient or 5% urea — never panthenol alone, never vs vehicle. Panthenol's independent contribution is not isolable from the vehicle/co-actives.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Petrolatum Established · regulator recognises it, mechanism established, trials Moderate Limited
Standing document
21 CFR § 347.50 - Labeling of skin protectant drug products. (US Code of Federal Regulations (Legal Information Institute, Cornell Law School)), opened Jul 27, 2026
Counted on the row
0 randomised trials; a systematic review or meta-analysis; independently replicated; findings consistent; no guideline endorsement recorded; best source tier 1.
Studied dose
up to 100%, leave-on, judge after ~1 wk
Dose note
Petrolatum is the definitional emollient and the standard emollient control arm across atopic-dermatitis trials. Guideline endorsement is for moisturizers as a CLASS; petrolatum is the archetypal member, not a stretch from it.
Vehicle
Petrolatum's benefit in eczema is occlusion/barrier repair and flare reduction/steroid-sparing, not an anti-inflammatory efficacy claim. The endpoint match is via flare-frequency and steroid-sparing outcomes (explicitly in the eczema claim definition), which the Cochrane review reports.
Checked
against its sources on Jul 27, 2026; next review Jul 17, 2027
Sources
Shea Butter Thin · trials Limited Limited
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 1.
Dose note
The claim is the clinical disease (EASI/SCORAD/POEM), distinct from the physicochemical 'dryness' claim, where shea's emolliency is on firmer ground.
Vehicle
Every shea-containing product in the Cochrane review is a complex formulation (e.g. La Roche-Posay Iso-Urea, Pierre Fabre Exomega, Jirabundansuk's spent-grain+shea+argan cream) — shea's contribution cannot be isolated from urea, oat, glycerol, petrolatum and other actives.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Squalane Thin No evidence No evidence
Counted on the row
trial count not established; findings insufficient; no guideline endorsement recorded; best source tier 2.
Dose note
No clinical AD study of a squalene/squalane-containing emollient has been conducted, so no functional dose for the eczema claim can be stated.
Vehicle
Squalane is a genuine emollient — but the emollient/occlusive benefit belongs to the DRYNESS and barrier (physicochemical) claims, graded in their own rows. That mechanism must NOT be transferred to the clinical eczema claim, which asks whether squalane changes a validated AD severity index.
Checked
no verification date on record
Sources
Hydroxyethyl Urea Thin · trials Moderate Limited
Counted on the row
1 randomised trial; a systematic review or meta-analysis; independently replicated; findings consistent; no guideline endorsement recorded; best source tier 1.
Studied dose
5–10%, leave-on, judge after ~2 wk
Dose note
In eczema, urea functions as a humectant emollient reducing flare frequency and prolonging eczema-free time, not as a lesion-clearing drug.
Vehicle
Urea is a natural moisturizing factor / humectant with mild keratolytic action at higher strengths; the AD evidence is about barrier-supporting emollient use (relapse/flare-frequency, corticosteroid-sparing) rather than acute lesion clearance.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources

Excess Oiliness 7 rows

Hectorite Thin · trials Limited Limited
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 3.
Dose note
The excess-oiliness endpoint is a SUSTAINED sebumeter/Sebutape reduction — physiologic sebum SUPPRESSION. Clay's actual action is transient surface-oil ABSORPTION (mattifying), which the vocabulary explicitly classifies as a sensory property, NOT this claim.
Vehicle
Rinse-off masks contact skin for minutes; any sebum drop is dominated by immediate absorption plus the wash step, not by a lasting change in sebaceous output.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Camellia Oleifera Thin · trials Moderate Limited
Counted on the row
1 randomised trial, 22 participants pooled; findings insufficient; no guideline endorsement recorded; best source tier 2.
Studied dose
3–5%, leave-on
Dose note
5% green tea extract in a multiple emulsion is not what is in a green-tea toner. The 27% sebum reduction figure belongs to a 5% formulation and should never be quoted for a product with trace-level extract.
Vehicle
Both human sebum trials used purpose-built multiple/W-O emulsions engineered around the extract, not conventional cosmetic bases. Catechins are notoriously unstable to oxidation, so an aged toner is not delivering what a freshly-prepared study emulsion delivered — a stability question no trial addresses.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Niacinamide Thin · trials Moderate Limited
Counted on the row
2 randomised trials, 130 participants pooled; findings mixed; no guideline endorsement recorded; best source tier 1.
Studied dose
2%, leave-on, judge after ~8 wk
Dose note
The ENTIRE instrumented sebum evidence base for topical niacinamide is 2%. No sebumeter/Sebutape trial exists at 4%, 5% or 10%. This INVERTS the pigment claim (where 2% failed and only 5% worked) — so a 5% product cannot borrow the 2% sebum data, and a 2% product cannot borrow the 5% pigment data. BOTH directions of extrapolation are unsupported.
Vehicle
A moisturizer base is itself a plausible confounder for casual sebum readings — surface lipids and applied emollient are hard to disambiguate on a sebumeter tape. This matters specifically because CSL, not SER, was the endpoint that reached significance in the Caucasian arm.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Retinoids Thin No evidence No evidence
Counted on the row
trial count not established; findings insufficient; no guideline endorsement recorded.
Dose note
The claim endpoint is instrumented, SUSTAINED sebum reduction (sebumeter/Sebutape) — not mattifying, and not follicular keratinization.
Vehicle
Topical retinoids normalize follicular keratinization; the dramatic sebum-gland suppression associated with 'retinoids' is an ORAL isotretinoin effect (a different intervention and route). Conflating the two is indirectness_intervention.
Checked
no verification date on record
Sources
none recorded
Retinol Thin · trials Limited Limited
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 3.
Studied dose
0.1–1%, leave-on, judge after ~12 wk
Dose note
In-vitro data show retinoids suppress sebocytes, but sebumeter-measured output reduction from topical retinol is not established.
Vehicle
The sebum-reduction evidence for retinoids is surrogate (in-vitro sebocyte suppression) or belongs to oral isotretinoin. For topical retinol, the only relevant study (Nowak 2024, n=20 split-face) applied retinol creams to both sides with no untreated/vehicle control, so a true sebostatic effect cannot be isolated.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Salicylic Acid Thin · trials Moderate Limited
Counted on the row
1 randomised trial, 17 participants pooled; findings insufficient; no guideline endorsement recorded; best source tier 3.
Dose note
A large, unbridged gap between studied dose and consumer dose. Every vehicle-controlled sebum datapoint comes from a 30% IN-OFFICE PEEL. US OTC monograph acne products are limited to 0.5–2%. NO vehicle-controlled instrumented sebum data exists at ANY leave-on consumer concentration. A dose-response for sebum has never been established.
Vehicle
THE DEGREASING CONFOUND IS DECISIVE HERE. A Sebumeter reads skin-surface lipid; unless the site is degreased to a standardized protocol it reports 'casual level' — total surface lipid — not sebum excretion rate. A lipophilic, sebolytic LEAVE-ON applied twice daily can lower a casual-level reading by dissolving surface sebum with ZERO effect on gland output. This eliminates the leave-on evidence entirely and leaves only the rinse-off peel, which is structurally immune to the artifact because product is not resident between visits. Separately: 'supramolecular' SA is a delivery system (complexed for slow release); its kinetics are not interchangeable with free SA in an aqueous or hydroalcoholic base, so the 30% peel result does not transfer to a 2% leave-on even after discounting for concentration. RE-REVIEW 2026-10-13: re-opened Wang 2025 full text (PMC11836920) — the sebum table confirms control n=8 vs experimental n=9 (pooled_n 17 stands). Note an internal inconsistency in that paper: section 3.4.2 opens 'Sebum levels on both cheeks and forehead in the experimental group were reduced at final visit (p > 0.05)' while the abstract and discussion call the reduction significant; between-group differences are reported significant only at selected visits and sites, and the nose showed none. This supports the existing imprecision and risk_of_bias flags.
Checked
no verification date on record
Sources
Zinc Gluconate Thin · trials Moderate Limited
Counted on the row
1 randomised trial, 14 participants pooled; findings insufficient; no guideline endorsement recorded; best source tier 3.
Studied dose
1%, leave-on
Dose note
The oral trials used zinc sulfate 440-660 mg/day SYSTEMICALLY. Oral supplementation is not a dose of a topical cosmetic — it is a different intervention entirely, and its results are not counted toward this row.
Vehicle
Decisive and unresolved. In the one topical trial the zinc is bound in an erythromycin-zinc COMPLEX, and the complex is known to alter erythromycin delivery — so the between-arm difference may reflect zinc improving the antibiotic's penetration rather than zinc acting on sebocytes. A cosmetic zinc PCA serum shares neither the salt, the complex, nor the vehicle.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources

Fungal Acne 6 rows

Honey Extract Thin No evidence No evidence
Counted on the row
trial count not established; findings insufficient; no guideline endorsement recorded.
Dose note
Endpoint = lesion clearance in CONFIRMED Malassezia folliculitis. No human trial measures this for honey.
Vehicle
Most honey is a Malassezia substrate concern rather than a treatment; any antimicrobial action (hydrogen peroxide, methylglyoxal in manuka) is documented for bacteria/wounds, not for follicular Malassezia clearance.
Checked
no verification date on record
Sources
  • No admissible human trial located for honey in Malassezia folliculitis (the sole database hit is a veterinary survey of canine dermatology) — reviewer note
Ketoconazole 1% Thin · trials Limited Limited
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 2.
Studied dose
1–2%, leave-on, judge after ~4 wk
Dose note
Endpoint = lesion clearance in CONFIRMED Malassezia folliculitis (mycology/diagnostic response). Ketoconazole is a genuine anti-Malassezia azole — the mechanistically correct drug, unlike keratolytic salicylic acid.
Vehicle
Topical penetration to the follicular depth where Malassezia resides is a real limitation of topical azoles, which is why oral azoles (itraconazole/fluconazole) are often used for extensive disease.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Salicylic Acid Thin No evidence No evidence
Counted on the row
trial count not established; findings insufficient; no guideline endorsement recorded.
Studied dose
0.5–2%, leave-on, judge after ~6 wk
Dose note
Endpoint = lesion clearance in CONFIRMED Malassezia folliculitis (mycology or diagnostic response). Standard first-line therapy is topical/oral azole antifungals (ketoconazole, itraconazole), not a keratolytic.
Vehicle
Salicylic acid's mechanism is keratolytic (comedolytic), not antifungal — it does not kill Malassezia, which is the causative organism of the condition this claim names.
Checked
no verification date on record
Sources
  • No admissible trial located — standard of care for Malassezia folliculitis is azole antifungal therapy (context note, not a treatment citation) — reviewer note
Selenium Sulfide 1% Thin · trials Limited Limited
Counted on the row
trial count not established; findings insufficient; guideline endorsement conditional; best source tier 1.
Studied dose
1–2.5%, leave-on, judge after ~4 wk
Dose note
Applied to face/trunk as a wash-off, not a leave-on. The higher-quality selenium sulfide RCTs are in pityriasis versicolor and dandruff — DIFFERENT conditions that do not transfer to Malassezia folliculitis.
Vehicle
Selenium sulfide is an antifungal/keratolytic wash-off; its role in Malassezia folliculitis is anti-Malassezia load reduction plus maintenance to prevent recurrence, but the folliculitis-specific evidence is observational.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Sulfur 10% Thin · trials Limited Limited
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 3.
Studied dose
2–10%, leave-on
Dose note
Fungal acne is Malassezia folliculitis, confirmed by mycology or diagnostic response. Ordinary (bacterial) acne evidence never transfers here — a key vocabulary rule.
Vehicle
Sulfur has genuine antifungal + keratolytic activity, and is a recognized agent for other Malassezia-associated conditions (pityriasis versicolor, seborrheic dermatitis), but that is a different clinical presentation from Malassezia folliculitis.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Zinc Pyrithione Thin · trials Limited Limited
Counted on the row
trial count not established; findings insufficient; guideline endorsement conditional; best source tier 1.
Studied dose
1–2%, leave-on, judge after ~4 wk
Dose note
Contact time is short (wash-off). In-vitro anti-Malassezia MICs are favorable, but that is a laboratory endpoint, not clinical lesion clearance.
Vehicle
Zinc pyrithione is an antifungal that lowers Malassezia load; its clinical trial base is in dandruff / seborrheic dermatitis of the SCALP, a different condition and site from facial/truncal Malassezia folliculitis.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources

Hyperpigmentation 13 rows

Alpha-Arbutin Thin · trials Limited Limited
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 3.
Studied dose
1–5%, leave-on, judge after ~12 wk
Dose note
Alpha-arbutin's mechanism (competitive tyrosinase inhibition) is established in vitro, but no monotherapy clinical trial with a validated non-melasma pigment endpoint (mexameter/colorimetry on PIH) could be confirmed.
Vehicle
Every human alpha-arbutin efficacy trial found is either on MELASMA (a distinct claim that does not transfer to general/post-inflammatory hyperpigmentation per the vocabulary) or in a multi-ingredient formulation from which the arbutin effect cannot be isolated.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Azelaic Acid Supported · named in a guideline, trials Moderate Emerging
Standing document
Treatment recommendations for acne-associated hyperpigmentation: Results of the Delphi consensus process and a literature review (Journal of the American Academy of Dermatology 2023;89(2):316-323 (PMID 36924935)), opened Jul 27, 2026
Counted on the row
1 randomised trial; findings insufficient; no guideline endorsement recorded; best source tier 3.
Studied dose
10%, leave-on, judge after ~12 wk
Dose note
Endpoint = measured reduction of NON-melasma pigment (mexameter/colorimetry). Melasma evidence is a DISTINCT claim and must not transfer here.
Vehicle
The confirmed vehicle-controlled trial used 20% cream; OTC 'brightening' products are typically ~10% in different vehicles, so a dose/vehicle indirectness applies to the catalog presentation.
Checked
against its sources on Jul 27, 2026; next review Jul 17, 2027
Sources
Glycolic Acid Thin · trials Moderate Limited
Counted on the row
2 randomised trials, 46 participants pooled; findings consistent; no guideline endorsement recorded; best source tier 2.
Studied dose
5–10%, leave-on, judge after ~8 wk
Dose note
THE DECISIVE GAP. Both counted trials tested 35-68% in-office PEELS. The catalog sells 5-10% leave-on. These are not two points on one dose ladder: they differ in concentration by roughly an order of magnitude, in contact time, in pH, and in who applies them. Notably, the 1997 trial's CONTROL arm already contained 10% glycolic acid gel — i.e. the OTC-strength exposure was present in BOTH arms and was not what was being tested.
Vehicle
Free-acid concentration and pH govern glycolic acid's activity, and neither trial was designed to isolate them. Both are add-on designs layered on a hydroquinone/tretinoin base, so the vehicle question for a standalone cosmetic leave-on is untouched.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Hydroquinone Supported · named in a guideline, trials Limited Limited
Standing document
Treatment recommendations for acne-associated hyperpigmentation: Results of the Delphi consensus process and a literature review (Journal of the American Academy of Dermatology 2023;89(2):316-323; PMID 36924935), opened Jul 27, 2026
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 3.
Studied dose
2%, leave-on, judge after ~12 wk
Dose note
The same tyrosinase-inhibition mechanism as for melasma applies, but the PIH-specific controlled evidence is far thinner than the mechanism or clinical habit would suggest.
Vehicle
Tyrosinase inhibition is not pigment-type-specific mechanistically, but the vocabulary (and GRADE) require the endpoint to be measured in the claimed condition. PIH pigment evidence and melasma pigment evidence are separate rows and do not transfer.
Checked
against its sources on Jul 27, 2026; next review Jul 17, 2027
Sources
Kojic Acid Thin · trials Limited Limited
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 2.
Studied dose
1–2%, leave-on, judge after ~12 wk
Dose note
This row is NON-melasma pigment (general/post-inflammatory), measured by mexameter/colorimetry or a validated pigment scale. Kojic acid's MELASMA evidence is a separate claim and must not transfer here.
Vehicle
Kojic acid is chemically unstable (oxidizes, browns); many formulas underdeliver the labeled dose, and clinical dyschromia studies bundle it with tranexamic acid/niacinamide so its independent effect is unmeasured.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Lactic Acid Thin · trials Moderate Limited
Counted on the row
1 randomised trial, 74 participants pooled; findings insufficient; no guideline endorsement recorded; best source tier 1.
Studied dose
8%, leave-on, judge after ~4 wk
Dose note
No dose-response for pigment is established. The leave-on cosmetic dose (5–12%) and the in-office peel doses (15–92%) are pharmacologically different interventions and must not be pooled. Note the ~10× gap between the only controlled dose (8%) and the peel literature. pH is not reported and could not be established — for an AHA, pH governs free-acid availability and therefore potency, so '8%' alone is not reproducible.
Vehicle
Stiller 1996 compared 8% L-lactic acid against its OWN vehicle cream — a genuine vehicle control, rare in this literature. IMPORTANT MECHANISTIC CONFOUND: lactic acid's pigment effect cannot be separated from its keratolytic/desquamative action, which already holds a separate established-mechanism row for rough-texture. Accelerated turnover sheds pigmented corneocytes and lightens appearance WITHOUT any tyrosinase-level activity — so a positive pigment reading here does not establish a pigmentary mechanism.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Glycyrrhetinic Acid Thin · trials Limited Limited
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 3.
Studied dose
4–20%, leave-on
Dose note
Do not let the melasma concentrations migrate onto this row. That migration is the exact failure mode this table exists to stop.
Vehicle
Not assessable — no qualifying trial exists to have a vehicle. Note that licorice's catalog presence is largely as dipotassium glycyrrhizate and glycyrrhetinic acid (soothing agents), which are not the same molecules as the pigment-active flavonoids (glabridin, liquiritin) studied in melasma.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Niacinamide Emerging · trials Moderate Emerging
Counted on the row
5 randomised trials, 336 participants pooled; independently replicated; findings consistent; no guideline endorsement recorded; best source tier 1.
Dose note
Dose threshold unchanged: 2% alone did not separate from vehicle in the prior record; 4-5% did. New and sponsor-run (P&G, Hakozaki 2026 IJCS): 2% at pH 2.5 outperformed 2% at neutral pH and vehicle on facial spots over 4 weeks, so formula pH may matter as much as percent. That is one sponsor's finding and is not counted. Effect is reversible after stopping.
Vehicle
All pivotal work uses niacinamide in a moisturizer base, vehicle-controlled split-face. CONFOUNDS: Hakozaki's 120-subject arm was 2% niacinamide + SUNSCREEN (sunscreen alone reduces pigment). Lee 2014 is niacinamide + TRANEXAMIC ACID (an active depigmenting agent in its own right) — niacinamide's contribution is unattributable. ADDED Oct 2026: Castanedo-Cazares used a plain moisturizer (Cetaphil) with and without 4% niacinamide, HPLC-confirmed >90% stable at 3 months. That is the cleanest vehicle contrast in the base. Kimball 2010 ran both arms on SPF 15 by day, so the effect is beyond sunscreen, not instead of it.
Checked
no verification date on record
Sources
Retinoids Thin · trials Moderate Limited
Counted on the row
1 randomised trial, 54 participants pooled; independently replicated; findings insufficient; no guideline endorsement recorded; best source tier 1.
Studied dose
0.1%, leave-on, judge after ~12 wk
Dose note
FAMILY ROW. The PIH evidence is Rx tretinoin's; it does not transfer to OTC retinol.
Vehicle
The non-melasma PIH endpoint evidence is Bulengo-Ransby's Rx tretinoin trial. Melasma is a separate claim and is not counted here.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Retinol Thin · trials Moderate Limited
Counted on the row
3 randomised trials, 44 participants pooled; findings consistent; no guideline endorsement recorded; best source tier 1.
Studied dose
0.1–1%, leave-on, judge after ~12 wk
Dose note
No dose-response relationship for retinol on non-melasma pigment could be established: there is only one tier-1 trial and it tested a single concentration.
Vehicle
Retinol is only converted to retinoic acid intracellularly, and the conversion efficiency is formulation- and stability-dependent. No located trial compared vehicles on a pigment endpoint, so vehicle effects on this claim are entirely unmeasured.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Tranexamic Acid Thin · trials Moderate Limited
Counted on the row
1 randomised trial, 60 participants pooled; findings insufficient; no guideline endorsement recorded; best source tier 3.
Studied dose
2–5%, leave-on, judge after ~12 wk
Dose note
The evidence base and the product category barely overlap. Most TXA/PIH research is systemic or injected; the catalog is entirely topical leave-on.
Vehicle
TXA is hydrophilic and penetrates poorly — the reason so much of the literature reaches for oral, intradermal, microneedling or laser-assisted delivery. Lin 2021 (topical TXA + picosecond laser) attributed its null result specifically to poor penetration. A 2-3% TXA serum is therefore not a weaker version of the injected evidence; it is a different question.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Tretinoin 0.025% Supported · regulator recognises it, trials Moderate Emerging
Standing document
REFISSA- tretinoin cream (Tretinoin Cream, USP (Emollient) 0.05%) — INDICATIONS AND USAGE (US FDA prescribing information via DailyMed (NLM)), opened Jul 27, 2026
Counted on the row
1 randomised trial, 54 participants pooled; independently replicated; findings insufficient; no guideline endorsement recorded; best source tier 1.
Studied dose
0.1%, leave-on, judge after ~12 wk
Dose note
Tretinoin is prescription-only. Its evidence must NOT be transferred to OTC retinol/retinaldehyde products, which are distinct interventions.
Vehicle
The endpoint-matching evidence is for post-inflammatory hyperpigmentation (within the general/PIH hyperpigmentation claim). Tretinoin's melasma trials are a DISTINCT claim and are not counted here.
Checked
against its sources on Jul 27, 2026; next review Jul 17, 2027
Ascorbic Acid Thin · trials Moderate Limited
Counted on the row
1 randomised trial, 25 participants pooled; findings insufficient; no guideline endorsement recorded; best source tier 3.
Dose note
No vehicle-controlled trial of vitamin C ALONE on established non-melasma lesional pigment was found. The L'Oreal Bayesian meta-analysis (31 RCTs, n=741) found vitamin C PREVENTS UV-induced pigmentation dose-dependently but added nothing in the DEPIGMENTATION phase. That is the authors' own conclusion: antipigmenting, not depigmenting.
Vehicle
VERIFIED (Pinnell 2001, pig skin): L-AA requires pH<3.5 to penetrate; maximal absorption at 20%; tissue saturates after 3 daily applications; half-life ~4 days. The same paper found magnesium ascorbyl phosphate, ascorbyl-6-palmitate and dehydroascorbic acid did NOT improve penetration over L-AA — yet several pigment trials use exactly those derivatives, so form is a live confound across this base. ADDED Oct 2026: in the only on-claim RCT (Hu 2024), the serum was a co-intervention after laser and the control was laser alone, not a vehicle, so vitamin C's own contribution is unattributable.
Checked
no verification date on record
Sources

Keratosis Pilaris 5 rows

Ammonium Lactate Thin · trials Limited Limited
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 3.
Studied dose
12–14%, leave-on, judge after ~4 wk
Dose note
Ammonium lactate is essentially pH-neutralized lactic acid, so the nearest efficacy signal comes from a lactic-acid KP trial, not from ammonium lactate itself.
Vehicle
Ammonium lactate is the neutralized ammonium salt of lactic acid; its keratolytic mechanism is established, but the salt/pH form is not identical to the free lactic acid tested in the nearest KP RCT.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Glycolic Acid Thin · trials Limited Limited
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 3.
Dose note
The isolating glycolic KP data is an in-office high-concentration peel, not the leave-on products the catalog sells.
Vehicle
No vehicle-controlled leave-on glycolic KP trial; the isolating study is an uncontrolled high-concentration peel cohort, and the other glycolic KP study is a combination serum with a questionnaire endpoint.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Lactic Acid Thin · trials Moderate Limited
Counted on the row
1 randomised trial; findings consistent; no guideline endorsement recorded; best source tier 3.
Studied dose
10%, judge after ~4 wk
Dose note
10% LA reached 66% lesion reduction over 12 weeks; higher ammonium-lactate strengths were not the tested presentation here.
Vehicle
Active-controlled (10% lactic acid vs 5% salicylic acid), no vehicle arm — both are keratolytics, both worked, LA numerically ahead.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Salicylic Acid Thin · trials Moderate Limited
Counted on the row
1 randomised trial; findings consistent; no guideline endorsement recorded; best source tier 3.
Studied dose
5%, judge after ~6 wk
Dose note
Tested head-to-head against 10% lactic acid in the same trial.
Vehicle
Active-controlled (5% salicylic acid vs 10% lactic acid), no vehicle arm; both keratolytics improved KP, salicylic acid slightly behind lactic.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Hydroxyethyl Urea Thin · trials Limited Limited
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 3.
Studied dose
10–20%, judge after ~2 wk
Dose note
Urea's action is concentration-dependent humectant + keratolytic; no controlled KP dose-finding exists.
Vehicle
The one dedicated urea KP study is OPEN-LABEL and NONCOMPARATIVE (no vehicle/control arm), so improvement cannot be separated from time, regression, or generic moisturization.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources

Large Pores 7 rows

Hectorite Thin No evidence No evidence
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded.
Dose note
No dosed-active concept; no trial defines a range for a pore endpoint.
Vehicle
Rinse-off contact time is minutes; any effect on pores is transient absorbent/optical, not a measured structural change.
Checked
no verification date on record
Sources
none recorded
Glycolic Acid Thin No evidence No evidence
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 3.
Studied dose
5–10%, leave-on, judge after ~8 wk
Dose note
There is no dose-response to describe because there is no response to describe.
Vehicle
Not assessable — no glycolic-isolating trial with a pore-imaging endpoint exists to have a vehicle.
Checked
no verification date on record
Sources
Niacinamide Thin · trials Moderate Limited
Counted on the row
1 randomised trial, 52 participants pooled; findings insufficient; no guideline endorsement recorded; best source tier 2.
Studied dose
4%, leave-on, judge after ~8 wk
Dose note
Single concentration, single trial. NO dose-response information exists for this claim at all. Nothing supports 2%, 5% or 10% products on pore endpoints.
Vehicle
Split-face with placebo on the contralateral side; the niacinamide-alone arm DOES isolate the ingredient — a genuine strength, rarer than expected in this literature.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Retinoids Thin No evidence No evidence
Counted on the row
trial count not established; findings insufficient; no guideline endorsement recorded.
Dose note
Claim endpoint is measured pore area/count by standardized imaging. Pore size is structurally fixed; only appearance is claimable, and the vocabulary flags this as a contested, easily-overstated claim.
Vehicle
Any pore-appearance benefit from retinoids is most plausibly a secondary consequence of photoaging/texture improvement rather than a direct pore-area effect; the direct standardized-imaging endpoint is what this claim requires.
Checked
no verification date on record
Sources
none recorded
Retinol Emerging · trials Moderate Emerging
Counted on the row
2 randomised trials, 120 participants pooled; findings consistent; no guideline endorsement recorded; best source tier 3.
Studied dose
0.1–1%, leave-on, judge after ~12 wk
Dose note
Pore size is structurally fixed; the honest endpoint is measured pore AREA/appearance by standardized imaging, not shrinkage.
Vehicle
Retinol's pore-appearance signal comes from small industry-linked split-face/comparative trials measured by imaging (ANTERA/VISIA); there is no independent vehicle-controlled OTC-retinol pore RCT of size.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Salicylic Acid Thin No evidence No evidence
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 3.
Studied dose
0.5–2%, leave-on, judge after ~6 wk
Dose note
20–30% is the in-office PEEL range and is a procedure, not a leave-on product. The two SA papers that imaged pores at all used a 2% hydrogel and a 30% peel respectively — neither maps to an OTC leave-on pore claim.
Vehicle
SA requires a functional pH (≤ ~4) to act as an exfoliant; a neutralized formulation is inert for any keratolytic mechanism. No pore trial reported vehicle pH, so even the mechanistic hand-wave is unverifiable in the published record.
Checked
no verification date on record
Sources
Tretinoin 0.025% Thin · trials Moderate Limited
Counted on the row
1 randomised trial, 120 participants pooled; findings insufficient; no guideline endorsement recorded; best source tier 3.
Studied dose
0.025%, leave-on, judge after ~12 wk
Dose note
Pore appearance was a secondary endpoint within a photoaging trial, not a primary standardized pore-imaging outcome.
Vehicle
Any tretinoin pore effect is plausibly downstream of dermal remodeling / reduced follicular plugging over months; pore SIZE is structurally fixed, so the honest claim is appearance, which the vocabulary flags as usually overstated.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027

Melasma 11 rows

Alpha-Arbutin Thin · trials Moderate Limited
Counted on the row
1 randomised trial, 10 participants pooled; findings insufficient; no guideline endorsement recorded; best source tier 3.
Dose note
Marketed 'alpha-arbutin' is usually ~2%; the 2008 trial did not state its arbutin concentration.
Vehicle
Between-actives comparison (arbutin vs synthetic ellagic acid vs plant ellagic extract), NO vehicle/placebo arm, and OPEN-LABEL (unblinded).
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Azelaic Acid Emerging · trials Moderate Emerging
Counted on the row
6 randomised trials, 673 participants pooled; a systematic review or meta-analysis; independently replicated; findings consistent; no guideline endorsement recorded; best source tier 1.
Studied dose
10%, leave-on, judge after ~12 wk
Dose note
All positive melasma trials used 20%; the OTC 10% products carry a dose-indirectness gap.
Vehicle
Consistent MASI/pigment reduction across multiple independent, mostly double-blind trials of 20% azelaic acid vs hydroquinone; azelaic is generic and off-patent, so the evidence base is not manufacturer-captured the way niacinamide's is.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Cysteamine Hydrochloride Supported · trials Moderate Promising
Counted on the row
2 randomised trials, 90 participants pooled; a systematic review or meta-analysis; independently replicated; findings consistent; no guideline endorsement recorded; best source tier 1; GRADE certainty moderate.
Studied dose
5%, leave-on
Dose note
Endpoint = MASI/mMASI change in diagnosed melasma. The 5% cream is the studied presentation.
Vehicle
Cysteamine is malodorous and unstable, so it is used as a short-contact product; tolerability (adverse events markedly higher than vehicle, RR ~12.6) is the practical limiter, not efficacy.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Hydroquinone Established · named in a guideline, trials Strong Promising
Standing document
Delphi consensus on melasma management by international experts and pigmentary disorders society (Journal of the European Academy of Dermatology and Venereology 2026 (published online 25 Sep 2025); PMID 40996222), opened Jul 27, 2026
Counted on the row
6 randomised trials, 673 participants pooled; a systematic review or meta-analysis; independently replicated; findings consistent; no guideline endorsement recorded; best source tier 1.
Studied dose
2%, leave-on, judge after ~12 wk
Dose note
The triple-combination cream (TCC) is more effective than HQ alone (Cochrane RR 1.58). 4% HQ monotherapy beat placebo for complete melasma clearance (40% vs 10%).
Vehicle
Hydroquinone is the reference depigmentant against which nearly every melasma comparator (azelaic acid, thiamidol, tranexamic acid) is measured. Its effect is not vehicle-dependent in the way vitamin C or the AHAs are.
Checked
against its sources on Jul 27, 2026; next review Jul 17, 2027
Kojic Acid Supported · named in a guideline, trials Moderate Limited
Standing document
Delphi consensus on melasma management by international experts and pigmentary disorders society (Journal of the European Academy of Dermatology and Venereology (published online 25 Sep 2025; PMID 40996222)), opened Jul 27, 2026
Counted on the row
1 randomised trial, 20 participants pooled; findings insufficient; no guideline endorsement recorded; best source tier 1.
Studied dose
1–2%, leave-on, judge after ~12 wk
Dose note
Kojic acid is chemically unstable in leave-on products; most positive melasma data are for combinations, not kojic alone.
Vehicle
Kojic acid is a tyrosinase inhibitor, but its clinical melasma evidence base is dominated by combination formulas; monotherapy signal is weak and comes from a single small single-blind trial.
Checked
against its sources on Jul 27, 2026; next review Jul 17, 2027
Glycyrrhetinic Acid Thin · trials Moderate Limited
Counted on the row
1 randomised trial, 20 participants pooled; findings insufficient; no guideline endorsement recorded; best source tier 2.
Dose note
Liquiritin is the licorice flavonoid tested; glabridin is the other commonly-marketed licorice active but appears only in combination products here.
Vehicle
Split-face vehicle-controlled design (liquiritin cream one side, vehicle the other) — a genuine internal control, the strongest feature of this row.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Niacinamide Thin · trials Moderate Limited
Counted on the row
2 randomised trials, 57 participants pooled; findings consistent; no guideline endorsement recorded; best source tier 3.
Studied dose
4%, leave-on, judge after ~8 wk
Dose note
Only ONE concentration (4%) and ONE duration (8 weeks) have ever been tested in diagnosed melasma — no dose-response and no long-term data exist. Melasma is relapse-prone and both trials stopped at 8 weeks, so durability is entirely unknown. Application frequency is not stated in Navarrete-Solís.
Vehicle
CRITICAL CONFOUNDER — BOTH trials mandated aggressive photoprotection on top of niacinamide. Navarrete-Solís instructed all patients to use SPF 50+ broad spectrum EVERY 3 HOURS and had NO vehicle arm (niacinamide vs hydroquinone split-face only), so both sides improved and neither can be separated from the sunscreen. Campuzano-García gave every arm SPF 50+, and its PLACEBO arm ALSO dropped significantly on MASI (9.1±1.4 → 7.1±1.2, p=0.03) — a direct in-trial demonstration that the sunscreen background alone moves the endpoint. NO published trial isolates 4% niacinamide against a vehicle without this photoprotection layer.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Retinoids Thin · trials Moderate Limited
Counted on the row
1 randomised trial; independently replicated; findings insufficient; no guideline endorsement recorded; best source tier 1.
Studied dose
0.1%, leave-on, judge after ~12 wk
Dose note
FAMILY ROW. Melasma evidence is Rx tretinoin's (monotherapy trial + triple-combination component); it does not transfer to OTC members.
Vehicle
Most retinoid melasma data is as a component of Kligman's triple-combination, where the retinoid is not isolated. The one isolable monotherapy trial is Rx tretinoin. Melasma is a distinct, recalcitrant entity — this evidence must not be read onto the general hyperpigmentation claim.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Tranexamic Acid Established · trials Strong Well-established
Counted on the row
4 randomised trials, 288 participants pooled; a systematic review or meta-analysis; independently replicated; findings consistent; no guideline endorsement recorded; best source tier 1; GRADE certainty high.
Studied dose
2–5%, leave-on, judge after ~12 wk
Dose note
The tier-1 GRADE rating is for TOPICAL, self-applied TXA on the melasma (MASI) endpoint — the correct finished-product class. Do not import the larger oral-TXA effect size to a topical product; the routes are distinct.
Vehicle
The single HIGH-certainty comparison (TXA vs placebo, MASI) was a 2.5% CATAPLASM (poultice), not a leave-on serum; the two moderate-certainty TXA-vs-hydroquinone trials used 5% cream, which better matches catalog leave-ons. A 2% serum in a fast-absorbing vehicle is not guaranteed to reproduce a 5% cream or a cataplasm — extrapolation across vehicle/concentration is a real but secondary caveat, not a reason to reject the tier-1 rating.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Tretinoin 0.025% Thin · trials Moderate Limited
Counted on the row
2 randomised trials, 68 participants pooled; a systematic review or meta-analysis; findings consistent; no guideline endorsement recorded; best source tier 1.
Studied dose
0.1%, leave-on, judge after ~12 wk
Dose note
Tretinoin monotherapy works slowly and modestly; it is more commonly deployed as one component of the Kligman triple-combination cream (hydroquinone + tretinoin + corticosteroid).
Vehicle
Tretinoin lightens melasma by accelerating epidermal turnover / pigment dispersal and reducing epidermal melanin; the pivotal trials showed histologic reduction in epidermal pigment, but onset is slow and retinoid dermatitis is common (67-88% of treated patients).
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Ascorbic Acid Thin · trials Moderate Limited
Counted on the row
1 randomised trial, 16 participants pooled; findings insufficient; no guideline endorsement recorded; best source tier 2.
Studied dose
4–5%, leave-on, judge after ~12 wk
Dose note
Endpoint is MASI/mMASI or a validated melasma measure in DIAGNOSED melasma — a distinct, relapse-prone entity; PIH/dark-spot evidence does not transfer here.
Vehicle
Ascorbic acid is famously unstable; trials seldom characterize the delivered/penetrated concentration, and the one clean topical trial had no vehicle arm — so absolute efficacy vs no treatment remains unestablished.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources

Nail Fungus 9 rows

Amorolfine Established · regulator recognises it, trials Moderate Limited
Standing document
Loceryl 5% w/v Medicated Nail Lacquer — Summary of Product Characteristics (SmPC) (electronic medicines compendium (emc) / MHRA-authorised product, Galderma (U.K.) Limited), opened Aug 11, 2026
Counted on the row
trial count not established; a systematic review or meta-analysis; findings mixed; guideline endorsement conditional; best source tier 1.
Studied dose
5%, leave-on, judge after ~24 wk
Dose note
A randomized dose-comparison found 5% superior to 2% once weekly, so the marketed strength is the studied strength. Trials run 6-12 months; anything shorter is untested.
Vehicle
The lacquer vehicle is the delivery system — it forms an occlusive film that carries drug into the nail plate. Evidence from lacquer trials cannot be transferred to creams, oils, or 'nail serums' containing amorolfine-adjacent actives.
Checked
against its sources on Aug 11, 2026
Ciclopirox Established · regulator recognises it, trials Moderate Promising
Standing document
Ciclopirox Topical Solution, 8% (Nail Lacquer) — CICLOPIROX solution (DailyMed (US National Library of Medicine) — FDA prescription drug label, Akorn Operating Company LLC), opened Aug 11, 2026
Counted on the row
trial count not established; a systematic review or meta-analysis; findings consistent; no guideline endorsement recorded; best source tier 1; GRADE certainty moderate.
Studied dose
8%, leave-on, judge after ~4 wk
Dose note
The newer 8% hydrolacquer (P-3051, water-soluble HPCH vehicle) outperformed both the conventional 8% lacquer and amorolfine 5% for complete cure in the Cochrane comparisons — the vehicle, not the concentration, is the differentiator.
Vehicle
Lacquer penetration of the nail plate is the entire clinical problem. Cochrane's ratings apply to the 8% lacquer/hydrolacquer presentations; nothing here transfers to creams, ointments, or tinea pedis (skin) evidence.
Checked
against its sources on Aug 11, 2026
Sources
Efinaconazole Established · regulator recognises it, trials Moderate Well-established
Standing document
JUBLIA (efinaconazole) topical solution, 10% (DailyMed (US National Library of Medicine) — FDA prescription drug label, Bausch Health US LLC), opened Aug 11, 2026
Counted on the row
2 randomised trials, 1,655 participants pooled; a systematic review or meta-analysis; findings consistent; no guideline endorsement recorded; best source tier 1; GRADE certainty high.
Studied dose
10%, leave-on, judge after ~48 wk
Dose note
Absolute complete-cure rates are modest even at full adherence: 15.2-17.8% vs 3.3-5.5% for vehicle at week 52. High-certainty evidence of a real but limited effect — the honest framing is 'the best topical, still far below oral terbinafine'.
Vehicle
Low-surface-tension solution engineered for nail-plate and subungual penetration; the vehicle is part of the drug's design. Evidence applies only to the 10% solution as studied.
Checked
against its sources on Aug 11, 2026
Sources
Tavaborole Established · regulator recognises it, trials Strong Well-established
Standing document
TAVABOROLE- tavaborole topical solution, 5% solution (DailyMed, U.S. National Library of Medicine — FDA drug label (ANDA 211297, generic of Kerydin)), opened Aug 11, 2026
Counted on the row
2 randomised trials, 1,198 participants pooled; a systematic review or meta-analysis; findings consistent; guideline endorsement strong; best source tier 1; GRADE certainty moderate.
Studied dose
5%, leave-on, judge after ~48 wk
Dose note
No OTC presentation exists. Absolute complete-cure rates are low — 6.5% and 9.1% vs 0.5% and 1.5% for vehicle — statistically robust but clinically modest, which is the known ceiling of topical onychomycosis therapy.
Vehicle
Alcohol-based dropper solution engineered for nail-plate penetration — the delivery problem is the whole disease; skin antifungal data never transfers to nails.
Checked
against its sources on Aug 11, 2026
Tea Tree Oil Thin · trials Moderate Limited
Counted on the row
2 randomised trials, 177 participants pooled; findings insufficient; no guideline endorsement recorded; best source tier 3.
Studied dose
2–100%, leave-on
Dose note
The claim's endpoint is mycological/complete cure in diagnosed onychomycosis. Neither confirmed trial establishes tea tree alone against vehicle on that endpoint.
Vehicle
Nail-plate penetration is the entire problem for this claim; per the vocabulary, athlete's-foot (tinea pedis) skin evidence must never transfer to nails, and no penetration-engineered tea tree vehicle has been trialed.
Checked
no verification date on record
Terbinafine Thin · trials Moderate Limited
Counted on the row
1 randomised trial; findings insufficient; no guideline endorsement recorded; best source tier 1.
Studied dose
1%, leave-on
Dose note
The dose ladder is discontinuous: nothing connects OTC 1% skin creams to the 10% nail-solution data. Evidence from one format does not transfer to the other.
Vehicle
Nail-plate penetration is the entire problem in onychomycosis. MOB-015's result comes from a purpose-built 10% solution vehicle; conventional cream vehicles do not deliver terbinafine through the nail plate, which is why the OTC label excludes nails. A catalog product's vehicle, not the ingredient name, decides whether this evidence applies.
Checked
no verification date on record
Sources
Tolnaftate 1% Thin · trials Moderate Limited
Counted on the row
1 randomised trial, 69 participants pooled; findings insufficient; no guideline endorsement recorded; best source tier 3.
Studied dose
1%, leave-on
Dose note
The one positive trial used Tolcylen, a 1% tolnaftate solution whose vehicle also contains urea, lactic acid, undecylenic acid and propylene glycol — the tested product, not tolnaftate per se.
Vehicle
Tolcylen suspends tolnaftate in a low-surface-tension penetrating vehicle carrying three other active-adjacent ingredients (urea, lactic acid, undecylenic acid); any effect cannot be attributed to tolnaftate alone. Conventional tolnaftate creams/powders have no nail-plate delivery mechanism, which is what the mandatory FDA label exclusion reflects.
Checked
no verification date on record
Undecylenic Acid Thin · trials Limited Limited
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 1.
Studied dose
10–25%, leave-on
Dose note
Widely sold as ~25% 'fungal nail' brush-on solutions — a monograph tinea active marketed at the one site the monograph explicitly excludes.
Vehicle
No delivery system in consumer undecylenic products addresses nail-plate penetration; brush-on polish formats coat the plate surface. The published nail literature for this ingredient consists of 1960s–70s case series (Mykosen, in German), contact-dermatitis case reports from nail solutions, and formulation papers — not controlled trials.
Checked
no verification date on record
Sources
Hydroxyethyl Urea Thin · trials Limited Limited
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 3.
Studied dose
20–40%, leave-on, judge after ~2 wk
Dose note
The trial dose with usable data is 40% under occlusive dressing — far above the 2–10% humectant use in ordinary skincare.
Vehicle
Urea's entire contribution is vehicle-scale physical action: 40% under occlusion chemically ablates infected nail keratin and improves antifungal penetration. It is a debridement adjunct and penetration enhancer, not a fungicide — no trial tests urea alone against the mycological-cure endpoint, because removing infected keratin without killing the dermatophyte is not expected to cure.
Checked
no verification date on record

Post-Procedure 7 rows

Allantoin Thin · trials Limited Limited
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 3.
Studied dose
0.5–2%, leave-on
Dose note
No post-procedure product isolates allantoin; the 6% SD-101 evidence is from a different condition entirely.
Vehicle
In catalog post-procedure creams allantoin is one soothing/keratolytic component among several (panthenol, centella, etc.), so any observed post-laser benefit cannot be attributed to allantoin — indirectness_intervention.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Aloe Barbadensis Leaf Extract Thin · trials Moderate Limited
Counted on the row
1 randomised trial; findings insufficient; no guideline endorsement recorded; best source tier 2.
Dose note
The dermabrasion study delivered aloe via a polyethylene-oxide dressing saturated with stabilized aloe gel.
Vehicle
Delivery differs sharply across the base: an occlusive gel/dressing during re-epithelialization is not the same exposure as a leave-on cream. Adverse contact-dermatitis reactions to aloe preparations after dermabrasion and chemical peel have also been reported, so tolerability is not uniformly benign.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Centella Asiatica Extract Thin · trials Moderate Limited
Counted on the row
1 randomised trial, 30 participants pooled; findings insufficient; no guideline endorsement recorded; best source tier 3.
Studied dose
0.05%, leave-on
Dose note
0.05% is far below the concentrations cosmetic marketing implies, and ECa 233 is a specific standardized extract (defined triterpene ratio), not generic centella extract. A product listing 'Centella Asiatica Extract' at an undisclosed concentration is not this intervention.
Vehicle
Gel vehicle, split-face against an otherwise identical placebo gel — the cleanest attribution available for this ingredient on any claim in this batch. Note the comparator is placebo gel, not no-treatment: the trial isolates centella, not 'gel after laser'.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Ceramides NP, AP, EOP Thin · trials Moderate Limited
Counted on the row
1 randomised trial, 30 participants pooled; findings insufficient; no guideline endorsement recorded; best source tier 3.
Dose note
Ceramide concentration is UNSTATED in both post-procedure trials. Nawaz 2025 discloses only an unquantified 'Lipid Complex' of ceramides + cholesterol + fatty acids. Mortensen 2001 used Locobase Repair, specified as 63% total lipids comprising cholesterol, ceramide and free fatty acids — the 63% is the WHOLE LIPID PHASE, not the ceramide fraction, which is never given.
Vehicle
*** THE DECISIVE FINDING — AND THE CONFOUND PREDICTED IN THE BRIEF MATERIALIZED EXACTLY. *** The single RCT (Nawaz 2025, randomized, double-blind, split-face, n=30, post-microneedling) compared the ceramide lipid complex against ITS OWN MATCHED VEHICLE — the identical cream minus the Lipid Complex. THE VEHICLE PERFORMED AS WELL AS OR BETTER THAN THE ACTIVE ON EVERY OBJECTIVE ENDPOINT. TEWL fell 15.0% (active) vs 16.3% (vehicle) at week 4 — vehicle NUMERICALLY SUPERIOR; erythema fell ~1.7–1.8% clinically and ~8.6% by VISIA on BOTH arms. Verbatim: 'no statistically significant differences were observed between the active and vehicle formulations.' Only the UNBLINDED satisfaction survey favoured the active. Mortensen 2001 independently reproduces the pattern against PETROLATUM: no significant difference in erythema, scaling, oedema, pain, duration of wound healing, or final cosmetic outcome.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Panthenol Thin · trials Moderate Limited
Counted on the row
1 randomised trial, 38 participants pooled; findings insufficient; no guideline endorsement recorded; best source tier 3.
Studied dose
0.5–2%, leave-on
Dose note
The 5% figure is the ONLY concentration with a controlled post-procedure trial behind it. This row will attach to 184 catalog products, in most of which panthenol is a minor humectant far below 5% and far down the INCI. The ointment vehicle is also inseparable from the tested effect.
Vehicle
The single admissible trial compared a dexpanthenol-containing OINTMENT against petroleum jelly — two different vehicles, not a matched vehicle control. The trial therefore cannot separate 'dexpanthenol' from 'this ointment base vs petrolatum'. Extrapolating it to panthenol at 1% in a lightweight post-procedure gel is unsupported.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Petrolatum Supported · mechanism established, trials Moderate Promising
Counted on the row
1 randomised trial, 922 participants pooled; independently replicated; findings consistent; no guideline endorsement recorded; best source tier 1.
Dose note
Petrolatum is the occlusive comparator against which topical antibiotic ointments are judged; it maintains a moist wound environment.
Vehicle
The active property is occlusion / moist-wound-healing support with an essentially inert, non-sensitizing vehicle; the trials establish equivalence to topical antibiotics for healing and infection with markedly lower contact-allergy and irritation.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Squalane Thin No evidence No evidence
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded.
Dose note
No procedure-recovery trial dosing squalane exists to define a range.
Vehicle
N/A — no finished-product trial exists for this endpoint.
Checked
no verification date on record
Sources
none recorded

Psoriasis 7 rows

Aloe Barbadensis Leaf Extract Emerging · trials Moderate Emerging · contested
Counted on the row
3 randomised trials, 100 participants pooled; findings conflicting; no guideline endorsement recorded; best source tier 2.
Studied dose
0.5%, leave-on
Dose note
Formulation varies wildly between the positive and negative trials — a 0.5% standardized extract vs an off-the-shelf commercial gel — which is itself part of why results conflict.
Vehicle
Aloe preparations are chemically heterogeneous (species, part, processing, concentration), so 'aloe vera' is not one intervention. This heterogeneity plausibly drives the conflicting trial results.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Ceramides NP, AP, EOP Thin · trials Limited Limited
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 3.
Dose note
Endpoint = change in PASI or plaque clearance. Ceramide creams are barrier-repair adjuncts, not plaque-clearing monotherapy.
Vehicle
Psoriatic skin has a documented ceramide DEFICIT, which is the rationale for ceramide replacement as barrier support — but a barrier-repair rationale is not the same as evidence of PASI reduction.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Coal Tar 0.5% Established · regulator recognises it, trials Strong Well-established
Standing document
§ 358.710 Active ingredients for the control of dandruff, seborrheic dermatitis, or psoriasis. (U.S. Food and Drug Administration — 21 CFR Part 358 Subpart H, Code of Federal Regulations via GovInfo), opened Jul 27, 2026
Counted on the row
trial count not established; independently replicated; findings consistent; guideline endorsement strong; best source tier 1.
Studied dose
0.5–5%, judge after ~4 wk
Dose note
Endpoint = PASI/plaque clearance in diagnosed psoriasis. Coal tar is a legacy antipsoriatic with a genuine disease-modifying (antiproliferative) effect, unlike the emollient-adjunct ingredients.
Vehicle
Efficacy is concentration- and contact-time-dependent; the strong evidence is for leave-on tar preparations, not brief-contact shampoo, though shampoo has its own scalp-psoriasis evidence.
Checked
against its sources on Jul 27, 2026; next review Jul 17, 2027
Lactic Acid Thin · trials Limited Limited
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 3.
Studied dose
12%, leave-on, judge after ~4 wk
Dose note
Psoriasis endpoint is PASI/plaque clearance. Do NOT confuse topical lactic acid with INDOLE-3-lactic acid, a gut/microbiome tryptophan metabolite studied systemically in mouse psoriasis models — a different molecule and route.
Vehicle
Lactic acid is a humectant and mild keratolytic; in psoriasis products it functions as the emollient/descaling VEHICLE base, not as a disease-active ingredient.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Petrolatum Thin · trials Limited Limited
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 3.
Dose note
Endpoint = change in PASI or plaque clearance in diagnosed psoriasis. Petrolatum is adjunctive emollient care, not a plaque-clearing monotherapy.
Vehicle
Petrolatum is, in most psoriasis RCTs, the VEHICLE or emollient CONTROL arm (e.g. tazarotene gel 'plus petrolatum' vs calcipotriol) — i.e. the thing the active is measured against, not the tested active. Any 'effect' attributed to it is essentially the vehicle response.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Salicylic Acid Established · regulator recognises it, trials Limited Limited
Standing document
§ 358.710 Active ingredients for the control of dandruff, seborrheic dermatitis, or psoriasis (US FDA — 21 CFR 358.710 (via Cornell Legal Information Institute)), opened Jul 27, 2026
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 3.
Studied dose
2–6%, judge after ~6 wk
Dose note
The psoriasis endpoint is PASI or plaque clearance in diagnosed psoriasis. Descaling (removing scale) is a keratolytic mechanism, NOT disease clearance — and salicylic acid can reduce the efficacy of, and should not be co-applied with, UV phototherapy or calcipotriol.
Vehicle
Salicylic acid's recognized role is adjunctive: softening and removing scale so a disease-modifying agent (steroid, vitamin-D analogue, tar, UV) can penetrate — it is a penetration enhancer, not the therapeutic active.
Checked
against its sources on Jul 27, 2026; next review Jul 17, 2027
Hydroxyethyl Urea Thin · trials Moderate Limited
Counted on the row
1 randomised trial, 10 participants pooled; findings insufficient; no guideline endorsement recorded; best source tier 2.
Studied dose
5–10%, leave-on, judge after ~2 wk
Dose note
Endpoint = PASI or plaque clearance. Urea's role is keratolytic scale-reduction and penetration enhancement for co-applied actives, not disease clearance.
Vehicle
Urea is most often used to ENHANCE penetration of calcipotriol/corticosteroids (a combination role, e.g. urea 20% + calcipotriol for nail psoriasis), so isolating its own contribution to plaque clearance is difficult.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027

Rosacea 12 rows

Allantoin Thin No evidence No evidence
Counted on the row
trial count not established; findings insufficient; no guideline endorsement recorded; best source tier 3.
Studied dose
0.5–2%, leave-on
Dose note
No trial isolates allantoin as the intervention for rosacea. The one rosacea RCT tests a COMPOUND gel whose declared main components are onion extract and quercetin (plus heparin sodium), with allantoin a minor conditioning ingredient.
Vehicle
Allantoin's established mechanism is soothing/keratolytic/barrier-conditioning, not vasoregulatory or anti-inflammatory at the level rosacea requires. Any erythema improvement in a compound product cannot be attributed to allantoin.
Checked
no verification date on record
Azelaic Acid Established · regulator recognises it, trials Strong Well-established
Standing document
FINACEA® (azelaic acid) Gel, 15% (US Food and Drug Administration prescribing information, via DailyMed (National Library of Medicine)), opened Jul 27, 2026
Counted on the row
trial count not established; a systematic review or meta-analysis; independently replicated; findings consistent; no guideline endorsement recorded; best source tier 1; GRADE certainty high.
Studied dose
15%, leave-on, judge after ~12 wk
Dose note
20% cream is the ACNE presentation. The 15% gel is not weaker than the 20% cream — the gel vehicle raises bioavailability, which is why the lower nominal % carries the rosacea indication. Reading these as a dose ladder is a category error.
Vehicle
Vehicle is decisive, not incidental. Extrapolating 15%-gel Rx trial efficacy to a 10% OTC leave-on in an unrelated vehicle is unsupported.
Checked
against its sources on Jul 27, 2026; next review Jul 17, 2027
Sources
Centella Asiatica Extract Thin · trials Moderate Limited
Counted on the row
1 randomised trial, 64 participants pooled; findings insufficient; no guideline endorsement recorded; best source tier 3.
Dose note
No concentration can be quoted for this claim. A rinse-off mask used 3x/week is a fundamentally different exposure from a leave-on cream, and the trial cannot tell us what a leave-on centella product does in rosacea.
Vehicle
Rinse-off mask, 3x weekly, applied ON TOP of oral minocycline 50mg BID plus a leave-on hydrating cream that BOTH arms received. Contact time is short and the active is co-applied with a systemic tetracycline of established efficacy — the vehicle and the background therapy, not centella, may carry the observed change.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Camellia Oleifera Thin No evidence No evidence
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded.
Studied dose
2%, leave-on
Dose note
No full peer-reviewed trial defines a green-tea dose against a validated rosacea endpoint.
Vehicle
EGCG is oxidation-prone and its topical stability/penetration are formulation-dependent; the single rosacea report used a 'polyphenon' hydrophilic cream whose full formulation and stabilization were never published in a peer-reviewed paper.
Checked
no verification date on record
Sources
none recorded
Ivermectin Established · regulator recognises it, trials Strong Well-established
Standing document
SOOLANTRA (ivermectin) cream, 1%, for topical use — INDICATIONS AND USAGE (U.S. FDA-approved prescription drug label (Galderma Laboratories), via DailyMed / U.S. National Library of Medicine), opened Jul 27, 2026
Counted on the row
trial count not established; a systematic review or meta-analysis; independently replicated; findings consistent; no guideline endorsement recorded; best source tier 1; GRADE certainty high.
Studied dose
1%, leave-on, judge after ~12 wk
Dose note
Endpoint = reduction in inflammatory (papulopustular) rosacea lesions / IGA. Ivermectin is Rx-only; a catalog 'rosacea' cosmetic would not contain it.
Vehicle
The evidence is specifically for the 1% cream; there is no OTC ivermectin presentation to extrapolate to.
Checked
against its sources on Jul 27, 2026; next review Jul 17, 2027
Glycyrrhetinic Acid Thin · trials Limited Limited
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 2.
Dose note
The rosacea endpoint is papulopustular lesion count OR a validated erythema assessment (CEA/IGA) in DIAGNOSED rosacea — not a colorimetric a* reading in an uncontrolled cosmetic study.
Vehicle
The cosmetic rosacea studies pair licochalcone A with 4-t-butylcyclohexanol (a TRPV1 antagonist), so any redness benefit cannot be attributed to licorice alone.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Metronidazole Established · regulator recognises it, trials Moderate Promising
Standing document
METROGEL® (metronidazole) topical gel (US Food and Drug Administration prescribing information, via DailyMed (National Library of Medicine)), opened Jul 27, 2026
Counted on the row
trial count not established; a systematic review or meta-analysis; independently replicated; findings consistent; no guideline endorsement recorded; best source tier 1; GRADE certainty moderate.
Studied dose
0.75–1%, leave-on, judge after ~8 wk
Dose note
Endpoint = reduction in inflammatory (papulopustular) rosacea lesions / IGA. Metronidazole is Rx-only; not a cosmetic-catalog active.
Vehicle
Evidence is for the 0.75–1% Rx gel/cream; no OTC presentation exists to extrapolate to.
Checked
against its sources on Jul 27, 2026; next review Jul 17, 2027
Niacinamide Thin · trials Limited Limited
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 3.
Dose note
Niacinamide is not a prescription agent for rosacea and has no established therapeutic dose for this indication. The 2% figure describes ONE commercial multi-ingredient moisturizer, not a validated dose.
Vehicle
The single study reporting a rosacea endpoint tested a complete commercial multi-ingredient moisturizer (Olay Total Effects 7x Fragrance Free), NOT niacinamide in a vehicle-controlled comparison. Any clinical change is inseparable from the base's own moisturizing/barrier action — the authors themselves attribute the erythema reduction to improved stratum corneum barrier rather than a direct anti-inflammatory effect of niacinamide. Baseline followed a 2-week washout on a DIFFERENT Olay moisturizer plus Olay cleansers, and concurrent prescription rosacea therapy was PERMITTED throughout.
Checked
no verification date on record
Sources
Panthenol Thin No evidence No evidence
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 1.
Studied dose
1–5%, leave-on
Dose note
COULD NOT ESTABLISH — no trial on this endpoint exists.
Vehicle
Not assessable — no clinical trial on this endpoint exists.
Checked
no verification date on record
Sources
Sulfur 10% Thin · trials Moderate Limited
Counted on the row
1 randomised trial; findings insufficient; no guideline endorsement recorded; best source tier 3.
Studied dose
2–5%, leave-on
Dose note
No trial isolates sulfur. The rosacea evidence is always the sulfacetamide-sulfur COMBINATION, in which sodium sulfacetamide (a sulfonamide antibacterial) is the more plausible active.
Vehicle
Attributing the combination's effect to sulfur specifically is unsupported. The available RCT is an active-comparator design (vs metronidazole), not vehicle-controlled, so it cannot even establish the combination's effect size over placebo.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Titanium Dioxide Thin No evidence No evidence
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 1.
Dose note
No controlled study of topical titanium dioxide on a rosacea lesion-count or erythema endpoint was found at any concentration.
Vehicle
Even thinner than zinc oxide: not even an uncontrolled case series on a rosacea endpoint was found. The only hit for 'titanium dioxide rosacea' on PubMed is a 2004 Cutis narrative review on UV and rosacea. A 1998 Cutis bilateral comparison (Nichols, Desai, Lebwohl) studied IRRITANCY of sunscreen ingredients in rosacea patients — a tolerability question that belongs to Fit, not an efficacy endpoint.
Checked
no verification date on record
Sources
Zinc Oxide Thin No evidence No evidence
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 3.
Dose note
No study has tested topical zinc oxide at any concentration against a control on a rosacea lesion-count or erythema endpoint.
Vehicle
The only rosacea-endpoint data found for anything called zinc oxide is an UNCONTROLLED 8-week case series (n=35) of nanodiamond-zinc oxide (ND-ZnO) cream/serum — a proprietary composite material in multi-ingredient finished products, not zinc oxide as it appears in a mineral sunscreen. Three of seven authors are employees of the material's manufacturer (B.A.I. Biosciences); the paper itself states the absence of a placebo control limits generalizability and calls for RCTs. Mean expert erythema score fell 0.86 -> 0.54. Not admissible as a controlled anchor; recorded for transparency. Separately: the BJD 2019 GRADE systematic review of rosacea interventions assesses ORAL zinc sulfate (Sharquie 2006; Bamford 2012) — an oral zinc salt, a different intervention and route; it does NOT transfer to topical zinc oxide. Sunscreen-vehicle trials (e.g. metronidazole 1% + SPF15 vs SPF15 vehicle, Tan 2002) put sunscreen in BOTH arms, so they test metronidazole, not the filter.
Checked
no verification date on record
Sources

Rough Texture 4 rows

Glycolic Acid Established · mechanism established Limited Established mechanism
Counted on the row
trial count not established; no guideline endorsement recorded; best source tier 2.
Studied dose
4–10%, leave-on, judge after ~8 wk
Dose note
Corneocyte desmosome disruption is a chemical mechanism. REQUIRES functional pH — neutralised to the salt it is a humectant, not an exfoliant.
Vehicle
The pH dependence is the whole claim. See inci-flag-rules -> antagonism_pairs.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
  • International Cosmetic Ingredient Dictionary and Handbook — declared function — Personal Care Products Council
  • Harry's Cosmeticology / Baran & Maibach Textbook of Cosmetic Dermatology — humectant, emollient and occlusive classification — standard reference text
Lactic Acid Established · mechanism established Limited Established mechanism
Counted on the row
trial count not established; no guideline endorsement recorded; best source tier 2.
Studied dose
5–12%, leave-on, judge after ~4 wk
Dose note
AHA; also humectant at lower concentrations — dose changes the claim.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
  • International Cosmetic Ingredient Dictionary and Handbook — declared function — Personal Care Products Council
  • Harry's Cosmeticology / Baran & Maibach Textbook of Cosmetic Dermatology — humectant, emollient and occlusive classification — standard reference text
Salicylic Acid Established · mechanism established Limited Established mechanism
Counted on the row
trial count not established; no guideline endorsement recorded; best source tier 2.
Studied dose
0.5–2%, leave-on, judge after ~6 wk
Dose note
Lipophilic BHA; keratolytic at functional pH.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
  • International Cosmetic Ingredient Dictionary and Handbook — declared function — Personal Care Products Council
  • United States Pharmacopeia / National Formulary — excipient monograph — USP-NF
Hydroxyethyl Urea Established · mechanism established Limited Established mechanism
Counted on the row
trial count not established; no guideline endorsement recorded; best source tier 2.
Studied dose
10%, leave-on, judge after ~2 wk
Dose note
At >10% urea's mechanism switches from humectant to keratolytic.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
  • International Cosmetic Ingredient Dictionary and Handbook — declared function — Personal Care Products Council
  • United States Pharmacopeia / National Formulary — excipient monograph — USP-NF

Sensitivity 1 row

Centella Asiatica Extract Standing unverified No evidence Not applicable
Counted on the row
0 randomised trials; findings mixed; no guideline endorsement recorded; best source tier 3.
Studied dose
0.05%, leave-on
Dose note
No dose-response for sensitivity can be established. The only trial stating a concentration is in the WRONG POPULATION; the only trial in the right population does not state a concentration.
Vehicle
Su 2025's formulation confound is disqualifying for an ingredient-level claim: centella is co-formulated with ceramide NP AND panthenol — both independently barrier-active — and with no control arm, the observed improvements are exactly what a ceramide-plus-panthenol cream would produce on its own.
Checked
no verification date on record

Sun Damage 23 rows

Avobenzone 3% Established · regulator recognises it, trials Moderate Emerging
Standing document
21 CFR § 352.10 Sunscreen active ingredients (US FDA — 21 CFR Part 352 (Sunscreen Drug Products for Over-the-Counter Human Use), via Cornell Legal Information Institute; part heading marked [STAYED INDEFINITELY], effective date stayed at 66 FR 67485), opened Sep 1, 2026
Counted on the row
1 randomised trial, 903 participants pooled; findings insufficient; no guideline endorsement recorded; best source tier 1.
Studied dose
up to 3%, leave-on
Dose note
CONCENTRATION IS THE WRONG UNIT FOR THIS CLAIM. FDA attaches the early-skin-aging claim to the finished product's TESTED Broad Spectrum + SPF>=15 performance (21 CFR 201.327), not to any filter percentage. The % here is the permitted maximum for the active, not a dose ladder; delivered protection is a formulation property (film, photostability, dispersion). Same policy as the zinc-oxide and titanium-dioxide rows.
Vehicle
For a UV filter the vehicle IS the product: film-forming, photostability (avobenzone needs a stabiliser), and rub-off resistance decide delivered SPF/UVA-PF. This is why the regulatory claim is gated on a performance test of the finished product, and why the engine's G4 gate reads the INCI for adequate UVA filtration rather than trusting the hero.
Checked
against its sources on Sep 1, 2026; next review Sep 1, 2027
Sources
Phenylbenzimidazole Sulfonic Acid Established · regulator recognises it, trials Moderate Limited
Standing document
21 CFR § 352.10 Sunscreen active ingredients (US FDA — 21 CFR Part 352 (Sunscreen Drug Products for Over-the-Counter Human Use), via Cornell Legal Information Institute; part heading marked [STAYED INDEFINITELY], effective date stayed at 66 FR 67485), opened Sep 1, 2026
Counted on the row
1 randomised trial, 903 participants pooled; findings insufficient; no guideline endorsement recorded; best source tier 1.
Studied dose
up to 4%, leave-on
Dose note
CONCENTRATION IS THE WRONG UNIT FOR THIS CLAIM. FDA attaches the early-skin-aging claim to the finished product's TESTED Broad Spectrum + SPF>=15 performance (21 CFR 201.327), not to any filter percentage. The % here is the permitted maximum for the active, not a dose ladder; delivered protection is a formulation property (film, photostability, dispersion). Same policy as the zinc-oxide and titanium-dioxide rows.
Vehicle
For a UV filter the vehicle IS the product: film-forming, photostability (avobenzone needs a stabiliser), and rub-off resistance decide delivered SPF/UVA-PF. This is why the regulatory claim is gated on a performance test of the finished product, and why the engine's G4 gate reads the INCI for adequate UVA filtration rather than trusting the hero.
Checked
against its sources on Sep 1, 2026; next review Sep 1, 2027
Sources
Ethylhexyl Triazone Established · regulator recognises it, trials Moderate Limited
Counted on the row
1 randomised trial, 903 participants pooled; findings insufficient; no guideline endorsement recorded; best source tier 1.
Studied dose
5%, leave-on
Dose note
CONCENTRATION IS THE WRONG UNIT FOR THIS CLAIM. FDA attaches the early-skin-aging claim to the finished product's TESTED Broad Spectrum + SPF>=15 performance (21 CFR 201.327), not to any filter percentage. The % here is the permitted maximum for the active, not a dose ladder; delivered protection is a formulation property (film, photostability, dispersion). Same policy as the zinc-oxide and titanium-dioxide rows.
Vehicle
For a UV filter the vehicle IS the product: film-forming, photostability (avobenzone needs a stabiliser), and rub-off resistance decide delivered SPF/UVA-PF. This is why the regulatory claim is gated on a performance test of the finished product, and why the engine's G4 gate reads the INCI for adequate UVA filtration rather than trusting the hero.
Checked
against its sources on Sep 1, 2026; next review Sep 1, 2027
Sources
Ferulic Acid Thin · trials Moderate Limited
Counted on the row
2 randomised trials; findings consistent; no guideline endorsement recorded; best source tier 1.
Studied dose
0.5–15%, leave-on
Dose note
Ferulic acid appears only as a ~0.5% stabilizer of the C+E antioxidant pair. No study applies ferulic acid as an isolated topical.
Vehicle
The measured photoprotection (halved erythema/sunburn-cell/thymine-dimer/p53 formation after UV challenge) is a property of the CEFer COMBINATION. Ferulic's own role is documented as chemically STABILIZING the vitamins, i.e. a formulation effect, not an independent action on skin.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Glycolic Acid Thin · trials Moderate Limited
Counted on the row
2 randomised trials, 75 participants pooled; independently replicated; findings consistent; no guideline endorsement recorded; best source tier 1.
Studied dose
5–25%, leave-on, judge after ~8 wk
Dose note
The everyday OTC dose (5%) has the cleaner placebo-controlled facial data; the 25% forearm study adds histologic mechanism.
Vehicle
Both anchor trials are genuinely PLACEBO-controlled (glycolic/AHA vs vehicle on the contralateral side or arm) — the strongest feature of this row.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Homosalate 13.5% Established · regulator recognises it, trials Moderate Limited
Standing document
21 CFR § 352.10 Sunscreen active ingredients (US FDA — 21 CFR Part 352 (Sunscreen Drug Products for Over-the-Counter Human Use), via Cornell Legal Information Institute; part heading marked [STAYED INDEFINITELY], effective date stayed at 66 FR 67485), opened Sep 1, 2026
Counted on the row
1 randomised trial, 903 participants pooled; findings insufficient; no guideline endorsement recorded; best source tier 1.
Studied dose
up to 15%, leave-on
Dose note
CONCENTRATION IS THE WRONG UNIT FOR THIS CLAIM. FDA attaches the early-skin-aging claim to the finished product's TESTED Broad Spectrum + SPF>=15 performance (21 CFR 201.327), not to any filter percentage. The % here is the permitted maximum for the active, not a dose ladder; delivered protection is a formulation property (film, photostability, dispersion). Same policy as the zinc-oxide and titanium-dioxide rows.
Vehicle
For a UV filter the vehicle IS the product: film-forming, photostability (avobenzone needs a stabiliser), and rub-off resistance decide delivered SPF/UVA-PF. This is why the regulatory claim is gated on a performance test of the finished product, and why the engine's G4 gate reads the INCI for adequate UVA filtration rather than trusting the hero.
Checked
against its sources on Sep 1, 2026; next review Sep 1, 2027
Sources
Mexoryl SX Established · regulator recognises it, trials Moderate Limited
Counted on the row
1 randomised trial, 903 participants pooled; findings insufficient; no guideline endorsement recorded; best source tier 1.
Studied dose
10%, leave-on
Dose note
CONCENTRATION IS THE WRONG UNIT FOR THIS CLAIM. FDA attaches the early-skin-aging claim to the finished product's TESTED Broad Spectrum + SPF>=15 performance (21 CFR 201.327), not to any filter percentage. The % here is the permitted maximum for the active, not a dose ladder; delivered protection is a formulation property (film, photostability, dispersion). Same policy as the zinc-oxide and titanium-dioxide rows.
Vehicle
For a UV filter the vehicle IS the product: film-forming, photostability (avobenzone needs a stabiliser), and rub-off resistance decide delivered SPF/UVA-PF. This is why the regulatory claim is gated on a performance test of the finished product, and why the engine's G4 gate reads the INCI for adequate UVA filtration rather than trusting the hero.
Checked
against its sources on Sep 1, 2026; next review Sep 1, 2027
Sources
Mexoryl XL Established · regulator recognises it, trials Moderate Limited
Counted on the row
1 randomised trial, 903 participants pooled; findings insufficient; no guideline endorsement recorded; best source tier 1.
Studied dose
15%, leave-on
Dose note
CONCENTRATION IS THE WRONG UNIT FOR THIS CLAIM. FDA attaches the early-skin-aging claim to the finished product's TESTED Broad Spectrum + SPF>=15 performance (21 CFR 201.327), not to any filter percentage. The % here is the permitted maximum for the active, not a dose ladder; delivered protection is a formulation property (film, photostability, dispersion). Same policy as the zinc-oxide and titanium-dioxide rows.
Vehicle
For a UV filter the vehicle IS the product: film-forming, photostability (avobenzone needs a stabiliser), and rub-off resistance decide delivered SPF/UVA-PF. This is why the regulatory claim is gated on a performance test of the finished product, and why the engine's G4 gate reads the INCI for adequate UVA filtration rather than trusting the hero.
Checked
against its sources on Sep 1, 2026; next review Sep 1, 2027
Sources
Niacinamide Thin · trials Moderate Limited
Counted on the row
1 randomised trial, 50 participants pooled; findings insufficient; no guideline endorsement recorded; best source tier 2.
Studied dose
5–10%, leave-on, judge after ~8 wk
Dose note
10% — heavily marketed and common in the catalog — has NO photodamage trial behind it. There is no dose-response data on this endpoint, so the 10% products are extrapolating from a 5% dataset upward, which the evidence does not license.
Vehicle
The Bissett trial delivered 5% niacinamide in a moisturizer vs the SAME moisturizer as control, which genuinely isolates the actives — a better-designed comparison than most cosmetic trials. Niacinamide is water-soluble and vehicle-tolerant; delivery is not the contested variable here. Sponsorship is.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Octinoxate Established · regulator recognises it, trials Moderate Emerging
Standing document
21 CFR § 352.10 Sunscreen active ingredients (US FDA — 21 CFR Part 352 (Sunscreen Drug Products for Over-the-Counter Human Use), via Cornell Legal Information Institute; part heading marked [STAYED INDEFINITELY], effective date stayed at 66 FR 67485), opened Sep 1, 2026
Counted on the row
1 randomised trial, 903 participants pooled; findings insufficient; no guideline endorsement recorded; best source tier 1.
Studied dose
up to 7.5%, leave-on
Dose note
CONCENTRATION IS THE WRONG UNIT FOR THIS CLAIM. FDA attaches the early-skin-aging claim to the finished product's TESTED Broad Spectrum + SPF>=15 performance (21 CFR 201.327), not to any filter percentage. The % here is the permitted maximum for the active, not a dose ladder; delivered protection is a formulation property (film, photostability, dispersion). Same policy as the zinc-oxide and titanium-dioxide rows.
Vehicle
For a UV filter the vehicle IS the product: film-forming, photostability (avobenzone needs a stabiliser), and rub-off resistance decide delivered SPF/UVA-PF. This is why the regulatory claim is gated on a performance test of the finished product, and why the engine's G4 gate reads the INCI for adequate UVA filtration rather than trusting the hero.
Checked
against its sources on Sep 1, 2026; next review Sep 1, 2027
Sources
Octisalate Established · regulator recognises it, trials Moderate Limited
Standing document
21 CFR § 352.10 Sunscreen active ingredients (US FDA — 21 CFR Part 352 (Sunscreen Drug Products for Over-the-Counter Human Use), via Cornell Legal Information Institute; part heading marked [STAYED INDEFINITELY], effective date stayed at 66 FR 67485), opened Sep 1, 2026
Counted on the row
1 randomised trial, 903 participants pooled; findings insufficient; no guideline endorsement recorded; best source tier 1.
Studied dose
up to 5%, leave-on
Dose note
CONCENTRATION IS THE WRONG UNIT FOR THIS CLAIM. FDA attaches the early-skin-aging claim to the finished product's TESTED Broad Spectrum + SPF>=15 performance (21 CFR 201.327), not to any filter percentage. The % here is the permitted maximum for the active, not a dose ladder; delivered protection is a formulation property (film, photostability, dispersion). Same policy as the zinc-oxide and titanium-dioxide rows.
Vehicle
For a UV filter the vehicle IS the product: film-forming, photostability (avobenzone needs a stabiliser), and rub-off resistance decide delivered SPF/UVA-PF. This is why the regulatory claim is gated on a performance test of the finished product, and why the engine's G4 gate reads the INCI for adequate UVA filtration rather than trusting the hero.
Checked
against its sources on Sep 1, 2026; next review Sep 1, 2027
Sources
Octocrylene Established · regulator recognises it, trials Moderate Limited
Standing document
21 CFR § 352.10 Sunscreen active ingredients (US FDA — 21 CFR Part 352 (Sunscreen Drug Products for Over-the-Counter Human Use), via Cornell Legal Information Institute; part heading marked [STAYED INDEFINITELY], effective date stayed at 66 FR 67485), opened Sep 1, 2026
Counted on the row
1 randomised trial, 903 participants pooled; findings insufficient; no guideline endorsement recorded; best source tier 1.
Studied dose
up to 10%, leave-on
Dose note
CONCENTRATION IS THE WRONG UNIT FOR THIS CLAIM. FDA attaches the early-skin-aging claim to the finished product's TESTED Broad Spectrum + SPF>=15 performance (21 CFR 201.327), not to any filter percentage. The % here is the permitted maximum for the active, not a dose ladder; delivered protection is a formulation property (film, photostability, dispersion). Same policy as the zinc-oxide and titanium-dioxide rows.
Vehicle
For a UV filter the vehicle IS the product: film-forming, photostability (avobenzone needs a stabiliser), and rub-off resistance decide delivered SPF/UVA-PF. This is why the regulatory claim is gated on a performance test of the finished product, and why the engine's G4 gate reads the INCI for adequate UVA filtration rather than trusting the hero.
Checked
against its sources on Sep 1, 2026; next review Sep 1, 2027
Sources
Oxybenzone Established · regulator recognises it, trials Moderate Limited
Standing document
21 CFR § 352.10 Sunscreen active ingredients (US FDA — 21 CFR Part 352 (Sunscreen Drug Products for Over-the-Counter Human Use), via Cornell Legal Information Institute; part heading marked [STAYED INDEFINITELY], effective date stayed at 66 FR 67485), opened Sep 1, 2026
Counted on the row
1 randomised trial, 903 participants pooled; findings insufficient; no guideline endorsement recorded; best source tier 1.
Studied dose
up to 6%, leave-on
Dose note
CONCENTRATION IS THE WRONG UNIT FOR THIS CLAIM. FDA attaches the early-skin-aging claim to the finished product's TESTED Broad Spectrum + SPF>=15 performance (21 CFR 201.327), not to any filter percentage. The % here is the permitted maximum for the active, not a dose ladder; delivered protection is a formulation property (film, photostability, dispersion). Same policy as the zinc-oxide and titanium-dioxide rows.
Vehicle
For a UV filter the vehicle IS the product: film-forming, photostability (avobenzone needs a stabiliser), and rub-off resistance decide delivered SPF/UVA-PF. This is why the regulatory claim is gated on a performance test of the finished product, and why the engine's G4 gate reads the INCI for adequate UVA filtration rather than trusting the hero.
Checked
against its sources on Sep 1, 2026; next review Sep 1, 2027
Sources
Retinoids Established · regulator recognises it, trials Moderate Promising
Standing document
Refissa [Tretinoin Cream, USP (Emollient) 0.05%] (US Food and Drug Administration prescribing information, via DailyMed (National Library of Medicine)), opened Jul 27, 2026
Counted on the row
8 randomised trials, 1,361 participants pooled; a systematic review or meta-analysis; independently replicated; findings consistent; no guideline endorsement recorded; best source tier 1.
Studied dose
0.02–0.1%, leave-on, judge after ~12 wk
Dose note
FAMILY ROW. Photoaging is the strongest retinoid-class claim, but the pooled evidence is overwhelmingly Rx tretinoin; OTC members are weaker, hence the downgrade from the class's Rx-anchored ceiling.
Vehicle
Measured fine/coarse wrinkle improvement (the photoaging endpoint). The meta-analysis pools Rx tretinoin trials; extending the full effect to OTC retinol is unsupported, so the class row is downgraded one tier for intervention indirectness.
Checked
against its sources on Jul 27, 2026; next review Jul 17, 2027
Retinol Thin · trials Moderate Limited
Counted on the row
3 randomised trials, 44 participants pooled; findings consistent; no guideline endorsement recorded; best source tier 1.
Studied dose
0.1–0.5%, leave-on, judge after ~12 wk
Dose note
Vehicle and stabilization matter for retinol bioavailability; the verified trial used 0.5% retinol cream over 12 weeks.
Vehicle
The one tier-1 trial is ACTIVE-controlled (retinol 0.5% vs bakuchiol 0.5%) with NO vehicle/placebo arm — both arms improved, so it demonstrates retinol reduces photodamage markers but cannot quantify the effect versus no treatment.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
SPF 15 UV Filter System Established · regulator recognises it, trials Moderate Limited
Counted on the row
1 randomised trial, 903 participants pooled; findings insufficient; no guideline endorsement recorded; best source tier 1.
Studied dose
10%, leave-on
Dose note
CONCENTRATION IS THE WRONG UNIT FOR THIS CLAIM. FDA attaches the early-skin-aging claim to the finished product's TESTED Broad Spectrum + SPF>=15 performance (21 CFR 201.327), not to any filter percentage. The % here is the permitted maximum for the active, not a dose ladder; delivered protection is a formulation property (film, photostability, dispersion). Same policy as the zinc-oxide and titanium-dioxide rows.
Vehicle
For a UV filter the vehicle IS the product: film-forming, photostability (avobenzone needs a stabiliser), and rub-off resistance decide delivered SPF/UVA-PF. This is why the regulatory claim is gated on a performance test of the finished product, and why the engine's G4 gate reads the INCI for adequate UVA filtration rather than trusting the hero.
Checked
against its sources on Sep 1, 2026; next review Sep 1, 2027
Sources
Titanium Dioxide Established · regulator recognises it, trials Moderate Limited
Standing document
§ 352.10 Sunscreen active ingredients. (U.S. Food and Drug Administration — 21 CFR Part 352 (Sunscreen Drug Products for Over-the-Counter Human Use), Code of Federal Regulations via GovInfo), opened Jul 27, 2026
Counted on the row
1 randomised trial, 903 participants pooled; findings insufficient; guideline endorsement conditional; best source tier 1.
Studied dose
up to 25%, leave-on
Dose note
CONCENTRATION IS THE WRONG UNIT FOR THIS CLAIM. FDA attaches the early-skin-aging claim to the finished product's TESTED Broad Spectrum + SPF>=15 performance, not to any filter percentage. A 20% TiO2 product that fails the broad-spectrum test carries no photodamage claim; a 6% TiO2 product that passes at SPF 30 does. Reading TiO2 % as a dose ladder for photodamage is a category error.
Vehicle
Vehicle determines whether the claim exists at all: TiO2's UV attenuation depends on particle size, dispersion quality and film-forming behaviour, which are formulation properties, not ingredient properties. This is why the regulatory claim is gated on a performance TEST of the finished product rather than on the INCI list.
Checked
against its sources on Jul 27, 2026; next review Jul 17, 2027
Sources
Tretinoin 0.025% Established · regulator recognises it, trials Moderate Well-established
Standing document
Refissa [Tretinoin Cream, USP (Emollient) 0.05%] (US Food and Drug Administration prescribing information, via DailyMed (National Library of Medicine)), opened Jul 27, 2026
Counted on the row
8 randomised trials, 1,361 participants pooled; a systematic review or meta-analysis; independently replicated; findings consistent; no guideline endorsement recorded; best source tier 1.
Studied dose
0.02–0.1%, leave-on, judge after ~12 wk
Dose note
Prescription-only. FDA-approved for fine wrinkles/mottled hyperpigmentation/tactile roughness of photoaged facial skin (Renova) as part of a comprehensive skin-care and sun-avoidance program.
Vehicle
Endpoint-matching evidence is measured wrinkle/photodamage improvement (the sun-damage/photoaging endpoint). This is the wrinkle-and-photodamage claim, distinct from the discrete-lentigines (dark-spots) claim, which is graded separately.
Checked
against its sources on Jul 27, 2026; next review Jul 17, 2027
Hybrid UV Filter System Established · regulator recognises it, trials Moderate Limited
Standing document
21 CFR § 352.10 Sunscreen active ingredients (US FDA — 21 CFR Part 352 (Sunscreen Drug Products for Over-the-Counter Human Use), via Cornell Legal Information Institute; part heading marked [STAYED INDEFINITELY], effective date stayed at 66 FR 67485), opened Sep 1, 2026
Counted on the row
1 randomised trial, 903 participants pooled; findings insufficient; no guideline endorsement recorded; best source tier 1.
Dose note
CONCENTRATION IS THE WRONG UNIT FOR THIS CLAIM. FDA attaches the early-skin-aging claim to the finished product's TESTED Broad Spectrum + SPF>=15 performance (21 CFR 201.327), not to any filter percentage. The % here is the permitted maximum for the active, not a dose ladder; delivered protection is a formulation property (film, photostability, dispersion). Same policy as the zinc-oxide and titanium-dioxide rows.
Vehicle
For a UV filter the vehicle IS the product: film-forming, photostability (avobenzone needs a stabiliser), and rub-off resistance decide delivered SPF/UVA-PF. This is why the regulatory claim is gated on a performance test of the finished product, and why the engine's G4 gate reads the INCI for adequate UVA filtration rather than trusting the hero.
Checked
against its sources on Sep 1, 2026; next review Sep 1, 2027
Sources
Diethylamino Hydroxybenzoyl Hexyl Benzoate Established · regulator recognises it, trials Moderate Limited
Counted on the row
1 randomised trial, 903 participants pooled; findings insufficient; no guideline endorsement recorded; best source tier 1.
Studied dose
10%, leave-on
Dose note
CONCENTRATION IS THE WRONG UNIT FOR THIS CLAIM. FDA attaches the early-skin-aging claim to the finished product's TESTED Broad Spectrum + SPF>=15 performance (21 CFR 201.327), not to any filter percentage. The % here is the permitted maximum for the active, not a dose ladder; delivered protection is a formulation property (film, photostability, dispersion). Same policy as the zinc-oxide and titanium-dioxide rows.
Vehicle
For a UV filter the vehicle IS the product: film-forming, photostability (avobenzone needs a stabiliser), and rub-off resistance decide delivered SPF/UVA-PF. This is why the regulatory claim is gated on a performance test of the finished product, and why the engine's G4 gate reads the INCI for adequate UVA filtration rather than trusting the hero.
Checked
against its sources on Sep 1, 2026; next review Sep 1, 2027
Sources
Ascorbic Acid Thin · trials Moderate Limited
Counted on the row
1 randomised trial, 19 participants pooled; a systematic review or meta-analysis; findings insufficient; no guideline endorsement recorded; best source tier 2.
Dose note
No dose can be attached to the only endpoint-matching RCT. The only concentration-response data (de Dormael: 10% strong > 7% moderate > 2% weak) comes from an ACUTE UV-CHALLENGE TANNING MODEL on back skin, and the sponsor itself notes 10% 'was assessed only once' and 'must be taken with caution'. There is no dose-response for clinical facial photodamage.
Vehicle
Vehicle is decisive and unreported in the one trial that matters. Pinnell 2001 (pig skin, PHARMACOKINETICS not a clinical dose) established L-AA needs pH<3.5 and absorption peaks ~20%; L-AA also oxidizes readily, so a labeled percentage does not describe what reaches skin. The four trials use four non-interchangeable vehicles: an undisclosed proprietary serum (Traikovich), anhydrous polysilicone gel (Fitzpatrick), SMA copolymer aqueous (de Dormael), ferulic-stabilized aqueous (Murray). None reports stability or assay data confirming delivered concentration. A shelf product at '15% vitamin C' cannot be assumed to reproduce any of these results, and results cannot be pooled across vehicles even in principle.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Sources
Tocopheryl Acetate Thin · trials Limited Limited
Counted on the row
0 randomised trials; findings insufficient; no guideline endorsement recorded; best source tier 3.
Dose note
Vitamin E is NOT a UV filter and carries no SPF. It is sometimes added to sunscreens as a non-filtering adjunct antioxidant, which is a different and much weaker proposition than filtration — see the Photodermatol review on exactly this question.
Vehicle
Same ester caveat as the aging row: the cosmetic form is usually tocopheryl acetate, while the antioxidant mechanism belongs to free α-tocopherol. Additionally, α-tocopherol is itself photolabile — it is consumed as it quenches radicals — so any photoprotective contribution is inherently self-limiting over a sun exposure, unlike a filter that keeps attenuating.
Checked
against its sources on Jul 17, 2026; next review Jul 17, 2027
Zinc Oxide Established · regulator recognises it, mechanism established, trials Moderate Limited
Standing document
§ 201.327 Over-the-counter sunscreen drug products; required labeling based on effectiveness testing (US FDA — 21 CFR 201.327 (via Cornell Legal Information Institute)), opened Jul 27, 2026
Counted on the row
1 randomised trial, 903 participants pooled; findings insufficient; no guideline endorsement recorded; best source tier 1.
Studied dose
10–25%, leave-on
Dose note
Concentration is the WRONG variable for a UV filter. What matters is the finished product's measured SPF and UVA-PF (ISO 24444 / 24443) — a well-formulated 12% zinc oxide can outperform a poorly-formulated 20%. Do not read the % as a dose ladder.
Vehicle
For a filter, the vehicle IS the product: film-forming behaviour, particle dispersion and rub-off resistance determine delivered protection far more than the nominal %. This is precisely why the physicochemical uv-protection claim is graded on a standardized test of the FINISHED PRODUCT rather than on the ingredient.
Checked
against its sources on Jul 27, 2026; next review Jul 17, 2027
Sources