Q.01

How is dermal absorption assessed for cosmetic ingredients?

In the EU, dermal absorption of a cosmetic ingredient is normally measured with an in vitro skin absorption study under OECD TG 428, run to the SCCS "Basic Criteria" (SCCS/1358/10). The SCCS says the combination of the two "is considered to be essential" for cosmetic ingredients [1][2]. The result feeds the systemic exposure dose (SED) and the margin of safety (MoS). If no adequate study exists, the SCCS uses a default of 50% [3].

Approach: I read the dermal absorption, SED and low-absorption sections of the SCCS Notes of Guidance, 12th revision (SCCS/1647/22) [4], Annex I of the consolidated EU Cosmetics Regulation 1223/2009, and the GB SAG-CS opinions. The GB opinions cite a 2023 version of the Notes of Guidance, which I did not have, so details may have changed since the 12th revision.

Legal basis

  • Annex I, Part A, section 8 of Regulation 1223/2009 says "All significant toxicological routes of absorption shall be considered as well as the systemic effects and margin of safety (MoS)…", and that leaving these out "shall be duly justified" [5].
  • The skin is the main route of exposure. The stratum corneum is treated as the layer that limits the rate of absorption [1].

Study methods

  • In vitro: OECD TG 428. In vivo: OECD TG 427, only for studies done before the EU animal-testing deadlines. There is no OECD guideline for in vivo human dermal absorption studies [1].
  • In vitro work should follow the OECD guidance on good in vitro method practice (GIVIMP, OECD GD 286) [6].

Main SCCS Basic Criteria [2][3]

  • Skin: human skin from a suitable body site is the gold standard, with pig skin as the alternative. Skin integrity "should be verified" and kept at normal human skin temperature.
  • Receptor fluid: at physiological pH, with the test substance shown to be soluble and stable in it.
  • Test material: it should match the substance used in the finished product.
  • Dose and vehicle: these should reflect in-use conditions, including contact time, and should include the highest concentration found in a typical formulation.
    • Normal application: 1-5 mg/cm² for solids and up to 10 µl/cm² for liquids.
    • Oxidative hair dyes: usually 20 mg/cm² for 30-45 minutes.
  • What to measure: the substance in every compartment:
    • the excess left on the skin surface (dislodgeable dose);
    • the stratum corneum (tape strips);
    • the living epidermis;
    • the dermis;
    • the receptor fluid.
  • Recovery: overall recovery should be 85-115%.
  • Sample size: 8 skin samples from at least 4 donors. Absorption is calculated per diffusion cell and then averaged.
  • Results below LOD/LOQ: use whichever bound, zero or the LOD/LOQ, gives the highest absorption.
  • Laboratory proficiency: check at least twice a year with reference compounds such as caffeine or benzoic acid.
  • Formulation effects: consider penetration enhancers and excipients such as liposomes.

How the result is interpreted

  • What counts as absorbed: the amounts in the dermis, the epidermis (without stratum corneum) and the receptor fluid. For substances with very low absorption, the epidermis may be left out if the data clearly show no movement from the skin into the receptor fluid [3].
  • Studies that meet all the criteria: the mean + 1 SD is used. If there are significant deviations or very high variability, the mean + 2 SD may be used. If deviation is too high, the study is rejected and the 50% default applies [3][6].
  • Calculating SD: use N; N-1 is allowed only above 30 data points [6].
  • Damaged skin: there is no standard protocol. The SCCS prefers intact skin, and damaged-skin data count only as supporting information [6].
  • Nanomaterials: the study must show whether the absorbed material was in particle form or dissolved [3].

Using the result in the SED [7][8]

  • Preferred form, an amount (µg/cm²): SED = DAa × 10⁻³ × skin surface area × daily application frequency / 60 kg body weight.
  • Alternative form, a percentage: SED = E_product × C/100 × DAp/100.
    • A percentage is only valid if the test doses mimic, without exceeding, real use.
    • In vitro tests below 2 mg/cm² are not technically feasible, while real use is usually 1 mg/cm² or less. A percentage from such a test can therefore underestimate systemic exposure [3].

Substances with very low absorption

  • Properties that may point to very low absorption [6]:
    • molecular weight above 500 Da;
    • a high degree of ionisation;
    • log Pow of -1 or below, or 4 or above;
    • topological polar surface area above 120 Ų;
    • melting point above 200 °C.
  • For these substances some studies may be waived. The minimum data set is:
    • measured physicochemical data;
    • local toxicity;
    • mutagenicity/genotoxicity;
    • a high-quality in vitro study that meets the Basic Criteria.
  • In that case the experimental mean, not mean + SD, is used [9].

Supporting methods only

  • In silico skin permeability (Kp) estimates can be used as supporting information, for example the ten Berge and Potts and Guy models, Vermeer Cosmolife and the NIOSH SkinPerm calculator [6][10].
  • PBPK model results are used quantitatively only if enough model detail is provided [7].

Other routes

  • Ingestion from oral-care products: 100% absorption is assumed if there are no data [10].

Great Britain SAG-CS opinions follow the same method:

  • Homosalate: 5.3% was taken from a study meeting OECD TG 428, GLP and the SCCS Basic Criteria, using mean + 1 SD. The panel kept this higher figure over a PBPK-derived 2.48% as more precautionary [11].
  • Daidzein: the panel agreed with the SCCS 50% default, because none of the earlier studies met the Basic Criteria [12].

One difference is body weight. The GB opinions I read use 70 kg [13], while the SCCS SED equations use a 60 kg default [8].

Q.02

What absorption defaults does the SCCS use?

The SCCS uses five default absorption values. They apply when no adequate study data exist, and measured data replace them whenever available:

Route / contextDefaultWhen it applies
Dermal50%No measured dermal absorption, or only inadequate data [3]
Oral (correcting an oral study's point of departure, PoD)50%No oral absorption data. The PoD is divided by 2 to give the systemic PoD (PoDsys) [14]
Oral, poorly bioavailable substances10%"Could be considered" where information suggests poor oral bioavailability [14]
Ingestion from oral-care products (toothpastes, mouthwashes)100%No experimental data provided [10]
Oral to inhalation extrapolationFactor of 2 (oral 50%, inhalation 100%, taken from ECHA 2012a)Extrapolating from oral to inhalation [14]

Approach: I searched the SCCS Notes of Guidance, 12th revision (SCCS/1647/22), for default absorption and bioavailability values. As noted last time, the GB opinions cite a 2023 version that I did not have, so later revisions may differ.

How the defaults are used

  • Dermal 50%: the SCCS calls this a "conservative value". It also applies when the study data are inadequate, so a study that fails the Basic Criteria falls back to 50% [3].
  • Oral 50%: this rests on the view that "not more than 50% of an orally administered dose is systemically available". Wherever oral absorption data exist they "should be used", and other kinetic data should also be considered [14].
  • Keep the two 50% values apart: a footnote says the oral 50% default must be distinguished from the separate dermal 50% default [14].
  • Very low dermal absorbers: these get no lower default. Instead, the experimental mean replaces mean + SD in a study that meets the Basic Criteria [9].

Other default assumptions

  • Nanomaterials: if systemic absorption is seen and absorption of nanoparticles cannot be ruled out, the SCCS "may apply a default approach and assume that 100% of the absorbed material was in nano form" [15].
  • TTC: the original TTC approach, which compared a skin-adjusted internal exposure against oral TTC values, implicitly assumed 100% oral bioavailability. The SCCS says this "in many cases is an overestimation". For proper route-to-route extrapolation, the NOAELs from the Munro database need correcting for oral absorption [16].
  • Inhalation: apart from the 100% used in the oral-to-inhalation factor, the inhalation section gives no standalone default. It explains that absorption depends on physical form, particle size, surface properties and the blood-to-gas partition coefficient [10].
Q.03