The Second Skin

A pharmaceutical transdermal patch is defined, at the regulator's desk, by five variables: the active compound, the vehicle that carries it, the occlusion that seals it, the duration it rests on skin, and the skin itself — site, temperature, hydration. Each variable must be disclosed. Each must be measured. None may drift without re-submission.

Activewear worn during exercise meets every one of those variables. What we will call, for this report, an Unregulated Transdermal Patch (UTP) is the class of garment that meets the physical specification of a drug-delivery system without meeting any of the measurement specification. The operational definition is narrow: five or more skin-penetration enhancers (heat, moisture, friction, occlusion, stretch) acting on sub-500-Dalton bioactive fluorochemistry, with none of the enhancer variables disclosed to the wearer and none of the chemistry identified on the label.

The class excludes realistic cases. Three worked non-examples show where the specification stops.

The class is the intersection: performance-spec garment, in motion, against skin.

This is the body report in a three-part series on PFAS in performance activewear. Report 072 takes up the regulatory architecture. Report 073 takes up the money. The material question comes first because the body is where the other two resolve.

The specification nobody wrote

A 25 cm² drug patch releases a defined active through defined skin under defined conditions for a defined interval. A pair of 7/8 leggings covers roughly 5,500 cm² — two hundred times the patch area — and is worn for the duration of a workout plus whatever follows before the wearer changes. The active ingredient, in the drug-patch case, is selected against a published rule: Bos and Meinardi's 500-Dalton threshold, which states that "virtually all common contact allergens… all known topical drugs used in transdermal drug-delivery systems are under 500 Dalton."¹ Molecules above 500 Da do not reliably cross intact stratum corneum. Molecules below it can.

The rule is not a heuristic; it is the gate pharmaceutical formulators design against. Short-chain PFAS — perfluorobutanoic acid (PFBA, 214 Da), perfluoropentanoic acid (PFPeA, 264 Da), perfluorobutane sulfonate (PFBS, 300 Da), 6:2 fluorotelomer alcohol (FTOH, 364 Da) — sit well inside the window pharmaceutical chemistry treats as absorbable. Perfluorooctanoic acid (PFOA, 414 Da) also sits inside it. Perfluorooctane sulfonate (PFOS, 500 Da) sits on the line.

The rule identifies which molecules can cross intact stratum corneum under idealised conditions. Whether specific PFAS congeners do cross is a separate empirical question, and it has now been directly measured. Ragnarsdóttir, Abdallah and Harrad (2024), applying fifteen PFAS to 3D human skin equivalent membranes with a methanol vehicle at 500 ng/cm² for 24 to 36 hours, found that 58.9% of applied PFPeA and 48.7% of applied PFBS permeated through the membrane.⁶ The rule said these molecules could cross; the measurement showed they did. Because the Bos–Meinardi rule was originally derived for neutral and weakly ionised compounds, and PFAS are charged amphiphiles, the rule is necessary but not sufficient on its own. The Harrad measurement is what converts it from an academic threshold into an empirical finding that applies to this compound class. The two belong in the same sentence, always.

The industry's own chemistry transition, moving from long-chain C8 compounds to shorter C6, C4 and ultrashort-chain replacements, has not moved PFAS away from the skin. It has moved them further into the Bos–Meinardi window. Sun and colleagues documented this trajectory in 2016: as regulators restricted PFOA and PFOS, industry substituted shorter homologues that cleared the regulatory bar on bioaccumulation while clearing the 500-Dalton bar on absorbability more decisively than the compounds they replaced.²

Two facts belong next to each other. The first is that short-chain PFAS are, by molecular weight and by measured permeation, candidates for dermal transit. The second is that nobody designing, specifying, selling or buying activewear has been required to treat them that way.

What skin does under exercise

Stratum corneum is not a static barrier. Its permeability is a function of temperature, hydration, the vehicle in contact with it, the duration of contact, and the integrity of the epidermal surface. Exercise moves each of these variables in the direction that favours absorption.

Vasodilation during exercise raises cutaneous blood flow by five to ten times baseline; maximum skin perfusion reaches six to eight litres per minute under heat stress.³ Sweat arrives from eccrine glands at rates that, on the female central upper back at 60% VO₂max, reach 223 grams per square metre per hour.⁴ A fabric layer pressed against that surface becomes an occlusion — the same mechanism a pharmaceutical patch uses to drive flux. Occlusion of hydrated skin increases permeability for a wide range of compounds; the effect has been characterised in dermatological literature for half a century.⁵

Friction adds a fifth variable. Repeated contact of stretch fabric against skin, at the elastic deformation rates typical of running and strength training, abrades and loads the stratum corneum in ways that laboratory measurements of static skin permeability do not capture. None of this is novel science. It is the same set of conditions that pharmaceutical regulators require to be held constant when a drug delivery patch is evaluated.

What crosses

The dermal absorption question for PFAS was treated, for most of the period during which these compounds entered consumer products, as background. It has now been directly measured.

Ragnarsdóttir, Abdallah and Harrad (2024) applied fifteen PFAS to 3D human skin equivalent membranes using a methanol vehicle at 500 ng/cm² for 24 to 36 hours. Under those conditions, permeation and retention split by chain length.⁶

Across the congener series, permeation correlated negatively with octanol-water partition coefficient (r = –0.97 for perfluorocarboxylates across log K_OW; r = –0.99 for perfluorosulfonates across log K_OW). Retention in skin tissue correlated positively with chain length (r = 0.97).

The methodology matters and must travel with the numbers. This is 3D human skin equivalent, not intact human skin. The vehicle was methanol, not sweat. The exposure was 500 ng/cm², not a measured activewear concentration. The interval was 24 to 36 hours, not the duration of a workout. The study establishes that these molecules cross skin-equivalent membranes under controlled laboratory conditions; it does not establish a flux rate for a specific garment on a specific body during a specific workout. That measurement does not yet exist.

What the study does establish is the direction of the relationship. Short-chain PFAS — exactly the compounds now dominating the global PFAS load in human serum and surface water — are the ones that permeate. Long-chain compounds remain in skin tissue. Shifting the chemistry towards short-chain homologues, which was framed as a safety improvement on the grounds of reduced bioaccumulation, shifts the profile of what reaches blood through skin upward.

The direction is upward. The magnitude, garment-in-use, is unknown.

The zones that overlap

The female central upper back reaches 223 g/m²/h at 60% VO₂max. The anterior and posterior thighs, the lumbar region, the anterior chest — these are also among the highest sweat-flux zones during exercise. They are the zones a compression fit covers. A sports bra covers the anterior chest. High-rise leggings cover the lumbar. The overlap is not incidental; it is the function the compression specification produces.

What this means, for the body report, is that the wear zones of maximum chemical opportunity — highest hydration, highest perfusion, highest occlusion — are also the zones the industry has specified with the most sophisticated finishes: durable water repellent (DWR) on outer surfaces, antimicrobial finishes on inner surfaces, thermoregulating finishes on the body-facing layer. Fluorinated chemistry has historically been the default technology for DWR and for a range of stain- and oil-resistance applications, and short-chain fluorotelomer chemistry has been the post-2015 default. Alternatives exist — silicones, paraffins, dendrimers — but fluorinated chemistry remains present on a significant share of performance activewear currently in circulation.

Mamavation's 2022 investigation, conducted on 18 January 2022, tested 32 pairs of athletic leggings from 27 brands using certified EPA-method total fluorine measurement. Three pairs from Lululemon tested above the detection limit for total fluorine.⁷ Total fluorine is a screening method, not a speciation method; it cannot identify which PFAS are present or at what concentration. It establishes presence, not dose. Lululemon's published Restricted Substances List restricts intentionally-added PFAS at a 50 ppm total organic fluorine threshold, and its 2022 Chemistry Management Approach document describes the PFAS phase-out as ongoing.⁸

The polymer exemption and what it cannot explain

Industry responses to PFAS-in-textile findings rest on a specific argument: that the fluorinated chemistry used in performance fabrics is bound into polymer networks, is not bioavailable, and is not the same class of compound as the short-chain PFAS detected in serum.

The argument to examine is not a single experimental claim. It is a regulatory rule. The polymer exemption, as written into the European and US regulatory frameworks, does not distinguish between substrates. It is a class-level carve-out: fluorinated chemistry incorporated into a polymer matrix is treated as non-bioavailable, full stop. That is the rule on which the defence rests.

Schellenberger and colleagues (2022) weathered commercial outerwear containing a C8 side-chain fluoropolymer on a Sydney rooftop for six months. Extractable PFOA in the fabric rose from 12 ng/g to 513 ng/g over the exposure period — a 42.75-fold increase.⁹ Van der Veen and colleagues (2022) subjected a C6 side-chain fluoropolymer on polyamide to ISO 6330:2012 domestic-wash protocol. Extractable 6:2 fluorotelomer alcohol rose from 87 μg/kg (aged fabric baseline, unwashed) to 430 μg/kg after five ISO 6330:2012 domestic-wash cycles on aged fabric at 40 °C — a 4.9-fold increase in the extractable fraction of a PFAS degradation product from the same fibre, measured across a realistic domestic wash protocol (10-cycle samples showed a further rise to 520 μg/kg).¹⁰

Both studies tested polyamide outerwear substrates, not performance knit activewear. That distinction matters for dose quantification. It does not rescue the regulatory rule. The rule claims that polymer-bound fluorinated chemistry is non-bioavailable as a class. The primary literature has now falsified the "stays bound" premise on two substrates under two realistic use conditions. Once the class-level claim fails on any substrate where it has been tested, the burden of evidence for continued application of the exemption to untested substrates — including performance knit — lies with the party asserting the exemption, not with the party questioning it. The exemption was not granted on the basis of activewear-specific evidence to begin with. It cannot be rescued by the absence of activewear-specific evidence against it.

What would restore the polymer-exemption defence for activewear is affirmative substrate-specific measurement showing that performance knit substrates behave differently from tested outerwear substrates under wash and wear conditions. No such study exists. Until it does, the defence operates on an evidentiary base the primary literature has already unsettled.

Three methods, three pictures

The disagreement between industry chemists and independent researchers on "how much PFAS is in this garment" is not, in the main, a disagreement about measurement quality. It is a disagreement about which measurement.

Three methods are in current use. Total fluorine (TF) measures all fluorine in a sample, organic and inorganic, targeted and untargeted. Extractable organofluorine (EOF) measures all organic fluorine that can be solvent-extracted from the sample. Targeted-list analysis measures the specific PFAS congeners on a pre-specified list — typically the 40 to 60 compounds for which analytical standards exist.

On the same garment, the three methods can produce numbers that differ by orders of magnitude. Total fluorine is always highest. Targeted analysis is always lowest. Extractable organofluorine sits between them. The ratio of targeted to total routinely ranges from less than one percent to around thirty percent. That is the spread a brand's "below restricted limits" declaration travels through — which method, at what limit of detection, against what threshold, determines what "below" means on the same garment.

No currently public test has reported all three methods applied to the same activewear garment side by side. The methods exist. The laboratory standards exist. The absence is not technical; it is disclosure. A wearer cannot point to a published example of the TF–EOF–targeted spread on the specific leggings in their drawer because no brand has voluntarily published one, and no regulator has required one. The measurement gap is the disclosure gap.

Zheng, Eick and Salamova (2023), sampling serum for total PFAS burden across the US population, found that ultrashort- and short-chain PFAAs account for 69% to 100% of the total PFAA load depending on the congener definition used.¹¹ That finding establishes class congruence between what is reaching human blood and the chemistry currently dominating consumer-product substitutions. It does not establish pathway attribution. The study cannot say how much of the short-chain burden came from textiles, how much from cookware, how much from water, how much from food packaging. It says that the compounds reaching blood are, as a class, the compounds that clear the Bos–Meinardi threshold most decisively.

Washing

The wash cycle is the activewear amplifier. Van der Veen's 4.9-fold increase in extractable 6:2 FTOH across the first five wash cycles on aged polyamide applies to the substrate tested; activewear substrates are not identical but share the key mechanical feature — a fluorinated side-chain polymer on a fibre surface subjected to mechanical agitation, heat and surfactants. What the wash cycle does to the garment, it also does to the water leaving the machine: a fraction of what is released goes into the wearer's next wear; a fraction goes into wastewater and, eventually, surface water. The environmental half of the accounting is Report 073's territory. The body half is this: the garment does not arrive at the wearer in a static state. Its fluorinated load shifts — downward in the garment, upward in the wash effluent, ambiguous at the skin — with each cycle.

Harm

The regulatory threshold for PFAS intake from any source, set by the European Food Safety Authority's CONTAM Panel in 2020, is a tolerable weekly intake of 4.4 nanograms per kilogram of body weight per week for the sum of PFOA, PFNA, PFHxS and PFOS, treated as equipotent.¹²

The endpoint that drove the tolerable weekly intake downward was not cancer.

It was a reduction in vaccine-antibody response in one-year-old children whose mothers carried higher serum PFAS concentrations through pregnancy. Children who had received routine tetanus and diphtheria vaccinations produced less of the antibodies those vaccinations are designed to elicit. The immune response did not fail; it was attenuated — measurable in the blood, traceable to a prenatal and early-life exposure the children did not choose. Grandjean and colleagues (2012), working in a Faroe Islands birth cohort of 656 singleton children enrolled and 587 completing follow-up through 2008, reported that elevated PFC exposures were associated with reduced humoral immune response to routine tetanus and diphtheria immunisations in children at ages five and seven, with serum PFC concentrations in the child at age five showing uniformly negative associations with antibody levels.¹³ That finding is the one the European regulator used to set the current threshold. The threshold is a nanogram-per-kilogram-per-week figure because that is where the paediatric antibody attenuation became statistically visible.

The C8 Science Panel — established under a 2005 class-action settlement, given full access to the community surrounding a DuPont plant in the Ohio River Valley, and funded to run independently — concluded in 2012 that there is a "probable link" between PFOA and six conditions: kidney cancer, testicular cancer, thyroid disease, ulcerative colitis, pregnancy-induced hypertension, and diagnosed high cholesterol.¹⁴ The finding rests on three peer-reviewed papers by Barry, Darrow and Steenland from that panel's work.

The US Agency for Toxic Substances and Disease Registry's 2021 Toxicological Profile for PFAS acknowledges that dermal contact is a contributing route of exposure to perfluoroalkyl compounds, alongside oral and inhalation routes.¹⁵ The acknowledgement is formal; it is not quantitative. No regulator has set a dermal exposure threshold for PFAS.

What the absence of measurement means

There is no published activewear-substrate study of PFAS release rate under exercise conditions. There is no regulatory threshold for dermal PFAS exposure. There is no requirement for a performance-fabric brand to disclose which PFAS congeners are present, at what concentration, measured by which method, at what limit of detection. There is no requirement to label a garment worn against skin during exercise differently than a garment worn as outerwear over clothes.

The class — the Unregulated Transdermal Patch — exists because the measurement does not.

The second and third reports in this series take up the question of why. The regulatory architecture that permits this gap is the subject of report 072. The economics of delay — including the 1,547-day gap, in the best-documented case, between the first regulatory signal on a specific PFAS compound and the market-wide consumer substitution away from it — is the subject of report 073. The body report comes first because the body is where the other two resolve.

What the counter-position looks like

Four defences are available to the industry, and each can be stated in its strongest form.

The first is that the polymer-bound argument still holds for performance knits even though the primary literature tests outerwear substrates. The response is that polymer-bound chemistry has been shown to leak on the substrates it has been tested on; extension to untested substrates cuts both ways, and the burden of evidence for the exemption lies with the party claiming it.

The second is that the Harrad dermal-permeation numbers used a methanol vehicle and 3D skin equivalent, and cannot be extrapolated to sweat-on-intact-skin. The response is that this is correct, and that the study is the best available evidence on the direction of the chain-length–permeation relationship. Methanol is not sweat; 3D skin equivalent is not intact human skin; 500 ng/cm² is not a measured garment concentration. The study does not close the dose question. It opens the direction question.

The third is that dermal contribution, even if real, is a small fraction of total PFAS intake compared with water and food. The response is that ATSDR acknowledges dermal contribution without quantifying it, that no regulator has quantified it, and that "small fraction of total" is not an answer to "can this specific exposure route be specified and measured?"

The fourth is that responsible brands have already committed to phasing PFAS out. The response is in the tenses — committed to, plans to, ongoing — and in the absence of method-and-limit-of-detection-specific thresholds against which an independent observer could verify the commitment.

None of these defences is dishonest. Each is the strongest version of a position the current evidence base permits. What none of them does is close the measurement gap.

What we can say

We can say the chemistry is in the Bos–Meinardi window. We can say the substrate and the body zone meet every physical specification a pharmaceutical transdermal patch must meet. We can say short-chain PFAS permeate 3D human skin equivalent membranes at 58.9% for PFPeA and 48.7% for PFBS under controlled laboratory conditions. We can say the polymer exemption has been shown to leak on two outerwear substrates under weathering and laundering. We can say the regulatory harm floor for the same class of compound is measured in nanograms per kilogram per week. We can say ATSDR has formally acknowledged that dermal contact may contribute.

We cannot say what the flux rate is for a specific garment on a specific body during a specific workout. We cannot say what proportion of a given wearer's serum burden is attributable to textile dermal exposure. We cannot say whether the flux is closer to "detectable but trivial" or closer to "comparable to dietary intake". Those measurements do not exist because no regulatory body has required them and no industry body has funded them.

What we can say, and what matters, is that the direction of every known variable points the same way. Upward. Into the body. Through the widest vasodilated capillary surface it has. The magnitude is the unmeasured quantity. The sign is not.

The Levers

Tier 1 — Specification.

• Disclosure at method and limit-of-detection level. For any garment marketed as PFAS-free or PFAS-below-threshold, a brand should state the method used (TF, EOF, or targeted), the limit of detection, and the specific threshold. A threshold without a method is not a threshold. Ask for the three numbers together.

• Substrate-specific measurement. The polymer-exemption literature tests polyamide outerwear. Performance knit activewear has not been equivalently tested. Independent measurement of fluorine release rate from activewear substrates under wash and wear conditions is the single highest-yield empirical question in this area.

• Dermal-route inclusion in regulatory thresholds. The EFSA TWI and the ATSDR MRLs do not currently allocate a component to dermal intake from textiles. A dermal allocation — even a provisional one set at a precautionary level — would convert the current measurement gap from an absence into a specification.

Tier 2 — Body.

• The five-variable audit. Before buying a garment marketed for exercise use, treat the five pharmaceutical skin-penetration enhancers as a checklist: Is this garment heat-retentive? Is it occlusive? Does it stay in contact with sweat? Does it cover high-sweat-flux zones? Is it worn for multi-hour durations? A yes to four or five places the garment in the UTP class.

• Change out. Where a garment meets the UTP specification and is worn for exercise, the single mechanical intervention that reduces dose is duration. Changing out of the garment promptly after exercise removes the body from the occlusion before the hydration state of the stratum corneum relaxes to baseline. This is not a solution. It is a dose-reduction measure available to a wearer who has no disclosure to act on.

What Would Change This Analysis

The analysis rests on three claims. Each has a specification under which it would be falsified.

Claim 1: Short-chain PFAS sit within the 500-Dalton window that pharmaceutical chemistry treats as absorbable through intact skin. Would change if: A revision of the Bos–Meinardi framework — or direct measurement on intact human skin — established that fluorinated short-chain compounds behave differently from the allergens and pharmaceuticals the rule was derived from.

Claim 2: The polymer exemption is not a permanent sink. Would change if: Substrate-specific studies on performance knit fabrics, under wash and wear conditions, showed zero or negligible release across realistic use lifetimes — and if independent replication confirmed the finding.

Claim 3: The dermal pathway is credible enough to warrant measurement, even though its magnitude is unquantified. Would change if: A direct in-vivo study — activewear garment, exercising participants, serum or urine PFAS measurement pre- and post-exposure — established that the contribution to body burden from this pathway is at least one order of magnitude below the contribution from drinking water, food and inhalation combined. Such a study does not exist.

The report will be revised in either direction when any of the three falsification conditions is met.