'Hypoallergenic' Has Had No Legal Meaning in the US Since 1977. Four of Our Own 'Allergen-Free' Products Contain a Baseline Allergen.
On 21 December 1977 the US Court of Appeals for the District of Columbia Circuit struck down the only serious attempt American regulators have made to define what "hypoallergenic" means on a cosmetic. The regulation had been on the books for two and a half years. It has never been replaced.
Which is why the word costs nothing to print. Not "means less than people think" — means nothing, legally, in the United States, and has meant nothing since a Tuesday in the Carter administration.
We do not use the word. We checked, because it would be embarrassing to write this and then find it on one of our own boxes. What we found instead was five products carrying a claim in the same family, and four of them contain something a dermatology clinic would patch-test you against.
The FDA proposed a rule in 1975 requiring anyone printing "hypoallergenic" to prove by comparative testing that their product caused fewer reactions than competitors. Almay sued. The DC Circuit vacated the rule as arbitrary and capricious, and 21 CFR Part 700 today has no section on the word at all — it runs 700.3, then jumps to bithionol, mercury, vinyl chloride. What has filled the gap is data showing the label does not predict the contents: 73.6% of 208 UK products in a 2024 audit contained at least one allergen or related chemical from the British baseline series, and 87.7% of 212 infant cosmetics in a 2019 Polish study contained at least one hapten. Our own catalogue is not clean either. Five products say some version of allergen-free; four of them list an ingredient on the same baseline series.
What the 1975 rule would have required
The proposed regulation was 21 CFR 700.100, published on 6 June 1975. It did something quite clever and quite doomed. Rather than define hypoallergenic by an absolute standard — no allergens present, or no reactions in a panel — the FDA defined it comparatively. To use the word, you had to run scientific studies showing your product produced fewer adverse reactions than competing products holding at least 10% of the market. And you had to print a specific sentence on the label: less likely to cause adverse reactions than some competing products.
Almay challenged it. The court agreed, and the reasoning is worth reading because it is not a technicality about procedure.
The judges went through the administrative record looking for the evidence behind the comparative definition and found three things: a dictionary that did not employ the chosen definition, an AMA report that recommended against using the word "hypoallergenic" at all, and an FTC consumer survey whose unreliability, the court wrote, is apparent on its face. Arbitrary and capricious. Regulation vacated.
The AMA detail is the one that sticks. The medical body the agency cited as support had said the honest answer was to abolish the term. The FDA instead tried to define it, on a record too thin to hold, and lost the word entirely. Nobody has picked it back up in forty-nine years.
What is in Part 700 now
We pulled the current text. Part 700 opens with § 700.3, Definitions, and Subpart B lists the specific things the agency has actually banned or restricted: bithionol, mercury compounds, vinyl chloride, certain halogenated salicylanilides, zirconium in aerosols, chloroform, methylene chloride, chlorofluorocarbon propellants, tamper-resistant packaging, prohibited cattle materials, sunscreen ingredients.
There is no § 700.100. There is no section on hypoallergenic, sensitive-skin, dermatologist-tested or allergy-tested claims anywhere in the part.
We wanted to tell you what Korea does about this, since most of what we sell is Korean and the MFDS does regulate substantiation for some claims. We could not open the relevant provision from this desk tonight — the Korean legislation portal we normally use would not return the text, and we are not going to summarise a statute we have not read. It is on the list for a follow-up piece.
What the audits keep finding
Siwaporn Hiranput and colleagues at Wrexham Maelor Hospital took 208 personal care products off Boots.com and Superdrug.com in 2022 and read every label against the British Society for Cutaneous Allergy baseline series. Their results ran in Clinical and Experimental Dermatology in April 2024.
153 products, 73.6%, contained at least one allergen or related chemical from that series. Twenty-one contained two. Four contained three. In leave-on products the most common single allergen was cetearyl alcohol. In rinse-off products it was parabens. Fragrance, counted as a related chemical, showed up in 85 products — 40.1%.
The paediatric numbers are worse. Karolina Dumycz, Katarzyna Kunkiel and Wojciech Feleszko analysed 212 cosmetics marketed for children under twelve months and published in Clinical and Translational Allergy in 2019. They found 41 distinct sensitisers, an average of 2.51 per product, and 87.7% of products containing at least one. The most common were cocamidopropyl betaine, tocopherol, propylene glycol, fragrances and lanolin. The majority of these products were labelled hypoallergenic, dermatologically tested or safe, and 85% of those contained haptens.
None of this means the products are dangerous. The point is narrower than that, and harder to escape: the word on the front does not predict the list on the back. An Goossens, reviewing 14,911 patients patch-tested in Leuven between 1990 and 2014 in the journal Cosmetics, found cosmetic allergy rising to about 25% of her tested population by the last five-year block, led by fragrance components and preservatives — methylisothiazolinone above all, which she describes as a worldwide epidemic. A patch-test population is by construction made of people whose skin has already gone wrong, so the base rate out in the street is a long way below 25%.
Five of ours say it, in four different ways
Our free_of field is where our suppliers' claims land, and searching it for anything allergy-shaped returns five products out of fifty-six.
| Product | Claim as printed | What is on the list |
|---|---|---|
| seapuri body mist | Parfum-free · Allergen-free | Tea tree leaf oil, row 32 — we grade it mid |
| foellie inner perfume | Declared-allergen-free (per Hwahae) | Fragrance (parfum blend), row 4 — we grade it mid |
| somi eye cream | Fragrance-free · Allergen-free | Cetearyl alcohol, row 4; tocopherol, row 26 |
| celleanser barrier cream | Fragrance-free · Allergen-free | Cetearyl alcohol, row 45 |
| towbtowb collagen mask | Fragrance-free · Allergen-free | Nothing on the series |
Take the five one at a time, because the four that fail fail differently, and the fifth does not fail at all.
The seapuri mist is the worst of it. Parfum-free and allergen-free on the front; tea tree leaf oil at position 32, plus lavender and clary sage extracts, all three of which we ourselves grade mid. The EU's own scientific committee has called tea tree oil a moderate sensitiser, which is why we flagged it in the ingredients we flag most. A botanical essential oil is not an absence of fragrance. It is fragrance with a different sourcing story.
The foellie claim is more interesting, and arguably not a lie. Declared-allergen-free, sourced to Hwahae, almost certainly means none of the 26 fragrance allergens the EU requires to be declared by name. The product still contains a parfum blend at row 4. So the claim is true and useless simultaneously: it tells you which allergens are absent from a list of 26, and says nothing about the rest of the blend. It is a perfume. We sell it as a perfume.
Then somi and celleanser, which both say allergen-free and both contain cetearyl alcohol — the single most frequent leave-on allergen in the Hiranput audit. We grade it low, and we would defend that grade: it is a fatty alcohol, it is in eleven of our products, and the reaction rate in the general population is small. But low is our judgment about frequency, not a claim that the ingredient is absent from any baseline series. The label says the second thing.
The towbtowb mask survives the check clean.
What our screening found
We ran the Hiranput method against ourselves. Across 1,920 ingredient rows in 56 products, 35 products — 62.5% — contain at least one ingredient from a baseline-series list we assembled from the published papers: named fragrance allergens, fragrance itself, cetearyl alcohol, parabens, lanolin, isothiazolinones, formaldehyde releasers, propylene glycol, tocopherol, cocamidopropyl betaine, tea tree oil. That is 82 rows. The most common by a distance is tocopherol at 17 products, then cetearyl alcohol and fragrance at 11 each, linalool at 7. What we did not find, anywhere in the catalogue, is a single paraben, a single isothiazolinone, a single formaldehyde releaser or a drop of lanolin — which is worth saying, because 46 of our 56 products carry a Paraben-free chip and every one of them is telling the truth about a class that is not in this catalogue at all. The ten products that do not carry the chip are equally paraben-free. That is what a claim looks like when it has stopped carrying information.
The practical takeaway
Treat hypoallergenic, allergen-free, sensitive-skin and dermatologist-tested as what they are: a statement of intent from a marketing department, unaudited by anybody. Not a lie, usually. Just not a fact you can act on.
If you have a known allergy, the front of the box is the wrong surface entirely. You want the ingredient list, and you want to search it for your specific allergen name and its synonyms — most of them have several. If you have never been patch-tested and your skin reacts to things you cannot pin down, that is a clinic question, and a baseline series test will give you a name in an afternoon that no amount of label-reading will.
And if you are shopping without a known allergy, the honest ranking is short: fewer fragrance components beats more, a shorter list beats a longer one at the same price, and the phrase on the front is worth exactly nothing next to the six lines at the bottom of the back. We took the same walk through our own fragrance-free claims and through clean, vegan and non-comedogenic, and the pattern does not change: the claims that survive contact with an ingredient list are the ones that name a specific molecule.
We are going to ask seapuri why a mist listing tea tree oil, lavender, clary sage and anise extracts is sold as allergen-free, and we will print the answer. In the meantime the product page is here, with the full list, which is more than the front of the bottle gives you. Which claims we think are worth reading is set out in this piece, and the products that scored best under our screening are on the awards page.
- Almay, Inc. v. Califano, 569 F.2d 674 (D.C. Cir., decided 21 December 1977). Full opinion read. Challenged regulation 21 CFR 700.100, published 6 June 1975.
- 21 CFR Part 700, current text as retrieved from the electronic Code of Federal Regulations, 7 September 2026. § 700.3 Definitions; Subpart B §§ 700.11–700.35. No § 700.100 and no hypoallergenic provision.
- Hiranput S, McAllister L, Hill G, Yesudian PD. "Do hypoallergenic skincare products contain fewer potential contact allergens?" Clinical and Experimental Dermatology 49(4):386–387 (2024). doi 10.1093/ced/llad436. 208 products; 153 (73.6%) with at least one baseline-series allergen or related chemical.
- Dumycz K, Kunkiel K, Feleszko W. "Cosmetics for neonates and infants: haptens in products' composition." Clinical and Translational Allergy (2019). doi 10.1186/s13601-019-0257-8. 212 products; 41 haptens; 87.7% with at least one.
- Goossens A. "Cosmetic Contact Allergens." Cosmetics 3(1):5 (2016). 14,911 patients patch-tested 1990–2014.
- British Association of Dermatologists, press release and annual meeting abstract, 26 June 2023 (Hiranput, Smith, Yesudian). Conference version of reference 3; product count identical, percentages reported differently.
- DHALIORA screening data: 56 products, 1,920 ingredient rows, full recount 2026-09-07. Five products carrying an allergen-shaped claim; 35 products with at least one baseline-series ingredient; 82 matching rows; zero parabens, isothiazolinones, formaldehyde releasers or lanolin.
