Material guide · 70
Material guide: isoeugenol — it just became the top fragrance allergen, because the ones above it were regulated away
· 28 min read
CAS 97-54-1. A 2020 analysis of German IVDK data from 2016-2018 reports that the allergen with the most positive reactions in fragrance mix I is no longer oakmoss absolute but isoeugenol. It did not become more dangerous; oakmoss and HICC were prohibited in the EU, and the same paper records fragrance mix I positives falling to a historical low of 5.4%. IFRA Amendment 49 caps it at 0.11% in Category 4, one ninetieth of the 9.9% allowed for the hexyl cinnamal of the previous piece. And the median dose across 954 public formulas exceeds that limit once diluted, because most of those formulas predate 2020.
Geier and Brans analysed German IVDK patch test data from 2016 to 2018 in Der Hautarzt in 2020, and one sentence deserves lifting out on its own:
Of fragrance mix I, the allergen with the most positive test reactions is no longer oakmoss absolute, but isoeugenol.
It moved into first place.
And it did not become more dangerous. The one ahead of it was taken away.
The short version
- In the 2016-2018 IVDK data, the top allergen in fragrance mix I changed from oakmoss absolute to isoeugenol (Geier and Brans 2020).
- The same paper records that positive reactions to fragrance mix I reached a historical low of 5.4% in 2018, with mix II at 3.2%. The decline is mainly due to reduced use of oakmoss containing atranol and chloroatranol, and of HICC — both prohibited for cosmetic use in the EU from August 2019.
- So the change at the top is a by-product of regulation working, not of anything happening to this material.
- The same paper gives another number: of the 26 fragrances subject to mandatory declaration, only 11 elicited allergic test reactions in more than 1% of patients tested.
- IFRA Amendment 49 caps it at 0.11% in Category 4 (fine fragrance), one ninetieth of the 9.9% allowed for the hexyl cinnamal of the previous piece.
- And the median dose across 954 public formulas exceeds that limit once diluted. Most of those formulas predate the 2020 amendment.
- Two assays describe its potency in different words: the guinea pig maximization test classified it as an extreme sensitiser, while a non-radioisotopic LLNA gave an EC3 of 12.7% (Takeyoshi et al. 2008).
About 8 minutes.
How the ranking changed
Laying out the sequence, because this stretches across several pieces here.
| When | What |
|---|---|
| 2003-2004 | IVDK tests 21,325 patients: oakmoss 2.0%, isoeugenol 1.1% |
| 2015 | "New oak moss" with reduced atranol and chloroatranol shown to elicit significantly less |
| August 2019 | Oakmoss containing atranol and chloroatranol, and HICC, prohibited for cosmetics in the EU |
| 2016-2018 | Isoeugenol becomes the top allergen in fragrance mix I; positives fall to a historical low of 5.4% |
Those middle two rows are the line we traced in the Veramoss piece.
Oakmoss's IFRA ceiling was pushed to 0.10% in Category 4 with a specification requiring atranol and chloroatranol each below 100 ppm, and the market shifted heavily toward synthetic substitutes. The IVDK figures show that shift landing in the clinic.
Geier and Brans state it plainly in their discussion: the decline of positive test reactions to the fragrance mixes is mainly due to the reduced use of oakmoss containing atranol and chloroatranol, and of HICC, and since both were prohibited in the EU from August 2019, a further decline in sensitisation frequencies can be expected.
This is an uncommonly clear case of a control measure visibly working. Its side effect: whatever remains on the bench moves up the ranking automatically.
So "isoeugenol is now number one" cannot be read alone. It is missing three words at the end: of what remains.
Only 11 exceed 1%
Geier and Brans have another sentence that says the same thing as the previous piece:
Only 11 of the 26 fragrances subject to mandatory declaration elicited allergic test reactions in more than 1% of the patients tested.
Eleven out of twenty-six. The other fifteen came in below 1% in the clinic.
And isoeugenol is the first of those eleven.
This is the label-versus-risk gap with a number on it. That list of 26 was produced by rule: the hexyl cinnamal of the previous piece ranks in the top three on labels and 0.1% in the clinic, and this material is the reverse.
The paper's own conclusion says as much: a differentiated consideration of the individual fragrances under allergological aspects is urgently required.
The two materials side by side
Putting the last round's material next to this one shows what the regulation is tracking:
| alpha-hexyl cinnamaldehyde | isoeugenol | |
|---|---|---|
| 2007 IVDK sensitisation rate | 0.1% (near bottom of 26) | 1.1% (5th) |
| 2016-2018 IVDK | — | top of fragrance mix I |
| IFRA Category 4 limit | 9.9% | 0.11% |
| IFRA Category 5A | 2.5% | 0.027% |
| IFRA Category 11A | 38% | 0.0090% |
| Corpus appearances | 157 (16.5%) | 68 (7.1%) |
| Corpus median dose | 8.00% | 1.276% |
| Suppliers | 56 | 56 |
The IFRA limits differ by ninety-fold and the clinical rates by elevenfold. Same direction.
Labelling frequency agrees with neither — hexyl cinnamal is top three on household ingredient lists and isoeugenol is not.
So three datasets describe three different things:
- Clinical sensitisation rates say who actually causes allergy.
- IFRA limits track that, plus an exposure estimate.
- Labelling frequency tracks who gets used a lot and must be declared.
Read the third as though it were the first and you will fear the wrong things. We've written about reading safety studies, and this pair is the cleanest demonstration of it.
A problem hiding in the corpus
Across the 954 public formulas, isoeugenol appears in 68 (7.1%), at a median dose of 1.276%, third quartile 3.00%, heaviest single formula 42.00%.
That median needs arithmetic.
- 1.276% of the concentrate
- Finished perfume usually dilutes to 10-20%
- So 0.13% to 0.26% in the finished product
- The IFRA Category 4 limit is 0.11%
The median formula, diluted to 20%, lands at roughly 2.3 times the limit; at 10% it is still slightly over.
Last round I demonstrated the opposite case in the same place: hexyl cinnamal's 8.00% looked as though it sat against the 9.9% ceiling, and after dilution it was an order of magnitude away. That was a false alarm.
This one is not.
The reason is not mysterious: 74% of that corpus comes from patents and published sources, and IFRA's history with this material is long — the standard's own text lists publication years of 1980, 1998, 2001, 2004, 2007, and then Amendment 49 in 2020. Most of those formulas were written before the current limit.
The practical upshot is direct: recalculate this one whenever you use an old patent formula as a template. It is the sort of material where copying across puts you over.
Potency: two assays, two different words
Takeyoshi and colleagues compared isoeugenol with two of its dimers in the Journal of Applied Toxicology in 2008, using both a non-radioisotopic local lymph node assay and a guinea pig maximization test.
The results:
| Substance | Guinea pig maximization | non-RI LLNA EC3 |
|---|---|---|
| isoeugenol | extreme | 12.7% |
| β-O-4-dilignol (dimer) | weak | >30% |
| dehydrodiisoeugenol (dimer) | moderate | 9.4% |
One material, called extreme by one assay and given an EC3 of 12.7% by the other.
Those do not contradict each other, and they are not the same ruler. The paper's own introduction calls isoeugenol a well-known moderate human sensitiser — which is not the same word as the guinea pig test's "extreme".
Worth remembering: a potency classification depends on which assay produced it and on that assay's grading rules. When you meet an adjective, ask which system it came from.
The paper's other finding is more hopeful: dimerisation reduced sensitisation potency, and isoeugenol dimers are not currently used as fragrance chemicals. The authors suggest dimerisation may yield a candidate cosmetic ingredient with low sensitisation risk. That is a direction rather than an available substitute.
Natural does not mean gentle
The occurrence field for this material runs long:
allspice leaf oil, allspice, basil, betel leaf oil @ 10.59%, blueberry, cassia, Ceylon cinnamon bark, Ceylon cinnamon leaf oil, clove bud oil @ 1.08%, clove leaf oil, clove oil, champaca concrete @ 0.10%…
It is everywhere in nature. Clove, cinnamon, allspice, basil, betel — all things you can buy in a kitchen or a market.
And it is currently the material with the most positive reactions in fragrance mix I.
We've asked whether naturals are gentler, and this is that piece's most direct illustration. How widespread a molecule is in nature carries no information about its capacity to cause allergy.
This also bears on buying essential oils. Clove bud oil contains 1.08% of it, so if your formula has clove oil in it, that thread is already there — the same situation as a jasmine absolute carrying salicylates.
Physical properties
| Property | Value |
|---|---|
| CAS | 97-54-1 |
| Formula | C₁₀H₁₂O₂, MW 164.20 |
| Appearance | colourless to yellow clear liquid (est) |
| Melting point | 14-18 °C |
| Boiling point | 266-267 °C |
| Flash point | > 100 °C |
| logP | 3.00 (est) |
| Water solubility | 165.9 mg/L at 25 °C (est) |
| Substantivity | 400 hours at 100% |
| Strength | medium |
| Suppliers | 56 |
| Odour | sweet, spicy, clove, woody, allspice, carnation, floral |
| GHS | H302, H315, H317, H319, H335 |
The 14-18 °C melting point is worth noting. We ran the numbers in the white particles piece: materials melting near room temperature are the ones that catch people out when the weather turns. This sits in that band.
Substantivity of 400 hours at 100% matches hexyl cinnamal, both high values measured neat. This is not a material that leaves early.
What this doesn't establish
I have not smelled it and made no comparisons. The whole piece sets database fields, the formula corpus, the IFRA text and two papers against each other.
"The ranking changed because of regulation" is Geier and Brans's own discussion, offered as an explanation for an observed decline rather than a demonstrated cause. Other factors — changing formulation habits, a changing patient mix — could contribute.
The IVDK population is dermatology-clinic patch test patients, not the general public. That population is enriched for people with skin problems.
"Top of fragrance mix I" means the ordering within mix I, not across all fragrances. HICC still leads in mix II.
The over-limit calculation uses a 10-20% dilution assumption. Real dilution rates vary by product, and I do not know which product category those patent formulas were aimed at. Category 4 is fine fragrance; a formula intended for Category 10B (limit 0.75%) yields a different conclusion.
IFRA limits are finished-product concentrations while the corpus percentages are shares of concentrate. Converting between them needs a dilution rate, and I used a common range rather than those formulas' actual values.
I cannot reconcile Takeyoshi's "extreme" and 12.7% into one number. I list both because they come from different assays and grading systems, and I am not in a position to judge which better reflects the human situation.
The dimer section is that paper's outlook, not an available product. It states that isoeugenol dimers are not currently used as fragrance chemicals.
The GHS classification is what the database records, not a complete regulatory determination. Work from the SDS for your own batch and the current IFRA standard.
References
J. Geier, R. Brans, Wie häufig ist Duftstoffallergie wirklich?, Der Hautarzt, 71(3), 197-204 (2020). PMID 31965209. doi:10.1007/s00105-019-04534-w
M. Takeyoshi, K. Iida, K. Suzuki, S. Yamazaki, Skin sensitization potency of isoeugenol and its dimers evaluated by a non-radioisotopic modification of the local lymph node assay and guinea pig maximization test, Journal of Applied Toxicology, 28(4), 530-534 (2008). PMID 17929237. doi:10.1002/jat.1305
Related: alpha-hexyl cinnamaldehyde, Veramoss / Evernyl, are naturals gentler, what IFRA limits actually limit.