Fragrance Allergy Was Nearly Twice as Common in the White Patients

Patch testing is how you find out what your skin is actually reacting to. Fifty or so allergens go on your back under adhesive, they stay for two days, and a dermatologist reads what happened at 48 and 96 hours.

The North American Contact Dermatitis Group has published the same race-comparison analysis three times across three decades on three non-overlapping cohorts. Those three papers are the best evidence there is, they do not entirely agree with each other, and the thing they are clearest about is not what most people assume.

The numbers, from the generation that published them

The 1998–2006 cohort: 19,457 patients tested across thirteen centers with 45 allergens. White patients 92.9 percent, Black patients 7.1 percent.

Final diagnoses were essentially the same in both groups — allergic contact dermatitis in 45.9 percent of white patients and 43.6 of Black patients, irritant contact dermatitis 13.0 and 13.3. Atopic dermatitis was higher in Black patients, 13.3 against 8.9, which matters because atopic dermatitis is itself a sensitization risk factor and a confounder in every comparison below it.

Higher in Black patients: p-phenylenediamine 7.0 percent against 4.4 (p less than 0.001). Bacitracin 11.6 against 8.3. Mercaptobenzothiazole 2.7 against 1.8. Thiuram 6.2 against 4.3. Mercapto mix 1.9 against 0.8.

Higher in white patients: fragrances 12.12 percent against 6.77. Formaldehyde 9.25 against 5.45. Formaldehyde-releasing preservatives 9.80 against 6.18. All at p below 0.0001.

Fragrance allergy was nearly twice as common in white patients. Which runs directly against how fragrance-free-for-sensitive-skin marketing tends to get racialized, and against the assumption most people bring to the subject.

One discipline note on the PPD figure. Seven percent against 4.4 is a 1.59-fold difference — 59 percent higher in relative terms, 2.6 percentage points in absolute terms. It is a real and consistent difference, not a categorical one, and both framings belong in the same sentence.

What replicates and what does not

Across all three generations — 1992–98, 1998–2006, and the newest covering 2007–2020 with 32,138 patients — PPD is higher in Black patients every single time. That is the most robust race-associated patch test finding in the North American literature.

Fragrance higher in white patients: consistent in the two most recent. Formaldehyde and its releasers higher in white patients: consistent in all three.

The rubber accelerators, and cobalt, are not stable. Cobalt was higher in Black patients in the first generation and highest in Asian patients in the third. Thiuram was higher in Black patients in the first two and highest in Asian patients in the third. So “Black patients are more allergic to rubber accelerators” is not something the data supports across cohorts. PPD and fragrance are the two you can stand behind.

The newest analysis is also the first to go beyond Black and white: white 86.5 percent, Asian 6.8, Black 4.8, Hispanic 2.0. Thirteen of fifty allergens differed significantly. Highest in white patients: balsam of Peru, neomycin, quaternium-15, fragrance mix II. Highest in Black patients: PPD and disperse dyes. Highest in Asian patients: nickel, cobalt, thiuram mix. Highest in Hispanic patients: propylene glycol.

Its per-allergen percentages sit in paywalled tables and I am not going to attach numbers to it. Direction and significance only.

What the researchers themselves say it means

This is the strongest card in the whole subject and it is quotable verbatim from three separate papers.

1992–98: the differences, “although possibly related to genetic factors based on race, are more likely related to differences in allergen exposure determined by ethnicity.”

2007–2020: the differences “may relate to sociocultural variations in allergen exposures.” And, flatly, in the limitations: “Race/ethnicity are social constructs.”

An independent testing center in 2025, with a different population: “Differing cultural practices may result in these variations.”

Three papers, two research groups, twenty-four years apart, all reaching for exposure rather than biology.

The PPD story is the clearest illustration. PPD’s dominant exposure routes are permanent oxidative hair dye and black henna temporary tattoos. Disperse dyes are textile dyes — aromatic amines that cross-react with PPD — and they were also highest in Black patients in the newest analysis. Two chemically related exposures, co-elevated, pointing at hair and textile products rather than at a property of skin.

And here is the honest limit. No study I could find has measured product use and sensitization rate in the same patients. The exposure explanation is an inference the researchers keep drawing, consistently, across decades — and nobody has tested it directly. That study appears not to exist, which is its own small commentary on what gets funded.

The claim I cannot source

It is widely said that the standard patch test series were developed on predominantly European populations. It is plausible on its face — the international and European contact dermatitis research groups originated in Europe in the 1960s and 70s — and I could not find a single published source that establishes it. So I am not printing it.

What is verifiable is stronger anyway.

A screening series is assembled from allergens that turn up frequently in the tested population. The tested population in North America’s flagship dataset was 89.5 percent white, then 92.9 percent, then 86.5 percent. Whatever the historical origin of the allergen list, the frequency data that determine which allergens earn a place on it are overwhelmingly derived from white patients.

The consequence is arithmetic rather than intentional: an allergen common in a group that makes up 4.8 percent of the reference dataset is systematically less likely to clear the threshold for inclusion.

Nor could I find any validation study showing whether the standard series performs differently — different yield, different miss rate — in Black against white patients. That has not been tested either.

Four point eight percent

Across the three NACDG papers, the Black share of the patch-tested population went 10.5 percent, then 7.1, then 4.8.

More than halved across three decades, in the reference dataset for a diagnostic procedure, while the US Black population did not halve.

It should be checked against changes in which centers contribute before anybody calls it a pure access effect. But it is a real trend across three published papers from the same group, and as far as I can tell nobody in the literature has remarked on it.

Reading the test on skin that does not go red the same way

A patch test reading is a visual judgment. The grades — doubtful, weak positive, strong positive — turn substantially on erythema, and erythema is the thing that is hardest to see on deeply pigmented skin.

A 2026 systematic review in the field’s own journal states it without hedging: interpretation of patch test reactions remains subjective, relying heavily on visual assessment and palpation, which presents challenges particularly in patients with skin of color, where erythema may be less apparent.

That review then examined 38 studies of the technologies meant to fix it — ultrasound, thermography, optical coherence tomography, reflectance confocal microscopy, 3D imaging — and concluded that many were limited to binary classification and lacked validation in diverse populations.

Thirty-eight studies of objective patch-test imaging, and the instruments built to solve the darker-skin reading problem have not themselves been validated in darker skin.

The first educational module I could find for training clinicians to read patch tests across Fitzpatrick I through VI was published in 2026. Forty providers, pre- and post-intervention self-reported confidence. That is a pilot, not a solution. It is evidence that the field started building the training this year.

One older pilot is worth naming because it will be misread otherwise. Testing with a soap irritant, white subjects showed significantly lower clinical thresholds for irritant contact dermatitis than African American subjects, with more severe features. The threshold was determined by clinical scoring — by looking. Whether that reflects a real difference in barrier response or a difference in how visible the reaction is on lighter skin is precisely the confound this section is about, and the study cannot separate them.

The thing on the hair dye box

Permanent hair dye packaging carries a self-test: dab some on, wait, see what happens.

A 2024 systematic review of hair dye allergens, in the field’s own journal, on that open application self-test: the lack of standardized conditions makes the efficacy of this process doubtful.

That is the allergy check most people ever perform, printed on the product, and the field’s assessment of it is that it may not work.

The same review notes that cross-reactivity between PPD and related aromatic amines means switching to an alternative dye is not reliably safe either — though it adds, fairly, that cross-reactivity detected on a patch test is not always clinically relevant.

Worth knowing about the stakes, from a pediatric case report: the aftermath of a severe PPD reaction was a permanent hypopigmented patch. The acute dermatitis resolves. The pigmentary scar does not. On darker skin that is the part that lasts.

Taking this into an appointment

If you have reacted to hair dye, black henna, or anything dark and dyed against your skin, the allergen to ask about by name is p-phenylenediamine — and to ask that disperse dyes go on the panel too, because they cross-react and they are the textile half of the same chemistry.

If you are being patch tested on deeply pigmented skin, it is reasonable to ask the reader to palpate as well as look, and to photograph the reactions. The field’s own systematic review says the reading is subjective and hardest here. That is not a rude question; it is the published position.

And the general shape of it: the overall rate of allergic contact dermatitis was the same in every group in every one of these studies. What differed was which allergens, and the researchers who collected the data have been saying for twenty-four years that the difference is in what people are exposed to. Nobody has yet run the study that would prove them right.

One place that exposure question turns clinical is the dermal pigmentation associated with hair dyes and cosmetics. The international panel that agreed on what to call it proposed classifying it by whether contact sensitization is present, which makes a patch test the first useful step rather than the last.

This article is educational and is not a substitute for care from someone who can look at your skin. Anything new, changing, painful or spreading should be assessed rather than researched.