The Darkening Is Thickened Skin. That Is Why It Will Not Scrub Off.

An online survey of 48,000 adults across 34 countries asked what pigmentary problems people have. Solar lentigines came first, reported by 27.5 percent.

Second, at 17.2 percent — 8,289 people — was axillary hyperpigmentation. Ahead of post-inflammatory pigmentation, ahead of dark circles, ahead of melasma and vitiligo.

The caveat has to travel with the number: this is self-report. Fewer than half of those reporting any pigmentary disorder had a physician diagnosis, and about two in five self-diagnosed, either before the survey or from the survey’s own definitions and images. It is not a clinically verified prevalence.

What it does establish is that this is not a niche complaint. And for something this widely experienced, the evidence underneath the explanations people are given is remarkably thin.

The deodorant claim, checked

The standard explanation is that deodorant or antiperspirant is staining or irritating the skin.

No study measures pigmentation as an outcome of deodorant or antiperspirant use. Not one that I could find.

What exists is two adjacent literatures. There is a real case-report literature on allergic contact dermatitis to deodorant ingredients — zinc ricinoleate in a natural deodorant, oxidized linalool, a silicone, a sorbitan ester — but every one of those reports dermatitis as the endpoint, not pigment. The closest thing to a deodorant-causing-underarm-disease primary source is a single patient with granulomatous dermatitis from a crystal deodorant, and even that is a different outcome.

And there is a genuine entity called pigmented contact dermatitis, demonstrated properly. A 1977 study of 53 patients patch-tested cosmetics, coal tar dyes, fragrances, lanolin and parabens, and found positive responses to one particular red dye in 22 of 38 tested. A 1989 case describes brown facial patches in a woman testing positive to lemon oil, geraniol and hydroxycitronellal — two of which were in her face powder — with the pigmentation disappearing within six months of avoiding them.

That is a clean demonstration that a cosmetic can cause hyperpigmentation and that it can resolve on withdrawal. It is also about face powder, on faces.

So the chain from deodorant to dermatitis to pigment is mechanistically plausible and supported for other product categories. It has not been demonstrated for deodorants. Anybody telling you to switch brands to fix your underarms is reasoning from analogy, which is fine, as long as they say that is what they are doing.

The one proper experiment on hair removal

Sixty-six healthy Thai women aged 30 to 40 went two weeks with no hair removal at all. Then each removed hair from one randomized armpit by her own established method — 21 shaved, 22 plucked, 21 waxed — with assessment at 30 minutes and again at 48 hours, plus swabs for inflammatory cytokines.

All three methods raised erythema and dryness. Shaving produced significantly less redness but significantly more dryness than the others at 30 minutes. Interleukin-1 alpha and its receptor antagonist rose; interleukin-8 fell.

And at 48 hours there were no differences in post-inflammatory pigmentation between the techniques, because there was no detectable pigmentation to differ.

The authors’ summary is that a single hair-removal treatment did not appear to induce pigmentation or further biophysical change. Their explanation for the gap between that and what people report is stated as an inference: pigmentation increases that individuals observe over time must result from cumulative effects.

Must is doing a lot of work. That is a hypothesis, and it is the obvious one, and nobody has tested it. There is no study of repeated hair removal over months with pigmentation as an endpoint.

Which is a shame, because the study design to answer it is not hard and this one already exists as a template.

How thin the rest of it is, measured

I ran the searches rather than describing an impression.

Threading and hair removal complications: two records in the entire indexed literature. Sugaring: two records. Waxing and hyperpigmentation: thirteen records, of which eleven turn out to be about something else entirely — a liposculpture burn, a veterinary folliculitis, congenital adrenal hyperplasia, an ulcer on a foot.

Two papers each, ever, for two of the most widely used hair-removal methods on earth. Any figure you read for how often threading causes pigmentation in darker skin does not have a primary source of that kind behind it.

The nearest thing to real-world data is an Indian tertiary-hospital series of 102 patients who presented with a skin problem after a salon procedure over twelve months. Acneiform eruptions in 26, post-waxing folliculitis in 17, facial hyperpigmentation in 15, molluscum in 11, plane warts in six, tinea in three.

Read that denominator carefully. The 102 are the people who came in with a problem, not the people who had procedures. Fifteen out of 102 is not a 14.7 percent rate of pigmentation among salon customers, and the authors do not claim it is. It is a description of what turns up in clinic.

Where there is a hard number, and it is a large one

The exception is shaving in tightly coiled facial hair, because somebody studied a population that had no choice about it.

A cross-sectional study at the National Police Academy in Dakar examined 655 police students, all of African descent with curly hair, all required to shave weekly.

Pseudofolliculitis barbae in 254 of them — 38.8 percent. Among the men, 242 of 554, which is 43.7 percent; among the women, 12 of 101.

Then the number that matters here: complications were noted in 90.1 percent of those cases, in the form of post-inflammatory hyperpigmentation in 87 percent and keloid scars in 3.1.

Eighty-seven percent of the 254 with the condition, not of the 655 shavers. But that is still the picture: where mandated shaving produces pseudofolliculitis in richly pigmented skin, pigmentation is not an occasional side effect. It is what normally happens next. Keloid-prone skin carried an odds ratio of 2.9 for developing it in the first place, and the association with acne keloidalis nuchae ran at 8.8.

The authors note, drily, that very few data are available for this condition, especially in sub-Saharan Africa. The largest study of it was possible because an institution required 655 people to shave.

The laser data, and a familiar pattern

Two 2025 studies in Sudanese women with Fitzpatrick types IV to VI are worth having.

The larger, 55 women treated on the chin with a long-pulse Nd:YAG laser, found that fluence significantly affected outcomes — with 25 joules per square centimeter, at the bottom of the range used, demonstrating effectiveness. Skin type, age and geographic location were not significant factors. No adverse events and no paradoxical hair growth were observed, and the authors’ conclusion is that low-fluence treatment is safe and effective in darker skin.

The smaller, ten women treated with a diode laser, reported one patient with pain and one superficial blister at 9 joules that resolved after the fluence was adjusted, with no pigmentary changes at all. Ten patients cannot establish the rate of a complication that may occur in a few percent of people, and that series should not be read as if it could.

But notice the convergence. In acne scarring, the finding that mattered was that halving the treatment density cut the pigment complication without costing efficacy. In hair removal, the effective setting was the low end of the fluence range. Two different procedures, same principle, and it is the same argument I made about density: in pigmented skin, the dial matters more than the device.

The cause with the best evidence, which is not cosmetic

Acanthosis nigricans is velvety, thickened, darkened skin in the folds — neck, armpits, groin. It has been studied properly, and the reason is that it was studied as a diabetes screening question rather than a beauty one, so the papers have real denominators.

The foundational study examined the back of the neck of every child in the sixth and eighth grades of the public schools of one Texas city: 1,412 children, personally examined during a state health survey. Not a clinic sample. Everyone.

Acanthosis nigricans in 7.1 percent overall. By group: two of 440 white non-Hispanic children, 19 of 343 Hispanic children, and 80 of 601 Black children. Those counts sum to 28 fewer than the total examined and the paper does not account for the difference, so I am reporting them as published.

Fasting insulin concentrations correlated with the presence and severity of the lesion. The authors’ conclusion in 1989 was that the high prevalence of this skin sign suggests that insulin resistance and hyperinsulinemia, with all of their serious medical implications, are also highly prevalent.

Later work with the same design keeps confirming it. A random sample of 2,205 Cherokee Nation members aged 5 to 40: acanthosis nigricans in 34.2 percent, hyperinsulinemia in 47.2, and the association surviving multivariate adjustment at p equals 0.0001. A study of 2,520 American Indian schoolchildren in the Northern Plains: 9.7 percent, with a relative risk of 9.8 for developing it in the year following, in children who were obese rather than normal weight. Southwestern primary care, 1,133 patients: 17 percent of children and 21 percent of adults, with a prevalence ratio for type 2 diabetes of 1.97 after controlling for age, body mass index and the number of other risk factors.

Two things follow from that set.

One: there is no single Native American prevalence figure. The studies located run from 9.7 to 38 percent across different communities and age ranges, and anybody quoting one number is picking one.

Two, and this is the sentence I would put on a poster: the apparent darkening is due to hyperkeratosis.

That is from the group’s own review, describing what the histology shows. The skin is thickened and roughened, and the thickening is what reads as dark. It is not surface soiling, and it is not primarily melanin sitting in the epidermis.

Which is why it does not come off. People scrub, and use exfoliating scrubs, and try lemon juice and baking soda, and the friction is itself one of the things that drives the change. The mechanism and the folk remedy are working in the same direction.

The authors are also careful about the ethnic reading of their own data. Acanthosis nigricans, they write, identifies a subgroup within an ethnic group who have the highest insulin concentrations and the most severe insulin resistance. It is not reducible to who you are. It marks who, within any group, is at highest risk.

What has been tried, and how little of it there is

A 2025 systematic review of axillary hyperpigmentation treatment searched four databases and found ten studies. It opens by stating that this is a common complaint with no standard treatment.

What has been tested: niacinamide, desonide, sweet orange extract, perilla leaf extract, Cyperus rotundus oil, glycolic acid and various serums, with desonide and the Cyperus oil showing the clearest effect and few side effects. Q-switched Nd:YAG laser across three studies, and intense pulsed light, both with significant improvement and minimal adverse events, with more pain reported from the laser. One comparison found pulsed light better than an alpha-hydroxy acid.

No pooled sample size. No phototype distribution. No effect sizes in the abstract.

Ten studies, for something 8,289 of 48,000 surveyed adults say they have.

One more study is worth knowing because it gives a time course. Twenty women of Fitzpatrick types III and IV were followed with a standardized pigmentation scale after underarm surgery. Pigmentation peaked between the third and sixth months and then declined over the following months, and in some patients resolved completely.

Post-inflammatory darkening in the underarm, after a defined injury, is a months-long process that can reverse. That is a genuinely useful thing to know while waiting.

Stop scrubbing, and check the neck

Stop scrubbing. If the skin is velvety and thickened rather than flat and stained, the darkness is in the texture and abrasion makes the texture worse.

If it is velvety and thickened, and especially if the same change is on the back of the neck, that is worth mentioning to a doctor — not because it is dangerous in itself, but because in every properly designed study it has tracked with fasting insulin, and the whole reason we have good prevalence data on it is that it was being used as a screening sign. What the neck sign adds on top of weight is genuinely split: two studies say it adds information and two say it does not, with the severity of the sign itself running about three quarters heritable.

If it is flat pigment following irritation from shaving or waxing, the published time course after a defined insult is a peak at three to six months and then a slow fade, which argues for changing the irritation rather than attacking the color.

And if somebody offers you a laser, the setting is the conversation. The best result in the darkest skin came at the bottom of the fluence range, not the top.

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.