Two Studies Say the Neck Tells You Something. Two Say It Only Tells You the BMI.

In 2021 a research team needed a way to set waist circumference cut-offs for metabolic risk in Pacific Islander children aged two to eight. They had 4,023 children and they needed a reference standard — something that indicated a child was in metabolic trouble, against which a tape measure could be calibrated.

They used the darkening on the back of the neck.

Acanthosis nigricans, as the criterion. Not a blood test. A skin sign, treated as the ground truth that a waist measurement had to predict.

Raised, furrowed, and not a stain

Velvety, thickened, darker skin, most often at the back of the neck, in the armpits, in the groin, and over the knuckles. It has a texture — that is the diagnostic part. Run a finger over it and it is raised, slightly furrowed, not flat.

Which is the whole reason it does not wash off. The color is not sitting on the surface. The epidermis is thickened and folded, and thickened skin looks darker for the same reason a folded sheet of paper looks darker at the crease. This is the identical mechanism behind underarm darkening, and it produces the same futile scrubbing.

The conventional account is that insulin, at high circulating levels, cross-reacts with growth factor receptors on keratinocytes and fibroblasts and tells the skin to proliferate. Which would make the neck a readout of the pancreas.

The evidence for that is more argued-over than the account suggests.

The case that it means something

A UK single-center study took 94 young people with acanthosis nigricans and matched them to 94 with obesity alone — same age, same body mass index, same sex.

The acanthotic group had a median fasting insulin of 215 picomoles per liter against 126. Mean fasting glucose 4.7 against 4.5. Median HOMA-IR — the standard calculated index of insulin resistance — 6.4 against 3.7.

Nearly double the insulin resistance at the same body mass index. Within that population, the sign had a positive predictive value of 81 percent.

An Israeli pediatric obesity clinic found something compatible. Of 230 children with obesity, only 21 percent were metabolically healthy, and the three best predictors of being in the unhealthy group were male sex, a rising BMI score over follow-up, and the presence of acanthosis nigricans at baseline — odds ratio 2.35.

The case that it does not

A community sample of adolescents with obesity in a UK trial found acanthosis nigricans in 63 percent of them, severe in 43 percent.

Presence of the sign was associated with fasting hyperinsulinemia — but not after adjusting for body mass index. Severe-grade acanthosis was associated with arterial stiffness measured by pulse wave velocity — but not after adjusting for body mass index, ethnic grouping and age.

The authors’ conclusion: presence and severity offered little additional information on cardiovascular risk beyond the degree of obesity itself, and the relevance of the sign should be interpreted with caution.

Then the study that went after the causal structure directly. 670 non-diabetic Mexican American children, aged 6 to 17, acanthosis nigricans in 33 percent, severity scored 0 to 5 and treated as a quantitative trait so it could be run through genetic and mediation modeling.

Two findings.

The heritability of the severity score was 0.75.

And the model that fitted best was not insulin resistance causing the skin change. It was body mass index explaining the insulin resistance and the skin change separately, in parallel, with no causal relationship between the two.

Which of these is right, and why the disagreement is the useful part

I do not think these results are in conflict so much as answering different questions.

The UK case-control matched on body mass index and found a difference. The community sample adjusted for body mass index and found the difference disappear. Matching and adjustment are not the same operation and they behave differently when the covariate is doing most of the work.

But notice what the UK authors flag in their own limitations. Their groups were well matched for age, body mass index and sex — and were not matched for ethnicity.

That is the confound this entire section exists to name, appearing in the paper’s own text. If acanthosis nigricans is more visible or more prevalent at a given metabolic state in some ancestries, then a study that matches on weight but not ancestry is partly measuring who is in which group.

And the heritability figure tells you that is not hypothetical. Three quarters of the variation in how severe this sign gets, in that cohort, tracked with relatedness.

What a 0.75 heritability does to the clinical logic

It cuts both ways, and both directions matter.

If whether you show the sign is substantially inherited, then its absence is weak reassurance. Two people with identical insulin resistance can have one visibly marked neck and one clear one, and the clear one gets sent home. The Pacific study half-notices this: its own closing line calls for analysis of why acanthosis nigricans appeared at lower-than-expected waist percentiles in that population.

And its presence in somebody who is not heavy is not nothing. In the same modeling, the sign correlated with C-reactive protein and inversely with HDL cholesterol and physical fitness score — it is traveling with cardiometabolic risk even where the causal arrow is unclear.

What it is not is a measure of how much someone eats, and it is not a measure of how they wash. It is skin thickening under hormonal instruction, in a person whose skin was always going to be more prone to it than somebody else’s.

Take it to a blood test, not a scrub

Treat it as a prompt, not a blemish.

The pigment is a consequence of the thickening, so the products aimed at pigment are aimed at the wrong layer, and the exfoliating approach is aimed at the right layer with the wrong force — the same trap as the underarms. Scrubbing thickened, irritated skin in a person prone to pigment deposition is a reliable way to make the color worse.

The conversation worth having is a metabolic one, and the useful framing is that you have noticed a physical sign and want to know what it is tracking in your particular case. In a matched study it carried an 81 percent positive predictive value for insulin resistance; in a community sample it added nothing beyond weight. Both of those are reasons to test rather than to guess.

And if the sign runs in your family, say so. In the one cohort where anybody measured it, that was three quarters of the story.

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.