Nobody Biopsied Anyone. The Number Is Still Everywhere.

Central centrifugal cicatricial alopecia starts at the crown and works outward. The scalp there goes smooth and slightly shiny, the follicles are destroyed, and what is gone does not come back. It is the most common scarring hair loss in Black women, and almost everything printed about it is shakier than it sounds.

Start with the prevalence figure, because it is the one that travels.

Five point six percent, and where it came from

Olsen and colleagues, JAAD 2011, 529 African American women. The sentence everybody quotes: “Extensive central scalp hair loss was seen in 5.6% of subjects.”

Three things about that study.

The women were recruited at six workshops held at four sites. That is a self-selected sample of people who came to a hair-loss event — not a population.

Nobody was biopsied. The authors say so themselves as a stated limitation: no scalp biopsy correlation with clinical pattern of hair loss. What was measured was extensive central scalp hair loss on a standardized photographic scale. Not CCCA. A pattern that CCCA produces, and that several other things produce too.

And the same study found no association between that hair loss and relaxer use, hot comb use, seborrheic dermatitis, a reaction to a hair care product, bacterial infection, or male pattern baldness in the father. The only positive association was a history of tinea capitis.

In 2019 the New England Journal of Medicine published the PADI3 genetics paper and wrote, in a parenthesis: “(The approximate prevalence of the disease is up to 5.6%.)” A photographic-scale finding became a disease prevalence in one bracket, in the most-cited journal in medicine, and it has been a disease prevalence ever since.

The competing estimates do not agree with it or with each other. Khumalo’s Cape Town community study, 874 adults examined directly, put CCCA at 1.9 percent overall and 2.7 percent in women. Kyei’s Cleveland study, 326 women recruited at two churches and a health fair, produced something closer to 16.5 percent. One point nine, five point six, sixteen point five. Three convenience samples, no biopsies, a ninefold spread.

There is no population-based, biopsy-confirmed prevalence figure for CCCA anywhere I could find.

The gene is real and it is not the answer

Malki and colleagues, NEJM 2019. Exome sequencing in 16 patients found one splice-site and three heterozygous missense mutations in PADI3 in five of them. Direct sequencing in a further 42 found variants in nine more. Combined: 14 of 58, or 24 percent.

Which means roughly three quarters of the CCCA patients in the study that discovered the gene did not have the gene.

PADI3 encodes an enzyme that modifies trichohyalin, a protein the hair shaft needs to form properly. The functional work is convincing — reduced expression, abnormal localization, decreased enzymatic activity, less PADI3 visible in patient scalp biopsies. The statistics are less so. The comparison against controls of African ancestry was post hoc, and once the analysis adjusted for people being related to each other, the p values moved to 0.03 and 0.04. That is a real finding sitting very close to the line.

The authors’ own framing is a gene-by-grooming interaction: the disease “usually manifests clinically after intense hair grooming.” Not a gene that acts alone.

Whether grooming causes it is genuinely unsettled, and the bigger studies say no

This is the part that gets people angry, and it should be handled carefully, because the answer is not the comfortable one in either direction.

Against a relaxer link: Olsen, 529 women, no association. Khumalo, 1,042 South African schoolchildren, 78.4 percent of the girls already chemically relaxing — and zero cases of CCCA identified. A 2023 review of fifteen studies found relaxers associated with CCCA in one study out of three that examined it.

For a link: a 2020 study of 74 cases against 96 matched controls reported that CCCA patients were 12.37 times more likely to have used relaxers. Treat that number with suspicion. An odds ratio of twelve for an exposure that half to two thirds of the source population has is not a plausible shape. The controls were women attending the same clinic for other kinds of hair loss, and no confidence intervals appear in the abstract.

What is not in dispute is that relaxers wreck the fiber. Khumalo’s amino acid analysis found cystine in the distal ends of relaxed hair had fallen to levels comparable with trichothiodystrophy — a congenital hair-fragility disease. That is a real and measurable harm, and it sits alongside a median relaxer pH of 12.36, above the threshold a workplace regulator calls corrosive. It is simply not the same claim as “relaxers cause scarring alopecia,” and the two largest datasets that looked for the second one came back negative.

The comorbidities keep failing to replicate

Type 2 diabetes is the association everybody repeats. In the Black Women’s Health Study, 5,389 respondents, the hazard ratio for severe central hair loss with type 2 diabetes was 1.68, rising to 2.05 with ten or more years of diabetes. Real signal — but self-reported hair loss on a photographic scale again, not biopsy.

Then a 2025 case-control study took 2,482 Black women with CCCA and propensity-matched them against 2,482 women with other scarring alopecias. Type 2 diabetes: 19.7 percent against 19.5, p = 0.59. Prediabetes, obesity, metabolic syndrome: no difference. Hypertension was lower in the CCCA group. The authors concluded CCCA was not associated with most metabolic disorders.

Uterine fibroids went the same way. The Johns Hopkins group reported an association in 2018. Their own follow-up, sequencing 16 CCCA scalps, found no shared gene expression between patients who had fibroids and patients who did not, and concluded there was no evidence of linked mechanobiology. The largest study using biopsy-proven cases — 427 against 1,281 matched controls — found no fibroid association at all. The defensible number is the pooled one from the 2026 meta-analysis of 23 studies: odds ratio 2.25, confidence interval 1.10 to 4.58. Not five times. Not a mechanism. An association that survives pooling and barely clears one.

And there is no trial

Standard care for CCCA is a topical steroid, intralesional triamcinolone, doxycycline and minoxidil. That regimen has never been tested in a completed, published, placebo-controlled randomized trial.

The one prospective multicenter study of exactly those four drugs is NCT04207931, target 250 patients, at Wake Forest. It opened on 30 April 2018. It is still recruiting. Eight years.

The only registered trial of a topical steroid in CCCA started in 2009 with a target of thirty patients. Its status on the registry is Unknown — meaning nobody has updated it since April 2010 and it has never reported.

The metformin result that went round the world in 2024 was a retrospective case series of twelve women, with transcriptomics on four of them, all of whom were already on other treatments. Nine of twelve reported less pain, itch or inflammation; six of twelve had some regrowth. That is a reason to run a trial. It was covered as a result.

A survey of dermatology residents and attendings at two academic programs put it plainly: no established standard of care exists.

What the delay costs

Here the data is good and it is bleak. A 2026 analysis of 1,034 scarring alopecia patients found the average time from first symptom to diagnosis was 3.54 years across all types — and 5.67 years for CCCA, against 2.88 for frontal fibrosing alopecia and lichen planopilaris.

Five and a half years, on a condition where the follicle is destroyed as you wait.

The reason it takes that long is in the same paper. CCCA mimics traction alopecia and seborrheic dermatitis. The histology overlaps with other lymphocytic scarring alopecias. Primary care awareness is limited. And the authors name, without softening it, distrust of healthcare and the perception that clinicians do not understand Afro-textured hair.

One more thing from that study that cuts against the way this condition is usually described: 14 percent of the CCCA patients were white, and they had a very long time to diagnosis too. Overwhelmingly a disease of Black women. Not exclusively.

The disease’s own name is part of the problem. A scoping review of 99 studies and 281 cases found documented patchy, occipital, parietal, frontal, temporal and trichorrhexis presentations, and concluded that the terminology “may not always be reflective of clinical presentation.” A condition called central and centrifugal, which frequently is neither, being looked for by clinicians who learned it as central and centrifugal.

Ask for the biopsy, by name

If the crown is thinning and the scalp there looks smoother than the rest, that is the appointment. Not next year. Trichoscopy distinguishes CCCA from traction alopecia and from seborrheic dermatitis — perifollicular halos rather than the fringe sign — and a biopsy settles it. Both are quick.

Ask by name. The five and a half years in that study were not spent doing nothing; they were spent being treated for something else.

Everything else in this piece — the gene, the relaxers, the diabetes, the fibroids, the metformin — is a field arguing with itself in public. None of it changes the one thing that is not in dispute, which is that scarred follicles do not grow hair and unscarred ones still can.

Which leaves the question of what surgery can do once the scarring is done. The flagship paper on transplantation into this condition reports two patients, with a condition attached that usually gets dropped — the piece on transplantation goes through it.

Two other things can produce a scarring patch that looks like this, and both are diagnosed rather than guessed. Scalp sarcoidosis is separated from the rest only on histology, and most of the published cases already had disease in the lungs or lymph nodes. And in adult women with a scaly scalp and thinning hair, a fungal culture is the cheap test almost nobody runs.

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.