The Rule Has Been Coming for Three Years and Has Never Been Written

There is a rule that would ban formaldehyde and formaldehyde-releasing chemicals in hair smoothing and straightening products in the United States. It has a regulation identifier, 0910-AI83. It has an agency, a CFR citation, a legal authority, and a named policy analyst with a phone number.

It has never been published. Not finalized, not withdrawn, not even proposed. A Federal Register API query on that identifier, run on 13 September 2026, returns a count of zero.

Which also means — and this matters if you are trying to follow it — there is no docket number. A regulations.gov docket comes into existence when a document publishes. Nothing has published. Any docket number you see attached to this rule in press coverage belongs to something else.

Five agendas, twenty-eight months backwards

The Unified Agenda is where federal agencies state when they intend to act. Here is what 0910-AI83 has promised, edition by edition.

Fall 2023: proposed rule in July 2024. Spring 2024: September 2024. Fall 2024: March 2025. Spring 2025: December 2025. The 2026 edition: November 2026.

Five consecutive targets, each one set after the previous one was missed, moving the date back by about 28 months in total while the text of the rule has never once appeared.

The 2026 entry still reads “Proposed Rule Stage” and “Previously published in the Unified Agenda.” The only substantive change across the five editions is that the abstract stopped saying “This proposed rule would prohibit” and started saying “This rule would prohibit,” which is a copy edit rather than a stage.

Worth knowing for context: under the current deregulatory executive order the rule carries a “Regulatory” designation, meaning it counts as burden-adding. Rules in that column do not move quickly.

What FDA has actually done in the fifteen years since this became an issue is two warning letters. Brazilian Blowout in 2011 and Van Tibolli Beauty Corp in 2015. Both companies took steps to correct. The agency’s consumer page says it continues to monitor.

What the air measurements show

The best exposure work is a 2011 simulation study: four brands — Brazilian Blowout, Coppola, Global Keratin, La Brasiliana — applied to human hair wigs on mannequin heads, with 88 personal, area and clearance air samples and real-time tracer gas to measure air exchange.

Airborne formaldehyde during blow-drying: mean concentrations from 0.08 to 3.47 parts per million. During flat-ironing: 0.08 to 1.05. Across the salon during treatment: 0.02 to 1.19.

Estimated eight-hour averages ran from 0.02 ppm for one La Brasiliana treatment a day to 0.44–0.75 ppm for four Brazilian Blowout treatments a day.

For scale, the OSHA permissible exposure limit for formaldehyde is 0.75 ppm over eight hours, with a short-term limit of 2 ppm over fifteen minutes and an action level of 0.5 ppm. Four treatments a day with the highest-content product reaches the eight-hour limit. Blow-dry peaks of 3.47 ppm are above the fifteen-minute ceiling.

Bulk formaldehyde in the products: Brazilian Blowout 11.5 percent, Global Keratin 8.3, Coppola 3.0, La Brasiliana zero. The authors’ conclusion names the loophole directly: professional hair smoothing treatments, even those labeled formaldehyde-free, have the potential to produce concentrations that meet or exceed occupational limits.

In real salons the numbers are lower and the finding is stranger. A study of 23 salons in Bauru, Brazil found product content of 3 to 11 percent — up to 54 times the 0.2 percent national limit. A study of 50 workers across six Brazilian salons measured 0.02 to 0.04 ppm in the two salons using formaldehyde-free products and 0.07 to 0.16 ppm in the four using formaldehyde-containing ones, and found significant differences between salons in urinary formic acid, in comet assay DNA damage index and frequency, and in karyorrhexis on the micronucleus test.

Those air concentrations are well below the OSHA limit. The DNA damage signals are still there. Fifty people, cross-sectional, so this is a signal rather than a verdict — but it is the kind of signal that usually prompts a larger study rather than a fifth delay.

Formaldehyde-free, tested

Maneli, Smith and Khumalo identified every Brazilian-keratin-type product on the South African market in 2012 — seven brands — and tested every component of each one three times by HPLC.

Six of the seven measured between 0.96 and 1.4 percent formaldehyde, against the 0.2 percent maximum the US Cosmetic Ingredient Review panel considers safe. Roughly five times over.

Five of those six were labeled formaldehyde-free.

The mechanism of the loophole is not a lie, exactly. The active species is methylene glycol — formaldehyde dissolved in water, chemically a different compound on an ingredient list, which releases gaseous formaldehyde when a flat iron hits it at 200 degrees. A product can be honestly labeled formaldehyde-free and still be a formaldehyde delivery device the moment it is used as directed. That is precisely why the unwritten FDA rule names both formaldehyde and formaldehyde-releasing chemicals, and restricts itself to products “applied to the hair and heated under their conditions of use.”

The cancer evidence, and the part coverage keeps getting wrong

The Sister Study, 2022. 33,947 women aged 35 to 74 with a uterus at enrollment, followed an average of 10.9 years, 378 uterine cancers.

Ever-use of straightening products in the previous twelve months: hazard ratio 1.80, confidence interval 1.12 to 2.88. Frequent use, more than four times in the year: 2.55, interval 1.46 to 4.45, trend p = 0.002. Hair dyes: nothing. Perms: nothing. In absolute terms, risk by age 70 went from 1.64 percent in never-users to 4.05 percent in frequent users.

Now the nuance almost every retelling dropped. About 60 percent of the straightener users in that study were Black women — and the association did not differ significantly by race.

The finding is not that these products are more carcinogenic in Black women. It is that Black women use them far more, and start younger, so the same relative risk falls on more people. Those are different claims and only one of them is in the paper.

Then the replication attempt, which is the larger study. The Black Women’s Health Study followed 44,798 women from 1997 to 2019 with actual lifetime exposure data rather than a single twelve-month recall. 347 uterine cancers. Heavy use against never or light use: hazard ratio 1.18, interval 0.81 to 1.71. Null.

In postmenopausal women only it turned positive — 1.60, 1.64 and 1.71 for moderate use, heavy use and twenty-plus years of use — with every lower bound sitting at 1.01 to 1.08. In premenopausal women the results were null.

So the bigger, all-Black, longer, better-exposure-data cohort found nothing overall and a fragile subgroup signal. That is a materially weaker picture than 2.55, and it is the one that should be quoted alongside it.

The fibroid follow-up is stranger still. In the Sister Study, straightener use between ages 10 and 13 was associated with young-onset fibroids in non-Hispanic white women — odds ratio 1.23, significant — despite use of 5 percent or less in that group. In Black women, over 70 percent of whom used straighteners, every estimate was null.

An exposure that shows an effect in the group barely exposed and none in the group overwhelmingly exposed is the opposite of a dose-response. The authors flag potential misclassification themselves and write that use “may be positively associated.” May be is doing the work.

And the newest Sister Study extension, looking at non-reproductive cancers, reported hazard ratios of 1.71 for thyroid cancer and 2.66 for pancreatic, on 225 and 138 cases — while stating in its own results that there was little evidence of dose response with increasing frequency of use.

Where that leaves it

The occupational exposure is documented and it is the strongest part of the file. People working eight-hour shifts in small rooms with flat irons and these products are breathing measurable formaldehyde, sometimes at the legal limit, and the DNA damage signals show up below it. If you work in a salon that does keratin treatments, the ventilation question is not fussy.

The consumer cancer evidence is real, contested, and smaller than the headlines. One cohort found a strong association on a single twelve-month recall. A larger one with lifetime exposure did not replicate it overall. The dose-response is missing in three separate analyses. That is a live scientific question, not a settled harm — and treating it as settled makes it easier, not harder, to dismiss when the replication lands badly.

What is not contested: five of six products labeled formaldehyde-free tested over the limit, the labeling loophole is structural rather than accidental, and the rule written to close it has been eighteen months away for three years running.

The reformulated products are moving to glyoxylic acid, and the independent literature there is already filling up with acute kidney injury case reports. The industry safety paper concluding otherwise carries two published corrigenda. That is the next file, and it is opening the way this one did.

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.