Glycerin Isn’t Sealing Your Teeth: The 27-Rinse Myth, the 2007 Diethylene Glycol Scare and the Test That Actually Matters

If your feed looks like ours this month, you have seen the claim. Glycerin, the clear syrupy ingredient in almost every tube of toothpaste, supposedly coats your teeth so thickly that it takes 27 rinses to come off, and while it is there, minerals cannot get back into the enamel. Google searches for “glycerin toothpaste” were flat all summer, hit a sharp peak on 3 September 2026, decayed for a week, and started climbing again on 10 September. “Glycerin-free toothpaste” is rising alongside it.

We sell glycerin by the quart and by the drum, in USP, technical and food-grade solutions, so this is not a story we can watch from a distance. What follows is what the molecule actually does, where the 27-rinse number came from, and the part of the glycerin story that genuinely deserved the alarm: a June 2007 morning when the U.S. Food and Drug Administration told the country to throw away toothpaste, because what was in the tube was not glycerin at all.

Where the 27-rinse number comes from

The claim is older than the platform it is spreading on. It traces to Gerard F. Judd, who described himself as a chemistry Ph.D. and self-published a book called Good Teeth, Birth to Death in the late 1990s, followed by letters to readers. The line that survives is that glycerin in toothpaste is so sticky it “takes 27 washes to get it off,” and that teeth under it cannot re-mineralise. Neither the book nor the letters attach a study to the number, and the sites that have gone looking for one since, including the natural-living blogs that would most like to find it, report that there is none.

That matters more than it sounds. Twenty-seven is a strangely precise figure for a claim with no measurement behind it, and precision is what makes it travel: it sounds like the output of an experiment. The honest position, and the one we would hold ourselves to on a Certificate of Analysis, is that a number without a method is not a result. It is a sentence.

Dental researchers have not run a large controlled trial on the question either way, and this article is not dental advice; if you have a question about your own teeth, ask a dentist. What we can do is the chemistry, because the chemistry does not need a trial. It needs a look at the molecule.

What glycerin actually is

Glycerin is the common name for glycerol, propane-1,2,3-triol: a three-carbon chain with a hydroxyl group on every carbon. Formula C₃H₈O₃, molecular weight 92.09, CAS 56-81-5. Those three hydroxyl groups are the whole personality of the molecule. Each one hydrogen-bonds to water, and to the other glycerin molecules around it, which is why it behaves the way it does on the bench.

The viscosity row is the one that fuels the myth. Glycerin feels thick and slippery between the fingers, so it is easy to imagine it as a varnish. But viscosity is a property of the pure liquid, not of the mixture. The moment glycerin meets water it starts dissolving into it, and a 70% glycerin solution has a viscosity of about 17 cP, down from 954. Add more water and it thins to nothing. The hydroxyl groups that make pure glycerin cling to itself are the same groups that make it let go the instant there is water to hold instead.

Why it is in toothpaste at all

Toothpaste is a wet paste that has to stay a wet paste, in a tube, in a warm bathroom, for a year. Water alone would evaporate through the cap and leave a brick. Glycerin is the humectant: Crest’s own explainer describes it as the ingredient that “helps retain the moisture of the paste so it doesn’t dry out in the tube.” It also sweetens slightly, carries flavour, and gives the paste its slip. The same property puts it in skin lotions, e-liquids, cough syrups, hand sanitiser gels, and the fog fluid we wrote about for Halloween season: everywhere a formulator needs something to hold on to water.

Why a rinse-proof film is the wrong picture

Imagine the coating the claim describes: a layer of glycerin on enamel that survives 27 separate floods of water, in a mouth that produces saliva continuously. For that to happen, glycerin would need to prefer the tooth surface to water so strongly that water could not pull it off. Nothing in its data sheet supports that. It is miscible with water in all proportions; it has no oily, water-repelling part of the molecule; it does not react with hydroxyapatite; and the one thing it does do with water is dissolve into it, eagerly, in either direction.

Compare the things that genuinely do coat teeth and resist rinsing: the pellicle your own saliva lays down, plaque, and the tars in tobacco smoke. Those are proteins, biofilms and hydrophobic residues, which is to say the opposite of a small polar sugar alcohol. If glycerin behaved like them, it would be useless in a lotion, because it would never release the moisture it holds. Formulators use it precisely because it does.

None of this says anything about whether a given toothpaste works. That is a question for the fluoride or hydroxyapatite it carries, and for the dentist. It says only that the mechanism the viral claim proposes is not how this molecule behaves in water. And it leaves the more interesting question standing: if glycerin is not the villain, why does “glycerin in toothpaste” still make people uneasy? Because once, it should have.

The one time glycerin in toothpaste was dangerous, it wasn’t glycerin

On 1 June 2007, FDA issued a public advisory with a blunt instruction: consumers should examine toothpaste for labelling that says it was made in China and, out of an abundance of caution, throw it away. The agency had detained one shipment at the border containing “about 3 percent DEG by weight,” then found and tested tubes at a distribution centre and a retail store; “the highest level found was between 3–4 percent by weight.” The retail product was not labelled as containing DEG. Fourteen brands went on an import alert the same day.

DEG is diethylene glycol, C₄H₁₀O₃: an industrial solvent and antifreeze ingredient that happens to be a colourless, hygroscopic, viscous liquid. Set a vial of it next to a vial of glycerin and you cannot tell them apart by eye. Somewhere upstream of those tubes, someone had not tried very hard to. FDA said what had triggered its sampling: “reports of contaminated toothpaste from China found in several countries, including Panama.”

Every glycerin poisoning FDA reviewed shares three gaps

FDA’s Guidance for Industry on testing glycerin for diethylene glycol was first issued in May 2007, in the middle of the toothpaste episode, and revised in May 2023 after a new wave of poisonings. Its background section is a history of the same accident repeating, and it is worth quoting at length because the pattern is the point.

We wrote up the 1937 case in full earlier this year, in Elixir Sulfanilamide: The 1937 Poisoning That Created the FDA. The 2022–23 wave, which the World Health Organization described as involving “diethylene glycol (DEG) and ethylene glycol (EG)”, “toxic chemicals used as industrial solvents and antifreeze agents,” is the reason FDA reopened the guidance. And a 2013 paper in Clinical Toxicology by CDC investigators of the Panama outbreak counts “more than 13 medication-associated, DEG mass poisonings” since 1937, “associated with more than 500 deaths.”

What FDA does next is the part nobody on social media is talking about. Having listed the incidents, the guidance lists what they have in common:

• The manufacturers “did not perform full identity testing on the glycerin raw material, including tests to quantify the amount of DEG present and to verify the purity of the glycerin received.”
• They “relied on the certificate of analysis (COA) provided by the supplier of the glycerin.”
• “The origin of the glycerin was not readily apparent from the COA. The COA obtained by the manufacturers of the liquid drug products was often a copy of a COA on the letterhead of the distributor from whom they had purchased the glycerin and not the COA provided by the original manufacturer of the glycerin. The chain of custody or distribution history of the glycerin was also not readily known, often because the glycerin might have been sold multiple times between its manufacture and its use.”

Read those three bullets as a distributor and they sting, because two of the three are about paperwork that passes through hands like ours. A certificate of analysis is a claim made by whoever signed it. Copied onto a second letterhead, it is the same claim with one less name attached. The 2022 wave, FDA notes, “resembles previous ones” in exactly this respect, down to a case where “the appearance of the label and COA of propylene glycol… suggested the component container’s content might differ from what the container label and COA stated.”

Colourless, viscous, hygroscopic, sweet enough to formulate into a children’s syrup: the reason DEG keeps getting sold as glycerin is that it looks like glycerin. And the reason it keeps getting through is that a CoA looks like a test.

The fix was an identity test with a limit in it

The regulatory answer to a substitute that passes by eye is a test that does not use eyes. Under 21 CFR 211.84, a drug manufacturer must run an identity test on each component before use, using the specific compendial test if one exists, on representative samples of each shipment of each lot. The United States Pharmacopeia rewrote its glycerin monograph in 2008 so that the identity test would catch DEG on its own: the revision became official on 15 May 2008 and replaced the old limit test with a gas chromatography procedure that quantifies diethylene glycol and ethylene glycol individually.

FDA’s 2023 guidance describes the result as a three-part identification. Test A is infrared absorption, which confirms that the material is glycerin but, in FDA’s words, “is not suitable for detection or quantitation of DEG or EG.” Test B is gas chromatography referencing the “Limit of Diethylene Glycol and Ethylene Glycol,” and “allows for quantitation of DEG and EG, if present, individually, down to the identified safety limit (NMT 0.10%).” Test C is an examination of the Test B chromatograms. The limit, not more than 0.10% of each, sits inside the identity test rather than in a separate purity section, which means a lot of glycerin that fails it is not merely impure. It is, for regulatory purposes, not glycerin.

Two more details from the guidance are worth carrying around. First, FDA now recommends that the representative sample “is of each container of each lot”, not a sample of the lot, because the agency “has seen wide variability of DEG and EG contamination from container to container”; in the 2022 Indonesian case, nine drum samples from the same supply came back with widely different levels over the limit. Second, FDA published its own gas chromatography–mass spectrometry screening method for DEG in toothpaste in 2007, so the analysis is not exotic. Any contract laboratory with a GC can run it.

What this means if you buy glycerin from us

We carry glycerin in three grades, and the honest thing to say first is that the grade name tells you what the material is sold to, not what you have tested it as.

Need the Certificate of Analysis for a lot? Just ask; we send it over at no charge. And then, if you are making a drug product, do what the guidance says and test it anyway. That is not us being modest about our paperwork. It is 21 CFR 211.84, and it exists because for eighty-six years the paperwork has occasionally been the problem. If you have never read a CoA critically, our guide to how to read a chemical Certificate of Analysis walks through what each line is and is not promising, and Understanding Chemical Grades covers what USP, technical and food-grade actually mean.

Common questions

Does glycerin in toothpaste really coat your teeth for 27 rinses?

There is no published measurement behind the number. It traces to a self-published book and letters by Gerard Judd from the late 1990s, which attach no study to the claim. Chemically, glycerin is miscible with water in all proportions and hygroscopic, so a film that resists 27 rinses is not how the molecule behaves; the property that makes it useful as a humectant is that water can always take it back. Whether a particular toothpaste works is a question for a dentist, not for this article.

Why is glycerin in toothpaste in the first place?

It is the humectant. Toothpaste has to stay a moist paste for a year in a tube, and water alone would evaporate through the cap. Glycerin holds moisture, adds slip and a slight sweetness, and carries flavour. The same property is why it is used in lotions, e-liquids, syrups and fog fluid.

What happened with toothpaste and diethylene glycol in 2007?

On 1 June 2007 FDA advised consumers to discard toothpaste labelled as made in China and issued an import alert on fourteen brands, after detaining a shipment containing about 3% diethylene glycol by weight and finding retail tubes at 3–4%. DEG is an antifreeze ingredient that had been substituted for glycerin upstream. FDA reported no U.S. poisonings from the toothpaste; the concern was chronic exposure and accidental swallowing, especially by children.

What is diethylene glycol and why is it dangerous?

Diethylene glycol (C₄H₁₀O₃) is an industrial solvent and antifreeze ingredient. It is a colourless, viscous, hygroscopic liquid that looks like glycerin, which is why it has repeatedly been sold as glycerin. WHO describes DEG and ethylene glycol as toxic chemicals that can be fatal even in small amounts; FDA’s guidance records outbreaks in 1937, Haiti in 1995–96, Panama in 2006 and at least seven countries in 2022–23, with more than 300 deaths in the most recent wave.

What does the USP glycerin identity test check?

Three things. Test A, infrared absorption, confirms the material is glycerin. Test B, gas chromatography, quantifies diethylene glycol and ethylene glycol individually against a limit of not more than 0.10% each. Test C examines the Test B chromatograms. Because the limit sits inside the identification test, glycerin that fails it does not meet the monograph’s identity, not merely its purity.

Which glycerin grade should a formulator buy?

USP Grade for anything that touches a drug product, a cosmetic or food contact, because it is sold to the monograph that carries the DEG and EG limit. Technical Grade for antifreeze blends, adhesives, fog fluid and other industrial uses. Pre-diluted food-grade solutions at 75%, 50% and 25% where the process would otherwise be diluting the 99% material on site. In every case, ask for the Certificate of Analysis, and if the material is going into a drug product, run the identity test on receipt as 21 CFR 211.84 requires.