The High Street Exemption

There is a date stamped on the gel-polish trade in Great Britain, and it falls on Valentine's Day. From the end of 14 February 2027, a bottle of UV gel polish containing the photoinitiator TPO can no longer be sold on a British shelf.¹ Until that day, it can. The same bottle came off shelves across the European Union seventeen and a half months earlier — on 1 September 2025. No grace period. No sell-through. No using up the stock already bought.²

One molecule, two shelves, two dates. The European Commission classed it too hazardous to keep selling. Great Britain will let it sit in a salon, and in your hand, for another year and a half. And there is nothing on the bottle — not the ingredient panel, not the "professional" line, not the words "dermatologically tested" — that tells you a neighbouring regulator pulled it from sale.

This is the kind of fact that arrives sounding like an alarm. It isn't one, quite. The truth is stranger and quieter than "banned chemical on British shelves," and getting it right means starting not with the regulation but with what the molecule actually does when the lamp comes on.

The molecule that destroys itself when it works

TPO is a photoinitiator. Its on-label name is Trimethylbenzoyl Diphenylphosphine Oxide — you would never recognise it as "TPO" scanning a bottle.³ Its whole job is to start a chemical reaction and be consumed by it. In the bottle it sits at somewhere between half a percent and five percent of the gel.⁴ Then the salon's UV or LED lamp comes on. The molecule absorbs the light, splits into reactive fragments, and those fragments trigger the chain reaction that turns a soft gel into a hard, glossy coating. In starting that reaction, the molecule is used up — its pieces locked into the cured polymer.

The European Union's own scientific committee measured what is left behind. Of the TPO applied to a nail, less than 0.2% is even theoretically extractable from the cured film — and the committee called that a worst-case figure. "TPO will be consumed rapidly during the polymerisation process," it wrote, and any trace remaining "will be trapped in the rapidly hardened polymer matrix."⁵ The thing being banned has, on a properly cured nail, largely ceased to exist by the time you leave the salon.

Could the small remainder cross into the body through the nail? Here the honest answer is that nobody knows, because nobody has measured it. The same committee was blunt about the gap in its own evidence: "There is no dermal absorption study or penetration study through the nail plate available for TPO," and "currently no standardized or validated test system exists to investigate penetration through the nail plate."⁶ What it could say is this. The nail plate behaves "more like a hydrogel than a lipophilic membrane" — it lets water-loving things through, not oily ones. And TPO is "a lipophilic substance and sparingly soluble in water," so the committee judged it "not a good candidate for penetrating the nail plate."⁶ So the wearer's uptake by this route is low to unmeasured: a barrier the molecule is poorly built to cross, sealing in a substance that is mostly gone.

This is the part the alarm version skips. It is also, for anyone who reads a label carefully, the part that matters most. It is the difference between "there is a banned chemical on my nails" and "there was a precursor in the bottle that the lamp turned into the coating."

"Banned in Brussels" is not "proven to harm you"

If the dose is so low, why is it banned at all? Because the European Union did not ban TPO on the strength of any finding that it was harming the people wearing it. It banned it on a classification.

In 2014, the EU's Scientific Committee on Consumer Safety assessed TPO and concluded it "is safe when used as a nail modelling product at a concentration of at maximum 5.0%," with a margin of safety it put at 1,515 — a wide one.⁷ That opinion was never overturned by new evidence of harm. What happened instead is bureaucratic and, once you see it, clarifying. Under the EU's Classification, Labelling and Packaging rules, TPO's hazard classification was upgraded. It became a category 1B reproductive toxicant — a substance presumed, largely on animal data, capable of damaging fertility. The upgrade came in a delegated regulation adopted on 19 October 2023, which took legal effect on 1 September 2025.⁸

From that moment, the EU's hazard-based machine took over, and it has no dimmer switch. The European Commission's own question-and-answer document on TPO states the rule plainly: "when the harmonised classification as CMR substance starts applying, the prohibition to use this substance must also apply from the same moment," and "for CMR 1A/1B substances, a ban is automatic unless an Article 15(2) derogation is granted."⁹ Nobody applied for a derogation for TPO before the ban took effect.⁹ The Commission was equally plain about the 2014 safe opinion: it "predates the new CMR classification, that upgraded the classification from Cat.2 to Cat.1. Once a CMR 1A/1B classification is adopted, earlier SCCS opinions no longer apply."⁹

Read that again, because it is the hinge of the whole story. The dose did not change. The exposure did not change. The <0.2% extractable from the cured film did not change. What changed was the legal class of the molecule — and the class, once changed, tripped an automatic ban regardless of the dose. The Commission's own guidance frames the prohibition as a direct consequence of that classification, not as a new finding of consumer harm: once a CMR 1A/1B class is adopted, the prohibition "must also apply from the same moment," and the earlier safety opinion "no longer" applies.¹⁰

This is worth being scrupulous about, because it is easy to have it both ways and dishonest to do so. A category 1B reproductive-toxicant classification is not nothing. It reflects real hazard properties in the underlying data, and it was serious enough to clear the bottle off every shelf in a 27-country market overnight. And it is not a measurement of what happens to the person wearing the cured film. Both things are true. The space between "serious enough to ban" and "not shown to harm you at this dose" is not a contradiction to paper over — it is the actual subject of this report.

It is also not an argument that the EU over-reacted or that Great Britain is being reckless. Two competent regulators can apply two defensible philosophies — the EU's precautionary, hazard-first approach; Britain's slower, evidence-led one — and reach the same destination by different routes. This report takes no side on which philosophy is right. The gap it is interested in sits underneath that whole debate, and it survives whichever way you come down on it.

The signal that didn't survive Brexit

Here is the gap. For decades, the most powerful safety claim any cosmetic ever made was one it never printed: its presence on a regulated shelf.

That sounds like a sentiment. It was a mechanism. The economist Anu Bradford named it the "Brussels Effect" — the EU's market is so large that manufacturers find it cheaper to make one formulation to the EU's standard and sell it everywhere than to run two.¹¹ For a British shopper before Brexit, this meant the EU's chemical floor was her floor too. It was built into the product upstream, in a formulation lab she would never see, on a label that would never mention it. She did not have to read the ingredient list against a prohibited-substances register, because the system was arranged so that she would never need to. When she picked a bottle off a British shelf, she was reading a signal she didn't know she was reading: this is for sale, therefore someone competent already checked.

The Food Standards Agency's own research describes exactly this reliance, in a different aisle. "Trust is the only way consumers can reduce the complexity and uncertainty they are faced with," it found, because "the average consumer cannot — even if they wanted to — substantiate whether a product contains the ingredients mentioned, or whether it has been produced according to the rules."¹² (That is food-system research, not cosmetics; the principle transfers, the specific numbers don't.) The point holds across the shop floor. People don't verify because they can't, and they don't have to, because the shelf does it for them. Trust is not laziness. It is the rational substitute for a check no individual can perform.

Then, on a dated, substance-by-substance basis, that substitute quietly stopped being reliable for Great Britain — and only for Great Britain. This is the precision the headline version gets wrong. Northern Ireland still follows EU cosmetics rules under the Windsor Framework, so Northern Ireland banned TPO on the EU's timeline, in September 2025. It is England, Scotland and Wales — Great Britain — that diverged.¹ Britain runs its own retained version of the cosmetics regulation. It amends its own annexes by its own statutory instruments, on the advice of its own scientific group, and does not automatically inherit the EU's changes.¹³ So when the EU removed TPO in September 2025, nothing happened in Great Britain — until a separate instrument, SI 2026/23, signed on 12 January 2026, re-enacted the ban with its own, later clock.¹

Picture the moment it bites. It is a Saturday in 2026. The bottle on the British counter says "professional" and "dermatologically tested." Eight hundred miles away, on a shelf in Dublin or Rotterdam, the identical bottle was pulled from sale a year ago — and behind the counter in Belfast, too. The British shopper has no way to know this. There is no label, no marking, no point-of-sale flag anywhere in the regime that says this product is on one side of a regulatory border its neighbours are on the other side of. The campaign group CHEM Trust, watching Great Britain fall behind on chemical controls, warned of "potential public outcry if GB starts becoming a dumping ground for substances or products that no longer comply with higher EU standards."¹⁴ But there is no error message. The floor she stood on moved, and the bottle in her hand looks exactly as it always did.

And here is the part that should be felt rather than merely noted: the reassurance words got louder over the same years the guarantee behind them got quieter. "Clean." "Professional." "Dermatologically tested." None of these is defined in law to mean anything about reproductive-toxicant content. The EU's own common criteria for cosmetic claims — retained in GB law — set out six principles for honest claims, but not one of them defines "clean" or "natural" or guarantees the absence of a classified substance.¹⁵ "Dermatologically tested" means a dermatologist supervised some test. It does not tell you what the test found, and it tests for skin irritation, not for what is locked in the polymer. The watchdog most often invoked here, the Competition and Markets Authority's Green Claims Code, polices environmental claims — "eco," "sustainable," "recyclable" — not the content of an ingredient list.¹⁶ So the words that occupy the slot where the safety signal used to live carry none of it. The lexicon thrived while the floor it implied dissolved.

Two lawful bottles, indistinguishable

There is a reason the EU-removed bottle is still on a British shelf rather than in a skip, and it is structural rather than sinister.

The EU instrument that banned TPO contains no sell-through provision for category 1A/1B substances. As the Commission put it: "From 1 September 2025 onwards, all placing on the market and making available in the course of commercial activity must stop, even for products already in stock or purchased before that date."¹⁷ On that morning, every bottle of TPO gel in the EU supply chain became dead stock — value, written off. Great Britain's instrument does the opposite. It sets a placing-on-market deadline of 15 August 2026, then a further window in which products already on the market "may continue to be made available" to the end of 14 February 2027.¹ (Industry guidance from the CTPA reads the statute as still permitting professionals to use stock they had already bought after that date; the instrument itself sets only the placing-on-market and making-available deadlines.¹)

That asymmetry is the whole economic engine, and it needs no euro figures to make its point. Value embedded in already-made TPO stock — destroyed in the EU — can be recovered through sale in Great Britain. The window is lawful, normal and proportionate; sell-through periods exist precisely so that saleable stock isn't needlessly burned. Naming it is not an accusation. But it produces a strange result on the shelf. A salon may now stock two gels side by side. One comes from a brand that reformulated early and ate the cost — the company GelBottle Inc, for instance, moved its range to a TPO-free alternative called TPO-L ahead of the British deadline, telling customers plainly that "current products remain safe to use" and "this is not a product recall."¹⁸ The other comes from a brand lawfully selling the retired TPO generation down the window. Both are legal. Both look identical. The only thing that distinguishes them is the ingredient list — which most people cannot decode against a prohibited list they have never seen.

Where the dose actually is

If the wearer's exposure to TPO is low to unmeasured, the natural question is whether anyone is exposed at all in a way that's documented. The answer is yes — but not the person the "consumer safety" framing centres.

The exposure that the nail-product literature actually documents does not run through the sealed, cured film on a careful client's nail. It runs through uncured product at the cuticle and the sides of the nail during application. The committee that assessed TPO noted that "without due care, there could be accidental skin contact at the cuticle and the side of the nails."¹⁹ And, structurally, it runs through the technician, who meets that uncured product not once but across thousands of applications. The documented hazard at that contact point is not the reproductive classification at all. It is skin sensitisation: TPO is classed a "moderate skin sensitizer,"²⁰ and the larger occupational problem is the family of uncured methacrylate monomers in the gel, which are potent contact allergens when polymerisation is incomplete.

This is well-mapped in named populations, and the mapping is consistent. A five-year study in Contact Dermatitis (Raposo and colleagues, 2017) examined 230 cases of allergic contact dermatitis from methacrylates in nail products — 55 nail technicians, 56 consumers, and 119 with mixed exposure — and found that the screening allergen, 2-hydroxyethyl methacrylate, identified 90% of cases.²¹ The exposure concentrates where uncured product meets repeatedly-exposed skin, and the people most exposed are the ones applying it for a living.

The discipline here is to keep two different hazards in two different rooms. The reproductive classification is what removed the bottle from the EU shelf. The technician's documented problem is contact allergy from uncured monomers at the cuticle — a different mechanism, a different part of the body, a different evidence base. As of mid-2026 I found no published biomonitoring study measuring systemic TPO uptake in salon workers at a level tied to a reproductive-risk threshold; environmental and hand-contact studies of salon photoinitiators are only now emerging, and that gap belongs on the page, not behind it. What can be said honestly is narrower and still real. The documented exposure pathway concentrates on the technician, repeatedly, all day. Not the documented harm — the documented exposure. And there is an unfortunate symmetry here: the professional setting that makes this exposure routine is also the strongest reason nobody asks about it. A licensed, skilled, uniformed setting reads as a safety endorsement, and authority cues suppress the very question that would close the gap.

This is not a one-off molecule, either. TPO is the type specimen of a standing process: a substance gets a harmonised hazard classification, the cosmetics regulation presumptively bans it, and a batch sweeps off EU shelves on a single date. The same instrument that took TPO also removed a nail-gel adhesion promoter, dimethyltolylamine, classed a carcinogen, on the same day.¹⁷ The contrast that proves the point runs the other way. Triphenyl phosphate, a nail-polish plasticiser easy to confuse with TPO, was banned by a separate EU instrument on endocrine grounds — and unlike TPO it has measured body uptake. A Duke University study found a urinary metabolite rising nearly sevenfold within hours of painting nails.²² Two nail chemicals, both "EU-banned," opposite profiles: one demonstrably enters the body, one seals itself into the coating. "Banned" is not one thing, and treating it as one thing is the imprecision this whole report exists to dismantle.

What would change this reading

The honest version of the opposing case is strong, and it deserves stating at full strength. It runs like this. Two functioning regulators assessed the same molecule and reached the same conclusion — ban it — by two legitimate routes and two reasonable timetables. The science on the substance is reassuring: consumed in its own cure, under 0.2% extractable, no measured nail penetration, a margin of safety the industry's science director put a thousandfold above any fertility-effect threshold.²³ Great Britain converges on the same bans and acts on its own evidence when it sees cause — the same instrument that ran seventeen months behind the EU on TPO also banned a UV filter, 4-MBC, on the advice of Britain's own scientific group, flagging endocrine and genotoxicity concerns.²⁴ So what is left but ordinary regulatory lag over a low-risk substance? A transition window is the most routine mechanic in product law. Demanding a "this differs from the EU" label at every till would be disproportionate for a divergence that, on this molecule, makes no measurable difference to the buyer.

That case is reasonable, and it is right about the substance. What it under-weights is one thing: the shelf was never a neutral fact. It was an inherited, relied-upon promise — a check the consumer could not perform, performed for her upstream, invisibly. A promise that fails silently fails differently from one that was never made. The lag is not the gap. The lag plus the silence is the gap.

So what would actually change this reading? A concrete thing. If Britain's product-safety office or the cosmetics trade body introduced a point-of-sale or labelling signal distinguishing divergence-status products, the missing-signal gap would close, and this report would say so plainly. On the substance, the conclusion would move if any of three studies appeared: one that finally measured TPO crossing the intact nail plate at salon dose; one that showed systemic uptake in salon air or technician biomonitoring above the industry's stated effect-margin; or one showing that under-curing — a too-short lamp cycle, an ageing bulb — leaves materially more than the worst-case 0.2% on the wearer's nail. None of those studies currently exists. The wearer's residual dose is low. Great Britain is converging, just more slowly, and on its own clock. The thing that has not closed is the signal.

What to do on Saturday

The next time you book a gel manicure, or pick up a "clean / professional / dermatologically tested" gel polish, there is one durable check and one question, and they aim at the same place.

The check is the ingredient list. Look for Trimethylbenzoyl Diphenylphosphine Oxide — or its chemical-name twin, Diphenyl(2,4,6-trimethylbenzoyl)phosphine oxide, or the number CAS 75980-60-8. If it is on the bottle, that bottle is lawful on a British shelf but is the EU-removed generation. You can cross-check any ingredient string for free against the EU's public CosIng database and its prohibited-substances list — the one reliable signal, available to anyone, that no reassurance word can substitute for.

The question is for the salon: which photoinitiator does your gel use — TPO, or TPO-L? That single question lands exactly where chemistry, dose and the trust reflex converge — and it is also where the documented exposure actually sits. Salons that switched early answer it readily. A few brands made the switch ahead of the deadline; GelBottle Inc is one documented example, named here as an early reformulator, not as an endorsement. The durable move is to check the ingredient list yourself, whatever the brand.

Be clear-eyed about what this check buys you, though, because the honest answer is not "it protects you from harm." On the available science, the cured film on your nail is largely benign. What the check buys you is the signal the shelf used to carry and quietly stopped carrying: the ability to know which side of the divergence you are standing on, and to make your own judgment instead of having a moved floor make it for you invisibly. That is agency, not danger-avoidance.

And the real fix is not something you can buy. No bottle on the shelf carries a divergence-status flag, because the regime does not require one. The durable answer is a point-of-sale signal — a small mark, a line on the panel, anything that tells a shopper in Reading what a shopper in Belfast already knows. Until that exists, the ingredient list is the interim lever, and the one name on it that means something specific is the one you now know to look for.