How Safe Is a Gel Manicure, Really?

Most coverage of gel manicure safety sorts the risks by how alarming they sound. Let’s sort them by how much human evidence sits behind them instead. The order changes completely, and the ingredient that just got banned across Europe barely makes the list.

Here’s the ranking, weakest evidence first.

Fourth: the UV lamp

This is the fear that launched a thousand posts, and it rests almost entirely on one paper.

In January 2023, Zhivagui and colleagues published a study in Nature Communications on cells irradiated by a nail polish dryer. They found oxidative DNA damage, mitochondrial dysfunction, and mutational signatures in three cell lines. The press translated this as gel manicures causing cancer.

What the study actually was: cells in a dish. Mouse fibroblasts, human foreskin fibroblasts, and keratinocytes in monolayer culture, with no stratum corneum above them, no melanin, and no repair-competent tissue architecture. The exposure was two to three sessions of twenty minutes. A real gel cure runs thirty to sixty seconds per coat. One analysis put the delivered dose at roughly 9 J/cm², which is not in the same universe as a manicure.

The finding that got almost no coverage is the one that matters most. The study found no cyclobutane pyrimidine dimers and no 6-4 photoproducts. Those are the canonical direct UV lesions that dominate the mutational landscape of real skin cancers. What it found instead were oxidative signatures that, as the authors themselves note, “generate low numbers of mutations in skin cancer.”

Two months after publication, the authors issued a correction to their own abstract, softening the claim that the effect on mammalian cells was previously unevaluated to a statement that it “remains unclear.”

Now the dose modelling. Measured emissions put a standard three-cycle LED session at about 1.87 J/cm², equivalent to roughly three and a half minutes of noon summer sun. One analysis calculated you’d need over 250 years of weekly sessions to match the exposure of a single course of narrowband UVB phototherapy. A 2014 JAMA Dermatology study measuring seventeen devices from sixteen salons concluded the carcinogenesis risk “remains small,” and a 2025 scoping review of the entire literature found cancer risks from typical use “acceptably low.”

Epidemiology, the study design that would actually answer this: there isn’t any. Not one cohort study, not one case-control study, after roughly two decades of mass exposure. What exists is a handful of case reports in the whole world literature, starting with two patients described in 2009 whose own authors wrote that “no strong conclusions can be made from this limited case series.” One of those two patients had used UV nail lights roughly eight times, all within a single year.

One caveat that’s genuinely useful. That 2014 study found lamp irradiance varying twenty-six-fold across salons, from 0.6 to 15.7 mW/cm², tracking bulb wattage. “A nail lamp” is not one exposure. Yours might be at either end.

The American Academy of Dermatology still recommends sunscreen or fingerless UV gloves before your appointment. That’s reasonable advice, not because the risk is established but because the mitigation costs nothing. Cheap insurance on an uncertain bet is just good sense.

Third: the nail plate

Your nails do get thinner and more brittle. This part is real, measurable, and mostly not the polish’s fault.

A 2026 review put numbers on it. Nail thickness dropped from 0.50 mm to 0.46 mm over six months of regular gel use, about eight percent. More interestingly, cysteine content fell 22% and methionine fell 36.5%. Those are the sulphur-bearing amino acids whose disulphide bridges give the nail plate its strength, so there’s a real biochemical mechanism here, not just mechanical wear. In a survey of over 2,000 users, 84.5% reported some side effect, most commonly brittleness and splitting.

But the dermatology literature is consistent about where the damage comes from, and it’s the removal. Three separate mechanisms, worth keeping apart:

Mechanical. The e-file drill is described as the most destructive removal method there is. Over-filing physically strips layers of nail plate. This is entirely operator-dependent and it’s the single largest controllable factor in the whole article.

Chemical. Long acetone soaks dehydrate and delipidate the plate, producing brittleness and that chalky white surface granulation. This is why the AAD tells you to soak fingertips only, or use acetone-soaked cotton wrapped in foil, rather than dunking your whole hands.

Allergic. Onycholysis and nail dystrophy driven by acrylate allergy, which is a completely different problem and gets confused with the other two constantly.

Which brings us to the actual story.

Second: HEMA, the one with an epidemic named after it

Hydroxyethyl methacrylate and its cousin di-HEMA trimethylhexyl dicarbamate are the acrylate monomers that make gel adhere and cure. They are also potent contact allergens, and this is the only risk in this article with a substantial human evidence base.

In August 2018 the British Association of Dermatologists issued a warning about a “UK artificial nail allergy epidemic.” It rested on a prospective audit across thirteen dermatology units in the UK and Ireland: of 4,931 patients patch-tested, 2.4% reacted to at least one methacrylate. North American data from 2019 to 2020 found 3.2% positivity to HEMA, a statistically significant increase over the previous decade. European multicentre figures range from 1.1% in a 2013 to 2015 cohort of 18,228 patients up to 3.7% in a Spanish series.

Those are patch-tested dermatology patients, an enriched population, so don’t read them as population prevalence. But the trend line is unambiguous and the mechanism is well understood.

Three things make this materially worse than a rash.

It’s a type IV hypersensitivity, which is conventionally permanent. You don’t grow out of it.

It cross-reacts. In one series, 66.3% of acrylate-allergic patients reacted to two or more acrylates. Sensitised patients have gone on to need cementless joint replacement because acrylic bone cement triggered prosthesis failure, and there’s a documented case of exactly that following acrylic nails. Dental acrylics are the other common collision, with reported oral lichenoid lesions and denture stomatitis. The BAD’s warning about “lifelong consequences for dental treatments and surgeries” is not hyperbole.

Gloves don’t reliably solve it. Methacrylates penetrate most glove materials, and methyl methacrylate gets through latex and vinyl within minutes. BAD’s Dr David Orton put it flatly: “wearing protective gloves is not enough.”

And here’s where the transatlantic gap is sharpest. Europe restricted HEMA in 2021 under Commission Regulation (EU) 2020/1682: professional use only, with mandatory “For professional use only” and “Can cause an allergic reaction” warnings. Great Britain carries the same restriction.

Note the word. Restricted, not banned. The reasoning is worth understanding, because it inverts the intuitive reading. The SCCS assessed HEMA in 2018 and found it safe at up to 35%, provided application stays on the nail plate and off the surrounding skin. The nail plate is a decent barrier and rapid curing limits how much free monomer is available. The risk comes from skin contamination and undercured product. So the EU didn’t decide HEMA was unsafe. It decided untrained people applying it at home couldn’t be relied on to apply it correctly.

In the United States there is no federal restriction on HEMA at all. At-home gel kits containing it, the exact product category Europe judged too risky for consumers five years ago, sit on American shelves and ship from American marketplaces every day.

One mechanistic detail with immediate practical value: lamps under 36 W may not fully polymerise the monomer, leaving free HEMA available to sensitise. An underpowered lamp is a sensitisation risk, which is an argument against the cheapest at-home kits specifically.

First: the people doing your nails

The strongest human signal in this entire subject isn’t about clients at all.

A 2016 meta-analysis of nineteen studies of cosmetologists and hairdressers found pooled odds ratios of 1.15 for infertility and 1.14 for fetal death. Modest, but consistent. A NIOSH analysis of the National Birth Defects Prevention Study found nail technicians working in early pregnancy roughly three times more likely to have infants with congenital heart defects.

Both findings need heavy caveats. The meta-analysis pooled manicurists with cosmetologists and used job title as an exposure proxy. The NIOSH figure rests on only 61 nail technicians, which means very wide confidence intervals, multiple comparisons, and no identified causal agent. NIOSH says so itself. This is hypothesis-generating work, not proof.

But look at the exposure conditions. NIOSH evaluated four salons and found that none of them had local exhaust ventilation at the manicure or acrylic stations. Formaldehyde exceeded the recommended exposure limit in six of twenty-four air samples. Acetone at one salon hit levels where the sampling media broke through, meaning true exposures were higher than measured. One salon was using nearly 100% methyl methacrylate, banned in that state. OSHA’s own hazard sheet notes that exposure to ethyl methacrylate “while pregnant may affect your child,” and NIOSH testing suggests exhaust ventilation can cut worker chemical exposure by at least half.

So the ventilation fix is known, cheap, and largely unimplemented.

Where TPO actually lands

Not on this list, which is the point.

TPO, the photoinitiator Europe banned in September 2025 and Britain began phasing out on 15 August 2026, was classified on rat oral-dose studies. No dermal or inhalation repeated-dose work, no human reproductive data, and a prior SCCS margin of safety of 1,515. Less than 0.2% of it can be extracted from the cured film. It is, on the evidence, the least concerning thing in the bottle.

Meanwhile HEMA, which has documented human sensitisation running into the low single digits of patch-tested patients and lifelong cross-reactive consequences, is restricted in Europe and untouched in America. And the occupational exposure question, the one with an actual human signal attached, is handled almost everywhere by voluntary guidance.

That’s the shape of the thing. Regulatory attention went to the ingredient with a reproductive hazard classification and no human evidence behind it. It has not gone to the ingredient with human evidence and no CMR classification, or to the ventilation in the room.

What to actually do

If you love your gels, you don’t need to quit. You need to change four things.

Ask what removal method the salon uses and decline the e-file. Soak fingertips rather than hands. Watch for redness, itching or swelling in the skin around the nail, not just on it, because that’s early sensitisation and it’s worth catching before it becomes permanent. And if you’re doing this at home, use a properly powered lamp and keep the product off your skin, because undercured monomer on skin is the specific thing Europe decided consumers couldn’t be trusted with.

Sunscreen on the hands beforehand if you like. It costs nothing and the evidence is genuinely unsettled.

And if you work in a salon, the highest-value thing in this article is a ventilated table.