Red Light Therapy for Eczema: What the Evidence Actually Shows

Affiliate disclosure: This post contains affiliate links. See our affiliate disclosure for details.

We went looking for the clinical evidence that red light therapy helps eczema and found almost nothing standing. Not thin evidence, not mixed evidence — one uncontrolled series from before most readers of this page were born. In the published human record there is no controlled trial of red or near-infrared light in atopic dermatitis at all — LED or laser. The only human evidence that exists is two uncontrolled case series from 1993, run by the same Japanese group, using an 830nm laser diode rather than the LED panels sold here.

That is an uncomfortable thing to write on a site that sells red light panels, and we own one and use it most days. But eczema is a chronic inflammatory skin disease, plenty of people reading this have been managing it for years, and the gap between what the marketing implies and what the literature contains is wider here than on any other red light topic we have covered.

So this article does the two useful things instead. It shows you exactly what the search returns, so you can judge the gap yourself. And it covers the part nobody writes about: if you are going to buy an at-home LED device anyway — which is a defensible thing to do — the specs that decide whether it works with your skin or against it are not the ones printed on the box.

Last reviewed: September 2026

---

Quick Comparison: At-Home LED Devices and Eczema

Device	Form factor	Price	Wavelengths	Eczema evidence	Cleared for
Celluma HOME Series 2	Contouring panel, no skin contact	~$795	465 + 640 + 880nm	None	Aging skin, pain management
Omnilux Contour Face	Flexible silicone face mask	~$395	633 + 830nm	None	Fine lines and wrinkles
CurrentBody Skin Series 2	Flexible silicone face mask	~$470	633 + 830 + 1072nm	None — names eczema in its contraindications	Wrinkles
Hooga HG300	Body panel, non-contact	~$199	660 + 850nm	None	Not claimed
LightStim for Acne	Handheld, contact	~$169	415 + 630nm	None — emits warmth by design	Mild to moderate acne

Prices checked 4 September 2026 against each brand's own US storefront. The Celluma HOME was $795 on celluma.com and $815 on Amazon the same day. Not one device in this table is cleared, approved or trialled for eczema, and the "eczema evidence" column is the honest state of the field rather than a knock on any particular product.

---

The entire human record, and the part our own search missed

In 2026 a scoping review in Dermatitis set out to catalogue every non-UV light treatment ever studied in human atopic dermatitis. Cespedes Zablah, Griffin and Lio searched Embase, Web of Science, PubMed, CINAHL and Google Scholar. They found fifteen studies: eight on heliotherapy and climatotherapy — sunshine and sea air, essentially — four on blue light, and three on everything else combined, being one low-level laser study, one intense pulsed light study and one pulsed dye laser study.

Red LED photobiomodulation, the thing sold in every panel and mask on this page, does not have a category. It has no dedicated study in that review's fifteen.

Near-infrared light does have human data, and it is not LED. Two case series published in 1993 by Morita and colleagues at Hyogo College of Medicine treated atopic dermatitis patients with a GaAlAs 830nm diode laser applied to the skin:

• Morita H, Kohno J, Tanaka S, Kitano Y, Sagami S. "Clinical Application of GaAlAs 830 nm Diode Laser for Atopic Dermatitis." Laser Therapy 1993;5(2):75–78. doi:10.5978/islsm.93-or-08. Eighty-one patients. Itch decreased in 63 of 81 (79%), skin eruption improved in 57 of 81 (71%), no side effects reported.
• Morita H, Kohno J, Hori M, Kitano Y. "Clinical application of low reactive level laser therapy (LLLT) for atopic dermatitis." Keio Journal of Medicine 1993;42(4):174–176. doi:10.2302/kjm.42.174, PMID 7907380. One hundred and twelve patients. Itch decreased in 79 of 112 (71%), skin eruptions improved in 69 of 112 (62%), no side effects, and MHC class II and ICAM-1 expression on epidermal cells fell after treatment.

Read the two abstracts side by side and they are near-identical — same protocol, same five numbered findings, same year, overlapping authors, with the patient count grown from 81 to 112. These are almost certainly one continuing series reported twice as it enrolled, not two independent datasets, so the honest total is a hundred-odd patients rather than 193.

More to the point, neither has a control arm. No sham, no blinding, no randomisation, no comparison group of any kind — patients were treated and then asked whether they felt better. That is the study design that the dog trial further down this page exists to warn you about, because the dogs' untreated paws improved just as much as the lasered ones. Thirty-three years on, nobody has run the controlled version in people, and nobody has run any version with an LED.

We ran the search ourselves on 4 September 2026 to check. In Europe PMC, a title search pairing atopic dermatitis or eczema with photobiomodulation, red light, low-level laser or light-emitting diode returns fourteen records; the relevant ones are mouse studies and a veterinary trial, and the rest are about nurse-led education programmes that matched on the word "led". Searching abstracts for 660nm alongside atopic dermatitis in the title returns zero. Searching for 850nm returns exactly one, and it is a study in NC/Nga mice.

That search missed the two Morita papers, and it is worth saying why, because it is the failure mode of every literature search you will read on a page like this. They are indexed as "low reactive level laser therapy (LLLT)" rather than the phrase "low-level laser" the query required, and the 1993 Laser Therapy paper is on J-STAGE rather than PubMed. A phrase-matched title search on a 33-year-old paper from a Japanese journal, written before the indexing conventions everyone now assumes, will not find it. We caught it on review and we would rather print the correction than the tidier version of the search.

That is the file. Here is what is actually in it.

The animal work is real, and it is in mice

The mechanistic case is not imaginary. Korean groups have published a run of studies in NC/Nga mice, the standard model for atopic dermatitis-like skin lesions: 850nm LED light inhibiting T-cell-induced cytokine expression in 2012, 850nm phototherapy combined with low-dose tacrolimus in 2013, low-dose LED therapy with a water bath in 2016, and a comparison of several wavelengths in Annals of Dermatology in 2017.

Those papers are the reason people assume red light should help eczema, and the reasoning is sound as far as it goes. Atopic dermatitis is inflammatory, photobiomodulation modulates inflammatory signalling, therefore. The problem is that "therefore" has been doing all the work for over a decade without anybody running the controlled trial in people.

The one controlled test in a living patient was in dogs, and it failed

The closest thing to a randomised controlled trial of low-level light on atopic dermatitis is a veterinary study, and it is worth reading precisely because of how well it was designed. The two human Morita series above have more patients and a far better-looking result — around 70 to 80 per cent reporting less itch — but they have no control arm, so what follows is what their design cannot rule out.

Stich, Rosenkrantz and Griffin, published in Veterinary Dermatology in 2014, enrolled thirty client-owned dogs with symmetrical pedal itch caused by canine atopic dermatitis. Each dog got low-level laser therapy on one paw and a placebo laser on the other — a split-body design, so every animal is its own control. Treatment was 4 J/cm², three times weekly for two weeks, then twice weekly for two more.

Both paws improved significantly from baseline. Neither paw improved more than the other. The difference between real laser and placebo laser came out at P = 0.0856 on the severity score and P = 0.5017 on the owner-rated itch score. The authors' conclusion is one sentence: "Low-level laser therapy is not an effective localized treatment for pedal pruritus in canine atopic dermatitis."

Both paws getting better while the treated paw does not beat the sham is the textbook shape of a placebo result. It is exactly what an uncontrolled before-and-after photo on a brand's website would show, and exactly why those photos are not evidence.

The itch study is the most interesting thing in the file

Itch is what people actually want fixed. There is one human randomised, sham-controlled trial of low-level light therapy on itch, and its result cuts against eczema specifically.

Lang-Illievich and colleagues, in PLoS One in 2024, ran a double-blind split-body study on seventeen healthy volunteers. They induced itch two ways: with histamine, and with Mucuna pruriens, the cowhage plant used in research precisely because its itch runs through non-histaminergic pathways. Six minutes of light therapy on one quadrant, sham on another.

On the histamine model it worked, and clearly: itch intensity down 13.9 points (95% CI 10.5–17.4, p = 0.001), with alloknesis and hyperknesis also reduced. On the cowhage model it did nothing at all — a difference of 0.8 points with a confidence interval running from −2.3 to 3.8, p = 0.61.

Now put that beside how eczema itch works. Oral antihistamines are not a recommended treatment for atopic dermatitis; the American Academy of Dermatology's 2023 guideline concluded there are insufficient data to recommend them. The reason is that atopic itch is driven substantially by non-histaminergic signalling, IL-31 among it — which is why the big blue-light trial we are about to discuss went to the trouble of measuring IL-31 in patients' blood.

So the only human itch trial of this technology succeeded on the mechanism that is not the eczema mechanism, and failed on the one that resembles it. That is not proof it cannot help. It is a specific, published reason to lower your expectations, and we have not seen it mentioned in a single consumer article on this topic.

---

The light that does have eczema data is not the light being sold to you

Two modalities have genuine atopic dermatitis evidence. Neither is a red LED panel.

Narrowband UVB is the phototherapy with guideline support

The AAD's 2023 guidelines for managing atopic dermatitis in adults make a conditional recommendation in favour of phototherapy, and note that most of the current literature reports on narrowband UVB — the most widely used form, with a safer track record than UVA1 or broadband UVB.

Narrowband UVB is 311–313nm. It is ultraviolet. It is delivered in a cabinet in a dermatology department, on a schedule, with cumulative-dose tracking, because the thing that makes it work is also the thing that carries long-term skin cancer risk. It is not a red light panel with a different setting, it is not something we sell, and it is not something anyone should be improvising at home.

If you have moderate or severe eczema that topicals are not controlling, a phototherapy referral is the conversation to have. It is the light treatment with the guideline behind it.

Blue light has four human studies, and the biggest one is mostly negative

Blue light is the interesting case, and it is where the honest version of this article gets genuinely useful.

AD-Blue is the largest and best-designed trial of non-UV light in atopic dermatitis. Buhl and colleagues, published in the Journal of the German Society of Dermatology in 2023, enrolled 87 patients across Göttingen, Marburg and Geneva in a double-blinded, three-armed randomised trial: 415nm versus 450nm versus sham. Full-body prototype cabinets, 40 mW/cm², fifteen minutes per body side, three times weekly for eight weeks.

The results deserve to be quoted rather than summarised. The 450nm arm produced a significant reduction in itch on a visual analogue scale (−1.6 ± 2.3; p = 0.023 against sham) but 415nm did not. The 415nm arm produced a significant drop in patient-oriented SCORAD (−11.5 ± 18.4; p = 0.028) but 450nm did not. And then: "None of the other outcome measures (EASI, SCORAD, IGA, DLQI) changed significantly." The authors' own conclusion is that full-body blue light "did not lead to an amelioration of any of the objective measures of AD."

The earlier, smaller studies were more encouraging. Keemss and colleagues in 2016 treated 21 patients with mild to moderate eczema using 453nm LED light locally, thirty minutes, three times weekly for four weeks; the local eczema severity index fell by 1.9 in treated areas against 1.3 in untreated ones. Becker and colleagues in 2011 gave 36 patients with severe atopic dermatitis full-body 400–500nm light daily for five days and reported EASI improvements of 41% and 54% at three and six months.

The pattern across the four is the one you see whenever a field matures: promising small open studies, then a properly blinded, sham-controlled trial that mostly does not replicate them.

---

The spec that decides everything: 450nm is not 415nm

Here is the part worth the price of the article, and it is the direct consequence of the trial data above.

Every blue-capable at-home device we could find runs its blue LEDs at 415nm. LightStim for Acne is 415 and 630nm. Omnilux Clear is 415nm blue with 633nm red. Dr Dennis Gross's FaceWare Pro is 415nm blue. That is not an accident or a cost saving — 415nm is the wavelength absorbed by the porphyrins inside acne bacteria, so it is the correct choice for the condition these devices are actually cleared to treat. Our blue light vs red light therapy guide goes through why that band is an acne tool and nothing else.

The eczema studies used 450nm and 453nm. In AD-Blue, run head to head in the same trial, 415nm was the arm that did not move itch.

So the reasoning that gets people to buy a combination mask for eczema — "blue light has eczema evidence, this mask has blue light" — skips the only comparison anyone has actually run, and lands on the wavelength that lost it. The single at-home device we found that gets anywhere near the trial band is the Celluma HOME at 465nm, and even there Celluma states that wavelengths and doses are treatment-dependent, so we cannot confirm the blue is firing in the modes the HOME unit runs.

There is a second gap underneath the first. The trials that showed anything used full-body cabinets at 40 mW/cm² for thirty minutes a session, three times a week for eight weeks, or a purpose-built local irradiator at a stated dose. A face mask covers your face. Eczema is usually on the insides of elbows, the backs of knees, the hands and the neck. Matching a wavelength you cannot verify, at a dose nobody publishes, over the wrong body area, is three assumptions stacked on a trial that mostly failed anyway.

---

Read the contraindications before you read the marketing

CurrentBody publishes more specification detail than anyone else in this category, which is the main reason to respect the brand, and its own contraindication list contains two lines that belong in every article on this subject:

"DO NOT use the device to treat moles, warts, open sores, cancerous lesions or any skin condition."

"DO NOT use this device if you suffer from Lupus erythematosus, photosensitive eczema or Albinism."

Take those in order. The first is the manufacturer of a leading LED mask stating plainly that the device is not for treating a skin condition. That is a regulatory position rather than a scientific claim, but it is the manufacturer's own framing of what they built, and it sits oddly against a purchase made specifically to treat a skin condition.

The second needs a careful reading, because it is easy to get wrong in both directions. Photosensitive eczema is not the same thing as atopic eczema. It is a photoaggravated or photoallergic dermatitis, a distinct condition in which light exposure is itself the trigger. If you have ordinary atopic dermatitis, that line is not automatically about you.

But the reason it matters is that plenty of people have never been told which one they have, and the overlap cases are real — photoaggravated atopic dermatitis exists. If your eczema reliably gets worse in summer, after sun exposure, or on sun-exposed skin while covered skin stays calm, that is the pattern to raise with a dermatologist before you buy a device that shines light on it. It costs one message.

The same list flags photosensitising medication, which is worth checking against your own prescriptions. Doxycycline is common enough in dermatology, and several eczema patients are on something.

---

Heat and contact: the two specs that actually matter here

If the wavelength argument is unresolvable, the physical design of the device is not, and for eczema it points somewhere specific.

Sweat is a documented eczema trigger, and eczema skin has a compromised barrier by definition — that is what the condition is. A device that warms the skin, or that seals a silicone shell against it for ten to thirty minutes, is creating occlusion and warmth on a surface that is already leaking water and already itching. That is not a hypothetical: it is the same reasoning that puts a hot bath on every eczema trigger list, and it is why we tell people in sauna for skin that heat can go either way on eczema and needs testing in short sessions.

The manufacturers differ on this more than you would expect, and they say so in their own words:

• Omnilux: "Omnilux Contour generates no heat or UV light."
• CurrentBody: "It does not generate heat and should not be uncomfortable."
• LightStim: "engineered to emit a soothing, gentle warmth."

For aging skin, that last one is a pleasant feature. For a barrier-compromised, itch-dominant condition where sweat is a trigger, it is the wrong product, and it is designed to be.

Contact is the second axis and almost nobody ranks on it. A mask presses against skin. A panel does not. If your eczema is weeping, fissured or freshly scratched, pressing a silicone shell onto it is a poor idea for reasons that have nothing to do with photons.

---

The devices, and what each one is actually for

A note on what these assessments are and are not. We own one red light panel, wall-mounted, and we have never used anything on this site to treat a skin condition. Everything below is manufacturer specifications and manuals read against each other on 4 September 2026, plus the published trial literature. Where a figure is not published, we say so rather than guess. None of these ratings is a claim that the device treats eczema, because none of them does.

Celluma HOME Series 2: the only one that never touches your skin

~$795 · 465 + 640 + 880nm · 14" x 6" treatment area · up to 30 min/day

The Celluma is not a mask. It is a flexible, shape-taking panel with a 14 by 6 inch treatment area, positioned about an inch off the skin rather than laid on it, and it comes with facial rest stands so it can hover. For eczema that matters twice over: no occlusion and no pressure on inflamed skin, and because it contours, it can wrap the inside of an elbow or the back of a knee — which is where eczema usually is and where no face mask will ever reach.

It is FDA cleared for aging skin and pain management. Not eczema. Its 465nm blue is the closest anything at home gets to the 450–453nm the eczema trials used, but Celluma states doses and wavelengths are treatment-dependent, so treat that as a maybe rather than a feature you are buying.

The 30-minute session is the longest here, which is a lot of exposure to build up to on reactive skin. Start well short of it.

Pros: No skin contact at all. Reaches flexures, not just the face. Genuine clinical pedigree. Closest wavelength to the trial band. Cons: Twice the price of a mask. Blue mode unverifiable. Long sessions. You have to set it up rather than strap it on. Best for: Body eczema on elbows, knees or hands, and anyone whose skin objects to anything resting on it.

Check price →

Hooga HG300: the cheapest way to cover body eczema

~$199 · 660 + 850nm · non-contact panel

If you have already decided to try red light on eczema, and your eczema is on your body rather than your face, a small panel is the rational form factor and this is the cheapest sensible one. It is red and near-infrared only, with no blue mode to talk yourself into, it never touches the skin, and at $199 it is roughly a fifth of the cost of being wrong.

What it does not have is any eczema evidence, any relevant clearance, or any published irradiance figure at a stated distance on the page we read. Its 850nm sits close to the 830nm of the 1993 Morita series, which is the nearest thing to a point in its favour and is not much of one — that was a laser diode in an uncontrolled study, and no LED has ever been tested on eczema in a person. Hooga's larger panels publish more. We covered the range in our Hooga red light review.

Pros: Cheapest device here by a distance. Non-contact. No heat against the skin. Covers body areas a mask cannot. Cons: No clearance claimed. No eczema evidence, like everything else here. Nothing near the 450–453nm blue band, and its 830nm-adjacent near-infrared borrows from a study with no control arm. Best for: Testing the idea on flexural eczema without spending $795.

Check price →

Omnilux Contour Face: the best-behaved mask for facial eczema

~$395 · 633 + 830nm · 132 LEDs · 10-minute sessions

If your eczema is facial and you want a mask, this is the one whose design fights you least. Omnilux states flatly that the Contour generates no heat or UV light, with no hedging, and the session is ten minutes rather than thirty — less contact time is less occlusion. Two wavelengths, 66 red and 66 near-infrared across 66 dual-chip bulbs, and no blue mode.

It is cleared for fine lines and wrinkles. Nothing else.

One caveat that matters more for eczema than for anything else: the silicone shell itself sits against the skin, and at least one reviewer on Omnilux's own product page describes an allergic reaction to it. On skin with a compromised barrier and a tendency to react, patch-test the jawline before a full-face session.

Pros: Unambiguous no-heat statement. Short session. No blue mode. The most-reviewed device in this group. Cons: No published irradiance. Silicone contact on inflamed skin is the wrong direction. Not cheap for two wavelengths. Best for: Facial eczema, if you have decided on a mask and want the least provocative one.

Check price →

CurrentBody Skin Series 2: the most honest spec sheet, and it names eczema

~$470 · 633 + 830 + 1072nm · 236 LEDs · 30 mW/cm² · 10 minutes, 3–5x weekly

CurrentBody publishes what everyone else withholds: the full LED breakdown (110 red at 633nm, 110 near-infrared at 830nm, 16 deep near-infrared at 1072nm), a stated power density of 30 mW/cm², a ten-minute treatment time and a recommended three to five sessions weekly. It states the mask does not generate heat. For a category this vague about numbers, that transparency is worth something.

It is also the only device here whose own contraindication list names eczema, and which tells you not to use it to treat any skin condition. We are quoting that against it in the section above, but it should count in the brand's favour that they published it at all — most of these pages simply do not have a contraindication list to read.

Pros: Best-documented specs in the category. Explicit no-heat claim. Publishes contraindications most brands omit. Cons: Most expensive mask here. Names photosensitive eczema in its warnings. The 1072nm band adds cost for a wavelength with no eczema data behind it. Best for: Buyers who want real numbers before spending, and who will read the warnings properly.

Check price →

LightStim for Acne: the one we would not buy for eczema

~$169 · 415 + 630nm · 36 LEDs · 3 minutes per area, once daily

Nothing is wrong with this device. It is FDA cleared to treat mild to moderate acne, it is the cheapest thing here, and for its actual indication it has a long track record and a sensible three-minute session.

It is simply the wrong tool for this condition, on both of the axes that matter. Its blue is 415nm — the arm that failed to move itch in AD-Blue — and LightStim states the light is "engineered to emit a soothing, gentle warmth", which is a deliberate design decision that runs straight into a sweat-triggered, barrier-compromised condition. It is also handheld and used in contact, three minutes per area at a time, so covering both arms means holding it against inflamed skin repeatedly.

If you have acne, this is a reasonable buy. If you have eczema, it is the one to skip.

Pros: Cheapest here. FDA cleared for its actual indication. Short sessions with clear guidance. Cons: Emits warmth by design. 415nm blue is the wavelength that failed the itch endpoint. Contact device, used area by area. Best for: Acne. Not eczema.

Check price →

---

Buyer's guide: how to think about this purchase

Spend the first $200 on a diagnosis, not a device

Eczema is a category, not a diagnosis. Atopic dermatitis, contact dermatitis, dyshidrotic eczema, nummular eczema and photosensitive dermatitis are managed differently, and one of them gets actively worse with light exposure. If you have never had it properly assessed, a consultation is a better first spend than any panel on this page, and it is the only way to rule out the version where a light device is contraindicated.

Ask about phototherapy before you buy a light

If topicals are not controlling it, narrowband UVB is the light treatment with the AAD's conditional recommendation behind it. It is a clinic course, it is monitored, and it may be covered where a consumer device never will be. Asking about it costs nothing and it is the actual evidence-based version of the thing you are shopping for.

Rank on heat and contact, then form factor, then wavelength

In that order, and it is the reverse of how these products are marketed. Find the manufacturer's own words about heat. Decide whether anything should be touching your skin at all. Then pick a form factor that reaches where your eczema actually is — which, for most people, is not the face. Wavelength comes last here only because no wavelength has been validated for this condition in a controlled human trial, so ranking on it is ranking on nothing.

Do not buy a device for the blue mode

The blue you can buy is 415nm. The blue with eczema data is 450–453nm, delivered in a full-body cabinet at a published dose, and even that trial missed every objective endpoint. A blue button on a face mask is not that treatment, and owning one makes it easy to talk yourself into thinking it is.

Understand what "FDA cleared" is doing in the copy

Not one device here is cleared for eczema. Clearance means substantial equivalence to a device already on the market — it is not approval, and it is not evidence that the thing works for your condition. Our guide to FDA-cleared red light therapy devices covers what the term does and does not tell you.

Keep doing the boring things

Emollients, trigger avoidance and whatever your dermatologist has prescribed are the foundation, and no light device substitutes for any of them. If a panel helps you at all, it will help on top of that, not instead of it.

---

FAQ

Does red light therapy help eczema?

There is no controlled human trial that answers this, which is itself close to the answer. A 2026 scoping review in Dermatitis searched five databases for non-UV light treatments in atopic dermatitis and found fifteen human studies — eight on heliotherapy, four on blue light, and three covering low-level laser, IPL and pulsed dye laser between them. Red LED photobiomodulation had no dedicated study. What does exist in humans is two uncontrolled case series from 1993 (Morita et al., Laser Therapy and Keio Journal of Medicine), in which 71 to 79 per cent of patients reported less itch after treatment with an 830nm laser diode — no sham arm, no blinding, no LED, and no follow-up in the three decades since. Red light is low-risk and the anti-inflammatory mechanism is plausible, so trying it is reasonable. Expecting a known effect is not, because nobody has run the study that would measure one.

Can red light therapy make eczema worse?

It can, through heat and occlusion rather than through the light itself. Sweat is a well-documented eczema trigger and eczema skin has a compromised barrier, so a device that warms the skin or seals a silicone shell against it for half an hour is working against you regardless of wavelength. Some manufacturers state their device generates no heat; at least one markets gentle warmth as a feature. Separately, if you have photosensitive or photoaggravated eczema — a distinct condition where light exposure is the trigger — light devices are contraindicated, and CurrentBody names it explicitly in its own warnings. If your eczema is reliably worse on sun-exposed skin, get that checked before buying anything.

Is red light or blue light better for eczema?

Blue has the evidence, red has none, and the blue evidence is weaker than it first looks. The largest trial, AD-Blue, randomised 87 patients to 415nm, 450nm or sham in full-body cabinets. The 450nm arm significantly reduced itch on a visual analogue scale; the 415nm arm significantly reduced patient-oriented SCORAD. Every objective measure — EASI, SCORAD, IGA, DLQI — was unchanged, and the authors concluded blue light did not improve any of them. Then note that every at-home blue device runs 415nm, the arm that did not move itch. See our blue light vs red light therapy comparison for what each band is genuinely for.

What light therapy actually works for eczema?

Narrowband UVB, delivered in a dermatology clinic. The American Academy of Dermatology's 2023 guidelines make a conditional recommendation in favour of phototherapy for adults with atopic dermatitis and note that narrowband UVB is the most widely used form, with the most literature and a safer profile than UVA1 or broadband UVB. It is ultraviolet light at 311–313nm, dosed and tracked because cumulative exposure carries risk. It is not something to approximate at home, and it is not what any device on this page emits.

Will a red light panel help with eczema itch?

The one controlled human trial of low-level light therapy on itch suggests probably not, for eczema specifically. Two uncontrolled 1993 case series did report itch relief in 71 to 79 per cent of atopic dermatitis patients (Morita et al.), but with no sham arm those numbers cannot be separated from what people report after any treatment they know they are receiving — and they used a laser diode, not an LED panel. Lang-Illievich and colleagues, in PLoS One in 2024, tested it on seventeen volunteers against two induced itch models. On histamine-induced itch it worked well. On Mucuna pruriens — cowhage, used in research because its itch runs through non-histaminergic pathways — it did nothing measurable at all. Atopic itch is understood to be driven substantially by non-histaminergic signalling, which is why antihistamines are not a recommended eczema treatment and why the AAD found insufficient data to recommend them. The model the light failed on is the one that resembles eczema.

Is there an FDA-cleared red light device for eczema?

Not among the at-home devices we checked, and we read each clearance. Omnilux Contour Face is cleared for fine lines and wrinkles. CurrentBody Skin Series 2 is cleared for wrinkles. Celluma HOME is cleared for aging skin and pain management. LightStim for Acne is cleared for mild to moderate acne. Hooga does not claim a clearance on the page we read. Clearance also means equivalence to an existing device rather than proof of benefit, so even a hypothetical eczema clearance would not be the same thing as trial evidence.

How long should I give it before deciding it is not working?

Eight to twelve weeks of consistent use, with a photograph taken in the same light on day one. Eczema fluctuates with season, stress, sleep, detergent and a dozen other things, so memory is a terrible instrument for judging it — that is exactly how the dog study got two paws improving on their own while the laser did nothing. If you have seen no change after three months, the honest read is that it is not doing anything for you, and the money is better spent on a dermatology appointment.

Can I use a red light mask on broken or weeping eczema?

No, and the manufacturers say so. CurrentBody's contraindications state plainly not to use the device on open sores, and the general instruction across these brands is to treat intact skin only. Pressing a silicone mask onto fissured or weeping skin is a poor idea for reasons that have nothing to do with light — infection risk and mechanical irritation being the obvious two. If your skin is broken, that is a treatment conversation, not a shopping one. Our red light therapy side effects guide covers what to watch for on reactive skin.

---

Our Verdict

If we had eczema and $800 to spend, none of it would go on a light device first. It would go on getting the diagnosis right and asking about a narrowband UVB course, because that is the light treatment with a guideline behind it and the only one anybody has properly tested.

If you have already decided to try an at-home device — and there are sane reasons to, starting with the fact that it is low-risk and cheaper than most things people try — buy the Celluma HOME Series 2 at around $795, and buy it for its shape rather than its spec sheet. It is the only device here that never touches your skin, and the only one that will wrap the inside of an elbow or the back of a knee, which is where eczema is on most people who have it. If $795 is not the right number, the Hooga HG300 at ~$199 gets you a non-contact body panel for a fifth of the price and no worse an evidence base, because the LED evidence base is the same for all of them: nothing.

The one thing we would want you to take away is the 415nm point. Every blue-capable device you can buy runs the wavelength that lost the head-to-head in the only trial that ran one. That is the gap between what these products are and what people buy them for, and it is invisible unless somebody reads the trial.

This article is research, not medical advice. Eczema covers several distinct conditions, one of which is triggered by light exposure. See a dermatologist before treating your skin with a light device.

See more of our coverage on the red light therapy hub, including red light therapy for skin and red light therapy for rosacea. More about how we work and what we actually own on our about page.