Home Learn Science
A century of research
A Nobel Prize in 1903, an accidental discovery in Budapest in 1967, a NASA programme, then twenty years of clinical trials. Here is that history with its sources — and, just as importantly, with what it does not yet prove.
Photobiomodulation — low-power red and near-infrared light — has been studied since 1967 and now has several meta-analyses of randomised trials behind it. The best-established results concern pain, muscle recovery, skin and hair. They remain heterogeneous: doses, distances and durations vary from one trial to the next.
Page reviewed on 08/09/2026. Antared devices are wellness products: they do not diagnose, treat or cure any disease.
From lab accident to controlled trial
Every milestone links to its original source. Scientific papers open on their DOI, the permanent identifier of the article — not on a second-hand summary.
Denmark's Niels Ryberg Finsen receives the Nobel Prize in Physiology or Medicine for treating lupus vulgaris — tuberculosis of the skin — with concentrated light. Phototherapy enters medicine through the front door.
NobelPrize.org ↗In Zur Quantentheorie der Strahlung, Albert Einstein describes stimulated emission: one photon can trigger a second, identical one. Nobody will build a laser for another forty-three years, but the theory is on paper.
Theodore Maiman fires a working ruby laser at Hughes Research Laboratories and publishes the result in Nature. Light becomes an instrument whose wavelength and intensity can be controlled.
Maiman, Nature 187, 493 (1960) ↗At Semmelweis University, Endre Mester sets out to find whether a low-power laser causes cancer in mice. It causes none — but the hair of the irradiated mice grows back faster than the control group's. Laser biostimulation has just been discovered by accident.
Mester & Mester, Photomed Laser Surg (2017) ↗LEDs originally developed to grow plants in orbit are tested on wound healing. Harry Whelan's team shows they upregulate tissue-regeneration genes. This is the decisive shift: unlike a laser, an LED can go into a consumer device.
Whelan et al., J Clin Laser Med Surg 21(2) ↗The Russian biophysicist identifies the photoacceptor: cytochrome c oxidase, the last link in the mitochondrial respiratory chain. She also shows the role nitric oxide plays in the reaction. For the first time, the effect has a mechanism rather than an observation.
Karu et al., Lasers Surg Med 36(4) ↗Roberta Chow and co-authors pool 16 randomised trials covering 820 patients with neck pain. Pain drops by nearly 20 mm on a 100 mm visual scale, and the effect holds up to 22 weeks after treatment ends. Side effects do not differ from placebo.
Chow et al., The Lancet 374(9705) ↗Twenty elite female basketball players, fourteen days, thirty minutes of whole-body irradiation each evening. The treated group improves its sleep score, its serum melatonin and its twelve-minute run performance. Small sample, but a clean result.
Zhao et al., J Athl Train 47(6) ↗Forty-four men with androgenetic alopecia, a 655 nm helmet, one session every other day for sixteen weeks. The active group gains 35 to 39 % in counted hairs versus placebo. The study is replicated in women the following year, with a 37 % gain.
Lanzafame et al., Lasers Surg Med 45(8) ↗136 volunteers, thirty sessions, two spectra compared against a control group. Skin roughness by profilometry, collagen density by ultrasound, photographs scored blind: all three measures improve. The broadband spectrum adds nothing over red alone.
Wunsch & Matuschka, Photomed Laser Surg 32(2) ↗Ernesto Leal-Junior pools thirteen randomised trials on muscle performance. Repetition counts rise, so does time to exhaustion — but only for red and infrared, applied before exercise, between 50 and 200 mW and 5 to 6 joules per spot.
Leal-Junior et al., Lasers Med Sci 30(2) ↗Michael Hamblin, of the Wellman Center for Photomedicine, synthesises fifty years of work: the photon displaces the nitric oxide that was blocking cytochrome c oxidase, electron transport resumes, ATP rises. This is still the reference cited today.
de Freitas & Hamblin, IEEE JSTQE 22(3) ↗A meta-analysis of ten trials on knee osteoarthritis concludes that laser added to rehabilitation exercise beats exercise alone on pain, stiffness and function. The important word is added: light complements movement, it does not replace it.
Ahmad et al., Physiotherapy 114 ↗The LIGHTSITE III trial follows 100 subjects with dry AMD, treated at 590, 660 and 850 nm against sham. Visual acuity in the treated group improves significantly. It is one of the most rigorous trials in the field — and it concerns a medical device, not a home appliance.
Boyer et al., Retina 44(3) ↗Six randomised trials, chronic lower-limb wounds, wavelengths from 620 to 950 nm and doses from 2.4 to 126 J/cm²: a review that sets the laser aside to look only at LEDs — what consumer devices are actually made of. The authors grade the evidence as very low under GRADE, and note that the highest-dose trial — 126 J/cm² — shows no benefit. It is the milestone closest to what we sell, and also the most cautious.
Miranda MB et al., Lasers Med Sci (2026) ↗The field now counts thousands of publications, but doses, distances and durations vary from one trial to the next to the point where comparison gets hard. That is exactly why we publish our parameters rather than promises — see how it works and our values.
What the trials measured
We give the sample size and what was actually measured — not just the conclusion. A trial on twenty people and a meta-analysis on eight hundred do not carry the same weight. The label at the top of each card gives the level of evidence, and the light source used — laser or LED — is stated whenever it changes how the result should be read.
Randomised controlled trial, 136 volunteers, 30 sessions. Collagen density measured by ultrasound and roughness by digital profilometry, with blinded photographic assessment.
Wunsch A, Matuschka K. Photomed Laser Surg 2014;32(2):93–100. DOI
Double-blind randomised trial, 42 women, 655 nm, 16 weeks at one session every other day. Hair counted photographically on a tattooed area, scored by a blinded investigator. No adverse events reported.
Lanzafame RJ et al. Lasers Surg Med 2014;46(8):601–607. DOI
Meta-analysis of 16 randomised trials against placebo or active treatment, 820 patients with acute or chronic neck pain. Side effects no different from placebo.
Chow RT et al. The Lancet 2009;374(9705):1897–1908. DOI
Meta-analysis of 13 randomised trials of sufficient methodological quality. Mean gain of 5.5 repetitions and 4.1 seconds to exhaustion, obtained mainly by applying light before exercise.
Leal-Junior ECP et al. Lasers Med Sci 2015;30(2):925–939. DOI
Meta-analysis of 14 studies on the blood marker of muscle damage. Clear effect for localised exercise, inconclusive for general exercise — a nuance most product pages leave out.
Machado AF et al. Clin J Sport Med 2020;30(3):267–274. DOI
Randomised trial on 20 athletes, 14 days, 30 minutes each evening. Pittsburgh Sleep Quality Index, serum melatonin and a 12-minute run. Small sample: treat it as a signal, not as definitive proof.
Zhao J et al. J Athl Train 2012;47(6):673–678. DOI
Meta-analysis of 10 randomised trials on knee osteoarthritis, all with good PEDro scores. Pain, stiffness and function improve when light is added to an exercise programme.
Ahmad MA et al. Physiotherapy 2021;114:85–95. DOI
Reference review from the Wellman Center for Photomedicine (Harvard) on the proposed mechanisms: photoacceptor, light-sensitive ion channels, signalling cascades. The starting point for any serious reading of the field.
de Freitas LF, Hamblin MR. IEEE J Sel Top Quantum Electron 2016;22(3). DOI
Meta-analysis of 18 randomised trials covering 670 skin wounds. Both wound-area reduction and complete healing rate were significantly higher under light, and pain was lower on the visual analogue scale (mean difference, 95 % CI: −2.52 to −0.19). These trials use a low-level laser, not LEDs — the same caveat as the 2003 NASA milestone. The authors themselves call for higher-quality trials.
Taha N et al. Cureus 2024;16(10):e72542. DOI
Systematic review of six randomised trials on chronic lower-limb wounds — the only paper on this page that keeps LEDs alone. Wound-area reduction, wound-bed quality and microcirculation all improved, particularly in diabetic foot ulcers. But the evidence is graded very low under GRADE: heterogeneous parameters, small samples. The highest-dose trial, at 126 J/cm², showed no benefit at all.
Miranda MB et al. Lasers Med Sci 2026;41(1). DOI
A review of mitochondrial redox signalling, filling in two things the 2016 review does not cover: light produces apparently opposite effects depending on whether the cell is healthy or stressed, and it can act as preconditioning — before exercise rather than after. This is the mechanistic argument behind a pre-workout session.
Hamblin MR. Photochem Photobiol 2018;94(2):199–212. DOI
Where we stand in all this
That distinction is not modesty. It is the only way to read everything above correctly.
We set our devices to the published parameters
Wavelength, irradiance and duration are chosen within the ranges found in the literature cited on this page — the 2026 LED review, for instance, describes trials between 620 and 950 nm. The spectrum is detailed on wavelengths, the dose on how it works.
We display those parameters, not promises
Every product page states the wavelength, the irradiance with its measurement distance and the session duration. That is what lets you compare a device against a published trial — and two devices against each other.
Our job is the quality-to-price ratio
We select and distribute red light therapy equipment. Our job is to put the best devices in your hands at the best price — not to produce science. The rules we hold ourselves to are written down.
We run no clinical trials
Antared is not a laboratory. None of the eleven papers cited here was run on an Antared device, and we claim none of them. A brand telling you its product is “clinically proven” without a trial reference on that product is telling you something else.
We do not turn a study into a promise
A result obtained with a laser on a chronic wound says nothing about what an LED panel will do in your evening session. The studies on this page describe a research field; they do not describe a result you will get.
The questions that keep coming up
The level of evidence depends on the domain. It is strongest for pain, with a meta-analysis of 820 patients published by The Lancet in 2009, and for knee osteoarthritis as an adjunct to exercise, with ten randomised trials in 2021. It is weakest for chronic wounds, where a 2026 review of six LED trials grades the evidence as very low under GRADE.
A large share of the historical literature uses a laser. In 2003, Harry Whelan's team showed that LEDs developed by NASA upregulated tissue-regeneration genes, which opened the way to consumer devices. The two sources are still not interchangeable: on this page, the light source is stated for every paper.
The 2026 LED review describes trials between 620 and 950 nm, at doses ranging from 2.4 to 126 J/cm². That range is wide because protocols are not standardised from one trial to the next. The dose figures we work from are detailed on how it works, and the spectrum on wavelengths.
No. Antared selects and distributes red light therapy devices: we are not a laboratory and we run no clinical trials. None of the eleven papers cited on this page was run on an Antared device. We set our devices to the parameters published in this literature, and we display those parameters.
Because the dose response is biphasic: too little energy does nothing, too much reduces the effect instead of amplifying it. The 2026 LED review illustrates this directly. Its highest-dose trial, at 126 J/cm², shows no benefit at all, where lower doses do.
From two weeks to four months, depending on what is measured: fourteen days for a sleep score and serum melatonin in 20 athletes, thirty sessions for collagen density in 136 volunteers, sixteen weeks for a hair count in 42 women. In every case the protocols rely on repeated sessions, never on a single one.
The double-blind trial on hair regrowth, run on 42 women over sixteen weeks, reports no adverse events. That finding concerns that specific trial, not general use. Antared devices are wellness products: they do not diagnose, treat or cure any disease. If you have symptoms, an ongoing treatment or are pregnant, a healthcare professional's advice takes precedence.
Put it into practice
We publish wavelength, irradiance and duration for every product, so you can compare them against what you have just read.
Personal · Desktop
Compact 7-spectrum panel, red through near-infrared, for a daily session that sits on your desk.
€299,00 DiscoverFace · Cordless
Flexible silicone mask, 288 LEDs across 4 channels including 630 nm red and 850 nm near-infrared.
€250,00 DiscoverFull body · Large format
Large-format panel for full-body routines, studios, and athletic recovery.
€1.500,00 DiscoverFrise sourcée de 1903 à 2026, onze études avec leur DOI, les doses employées dans les essais et leurs limites. Aucune n'a porté sur un appareil Antared.
