Red Light Therapy (Photobiomodulation): What the Evidence Shows

What is red light therapy?

Red light therapy is the popular name for photobiomodulation (PBM): the use of red or near-infrared light at low power densities to change how cells and tissues behave [S1]. Red light is visible. Near-infrared light lies just beyond red and is invisible to the eye.

Three things are often confused:

• Photobiomodulation devices — low-power lasers or light-emitting diodes (LEDs) used to change cell activity without heating or cutting tissue. Older papers call this low-level laser therapy (LLLT) [S2].
• Surgical and cosmetic lasers — high-power devices that cut, heat or ablate tissue on purpose. They work differently and their evidence does not transfer.
• Ordinary lamps and red night lights — household light sources are built for seeing, not for delivering a measured dose to tissue. Using red light in the evening to protect sleep is a separate topic, covered in the article on circadian lighting.

How is it thought to work?

The leading model is that light is absorbed in the mitochondria, the parts of the cell that produce energy, mainly by an enzyme called cytochrome c oxidase [S1]. One proposal is that light frees nitric oxide bound to this enzyme, which restores electron transport in the mitochondria [S1]. Light- or heat-sensitive ion channels have also been proposed, and the effects appear to differ between healthy and stressed cells [S1].

This is a model, built mostly from cell and animal experiments. It explains why a biological effect is plausible. It does not show that any particular device, dose or outcome works in people. The article on mitochondria and red light covers the mitochondrial side in more depth.

How is the dose measured?

A light dose has several parts, and each one matters [S2]:

• Wavelength — the colour of the light: visible red or invisible near-infrared. Different wavelengths reach different depths.
• Irradiance (power density) — how much light power falls on each square centimetre of skin, usually in milliwatts per square centimetre (mW/cm²).
• Time — how long the skin is exposed.
• Fluence (energy density) — irradiance multiplied by time, in joules per square centimetre (J/cm²). This is the "dose" most trials report.
• Distance — irradiance falls as you move away from the light source, so the same panel delivers a very different dose at different distances.
• Pulsing and timing — whether the light is continuous or pulsed, and how often sessions are repeated.

Biphasic dose response. Studies of photobiomodulation frequently describe a biphasic response, often drawn as the Arndt–Schulz curve: low levels of light work better than higher levels, and too much light can lose the effect or inhibit it [S2]. Where the turning point lies in human tissue is not known.

Why home dose rarely matches trial dose. Clinical trials use set protocols, often with lasers applied to specific spots for a specific energy per spot [S4] [S6]. Home panels and masks are usually described by the maker's irradiance figure, often measured close to the LEDs. When researchers measured hand-held home LED devices, output was heterogeneous and the makers' dosing directions did not match accurate dose delivery [S10]. Across dermatology LED studies, fluences and wavelengths varied widely even for the same condition, and no study validated the dose the patient actually received [S9]. In practice, a home user rarely knows the dose they get.

What does the evidence show?

Oral mucositis in cancer care. Guideline. This is one of the best-studied uses. Oral mucositis is painful inflammation and ulceration of the mouth lining caused by some cancer treatments. The international MASCC/ISOO guideline recommends photobiomodulation to prevent oral mucositis in defined groups: people having a stem cell transplant, and people having head and neck radiotherapy with or without chemotherapy [S3]. It names specific clinical protocols and says the clinician should follow all their parameters. For treating mucositis that has already developed, the evidence was not enough for a guideline [S3]. This is hospital care with clinical devices, not a reason to use a home panel.

Joint and neck pain. Context-dependent. In knee osteoarthritis, a meta-analysis of 22 randomised placebo-controlled trials with 1,063 participants [S4] found that low-level laser therapy reduced pain and disability compared with placebo, with the clearest effect at the doses recommended by a laser therapy association [S4]. The authors note that major knee osteoarthritis guidelines did not recommend it at the time [S4]. In neck pain, a meta-analysis of 16 randomised controlled trials with 820 patients [S5] found pain relief compared with placebo, with mild side effects similar to placebo [S5]. Both analyses concern clinical lasers with defined doses, not consumer LED panels.

Exercise performance and recovery. Emerging. A meta-analysis of small trials found that light applied before exercise improved some performance measures, but trial designs were too different to pool markers of muscle damage and recovery [S6]. Many of the trials come from one research group, and the first author has disclosed research support from a laser maker (see the note under that source).

Hair loss. Context-dependent. In a meta-analysis of treatments for male pattern hair loss (androgenetic alopecia), low-level laser therapy was superior to placebo for hair growth in men, alongside minoxidil and finasteride [S7]. Heterogeneity was high, and the laser analysis covered men only [S7]. These were specific laser devices, such as laser combs and caps, used to a set schedule.

Skin ageing, acne and wound healing. Emerging and mixed. In a controlled trial funded by the light-source maker, people treated with red or broadband light rated their complexion better, and skin roughness and ultrasound-measured collagen density improved more than in controls [S8]. Most of the control group were company employees who were not randomised. This is a single study. A methodological analysis of dermatology LED trials, covering acne, wrinkles, wound healing and psoriasis, found mixed evidence of efficacy, highly heterogeneous outcomes, and only a third of studies with both a control group and blinded assessment [S9].

Sleep, mood, cognition and dementia. Emerging, mostly early. Transcranial photobiomodulation shines near-infrared light on the head. A systematic review found 35 studies [S11] in people, most reporting better cognitive scores, including small studies in memory complaints, mild cognitive impairment and dementia [S11]. Only half of the clinical trials were randomised, and a series of stroke trials was stopped at a late stage for lack of statistical significance [S11]. Controlled evidence for better sleep or mood from red light devices is too thin to draw a conclusion.

Testosterone, fat loss and "mitochondrial health". Unknown. Claims that red light panels raise testosterone, burn fat or improve mitochondrial health in general rest mainly on cell and animal research and on the mitochondrial model above [S1]. We found no consistent body of controlled human trials for these outcomes. Until such trials exist, the honest answer is that it is not known.

Is a more powerful panel better?

Marketing often implies that more power means better results. The dose research says otherwise: photobiomodulation frequently shows a biphasic response in which lower doses act better than higher ones [S2], and in knee osteoarthritis the clearest effects came at specific recommended doses [S4]. A higher irradiance mainly shortens the time needed to reach a dose; it does not make an unknown dose more effective, and the figures quoted by makers often do not match the dose delivered [S10].

Can red light stop Alzheimer's disease?

Transcranial photobiomodulation is sometimes promoted as a cure for Alzheimer's disease. The human evidence is early: the studies are small, only half of the clinical trials were randomised, and the reviewers call for further investigation [S11]. No controlled trial has shown that light stops or reverses dementia.

Are home devices medical devices?

Regulators treat light devices according to their intended use. In the United States, many devices reach the market through clearance: an FDA order finding a device substantially equivalent to one already legally on the market [S12]. Clearance is not the same as premarket approval, and it applies only to the specific device and the specific use named in it [S12]. For example, a laser comb was FDA-cleared for androgenetic alopecia [S7]. A device being "registered" or "listed" with a regulator is a separate, lower step and is not an assessment of the device's claims. Check what a given clearance actually covers before reading it as evidence for other uses.

Safety

Photobiomodulation devices are generally low-risk when used as directed, and side effects in pain trials were similar to placebo [S5]. Some cautions apply in general:

• Protect your eyes. Do not look into the LEDs or lasers. Near-infrared light is invisible, so a panel can look dim while its output is high, and invisible light does not trigger the blink reflex. Wear the protective goggles supplied, especially with large panels and face masks, and follow the maker's distance and time limits.
• Light-sensitising medicines and conditions. Some medicines and skin products make skin more sensitive to light, and some medical conditions involve light sensitivity. If this may apply to you, ask a doctor or pharmacist before use.
• Pregnancy. Safety in pregnancy has not been studied well enough to say. Talk to your doctor first.
• Skin after procedures, and skin lesions. Ask the clinician who treated you before using light on skin after peels, laser treatment or surgery. Do not shine light on a changing mole or an undiagnosed skin lesion; have it checked first.
• Cancer care. Oral mucositis protocols are delivered by clinical teams with set parameters [S3]. People with cancer should not substitute a home device for this care.
• Heat and burns. Stop if the skin becomes painfully hot or red.

What it does not tell you

• Different doses. Wavelengths, power, time and distance vary widely between studies, so results from one protocol do not transfer to another [S2] [S9].
• Lasers versus home LEDs. The strongest results come from clinical lasers and set protocols [S3] [S4] [S5] [S7]; most consumer panels have not been tested in their own trials.
• Small samples. Many trials are small; dermatology LED studies had a median of a few dozen participants [S9].
• Manufacturer funding. In one analysis of dermatology LED research, 10 of 27 studies (37%) [S9] were sponsored by the device maker, and a key skin trial was fully funded by one [S8].
• Hard to blind. Visible light and warmth make it hard to hide which group is treated, so self-reported outcomes are open to expectation effects [S9].
• Mechanism is not outcome. The mitochondrial model explains why effects are possible [S1]; it does not prove any specific benefit in people.
• Long-term effects of frequent home use are largely unknown.

In ONDA

ONDA does not include a red light feature, does not measure light exposure or light dose, and does not assess skin, hair, pain or cognition. ONDA reviews of red light panels and face masks are listed in the red light therapy reviews; they score devices on specifications and evidence, which is not the same as clinical proof for any use.