Red light therapy: what the human research actually supports

"Photobiomodulation" (the formal term for red and near-infrared light therapy) has a real research base. It's also one of the most over-marketed wellness categories. The gap between what specific studies show and what devices claim is large.

Where the evidence is strongest#

Skin. Red and near-infrared LED panels have FDA clearances for specific dermatological indications. RCTs in skin conditions show:

• Modest improvements in mild acne (around 630 nm), clinical trials and a 2024 systematic review in JAMA Dermatology.
• Wound healing acceleration in some surgical and burn contexts.
• Photoaging: small studies report improvements in fine lines and skin texture.

The skin literature is the cleanest. Effects are real but modest; devices for these indications are typically regulated as medical devices.

Where the evidence is weaker#

Muscle recovery and performance. A series of small trials and a 2018 systematic review and meta-analysis (Vanin et al.) found low- to moderate-quality evidence of modest benefits from photobiomodulation on muscle performance and fatigue; effects specifically on post-exercise soreness are less well established. Effect sizes are modest and trial quality varies.

Hair growth. Several trials in androgenetic alopecia have reported modest improvements with low-level laser therapy devices; the effect is small relative to licensed medicines (minoxidil, finasteride) for that indication.

Wound healing in chronic wounds. Some evidence in clinical populations; modest in size.

Where the evidence is thinnest#

Thyroid function, mitochondrial health in general, longevity, energy. Most of these claims come from in vitro mitochondrial studies (cytochrome c oxidase absorption peak at certain wavelengths), animal studies, or small human pilots. The leap from mechanism to "shine this device at yourself for 10 minutes a day to optimise mitochondria" is large.

Wavelengths and devices#

The wavelengths most studied are around 630–660 nm (red, penetrates 1–2 mm) and 800–880 nm (near-infrared, penetrates further). Most studied effects are at specific wavelengths chosen to match mitochondrial chromophore absorption.

Consumer device variability is enormous:

• Power output (irradiance, measured in mW/cm²) varies by orders of magnitude between products with similar marketing.
• LED arrays at low power can advertise the right wavelength but deliver too little energy to reach studied effect sizes.
• Devices with FDA or CE clearances for specific indications are more credible than generic wellness panels.

What to expect#

For documented effects (acne, modest skin improvements, minor wound healing), realistic expectations are modest. The biggest disappointments come from people expecting transformative effects on energy or systemic metabolism.

A reasonable rule: if the trial that supports a claim is in vitro or in mice, treat human application as untested. If it's a small pilot, treat it as preliminary.

Safety#

Red and near-infrared light at consumer doses is generally well-tolerated. Eye protection is sensible: looking directly into bright LED panels for extended periods isn't great for retinal exposure regardless of wavelength. Pregnancy and active malignancy are typically exclusion criteria in the clinical trials.

The honest summary#

Photobiomodulation is real biology with modest, documented effects in specific indications, mostly dermatological. Most consumer marketing extrapolates well beyond what published human trials support. For most readers, the most reliable lights for health remain free: outdoor sunlight in the morning, modest dimming in the evening.