Older castle keeper holding a red lantern close to a drawbridge hinge, surrounded by rusted chains.
Back to the library

Pain & inflammation

Red light for joint and muscle pain

Near-infrared light applied to a sore joint has a reasonable body of placebo-controlled evidence, with the clearest picture in knee osteoarthritis and neck pain. The pooled trials reported lower pain scores than placebo, with bigger differences when the dose matched published recommendations. Almost all of that evidence comes from a handheld probe held against the joint, not from a full-body panel, and the two deliver very different amounts of energy to a small area. For members the honest position is that the underlying effect looks real at the right dose, while panel-delivered light to a joint remains largely untested.

Evidence so far
Moderate
Reading timeAbout 12 minutes

Start here

Why this is being studied

This use is about aching joints and tired or sore muscle, and the light that matters here is near-infrared at roughly 780 to 905 nm rather than visible red. Near-infrared scatters less and so travels further, reaching subcutaneous tissue, superficial muscle and the capsule around a joint. The proposal is that light absorbed by mitochondria in inflamed tissue may support a shift away from inflammatory signalling and towards repair, and this is the use where dose has been studied most carefully.

Where the evidence stands

Flag text

The mechanism

What is going on, and why this might help

What is going on in a sore joint or muscle

An osteoarthritic knee is more than worn cartilage. The joint lining is inflamed, the surrounding muscle guards and weakens, and the small nerve fibres carrying pain become sensitised. Overworked muscle shows a related picture: inflammatory signalling, oxidative stress and mitochondria working below par. In both cases there are living cells under strain rather than tissue that is simply gone.

What the light is absorbed by

Cytochrome c oxidase, the last enzyme of the mitochondrial chain, absorbs in the red and near-infrared. Under stress, nitric oxide competes with oxygen at that enzyme and slows respiration. The leading model is that absorbed photons displace the inhibitory nitric oxide, so oxygen use and ATP production rise briefly, the released nitric oxide relaxes local blood vessels and increases blood flow, and a small burst of reactive oxygen species acts as a signal that shifts the cell's redox state. Laboratory and animal work reports lower pro-inflammatory cytokines, less swelling and a shift in immune cell behaviour towards resolution. Light-sensitive ion channels are also proposed as a route by which longer near-infrared wavelengths may affect pain signalling.

The biphasic dose response

This is the use where dose matters most visibly. The pooled knee trials that used 4 to 8 joules per point at 785 to 860 nm, or 1 to 3 joules per point at 904 nm, reported roughly double the effect of trials outside that window. The published dose curve rises, plateaus and then falls, so overshooting is not neutral. Recommended schedules run at two or three sessions a week across four to eight weeks.

How far the light actually goes

Near-infrared at 800 to 905 nm scatters less than red and reaches subcutaneous tissue, superficial muscle and the joint capsule, measured in millimetres to a couple of centimetres in favourable conditions. That is enough for a knee, an elbow or a shin, and not enough for a hip or an organ. The bigger gap is delivery: a probe concentrates several joules into a spot a centimetre across, while a panel spreads a much lower intensity across the whole body. That difference, rather than the biology, is why panel-delivered light to joints is graded early.

Explainer illustration

The studies

Every study is laid out the same way: what the researchers wanted to find out, who took part, what they actually did, what they found, why it might work and where the limits are. No jargon, and a link to the original if you want to read it yourself.

Each study is laid out the same way: what the researchers asked, who took part, what they did, what they found, why it might work, and where the limits are. The clinician layer at the end of each one carries the numbers.

Paper 1

Efficacy of low-level laser therapy on pain and disability in knee osteoarthritis: systematic review and meta-analysis of randomised placebo-controlled trials

Stausholm et al. · BMJ Open · 2019 · Systematic review and meta-analysis of 22 randomised placebo-controlled trials, n=1,063

Read the paper

What they asked

Knee osteoarthritis is one of the commonest causes of long-term pain in adults, and most guidelines do not recommend light therapy for it because earlier reviews found inconsistent results. This Norwegian group suspected the inconsistency came from dose rather than from the light. They asked whether red or near-infrared light reduces knee pain and disability compared with placebo, and whether the answer changes once you separate trials that used recommended doses from those that did not.

Who took part

The review pooled 22 randomised trials covering 1,063 adults with diagnosed knee osteoarthritis, mostly middle-aged or older and many with years of knee pain behind them. Every trial compared active light with a placebo device that looked and sounded identical but emitted no therapeutic light. Trials came from several countries and used a mix of laser and LED devices in the red and near-infrared range. The protocol was registered in advance, which limits the temptation to change the question once the results are visible.

What they did

The team searched the main medical databases, extracted pain scores on a 0 to 100 mm scale plus disability scores, and assessed each trial for risk of bias. They then read the dosing details of every trial and classified it as inside or outside the recommended window, which for the knee means 4 to 8 joules per point at 785 to 860 nm or 1 to 3 joules per point at 904 nm. Results were pooled separately for the end of the course and for follow-up visits weeks later.

What they found

Across all trials, pain was about 14 mm lower with active light than placebo at the end of treatment. In the trials using recommended doses the gap was larger: roughly 19 mm at the end of the course and up to about 32 mm two to four weeks afterwards. Trials outside the dose window showed much smaller effects. Disability scores were also lower in the active groups, and the effect grew rather than faded at follow-up. No adverse events were reported in any included trial.

What it means

This is the strongest single piece of evidence that near-infrared light may support a reduction in joint pain, and it is unusual in showing that dose explains much of the disagreement in the older literature. The detail that matters most for members is delivery: these were probes pressed against the knee, concentrating several joules into each small point. A panel across the room spreads a much lower intensity over the whole body.

Limits

A meta-analysis inherits the weaknesses of what it pools, and many of these trials were small with follow-up rarely beyond three months. The dose categories were applied after the fact, which is reasonable but not the same as a trial designed to compare doses head to head. Some authors are long-standing researchers in the field and two had links to the body whose dose recommendations were being tested. Blinding is imperfect because active devices can feel warm. The findings apply to the knee treated with a probe, not to other joints or to panels.

Paper 2

Effects of low-level laser therapy on pain in patients with musculoskeletal disorders: a systematic review and meta-analysis

Clijsen et al. · European Journal of Physical and Rehabilitation Medicine · 2017 · Systematic review and meta-analysis of 18 randomised trials, n=1,462

Read the paper

What they asked

Knee osteoarthritis is only one of many painful musculoskeletal problems people bring to a physiotherapist. This Swiss and Belgian team asked the broader question: across shoulders, necks, backs, tendons and knees, does low-level laser reduce pain compared with placebo, and how much does following published dose recommendations change the answer.

Who took part

Eighteen randomised trials covering 1,462 adults with a range of musculoskeletal disorders were pooled. Conditions included tendinopathies, neck and shoulder pain and osteoarthritis, and participants were generally adults living with pain rather than people recovering from acute injury. Every trial used a placebo-controlled design.

What they did

The reviewers searched the literature, extracted pain outcomes and converted them to a common standardised scale so that trials using different questionnaires could be compared. They assessed methodological quality and, as in the knee review, separated trials that had followed the World Association for Photobiomodulation Therapy dose recommendations from those that had not.

What they found

Pooled across all trials, pain was lower with active light by about 0.85 standardised units, a moderate to large difference by the conventions of this kind of analysis. In the subgroup of trials that followed the recommended dosing, the pooled effect roughly doubled. The reviewers noted wide variation between trials, which is expected when several different conditions and devices are combined.

What it means

Taken with the knee review, this supports two things: that the effect is not confined to a single joint, and that dose is the single most important variable separating trials that report something from trials that do not. It also underlines how much of this literature comes from clinical settings where a trained person holds a device against a specific spot for a specified number of joules.

Limits

Pooling several different conditions into one number makes the result harder to apply to any individual situation, and the variation between trials was substantial. Standardised effect sizes are statistically convenient but do not translate neatly into how much better someone would feel. Many trials were small, and short follow-up is again common. As with all of this literature, it describes probe-delivered light, not panels.

Paper 3

Efficacy of low-level laser therapy in the management of neck pain: a systematic review and meta-analysis of randomised placebo or active-treatment controlled trials

Chow et al. · The Lancet · 2009 · Systematic review and meta-analysis of randomised placebo or active-controlled trials

Read the paper

What they asked

Neck pain is extremely common and often persistent, and the available options are limited. This review, published in a general medical journal rather than a specialist one, asked whether low-level laser reduces neck pain compared with placebo or with other active treatments, and crucially whether any relief outlasts the course of sessions.

Who took part

The review pooled randomised trials of adults with acute or chronic neck pain, drawn from several countries. Participants were people with mechanical or chronic non-specific neck pain rather than those with a defined injury or surgical problem. Trials used placebo devices identical in appearance to the active ones, and some also compared against other treatments.

What they did

The authors searched the literature systematically, assessed trial quality, and pooled pain outcomes both immediately after a course of treatment and at later follow-up points, out to several months. They also examined the dose and wavelength used in each trial.

What they found

The review reported reduced neck pain with active laser compared with placebo immediately after treatment, and reported that relief in the chronic neck pain trials persisted at follow-up for up to 22 weeks after the course ended. Adverse effects were mild and comparable to placebo.

What it means

The lasting element is the interesting part. Pain relief that continues for months after the sessions stop fits a slow tissue process better than a simple analgesic effect, which is consistent with the pattern seen in the knee review where the largest differences appeared weeks after the course finished. It is also the reason this field talks about courses of sessions rather than single appointments.

Limits

The review is now well over a decade old and included trials that varied widely in device, dose and site of application. Neck pain is a broad label covering several different problems, so pooling can obscure who benefited. Blinding difficulties apply here as elsewhere. And again, every one of these trials used a focused laser applied to specific points on the neck by a practitioner.