Photobiomodulation Therapy for Knee Pain: Clinical Evidence and Guidelines

Umbrella Reviews, Network Meta-Analyses, and Systematic Reviews

Umbrella Reviews and Network Meta-Analyses

The 2025 umbrella review by Son et al., published in Systematic Reviews, evaluated photobiomodulation across every health condition for which trial data exists, synthesizing 15 meta-analyses that covered 204 RCTs and over 9,000 participants. Knee OA disability improvement came out at moderate certainty (eSMD 0.65), one of the strongest evidence domains the review identified for the therapy anywhere.

The anchor of the modality comparisons is the 2025 network meta-analysis by Lan et al., published in Aging Clinical and Experimental Research. It pooled 32 RCTs and 2,078 participants and ran low-level light therapy directly against electrical stimulation, extracorporeal shockwave therapy, thermotherapy, cryotherapy, and ultrasound. Low-level light therapy ranked first for pain reduction on both VAS and WOMAC pain, and first for joint function improvement (SUCRA 79.8). Shockwave therapy came second for pain; ultrasound showed no significant benefit at all.

A second comparison reads slightly differently. Cao et al. (2024), analyzing 24 RCTs and 2,582 patients across non-pharmacological treatments, ranked laser therapy third for VAS and WOMAC total, behind shockwave and needle-knife therapy, while placing it first for stiffness reduction. The two analyses disagree on exact rank, which is what you would expect from different comparator sets, but both put laser therapy in the top tier.

Systematic Reviews and Meta-Analyses

Li et al. (2023), in Archives of Physical Medicine and Rehabilitation, pooled 14 RCTs and went after the most clinically relevant question in the field: does adding photobiomodulation to the exercise programs guidelines already recommend produce anything extra? Across WOMAC total, WOMAC pain, WOMAC function, VAS pain, and knee range of motion in older adults, the combination beat exercise alone. The answer is yes.

Stausholm et al. (2019), in BMJ Open, covered 22 placebo-controlled trials and established that photobiomodulation produces significant, dose-dependent pain relief in knee osteoarthritis. The dose-dependence is the part that carries clinical weight, because it means the parameters (wavelength, irradiance, dose per point) decide whether outcomes clear the bar for significance. Trials dosing below the World Association for Photobiomodulation (WALT) guidelines produced weaker results; trials meeting or exceeding them produced significant ones.

Oliveira et al. (2024), writing in Physical Therapy, analyzed 10 RCTs and 542 participants and confirmed significant reductions in pain intensity and improvements in disability. The authors graded the certainty of evidence as very low, citing bias risk in the included studies, and recommended photobiomodulation as a complement to established therapies rather than a standalone treatment. That recommendation runs with the grain of the whole literature: the therapy is at its strongest bolted onto active rehabilitation. A 2021 systematic review by Vassão et al. in Lasers in Medical Science, focused specifically on the photobiomodulation-plus-exercise combination, found consistent positive effects on pain and functional capacity, which points the same way.

Landmark Individual RCTs

Alqualo-Costa et al. (2021): Largest Knee-Specific RCT

The largest knee-specific trial reads as a clean four-arm comparison. This double-blind, placebo-controlled study in Clinical Rehabilitation enrolled 168 patients with knee OA and tested photobiomodulation alone, interferential current alone, the two combined, and placebo. Photobiomodulation alone reduced pain at rest versus placebo at every follow-up point: immediately after treatment, at 3 months, and at 6 months. The combined photobiomodulation-plus-interferential-current arm produced the strongest results. The device was an 850nm LED, which makes the trial a direct demonstration that LED-based near-infrared delivery sustains clinical outcomes at this sample size.

Stausholm et al. (2022): 12-Month Follow-Up

This randomized, placebo-controlled trial in the Journal of Clinical Medicine combined 904nm laser therapy with strength training in 50 patients over 8 weeks, then followed them for a year. Both groups achieved substantial pain reduction, and the trial found no significant between-group difference on its primary pain outcomes. The laser group did produce two clinically important between-group findings: significantly reduced use of analgesics and NSAIDs, and significantly better performance on the sit-to-stand functional test versus placebo, both persisting at 12 months. The authors flagged a baseline imbalance, with the placebo group reporting more intense pain at the start and therefore more room to improve, which complicates the pain comparison. For a patient worried about long-term medication reliance, the durable drop in analgesic use is the finding that lands.

Dos Santos Maciel et al. (2025): WALT Protocol Validation

This double-blind, three-group RCT enrolled 65 patients and ran 790nm laser at WALT dosing (4 J per point). The photobiomodulation group achieved significant pain reduction and WOMAC improvement against both the sham and control groups, which confirms in a rigorous design what the dose-finding literature predicts: hit the established parameters and the outcomes follow.

Pinto et al. (2022): Individualized Dosimetry

This double-blind pilot took 31 patients and customized the dose to each one's BMI and skin color. Pain dropped significantly at sessions 4, 5, and 10 and stayed lower at 6 weeks post-treatment versus placebo, quality of life improved significantly, and the treatment group also showed gains in dopamine level and microcirculation. It is a small pilot, so the findings are preliminary, but it plants a clinically interesting flag: tailoring the dose to the patient may beat one-size-fits-all protocols.

Guidini Lima et al. (2022): Strength Training Augmentation

This placebo-controlled RCT of 28 patients found 808nm photobiomodulation combined with strength training significantly more effective than strength training alone across pain, WOMAC scores, knee flexion, and quadriceps strength, with effects holding at 30-day follow-up.

Gavish et al. (2021): Patellofemoral Pain (Non-OA)

This double-blind, sham-controlled RCT in combat soldiers with anterior knee pain found that photobiomodulation plus physiotherapy significantly reduced pain versus sham plus physiotherapy at 4 weeks, with Kujala functional scores improving significantly only in the treatment group at 3-month follow-up. It is the first sham-controlled RCT of photobiomodulation for non-OA anterior knee pain, which pushes the evidence base past osteoarthritis for the first time.

Head-to-Head Comparisons with Other Physical Modalities

Pasin et al. (2025) ran a four-arm RCT in 120 knee OA patients comparing LLLT, extracorporeal shockwave therapy (ESWT), pulsed electromagnetic field therapy (PEMF), and a control. Both ESWT and LLLT produced significant gains in pain (VAS), function (WOMAC), quality of life (SF-36), and mobility (TUG), with no significant difference between the two and both beating PEMF.

Şen et al. (2025) enrolled 90 patients and compared high-intensity laser therapy and LLLT, both combined with exercise. Both arms improved significantly in VAS, WOMAC, and flexion ROM, with no significant between-group difference on most measures, and both showed within-group increases in ultrasound-measured femoral cartilage thickness. That thickness signal needs a careful read, because the trial had no placebo arm, so the change cannot be separated from the concurrent exercise. The one placebo-controlled low-level laser trial that measured the same endpoint, Stausholm et al. (2022), found no significant effect on cartilage thickness. The thickness-imaging evidence for cartilage preservation is therefore mixed. The stronger cartilage-protection evidence is biochemical, reduced MMP-3, MMP-13, and CTX-II measured in patients, rather than dimensional.

Post-Surgical Evidence

Bahrami et al. (2023) enrolled 45 patients after total knee arthroplasty into a three-arm RCT and found markedly higher range of motion in the laser group, 116.8° against 92.3° for controls at 3 months, alongside significantly lower swelling, reduced opioid consumption, and higher functional scores. Chia et al. (2025) confirmed significant post-TKA swelling reduction using objective bioimpedance measurement. For a therapy that costs nothing past the device, carries no drug interactions, and runs alongside standard post-surgical rehabilitation, that is a clinically meaningful pairing of trials.

Dosing Guidelines and Why Parameters Matter

The World Association for Photobiomodulation publishes recommended dosing for knee OA at both 780–860nm and 904nm. The Stausholm et al. (2019) meta-analysis confirmed that trials adhering to those guidelines produced significantly stronger outcomes than trials running suboptimal parameters. Dose-dependence is the clinical takeaway that does the most work in this literature: inadequate irradiance is the most common reason a trial lands in the negative column.

For infrared specifically, the 2024 Fan et al. network meta-analysis identified 904–905nm as the most effective wavelength, with 785–850nm also significantly outperforming sham. Together these dose-finding studies define the parameters effective treatment actually requires. For how LED and laser devices deliver those parameters, see LED vs. Laser for Knee Pain. For the complete safety profile, see Photobiomodulation Safety for Knee Pain.

Limitations of the Evidence

The evidence is strong in volume and uneven in quality. The 2024 Oliveira review rated its pooled certainty very low on bias risk, and the umbrella review found moderate certainty for knee OA disability but lower certainty for most other photobiomodulation applications. The biggest positive signals come from trials that combined photobiomodulation with exercise, which makes the light therapy's independent contribution hard to isolate cleanly. Dose heterogeneity across studies, with wavelength, irradiance, and session counts all varying, is at once the explanation for inconsistent results and the obstacle to a single standardized protocol.

The cartilage question is where the honesty has to be most precise. Direct ultrasound imaging of cartilage thickness is mixed: the Şen 2025 low-level laser arm showed a within-group increase without a placebo control, while the placebo-controlled Stausholm 2022 trial found no significant effect. The firmer cartilage-protection evidence is biochemical, reduced cartilage-degrading enzymes (MMP-3, MMP-13) and breakdown markers (CTX-II) in patients, not dimensional. And no dedicated long-term trial has tested whether sustained photobiomodulation delays progression to joint replacement. The case for pain and function relief is well-supported; the disease-modification case is still an open and active research question.

Conclusion

The arc of the evidence, across more than two decades, multiple countries, and independent research teams, points consistently in one direction. Photobiomodulation reduces knee pain, improves joint function, reduces medication dependence, and produces benefits that outlast the treatment period by months. With a 2025 network meta-analysis ranking it first among all physical modalities tested and a 2025 umbrella review of 204 RCTs placing knee OA disability among the strongest evidence domains, the live question has shifted from whether it works to how to run it well. Dose matters. Wavelength matters. Treatment consistency matters. For patients in the gap between medication management and surgical intervention, the clinical evidence establishes photobiomodulation as among the most effective non-pharmacological options available, used, as the trials show again and again, alongside exercise rather than in place of it. The broader evidence on red light therapy for osteoarthritis extends this picture to the other joints the disease affects.