REF-1767 · Nascimento et al., 2024 · International Journal of Exercise Science, · Paper · PubMed · Free full text
Red and near-infrared light from laser and LED clusters placed over the thigh and calf muscles in the minutes before a run did not improve running time-trial or time-to-exhaustion performance against placebo: SMD 0.13 (95% CI −0.03 to 0.29), p = 0.11, in a meta-analysis of 12 randomised trials (312 healthy adults, 11 of 12 trials men only), with no dose-response across total doses of 300-3000 J
Main caveat: Mostly small crossover trials (12-26 people in 10 of 12); 11 of 12 trials in men; the most trained sample is one group of young football players.
Nascimento and colleagues (University of Northern Paraná, Brazil) asked whether photobiomodulation, red to infrared light (600-1000 nm in their definition) from low-level lasers and/or light-emitting diodes, improves running performance, whether it matters if the light is added to a training programme, and whether a larger dose does more.
They searched five databases to June 2023 and kept randomised trials in healthy adults that compared active light with a placebo device on a running time trial or time to exhaustion. Twelve trials published 2012-2020 qualified, ten of them crossovers.
In every trial the light came from a cluster of laser diodes, LEDs or both, at 640-950 nm, aimed at sites on the quadriceps, hamstrings and calf (one trial also the gluteus maximus), from immediately to 20 min before the run; total doses for both legs ranged from 300 to 3000 J, most around 600-1000 J.
Pooled, active light did not beat placebo (SMD 0.13, 95% CI −0.03 to 0.29, p = 0.11), neither on its own nor alongside training, and the meta-regression found no dose-response (p = 0.82).
The authors conclude that the light does not improve time-trial or time-to-exhaustion running, while noting that the point estimates mostly leaned towards light and that doses above 1000 J and more training studies are untested ground.
Participants
12 trials; sample sizes 12 to 96 per trial (one parallel trial of 96, 48 per arm); 312 participants by the sum of Table 2, which matches the Discussion's 91 + 221.
Ages about 18-34; 11 trials men only, 1 trial men and women.
Training status by Table 2: untrained men (2 trials), physically or moderately active men (4), healthy men (1), recreational runners (2), male runners or triathletes (1), high-level football players (1), men and women of unstated status (1).
Protocol
Search: PubMed/Medline, Embase, Scopus, SPORTDiscus and Web of Science to June 2023, filtered to randomised trials, no language limit; one author searched and a second checked.
Inclusion: randomised trials, parallel or crossover, of photobiomodulation by laser and/or LED against placebo in healthy adults, with a running time trial or time to exhaustion. Excluded: older adults, supplement or steroid users, light combined with another therapy (cryotherapy, massage), and indirect outcomes such as lactate, VO2max or running economy.
What was applied (Table 2): clusters of 5 laser diodes (810 nm), of 12 super-pulsed laser and LED diodes (640, 875 and 905 nm), of 5 lasers plus 28 LEDs (670-950 nm), or of 69-152 LEDs (660, 850 or 880 nm); 5 to 17 sites over the quadriceps, hamstrings and gastrocnemius or plantar flexors; total dose for both legs 300-3000 J; timing immediately before, 5 min, 5-10 min or 20 min before the run.Whether the device touched the skin is not stated in the text we read. In one training trial the light was given before and after each session for 4, 8 and 12 weeks.
Risk of bias: PEDro scale (0-10; 9-10 excellent, 6-8 good, 4-5 moderate, 0-3 poor) by two raters, κ = 0.716 (95% CI 0.52-0.88).
Analysis: random-effects SMD with 95% CI in Comprehensive Meta-Analysis 2.2; I²; training (present or absent) as a moderator by Q-test; meta-regression on total dose; funnel plot and Duval and Tweedie trim and fill.
Funding and interests: one author held a research productivity fellowship from Brazil's National Council for Scientific and Technological Development (CNPq); the text we read carries no conflict-of-interest statement.
Key Results
| Comparison, active light vs placebo | SMD (95% CI) | p |
|---|---|---|
| All trials (19 outcomes) | 0.13 (−0.03 to 0.29) | 0.11 |
| Light alone, no training (Results text) | 0.14 (−0.05 to 0.33) | 0.15 |
| Light plus running training (Results text) | 0.11 (−0.18 to 0.39) | 0.46 |
| Training vs no training, subgroup difference | — | 0.797 |
| Meta-regression on total dose (slope) | — | 0.82 |
The abstract gives different subgroup numbers: with training SMD 0.16 (p = 0.38), without training 0.11 (p = 0.25); the Results text and the abstract disagree, and the Discussion quotes the subgroup p as 0.7. None of the versions is significant.
Heterogeneity: reported as I² = 0.0 (the text attaches p < 0.001 to it, which we cannot interpret).
Publication bias: trim and fill trimmed no studies.
Quality: mean PEDro 7.9; no trial rated poor or moderate.
Trial by trial: by the Discussion's count 7 trials (221 participants) reported a possible benefit of light on test duration and 5 (91 participants) none; Table 2 marks one of those five as better with light on its 5-km time trial. One dose-comparison trial in recreational runners found benefit at 420 and 840 J but not at 1680 J (Table 2; the Discussion says only 840 J helped); another in active men found no benefit at any of three doses.
Training trials: in 96 men and women, light before and after each session of a 3-times-weekly running programme lengthened time to exhaustion in a progressive test against placebo, by the authors' account; in 30 untrained men after 8 weeks there was no difference in time to exhaustion (Table 2 marks the 5-km time trial in this trial as better with light, the Discussion calls it null).
Practical Application
Red or near-infrared light on the legs before a run is not shown to make the run faster or longer: across 12 placebo-controlled trials the pooled effect was SMD 0.13 (95% CI −0.03 to 0.29), not significant, with light alone or added to training. The light in these trials was a laser or LED cluster applied over the thigh and calf muscles from immediately to 20 min before the test, not a lamp in the room or a whole-body bed.
More light was not better within the doses tested. Total doses ran from 300 to 3000 J across both legs, and the meta-regression found no dose-response (p = 0.82); one trial suggested a middle dose did better than the highest.
Inference, not tested here: the interval runs from about zero to a small gain (upper bound SMD 0.29), so a small benefit is not ruled out, but the base has no grounds to recommend a device for running performance; say "tested in 12 trials and not shown", never "a myth" (SYNTH-83). Harms were not reported.
Training blocks with light are barely tested: two trials, one positive and one null, so whether weeks of light treatment add to a running programme is open.
Limitations
Mostly small crossover trials (12-26 people in 10 of 12); 11 of 12 trials in men; the most trained sample is one group of young football players.
The abstract and the Results text give different subgroup estimates, the Discussion a third p value, and Table 2 and the Discussion disagree on two trials' results; the overall estimate is the same in both.
The text cites an earlier meta-analysis of 12 crossover trials with a pooled running SMD of 0.17 (not significant) without naming its authors in the text we read; its place in the alphabetical reference list (reference 7, between Dellagrana and Duval) and its crossover-only design fit REF-1766 Dutra 2022, so many of the same trials are likely in both and the two are not independent confirmations (our reading).Several included trials share lead authors (Dellagrana ×2, Miranda ×2, Peserico ×2) and the same devices; one included trial is by Malta et al., a surname also on the 2022 Sports Medicine meta-analysis REF-1766 Dutra 2022 (initials not given here, so the same person is our reading).
Per-trial PEDro scores, device makers and the trials' own funding are not given in the text we read; whether the cluster touched the skin is not stated.
Only time trials and time to exhaustion; no race, no outcome beyond a single test, and only 6 trials used total doses above 1000 J.
How this entry was checked and changed
Why this level
a meta-analysis of randomised trials whose risk of bias we read in the full text (PEDro mean 7.9, no trial below 6), held at A- because the trials are mostly small crossovers (12-26 people in 10 of 12) and only two tested light added to weeks of training
Content correction 2026-11-11
2026-11-11: added from the full text; level A- (meta-analysis of mostly small randomised crossovers, PEDro mean 7.9, read in full).