Some of it has real trials behind it. The device in the ad is a different question.
Short answer: for knee arthritis, light from a clinic laser, at a measured dose, applied over weeks, eased pain more than a placebo laser in pooled trials — modestly, and with researchers still disagreeing about how sure to be. Laser on acupuncture points did nothing. For skin there is one decent trial. None of that tells you whether a particular handheld device works.
The trials that found an effect used lasers with a stated wavelength and power, delivering a measured amount of energy to each spot — the 2019 review names 4–8 joules per spot at 785–860 nm, which is near-infrared and mostly invisible, not the red you can see. A device that does not state its wavelength and output cannot be matched to those doses, so the trials say nothing about it either way.
A Facebook advert for a red-light “pen” promises to erase arthritis pain. We set each of its claims beside the record: the InfraBeam pen’s claims, checked.
Red light therapy · evidence checked 2026-10-24
Grades are calculated from the studies listed, by fixed rules.
Clinic lasers at measured doses, applied to the knee, over weeks. The 2019 review found pain fell more than with placebo at 4–8 J per spot (785–860 nm, near-infrared — not visible red); the 2024 review agreed but rated the certainty "very low"; the older Cochrane review found conflicting trials. Major arthritis guidelines did not recommend it as of 2019.
A blinded trial of 282 people: laser acupuncture did no better than a sham laser for pain or function at 12 weeks.
One randomised trial against untreated controls (no sham light), 30 full-face sessions with large panels — not a pen. Photo ratings were blinded; participants were not.