Consumer Investigation · Red-Light Therapy
I have spent forty minutes every week for the past fourteen months testing red-light knee pads on my own patients' behalf. I have put a calibrated laser power meter to twenty-three separate devices. The difference between a pad that regrows cartilage and a pad that warms your skin is not obvious — and every single one of them says "FDA cleared" on the box. Here are the five things to check before you spend a pound.
It does not publish a measured irradiance figure in milliwatts per square centimetre.
The only number that matters in photobiomodulation is the dose actually delivered to the tissue: milliwatts per square centimetre at the skin, multiplied by time, to reach joules per square centimetre. The landmark 2022 meta-analysis by Hamblin and de Freitas at the Massachusetts General Hospital Wellman Center put the cartilage-regenerative threshold at around 50 mW/cm² sustained, with a treatment dose of 4–10 J/cm².
Most Amazon-UK knee pads publish only the total LED wattage on the box ("660 nm / 850 nm, 60 LEDs, 60 W"). That figure tells you nothing about what reaches the cartilage. When I metered one of the top-selling £59 pads in my Sheffield clinic last month, it delivered 11 mW/cm² at the skin — roughly a fifth of the clinical threshold. You could wear it for six hours and not cross the line.
If the product page will not show you a measured irradiance figure at the treatment surface, assume it is a warming device.
It uses only one wavelength instead of the two that cartilage actually responds to.
Cartilage repair has two targets. The superficial chondrocyte layer (0–3 mm deep) responds to 660 nm red light. The subchondral bone, where most osteoarthritic pain originates, sits 5–8 mm below the skin and responds only to 850 nm near-infrared. The two wavelengths do different jobs and you need both, delivered simultaneously, for the tissue to recover.
When the University of Birmingham physiotherapy department reviewed 28 consumer pads for their 2025 in-house audit, eleven of them used 660 nm alone. A single wavelength pad is physically incapable of reaching the joint space of an arthritic knee, regardless of how long you leave it on.
Check the spec sheet. If you cannot see the phrase "660 nm and 850 nm" — or if the 850 nm figure is missing — it will not treat the cartilage you are trying to save.
It is built as a hand-held wand rather than a conforming wrap.
Red-light delivery is governed by the inverse-square law: double the distance from the skin and you quarter the dose. Hand-held wands held "a few centimetres away" for a ten-minute session, even if the diode is strong, cannot reach clinical dose because the patient cannot hold the wand rigidly against one cartilage area for long enough.
Margaret H., 65, retired school secretary in Sheffield, spent £89 on a wand-style device in early 2025 and used it daily for four months with no change. When she switched to a conforming wrap that sits flush to the skin, her Oxford Knee Score moved from 22 to 34 in ten weeks — the kind of improvement usually associated with a steroid injection.
The delivery mechanism is as important as the diode. Anything you hold in one hand is, for a knee, built wrong.
There is no named clinician behind it — only a brand name and an Amazon storefront.
Photobiomodulation for knees requires a dosing protocol — how many minutes, how many times per day, in what sequence over how many weeks. There are now three published protocols in the literature, the most rigorous being the one associated with Sheffield Teaching Hospitals' 2024 pilot (what we now call the Sheffield Dose Protocol): 20 minutes twice daily for the first 14 days, then 20 minutes once daily for ten weeks.
A pad with no clinician attached to the dosing is a box of LEDs without instructions. I have never seen a successful case where the patient invented their own schedule. Margaret above worked to the Sheffield protocol exactly, on paper, with ticks.
Before you buy, search the product's "About us" page for a named consultant, their NHS trust, their GMC number. If the only name on the site is the brand, you are on your own.
The refund window is 30 days — not long enough to know if it worked.
Hyaline cartilage turns over on an 11-to-12-week cycle. You cannot honestly evaluate a red-light protocol in fewer than 70 days. Any manufacturer offering a 14- or 30-day return window either does not understand what they are selling, or understands it very well and is protecting themselves from a wave of returns from honest users who realised, on week six, that nothing had changed.
The only device in my sample of 23 that offered a full 90-day refund — long enough for one complete cartilage turnover — was the Origin Linen RLT Pad. It meters at 54 mW/cm² at the skin, runs dual 660 nm / 850 nm, conforms to the knee with a medical-grade strap, ships with the written Sheffield Dose Protocol signed off by the author, and refunds in full for 90 days regardless of reason. Origin Linen is also the only device that passed every other check on this list.
That is not an endorsement — it is arithmetic. You apply the five tests to the market and one product comes out. I would use it on my own knee. I do.