A knee brace works for arthritis, barely for runner's knee, and not at all for a meniscus tear. Buy by diagnosis

Amazon sells one product called a knee brace, in a thousand listings, for every knee. The listings all say the same things: meniscus, ACL, arthritis, running, pain relief. The evidence does not. It says a brace helps an arthritic knee a little and with a guideline behind it, helps a runner’s kneecap about half a point on a scale of ten if at all, and has never been tested on the meniscus tear most people over forty are told they have. So the way to buy one is by diagnosis, and the first job of this page is to split the one product back into the four it actually is.

Four things called a knee brace

• A compression sleeve is a tube of knit or neoprene. It warms, compresses, and gives the knee something to feel against, which is where its effect on pain seems to come from. Modvel, Bracoo and the Bauerfeind GenuTrain are sleeves, from ten dollars to a hundred.
• A hinged or stabilised brace is a sleeve or wrap with rigid or spring bars down each side and usually a gel ring around the kneecap. It resists sideways bending. NEENCA’s hinged brace and the DR. BRACE wrap are this type.
• An unloader, or offloader, brace has a rigid frame and a hinge that bends the knee slightly away from the worn side, shifting load from the inside (medial) compartment, where most arthritis is, to the outside. This is the type with the trials. Ovation’s Game Changer is the Amazon version; the ones in the trials were custom-fitted.
• A patella strap is a padded band under the kneecap that presses the patellar tendon. Bodyprox’s is the one everyone buys. It is for tendon pain and runner’s knee, not for arthritis or instability.

Everything else is a variation: open-patella versus closed, wrap versus pull-on, plus-size, copper-infused (the copper does nothing the fabric does not), “infrared” (a claim we skip; see our heating pad page on the word), heated, and so on.

Knee brace for arthritis: the one with the evidence

Knee osteoarthritis is where bracing has been tested properly, and where it has a guideline. The AAOS 2021 guideline on knee osteoarthritis says a brace “could be used to improve function, pain and quality of life”, at moderate strength, on ten trials, with the harm summarised as skin irritation and discomfort. The American College of Rheumatology’s 2019 guideline goes further: a tibiofemoral brace is strongly recommended for tibiofemoral osteoarthritis, a patellofemoral brace conditionally for patellofemoral disease. The OARSI 2019 guideline did not recommend bracing, on what it judged poor-quality evidence. Three guidelines, two for, one abstaining: that is as much agreement as a thirty-dollar product on this shelf ever gets.

The trials behind it are worth knowing because they decide which brace to buy. Kirkley and colleagues, in the Journal of Bone and Joint Surgery in 1999, randomised 119 people with varus medial-compartment arthritis to medical treatment alone, a neoprene sleeve, or a custom valgus unloader brace. At six months both the sleeve and the unloader beat no brace on pain, function and walking tests, and the unloader beat the sleeve on pain after a six-minute walk and a stair climb. That is the whole hierarchy in one trial: something is better than nothing, and the unloader is better than the sleeve for the loaded moments. Moyer’s 2015 meta-analysis of six trials put the pooled pain effect at 0.33 standard deviations and function at 0.22, with the effect against no brace about half a standard deviation, and noted compliance of 45 to 100% and minor complications in up to a quarter of users. Brouwer’s 12-month trial of 117 people found the brace group walked further but 25 of 60 abandoned the brace, mostly for lack of benefit. The Cochrane review of the same literature called it inconclusive, with the long-term effect unknown.

Sleeves have their own evidence. Cudejko’s 2018 meta-analysis of soft braces found pain reduced by about half a standard deviation and function by 0.39, in low-quality trials, and a follow-up study found the effect ran through reduced pain and less felt instability, not through the improved joint-position sense the sleeves are sold on. A tight sleeve and a loose one made no difference to instability.

What that means at the checkout: if you have medial-compartment arthritis and want the most relief when walking and climbing stairs, the unloader is the type the evidence favours, and it is also the type a quarter of people take off. If you want something you will actually wear all day, a sleeve gets most of the way for a fraction of the bulk. And the thing that works better than either is the exercise and weight management the same guidelines put first.

Knee brace for a meniscus tear: there is no trial

The queries behind this page are full of meniscus tears, and the listings all claim them. We searched for a trial of bracing for a degenerative meniscus tear and found none, and neither did the Cochrane and AAOS reviewers. What the meniscus literature does contain is three large trials comparing arthroscopic surgery with exercise or sham: Sihvonen’s FIDELITY trial (146 people, surgery no better than sham surgery at 12 months), Kise’s trial in the BMJ (140 people, exercise equal to surgery at two years and better for thigh strength), and the ESCAPE trial in JAMA (321 people, physiotherapy non-inferior to surgery, with 29% crossing over).

So the honest answer to “best knee brace for a meniscus tear” is that the treatment is the exercise programme, and a brace is a comfort device that may let you do it with less fear. A sleeve or a stabilised wrap for the swollen, wobbly-feeling weeks is reasonable; a brace bought as the treatment is a month lost. A knee that locks or cannot straighten is the exception, and it is a clinician’s, not ours.

Knee brace for running and patellofemoral pain

The Cochrane review of knee orthoses for patellofemoral pain pooled five trials and 368 people, all at high risk of bias, and graded every outcome very low certainty. Short-term pain was 0.46 points lower on a 0-to-10 scale with an orthosis, not statistically significant and below any threshold for mattering; function was no different. A brace versus a sleeve, a patella strap versus a sleeve, and a sleeve with a patellar ring versus a plain one all performed the same. In one sleeve trial, 16 of 44 knees (36%) reported discomfort or skin abrasion.

That is the evidence for the strap and the sleeve on a runner’s knee: nothing you could call an effect, and a one-in-three chance of a sore patch. Some runners swear by a strap, and the mechanism (changing where the tendon loads) is plausible for patellar tendinopathy specifically. It is cheap enough to find out. It is not a reason to skip the hip and quad strengthening the same review points to.

For patellofemoral arthritis, a different diagnosis, a six-week brace trial of 126 people found pain 1.3 points lower on a 10 cm scale and smaller bone-marrow lesions on MRI, which is the ACR’s reason for its conditional recommendation above.

Hinged braces, instability and sport

A hinged brace resists the knee bending sideways, which is what you want after a medial collateral ligament sprain or while a wobbly knee rebuilds the muscle around it. There is no good trial of the Amazon-class hinged brace for anything; the evidence is for rigid, fitted functional braces after ligament surgery, which is a different product and a surgeon’s prescription.

For preventing injury in sport, the evidence is against routine use. Sitler’s West Point study of 1,396 cadets playing intramural football found fewer knee and MCL injuries in braced players, but only among defensive players and with no reduction in severity. A 2010 review of one trial and five cohorts found one with fewer MCL injuries and two with more knee injuries in braced players, and concluded routine prophylactic bracing is unsupported. A brace is not a substitute for strength, landing mechanics and sleep.

Sizing: the thing that decides whether you wear it

The braces here are sized by circumference, and the number of inches above the kneecap is the maker’s choice. Bauerfeind’s listing has you measure the thigh and the calf a set distance from the centre of the knee with it slightly bent. NEENCA’s listings use four, five and a half, or seven inches above the knee centre depending on the model. The Aircast cuff on our ice machine page uses six. Bodyprox’s strap fits a 13 to 18 inch knee circumference, listing-stated. Modvel sells six sizes and shows the chart on the listing.

Two rules. Measure, on the day, standing, with a soft tape; do not guess from trouser size. And when the measurement sits between two sizes, size up for a sleeve you will wear for hours and down for a strap, because the complaints in every trial and most reviews are slipping, pinching and rubbing, and all three are sizing. A brace that has to be hitched up every fifty steps is a brace you will stop wearing, which is the compliance number in the trials.

The best knee braces by diagnosis: our picks

Every spec below is from the Amazon listing on 26 October 2026, read from a US delivery session. We have not worn any of them. Most sit in Amazon’s health category; the differences in Amazon’s commission tiers do not change what we would tell you to buy.

The sleeve for most knees

The stabilised braces

The unloader for medial arthritis

The strap for a tendon

How to wear a knee brace

1. For the activity, then off. Stairs, the walk, the run. The arthritis trials wore unloaders for hours a day over months with no loss of strength, and a quarter quit from discomfort; nobody tested sleeping in one.
2. Over clean, dry skin, not over a cut or a rash. Skin irritation and abrasion were the recorded harm in every trial that looked.
3. Check the skin after, especially under straps and at the top edge of a sleeve. Redness that persists means the size or the hours are wrong.
4. Not tight enough to mark the calf. A sleeve that leaves a deep ring is compressing the wrong thing. One published case report links an infrapatellar strap to a localised clot with pulmonary emboli in an athletic surgeon; it is one case, and it is the reason “tighter is better” is wrong.
5. Do the exercises. For every diagnosis on this page, the trials put strengthening ahead of the brace. The brace is what lets you do them.

Knee brace vs the rest of the shelf

For the same knee:

• A heating pad on a stiff arthritic knee before activity is the low-tech option with the better-evidenced mechanism for stiffness.
• An ice machine is what the knee gets after surgery; a brace is what it gets after.
• A foam roller on the quads and IT band is the runner’s-knee tool the same Cochrane reviewers would rather you used, with the strengthening.
• Compression boots are the sleeve’s big sibling for the whole leg, and no better evidenced.
• A TENS unit is the other thirty-dollar knee-pain device, with contested evidence.
• A massage gun stays on the muscle above and below the joint, never on the joint.

What we have not measured

Every material, size and feature here is listing-stated as of 26 October 2026. We have not worn, measured or washed any of these braces.

The measurements this page owes are cheap. A pressure sensor between brace and skin at the thigh, kneecap and calf for each sleeve in its stated size, standing and at 90° of flexion, to put a number in mmHg on “compression”, which no listing states and which decides both comfort and circulation. The same sensor under the unloader’s condyle pad at full extension, because the offloading claim is a force claim and the listing gives none. A tape measure on each sleeve after ten washes against its stated size. And a timed walk with each brace on a smooth floor to count how often it slips, which is the thing the trials say people quit over. All n=1, on one knee that is not yours.

What we would buy

For a knee with arthritis that wants comfort all day: the Modvel sleeve, sized from the chart, or the Bauerfeind GenuTrain if the fit is worth the price to you.

For medial arthritis where stairs and long walks are the problem: the Ovation Game Changer unloader, with a clinician’s confirmation of which compartment first.

For a sprained inner ligament or a knee that feels like it might give: the NEENCA hinged brace, or the DR. BRACE wrap if the knee swells and shrinks through the day.

For a sore patellar tendon on a runner: the Bodyprox strap, with the hip and quad work that actually fixes it.

For a meniscus tear: whichever of the sleeves you will wear while you do the exercise programme, which is the treatment. Our how we test page says what every figure here is and is not.