Knee

What Actually Prevents Knee Osteoarthritis: The Evidence on Weight, Strength, Running and Supplements

The single biggest risk factor is one you can change. Running does not wear your knees out. And the supplements do not work. Here is what the research on the ageing knee genuinely supports, and where it runs out.

Written by Dr Isa Waheed, MBBS MFSEM

Published 2 September 2026Last reviewed 2 September 202613 min read38 studies reviewed

Knee osteoarthritis is the condition most people have in mind when they worry about their knees "wearing out". It is the destination that many other knee problems lead towards if they are left alone, and it is, as the evidence below shows, far more preventable than the phrase "wear and tear" suggests.

This is the first of four detailed guides that sit under our knee hub. It covers the degenerative knee: osteoarthritis itself, plus the four conditions that most often travel with it in adults over 40, which are Baker's cyst, pes anserine bursitis, cartilage lesions, and the condition still widely called spontaneous osteonecrosis. If you want the short version across the whole knee, start with the hub. If you want to know what actually changes your risk of an arthritic knee, read on.

Key takeaways

Excess weight is the largest modifiable risk factor by a distance. Being overweight roughly doubles the odds of knee osteoarthritis and obesity raises them further, and about a quarter of new knee pain in people over 50 is attributable to it.
Weight loss works in a dose-dependent way once symptoms exist: every 1% of body weight lost improves pain and function by around 2%, and losing a quarter of body weight improves them by about half.
Weak quadriceps raise the odds of developing knee osteoarthritis, and strength is something you can change.
Recreational running does not cause knee osteoarthritis and may modestly protect against it. Only competitive or elite volumes carry higher risk, and that signal may partly reflect injury rather than running.
A previous knee injury nearly triples the odds, and removing meniscal tissue after injury raises them further. Preventing injury is preventing arthritis.
Glucosamine, chondroitin and vitamin D do not change the disease. The best trial evidence shows no meaningful effect on pain or joint structure.

How big the problem is, and why it is not just wear and tear

The lifetime risk of symptomatic knee osteoarthritis is about 60% in people with obesity [1]. Among adults aged 50 and over, roughly a quarter of new knee pain, 24.6%, is attributable to being overweight or obese, and a further 5.1% to a previous knee injury [2].

That is the case against "wear and tear" in two numbers. The largest drivers are not years of use but load and damage, and both of those can be changed.

Weight, the factor that outweighs everything else

The evidence here is unusually consistent. In the cornerstone review of risk factors in adults over 50, being overweight raised the odds of knee osteoarthritis onset (OR 1.98) and obesity raised them further (OR 2.66) [2]. A second pooled analysis put the odds with obesity at about 2.63 times those of normal weight, and noted that some large cohorts have reported a sevenfold to eightfold increase at a BMI of 30 or above [3]. The mechanism works through both mechanical load and metabolic effects [3], which is part of why the numbers are so large.

What is less often said is that this works in reverse. In people who already have knee osteoarthritis, weight loss improves symptoms in a dose-dependent way: every 1% of body weight lost improved WOMAC pain, function and stiffness by roughly 2%, and a 25% reduction gave about a 50% improvement [1]. The most effective approaches combined a low-calorie diet or bariatric surgery with exercise [1]. There is no threshold to clear before it starts helping.

Quadriceps strength

Across 11 studies and 46,819 people, weaker knee extensor muscles raised the odds of developing knee osteoarthritis. For symptomatic osteoarthritis the odds were 1.85 in women and 1.43 in men, with a similar picture for osteoarthritis on X-ray [4]. The authors' own conclusion was that optimising quadriceps strength may help prevent it [4].

This matters because, of the handful of factors that consistently predict knee osteoarthritis, muscle strength is one of the few that is entirely within your control. It also sits at the centre of what happens after injury, which we cover below.

Running does not wear your knees out

This is the myth the evidence most clearly overturns. Pooling the available studies, the prevalence of knee osteoarthritis was 3.5% in recreational runners, 10.2% in non-running controls, and 13.3% in competitive runners [5]. Recreational running carried slightly lower odds of osteoarthritis than not running (OR 0.86), while competitive or elite-level running carried higher odds (OR 1.34) [5].

Runners are a self-selecting group, so the protective effect should be read cautiously, but the direction is clear: ordinary running is not the problem. The competitive-runner signal may also partly reflect previous injury rather than running itself [5].

The cartilage data say the same thing at the level of a single run. Cartilage thickness and volume fell by roughly 3 to 5% immediately after running but recovered within about 90 minutes, and existing cartilage defects were unchanged 48 hours later [6]. Healthy cartilage compresses under load and rebounds. That is what it is for.

Injury is the other big lever

A traumatic knee injury substantially raises the odds of later osteoarthritis, and this holds across cruciate ligament, collateral ligament, meniscal, cartilage, kneecap dislocation, fracture and multi-structure injuries [7]. In the over-50 review, a previous knee injury raised the odds of osteoarthritis 2.83 times [2].

What happens after the injury matters as much as the injury itself. After ACL reconstruction, the odds of structural osteoarthritis were higher with an associated cartilage injury (OR 2.31), with partial meniscectomy (OR 1.87), and highest with total medial meniscectomy (OR 3.14) [7]. The more meniscal tissue removed, the higher the odds.

The practical consequences, preventing the injury in the first place through neuromuscular training and preserving the meniscus if surgery is needed, are covered in detail in our sports knee injury guide.

Work and physical load

Physically demanding work raises knee osteoarthritis risk, and the effect is gradable. Regular heavy lifting of more than 10 kg per week raised the odds (OR 1.52), as did squatting or kneeling (OR 1.69), prolonged standing for over 2 hours a day (OR 1.22) and frequent walking (OR 1.40) [8]. These effects were magnified by previous knee injury and by a BMI above 25 [8], which is the same combination of load and damage that runs through this whole guide.

Kneeling and heavy lifting are not always avoidable, but the amount is often more negotiable than it looks, and the compounding effect of excess weight is the part you can act on directly.

The supplements

Glucosamine and chondroitin are among the most widely bought supplements for joints, so it is worth being precise about the evidence.

Pooling 10 large randomised trials in 3,803 people with hip or knee osteoarthritis, glucosamine, chondroitin and their combination did not produce a clinically meaningful effect. Pain reductions were about 0.3 to 0.5 cm on a 10 cm scale, short of the 0.9 cm minimal important difference, and there was no effect on joint-space narrowing [9]. Industry-independent trials showed smaller effects than commercially funded ones [9].

There is one observational signal pointing the other way. Over six years, combined glucosamine and chondroitin use was associated with reduced cartilage volume loss, significant in those who took it for two years or more [10]. But this is observational rather than randomised, and trial evidence is inconsistent, so it is a signal to verify rather than a recommendation [10]. People who choose to take supplements differ in many ways from people who do not, and an observational study cannot separate the supplement from the person.

Vitamin D fares no better for prevention. Supplementation produced a small improvement in WOMAC pain and function but had no effect on tibial cartilage volume, and the authors concluded there is a lack of evidence to support vitamin D for preventing structural progression of knee osteoarthritis [11].

The honest summary is that no supplement has been shown to change the disease. The money is better spent on the things above.

What you cannot change, and what it does and does not mean

Some risk is fixed. Female sex raised the odds of knee osteoarthritis (OR 1.68) in the over-50 review [2]. Twin studies estimate heritability of around 30 to 46% for osteoarthritis in women, varying by joint [12], and the genetic influence appears to be site-specific, with little evidence of a single shared pathway driving hand, hip and knee disease together [13]. Diabetes raised osteoarthritis risk (OR 1.46), but among 12 studies that adjusted for body weight, 7 found diabetes an independent factor and 5 found no association, so part of that link is explained by shared obesity [14].

Alignment is more nuanced than it is usually presented. Knee malalignment, whether bow-legged or knock-kneed, is an independent risk factor for the progression of established osteoarthritis, supported by strong cohort and MRI evidence, but the evidence that it causes osteoarthritis to develop in the first place is limited [15]. If you have a knee that is already arthritic, alignment matters. If you are trying to prevent one, it matters less than the factors above.

And one factor that is often assumed to matter does not: smoking was not associated with knee osteoarthritis (OR 0.92) [2]. There are plenty of other reasons not to smoke, including its association with symptomatic cartilage defects covered below, but knee arthritis is not one of the clearer ones.

Fixed risk is a reason to act on the modifiable factors, not a reason to give up. A family history shifts your starting point. Weight, strength and injury shift the trajectory.

The conditions that travel with the ageing knee

Four other diagnoses turn up repeatedly in adults over 40, and understanding them mostly comes back to the same underlying story.

Baker's cyst

A Baker's cyst is a fluid-filled swelling behind the knee. In adults it is almost always a sign of something else going on inside the joint, not a disease in its own right [16, 19]. Among people having knee arthroscopy, cysts were found in 20% and in none of the symptom-free controls, and those with a cyst had far higher rates of medial meniscal tears (70% versus 19%) and cartilage lesions (85% versus 28%) [16]. Cartilage damage was the main reason cysts persisted [16]. On MRI, the probability of a cyst rose from about 8 to 10% with any one of joint effusion, meniscal tear or degenerative arthritis to 38% with all three present [17]. In osteoarthritic knees, roughly one in five has a cyst and the large majority cause no symptoms [18].

The prevention message is therefore the same as for osteoarthritis: treat the joint, not the cyst. In children the picture is different. There the cyst is usually primary, arising from the bursa without joint disease [19], and about 85% shrink or disappear without treatment [20].

Pes anserine bursitis

Pain on the inner side of the knee just below the joint line is often labelled pes anserine bursitis. On ultrasound, bursal swelling was found in 20% of osteoarthritic knees, was more common in women, and grew larger as osteoarthritis severity and age increased [21]. Imaging often fails to show true inflammation of the bursa or tendons, which is why many authors prefer the looser term anserine syndrome and treat it as a clinical pattern rather than a defined disease [22].

The alignment evidence conflicts. A small case-control study linked knock-knee (valgus) alignment to the condition, at about five times the odds [23], while a later study of osteoarthritic knees linked it to bow-leg (varus) alignment instead [24]. The condition is common in type 2 diabetes and almost only in women, but the pain is usually accompanied by structural osteoarthritis changes such as medial meniscus damage, with meniscal protrusion strongly predicting knee pain [25]. In practice this is usually the inner compartment of an arthritic knee announcing itself, and the prevention is the prevention of osteoarthritis.

Cartilage lesions

Cartilage damage is far more common than symptoms. In a pain-free population aged 40 to 79, moderate or worse cartilage damage was present on MRI in about 65%, and a BMI of 25 or above carried about three times the odds of moderate damage [26]. Only current weight was associated, not past weight or weight change [26], which is an encouraging finding: it suggests the load your cartilage is under now is what matters.

In a series of over 25,000 knee arthroscopies, cartilage lesions were seen in 60% of knees and were part of another injury in 70% of cases, most often alongside a medial meniscus tear or an ACL injury [27]. The natural history of an individual lesion is not well established and not every lesion progresses [28], which is a reason not to over-treat incidental findings on a scan.

Two things do predict trouble. In athletes who had lost meniscal tissue, 85% had a cartilage lesion, and the less meniscus remaining, the worse the cartilage [29]. And among people with symptomatic cartilage defects, higher body weight and smoking were both more common [30].

Spontaneous osteonecrosis of the knee

This condition has been substantially reframed. What was long described as a loss of blood supply to bone is now understood in most cases to be a subchondral insufficiency fracture: bone just beneath the cartilage, almost always on the inner femoral condyle, fails under load, and the changes once read as osteonecrosis are largely a consequence of that fracture rather than its cause [31, 32]. The meniscus was implicated in around 81% of studies [31].

The people it affects fit the pattern of the overloaded inner knee. In one large series, 68% were women with a mean age in the sixties, higher BMI was associated, around 89% of cases were medial, and medial meniscus tears were found in roughly 77% [33]. In another, 71% were women and the condition progressed to knee replacement in roughly a third of cases [34]. Medial meniscus posterior root and radial tears were closely linked, most accompanied by meniscal extrusion of 3 mm or more [35], and the condition is recognised after arthroscopy, particularly meniscectomy [36], which again points to removing meniscal tissue as a driver.

Whether weak bone contributes is contested. In one series most patients were not osteoporotic, meniscal extrusion predominated, and average bone density was actually above normal for age [37], while a smaller study found some older women with the condition had lower bone density [38]. The prevention message is not primarily about bone. It is about protecting the meniscus and controlling the load through the inner compartment, which means weight, strength, and not removing meniscal tissue unless there is no alternative.

What changes your risk of knee osteoarthritis

Factor	Direction	Effect	Can you change it?	Evidence
Overweight	Raises risk	OR 1.98	Yes	Tier 2
Obesity	Raises risk	OR 2.63 to 2.66; up to 7 to 8x in some cohorts	Yes	Tier 2
Weight loss (once symptomatic)	Improves symptoms	About 2% better per 1% of weight lost	Yes	Tier 1
Weak quadriceps	Raises risk	OR 1.43 (men) to 1.85 (women)	Yes	Tier 2
Previous knee injury	Raises risk	OR 2.83	Partly	Tier 2
Meniscectomy after injury	Raises risk	OR 1.87 partial; 3.14 total medial	Partly	Tier 2
Recreational running	Slightly lowers risk	OR 0.86	Yes	Tier 2
Competitive or elite running	Raises risk	OR 1.34	Partly	Tier 2
Heavy occupational load	Raises risk	OR 1.22 to 1.69	Partly	Tier 2
Female sex	Raises risk	OR 1.68	No	Tier 2
Family history	Raises risk	Heritability 30 to 46% in women	No	Tier 3
Diabetes	Raises risk	OR 1.46, partly via obesity	Partly	Tier 2
Malalignment	Drives progression, not onset	Independent factor for progression	No	Tier 2
Glucosamine and chondroitin	No meaningful effect	0.3 to 0.5 cm vs 0.9 cm threshold	Intervention	Tier 1
Vitamin D	No structural effect	No change in cartilage volume	Intervention	Tier 1
Smoking	No effect	OR 0.92	Yes	Tier 2

Where the evidence runs thin

Most of what we know about who develops knee osteoarthritis comes from observational cohorts, which are good at finding associations and less good at proving cause. The protective effect of recreational running, for instance, could partly reflect the fact that people with healthy knees keep running. The weight loss data are strong for symptoms but thinner for preventing the disease from starting in the first place. The supplement trials are strong and negative. And for the four related conditions, the evidence is mostly cross-sectional and small, which is why our confidence in the underlying mechanism, overload of the inner knee, is higher than our confidence in any single figure.

What we would actually do

Keep your weight in a healthy range, and if you already have knee pain and excess weight, treat weight loss as a treatment with a dose-response curve. Train your quadriceps and keep training them. Run if you enjoy it. If you play pivoting sport, do a neuromuscular warm-up, because preventing the injury is preventing the arthritis. If you injure a meniscus, ask about repair before removal. Save your money on supplements. And if you are told you have a Baker's cyst, cartilage wear on a scan, or anserine bursitis, understand that these are usually the arthritic knee describing itself, and the prevention is the same.

References

Frequently asked questions

Does running cause knee arthritis?

The pooled evidence says no. Recreational runners had a lower prevalence of knee osteoarthritis than non-runners (3.5% versus 10.2%), and a single run causes only small, temporary cartilage changes that recover within about 90 minutes. Only competitive or elite volumes carried higher odds, and that may partly reflect injury rather than running.

Do glucosamine and chondroitin work?

Not in a way that matters. Pooling 10 large trials, the effect on pain fell short of the threshold for a clinically meaningful difference, and there was no effect on joint structure. One observational study suggested slower cartilage loss with long-term use, but observational evidence cannot separate the supplement from the type of person who takes it.

How much weight do I need to lose to help my knees?

There is no threshold to clear. In people with knee osteoarthritis, every 1% of body weight lost improved pain and function by about 2%, and a 25% loss improved them by about half. Combining diet with exercise was the most effective approach.

Can I prevent knee arthritis if it runs in my family?

You can shift your risk substantially. Heritability is estimated at around 30 to 46% in women, which means most of the variation is not genetic. Weight, quadriceps strength and avoiding or properly managing injury act on the part you control.

Is a Baker's cyst dangerous?

Usually not, and in adults it is almost always a sign of something else inside the knee, most often a meniscal tear or cartilage damage. Treating the underlying joint problem is what resolves it. In children, cysts are usually primary and about 85% disappear without treatment.

Put the evidence to work

Movement Age turns research like this into a single score for your body, and shows you where to focus. Join the waitlist for early access.

Join the Movement Age waitlist

About the author

Dr Isa Waheed

MBBSMFSEMBSc (Hons)DipMSKDipExMedDipTCPGCertFHEA

NHS doctor and sport and exercise medicine clinician, translating injury prevention research into guidance people can act on.

Movement Age

Get your Movement Age

Join the waitlist for early access to the score that benchmarks your spine, joints, and muscles against your actual age.

Educational information only. Not medical advice and not a substitute for assessment by a qualified clinician. Seek prompt medical assessment for a knee that locks, gives way, swells significantly, or cannot bear weight, or for any sudden or severe knee injury.