Does running ruin your knees? What the research actually found

It is the most common reason people give for never starting. The honest answer is more interesting than either side of the argument usually admits: running loads the knee three times harder than walking does, and runners still do not get more arthritis than people who sit still. Here is why, and where the reassurance runs out.

The biggest study, read properly

The most-cited analysis pooled 17 studies covering 114,829 people and compared hip and knee osteoarthritis across three groups:

• Recreational runners: 3.5% (95% CI 3.4 to 3.6)
• Sedentary controls: 10.2% (95% CI 9.9 to 10.6)
• Competitive runners: 13.3% (95% CI 11.6 to 15.2)

Those raw numbers are where the popular headline comes from, and we have to be straight about something: the adjusted odds ratios do not reach statistical significance. Recreational runners against controls came out at 0.86 (95% CI 0.69 to 1.07), and competitive runners at 1.34 (95% CI 0.97 to 1.86). Both intervals cross 1. What was significant was the difference between the two runner groups, and one other finding: less than 15 years of running exposure, OR 0.6 (95% CI 0.49 to 0.73).

The authors’ own caveat is the one to keep: “it was not possible to determine whether these associations were causative or confounded by other risk factors, such as previous injury.”

So the defensible claim is not “running protects your knees.” It is “running has not been shown to harm them, and runners have less arthritis than non-runners in every raw comparison anyone has done.” That is still a good answer to the question people are actually asking.

What happens when you follow runners for years

Prevalence snapshots have an obvious problem: people whose knees hurt stop running, so any group of current runners is a group of survivors. The studies that followed people over time are better, and there are not many.

The most-cited one ran from 1984 to 2002 — a mean follow-up of about eleven to twelve years — with serial knee X-rays on 45 distance runners and 53 controls who were 50 to 72 at the start. Final arthritis prevalence: 20% in runners, 32% in controls (p = 0.25). Severe arthritis: 2.2% versus 9.4% (p = 0.21). Neither was statistically significant, and what did predict worse X-rays was existing damage and elapsed time — not how much anyone ran.

That study gets quoted as though it settles the question. It does not, and the authors were candid about why. Across all 98 people there were six cases of severe arthritis and three knee replacements in total — far fewer than expected, and nowhere near enough to detect anything but a very large effect. The runners had worse knees than controls at the start (p = 0.018). And the exposure contrast dissolved during the study: the runners’ weekly running fell by 56%, with nearly half stopping altogether, while the controls took up brisk walking. A p-value of 0.25 in that setting means “we could not tell,” not “there is no difference.”

A larger and cleaner prospective study — 1,279 people from the Framingham cohort, followed about nine years, looking at both compartments of the knee — reached a genuinely neutral verdict: “recreational exercise neither protects against nor increases risk of knee OA.”

A 2023 meta-analysis of nine studies covering 12,273 people found no difference between runners and controls: OR 0.97 (95% CI 0.56 to 1.68), and — importantly — no dose-response. Running more than 48 km a week came out at OR 0.62 (0.35 to 1.10) against the lowest-volume group. Five of the nine studies were at high or very high risk of bias, and the authors called for proper prospective work.

The most honest summary in this entire literature comes from a separate review: “With this evidence, it is not possible to determine the role of running in knee OA.”

Why three times the force does not equal three times the damage

This is the part that resolves the apparent paradox, and it is worth understanding.

Researchers measured knee joint contact forces in the same fourteen people walking and running. Peak force: 8.02 times body weight running, 2.72 walking — running is three times harder on the knee, exactly as your intuition says.

But then they calculated load per kilometre covered: 0.80 body weights per metre running, 0.75 walking. Not significantly different (p = 0.098).

Running is harder per step, but each step covers much more ground and spends far less time on the floor. Cover a given distance and the total load on the knee is about the same either way. Running the distance faster actually lowers the load per kilometre.

What actually happens to cartilage

Four different timescales, four different answers.

During a run, cartilage compresses — and no more than walking does. In 25 healthy people doing 30 minutes of each, medial femoral cartilage deformed 8.9% running versus 6.7% walking, a difference that was not statistically significant (p = 0.28). A separate study of 40-minute treadmill runs found changes under 3%, with no time point differing from baseline.

After a marathon, the picture is less comfortable, and this is the strongest counter-evidence in the article. Ten marathoners scanned before, within 48 hours after, and again at 10 to 12 weeks showed no structural damage — but compositional MRI markers rose significantly, and one of them (T1ρ) was still elevated at three months, worst in the kneecap and inner compartment. The authors described those areas as at higher risk of degeneration. Nobody knows whether that matters.

Over a very long distance, cartilage appears to adapt. Runners in a 4,486 km race across Europe were scanned in a mobile MRI unit throughout. Markers rose sharply over the first 1,100 km — up to 44% in places — then stopped rising, and began recovering after 3,500 km, while the athletes were still running 60 to 70 km a day. No new cartilage lesions appeared, and nobody withdrew with knee problems.

And cartilage responds to training like other tissue does. In a randomised trial, 30 people who had had part of a meniscus removed years earlier were assigned to supervised exercise three times a week for four months, or nothing. Cartilage quality on dGEMRIC scanning improved in the exercise group and declined in the controls (p = 0.036), with a dose-response between activity increase and cartilage change (rS = 0.70). The authors concluded that “adult human articular cartilage has a potential to adapt to loading change.”

Be careful with that last one: 30 people, four months, a biochemical surrogate rather than a clinical outcome, and they were post-surgical patients doing supervised gym work, not runners. It supports “cartilage is living, adaptable tissue.” It does not prove running builds cartilage.

One caveat covering all of the above: these MRI markers “lack direct correlation with pain or function loss” and there is no standardised way of measuring them. Three things get conflated constantly and should not be — temporary squashing, chemical change of unknown meaning, and actual structural damage. Only the first two have ever been shown in runners.

The elite exception, and what probably explains it

Competitive and elite runners genuinely do show more knee arthritis. One analysis of athletes across sports put elite-level distance running at OR 3.3. A Finnish study of 117 former top-level male athletes found radiographic knee arthritis in 3% of shooters, 14% of distance runners, 29% of footballers and 31% of weightlifters.

The obvious reading is that high mileage wears the joint out. The evidence points somewhere else. In that Finnish study, the regression identified previous knee injury (OR 4.73), body mass at 20, and years of heavy or kneeling work as the risks — running itself did not emerge as an independent factor.

And the size of the injury effect dwarfs everything else here. Pooling 53 studies covering about a million people:

• Previous ACL injury: OR 4.2 (95% CI 2.2 to 8.0)
• Previous meniscal injury: OR 6.3 (95% CI 3.8 to 10.5)
• Both: OR 6.4 (95% CI 4.9 to 8.3)

Set that beside the largest estimate for competitive running — 1.34, not statistically significant — and the comparison makes itself. A previous knee injury carries roughly four to six times the risk. Competitive running carries an effect that cannot be distinguished from zero. Elite careers accumulate injuries, and that is the most plausible explanation for the elite excess. Nobody has cleanly separated the two, and the researchers say so.

(Those injury estimates come with severe heterogeneity between studies — I² as high as 95% — so treat the exact numbers as approximate and the direction as solid.)

What if your knees already hurt?

Two separate questions here.

Can you keep running with diagnosed knee arthritis? A study of 1,203 people over 50 with existing knee OA, followed for four years and adjusted for age, sex, BMI, arthritis severity and injury history, found runners were no worse off: X-ray worsening OR 0.9 (0.6 to 1.3), new frequent knee pain OR 0.9 (0.6 to 1.6), and resolution of existing pain OR 1.7 (1.0 to 2.8). The authors concluded that self-selected running “need not be discouraged in people with knee OA.” Note “self-selected” — these were people who could already tolerate it. It is not a licence to prescribe running to anyone with arthritis.

Should you exercise if you have knee arthritis? Here the evidence is genuinely strong, and it is the best news on this page. A Cochrane review of 54 studies found land-based exercise reduced pain (SMD −0.49, high-quality evidence, 44 trials, 3,537 people) and improved function (SMD −0.52). The authors described the effect as “comparable with estimates reported for non-steroidal anti-inflammatory drugs.” Across all those trials, no study reported a serious adverse event.

Both the American College of Rheumatology and OARSI make exercise a core, strongly recommended treatment for knee osteoarthritis. Neither names running specifically — they recommend exercise, and that distinction is worth keeping straight.

The case against, stated fairly

We would rather you had this than found it elsewhere:

1. A 21-year Finnish study points the other way. Tracking hospital admissions for lower-limb arthritis in 2,049 former elite male athletes against 1,403 controls, endurance athletes came out at OR 1.73 (95% CI 0.99 to 3.01) — and after adjustment for age, occupation and body mass at 20, the estimates across sport types rose to 2.37 to 2.68. This study is rarely quoted in the reassuring summaries.
2. One marathon changed cartilage chemistry for at least three months. The markers had not returned to baseline at 12 weeks.
3. Elite distance running does show elevated arthritis — OR 3.3 in one analysis — even if injury history probably explains it.
4. Much of the apparent benefit may be about body weight, not joints. In a cohort of 74,752 runners, the reduction in arthritis risk fell from 16.5% to 8.6% and lost significance (p = 0.21) once BMI was accounted for; the hip replacement finding went the same way. Runners are lighter, and lighter people get less knee arthritis. That is a real benefit — but it is not evidence that running protects cartilage.
5. Heavier runners showed more cartilage change after a marathon, correlating with BMI at R = 0.70. That is more actionable than any advice about pace.
6. Almost all the reassuring data is cross-sectional and drawn from people still running — a survivor group by definition. If your knees hurt, you stop, and you stop being counted as a runner.
7. Every null here is an underpowered null. An odds ratio of 0.97 with an interval from 0.56 to 1.68 is consistent with meaningful benefit and meaningful harm. Absence of evidence of harm keeps being reported as evidence of absence.

An Australian study is currently following 214 runners with ten-year knee MRI specifically to answer this properly. It will not report until the 2030s. Until then, anyone telling you the question is settled — in either direction — is going beyond the evidence.

What to actually do

1. Start, if the only thing stopping you is your knees. The evidence does not support the fear, and the sedentary comparison group has more arthritis in every raw count.
2. Take a previous knee injury seriously. It is the one risk factor here that is large, replicated and unambiguous. If you have had an ACL or meniscus injury, get proper rehabilitation before building mileage — that matters far more than how far you run.
3. Manage body mass where you can. It predicted worse X-rays over 18 years and predicted cartilage change after a marathon. It did not depend on how much anyone ran.
4. Build up gradually and let the tissue adapt. The ultramarathon data suggests cartilage responds to progressive load the way muscle and bone do. It does not suggest it responds well to a sudden jump.
5. Do the strength work. Not because it protects cartilage — that specific claim is not supported — and not on a promise of injury prevention, which has not held up in running-specific trials. Do it because it makes you a more durable, more economical runner.
6. If a knee hurts for more than a couple of weeks, get it looked at rather than running through it. The risk in this whole area is injury, and injuries get worse when ignored.

References

• Alentorn-Geli E, et al. The association of recreational and competitive running with hip and knee osteoarthritis. J Orthop Sports Phys Ther, 2017;47(6):373–90. doi:10.2519/jospt.2017.7137
• Chakravarty EF, et al. Long distance running and knee osteoarthritis: a prospective study. Am J Prev Med, 2008;35(2):133–8. doi:10.1016/j.amepre.2008.03.032
• Burfield M, Sayers M, Buhmann R. The association between running volume and knee osteoarthritis prevalence. Phys Ther Sport, 2023;61:1–10. doi:10.1016/j.ptsp.2023.02.003
• Timmins KA, et al. Running and knee osteoarthritis: a systematic review and meta-analysis. Am J Sports Med, 2017;45(6):1447–57. doi:10.1177/0363546516657531
• Miller RH, et al. Why don’t most runners get knee osteoarthritis? A case for per-unit-distance loads. Med Sci Sports Exerc, 2014;46(3):572–9. doi:10.1249/MSS.0000000000000135
• Harkey MS, et al. Ultrasonographic assessment of medial femoral cartilage deformation acutely following walking and running. Osteoarthritis Cartilage, 2017;25(6):907–13. doi:10.1016/j.joca.2016.12.026
• Luke AC, et al. High-field magnetic resonance imaging assessment of articular cartilage before and after marathon running. Am J Sports Med, 2010;38(11):2273–80. doi:10.1177/0363546510372799
• Schütz U, et al. Chondral adaptation in the knee of ultra-marathon runners. Sci Rep, 2020;10:8157. doi:10.1038/s41598-020-64994-2
• Roos EM, Dahlberg L. Positive effects of moderate exercise on glycosaminoglycan content in knee cartilage: a randomized controlled trial. Arthritis Rheum, 2005;52(11):3507–14. doi:10.1002/art.21415
• Wang Z, et al. Higher body mass index is associated with biochemical changes in knee articular cartilage after marathon running. Orthop J Sports Med, 2020;8(8):2325967120943874. doi:10.1177/2325967120943874
• Driban JB, et al. Is participation in certain sports associated with knee osteoarthritis? J Athl Train, 2017;52(6):497–506.
• Kujala UM, et al. Knee osteoarthritis in former runners, soccer players, weight lifters, and shooters. Arthritis Rheum, 1995;38(4):539–46.
• Poulsen E, et al. Knee osteoarthritis risk is increased 4–6 fold after knee injury. Br J Sports Med, 2019;53(23):1454–63. doi:10.1136/bjsports-2018-100022
• Lo GH, et al. Running does not increase symptoms or structural progression in people with knee osteoarthritis. Clin Rheumatol, 2018;37(9):2497–504. doi:10.1007/s10067-018-4121-3
• Kujala UM, Kaprio J, Sarna S. Osteoarthritis of weight bearing joints of lower limbs in former elite male athletes. BMJ, 1994;308(6923):231–4. doi:10.1136/bmj.308.6923.231
• Felson DT, et al. Effect of recreational physical activities on the development of knee osteoarthritis in older adults of different weights: the Framingham Study. Arthritis Rheum, 2007;57(1):6–12. doi:10.1002/art.22464
• Williams PT. Effects of running and walking on osteoarthritis and hip replacement risk. Med Sci Sports Exerc, 2013;45(7):1292–7. doi:10.1249/MSS.0b013e3182885f26
• De Oliveira Silva D, et al. Training, Running Injury and Longitudinal knee health (TRAIL) cohort: protocol. BMJ Open, 2025;15(9):e101625. doi:10.1136/bmjopen-2025-101625
• Fransen M, et al. Exercise for osteoarthritis of the knee. Cochrane Database Syst Rev, 2015;1:CD004376. doi:10.1002/14651858.CD004376.pub3