Dr Matthew BroadheadOrthopaedic Surgeon — Sydney & Coffs Harbour

Journal Club · August 2026

Does running make knee arthritis worse?

Over four years, people aged 50+ with knee osteoarthritis who ran were no more likely to show X-ray worsening or new pain than those who did not — and were somewhat more likely to report their pain had improved.

Two questions come up constantly in clinic, and they deserve a straight, evidence-based answer. If I already have arthritis in my knees, am I grinding them down every time I run? And if I have a knee replacement, will I ever run again?

The honest position is that the evidence is much better for the first question than the second. This entry deals mainly with the first, and is clear about how thin the ground gets on the second.

What the study looked at

The researchers used the Osteoarthritis Initiative, a large American study that followed people with knee arthritis for years with repeated X-rays and symptom questionnaires. They took 1,203 people aged 50 and over who already had arthritis visible on X-ray, and asked which of them ran. 138 (about 11.5%) did.

They then compared runners with non-runners over four years on three things: whether the arthritis got worse on X-ray, whether new persistent knee pain developed, and whether existing pain improved.

What they found

  • X-ray worsening: 15.6% of runners versus 19.2% of non-runners (adjusted odds ratio 0.9, 95% CI 0.6–1.3).
  • Joint-space narrowing — the gap between the bones closing up: 19.5% versus 23.6% (adjusted OR 0.8, 95% CI 0.5–1.2).
  • New frequent knee pain: 26.8% versus 29.0% (adjusted OR 0.9, 95% CI 0.6–1.6).
  • Pain that improved: 50.0% of runners versus 39.1% of non-runners (adjusted OR 1.7, 95% CI 1.0–2.8).

Read those carefully. The first three show no signal of harm — the runners were, if anything, slightly better off, but the confidence intervals all comfortably include “no difference”, so the fair reading is no evidence that running made things worse. The fourth, on pain improving, is the only one pointing to possible benefit, and even there the confidence interval just touches 1.0, which makes it borderline rather than established.

What it means for you

For someone over 50 with arthritis on their knee X-ray who is already running and enjoying it, this study offers reassurance: there was no sign that continuing to run accelerated the arthritis or worsened symptoms over four years. Running need not be given up simply because a scan shows wear.

Two things it does not say. It does not say running will fix your arthritis — the pain finding is too borderline for that. And it does not say that someone with painful arthritis who currently does not run should take it up as treatment; everyone in this study chose their own activity, and people generally choose what their knees tolerate.

And running after a knee replacement?

This is the weaker half of the answer, and it is worth being upfront about that. A 2025 systematic review in Musculoskeletal Surgery found only six studies in the entire published literature on running after total knee replacement — and much of that paper is devoted to proposing a new rehabilitation protocol that is still enrolling patients, rather than reporting established results.

What can reasonably be said is that running after a knee replacement remains uncommon and debated. Low-impact activity — walking, cycling, swimming, golf — is what is routinely encouraged, because the long-term concern with repeated impact is wear of the bearing surface and loosening of the implant. Reported complication rates in the small published series have been low, but the number of patients is small and the follow-up is short, which is precisely why nobody can yet quote you a reliable long-term figure.

In practice this is an individual conversation: it depends on the implant used, how it was fixed, your weight and fitness, what running you did before, and what you are hoping to get back to. It is a reasonable thing to raise before surgery, not after.

A caveat worth knowing

The central weakness of this study is the order in which things were measured. Participants were asked about their running history on a survey completed at the 96-month visit — after the four-year outcome window had already closed. So people were recalling their activity, and there is no way to be certain the arthritis did not come first and the running habit second. People whose knees hurt more tend to stop running; that alone could make runners look healthier without running being the cause.

There was also no record of how far or how fast anyone ran, most of the runners were men, and the study is about self-selected recreational running — people doing the amount their own knees allowed. It cannot tell you what would happen if someone with painful arthritis were instructed to run a set weekly distance. This is reassurance about continuing an activity you already tolerate, not a prescription.

Assessing this studyA standardised appraisal for clinicians and registrars — study design, strength and validity.
Study design
Nested observational cohort study within the Osteoarthritis Initiative (2004–2014). 1,203 participants aged 50 or over with radiographic knee osteoarthritis (Kellgren-Lawrence grade ≥2), of whom 138 (11.5%) were classified as runners; outcomes assessed over 48 months.
Level of evidence
OCEBM Level 3 (prognosis) — an observational cohort, downgraded because the running exposure was ascertained retrospectively, at the 96-month visit, after the outcome window had already closed.
Are the results valid?
The main threat is that running was self-selected and self-reported. Participants were classed as runners if running or jogging was among their three most frequent activities at age 50 or over, recorded on a survey completed after the outcomes were measured — so recall bias and reverse causation (people whose knees hurt more simply stop running) cannot be excluded, and the authors say so. There was no information on distance, pace or weekly volume, runners were predominantly male (69.6%), and the number of events among runners was too small for any dose-response analysis. Analyses were adjusted for age, sex, BMI, injury, baseline KL score and contralateral knee status; about 8% of data were imputed.
What are the results?
Over 48 months, radiographic worsening (KL grade) occurred in 15.6% of runners versus 19.2% of non-runners (adjusted OR 0.9, 95% CI 0.6–1.3); medial joint-space narrowing in 19.5% versus 23.6% (adjusted OR 0.8, 95% CI 0.5–1.2); and new frequent knee pain in 26.8% versus 29.0% (adjusted OR 0.9, 95% CI 0.6–1.6). Improved knee pain was reported by 50.0% of runners versus 39.1% of non-runners (adjusted OR 1.7, 95% CI 1.0–2.8) — a confidence interval whose lower bound touches 1.0, so this result is borderline rather than firmly positive. All confidence intervals for the structural outcomes comfortably include no effect, so this is an absence-of-harm signal, not a demonstrated benefit.
Do they apply to our patients?
Applies to people aged 50 and over with established radiographic knee osteoarthritis who are already running recreationally at self-selected volumes. It does not speak to high-volume or competitive running, to those with severe symptomatic disease who have already stopped, or to running after joint replacement — a separate question the paper does not address.
Bottom line
In over-50s with knee osteoarthritis, self-selected recreational running was not associated with faster X-ray progression or more pain over four years and may be associated with less pain, but because running was self-selected and recorded retrospectively this supports not discouraging running rather than prescribing it.

Framework: study design graded with the Oxford Centre for Evidence-Based Medicine (OCEBM) Levels of Evidence; validity and applicability appraised in the CASP tradition.

This is a plain-language summary of published research, provided for general education. It is not medical advice and does not describe Dr Broadhead's own results. Whether any finding applies to you depends on your individual circumstances — please discuss your care with your treating team.

← All Journal Club entries

CallEmail