DOES RUNNING WRECK YOUR KNEES AFTER 50?

"Running is bad for your knees" is one of the most repeated pieces of advice a runner past 50 hears. The largest cohort study built specifically to answer this question ran for nearly two decades and started with subjects the same age as WPP's audience.

The Study Built to Answer This Exact Question

Chakravarty and colleagues at Stanford published the result in 2008 in the American Journal of Preventive Medicine. They enrolled 45 long-distance runners and 53 non-running controls in 1984, mean age 58, ranging from 50 to 72, and tracked their knees with serial X-rays through 2002.

By the end of the study, the runners did not have more knee osteoarthritis than the controls — 20% versus 32%, a difference that did not reach statistical significance (p=0.25) — and they did not have more severe cases, 2.2% versus 9.4% (p=0.21). What did predict worse knee damage at the final X-ray: higher starting BMI, more existing damage at the first X-ray, and more years in the study. Runner-versus-control status was not a significant predictor.

The authors' own conclusion: long-distance running among healthy older individuals "was not associated with accelerated radiographic OA." That is the direct answer to the question this population actually asks — not "does running affect the knee somehow," but "if I keep running into my sixties and seventies, will my knees get worse faster than if I stopped."

The Larger Reviews Point the Same Direction

Timmins, Leech, Batt and Edwards pooled 15 studies in a 2017 systematic review and meta-analysis in the American Journal of Sports Medicine. The clearest number to come out of it: runners had lower odds of needing knee-replacement surgery for osteoarthritis than non-runners — pooled odds ratio 0.46 (95% CI 0.30–0.71). The authors were careful about the rest: evidence on OA diagnosis itself was mixed across the included studies, and they concluded it was "not possible to determine the role of running in knee OA" from the diagnostic data alone. The surgery finding is the one result in their review that reached statistical significance and held together, and it points toward protection, not harm.

A more recent systematic review, Dhillon and colleagues in 2023 in Orthopaedic Journal of Sports Medicine, compared 7,194 runners to 6,947 non-runners across 17 studies — mean age 56.2 in the runner group, 61.6 in the non-runner group. Knee pain was significantly more common in the non-runners (p<.0001), and most of the included studies found no difference in radiographic OA or MRI cartilage thickness between the two groups. One study found non-runners were more likely to progress to total knee replacement, 4.6% versus 2.6%. The authors' own stated conclusion was that running "is not associated with worsening" outcomes in the short term and "may be protective against generalized knee pain" — their hedge, not this post's.

Where the Real Risk Actually Sits

None of this means volume is irrelevant. A 2017 meta-analysis by Alentorn-Geli and colleagues in the Journal of Orthopaedic & Sports Physical Therapy, pooling 25 studies and over 125,000 people, found a split by intensity: competitive runners — professional or international-level athletes — had a higher prevalence of hip and knee OA (13.3%) than sedentary controls (10.2%), while recreational runners had a lower prevalence (3.5%) than either group. Running exposure under 15 years was associated with lower odds of OA than the control group; the authors could not determine whether the pattern was caused by running itself or by other factors bundled with a competitive athletic career, most notably prior injury.

That's the actual dividing line in this evidence, and it's a volume-and-injury-history question, not a running-versus-not-running question. Every study above is observational — none of them randomized people into "run" or "don't run" for two decades — so what they show is association, not proof of a mechanism. The consistent pattern across four separate research groups, though, is that ordinary recreational running does not show up as a knee-osteoarthritis risk in the data, and non-runners report knee pain at least as often.

The Protocol

• Recreational mileage is the population these studies actually describe. The protective signal is strongest for people running for fitness, not racing at a competitive level for decades.
• Prior injury is the variable the authors could not separate out. In the Alentorn-Geli analysis, the higher OA prevalence in competitive runners may have been driven by injury history rather than running itself. Rehabilitation and return to running after an injury is a decision for your physician or physical therapist.
• In the Stanford cohort, starting BMI predicted knee damage; running status did not. Higher starting BMI and more initial radiographic damage predicted worse knees at the final X-ray.
• What these studies are: observational cohorts and meta-analyses. They show that recreational running was not associated with more knee OA in these populations, not that running protects anyone's knees.

Four separate research groups, using different methods, reported that recreational running was not associated with more knee osteoarthritis in the people they studied. If you already have diagnosed knee OA or a recent injury, how much to run is a conversation for your own physician.

Battle Hard. — Will Power