Loading the knee slowed injury-related OA — in male rodents

After a knee injury, post-traumatic osteoarthritis is a common long-term consequence, and altered joint loading is thought to be part of the story. That leaves clinicians and patients with a genuinely difficult question: should weight-bearing activity be restricted to protect the joint, or encouraged to keep cartilage and bone healthy? Human trials are hard to run because osteoarthritis develops slowly and joint tissue cannot be sampled directly. Rodent models allow direct histological assessment of cartilage and bone, which is why this review turned to animal data to look for a signal about how loading should be modulated after injury.

The authors searched PubMed, Cochrane Library, Embase and CINAHL through February 2025 for studies in rodents with knee post-traumatic osteoarthritis that compared locomotor exercise, on a treadmill or wheel, against no exercise. Outcomes had to include OARSI or Mankin osteoarthritis scores, bone quality, osteophyte condition, or cartilage quality and morphology. Twenty-two studies met criteria, yielding 156 effect sizes, all from male animals. Risk of bias was rated with the CAMARADES checklist, giving 5 studies low risk and 14 moderate risk. Where heterogeneity was significant, moderators such as locomotion speed, session duration, session count, intervention length, timing of exercise initiation, follow-up duration, species and osteoarthritis model were examined.

Locomotion significantly reduced osteoarthritis severity on composite OARSI or Mankin scores, with an effect size of 0.927 (95% CI 0.590 to 1.264), where a positive value means less disease. Bone quality showed an effect size of 0.379 with a confidence interval crossing zero. Isolated measures were small to moderate but non-significant: osteophyte condition 0.379, cartilage quality 0.430 and cartilage morphology 0.464, all with intervals touching or crossing the null. Moderator analyses found larger benefits for composite scores with shorter individual sessions, more sessions, longer overall intervention and longer follow-up. Bone quality benefits were greater with longer follow-up and later initiation of exercise after injury.

This challenges a purely mechanical view in which loading a damaged joint inevitably accelerates degeneration, at least in animal models. It also suggests a measurement lesson: composite joint scores appear more responsive to exercise than isolated cartilage or osteophyte outcomes, which may explain inconsistent findings across individual studies. The heterogeneity was substantial and most studies were only at moderate risk of bias, so precision is limited. The most obvious gap is sex: every effect size came from male animals, so nothing here speaks to females, and translation from rodent knees to human joints remains an assumption rather than a finding.

This is a reason to think that movement after knee injury is not automatically damaging to the joint, and may be protective — but the evidence is from rodents, not people. The pattern that more frequent, shorter bouts sustained over a longer period were associated with bigger benefits is a hypothesis worth watching in human research, not a training instruction. Because all the animals were male, even the animal conclusion has a large blind spot. Anyone managing a knee injury should treat this as background biology rather than guidance about their own loading.