Overloading and rushed return

The largest meta-analysis on running and joints puts recreational runners below sedentary people for hip and knee osteoarthritis — 3.5% versus 10.2%. It is the competitive, elite-volume end of the curve that climbs to 13.3%. That U-shape frames this whole entry: moderate load is associated with the healthiest joints, and the precaution lives at the extremes — spikes, sharp pain, and going back too soon after injury.

Why rate matters more than magnitude

Cartilage tolerates enormous cyclic load but is rate-sensitive. In explant models, injurious compression delivered at high strain rates suppresses proteoglycan and protein synthesis, degrades the tissue mechanically, and kills chondrocytes — while the same energy delivered slowly does far less damage. In living knees, a single marathon measurably shifts cartilage matrix composition on quantitative MRI (T1rho rising from 37.0 to 38.9 ms), and the change is still present at three months. The tissue's recovery timescale is months, not days.

The best-documented harm: rushing back after injury

The clearest cartilage disaster in this literature is the injured joint pushed straight back to full sport. At ten or more years, radiographic knee OA affects 0–13% of people after an isolated ACL injury but 21–48% when a meniscal injury comes with it. On average, roughly half of ACL- or meniscus-injured patients have knee OA 10–20 years later — typically in their 30s and 40s — and reconstruction has not been shown to prevent it. These studies cannot separate the injury itself from the decision to keep playing on it, but together with the months-long MRI recovery windows they argue for graded, criteria-based return rather than calendar-based.

Spikes, not volume

In team-sport data, injury risk rises when the acute workload spikes far above the chronic base an athlete is prepared for — while consistently high chronic workloads are protective. Be careful with the numbers, though: a rigorous critique found no solid evidence for the popular acute:chronic workload ratio thresholds, and the underlying injury endpoints are mostly muscle and tendon, not cartilage. What survives the critique is the qualitative principle: build gradually, don't spike.

Sharp pain, not any pain

This is the line that keeps the precaution honest. The one randomized trial to test continued sport during rehabilitation — Achilles tendinopathy, 38 patients — allowed loading with pain up to 5/10 provided it settled by the next day, and found no worse outcomes than active rest at any point over 12 months. Moderate, monitored pain is not the enemy. The signals to respect are sharp pain, escalating pain, and pain that is worse the next morning. A rule of "stop at any pain" would contradict the best evidence available.

That trial is tendon, not cartilage, and small — but it is the only randomized data on pain-guided loading, and it defines a workable ceiling.

What I take from it

Progression, not abstinence. The meta-analysis places recreational runners below sedentary controls, and the mirror-image precaution — immobilization — shows cartilage thinning within weeks of unloading. So: build load gradually without spikes; treat pain up to ~5/10 that resolves by the next morning as acceptable and sharp or next-day-worse pain as a stop signal; and after a joint injury, earn the return with criteria, not a date on the calendar. Whether structured load management changes long-term cartilage outcomes — not just injury counts — has never been tested, so this stays inference from converging lines, honestly labeled.