Load management

Tissue adapts to the loads it is progressively exposed to; harm follows spikes in load relative to preparedness, not load itself. Two things are established at the two ends of the range. Complete unloading thins cartilage measurably within weeks to months, and structured exercise does not raise biomarkers of cartilage breakdown — in the one trial to vary intensity in osteoarthritic knees, the hardest-working arm saw them fall furthest. Pain during sensible loading does not equal damage. The question is how to dose it, and the best-tested answer is a monitoring system rather than a fixed program.

The pain-monitoring trial

In a randomized trial of 38 people with Achilles tendinopathy, one group was allowed to keep training with pain up to 5/10 — provided it settled by the next morning and wasn't worsening week to week — while the other rested for six weeks first. The continue-training group recovered just as fully (VISA-A 57→85 vs 57→91, no significant differences), with no harm. This is the origin of the "acceptable pain, 24-hour rule" used throughout this section.

The rules, as used on this site

1. Traffic-light pain scale. 0–2/10 is safe, 3–5/10 is acceptable, above 5/10 is too much.
2. The 24-hour rule. Pain and swelling should return to baseline by the next morning. If they haven't, the dose was too high — reduce, don't stop.
3. Progress one variable at a time — load, volume, or frequency — and avoid spikes relative to your recent weeks of training.
4. Total rest is itself a harmful dose. Complete unloading thins cartilage measurably within weeks to months.

The supporting dose data agree with the structure. Across 217 randomised trials and 15,684 people with knee osteoarthritis, every exercise modality tested beat doing nothing on at least one outcome, and aerobic exercise ranked first overall; in a separate network of 46 resistance-training trials, all seven training types relieved pain against control. In an OA cohort each additional 1,000 daily steps predicted a 16–18% lower risk of functional decline. What no trial supports is a threshold below which exercise stops counting — the consistent finding is that structured, supervised, regular loading works and sporadic loading works less well.

The limits of this extrapolation

The pain-monitoring RCT is a tendon study. No trial validates the 5/10-and-settles thresholds for cartilage lesions or OA — applying them there is principled extrapolation, which is exactly why this entry is promising rather than strong. The acute:chronic workload ratio popularized in sports science has been criticized on statistical grounds; treat "avoid spikes" as a heuristic, not a metric to compute. And effusion and irritability after cartilage repair may not follow tendon-style pain rules — post-operative protocols override these generic rules entirely (see the post-op rehab entry).

Safety and red flags

Monitored loading produced no worse outcomes than rest in the RCT, exercise shows no cartilage-damage signal on imaging across 9 RCTs, and degradation biomarkers mostly fall or stay flat across 12 RCTs. But some signals override the model rather than feed into it: increasing swelling, night pain, or mechanical locking warrant clinical evaluation, not another week of titration. That escalation advice is standard clinical practice, not trial-derived.

Open questions

An RCT of pain-monitoring rules in knee OA or post-cartilage-repair populations is the obvious gap — along with whether morning-after effusion, rather than pain, is the better dial for cartilage patients.