A systematic review with dose-response meta-analysis pooled 20 randomised trials of isolated isotonic resistance training in knee osteoarthritis, covering 791 participants. Trials mixing resistance work with other exercise were excluded, so this is about the resistance component alone.
The relationship was non-linear. Estimated pain reduction peaked at around 6,000 total repetitions, functional improvement at around 2,600 repetitions of moderate intensity, and strength gain at around 1,600. Very light to moderate intensity gave the largest pain reductions; moderate intensity the best function; moderate to vigorous intensity produced similar strength gains. Both home and centre-based programmes helped, with centre-based more pronounced.
The honest caveat is the certainty: low to very low throughout, on 791 patients across 20 small trials. These numbers are a shape, not a prescription. What they argue against is the common instruction to simply do more, and the assumption that a patient who is not improving needs a harder programme.
In Indian practice, where formal physiotherapy is often a few supervised sessions before the patient continues alone, the finding that home programmes work — just less well — is the practical half. It supports spending the supervised sessions on technique and a specific volume target rather than on delivering the exercise itself.
- Set a repetition target rather than telling a patient to do as much as they can tolerate.
- For pain as the main complaint, favour light to moderate intensity at higher volume over heavy loading.
- Use the supervised sessions to teach technique and set the dose, then transfer to a home programme.
- Reassess at a fixed point rather than escalating intensity because progress feels slow.
- Treat the certainty as low: this guides the shape of a programme, it does not settle it.
The statistics, in plain English
A mean difference of -3.52 for pain is on the scale the trials used, so read it as a direction and a peak rather than a number to quote to a patient. The important word is non-linear: benefit rose with volume, flattened, and did not keep climbing, which is why a peak can be named at all. GRADE certainty of low to very low means the true effects could differ substantially from these estimates once better trials are done, and with 791 patients spread across 20 trials the average study is small. Take the ordering of the findings as more reliable than any single figure.
Read the rest in the app
You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for rheumatology, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free