Effectiveness of exercise therapy for osteoarthritis: an overview of systematic reviews and randomised controlled trials, 2026, Schleimer et al

[Mij]

Massive review suggests exercise may do little for osteoarthritis pain

Exercise may not be the osteoarthritis game-changer many hoped it was

Date: February 25, 2026
Source: BMJ Group

Summary: A sweeping new analysis of the evidence suggests that exercise therapy — long promoted as a first-line treatment for osteoarthritis — may offer only small and short-lived relief, and in some cases might be no better than doing nothing at all. After reviewing dozens of clinical trials involving more than 13,000 participants, researchers found that benefits for knee osteoarthritis pain were minimal and tended to shrink in larger or longer-term studies.

A sweeping (umbrella) systematic review and pooled analysis published in the open access journal RMD Open suggests that exercise therapy may provide only minimal and short lived relief from osteoarthritis symptoms. In some cases, the benefits may be little different from receiving no treatment at all.

Study

​

 

[Mij]

Conclusion and relevance Evidence on exercise for osteoarthritis remains largely inconclusive, suggesting negligible or short-lasting small effects comparable to, or less effective than, other treatments. These findings question its universal promotion and highlight the need to revisit research priorities and clinical discussions around its worthwhileness.

 

[Mij]

We know all about that don't we.

 

[Jonathan Edwards]

Did anyone really think it did anything? I guess some did. I don't remember seeing any decent evidence.

 

[rvallee]

benefits were ... minimal and tended to shrink in larger or longer-term studies

It's literally always the case. Literally. Always. Funny how that works. I'm thinking we need the concept of the gullibo effect, for people who accept too-good-to-be-true bullshit findings just because they want them to be true. This makes up the bulk of all non-pharmaceutical treatments. Still with the old wanting-magic-to-be-true.

Over-stating the benefits of exercise will end up doing more harm than good in the long run. It is good, at not being unfit. Otherwise, it all turns to be as simple as: move your body, not too little, not too much, making it fun is more important than doing it hard, there is no right way, but plenty of wrong ones.

And that's despite such evaluations always significantly under-estimating biases:

Overall, no study was rated as low risk of bias, 20 (71%) studies were at some concerns, and eight (29%) studies were rated as high risk of bias.

Always the same thing. Maybe biasebo? It's all just bias and people just overlook conflicting evidence.

Common sources of bias were the measurement of the outcome (in none of the studies were patients blinded) and the selection of the reported results (eight trials were retrospectively registered).

When you describe both the problem and the standard way things are done at the same time. Oops?

Keeping in mind that most of the conclusions are of the same language and nature as the ME/CFS exercise review, stating that the interventions "probably help", but the review speaks of very low certainty, even though the ME/CFS exercise trials are even worse in all aspects.

Further, the decision to re-analyse certain reviews was based on the certainty of evidence. However, clear criteria for defining conclusive evidence are still lacking29 and the GRADE approach has been criticised for its subjectivity113 and low validity in predicting the stability of an effect.114 Defining conclusiveness of evidence could prevent research waste and shift the focus towards more important research areas.

You don't say? We have said so as well.

But, their conclusions? You guessed it. You know you did:

Given the stability of the mean estimates, but large heterogeneity in the observed effects of exercise, future research could focus on identifying subgroups of patients that better respond to certain exercise types or doses.

More research is needed. Of course.

Future pragmatic clinical trials should focus on comparative effectiveness of exercise and physical activity interventions (both against other treatments and across different types/doses of exercise)

That should waste a decade or two keeping up this sham.

There appears to be no consideration of what it means for so much clinical evidence to be of such poor quality. No, rather they encourage it to continue. Because biasebo.