Exercise for osteoarthritis shows minimal benefit in the most comprehensive evidence review

Exercise is embedded in virtually every major clinical guideline for osteoarthritis. Rheumatology societies, physiotherapy organizations, and primary care frameworks consistently list it as a first-line treatment. That consensus rests on a substantial body of trial data. A new overarching review published in RMD Open examines what that data actually shows when pooled and assessed rigorously together.

The researchers conducted an umbrella review - analyzing both existing systematic reviews and individual randomized trials. They identified 5 systematic reviews covering 8,631 participants and 28 randomized clinical trials covering 4,360 participants. The trials addressed knee and hip osteoarthritis (23 trials), hand osteoarthritis (3 trials), and ankle osteoarthritis (2 trials). Effect sizes were pooled and compared against placebo, no treatment, usual care, painkillers, steroid injections, hyaluronic acid injections, and surgical options.

What the pooled data shows

For knee osteoarthritis - the most heavily studied type - pooled analysis showed small, short-lived improvements in pain compared with placebo or no treatment. The certainty of evidence was rated very low, and in larger and longer-term trials the effect sizes were smaller still. For hip osteoarthritis, moderate certainty evidence indicated negligible effects. For hand osteoarthritis, small effects at low certainty.

When exercise was compared to other treatments rather than to no treatment, outcomes were comparable across a wide range: patient education, manual therapy, painkillers, corticosteroid injections, and hyaluronic acid injections all produced similar results. In individual trials examining longer-term outcomes, exercise was less effective than bone-remodeling surgery (osteotomy) and joint replacement.

"We found largely inconclusive evidence on exercise for osteoarthritis, suggesting negligible or, at best, short-lasting small effects on pain and function compared with placebo or no treatment," the researchers concluded. "These effects appear less pronounced in larger and longer-term trials."

Why smaller effects in larger trials matters

The pattern of effects shrinking in larger, longer trials is informative. Small trials are prone to biases - enthusiastic investigators, selected patient populations, regression to the mean - that inflate apparent effects. When those biases are diluted across larger, more representative studies, smaller true effects become visible. The trend in the osteoarthritis exercise data toward smaller effects in the trials best positioned to detect real outcomes is precisely what would be expected if the true effect were minimal.

The researchers explicitly state that exercise is not harmful or without value. It provides well-documented benefits for cardiovascular health, metabolic function, mental health, and overall mortality. For some individual patients, exercise may meaningfully reduce osteoarthritis pain even if the population-average effect is small. Exercise carries essentially no serious safety risks, is inexpensive, and does not foreclose other treatment options.

What the review questions is the universality of the recommendation - the practice of prescribing exercise to all osteoarthritis patients as the default first response, before considering whether they are likely to respond and what alternatives might work better for them.

Limitations of the umbrella review

The researchers acknowledge several constraints. They prioritized certain systematic reviews in their selection and may have excluded others that were relevant, though additional analysis of excluded reviews showed similar results. Head-to-head comparisons between exercise and specific alternative treatments were limited in the included studies. There was considerable variation in participant characteristics - particularly symptom severity - across trials, making it possible that subgroups respond differently than pooled averages suggest. Some trials allowed concurrent treatments alongside exercise, which could distort apparent effects.

"Clinicians and patients should engage in shared decision-making, weighing the worthwhileness of exercise effects on pain and function alongside secondary health benefits, safety, low-cost profile, care stage, and alternative treatment options," the researchers advise.

The practical implication is a reconsideration of exercise automatic elevation to universal first-line status in osteoarthritis - and a recognition that for patients whose primary concern is pain relief and functional improvement, other options deserve equal consideration from the start rather than as fallbacks after exercise has proved insufficient.