New analysis questions its universal standalone first-line status, urging shared decision-making over blanket recommendations.
An umbrella review of the evidence base for exercise therapy in osteoarthritis has concluded that its effects on pain and physical function are likely minimal, short lived and in some cases no better than no treatment.
The findings have prompted researchers to call for reconsideration of its universal positioning as first line care.
“Evidence on exercise for osteoarthritis remains largely inconclusive, suggesting negligible or short-lasting small effects comparable to, or less effective than, other treatments,” they wrote in the BMJ Group’s RMD Open.
“These findings question its universal promotion and highlight the need to revisit research priorities and clinical discussions around its worthwhileness.”
The analysis synthesised data from systematic reviews and randomised controlled trials across knee, hip, hand and ankle osteoarthritis.
The authors reported findings that included small and transient improvements in knee osteoarthritis pain compared with placebo or no intervention, and diminishing effects in larger or longer-term studies.
Exercise is consistently recommended as an initial management strategy for osteoarthritis in clinical guidelines worldwide.
However, according to the researchers, an expanding body of evidence has questioned both the magnitude and durability of benefit.
While multiple systematic reviews have assessed exercise in isolation, the researchers noted that there has been no overarching synthesis comparing exercise with placebo, usual care, pharmacological therapies, injections, manual therapy, patient education and surgical interventions.
To address this gap, the investigators searched research databases for systematic reviews and randomised trials published up to November 2025. Five systematic reviews, comprising 8631 participants, and 28 randomised clinical trials involving 4360 participants were included.
Most trials focused on knee or hip osteoarthritis (23), with smaller numbers examining hand (three) and ankle (two) disease.
Pooled analyses indicated that exercise was associated with small, short-term reductions in knee osteoarthritis pain compared with placebo or no treatment, although the certainty of evidence was rated very low.
Effects were smaller still in larger or longer-duration trials. For hip osteoarthritis, moderate-certainty evidence suggested negligible benefit, while small effects were observed in hand osteoarthritis.
Comparative analyses across interventions showed broadly similar outcomes between exercise and patient education, manual therapy, analgesics, intra-articular steroid or hyaluronic acid injections and arthroscopic knee surgery.
In individual trials of specific patient groups, exercise appeared less effective over the longer term than corrective osteotomy or joint replacement surgery.
The authors acknowledged several limitations. They prioritised inclusion of selected systematic reviews and may have excluded others of potential relevance, although supplementary analyses of additional reviews suggested similar effect estimates.
Direct head-to-head comparisons were uncommon, participant populations varied widely in symptom severity, and some trials permitted concomitant treatments alongside exercise, potentially diluting or confounding effects.
Despite this, the researchers concluded that evidence supporting exercise therapy as a stand-alone, universally applicable first line intervention for osteoarthritis was weak.
They reported negligible or, at best, short-lasting small effects on pain and function across different joint sites compared with placebo or no treatment, with diminishing benefits in larger and longer-term trials.
However, they did emphasise that the findings did not negate the broader health benefits of physical activity, nor did they preclude individual patient preference for exercise-based approaches.
They said their findings did have implications for practice, as they questioned the “universal promotion of exercise therapy as the sole focus in first-line treatment to improve pain and physical function in all patients with osteoarthritis”.
“Varying certainty evidence indicates that exercise yields outcomes comparable to common non- pharmacological and pharmacological treatments and arthroscopic surgery,” they concluded.
“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.
“Low-to-moderate certainty evidence also suggests that exercise is less effective than knee osteotomy and total knee and hip replacement in selected patients, with more pronounced effects for hip osteoarthritis, aiding informed decision-making between patients and surgeons.”
