Exercise for Knee and Hip Osteoarthritis: Who Benefits, and How Much?

Exercise is recommended as first-line treatment for knee and hip osteoarthritis in every major clinical guideline, yet clinicians have long wondered whether certain patients — older, heavier, more structurally advanced, or in more pain — respond differently. Holden and colleagues addressed this using individual participant data (IPD) meta-analysis, published in The Lancet Rheumatology in 2023 through the international OA Trial Bank collaboration.

Unlike conventional meta-analyses that pool study-level averages, IPD analysis re-analyses the raw data of individual participants across trials, giving far greater power to detect whether specific patient characteristics moderate treatment response. The team pooled data from randomised controlled trials comparing therapeutic exercise with non-exercise control conditions in people with knee osteoarthritis, hip osteoarthritis, or both.

Two findings stand out. First, therapeutic exercise produced a small, positive overall effect on pain and physical function compared with non-exercise controls — but the authors were candid that the average effect is of questionable clinical importance, particularly in the medium and long term. Second, the analysis did not identify patient subgroups for whom exercise should be withheld: no characteristic emerged as a strong, consistent moderator of response. In other words, exercise helps modestly on average, and there is no evidence-based reason to deny it to any particular group.

The practical interpretation is nuanced. Rather than undermining exercise therapy, these results have sharpened the field's focus on implementation: individualised dosing, supervised progression, adherence support, weight management where relevant, and long-term behaviour change, since short courses of generic exercise are unlikely to deliver durable benefit on their own. Exercise remains the foundation of osteoarthritis care — the challenge is delivering it well enough, for long enough, to matter.

Key findings

• The study pooled individual participant data from randomised controlled trials comparing therapeutic exercise with non-exercise controls in people with knee and/or hip osteoarthritis, coordinated through the OA Trial Bank.
• Therapeutic exercise showed a small, positive overall effect on pain and physical function compared with non-exercise controls.
• The authors noted that the average effect size is of questionable clinical importance, particularly at medium- and long-term follow-up — an honest framing that has fuelled ongoing debate about how exercise is dosed, delivered, and maintained.
• No patient characteristics emerged as strong, consistent moderators that would justify withholding exercise from particular subgroups, supporting exercise as a broadly applicable first-line treatment.
• The findings argue for improving how exercise is implemented — adherence, progression, and long-term maintenance — rather than abandoning it.

Clinical relevance

Osteoarthritis is among the most common presentations in rehabilitation settings, and this analysis supports offering exercise to essentially all patients while pushing clinics to focus on adherence, progression, and long-term support rather than short generic programs.

Prepared by the Volt Health team as an educational summary of the cited publication. It is not medical advice; assessment and treatment decisions should always be made with a qualified health professional.