Cochrane: exercise eases hip arthritis only slightly

Hip osteoarthritis is a major public health issue, and international treatment guidelines recommend exercise as a first-line way to manage symptoms. That recommendation has stood for years, but the underlying trial evidence has grown considerably since this Cochrane review was first published in 2009 and last updated in 2014. The open question is not whether exercise is plausible — it is whether the measured benefit for pain, function and quality of life is large enough to matter to a patient, and whether it adds anything when layered on top of the other care people already receive.

The reviewers searched CENTRAL, MEDLINE, Embase and two trial registries from February 2013 to 5 February 2025, including randomised controlled trials of adults with hip osteoarthritis. Three comparisons were eligible: exercise versus attention control or placebo, exercise versus no treatment, usual care or limited education, and exercise plus a co-intervention versus that co-intervention alone. Perioperative programmes and interventions involving vibration therapy, gait aids or gait retraining were excluded, as were head-to-head comparisons of exercise types. Eighteen studies with 1368 participants qualified. Critical outcomes were pain, physical function and quality of life; important outcomes were participant-reported treatment success, withdrawals and adverse events. Continuous estimates were converted to mean differences on a 0 to 100 scale, and certainty was rated using GRADE.

Against no treatment, usual care or limited education, exercise probably reduced pain slightly (MD -7.19 points, 95% CI -10.70 to -3.68; 9 studies, 449 participants) and probably improved physical function slightly (MD -8.79 points, 95% CI -12.00 to -5.41), both at moderate certainty — yet the reviewers state these gains are unlikely to be clinically meaningful. Quality of life probably changed little (MD 2.31 points, 95% CI -1.15 to 5.91). Against attention control or placebo, exercise may have had little to no effect on pain (MD -6.31 points, 95% CI -12.98 to 0.35) while slightly improving function (MD -7.44 points, 95% CI -13.86 to -1.01), from just 2 studies and 123 participants. Adding exercise to an existing co-intervention probably did little for pain, function or quality of life across 7 studies and 751 participants, though it probably reduced adverse events slightly (RR 0.75, 95% CI 0.58 to 0.97).

This update complicates rather than overturns the guideline position: exercise remains safe and modestly helpful, but the effect sizes are smaller than the enthusiasm around exercise therapy might imply, and they shrink towards nothing when exercise is added to other care. Most included studies were small and unblinded, which limits certainty and means placebo and attention effects are hard to separate from the exercise itself. Evidence on quality of life and participant-reported treatment success was sparse or absent in some comparisons. What is still unknown is which exercise types, doses or patient subgroups produce benefits people actually notice, since head-to-head comparisons of exercise forms were outside this review's scope.

This is a reason to hold realistic expectations rather than to abandon movement: for hip osteoarthritis, structured exercise appears to nudge pain and function in the right direction without obvious harm, but on average the shift is small. The finding that exercise added on top of other care probably changes little suggests overlapping rather than additive benefits. Reassuringly, in that same comparison exercise probably reduced adverse events slightly, which argues against the fear that loading an arthritic joint makes things worse. Anyone weighing options can reasonably read this as evidence that exercise is a defensible, low-risk component of care rather than a decisive fix.'