Comparative Efficacy of Various Exercise Modalities and Doses in Reducing Pain and Improving Function in Older Adults With Knee or Hip Osteoarthritis: A Bayesian Network and Dose-Response Meta-analysis
*Systematic review and Bayesian network meta-analysis of randomized trials; Level 1a.
Citation
Tian S, Liang Z, Yu Y, Che W, Wang X. Comparative Efficacy of Various Exercise Modalities and Doses in Reducing Pain and Improving Function in Older Adults With Knee or Hip Osteoarthritis: A Bayesian Network and Dose-Response Meta-analysis. Archives of Physical Medicine and Rehabilitation. 2026;107:784–96. doi:10.1016/j.apmr.2025.08.013
Background
Exercise is widely recommended for osteoarthritis, but clinicians lack clear guidance on which exercise type and how much exercise works best for older adults.
Patients
Adults aged 60 years or older with knee and/or hip osteoarthritis. Excluded: exercise combined with other therapies (for example, heat or electrical stimulation) and programs shorter than 4 weeks.
Intervention
Aerobic, resistance, flexibility, yoga, tai chi, water-based exercise, and mixed programs.
Control
No exercise/usual care/waitlist, or a different exercise type.
Outcome
Pain and physical function.
Follow-up Period
End of intervention (4–104 weeks).
Results
| Outcome | Exercise type vs no exercise | Effect (standardized mean difference, 95% confidence interval) |
|---|---|---|
| Pain (primary) | Aerobic exercise | -1.19 (-1.66 to -0.74) |
| Function (primary) | Flexibility training | -1.07 (-1.68 to -0.47) |
Standardized mean difference combines different scales; more negative numbers mean less pain or better function.
A “U-shaped” pattern was found for weekly exercise amount (energy-use units called metabolic equivalent minutes per week): minimum helpful amount was about 240; best average amounts were about 580 for pain and 450 for function; benefit declined at higher amounts.
Limitations
Evidence certainty ranged from very low to moderate, with many trials at risk of bias and wide differences between studies. Possible publication bias for pain. Results reflect short-term effects only, and the “energy-use” dose measure may miss benefits unique to some exercise styles.
Funding
Chinese national social science fund and a provincial youth innovation program.
Clinical Application
For older adults with knee/hip osteoarthritis, prioritize aerobic exercise for pain and flexibility work for function; aim for moderate weekly volume (~240–600 energy-use units), avoiding extremes.
Discussion
Sign in to join the discussion.
In this Bayesian network and dose-response meta-analysis of RCTs, aerobic exercise showed the largest pain reduction versus control (SMD −1.19, 95% CI −1.66 to −0.74); given low/very-low-to-moderate certainty and indirect comparisons, should this change your exercise prescriptions for older adults with knee/hip OA? In this article, pain and function were measured using different scales across trials, so the authors reported standardized mean differences (SMDs), such as SMD −1.19 for aerobic exercise versus control for pain. What is the best interpretation of an SMD in this context?