OBJECTIVE: To compare the effects of aerobic exercise, resistance exercise, and combined interventions on pain, function, and quality of life (QoL) in patients with knee osteoarthritis (OA) at the short term (=12 weeks) and long term (>12 weeks). DESIGN: Intervention systematic review with Bayesian network meta-analysis (NMA). LITERATURE SEARCH: MEDLINE, PubMed, EMBASE, CENTRAL, CINAHL, PEDro, AMED, and SPORTDiscus were searched from inception until July 2025. STUDY SELECTION CRITERIA: We included randomized controlled trials of patients with knee OA that evaluated exercise interventions aimed at improving pain, function, and QoL. DATA SYNTHESIS: Using random-effects models, the NMA synthesized effect sizes and ranked interventions by the surface under the cumulative ranking curve (SUCRA), with evidence certainty assessed via the Grading of Recommendations Assessment, Development and Evaluation (GRADE) framework. RESULTS: Thirty-five trials (n = 3808 patients) were included. For pain, short-term combined aerobic and resistance exercise had the largest effect among the interventions (standardized mean difference [SMD] = -0.937, low certainty). For function, short-term combined aerobic and resistance exercise had the largest effect estimates (SMD = -0.590, low certainty), followed by long-term aerobic exercise (SMD = -0.540, very low certainty). Short-term combined aerobic and resistance exercise had the largest effect (SMD = 0.663, low certainty) for improving QoL. CONCLUSION: Aerobic and resistance exercise, alone or in combination, and delivered over the short- or long-term, improved pain, function, and QoL for patients with knee OA. Short-term combined aerobic and resistance exercise had clinically meaningful improvements for pain, with relatively favorable effects on function and QoL. Evidence regarding the clinical significance of long-term exercise interventions was limited. J Orthop Sports Phys Ther 2026;56(11):724-734. Epub 24 July 2026. doi:10.2519/jospt.2026.13875.
| Discipline Area | Score |
|---|---|
| Physician | ![]() |