Objective: To validate the effectiveness of six exercise therapies in treating low back pain using Meta-analysis methods, and to propose optimal exercise duration, frequency, and cycle.
Methods: Databases such as PubMed, EMBASE, Cochrane Library, China National Knowledge Infrastructure, VIP Data, and SinoMed were searched. The RevMan 5.4 tool was utilized to conduct subgroup analyses on eight moderating variables, including types of exercise, duration, frequency, cycle, sample size, study quality, outcome indicator, and comparisons of different intervention methods with the control group from 42 included studies. Random effects models were employed to test for overall effects, heterogeneity, and bias.
Results: The overall effect size for six exercise therapies for low back pain was significant (SMD= -1.21, P < 0.00001). Subgroup analyses showed yoga had the largest effect (SMD= -1.97, P = 0.0001). Exercise duration ≤ 30 min (SMD= -1.31, P < 0.0001), frequencies > 4 times/week (SMD= -1.56, P < 0.00001), and cycles ≤ 4 weeks (SMD= -1.61, P < 0.00001) were most effective. Sample sizes of 30~60 cases (SMD= -1.36, P < 0.00001) and studies with moderate bias risk (SMD= -1.37, P < 0.00001) also showed large effects. The Oswestry Disability Index scores demonstrated the most significant effect size (SMD= -3.35, P < 0.00001). The effect size of the physical factors in the control group was the largest (SMD= -1.85, P < 0.00001).
Conclusion: All six exercise therapies effectively alleviated low back pain, with yoga showing the best results. The optimal exercise intervention protocol involved exercise duration not exceeding 30 min per session, frequency of more than 4 times per week, and cycle not exceeding 4 weeks. Additionally, exercise interventions exhibited the most significant improvements in Oswestry Disability Index scores for low back pain.
Keywords: Effect size; Exercise intervention; Low back pain; Meta-analysis; Randomized controlled trial.
© 2025. The Author(s).