Abstract
OBJECTIVE: To compare the effectiveness of exercise therapy with surgery for lumbar spinal stenosis. METHODS: Five English databases PubMed, the Cochrane Library, Web of science, OVID and PEDro database were searched for randomized controlled trials comparing surgical procedures with exercise therapy for lumbar spinal stenosis. Information on patients, study design, inclusion criteria, intervention and follow-up, outcomes, treatment details and adverse events were extracted. Meta-analysis was performed using Review Manager Version 5.3. RESULTS: Two randomized controlled trials and one mixed design trial with a total of 897 patients were included. The pooled results showed a significant difference between exercise and surgery in Oswestry Disability Index at two years (MD= 3.85, 95%CI: 0.48 to 7.22; P=0.03), but no significant difference at six months (MD= 2.18, 95%CI: -2.80 to 7.17; P=0.39) and one year (MD= 4.26, 95%CI: -1.79 to 10.32; P=0.17). In terms of physical function of 36 Items Short Form Health Survey, there were no significant differences between exercise and surgery at six months (MD= -2.23, 95% CI: -7.46 to 2.99; P=0.40), one year (MD= -2.17, 95% CI: -7.44 to 3.10; P=0.42) and two years (MD= -0.67, 95% CI: -6.16 to 4.82; P=0.81). CONCLUSION: In brief, the current evidence demonstrated a trend that exercise therapy had a similar effect for lumbar spinal stenosis compared with decompressive laminectomies. However, for the small sample size and low methodology quality of the included trials, some rigorously designed and large-scaled RCTs need to be performed to confirm the conclusion.