Objectives: The aim of this study was to systematically assess the long-term (≥ 6 months) benefits of epidural steroid injection therapies for patients with low back pain.
Methods: We identified randomized controlled trials by database searches up to October 2011 and by additional hand searches without language restrictions. Randomized controlled trials on the effects of epidurals for low back pain with follow-up for at least 6 months were included. Outcomes considered were pain relief, functional improvement in 6 to 12 months after epidural steroid injection treatment and the number of patients who underwent subsequent surgery. Meta-analysis was performed using a random-effects model.
Results: Twenty-nine articles were selected. The meta-analysis suggested that a significant treatment effect on pain was noted at 6 months of follow-up (weighted mean difference [WMD], -0.41; 95 percent confidence interval [CI], -0.66 to -0.16), but was no longer statistically significant after adjusting for the baseline pain score (WMD, -0.19; 95 percent CI, -0.61 to 0.24). Epidural steroid injection did not improve back-specific disability more than a placebo or other procedure. Epidural steroid injection did not significantly decrease the number of patients who underwent subsequent surgery compared with a placebo or other treatments (relative risk, 1.02; 95 percent CI, 0.83 to 1.24).
Conclusions: A long-term benefit of epidural steroid injections for low back pain was not suggested at 6 months or longer. Introduction of selection bias in the majority of injection studies seems apparent. Baseline adjustment is essential when we evaluate pain as a main outcome of injection therapy.