Background: Use of epidural corticosteroid injections is increasing.
Purpose: To review evidence on the benefits and harms of epidural corticosteroid injections in adults with radicular low back pain or spinal stenosis of any duration.
Data sources: Ovid MEDLINE (through May 2015), Cochrane Central Register of Controlled Trials, Cochrane Database of Systematic Reviews, prior systematic reviews, and reference lists.
Study selection: Randomized trials of epidural corticosteroid injections versus placebo interventions, or that compared epidural injection techniques, corticosteroids, or doses.
Data extraction: Dual extraction and quality assessment of individual studies, which were used to determine the overall strength of evidence (SOE).
Data synthesis: 30 placebo-controlled trials evaluated epidural corticosteroid injections for radiculopathy, and 8 trials were done for spinal stenosis. For radiculopathy, epidural corticosteroids were associated with greater immediate-term reduction in pain (weighted mean difference on a scale of 0 to 100, -7.55 [95% CI, -11.4 to -3.74]; SOE, moderate), function (standardized mean difference after exclusion of an outlier trial, -0.33 [CI, -0.56 to -0.09]; SOE, low), and short-term surgery risk (relative risk, 0.62 [CI, 0.41 to 0.92]; SOE, low). Effects were below predefined minimum clinically important difference thresholds, and there were no longer-term benefits. Limited evidence showed no clear effects of technical factors, patient characteristics, or comparator interventions on estimates. There were no clear effects of epidural corticosteroid injections for spinal stenosis (SOE, low to moderate). Serious harms were rare, but harms reporting was suboptimal (SOE, low).
Limitations: The review was restricted to English-language studies. Some meta-analyses were based on small numbers of trials (particularly for spinal stenosis), and most trials had methodological shortcomings.
Conclusion: Epidural corticosteroid injections for radiculopathy were associated with immediate reductions in pain and function. However, benefits were small and not sustained, and there was no effect on long-term surgery risk. Limited evidence suggested no effectiveness for spinal stenosis.