There were 11 trials with a total of 907 patients. Two trials were excluded, one because the response data could not be extracted from the cross-over design and the other because the follow-up period was not stated. Ten trials examined the effect of epidural versus placebo in the short-term and 5 trials examined the long-term effect.
The quality assessment of trials was generally good with 5 studies scoring the maximum nine points.
Short-term effect of epidural steroids versus placebo on pain relief.
The pooled OR for >75% relief was 2.61 (95% CI: 1.80, 3.77) in favour of epidural. OR for complete relief was 2.79 (95% CI: 1.92, 4.06). The OR for caudal epidural steroids was 3.80 (95% CI: 1.36, 10.6) and the OR for lumbar steroids was 2.43 (95% CI: 1.77, 3.74).
Long-term effect of epidural steroids versus placebo on pain relief.
The pooled OR for >75% relief was 1.87 (95% CI: 1.31, 2.68).
4 trials did not mention adverse events. Pooled data from the other trials (431 patients) reported 11 dural taps (2.5%), 10 transient headaches (2.3%), transient increase in pain (1.9%) and one patient complained of irregular periods (0.2%). No meaningful comparisons were made, however, between active and control group.