Is epidural injection of steroids effective for low back pain?

Low back pain is a problem that is common and costly to society, and its effective management remains a challenge.¹,² Exercise programmes combined with early return to normal activities have been shown to be beneficial in chronic low back pain.^2-3 Other interventions may also have a beneficial effect, and epidural injection of steroids represents one such alternative. This editorial examines the evidence to determine whether such treatment is justified.

In clinical practice a structured approach at the initial consultation facilitates the evaluation of patients with low back pain.⁴ The presence of “red flag” signs indicates possible serious underlying pathology and warrants urgent referral to a specialist unit.⁵ Usually, “leg pain dominant” features indicate lumbosacral nerve root irritation or entrapment, whereas “back pain dominant” features indicate a biomechanical cause. Mechanical back pain may often be associated with some nerve root irritation, or present as an exacerbation of chronic symptoms.

The therapeutic effects of epidural steroid injections are attributed to an inhibition of the synthesis or release of pro-inflammatory substances.⁶ Corticosteroid delivered into the epidural space is able to attain high local concentrations.⁷ Reports on thousands of patients indicate that epidural corticosteroid injections are relatively straightforward and safe.⁸

Recent reviews on epidural steroid injection have shown variable results. In a review of 12 randomised clinical trials half showed that epidural steroid was more effective, whereas the other half reported it to be no better or worse than the reference treatment.⁹ A meta-analysis of 11 placebo controlled trials showed an improvement (75% reduction of pain) both in the short term (1-60 days, odds ratio 2.61, 95% confidence interval 1.8 to 3.77), as well as in the long term (12 weeks-1 year, 1.87, 1.31 to 2.68).¹⁰ The conclusions drawn were that epidural steroid is effective in the management of sciatica accompanying low back pain. Analysis by the numbers needed to treat gives a measure of the clinical benefit of this study. The numbers needed to treat were six (95% confidence interval 4 to 12) for short term benefit and 11 (6 to 90) for long term benefit.¹¹

A recent Cochrane review of 21 randomised trials of all types of spinal injection therapy for low back pain has shown a leaning towards a positive effect, although this benefit is not supported by unequivocal evidence.¹² Only four randomised placebo controlled trials of epidural injection were identified (Beliveau 1971; Cuckler 1985; Bush and Hillier, 1991; Carette 1997).¹⁰ All report pain relief in more patients with active treatment compared with placebo. However, the pooled data did not show conclusive benefit. The relative risk for short term improvement was 0.93 (0.79 to 1.09) and 0.92 (0.76 to 1.11) in the long term. ¹²

The beneficial effects of epidural injection might not have been apparent for several reasons.¹² Firstly, the sample size in these studies may have been too small to detect clinically relevant differences. Secondly, the recovery rate from placebo was substantial. This might have been because of the natural course of recovery of back pain. Thirdly, the volume of fluid and technique used has varied between studies. The effect of this as a confounding factor remains unknown. Fourthly, low back pain is not a discrete entity, and the diagnosis of the origin of pain can be extremely difficult. These trials may have included a heterogeneous population rather than a specific disease group. Fifthly, epidural steroid injection may be ineffective.

A need exists for well designed trials of adequate size to determine the effectiveness of epidural injection in back pain. Such trials should compare the effect of injection of anaesthetics with or without steroid against placebo and should have homogeneous groups of patients, baseline comparability of groups, intention to treat analysis, and clinically relevant outcome measures.¹²

What then are the implications for clinical practice? Epidural injection therapy has not yet been shown to be effective, nor has it been shown to be ineffective.¹² Side effects are relatively minor, and a tendency exists towards an outcome favouring injection therapy. On the basis of our longstanding clinical experience we suggest that epidural steroid injection may have a role in specific clinical situations. Low back pain that has not resolved within three months leads to greater long term morbidity.¹² Epidural injection therapy may provide a useful adjunct to recovery in patients whose symptoms have extended beyond three months in the absence of recognised indicators of chronicity (“yellow flags”),⁴ and who may have radicular symptoms. The evidence for and against epidural steroid injection should be clearly explained to allow patients to make an informed choice regarding their treatment.