Context

Epidural steroid injections are commonly used for the treatment of lumbar radiculopathies which have failed other conservative measures such as physical therapy or non-steroidal anti-inflammatories. Although the use of such injections has risen considerably over recent years, their efficacy remains unclear. This randomised trial compared disability following epidural steroids, epidural saline or sham injections.

Findings

There were no statistically significant differences between any of the three groups at any time points; and while each group improved somewhat for both primary and secondary measures, these improvements failed to reach accepted levels of clinical significance.

Commentary

This study had several limitations: (A) based upon the authors' power calculations, the study is slightly underpowered. Insufficient power is always concerning when no differences are noted between groups; but the numbers in this study are ‘close’ and no trends were identified, suggesting likely validity; (B) patients were included based solely upon clinical symptoms, signs and examination; not MRI findings (that could impact treatment efficacy), although all enrolled patients did receive MRI or CT; (C) only one type of steroid in one dose was tested. Many other dosages, types and combinations are used by different centres; so generalisability might be a concern; and (D) the approach used was caudal and a few previous randomised controlled trials (RCTs) have demonstrated that trans-foraminal injections might provide superior results (although these RCTs were of poor quality).

Characteristics of previous RCTs of lumbar epidural steroid injections for radiculopathy vary widely – including the length of follow-up, steroid used, control group and outcome measures, so definitive answers are difficult to provide. Studies that include acute radiculopathies, short-term follow-up, and emphasise radicular pain as the outcome measure tend to find a positive response; whereas those assessing more chronic symptoms, longer follow-up, clinically important differences and functional outcomes tend to find no response – the current study being an example. This study, however, has some strengths that set it apart: (A) the inclusion of sham and saline epidural injection groups provides better controls than those used in previous studies, (B) the multiple and extended time points for data collection help to exclude short-term or delayed-onset effects, (C) the authors include a wide range of blinded specialists and this diversity affords some comfort. There is a sense from previous studies that one could predict the findings and conclusions based solely upon the occupation of the authors; that is, pain management physicians finding significantly positive impact, and surgeons finding none. Protection of turf and the profit motive cannot be denied; and bias need not be at the point of ‘interpretation of the data’, but at the point of ‘establishment of methods’ – the answers one gets depend most tightly upon the questions one asks.

Sir Karl Popper famously stated that science advances by falsification – weeding out the theories that are contradicted by evidence or, in the case of medicine, the ineffective therapies. He also recognised that such weeding is subject to the perspective of the person doing the weeding. Those who envision epidural injections for chronic radiculopathy as at best a psychological ointment (‘We did everything prior to going forward with surgery’) will take the study's results as further confirmation of their position against their use; while believers will suggest that alternative steroids or techniques do the trick and that the results are not generalisable to their practice. Thus, the implications will likely (again) fall to payers, who, if convinced of falsification or demand verification, of the procedure might deny reimbursement. Only a middle ground (‘We do not yet have enough evidence either way’) sustains payments and, indirectly, the use of such injections for this indication.