When should you order an MRI scan before performing an epidural injection for lower back pain?

A recent postgraduate education release¹ is entitled: ‘MRI Before Epidural Injection Not Beneficial in Back Pain’ referring to a recent study comparing treatment decisions with and without magnetic resonance imaging (MRI) results.² But, there should be an addendum: ‘in Certain Circumstances.’ Let us consider the facts.

It is well accepted that approximately 50% of lumbar MRI scans in asymptomatic 50 year olds demonstrate potential discogenic pathology.³ Therefore, the indiscriminate use of MRI in managing non-specific lower back pain is generally unlikely to aid clinical management and is thus not cost effective.

The indication for performing an epidural steroid injection is primarily to address dural inflammation. Sciatica is the most common condition presenting with clinical symptoms and signs that can confidently be accounted for by this aetiology: in other words, nerve root compromise. In healthy middle aged adults, sciatica is generally accounted for by posterolateral L4/5 or L5/S1 intervertebral disc protrusions/herniations.

We were able to demonstrate the efficacy of non-image-guided caudal epidural steroid injections in managing intractable sciatica, diagnosed on clinical grounds (without an MRI scan) in a controlled prospective study.⁴ Subsequently, we managed 165 patients with sciatica, diagnosed both clinically and with correlating computerized tomography (CT) evidence of nerve root compromise. Eighty-six percent of patients responded well to epidural steroid injections, making a good natural recovery without the need for decompressive surgery.⁵ These patients received an average of three epidural steroid injections, 76% of which were non-image-guided caudal epidural steroid injections but 24% were image-guided transforaminal steroid injections (i.e. dorsal root ganglion blocks).

Taking these two studies into account, one could argue that a significant number of patients presenting with sciatica (diagnosed on strict clinical criteria and without a CT or MRI scan demonstrating correlating pathology) will recover with non-image-guided epidural steroid injections irrespective of MRI scan findings. Implementing this philosophy could reduce costs.

In contrast, there are a number of patients presenting with sciatica who have more sinister pathology or who have, for instance, far lateral disc herniations compromising the nerve root either within or lateral to the intervertebral foramen. This pathology will not be addressed by a non-image-guided caudal or interlaminar injections. Furthermore, it is well established that non-image-guided injections have an incidence of mis-tracking for various reasons (e.g. extra-sacral spread or intravascular spread).

We audited our consecutive series of 535 patients who underwent image-guided injections through the 12 months of 2009, and we found that on average these patients had 1.5 injections, i.e. 50% less than in our 1992 series.⁵

In conclusion, a high proportion of carefully selected patients with sciatica will get better without MRI scans and with non-image-guided epidural steroid injections. In contrast, if one utilizes MRI to demonstrate a specific target and then uses imaging to precisely reach that target, patients are likely to get better with half the number of injections and, by implication, more quickly too. Which would you prefer if you were the patient?