Do epidural steroid injections relieve radicular lumbosacral pain between two and six weeks after the injection better than placebo? And, in general, does epidural steroid injection affect function, the need for surgery, or provide long-term relief beyond three months? These were among the questions considered by the AAN Technology and Therapeutics Assessment Subcommittee, which reviewed and released their assessment of the best available evidence Mar. 6 in Neurology (68:723–729).
Charles Argoff, MD, assistant professor of neurology at the New York University School of Medicine, spoke to Neurology Today about the evidence and, in an accompanying sidebar, offers a case example exemplifying how a physician might apply the new information to patients. Lead author Carmel Armon, MD, chief of neurology at Baystate Medical Center in Springfield, MA, also answered questions here about the guidelines.
WHAT CONDITION DOES EPIDURAL STEROIDAL INJECTION BENEFIT BASED ON THE AVAILABLE EVIDENCE CITED IN THE PAPER?
Available evidence suggests the injections may result in some improvement in radicular lumbosacral pain between two and six weeks after the injection.
HOW STRONG IS THE BENEFIT?
Based on the strongest papers cited in our review, this remains a treatment whose ultimate value is unknown. Some health-care professionals or others may use this review to say the value of this procedure is overstated. Others who do this procedure, as well as patients whose lives have been so positively changed by this procedure, may hope that further studies better define the true value.
In my own practice, I have seen patients incapacitated by their pain return to their work and sports activities two weeks after receiving these injections. Our review of the literature, however, “found the average effect difference (advantage of steroids over control treatment) was small, usually falling short of the value proposed as a clinically meaningful average difference.”
WHY DO YOU THINK THE REVIEW FOUND THE BENEFIT FELL SHORT OF THE CLINICALLY MEANINGFUL AVERAGE DIFFERENCE?
One reason is that when the results of a study are reported in terms of average improvement, the responses of patients who benefited are diluted by responses of those who did not. A better way of reporting the results is percentage of patients who achieved a clinically meaningful response. However, none of the studies provided data in that way.
HOW MANY PLACEBO-CONTROLLED TRIALS HAVE BEEN DONE?
I don't have the total number, but we identified four articles evaluating the efficacy of epidural injections (without fluoroscopic guidance) that had placebo controls and met additional criteria. We also considered two articles identified as the best two by an Institute of Clinical Systems Improvement review of fluoroscopically guided transforaminal epidural injections.
DID ANY GROUP OF PATIENTS APPEAR TO BENEFIT THE MOST?
At this time, there are no studies that identify such patients. Further studies that more clearly define different subtypes of patients with low-back pain and radiculopathy should be done to learn more about which patients are best suited for this treatment.
ARE THERE PATIENTS FOR WHOM THE EVIDENCE SHOWED NO BENEFIT?
The review found insufficient evidence to recommend its use in the treatment of radicular cervical pain. Yet occasionally, such patients do seem to respond to this treatment.
Can these injections be recommended if occasionally a patient would benefit? If a patient has not benefited either from exercise, lifestyle changes, proper nutrition, acupuncture, non-opioid therapies, is there any role for the injections in the absence of more formal data in non-radicular lumbosacral pain? It's hard to suggest such treatments given the data.
Admittedly, the treatment of non-radicular lumbosacral pain can be quite challenging regardless of your specialty, whether neurology, anesthesiology or physiatry. We just don't have the kind of data to make a stronger statement.
DID THE STUDY FIND THE THERAPY IS USED TOO OFTEN IN THE WRONG PATIENTS?
While it may be that these procedures are overdone, one take-home message from the review is that we have much to learn about who is a properly selected patient.
WHAT LIMITATIONS DID THE REVIEW FIND FOR PATIENTS WHO BENEFITED FROM THE TREATMENT?
In general, epidural steroidal injection for radicular pain does not affect the average impairment of function or the need for surgery, and it does not provide long-term relief beyond three months. The routine use for these indications is not recommended, according to the evidence.
HOW ABOUT TIMING? WHEN DO THE INJECTIONS BECOME EFFECTIVE, AND HOW OFTEN DO YOU ADMINISTER THERAPY?
The review found the therapy was not effective for the first 24 hours. And doctors do typically advise patients there won't be any effect for 24 or more hours. It found some efficacy at two- to six-weeks, but either no difference or rebound pain at three months and six months. There was no difference in treatment groups at one year.
Many have advocated doing as many as three injections in six months. We still need additional data to determine an optimal number of treatments. As for when to stop treatment, despite the findings of the review, there are more than a few patients who have found it beneficial to have these injections three or four times a year.
When someone is in agony, with occasional exacerbations — and they need to be taken down several notches in order to start functioning and moving while the exacerbation is resolving — then you can argue there may be a role for this therapy among radicular lumbosacral pain patients. There's also no guarantee that surgery or any other treatment will provide that person with long-term benefit; this reality emphasizes how complicated it is to take care of these conditions.
Perhaps this review points out that if someone is experiencing radicular lumbosacral pain and needs to get through a difficult acute or subacute period, injections may benefit such patients.
You cannot view this treatment as the only treatment a person might require. But nevertheless, who among us wouldn't want that temporary, dramatic benefit that some patients experience with this therapy?
OF TECHNIQUES FOR ADMINISTERING THE INJECTIONS – INCLUDING TRANSFORAMINAL AND TRANSLAMINAR TECHNIQUES, WITH OR WITHOUT FLUOROSCOPIC GUIDANCE – DID THE REVIEW FIND ONE MORE EFFECTIVE THAN ANOTHER?
The review found there was insufficient evidence to rate one technique better than another.
BASED ON THIS ASSESSMENT, HOW SHOULD NEUROLOGISTS VIEW THIS THERAPY?
Perhaps we had unrealistic expectations of this treatment. Perhaps what these data provide is a more realistic view of these injections in the management of radicular lumbosacral pain. They are a temporary aid designed to facilitate enough recovery so that the patient could then participate in a more comprehensive treatment program.
What this paper highlights is that this modality cannot maximize pain relief on its own. We may need to explore its potential as part of a comprehensive pain management strategy in combination with other approaches.
ARE THERE ANY CONTRAINDICATIONS OR SIDE EFFECTS, EVEN AMONG RADICULAR LUMBOSACRAL PAIN PATIENTS?
The most common complication is a transient headache, whether or not associated with identifiable dural puncture. More serious complications included several cases of aseptic meningitis, arachnoiditis, and conus medullaris syndrome, typically after multiple subarachnoid injections. No neurological complications were reported among a series of 1,035 patients who received epidural steroid injections while on antiplatelet therapy.
SHOULD THIS GUIDELINE BE COMPARED TO COCHRANE REVIEW?
Strictly speaking, the guideline is not completely comparable to a Cochrane review. In fact, there is no Cochrane review of epidural steroid injections. A Cochrane review considers only Class I evidence (according to the Cochrane classification system), whereas this review was willing to consider Class II-III evidence if it arose out of adequately designed studies. [See “Classification of Evidence.”]
However, we are somewhat similar to a Cochrane review, in that we looked for well-designed studies with placebo controls, masking of assessor from outcome, and randomization to treatment arms, hoping to find Class I evidence.
WHAT'S THE NEXT STEP?
The chief recommendation is to do additional studies that may provide data that are easier to interpret, in the context of real-world experience.
APPLYING THE GUIDELINES: A CASE HISTORY
Charles Argoff, MD, assistant professor of neurology at New York University School of Medicine, shares here a typical case history exemplifying how a physician might apply the recommendations to a classic patient.
A 54-year-old man is playing tennis when he notes the sudden onset of pain in his lower back and right leg after a twisting injury. The pain radiates from the back to the right buttock, thigh, calf, and foot. He feels numbness in the foot immediately after the injury.
After icing his back and taking ibuprofen that evening, he visits his internist the next morning with unchanged symptoms. He can barely walk and he says the severity of pain is 9- on a 10-point scale. He has had no new bowel or bladder symptoms. He is given home exercises to do as well as a prescription for diclofenac (75mg po twice daily) and metaxolone (800mg po four times daily). He is a high school mathematics teacher and he cannot work. A neurologist notes a slightly diminished right ankle jerk, but he is otherwise neurologically intact. Straight-leg raising is limited at 45 degrees on the right. MRI of the lumbosacral spine demonstrates a large right-sided herniated disc at L5-S1 with right S1 nerve root compression. He has no other significant disc disease.
The patient is referred for physical therapy and his medication regimen is changed to include oxycodone/APAP (5/325) four times daily as needed. After three additional weeks, he still has pain (7- on a 10-point scale) and he has not returned to work.
As the guidelines suggest, this patient might benefit from this therapy. He has radicular lumbosacral pain. Other interventions have not greatly reduced his pain. And he illustrates the typical response, as noted in the guidelines, in that pain relief was not immediate; in fact, the pain followed the pattern of best relief in the second to sixth week interval.
In this case, after lumbar epidural steroid injections, the patient's pain level dropped to 2 of 10 within three days. He continues on the medical and physical therapy regimen, needs less oxycodone/APAP, and after two additional weeks he returns to work complaining of 1/10 pain. He resumes playing tennis several weeks later. He sees his neurologist several months later for follow-up and he is feeling almost normal.