Steroid Injections: Can They Actually Make Things Worse?

Steriod injections have long been the go-to treatment for common musculoskeletal ailments such as low back pain and tennis elbow. Here, the conventional wisdom around these injections gets a closer look.

I recently wrote here about epidural steroid injections (ESI) and the dearth of research to support this modality in the treatment of low back pain. Since then, 2 studies^1,2 have been published that indicate that ESI may not only be of limited efficacy but actually might also impede improvement.

ESI for lumbar spinal stenosis
The first study involved ESI for patients with lumbar spinal stenosis (LSS).¹ The multisite study compared 69 patients with LSS who underwent ESI with 207 who also had LSS but did not receive ESI. All patients had symptoms considered secondary to LSS consisting of either neurogenic claudication or radicular leg pain with associated neurologic signs that had persisted for at least 12 weeks. There were no relevant significant differences between the groups at baseline. The patients were followed for 4 years.

Patients who had received ESI and subsequently underwent surgery for their LSS were found to have an average 26-minute increase in operative time and nearly 1 full additional inpatient day compared with those who had surgery but did not undergo ESI at the outset of the study.¹ Also, it was found that those who had ESI were more likely to undergo surgery than those who had not.

For those patients who did not undergo surgery, those who had ESI had less improvement on scales measuring pain and physical function than the non-ESI group at the end of the 4-year study period.¹

The study authors provide several possible explanations for these results. ESI might actually exacerbate stenosis by introducing additional material (ie, steroid, anesthetic) into an already compromised site or in some other way damage nerve roots. ESI may also provide some temporary relief that might allow patients to participate in activity that might lead to an increase in pain.

The additional surgical and recovery time required by patients who had received ESI, the authors suggest, may be the result of adhesions or scarring caused by the original injection procedure that made the surgery more difficult, which, in turn, necessitated an extended recovery period.¹Steroid injections for lateral epicondylalgia
The second study examined the use of corticosteroid injections for lateral epicondylalgia, more popularly referred to as “tennis elbow.”² A group of 165 patients who had the condition for at least 6 months was divided into 4 experimental groups: steroid or placebo injection alone or one of these injections plus physical therapy. The physical therapy consisted of 8 weekly sessions of local elbow manipulation and exercise and home exercises the participants were to perform daily.²

At 4-week follow-up, the groups receiving the steroid injections were more likely to report complete recovery or much improvement compared with those who received the placebo injection.²

At the 26-week and 1-year follow-ups, however, it was found that those who underwent the steroid injection were less likely to have complete recovery or much improvement and more likely to have recurrence of the pain compared with those who underwent the placebo injection. Whether or not the patients underwent physical therapy had little effect on the outcomes at 26 weeks and 1 year.² There was at least one apparent benefit to physical therapy: patients who received it were less likely to use analgesic medications, including nonsteroidal anti-inflammatory drugs.

This study indicates that while steroid injections may provide some short-term benefit for tennis elbow, in the long run they not only have limited benefit but may actually have a negative effect. Also, the results fail to support the conventional view that steroid injections make benefits from physical therapy more likely because the injections make it easier (less painful) for patients to participate.

The failure of steroid injections to provide long-term relief indicates that the pain in tennis elbow is not primarily related to an inflammatory process. Why did patients who received these injections not fare as well as those who did not? The study authors conjecture that the steroid might actually exacerbate the underlying condition or that the short-term relief received from the injections might result in excessive use of the affected arm before there is any actual improvement in this condition.

Although physical therapy did not affect the outcome measures used by the study, the fact that it did reduce the use of analgesic medications suggests that it did actually have an impact.

Conventional wisdom?
Both of these studies call into question the still widespread belief that steroid injections are a first-line treatment for the management of low back pain and other pain conditions presumed to be caused by inflammation even where there is little evidence that this is the etiology of the pain. The limited efficacy of the injections observed in these 2 studies is not that surprising in light of similar findings in previous studies. What is new is that steroid injections may not only not provide relief but may actually exacerbate certain disorders and the associated pain. 

Obviously findings from a pair studies are not sufficient to make any final statements about any therapy. However, these two in particular should at least make us continue to critically evaluate the appropriateness of using steroid injections. And we certainly need to refrain from suggesting to patients suffering pain that the possible benefit they may gain from steroid injections is well supported by research.

References
1. Radcliff K, Kepler C, Hilibrand A, et al. Epidural steroid injections are associated with less improvement in patients with lumbar spinal stenosis. Spine. 2013;38:279. (Abstract)

2. Coombes BK, Bisset L, Brooks P. Effect of corticosteroid injection, physiotherapy, or both on clinical outcomes in patients with unilateral lateral epicondylalgia. JAMA. 2013;309:461-469. (Abstract)