“For straightforward cases, an MRI won’t help most people with sciatica who are referred for epidural steroid injections,” first author Dr. Steven P. Cohen, of Johns Hopkins School of Medicine, Baltimore, Maryland, told Reuters Health by email.
Considering how common this treatment is, not routinely ordering an MRI beforehand “would save tens of millions,” Dr. Cohen said.
But current radiologic guidelines endorse MRI before epidural steroid injection, he and his colleagues noted online today in an Archives of Internal Medicine paper. They say the rationale is that MRI helps in three respects: ruling out conditions such as tumors, fracture or instability for which injection would not be safe; choosing the safest level and approach based on anatomy; and improving outcomes by targeting the injection to the appropriate site.
And the results of the new study weren’t entirely clear-cut. The researchers do say that a “trend toward superiority” in the MRI group leaves open the possibility that obtaining an MRI might improve short-term outcomes in some patients.
The study included 132 patients with lumbosacral radiculopathy referred for epidural steroid injection. The patients were mostly women (57%) with an average age of 52 years and moderate to severe leg pain and dysfunction. Roughly a third were taking opioids.
Each patient had an MRI, but in one group (n = 65), the treating physician was blinded to the results, while in the other (n = 67), the treating physician used the MRI results to guide the treatment decision and injection approach.
When the doctors could review the scans in advance, five patients (7%) did not receive the steroid injection. Leg and back pain scores were “slightly lower” in this group at one month compared with MRI-blinded patients, but not significantly so.
There were no between-group differences in disability at any time point, and by three months, both groups had roughly similar improvements in leg pain, back pain, and perceived treatment effect.
Overall the numbers of patients with positive outcomes were similar at all time points; at three months, the proportions were 35.4% when doctors didn’t know the MRI results and 40.7% when they did.
None of the five patients for whom the treating physician decided not to perform epidural steroid injection obtained significant benefit from the alternative treatment, the investigators note.
As part of the study, when the treating physician was blinded to the MRI results, an independent physician proposed a treatment plan after reviewing the MRI. When that plan was compared to the actual treatment received, 66% of decisions were discordant.
However, in more than 80% of those cases, the independent physician who had seen the MRI would have given a different type of injection, or given it at a different level. In no case would the independent physician have withheld the injection entirely.
An editorial says, “Although the overall results of this study were largely negative, they suggest a small benefit to using MRI to guide epidural steroid injection planning in patients with lumbar radiculopathy.”
The authors of the editorial — Dr. Janna Friedly from the University of Washington, Seattle, and Dr. Richard A. Deyo of Oregon Health and Science University, Portland — also point out that MRI may have reduced the total number of injections required and may have improved outcomes in a subset of patients.
“Given these potential benefits as well as concerns related to missing important rare contraindications to epidural steroid injection, it seems premature to counsel against guideline recommendations for obtaining MRI prior to consideration of epidural steroid injections,” Dr. Friedly and Dr. Deyo conclude.
In the meantime, they think “cost savings and clinical benefit might result from discouraging epidural injections in conditions for which efficacy is unclear, such as spinal stenosis and low back pain without radiculopathy.”
“More prudent patient selection may help offset the costs of obtaining advanced imaging prior to injection for those in whom a treatment benefit is more likely.…Ultimately, we need better data on the efficacy of epidural steroids for conditions such as spinal stenosis, and on the cost-effectiveness of injections performed with and without MR,” the editorialists say.
Arch Intern Med 2011.