This Viewpoint lists some recommendations to improve epidural steroid injections that were developed by a multidisciplinary working group.

Epidural steroid injections are used to treat patients with low back or neck pain with a radicular component. Approximately 4800 epidural injections per 100 000 Medicare patients were performed in 2011, a total of 2.3 million procedures among Medicare patients alone.1 This represents a substantial increase over the previous rate of 2100 per 100 000 Medicare beneficiaries in 2000.

The 2 approaches used to access the epidural space are (1) the interlaminar approach, in which the tip of the needle is placed in the posterior epidural space similar to epidural catheter placement in surgery and obstetrics, and (2) the transforaminal approach, in which the tip of the needle is placed in an intervertebral foramina where the spinal nerve exits the spinal canal. With the interlaminar approach, most of the injected drug remains in the posterior epidural space, whereas with the transforaminal approach, the drug is placed in close proximity to the (inflamed) spinal nerve and dorsal root ganglion and spreads into the lateral and anterior epidural space, the interspace between the spinal nerve and the herniated disk. Practitioners usually use the transforaminal approach when a single nerve root in one extremity is affected from a single lateral herniated disk and use the interlaminar approach when several spinal nerves are involved in one leg or in both legs, as in the case of central disk herniation.

Most physicians who perform epidural steroid injections use a particulate steroid (those commonly used include methylprednisolone acetate, triamcinolone acetonide, or betamethasone sodium phosphate/betamethasone sodium acetate) instead of a nonparticulate (dexamethasone sodium phosphate) because early studies suggested that the duration of pain relief was longer with the particulates. Recent studies, however, suggest the duration of relief to be comparable.2

Published reports seem to demonstrate the safety of epidural steroid injections, with only mild and transient adverse effects, in large clinical trials. However, rare occurrences of catastrophic central nervous system injuries, including paraplegia, quadriplegia, medullary infarct, cerebellar infarct, and death after epidural steroid injections, have been reported as isolated case reports. Epidural injections in the cervical epidural space, especially when performed while the patient is under sedation and without appropriate precautionary steps, such as injection of radiographic contrast medium under fluoroscopy, have resulted in spinal cord injury.

Transforaminal injections of particulate steroid, such as methylprednisolone, triamcinolone, or betamethasone, have resulted in cerebrovascular occlusion, likely from intravascular injection of the particulate steroid through 1 of the arteries (ascending or deep cervical artery or the radicular artery accompanying the spinal nerve) that communicates with the anterior spinal artery, resulting in a segmental cord infarct.3,4 Injuries from nonparticulate steroids, such as dexamethasone, have been associated with temporary events such as blindness and lower extremity paralysis.5,6 In addition, fungal meningitis has occurred from the injection of contaminated compounded methylprednisolone acetate.7

A multidisciplinary working group, consisting of specialists who had previously published research related to epidural steroid injections, discussed the adverse effects posed by the procedures and recommended safety improvements. Although the US Food and Drug Administration (FDA) Safe Use Initiative coordinated the working group’s deliberations, it neither created nor influenced the final recommendations.

The recommendations of the working group were voted on by representatives of an initial list of several national organizations. The group’s proposals were later discussed, revised, and voted on by representatives and boards of directors of an expanded group of national specialty organizations and medical societies, which included the following disciplines: anesthesiology, pain medicine, physical medicine and rehabilitation, neurosurgery, orthopedic surgery, and radiology. The recently published recommendations8 included several important suggestions for improving the safety of epidural steroid injections (explanations, if needed, are in parentheses).

1. All cervical and lumbar interlaminar epidural steroid injections should be performed using image guidance, with appropriate anteroposterior, lateral, or contralateral oblique views and a test dose of contrast medium. (There has been a case report of lower extremity paralysis after lumbar interlaminar injection without fluoroscopy and a case report of paraplegia after thoracic interlaminar injection when fluoroscopy was used but contrast was not injected.)

2. Cervical and lumbar transforaminal epidural steroid injections should be performed by injecting contrast medium under real-time fluoroscopy or digital subtraction imaging, before injecting any substance that may be hazardous to the patient. (The use of digital subtraction imaging has been shown to be more effective in detecting intravascular injection than syringe aspiration alone.)

3. Cervical interlaminar epidural steroid injections are recommended to be performed at C7-T1, but preferably not higher than the C6-7 level. (The cervical epidural space is widest at the C6-T1 levels. Gaps in the ligamentum flavum are more frequent with ascending cervical levels.)

4. No cervical interlaminar epidural steroid injection should be undertaken, at any segmental level, without preprocedural review of prior imaging studies demonstrating sufficient epidural spatial dimensions for needle placement at the target level.

5. Particulate steroids should not be used in therapeutic cervical transforaminal injections. (Injuries following nonparticulate injections were temporary, whereas paraplegias after particulate steroids were permanent. If the nerve root involved is at a higher level, ie, C5, most pain medicine physicians perform an interlaminar injection at C6-7 or C7-T1, insert a catheter, and advance it to C5. For diagnostic injections, to help the surgeon identify the affected nerve root, pain physicians perform transforaminal injections using local anesthetic, with or without a nonparticulate dexamethasone.)

6. A nonparticulate steroid (eg, dexamethasone)6 should be used for the initial injection in lumbar transforaminal epidural injections.

7. There are situations in which particulate steroids could be used in the performance of lumbar transforaminal epidural steroid injections. (This is because the lumbar transforaminal area is wider than in the cervical regions. If relief from a nonparticulate steroid is of short duration, some physicians will inject a steroid containing smaller particles, either betamethasone or triamcinolone.)

The risks associated with epidural steroid injections reported to the FDA or published in the medical literature were discussed at a meeting of the FDA’s Anesthetic and Analgesic Drug Products Advisory Committee on November 24 to 25, 2014. The advisory committee members voted 15 to 7 (with 1 abstention) that they could envision a situation in which epidural steroid injections should warrant a contraindication warning in the labeling of injectable corticosteroids.9 Most of the 15 committee members described the use of particulate steroids injected by the transforaminal route into the cervical spine as the likely situation that would require a contraindication. As of March 26, 2015, the FDA has yet to take formal action on the recommendations provided by the advisory committee.

Epidural steroid injections will continue to be used for short-term relief of radicular pain. More research is needed to determine whether nonparticulate steroids have equal efficacy compared with particulate steroid formulations and to know the lowest dose that will lead to meaningful reduction in pain while minimizing adverse effects, such as elevations in blood glucose. The recommendations of the working group and the national organizations8 are intended to help reduce or eliminate the occurrence of these rare but devastating neurological injuries associated with epidural steroid injections.

Corresponding Author: Honorio T. Benzon, MD, Department of Anesthesiology, Northwestern University Feinberg School of Medicine, 251 E Huron St, Feinberg Pavilion, Ste 5-704, Chicago, IL 60611 (h-benzon@northwestern.edu).

Published Online: March 30, 2015. doi:10.1001/jama.2015.2912.

Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Benzon reported that he is the deputy editor in chief of Pain Practice, and is a member of the board of directors of the American Society of Regional Anesthesia and Pain Medicine. Dr Huntoon reported that he is the editor in chief of Regional Anesthesia and Pain Medicine, a director of the North American Neuromodulation Society and has been a member of the board of directors of the American Society of Regional Anesthesia and Pain Medicine. Dr Rathmell reported that he is the executive editor of Anesthesiologyand a director of the American Board of Anesthesiology.

Additional Contributions: We thank Salma Lemtouni, MD, MPH, of the FDA Safe Use Initiative for indefatigably coordinating the teleconferences of the working group; representatives of the national organizations; John Whyte, MD, chief of the FDA Safe Use Initiative for his review and revisions on the involvement of the FDA; and Laurie Haughey, Professional Affairs & Stakeholder Engagement of the FDA, for her review and revisions on the involvement of the FDA, providing the latest official statement of FDA and the URL to the official report of the FDA Advisory meeting.