Circulation 36,825 • Volume 19, No. 1 • Spring 2004   Issue PDF

The Safe Use of Epidural Steroid Injections

Stephen E. Abram, MD

Injection of suspensions of corticosteroids into or adjacent to the spinal canal is performed on a regular basis in the United States. Translaminar epidural steroid injections are performed at the lumbar, thoracic, and cervical levels. Caudal injections are done by a “single shot” needle technique as well as by fluoroscopically-guided catheters that are directed toward a particular nerve root. Transforaminal steroid injections (also known as selective nerve root blocks) are performed at nearly every spinal level. Facet joint injections of steroid and local anesthetic are performed routinely in the lumbar and cervical regions. It is not uncommon for injected material to flow into the epidural space, either through rupture of the capsule during injection or via existing capsular tears.

Neuraxial steroid injections are generally considered to be safe by physicians in the U.S. On the other hand, sensationalized reporting of a patient’s claim of arachnoiditis resulting from a steroid epidural nearly resulted in abandonment of epidural steroid injections in Australia. As it turned out, the arachnoiditis had been documented on MRI prior to the procedure.

A thorough review of the literature published in 1996 reported a very low incidence of serious complications.1 Unfortunately, a literature search may not be the best method to determine the true incidence of complications. A somewhat more realistic survey of complications can be found by reviewing the report on the ASA Closed Claims Project published by Fitzgibbon et al., who found that 42% of all claims associated with chronic pain treatment were for complications related to epidural steroid injections (N=114) or facet injections (N=4).2

While closed claims analysis helps to identify the types of complications associated with neuraxial steroid injections, it fails to provide a reasonable estimate of the incidence of those complications. We do not know how many and what type of complications occurred with no malpractice claim filed, nor do we know the denominator, i.e., the total number of procedures performed. Another confounding factor is the delay in processing malpractice claims. There is often a delay of months to years in filing a claim, and it may take several years before a claim is litigated or settled. Therefore, we are unaware of most of the complications that have occurred in the past several years, during which time practice may have changed appreciably. For instance, it is likely that a higher proportion of epidural injections are now being done under fluoroscopic guidance, and more patients are treated with a transforaminal approach.

I will briefly review the types of complications that have been reported following neuraxial steroid injections and discuss methods for minimizing risk. I have reviewed published literature, including that obtained from a Medline search subsequent to our 1996 complications review, information from the Closed Claims Analysis, as well as several cases I personally reviewed as an expert witness that have since been settled.

Nerve Injury

Direct Spinal Cord Injury

I have reviewed 2 cases in which spinal cord injury occurred following injection. One case involved a cervical epidural injection performed under fluoroscopy in a deeply sedated patient. The patient suffered a cardiac arrest and was resuscitated, but had severe brain and spinal cord injury and expired following removal of life support. MRI showed injury to the brainstem and cervical spinal cord. The second case involved a lumbar epidural steroid injection done without fluoroscopy. The patient was deeply sedated because she was “allergic” to local anesthetics. Following the procedure she had severe motor and sensory loss in one leg. MRI showed a lesion in the conus.

Hodges et al.3 reported two cases of nerve injury following cervical epidural steroid injections, both performed in heavily sedated patients using fluoroscopy. In both cases, dural puncture occurred and the needles were repositioned prior to injecting steroids. One patient experienced new persistent painful paresthesias in the upper extremity. The other had new persistent painful paresthesias in one arm and one leg. In both patients, injury to the cord was evident on MRI.

Brouwers et al.4 reported a spinal cord infarction following a C-6 nerve root injection with iotrolan, bupivacaine, and triamcinolone hexacetonide. The cord injury was documented on MRI scan and resulted in the patient’s death. Baker et al.5 cited this case and 6 other cases of cord injury after transforaminal injections that could not be reported because of pending litigation. They suggested that the mechanism of injury was likely the injection of steroid suspension into a radicular artery with embolization of the spinal cord. They documented visualization of a spinal radicular artery following injection of radiographic dye during a C6-7 transforaminal injection.


I could find only 2 reported cases of neurologic injury associated with hematomas resulting from epidural steroids. One was a report of a subdural hematoma following a cervical epidural steroid injection. The patient had been taking Fiorinal®, which was stopped 2 weeks before the procedure. She developed quadriplegia, recovered partially following surgical decompression, but developed meningitis and died.6 The other case also occurred after cervical epidural steroid injection. The patient had received 6 previous epidurals over a 2-year period. The patient became quadriplegic, but eventually recovered following extensive decompression surgery.7 I reviewed the case of a patient who developed upper extremity weakness associated with an epidural hematoma following a cervical epidural steroid injection. Surgical decompression relieved his symptoms, but he developed a recurrent hematoma and experienced permanent upper extremity weakness despite a second operation. He had been on no anticoagulant or antiplatelet drugs. I also reviewed the case of a patient who developed motor and sensory loss after a lumbar epidural steroid injection. He had undergone myelography the day before, which was interpreted as disc herniation. At surgery, he was found to have an epidural angiofibroma and a significant subarachnoid hemorrhage. His recovery was complete. This case occurred prior to the routine use of MRI.

Horlocker et al.8 prospectively assessed 1035 patients undergoing a total of 1214 epidural steroid injections for the development of neurologic dysfunction associated with hematoma formation. Blocks were performed at the cervical level in 107 procedures, thoracic in 15, lumbar in 988, and caudal in 104. A history of bleeding or bruising was elicited in 176 patients, and 383 patients were taking NSAIDS, with aspirin being the most prevalent. None were taking clopidogrel or ticlopidine. No patients experienced neurological dysfunction requiring assessment for a hematoma.

The Closed Claims Project reported a total of 14 claims of spinal cord injury following epidural steroid injection and 1 following cervical facet injection. The report did not specify the mechanism of spinal cord injury in most cases, although it was stated that 2 cases involved hematomas in patients receiving anticoagulants. An additional 14 patients had non-spinal cord nerve injury, but the exact nature of these was not reported.


In our 1995 literature review, we found only 2 cases of epidural abscess reported following epidural steroid injection, both in diabetic patients.1 One patient recovered uneventfully, the other died. Knight et al.9 subsequently reported an additional 6 cases, most of which were in diabetics. We found reports of 2 cases of meningitis after epidural steroids. In 1 case a dural puncture was documented, and in the other case dural puncture could not be ruled out. The Closed Claims Project reported an additional 12 cases of meningitis and 7 cases of epidural abscess as well as 2 cases of osteomyelitis. No details were reported for any of these cases.

Inflammatory Complications

A few case reports of adhesive arachnoiditis were reported following multiple intrathecal injections of methylprednisolone acetate in patients with multiple sclerosis.1 This led to warnings in the neurology literature about the potential hazards of epidural steroid injections despite the fact that no cases have been reported after epidural injections. Indeed, I have found no reports of arachnoiditis after only 1 intrathecal steroid injection. There have been several reports of aseptic meningitis after intrathecal deposteroid injections. Symptoms include fever, nausea and vomiting, lower extremity pain, and, in one report, seizures. CSF examination shows elevated leucocytes and protein and decreased glucose. Cultures are negative. To my knowledge, no cases have gone on to permanent neurologic dysfunction or increased pain. There are no reported cases after epidural steroid injections.

Other Complications and Side Effects

Systemic steroid-induced side effects, including fluid retention, hypertension, congestive heart failure, facial edema, buffalo hump, supraclavicular fat pads, easy bruising, and scaly skin can occur after depo steroid injections. In most cases these changes are dose-related, occurring mainly in patients who receive multiple injections. Cushingoid symptoms can be long-lasting, even when injections are discontinued.10 Hyperglycemia is commonly seen for several days after the procedure in diabetic patients. Exacerbation of radicular symptoms is common but rarely prolonged. Exacerbation of epidural lipomatosis, severe enough to require surgical decompression, was reported following a series of 3 epidural steroid injections.11

Nine cases of death or brain damage were reported in the Closed Claims Project.2 Five were the result of unintended intrathecal local anesthetic injection, while 3 involved delayed respiratory depression from the addition of morphine to the epidural. A severe allergic reaction accounted for the other case.

Safety Recommendations

While risks are probably small, catastrophic complications can occur after steroid epidurals. Following are suggestions that may reduce risks to patients and may help protect physicians from negligence claims:

  1. Provide detailed informed consent. Inform diabetic patients about increased risk of infection as well as the probability of hyperglycemia. Discuss post-procedure diabetes care with the patient and with the primary care physician if pre-procedure management is difficult.
  2. Take a careful history; ask about the use of antiplatelet drugs and anticoagulants. Include lay terms such as “blood thinners” and “heart medications.” Avoid epidurals in patients on newer antiplatelet drugs such as clopidogrel, ticlopidine, and low molecular weight heparin. Ask about recent bacterial infections. Look for contraindications to corticosteroids.
  3. Perform a physical examination. Document preexisting neurologic abnormalities. Look for skin bruising.
  4. Do not perform the procedure for improper indications. The procedure is most effective for patients with well-documented radiculopathy. It is generally ineffective for axial low back or neck pain.
  5. Consider the use of fluoroscopy. Check for allergy to contrast materials. Inject radiographic dye “live” to rule out intravascular, and particularly, intra-arterial injection. This should definitely be done for transforaminal injections.
  6. Rule out intrathecal needle placement with a local anesthetic test dose. Abandon the procedure if dural puncture is evident. Do not attempt the procedure at another level at that time. Allow time for the dural puncture to heal before reattempting.
  7. Minimize the amount of steroid used. There is probably no reason to use more than 80 mg methylprednisolone acetate or its equivalent for epidural injection. Lower doses are appropriate for transforaminal injections. Wait at least 2 weeks before considering a repeat injection of steroids at any site. Do not routinely perform a “series” of injections, but tailor therapy to the patient’s response. A single injection may be adequate.
  8. Limit the total local anesthetic to an amount that is safe if delivered intrathecally. Provide close monitoring initially, which should be continued until total recovery if an intrathecal injection occurs. Use a short-acting local anesthetic, especially in outpatients.
  9. Avoid the addition of epidural opioids, especially morphine.

The performance of epidural and transforaminal steroid injections is part of the practice of medicine, and should only be performed by those who are actively (not necessarily exclusively) involved in the practice of pain medicine. Patients who are candidates for these injections deserve careful assessment and attention to technical detail during the procedure.

Dr. Abram is a Professor in the Department of Anesthesiology at the Medical College of Wisconsin.


  1. Abram SE, O’Connor TC. Complications associated with epidural steroid injections: a review. Reg Anesth 1996;21:149-62.
  2. Fitzgibbon DR, Posner KL, Caplan RA, et al. Chronic pain management: American Society of Anesthesiologists Closed Claims Project. Anesthesiology 2004;100:98-105.
  3. Hodges SD, Castleberg RL, Miller T, et al. Cervical epidural steroid injection with intrinsic spinal cord damage: two case reports. Spine 1998;23:2137-42.
  4. Brouwers PJ, Kottink EJ, Simon MA, Prevo RL. A cervical anterior spinal artery syndrome after diagnostic blockade of the right C6-nerve root. Pain 2001;91:397-9.
  5. Baker R, Dreyfuss P, Mercer S, Bogduk N. Cervical transforaminal injection of corticosteroids into a radicular artery: a possible mechanism for spinal cord injury. Pain 2003;103:211-5.
  6. Reitman CA, Watters W. Subdural hematoma after cervical epidural steroid injection. Spine 2002;27:E174-6.
  7. Williams KN, Jackowski A, Evans PJD. Epidural hematoma requiring surgical decompression following repeated cervical epidural steroid injections for chronic pain. Pain 1990;42:197-9.
  8. Horlocker TT, Bajwa Z, Ashraf Z, et al. Risk assessment of hemorrhagic complications associated with nonsteroidal anti-inflammatory medications in ambulatory pain clinic patients undergoing epidural steroid injections. Anesth Analg 2002;95:1691-7.
  9. Knight JW, Cordingley JJ, Palazzo MG. Epidural abscess following epidural steroid and local anesthetic injection. Anaesthesia 1997;52:576-8.
  10. Tuel SM, Meythaler JM, Cross LL. Cushing’s syndrome from methylprednisolone. Pain 1990;40:81-4.
  11. Sandberg DI, Lavyne MH. Symptomatic spinal epidural lipomatosis after local epidural corticosteroid injections: case report. Neurosurgery 1999;45:162-5.