Epidural steroids injections (ESI) have been utilized for the treatment of pain due to lumbar disc injuries since the 1950s. They are frequently used in the treatment of radicular pain which is a neuropathic pain syndrome along the sensory distribution of a dermatome of one or more spinal nerves. Typically, the pain is due to nerve root irritation from mechanical compression by an intervertebral herniated disc and resultant inflammation. Additional etiologies leading to mechanical spinal nerve compression include spondylosis, spondylolisthesis, and ligamentum flavum hypertrophy resulting in neuroforaminal stenosis. In persons with lumbosacral radiculopathy, more than half of the patients report interference in their activities of daily living, and a quarter continues to have severe pain that does not respond well to oral pharmacological treatments.
Approximately 14% of patients with lumbosacral radiculopathy will eventually require surgery for severe pain that may or may not be associated with neurological deficit. However, in many cases, radicular pain secondary to intervertebral disc herniation improves with medical and rehabilitative treatment. Steroid injection into the epidural space is used to achieve inflammation reduction, pain relief, and limit the need for medications or surgery.
The 3 routes for epidural steroid delivery include the transforaminal, interlaminar, and caudal approach. Utilization of fluoroscopy or less commonly, computed-tomography (CT) guidance to ensure needle placement in the proper position using contrast flow has become the standard of care. The efficacy of epidural steroid injections in lumbosacral radiculopathy has been widely studied and based on a recent systematic review that included 8 randomized controlled trials there is a strong recommendation based on moderate quality evidence that transforaminal ESIs can be used to reduce pain at 3 months. However, a clear improvement in physical disability and reduction in surgery was not consistently seen in the literature.
Cervical ESIs similarly have been shown to provide effective short-term relief for radicular cervical pain, but long-term outcomes have not been well studied. There is limited high-quality evidence that cervical ESIs can provide benefit in radiculitis secondary to disc herniation and discogenic pain. Most of the literature focuses on interlaminar cervical ESI as opposed to transforaminal cervical ESI. Though rare, the morbidity associated with ESIs can be catastrophic.
Common causes of nerve root compression that are treated with ESIs are an intervertebral herniated disc, spondylosis, spondylolisthesis, and ligamentum flavum hypertrophy which can result in neuroforaminal stenosis.
Cervical pain is the fourth leading cause of disability in the United States. The incidence of cervical radiculopathy is 1 in 1000 adults per year. The rate of cervicothoracic ESIs has increased substantially over the past decade with state-to-state utilization varying from 5 to 40 per 1000 Medicare enrollees. Low back pain ranks fifth among reasons for seeking medical care in the United States at an economic cost of 100 million to 50 billion dollars. It is estimated that 40% of patients presenting with low back pain have associated lumbosacral radicular pain. Estimates of the annual prevalence of low back pain with radicular symptoms described as pain coming from the back traveling below the knee varied from 9.9% to 25%.
Radicular pain is theorized to result from mechanical nerve compression leading to the production of inflammatory molecules and neurochemical mediators. In intervertebral disc herniations, these mediators include phospholipase A2, substance P, vasoactive intestinal peptide and calcitonin gene-related peptide which are released from the nucleus pulposus. These molecules have a role in sensitizing the free nerve endings, nerve root and dorsal root ganglion precipitating neuropathic pain. Their presence may additionally produce changes in ion functioning of sensory neurons that increase pain. Delivery of corticosteroids to the site of inflammation is hypothesized to interrupt this inflammatory cascade.
Also known as subpedicular or supraneural, the safe triangle is formed by the inferior margin of the pedicle (superiorly), a line drawn inferiorly from the anterior margin of the pedicle (anteriorly) and the exiting nerve (as the hypotenuse). The traditional needle target for a transforaminal approach is the epidural space just caudal to the inferior margin of the pedicle, referred to as the 6 o'clock position. Positioning more medially increases the risk of dural puncture, and placement too laterally will not achieve retrograde epidural flow demonstrated with contrast injection and will only perform achieve selective nerve root block. This approach was initially described as a method to theoretically minimize the risk of nerve injury, intrathecal puncture, or vascular injection. Recent literature has questioned the safety of this technique due to the positioning of radiculomedullary arteries in the anterosuperior portion of the foramen and the reported cases of paraplegia associated with this technique.
This approach is also referred to as the modified, safe triangle approach with the tip in the posterior portion of the neural foramen and slightly inferior to the safe triangle needle position. Lateral oblique fluoroscopy view targets the same safe triangle view; however, needle is not advanced into the anterior epidural space but instead the tip remains in the posterior aspect of the foramen on the lateral view. This approach can be useful in cases of severe foraminal stenosis or if the patient has nerve pain during the procedure and anterior epidural space cannot be accessed. A single center prospective study of 50 patients showed less nerve pain with the posterolateral approach compared to the safe triangle approach with no differences in outcome measures.
The triangle is bounded by the superior endplate or the inferior vertebral body (inferiorly), the exiting nerve (hypotenuse), and by the superior articulating process (posteriorly). Initially, this technique was described as a safe and convenient method to perform discectomy and discograms. The target is the inferior third of the neural foramen at the level of the intervertebral disc. If the needle is advanced overly far ventrally during epidural injection, it may be inserted inadvertently into the disc. Avoiding advancing the needle medial to the mid-pedicular line reduces the risk of intradural puncture. Several studies, including a single center prospective study, have found that the Kambin approach reduces the risk of intravascular injections compared to the safe triangle; however, the rate of intradiscal injections is higher.
The interlaminar approach can be preferable to the transforaminal approach in cases of severe neuroforaminal stenosis or in the cervical spine where the risk of intravascular injection is higher and the consequences severe. The patient is typically positioned prone. Sedation is restricted to cases of severe anxiety to allow the patient to alarm the physician in cases of pain or paresthesias which can be a warning of a malpositioned needle. The segmental level is identified in the AP view and after appropriate local anesthesia, an 18- to 22-gauge Tuohy needle is introduced using a paramedian approach targeting the superior lamina of the lower vertebra in the segment. A loss of resistance syringe containing air, saline or contrast is used while advancing the needle tip millimeters at a time until the loss of resistance is felt signaling penetration through the ligamentum flavum into the epidural space. A lateral view then confirms the depth of insertion.
The obvious risk in this approach in the advancement of the needle too far anterior puncturing the dura or damaging the cord directly. This may also introduce the possible complication of an epidural hematoma or abscess, which may compress the cord. Avoiding sedation when possible allows for the patient to alarm the physician of pain or paresthesias from needle advancement into the cord. The epidural space is narrow in the cervical spine compared to the lumbar spine and continues to narrow significantly in the "cervicogenic bulge." Therefore interlaminar injections are preferentially performed below C6 which can be sufficient for pathology above the level due to the cephalo-caudal spread of medication of about 2 levels.
The primary role of ESIs is to treat cervical or lumbosacral radicular pain, and in certain cases, bypass the need for operative intervention. Previous investigations into the efficacy of ESIs including case series, retrospective studies, prospective clinical trials, and systematic reviews have limitations including heterogeneous methods and broad inclusion criteria. Published reviews often include a variety of ESI delivery routes including transforaminal, interlaminar, and caudal and also include a wide variety of indications in addition to radicular pain, such as axial back pain, failed back syndrome, and spinal stenosis.
Recent studies, including several reviews and meta-analyses, have shown that transforaminal and interlaminar epidural steroid injections can provide reliable pain relief for patients with low back pain associated with radicular symptoms, but not axial back pain. In some cases, ESIs can provide long-term benefits lasting up to 12 months and surgery sparing effects. For radicular pain secondary to disc herniation the evidence is good with steroids combined with local anesthetics and fair with local anesthetics alone. Theoretically, the transforaminal route is superior to interlaminar and caudal approaches in delivering medication in proximity to the exiting spinal nerve, anterior epidural space, and dorsal root ganglion. However, comparison of the transforaminal versus interlaminar approach demonstrate similar efficacy at 6 months but an early benefit at 2 weeks for transforaminal. Other studies have shown that interlaminar ESI for primarily axial pain regardless of etiology lacks effect, except for possibly short-term pain relief. Literature reviews investigating the dose of steroid in ESIs do not show a dose-dependence in analgesic effect in chronic low back pain or lumbosacral radicular pain.
The studies for cervical ESIs have similarly heterogeneous methods and outcome measures. There is more evidence for interlaminar cervical ESIs than for transforaminal. The evidence demonstrates durable pain relief and improved disability measures at 12 to 24 months. Several of these studies required multiple injections, and there was similar efficacy between the epidural steroid injection with local anesthetic versus local anesthetic alone. The indications included radicular pain, disk herniation, cervical spinal stenosis, and cervical post-surgery syndrome.
Major procedure-related complications and drug-related systemic effects of ESI requiring hospitalization including spine infection, hematoma, and sepsis are very rare. This was estimated to be 0.46% in a large single-center European study. Though the incidence of permanent neurological complications from epidural steroid injections is rare, great care is taken to avoid critical vascular and neurological structures in the vicinity of the needle during injection. Precautions to avoid an intra-arterial injection into the radicular arteries supplying the spinal cord include aspiration before injection, visualization of contrast flow, and digital subtraction angiography, and an anesthetic test dose.
Particulate steroids have been implicated in multiple case reports of permanent neurological damage after cervical transforaminal ESI. Intra-articular injection of steroid into a radiculomedullary artery supplying the spinal cord can cause occlusion of the anterior spinal artery and subsequent cord infarction and injection into the vertebral arteries have been known to cause posterior circulation stroke. Ultrasound has been used to demonstrate the incidence of perineural blood vessels in the projected pathway for selective nerve roots blocks of C5 to C7, ranging from 5.5% to 13.5%. When members of the American Pain Society were given a survey, 287/1340 physicians responded and reported 78 complications of transforaminal cervical ESI which included 16 vertebrobasilar brain infarcts, 12 cervical spinal cord infarcts, and 2 combined brain/spinal cord infarction, and 13 patients died. There are at least 18 cases of paralysis in the literature following TFESI in the lumbar spine from T12-SI despite the use of CT and fluoroscopic guidance.
Different approaches have been described to safely deliver medication to the target area with avoidance of the neurovascular structures. There is controversy on the safest approach to minimize the chance of neurological injury. In a review of 18 cases of paralysis following lumbar transforaminal ESI where needle position could be determined, the needle was in the superior portion of the neural foramen in 77.7% of cases, in the mid-zone in 22.2%, and in no case was the needle in the inferior portion of the neural foramen. The needle was more commonly anterior (71.4%) and less commonly posterior (28.5%). Confirmation of the needle tip is provided by contrast injection that will produce a epidurogram. If the needle tip is not placed far enough within the foramen then epidural flow may not be obtained. In this case, only the nerve is outlined and referred to as a selective nerve block.
In the spinal column, the neural foramen is bounded anteriorly by the intervertebral disc, the superior and the inferior endplates of the vertebral bodies; posteriorly by the superior and inferior articular processes, IAP, which articulate as the facet joint; the adjacent pedicles form the roof and floor at that level. The dorsal and ventral roots join to form a spinal nerve within the neuroforamen. Each spinal nerve divides into a ventral ramus and smaller dorsal ramus as it exits the foramen.
The neuroforamen also contains arteries, veins, and epidural fat. The spinal nerves are supplied in a segmental manner by radicular arteries which travel with each nerve root. In the thoracic and lumbar spine, radicular arteries arise from branches of the aorta and intercostal arteries. The intercostal and lumbar arteries divide into a dorsal segment supplying the paraspinal muscles, a somatic branch that travels ventral to the spinal cord that feeds the dura, and a radicular artery that supplies the spinal nerve and nerve roots within the neural foramen. The ascending cervical, deep cervical and vertebral arteries give rise to the radicular and spinal medullary arteries which traverse the cervical neuroforamina.
A major concern during transforaminal epidural steroid injections is a careful avoidance of the radicular arteries. Specific needle approaches and protocols are designed to minimize and detect an intravascular injection discussed in the section "Complications/Safety." The largest radiculomedullary artery is the artery of Adamkiewicz, which anastomoses with the anterior spinal artery and supplies the thoracolumbar spinal cord in adults. The origin of this artery is variable but generally takes off at T9, but can arise as low as L5. It is left sided in 68% to 85% of cases and lies in the superior or mid-portion of the neuroforamen in the majority of cases. The vertebral arteries course between the transverse foramina of the C2 to C6 vertebrae typically anterior to the cervical facet joints and the ventral ramus of the cervical nerve root though variations exist.
Choice of Steroid
Previously particulate steroids (triamcinolone, methylprednisolone, and betamethasone acetate) have been favored by some providers due to their theorized "depot effect," or increasing the duration the medication is deposited at the site of pathology. The same particulate effect that theoretically provides an analgesic benefit poses a safety concern as the particulates can aggregate to form particles larger than the size of a red blood cell (RBC). This is in contrast to non-particulate steroids like dexamethasone which are 10-fold smaller than an RBC and do not aggregate under light microscopy.
Animal studies show that intravascular injection into the vertebral artery of particulate steroids cause neurological injury, whereas non-particulate steroids do not. Furthermore, the theoretical analgesic advantage of particulate steroids has not been established in the literature with regards to pain reduction or functional outcomes. Perhaps for this reason, non-particulate steroids are increasingly favored in transforaminal ESIs.
Improving the outcomes of epidural steroid injections in the treatment of acute or chronic pain requires careful patient selection. ESIs have been shown to be efficacious in the treatment of radicular pain and not axial pain. Physicians have an ethical responsibility to utilize ESIs in cases of radicular pain that have a history and physical exam findings concordant with MRI evidence of nerve root pathology at the level to be treated. The evidence is lacking for ESIs in the treatment of axial back pain. Most patient with back pain, including those with disc herniation with radicular pain, will improve with time or conservative treatments including oral analgesics, various treatment modalities, and physical therapy. Thus, it is the role of the provider to place ESIs in the context of the individual's disease and form a patient-centered treatment plan.