Introduction

Back pain is the leading cause of disability worldwide, with an estimated financial burden exceeding $100 billion annually in the United States alone. Most cases of back pain seen in the primary care setting are acute and axial, attributable to muscular and ligamentous strain and spasm. But patients who are referred to the pain medicine specialist typically suffer from pain that is chronic and more complex in nature. Chronic spine pain arises from an array of often overlapping etiologies and, in many cases, exhibits a radicular component, with epidemiologic studies estimating approximately 40% to be predominantly neuropathic in nature. Although randomized, controlled studies have demonstrated that epidural steroid injections (ESIs) may alleviate axial spine pain, it is widely acknowledged that it is most effective in the treatment of radicular pain, particularly pain arising from a herniated disc.

Radicular pain, also known by the ambiguous and misleading term sciatica, can be caused by several distinct conditions, but by far its most common causes are herniated disc and spinal stenosis. Disc herniation, or the extrusion of disc material from the nucleus pulposus beyond the limits of the intervertebral disc space, results in the liberation of inflammatory mediators near spinal nerve roots. A combination of chemical and mechanical factors may serve to irritate adjacent exposed nerve root(s), causing radiating pain that is often described as tingling, burning, or electrical in nature. The prevalence of disc disruption in patients with chronic low back pain is high and has been reported to be near 40%. Whereas 95% of all disc herniation occurs in the lumbar spine at L4-5 or L5-S1, the cervical spine can also be affected, most commonly at C6-7.

Spinal stenosis, or the progressive narrowing of the spinal canal, may occur alone or in conjunction with disc herniation. Spinal stenosis can be congenital (i.e., short pedicles) but is most frequently caused by degenerative changes of the spinal anatomy, which may include thickening and buckling of the ligamentum flavum, facet hypertrophy, osteophyte formation, or a combination of these conditions. A narrowed central spinal canal results in compression-induced ischemia of the spinal cord (in the neck) and/or cauda equina. The incidence of spinal stenosis increases with age, with the prevalence among patients aged 60–69 years documented to be as high as 47% for mild to moderate stenosis and 19% for severe stenosis (defined as a spinal canal anteroposterior diameter of less than 12 mm). The most common vertebral levels affected by spinal stenosis are L4-5 in the lumbar spine and C5-6 in the neck.

Physicians have performed therapeutic epidural injections for well over 100 years in an effort to treat a wide scope of conditions. Until the first epidural injection of corticosteroids was performed in 1952, these procedures utilized an injectate of plain local anesthetic (LA) solution. In this case study published by Robecchi and Capra, hydrocortisone was injected through the S1 posterior sacral foramen in an effort to treat lumbar radicular pain. Since that time, the procedure has evolved, and the sacral route, though still utilized, has largely been supplanted in favor of the more targeted and accessible caudal, lumbar, and cervical techniques. Today, ESIs are the most frequently performed procedures by interventional pain physicians worldwide and have been the subject of dozens of reviews and guidelines. Despite its extensive study, the resulting body of data has been largely conflicting in terms of procedural efficacy, polarizing health care professionals on its utility and casting a shadow on the future of the procedure.

Technique

The performance of interlaminar ESI (ILESI) requires directing a needle at the midline or paramedian interlaminar space closest to the site of pathology. The needle must pass through the skin, subcutaneous fat, supraspinous ligament, interspinous ligament, and ligamentum flavum to enter the epidural space. The loss-of-resistance (LOR) technique, where manual pressure is applied to the plunger as the needle is advanced through the spinous ligaments, exploits the palpable difference in resistance between the thick fibrous ligamentum flavum and the vacuous potential epidural space. Typically, a 17- or 18-gauge Tuohy needle is used. Smaller 20-gauge needles have been advocated to optimize patient comfort and decrease the risk of spinal headache, but they have been found to be more technically difficult to use and less accurate than their larger counterparts.

Studies analyzing the accuracy of epidural needle positioning have found that extradural injection occurs in up to 30% of patients in whom blind placement is attempted, even in the hands of experienced proceduralists, compared with very accurate and precise placement using confirmatory fluoroscopy. A multicenter, retrospective analysis of cervical epidurograms by Stojanovic et al. observed a 53% rate of false LOR during the first attempt to enter the epidural space and determined that the use of epidurography can improve the accuracy of needle placement and medication delivery. Whereas blind ESI is still performed in some places, the use of fluoroscopic guidance has been found to provide superior accuracy and safety and has become the standard of care.

Even without the use of real-time fluoroscopic guidance, the inherent margin of safety of lumbar ILESI is relatively high due in part to the diameter of the lumbar epidural space (4–7 mm) and the free mobility of the cauda equina ( Fig. 62.1 ). However, at the cervical level the small diameter of the cervical epidural space (1–4 mm) and other factors increase the risk for catastrophic complications associated with inadvertent dural puncture. The 2015 Multi-society Pain Workgroup guidelines state that cervical ESI should not be attempted above C6-7. Because of this consideration, some physicians employ the “hanging drop” or “infusion drip” technique. In this method, whereby a column of saline is attached to the epidural needle via intravenous tubing, entry into the epidural space is theoretically confirmed when the negative pressure of the potential space causes the meniscus of the hanging fluid to drop into the tubing.

Lateral fluoroscopic view demonstrating contrast spread during an L5-S1 interlaminar epidural steroid injection.

Despite the notional advantages of this technique, studies have not borne out any evidence of increased safety as compared with LOR. This finding may stem from the fact that the epidural pressure on which this technique depends is not intrinsically negative. A study comparing 30 cervical epidural pressures found them to be highly dependent upon patient positioning, with the vast majority being positive in both the sitting and prone positions.

In the neck, the ligamentum flavum is incompletely fused at the midline, with an incidence of gaps between C3 and T2 of 83%–100%. The absence of this fibrous harbinger of the epidural space has been theorized to account for incidents of inadvertent subarachnoid puncture even with proper use of the LOR technique. A prospective, randomized comparative study by Joo and colleagues compared the pressure changes throughout the pathway to the cervical epidural space between the midline and paramedian approaches. The investigators found that an abrupt pressure decrease at the moment of exiting the ligamentum flavum was more frequently observed when the paramedian approach was being used and concluded that this is a superior technique when employed for cervical ESI ( Fig. 62.2 ).

Lateral fluoroscopic view demonstrating contrast spread during a C5-6 cervical epidural steroid injection. For safety concerns and because of the increased spread in the cervical epidural space compared with the lumbar region, higher levels (e.g., C6-7 and C7-T1) should generally be targeted.

The caudal technique involves introducing a needle at the sacral hiatus via the sacrococcygeal ligament into the epidural space ( Fig. 62.3 ). The hiatus may be located by palpation or lateral fluoroscopic imaging, and the needle is advanced until it is felt to penetrate the membrane that encloses the hiatus with an abrupt LOR. Advantages attributed to this approach are its ease of use, particularly in cases of altered anatomy, and a decreased risk of dural puncture and consequent intrathecal injection.

Lateral fluoroscopic view demonstrating caudal epidural steroid injection. The caudal approach is often employed when an epidural steroid injection is performed after low back surgery to reduce the risk of dural puncture and in performing epidural lysis of adhesions.

Because of the distance from the site of injection to the site of lumbar pathology, caudal injections require a substantially larger volume of fluid to be injected as compared with lumbar ESI. Whereas historical volume recommendations vary from 10 to 64 mL, a study by Kim and colleagues used fluoroscopy to determine that 10 mL of injectate was sufficient to reach the L3-4 interspace. The investigators observed that after the initial injection of 10 mL, serial injections of up to 40 mL achieved additional spread of only one level cephalad. Despite the purported ease of the caudal technique, studies evaluating the accuracy of landmark-guided injections have demonstrated failure of the injectate to reach the epidural space in up to 35% of cases, signifying the need for fluoroscopic guidance with this technique as well. Each of the three ESI methods is associated with unique advantages and disadvantages that must be understood and tailored to the individual patient ( Table 62.1 ).

Injectate Composition

The most commonly employed and best-studied steroids used in ESI are methylprednisolone and triamcinolone diacetate. The therapeutic ESI dose range for both medications is typically 40–80 mg, although three randomized trials comparing 40 mg with 80 mg all found no difference in effectiveness. No large-scale studies have been performed comparing triamcinolone and methylprednisolone, although one small, nonrandomized study found methylprednisolone to be more effective. An area of intense interest, especially for transforaminal (TF) ESI, centers around the use of nonparticulate steroids. Although some studies have found comparable benefit between depo-steroids and soluble dexamethasone, the two largest studies have demonstrated greater effectiveness for depo-steroids. The reduced risk of embolic phenomenon and damage to the blood-brain barrier cited for the use of nondepo-steroids does not apply to interlaminar injections, and the Multi-Society Pain Workgroup guidelines endorse the use of depo-steroids as a first-line treatment for interlaminar ESI.

Although one might surmise that higher injectate volumes may dilute the steroid concentration from epidural injections targeted to the level of pathology and result in inferior outcomes, this is not the case. A systematic review by Rabinovitch et al. found that higher volumes resulted in significantly better outcomes in the short (6 weeks to 3 months) and intermediate (3 months to 1 year) terms than lower volumes, controlling for the dose of steroid. Studies evaluating the spread of lumbar ILESI injectate have found that volumes of 3–5 mL reach a median of two vertebral levels. A study of cervical ESI by Stojanovic et al. found that a 2-mL injectate achieved an average spread of 3.14 vertebral levels. The greater degree of spread observed is likely a result of the smaller volume of the cervical epidural space compared with that of more caudal levels. Goel and colleagues found that the injection of 2–4 mL in the cervical epidural space resulted in an average spread of 3.75 vertebral levels, whereas a study by Lee et al. found that injecting contrast volumes of 2.5, 5, and 10 mL resulted in spread to 8.3, 11, and 13.6 levels, respectively. If a caudal route is selected, larger volumes (approximately 10–20 mL) should be used to ensure adequate spread to the levels of pathology. In a study of caudal epidural flow patterns by Cleary et al., an injectate volume of 20 mL reached a median segmental level of L3 with a range of T9 to L5, leading investigators to conclude that the caudal route should not be recommended for the treatment of disease pathology above L3.

Mechanism(S) of Action

Numerous theories have been proposed to explain the analgesic mechanism of action of ESI. The oldest and best supported among these attributes epidural analgesia to the antiinflammatory properties of corticosteroids. Inflammation and nociception share two key biochemical pathways, the lipoxygenase and cyclooxygenase pathways, which result in the formation of leukotrienes, prostaglandins, and thromboxanes, respectively. These biochemical mediators serve to promote both direct nociception and inflammation-induced pain. The inflammatory enzyme phospholipase A2 converts membrane phospholipids into arachidonic acid, an important substrate in both pathways that is found in high concentrations in intervertebral discs. This finding could explain why conditions causing the extrusion of disc contents result in an inflammatory-mediated chronic pain syndrome.

Exogenous glucocorticoids directly inhibit the lipoxygenase pathway, resulting in diminished downstream formation of leukotrienes. In addition, cells exposed to glucocorticoids release lipomodulin, a phospholipase A2-inhibitory glycoprotein that downregulates the formation of arachidonic acid and consequently its downstream pronociceptive and inflammatory metabolites. Steroid preparations inhibit the inflammatory process at a step antecedent to that blocked by nonsteroidal antiinflammatory drugs (NSAIDs). This has a theoretical advantage in patients with a chemical rather than a mechanical radicular pain syndrome and negative radiologic studies.

In addition to the antiinflammatory theory, several alternate or complementary mechanisms of action have been postulated. Corticosteroids have been found to directly suppress the ectopic discharge of injured neurons. Indirectly, exogenous glucocorticoids have been shown to reduce collagen deposition and its subsequent scar formation. A lower scar tissue burden has the theoretical benefit of reducing mechanical pressure on posttraumatic nerve roots. The addition of LA, which has been utilized in the majority of clinical trials evaluating epidural steroids, may alleviate pain by enhancing blood flow to ischemic nerve roots, which may be particularly important in individuals with neurogenic claudication. In addition, as noted earlier, high volumes of injectate may alleviate pain by washing out inflammatory mediators such as cytokines and chemokines.

Another documented effect of steroid administration is the reduction of capillary wall permeability, leading to a reduction in the formation of edema. On the genomic level, glucocorticoids bind to an intracellular receptor, forming a glucocorticoid-receptor complex that translocates to the cell nucleus and transcriptionally activates (transactivation) or represses (transrepression) genes involved in the inflammatory pathway. This mechanism has been shown to impact the expression of more than 6500 genes.

Efficacy

Despite extensive study of the efficacy of ILESI, the literature has failed to produce a consensus opinion on the subject. Even when studies that are retrospective, unblinded, and/or lacking a control group are excluded, efficacy data remain frustratingly variable ( Table 62.2 ). The lack of unanimity can be at least partially attributed to the enormous challenges investigators face in designing protocols.

Even in higher-quality studies, patient populations are often poorly defined and lack homogeneity. Subjects suffering from acute, chronic, and postsurgical back pain secondary to various etiologies are often pooled into collective cohorts without controlling for their significant differences. Treatment protocols, measurement tools, follow-up observations and outcome criteria all lack standardization, so interpretation of the mixed data often falls to expert opinion.

This too is fraught with uncertainty, as different investigators reviewing the same literature have come to contradictory conclusions. In some cases, this discrepancy can be accredited to bias, such as the fact that those with an interventional background are 3 times more likely to ascribe success to the procedure than those who do not perform ESI. In other cases, efficacy determination is closely tied to the reviewer’s definition of procedural success.

Pain relief, though the primary goal of ESI, is but one measure of efficacy. Other surrogate markers include functional capacity, opioid reduction, return to work, and prevention of surgery. Studies evaluating improvement in functional capacity have met with mixed results, and those finding relative improvement in ESI groups showed that patients with the worst self-reported health-related quality of life had the greatest improvement. Investigators searching for a reduction in opioid intake following ESI have also found conflicting data, as both ESI and control groups have been shown to experience a decline in analgesic intake.

In controlling for etiology, randomized studies of comparative effectiveness have generally found that ESI reduces opioid consumption for within-group but not between-group comparisons in patients suffering from herniated disc, failed back surgery syndrome, discogenic spine pain, and lumbosacral spinal stenosis. Whereas a majority of clinical trials have failed to report a significant difference between return-to-work rates or missed work days when ESI and control groups are compared, some studies suggest that in well-selected patient groups, ESI may improve work status. In a review of 26 randomized controlled trials (RCTs), moderate evidence was found in support of ESI as a surgery-sparing intervention for up to 1 year but not thereafter.

Comparative-effectiveness studies are more clinically relevant than efficacy studies and are a cornerstone of the Affordable Care Act’s research priorities. In a randomized double-blind comparative efficacy study, ESI performed via the IL and TF routes were found to be superior to gabapentin in some outcome measures—such as a reduction in worst leg pain at 1 month—but the differences were modest thereafter. In an open-label, comparative-effectiveness study comparing a series of cervical ILESIs, conservative treatment consisting of first-line adjuvants and physical therapy or a combination of the two approaches in patients with cervical radiculopathy, the combination group fared better than the two stand-alone treatments for up to 6 months. Whereas all randomized trials that occurred before 1999 focused solely on IL procedures, many subsequent studies feature either solely TFESI or a combination of the two.

Wilson-MacDonald and colleagues performed a randomized, controlled trial evaluating 93 patients with radiologic evidence of disc herniation, spinal stenosis, or a combination of the two. The experimental group received ILESI of 40 mg bupivacaine 0.5% with 80 mg methylprednisolone, while the control group received the same injectate as an intramuscular injection. Patients randomized to the ILESI group experienced a statistically significant improvement in nerve root symptoms at up to 35 days, but long-term relief and reduction in need for surgery was not demonstrated at 2-year follow-up.

In another randomized, controlled study by Arden et al. in which 228 patients with unilateral sciatica were randomized to receive a series of three ILESIs consisting of either 80 mg of triamcinolone acetonide and 10 mL of bupivacaine (at weeks 0, 3, and 6) or placebo injections with 2 mL normal saline, patients in the experimental group showed significant improvement in functionality as compared with the placebo group at 3 but not 6 weeks. In studies comparing caudal or conventional ILESI to epidural saline injection, a similar short-term improvement in symptoms has been found. In treating radiculopathy associated with spinal stenosis, results suggest that patient symptom improvement is optimized when ILESI is performed at the intervertebral level of maximal stenosis.

Ilesi vs. Tfesi

Controversy continues regarding the most effective method of ESI. Despite the wealth of literature dedicated to ESI, few randomized trials have compared the two methods head to head. Most studies have been retrospective or technically focused and have yielded conflicting data. Yet the preponderance of evidence gleaned from these studies seems to favor TFESI over ILESI for short-term pain reduction in patients with disc herniation and spinal stenosis.

A recent randomized, double-blind outcome study by Gharibo and colleagues on 42 patients with nonstenotic radicular pain generated data suggesting that patients may experience greater subjective relief, as determined by the postinjection numeric rating scale (NRS), from TFESI but that more objective and subacute therapeutic effects (disability, activity, opioid use) are similar at the 10–16-day follow-up period. In another nonblinded prospective, randomized study of 64 patients with chronic radiculopathy by Rados et al., the investigators compared ILESI and TFESI in patients with chronic unilateral radicular pain. They found that using both routes of epidural injections provided significant improvements in patient function and pain relief, but no statistically significant difference could be found between the two techniques at 24 weeks.

A systematic review evaluating five prospective and three retrospective studies comparing transforaminal and interlaminar approaches found that both ILESI and TFESI were effective in reducing pain. At 2 weeks, data from the combined prospective studies favored the TFESI group by a little more than 10%, with no differences noted at 6 months. For functional improvement, there were no clinically meaningful differences between the different routes of epidural steroid administration. In a large, systematic review and meta-analysis by Chou et al., 38 placebo-controlled studies were analyzed; the authors found that ESIs for radiculopathy were associated with immediate reductions in pain and improvement in function, but the results were small and not sustained. Data suggested no differences between TFESI and ILESI based on limited evidence and no clear benefits for spinal stenosis.

The common use of epidurography in performing ESI has allowed practitioners to become more mindful of the pattern of injectate spread, which is sometimes used as a surrogate outcome measure for improvement. Although the TFESI technique has been touted to deposit injectate more closely to the site of unilateral, single-level pathology (i.e., the ventral epidural space of the affected nerve root), the use of ILESI allows for bilateral spread to multiple spinal levels. Of note, the use of a parasagittal approach has been shown to both increase the incidence of ventral epidural spread and improve treatment outcomes in a randomized double-blind comparative effectiveness trial compared with the traditional, midline interlaminar approach.

A review of the literature reveals that a disproportionate percentage of ESI complications are related to the transforaminal approach, with the greatest risk occurring with cervical TFESI. The TFESI method, especially when particulate steroids are used, carries certain unique risks and has been implicated in permanent, severe complications such as spinal cord infarction, paralysis, and inadvertent disc injection in the presence of lateral disc herniations. Although the incidence of postdural puncture headache is reduced by the TFESI approach, it does not decrease the risk of many other complications compared with ILESI. In fact, the risk of many complications, including catastrophic embolic events, is increased in using the TFESI method; some feel that ILESI should therefore always be used as a first-line treatment.