Chapter 65. Epidural Steroid Injections

The administration of corticosteroids into the epidural space to relieve both acute and chronic pain of spinal origin has been utilized for over 40 years. In 1957 Lievre et al reported the first use of epidural hydrocortisone for the relief of back pain and sciatica.1 Since that time numerous investigators have argued both for and against the efficacy of this technique versus alternative treatment modalities in providing significant lasting relief for patients suffering from low back and leg pain. In 1986 Benzon’s review of the literature on lumbar epidural steroid injections concluded that low back pain of mechanical origin, especially accompanied by signs of nerve root irritation, may respond to epidural steroid injection.2 In 1988 Rosen et al, performed a retrospective analysis of the efficacy of epidural steroid injections, studying 40 patients treated for low back pain and sciatica secondary to spinal stenosis or lumbar herniated disc. They concluded that 50% of the patients with radicular symptoms may receive temporary relief with epidural steroid injection (ESI); long-term relief, however, occurs in less then 25% of patients treated.3 In most studies the principal indications for administering epidural steroid injections included symptoms of low back pain or leg pain alone or in combination with varying selection criteria (e.g., previous surgery, duration of symptoms, patient age).

Most studies to date have been open trials of varying duration. There have been several investigations, however, that have attempted to control the variables that may influence the outcome and subsequently yield more accurate results regarding the efficacy of ESI. In a prospective, randomized, double-blind study, Dilke, Burry, and Grahame4 compared patients whose cause and duration of low back pain and radiculopathy were similar, received the same steroid medication via epidural route, and had not been treated previously. They found significantly better results in the patients receiving steroid; 21 of 35 patients (60%) compared with 11 of 36 patients (31%) in the placebo group. At the 3 month follow-up, fewer patients in the steroid group had severe residual pain (1/44 vs 6/38 patients) and a greater number of patients who received placebo injections (14 vs 3 patients) were still not working.

Overall immediate success rates of ESI vary from 25% to 89%,5–7 with long-term results falling to a mean of 80% by 6 months8 and 56% by 2 years.9 Ryan and Taylor found that 77% of their patients experiencing symptoms for less than 2 weeks obtained complete relief from epidural steroid injection, whereas response rates varied from 72%, 60%, and 43% for patients experiencing pain for 4 weeks, 6 weeks, and >6 weeks, respectively.6 White et al found that epidural steroid injections were most effective in patients with nerve root irritation manifesting as radicular pain, dermatomal hypesthesia, weakness of muscle groups innervated by the affected nerve roots, decreased deep tendon reflexes, and diminished straight leg raise. They noted success rates at 6 months for 34% of their patients with acute pain and 12% of their patients with chronic pain, whereas their patients’ overall success rate at 2 years was 1.3%.10 More recently, Koes and Scholten examined 12 randomized clinical trials looking at the efficacy of ESIs for low back pain and sciatica. They found that 50% of the trials reported positive outcomes and the other half reported negative results. Additionally, they found that most of the studies involved had significant design flaws that consequently invalidated the outcomes. Subsequently, they called for additional studies, taking into account methodologic shortcomings and focusing on determining which patients would most likely respond to epidural steroid injections.11 Watts and Silagy performed a formal meta-analysis on the same literature and concluded that there was evidence for efficacy and safety in the use of epidural steroid treatment for sciatica while acknowledging the shortcomings of meta-analysis reviews on small trials.12 Carette et al reported a randomized, placebo-controlled trial of epidural steroids in 58 patients with documented disc herniation and found significant pain relief and improved functioning at longer follow-up intervals. As in previous similar studies, however, significant methodologic flaws, including noncomparable placebo control and inadequate study group size, likely affected outcome.13

There is widespread agreement that a large, randomized, blinded, placebo and active treatment-controlled trial assessing the efficacy of ESI is needed. Additionally, similar studies designed to define appropriate candidates for injection, ideal number of injections, benefit of fluoroscopic assistance versus blind technique, ideal volume and content of injectate, maximum number of injections, and amount of steroid patients should receive in a given period are also warranted (Table 65-1).

The mechanism by which administering corticosteroids into the epidural space provides relief is largely thought to be by reducing inflammation. The ability of phosopholipase A2 to induce membrane injury and produce edema has been well documented.14 Extracts from herniated lumbar discs have been found to contain 20 to 10,000 times more phospholipase activity than those obtained from any other human source.15 Steroids decrease inflammation by inducing the biosynthesis of a phospholipase A2 inhibitor, which prevents prostaglandin generation.16–17 Subsequently, epidural steroids function to inhibit the activity of phospholipase A2 and therefore block the rate-limiting step in the production of prostoglandins and leukotrienes, which, when liberated, have been shown to sensitize small neurons and enhance pain generation.18–19 Additionally, prostaglandins of the E series have been found to cause hyperalgesia.20 Green et al examined intramuscular administration of dexamethasone in patients with pain from lumbar disc disease and found that 80% of patients obtained good relief; pain returned as doses were tapered, but not to original intensity.21 The advantages of ESI over intramuscular administration of steroid include deposition of drug into the affected area, use of a much smaller dose to achieve the desired result with lower risk of side effect, and longer duration of relief.

There has been long-standing disagreement among authors as to how many ESIs should be given in a certain period and what criteria should be used to determine continuing or withdrawing therapy. Some authors recommend performing a series of 1 to 3 injections even if no relief is obtained from the first procedure,22 considering that some patients fail to respond initially but do receive significant lasting relief from subsequent injections. We will typically repeat an ESI even if no relief is obtained from the first injection on this basis. Should the patient fail to obtain relief from two consecutive injections, no further ESIs are offered. Likewise, if a patient experiences complete lasting relief after an initial injection, there is no need for further injections. If relief is partial, a series of up to three ESIs, each performed approximately 2 weeks apart, will be completed. At that time the patient is reassessed, remaining pain and dysfunction documented, and alternative pain management modalities considered. There is no need to give more than three consecutive injections, since additional relief has not been documented with higher numbers of injections.23 We perform a maximum of three ESIs within a 6-month period. Should the patient’s symptoms return after this time, a second series of ESIs is considered. This decision is primarily determined by the degree and duration of relief as well as improvement in function experienced after the first series of injections. Several factors have been found to be associated with a higher response rate to ESI: These include patients who have not undergone previous surgery,24,25 patients with acute versus chronic complaints, diagnosis of nerve root inflammation, younger patients rather then older patients, and location of injection (injectate placed at the level of the affected nerve root).26

Many questions still remain regarding the anatomy of the epidural space such as distribution of fat, the extent of areas devoid of contents, the location of epidural veins, and the presence and location of fibrous structures. By nature, exploration of this area by dissection disrupts and destroys delicate tissue and septae, which disturbs the true anatomic relationships. Likewise other objective approaches to studying epidural anatomy (epiduraloscopy, magnetic resonance imaging) distort the structures by introducing contrast or air into the epidural space, or frankly do not provide the in-depth detail to answer these questions. Hogan examined epidural anatomy using cryomicrotome section with regard to influence of age, vertebral level, and disease. This technique creates minimal artifact and is therefore ideally suited to delineate details of tissue relationships in the epidural space. He found variations in epidural anatomy due to all of these factors, and concluded that they may play a role in altering the ease of epidural entry and passage of both catheters and injected solutions.27

The epidural space exists outside the dural sac but within the vertebral canal. It is bounded anteriorly by the vertebral bodies and intervertebral discs, posteriorly by the laminae and ligamenta flavum, and laterally by the pedicles. The posterior compartment of the epidural space is filled by a fat pad that is triangular in shape on axial section. It lies between the dura and ligamenta flavum, but it extends slightly under the caudal-most portion of the laminae above. It does not adhere to these structures, which allows movement of the dura within the canal during spinal flexion.28–32 The lateral epidural compartment forms just medial to each intervertebral foramen and is filled with segmental nerves, vessels, and fat. The intervertebral foramen are widely patent (except in cases of severe stenosis) and allow free flow of solution injected into the epidural space. The anterior epidural space is almost entirely occupied by a nearly confluent internal vertebral plexus, from which the basivertebral vein originates as it penetrates into the vertebral body. Above the L4-5 disc level, the anterior epidural compartment is obliterated by the attachment of the posterior longitudinal ligament to each disc; below this level, the anterior epidural space widens to a capacious fat-filled cavity.27 When performing an lumbar ESI, the needle must pierce the skin and subcutaneous tissues, as well as the supraspinous and interspinous ligaments and finally the ligamentum flavum prior to passing into the posterior epidural space (Fig. 65-1). Conversely, there is a minimal to absent interspinous ligament in the cervical vertebral column, especially above the C7-T1 interspace. At the cervical level, with negligible interspinous ligament, entry into the epidural space could be achieved without passage through either interspinous ligament or ligamenta flavum and therefore without loss of resistance. This factor, combined with the very shallow or absent posterior compartment, makes epidural needle placement especially challenging at cervical levels above C7-T1.

The epidural space behaves like a Starling resistor,33 with pressures decreasing to a nonzero plateau after injection,34 independent of volume.35 As in the joint spaces and pleural cavities, the rigid enclosure of the posterior epidural space allows the tissues to generate a subatmospheric tissue fluid pressure.36 Entrance of a needle into the epidural space further decreases epidural space pressures. Subsequently, these changes in pressure create the force that draws the hanging drop of saline into the needle hub upon entrance into the epidural space. When using a midline approach to perform both a lumbar ESI and cervical ESI in an adult, the depth from the skin to the ligamentum flavum varies from 4 to 6 cm.37 Likewise, when drugs are injected into both the lumbar and cervical epidural spaces, they are absorbed by the epidural fat, taken up by the epidural vessels, and diffuse through the dura mater along adjacent nerve roots. Subsequently, the drug should be administered as close to the affected area as possible to maximize the amount of drug reaching the desired nerve root(s).

Although there are many variations in technique used to access the epidural space, the rate of success is primarily dependent on the experience of the physician. Epidural injections can be safely administered in the sitting, prone, and lateral decubitus positions, with or without the assistance of fluoroscopic guidance. Additionally, confirmation of entrance into the epidural space can be achieved by loss of resistance approach, “hanging drop” technique, or administration of contrast medium yielding an epidurogram. Whichever technique is chosen, every effort should be taken to maximize patient comfort, to use the technique most familiar to the physician, and to assure proper placement of the injectate into the epidural space at the desired level. Likewise, all precautions should be taken to maximize patient safety before any invasive procedure. Typically, we will ask our patients not to eat or drink for 6 hours prior to their injections (except taking necessary medications with a sip of water) and to refrain from taking nonsteroidal anti-inflammatory drugs (NSAIDs) for at least 5 days before the scheduled procedure.

Cervical Epidural Steroid Injection

Cervical epidural steroid injections (CESIs) are most commonly indicated in patients suffering from neck and arm pain secondary to cervical disc abnormality, spinal stenosis, spondylosis, and facet arthropathy with significant foraminal narrowing, which produces nerve root irritation and resultant radiculopathy.

During the procedure, the patient is continuously monitored with electrocardiogram (ECG) and for blood pressure and O2 saturation; a 20-gauge intravenous catheter is placed in a peripheral vein. With the patient in the sitting position and his or her elbows resting on each thigh to relax the shoulders, the neck is flexed so the chin approximates the chest with the forehead resting on a pillow support (Fig. 65-2). Taking a few moments to ensure proper positioning for the procedure in most cases allows for easy access into the epidural space and assures patient comfort. The posterior aspect of the neck is prepared and draped in usual sterile fashion and the desired interspace is identified and appropriately marked at the corresponding skin level. Next, 3 to 5 mL of 1% lidocaine is administered to the skin, subcutaneous and interspinous structures at this level, after which a 20-gauge 5-cm winged needle, or an 18-gauge 10-cm Tuohy needle, is introduced and advanced until it is firmly seated in the interspinous ligament. To use the “hanging drop” technique, the stylet of the needle is removed and preservative-free saline is used to fill the needle until a drop of saline sits above the hub of the needle. The wings of the needle are then held firmly between the thumb and index fingers while the hands rest on the patient’s shoulder to ensure needle stability and a slow, smooth advance of the needle into the epidural space. When the epidural space is entered, the negative pressure within the space allows the drop of saline at the hub of the needle to be drawn into the space. Alternatively, the “loss of resistance” technique to saline or air may be used to confirm needle placement in the epidural space. After negative aspiration for blood or cerebrospinal fluid (CSF) is performed, 3 to 5 mL of preservative-free normal saline mixed with 50 mg of methylprednisolone or 80 mg of triamcinolone is injected. The needle is then flushed and withdrawn, the neck is cleaned and dressed, and the patient is brought to the post-procedure room for monitoring. Prior to discharge patients are instructed to expect some soreness at the injection site, which may last a day or two. They are also told to call the office should they experience redness or increasing tenderness at the injection site, fever or chills, or significant increase in pain lasting greater than 1 day following their procedure.

Lumbar Epidural Steroid Injection

Although a variety of positions can be used to access the epidural space at the lumbar level, concern for both patient comfort and delivery of medication to the appropriate level should be primary factors in determining how each individual procedure is performed. I prefer to use the prone position with fluoroscopic guidance; however, many times patients who have had multiple lumbar surgeries or are experiencing severe pain and muscle spasms find it impossible to maintain this position for the duration of the procedure. Likewise, some younger patients will present having had a previous bad experience when a lumbar epidural steroid injection (LESI) was attempted in the sitting position that had to be halted because of a vasovagal event. Use of fluoroscopic guidance not only ensures placement of the injectate at the desired level but also significantly reduces the duration of the procedure. Additionally, direct visualization eliminates the “poking around” some patients report having experienced on previous injections, which despite obtaining good pain relief causes anxiety and hesitation for the patient contemplating an additional injection. With continuous ECG, blood pressure, and O2 saturation monitoring, a 20-gauge intravenous catheter is placed in a peripheral vein and the patient is placed in the prone position with pillow supports beneath his or her lower abdomen and pelvis to augment lumbar positioning for the procedure. The arms are placed overhead or hanging off the side of the table to ensure patient comfort and maintain a neutral lumbar position (Fig. 65-3). The lumbar area is prepared and draped in usual sterile fashion and fluoroscopic guidance is used to identify the appropriate interlaminar space (i.e., L4-5), which is then marked at the corresponding skin level. Next, 3 to 5 mL of 1% lidocaine is used to infiltrate the skin, subcutaneous and deeper structures, after which an 18-gauge 10-cm Tuohy needle is seated in the interspinous ligament. A 10-mL loss of resistance syringe filled with 3 to 5 mL of air or saline is attached to the needle, which is then slowly advanced under fluoroscopic assistance until the epidural space is entered, as evidenced by a positive loss of resistance. Alternatively, 3 to 5 mL of a mixture of 150 isovue and 1% lidocaine can be injected to produce an epidurogram at the desired level. We routinely use 3 to 5 mL of preservative-free saline mixed with 80 mg triamcinolone; however, a local anesthetic (1% MPF lidocaine) is occasionally used for patients who have acute low back pain with severe incapacitating symptoms. The local anesthetic is helpful in transiently decreasing the pain and spasm that typically occur during episodes of acute pain. The needle is then flushed and withdrawn, the low back is cleaned and dressed, and the patient is brought to the post-procedure room for monitoring. Before leaving, patients are given routine discharge instructions and encouraged to call should they have any questions or develop any concerns following their procedure.

Methylprednisolone and triamcinolone are the steroid preparations most commonly employed as injectates for intraspinal and epidural injections. Methylprednisolone comes in concentrations of 40 mg/mL and 80 mg/mL; triamcinolone is available in concentrations of 25 mg/mL and 40 mg/mL. A common dose used for epidural injection is 50 to 80 mg. Each steroid preparation contains additives including polyethylene glycol, myristyl-gamma-picolinium chloride, sodium chloride, polysorbate 80, and 0.9% benzyl alcohol. Although potentially neurotoxic at higher levels, studies have supported the safety of these diluents with no documented cases of direct neurotoxicity following administration into the epidural space (Table 65-2).

Complications

For the most part complications occurring after epidural steroid injections are rare and most commonly a result of technical error (i.e., postdural puncture headache, intravascular injection of local anesthetic, increased pain due to large volume of injectate used). Vasovagal events can also occur, and in our experience are more common in younger patients at the time of first injection, when the procedure is performed in the sitting position and when ambient room temperature is increased (70° to 74° degrees vs <70°).

Systemic steroid effects are rare,38–40 although plasma cortisol suppression by epidural steroids have been observed to last up to 3 weeks after injection.41 Subsequently, in a patient undergoing surgery within 5 weeks of receiving an ESI, exogenous steroids should be provided. Rare cases of Cushing’s syndrome38,39 have been reported, as has a case of congestive heart failure,42 following administration of higher doses of steroids. Subsequently, a recommended maximum dose of 3 mg/kg for methylprednisolone has been determined to prevent salt and water retention.43

There have been several reports of neurologic sequelae following neuroaxial steroid injections, all of which have involved subarachnoid injections or attempted epidural injections in which subarachnoid placement could not be ruled out.44 Nelson et al cited two cases of adhesive arachnoiditis among 23 patients who had undergone 83 subarachnoid injections and 1 case of aseptic meningitis developing several hours after subarachnoid administration of methylprednisolone.45

Several cases of epidural abscess occurring after epidural administration of both methylprednisolone46 and triamcinolone47 have also been reported. Simon et al48 described a case of allergic reaction to an epidural injection of triamcinolone diacetate, which manifested as skin rash, pruritis, dyspnea, abdominal pain, and diarrhea. Repeat challenge with a small alloquot of the same drug produced a similar response that required treatment with epinephrine and antihistamine. Epidural hematoma requiring surgical decompression following repeated cervical epidural steroid injections has been reported by Williams et al.49 This complication occurred in a patient who was taking indomethacin on a long-term basis and who had previously received 6 CESIs all without event. Additionally, intraoccular hemorrhage has been reported,50 as a case of transient paralysis secondary to brief neurocompressive injury with spontaneous resolution within 3 hours following ESI.51

Other minor side effects reported include transient headache,52 stiff neck,53 and facial flushing and generalized erythema.54 There appears to be little risk of serious complications associated with the use of ESI when total dose of medication given and the number of injections performed and time are kept to a minimum. Likewise, as the physician performing the procedure gains experience, those complications directly related to technique will likely diminish as well.

Epidural steroid injections can be effective in providing significant pain relief, particularly in patients experiencing low back, neck, or extremity pain secondary to nerve root irritation. The effectiveness of ESI in providing long-term relief in painful conditions as a result of bony spinal abnormalities has yet to be determined. The addition of steroid to an epidural local anesthetic or saline injection gives better results when compared with local anesthetic or saline alone.6 The mechanism by which ESI provides relief appears to be by an anti-inflammatory mechanism secondary to the inhibition of phospholipase A2 with resultant blockade of prostaglandin and leukotreine synthesis. A series of 1 to 3 injections is typically performed within a 6-month period, with additional injections considered after duration of relief and improvement in function have been determined. There does not appear to be a significant difference in the degree of relief provided by either methylprednisolone or triamcinolone. Plasma cortisol level is depressed following an ESI but returns to baseline within 3 to 5 weeks. ESIs are not without risk; however, the overwhelming majority of reported complications are directly related to technique with rare serious complications occurring after unintentional subarachnoid administration of steroid.