Fluoroscopic guidance and contrast enhancement are essential for accuracy when performing epidural injections (43). Published data show that even in experienced hands, epidural injections without fluoroscopic and contrast-enhanced guidance (i.e., “blind injections”) often result in inaccurate placement (Table 68-3) (43). These misplacements include the needle being inadvertently positioned into the subarachnoid, intravascular (Table 68-4), or subcutaneous regions (caudal approach) or fascial plane superficial to the ligamentum flavum for interlaminar ESI. Misplacement into the subarachnoid or intravascular regions has major potential safety implications, particularly for those injections that include local anesthetics as part of the injectate. Use of detection of flash back of blood in the needle hub to gauge the intravascular placement of needle is not a reliable substitute for looking for a vascular pattern after contrast injection (44). Although injection accuracy should also logically affect efficacy, there is very limited data on the efficacy of fluoroscopic-guided ESIs compared with blind ones. One such study demonstrated that fluoroscopic- guided transforaminal ESIs provided better pain relief than blind interlaminar ESIs (45). ESIs using fluoroscopic guidance have also been shown to reduce procedure-related complications compared to non–image-guided injections (46–48).
As a result of these factors, it is recommended that ESIs be performed under fluoroscopic guidance and with radiographic contrast, documenting appropriate needle placement in order to improve their accuracy, and by extension their safety and efficacy (49).
Efficacy of Epidural Injections
Recent studies have demonstrated good efficacy of lumbar ESIs when proper needle placement is confirmed by using fluoroscopic guidance and radiographic contrast (50,51). A meta-analysis of 12 published randomized controlled trials concluded that ESIs are effective (52). In a systematic review of randomized trials on lumbar epidural injections, Abdi et al. concluded that there was moderate evidence that caudal and TEIs are effective in providing long-term (>6 weeks) pain relief and limited evidence for the effectiveness of lumbar interlaminar ESIs (30). Other studies have suggested that 60% to 75% of patients receive some relief after ESIs (53,54). Benefits include relief of radicular pain and low back pain (generally relieving leg pain more than back pain), improved quality of life, reduction of analgesic consumption, improved maintenance of work status, and a decreased need for hospitalization and surgery in many patients (27,50–56). One study showed no difference in analgesic use in patients with sciatica who had received three ESIs (58). Another study reported that patients were more likely to start taking opioids and more likely to receive surgery after receiving multiple (>3) injections than patients receiving fewer injections (59). However, the population of patients receiving multiple steroid injections was more likely to have had more advanced disease such as spinal stenosis. A prospective cohort study was conducted on cervical TEIs for both neck pain and radicular pain from herniated discs or spondylosis. Twenty-one such patients awaiting surgery received cervical TESIs 2 times, at 2-week interval with 12 months follow-up. All patients had reduction in neck and radicular pain, and five of these patients cancelled the surgery (59). In contrast, a prospective randomized study involving 20 patients with cervical radicular pain confirmed by selective nerve root block (SNRB) and with magnetic resonance imaging (MRI) evidence of corresponding segmental pathology demonstrated that there was no difference in radicular pain reduction between steroid/local anesthetic and saline/local anesthetic groups at 3-week follow-up (60). A limitation of this study, however, was that it only involved small numbers of patients and that it is unknown whether saline/local anesthetic is a true control.
There are more studies in support of ESIs for low back pain (7,8,42,53–55) than there are negative studies (56). Problems with some of these supportive studies, however, include the fact that most of these studies did not use fluoroscopy and radiographic contrast to document accurate placement of the injected substance into the epidural space. Furthermore, many of these injections were not performed at the presumed level of pathology, even though this has been demonstrated to be critical to the success of ESIs (61). These methodologic problems are likely contributing factors to the mixed assessment that ESIs have received. A review of six prospective randomized clinical trials of fluoroscopic-guided transforaminal ESIs, selective nerve blocks, or periradicular nerve injections concluded that there is moderate (level III) evidence that TESIs are safe and effective in reducing radicular pain. However, more prospective, randomized, placebo-controlled studies using sham procedures are needed to provide more conclusive evidence for the efficacy of TESIs in treating lumbar radicular symptoms (62). A recent review article concluded that with proper patient selection, ESIs are a reasonable alternative to surgery for short-term pain relief, reduced medication use, and increased patient activities while awaiting natural recovery (63).
Aside from technical considerations, response to ESIs has been shown to be related to several other factors such as the type and quantity of steroid preparation used, volume of injectate, underlying pathophysiology, and the duration of symptoms (23,26,28). In general, radicular pain or radiculopathy induced by herniated nucleus pulposus appears to respond better to corticosteroid injection than that induced by spinal stenosis. There is essentially no literature that correlates the type of disc herniation with the response of ESIs. It is the authors’ collective experience and observations that patients with large lumbar disc herniations obliterating the neuroforamen or extraforaminal herniations often have less benefit from ESIs. One study demonstrated that radiculopathy induced by the combination of spinal stenosis and disc herniation has less favorable outcome with ESI. In lumbar spinal stenosis, the efficacy of ESI correlated with the degrees and the levels of stenosis categorized by MRI (64). Patients with single-level lumbar spinal stenosis generally respond better than those with multilevel lumbar spinal stenosis. ESIs provide better efficacy in reducing pain and opioid consumption for patients with mild to moderate rather than severe stenosis. But a prospective cohort study with 12-month follow-up in patients with severe degenerative lumbar spinal stenosis found that fluoroscopicguided and contrast-enhanced caudal ESIs reduced bilateral radicular pain and improved standing and walking tolerance (65). In contrast to radiculopathy due to herniated discs and/ or spinal stenosis, radiculopathy caused by epidural scar tissues or trauma such as nerve root stretch injury often responds poorly to ESI.
A recent prospective, randomized study on lumbar TESIs demonstrated positive efficacy in treating radicular low back pain. The success rate for TESI is 84%, compared to 48% with trigger point injection, at 1.4 years of follow-up (66). Another prospective, randomized controlled clinical trial compared perineural (transforaminal) epidural injection with conventional posterior (interlaminar) epidural injection with steroid, and perivertebral injection with local anesthetic as a control group (27). The result demonstrated that perineural injection was the most effective approach. Both perineural and conventional epidural injection with steroid were better than that with saline alone (27).
Uncontrolled studies have generally reported favorable outcome of cervical epidural injections for cervical radiculopathy with structural abnormalities such as cervical disc herniation (66,67) and spondylosis (68). However, the prospective, randomized, blinded and controlled clinical trials on the outcome of cervical and thoracic epidural injections have not been reported yet in the peer-reviewed literature. At the time of this writing, there have been no prospective randomized trials on thoracic ESIs that have been published in the peer-reviewed literature.
Patients should be educated that ESI alone may not be the only solution to give them long-term benefits. ESI is just one of many nonoperative treatments used to treat low back pain or radicular symptoms. Other treatments may include shortterm bed rest; medications (e.g., analgesics, muscle relaxants); a properly designed program of physical therapy; and management of any psychological, financial, marital, and work-related problems. A comprehensive treatment approach is likely to produce better outcomes for patients with low back pain than any single modality used in isolation (23,26,51). Recently published research on the outcome of ESIs has supported this notion of multifaceted treatment (50,69).
Recent Advances and Investigations on the Management of Radicular Pain
A study was performed on lumbar TEIs for radiculopathy using autologous conditioned serum (ACS) containing enriched IL-1 antagonist. The ACS group showed statistical superiority over both triamcinolone groups (5 and 10 mg) with regard to the VAS score for pain from week 12 to the final evaluation at week 22, statistical superiority at week 22 compared to the triamcinolone 5 mg group, and no significant difference compared to the 10 mg triamcinolone group. This is an exciting finding, as autologous blood is not associated with the same side effect concerns associated with corticosteroids, and theoretically can be used more frequently than corticosteroids. Additional studies are needed to confirm this finding (70).
An animal study examined the potential benefits from anti-TNF-[α] therapy in reducing neurotoxic effects induced by the nucleus pulposus on neuronal tissues (71). Two openlabel human clinical trials, one using intravenous infliximab (a monoclonal antibody against TNF-[α]) and the other using (etanercept) a soluble TNF-[α] receptor antagonist, in patients with sciatica from disc herniation demonstrated significant efficacy in pain reduction (72,73). Although these basic science and human studies initially implied potential clinical use of anti-TNF-α medication as a treatment for patients with radiculopathy due to disc herniation, there were disappointing long-term findings related to the evaluation of the efficacy of an anti-TNF-[α] treatment versus a placebo injection in disc herniation-induced sciatica in a randomized controlled setting. Specifically, 3-month results showed no difference in the patient-reported symptoms or in the more objective outcomes (SLR, days on sick leave, discectomies) between intravenous infliximab 5 mg/kg and placebo (74). The 1-year results also confirmed the earlier findings (75). Clearly, further studies using multiple intravenous infusion of the anti-TNF-α agents or epidural injection of the similar substances are necessary to clarify any efficacy of anti-TNF-α treatment in radicular pain.
Safety and Complications of Epidural Injections
A retrospective cohort study reviewing the immediate complications of 2,217 patients who received selective lumbar nerve root blocks under fluoroscope, reported a 5.5% minor complication rate (76). When performed by a skilled, experienced clinician within an appropriate setting and on carefully selected patients, the chance of a significant complication from an ESI is remote (23,25,26,77–83). Like any procedure that punctures the skin, bleeding and soft tissue infection are potential but rare risks. More common risks of epidural injection are acute back pain, postural puncture headache (0.5% to 1% for lumbar interlaminar and 0.6% for caudal epidural injections), nausea, vomiting, dizziness, vasovagal reactions, and epidural hematoma (0.001%) (23,25,26,29). Nerve root injury, arachnoiditis, and meningitis also have been reported but are very rare. Lumbar transforaminal and caudal epidural injections were associated with 9.6% and 15.6% of minor complications, respectively (82,83). Anterior spinal cord syndrome has been reported after lumbar transforaminal ESIs, presumably due to inadvertent needle contact with, or local anesthetics induced spasm of, the artery of Adamkiewvicz. Although the artery of Adamkiewvicz is usually located on the left side from T9 to the L3 segments, anterior cord syndrome has been reported in transforaminal ESI as low as the S1 level. Other possible mechanism of spinal cord injury related to ESIs include embolism of the injected corticosteroid particles causing spinal cord ischemia. Although rare, spinal cord injuries have been reported due to cervical and thoracic epidural injections from direct needle trauma, presumed radicular artery spasm, or steroid particle embolization (84–87). Therefore, transforaminal ESIs, especially cervical TEIs, should be performed by the most skilled and highly experienced injectionists. Real-time fluoroscopic imaging during contrast injection should be employed. Digital subtraction angiography may provide an additional safety margin for the prevention of inadvertent intra-arterial injection. Dexamethasone should be chosen as the steroid for transforaminal ESIs due to its small particle size among the various corticosteroid preparations. A recent study has demonstrated that dexamethasone and the larger particulate-sized methylprednisolone (88) have essentially the same efficacy in cervical epidural injections (89).
Corticosteroid-induced adrenal insufficiency has been reported. For example, mild hypothalamic-pituitary-adrenal (HPA) axis suppression has been reported from 1 to 3 months after receiving a total of three ESIs (once weekly) with 80 mg of triamcinolone (Aristocort) in 7 mL of 1% lidocaine (90).