Summary of Evidence
Acute LBP, defined as LBP with less than 4 weeks duration, has an excellent prognosis to recover within the first 4 to 6 weeks. Initially, treatments include non-invasive therapies, including pharmacologic and non-pharmacological measures. Patients with acute or subacute nonspecific LBP typically would not be considered candidates for interventional therapies, as they could still improve with non-invasive therapies, and there is little evidence on the efficacy of invasive therapies in this circumstance.24 The American College of Physicians (ACP) offers a strong recommendation for non-pharmacological measures as the first line treatment. Recommended non-pharmacological measures include conservative measures such as superficial heat, exercise, multidisciplinary rehabilitation, acupuncture, mindfulness-based stress reduction having moderate-quality evidence. For pharmacologic therapy, they recommend NSAIDs as first line treatment, and muscle relaxants as second-line treatment.25 Systemic steroids may provide benefits with 3 randomized control trials (RCTs) showing benefit, especially in the short term. The benefit of anticonvulsants and antidepressants is overall low quality, but there is potential benefit in some patients.26 A systematic review comparing ESIs to conservative treatment for patients with lumbosacral radicular pain reported that ESIs were more effective for alleviating lumbosacral radicular pain but not function in both short- and intermediate-term compared to conservative treatment; however, this effect was not maintained at long-term follow-up.27 It is standard practice to utilize conservative treatment measures for the management of acute LBP. When back pain persists over 12 weeks, it is considered CLBP, and spontaneous recovery is less likely. The SMEs voted 3/5 (range 2-5) that there is evidence to support periods of conservative management prior to treatment with epidural injections. The panel discussed time for conservative care is related stronger to the natural history of the disease than the evidence, as it has not been well studied. The panel voted low confidence (2/5; range 1-5) in terms of evidence to support documentation of failure of at least 2 classes of medication prior to ESI. The use of medication is more controversial due to the risk associated with non-steroidal anti-inflammatory, specifically gastrointestinal complications, and the risk of opioid addiction with opioid use. Other medications remain available, but there is not strong evidence to support 1 specific category of medication.
Patients with LBP should be evaluated for the underlying condition, and targeted treatments implemented. History and physical examination, including a neurological exam to evaluate for radicular pain, should be performed. Imaging should be targeted to identify specific differential diagnoses and is not typically indicated in the early evaluation of LBP.28 The American College of Radiology (ACR) offers evidence-based criteria for imaging for LBP.29 According to the ACR criteria, patients who have acute, subacute, or chronic LBP or radiculopathy with persistent or progressive symptoms during or following 6 weeks of conservative management are considered candidates for surgery or intervention. Imaging may be considered for patients who have had no improvement in their back pain after 6 weeks of medical management and physical therapy. Imaging is classified as “may be appropriate”, with MRI of the lumbar spine without contrast classified as “usually appropriate” and CT if MRI is unavailable as needed for diagnostics. The American College of Occupational and Environmental Medicine guidelines state that while diagnostic tests are not indicated for the majority of LBP, the literature supports CT or MRI.30 The ESIs are utilized as an option for a specific diagnosis, and therefore a thorough evaluation to achieve diagnosis is a necessary step prior to consideration of ESIs. In most cases, by the time ESIs are considered, imaging has been completed as part of the evaluation. Any prior imaging should be reviewed for anatomy prior to interventional procedures. The SMEs voted low confidence (2/5; range 1-4) that radicular pain should be concordant with a radiologist’s interpretation of an advanced diagnostic study demonstrating compression of the involved nerve root. The concern was that the imaging study is a static image in a dynamic process and the compression may be missed in the supine position even if present. The SMEs felt there was benefit in obtaining imaging to aid in diagnosis, such as to reveal anatomical features that may guide treatment options or alter approach, selection of needle size and appropriate segmental level. There was also concern that imaging would not reveal non-anatomical sources like chemical radiculitis.
Epidural administration of corticosteroids has not been approved by the FDA and is classified as “off-label” use. While complications with ESIs are rare, they can be devastating. In 2014, the FDA issued a drug safety communication about the epidural injection of glucocorticoids, noting the potential for rare but serious adverse effects, and the effectiveness has not been established.31 The warning followed 2 separate outbreaks of fungal meningitis 2002 and 2012-2013 linked to ESIs and multitudes of case reports on related complications. After the FDA warning, there was a flurry of papers to document safety and provide protocols and/or guidelines to improve safety. The reported risk of ESIs includes loss of vision, stroke, dural puncture, infection, spinal cord injury, paralysis, and death. In addition to IA injections of particulate steroids, arachnoiditis, nerve damage, osteomyelitis, hemorrhage, and epidural abscess have been reported.1,32 In a 2015 review on trends in safety and complications, the author explains the reasons for these risks stating “significant neurological complications associated with transforaminal injections include stroke and spinal cord injury, which arise due to the need to place the needle within the intervertebral foramina. This space contains radicular arteries that perfuse the spinal cord and course adjacent to the targeted spinal nerve root, creating a challenge when variations in anatomical structure are present."1 A 2018 survey of 249 interventional pain management physicians (13.8% response rate) demonstrated a high level of variability and lack of standardized practices. They reported that in ILESI practice patterns, there is a lack of standardization in needle sizes, use of imaging, and choice of injectant.33 The lack of evidence-based standardized practice makes it difficult to interpret the literature since these variables exist across existing studies.
A 2015 review on key safety when administering ESIs states the FDA identified 131 neurological adverse events, including 41 cases of arachnoiditis and 700 cases of fungal meningitis following injection of contaminated steroids. They further explain that most complications were related to TFESIs, specifically cervical TFESIs. They conclude that proper technique can avoid injury, and lumbar TFESIs can be performed safely; however, they warn that cervical TFESIs “must not be performed until appropriate evidence develops and safe preparations of steroids are available."32 A 2015 multidisciplinary workgroup stated concern regarding the injection of ESIs under sedation and without appropriate precautionary steps, such as injection of radiographic contrast medium under fluoroscopy, resulting in spinal cord injury. They also expressed concern for using particulate steroids, such as methylprednisolone, triamcinolone, or betamethasone which have been linked to cerebrovascular occlusion associated with cervical TFESIs.14 This review called for more standardized safety protocols and research in best practices.
A 2016 multi-institutional study on adverse event rates associated with TFESI and ILESI reported no major adverse events in 16,638 consecutive procedures in all spine segments (14,956 TFESI; 1,682 ILESI). The most common adverse event was a vasovagal reaction in 1.2% of procedures (p=0.004). Dural punctures occurred in 0.06% of procedures, more commonly after ILESI (0.2% vs 0.04%, p=0.006). Central steroid response (sleeplessness, flushing, non-positional headache) was seen in 2.6% of TFESI and ILESI patients. Patients that reported increased pain included 2.1% of TFESI and 1.8% of ILESI patients. No long-term sequelae were seen from any immediate or delayed minor adverse event at the time of the procedure or follow-up.34 A 2012 prospective study on 10,261 fluoroscopic guided epidural procedures included 2,376 cervical interlaminar epidural injections, 301 thoracic interlaminar epidural injections, 1,450 lumbar interlaminar epidural injections, 1,395 caudal epidural injections, 1,310 lumbar transforaminal epidural injections, and 839 caudal epidural L adhesiolysis procedures. The authors reported an adverse rate of overall intravascular penetration of 4.3%, local bleeding of 63%, 0.5% rate of dural punctures with 0.05% post-lumbar puncture headache, 0.85% transient nerve irritation of 0.08% as well as transient spinal cord irritation and other minor complications, but no major complications.35
The WIP Benelux Work Group developed evidence-based safety guidelines based on existing literature. In reviewing major complications, they reported that the majority involved direct nerve trauma or spinal cord injury. To reduce the risk of nerve damage, they recommended avoidance of deep sedation so the patient can alert the provider to any paresthesia during needle placement to minimize this risk.16 The SMEs agreed that evidence supports that ESI should not be performed with moderate sedation or general anesthesia with a score of 3/5 (range 1-5). The American Society of Anesthesiologists (ASA) provides definitions on the continuum of the depth of sedation. Minimal sedation anxiolysis is defined as a drug-induced state during which patients respond normally to verbal commands. Although cognitive function and physical coordination may be impaired, airway reflexes, and ventilatory and cardiovascular functions are unaffected.36 Progressing depth of sedation beyond minimal is moderate or “conscious,” deep and general. Monitored Anesthesia Care or “MAC” is not a depth of sedation, but a specific anesthetic service allowing a deeper level of analgesia than can be provided moderate sedation.
The impact of glucocorticoids injected into the epidural space is not fully defined. Evidence supports that even a single injection risk includes bone demineralization and increased risk of fractures in postmenopausal women, suppression of the hypothalamic-pituitary adrenal axis, immune dysregulation, and hyperglycemia in patients with diabetes mellitus.1,16 A small prospective study in which 28 postmenopausal women who underwent a single ESI (triamcinolone 80 mg) had bone mineral density (BMD) measurements completed and showed a significant decline in the hip BMD of 0.018 g/cm2 (0.028 ± 0.007, P=0.002) at 6 months compared with baseline as compared to age-matched control population and concluded there is a potential deleterious effect of ESI with steroids on BMD.37 A large retrospective cohort study matched 3,000 patients who had ESIs to 3,000 similar non-injected patients. Based on their analysis, they conclude each successive injection increased the fracture risk by a factor of 1.21 (95% confidence interval [CI], 1.08 to 1.30) after adjustment for covariates (p=0.003), concluding a 21% increase risk of fracture risk per ESI injection.1,38
In contrast, a 2,000 prospective study on the effect of corticosteroids and dose relationship on weight gain and BMD receiving steroids through the neuraxial block included ESI and facet joint injections. Data were collected on 123 patients divided into 2 groups. One received injection with LA plus steroids, and the other LA alone. A BMD study was conducted at baseline and repeated at 3, 6, and 12 months and concluded no significant change in BMD at 12 months from baseline.39 However, the ESI doses in this study were lower than routinely used, so results must be interpreted cautiously. The lack of standardized injectant and dosing make it difficult to determine if there is a safe threshold in this population. A 2019 retrospective paper reported on 172 postmenopausal women with osteoporosis and with half receiving ESIs. The mean number of ESIs was 6.2, and the mean cumulative administered dose of glucocorticoids (dexamethasone) was 31 mg. The incidence of fractures in the medication only (for back pain) and ESI groups were 22% and 24%, respectively, in the thoracolumbar spine, and 2% and 5%, respectively, in the hip joint. There was no significant difference in the incidence of osteoporotic fractures between the groups suggesting a maximum cumulative dexamethasone dose of 31 mg could be safely used in postmenopausal women with osteoporosis.40 This was limited by retrospective design and a small sample size. The WIP Benelux Guidelines cite each ESI increases the fracture risk by 31% and recommends keeping the corticosteroid exposure to a minimum, especially for high-risk individuals such as the elderly and women with prior history of osteoporosis or osteopenia.16
There are reported cases of suppression of the hypothalamic-pituitary-adrenal (HPA) axis after ESI causing Cushing syndrome. While the incidence of Cushing syndrome is low, HPA access suppression is frequent and can continue for 3 to 6 weeks. A diabetic patient’s elevation in glucose levels may be seen after ESI persisting for 2 to 6 days after the initial injection with a risk of hyperglycemia. There is also concern about the impact of corticosteroids on the immune system, and dose-dependent suppression of the immune system has been reported after ESI.16 In a retrospective study, it was found that corticosteroids reduced the effectiveness of the vaccine and patients were at increased risk for developing influenza after a major joint injection with corticosteroids (relative risk, 1.52; 95% CI, 1.20-1.93).41 This led to recommendations to avoid elective procedures and reduce steroid dosing for necessary procedures in elderly and high-risk patients during the COVID-19 pandemic.42 SMEs agreed with high confidence (4/5; range 2-5) that there is evidence to support a maximal steroid dose for corticosteroid limits for the injection. The precise maximum is not established with variations in the literature, but there is consistent evidence to use the lowest effective dose.16
Bleeding risks are a significant concern for patients undergoing ESIs. An epidural hematoma can create an expanding volume that compresses the spinal cord and/or nerve roots and can result in potentially catastrophic neurological complications. This risk is a possibility for hematologically normal patients but significantly higher if on anti-coagulation. The practice of holding anticoagulation to perform ESIs is controversial. While holding anticoagulation may potentially reduce the risk of bleeding, there is an increased risk of thromboembolic events during the time patients are off anti-coagulants. A 2012 survey of 325 interventional pain physicians reported the number of bleeding complications after ESI was similar regardless of whether anticoagulation was held. Still, the reported number of thromboembolic events was 17-times greater when anticoagulants were held compared with when not held.43 In the American Society for Interventional Pain Physicians (ASIPP) updated Comprehensive Evidence-Based Guidelines Part II: Guidance and Recommendations, the authors report44:
- Good evidence for risk of a thromboembolic phenomenon in patients with antithrombotic therapy if discontinued, spontaneous epidural hematomas with or without traumatic injury in patients with or without anticoagulant therapy to discontinue or normalize the international normalized ratio (INR) with warfarin therapy, and the lack of necessity of discontinuation of NSAIDs, including low dose aspirin prior to performing interventional techniques.
- Fair evidence with excessive bleeding, including epidural hematoma formation with interventional techniques when antithrombotic therapy is continued, the risk of a higher thromboembolic phenomenon than epidural hematomas with discontinuation of antiplatelet therapy before interventional techniques and to continue phosphodiesterase inhibitors (dipyridamole, cilostazol, and Aggrenox®).
- Limited evidence to discontinue antiplatelet therapy with platelet aggregation inhibitors to avoid bleeding and epidural hematomas and/or continue antiplatelet therapy (clopidogrel, ticlopidine, prasugrel) during interventional techniques to avoid cerebrovascular and cardiovascular thromboembolic fatalities.
- Limited evidence about dabigatran (Pradaxa®) and rivaroxaban (Xarelto®) to discontinue to avoid bleeding and epidural hematomas and are continued during interventional techniques to avoid cerebrovascular and cardiovascular thromboembolic events.
The American Society of Regional Anesthesia (ASRA) published a 2018 systematic review of 14 articles for evidence regarding risks associated with either continuing or ceasing anticoagulant and antiplatelet medication in patients having image-guided interventional spine procedures using GRADE methodology. Interlaminar procedures carried a greater risk of hemorrhagic complications regardless of whether the anticoagulants were ceased or continued, while other procedures did not carry that risk and could be performed while continuing anticoagulation. Three articles reported adverse effects of ceasing anticoagulants, with serious consequences, including death.45 However, this report was limited to case studies, case series, several observational studies, 1 large retrospective study, and 1 large prospective study with broad heterogenicity, multiple different neuraxial procedures included, and nonuniformed protocols making it difficult to draw clear conclusions.
A summary of current literature supports that withholding anticoagulation medication to reduce the risk of major bleeding such as hematoma from ESIs may increase the risk of thromboembolic events. A 2015 review concludes, “the decision should be considered in the context of a clear understanding of the patient’s anatomy, established guidelines and the perspectives of the patient and prescriber of the anticoagulation medication(s).”1
The SME panel agreed on the absolute contraindications included in the LCD with voting of 3 to 5 for each of the listed contraindications. However, the panel had low confidence score 2/5 (range 1-4) that the evidence supports the continuation of anticoagulation for epidural injections. The wide range demonstrates the continued controversy which is not resolved among variations in societal guidance.
Image Guidance Procedures
The use of imaging during the performance of ESIs is considered standard. A 2016 systematic review to determine the effectiveness and risks of non-image-guided lumbar ILESIs reviewed 92 primary publications using GRADE methodology and concluded that these procedures should be performed with image guidance and procedures without imaging should be restricted to the rare settings where fluoroscopy is not available.13 A 2015 multi-disciplinary workgroup to improve the safety of ESIs included a recommendation that all cervical and lumbar ILESIs should be performed with image guidance plus a test dose of contrast medium. They further state that cervical and lumbar TFESIs should be performed by injecting contrast medium under real-time fluoroscopy or digital subtraction imaging.14 Multiple other societies also endorse imaging requirements for ESIs. The North American Spine Society (NASS) Choosing Wisely recommendations include “elective spinal injections, such as epidural steroid injections, should be performed under imaging guidance using fluoroscopy or CT with contrast enhancement (unless contraindicated) to ensure correct placement of the needle and to maximize diagnostic accuracy and therapeutic efficacy."46 The NASS published guidelines suggest that while there is some evidence to support ultrasound-guided lumbar TFESI, this data is limited and, due to the risk of serious complications with ESIs, it is not considered standard care.47,48
Evaluation of Success
Once an ESI is performed, the patient must be assessed to determine if the injection was successful. This requires a clear definition of success. Success should include a reduction in pain symptoms and/or as improvement in function. Most research protocols studying the efficacy of ESIs used the ODQ or Roland-Morris Disability Questionnaire (RMDQ) to measure function, VAS or NRS to measure pain, and a variety of questionnaires to measure the quality of life, depression and other outcomes. Most studies used a threshold of 50% improvement from baseline to consider the ESI successful. While questionnaires may be burdensome in clinical practice, a standard measure for improvement is important to determine if the treatment is effective. Further research for simple assessment tools is needed.
A 2020 prospective observational study for 16 patients >65 years old with symptomatic lumbar stenosis with radiculopathy who selected ILESI or medical management were followed with the Short Physical Performance Battery (SPPB) and numeric pain score. The SPPB score includes a 4-meter walk test, chair stand time, and balance score. There was statistically significant improvement reported in SPPB score at 1 month for the injection group, but not in the medication group. Most studies on ESIs use metrics that include improvement in pain scores and/or questionnaires. There are few studies that measure functional outcomes. The RMDQ is a validated measure of disability and patient-reported function but does not include objective or observable measures. This study, while too small in number for conclusive results, offers a functional parameter to measure improvement after ESI, stating the SPPB is a quick and feasible screening tool. They state, “the SPPB is an easy-to-use tool to measure changes in physical function in older adults and could easily be integrated into the outpatient pain clinic setting. Even a 1-point change in an SPPB score and its subsets is clinically meaningful and correlates with decreased mortality and disability.”49
Success may also be defined as avoidance of surgery. A small double-blinded RCT, looked at 21 patients previously identified as operative candidates who delayed surgery and underwent ESIs. Seventeen of the 21 did not undergo operative intervention at the 5-year mark. The authors concluded that injections can avoid operative intervention.50 Limitations are the very small sample size and risk of selection bias. Throughout the evidence review below there are studies which do not show a reduction in need for surgery, while others demonstrate a trend in reduction of surgical management, with no moderate- or high-quality literature to confirm this observation.
The SMEs were asked about their confidence that the clinical literature supports that ESIs provided at least 50% pain relief and voted 3/5 (range 1-5). The panel expressed high confidence in using a measure of function for ESIs success with a score of 4/5; range 2-5. The panel discussed that a clinical success must be a combination of both improvement in pain and/or function and that standardized scales used in the setting of research may not be practical for clinical practice. The panel does acknowledge the need for some standardized measurement of successful outcomes of the procedures.
A 2014 Cochrane review on MBR for CLBP included 41 RCTs (n=6858). It concluded there is moderate-quality evidence that patients with CLBP receiving MBR are likely to experience less pain and disability than those receiving usual care or physical treatment. They also correlated a positive influence on work status.10
The ASA Task Force of Pain Management recommends ESIs with or without LA as part of a multimodal treatment regime to provide pain relief in selected patients with radicular pain or radiculopathy.12
Lumbosacral Radicular Pain
The effectiveness of ESIs has been a source of controversy. Multiple studies and systematic reviews reported favorable results, but several reports offered conflicting evidence. A 1997 RCT published in the New England Journal of Medicine with 158 patients with disc herniation and pain were randomized to ESI with glucocorticoid or placebo. At 3 weeks, the Oswestry score had improved by a mean of -8.0 in the methylprednisolone group and -5.5 in the placebo group (95% CI for the difference, -7.1 to 2.2), there was an improvement in leg pain in the steroid group at 6 weeks, but no significant difference between the groups at 3 months. At 12 months, the cumulative probability of back surgery was 25.8% in the methylprednisolone group and 24.8% in the placebo group (P=0.90). They concluded there was short-term improvement from ESIs for herniated disc but without significant functional benefits or reduction in the need for surgery.51 In 2015, the Agency for Healthcare Research and Quality (AHRQ) released a technology report consisting of a systematic review of RCT of patients with lumbosacral radiculopathy, spinal stenosis, non-radicular back pain, or chronic post-surgical back pain. Quality of evidence was assessed for risk of bias using Cochrane Back Review Group criteria. Meta-analysis was performed and stratified by time. Seventy-nine RCTs of ESIs were included. Thirty placebo-controlled trials evaluated epidural corticosteroid injections for radiculopathy. For radiculopathy, epidural corticosteroids were associated with a greater immediate-term reduction in pain (weighted mean difference on a scale of 0 to 100, -7.55 [95% CI, - 11.4 to -3.74]; strength of evidence [SOE], moderate), function (standardized mean difference after exclusion of an outlier trial, -0.33 [CI, -0.56 to -0.09]; SOE, low), and short-term surgery risk (relative risk, 0.62 [CI, 0.41 to 0.92]; SOE, low). For ESIs compared to placebo, they concluded that the magnitude of effects on pain and function was small. They did not meet predefined thresholds for minimum clinically important differences, lacked longer-term benefit, and did not demonstrate effectiveness (SOE: insufficient to low). The authors concluded that corticosteroid injections for radiculopathy were associated with immediate but short-term benefits and did not reduce the longterm risk of surgery. Limitations of this paper included methodological shortcomings of the RCTs evaluated.52
A 2016 systematic review and meta-analysis included RCTs comparing TFESI with steroids (with or without LA to LA or saline in adults with lumbosacral radicular pain secondary to herniated intervertebral discs. They reviewed data on NRS scores for pain, validated scores for measuring physical disability and quality of life, and incidence of surgery measured at 1 month to 2 years after the interventions were meta-analyzed. Evidence was rated with GRADE. Eight studies were included, including 771 patients (336 in steroid group and 405 in the comparator group). They concluded that TFESI has no impact on physical disability or incidence of surgery. They reported the following:
- Patients who received TFESI steroids reported a significant, but clinically modest, reduction in mean pain scores (0-10 scale) compared with LA/saline (-0.97 points; 95% CI, -1.42 to -0.51 points; P<0.0001, I2 = 90%; GRADE weak recommendation; moderate-quality evidence) at 3 months after the interventions.
- The TFESI with steroids did not decrease physical disability at 1 to 3 months after the intervention (GRADE strong recommendation; high-quality evidence)
- The TFESI did not decrease the incidence of surgery at 12 months after the intervention (GRADE strong recommendation; moderate-quality evidence) compared with LA/saline.
- The TFESI steroids provide modest analgesic benefit at 3 months in patients with lumbosacral radicular pain secondary to herniated intervertebral discs.
Limitations were the high degree of heterogeneity among the publications included in this meta-analysis.53
In the same year, a systematic review of 52 RCTs with placebo-control or active-control design looked at the efficacy of ESIs in the management of chronic spinal pain. The quality of each article was assessed by Cochrane review criteria, as well as the Interventional Pain Management Techniques-Quality Appraisal of Reliability and Risk of Bias Assessment (IPM-QRB). The authors concluded there was Level II evidence for managing lumbar disc herniation or radiculitis for long-term improvement either with caudal, interlaminar, or transforaminal epidural injections with no significant difference in approaches. They challenged the methodology of the contrasting report and stated the limitations are the paucity of high-quality randomized trials literature which continues to confound the evidence.54
A 2016 comparative systematic review and meta-analysis included 31 RCTs that compared the efficacy of ESIs for lumbar and spinal stenosis performed with saline with steroids, LA alone, or steroids with a LA. The review included 9 placebo-controlled trials evaluating ESIs, either with sodium chloride solution or bupivacaine, compared to placebo injection, and 12 studies comparing LA alone to LA with steroid. The meta-analysis of 5 studies utilizing sodium chloride or bupivacaine with steroid showed a lack of efficacy. Comparing lidocaine to lidocaine with steroids in 7 studies showed significant effectiveness from baseline to long-term follow-up, which was considered 3 months. Meta-analysis showed similar effectiveness for pain and function without non-inferiority of lidocaine compared to lidocaine with steroid at 3 months and 12 months. The duration of follow-up varied among the included trials with a minimum of 3 months, so the 12-month outcome data were from the limited number of trials extended for that duration. They concluded ESIs for radiculopathy or spinal stenosis with sodium chloride solution or bupivacaine were shown to be ineffective. Lidocaine alone or lidocaine in conjunction with steroids was significantly effective.55
A 2017 systematic review and a meta-analysis using GRADE methodology evaluated 71 publications, including observational and RCTs addressing fluoroscopically guided lumbar ILESIs to determine the effectiveness and risk of a variety of etiologies, including lumbar radiculopathy. They evaluated the body of literature for specific underlying etiologies. They reported most evidence regarding lumbar ILESI was observational and low-quality. They concluded radicular pain due to lumbar disc herniation provides a statistically significant short-term improvement in pain.17 A 2019 Korean review using GRADE concluded a high level of evidence [LoE] for ESI for radicular pain with the strength of recommendation [SoR] as strong. They favored TFESI over ILESI (LoE :moderate, SoR: weak) and CESI (LoE:moderate, SoR: strong).26
Few studies address long-term outcomes, with most going to 12 months as the longest duration. A 2018 prospective cohort study contacted patients who had received single-level lumbar TFESIs for herniated nucleus pulposus to evaluate long-term outcomes. Of the original 78 patients enrolled, they could contact 39 (50%) at 5 years for follow-up. They found that despite a high success rate at 6 months, most subjects experienced a recurrence of symptoms at some time during the subsequent 5 years. This concludes that lumbar disc herniation is a disease that can be effectively treated in the short-term by TFESI or surgery, but long-term recurrence rates are high regardless of treatment received.56
In 2020, a Cochrane database systematic review reported 25 clinical trials with 2,470 participants, comparing epidural corticosteroid injections to placebo for lumbosacral radicular pain. Using GRADE methodology, 8 trials were considered high-quality.57 They report the overall body of literature ranged from very low to moderate. The study concludes there is limited evidence for the use of ESIs in people with lumbosacral radicular pain as the treatment effects are small, mainly evident at short-term follow-up, and may not be considered clinically important by patients and clinicians (i.e., mean difference lower than 10%). No minor or major adverse events were reported at short-term follow-up. The report limitations included insufficient information on how or when adverse events were assessed for immediate and long-term follow-up.
Key findings reported include:
- The ESIs were probably slightly more effective compared to placebo in reducing leg pain at short-term follow-up (mean difference [MD] -4.93, 95% CI -8.77 to -1.09 on a 0 to 100 scale; 8 trials, n=949; moderate-quality evidence [downgraded for risk of bias]).
- For disability, epidural corticosteroid injections were probably slightly more effective compared to placebo in reducing disability at short-term follow-up (MD -4.18, 95% CI -6.04 to -2.17, on a 0 to 100 scale; 12 trials, n=1367; moderate-quality evidence [downgraded for risk of bias]).
- The treatment effects are small, however, and may not be considered clinically important by patients and clinicians (i.e., MD lower than 10%).
- There is uncertain evidence if ESI makes no difference compared to placebo injection in the frequency of minor adverse events (risk ratio [RR] 1.14, 95% CI 0.91 to 1.42; 8 trials, n=877; very low-quality evidence [downgraded for risk of bias, inconsistency, and imprecision]).
A 2020 systematic review by Smith et al. evaluated the effectiveness of lumbar transforaminal injection of steroid for the treatment of radicular pain. This review used GRADE methodology and reviewed 32 observational cohort studies, 9 pragmatic trials, and 2 explanatory trials published between 1997-2017. Using a reduction of pain of =50%, the authors reported the success rate across the included studies. They reported a success rate for disc herniations of 63%, 74%, 64% and 64%, at 1, 3, 6, and 12 months respectively. The authors stated high-quality evidence for the effectiveness of TFESI, based on outcomes of multiple RCTs and high-quality observational studies. Most studies in this review showed treatment benefits lasting 3 to 6 months, with some studies suggesting benefits at 1 or even 2 years post-injection. However, they attribute the 1 to 2-year benefit likely related to the natural history of lumbar radicular pain rather than a direct effect of the corticosteroids. One of the challenges with this body of literature is that the follow-up interval is highly variable among the studies, and co-interventions are prominent, so it is difficult to determine the true durability of the intervention. The literature reviewed showed a trend towards a reduction in surgery in patients who reported success with ESIs.58
While the overall clinical experience of the SMEs leaned towards a reduction in surgery with the use of ESIs they agreed this was not consistently reproduced in the literature with a score of 2/5 (range 2-5). Overall, the panel felt that there was some evidence to support that ESIs reduced the need for opiates with a score of 3/5 (range 1-4), however studies to specifically address this are lacking.
Lumbar Spinal Stenosis
The evidence surrounding the use of ESIs for lumbosacral spinal stenosis is conflicting. A 2014 double-blind, multisite, RCT funded by the AHRQ reported on 400 patients with lumbar central spinal stenosis with moderate to severe pain (>4/10 at baseline) and a disability score of 7 or higher on the RMDQ to receive ESI with glucocorticoids plus lidocaine or lidocaine alone. The patients received 1 or 2 injections and evaluation at 6 weeks from the first injection. At 6 weeks, there was no significant difference between the groups' RMDQ scores (adjusted difference in the average treatment effect between the glucocorticoid–lidocaine group and the lidocaine-alone group, -1.0 points; 95% CI, -2.1 to 0.1; P=0.07) or leg pain scores (adjusted difference in the average treatment effect, -0.2 points; 95% CI, -0.8 to 0.4; P=0.48). They concluded ESI with LA and glucocorticoids offered minimal or no short-term benefit compared to ESI with LA alone.59 This was met by support for surgical management or trial of conservative measures for lumbar spinal stenosis and criticism from members of the interventional pain management community.60 A 2015 systematic review and meta-analysis, including 8 trials on ESIs for spinal stenosis, concludes there were no clear effects of ESIs for spinal stenosis with the strength of evidence rated low to moderate.52 The 2019 Minimally Invasive Spine Treatment (MIST) Guidelines for lumbar spinal stenosis from the minimally invasive spine treatment consensus group evaluated 9 studies (2 RCTs; 7 observational studies, 4 prospective studies and 3 retrospective studies) of minimally invasive spine treatments and reported there is Level I evidence for percutaneous image-guided lumbar decompression as superior to lumbar ESIs. The consensus group developed treatment algorithms to guide management.61
On the contrary, there are several papers that support the use of ESI for spinal stenosis. A 2-year observational study on the use of TFESI for lumbar spinal stenosis followed 68 consecutive patients at 6 weeks, 1 and 2 years, to avoid decompressive surgery. Of these patients, 32% opted for surgery, 24% had repeat injections, and 44% were satisfied with non-surgical management at 2 years. The authors conclude TFESI is a reasonable treatment for lumbar spinal stenosis, provides long-term relief in a high proportion of patients, and can reduce the need for surgical management.62 There are prospective studies supporting this use.63 A 2015, double-blinded RCT compared the effectiveness of lumbar ILESI with LA only to lumbar ILESI with corticosteroids and LA with central spinal stenosis. Assessments to measure success were defined as = 50% pain relief at 3,6,12, 18, and 24 months. They reported 72% (95% CI: 61-83%) of patients receiving lidocaine and corticosteroids and 73% (95% CI: 62-85%) receiving lidocaine alone had significant pain relief at 24 months; concluding ESI with LA with or without steroids were equally effective and beneficial.64 However, overlapping CIs suggest the treatments are equivalent at those timeframes, so it does not provide evidence for long-term improvement. In the 2015 systematic review of 52 RCTs with placebo-control or active-control design looking at the efficacy of epidural injections in managing chronic spinal pain, they concluded Level II evidence for caudal and lumbar interlaminar epidural injections with Level III evidence for lumbar transforaminal epidural injections for lumbar spinal stenosis. The evidence is Level II for cervical spinal stenosis management with an interlaminar approach.54
Even further conflicting papers support a short-term, but not long-term, benefit. A 2017 multicenter, double-blind, RCT compared epidural injections of corticosteroid plus lidocaine versus lidocaine for lumbar central spinal stenosis alone at 16 clinical sites. Participants had the option of blinded crossover after 6 weeks to receive the alternate treatment. Disability was measured using the RMDQ. The authors report at 12 months, both treatment groups maintained initial observed improvements, with no significant differences between groups on the RMDQ (adjusted MD, -0.4; 95% CI, -1.6 to 0.9; P=0.55), leg pain (adjusted MD, 0.1; 95% CI, -0.5 to 0.7; P=0.75), opioid use (corticosteroid plus lidocaine: 41.4% vs lidocaine alone: 36.3%; P=0.41), or spine surgery (corticosteroid plus lidocaine: 16.8% vs lidocaine alone: 11.8%; P=0.22). They did not observe a difference among participants who crossed over at 6 weeks. They conclude for lumbar spinal stenosis symptoms, epidural injections of corticosteroid plus lidocaine offered no benefits from 6 weeks to 12 months beyond that of injections of lidocaine alone in terms of self-reported pain and function or reduction in the use of opioids and spine surgery. However, in patients with improved pain and function 6 weeks after the initial injection, these outcomes were maintained at 12 months. They did not receive additional benefits after 3 weeks based on the injectant type. Additionally, if there was no improvement during the first 6 weeks, repeated injections of either type provided no added long-term benefit.65 A 2017 systematic review and a meta-analysis using GRADE methodology reported on 4 observational studies; 1 RCT concluded the evidence suggests a lack of effectiveness of lumbar ILESI in treating primary axial pain related to spinal stenosis, but does suggest significant short-term improvements in radicular pain with stenosis.17
A 2019 meta-analysis reviewed non-surgical medical treatment in the management of pain due to lumbar disc prolapse. Fifty-eight studies in global effects and 74 studies in pain intensity analysis were included, with 65.5% rated as high risk for bias. They found CESI and TFESI had higher short-term scores at follow-up while biologicals and manipulation had a higher score at long-term follow-up for pain intensity; however, no treatment was found to be superior when comparing multiple outcomes and periods. The TFESI was reported to be an effective short-term treatment. They reported a lack of high-quality studies as a limitation.66,67
There is little evidence showing a benefit for long-term outcomes. A 2020 retrospective cohort of 54/90 patients who underwent TFESI for controlling lumbar spinal stenosis induced radicular pain were interviewed at least 5 years after their initial procedure. Half of the patients were receiving repetitive TFESIs every 2 to 6 months or were taking oral pain medication; 25% had undergone surgical intervention. The authors report poor long-term pain relief with surgery as well and conclude ESIs offer a safe option for controlling radicular pain associate with spinal stenosis.68
A 2020 prospective comparative analysis of cost and quality of life compared lumbar ESIs to conservative management alone at 3 and 6 months. One hundred forty-one patients were included, and the authors concluded at 3 months, ESIs provided similar improvements in quality of life outcomes relative to medical management and at similar costs. At 6 months, neither ESIs nor conservative management provided significant quality of life outcomes.69
In the 2020 systematic review by Smith et al. for lumbar spinal stenosis, using a =50% pain relief criteria, the studies (range) reported success rates of 49% (43–55%) at 1 month, 48% (35–61%) at 3 months, 43% (33–53%) at 6 months, and 59% (45–73%) at 1 year. The author offers low-quality evidence according to GRADE methodology and notes the lack of corroboration from appropriately controlled studies.58
The MIST Guidelines on lumbar spinal stenosis consensus group concludes there is ample evidence to support the use of ESIs in the management of symptomatic lumbar spinal stenosis. Depending on the duration and extent of relief, these options can be repeated or continued to more surgical treatment solutions (grade B, level II-2, consensus strong).61
Overall, the SMEs felt there was evidence to support a benefit of ESIs for spinal stenosis voting 4/5 for lumbar central spinal stenosis (range 2-5), foraminal stenosis (range 1-5) and subarticular stenosis (range 1-5).
Herpes zoster may lead to severe zoster-associated pain (ZAP) and post neuralgia pain (PNP) to the degree of interfering with activities of daily living. Several studies have explored the role of ESIs in the management of ZAP. A 2017 retrospective review of 137 patients with ZAP who underwent TFESI for management of ZAP was included. They compared those who had the TFESI within the first 30 days to those who had the procedure between 30-90 days. They concluded that early application of TFESIs in the acute phase was a useful option to control ZAP-related pain. Those who received the injection earlier had a significantly shorter time to pain relief with a lower incidence of postherpetic neuralgia (PHN).6 This was limited by the retrospective nature of the study, lack of control group, or comparison to other treatment options such as antivirals, topical analgesic, or pharmacological treatment. There are several trials evaluating continuous epidural infusion for the management of ZAP, but no studies that compare infusion to injections.
The SMEs had low confidence voting 2/5 (range 1-3) that there is evidence to support the use of ESIs for PHN and 3/5 (range 1-5) for acute herpes ZAP.
Cervicogenic Headaches and Cervicobrachialgia
The CESI is considered by many interventional pain management specialists to be a reasonable option for patients who have failed conservative treatments for cervicogenic headache. A 2014 review on CESI for cervicogenic headache found no RCTs, and 2 prospective studies, with 36 patients in the studies combined, that support a role of CESI for this indication; however, there is a paucity of literature to support this indication.18
Cervicobrachial is used to describe pain in the neck radiating to the arm due to nerve root compression typically caused by disc herniation or spinal stenosis. A non-controlled randomized trial in 1993 on CESI for cervicobrachialgia randomized 25 patients to CESI and 17 to steroid/lidocaine injection into the posterior neck muscles with 1 to 3 injections at 2-week intervals and followed them for 1 year. One week after the last injection, they found that pain relief was rated as very good and good in 76% of the patients in the CESI group, compared to 35.5% of the patients in the posterior neck muscle injection group.70
The SMEs voted 3/5 (range 1-5) that there is evidence to support the use of ESIs for cervicogenic headaches and 4/5 (range 1-5) for cervicobrachialgia. The SMEs indicated that cervicobrachalgia is caused by disc herniation and spinal stenosis, where ESIs are beneficial.
Non-Specific Back Pain
There is a lack of literature on the use of ESIs for non-specific LBP. Most of the literature for non-specific pain support exercise therapy and multi-modality conservative treatments. A 2017 systematic review reported on 71 publications addressing lumbar ILESIs regardless of etiology. They found low-quality evidence using GRADE methodology and concluded lack of effectiveness of lumbar ILESIs for treatment of primary axial pain regardless of etiology. They report that most studies on radicular pain due to lumbar disc herniation and stenosis reports statistically significant short-term improvement.17
The SMEs voted not confident 1/5 (range 1-2) that there is evidence to support the use of ESIs for non-specific back pain, widespread diffuse back pain, and non-organic back pain.
Post-Surgical Back Pain
A 2018 systematic review and meta-analysis of RCTs comparing outcomes of epidural steroids versus placebo after lumbar discectomy in lumbar disc herniation reported on 12 studies with 1,006 patients (502 in ESI group and 504 in placebo group) and found an unstandardized MD of VAS of back pain at 1 week and 1 month, leg pain at 1 week and 1 month, morphine consumption and hospital stay was - 0.53 (95% CI - 1.42, 0.36) score, - 0.89 (95% CI - 1.36, - 0.42) score, - 0.63 (95% CI - 0.75, - 0.50) score, - 0.47 (95% CI - 0.78, - 0.15) score, - 8.47 (95% CI - 16.16, - 0.78) mg and - 0.89 (95% CI - 1.49, - 0.30) days lower when compared to placebo after lumbar discectomy in patients with lumbar disc herniation. Ten studies compared complications and found no significant difference between the 2 groups (0.92; 95% CI 0.47, 1.83), and reported no significant difference in complications for epidural steroid application after lumbar discectomy in lumbar disc herniation.71
A 2018 case-controlled series of 60 patients by Akbas et al. compared the 3 different approaches (TFESI, CESI, ILESI) for post lumbar surgery syndrome and concluded that VAS, ODQ, and patient satisfaction scores were comparable across the 3 groups at all time intervals (P>0.05), relative to baseline, there were significant decreases in pain relief scores (VAS and ODQ) and functional assessment expressed by patients’ satisfaction across all time intervals and in all 3 groups (P<0.01).72
A 2020 systematic review by Smith et al. states a paucity of literature for TFESI for failed back surgery syndrome.58
The SMEs voted 3/5 (range 1-5) that there is evidence to support the use of ESIs for post-laminectomy pain syndrome.
Cervical Radicular Pain
Intervertebral disc-related pain can be caused by disc degeneration or herniation in the cervical spine region. Cervical disc herniation is a common source of cervical radicular pain. The natural history of this condition is spontaneous resolution over time. The NASS Guidelines from 2011 reported 43% of patients’ symptoms resolve in the first few months of experiencing mild or intermittent symptoms, and 27% continue to have chronic pain.73 Safety issues associated with cervical epidural injections include the risk of spinal cord injury, vascular injury, spinal infarct, and ischemia. Cervical TFESI has been associated with a higher risk of neurovascular complication and possible infarction off the spinal cord, brain stem, cerebrum, or the cerebellum. The interlaminar technique is associated with increased risk of direct trauma to the spinal cord.16
A 2015 systematic review reports on the long-term efficacy of cervical ILESI and TFESI to treat neck pain. Seven manuscripts between 1966-2014 using IPM-QRB methodology were included. There were 4 papers rated high quality, all authored by the same author who was also the lead author on this systematic review. They reported Level II evidence for the efficacy of cervical interlaminar epidural injections with LA with or without steroids, based on at least 1 high-quality relevant RCT in each category for disc herniation, discogenic pain without facet joint pain, central spinal stenosis, and post-surgery syndrome.74 In the 2015 systematic review, also published by this group, on the efficacy of epidural injections for chronic spinal pain, they rate the evidence Level II for long-term management of cervical disc herniation with ILESI, Level II for cervical spinal stenosis management with an interlaminar approach, Level II for post cervical surgery syndrome treated with cervical ILESI, and Level II to III in managing thoracic disc herniation with an interlaminar approach.54 These systematic reviews rated a higher level of evidence than an earlier review on the same subject with overlapping authors but using a different methodology for the assessment of the literature, and the literature was older,75 but similar to the 2021 consensus reports from ASIPP Guidelines for methodology.76
A 2014 multi-centered RCT of 169 patients with cervical radicular pain compared ESIs, conservative treatment, or combination treatment for cervical radicular pain over 6 months. Conservative treatment was pharmacotherapy with gabapentin and/or nortriptyline, and physical therapy (PT), as indicated, and combination therapy was ESI and pharmacotherapy with gabapentin and/or nortriptyline plus PT. At 1-month, arm pain scores were 3.5 (95% CI, 2.8 to 4.2) in the combination group, 4.2 (CI, 2.8 to 4.2) in ESI patients, and 4.3 (CI, 2.8 to 4.2) in individuals treated conservatively (P=0.26). Combination group patients experienced a mean reduction of -3.1 (95% CI, -3.8 to -2.3) in average arm pain at 1 month versus -1.8 (CI, -2.5 to -1.2) in the conservative group and -2.0 (CI, -2.7 to -1.3) in ESI patients (P=0.035). For neck pain, a mean reduction of -2.2 (95% CI, -3.0 to -1.5) was noted in combination patients versus -1.2 (CI, -1.9 to -0.5) in conservative group patients and -1.1 (CI, -1.8 to -0.4) in those who received ESI (P=0.064). Three-month post-treatment, 56.9% of patients treated with combination therapy experienced a positive outcome versus 26.8% in the conservative group and 36.7% in ESI patients (P=0.006). The authors reported a lack of significant difference between the groups, with the combination therapy providing better improvement. These results support an interdisciplinary approach to neck pain may improve outcomes; confirmatory studies are needed.77
A 2014 prospective observational study of 143 patients who received cervical paramidline ILESIs reported initial improvements at 2 weeks in 115 of 143 patients (80.8%). Patients with paresthesia only and no pain showed significantly fewer improvements after ESIs (11/19, 57.9%) than patients with pain (104/124, 83.9%) (p=0.013). The authors conclude that fluoroscopic paramidline ILESIs can effectively manage cervical radiculopathy, irrespective of the cause or zone of nerve root compression, and patients with paresthesia only experienced fewer improvements.78
A 2018 randomized trial compared cervical interlaminar epidural injections in post-surgery syndrome with LA with steroids to LA alone using a = 50% improvement on NRS and functional status improvement using the Neck Disability Index (NDI). The authors report similar improvements in both groups, with 69% of patients receiving LA and 71% of subjects receiving LA and steroids showing significant improvement at the end of 2 years. They report an average number of procedures of 5 to 6 with an approximate 12 weeks of significant improvement per procedure. Limitations of the study were multiple co-interventions, and variability in the frequency and number of interventions, and lack of true control group.79
A 2020 systematic review and meta-analysis on the effectiveness of fluoroscopic guided cervical TFESI for the treatment of radicular pain included 6 randomized and 11 nonrandomized comparative studies with pain improvement at least 50% from baseline. Using GRADE methodology, they reported approximately 50% of patients experienced = 50% pain reduction at short- and intermediate-term follow-up. However, they acknowledged the literature was very low quality according to GRADE criteria and the lack of existing studies with adequate control groups.80
A 2020 systematic review and meta-analysis on the effectiveness of epidural injections with LA with or without steroids for management of chronic neck pain concluded that similar pain relief and function were achieved between the 2 groups. This was limited by the lack of RCTs, lack of heterogenicity in the included studies with only 1 group publishing on this topic, and lack of control groups within the RCTs.81
A 2020 retrospective report on the proportion of patients requiring surgery after CESIs reported 11.2% of patients underwent surgery within 6 months of CESI, 14.5% by 1 year, and 22.3% by 5 years. Patients with stenosis and herniation were more likely to undergo surgery than those with radiculopathy. They report repeat injections reduced risk for subsequent surgery with a mean of 1.7 to 5.5 injections in follow-up over 2 years. The study is limited by its retrospective nature, varying numbers of repeat injections and duration of time between injections, and lack of specific indications for injections or subsequent surgery or surgical candidacy. It is difficult to conclude that the steroid injections reduce the risk for subsequent surgery as the natural disease course points towards improvement alone.82
Multiple papers review the safety concerns surrounding cervical ESIs. The risk of spinal cord injury increases as the epidural space narrows moving cranially, so many authors and recommendations conclude that cervical injections should be restricted to C6-C7 or C7-T1. Additional safety measures include a review of pre-procedural imaging and avoiding particulate steroids in cervical injections.1,14 A report addressing key safety issues states, “overall cervical and thoracic transforaminal epidurals constitute 2.4% of all epidural injections and less than 5% of all transforaminal epidural injections; however, they contribute to over 99% of the complications related to intra-arterial injection of particulate steroids” and recommends against the use of cervical TFESI injections.32
The SME panel was asked their confidence for evidence to support cervical ESIs and they voted 4/5 (range 1-5).
Thoracic Epidural Injections
A 2012 systematic review of thoracic ILESI used the United States Preventive Task Force (USPTF) rating of evidence methodology and reviewed 2 studies, 1 RCT and 1 observational study. The review concluded the evidence for thoracic epidural injection treating chronic thoracic pain is considered fair and limited for post-thoracotomy pain.83 The USPTF grading has been revised since 2012, but at that time fair was considered sufficient evidence to determine the effect on health outcome, but the strength of evidence was limited, and limited/poor was insufficient evidence to assess.
A 2014 RCT without a placebo group with 110 patients was randomized to 2 groups; 1 group received LA (n=55) and the other group LA with steroids (n=55). They reported most patients had 5 to 6 procedures over 2 years and reported improvement in 71% in LA alone and 80% in LA plus steroids. Limitations included lack of placebo group and variability in timing and number of procedures.84
The SME panel voted 4/5 (range 2-5) that the evidence supports the benefits of ESIs outweighs risk for thoracic radicular pain. The panel discussed that thoracic radicular pain is far less common and has not been well studied.
Frequency, Laterality and Multiple Level Injections
Existing evidence and societal guidance support conservative management for acute back pain. There is little data regarding the frequency of ESIs for patients who are candidates for intervention. A 2018 prospective cohort study of 102 patients who had ESI for radicular symptoms related to disc herniation (n=57) or cervical disc herniation (n=45) and had a second injection administered for persistent pain were followed for 1 year. Seventeen patients had a second injection for lumbar herniation (29.8%) and 7 for cervical herniation (15.6%) at a mean of 65.3 days (standard deviation [SD] 46.5) and 47 days (SD 37.2), respectively. All but 1 patient who had surgery responded satisfactorily with a mean VAS for leg pain of 8.8 mm (SD 10.3) and a mean VAS for arm pain of 6.3 mm (SD 9) 1 year after the second injection, respectively. They concluded repeat injections were acceptable treatment in symptomatic patients without satisfactory relief after the first injection.85 Limitations of this study conducted in Europe were different standards for ESI and repeat injection, risk of recall bias, and varying times to the second injection. In a study by the same authors, 1,002 consecutive patients were prospectively assessed on short-term response to ESIs for various indications. They determined 10-day outcomes of the second injection were as good as the outcome of the primary injection for a lumbar herniation: (72.2% better, 24.7% the same, and 3.1% worse following 295 primary injections compared with 70.1% better, 24.8% the same, and 5.1% worse after 117 second injections; p=0.593). For cervical herniation, the improvement was reported in 63.1%, the same in 28.6%, and worse in 8.3% after 84 primary injections compared with 52.2% better, 34.8% the same, and 13% worse after 23 second injections; p = 0.602). This study was also a European population and did not use validated outcome scores.86 Other studies have also reported improvements with second injections, but high variability in timing and number of injections and inclusion of other spinal procedures challenge interpretation. A 2016 prospective observational study included 184 patients who underwent TFESI for axial neck or radicular arm pain due to herniated disc or spinal stenosis. A group of partial responders (n=108) was scheduled to undergo repeat injection at two-three weeks. The other group (n=76) of partial responders received repeat injections for aggravation of pain. They concluded the scheduled repeat injection group showed a significantly longer time to reinjection and improved clinical benefits.67 This was limited by a lack of a control group, lack of standardization in the total number of injections, and unclear randomization.
A 2014 retrospective observational study evaluated 933 patients with a total of 2,087 TFESI. Injections were repeated at intervals of at least 2 weeks for a duration of up to 1 year from index injection. Of the total patients in the database (933/4,161) 22.4% received repeat injections during the study period, and the cohort showed a significant decrease in improvement with subsequent injections (p=0.0311). This study is limited by retrospective design and high variability in the duration between injections and steroid preparations. They found most patients did not require repeat injections for treatment of lumbar radicular pain but did find those with an incomplete initial response who received early repeat injections within 3 months can achieve a cumulative benefit.87
On the contrary, several authors reported no difference in long-term outcomes in patients receiving a second injection nor a reduction in the need for surgery.88,89 The 2005 Wessex Epidural Steroids Trial (WEST) study was a multicenter, double-blinded randomized placebo-controlled trial with 228 patients with sciatica. The patients were randomized to either 3 lumbar ESIs with steroid or placebo (saline) at intervals of 3 weeks. At 3 weeks, they discovered that the ESI group had a transient benefit over the placebo group (75% improvement in ODQ, 12.5 vs. 3.7%; number needed to treat, 11.4). No benefit was demonstrated from 6 to 52 weeks. They concluded that ESIs did not improve physical function, hasten return to work or reduce the need for surgery and did not find a benefit of repeated ESIs over single injection. They advocate a multidisciplinary approach for management.90 Limitations were the use of 75% cut off for improvement while most studies use 50%, and assessment of disability but not pain assessment.
A 2013 systematic analysis identified 9 studies with = 50% pain relief after TFESIs for radicular pain after a single injection and reported 94% required a single injection and 4% required a second injection to achieve this level of pain relief. The use of 3 or 4 injections was rare.7
A 2014 retrospective observational study investigated if repeat lumbar TFESIs resulted in pain relief which has waned since index injection, and if a cumulative benefit could be achieved by repeat injections within 3 months of the index injection. Four thousand one hundred sixty-one patients received single-level TFESIs for radicular pain with or without radiculopathy. Between 2 weeks and 1 year from index injection, 933 subjects (22.4%) had repeat injection within 1 year. The data indicates that 18% of subjects had 2 injections, 3.6% had 3 or more injections. They concluded there was cumulative benefit of repeat injections within 3 months of the index injection.87 Limitations of the study included the retrospective design, a high number of patients lost to follow-up and no standardization in terms of timing of the repeat injection(s).
Using quality assurance databases from radiology and 2 medicine and rehabilitation (PM&R) practices, 16,638 consecutive procedures in all spine segments (14,956 TFESI; 1,682 ILESI) were evaluated for major and minor adverse safety outcomes. The most frequent complication was a vasovagal reaction in 1.2% and dural puncture in 0.06%. They reported no major adverse reactions, no long-term sequelae from any intermediate or delayed minor adverse events. They concluded both TFESI and ILESI are safely performed with low immediate and delayed adverse event rates when informed by evidence-based procedural guidelines. Of note, multiple patients in this population had bilateral procedures or 2-level injections without additional safety concerns.34
A 2016 study was conducted for 184 patients who underwent TFESI for axial neck and radicular arm pain with single ESI. Group A (n=108) was comprised of partial responders defined as a NRS = 3 and received scheduled repeat injections at 2 to 3 weeks after the first injection. Group B (n=76) was comprised of partial responders who did not receive a repeat injection but received intermittent injections for aggravation of pain. Total number of injections within 1 year and NRS score <3 were recorded, and they concluded Group A showed a significantly longer time to re-injection and a longer time between injections. The average number of injections was 2.48 in Group A, with 4 patients receiving more than 3 injections and 2.98 in Group B with 16 patients receiving >3 injections.91 This study is limited by retrospective design and selection bias as patients who did not follow-up for a year or choose surgery were excluded.
SMEs were asked about their confidence that the evidence demonstrates ESIs provided relief for a minimum of 6 weeks after injection with score of 4/5 (range 1-5).
- The panel was split regarding repeating ESIs if the initial treatment did not result in substantial pain relief with the majority feeling that repeating in a different level would be appropriate (3/5; range 1-5), but not at the same level that did not achieve improvement with ESI.
- When asked if there is literature on a safe or harmful number of ESIs per year 7/11 SMEs voted no. There was not a consensus among the panel in terms of the number of ESIs that would be considered safe or harmful in a 12-month time period with a range of 3-4 in 12 months suggested.
- In response to the question concerning the literature providing input on the safety of multiple levels of ESIs performed in the same session, 6 out of 11 voted no. When asked the number of levels they felt would be safe the range was 1 to 3 per session.
- Half (5/10) did not feel the literature provides input on a safe duration of time between the administration of separate ESIs in the same spinal region and the recommended duration ranged from 2 weeks to 3 months. One SME stated, “the need for individual assessment and response to prior injections as well as goals of treatment management versus resolution are factors that can influence this decision”.
- SMEs were asked about their level of confidence in the evidence to support repeat epidural injections for long-term management of chronic back pain >6 months and there was not a consensus. Comments included lack of supporting evidence for ESIs for chronic pain management. This must be balanced against patients who do have good results with injections and wish to avoid surgery when clinical benefit is evident.
ESIs usually contain LA and glucocorticoids. The optimal combination and dosing have not been determined. The role of epidural nonsteroidal injections remains controversial. A 2013 systematic review and meta-analysis looked at the control arm of 43 studies which included saline, LA, and injections into muscle or ligament (sham). In the indirect comparison, they concluded E epidural non-steroidal injections achieved positive outcomes (RR, 2.17; 95% CI, 1.87-2.53) and provided greater pain reduction scores (MD, -0.15, 95% CI, -0.55 to 0.25). Indirect comparison does not qualify as the same level of evidence as a randomized comparison, and the study was limited by only 1 included study being ranked as high quality using GRADE methodology and inadequate numbers to detect in effect by size, but suggest that the nonsteroidal injections were not entirely placebo.92 A 2015 systematic review of RCTs using Cochrane review criteria and the ASIPP criteria for assessing interventional techniques looked at 31 trials seeking evidence on LAs, saline, steroids, and other solutions. They conclude equal efficacy for LA with steroids and LA alone in multiple spinal conditions. For disc herniation, they report superiority of LA with steroids over LA alone.93 A 2021 systematic review of RCTs by Manchikanti et al. compared sodium chloride solution alone, steroids alone, or sodium chloride solution with steroids in managing spine pain secondary to disc herniation or spinal stenosis. The authors reported utilizing a single-arm analysis that both epidural saline and epidural steroids with saline were effective in reducing 20% of pain, however, only reducing disability score by 10 to 20%.94 Several 2020 systematic reviews and meta-analysis looking at the difference in efficacy between lidocaine alone versus lidocaine and steroids in the management of lumbar disc herniation or spinal stenosis concluded there were similar effects associated with lidocaine alone or in combination with steroids.95,96
This study contradicts multiple other studies that showed steroids were superior to saline or other placebos. The Friedly et al. multicenter, double-blinded RCT compared epidural injections of corticosteroids plus lidocaine versus lidocaine alone in 400 patients with confirmed spinal stenosis. In this study, patients had the option of blinded crossover after 6 weeks to receive the alternate treatment. However, fewer participants randomized to corticosteroids plus lidocaine (30%, n=60) versus lidocaine alone (45%, n=90) crossed over in 6 weeks (p=0.03) and 93% indicating a lack of effectiveness as a reason to cross-over. Using an intention to treat (ITT) analyses, a small, but consistent difference favoring steroids plus LA over LA alone was observed. At 12 months there was no significant difference between the group. This study did not show repeat injections of either type offering additional benefit if injections in the first 6 weeks did not improve pain.65 A systematic review by Bicket et al. evaluating control injections in RCTs reports that ESIs may provide a benefit compared to non-ESIs while acknowledging this was based on few, low-quality studies directly comparing controlled treatment and short-term outcomes. However, this review with 3,651 patients from 43 studies represents the largest analysis comparing ESI with a steroid to a nonsteroidal alternative including, LA alone, etanercept, saline, intramuscular or ligamentous injections, and dry needling. The authors concluded the benefit from ESIs was limited but suggested it may not constitute a placebo effect.92
While systematic reviews and meta-analysis have suggested a role in non-steroidal injections, these studies are limited because they rely on previously conducted RCTs where the research question was not specific to determining the effectiveness of these nonsteroid injections. The data set utilized is subject to significant heterogenicity (I2 >50-99%), variability of the patient population, and small sample sizes. Studies dedicated to the investigation of the nonsteroid injections are necessary as well as studies with longer-term follow-up periods to understand if there is a role for nonsteroidal injections in lieu of ESIs. There is not enough evidence to be confident that nonsteroid injections are equally effective to steroid injections based on the current body of literature.
A 2016 study reported improvement in pain with particulate compared to non-particulate steroids.97 Spinal cord ischemia and posterior circulation infarction have been reported after cervical ESIs. Concern arose that this risk was greater for particulate compared to non-particulate steroids prompting further investigations and recommendations. In 2011 the FDA required a label change for triamcinolone stating it should not be used for ESI.98 A 2017 systematic review and meta-analysis comparing particulate steroids to non-particulate counterparts conclude that particulate steroids are not better in relieving pain compared to their non-particulate counterparts but may offer an improved safety profile. They conclude with the recommendation to consider switching to non-particulate steroids.99 Another 2017 systematic review by Mehta et al. agreed concluding, no benefit to particulate steroids and recommending non-particulate steroids with performing cervical TFESI, with Grade of Recommendation: B. For lumbar TFESI they state particulate vs. non-particulate as equivocal with Grade of Recommendation: B for pain reduction and C for function with an overall recommendation for non-particulate steroids for lumbar TFESI.100
Steroid dosing, dilution, and additives are not standardized and are another area of controversy. There is evidence that different dilutions such as sodium chloride and local anesthetic can alter the steroid's particle size and distribution. There is concern about the toxicity of additives (such as benzyl alcohol and polyethylene glycol). Based on the potential risk associated with systemic corticosteroid absorption the WIP Benelux Workgroup recommends using the lowest effective dosing, which amounts to 40 mg for methylprednisolone acetate (MPA), 10-20 mg for triamcinolone acetate, and 10 mg (10 mg/mL) for dexamethasone phosphate. They recommend limiting the cervical interlaminar and lumbar transforaminal volume to 4 mL and inject slowly.16
Etanercept is a soluble p75 tumor necrosis factor fusion protein administered subcutaneously for inflammatory arthritis and other rheumatological conditions. While it has been explored in several studies with preliminary positive results, it is not FDA approved for injection into the epidural space, and therefore is not considered medically reasonable and necessary. Additional agents such as platelet-rich plasma, amniotic fluid, gabapentin, and others have been investigated, but there is not sufficient evidence to support use and are considered investigational.
Since injectants may have a bilateral effect or spread to adjacent levels, diagnostic interlaminar or caudal ESIs are seldom used. Diagnostic TFESIs are sometimes used to determine the level of radicular nerve root pain to differentiate radicular from non-radicular pain, to evaluate a discrepancy between image studies and clinical findings, to identify the source of pain in the prevalence of multi-level nerve root compression, and to help identify the level of pathology at a previous operative site. Selective spinal nerve blocks may also be used to identify the source of pain. The cumulative steroid dose and long-term safety has not been studied in the setting of multiple injections administered in the same session.
The SMEs were split in terms of multiple procedures during a single session, with half of the experts (6/11) voting that evidence supported the administration of ESIs at the same time as other interventional procedures. However, there was no additional supporting literature provided, and others brought up the concern of lack of diagnostic specificity when multiple procedures are performed in the same session.
North American Spine Society (NASS)
The 2020 NASS Evidence-Based Clinical Guidelines for Multidisciplinary Spine Care: Diagnosis & Treatment of Low Back Pain concludes that there is insufficient evidence to make a recommendation for or against the use of caudal or interlaminar ESIs in patients with LBP with Grade of Recommendation: I (good evidence for or against recommending intervention).15
The 2020 NASS Epidural Steroid Injections & Selective Spinal Blocks offers evidence-based coverage recommendations when possible and in the absence of strict evidence-based criteria recommendations are based on multidisciplinary experience and expertise of the authors. Evidence in this coverage policy are not graded11:
- ESIs can be performed to provide relief of radicular or referred pain when 2/4 are present: pain is severe enough to cause a degree of functional and/or vocational impairment or disability; pain duration of at least 4 weeks, and/or inability to tolerate or failure to respond to 4 weeks of noninvasive care, objective findings of radiculopathy or sclerotomal referred pain pattern are present and documented on examination, and advanced imaging (CT or MRI) demonstrates a correlative region of nerve involvement. Criteria for exemption of 4-week delay is moderate to severe pain with functional loss at work and/or home, pain unresponsive to outpatient medical management, inability to tolerate non-surgical non-injection care to a coexisting medical condition or severe pain or prior successful injection therapy for same condition achieving greater than 50% relief.
- ESIs are contraindicated for axial or nonspecific pain without radiating pain, most cancer or strong clinical suspicion of cancer, infection and compression lesions of the spinal cord, conus medullaris or cauda equina. Relative contraindications to the performance of ESIs may include coexisting medical conditions such as uncontrolled bleeding disorders, poorly controlled diabetes (if corticosteroids are going to be used), immune system impairment, history of severe allergic reaction to components, etc. In these situations, the risk/benefits of the procedure should be considered in the medical decision-making process.
- Procedural based recommendations include the use of image guidance, ILESIs should not be performed above C7, injection of contrast media under real time fluoroscopy or digital subtraction imaging, use of a non-particulate steroid with the exception of thoracic and cervical spine, extension tubing for provider safety, and appropriate provider training.
- Suggested frequency of ESIs were based on author's expert opinion included no more than 4 ESIs and/or suprascapular nerve blocks (SSNBs) should be performed in a 6-month period of time, no more than 6 ESIs and/or SSNBs should be performed in a 12-month period of time regardless of the number of levels, no more than 2 TFESIs should be performed at a single setting (e.g., single level bilaterally or 2 levels), and for CESIs or ILESIs, only 1 level per session may be performed and NOT in conjunction with a TFESI injection.
- Local anesthesia is sufficient for most cases. Occasional minimally to moderate conscious sedation is appropriate and if MAC is utilized the need for such stations should be clearly documented in the medical record.
- They report that studies found that most patients who respond to ESIs do so with 3 or less injections for a specific episode of back and radicular pain. They provide non-evidence-based criteria for instances when they consider it medically appropriate to perform more than 3 injections. These circumstances include performance of SSNBs for surgical planning after failed ESIs, the presence of new injuries after resolution of prior conditions or after interval surgery since prior ESIs, prior injections were done without fluoroscopy or were inaccurately placed, exacerbation of symptoms that responded well to prior ESIs, and patients who responded well to prior ESIs that are not surgical candidates due to comorbid medical conditions.
US Department of Veterans Affairs (VA)/US Department of Defense (DoD)19
The 2017 VA/DoD Clinical Practice Guideline for Diagnosis and Treatment of Low Back Pain utilized GRADE methodology. The original guidelines were developed in 2007. In 2017 the literature was reviewed and recommendations were added (New-added).
- For the long-term reduction of radicular LBP, non-radicular LBP, or spinal stenosis, we recommend against offering spinal ESIs. (Strong against | Reviewed, New-added)
- For the very short-term effect (less than or equal to 2 weeks) of reduction of radicular LBP, we suggest offering ESI. (Weak for | Reviewed, New-added)
- For the treatment of LBP, we suggest against offering intra-articular facet joint steroid injections. (Weak against | Reviewed, New-added)
American Society of Interventional Pain Physicians (ASIPP)
The ASIPP released updated 2021 Epidural Interventions in the Management of Chronic Spinal Pain: ASIPP Comprehensive Evidence-Based Guidelines.76 This extensive evidence synthesis utilized a modified approach to the grading of evidence101 and a recommendations grade table published in the British Medical Journal in 2001. This grading system utilizes the type of study and the number of studies. In contrast, the GRADE system, defined in definitions above and commonly used in guideline development, focuses on the quality of the studies with up or downgrading based on methodology. The quality of each article was assessed by IPM-QRB. Recommendations were made based on the National Guideline Clearing Extent Adherence to Trustworthy Standards instrument. While many of the recommendations were moderate to strong and based on Level I-II evidence paucity of literature with many of these areas is based on 1 study.
- Disc Herniation: Based on relevant, high-quality fluoroscopically guided epidural injections, with or without steroids, and results of previous systematic reviews, the evidence is Level I for CESIs, lumbar ILESIs, and lumbar TFESIs and cervical ILESIs with a strong recommendation for long-term effectiveness.
- For thoracic disc herniation, based on 1 relevant, high-quality RCT of thoracic epidural with fluoroscopic guidance, with or without steroids, the evidence is Level II with moderate to strong recommendation for long-term effectiveness.
- For spinal stenosis, the evidence is based on 1 high-quality RCT in each category. The evidence is level II-III for fluoroscopically guided caudal epidural injections with moderate-to-strong recommendation and Level II for fluoroscopically guided lumbar and cervical interlaminar epidural injections with moderate-to-strong recommendation for long-term effectiveness.
- The evidence for lumber TFESIs is Level III-IV with moderate recommendation with fluoroscopically guided lumbar TFESIs for long-term improvement.
- For axial discogenic pain, the evidence without facet joint pain or sacroiliac joint pain in the lumbar, and cervical spine with fluoroscopically guided caudal, lumbar and cervical ILESIs, based on 1 relevant high-quality RCT in each category is Level II with moderate to strong recommendation for long-term improvement, with or without steroids.
- The evidence for lumbar and cervical post-surgical syndrome based on 1 relevant high-quality RCT with fluoroscopic guidance for caudal and cervical ILESIs, with or without steroids, is Level II with moderate to strong recommendation for long-term improvement.
The ACR, the American Society of Neuroradiology (ASNR), the American Society of Spine Radiology (ASSR), the Society of Interventional Radiology (SIR), and the Society of NeuroInterventional Surgery (SNIS)
ACR–ASNR–ASSR–SIR–SNIS Practice Parameter for the Performance of Image-Guided Epidural Steroid Injection (2019) is a Collaborative guideline.
- When an ESI is performed, success is defined as the achievement of significant pain relief, reduced disability, and/or improved quality of life. These should be measured by at least 1 of the relevant and validated measurement tools, such as the 10-point numerical pain rating scale score or a VAS score (Roland-Morris Back Pain score, ODI, The Short Form (36) Health Survey, or similar outcome tool to measure pain, disability, and/or quality of life). It is generally accepted that a minimum of 20% change in pain scores is clinically meaningful, based upon previous trials and FDA requirements.
- They list the following absolute contraindications:
- Coagulopathy not correctible
- Concurrent systemic infection
- Infectious spondylitis
- Acute spinal cord compression
- Myelopathy or cauda equina syndrome
- Inability to obtain informed consent
- Infection at the skin puncture site
- They list the following relative contraindications:
- Uncorrected anticoagulation therapy – ILESIs and TFESIs are considered intermediate-risk procedures with a moderate risk of bleeding.
- Hypersensitivity to administered agents – Allergy to contrast may be treated with premedication with antihistamine agents or an alternative approach (such as using CT guidance with air as the contrast medium may be considered).
- Pregnancy – Although such interventions may be performed without image guidance in pregnant patients, there is a 30% rate of incorrect placement . Other options include MRI-guided injections and ultrasound-guided injections as image-guided procedures have a significantly greater margin of safety and should be utilized when feasible.
- Hepatitis – When performing neuraxial blockade in hepatitis C patients, thrombocytopenia must be excluded in order to avoid hematoma formation and its associated neurologic complications.
- Uncontrolled diabetes mellitus- Insulin-dependent diabetics are at risk of elevated blood sugars after steroid injections.
- Congestive heart failure – The steroid may lead to fluid retention.
- Immunosuppressed state- Preprocedural antibiotics may be considered.
- Patient improving on medical and physical therapy.
- Severe spinal canal stenosis
- No response to previous well-performed ESI
- Complication to steroid therapy (Cushing’s etc.)
Agency for Healthcare Research and Quality (AHRQ)
The AHRQ 2015 Technology Assessment for Pain Management Injection Therapies for Low Back Pain2
- The 2015 AHRQ technology assessment played a pivotal role in epidural injection for pain. While this report has not been updated since 2015 due to loss of funding for this project, this report remains significant as it was funded by AHRQ, so there was low-risk of outcome reporting bias, and leads to questions regarding the effectiveness and utility of ESIs for pain. The report includes a systematic review of RCTs of patients with lumbosacral radiculopathy, spinal stenosis, non-radicular back pain, or chronic post-surgical back pain. Quality of evidence was assessed for risk of bias using the Cochrane Back Review Group criteria. Meta-analysis was performed and stratified by time. A total of 79 RCTs of epidurals were included. The results were: “For ESIs versus placebo interventions for radiculopathy, the only statistically significant effects were on mean improvement in pain at immediate-term follow-up (weighted mean difference [WMD] -7.55 on a 0 to 100 scale, 95% CI -11.4 to -3.74) (strength of evidence [SOE]: moderate), the mean improvement in function at immediate-term follow-up when an outlier trial was excluded (standardized mean difference [SMD] -0.33, 95% CI -0.56 to -0.09) (SOE: low), and risk of surgery at short-term follow-up (relative risk [RR] 0.62, 95% CI 0.41 to 0.92) (SOE: low). The magnitude of effects on pain and function was small, did not meet predefined thresholds for minimum clinically important differences, and there were no differences in outcomes at longer-term follow-up. Evidence on effects of different injection techniques, patient characteristics, or comparator interventions estimates was limited and did not show clear effects. Trials of epidural corticosteroid injections for radiculopathy versus non-placebo interventions did not clearly demonstrate effectiveness (SOE: insufficient to low). Evidence was limited for epidural corticosteroid injections versus placebo interventions for spinal stenosis (SOE: low to moderate) or nonradicular back pain (SOE: low), but showed no differences in pain, function, or likelihood of surgery. Serious harm from injections were rare in randomized trials and observational studies, but harm reporting was suboptimal (SOE: low).” The authors concluded that corticosteroid injections for radiculopathy were associated with immediate but short-term benefits and did not reduce the long-term risk of surgery. They also conclude that ESI is not effective for spinal stenosis or non-radicular back pain.
The American College of Physicians (ACP)
- The ACP created the Clinical Practice Guidelines for Noninvasive Treatments for Acute, Subacute, and Chronic Low Back Pain (2017) using the ACP grading system methodology and based on published systematic reviews and a systematic review of RCTs through April 2015 to develop the following recommendations25:
- Recommendation 1: Given that most patients with acute or subacute LBP improve over time regardless of treatment, clinicians and patients should select nonpharmacologic treatment with superficial heat (moderate-quality evidence), massage, acupuncture, or spinal manipulation (low-quality evidence). If pharmacologic treatment is desired, clinicians and patients should select NSAIDs or skeletal muscle relaxants (moderate-quality evidence). (Grade: strong recommendation).
- Recommendation 2: For patients with CLBP, clinicians and patients should initially select nonpharmacologic treatment with exercise, multidisciplinary rehabilitation, acupuncture, mindfulness-based stress reduction (moderate-quality evidence), tai chi, yoga, motor control exercise, progressive relaxation, electromyography biofeedback, low-level laser therapy, operant therapy, cognitive behavioral therapy, or spinal manipulation (low-quality evidence). (Grade: strong recommendation).
This report was updated in 2017 with some changes in recommendations for initial pharmacological management, but other conclusions were relatively unchanged.102
American Society of Anesthesiology (ASA)
- The American Society of Anesthesiologist Task Force developed Practice Guidelines for Chronic Pain Management, which were last updated in 2010. At that time, they stated, “ESI, with or without local anesthetics, may be part of a multimodal treatment regimen to provide pain relief in selected patients with radicular pain or radiculopathy. Shared decision-making regarding epidural steroid injections should include a specific discussion of potential complications, particularly regarding the transforaminal approach. Transforaminal epidural injections should be performed with appropriate image guidance to confirm correct needle position and spread of contrast before injecting a therapeutic substance; image guidance may be considered for interlaminar epidural injections."12
American Academy of Neurology (AAN)
The AAN published a technology assessment and reported in March 2007 and reaffirmed in July 2010 on ESIs to treat radicular lumbosacral pain.12,20 Their recommendations and conclusions include:
- ESIs may result in some improvement in radicular lumbosacral pain when determined between 2 and 6 weeks following the injection, compared to control treatment (Level C, Class I–III evidence). The average magnitude of the effect is small, and the generalizability of the observation is limited by the small number of studies, limited to highly selected patient populations, the few techniques and doses studied, and variable comparison treatments.
- In general, ESIs for radicular lumbosacral pain have shown no impact on average impairment of function, on the need for surgery, or long-term pain relief beyond 3 months. Their routine use for these indications is not recommended (Level B, Class I–III evidence).
- Data on the use of ESIs to treat cervical radicular pain are inadequate to make any recommendation (Level U).
American Association of Neurological Surgeons (AANS) and the Congress of Neurological Surgeons (CNS)
The AANS and the CNS guidelines update for performance of fusion procedures for degenerative disease of the lumbar spine reviewed injection therapies for LBP in Part 13 of their guideline update and reported no new evidence that changes recommendations from their original publication in a 2014 update.103 The guidelines were based on a review of literature through 2011 and focuses on 3 systematic reviews and an updated review of RCTs. They conclude, “the medical literature continues to fail to support using lumbar epidural L injections for long-term relief of chronic back pain without radiculopathy. There is limited support for the use of lumbar epidural injections for short-term relief in selected patients with chronic back pain.”
- Lumbar ESIs: Grade C Lumbar ESIs are an option for the short-term relief of CLBP without radiculopathy in patients with degenerative disease of the lumbar spine (Level III evidence).
- CESIs are an option for decreasing LBP of greater than 6 weeks’ duration, without radiculopathy, in patients with degenerative disease of the lumbar spine (Level III)
Spine Intervention Society (SIS)
The SIS Patient Safety Committee published 2020 guidance on the frequency of ESIs8:
- After an ESI, a period of up to 14 days may be needed to assess the clinical response. Systemic effects on the HPA axis may last 3 weeks or longer. These factors must be considered when determining if or when another ESI is indicated.
- There is no evidence to support the routine performance of a “series” of repeat injections without regard to the clinical response to an initial injection.
- Two separate studies suggest that repeat injections may improve outcomes in patients with a partial response to a first (index) ESI. Repeat ESI at greater than 2 weeks and less than 1 year from the index injection has also been shown to result in a statistically and clinically significant decrease in pain, and patients with acute to subacute symptoms recover all prior benefit with a statistically significant cumulative benefit.
The American Society of Regional Anesthesia and Pain Medicine, the European Society of Regional Anesthesia and Pain Therapy, the American Academy of Pain Medicine, the International Neuromodulation Society, the North American Neuromodulation Society, and the World Institute of Pain provided updated guidelines for Interventional Spine and Pain Procedures in Patients on Antiplatelet and Anticoagulant Medications (Second Edition)104
This updated guideline addresses "Pain Procedures Classification According to the Potential Risk of Serious Bleeding" and indicates:
- Intermediate risk defined as patients with a high risk of bleeding (e.g., old age, history of bleeding tendency, concurrent uses of other anticoagulants/antiplatelets, liver cirrhosis or advanced liver disease, and advanced renal disease) undergoing low- or intermediate-risk procedures should be treated as intermediate or high risk, respectively includes: ILESIs (cervical, lumbar, sacral and thoracic) TFESIs (cervical, lumbar, sacral and thoracic).
- Recommends review of radiological imaging prior to performance of the procedure to assess for central and foraminal stenosis, disk herniations that compromise canal diameter, ligamentum flavum hypertrophy, epidural fibrosis, and previous surgical scarring, which can alter the level of procedural difficulty.
- Provides summary and guidance for multiple drugs including aspirin, NSAIDs, coumadin, heparin, low molecular weight heparin, fibrinolytic agents, etc.
The National Institute for Health and Care Excellence (NICE)28 provided guidelines for “low back pain and sciatica.” Assessment and management recommendations include:
- Support use of epidural injections for acute and severe sciatica
- Recommends imaging if it may change future management, including consideration for epidural or spinal surgery.