Nyucardiacsurgery.com

Techniques in Regional Anesthesia and Pain Management (2009) 13, 266-271 Epidural steroid injections: An update on mechanisms of
injury and safety
Christopher Gharibo, MD, Caroline Koo, MD, Jennifer Chung, MD, Alex Moroz, MD
From the Departments of Anesthesiology and Rehabilitation Medicine, New York University Medical Center, New YorkUniversity School of Medicine, New York, New York. KEYWORDS:
Epidural steroid injections (ESIs) are the most commonly performed intervention in the United States to manage chronic and subacute low back and neck pain with radiculopathy. ESIs have been used for decades for the treatment of discogenic and osteoarthritic radicular conditions originating from the cervical, thoracic, and lumbar spine, as well as spondylosis, nonspecific radiculitis, and spinal stenosis.
With the ever-increasing use of epidural steroids, there has been a disproportionate increase in popularity of transforaminal ESIs in particular. Since 2002, there has been a growing body of largelytransforaminal epidural steroid case report literature that describes paralysis, stroke, and death thatimmediately follows the performance of these procedures. These complications are thought to berelated to a combination of factors, which may include the technique used, underlying pathophysiologythat is being treated, anatomical variations in the blood supply, as well as the specific injectate used.
This article discusses the pathogenesis of these complications and puts the role of steroids in their causation into perspective.
2009 Elsevier Inc. All rights reserved.
Background
The therapeutic effects of epidural steroids are a combi- nation of the primary physiological changes that result from Epidural steroid injections (ESIs) are the most commonly the procedure and the secondary results arising from the performed intervention in the United States to manage enhanced pain control that allow integration of other ther- chronic and subacute low back pain. ESIs have been used apeutic modalities. A spectrum of direct effects may provide for decades for the treatment of discogenic and osteoar- therapeutic benefit. For example, most investigators believe thritic radicular pain originating from the cervical, tho- the predominant benefit of epidural steroids is their anti- racic, and lumbar spine, as well as for spondylosis, non- inflammatory effect, although their neurolytic effect on un- specific radiculitis, and spinal Other reported myelinated C-fibers has also been demonstrated.Various uses of epidural steroids include the treatment of pain publications between controlled and uncontrolled trials de- from post-herpetic or post-traumatic neuralgia, muscle scribe success rates varying between 19% and contraction headaches, or subacute inflammatory spinal The epidural space entry can be accomplished via inter- pain syndrome that has not responded to more conserva- laminar, transforaminal, or caudal approaches. Interlaminar and transforaminal epidural steroids in the lumbar and cer-vical spine are the most commonly performed types of ESIs.
The type and amount of steroid to inject as well as thespecific mixture that may contain normal saline or local Address reprint requests and correspondence: Christopher Gharibo,
anesthetic has always been based largely on personal pref- MD, 317 East 34th Street, Suite 902, New York, NY 10016.
erences rather than published evidence. There is no consen- 1084-208X/$ -see front matter 2009 Elsevier Inc. All rights reserved.
doi:10.1053/j.trap.2009.06.025 sus on the optimum steroid type or standard steroid dose in country, the evidence for transforaminal efficacy and supe- epidural injections (e.g., Depo-Medrol, Aristocort, Kenalog riority over interlaminar ESI is Thus, any discus- sion of risk versus benefit of interlaminar versus transfo- The interlaminar approach is the most commonly used raminal epidural steroids must weigh the growing body of approach, as its performance is well-known by the greatest transforaminal complication literature against the lack of number of physicians. The interlaminar ESI has the advan- prospective, randomized literature evidence of superior ef- tage of providing multilevel and bilateral, albeit dorsal and ficacy when compared with the interlaminar approach.
not transforaminal, spread of the injectate in the epiduralspace that may make the interlaminar technique the morelogical approach in a patient with multilevel spinal disease.
The interlaminar ESI, although performed with or without Case reports and the resultant steroid
fluoroscopic guidance, is largely considered a safe and ef-fective procedure.
controversy
An alternative approach to the epidural space is the In essentially all reported cases, the common thread is that transforaminal injection. In contrast to the interlaminar the procedural physicians had longstanding experience in technique, the transforaminal approach always uses fluoros- performing spinal steroid injections. Furthermore, despite a copy to guide a needle adjacent to or into the intervertebral high level of technical competence, accepted technique, and foramen on one side. The steroid solution is injected directly proper injectate confirmation (supported, for example, by onto the dural sleeve and spreads transforaminally into the biplanar fluoroscopy, negative repeat aspirations, and injec- epidural space, thus covering a greater area of nerve root tion of myelographic contrast), sudden or delayed-onset inflammation than can be achieved by the interlaminar ap- neurological complications occurred that have resulted in proach. In addition, the medication is deposited more an- spinal cord infarctions, paraplegia, tetraplegia, brain hernia- terolaterally in the epidural space, close to the area of greatest inflammation and pain production with, for ex- Most of these cases occurred after cervical transforami- ample, more common posterolateral herniations. A trans- nal ESIs, although similar adverse outcomes have been foraminal peridural nerve root sleeve and central canal described after lumbar transforaminal ESIs in operated epidural steroid spread has the potential to treat the inflammation at the injected nerve roots and the dorsal root For example, prompt onset of quadriplegia after an ap- ganglions, anterolateral dura, posterior longitudinal liga- parently unremarkable selective cervical transforaminal ESI ment, and at the nerve supply of the intervertebral disc.
has been reported. This was due to a massive cerebellar and Therefore, relief of axial and/or radicular pain can occur.
occipital cortex infarction that was followed by brainstem This arguably technically improved injection at the site herniation and death. The potential role of particulate cor- of the lesion and the improved spread of the steroid may ticosteroids and its potential for microvascular embolization therefore be more effective. The transforaminal epidural was entertained. Because a radicular artery embolization steroids may also have utility in diagnosis of clinically would not result in the type of infarction that has occurred significant foraminal stenosis and prognosticating surgical in this case, a vertebral artery or terminal branch injury or With the ever-increasing use of epidural steroids, there Several cases of paraplegia after lumbar transforaminal has been an increase in popularity of transforaminal ESIs ESIs in operated lumbar spines have been reported by in particular. There are several reasons for the increased Houten and Errico. The proposed mechanism was vascular popularity of transforaminal epidural steroids. Although randomized, prospective, comparative studies looking at Despite using standard of care technique, immediate ma- outcomes between fluoroscopic interlaminar and transfo- jor complications still occur. Various conclusions have been raminal epidural steroids are lacking, there is a general reached by authors of these and other case reports that opinion that a transforaminal epidural steroid provides su- ultimately implicate nerve ischemia and infarction as the perior steroid spread in terms of laterality, anterior epidural final common pathway that resulted in the reported neural injury and, in some cases, brain edema, herniation, and Since 2002, there have been a growing number of trans- foraminal epidural steroid case reports that describe paral- The exact mechanism or combination of mechanisms ysis, stroke, and even death that follow these procedures.
that may contribute to such neural injury may be due to one They are thought to be related to a combination of factors that may include the approach used, underlying pathophys-iology that is being treated, anatomical variation of the 1. Sustained compressive effect of the injectate that ex- blood supply, as well as the specific injectate ceeds the local arterial pressure or neural perfusion pres- Prospective, randomized trials comparing outcomes after sure in the area injected producing neural ischemia.
fluoroscopic interlaminar and transforaminal ESIs are lack- 2. Mechanical needle injury to the vasculature that disrupts ing. Despite abundant anecdotal experience around the Techniques in Regional Anesthesia and Pain Management, Vol 13, No 4, October 2009 3. Inflammatory arterial vascular irritability that predis- used during injection so as not to miss simultaneous epi- poses the local vasculature to vasospasm from an ad- vancing needle or from the mechanical effect of the Any performance of transforaminal ESI potentially puts the radicular artery at risk of puncture, vasospasm,disruption, or occlusion. This fact is made more problem- A comparative model to this theory could be Prinzmetal’s atic by the variations in the location of the radicular artery angina (vasospasm of the coronary circulation). Such vaso- within the cervical or lumbar intervertebral foramina.
spasm can produce sufficient ischemia to produce myocar- As documented by multiple investigators, there is wide dial infarction, and by the same mechanism, an inflamed anatomical variation of the vascular presence within the epidural space and vasculature can be provoked into vaso- In essence, no absolute, valid, and consistent spasm by an advancing needle or injectate producing in- anatomical advice can be given with respect to needle place- farction in the more susceptible noncollateralized neural ment, and larger anatomical studies are needed to ascertain the exact extent of the anatomical vascular variability within 4. Undetected embolization of a particulate steroid intra- the neural foramina. Due to this lack of anatomical consis- tency, there is little cogent information regarding level of 5. Previous spine surgery appears to be an independent risk entry, extent of entry into the foramen, and vertical location factor for spinal cord Almost all the cases of lumbar transforaminal epidural steroids resulting in spi-nal cord infarction were patients who had prior lumbarspine surgery. This brings up the possibility of alteredanatomy, vascular relocation, neovascularization, and scar tissue vascularization that anastomoses with theexisting spinal or radicular blood supply.
The spinal cord receives its vascular supply from one ante- 6. Although least likely, an intraosseous injection (due to rior spinal artery and two posterior spinal arteries.
osteopenia) with a particulate steroid has been demon- A single anterior spinal artery feeds the anterior two- strated to travel to the inferior vena cava. It is question- thirds of the spinal cord along its entire length and receives able whether such communication can occur with the its supply near the cervicomedullary junction from the ver- arterial circulation leading to cerebrovascular or spinal tebral arteries via two anterior spinal branches. The spinal embolization or whether venous embolism of steroid cord and the nerve roots also receive their blood supply particles would be of sufficient caliber to result in neural from radicular arteries, which take off from the aorta and travel at each of the vertebral levels through the neuralforamen along with each nerve root bilaterally. These ra- The pathogenesis of the final injury may be due to any dicular arteries primarily supply the nerve root, entering the one of the above factors alone or a combination of them.
intervertebral foramina just inferior to the exiting spinal There has been much discussion of types of steroids injected nerve and travel a tortuous path in the anterior inferior in the epidural space and specific focus on particulate ste- aspect of the foramina. Further, in a cadaveric dissection, it roid size in relation to the inner blood vessel diameter in the was found that radicular arteries that arise from vertebral region. However, when the particulate steroid embolization arteries tend to course in the anteromedial portion of the is put into perspective, it is only one of many potential foramen. The most likely area in which an injection would interfere with radicular vasculature was found to be in the In all the reported cases, the exact nature of the events that led to these complications (vasospasm, disruption, em- There are on average three larger radicular arteries that bolic occlusion) is unknown. Therefore, even with strict supply the anterior spinal cord by giving off branches su- adherence to proper technique arterial spasm, arterial dis- periorly and inferiorly to feed the anterior spinal artery ruption or embolization is still a rare but real possibility.
called radiculomedullary arteries. These radiculomedullary The commonly used technique of detecting a flashback arteries often predominate in the cervical spinal region. At of blood in the needle hub is only 44.7% sensitive when the thoracolumbar levels, these arteries take off from inter- costals and lumbar artery branches, which enter the foram- In a separate study involving 191 fluoroscopically guided ina and continue on to supply the anterior spinal artery.
lumbosacral transforaminal ESIs, using intermittent fluoros- Below the level of T8, there is often one large radiculomed- copy to identify proper segmental positioning followed by ullary branch that supplies the anterior spinal artery, called live anteroposterior fluoroscopy during injection of contrast, the artery of Adamkiewicz. Although its origin is variable, there was a reported 8.9% incidence of simultaneous epi- 85% of the time it takes off from the left side between the dural and vascular injection pattern, as well as a 4.2% levels of T9 and L2, at times as low as at lower lumbar incidence of vascular injection alone, totaling a 13.1% in- levels, rarely as low as and most often at cidence of vascular The study recommended There have been noted cases where the artery of Adamk- that live fluoroscopy, as opposed to intermittent imaging, be iewicz has a higher thoracic take-off point and supplies the anterior spinal artery via an iliac radiculomedullary ischemia or infarction of the supplied area, potentially ex- plaining the reported complications.
The anterior spinal artery may also receive blood supply There are little data that compare outcomes with partic- from ascending and deep cervical arteries that anastomose ulate and nonparticulate steroids. Donnell and coworkers with the vertebral and other arteries. The ascending cervical compared 8 mg dexamethasone with 80 mg triamcinolone artery often takes off either from the inferior thyroid artery mixed with 2 cc 1% lidocaine. All patients received 2 level or straight from the thyrocervical trunk or subclavian artery.
lumbar transforaminal ESIs. The average change in the Of note, in one case, it was found to be large enough to be visual analog scale was analyzed at the end of treatment.
cannulated by a 22-gauge needle. The ascending cervical The triamcinolone group had an average visual analog scale artery can then form anastomotic connections with the ver- decrease of 2.87 versus 1.21 for the dexamethasone group tebral, deep cervical, or occipital arteries. The deep cervical (P ϭ 0.0037). The pain completely resolved in 15 patients artery is one of two divisions of the costocervical trunk that receiving triamcinolone versus 4 of those patients who re- takes off from the posterior subclavian artery; the other ceived dexamethasone and experienced complete pain re- division is the superior intercostal artery. This deep cervical artery will provide spinal branches from levels C7 to T1, the The epidural steroids efficacy literature has almost ex- cervical radiculomedullary arteries. At the point of cervical clusively used particulate steroids in achieving favorable enlargement (C5 or C6), there is a larger medullary branch outcomes. Prospective, randomized, comparative analyses that joins with the anterior spinal artery. These radiculomed- of particulate and nonparticulate steroids are lacking. In the ullary arteries are of great clinical importance as they tra- absence of data to the contrary, use of nonparticulate ste- verse the entire length of the intervertebral foramen medi-ally and can be compromised during injection.
roids in ESIs likely compromises efficacy and may be no The posterior one-third of the spinal cord is supplied by safer than the more popular and standard of care particulate two posterior spinal arteries, which are smaller and less steroids due to other mechanisms of injury that are still contiguous compared with the single anterior spinal artery, present when nonparticulate steroids are used.
and these travel in the posterolateral areas of the spinal cord.
Conclusions and considerations to improve
epidural steroid injection safety
Comparison of particulate and nonparticulate
steroid injectate formulations
While making recommendations to enhance safety, itshould be noted that there is incomplete evidence to The commonly injected synthetic corticosteroids include formulate guidelines for a specific type of technique, prac- betamethasone sodium phosphate (BSP), BSP and beta- tice modification, or type of steroid used. Given the poten- methasone acetate (BSP-BA), triamcinolone acetonide tial for various types of vascular neural injury that range (TA), dexamethasone sodium phosphate (DSP), and meth- from vasospasm to mechanical injury to particulate matter ylprednisolone acetate They are all derivatives embolization and the essential absence of efficacy or safety of the cortisol analog prednisolone and have greater anti- data on nonparticulate steroids, such as dexamethasone, it inflammatory properties than The half-life of would be inappropriate to advocate use of one category of these formulations ranges from 36 to 72 hours, which is steroid over another or one particular type of steroid over related to particulate steroid size. The solutions with larger another. It is also notable that neural injury that is explained particles, and therefore a presumably longer half-life, would by vascular injury was also described in the absence of seem to be the preferred choice of injectate. These depot formulations would provide several weeks of sustained anti- The sustained duration of the anti-inflammatory effect of inflammatory effect over the injected area and therefore particulate steroids is postulated to be responsible for the presumably a greater degree of therapeutic effect.
In a microscopic comparative analysis between the var- weeks and months of improvement our patients report with ious formulations, there were significant differences found epidural steroids. Nonparticulate steroids such as dexameth- between the synthetic compounds. It is apparent that the asone do not provide sustained anti-inflammatory effects BSP formulations have the fewest proportions of the largest and have virtually no data supporting their use in ESIs.
sized particles in excess of 50 ␮m, whereas MA has the Because the nonparticulate steroids would be promptly ab- greatest amount of large particles. Of the two injectates with sorbed systemically and dissipate from the injected area, intermediate amounts of large particles—TA and DSP—TA they are likely no better than if they were injected intrave- tended to aggregate into larger particles measuring greater Back bleeding is unreliable in detecting intravascular Some have suggested that the larger the particulate size, entry. Live fluoroscopy while the myelographic contrast is the more likely it is that vessels which lie in the path of a being injected can be used to exclude intra-arterial needle transforaminal injection will become occluded and cause location before injection of particulate Techniques in Regional Anesthesia and Pain Management, Vol 13, No 4, October 2009 Contrast injection and digital subtraction analysis is an movement after final placement and therefore potential en- imaging technology that can enhance the detection of intra- vascular contrast uptake during transforaminal ESIs.
Should the number of attempts in a transforaminal injec- Real-time digital subtraction technology digitally “sub- tion be limited in presence of vascular spread? The answer tracts” the baseline radiograph from serial images. It has to this question is likely affirmative. Although the authors been demonstrated to reduce the incidence of intravascular are unable to provide a specific number, one or two in- injection.It is important to note that, although the inci- stances of arterial spread should probably result in termina- dence of detection of intravascular injection is statistically tion of procedure without further attempts.
significantly higher with digital subtraction analysis, it is ESIs are considered by many to be an important part of still not 100%. For example, the needle could move after the the multidisciplinary plan of care in the radiculopathy pa- real-time imaging before the injectate is given. Furthermore, tient. As discussed in this article, limiting the resulting use of this imaging technology may be providing us with a controversy over the case reports to the type of steroid capability that does not address other mechanisms of isch- injected oversimplifies the pathophysiology complexity and emic insult, such as mechanical injury to the blood supply or ignores many of the other factors that can also contribute to similar injury and the role of the transforaminal approach in Hodges and coworkers emphasize the dangers of seda- tion when performing cervical Sedation increases In our efforts to maximize epidural steroid efficacy and the possibility that the patient may not be able to verbalize safety, the type of steroid injected should be part of the pain or other abnormal sensations that might occur during greater discussion that includes discussion of relative safety of interlaminar versus transforaminal techniques and the Despite general discussion in the literature about exclud- multiple plausible alternative mechanisms of neural injury ing patients with preexisting vascular disease or spinal sur- gery from getting a transforaminal epidural steroid, there isno real information to base this conclusion on. It would be References
unfortunate to exclude and limit the use of these techniquesto nonoperated spines, for example. What would we do with 1. Brouwers P, Kottink E, Simon M, et al: A cervical anterior spinal those patients with disabling recurrent disc herniations and artery syndrome after diagnostic blockade of the right C6 nerve root.
stenotic adjacent segment disease? Furthermore, vascular anomalies occur even in nonoperated patients with lumbo- 2. Rathmell J, Aprill C, Bogduk N: Cervical transforaminal injection of steroids. Anesthesiology 100:1595-1600, 2004 Another safety factor that has been mentioned is use of 3. Devor M, Govrin-Lippman R, Raber P: Corticosteroids suppress ec- topic neural discharge originating in experimental neuromas. Pain blunt-tip needles that may reduce the probability of arterial entry, although this has never been proven to be the actual 4. Johansson A, Hao J, Sjolund B: Local corticosteroid application blocks source of the problem. Furthermore, it can’t be stated with transmission in normal nociceptive C-fibres. Acta Anaesthesiol Scand certainty that a blunt needle would eliminate intravascular 5. Benzon HT: Epidural steroid injections for low back pain and lumbo- entry or prevent vasospasm or vessel injury.
sacral radiculopathy. Pain 24:277-295, 1986 There is absence of comparative outcome studies that 6. Derby R, Kine G, Saal J, et al: Response to steroid and duration of favor one type of ESI approach over another. Therefore, radicular pain as predictors of surgical outcome. Spine 17:S176-S178, given the higher incidence of vascular complications re- ported in the literature with the performance of transforami- 7. Houten J, Errico T: Paraplegia after lumbosacral nerve root block: A report of three cases. Spine J 2:70-75, 2002 nal ESIs, the interlaminar ESI can be considered a safer 8. Huntoon M, Martin D: Paralysis after transforaminal epidural injection alternative. An argument can also be made to provide the and previous spinal surgery. Reg Anesth Pain Med 29:494-495, 2004 first injection consistently as an interlaminar ESI. If the 9. Rozin L, Rozin R, Koehler S, et al: Death during transforaminal patient response is inadequate in 1-2 weeks, only then can a epidural steroid nerve root block (C7) due to perforation of the leftvertebral artery. Am J Forensic Med Pathol 24:351-355, 2003 10. Baker R, Dreyfuss P, Mercer S, et al: Cervical transforaminal injection When a transforaminal ESI is being performed, consid- of corticosteroids into a radicular artery: A possible mechanism for eration should be given to the extent of the needle entry into spinal cord injury. Pain 103:211-215, 2003 the neuroforamen. The needle tip can be placed at the 11. Ludwig M: Spinal cord infarction after cervical transforaminal epi- entrance to the foramen, paraforaminally, or as far away dural injection: A case report. Arch Phys Med Rehabil 84:E37, 2003 12. Karasek M, Bogduk N: Temporary neurologic deficit after cervical from the foramen as possible while also allowing the injec- transforaminal injection of local anesthetic. Pain Med 5:202-205, 2004 tate to disperse on to the dural sleeve so the steroid can 13. Thomas E, Cyteval C, Abiad L, et al: Efficacy of transforaminal versus spread into the foramen and the anterior epidural space.
interspinous corticosteroid injection in discal radiculalgia: A prospec- During the performance of a transforaminal ESI, the tive randomized, double-blind study. Clin Rheumatol 22:299-304,2003 needle should be completely immobilized during and after 14. Watts RW, Silagy CA: Meta-analysis and the efficacy of epidural contrast injection as well as while connecting the syringe or corticosteroids in the treatment of sciatica. Anaesth Intens Care 223: the microbore extension to eliminate any subsequent needle 15. Hodges SD, Castelberg RL, Miller T, et al: Cervical epidural steroid 25. Gillilan LA: The arterial blood supply of the human spinal cord.
injection with intrinsic spinal cord damage: Two case reports. Spine 26. Tiso RL, Cutler T, Catania JA, et al: Adverse central nervous system 16. Nelson D, Landau W: Intraspinal steroids: History, efficacy, acciden- sequelae after selective transforaminal block: The role of corticoste- tality, and controversy with review of United States Food and Drug Administration reports. J Neurol Neurosurg Psychiatry 70:433-443, 27. Donnell C, Cano W, Eramo G: Comparison of of triamconolone to dexamethasone in the treatment of low back and leg pain via 17. Conners J, Wojack J: Interventional Neuroradiology: Strategies and lumbar transforaminal epidural steroid injection. Spine J 8:1S- Practical Techniques. Philadelphia, PA, W. B. Saunders, 1999 18. Bagshawe K, Magrath I, Golding P: Intrathecal methotrexate. Lancet 28. McMillan MR, Crompton C: Cortical blindness and neurologic injury complicating cervical transforaminal injection for cervical radiculop- 19. Wybier M: Transforaminal epidural corticosteroid injections and spi- nal cord infarction. Joint Bone Spine 75:523-525, 2008 29. Pooley RA, McKinney JM, Miller DA: The AAPM/RSNA physics 20. Hoeft MA, Rathmell JP, Monsey RD, et al: Cervical transforaminal tutorial for residents: Digital fluoroscopy. Radiographics 21:528-531, injection and the radicular artery: Variation in anatomical location within the cervical intervertebral foramina. Reg Anesth Pain Med 30. Siglar J, Garvan C, McLean J, et al: Digital subtraction: Does it matter? The rate of intravascular uptake in cervical epidural steroid 21. Smuch M, Fuller B, Yoder B, et al: Incidence if simultaneous epidural and vascular injection during lumbosacral transforaminal epidural in- injections with and without digital subtraction. Archives of Physical Medicine and Rehabilitation, Academy Annual Assembly Assembly 22. Huntoon MA: Anatomy of the cervical intervertebral foramina: Vul- nerable arteries and ischemic neurologic injuries after transforaminal 31. Etminan M, Takkouche B, Isoma FC, et al: Risk of ischemic stroke in epidural injections. Pain 117:104-111, 2005 people with migraine: Systematic review and meta-analysis of obser- 23. Gilman G, Goodman L, Rall T, et al: The Pharmacologic Basis of Therapeutics. New York, Macmillan, 1985, pp 1473-1477 32. Kurunlahti M, Karppinen J, Haapea M, et al: Three year follow-up og 24. Huntoon MA: Anterior spinal artery syndrome as a complication of lumbar artery occlusion magnetic resonance angiography in patients transforaminal epidural steroid injections. Semin Pain Med 2:204-207, ith sciatica: Association between occlusion and reported symptoms.

Source: http://nyucardiacsurgery.com/pmr/news/ESI%20MOI.pdf