Anatomy of the cervical intervertebral foramina: vulnerable arteries and ischemic neurologic injuries after transforaminal epidural injections

Anatomy of the cervical intervertebral foramina: vulnerable arteries and ischemic neurologic injuries after transforaminal epidural injections

Pain 117 (2005) 104–111 www.elsevier.com/locate/pain Anatomy of the cervical intervertebral foramina: vulnerable arteries and ischemic neurologic inj...

702KB Sizes 0 Downloads 44 Views

Pain 117 (2005) 104–111 www.elsevier.com/locate/pain

Anatomy of the cervical intervertebral foramina: vulnerable arteries and ischemic neurologic injuries after transforaminal epidural injections* Marc A. Huntoon Department of Anesthesiology, Mayo Clinic, 200 First Street SW, Rochester, MN 55905, USA Received 29 September 2004; received in revised form 19 May 2005; accepted 27 May 2005

Abstract Cervical transforaminal epidural steroid injections are performed for the treatment of radicular pain. Multiple recent case reports have raised safety concerns regarding neurologic deficits such as anterior spinal artery syndrome and cerebellar injury after these injections. To investigate the potential causes of these injuries, an anatomic study was conducted. In this study of 10 embalmed cadavers, the cervical intervertebral foramina were examined to determine if the ascending or deep cervical arteries supplied radicular or segmental medullary arteries potentially susceptible to cannulation or needle trauma during transforaminal injection. In two specimens, dissection was carried down to the spinal cord, demonstrating the anterior spinal, radicular, and segmental medullary arteries. Of 95 intervertebral foramina dissected, 21 had an arterial vessel proximal to the posterior aspect of the foraminal opening. Seven of these 21 were spinal branches that entered the foramen posteriorly, potentially forming radicular or segmental medullary vessels to the spinal cord. One additional ascending cervical artery formed a segmental medullary artery that joined the anterior spinal artery. This would only be injured by anterior needle misplacement. Of the seven foraminal branches, three were included in the deep dissections. Two contributed to segmental medullary arteries and one to a radicular artery. Variable anastomoses between the vertebral and cervical arteries were found. Therefore, it is possible to introduce steroid particles into the vertebral circulation via the cervical arteries. Critical arteries are located in the posterior aspect of the intervertebral foramen and may be vulnerable to injection or injury during transforaminal epidural steroid injection. q 2005 International Association for the Study of Pain. Published by Elsevier B.V. All rights reserved. Keywords: Anterior spinal artery syndrome; Cervical spine; Corticosteroids; Epidural injection; Spinal cord injury; Transforaminal injection

1. Introduction Conservative care of cervical radiculopathy may include epidural administration of corticosteroids and local anesthetics. Techniques include interlaminar and transforaminal epidural injections (Bush and Hillier, 1996; Rowlingson and Kirschenbaum, 1986; Slipman et al., 2000; Valle´e et al., 2001). Initial reports of experience with cervical transforaminal epidural steroid (CTES) injections indicated

Abbreviations: CTES, cervical transforaminal epidural steroid; SAP, superior articular process. * Presented in part at the annual meeting of the American Society of Anesthesiologists, Las Vegas, Nevada, October 2004. E-mail address: [email protected].

improvements in radicular symptoms with no reported injection-related neurologic injuries (Bush and Hillier, 1996; Slipman et al., 2000; Valle´e et al., 2001). Severe neurologic injuries have been reported after interlaminar epidural injections, caused by needle trauma to the spinal cord or nerves (Abram and O’Connor, 1996; Botwin et al., 2003; Bromage and Benumof, 1998; Hodges et al., 1998). Therefore, some authors considered the possibility that the CTES technique might be safer than the interlaminar approach (Manchikanti, 1999). Several recent reports of anterior spinal artery syndrome or brain injury occurring during CTES injections, however, have raised new safety concerns (Baker et al., 2003; Brouwers et al., 2001; Karasek and Bogduk, 2004; Ludwig and Burns, 2003; Rathmell et al., 2004; Rozin et al., 2003;

0304-3959/$20.00 q 2005 International Association for the Study of Pain. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.pain.2005.05.030

M.A. Huntoon / Pain 117 (2005) 104–111

Tiso et al., 2004). One CTES technique involves introducing the needle laterally in the neck and directing it toward the superior articular process (SAP) at the posterior aspect of the intervertebral foramen. The spine is viewed fluoroscopically along the intervertebral canal to optimize imaging of the intervertebral foramen. After the needle approaches or touches the SAP, it is then redirected to enter the dorsal midpoint of the foramen. An anteroposterior fluoroscopic view should confirm that the needle does not extend beyond the midsagittal plane of the cervical articular pillar (Rathmell et al., 2004). A potential hazard to the safe performance of the CTES procedure would be the presence of a vessel that communicates with the anterior spinal artery in the posterior aspect of the foramen. However, the incidence of such critical arteries has not yet been reported. Reported complications of CTES, such as anterior spinal artery syndrome, have been attributed to vertebral artery injections, presuming incorrect needle placement (Rozin et al., 2003). Other hypotheses for the cause of these episodes include movement of the needle after contrast injection has demonstrated satisfactory placement, vasospasm caused by the injected corticosteroid or local anesthetic, or particulate embolism causing downstream occlusion of critical anterior spinal arterial feeder vessels (Baker et al., 2003; Tiso et al., 2004). Because of the increasing reports of adverse events and the suggestion that many incidents are still not being reported (Baker et al., 2003; Rathmell et al., 2004), radiologic anatomy and spinal cord blood supply must be correlated. The present study was designed to examine the anatomy of the cervical intervertebral foramina to illustrate and define potentially vulnerable arteries. This has led to hypotheses as to the cause of neurologic injuries occurring during CTES injection.

2. Methods This study was approved by the Mayo Foundation Institutional Review Board. Ten embalmed cadavers were dissected to determine the relationships of the vertebral, deep, and ascending cervical arteries to the intervertebral foramina. Spinal branches originating from the deep and ascending cervical arteries and any anastomoses with the vertebral artery were sought. Arterial origin, outer diameter, location relative to the foraminal opening, and any resulting vessel vulnerability to a posteriorly placed periradicular or transforaminal needle were specifically examined. In 2 of the 10 cadavers, fluoroscopy was used to place needles using the previously described CTES technique (Rathmell et al., 2004). In the other cadavers, a needle was placed during the dissection in the clinical target area at the posterior aspect of the foramen. A vulnerable vessel was defined as an artery 2 mm or less from the pathway of the needle placed in the posterior aspect of the intervertebral foramen, dorsal to the exiting spinal nerve and anterior to the SAP. Only those vessels within the 2-mm distance were considered. Vessel outer diameter was measured at its origin unless the vessel entered the foramen, in which case

105

the measurement was taken as close as possible (without damaging tissues) to the external foraminal opening. Dissections were performed to isolate the subclavian artery and its branches and the roots of the brachial plexus. The branches of the subclavian artery, including the vertebral, ascending, and deep cervical arteries, were then dissected. Muscle attachments, fascia, and fatty tissues were removed to better visualize the vessels and their relationship to the posterior aspect of the intervertebral foramina. The foraminal areas examined included the clinically relevant C3-4 to C7-T1. Two additional deep dissections were then performed on cadavers 9 and 10 to demonstrate the anterior spinal cord arterial supply. The deep dissections were performed via an anterior approach. The vertebral arteries were located by carefully removing surrounding transverse foraminal bone and fascia, and examining the arteries for spinal branches. The ascending and deep cervical arteries were similarly dissected and any spinal branches noted. The anterior spinal elements including the vertebral bodies and prevertebral fascia were carefully removed. The dura was exposed and incised to demonstrate the spinal cord anterior surface. The ventral rami from C4 to C8 were dissected and retracted to demonstrate the position of any ascending or deep cervical vessels that had entered the external opening of the intervertebral foramen. The formation of any radicular or segmental medullary vessels from the vertebral, ascending cervical, or deep cervical arteries was noted.

3. Results Seven of the cadavers were female and three were male. The meanGSD age at time of death was 82.4G10.5 years. One cadaver (8) received from mortuary had extensive silica gel present and tissue damage from dissection of the great vessels on the right side of the neck. This side was excluded from the dissection. Thus, 10 left and 9 right sides were dissected and five intervertebral foramina dissected on each side—a total of 95 cervical foramina. The outer diameters of vessels from all cadavers are shown in Table 1. 3.1. Vertebral artery In this study, the vertebral arteries commonly arose from the subclavian artery but occasionally from the ascending aorta. The vertebral arteries entered the foramina transversaria at C6 in all but one case in which the right vertebral artery entered at C5. The average diameter of the vessel was 4.0 mm. In four instances, an anastomosis with an ascending cervical artery from the thyrocervical trunk or subclavian artery was found. Three of these anastomoses occurred proximal to the vertebral artery entering the foramina transversaria. One of the anastomoses demonstrated in a deep dissection (cadaver 9) occurred at the entry to the spinal canal (Fig. 1A). The vertebral artery was the parent vessel for several segmental medullary vessels, which generally were smaller than 1 mm and branched from the medial surface of the vessel (Fig. 1B). No vertebral artery or

106

M.A. Huntoon / Pain 117 (2005) 104–111

Table 1 Outer diameter of vessels in dissected cadavers Side

Right

Left

Average

Cadaver/ sex

Vessel outer diameter (mm) Vertebral artery

Ascending cervical artery

Deep cervical artery

Thyrocervical artery

1/F 2/M 3/F 4/F 5/M 6/F 7/F 8/M 9/F

3.0 4.0 3.5 3.2 3.5 4.0 4.5 D 4.5 5.5 3.0 4.0 3.5 4.0 4.0 3.2

7/F 8/M 9/F 10/F

5.0 3.0 5.0 5.5

0.6 0.8 1.2 1.0

4.0

1.0

1.8 1.0 1.4 1.0 0.8 1.8 A D 1.8 2.2 2.6 2.0 1.2 0.8 1.0 0.8 1.2 2.0 2.0c A 2.8 2.0 1.8 1.5

0.5a

10/F 1/F 2/M 3/F 4/F 5/M 6/F

1.0 1.2 1.0 1.0 1.0 0.8 1.2 D 1.0 1.2a 2.0 1.0 1.0b 1.0 0.8 0.8 1.2

2.0a

2.0a

A, absent; D, damaged. a Anastomosed with vertebral artery. b Multiple smaller branches !0.5 mm. c There were five distinct branches from a costovertebral trunk; the two largest were 1.2 and 1.0 mm and both were vulnerable at C6, C7, and C8.

branch met the criteria for vulnerability relative to the posterior foraminal needle placement. 3.2. Ascending cervical artery The ascending cervical artery was typically derived from the inferior thyroid or thyrocervical trunk. Occasionally it came directly from the subclavian artery. The ascending cervical arteries had an average outer diameter of 1.0 mm. The artery typically ascended on the anterior tubercles of the transverse processes and ascended roughly parallel to the more lateral phrenic nerve (Fig. 2A and B). It was noted that if the ascending cervical artery supplied a spinal branch, it typically occurred at the C3-4 or C4-5 foramen, and the spinal branch entered the posterior/inferior aspect of the external foraminal opening (Fig. 2A and B). In cadaver 10, a 2.0-mm (at origin) ascending cervical artery derived from the inferior thyroid artery entered the external foramen (0.8 mm) at C3-4 and eventually supplied a segmental medullary artery communicating with the anterior spinal artery (Fig. 2B). In its course, this artery passed dorsal/inferior to the C4 ventral ramus, also supplying a branch to the ventral ramus of C4. In another cadaver (5), an ascending cervical artery sent a spinal branch dorsal to the exiting C5 ramus. The ascending cervical branch

Fig. 1. (A) The anterior elements of the vertebrae and dura are removed. The right vertebral artery (RVA, top middle) is retracted with a needle. The C5, C6, and C7 ventral rami are labeled. The ascending cervical artery (AC, large arrow) forms an anastomosis (small arrow) at C4-5. The anterior spinal artery (ASA) is shown just above the lettering. (B) Detail of (A). Left of photo is inferior and right is superior. The right subclavian artery (SCA) gives off an AC. The carotid artery (CA), vagus nerve (VN, arrowhead), and right vertebral artery (VA) are shown at top left. A segmental branch arises from the AC (medium arrows) and becomes a large segmental medullary artery (SMA) to the ASA (small arrow). A smaller SMA from the left vertebral artery joins the ASA at the same level (far right, medium arrow).

itself entered the external foraminal opening. The small vertebral artery on this same side (3.5 mm) had no medially projecting spinal branches between C3-4 and C7-T1. 3.3. Deep cervical artery The deep cervical arteries in this study were usually slightly larger than the ascending cervical arteries, with an average diameter of 1.5 mm. The deep cervical arteries commonly arose as a single or multiple vessel from the costocervical trunk but occasionally as a single or multiple vessel directly from the subclavian artery. Two of the sides dissected did not have a definable deep cervical vessel, and four cadavers (6, 7, 9, and 10) had 2, 5, 3, and 2 deep cervical branches on one side, respectively. For the calculation of average size, therefore, the largest vessel of a deep cervical type was counted, and the denominator was 17 because of the two sides without a deep cervical vessel.

M.A. Huntoon / Pain 117 (2005) 104–111

107

Fig. 2. Cadaver 10. (A) The anterior elements of the vertebrae and foramina tranversaria have been removed to show the anterior spinal cord surface. The right subclavian artery gives off a thyrocervical trunk (TT, large arrow) at bottom left. From the TT, an ascending cervical artery is formed (arrowheads) and ascends medial to the more lateral phrenic nerve (PN, small arrow). The right vertebral artery (RVA) is shown at its origin from the subclavian at bottom and also more cephalad where it has been cut. (B) Detail of (A). A probe at mid left pushes the C4 ventral ramus cephalad, with the tip of probe elevating the ascending cervical artery (AC) branch after it enters the foramen. The anterior portion of the C4-5 zygapophysial joint (Z Joint) is shown after removal of transverse foraminal bone. The AC (tailed arrow) is shown ascending toward the foramen and forming a segmental artery in the foramen. The anterior spinal artery (ASA, large arrow) is at top right in the anterior median sulcus. Small arrows depict the course of the segmental medullary artery coursing ventral to the C4 root as it joins the ASA.

The deep cervical artery often gave branches only to the roots of the brachial plexus or adjacent muscles. In five instances, the vessel formed large spinal branches and entered the posterior aspect of the foramen directly posterior to the exiting ventral ramus (Figs. 3 and 4). These deep cervical spinal branches always entered at either C5-6, C6-7 or C7-T1. In one cadaver (10), two deep cervical arteries contributed to the formation of an anastomosis, from which a segmental medullary artery emerged. The deep cervical artery was commonly more anterior than is usually depicted and often coursed near the posterior aspect of the external foraminal opening (Figs. 3 and 5). In many cases, however, the artery was seen to migrate more posteriorly between zygapophysial joints (Fig. 5) into its more classic location between the semispinalis capitis and cervicis muscles. 3.4. Vessel proximity to foramina In 21 of the 95 foraminal areas examined, the parent ascending or deep cervical artery, or a large branch of it, was

Fig. 3. Top of photo is superior. An oval approximates the area of the intervertebral foramen. At left of oval, the superior articular process (SAP, small arrow) is shown. Inferior to the oval, the right deep cervical artery (large arrow) is shown entering the posterior aspect of the C5-6 foramen. At top middle, the C6 ventral ramus is retracted cephalad. The vertebral artery, only marginally larger than the deep cervical artery, is shown at right. This right vertebral artery did not supply any segmental vessels between C3-4 and C7-T1.

108

M.A. Huntoon / Pain 117 (2005) 104–111

Fig. 5. A 22-gauge needle (middle) is placed in the posterior aspect of the intervertebral foramen at C5-6 at a more anterior angle than is typical for cervical transforaminal epidural steroid injection. The left C3-4 and C4-5 zygapophysial joints (Z Joints) are shown at bottom right. At left, (inferior) the subclavian artery (SCA) and its branches, including the thyrocervical trunk (TT), transverse cervical (TC), and deep cervical (DC, large arrow), are shown. The most medial branch from the SCA is the vertebral artery (RVA) shown at top right, anterior and medial to the C6 ramus (C6). The DC could easily be punctured only at the external foraminal opening, and in this case did not supply the spinal cord.

Fig. 4. An illustration demonstrating ascending cervical and deep cervical arteries anastomosing with the vertebral artery posterior to the spinal nerves. Note that the ascending cervical artery enters the foramen at C3-4 or C4-5, whereas the deep cervical artery enters the more distal C5-6, C6-7, or C7-T1 foramina. (By permission of Mayo Foundation for Medical Education and Research.)

within 2 mm of the needle path for a CTES procedure. Thirteen of those 21 vessels potentially could be penetrated during a CTES injection, but only if the needle was not sufficiently advanced into the foramen. Of note, these 13 vessels did not wholly enter the foramen, did not contribute major spinal branches, and had no demonstrable communication with the spinal circulation except smaller twigs to the ventral rami. For another spinal branch, the ascending cervical artery coursed anteriorly to the ventral rami, and its segmental branch entered the spinal canal at C6-7. In the seven other instances, in which a vessel passed into the dorsal aspect of the foramen, cannulation during correct CTES needle placement was possible. Of these seven spinal branches, two were ascending cervical arteries that entered the foramina at C3-4 or C4-5 (1 each), and five were deep cervical arteries that entered at C5-6 (1), C6-7 (2), or C7-T1 (2). The deep dissection on cadaver 10 demonstrated three vessels that passed into the foramen posteriorly. On the right side, a 2.0-mm ascending cervical artery came off the inferior thyroid artery, ascended, and entered the posterior

aspect of the C3-4 foramen. This artery then supplied the ventral ramus of C4 and continued to form a segmental medullary feeder vessel that joined the anterior spinal artery (Fig. 2A and B). At the point where a CTES needle would be placed, this vessel was still 0.8 mm in diameter. This segmental vessel would be vulnerable to cannulation. On the opposite side, the costocervical trunk passed inferior to the anterior scalene muscle, then sent two deep cervical arteries into the posterior aspects of the C6-7 and C7-T1 foramen. The C6-7 branch further divided to send a branch to the C5-6 foramen. These arteries supplied the ventral rami and formed other small branches, coursing through the posterior aspect of the foramen (Fig. 6A and B). Near the target area for CTES injection, these two deep cervical branches were 1.2 mm in diameter. The vessels coursed toward the spinal cord to contribute to an anastomosis of several small branches, forming anterior radicular branches that supplied roots, spinal cord pial vessels, and a lateral longitudinally oriented artery. A definitive anterior spinal artery could not be confirmed medial to where this longitudinal artery was noted. A segmental branch from this longitudinal vessel later joined the anterior spinal artery just cephalad to C5-6.

4. Discussion The ascending and deep cervical arteries have been previously described to have spinal branches that anastomose with branches from the vertebral artery or contribute radicular or segmental medullary branches to the anterior spinal artery (Gillilan, 1958). These vessels have not been

M.A. Huntoon / Pain 117 (2005) 104–111

Fig. 6. Left side of cadaver 10. Left is inferior, right is superior. (A) The cut end of the left vertebral artery (LVA) is shown at top left. The subclavian artery (SCA, bottom left) gives off a costocervical trunk (CT) and two deep cervical arteries (DC). The DC arteries and branches course through the posterior aspect of the intervertebral foramina of C5-6, C6-7, and C7-T1. The C7 and C8 ventral rami are pulled medially toward the spinal cord. A 25-gauge needle (bottom left) and 22-gauge needle (bottom right) enter the foramen; the 25-gauge needle is cannulating the DC artery branch. Arrowheads depict the course of two of the DC arteries, with the most medial arrowhead at the target area for transforaminal injection. The branches of C5-6 contribute to an anastamosis of small arteries that supply a segmental medullary artery to the anterior spinal artery. (B) The C7 and C8 ventral rami are shown in their natural position with needles in place. The more superior 22-gauge needle (right) was too large to cannulate the DC artery at the midsagittal area of the foramen. The vertebral artery (VA), subclavian artery (SCA), and costocervical trunk (CT) are shown top left. Arrowheads show the continuation of the DC arterial branches at approximately the level of a transforaminal epidural needle placement. The anterior surface of the spinal cord is clearly seen at the top.

previously reported to enter the intervertebral foramina posteriorly. In the present study, seven of the eight spinal arterial branches from the ascending or deep cervical arteries entered the foramina posteriorly. In the two deep dissections, three of these arterial vessels would have been easily cannulated during CTES. Two other cadavers (4 and 7) had extremely tiny or absent medially projecting spinal segmental branches from the vertebral artery on the same side that ascending or deep cervical arteries gave off spinal branches. These dissections indicate that the ascending and deep cervical arteries were also potentially the main supply to the cervical anterior spinal artery. The presence of

109

significant collaterals from above C3-4 or below C7-T1 could be neither confirmed nor refuted. The ascending cervical artery is usually described as a small branch continuing superiorly after the inferior thyroid artery projects medially, ascending on the anterior tubercles and supplying one or two spinal branches (Williams, 1995). The ascending cervical artery is depicted as sending these spinal branches at the same level as the deep cervical artery (Standring, 2005). This was not seen in the present dissections. The deep cervical artery is classically described as arising from the costocervical trunk and passing between the seventh cervical transverse process and the neck of the first rib. It ‘ascends between the semispinalis capitis and cervicis to the second cervical level. A spinal branch enters the vertebral canal between the seventh cervical and first thoracic vertebrae’ (Williams, 1995). The present study confirmed that a spinal branch does occasionally enter not only at C7-T1 but also at C5-6 or C6-7. The arteries in these dissections were frequently found to be more anterior than classically described, particularly in their early course. Anterior spinal artery syndrome and cerebellar ischemia are devastating complications of CTES injection. The solitary and often small or discontinuous anterior spinal artery is susceptible to decreased perfusion without reinforcing arteries (Gillilan, 1958). In many animal models, cervical segmental medullary vessels enter at nearly every spinal level from the vertebral artery. Humans, however, may have only one or two of these vessels feeding into the anterior spinal artery. Primate research showed that occlusion of the segmental medullary arteries leading to

Fig. 7. An illustration of a cervical transforaminal needle cannulating a segmental artery contributed by the ascending cervical artery. Steroid particles (purple) are shown coalescing in the anterior spinal artery. (By permission of Mayo Foundation for Medical Education and Research.)

110

M.A. Huntoon / Pain 117 (2005) 104–111

the anterior spinal artery resulted in severe damage to the anterior two-thirds of the cord both above and below the lesion (Yoss, 1950). If the ascending and deep cervical arteries both supply arteries that are critical to spinal perfusion, we may need to reexamine conventional assumptions regarding the relative contributions of the vertebral artery and ascending and deep cervical arteries. Anterior cord ischemic events occurring after seemingly appropriate CTES injections argue that the importance and location of these cervical vessels have been underestimated. It could be argued that arteries at the external opening of the intervertebral foramen are immaterial to ultimate needle position for CTES; the vertebral artery is also at the external opening. Notably in this study, no vertebral artery-derived radicular or segmental vessels were identified near classically placed CTES needles. Avoidance of penetration of the vertebral artery is the reason the CTES technique advocates needle placement in the posterior aspect of the foramen. This study demonstrated that two-thirds of the extraforaminally located ascending and deep cervical arteries did not supply a spinal artery. The variability in anatomy would make it unclear in which circumstances the physician should abandon the procedure if a blood vessel is penetrated. Some physicians may not completely advance the needle into the foramen because of a fear of cannulating a segmental vessel, thereby performing an inadequate injection. This study demonstrates that the parent vessels for those supplying the anterior spinal cord are more likely to be cannulated because of their larger size at the external foraminal opening. Even if only one instance of an ascending cervical artery joining the anterior spinal artery in the posterior aspect of the foramen was present in these 10 cadavers, it would account for at least a 1% risk of a potential for anterior spinal artery ischemic injury. The two deep cervical branches in cadaver 10 that contributed radicular and segmental vessels might increase the potential risk of neurologic injury to 3%. If the spinal branches in the four incompletely dissected foramina contributed to the anterior spinal blood supply, the risk could be as high as 7%. The presence of a critical vessel in the foramen does not necessarily mean that it will be cannulated. Nor is it certain that accidental cannulation of one of these vessels would always result in neurologic injury. Some blood vessels supply non-critical tissues (e.g. muscle) with extensive anastomoses, whereas other vessels are critical to spinal perfusion and may have few, if any, collaterals. Obviously, findings from 10 cadaver specimens cannot quantitate the actual risk, but they can help physicians performing these procedures to develop very strict criteria for their safe conduct. Possible mechanisms of injury from a CTES injection besides needle trauma-induced vascular injury have been suggested by other authors (Baker et al., 2003; Rathmell et al., 2004; Tiso et al., 2004). For example, it is known that corticosterone may have effects on vascular endothelium, which may lead to vasospasm (Rorie, 1982). Vasospasm is

one possible cause of neurologic injury after CTES. It is also possible that particulate depot steroid embolization could jeopardize blood flow to the anterior spinal artery (Fig. 7) (Tiso et al., 2004). Several procedural modifications may help to minimize the risks during CTES. Conventional fluoroscopy and realtime computed tomography cannot consistently demonstrate the presence or absence of radicular or segmental vessels (Baker et al., 2003; Somayaji et al., 2005). The inherent variability in anatomy, wherein some vessels harmlessly pass by the external foraminal opening and others form a spinal branch, makes it difficult to know when to proceed with CTES injection when a vessel is punctured. If CTES injections are performed, digital subtraction radiography should be considered (Baker et al., 2003). It has been shown that a larger 22-gauge pencil point needle with side hole may be less likely to puncture or cannulate a critical vessel (Figs. 4 and 6A and B) (Heavner et al., 2003). The vessels in this study maintained very close contact to the caudad aspect of the foramen. No area of the foramen appeared to be ‘safer’ than another with respect to potential arterial puncture or injection. In summary, ascending and deep cervical arterial branches enter the external opening of the posterior intervertebral foramen near the classic target area for transforaminal epidural injections. These branches occasionally supply anterior radicular and segmental medullary arteries to the spinal cord. Because these arteries are contributors to anterior spinal artery flow, injection into or injury to these vessels may explain the occurrence of ischemic neurologic events. These results should promote caution in the use of these techniques.

Acknowledgements I thank Peter Wilson, MBBS, PhD and Duane Rorie, MD, PhD for their review of both the manuscript and the cadaver images, Rita E. Anderson for help with typing and organizing the manuscript, and Michael A. King and M. Alice McKinney for the illustrations.

References Abram SE, O’Connor TC. Complications associated with epidural steroid injections. Reg Anesth 1996;21:149–62. Baker R, Dreyfuss P, Mercer S, Bogduk N. Cervical transforaminal injection of corticosteroids into a radicular artery: a possible mechanism for spinal cord injury. Pain 2003;103:211–5. Botwin KP, Castellanos R, Rao S, Hanna AF, Torres-Ramos FM, Gruber RD, Bouchlas CG, Fuoco GS. Complications of fluoroscopically guided interlaminar cervical epidural injections. Arch Phys Med Rehabil 2003;84:627–33. Bromage PR, Benumof JL. Paraplegia following intracord injection during attempted epidural anesthesia under general anesthesia. Reg Anesth Pain Med 1998;23:104–7.

M.A. Huntoon / Pain 117 (2005) 104–111 Brouwers PJ, Kottink EJ, Simon MA, Prevo RL. A cervical anterior spinal artery syndrome after diagnostic blockade of the right C6-nerve root. Pain 2001;91:397–9. Bush K, Hillier S. Outcome of cervical radiculopathy treated with periradicular/epidural corticosteroid injections: a prospective study with independent clinical review. Eur Spine J 1996;5:319–25. Gillilan LA. The arterial blood supply of the human spinal cord. J Comp Neurol 1958;110:75–103. Heavner JE, Racz GB, Jenigiri B, Lehman T, Day MR. Sharp versus blunt needle: a comparative study of penetration of internal structures and bleeding in dogs. Pain Pract 2003;3:226–31. Hodges SD, Castleberg RL, Miller T, Ward R, Thornburg C. Cervical epidural steroid injection with intrinsic spinal cord damage: two case reports. Spine 1998;23:2137–42. Karasek M, Bogduk N. Temporary neurologic deficit after cervical transforaminal injection of local anesthetic. Pain Med 2004;5:202–5. Ludwig MA, Burns S. Spinal cord infarction after cervical transforaminal epidural injection: a case report [abstract]. Arch Phys Med Rehabil 2003;84:A37. Manchikanti L. Cervical epidural steroid injection with intrinsic spinal cord damage. Spine 1999;24:1170–2. Rathmell JP, Aprill C, Bogduk N. Cervical transforaminal injection of steroids. Anesthesiology 2004;100:1595–600. Rorie DK. Metabolism of norepinephrine in vitro by dog pulmonary arterial endothelium. Am J Physiol 1982;243:H732–7. Rowlingson JC, Kirschenbaum LP. Epidural analgesic techniques in the management of cervical pain. Anesth Analg 1986;65:938–42.

111

Rozin L, Rozin R, Koehler SA, Shakir A, Ladham S, Barmada M, Dominick J, Wecht CH. Death during transforaminal epidural steroid nerve root block (C7) due to perforation of the left vertebral artery. Am J Forensic Med Pathol 2003;24:351–5. Slipman CW, Lipetz JS, Jackson HB, Rogers DP, Vresilovic EJ. Therapeutic selective nerve root block in the nonsurgical treatment of atraumatic cervical spondylotic radicular pain: a retrospective analysis with independent clinical review. Arch Phys Med Rehabil 2000;81: 741–6. Somayaji HS, Saifuddin A, Casey AT, Briggs TW. Spinal cord infarction following therapeutic computed tomography-guided left L2 nerve root injection. Spine 2005;30:E106–8. Standring S, editor. Gray’s anatomy: the anatomical basis of clinical practice. 39th ed. Edinburgh, UK: Churchill Livingstone; 2005. p. 784–6. Tiso RL, Cutler T, Catania JA, Whalen K. Adverse central nervous system sequelae after selective transforaminal block: the role of corticosteroids. Spine J 2004;4:468–74. Valle´e JN, Feydy A, Carlier RY, Mutschler C, Mompoint D, Vallee CA. Chronic cervical radiculopathy: lateral-approach periradicular corticosteroid injection. Radiology 2001;218:886–92. Williams PL, editor. Gray’s anatomy: the anatomical basis of medicine and surgery. 38th ed. Edinburgh: Churchill Livingstone; 1995. p. 1535–6. Yoss RE. Vascular supply of the spinal cord: the production of vascular syndromes. Med Bull (Ann Arbor) 1950;16:333–45.