Complications and side effects of cervical and lumbosacral selective nerve root injections

277

Complications and Side Effects of Cervical and Lumbosacral Selective Nerve Root Injections Christopher W. Huston, MD, Curtis W. Slipman, MD, Cyndi Garvin, PhD ABSTRACT. Huston CW, Slipman CW, Garvin C. Complications and side effects of cervical and lumbosacral selective nerve root injections. Arch Phys Med Rehabil 2005; 86:277-83. Objective: To determine the incidence of complications and side effects of cervical and lumbosacral selective nerve root injections (SNRIs). Design: Prospective, nonrandomized controlled trial of consecutive patients with independent interviews for immediate, 1-week, and 3-month follow-ups. Setting: Tertiary, academic spine center. Participants: Of 160 patients eligible for SNRI, 151 underwent 306 SNRIs (study group). Sixty patients who had not undergone any intervention served as the control group. Interventions: Diagnostic and therapeutic fluoroscopically guided cervical and lumbosacral SNRIs. Lidocaine (Xylocaine) was used in the diagnostic injections and a mixture of lidocaine and betamethasone was used in the therapeutic injections. Main Outcome Measures: Complications and side effects experienced during the procedure were recorded by the interventionalist. Side effects and complications experienced immediately, 1 week, and 3 months after injection were determined through independent interviews using a questionnaire format. The control group was independently interviewed using the same questionnaire. Results: There were no major complications, such as death, paralysis, spinal nerve injury, infection, or allergic reaction, during the study. Ninety-one percent of subjects had no side effects during the procedure. A positive response on interview was reported by 39.4% of the study subjects immediately after the procedure. Lumbosacral SNRI side effects were as follows: increased pain at the injection site (17.1%); increased radicular pain (8.8%); lightheadedness (6.5%); increased spine pain (5.1%); nausea (3.7%); nonspecific headache (1.4%); and vomiting (0.5%). Immediate side effects from cervical SNRI were as follows: increased pain at injection site (22.7%); increased radicular pain (18.2%); lightheadedness (13.6%); increased spine pain (9.1%); nonspecific headache (4.5%); and nausea (3.4%). At the 1-week interview, comparison of the study and control groups showed statistical significance for the following: pain at injection site, cervical (P⫽.001) and lumbar (P⫽.005); nonspecific headache, cervical (P⫽.019); and nonspinal headache, cervical (P⫽.002). At 3 months, 2 subjects complained of

From The Orthopedic Clinic Association, Scottsdale, AZ (Huston); Penn Spine Center, Department of Rehabilitation Medicine, University of Pennsylvania Medical Center, Philadelphia, PA (Slipman); and College of Medicine, University of Florida, Gainesville, FL (Garvin). Presented to the American Academy of Physical Medicine and Rehabilitation, October 10 –13, 1996, Chicago, IL. No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit upon the author(s) or upon any organization with which the author(s) is/are associated. Reprint requests to Christopher W. Huston, MD, The Orthopedic Clinic Association (TOCA), 9377 E Bell Rd, Ste 231, Scottsdale, AZ 85260, e-mail: [email protected]. 0003-9993/05/8602-8369$30.00/0 doi:10.1016/j.apmr.2004.02.018

increased neck pain and 1 reported heartburn and fluid retention. Two lumbar SNRI subjects were lost to follow-up. Conclusions: There were no major complications with cervical and lumbosacral SNRIs, although various minor side effects occurred. Key Words: Injections; Nerve root disorder; Postoperative complications; Radiculitis; Rehabilitation. © 2005 by American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation PIDURALS HAVE BEEN USED to treat radicular pain E for more than 50 years. Potential complications and side effects with lumbar, caudal, and cervical epidural steroid injections have been reported and include dural puncture,1-5 spinal headache,3,5 transient hypotension,6,7 Cushing’s syndrome,8 bacterial meningitis,9 chemical meningitis,10 epidural abscess,11,12 angina,6 sinus arrhythmia,6 respiratory distress from spinal anesthesia,6 sacral pain,7 transient increased sciatica,1,13 numbness,7 subjective weakness,7 transient headache,4 and transient dizziness.4 Complications of cervical epidural injections include dural puncture, bloating, nausea and vomiting, facial flushing, fever, nerve root injury,14 pneumocephalus,15 epidural hematoma,16,17 stiff neck,18 Cushing’s syndrome,19 transient paresthesias, hypotension, and respiratory insufficiency.20 With advances in technology, blinded interlaminar epidurals are now being replaced by fluoroscopically guided selective nerve root injections (SNRIs), which have the advantage of directly targeting the involved spinal nerve and dorsal root ganglion within the foramen.21 The potential benefits of this intervention have been reported22-25; however, the associated risks have not been thoroughly investigated. This lack of scientific inquiry is important because the decision to proceed with an SNRI should be based not only on the efficacy of the treatment but also on its potential risk. Kikuchi et al26 reported no complications in 75 patients who underwent a cervical SNRI. They list potential complications of major blood vessel puncture, dural puncture with spinal injury, and pneumothorax. Krempen and Smith27 performed lumbosacral SNRIs on 85 patients and encountered no complications. They hypothesized that nerve root sleeve tear, infection, and viscus perforation are possible complications. Schutz et al28 reported the same potential complications, but none occurred in their 23 patients. Kikuchi and Hasue,29 in a retrospective review of 119 patients, reported no complications. None of the cited studies systematically investigated the potential side effects and complications. Slipman et al30 retrospectively evaluated 888 spinal injections that included cervical and lumbar SNRIs and found a 0.9% incidence of side effects. Botwin et al31 retrospectively evaluated side effects with lumbar SNRI and reported increased back pain, radicular pain, nonpositional headaches, vasovagal reaction, intraoperative hypertension, nausea, dizziness, facial flushing, and transient leg paresis. Retrospective studies may underestimate potential side effects and complications. Additionally, one should Arch Phys Med Rehabil Vol 86, February 2005

278

COMPLICATIONS OF SELECTIVE NERVE ROOT INJECTIONS, Huston Table 1: Subject Demographics Group

SNRI

Cervical SNRI

Lumbar SNRI

Control

P

Subjects Men Women Mean age (y) (range) SNRI performed Previous spine surgery

151 69 82 47.4 306 30.5%

37 14 23 41.7(26–70) 89

114 55 59 49.2(24–86) 217

60 18 42 51.5 (15–90)

NS .042 NS

25.4%

NS

Abbreviation: NS, not significant.

not assume that the complications and side effects seen with interlaminar epidurals will be similar to SNRIs. The technical aspects of interlaminar and caudal epidural injections differ markedly from those required to perform an SNRI. Interlaminar and caudal epidural injections use a larger volume of injectate— contrast, local anesthetic, and corticosteroid. SNRI targets the nerve root sheath at the foramen and, therefore, is target specific.21 Because of these differences in injectate volume and site of injection, not all complications encountered during epidural interlaminar and caudal injections can be extrapolated to SNRIs. Furthermore, SNRI may have complications that are not encountered with nonspecific interlaminar epidural injections. A comprehensive Medline search of the English literature failed to identify any prospective studies that evaluated the complications and side effects of cervical and lumbosacral SNRIs. We designed a prospective study with independent evaluators and a control group to determine the incidence of such complications and side effects. METHODS Consecutive subjects undergoing SNRIs and a control group not undergoing an injection procedure were studied prospectively. Using information from the medical literature, we developed a questionnaire designed to elicit potential complications and side effects. Independent interviewers queried patients and accumulated data. The study was approved by the institutional review board of the University of Pennsylvania. During 1995 and 1996, consecutive patients were asked to participate in our study at the time of consent for an SNRI. All SNRIs were performed by either the attending physician (CWS) or fellow (CWH). All injections were done under fluoroscopic guidance with accurate needle placement, and injectate flow was confirmed with 0.5 to 2.0mL of contrast. Diagnostic injections were performed with 0.5mL of 1% lidocaine (Xylocaine) or 1.0mL of 2% lidocaine in cervical and lumbosacral SNRIs, respectively. Therapeutic cervical SNRI contained 1.0mL of betamethasone mixed with 0.5mL of 1% lidocaine. Therapeutic lumbar SNRI contained 2.0mL of betamethasone mixed with 2.0mL of 1% lidocaine. Any complication that occurred during the procedure was recorded by an interventionalist. In the recovery room, a registered nurse, who was not involved in the procedure itself, interviewed the subjects immediately after the procedure to determine what, if any, side effects and complications occurred. The standardized questionnaire we developed was used and the answers recorded. At 1 week, a physician, who was not a member of the treating team, interviewed all subjects by telephone. The standardized questionnaire, with 24 categories covering various symptoms, was used each time the side effects and complications were evaluated. Any symptom, regardless of its duration or severity or intensity, was recorded. At 3 months, an independent physician interviewed the subjects Arch Phys Med Rehabil Vol 86, February 2005

about their experiences with the 24 categories and about their individual complaints. All complaints were recorded. At 3 months, if a subject complained of having increased pain since the procedure, he/she was asked whether he/she would have consented to the injection based on his/her experience to date. Consecutive new patients who had not undergone any invasive intervention at the Spine Center served as the control group. Patients in that group were interviewed once by questionnaire by one of the independent physicians. All data were analyzed using SAS software, version 6.2.a Data sets were created and entered into Excelb spreadsheets. All data elements were checked for completeness and implausible values. Categoric data analyses were conducted using chi-square tests of proportions. In cases where sparseness of cells was observed, we used the Fisher exact test. We used the Wilcoxon rank-sum test to test for group differences on numerical measures. RESULTS One hundred fifty-one of 160 consecutive patients who underwent 306 SNRIs comprised the study group (9 refused to participate). Table 1 presents the study and control group demographics. The mean age of the study group was 47.4 years and was composed of 54.3% women and 45.7% men. In the study group, 114 subjects (59 women, 55 men; mean age 49.2y; range 24 – 86y) underwent 217 lumbosacral SNRIs. Thirty-seven subjects (23 women, 11 men; mean age, 41.7y; range, 26 –70y) received 89 cervical SNRIs. The control group included 60 new patients (42 women, 18 men; average age, 51.5y; range, 15–90y). Previous spine surgery had been performed on 25.4% of the control group and on 30.5% of the study group. A comparison between the 2 groups revealed no statistically significant difference in age and previous spine surgery. Statistically, the control group had more women (P⫽.042). Any given subject may be predisposed to more complaints than other subjects. Each time an injection is performed there are associated risks. Tables 2 and 3 present the side effects and complications presented by each subject per injection. Procedural Complications Of the 151 subjects who received the 306 injections, 91% experienced no complications or side effects during the procedure. The incidence of complication or side effect per injection was separately determined for lumbar and cervical SNRIs. Lumbar SNRI produced lightheadedness in 2.8% of patients, nausea in 1.4%, increased radicular pain in 0.9%, nonspecific headache in 0.5%, and vasovagal reaction in 0.5%. Cervical SNRI resulted in lightheadedness in 2.2% of patients, dural puncture in 1.1%, and nausea in 1.1%.

279

COMPLICATIONS OF SELECTIVE NERVE ROOT INJECTIONS, Huston Table 2: Complications and Side Effects Per Subject Complaint (% of Subjects)

Procedural

Subjects Any complaints Vasovagal Dural puncture Increased spine pain Increased radicular pain Increased pain at injection site Increased pain Lightheadedness Nausea Numbness: distribution of nerve block Numbness: lower extremity Weakness: upper extremity Headache: nonspecific, not spinal Headache: increased with standing Headache: not increased with standing Fluid retention Agitation Insomnia: pain related Insomnia: not pain related Weight gain Fatigue/malaise Esophagitis/gastritis: heartburn—epigastric or chest Facial or chest flushing Hearing loss

151 7.2 0.7 0.7 1.3

Immediate Postprocedure

1 Week

3 Months

Controls

P*

151 39.4

151 80.1

149 2.0

60 96.6

.003

11.6 22.6 32.9

36.8 37.4 30.3 14.8 19.4 17.4 5.8 11.0 1.9 7.7 5.2 18.1 8.4 16.8 11.0 9.0 7.1 18.7

1.3

33.9 35.6 8.4 22.0 27.1 10.2 NA 32.2 8.5 1.7 1.7 11.9 23.7 52.5 39.0 40.7 0.0 44.1

4.6 2.6

16.1 7.7 9.0 4.5

0.7

5.2

23.9 19.4 1.3

0.7

0.7

28.8 13.6 6.8

NS NS .001 NS NS NS NA NS .024 NS NS .002 .001 .001

.001 NS NS

NOTE. Control subjects were only compared with 1-week complaint using chi-square analysis; if cells were small, then the Fisher exact test was used. Blanks in table indicate no incidence of the complaint. Abbreviations: NA, not available; NS, not significant. *Significance for P⬍.05.

Immediate Postprocedure Interview The interview in the recovery room immediately after the procedure yielded an affirmative response to at least 1 question from 39.4% of the subjects. Table 2 shows the incidence of complications per subject and table 3 shows the incidence per injection. For lumbar SNRI, postprocedure interview complaints per injection were: increased pain at the injection site, 17.1%; increased radicular pain, 8.8%; lightheadedness, 6.5%; increased spine pain, 5.1%; nausea, 3.7%; nonspecific headache, 1.4%; and vomiting, 0.5%. For cervical SNRI, immediate complications and side effects per injection were: increased pain at injection site, 22.7%; increased radicular pain, 18.2%; lightheadedness, 13.6%; increased spine pain, 9.1%; nonspecific headache, 4.5%; and nausea, 3.4%. One-Week Interview No subject was lost to follow-up at the 1-week interview. Statistically, comparisons between controls and study subjects could only be made in incidence per subject (tables 2, 3) and not incidence per injection. We compared the control group with the total study group, cervical SNRI group, and lumbar SNRI group. There was no statistical significance for a higher incidence of complications or side effects in the total study group compared with the control group. Pain at the injection site occurred in 30.3% of the study group subjects. Unfortunately, because the control group did not have injections, no comparison was possible. The incidence of increased pain at the injection site was lower than the incidence of increased spine pain, and this latter side effect

was not statistically significant when compared with the control group. The cervical SNRI group, when compared with the control group, showed a statistically significant increase in pain at the injection site (P⫽.001), nonspecific headache (P⫽.019), and nonspinal headache— headache not increased with standing (P⫽.002). Pain at the injection site was increased for the lumbar SNRI group (P⫽.005). For all other potential complaints, lumbar SNRI had no statistical significance or had a lower incidence. Three-Month Interview Two lumbar SNRI subjects were lost to follow-up, leaving 149 subjects for the 3-month analysis. Only 3 subjects responded affirmatively to any of the interview questions. One subject who had lumbar SNRI complained of persistent heartburn and fluid retention, which began on the day of the injection. Two cervical SNRI subjects complained of persistent increased spine pain. They were asked whether, considering their outcome, they would have undergone the procedure had they known they might have increased pain. One subject would and the other would not have had the procedure. DISCUSSION We found no serious complications in the 151 subjects who underwent 306 SNRIs. More specifically, there were no deaths, paralysis, infection, or nerve root injuries. Various side effects did occur. Arch Phys Med Rehabil Vol 86, February 2005

280

COMPLICATIONS OF SELECTIVE NERVE ROOT INJECTIONS, Huston Table 3: Complaints—Cervical and Lumbar Procedure % Injection Complaint

Subjects No. of injections Any complaints Vasovagal Dural puncture Increased spine pain Increased radicular pain Increased pain at injection site Increased pain: not specific Lightheadedness Nausea Vomiting Numbness: distribution of nerve block Numbness: contralateral to injection Numbness: lower extremity Weakness: upper extremity Weakness: contralateral lower extremity Weakness: distribution of nerve injection Headache: nonspecific, not spinal Headache: increased with standing Headache: not increased with standing Fluid retention Agitation Insomnia: pain related Insomnia: not pain related Weight gain Fatigue/malaise Esophagitis/gastritis: heartburn— epigastric or chest Facial or chest flushing Hearing loss

C

37 89

Immediate % Injection

L

C

L

114

37 89

114

1 Week % Subject C

37 89 86.5

L

114

3 Months % Subject C

37 89

L

112

78.1

P (vs Control) C

L

P (C vs L)

96.6

NS

.001

NS

33.9 35.6 8.5 22.0 27.1 10.2

NS NS .001 NS NS NS

NS NS .005 NS .035 NS

NS .039 .005 NS .005 NS

Control

60

0.5 1.1

2.2 1.1

2.8 1.4

9.1 18.2 22.7

5.1 8.8 17.1

13.6 3.4

6.5 3.7 0.5

11.4

2.3

2.3 1.1

5.1

43.2 51.4 40.5 13.5 35.1 24.3

33.3 32.5 26.3 14.0 14.0 14.0

5.4

0.9

5.4

NA

NS

32.2 8.5 0.5 0.5

0.5

4.5

1.4

13.5

5.3

1.7

.019

NS

NS

5.4

5.3

1.7

NS

NS

NS

37.8 8.8 24.3 0.0

11.4 5.4 14.9 0.9

.002 .019 .006 .001

NS .007 .001 .001

.001 NS NS NS

32.4

14.0

11.9 23.7 52.5 39.0 40.7 0.0 44.1

NS

.001

.012

32.4 24.3

21.9 16.7

28.8 13.6 6.8

NS NS

NS NS

NS NS

0.9 0.9

NOTE. P values for control and cervical (C) versus lumbar (L) are for 1-week data only.

We assumed that affirmative responses to whether subjects experienced side effects would be related to the procedure. The study bias was to overreport rather than underreport any complication or side effect. For the immediate and 1-week interviews, the interviewer was instructed to record an affirmative response if the subject reported any increase in an existing complaint or a new complaint, no matter how mild or transient. To be recorded as a side effect after 3 months, a complaint had to have been persistent since the injection. Although we believed this assumption was reasonable because it related to procedural, immediate postprocedure, and 3-month reporting, we were concerned about the 1-week reporting. Many side effects—insomnia, headache, indigestion— are common. People with chronic low back pain, who are apt to be depressed, are more likely to experience a nocebo effect and answer “yes” when asked about various somatic complaints during the interview.32 This conclusion was supported by the fact that 97% of our control subjects responded affirmatively to one of the side effects listed on the questionnaire. Arch Phys Med Rehabil Vol 86, February 2005

To control for the nocebo effect, we determined the statistical significance between the control and 1-week study group. The control group, which had no intervention at the time of interview, was not followed and was not compared with the 3-month results. The control group was composed of new patients with problems similar to those of the SNRI group, but who had not received any intervention. Withholding treatment from the control group for 3 months was not, in our opinion, ethical. This does limit the study in that there was possible overreporting of side effects at 3 months in the study group. Side effects that occurred during the procedure included lightheadedness and nausea in less than 3% of the subjects. Only 1 vasovagal event occurred during the 306 injections (0.3%). This patient responded to Trendelenburg’s position and did not require fluid or pharmacologic resuscitation. In a retrospective study of lumbar SNRI, hypotension requiring treatment and vasovagal reaction occurred in 0.3% of injections.31 Cervical epidural steroid injections in retrospective studies caused nausea in 8%33 and vasovagal reactions or hypotension

COMPLICATIONS OF SELECTIVE NERVE ROOT INJECTIONS, Huston

in 1.4% to 13% of subjects.20,33,34 Hypotension in blinded lumbar epidural steroid injections occurred in 2.7% of patients in another study.6 Cardiopulmonary arrest has been reported with lumbar epidural steroid injection.35 In our study, 1 dural puncture occurred (.33%). Incidence of dural punctures with cervical epidural steroid injections ranges from 0.25% to 2.00%.34,36-38 For lumbar epidural steroid injections, the incidence ranges from 5.2% to 6.0%.1,39 With a dural puncture, the procedure is discontinued to avoid subarachnoid injection of corticosteroid or lidocaine. Inadvertent subarachnoid injection of local anesthetic may result in spinal anesthesia and subsequent respiratory depression, hypotension, and syncope.39,40 Dural puncture with inadvertent subarachnoid steroid injection has been thought to result in arachnoiditis.41-43 The effect of betamethasone (Celestone Chronodose) has been evaluated in sheep and found to cause arachnoiditis at suprapharmacologic dosages.44 Epidural steroid injection with methylprednisilone and triamcinolone has been evaluated in animals with no arachnoiditis or nerve root injury.43,45 We aborted the procedure when the dural puncture occurred, to avoid the risk of arachnoiditis and spinal anesthesia. Even when the procedure is discontinued, dural puncture may result in a spinal headache. Jurmand39 reported 183 dural punctures with 3544 injections. Headache occurred in only 33 subjects.39 In 2 other studies, 3 dural punctures occurred in 58 subjects who were receiving 113 cervical epidurals37 and 2 dural punctures occurred in 204 cervical epidural procedures in 142 subjects, with none reporting a headache.38 In a prospective study,36 2 dural punctures occurred in 790 cervical epidural procedures in 215 subjects, both of which lead to spinal headache. Our subject did not experience a spinal headache. Immediately after the procedure, a nonpositional headache occurred with 1.4% and 4.5% of the lumbar and cervical SNRIs, respectively. In a retrospective study31 of lumbar SNRI, nonpositional headache occurred in 3.1% of subjects. Transient headaches that occur during interlaminar or caudal epidurals have been attributed to rapid injection of medication.46,47 Another hypothesis is that an alteration occurs in cerebrospinal fluid pressures with the injection. The incidence of these transient headaches in lumbar and cervical epidurals has not been reported, to our knowledge. To minimize headache, the injectate should be introduced slowly into the epidural space. Immediately after SNRI, patients may experience increased axial or radicular pain at the injection site. The injection site was reported to be a source of pain in 17.1% and 22.7% of lumbar and cervical injections, respectively. Increased radicular pain occurred more commonly in cervical (18.2%) than in lumbar (8.8%) SNRIs. Incidence of spine pain postinjection was also more common in cervical (13.6%) compared with lumbar (5.1%) SNRIs. Retrospectively, lumbar SNRI had a 2.4% incidence of increased back pain and a 0.6% incidence of increased radicular pain.31 In a prospective study38 of cervical epidurals, 13.2% of the subjects complained of neck stiffness. Increased pain at the injection site was noted with 2.1% of the cervical epidurals.33 Nerve root injury occurred in 2 of 58 subjects who had cervical epidurals.37 Transient parasthesias was noted with one third of cervical epidurals.20 As previously stated, to our knowledge, no study in the English medical literature has prospectively evaluated the incidence of complications from lumbar epidurals. In a review18 of complications with lumbar epidurals, a weighted mean incidence of 1% for increased pain was reported. An increase in radicular pain with epidural injections has been attributed to

281

rapidly injecting medication or to use of large volumes of injectate.1,47 At 1-week evaluation, the SNRI group reported increased pain at the injection site; the control group could not be compared. The cervical SNRI group had a higher incidence of transient nonspecific and nonspinal headaches than did the control group. As previously stated, the exact cause of these headaches is unknown. At 3 months, 2 subjects reported increased spine pain, with 1 willing to repeat the injection. The patient who would not repeat the injection experienced resolution of his radicular pain secondary to herniated nucleus pulposus after cervical SNRI. However, he was dissatisfied with the continuing neck pain. Another subject complained of gastric distress that occurred within the first week after SNRI and was persistent. This subject was treated for gastroesophageal reflux disease. The risk of ulcer development with corticosteroid use is controversial.48 Steroid inhibition of phospholipase A2 increases the risk of ulcers, whereas inhibition of leukotrienes may be protective.49,50 Clinical studies50-52 have shown only rare development or no significant ulcer development in steroid usage. However, the risk of ulcer development with steroid use increases with concomitant nonsteroidal anti-inflammatory agents.52 This drug interaction should be considered in those who undergo steroid injection to decrease risk of gastrointestinal side effects. Prospectively, our subjects were interviewed for symptoms of Cushing’s syndrome. None of our 149 subjects had Cushing’s syndrome. This suggests that the incidence of Cushing’s syndrome at 3-month follow-up is less than 0.7% (1/149). Cushing’s syndrome has occurred transiently in lumbar epidural injections with dosages of 80mg of methylprednisolone given at 1-week intervals.8 In another case report,19 Cushing’s syndrome occurred after a cervical epidural injection of 60mg of methylprednisolone; the syndrome resolved after 12 months. Equivalent dosages of corticosteroid are used in SNRIs, and there is a risk that Cushing’s syndrome will result. To our knowledge, the incidence of Cushing’s syndrome after SNRI or epidural is unknown. None of our study subjects were diabetic; therefore, we cannot comment on the incidence of hyperglycemia postinjection. However, it is a known risk factor and should be addressed appropriately. Technical differences between SNRI and interlaminar epidurals may predispose subjects to different side effects. As noted, the incidence of hypotension and vasovagal reactions was lower for SNRIs. SNRI that targets the involved spinal nerve and dorsal root ganglion has the advantage of requiring less than half the amount of local anesthetic needed for nonspecific epidural injections. Volumes of lidocaine and bupivacaine in cervical epidurals may range from 2 to 20mL,33-37 and in lumbar or caudal epidurals from 5 to 30mL.18 Larger volumes of local anesthetic may increase the risk for cardiovascular depression, particularly in hypovolemic patients.53,54 Additionally, inadvertent intravascular injections with larger dosages of anesthetic agent could result in cardiopulmonary collapse.55 Use of fluoroscopic guidance with injection of nonionic contrast to confirm needle location reduces the risk of intravascular injection.7,56,57 Dural punctures were higher in the epidural studies and may be related to the fact that the procedure was not being performed under fluoroscopic guidance. SNRIs may have risks not associated with epidurals. By targeting the spinal nerve, there is a risk of nerve injury. In our study, there were no spinal nerve injuries. However, nerve injury has occurred with cervical epidurals.14,37 We found a Arch Phys Med Rehabil Vol 86, February 2005

282

COMPLICATIONS OF SELECTIVE NERVE ROOT INJECTIONS, Huston

higher incidence of pain at the injection site and of transient radicular pain. There is the potential for injury to the artery of Adamkiewicz with subsequent paralysis; our study had no cases of paralysis. SNRI has the advantage of targeting the pathology site in the treatment of radiculitis and may lead to improved outcomes. Overall, the procedure does not appear to place the patient at more risk than interlaminar epidurals. Future research with a prospective comparative trial of SNRIs and interlaminar epidurals is needed to answer this question definitively. Also, large multicenter trials are needed to determine the incidence of less common complications and side effects. CONCLUSIONS Our prospective study showed that the risk of serious complications after SNRI is low, and although various minor side effects may occur, typically they are transient. References 1. Dilke TF, Burry HC, Grahame R. Extradural corticosteroid injection in management of lumbar nerve root compression. Br Med J 1973;2:635-7. 2. Klenerman L, Greenwood R, Davenport HT, White DC, Peskett S. Lumbar epidural injections in the treatment of sciatica. Br J Rheumatol 1984;23:35-8. 3. Ridley MG, Kingsley GH, Gibson T, Grahame R. Outpatient lumbar epidural corticosteroid injection in the management of sciatica. Br J Rheumatol 1988;27:295-9. 4. Swerdlow M, Sayle-Creer W. A study of extradural medication in the relief of the lumbosciatic syndrome. Anaesthesia 1970;25: 341-5. 5. Williamson JA. Inadvertent spinal subdural injection during attempted spinal epidural steroid therapy. Anaesth Intensive Care 1990;18:406-8. 6. Berman AT, Garbarino JL, Fisher SM, Bosacco SJ. The effects of epidural injection of local anesthetics and corticosteroid on patients with lumbosciatic pain. Clin Orthop 1984;Sep(188): 144-51. 7. White AH, Derby R, Wynne G. Epidural injections for the diagnosis and treatment of low-back pain. Spine 1980;5:78-86. 8. Stambough JL, Booth RE, Rothman RH. Transient hypercorticism after epidural steroid injection. J Bone Joint Surg Am 1984;66: 1115-6. 9. Dougherty JH Jr, Fraser AR. Complications following intraspinal injections of steroids. J Neurosurg 1978;48:1023-5. 10. Gutknecht DR. Chemical meningitis following epidural injections of corticosteroid [letter]. Am J Med 1987;82:570. 11. Chan ST, Leung S. Spinal epidural abscess following steroid injection for sciatica. Case report. Spine 1989;14:106-8. 12. Goucke CR, Graziotti P. Extradural abscess following local anaesthetic and steroid injection for chronic low back pain. Br J Anaesth 1990;65:427-9. 13. Snoek W, Weber H, Jorgensen B. Double blind evaluation of extradural methylprednisolone for herniated lumbar discs. Acta Orthop Scand 1977;48:635-41. 14. Purkis IE. Cervical epidural steroids. Pain Clin 1986;1:3-7. 15. Simopoulos T, Peeters-Asdourian C. Pneumocephalus after cervical epidural steroid injection. Anesth Analg 2001;92:1576-7. 16. Williams KN, Jackowski A, Evans PJ. Epidural haematoma requiring surgical decompression following repeated cervical epidural steroid injections for chronic pain. Pain 1990;42:197-9. 17. Stoll A, Sanchez M. Epidural hematoma after epidural block: implications for its use in pain management. Surg Neurol 2002; 57:235-40. Arch Phys Med Rehabil Vol 86, February 2005

18. National Health and Medical Research Council. Epidural use of steroids in the management of back pain. Canberra (Aust): NHMRC; 1994. 19. Tuel SM, Meythaler JM, Cross LL. Cushing’s syndrome from epidural methylprednisolone. Pain 1990;40:81-4. 20. Catchlove RF, Braha R. Clinical reports: the use of cervical blocks in the management of chronic head and neck pain. Can Anaesth Soc J 1984;31:188-91. 21. White AH. Injection techniques for the diagnosis and treatment of low back pain. Orthop Clin North Am 1983;14:553-67. 22. Saal JS, Saal JA, Yurth EF. Nonoperative management of herniated cervical intervertebral disc with radiculopathy. Spine 1996; 21:1877-83. 23. Lutz GE, Vad VB, Wisneski RJ. Fluoroscopic transforaminal lumbar epidural steroids: an outcome study. Arch Phys Med Rehabil 1998;79:1362-6. 24. Vad VB, Bhat AL, Lutz GE, Cammisa F. Transforaminal epidural steroid injections in lumbosacral radiculopathy: a prospective randomized study. Spine 2002;27:11-6. 25. Riew KD. The effect of nerve-root injections on the need for operative treatment of lumbar radicular pain: a prospective, randomized, controlled, double-blind study. J Bone Joint Surg Am 2000;82:1589-93. 26. Kikuchi S, Macnab I, Moreau P. Localisation of the level of symptomatic cervical disc degeneration. J Bone Joint Surg Br 1981;63: 272-7. 27. Krempen JF, Smith BS. Nerve root injection: a method for evaluating the etiology of sciatica. J Bone Joint Surg Am 1974;56: 1435-44. 28. Schutz H, Lougheed WM, Wortzman G, Awerbuck BG. Intervertebral nerve-root in the investigation of chronic lumbar disc disease. Can J Surg 1973;16:217-21. 29. Kikuchi S, Hasue M. Combined contrast studies in lumbar spine diseases. Myelography (peridurography) and nerve root infiltration. Spine 1988;13:1327-31. 30. Slipman CW, Meyers JS, Chou LH, Sterenfeld EB, Abrams S. Complications of fluoroscopically-guided spinal injections [abstract]. Arch Phys Med Rehabil 1995;76:1032. 31. Botwin KP, Gruber RD, Bouchlas CG, Torres-Ramos FM, Freeman TL, Slaten WK. Complications of fluoroscopically guided transforaminal lumbar epidural injections. Arch Phys Med Rehabil 2000;81:1045-50. 32. Barsky AJ, Saintfort R, Rogers MP, Borus JF. Nonspecific medication side effects and the nocebo phenomenon. JAMA 2002; 287:622-7. 33. Shulman M. Treatment of neck pain with cervical epidural steroid injection. Reg Anesth 1986;11:92-4. 34. Ferrante FM, Wilson SP, Iacobo C, Orav EJ, Rocco AG, Lipson S. Clinical classification as a predictor of therapeutic outcome after cervical epidural steroid injection. Spine 1993;18:730-6. 35. Baron JF, Decaux-Jacolot A, Edouard A, Berdeaux A, Samii K. Influence of venous return on baroreflex control of heart rate during lumbar epidural anesthesia in humans. Anesthesiology 1986;64:188-93. 36. Waldman SD. Complications of cervical epidural nerve blocks with steroids: a prospective study of 790 consecutive blocks. Reg Anesth 1989;14:149-51. 37. Reiz S, Nath S. Cardiotoxicity of local anaesthetic agents. Br J Anaesth 1986;58:736-46. 38. Cicala RS, Westbrook L, Angel JJ. Side effects and complications of cervical epidural steroid injections. J Pain Symptom Manage 1989;4:64-6. 39. Jurmand SH. Corticotherapie peridurale des lombalgies et des sciatiques f’origine discale. EPU Rhum 1972;24:5061-70. 40. Brown FW. Management of diskogenic pain using epidural and intrathecal steroids. Clin Orthop 1977;Nov-Dec(129):72-8.

COMPLICATIONS OF SELECTIVE NERVE ROOT INJECTIONS, Huston

41. Bernat JL. Intraspinal steroid therapy. Neurology 1981;31:16871. 42. Nelson DA. Dangers from methylprednisolone acetate therapy by intraspinal injection. Arch Neurol 1988;45:804-6. 43. Cicala RS, Turner R, Moran E, Henley R, Wong R, Evans J. Methylprednisolone acetate does not cause inflammatory changes in the epidural space. Anesthesiology 1990;72:556-8. 44. Latham JM, Fraser RD, Moore RJ, Blumbergs PC, Bogduk N. The pathologic effects of intrathecal betamethasone. Spine 1997;22: 1558-62. 45. Delaney TJ, Rowlingson JC, Carron H, Butler A. Epidural steroid effects on nerves and meninges. Anesth Analg 1980;59:610-4. 46. Goebert HW, Jallo SJ, Wasmuth CE. Painful radiculopathy treated with epidural injections of procaine and hydrocortisone acetate: results in 113 patients. Anesth Analg 1961;40:130-4. 47. Burn JMB, Langdon L. Lumbar epidural injections for the treatment of chronic sciatica. Rheum Phys Med 1970;10:368-74. 48. Gotzsche PC. Steroids and peptic ulcer: an end to the controversy: J Intern Med 1994;236:619-32. 49. Pecora PG, Kaplan B. Corticosteroids and ulcers: is there an association? Ann Pharmacother 1996;30:870-2. 50. Conn HO, Poynard T. Corticosteroids and peptic ulcer: metaanalysis of adverse events during steroid therapy. J Intern Med 1994;236:599-601.

283

51. Messer J, Reitman D, Sacks HS, Smith H Jr, Chalmers TC. Association of adrenocorticosteroid therapy and peptic-ulcer disease. N Engl J Med 1983;309:21-4. 52. Piper JM, Ray WA, Daugherty JR, Griffin MR. Corticosteroid use and peptic ulcer disease: role of nonsteroidal anti-inflammatory drugs. Ann Intern Med 1991;114:735-40. 53. Bonica JJ, Berges PU, Morikawa K. Circulatory effects of peridural block: I. Effects of level of analgesia and dose of lidocaine. Anesthesiology 1970;33:619-26. 54. Bonica JJ, Kennedy WF, Akamatsu TJ, Gerbershagen HU. Circulatory effects of peridural block: III. Effects of acute blood loss. Anesthesiology 1972;36:219-27. 55. Lee PK, Kim JM. Lumbar epidural blocks: a case report of a lifethreatening complication. Arch Phys Med Rehabil 2000;81:1587-90. 56. Renfrew DL, Moore TE, Kathol MH, El-Khoury GY, Lemke JH, Walker CW. Correct placement of epidural steroid injections: fluoroscopic guidance and contrast administration. Am J Neuroradiol 1991;12:1003-7. 57. Furman MB, O’Brien EM, Zgleszewski TM. Incidence of intravascular penetration in transforaminal lumbosacral epidural steroid injections. Spine 2000;25:2628-32. Suppliers a. SAS Institute Inc, 100 SAS Campus Dr, Cary, NC 27513. b. Microsoft Corp, One Microsoft Way, Redmond, WA 98052.

Arch Phys Med Rehabil Vol 86, February 2005