Patients with preexisting spinal canal pathology, including spinal stenosis and lumbar disk disease (with or without prior spine surgery), are often not considered candidates for neuraxial blockade. For example, previous case reports have specifically attributed the presence of spinal stenosis or lumbar disk disease (with compressive radiculopathy) to an increased risk of neurologic complications. The mechanisms of injury are presumed to be ischemic,1–3 mechanical trauma,4,5 local anesthetic toxicity,6,7 or a multifactorial etiology. Large series and an increasing number of case reports suggest that undiagnosed spinal stenosis may be a significant contributor to neurologic complications after neuraxial block.3,5,6,8 Additional surgical factors such as intraoperative positioning, prolonged tourniquet ischemia, or high tourniquet inflation pressure may contribute in a synergistic (rather than simple additive) manner.9–11 Furthermore, it is unclear whether patients who have undergone prior spine surgery (e.g., laminectomy, diskectomy, spinal fusion) for these conditions are at additional risk of neurologic injury or block failure secondary to anatomic alterations or scarring of the central canal. Previous investigations have included small numbers of patients, typically with a history of extensive instrumentation or fusion who underwent epidural catheter placement for labor analgesia.12–15 However, the applicability of these results to other patient populations is unknown.
Therefore, the primary aim of this investigation was to examine the overall success and neurologic complication rates among patients with spinal stenosis or lumbar disk disease, with or without a history of prior spinal surgery, undergoing neuraxial blockade.
After IRB approval and patient consent, the medical records of all patients with a history of spinal stenosis or lumbar disk disease (i.e., compressive radiculopathy) who underwent a subsequent spinal or epidural anesthetic during a 15-year study period were retrospectively reviewed. Patients were identified if “spinal stenosis” or “lumbar disk disease” was entered on their Master Diagnosis List within the Mayo Clinic database. Neurologic diagnoses were limited to abnormalities of the spinal canal and did not include patients with primary central nervous systems disorders such as multiple sclerosis, amyotrophic lateral sclerosis, or postpolio syndrome. All neurologic diagnoses were confirmed by clinical examination and radiographic imaging by a neurologist or neurosurgeon before study inclusion. During the 15-year study period, multiple anesthesiologists (>15 faculty members) participated in the care of all patients.
Demographic data including age, gender, height, and weight were collected. The date of each neurologic diagnosis and details such as the presence of (1) motor deficits, (2) sensory deficits, (3) paresthesias or dysesthesias, or (4) hyperreflexia at the time of their spinal or epidural anesthetic were collected for each patient. Neurologic symptoms at the time of their procedure were further classified as (1) acute (exacerbation of symptoms within the last 30 days), (2) subacute (exacerbation of symptoms within the last 1 to 6 months), or (3) chronic/stable (no change in symptoms within the last 6 months).
Indications for neuraxial anesthesia or analgesia included surgical anesthesia, labor analgesia, or postoperative analgesia only. If a surgical indication occurred, it was classified as (1) orthopedic, (2) urologic, (3) general/abdominal, (4) cesarean delivery, or (5) other. Neuraxial blockade was categorized as (1) epidural, (2) single-injection spinal, (3) continuous spinal, or (4) combined spinal–epidural. Details of each neuraxial technique, including awake placement (yes or no), approach (midline, paramedian, both), number of attempts, and local anesthetic(s) used were collected. The use of epinephrine or other local anesthetic additives was also noted. Technical complications occurring at the time of block placement—such as difficulty identifying the epidural space, difficulty advancing an epidural or subarachnoid catheter, traumatic block placement (evidence of blood), unplanned dural puncture, difficulty obtaining cerebrospinal fluid, paresthesia elicitation, or unintended “total” or “high” spinal—were identified. Block efficacy was categorized as (1) satisfactory (surgery performed without additional intervention), (2) unilateral anesthesia or analgesia, (3) segmental or incomplete anesthesia or analgesia, or (4) no block/block failure.
New or progressive postoperative neurologic deficits (motor or sensory deficits, painful paresthesias, or bowel or bladder dysfunction) were identified in the daily progress notes of the primary surgical service or the anesthesia acute pain service. Complications were also noted during the patient's 2-, 4-, or 6-week surgical follow-up visit. The presence of infectious (neuraxial abscess) or hematologic (neuraxial hematoma) complications was also documented. All complications were followed until complete resolution or until the last documented date of evaluation.
Data are summarized using mean ± SD for continuous variables and percentages for categorical variables. The percentage of failed blocks and technical and neurologic complications was compared in patients with and without a history of spine surgery using the χ2 test or Fisher's exact test as appropriate. Patients with and without neurologic complications were compared to identify risk factors using Fisher's exact test. P < 0.05 was considered statistically significant.
Nine hundred thirty-seven (n = 937) patients were identified as having spinal stenosis or lumbar disk disease and undergoing subsequent neuraxial anesthesia or analgesia; 207 of these had a history of spinal surgery before undergoing subsequent spinal or epidural anesthesia (Table 1). Of these, 210 (22%) patients had a coexisting peripheral neuropathy in addition to their spinal canal pathology. At the time of neuraxial blockade, painful paresthesias or dysesthesias were the most common neurologic deficit, followed by nonspecific sensory and motor deficits (Table 1). The majority of patients (n = 335; 51%) had active neurologic signs or symptoms (motor or sensory deficits, dysesthesias/paresthesias, or hyperreflexia) at the time of surgery and subsequent neuraxial blockade.
The type of neuraxial blockade included spinal anesthesia in 545 (58%) patients, epidural anesthesia or analgesia in 358 (38%) patients, continuous spinal anesthesia in 24 (3%) patients, and a combined spinal–epidural technique in 10 (1%) patients. Eighty of the patients undergoing epidural anesthesia or analgesia noted above received postoperative epidural analgesia with local anesthetic infusions.
One hundred ninety-three (24%) patients underwent block placement precisely at the level of spinal canal pathology, and >75% of patients had block placement within 2 vertebral levels (Table 2). In addition, the majority of patients with a history of spine surgery (n = 165; 83%) underwent subsequent neuraxial anesthesia or analgesia within 2 vertebral levels of their prior spinal surgery.
There was no significant difference in block efficacy when comparing those patients with or without a history of spinal surgery (Table 3). Nine hundred eleven (97.2%) patients had a satisfactory block, 10 (1.1%) patients reported a patchy or segmental block, and 16 (1.7%) patients experienced complete block failure. Overall, there were 107 (11.4%) technical complications. There was no significant difference in technical complications comparing those patients with and without a history of spinal surgery (Table 3).
There were 10 (1.1%; 95% confidence interval [CI] 0.5%–2.0%) patients who experienced new deficits (3 patients) or worsening of preexisting symptoms (7 patients) postoperatively [Table 4 (cases 1 to 4), and Table 5 (cases 5 to 10)]. Three (1.4%) complications occurred in patients with a history of spinal surgery, and the remaining 7 (1.0%) in patients without prior surgical intervention. Eight of 10 patients with complications underwent an orthopedic surgical procedure; in 5 of the 6 cases in which the surgery was a unilateral lower extremity procedure, the postoperative deficit involved the operative side (cases 3, 4, and 8 to 10). Likewise, in both patients undergoing bilateral orthopedic procedures who developed bilateral deficits, the outcome was worse on the previously affected side (cases 2, 6). From exploratory analyses, the frequency of complications was found to be significantly higher in patients with a preoperative diagnosis of compressive radiculopathy versus other neurologic diagnoses (9/530 vs. 1/407; Fisher exact test P = 0.049) and also in patients with multiple versus single preoperative neurologic diagnoses (6/180 vs. 4/757; Fisher exact test P = 0.005).
A surgical cause was presumed to be the primary etiology in 4 (40%) of 10 patients (Table 4). Specifically, direct nerve trauma/stretch was the principal mechanism in 3 patients (cases 1, 2, 4), tourniquet ischemia in 2 patients (cases 2, 4), and ischemia from a popliteal cyst in 1 patient (case 3). However, it is possible that neurotoxicity produced by local anesthetic injection at the site of existing root compromise may have contributed in 3 cases (cases 2 to 4), because the postoperative deficits were a worsening of existing deficits. In 3 (75%) of the 4 patients, the neurologic deficits returned to baseline (complete resolution did not occur); in the remaining patient (25%), the symptoms persisted for 3 years (time of last patient follow-up).
The primary etiology of the remaining 6 (60%) complications was judged to be nonsurgical (including anesthetic-related factors) (Table 5). Contributing factors were identified in 3 of these patients and included a previously undiagnosed L2 to L3 ependymoma that became symptomatic after failed epidural anesthesia (likely resulting from needle trauma/bleeding of the neoplasm or alterations in cerebrospinal fluid flow) (case 5), compartment syndrome (with a delay in the diagnosis and worsening of outcome attributed to epidural infusion) (case 6), and neuropathy of critical illness (case 10). The remaining 3 patients, each with a preoperative diagnosis of spinal stenosis, underwent an uneventful surgery under spinal anesthesia. In these 3 patients (cases 7 to 9), the spinal anesthetic, in the setting of a preexisting neurologic condition, was determined to be the primary etiology, because no positioning or surgical factors were identified. Thus, in 5 of 6 patients with a neurologic complication attributed to a nonsurgical primary etiology, the neuraxial anesthetic was presumed to be the primary mechanism (cases 7 to 9) or a contributing factor in the outcome (cases 5, 6). Decompressive laminectomy was performed in 2 of 6 patients with deficits associated with a nonsurgical etiology (1 for ependymoma; 1 for radiculopathy/spinal stenosis). However, complete resolution was documented in only 2 (33%) of the 6 patients (1 of whom underwent laminectomy; case 5).
Patients with preexisting neurologic conditions are typically considered to be at increased risk of neurologic complications after spinal or epidural techniques. Previous studies have investigated the risk associated with multiple sclerosis or polio16 or diabetic sensorimotor neuropathy17 and concluded that the risk of severe postoperative neurologic injury in these patients is relatively uncommon but appears to be higher than that reported for the general population.6,18 Our series is the first to characterize the frequency and severity of neurologic events after neuraxial blockade in patients with previously diagnosed spinal stenosis or lumbar disk disease (with or without prior spinal decompressive surgery). Our major findings suggest that this patient population is at increased risk for worsening of preexisting or the development of new neurologic deficits postoperatively when compared with the general population,6,18 and that the presence of multiple neurologic diagnoses (radiculopathy, spinal stenosis, peripheral neuropathy) increases the risk.
Our results also demonstrated that a history of spinal surgery did not increase the risk of technical or neurologic complications or affect block success. Previous spinal surgery has often been considered a relative contraindication to the use of neuraxial blockade. Many of these patients experience chronic back pain and are reluctant to undergo epidural or spinal anesthesia, fearing exacerbation of their preexisting back complaints. Several postoperative anatomic changes make needle or catheter placement more difficult and complicated after major spinal surgery. The presence of adhesions or obliteration of the epidural space from scar tissue may increase the incidence of dural puncture or decrease the spread of local anesthetic within the epidural space, producing an incomplete or failed block. Needle placement in an area of the spine that has undergone bone grafting and posterior fusion may not be possible with midline or lateral approaches; needle insertion can be accomplished only at unfused segments.
Previous series investigating the efficacy and safety of neuraxial block after spinal surgery typically involved parturients with a history of Harrington rod instrumentation and fusion. For example, the largest series by Daley et al.19 included 18 patients with previous Harrington rod instrumentation who underwent 21 attempts at epidural anesthesia for obstetric analgesia. Continuous lumbar epidural anesthesia was successfully established in 20 of 21 attempts, but only 10 procedures were performed easily on the first attempt. The remaining 11 patients required larger amounts of local anesthetics or complained of a patchy block or both. There were no side effects except for low back pain in 2 patients with multiple attempts at catheter placement. Similar results were reported by Crosby and Halpern12 and Hubbert.14 Thus, historically, it was concluded that epidural anesthesia may be successfully performed in patients who have had spinal surgery, but successful catheter placement may be possible on the first attempt in only 50% of patients, even by an experienced anesthesiologist. Although adequate epidural anesthesia was eventually produced in 40% to 95% of patients, there appeared to be a higher incidence of traumatic needle placement, unintentional dural puncture, and unsuccessful epidural needle or catheter placement, especially if spinal fusion extends to between L-5 and S-1. However, the small numbers of patients and an indication of “labor analgesia” made extrapolation to patients with other surgical indications undergoing a nonepidural technique difficult.
A single study by Berkowitz and Gold20 evaluated the success rate and complications in 33 patients with prior laminectomy who underwent tetracaine spinal anesthesia. On the basis of their 100% success rate and lack of neurologic complications, including evidence of a prolonged block, the authors concluded that there was “no logical basis for avoiding spinal anesthesia in postlaminectomy patients.” Our 97.6% block success rate, trauma frequency of 3.9%, and paresthesia elicitation rate of 4.3% in patients with prior spine surgery support the conclusions of Berkowitz and Gold regarding block efficacy after laminectomy. Furthermore, their frequency of technical complications was similar to those for patients in our series (Table 3) and previous series from our outcomes database.21 Unfortunately, the frequency of persistent postoperative neurologic deficits in the current retrospective series (1.1%; 95% CI 0.5%–2.0%) is much higher than was expected; prospective epidemiologic investigations have reported frequencies between 1:1000 and 1:10,000.6,22 Although the neuraxial block was not the primary etiology of all 10 neurologic deficits in our series, it may have been a contributing factor in nearly all the cases because of the “double crush” phenomenon.
The double crush syndrome, first proposed by Upton and McComas11 in 1973, is a general term referring to the coexistence of dual compressive lesions along the course of a nerve, where the presence of a proximal lesion renders the nerve vulnerable to further injury with a more distal compression. Importantly, the 2 minor (perhaps even subclinical) insults synergistically result in a clinical and potentially permanent nerve injury.23 Although initially described in patients with concomitant cervical radiculopathies and median or ulnar neuropathies, the term has been extended to include injury from noncompressive mechanisms, such as toxic (chemotherapeutic agents, local anesthetics),24 metabolic (aging, diabetic sensorimotor neuropathy),17,25,26 ischemic (tourniquet inflation),27 and traumatic (surgical traction, needle/catheter placement)28 etiologies. The presence of a preexisting neurologic condition, whether neuraxial or peripheral in nature, would represent an additional mechanism by which minor or subclinical symptoms may interact to result in a new or worsened neurologic deficit.6,17 Our results suggest an additive effect of multiple preexisting neurologic conditions; we reported that neuraxial block performed in patients with a preoperative diagnosis of a compressive radiculopathy or multiple neurologic diagnoses significantly increased the risk of neurologic complications postoperatively. Previous surgical treatment did not affect the frequency.
The performance of a neuraxial technique in patients with a preexisting neurologic condition may predispose the patient to further functional worsening. The mechanism(s) of nerve injury after neuraxial block in patients with spinal stenosis or lumbar disk disease is presumed to be the additive effects of local anesthetic toxicity and neural ischemia (from the volume effects of the injectate/infusate)1,6,8,29 with epidural techniques and local anesthetic neurotoxicity with spinal blockade. The neurotoxic effects may be enhanced by maldistribution created by degenerative changes within the spinal column.30 Spinal cord ischemia from microcirculatory derangement has also been proposed.3,31 Perioperative positioning may also contribute to injury; postoperative cauda equina syndrome and paresis have been reported in previously asymptomatic patients with spinal stenosis after uneventful spinal6,8 or general anesthesia.32 Often these cases involve an orthopedic procedure.8,32
Patients undergoing orthopedic surgical procedures may be especially prone to the double crush syndrome. Two or more insults along the course of a nerve may readily occur, because the surgical procedure is often associated with a defined postoperative neuropathy (due to surgical neurapraxia), and tourniquet inflation is routinely used to assure a bloodless field. For example, total knee replacement is associated with an overall 2.2% frequency of peroneal nerve palsy (from surgical traction to the peroneal nerve at the fibular head), which increases to 7.7% with prolonged tourniquet inflation (producing ischemia to both the tibial and peroneal components of the sciatic nerve at thigh level).9,27 A preexisting L5/S1 radiculopathy may render the sciatic components even more sensitive to ischemia and traction, as well as to local anesthetic neurotoxicity. Although an orthopedic procedure was not a risk factor in our series, in 7 of the 8 orthopedic cases with complications, surgery was to the previously affected extremity. Although the relative roles of patient, surgical factors, and anesthetic factors remain undetermined, our results suggest that the overall risk may be high in this patient population.
We conclude that patients with preexisting spinal canal pathology have a higher incidence of neurologic complications after neuraxial blockade than that previously reported for patients without such underlying pathology. However, in the absence of a control group of surgical patients with similar anatomic pathology undergoing general anesthesia, we cannot determine whether the higher incidence of neurologic injury is secondary to the surgical procedure, the anesthetic technique, the natural history of spinal pathology, or a combination of factors and the relative contributions of each.
1. Usubiaga JE, Wikinski JA, Usubiaga LE. Epidural pressure and its relation to spread of anesthetic solutions in epidural space. Anesth Analg 1967;46:440–6
2. Hooten WM, Hogan MS, Sanemann TC, Maus TJ. Acute spinal pain during an attempted lumbar epidural blood patch in congenital lumbar spinal stenosis and epidural lipomatosis. Pain Physician 2008;11:87–90
3. de Seze MP, Sztark F, Janvier G, Joseph PA. Severe and long-lasting complications of the nerve root and spinal cord after central neuraxial blockade. Anesth Analg 2007;104:975–9
4. Tetzlaff JE, Dilger JA, Wu C, Smith MP, Bell G. Influence of lumbar spine pathology on the incidence of paresthesia during spinal anesthesia. Reg Anesth Pain Med 1998;23:560–3
5. Stambough JL, Stambough JB, Evans S. Acute cauda equina syndrome after total knee arthroplasty as a result of epidural anesthesia and spinal stenosis. J Arthroplasty 2000;15:375–9
6. Moen V, Dahlgren N, Irestedt L. Severe neurological complications after central neuraxial blockades in Sweden 1990–1999. Anesthesiology 2004;101:950–9
7. Yuen EC, Layzer RB, Weitz SR, Olney RK. Neurologic complications of lumbar epidural anesthesia and analgesia. Neurology 1995;45:1795–801
8. Kubina P, Gupta A, Oscarsson A, Axelsson K, Bengtsson M. Two cases of cauda equina syndrome following spinal–epidural anesthesia. Reg Anesth 1997;22:447–50
9. Horlocker TT, Hebl JR, Gali B, Jankowski CJ, Burkle CM, Berry DJ, Zepeda FA, Stevens SR, Schroeder DR. Anesthetic, patient, and surgical risk factors for neurologic complications after prolonged total tourniquet time during total knee arthroplasty. Anesth Analg 2006;102:950–5
10. Warner MA, Warner DO, Harper CM, Schroeder DR, Maxson PM. Lower extremity neuropathies associated with lithotomy positions. Anesthesiology 2000;93:938–42
11. Upton AR, McComas AJ. The double crush in nerve entrapment syndromes. Lancet 1973;2:359–62
12. Crosby ET, Halpern SH. Obstetric epidural anaesthesia in patients with Harrington instrumentation. Can J Anaesth 1989;36:693–6
13. Feldstein G, Ramanathan S. Obstetrical lumbar epidural anesthesia in patients with previous posterior spinal fusion for kyphoscoliosis. Anesth Analg 1985;64:83–5
14. Hubbert CH. Epidural anesthesia in patients with spinal fusion. Anesth Analg 1985;64:843
15. Kardash K, King BW, Datta S. Spinal anaesthesia for caesarean section after Harrington instrumentation. Can J Anaesth 1993;40:667–9
16. Hebl JR, Horlocker TT, Schroeder DR. Neuraxial anesthesia and analgesia in patients with preexisting central nervous system disorders. Anesth Analg 2006;103:223–8
17. Hebl JR, Kopp SL, Schroeder DR, Horlocker TT. Neurologic complications after neuraxial anesthesia or analgesia in patients with preexisting peripheral sensorimotor neuropathy or diabetic polyneuropathy. Anesth Analg 2006;103:1294–9
18. Auroy Y, Benhamou D, Bargues L, Ecoffey C, Falissard B, Mercier FJ, Bouaziz H, Samii K. Major complications of regional anesthesia in France: The SOS Regional Anesthesia Hotline Service. Anesthesiology 2002;97:1274–80
19. Daley MD, Rolbin SH, Hew EM, Morningstar BA, Stewart JA. Epidural anesthesia for obstetrics after spinal surgery. Reg Anesth 1990;15:280–4
20. Berkowitz S, Gold MI. Spinal anesthesia for surgery in patients with previous lumbar laminectomy. Anesth Analg 1980;59:881–2
21. Horlocker TT, McGregor DG, Matsushige DK, Schroeder DR, Besse JA. A retrospective review of 4767 consecutive spinal anesthetics: central nervous system complications. Anesth Analg 1997;84:578–84
22. Auroy Y, Narchi P, Messiah A, Litt L, Rouvier B, Samii K. Serious complications related to regional anesthesia: results of a prospective survey in France. Anesthesiology 1997;87:479–86
23. Prielipp RC, Warner MA. Perioperative nerve injury: a silent scream? Anesthesiology 2009;111:464–6
24. Hebl JR, Horlocker TT, Pritchard DJ. Diffuse brachial plexopathy after interscalene block in a patient receiving cisplatin chemotherapy: the pharmacologic double crush syndrome. Anesth Analg 2001;92:249–51
25. Moghtaderi A, Izadi S. Double crush syndrome: an analysis of age, gender and body mass index. Clin Neurol Neurosurg 2008;110:25–9
26. Yee TC, Kalichman MW. Effects of aging on nerve conduction block induced by bupivacaine and procaine in rats. J Peripher Nerv Syst 1997;2:175–9
27. Horlocker TT, Cabanela ME, Wedel DJ. Does postoperative epidural analgesia increase the risk of peroneal nerve palsy after total knee arthroplasty? Anesth Analg 1994;79:495–500
28. Hebl JR, Horlocker TT, Sorenson EJ, Schroeder DR. Regional anesthesia does not increase the risk of postoperative neuropathy in patients undergoing ulnar nerve transposition. Anesth Analg 2001;93:1606–11
29. Jacob AK, Borowiec JC, Long TR, Brown MJ, Rydberg CH, Wass CT. Transient profound neurologic deficit associated with thoracic epidural analgesia in an elderly patient. Anesthesiology 2004;101:1470–1
30. Hampl KF, Schneider MC, Drasner K. Toxicity of spinal local anaesthetics. Curr Opin Anaesthesiol 1999;12:559–64
31. Markman JD, Gaud KG. Lumbar spinal stenosis in older adults: current understanding and future directions. Clin Geriatr Med 2008;24:369–88
© 2010 International Anesthesia Research Society
32. Lewandrowski KU, McLain RF, Lieberman I, Orr D. Cord and cauda equina injury complicating elective orthopedic surgery. Spine 2006;31:1056–9