Share this article on:

Small Risk of Serious Neurologic Complications Related to Lumbar Epidural Catheter Placement in Anesthetized Patients

Horlocker, Terese T. MD; Abel, Martin D. MD; Messick, Joseph M. Jr MD; Schroeder, Darrell R. MS

doi: 10.1213/01.ANE.0000057600.31380.75

Previous studies have identified pain during needle/catheter placement or during the injection of local anesthetic as a risk factor for the development of persistent paresthesias after regional anesthetic techniques. The performance of regional blockade on anesthetized patients theoretically increases the risk of postoperative neurologic complications, because these patients are unable to respond to painful stimuli. In this study, we evaluated the frequency of neurologic complications in 4298 thoracic surgical patients undergoing lumbar epidural catheter placement while under general anesthesia. Catheters were placed immediately after the induction and tracheal intubation or on completion of the surgical procedure, before emergence. Most epidural catheters (4220, or 98.2%) were used solely for postoperative analgesia; only 78 (1.8%) epidural catheters were used for intraoperative anesthesia. In 4239 (98.6%) patients, an opioid alone was administered. The remaining 56 (1.3%) patients received a local anesthetic or local anesthetic/opioid mixture epidurally. Analgesia was graded as excellent or good in 92.2% of patients. Side effects included sedation in 455 (10.6%), nausea or emesis in 328 (7.6%), pruritus in 116 (2.7%), and respiratory depression (pH ≤7.3 and PaCO2 ≥50 mm Hg) in 308 (7.2%) patients. The mean duration of epidural analgesia was 2.4 ± 0.8 days (range, 0–10.7 days). There were no neurologic complications, including spinal hematoma, epidural abscess or catheter site infections, radicular symptoms, or persistent paresthesias (95% confidence interval, 0%–0.08%). In one patient, the epidural catheter broke during removal, and a portion was retained. The patient was notified; no long-term sequelae were noted. Six patients developed new neurologic symptoms or postoperative worsening of a previous neurologic condition unrelated to epidural catheterization. We conclude that the risk of neurologic complications associated with lumbar epidural catheter placement in anesthetized patients is small. However, the relative risk of this practice, compared with epidural catheter placement in awake patients, is unknown.

IMPLICATIONS: We report no neurologic complications in 4298 patients undergoing epidural catheter placement while under general anesthesia. Although the risk of neurologic complications associated with lumbar epidural catheter placement in anesthetized patients is small, the relative risk compared with epidural catheterization in awake patients is unknown.

Departments of *Anesthesiology and

†Health Sciences Research, Mayo Clinic, Rochester, Minnesota

Address correspondence and reprint requests to Terese T. Horlocker, MD, Department of Anesthesiology, Mayo Clinic, 200 First St. Southwest, Rochester, MN 55905. Address e-mail to

Accepted January 08, 2003

Perioperative nerve injuries have long been recognized as a complication of regional anesthesia. Risk factors contributing to neurologic deficit after regional techniques include neural ischemia, traumatic injury to nerves during needle or catheter placement, infection, and local anesthetic toxicity (1,2). A prospective survey in France evaluated the incidence and characteristics of serious complications after 103,730 regional anesthetics (1). Neurologic complications related to the regional technique occurred in 6 (0.02%) of 30,413 (95% confidence interval, 0.004%–0.036%) patients undergoing epidural anesthesia. In all five cases of radiculopathy after epidural block, the patient reported either a paresthesia during needle insertion or pain with injection. It is important to note that the postoperative (persistent) paresthesia had the same topography as the paresthesia elicited during block performance. Similar observations were reported by Cheney et al. (2) in a review of the American Society of Anesthesiologists Closed Claims Project database. Ninety-three percent of the 67 lumbosacral nerve root injuries were associated with the administration of spinal or epidural anesthesia, including one patient who underwent multiple attempts at lumbar epidural block while under general anesthesia. Associated risk factors were paresthesia during needle/catheter placement, pain during injection of local anesthetic, or multiple attempts to perform a block.

The performance of regional blockade in an anesthetized patient is controversial because of the possibility of unrecognized nerve injury from needle/catheter placement and local anesthetic injection. Supportive data are limited, however. Cases are typically reported individually (3–5) and after resolution of associated medicolegal issues (2). In addition, although several large studies have evaluated the risk of performing regional blockade on anesthetized children, these results are not immediately applicable to the adult population because of the differences in anatomy, block selection, and local anesthetic solution injected (6–9). The risk of nerve injury in adult patients undergoing regional techniques while anesthetized or heavily sedated remains undetermined. This study evaluates the frequency of neurologic complications after lumbar epidural catheter placement in anesthetized adult patients.

Back to Top | Article Outline


After IRB approval, we identified through the Acute Pain Service (APS) database all patients who received postoperative epidural analgesia after abdominal or thoracic surgical procedures between January 1991 and September 1997. Only patients who underwent lumbar epidural catheter placement while anesthetized were included. Epidural catheters were placed immediately after the induction and tracheal intubation or on completion of the surgical procedure, before emergence.

The APS database was prospectively generated by daily entry of information of patients receiving care by the APS. At the time of epidural placement, patient demographics, surgical site and pathologic diagnosis, and level of epidural catheter placement were recorded in the APS database. The APS visited patients as often as needed to provide care, but at least twice daily. During each patient visit, physiologic variables, patient comfort level, epidural solution (local anesthetic, opioid, or combination), infusion rate, any changes in the epidural solution/rate of infusion, pertinent laboratory data (arterial blood gases and coagulation tests), and the use of adjunctive analgesic drugs were noted. Side effects such as sedation, nausea, pruritus, and respiratory depression (PaCO2 ≥50 mm Hg and pH ≤7.30 in a spontaneously breathing, nonventilated patient) were recorded. The treatment of side effects and any epidural-related complications were noted. The epidural catheter was aspirated and the site inspected once daily. Patients were examined daily for evidence of neurologic complications resulting from infection, needle- or catheter-induced trauma, or spinal bleeding. The date and reason for catheter removal were recorded. An anonymous questionnaire was distributed to patients on catheter removal. Patients were asked to rate the quality of analgesia that had been provided as excellent, good, fair, or poor.

These prospectively collected data were retrospectively reviewed to evaluate the frequency of neurologic complications among study patients. To further ensure that all neurologic complications were discovered, the medical records of the study patients were studied to determine which patients had undergone a neurologic consultation within 30 days of surgery. The records of these patients were reviewed to establish the nature and etiology of the event precipitating the neurologic consultation. Data are reported as mean ± SD (range).

Back to Top | Article Outline


There were 4298 patients: 1707 (39.7%) women and 2591 (60.3%) men. The mean age was 57.7 ± 18.6 yr (range, 6 mo–94 yr). There were 4353 surgical procedures. (Some patients underwent more than one surgical procedure with the same epidural catheterization.) Of these, 4108 (94.4%) procedures were primarily thoracic, whereas 245 (5.6%) were primarily cardiac. Patients who underwent an Ivor-Lewis esophagogastrectomy (459, or 10.5%) required an abdominal incision in addition to a thoracotomy. Patients undergoing thoracotomy (4054, or 94.3%) typically received heparin 5000 U subcutaneously 2 h before surgery and then every 12 h while immobilized. Of the 2782 (64.7%) patients in whom the diagnosis of malignancy was made, there were 1691 (39.3%) patients with primary lung cancer and 593 (13.8%) with metastatic lung disease. In the remaining 1516 (35.3%) patients, surgery was performed for benign processes.

Epidural catheterization occurred at the lumbar level in all patients. Most epidural catheters (4220, or 98.2%) were used solely for postoperative analgesia; only 75 (1.7%) epidural catheters were used for intraoperative anesthesia. Solutions administered epidurally are reported in Table 1. A solution containing opioid (fentanyl or morphine) alone was used in 4239 (98.6%) patients. In the remaining 56 (1.3%) patients, a local anesthetic or local anesthetic/opioid mixture was administered epidurally.

Table 1

Table 1

Side effects were reported in 1207 (28.1%) patients and included sedation in 455 (10.6%), nausea or emesis in 328 (7.6%), pruritus in 116 (2.7%), and respiratory depression in 308 (7.2%) patients (Table 2). A total of 34 (0.8%) patients required reintubation of the trachea for respiratory failure. However, in only one case was respiratory failure attributed to sedation and/or respiratory depression from the epidural. Side effects were treated with antihistamines, antiemetics, anticholinergics, naloxone, an adjustment in the epidural concentration or infusion rate, or a combination of therapies (Table 2). Scopolamine was often used as a primary treatment of postoperative nausea and vomiting (10). Of the 4259 patients who returned the APS questionnaire, analgesia was graded as excellent or good in 92.2% of patients.

Table 2

Table 2

The mean duration of epidural analgesia was 2.4 ± 0.8 days (range, 0–10.7 days). The epidural catheter was removed after the anticipated duration of analgesia in 3080 (71.7%) patients. Other reasons for removal of the epidural catheter were technical problems (dislodgement, occlusion, migration, or leakage of the catheter) in 420 (9.8%) patients, inadequate analgesia in 117 (2.7%) patients, and fever in 90 (2.1%) patients (Table 3).

Table 3

Table 3

There were no documented neurologic complications, including spinal hematoma, epidural abscess, or catheter site infections, or radicular symptoms related to epidural catheterization. In one patient, after an apparently uneventful catheter removal, the catheter tip was noted to be “not intact.” The patient was notified of the retained catheter fragment. No long-term sequelae were noted, although the patient died 18 mo later of her underlying malignancy. Six patients developed new neurologic symptoms or postoperative worsening of a previous neurologic condition. In all six patients, formal neurologic consultation concluded that epidural catheterization was not the primary cause of the neurologic deficits (Table 4).

Table 4

Table 4

Back to Top | Article Outline


The safety and efficacy of epidural opioid analgesia have been well documented (11). During the study period, patients in our institution scheduled for thoracic surgery underwent lumbar epidural catheter placement either before induction, under general anesthesia or after surgery in the recovery room; the timing and site of epidural catheter placement were at the discretion of the attending anesthesiologist. None of the 4298 patients in our series who underwent lumbar catheter placement while anesthetized developed a neurologic complication related to lumbar epidural catheter placement or epidural infusion.

The performance of regional blockade on anesthetized patients theoretically increases the risk of perioperative neurologic complications, because these patients are unable to respond to the paresthesias related to needle or catheter placement or to the pain of intraneural injection (1,2). However, most children who undergo regional anesthetic techniques are either heavily sedated or under general anesthesia (6–9). The largest prospective study evaluating the morbidity of regional anesthesia in children was performed by Giaufre et al. (6). There were 24,409 regional blocks; 89% were performed under general anesthesia. Neuraxial blocks, most of which were caudal blocks, accounted for more than 60% of all regional anesthetics. All 23 complications occurred after neuraxial block, for an overall incidence of 1.5 per 1000 central blocks. Half of the 23 reported complications (four spinals, six intravascular injections, and two transient paresthesias) may have been influenced by the patient’s alertness during the performance of the neural blockade. None resulted in long-term sequelae or medicolegal action. However, because detailed sensory and motor examinations were not conducted and because pediatric patients may not describe paresthesias, it is possible that cases of minor neurologic injury were missed. Alternatively, nerve injuries may resolve more rapidly and completely in children, resulting in a decreased incidence after regional techniques. This is supported by the ASA Closed Claims Project, which notes a striking paucity of claims for pediatric nerve injury (1% of all claims) in comparison to nerve injury in adults (16% of all claims) (2,12). The decreased incidence of neurologic claims in children, regardless of the etiology, makes extrapolation of the pediatric literature to adults problematic.

Although the risk of nerve injury associated with regional anesthesia and analgesia in anesthetized adults has not been formally evaluated, several studies have investigated the safety of subarachnoid needle placement for the administration of intrathecal opioids or removal of cerebrospinal fluid among these patients (13,14). Gwirtz et al. (13) assessed the efficacy and safety of 5969 intrathecal opioid injections, 98% of which were performed at the end of surgery while the patient was still anesthetized. Patients received preservative-free morphine (with or without fentanyl) at the L3-4 or L4-5 interspace. In 2.5% of patients, a local anesthetic was administered with the opioid. Patients were observed for 24 hours. There were no neurologic complications. Likewise, there were no neurologic deficits attributed to spinal drainage in a series of 478 patients reported by Grady et al. (14). All malleable needles or spinal drainage catheters were inserted at the lumbar level under general anesthesia after tracheal intubation. Although more than 90% of patients underwent either cerebrospinal fluid removal or air injection, no intrathecal medications were administered.

There are similarities in our methodology and the methodologies of Gwirtz et al. (13) and Grady et al. (14) that may have contributed to the absence of serious neurologic complications among these three different populations. The site of needle or catheter insertion was at the lumbar level in nearly all patients. However, because clinicians are often inaccurate in their assessment of vertebral level and because there is natural variability in the levels at which the conus medullaris and dural sac end, it is probable that needle or catheter placement occasionally occurred at a level cephalad to the termination of the spinal cord (15). The likelihood of nerve injury due to neurotoxicity was also minimized; a local anesthetic was administered in <2% of patients in the combined series, and the remaining patients received opioids or no medication intrathecally or epidurally. The small potential for opioid neurotoxicity has been well documented in the laboratory and in extensive clinical use (16). For example, cauda equina syndrome has been reported after uneventful spinal anesthesia (1). However, there are no cases of neurologic impairment related to neuraxial opioids administered in awake or anesthetized patients (16). Although nerve trauma, interneuronal injection, and neurotoxicity are the etiologies of nearly all neurologic complications (1), the relative contribution of each to postoperative neurologic dysfunction is unknown. We speculate that although paresthesias may be unknowingly elicited during needle and catheter placement in an anesthetized patient, the injection of a nonneurotoxic (opioid) drug is less likely to accentuate neuronal injury and result in clinically significant neurologic complications.

Although patients were evaluated daily for neurologic complications, it is possible that some neurodeficits may have been overlooked, particularly if the nerve injury was minor or of short duration. However, it is unlikely that a significant complication would have remained undiagnosed. All neurologic complications in the study by Auroy et al. (1) were noted within 48 hours. This would have been during the period of epidural catheterization in our patients. Moreover, the frequency of our other side effects, such as sedation, nausea, and respiratory depression, was similar to those of previous prospective analyses (11,17). The reliability of our database is also supported by the report of the retained epidural catheter tip. Additionally, cross-referencing our APS database with the neurologic consultation files allowed us to verify that there were no major regional analgesia-related complications. Rather, postoperative neurologic complications were associated with progression of a preexisting neurologic condition, perioperative positioning, or the surgical procedure.

It is essential to note that even with a zero incidence of neurologic complications among our 4298 patients, the possibility of a serious neurologic complication may still be as frequent as 0.08% (95% confidence interval). Although this frequency appears comparable to those of previous studies (1,17), the lack of a comparative database (epidural catheter placement in awake patients for postoperative opioid analgesia) makes it difficult to definitively evaluate the relative risks and benefits of the practice. It is also important to note that there were no spinal hematomas among the 4108 patients who underwent a thoracotomy and received standard (unfractionated) subcutaneous heparin 5000 U twice daily as thromboprophylaxis. The presence of an indwelling epidural catheter in combination with the administration of standard subcutaneous heparin is supported by previous reviews and recommendations (18).

In summary, our study documented the small frequency of serious neurologic complications associated with epidural catheter placement for postoperative opioid infusion in anesthetized patients. Although our results are reassuring, they must be cautiously interpreted. Our outcomes should not be extrapolated to the performance of thoracic epidural block or the administration of epidural local anesthetic solutions, because our methods address neither. Furthermore, these results should not be applied to peripheral or plexus blocks; four cases of permanent spinal cord injury in patients who underwent interscalene block while under general anesthesia have been reported (4). Therefore, the decision to perform a regional anesthetic on any heavily sedated or anesthetized patient should take into account whether the benefits of this approach are greater than the risk of a rare and catastrophic, but potentially preventable, outcome.

1. Auroy Y, Narchi P, Messiah A, et al. Serious complications related to regional anesthesia. Anesthesiology 1997; 87: 479–86.
2. Cheney FW, Domino KB, Caplan RA, Posner K. Nerve injury associated with anesthesia: a closed claims analysis. Anesthesiology 1999; 90: 1062–9.
3. Bromage PR, Benumof JL. Paraplegia following intracord injection during attempted epidural anesthesia under general anesthesia. Reg Anesth Pain Med 1998; 23: 104–7.
4. Benumof JL. Permanent loss of cervical spinal cord function associated with interscalene block performed under general anesthesia. Anesthesiology 2000; 93: 1541–4.
5. Lang SA. Postoperative analgesia following total knee arthroplasty: a study comparing spinal anesthesia and combined sciatic femoral 3-in-1 block. Reg Anesth Pain Med 1999; 24: 97.
    6. Giaufre E, Dalens B, Gombert A. Epidemiology and morbidity of regional anesthesia in children: a one-year prospective survey of the French-Language Society of Pediatric Anesthesiologists. Anesth Analg 1996; 83: 904–12.
    7. Wood CE, Goresky GV, Klassen KA, et al. Complications of continuous epidural infusions for postoperative analgesia in children. Can J Anaesth 1994; 41: 613–20.
      8. McNeely JK, Farber NE, Rusy LM, Hoffman GM. Epidural analgesia improves outcome following pediatric fundoplication: a retrospective analysis. Reg Anesth 1997; 22: 16–23.
        9. Tobias JD, Lowe S, O’Dell N, Holcomb GW. Thoracic epidural anaesthesia in infants and children. Can J Anaesth 1993; 40: 879–82.
          10. Kranke P, Morin AM, Roewer N, et al. The efficacy and safety of transdermal scopolamine for the prevention of postoperative nausea and vomiting: a quantitative systematic review. Anesth Analg 2002; 95: 133–43.
          11. Sandler AN, Stringer D, Panos L, et al. A randomized, double-blind comparison of lumbar epidural and intravenous fentanyl infusions for postthoracotomy pain relief: analgesic, pharmacokinetic, and respiratory effects. Anesthesiology 1992; 77: 626–34.
          12. Morray JP, Geiduschek JM, Caplan RA, et al. A comparison of pediatric and adult anesthesia closed malpractice claims. Anesthesiology 1993; 78: 461–7.
            13. Gwirtz KH, Young JV, Byers RS, et al. The safety and efficacy of intrathecal opioid analgesia for acute postoperative pain: seven years’ experience with 5969 surgical patients at Indiana University Hospital. Anesth Analg 1999; 88: 599–604.
            14. Grady RE, Horlocker TT, Brown RD, et al. Neurologic complications of spinal drainage. Anesth Analg 1999; 88: 388–92.
            15. Hogan QH. Tuffier’s line: the normal distribution of anatomic parameters [letter]. Anesth Analg 1994; 78: 194–5.
            16. Hodgson PS, Neal JM, Pollock JE, Liu SS. The neurotoxicity of drugs given intrathecally (spinal) [review]. Anesth Analg 1999; 88: 797–809.
            17. Scott DA, Beilby DS, McClymont C. Postoperative analgesia using epidural infusions of fentanyl with bupivacaine: a prospective analysis of 1,014 patients. Anesthesiology 1995; 83: 727–37.
              18. Liu SS, Mulroy MF. Neuraxial anesthesia and analgesia in the presence of standard heparin. Reg Anesth Pain Med 1998; 23: 157–63.
              © 2003 International Anesthesia Research Society