Can Direct Spinal Cord Injury Occur Without Paresthesia?

A Report of Delayed Spinal Cord Injury After Epidural Placement in an Awake Patient
Ban C. H. Tsui, MD, MSc, FRCP(C)*, and Kevin Armstrong, MD, FRCP(C) *Department of Anesthesiology and Pain Medicine University of Alberta, Edmonton, Alberta, Canada, and Department of Anesthesia, University of Western Ontario, London, Ontario, Canada.

We discuss the etiology of a delayed spinal cord injury after epidural anesthesia without paresthesia. The description of such a case in an awake, adult patient who underwent a Whipple resection is provided. An epidural was performed at approximately the T8-9 interspace with the patient in the sitting position after 1 mg of midazolam was administered. On the first attempt, a dural puncture occurred. The patient did not report any paresthesia or pain. The needle was withdrawn and a second attempt was made one interspace lower. At this level, the epidural catheter was advanced into the epidural space uneventfully. Postoperatively,

the patient suffered decreased motor function in the right leg. Magnetic resonance imaging revealed high signal intensity within the spinal cord, indicating cord edema compatible with direct needle trauma. An extradural fluid collection consistent with a hematoma was also noted. Although it may be impossible to confirm if the spinal cord injury was a result of direct needle trauma, hematoma, or a combination of needle trauma and hematoma, these events clearly raise the important question of whether an awake patient will always report paresthesia secondary to spinal cord trauma. (Anesth Analg 2005;101:12124)

eports of spinal cord injury occurring while performing epidural anesthesia in unconscious patients have generated debate over the safety of performing epidural anesthesia in awake versus unconscious patients (1 7). The main argument against the performance of central neural blockade in anesthetized or heavily sedated patients is that unconscious subjects are unable to report paresthesia. However, is the converse true? Will the conscious patient always report a paresthesia in response to spinal cord or nerve root trauma?

Case Report
An 81-yr-old male (69 kg) was scheduled to have a Whipple resection for pancreatic cancer. With the exception of pronounced thoracic kyphosis, the patients physical and neurological examinations were unremarkable.
Accepted for publication February 15, 2005. Supported, in part, by Education and Research Fund, Department of Anesthesiology and Pain Medicine, University of Alberta, Edmonton, Canada and Department of Anesthesia, University of Western Ontario, London, Ontario, Canada. Ban C.H. Tsui is a recipient of Clinical Investigatorship Award, Alberta Heritage Foundation for Medical Research, Alberta, Canada. Address correspondence and reprint requests to Ban C.H. Tsui, MSC, MD, FRCP(C), Department of Anesthesiology and Pain Medicine, 8 120 Clinical Sciences Building, Edmonton, Alberta, Canada T6G 2G3. Address e-mail to btsui@ualberta.ca. DOI: 10.1213/01.ANE.0000175764.16650.85

An initial epidural needle was placed at the T8-9 interspace with the patient awake and in the sitting position after 1 mg of midazolam was administered. A right paramedian epidural was performed using a 17-gauge Tuohy needle with a loss of resistance (LOR) to air. On the first attempt, dural puncture occurred as free-flowing cerebrospinal fluid (CSF) was seen in the syringe. No air was injected. A second single pass attempt was successfully made one interspace lower and a soft-tipped epidural catheter was threaded easily into the epidural space. The anesthesiologist recorded that there was no pain or discomfort during the procedure. No blood or CSF was returned through the catheter. Three mL of 2% lidocaine was injected through the catheter and general anesthesia was then induced. Intraoperatively, 2 5-mL boluses of 0.125% bupivacaine plus 5 g/mL of fentanyl were given and a 12 mL/h infusion of the same solution was administered throughout the procedure. Two episodes of hypotension occurred: 80/50 mm Hg at 75 min and 75/45 mm Hg at 150 min postinduction. Both episodes responded readily to ephedrine (5 and 10 mg respectively) and were 510 min in duration. Estimated blood loss was 900 mL, which was replaced with 2000 mL of Ringers lactate solution and 1000 mL of Pentaspan. Surgery lasted approximately 4 h with no obvious surgical complications. The patient was tracheally extubated in the operating room and was admitted to the intensive care unit in stable condition for postoperative monitoring. Postoperatively, a formal neurological examination was not performed until postoperative day (POD) 4 because of a decreased level of consciousness. On POD 1 the patient was
2005 by the International Anesthesia Research Society 0003-2999/05

1212

Anesth Analg 2005;101:12124

ANESTH ANALG 2005;101:12124

CASE REPORTS

1213

Figure 2. Magnetic resonance image of the spinal cord. Black arrows indicate the spinal cord and E indicates the edema localized on the right side of the spinal cord.

Figure 1. Sagittal view of magnetic resonance imaging (MRI). White arrows indicate a high T2 signal intensity within the spinal cord extending from the mid T9 to mid T10 vertebral bodies. Black arrows indicate high signal intensity extradurally.

Postlaminectomy, motor and sensory function was still profoundly depressed. This partially resolved over time. After 8 wk the patient was dependent on a wheelchair for ambulation although he was able to bear weight on his right leg and perform self-transfer from his wheelchair to bed. Sensation had returned to normal. At 12 wk, he could ambulate with a walker but with impaired proprioception.

significantly obtunded but he moved all four limbs spontaneously and purposefully. A drug error was then noted: 5 g/mL of epidural fentanyl and 0.125% bupivacaine was being infused instead of 2 g/mL. This error was corrected by changing the epidural solution. On POD 2, the patient was drowsy and the epidural infusion was reduced from 12 mL/h to 9 mL/h. On POD 3, the patient was more alert but appeared to have more difficulty when mobilizing for any activity. The epidural infusion rate was then further reduced to 6mL/h. On POD 4, 2.5 h after the infusion was stopped, a neurological examination revealed little or no movement (0/5) and no sensation to pinprick in the right leg. Motor function and sensation were overall intact (4/5) but with decreased reflexes in his left leg. The epidural catheter was immediately removed intact. Severe winter weather made travel unsafe. As a result the emergent neurological consultation and urgent magnetic resonance imaging (MRI) did not take place as requested. On POD 5, neurological consultation confirmed that there was a deficit in the right leg with 1/5 motor function and decreased temperature, vibration, and position sensation. Spinal cord edema, a high T2 signal intensity on MRI, compatible with direct trauma was noted from the mid T9 to the mid T10 vertebral body (Figs. 1 and 2). Moreover, a posterior extradural fluid collection from the T8-9 disk level extending inferior to the T11 vertebral body consistent with a hematoma was present. Its maximal anterior posterior dimension was 7 mm. There were compression fractures at T8 and T11. The neurosurgical team was consulted and subsequently performed an emergency laminectomy and evacuation of the hematoma 7 h after the MRI results. Under direct surgical examination, the hematoma did not appear to cause significant compression.

Discussion
This report describes the case of delayed spinal cord injury from epidural anesthesia without reported paresthesia in an awake patient. Although it may be impossible to confirm the etiology of the spinal cord injury in this case, these events raise the important question of whether an awake patient will always be able to report paresthesia secondary to spinal cord trauma. Spinal cord damage can result from needle or catheter trauma, local anesthetic toxicity, epidural hematoma, ischemia from an arterial injury, or severe hypotension. Although present, an epidural hematoma was unlikely to produce the patients neurological deficit, as MRI did not suggest significant compression of the spinal cord, a fact confirmed at laminectomy. As the clinical sign of spinal cord ischemia is normally bilateral paresis of the legs, the fact that this patient had a unilateral deficit of the right side is not consistent with spinal cord ischemia from a vascular insult (8,9). Spinal cord edema confined to the right side suggests a small likelihood of injury to the Artery of Adamkiewicz, as it is predominantly left sided (10). In addition, there was clear MRI evidence of edema mainly localized on the right side of the spinal cord (Fig. 2). Although the initial epidural infusion of 12 mL/h with bupivacaine 0.125% plus 5 g/mL fentanyl was larger than usual, it is highly unlikely that

1214

CASE REPORTS

ANESTH ANALG 2005;101:12124

drug toxicity would account for this unilateral right side spinal cord edema. However, the large opioid dose certainly could have contributed to the patients sedation in the postoperative period. This rapid infusion rate would make intrathecal catheter placement very unlikely, as there was no sign of a total spinal. The fact that a soft catheter was used further reduces the likelihood of direct catheter trauma. In contrast, the dura was clearly punctured around the level of spinal cord edema on the first attempt of epidural needle placement. This certainly raises the possibility that the lesion on MRI resulted from an unrecognized spinal cord puncture during the epidural needle placement while using a right-sided paramedian approach. Given that the epidural was performed with the patient awake and in a standard fashion, the questions are raised whether paresthesia is a reliable sign of spinal cord encroachment and whether LOR alone is sufficient to avoid adverse sequelae. Paresthesia associated with spinal cord injury can occur at the time of needle placement but it also has been reported to develop only at the time of injection or secondary to irritation, edema, or hematoma (11,12). In addition, pain is more common in extraaxial lesions affecting the nerve roots or blood vessels that are innervated by sensory neurons mediating pain (13). In contrast, because there are no pain receptors within the spinal cord itself (or the brain), intraaxial lesions may be painless (13). This allows percutaneous cervical cordotomy to be performed in awake patients (14,15). During the procedure, the cervical cord is typically punctured multiple times with a 22gauge needle electrode, and yet the patient generally describes neither pain nor paresthesia (16). In addition, pain reported from dural puncture is rare in clinical practice. This point is also well illustrated in this case report, as our patient did not complain of pain despite a clinically obvious dural puncture. In this case, it is difficult to determine the primary cause of the injury and how a delayed diagnosis as a result of human error, lack of vigorous neurological monitoring, and uncontrollable environmental conditions contributed to the extent of the damage. Clearly, the lack of vigorous neurological monitoring contributed to a delayed diagnosis and possibly reduced the opportunity to correct any reversible defect that may have occurred. Thus, vigorous postoperative neurological monitoring should be implemented in all patients receiving neuraxial analgesia. Preoperative review of diagnostic imaging of the vertebral column may be helpful in delineating the extent of preexisting pathology. This case reminds clinicians that we should not simply assume paresthesia will always occur and be reported if a needle encroaches upon the spinal cord.

Although the use of electrophysiological monitoring has become common practice in spinal surgery (17) and the idea of using electrical stimulation during epidural needle advancement to provide an additional monitoring technique has been recently introduced (18,19), there is still no clear evidence that direct thoracic epidural placement can be performed without risk in either awake or anesthetized patients.

References
1. Bromage PR, Benumof JL. Paraplegia following intracord injection during attempted epidural anesthesia under general anesthesia. Reg Anesth Pain Med 1998;23:104 7. 2. Fischer HB. Performing epidural insertion under general anaesthesia. Anaesthesia 2000;55:288 9. 3. Kao MC, Tsai SK, Tsou MY, et al. Paraplegia after delayed detection of inadvertent spinal cord injury during thoracic epidural catheterization in an anesthetized elderly patient. Anesth Analg 2004;99:580 3. 4. Kasai T, Yaegashi K, Hirose M, Tanaka Y. Spinal cord injury in a child caused by an accidental dural puncture with a singleshot thoracic epidural needle. Anesth Analg 2003;96:657. 5. Krane EJ, Dalens BJ, Murat I, Murrell D. The safety of epidurals placed during general anesthesia. Reg Anesth Pain Med 1998; 23:433 8. 6. Mayall MF, Calder I. Spinal cord injury following an attempted thoracic epidural. Anaesthesia 1999;54:990 4. 7. Rose JB. Spinal cord injury in a child after single-shot epidural anesthesia. Anesth Analg 2003;96:3 6. 8. Kuker W, Weller M, Klose U, et al. Diffusion-weighted MRI of spinal cord infarction: high resolution imaging and time course of diffusion abnormality. J Neurol 2004;251:818 24. 9. Weidauer S, Nichtweiss M, Lanfermann H, Zanella FE. Spinal cord infarction: MR imaging and clinical features in 16 cases. Neuroradiology 2002;44:8517. 10. Kawaharada N, Morishita K, Hyodoh H, et al. Magnetic resonance angiographic localization of the artery of Adamkiewicz for spinal cord blood supply. Ann Thorac Surg 2004;78:846 51. 11. Simon SL, Abrahams JM, Sean GM, et al. Intramedullary injection of contrast into the cervical spinal cord during cervical myelography: a case report. Spine 2002;27:E274 7. 12. Hamandi K, Mottershead J, Lewis T, et al. Irreversible damage to the spinal cord following spinal anesthesia. Neurology 2002; 59:624 6. 13. Kandel ER. The perception of pain. In: Kandel ER, Schwartz JH, Jessel TM, eds. Principles of neural science. New York: McGraw-Hill Health Professions Division, 2000:47291. 14. Jackson MB, Pounder D, Price C, et al. Percutaneous cervical cordotomy for the control of pain in patients with pleural mesothelioma. Thorax 1999;54:238 41. 15. Lahuerta J, Bowsher D, Lipton S, Buxton PH. Percutaneous cervical cordotomy: a review of 181 operations on 146 patients with a study on the location of pain fibers in the C-2 spinal cord segment of 29 cases. J Neurosurg 1994;80:975 85. 16. Pounder D, Elliott S. An awake patient may not detect spinal cord puncture. Anaesthesia 2000;55:194. 17. Raynor BL, Lenke LG, Kim Y, et al. Can triggered electromyograph thresholds predict safe thoracic pedicle screw placement? Spine 2002;27:2030 5. 18. Tsui BC, Wagner A, Cunningham K, et al. Threshold current of electrical stimulation in intrathecal space using insulated needles in pediatric patients. Anesth Analg 2005;100:6625. 19. Tsui BC, Emery D, Uweira RE, Finucane B. The use of electrical stimulation to monitor needle placement in a porcine model. Anesth Analg 2005;100:16113.