SPINAL cord ischemia resulting in postoperative paraplegia is a devastating complication after thoracoabdominal aortic aneurysm (TAAA) surgery. Numerous strategies have been devised to protect the spinal cord from ischemia by maintaining spinal cord perfusion. Although controversial, one of the more popular clinical strategies, cerebrospinal fluid (CSF) drainage, theoretically increases spinal cord blood flow by decreasing CSF pressure. 1This article describes a case report of a patient with a thoracic epidural catheter in place who developed paraplegia immediately after removal of a lumbar CSF drainage catheter during the postoperative period after TAAA surgery, suggesting that the technique may have been of benefit to this patient.

Case Report

The patient was a 74-yr-old man with an asymptomatic complex TAAA (Crawford classification III). The superior aspect of the aneurysm began in the midthoracic region (6.5 cm maximum diameter) and tapered to normal diameter at the superior mesenteric artery. A second aneurysm (5.0 cm maximum diameter) began at the renal arteries and extended into the left common iliac artery. Medical history was significant for hypertension. Medications taken included clonidine, amlodipine, and hydrochlorothiazide. On the evening before surgery, a thoracic epidural catheter was inserted atraumatically (first attempt) at T8. The catheter was advanced 5 cm into the epidural space.

On the morning of surgery, in the operating room, a lumbar intrathecal catheter was inserted atraumatically (first attempt) at L3 via  a 14-gauge Tuohy needle. The catheter was advanced 5 cm into the intrathecal space. The CSF drainage catheter remained open throughout the intraoperative period and was adjusted to maintain the CSF pressure at 10 cm H2O with a Becker®external drainage and monitoring system (Medtronic, Inc., Goleta, California). General anesthesia was then induced with intravenous midazolam, fentanyl, and pancuronium. Maintenance general anesthesia consisted of supplemental midazolam and fentanyl combined with inhaled isoflurane. The trachea was intubated with a left-sided double-lumen endotracheal tube and its proper position was verified via  fiberoptic bronchoscopy. A central venous catheter was then inserted into the right internal jugular vein (through which a pulmonary artery catheter was eventually placed) followed by insertion of a right femoral artery catheter. Transesophageal echocardiography revealed mild hypokinesis of the left ventricular anterolateral wall, mild aortic insufficiency, and mild tricuspid insufficiency.

After adequate TAAA exposure via  left thoracotomy and after 300 U/kg intravenous heparin, partial cardiopulmonary bypass (CPB) was initiated via  femoral arterial and venous cannulae at approximately 1 l · min1· m2. Anticoagulation was assessed via  activated clotting time (ACT) determinations (lowest 664 s, highest 855 s). Methylprednisolone (2 g intravenous) was administered before initiation of CPB and 12.5 g mannitol was added to the standard crystalloid CPB prime. Dopamine (3 μg · kg1· min1intravenous) was administered throughout the intraoperative period. The proximal aortic cross-clamp was applied in the midthoracic region, and the distal aortic cross-clamp was applied just inferior to the renal arteries. A Dacron graft (28 cm) was inserted and the celiac trunk, superior mesenteric artery, and renal arteries were all reimplanted. No intercostal arteries were reimplanted. The patient’s temperature was allowed to passively decrease (lowest esophageal temperature, 35.5°C). Following hemostasis, separation from CPB was accomplished without support of vasoactive medications. The ACT decreased to baseline (120 s) after protamine administration (250 mg). Total cross-clamp and CPB times were 38 and 48 min, respectively.

The patient was hemodynamically stable throughout the intraoperative period. Nitroglycerin was required for hypertension with initial application of the proximal cross-clamp. During cross-clamp application, proximal mean arterial pressure was maintained above 90 mmHg and distal mean arterial pressure was maintained above 60 mmHg. The CSF catheter drained 173 ml throughout the intraoperative period. During closure, an infusion of bupivacaine (1.0%) and fentanyl (5 μg/ml) was initiated at 10 ml/h via  the thoracic epidural catheter. Total intraoperative intravenous fluids were 4,200 ml crystalloid, 1,500 ml colloid, 1,250 ml “cell-saver” blood, 2 units fresh frozen plasma, and 4 units packed erythrocytes. Estimated blood loss was 2,000 ml. Total intraoperative urine output was 710 ml.

Within a few hours of arrival to the intensive care unit (ICU), the patient was awake and neurologically intact. The CSF drainage catheter remained open and was adjusted to maintain the CSF pressure at 10 cm H2O. Tracheal extubation occurred 6 h after ICU arrival. Soon thereafter, the patient complained of left leg weakness and inability to flex either knee. The epidural infusion was stopped and over the next hour, full motor strength and function of both lower extremities returned. Because of increased incisional pain during this same time period, the epidural infusion was reinitiated at 10 ml/h.

At 0900 on the first postoperative day, the CSF drainage catheter was removed. At this time the patient had full motor strength and function of both lower extremities and was ambulating without difficulty. During these initial ICU hours (ICU arrival to catheter removal), the patient was hemodynamically stable. The mean arterial blood pressure was never less than 70 mmHg and small amounts of intravenous nitroprusside were required to maintain the systolic blood pressure less than 130 mmHg. During these initial 16 ICU hours, the CSF catheter drained 64 ml to maintain the CSF pressure at 10 cm H2O. Within 30 min of catheter removal, the patient reported bilateral lower extremity weakness during ambulation. One hour after catheter removal, he could not move his left lower extremity and soon thereafter(30 min) could not move either lower extremity at all. The epidural infusion was decreased to 7 ml/h. Seven hours after catheter removal, the patient was unable to move either lower extremity against gravity. The epidural catheter was repositioned (pulled back) and the infusion was decreased to 6 ml/h. Ten hours after catheter removal, the patient was unable to move either lower extremity. The epidural infusion was stopped. Twelve hours after catheter removal, bilateral paralysis persisted, prompting emergency magnetic resonance imaging, which revealed no signs of extradural mass or epidural hematoma yet demonstrated spinal cord infarction in the anterior spinal artery distribution. Observed on axial images were bilateral areas of abnormal signal within the spinal cord between the conus to the level of approximately T8, where axial images were no longer obtained. These “snake eye”-shaped lesions demonstrated in axial views (fig. 1) are typical of cord ischemia and infarction. 2During the 12 h after CSF catheter removal, the patient was hemodynamically stable. Intensive therapy after emergent magnetic resonance imaging (reinsertion of CSF drainage catheter, intravenous dopamine, intravenous heparin, intravenous dexamethasone) failed to improve the neurologic deficit. By postoperative day 5, sensory level to pin prick had stabilized at T10 and the patient had no motor function in either leg. Posterior cord function (position sense and vibratory sensation) remained intact. On postoperative day 7, the patient experienced sudden cardiac arrest. Following emergent surgical exploration, the proximal and distal ends of the Dacron graft were found to be intact, yet large sections of the small and large intestine were infarcted. The patient died in the operating room. An autopsy was not performed.

Fig. 1. Representative axial image of the spinal cord. The top of the image is anatomically anterior whereas the bottom of the image is anatomically posterior. The “snake eye”-shaped lesion demonstrated within the spinal cord is typical of cord ischemia and infarction. 2 

Fig. 1. Representative axial image of the spinal cord. The top of the image is anatomically anterior whereas the bottom of the image is anatomically posterior. The “snake eye”-shaped lesion demonstrated within the spinal cord is typical of cord ischemia and infarction. 2 


Autopsy studies indicate the prevalence of asymptomatic thoracic aneurysm is approximately 4–6 per 1,000 patients, with the incidence increasing with age. 3TAAA is extremely rare and represents approximately 5% of all asymptomatic thoracic aneurysms. 3Recent recommendations regarding TAAA management have emphasized individualized treatment based on balancing a patient’s calculated risk of rupture with their anticipated risk of postoperative death or paraplegia. 4With current techniques, elective TAAA resection can be accomplished with operative mortality rates of approximately 5–10%. 3,4Major perioperative morbidity associated with TAAA resection includes myocardial infarction, respiratory failure, renal failure, stroke, and paraparesis or paraplegia. The incidence of postoperative permanent spinal cord injury remains high (10–30% depending on type of TAAA), likely a result of the remarkable variability among patients regarding spinal cord blood supply. 5The tenuous collateral anastomosis of the anterior spinal artery in the midthoracic region places segments of the spinal cord in jeopardy during aortic occlusion or hypotension. Damage may result from either actual surgical dissection of the artery of Adamkiewicz (because the origin is unknown) or exclusion of the origin of the artery by the aortic cross-clamps.

Many techniques have been investigated in hopes of improving spinal cord protection during TAAA surgery, none of which have proven clinically reliable. 6,7These various techniques include distal perfusion methods (arterial and venous), hypothermia (general and regional), monitoring spinal cord somatosensory and motor evoked potentials, pharmacologic agents (vasodilators, naloxone, etc. ), intercostal artery reimplantation, profound hypothermic circulatory arrest, endovascular stents, and ischemic preconditioning. One of the most intensely investigated and clinically utilized techniques has been CSF drainage, which decreases CSF pressure and theoretically increases spinal cord blood flow by increasing spinal cord perfusion pressure. Animal studies 8indicate that CSF drainage (to a CSF pressure of 10 mmHg) prevents paraplegia yet human studies 9have been less successful. Following a recent systematic overview of evidence involving use of CSF drainage in TAAA surgery for prevention of paraplegia, Ling and Arellano conclude that the benefits, if any, are unsubstantiated. 1 

Almost half of all episodes of spinal cord injury related to TAAA resection occur during the postoperative period (for example, delayed neurologic deficits) and may occur anywhere from 12 h to 21 days after surgery. 10,11Although many factors may play a role in delayed neurological deficit formation (underlying atherosclerotic vascular disease, spinal artery spasm, etc. ), the two determinants of spinal cord perfusion pressure (CSF pressure and blood pressure) are thought to play pivotal roles. 10,11Postoperative episodes of increased CSF pressure (from an epidural injectate), and decreased blood pressure have been linked to development of delayed neurologic deficits in certain patients. 10,11To the authors’ knowledge, this case report represents the first description of delayed neurologic deficit possibly caused by cessation of CSF drainage. We theorize that with cessation of CSF drainage, CSF pressure may have increased to a point that initiated spinal cord ischemia/infarction (the patient’s blood pressure did not change during this time). Although we cannot definitively conclude that the supposed increase in CSF pressure was the sole mechanism of the delayed neurologic deficit, the timing of the postoperative clinical events, along with the presence of hemodynamic stability during the same time period, suggest an etiologic role. We theorize that the second CSF catheter failed to improve the neurologic deficit because irreversible damage to the spinal cord had already occurred. On the other hand, the possibility exists that this patient suffered a coincidental embolic event that resulted in spinal cord damage, a possibility supported by the subsequent development of bowel ischemia. Thus, it remains possible that the neurologic deficit would have occurred despite removal of the CSF catheter or despite any CSF catheter at all.

The postoperative management of an epidural (lumbar or thoracic) catheter in a patient after TAAA surgery is certainly challenging and somewhat controversial. The goal is to attain adequate analgesia without detrimental effects (hypotension, respiratory depression, etc. ). Furthermore, in this unique patient population, concern always exists regarding hematoma formation and masking a developing neurologic deficit. One could argue that in this patient, the thoracic epidural infusion should have been immediately discontinued when the neurologic deficit was first noted and that the MRI should have been performed earlier. In hindsight, these statements are probably true. Perhaps with earlier detection, prognosis would have been altered.

In conclusion, development of paraplegia after TAAA surgery remains a clinically important problem, the etiology of which is multifactorial. No one technique has been proven to substantially reduce incidence. Therefore, a multimodal approach to prevention of paraplegia should be adopted and tailored to each individual patient. CSF drainage may be of benefit to some patients, not only during aortic cross-clamp application intraoperatively, but also during the immediate postoperative period (the length of which remains to be determined).


Ling E, Arellano R: Systematic overview of the evidence supporting the use of cerebrospinal fluid drainage in throacoabdominal aneurysm surgery for prevention of paraplegia. A nesthesiology 2000; 93: 1115–22
Osborn AG: Diagnostic Neuroradiology, nonneoplastic disorders of the spine and spinal cord. Edited by Osborn AG. St. Louis, Mosby, 1994, pp 820–75
Kouchoukos NT, Dougenis D: Medical progress: Surgery of the thoracic aorta. N Engl J Med 1997; 336: 1876–88
Coselli JS, LeMaire SA, Miller CC, Schmittling ZC, Köksoy C, Pagan J, Curling PE: Mortality and paraplegia after thoracoabdominal aortic aneurysm repair: A risk factor analysis. Ann Thorac Surg 2000; 69: 409–14
Biglioli P, Spirito R, Roberto M, Grillo F, Cannata A, Parolari A, Maggioni M, Coggi G: The anterior spinal artery: The main arterial supply of the human spinal cord: A preliminary anatomic study. J Thorac Cardiovasc Surg 2000; 119: 376–9
Svensson LG: An approach to spinal cord protection during descending or thoracoabdominal aortic repairs. Ann Thorac Surg 1999; 67: 1935–6
Safi HJ, Miller CC: Spinal cord protection in descending thoracic and thoracoabdominal aortic repair. Ann Thorac Surg 1999; 67: 1937–9
Bower TC, Murray MJ, Gloviczki P, Yaksh TL, Hollier LH, Pairolero PC: Effects of thoracic aortic occlusion and cerebrospinal fluid drainage on regional spinal cord blood flow in dogs: correlation with neurologic outcome. J Vasc Surg 1988; 9: 135–44
Crawford ES, Svensson LG, Hess KR, Shenaq SS, Coselli JS, Safi JH, Mohindra PK, Rivera V: A prospective randomized study of cerebrospinal fluid drainage to prevent paraplegia after high-risk surgery on the thoracoabdominal aorta. J Vasc Surg 1990; 13: 36–46
Fitzgibbon DR, Glosten B, Wright I, Tu R, Ready B: Paraplegia, epidural analgesia, and thoracic aneurysmectomy. A nesthesiology 1995; 83: 1355–9
Crawford ES, Mizrahi EM, Hess KR, Coselli JS, Safi HJ, Patel VM: The impact of distal aortic perfusion and somatosensory evoked potential monitoring on prevention of paraplegia after aortic aneurysm operation. J Thorac Cardivasc Surg 1988; 95: 357–67