Interventions for Reversing Delayed-Onset Postoperative Paraplegia After Thoracic Aortic Reconstruction

Transcription

1 Interventions for Reversing Delayed-Onset Postoperative Paraplegia After Thoracic Aortic Reconstruction Albert T. Cheung, MD, Stuart J. Weiss, MD, PhD, Michael L. McGarvey, MD, Mark M. Stecker, MD, PhD, Michael S. Hogan, BS, Alison Escherich, MPH, and Joseph E. Bavaria, MD Departments of Anesthesia, Neurology, and Surgery, University of Pennsylvania, Philadelphia, Pennsylvania, and Royal College of Surgeons in Ireland, Dublin, Ireland Background. Delayed postoperative paraplegia is a recognized complication of thoracic (TAA) or thoracoabdominal aortic aneurysm (TAAA) repair. The purpose of this study was to evaluate the effectiveness of interventions to treat delayed-onset paraplegia. Methods. Between January 1, 2000 and August 31, 2001, 99 patients underwent surgical repair of TAA, Crawford type I, II, or III TAAA. Standard intraoperative management included distal aortic perfusion and cerebrospinal fluid (CSF) drainage unless contraindicated. Therapeutic interventions to treat delayed paraplegia included lumbar CSF drainage and vasopressor therapy. Results. Three of the 99 patients had paraplegia upon awakening. Delayed-onset paraplegia occurred in 8 patients, 2 of whom had recurrent episodes. In those 8 patients, the initial episode occurred at a median of 21.6 hours (range 6.4 to hours) after surgery and the second episode averaged 176 hours after surgery. At the onset of paraplegia, the average mean arterial pressure was 74 mm Hg and CSF pressure was 14 mm Hg. Three of the 8 patients had a functioning CSF catheter at the onset and the other 5 patients had catheters subsequently placed. Therapeutic interventions increased blood pressure to a mean arterial pressure of 95 mm Hg and decreased CSF pressure to 10 mm Hg. Five of the 8 patients with delayed-onset paraplegia made a full neurologic recovery and 3 had partial recovery. Conclusions. Patients with delayed-onset paraplegia had an increased chance of recovery as compared with those patients in whom paraplegia was diagnosed upon emergence from anesthesia. Acute interventions directed to increase spinal cord perfusion by increasing systemic blood pressure and decreasing CSF pressure were effective for the reversal of delayed onset of paraplegia after TAA or TAAA repair, resulting in an overall 3% incidence of permanent paraplegia and 3% incidence of residual paraparesis. (Ann Thorac Surg 2002;74:413 21) 2002 by The Society of Thoracic Surgeons Paraplegia is a well-recognized and morbid complication of thoracic (TAA) or thoracoabdominal aortic aneurysm (TAAA) repair [1 8]. In most reported cases of postoperative paraplegia, it has been difficult to identify the exact time of onset, the factors that may have contributed to the development of spinal cord ischemia, or whether therapeutic interventions were effective for the treatment of this complication. In many of the cases, the delayed onset of paraplegia temporally followed episodes of hypotension, suggesting that the neurologic deficit was related in part to hypoperfusion [1, 9, 6, 10]. Beginning in January 2000, a multidisciplinary team was assembled at the University of Pennsylvania to address specifically the clinical management of postoperative paraplegia after TAA and TAAA repair. The Presented at the Thirty-eighth Annual Meeting of The Society of Thoracic Surgeons, Fort Lauderdale, FL, Jan 28 30, Address reprint requests to Dr Cheung, Division of Cardiothoracic and Vascular Anesthesia, University of Pennsylvania, 3400 Spruce St, Ravdin 4 Courtyard, Philadelphia, PA ; cheunga@ uphs.upenn.edu. objective was to detect and diagnose postoperative paraplegia at its onset in order to institute immediate therapeutic interventions to improve spinal cord perfusion. This study assessed the effectiveness of interventions to treat postoperative paraplegia to test the hypothesis that patients with delayed-onset postoperative paraplegia had the potential to recover neurologic function. Material and Methods All patients undergoing TAA or TAAA repair from January 1, 2000 to August 31, 2001 were prospectively entered into a clinical database. Patients in this database with postoperative paraplegia were identified and analyzed. TAAA were classified according to Crawford type I to IV. In addition, aneurysms isolated to the thoracic aorta (TAA) were included. Descending aneurysms with concomitant distal arch extension were also included (Table 1). Patients with Crawford type IV aortic aneurysms repaired without using extracorporeal circulation were not included in the prospective database. Aneu by The Society of Thoracic Surgeons /02/$22.00 Published by Elsevier Science Inc PII S (02)

2 414 CHEUNG ET AL Ann Thorac Surg REVERSAL OF POSTOPERATIVE PARAPLEGIA 2002;74: Table 1. Patient Demographics Total 99 patients n Redo Dissection d TAA a 40 13% 10% Stent Grafts 12 8% Thoracoabdominal b 40 38% 20% Extent I TAAA 19 21% Extent II TAAA 9 44% Extent III TAAA 12 0% Distal arch TAA/TAAA c 19 26% 47% (hypothermia with HCA/RCP) Distal Arch TAA 9 67% Distal Arch TAAA 10 30% a TAA were defined as aneurysms within the chest between the left subclavian artery and the diaphragm. They included 4 patients with combined ascending aortic aneurysm and descending thoracic aortic aneurysm replacement. b TAAA were classified according to the standard Crawford criteria. c Included 2 patients with total thoracic aortic replacement. d Only 2 patients had acute dissections. TAA thoracic aortic aneurysm; TAAA thoracoabdominal aortic aneurysm; HCA hypothermic circulatory arrest; RCP retrograde cerebral perfusion. rysm etiology included atherosclerotic disease, chronic dissection, and saccular aneurysms. All patients had a narcotic-based general anesthetic with inhaled isoflurane in oxygen. A 0.7-mm ID lumbar cerebrospinal fluid (CSF) drainage catheter (MoniTorr ICP; Clinical Neuro Systems, Exton, PA) was inserted an average of 7 cm into the subarachnoid space via a 14-G Tuohy needle at the L3 L4 vertebral interspace after the induction of general anesthesia. CSF was drained for a lumbar CSF pressure 12 mm Hg during operation. Epidural or subarachnoid narcotics or local anesthetics for analgesia were not administered until after postoperative neurologic assessment. The mean arterial pressure (MAP) was maintained between 75 and 85 mm Hg during general anesthesia before extracorporeal circulation. All patients were exposed via a posterolateral thoracotomy. The incision was extended past the costal cartilages, lateral to the rectus muscle, for thoracoabdominal aneurysms. Circulation management consisted of either left atrial to femoral artery (or distal aortic) partial left heart bypass with moderate core cooling to 32 C (LA-FA bypass) or a hypothermic technique utilizing full cardiopulmonary bypass (CPB) via the left chest with an open proximal anastomosis (hypothermic technique). LA-FA bypass was used in most cases with a staged segmental reconstruction of the aorta. LA-FA bypass flow rates averaged 2.5 L per minute, adjusted to achieve a target distal aortic perfusion pressure of at least 60 mm Hg, while maintaining a proximal aortic pressure of at least 90 mm Hg. During the mesenteric anastomosis, perfusion cannulae to the renal arteries and superior mesenteric artery were utilized off the cardioplegia line from the CPB circuit. Intercostal arteries were not selectively perfused during LA-FA bypass. The hypothermic technique was used if a concomitant distal arch aneurysm needed resection or if LA-FA bypass could not be performed. Circulation management during an open proximal anastomosis at the distal aortic arch consisted of hypothermic circulatory arrest (HCA) at a mean lowest nasopharyngeal temperature of 14.5 C, together with total-body retrograde cerebral perfusion (RCP) via the superior vena cava at 12 C, in a slight Trendelenberg position with a target central venous pressure of approximately 15 mm Hg. This usually translated into a RCP flow of 300 to 500 cm 3 per minute. After completion of the open proximal anastomosis, arterial circulation was reinitiated via the Dacron graft and rewarming was begun. Lumbar CSF drains were used in both LA-FA bypass and hypothermic cases. Mannitol 25 g, methylprednisolone 1 g, magnesium 2 g, and lidocaine 200 mg were administered upon initiation of LA-FA bypass or CPB. Intercostal arteries were reimplanted in all patients with dissecting aortic aneurysms and selectively in patients with atherosclerotic aneurysms if a large patch of intercostal arteries was identified between the T7 and L1 vertebral levels. Patients were admitted to a surgical intensive care unit (SICU) after operation. Vital signs, core temperature, cardiac output, arterial pressure, central venous pressure, pulmonary artery pressures, and the lumbar CSF pressure were recorded at 15- to 60-minute intervals. The MAP was maintained in a range of 75 to 85 mm Hg using vasopressors or vasodilators (nicardipine) depending on the perceived strength of the arterial anastamosis and the risk of bleeding. Lumbar CSF was drained in 10-mL aliquots to maintain a lumbar CSF pressure 12 mm Hg. A limited neurologic assessment was performed on an hourly basis until recovery from general anesthesia permitted a complete neurologic examination. In the absence of a neurologic deficit, patient-controlled epidural analgesia with fentanyl 3 g/ml and bupivacaine 0.05% was administered by epidural catheter. The lumbar CSF drainage catheter was occluded 24 hours after operation and removed at 48 hours after operation in the absence of a neurologic deficit. If a neurologic deficit was detected after emergence from general anesthesia, a full neurologic examination was performed emergently by a neurologist. A standard neurologic examination was performed at the time of onset of any neurologic deficit by a board-certified neurologist. This examination was performed according to the Coalition-revised Neurologic and Mental Status Exam of the American Academy of Neurology. Subsequent examinations were performed serially and during any periods of improvement or decline in neurologic function relative to the initial examination. Special attention was directed toward diagnosing spinal cord ischemia to exclude brain ischemia or peripheral nerve injury as etiologies. The documented neurologic examinations were reviewed and scored retrospectively. The lowerextremity motor component of the American Spinal Injury Association (ASIA) Standard Neurologic Classification of Spinal Injury 50-point scale was used to score the maximum neurologic deficit, the best neurologic function after treatment, and the severity of the neurologic deficit at the time of onset [11]. The scale graded only motor function in each of five muscle groups in each lower extremity: hip flexors, knee extensors, ankle dorsiflexors,

3 Ann Thorac Surg CHEUNG ET AL 2002;74: REVERSAL OF POSTOPERATIVE PARAPLEGIA 415 Table 2. Episodes of Paraplegia in Patients After Thoracic Aortic Reconstruction At Diagnosis At Recovery Patient No. Aneurysm Type Onset a (Hours) MAP (mm Hg) CSF (mm Hg) MAP (mm Hg) CSF (mm Hg) 1 TAA b TAA b f 87 f n.a. 94 f n.a. 3 TAAA II c 17.0 g h h 4 TAA d,e n.a. 83 n.a f 96 f 13 f 93 f 12 f 5 TAAA I c,e 16.0 g h h 6 TAAA II c TAAA I c,e 6.0 g h h 8 TAAA II c TAA d,e TAAA III c TAAA I c a Time after arrival to SICU after operation. b TAA were defined as aneurysms within the chest between the left subclavian artery and the diaphragm. c TAAA classified according to the Crawford criteria repaired using left atrial-to-left femoral artery partial cardiopulmonary bypass. d TAA or TAAA with distal aortic arch involvement repaired using deep hypothermic circulatory arrest. e Dissection present. f Second recurrent episode of postoperative paraplegia in the same patient. g Paraplegia detected immediately after emergence from general anesthesia. h No recovery from paraplegia CSF cerebrospinal fluid; MAP mean arterial pressure; n.a. not available; Patient No. patient number, TAA Thoracic aortic aneurysm; TAAA thoracoabdominal aortic aneurysm and Crawford type. long toe extensors, and ankle plantar flexors. The motor strength of each muscle group was rated on a scale of 0 to 5, with 0 total paralysis, 1 palpable or visible contraction, 2 active movement, gravity eliminated, 3 active movement against gravity, 4 active movement against some resistance, and 5 active movement against full resistance. A total score of 0 indicated paralysis of both lower extremities. Paraparesis was defined as weakness in a lower extremity muscle group (motor score of 1 through 4), incomplete paraplegia, or unilateral paraplegia. The presence or absent of a sensory deficit was also recorded. After verification of paraplegia or paraparesis by a neurologist and a diagnosis consistent spinal cord ischemia, the MAP was increased to 90 mm Hg using intravenous infusion of phenylephrine or norepinephrine. In patients with a functioning lumbar CSF drain in place, CSF was drained if the lumbar CSF pressure was greater than 10 mm Hg. In patients without a functioning lumbar CSF drain, a lumbar CSF drain was inserted emergently if there was no immediate improvement in neurologic function after augmentation of the arterial pressure. If there was no evidence of neurologic recovery at a MAP of 90 mm Hg, the MAP was augmented further to 95 mm Hg, then 100 mm Hg. The estimated spinal cord perfusion pressure (SCP) was determined by subtracting the transduced lumbar CSF pressure from the MAP (SCP MAP lumbar CSF pressure). Serial neurologic examinations were performed until resolution or stabilization of the postoperative neurologic deficit. The independent Student s t test, 2 square test, or Fisher s exact test were used to test for differences between groups of patients. The paired Student s t test was used to test for changes in blood pressure, body temperature, or neurologic score at different time points within a selected group of patients. A p value 0.05 was considered significant. Results During the study period, 99 patients underwent thoracic or thoracoabdominal aortic replacement. This cohort was 55% male and had a mean age of years (Table 1). Included in the cohort were 19 patients with combined thoracic and thoracoabdominal aneurysms with distal arch extension requiring an open proximal anastomosis utilizing HCA techniques. Intercostal artery implantation rate was near 100% in patients with dissecting aortic aneurysms, but only 25% in patients with atherosclerotic aneurysms. The percentage of patients that had aortic dissection and underwent redo operations was listed in Table 1. Overall, 11 patients (11%) had postoperative paraplegia or paraparesis. Postoperative neurologic assessment was not possible in 3 patients, all with TAAA managed with LA-FA bypass who died before emergence from general anesthesia. The incidence of paraplegia or paraparesis according to aneurysm type and circulation management strategy was 5% (2/40) in patients with isolated TAA, 17.5% (7/40) in patients with TAAA managed with LA-FA bypass, and 10% (2/19) in patients with TAA or TAAA with distal aortic arch extension managed with HCA (Table 2). The overall incidence of permanent paraplegia was 3% (3/99) and was only observed in the subgroup with TAAA managed with LA-FA bypass (Table 2). Of the 11 patients with postoperative paraplegia, 5

4 416 CHEUNG ET AL Ann Thorac Surg REVERSAL OF POSTOPERATIVE PARAPLEGIA 2002;74: Table 3. Neurologic Score a and Findings in Patients With Postoperative Paraplegia Patient No. Neurological Score a at Emergence from GA b Neurologic Score a and Findings at Diagnosis c Neurological Score a at the Time of Maximum Deficit c Neurological Score a and Findings at Recovery c (LLE paraplegia, Nl sensation) (Mild LLE weakness and Nl sensation) (LLE paraplegia, Sensory loss) (Nl sensation) (BLE paraplegia and sensory loss) 0 0 (No recovery) (BLE paraplegia, and sensory loss) 0 50 (Nl sensation) (BLE paraplegia and sensory loss) 0 0 (No recovery) (BLE paraplegia and sensory loss) 0 50 (Nl sensation) (BLE paraplegia and sensory loss) 0 0 (No recovery) (LLE paraplegia and sensory loss) (Mild LLE weakness and Nl sensation) (Mild RLE paraparesis, severe (Nl sensation) LLE paraparesis, and sensory loss) (BLE paraplegia and sensory loss) 0 40 (Mild LLE weakness and Nl sensation) (Moderate RLE paraparesis, mild LLE paraparesis, and sensory loss) (Nl sensation) a Motor score of the American Spinal Injury Association (ASIA) Standard Neurological Classification of Spinal Cord Injury (see text). b ASIA scores based on the initial neurologic assessment by the intensive care unit staff upon emergence from general anesthesia. c ASIA scores based on the full neurologic examination by a neurologist. BLE both lower extremities; GA general anesthesia; Nl normal; LLE left lower extremity; RLE right lower extremity. had full recovery of neurologic function and 3 had incomplete recovery with residual deficits (Table 3). The average age of patients with postoperative paraplegia or paraparesis was 74 9 (SD) years. Eight out of the 11 patients had a normal neurologic examination after emergence from general anesthesia and subsequently developed paraplegia or paraparesis. In the 8 patients who developed delayed-onset paraplegia or paraparesis, the average time to first detection of the postoperative neurologic deficit was (SD) hours after patient arrival to the intensive care unit (ICU) with a median of 21.6 hours and a range from 6.4 to hours. A second recurrence of postoperative paraparesis occurred in 2 patients (patient nos. 2 and 4) at hours (5.6 days) and hours (9.1 days), respectively, after arrival to the ICU. Three out of the 11 patients (patient nos. 3, 5, and 7) had paraplegia or paraparesis on the initial neurologic examination upon emergence from general anesthesia at an average time of (SD) hours (range 6.0 to 17.0 hours) after arrival to the ICU (Table 3). Two out of the 3 patients with paraplegia upon emergence had no recovery of neurologic function and died during hospitalization (patient nos. 3 and 5). The 1 patient with paraparesis upon emergence from general anesthesia had subsequent worsening of neurologic function (patient no. 7) and also died during hospitalization. Five patients had complete recovery (patient nos. 2, 4, 6, 9, and 11) with no residual neurologic deficits, and 3 patients had incomplete recovery with mild residual lower extremity weakness (Table 3). There was no operative mortality in patients who recovered from delayedonset paraplegia or paraparesis. A lumbar CSF drainage catheter was inserted after the induction of general anesthesia and used during operation in 9 out of 11 patients with postoperative paraplegia or paraparesis. The 2 patients who did not have a lumbar CSF drain inserted before operation had paraplegia upon emergence from general anesthesia. Lumbar CSF drains were not inserted before operation in these 2 patients (patient nos. 3 and 5) because of previous lumbar spine surgery in one patient (patient nos. 3) and hypotension after the induction of general anesthesia in the other (patient no. 5). Lumbar CSF catheters were emergently inserted in these 2 patients (patient nos. 3 and 5) after operation upon diagnosis of paraplegia, but neither patient recovered neurologic function in response to medical intervention. Delayed postoperative paraparesis developed in 5 patients after removal of the lumbar CSF catheter in the postoperative period (patient nos. 2, 4, 8, 10, and 11) and subsequently had emergent reinsertion of a lumbar CSF catheter for the treatment of paraparesis in the postoperative period. One patient (patient no. 2) had emergent reinsertion of a lumbar CSF drainage catheter for a second episode of postoperative paraparesis. Another patient (patient no. 4) had recovery of neurologic function with blood pressure augmentation alone after an initial episode of delayed-onset paraplegia after the removal of the lumbar drain. A lumbar CSF drain was emergently reinserted, however, for a second episode of delayed-onset paraparesis in that same patient (patient no. 4). The mean lumbar CSF pressure upon diagnosis of paraplegia or paraparesis was 14 3 (SD) mm Hg, and at the time of neurologic recovery was 10 3mmHg(p 0.001). The mean patient temperature at the time diagnosis was F, and at the time of recovery was F (p 0.87). An acute decrease in arterial pressure preceded the onset of postoperative paraplegia or paraparesis in 3

5 Ann Thorac Surg CHEUNG ET AL 2002;74: REVERSAL OF POSTOPERATIVE PARAPLEGIA 417 Fig 1. Systolic (downward-pointing triangles), diastolic (upward-pointing triangles), and mean (circles) arterial pressures in the period surrounding the onset (E) and recovery (R) from paraplegia after thoracic aortic reconstruction in 3 patients (patient nos. 1, 9, and 10). A decrease in arterial pressures preceded the onset of paraplegia. Arterial pressures were augmented by the administration of intravenous phenylephrine or norepinephrine (lower panels). Vasopressor requirements decreased after recovery from paraplegia. patients (patient nos. 1, 9, and 10) who had subsequent recovery of neurologic function (Fig 1). Vasopressor therapy to augment the MAP was administered to treat 7 out of the 8 patients with delayed-onset postoperative paraplegia. Arterial pressure was augmented in 1 patient by discontinuing nicardipine (patient no. 11). In the 10 episodes of delayed-onset paraplegia in the 8 patients (Table 2), the MAP was increased significantly from a mean of (SD) mm Hg at the time of diagnosis to a mean of 95 9 (SD) mm Hg at the time of recovery (p 0.003). The calculated SCP (Table 2) increased from a mean value of (SD) mm Hg at the time of diagnosis to 87 9 mm Hg at the time of recovery (p 0.002). Vasopressor agents administered to increase MAP were phenylephrine (n 5), norepinephrine (n 4), dopamine (n 2), dobutamine (n 1), and epinephrine (n 5). The maximum infusion dose of phenylephrine in the 5 patients who were treated with this agent averaged (SD) g/min, with range of 70 to 300 g/min. The maximum infusion dose of norepinephrine in the 4 patients who were treated with this agent averaged 11 2 (SD) g/min, with range of 8 to 13 g/min. At the time neurologic function recovered, vasopressor therapy was no longer needed to maintain a MAP 90 mm Hg and was weaned off (Fig 1). Postoperative bleeding was not observed as a consequence of arterial pressure augmentation used for the treatment of delayed-onset paraplegia. Patients who recovered from delayed postoperative paraplegia had minimal residual neurologic deficits (Table 3). The average ASIA lower-extremity motor score at diagnosis was (SD), and was similar to the average score of (SD) at the time of maximum neurologic deficit (p 0.35). Sensory deficits at the time of diagnosis were present in 10 out of the 11 patients with postoperative paraplegia. In the patients who recovered from paraplegia, the average score at recovery was 48 4 (SD), and was significantly greater than the score at the time of diagnosis or at the time of maximum neurologic deficit (p 0.005).

6 418 CHEUNG ET AL Ann Thorac Surg REVERSAL OF POSTOPERATIVE PARAPLEGIA 2002;74: Comment Postoperative paraplegia has been a well-recognized complication of operations to replace the descending thoracic or abdominal aorta and has an estimated incidence that ranges between 2.7% and 20% [3, 7]. Spinal cord ischemia and subsequent infarction as a consequence of temporary or permanent interruption of the vascular supply to the spinal cord during operation has been believed to be the major cause of postoperative paraplegia in this patient population. Distal aortic perfusion, deliberate hypothermia, reimplantation of intercostal arteries, lumbar CSF drainage, intraoperative neurophysiologic monitoring, and pharmacologic approaches to protect the spinal cord from ischemic infarction have all been tried, but these techniques used alone or in combination have not been completely effective at preventing or treating this complication. In contrast to paraplegia detected immediately upon emergence from general anesthesia, the delayed onset of paraplegia was a sign that the vascular supply to the spinal cord was not irreversibly damaged as a consequence of operation. Reports of full or partial recovery after delayed onset of paraplegia after TAA repair in the literature support reversible spinal cord ischemia as a pathophysiologic mechanism of this complication in some cases [3, 10, 12, 13]. Postoperative events such as hypotension, increased CSF pressure, thrombosis, hematoma, or embolization have the potential to cause spinal cord ischemia in patients with a compromised blood supply to the spinal cord as a consequence of TAA repair [1, 3, 8 10, 12 14]. Based on this pathophysiology, it was reasonable to predict that some patients experiencing delayed-onset paraplegia would respond to interventions directed at increasing spinal cord perfusion. The observation that all patients who had delayed-onset paraplegia after operation recovered neurologic function in response to interventions to augment spinal cord perfusion suggested that the majority of events were a consequence of spinal cord ischemia rather than infarction. In contrast, patients who were paraplegic upon emergence from anesthesia after operation did not recover neurologic function and probably suffered irreversible damage to the vascular supply of the spinal cord injury as a consequence of the operation or undetected spinal cord ischemia that evolved to infarction shortly after operation. The high mortality observed in patients with permanent paraplegia after operation was consistent with other clinical series [4, 7]. Emergency treatment of paraplegia with vasopressor therapy to augment the arterial pressure and lumbar CSF drainage to decrease the lumbar CSF pressure was immediately initiated with the objective to increase the net spinal cord perfusion pressure. The reason for this approach was based on the premise that hypoperfusion was the cause of spinal cord ischemia in the majority of cases. Lumbar CSF drainage to prevent postoperative paraplegia and for the treatment of delayed-onset paraplegia after aortic surgery has been reported, but it was not always successful and the technique remains controversial [7, 16 18]. The rationale for using vasopressor therapy to treat postoperative spinal cord ischemia was not novel. The clinical effectiveness of this strategy, however, has not been tested, in part, because of the perceived risk of bleeding associated with hypertension in this patient population. Norepinephrine and epinephrine were chosen together with phenylephrine in an effort to increase both the arterial pressure and cardiac output. Multiple vasopressor agents at relatively high doses were required often to generate a MAP greater than 95 mm Hg. Recovery of neurologic function coincided with vasopressorinduced increases in the arterial pressure. Vasopressor requirements also decreased during recovery. Hypotension may have been a factor contributing to the onset of spinal cord ischemia in some patients. It was possible also that hypotension preceding the onset of neurologic dysfunction may have also been an early sign of spinal cord ischemia and caused by autonomic dysfunction in a manner similar to the syndrome of spinal or neurogenic shock. Ischemia-induced autonomic dysfunction would explain also the need for high doses of vasopressor therapy during the event and the spontaneous recovery of arterial pressure after recovery of neurologic function (Fig 1). The relative contribution of lumbar CSF drainage and MAP augmentation with vasopressor therapy toward treatment success could not be separated out with certainty. Review of information from individual cases suggested that both interventions were important. Vasopressor therapy increased the MAP by an average of 21 mm Hg. Even though lumbar CSF drainage decreased the lumbar CSF pressure an average of only 4 mm Hg, the small change produced a relatively large proportional change in the estimated SCP pressure. The 4-mm Hg decrease in lumbar CSF pressure contributed 19% to the increase in the estimated SCP pressure. It was also likely that the actual lumbar CSF pressure at the onset of paraplegia was much greater than the values recorded because CSF was drained immediately upon emergent insertion of lumbar CSF catheters before the lumbar CSF pressure could be measured and recorded. The practice of delaying removal of the lumbar CSF drainage catheter until 48 hours after operation was justified because 75% of the episodes of delayed-onset paraplegia occurred within 48 hours after operation and the median onset time of paraplegia was 21.6 hours after operation. Routine postoperative neurologic assessment and evaluation of patients at risk were necessary to detect the onset of paraplegia. The importance of early diagnosis of neurologic dysfunction was evident in the cases where paraplegia was not detected until emergence from general anesthesia. In those cases where paraplegia was not detected at its onset, there was no recovery of neurologic function despite treatment. The treatment algorithm relied specifically on a clinical diagnosis of spinal cord ischemia. Magnetic resonance imaging or other imaging studies were not used for the initial evaluation because they would delay the start of treatment. It was recognized that some episodes of paraplegia could be caused by epidural hematoma as a consequence of anticoagulation

7 Ann Thorac Surg CHEUNG ET AL 2002;74: REVERSAL OF POSTOPERATIVE PARAPLEGIA 419 and lumbar CSF catheter insertion, but this complication has been rare compared with spinal cord ischemia [14]. Diagnostic imaging studies were reserved for patients who did not improve in response to blood pressure augmentation or lumbar CSF drainage. It was also important to note that the neurologic presentation at the onset of paraplegia was not always consistent with spinal cord ischemia in the distribution of the anterior spinal artery. Almost all patients had sensory deficits on presentation, and motor weakness was not always symmetrical. Finally, ongoing neurologic assessment was necessary during treatment. If there was no recovery of neurologic function with the initiation of treatment, the MAP was augmented above 95 mm Hg until a response was observed. Although the ASIA scores were not recorded at the time of the initial examination, studies suggest that conversion of neurologic findings documented in medical records were reliable and valid [19]. Technical improvements in intraoperative and circulatory management of patients undergoing TAA or TAAA repair may decrease the risk of intraoperative spinal cord infarction, but the risk of postoperative spinal cord ischemia remains problematic. Fortunately, our experience suggested that early detection and treatment of delayed-onset postoperative paraplegia led to recovery of almost full neurologic function in all patients and prevented the evolution of spinal cord ischemia to irreversible spinal cord infarction. The case series challenged earlier presumptions that postoperative paraplegia was an unavoidable, unpredictable, and untreatable complication of aortic reconstruction [7, 8, 20, 21]. Although it was not possible to determine with certainty the precise mechanism of postoperative paraplegia and the relative contribution of individual treatment interventions, the additive effects of lumbar CSF drainage and MAP augmentation to increase spinal cord perfusion pressure correlated temporally with the recovery of neurologic function. Considering the severity of this complication and its associated morbidity, present efforts are being directed towards recovering patients early after general anesthesia to permit early detection of spinal cord ischemia in order to initiate the treatment algorithm without delay while accumulating additional evidence to support its clinical effectiveness. We acknowledge Elizabeth Hoel and William Moser for their assistance in managing the aortic surgical database used in this study. We also wish to acknowledge the nursing staff of the Cardiothoracic Surgical Intensive Care Unit for their dedication to the project and attention to the early identification of patients with postoperative paraplegia. The study was unfunded. References 1. Fitzgibbon DR, Glosten B, Wright I, Tu R, Ready LB. Paraplegia, epidural analgesia, and thoracic aneurysmectomy. Anesthesiology 1995;83: Hollier LH, Money SR, Naslund TC, et al. Risk of spinal cord dysfunction in patients undergoing thoracoabdominal aortic replacement. Am J Surg 1992;164: Estrera AL, Rubenstein FS, Miller CC, III, et al. Descending thoracic aortic aneurysm: surgical approach and treatment using the adjuncts cerebrospinal fluid drainage and distal aortic perfusion. Ann Thorac Surg 2001;72: Cambria RP, Davison JK, Carter C, et al. Epidural cooling for spinal cord protection during thoracoabdominal aneurysm repair: A five-year experience. J Vasc Surg 2000;31: Coselli JS, LeMaire SA, Miller CC, III, et al. Mortality and paraplegia after thoracoabdominal aortic aneurysm repair: a risk factor analysis. Ann Thorac Surg 2000;69: Guerit JM, Witdoeckt C, Verhelst R, et al. Sensitivity, specificity, and surgical impact of somatosensory evoked potentials in descending aorta surgery. Ann Thorac Surg 1999;67: Crawford ES, Mizrahi EM, Hess KR, et al. The impact of distal aortic perfusion and somatosensory evoked potential monitoring on prevention of paraplegia after aortic aneurysm operation. J Thorac Cardiovasc Surg 1988;95: Heller LB, Chaney MA. Paraplegia immediately following removal of a cerebrospinal fluid drainage catheter in a patient after thoracoabdominal aortic aneurysm surgery. Anesthesiology 2001;95: Kasirajan K, Dolmatch B, Ouriel K, Clair D. Delayed onset of ascending paralysis after thoracic aortic stent graft deployment. J Vasc Surg 2000;31: Widmann MD, DeLucia A, Sharp J, Richenbacher WE. Reversal of renal failure and paraplegia after thoracoabdominal aneurysm repair. Ann Thorac Surg 1998;65: worksheet.pdf. Accessed January 15, Ortiz-Gomez JR, Gonzalez-Solis FJ, Fernandez-Alonso L, Bilbao JI. Reversal of acute paraplegia with cerebrospinal fluid drainage after endovascular thoracic aortic aneurysm repair. Anesthesiology 2001;95: Hill AB, Kalman PG, Johnston KW, Vosu HA. Reversal of delayed-onset paraplegia after thoracic aortic surgery with cerebrospinal fluid drainage. J Vasc Surg 1994;20: Weaver KD, Wiseman DB, Farber M, et al. Complications of lumbar drainage after thoracoabdominal aortic aneurysm repair. J Vasc Surg 2001;34: Azizzadeh A, Huynh TT, Miller CC III, Safi HJ. Reversal of twice-delayed neurologic deficits with cerebrospinal fluid drainage after thoracoabdominal aneurysm repair: a case report and plea for a national database collection. J Vasc Surg 2000;31: Ling E, Arellano R. Systematic overview of the evidence supporting the use of cerebrospinal fluid drainage in thoracoabdominal aneurysm surgery for prevention of paraplegia. Anesthesiology 2000;93: Shenaq SA, Svensson LG. Paraplegia following aortic surgery. J Cardiothorac Vasc Anesth 1993;7: Svensson LG, Hess KR, D Agostino RS, et al. Reduction of neurologic injury after high-risk thoracoabdominal aortic operation. Ann Thorac Surg 1998;66: Kasner SE, Chalela JA, Luciano JM, et al. Reliability and validity of estimating the NIH stroke scale score from medical records. Stroke 1999;30: Doppman JL. Paraplegia after surgery for thoracoabdominal aneurysms: Russian roulette for the vascular surgeon. Radiology 1993;189: Rosenthal D. Spinal cord ischemia after abdominal aortic operation: is it preventable? J Vasc Surg 1999;30:391 7.

8 420 CHEUNG ET AL Ann Thorac Surg REVERSAL OF POSTOPERATIVE PARAPLEGIA 2002;74: DISCUSSION DR NICHOLAS T. KOUCHOUKOS (St. Louis, MO): Dr Cheung and his colleagues have reported their results with a standard protocol in a recent time interval for the detection and treatment of delayed paraplegia following operations on the descending thoracic and thoracoabdominal aorta. This protocol involved emergent examination of the patient when a neurological deficit was detected, increase in the mean arterial pressure with vasopressors, and drainage of cerebrospinal fluid. Eleven patients, or 11% of the total, developed postoperative paraplegia. Five of the 8 patients with delayed paraplegia had complete recovery. It is clear from this study that prompt diagnosis and treatment of delayed paraplegia are essential for optimal outcomes, and the authors are to be commended for implementing an aggressive multidisciplinary approach to the management of this complication. What is less clear is the precise etiology of the delayed paraplegia and the relative importance of vasopressor therapy and cerebrospinal fluid drainage in its management. The mechanisms for the development of delayed paraplegia remain incompletely defined. These include a marginal spinal cord blood supply, increased cerebrospinal fluid pressure, intrathecal or epidural hemorrhage from instrumentation or anticoagulation, and delayed death of ischemic neurons from necrosis and apoptosis. As the preferred method for intraoperative management of extensive thoracic and thoracoabdominal aneurysms has evolved from the technique of simple aortic clamping to the use of some form of distal perfusion, usually with a degree of hypothermia, it appears that the prevalence of delayed paraplegia has increased. What is less certain is whether the overall prevalence of paraplegia has been reduced. It is not clear from Dr Cheung s presentation whether the prevalence of permanent paraplegia, occurring early or late, has been reduced in the subgroups of patients, those Crawford extent I and II disease, who have the highest rates of paraplegia. It is also unclear whether the two different techniques of distal perfusion used in this study, partial left heart bypass with moderate hypothermia, or total cardiopulmonary bypass with deep hypothermia and circulatory arrest, were associated with similar or differing rates of paraplegia. Whereas the treatment of delayed paraplegia is clearly important and often successful, prevention of this dreaded complication remains an equally important objective. (Slide) In our experience with 106 operative survivors whose neurologic status could be evaluated following Crawford extent I, II and III thoracoabdominal aneurysm repair using profound hypothermic cardiopulmonary bypass without cerebrospinal fluid drainage or measurement of cerebrospinal pressure, paraplegia occurred in 4 patients, just under 4%. It was delayed in onset in only 1 patient. In this patient, it was associated with hypotension that resulted from perforation of the colon and peritonitis on the 10th postoperative day. In an additional 76 patients who have had operations on the descending thoracic aorta using the same technique, no paraplegia has occurred. One patient developed a transient paraparesis. Thus, the overall prevalence of paraplegia in using this technique was 2.2%. We believe that more optimal protection of the spinal cord is provided intraoperatively by profound hypothermia and that this renders the spinal cord more tolerant to changes in blood pressure and increases in cerebrospinal fluid that may occur in the postoperative period. I have several questions for Dr Cheung. Did any patients die before their neurologic status could be evaluated? If so, the denominator would be decreased and the prevalence of paraplegia would be higher. Did you observe different rates of paraplegia, both early and delayed, between the patients with descending thoracic and thoracoabdominal aortic disease? In your abstract, which included 42 patients with type I, II, and III aneurysms, 10 of these patients developed paraplegia, and thus the prevalence would be 24%, which is relatively high. Did you observe different rates of paraplegia between the patients managed with atriofemoral mildly hypothermic bypass and those with hypothermic circulatory arrest? Do you have any concern about the use of catheters for drainage of cerebrospinal fluid in partially or fully heparinized patients? And finally, did you use any neuroprotective agents intraoperatively, such as barbiturates or corticosteroids, and was implantation of intercostal arteries an important part of the intraoperative protocol? This is an important study with important implications for the management of patients with postoperative paraplegia, and I congratulate the authors. Thank you. DR CHEUNG: Thank you, Dr Kouchoukos, for the excellent and outstanding comments and questions. Let me try to reply to some of these questions. Unfortunately, I do not have the data on the number of patients who died prior to the ability to perform a neurologic assessment. Our complication rate was based only on patients who had a definitive diagnosis of spinal cord ischemia by neurologic examination, and certainly there could have been patients who died with paraplegia that could not be examined and where the diagnosis could not be established. Query of our database revealed 3 patients who died before emergence from anesthesia and before a neurologic examination could be performed. In terms of the incidence of paraplegia by subgroups, we actually had an incidence of paraplegia of 5% in patients with isolated aneurysm or disease of the thoracic aorta that were managed by LA-FA bypass or distal perfusion. We had a 10% incidence of paraplegia in the patients managed with deep hypothermic circulatory arrest, those with proximal extension into the distal aortic arch. The highest incidence of paraplegia, as mentioned, was 17.5% in the patients with Crawford type I, II, and III thoracoabdominal aortic aneurysms. It is important to note that although our overall incidence of paraplegia was 11%, 8 out of the 11 patients with paraplegia had a complete or nearly complete recovery with a satisfactory functional outcome. Therefore the incidence of permanent paraplegia was only 3% in our study population. CSF drainage complications have been a concern for us, especially in fully anticoagulated patients. After insertion of the lumbar CSF drain, patients were fully anticoagulated for LA-FA distal aortic perfusion or deep hypothermic circulatory arrest. Out of our series of over 100 patients, we have not had a single episode of epidural hematoma, although we have had two episodes of retained CSF catheters breaking upon trying to remove these catheters, and we have had two incidents of meningitis and persistent CSF leaks. Although we recognize that epidural hematoma is a potential cause of postoperative paraplegia, our experience suggests that this is a rare complication compared with spinal cord ischemia as a consequence of operation. For this reason, the prompt treatment for presumed spinal cord ischemia should not be delayed in order to perform

9 Ann Thorac Surg CHEUNG ET AL 2002;74: REVERSAL OF POSTOPERATIVE PARAPLEGIA 421 imaging studies to rule out epidural hematoma. The patients who did not improve within a short period of time underwent emergent MRI to rule out an epidural hematoma and confirm the diagnosis of spinal cord infarction. In terms of neuroprotective agents, we routinely use methylprednisolone, mannitol, magnesium, and lidocaine administration intraoperatively prior to the institution of extracorporeal circulation. I am not sure if these drugs make a big difference or not, but that is part of our practice. Considering that evidence supporting the efficacy of these agents is indeterminate, we believe that there is little downside to administering these agents so long as there is no danger in doing so. In terms of postoperative use of pharmacologic agents, we have demonstrated initial success in our series with just blood pressure augmentation and CSF drainage. We reserve the use of pharmacologic agents such as high-dose steroid therapy that has been used in spinal cord injury patients as a third-line approach in patients with refractory paraplegia. We have not experienced success with high-dose steroid therapy in refractory paraplegia, but our numbers are very small (3/99 or 3%). To address your questions regarding operative technique, including the reimplantation of the intercostals arteries, I would like to defer to Dr Bavaria. DR BAVARIA: We have near 100% reimplantation on dissection cases, both acute and chronic, especially the chronic ones, and probably closer to only 25% reimplantation rates on the atherosclerotic aneurysms. Requirements for Recertification/Maintenance of Certification in 2003 Diplomates of the American Board of Thoracic Surgery who plan to participate in the Recertification/ Maintenance of Certification process in 2003 must hold an active medical license and must hold clinical privileges in thoracic surgery. In addition, a valid certificate is an absolute requirement for entrance into the recertification/maintenance of certification process. If your certificate has expired, the only pathway for renewal of a certificate is to take and pass the Part I (written) and the Part II (oral) certifying examinations. The American Board of Thoracic Surgery will no longer publish the names of individuals who have not recertified in the American Board of Medical Specialties directories. The Diplomate s name will be published upon successful completion of the recertification/maintenance of certification process. The CME requirements are 70 Category I credits in either cardiothoracic surgery or general surgery earned during the 2 years prior to application. SESATS and SESAPS are the only self-instructional materials allowed for credit. Category II credits are not allowed. The Physicians Recognition Award for recertifying in general surgery is not allowed in fulfillment of the CME requirements. Interested individuals should refer to the 2003 Booklet of Information for a complete description of acceptable CME credits. Diplomates should maintain a documented list of their major cases performed during the year prior to application for recertification. This practice review should consist of 1 year s consecutive major operative experiences. If more than 100 cases occur in 1 year, only 100 should be listed. Candidates for recertification/maintanance of certification will be required to complete all sections of the SESATS self-assessment examination. It is not necessary for candidates to purchase SESATS individually because it will be sent to candidates after their application has been approved. Diplomates may recertify the year their certificate expires, or if they wish to do so, they may recertify up to two years before it expires. However, the new certificate will be dated 10 years from the date of expiration of their original certificate or most recent recertification certificate. In other words, recertifying early does not alter the 10-year validation. Recertification/maintenance of certification is also open to Diplomates with an unlimited certificate and will in no way affect the validity of their original certificate. The deadline for submission of applications for the recertification/maintenance of certification process is May 10 each year. A brochure outlining the rules and requirements for recertification/maintenance of certification in thoracic surgery is available upon request from the American Board of Thoracic Surgery, One Rotary Center, Suite 803, Evanston, IL 60201; telephone number: (847) ; fax: (847) ; ; abts_ evanston@msn.com. This booklet is also published on the website: by The Society of Thoracic Surgeons Ann Thorac Surg 2002;74: /02/$22.00 Published by Elsevier Science Inc