- Email: [email protected]
Spinal cord ischemia after abdominal aortic operation: Is it preventable?
ORIGINAL ARTICLES For the Southern Association for Vascular Surgery
Spinal cord ischemia after abdominal aortic operation: Is it preventable? David Rosenthal, MD, for the Southern Association for Vascular Surgery, Atlanta, Ga Purpose: Spinal cord ischemia after operation on the abdominal aorta is a rare event that is attributed to variations in the spinal cord blood supply. The purpose of this study was to evaluate the possible causes of this devastating event. Methods: A survey of patients among the members of the Southern Association for Vascular Surgery was performed, and 18 patients were identified with spinal cord ischemia manifested by paraplegia or paraparesis after abdominal aortic operation. Results: Preoperative computed tomographic, magnetic resonance, and aortographic results did not visualize the greater radicular artery (Adamkiewicz’s artery) in any patient. Eleven patients underwent resection of infrarenal abdominal aortic aneurysms (AAAs): seven of these patients had tube grafts, three had aortobifemoral grafts, and one had an aortobiiliac graft. Five other patients underwent placement of aortobifemoral grafts, and one patient underwent aortobiiliac graft placement for occlusive disease. One patient underwent suprarenal AAA resection with an interposition graft to a previous aortobiiliac graft. The mean operative time was 3 hours and 39 minutes (range, 2 hours and 45 minutes to 6 hours and 30 minutes), with a mean aortic cross-clamp time of 48 minutes (range, 24 to 97 minutes). Sixteen aortic cross-clamps were placed infrarenally and two suprarenally (one in a case of ruptured AAA, the other a suprarenal AAA). Seventeen proximal anastomoses were end to end. The average minimum systolic blood pressure during the aortic cross-clamping was 96 mm Hg (range, 80 to 130 mm Hg). All the patients had internal iliac artery flow preserved with either prograde perfusion (10 patients) or retrograde perfusion (eight patients), and one patient underwent unilateral internal iliac artery ligation because of aneurysmal disease. One aortobifemoralgraft limb necessitated thrombectomy, but no cases of massive peripheral embolization occurred. When paraplegia was suspected after operation (6 to 20 hours after surgery), five patients underwent lumbar drainage. No clinical improvement was noted. Conclusion: Interference with pelvic blood supply from prolonged aortic cross clamping, intraoperative hypotension, aortic embolization, and interruption of internal iliac artery circulation have all been suggested as possible causes of spinal cord ischemia. In this survey, none of these factors proved to be significant as the sole cause of spinal cord ischemia. In the performance of an aortic operation with an end-to-end proximal anastomosis in the presence of severe external or internal iliac artery disease, there may be an increased incidence of spinal cord ischemia despite appropriate surgical techniques to ensure internal iliac perfusion. Spinal cord ischemia after abdominal aortic operations appears to be a tragically unpredictable, random, and unpreventable event. (J Vasc Surg 1999;30:391-9.)
From the Department of Vascular Surgery, Georgia Baptist Medical Center, Medical College of Georgia. Presented at the Twenty-third Annual Meeting of The Southern Association for Vascular Surgery, Naples, Fla, Jan 27–30, 1999. Reprint requests: David Rosenthal, MD, 315 Boulevard NE, Ste 412, Atlanta, GA 30312. Copyright © 1999 by the Society for Vascular Surgery and International Society for Cardiovascular Surgery, North American Chapter. 0741-5214/99/$8.00 + 0 24/6/100787
Spinal cord ischemia after operation on the abdominal aorta is a rare but devastating event that continues to haunt the vascular surgeon. Szilagyi et al,1 in a landmark 1978 publication, noted only a 0.25% incidence of paraplegia or pariparesis after 3614 abdominal aortic operations. Classically, the symptoms of spinal cord ischemia of the lower extremities include sensory and motor deficits associated with bladder/rectal incontinence with conserva391
392 Rosenthal
tion of vibratory and proprioceptive sensation, thus reflecting damage to the anterior horn of the spinal cord. The causative factors associated with spinal cord ischemia include interruption of the greater radicular artery (Adamkiewicz’s artery), prolonged aortic occlusion, intraoperative hypotension, atheromatous embolization, and interruption of the internal iliac artery circulation.2-8 The purpose of this report, the largest series to date, was to evaluate the possible causes of spinal cord ischemia after abdominal aortic operations and to determine whether it is preventable. MATERIALS AND METHODS A survey was performed among members of the Southern Association for Vascular Surgery, and 16 surgeons reported 18 patients who had spinal cord ischemia develop manifested by paraparesis or paraplegia after abdominal aortic operations. Of these 18 patients, 15 underwent preoperative computed tomographic (CT) scanning of the abdomen, 17 underwent aortography, and one underwent magnetic resonance (MR) scanning. Eleven patients underwent resection of infrarenal abdominal aortic aneurysms (AAAs). Seven of these underwent tube graft placement, three underwent aortobifemoral grafting, and one underwent placement of an aortobiiliac graft. One patient underwent suprarenal AAA resection with an interposition graft to a previous aortobiiliac graft. Six other patients underwent operation for aortoiliac occlusive disease: five had aortobifemoral graft placement and one patient had aortobiiliac graft placement. Hospital charts, operative notes, anesthesia records, radiologic reports, and follow-up data from all patients were retrospectively reviewed in an attempt to identify the factors that might have contributed to the occurrence of spinal cord ischemia. These possible factors included: type of operation, aortic crossclamp location and duration, presence of hypotension, use of heparin therapy, operative duration, maintenance of internal or external iliac artery perfusion, and intraoperative embolization. The postoperative neurologic examinations, electromyography, CT scanning, or MR imaging studies results, and the clinical course were reviewed to determine the nature and extent of the neurologic injury. RESULTS Fourteen of the 18 patients had paraplegia develop associated with bowel and bladder dysfunction, and four patients had paraparesis manifested by unilateral limb motor and sensory deficits. All the symptoms were noted within 20 hours after operation.
JOURNAL OF VASCULAR SURGERY September 1999
The intraoperative factors that may have contributed to spinal cord ischemia were evaluated (Table I). The aortic cross clamp was placed infrarenally in 16 of 18 patients, with a mean cross-clamp time of 48 minutes (range: 24 to 97 minutes). Seventeen proximal aortic anastomoses were performed end to end, and one aortobifemoral graft had an end-to-side aortic anastomosis. The mean intraoperative blood loss was 1250 mL (range, 300 to 9000 mL), and a cell saver was used during all the operations. Hypotension, defined as a systolic blood pressure less than 100 mm Hg, occurred in eight patients. The average minimum systolic blood pressure during aortic cross clamping in these eight patients was 96 mm Hg (range, 80 to 130 mm Hg). Intravenous heparin therapy was administered during 17 operations: one patient who had a ruptured AAA did not undergo heparin therapy. The mean duration of operation was 3 hours and 39 minutes (range, 2 hours and 45 minutes to 6 hours and 30 minutes). In 11 patients, preoperative arteriographic results identified patent external and internal iliac arteries, and six patients had either unilateral or bilateral occlusive disease (>50% reduction in diameter) of the internal iliac vessels. On the basis of the preoperative aortographic findings and the type of operation, internal iliac perfusion was maintained with either prograde flow (10 patients) or retrograde flow (eight patients). Two patients who had undergone aortobifemoral graft placement had inferior mesenteric artery reimplantation to protect the pelvic circulation: one of these patients had severe bilateral internal iliac artery occlusive disease and the other had an internal iliac artery aneurysm oversewn. No cases of massive embolization occurred, although one aortobifemoral graft limb necessitated thrombectomy. When paraplegia was suspected after operation (16 to 20 hours after surgery), five patients underwent urgent lumbar drainage. However, no clinical improvement was noted. Eleven patients underwent CT scanning, and nine underwent nuclear MR scanning of the spinal cord. One MR scan showed a zone of high-signal intensity consistent with edema after cord infarction. Eight patients died within 30 days of operation. Death was attributed to cardiac causes in five patients and to visceral or somatic ischemia that led to multi-system organ failure in three patients. Ten patients survived: four patients remained paraplegic, three patients with postoperative paraparesis were able to ambulate with assistance (ie, walker, leg brace) at discharge, and three patients were lost to follow-up.
JOURNAL OF VASCULAR SURGERY Volume 30, Number 3
DISCUSSION Ischemic injury to the spinal cord after abdominal aortic reconstruction is fortunately rare. Since the initial clinical report by McCune9 in 1956, we are aware of only 111 reported English-language cases.10-36 Of these 111 patients, 70 (63%) underwent operations for AAAs (39 ruptured, 31 elective) and 41 (37%) underwent operations for aortoiliac occlusive disease. The early, elegant work of Adams and Geertruyden9 indicated that spinal cord ischemia most frequently occurred after operation for ruptured AAA because these patients had significant periods of hypotension, often underwent suprarenal or supraceliac cross clamping, and usually did not undergo anticoagulation therapy. Of interest, however, was the fact that nearly two thirds (64%; 72 of 111) of the reported patients who had spinal cord ischemia develop had undergone elective abdominal aortic operations. This finding is similar in our series, in which 17 of 18 patients underwent elective operations, and indicates that the risks associated with emergent operations (ie, hypotension, suprarenal or celiac cross clamping) may not be such causative factors of spinal cord ischemia as previously believed. The location and duration of the aortic cross clamp have been suggested as possible causes of spinal cord ischemia. In this review, 16 aortic cross clamps were placed infrarenally. All the patients, except one (with a ruptured AAA), underwent anticoagulation therapy, and the mean aortic crossclamp time was 48 minutes. Forty-five minutes of cross-clamp time has been shown to be a safe limit, although no distinct limit has been defined.37 The location and duration of the aortic cross clamp in our series could not, therefore, be indicted as probable causes of spinal cord ischemia. Hypotension plays an important role in the cause of spinal cord ischemia. Several postmortem study results have shown that ischemia may occur without occlusion of the anterior spinal artery.38 If the distal spinal cord blood supply is compromised, severe hypotension alone may surpass the safe limits of cord perfusion. The average minimum blood pressure in our study among eight patients who were hypotensive was 96 mm Hg (range, 80 to 100 mm Hg). Lynch et al,39 however, found an increased risk of central nervous system ischemia when the mean arterial pressure fell below 55 mm Hg, a number that correlated well with the loss of brain and spinal cord autoregulation observed experimentally at 50 mm Hg. Certainly, hypotension should be avoided during the perioperative period, especially in patients with severe arterial occlusive disease, but hypoten-
Rosenthal
393
Fig 1. Intrinsic and extrinsic blood supply to spinal cord. Unshaded area is perfused by anterior spinal artery, shaded area by posterior spinal arteries. (Plecha EJ, Seabrook GR, Freischlag JA, et al. Neurologic complications of reoperative and emergent abdominal aortic reconstruction. Ann Vasc Surg 1995;9:95-101. Reprinted with permission).
sion alone in our patients was not believed to be a cause of spinal cord ischemia. The etiology of spinal cord ischemia after abdominal aortic operation is likely multifactorial, but the fundamental cause is always an alteration in blood supply to the spinal cord. The blood supply to the cord consists of an intrinsic and extrinsic network of arteries.4,7,36 The intrinsic blood supply is composed of three longitudinal systems: one anterior and two posterior spinal arteries (Fig 1). Intramedullary anastomoses between the anterior and posterior spinal systems are nonexistent except along the circumference, where the arterial vasocoronae coalesce. The anterior spinal artery extends the full 40-cm to 45cm length of the spinal cord, but it can be so narrow at the distal thoracic level that it cannot compensate for an interrupted radicular artery, thus increasing the risk of infarction after loss of a single radicular artery.37 The anterior spinal artery is, therefore, the major independent source of spinal cord perfusion, and inadvertent interruption of the largest of its segmental radicular arteries (the greater radicular artery or the artery of Adamkiewicz) has been implicated as the principal cause of spinal cord ischemia.4
394 Rosenthal
Fig 2. Anatomy of blood supply to distal cord and lumbosacral plexus. (Gloviczki P, Cross SA, Stanson AW, et al. Ischemic injury to the spinal cord or lumbosacral plexus after aorto-iliac reconstruction. Am J Surg 1991;162:1316. Reprinted with permission).
The extrinsic blood supply to the spinal cord is made up of segmental medullary feeder arteries that supply the anterior median spinal artery and the spinal arteries, which are present at each vertebral level (Fig 2). The anterior spinal artery is fed by these segmental medullary arteries, which originate from the subclavian, intercostal, and lumbar arteries. Perfusion of the distal spinal cord is derived from the lumbar, iliolumbar, and lateral sacral arteries that anastomose with the intrinsic spinal arteries at the level of the conus medullaris. The greater radicular artery (Adamkiewicz’s artery) is located between T9 and T12 in 75% of the cases, but, if it originates abnormally high or is chronically compromised,
JOURNAL OF VASCULAR SURGERY September 1999
these lower lumbar arteries that anastomose with the intrinsic network may contribute significantly to the blood supply of the terminal spinal cord.7 In general, when the greater radicular artery is open and of normal size, the pelvic blood supply is of minor importance. When the greater radicular artery is compromised, however, the pelvic blood supply may be critically important. Indeed, Picone et al4 identified several factors that may lead to spinal cord ischemia as the result of interference of pelvic blood flow and therefore should be avoided: (1) internal iliac devascularization with resultant severe pelvic oligemia, (2) significant hypotension, and (3) massive embolization to the pelvis. These authors suggested that intraoperative technical factors that may lead to these problems be avoided by revascularization of the internal iliac artery with a separate graft, by the avoidance of hypotension, and by antegrade flushing and irrigation of the graft. The authors also recommended that if an internal iliac artery aneurysm must be oversewn, it should be done so from within the aneurysm to preserve the collateral branch flow to the lateral sacral and ilial lumbar arteries. It was of interest to note that among the 10 patients in our series who underwent aortobifemoral (n = 8) or aortobiiliac (n = 2) grafting, only one proximal aortic anastomosis was end to side, which theoretically might preserve pelvic flow. Of these 10 patients, however, only four had internal iliac occlusive disease that was identified before surgery (Table I). In the performance of an aortic operation with an endto-end proximal anastomosis in the presence of significant iliac artery occlusive disease, the better part of surgical wisdom is to ensure internal iliac artery perfusion. If iliac perfusion is uncertain or seems inadequate, a separate internal iliac artery graft likely is appropriate. On the basis of the anatomy of the blood supply to the distal spinal cord, nerve roots, and lumbosacral plexus, correlated with the observed neurologic deficit, Gloviczki et al7 identified six types of ischemic neurologic injury. Type I cord injuries are associated with global ischemia to the distal cord and conus: these injuries are characterized by flaccid paraplegia and poor long-term recovery. A type II injury, or anterior spinal artery syndrome, is characterized by flaccid paraplegia with loss of pain and temperature sensation, with preservation of proprioception and vibratory sensation as the result of the posterior spinal artery circulation. A type II injury is frequently seen after thoracic or thoracoabdominal reconstruction, and the prognosis is guarded, with minimal long-term
JOURNAL OF VASCULAR SURGERY Volume 30, Number 3
Rosenthal
395
Table I. Risk factors for spinal cord ischemia after abdominal aortic operations Operation AAA
Location/duration Hypotension (systolic, Heparin of aortic cross clamp <100 mm Hg) (IU) Infrarenal/30 minutes
Operation time
Internal iliac artery status
None
5000
3 hours, 10 minutes
Patent bilaterally
AAA/ABF Infrarenal/95 minutes and AR Suprarenal AAA SupraSMA/59 minutes bilateral renal reimplant AAA/ABI Infrarenal/48 minutes
90 mm Hg × 40 min None
7500
6 hours, 30 minutes
5000
3 hours, 10 minutes
Bilateral 50% stenosis 50% unilateral stenosis
None
5000
3 hours, 39 minutes
Patent bilaterally
AAA/ABF/ IMA graft
None
5000
3 hours
85 mm Hg × 35 minutes
9000
6 hours, 24 minutes
Patent bilaterally (internal iliac aneurysm oversewn) Unilateral occlusion
5000
3 hours
Patent bilaterally
5000
3 hours, 30 minutes
Patent bilaterally
No
3 hours
Unknown
4000
2 hours, 45 minutes
Patent bilaterally
Infrarenal/60 minutes
ABF with IMA, Infrarenal/57 minutes SMA thrombectomy, bowel resection AAA Infrarenal/63 minutes AAA
Infrarenal/30 minutes
Ruptured AAA
Infrarenal/45 minutes
ABF
Infrarenal/25 minutes
90 mm Hg × 6 minutes 80 mm Hg × 15 minutes 90 mm Hg × 6 minutes None
ABF/IMA reimplant
Infrarenal/31 minutes
None
5000
3 hours, 30 minutes
AAA/ABF
Infrarenal/97 minutes
5000
5 hours, 40 minutes
ABF/FP
Infrarenal/38 minutes
80 mm Hg × 14 minutes None
5000
4 hours, 20 minutes
Severe bilateral iliac occlusive disease Occluded bilaterally Patent bilaterally
AAA
Infrarenal/48 minutes
None
6000
4 hours, 30 minutes
Patent bilaterally
AAA/ABF
Infrarenal/24 minutes
90 mm Hg × 8 minutes None None
6000
3 hours, 30 minutes
5000 5000
4 hours, 10 minutes Unknown
50% stenosis bilaterally Patent bilaterally Patent bilaterally
5000
3 hours, 15 minutes
Unknown
ABF Infrarenal/40 minutes ABF/bilateral Infrarenal/49 minutes RA grafts/FP AAA/ABI Suprarenal/unknown
<80 mm Hg × 12 minutes
Neurologic deficit Paraplegia with bowel/ bladder dysfunction Paraparesis Paraplegia with bowel/ bladder dysfunction Paraplegia with bowel/ bladder dysfunction Paraplegia with bowel/ bladder dysfunction Paraplegia with bowel/ bladder dysfunction Paraplegia with bowel/ bladder dysfunction Paraplegia with bowel/ bladder dysfunction Paraplegia intact sensation Paraplegia with bowel/ bladder dysfunction Paraparesis Paraparesis Paraplegia with bowel/ bladder dysfunction Paraplegia with bowel/ bladder dysfunction Paraparesis Paraparesis Paraplegia with bowel/ bladder dysfunction Paraplegia with bowel/ bladder dysfunction
AAA, Abominal aortic aneurysm repair; ABF, aortobifemoral graft placement; AR, aorto renal; ABI, aortobiiliac graft placement; IMA, inferior mesenteric artery repair; SMA, superior mesenteric artery repair; FP, femoropopliteal bypass grafting; RA, renal artery repair.
improvement. Type III and IV injuries involve the nerve roots and lumbosacral plexus. Most cord functions are therefore preserved but with neurologic asymmetric motor and sensory deficits with or without bladder and bowel incontinence. Because the injury is at the axonal level, patients with type III and IV injuries have favorable prognoses. A type V injury is caused by segmental spinal cord infarction and may be seen after embolization to the cord at a specific level or hypoperfusion in a watershed region, and type VI injuries occur as a result of interruption of the posterior spinal arteries blood supply to the posterior columns. These injuries are rare and are characterized
by loss of proprioception and vibratory sensation alone. The prognoses for type V and VI injuries are favorable. The neurologic deficits in our patients appear to have been type II deficits (anterior spinal artery syndrome) in 13 patients. Type III or IV injuries occurred in four patients, and type V occurred in one patient. The type III, IV, and V injuries occurred in patients who had undergone aortobifemoral bypass grafting with end-to-end proximal anastomoses and may have been the result of inadequate retrograde pelvic arterial flow. In an editorial in 1993, Szilagyi40 reported that the approximate incidence of paraplegia after AAA
396 Rosenthal
was 1 in 400 and after aortoiliac reconstruction for occlusive disease was 1 in 5000. Szilagyi stated that the greater radicular artery may be “harmlessly occluded much more frequently than it is occluded with serious consequences”40 and attributed this to ample pelvic collateral flow, which prevents damaging ischemia to the spinal cord. Spinal cord ischemia after abdominal aortic operations is likely multifactorial and may involve interference with pelvic blood supply, prolonged aortic cross clamping, intraoperative hypotension, and aortic embolization. As stated earlier, however, none of these factors could be indicted in our study as a sole cause of cord ischemia. It cannot, however, be refuted that the revascularization of one internal iliac artery, the avoidance of profound hypotension and its associated low-flow state to the spinal cord, the avoidance of prolonged supraceliac cross clamping, the intraoperative use of heparin therapy, and the gentle surgical techniques to avoid embolization are all sound operative principles that minimize the potential for spinal cord injury. If the surgeon must perform an end-to-end proximal aortic anastomosis in the presence of severe iliac occlusive disease, awareness of these principles will hopefully decrease the incidence of spinal cord ischemia. Nevertheless, spinal cord ischemia after abdominal aortic operations remains a tragically unpredictable, random, and unpreventable event. REFERENCES 1. Szilagyi DE, Hageman JH, Smith RF, Illiott JP. Spinal cord damage in surgery of the abdominal aorta. Surgery 1978; 83:38-56. 2. Zuber WF, Gaspar MR, Rothschild PD. The anterior spinal artery syndrome—a complication of abdominal surgery: report of five cases and review of the literature. Am Surg 1970;172:909-15. 3. Ferguson LR, Bergan JJ, Conn J Jr, Yao JST. Spinal ischemia following abdominal aortic surgery. Ann Surg 1975;181: 2677-82. 4. Picone AL, Green RM, Ricotta JR, May AG, DeWeese JA. Spinal cord ischemia following operations on the abdominal aorta. J Vasc Surg 1986;3:94-103. 5. Iliopoulos JI, Howanitz PE, Pierce GE, Kueshkerian SM, Thomas JH, Hermreck AS. The critical hypogastric circulation. Am J Surg 1987;154:671-5. 6. Dimakakos P, Arapoglou B, Katsenis K, et al. Ischemia of the spinal cord following elective operative procedures of the infrarenal aorta. J Cardiovasc Surg 1996;37:243-7. 7. Gloviczki P, Cross SA, Stanson AW, et al. Ischemic injury to the spinal cord or lumbosacral plexus after aorto-iliac reconstruction. Am J Surg 1991;162:131-6. 8. Senapati A, Browsf NL. Gluteal necrosis and paraplegia following postoperative bilateral internal iliac artery occlusion. J Cardiovasc Surg 1990;31:194-6. 9. McCune WS. Discussion. In: Adams HD, Geetruyden HH, editors. Neurologic complications of aortic surgery. Ann Surg 1956;144:574:610-6.
JOURNAL OF VASCULAR SURGERY September 1999
10. Bates T. Paraplegia following resection of abdominal aneurysm. Br J Surg 1971;58:913-6. 11. Coupland GAE, Reeve TS. Paraplegia: a complication of excision of abdominal aortic aneurysm. Surgery 1964; 878:19-23. 12. Di Chiro G, Doppman JL. Paraplegia after resection of aneurysm. N Engl J Med 1969;281:79-83. 13. Edmondson HT, Gindin RA. Paraplegia as a complication of abdominal aortic resection. Am Surg 1970;36:383-7. 14. Golden GT, Sears HF, Wellons HA Jr, Muller W Jr. Paraplegia complication resection of aneurysms of the infrarenal abdominal aorta. Surgery 1973;73:91-5. 15. Gump FE. Paraplegia after resection of aneurysm. N Engl J Med 1969;281:798-801. 16. Hara M, Lipin RJ. Spinal cord injury following resection of abdominal aortic aneurysm. Arch Surg 1960;80:419-22. 17. Hogan EL, Romanul FCA. Spinal cord infarction occurring during insertion of aortic graft. Neurology 1966;16:67-74. 18. Jellinger K. Spinal cord arteriosclerosis and progressive vascular myelopathy. J Neurol Neurosurg Psychiatry 1967; 30:195-8. 19. Lake PA. Paraplegia after resection of aneurysm. N Engl J Med 1969;281:798-801. 20. Mehrez IO, Nasbeth DC, Hogan EL, Deterling RA. Paraplegia following resection of abdominal aortic aneurysm. Ann Surg 1962;156:890-8. 21. Pasternak BM, Boyd DP, Ellis FH Jr. Spinal cord injury after procedures on the aorta. Surg Gynecol Obstet 1972;135: 29-34. 22. Reich MP. Paraplegia following resection of an abdominal aortic aneurysm: report of a case of atheromatous embolization to the anterior spinal artery. Vasc Surg 1968;2:230-4. 23. Sher MH, Healy EH. Paraplegia following infrarenal aneurysmorrhaphy. Vasc Surg 1971;5:171-4. 24. Skillman JJ, Zervas NT, Weintraub RM, Mayman C. Paraplegia after resection of aneurysms of the abdominal aorta. N Engl J Med 1969;281:422-4. 25. Lynch C, Weingarden SI. Paraplegia following aortic surgery. Paraplegia 1982;20:196-200. 26. Mills RP, Robbs JV, Duncan HJ. Paraplegia following abdominal aortic surgery. S Afr Med J 1983;63:930-2. 27. Friedman SG, Moccio CG. Spinal cord ischemia following elective aortoiliac reconstruction. Ann Vasc Surg 1988;2:295-7. 28. Michaels L. Case report: spinal cord ischemia associated with repair of a ruptured abdominal aortic aneurysm. Stroke 1972;3:238-42. 29. Connolly JE. Prevention of paraplegia secondary to operation on the aorta. J Cardiovasc Surg 1986;27:410-7. 30. Nielson BP. Ischemic injury to the spinal cord as a complication to abdominal aortic surgery. Acta Chir Scand 1985; 151:433-5. 31. Perry MO. Spinal cord injury following aortorenal bypass. Cardiovascular Surg 1979;20:261-4. 32. Salam AA, Sholkamy SM, Chaikof EL. Spinal cord ischemia after abdominal aortic procedures: is previous colectomy a risk factor? J Vasc Surg 1993;17:1108-10. 33. Graci RR, Matton KL. Anterior spinal artery syndrome following abdominal aortic aneurysmectomy. Arch Surg 1977; 112:813-5. 34. Gunderren J, Liungqvist V, Bergentz SE. Paraplegia after resection of a ruptured aortic aneurysm. Vasa 1980;9:304-5. 35. Jacoss PB, Dicke HW. Ischemic damage to the spinal cord following surgery of the abdominal aorta. Neth J Surg 1984; 36:1-5.
JOURNAL OF VASCULAR SURGERY Volume 30, Number 3
36. Plecha EJ, Seabrook GR, Freischlag JA. Neurologic complications of reoperative and emergent abdominal aortic reconstruction. Ann Vasc Surg 1995;9:95-101. 37. Defraigne JO, Otto B, Sakalihasan N, Limet R. Spinal ischemia after surgery for abdominal infrarenal aortic aneurysm: diagnosis with nuclear magnetic resonance. Acta Chir Belg 1997; 97:250-6. 38. Singh V, Silver JR, Weply NC. Hypotensive infarction of the spinal cord. Paraplegia 1994;32:314-22.
DISCUSSION Dr Elliot L. Chaikof (Atlanta, Ga). In a sobering presentation, Dr Rosenthal and colleagues have highlighted that spinal cord ischemia after abdominal aortic reconstruction is somewhat akin to playing Russian roulette: the gun is always loaded and ready to fire, and little or no forewarning is provided. In medicine, we take comfort from the time-honored practice of analyzing our outcomes, particularly those that are less than ideal. Implicit to this approach is the assumption that in experience there are always lessons to be learned and more appropriate paths that can be pursued the next time around. Unfortunately, the patterns of life’s experiences are occasionally not very neat, and, in this regard, spinal cord ischemia after abdominal aortic surgery appears to be excluded from this paradigm. Second guessing is common after these rare but catastrophic events. However, most recommendations, including the avoidance of systemic hypotension or prolonged suprarenal clamping, are so general and inherent to the standard practice of aortic surgery as to have little practical value. I have always held to the belief that the preservation of pelvic circulation is likely the most important factor in avoiding neurologic complications after abdominal aortic reconstruction. Although no one would deny the potential significance of pelvic ischemia, this could not be identified as a dominant factor in this collected series. In fact, in at least seven of these 18 patients, the operation was routine in every possible way, without systemic hypotension or significant iliac artery disease and with infrarenal clamping of short duration. An additional six patients had relatively mild hypotensive episodes during surgery but, in all cases, for periods of less than 15 minutes. This report may find its greatest impact in the medicolegal arena, and I believe it is important that events as rare and unpredictable as these not dictate established practice patterns. Nonetheless, the underlying message in this review is that bad things sometimes happen to good people during surgery, despite everyone’s best efforts. This knowledge held in the back recesses of our consciousness should give us some pause and require that recommendations for surgical intervention always be provided in a most considered manner. As a final comment, I believe it is worth noting that this report is an important outgrowth of Dr Clagett’s rec-
Rosenthal
397
39. Lynch DR, Dawson TM, Raps EC, Galetta SL. Risk factors for the neurologic complications associated with aortic aneurysms. Arch Neurol 1992;49:284-8. 40. Szilagyi DE. A second look at the etiology of spinal cord damage in surgery of the abdominal aorta. J Vasc Surg 1993; 17:1111-3.
Submitted Feb 1, 1999; accepted Jun 17, 1999.
ommendations to our Association during his 1998 presidential address. The collective experience of a society whose members practice in both community and academic settings can indeed be harnessed to address some very significant clinical problems. In sharing their experience, Dr Rosenthal and the participating members of the Southern Assocation for Vascular Surgery have done us all an important service. I have three questions for Dr Rosenthal. First, eight of the 18 patients died within 30 days of operation from somatic or visceral ischemia. Can this be taken as an indication that, in this series, atheroembolization is the likely cause of most cases of spinal ischemia? Second, anesthesiologists, on occasion, remind us that an epidural hematoma can complicate catheter placement in the patient who undergoes aortic surgery. Were you able to exclude epidural hematoma as a causative factor of spinal ischemia in this series? Finally, how much information should we provide to our patients on a routine basis regarding the risks of rare and random catastrophic events? I would like to thank the Association for the privilege of discussing this report. Dr David Rosenthal. The first question was in reference to atheromatous emboli, a common cause of spinal cord ischemia. In actuality, of the eight deaths, five were caused by cardiovascular sequelae and three, as Dr Chaikof mentioned, were related to somatic or visceral ischemia. These, however, were not the massive, catastrophic embolization problems that other have reported in which profound ischemia to the pelvis and lower extremities occurs. I believe three of the deaths may have been embolic related, but there is no way of proving this. The second question was in reference to epidural hematoma, which is an appropriate question. As I believe, all the patients for elective surgery had epidural catheters. All of the patients in the study underwent radiologic diagnostic studies when spinal cord ischemia was suspected. Eleven had computed tomographic scans and nine had magentic resonance scans, but there was no evidence of epidural hematoma in any of these patients. Interestingly, only one patient, on a magentic resonance scan, had edema of the cord that indicated ischemia. Certainly, if a patient awakens from an aortic operation and has lumbar pain and a persistent motor blockade
398 Rosenthal
when there is recovery from sensory anesthesia, we should be alerted that an epidural hematoma may be present. Because early intervention is the key to success in the management of an epidural hematoma, urgent neurologic assessment with computed tomographic or magentic resonance scanning is mandatory, followed by urgent operation to evacuate the hematoma. The last question related to the concern of the spinal cord ischemia being part of our informed consent, and, as usual, Dr Chaikof has hit to the heart of the matter. I personally have never mentioned this as a possible risk to a patient who was about to undergo aortic operation, but at my hospital in Georgia the standard hospital consent form lists paraplegia as an operative risk. As Dr Chaikof also mentioned, this raises a possible medicolegal problem for all of us, which possibly should be addressed by our National Vascular Societies Ethics Council to offer all of us structured guidelines. Dr G. Melville Williams (Baltimore, Md). Dr Rosenthal, I really appreciated your in-depth analysis, and I just stand to request more information on the diagnostic evaluation of these patients after the event. It has been interesting to me, in doing thoracoabdominal procedures, that we have evolved from the dense paraplegia that happened all too frequently, so that the patient would awake paralyzed, to a more subtle form in which the paraplegia occurs days later and may even be a monoplegia or some other weird syndrome. I have tried to figure this out with magnetic resonance imaging, magnetic resonance angiography and all this business of the spinal cord blood supply, and what I plead for is a thorough and intense neurologic evaluation by every method we now know to try to characterize exactly what goes wrong in these patients. I am persuaded that a lot happens with nerve root ischemia or injury as opposed to the spinal cord itself. I do not know how else you can explain the paraparalysis limited to one extremity. And even when we do magnetic resonance imaging studies, there is controversy in the interpretation. There can be a little streak that someone sees and says, “Ah ha, that is indicative of an ischemic insult.” So, I think one of the things as a society we ought to demand, and I think it is going to help us in every medicolegal adventure that we get into with these kind of patients, is a complete neurologic evaluation with every known method to try to investigate just what is ischemic and what went wrong. Dr Rosenthal. Dr Williams, I could not agree more. In this retrospective study, you saw how variable the studies were: some patients had electromyographys, some nerve conduction velocity studies, some the computed tomographic or magnetic resonance scans. From what I learned in my readings, it seems that if you are going to obtain only one study, a magnetic resonance scan is the most appropriate, but to gain more knowledge on the pathophysiology of what has occurred, a complete battery of studies should be analyzed. Dr Roger T. Gregory (Norfolk, Va). Dr Rosenthal, I
JOURNAL OF VASCULAR SURGERY September 1999
thoroughly enjoyed that presentation. Fortunately, we were not part of this group. Paraplegia has not been seen by our group in Norfolk in more than 30 years. We have, however, seen three patients with lumbosacral plexopathy after ruptured aneurysms, which we reported a few years ago. These were type III and type IV aneurysms or a type IV plexopathy. All of these were in ruptured aneurysms. Aggressive therapy and evaluation, as Dr Williams has just suggested, however, resulted in all three patients in an ambulatory status a few months later. Therefore, I think it is incumbent on all of us, and I would certainly stress your point, to evaluate these patients thoroughly, because those with plexopathy have actually a very reasonable prognosis. Dr Rosenthal. It is interesting, Dr Gregory, that we did a very extensive literature review, which we mention in the manuscript, and as far as I could tell, in the English language, there have been 107 case reports of paraplegia or paraparesis in the literature dating back to 1956, when the first patient was identified. All of the early reports, in the first decade, were on ruptured abdominal aortic aneurysms. But the thing that struck me when I was doing the literature search and review was that in the last several decades, more than two thirds of all these reports are in patients for elective surgery and paraphrase what Dr Chaikof said. I think that when you get that paraparesis, it is because of one of two factors: it is either microembolization to the conus region, or sometimes when you get these patients who have a massive hematoma tracking down the gutter towards the pelvis, you can get literally ischemia of the spinal roots as they are coming out of the pelvis itself. But those are the patients that get reclamation of function. Dr Hazim J. Safi (Houston, Tex). I enjoyed your presentation, and I think this is going to be a landmark paper. We have seen two kinds of postoperative paraplegia or neurological deficit—immediate and delayed—and we found that delayed paraplegia responds to cerebrospinal fluid drainage. I would like to know if you looked into the phenomenon of immediate delayed paraplegia. Dr Rosenthal. All of the patients were suspected of having spinal cord ischemia within 20 hours of operation. The delayed cord ischemia symptoms that you and Dr Williams see with the thoracoabdominal operations as the result of a possible hyperemia response may well respond to cerebrospinal fluid drainage, as others have recommended. Six of the 18 patients, I believe, had cerebrospinal fluid drainage as soon as the neurologic event was suspected, but tragically their clinical outcome did not improve. Thank you. Dr Gary M. Gross (Huntsville, Ala). This complication is a rare event. A survey of the Society is a way, perhaps the only way, of finding enough of these events to be worth studying. I am glad that the Lifeline Foundation is considering supporting some of these survey studies. My question is about an end-to-side case. Presumably, the existing pelvic circulation and the collaterals were preserved. What was the suspected cause in that case?
JOURNAL OF VASCULAR SURGERY Volume 30, Number 3
Dr Rosenthal. I do not know. That patient had undergone an elective aortobifemoral graft placement, with one internal iliac artery patent on a preoperative arteriography, yet the patient awoke from anesthesia paraplegic. You would have to think this was caused by either a low anterior spinal artery takeoff or pelvic oligemia. It is interesting that when we were gathering data for the study, I literally took 10 sheets of paper and hung them on a basement wall to make a flow chart in the hope we could finally find an answer to this devastating problem. We have 18 patients, this is the largest series reported, and despite all our efforts, many of the same questions remain unanswered. Dr S. Edwin Duncan (Tyler, Tex). Dr Rosenthal, I enjoyed the paper. The question I had was that some of the old literature talked about the significance of large, maybe not paired, lumbar vessels that were a possible etiology. Did you notice such a lumbar artery, and would you consider reimplantation of some significant honker lumbar? And I just wondered, in any of your review of these things, was that mentioned as something that was a consideration? Is that something we need to be aware of? Is that another risk factor to pay attention to? Thank you. Dr Rosenthal. In the literature that I reviewed there is anecdotal information on the large lumbar artery that does not bleed back, similar to what we see during thoracoabdominal aneurysm repair. Obviously, most of the lumbars that we are dealing with in the infrarenal aorta we routinely oversew because of vigorous back bleeding. In our data retrieval form, and in the operative notes I reviewed, I specifically looked for mention of lumbar vessels but could not find anything consistent. Certainly, after having been part of this study, if I come across an inordinately large patent lumbar artery during an aneurysm
Rosenthal
399
operation that is not back bleeding, I would strongly consider reimplantation. Dr Jose Alvarez (Miami, Fla). Did you notice how many of these patients had epidural anesthesia, because most of them underwent elective aneurysmectomies, and was there any role for this intervention in the course of their paraplegia now that they underwent heparin therapy and we know about that problem? Dr Rosenthal. This is similar to Dr Chaikof’s question and epidural hematoma. As I mentioned, each patient underwent a radiologic study, either computed tomographic or magnetic resonance scanning, but no epidural hematoma was identified.
APPENDIX Contributing members Colby P. Atkins, MD, Lexington, Ky Robert A. Brigham, MD, West Reading, Pa T.J. Bunt, MD, Bullhead City, Ariz Matthew P. Campbell, MD, Houston, Tex G. Patrick Clagett, MD, Dallas, Tex Steven J. Eskind, MD, Nashville, Tenn Robert W. Feldtman, MD, Houston, Tex D. Nelson Gwaltney, MD, Bristol, Tenn William D. Jordan, Jr, MD, Birmingham, Ala Herman W. Kaebnick, MD, Louisville, Ky Irving L. Kron, MD, Charlottesville, Va Samuel R. Money, MD, New Orleans, La Andrew J. Olinde, MD, Baton Rouge, La J. Mark Rheudasil, MD, Atlanta, Ga Edward E. Rigdon, MD, Jackson, Miss Thomas H. Schwarcz, MD, Lexington, Ky
Spinal cord ischemia after abdominal aortic operation: Is it preventable? David Rosenthal, MD, for the Southern Association for Vascular Surgery, Atlanta, Ga Purpose: Spinal cord ischemia after operation on the abdominal aorta is a rare event that is attributed to variations in the spinal cord blood supply. The purpose of this study was to evaluate the possible causes of this devastating event. Methods: A survey of patients among the members of the Southern Association for Vascular Surgery was performed, and 18 patients were identified with spinal cord ischemia manifested by paraplegia or paraparesis after abdominal aortic operation. Results: Preoperative computed tomographic, magnetic resonance, and aortographic results did not visualize the greater radicular artery (Adamkiewicz’s artery) in any patient. Eleven patients underwent resection of infrarenal abdominal aortic aneurysms (AAAs): seven of these patients had tube grafts, three had aortobifemoral grafts, and one had an aortobiiliac graft. Five other patients underwent placement of aortobifemoral grafts, and one patient underwent aortobiiliac graft placement for occlusive disease. One patient underwent suprarenal AAA resection with an interposition graft to a previous aortobiiliac graft. The mean operative time was 3 hours and 39 minutes (range, 2 hours and 45 minutes to 6 hours and 30 minutes), with a mean aortic cross-clamp time of 48 minutes (range, 24 to 97 minutes). Sixteen aortic cross-clamps were placed infrarenally and two suprarenally (one in a case of ruptured AAA, the other a suprarenal AAA). Seventeen proximal anastomoses were end to end. The average minimum systolic blood pressure during the aortic cross-clamping was 96 mm Hg (range, 80 to 130 mm Hg). All the patients had internal iliac artery flow preserved with either prograde perfusion (10 patients) or retrograde perfusion (eight patients), and one patient underwent unilateral internal iliac artery ligation because of aneurysmal disease. One aortobifemoralgraft limb necessitated thrombectomy, but no cases of massive peripheral embolization occurred. When paraplegia was suspected after operation (6 to 20 hours after surgery), five patients underwent lumbar drainage. No clinical improvement was noted. Conclusion: Interference with pelvic blood supply from prolonged aortic cross clamping, intraoperative hypotension, aortic embolization, and interruption of internal iliac artery circulation have all been suggested as possible causes of spinal cord ischemia. In this survey, none of these factors proved to be significant as the sole cause of spinal cord ischemia. In the performance of an aortic operation with an end-to-end proximal anastomosis in the presence of severe external or internal iliac artery disease, there may be an increased incidence of spinal cord ischemia despite appropriate surgical techniques to ensure internal iliac perfusion. Spinal cord ischemia after abdominal aortic operations appears to be a tragically unpredictable, random, and unpreventable event. (J Vasc Surg 1999;30:391-9.)
From the Department of Vascular Surgery, Georgia Baptist Medical Center, Medical College of Georgia. Presented at the Twenty-third Annual Meeting of The Southern Association for Vascular Surgery, Naples, Fla, Jan 27–30, 1999. Reprint requests: David Rosenthal, MD, 315 Boulevard NE, Ste 412, Atlanta, GA 30312. Copyright © 1999 by the Society for Vascular Surgery and International Society for Cardiovascular Surgery, North American Chapter. 0741-5214/99/$8.00 + 0 24/6/100787
Spinal cord ischemia after operation on the abdominal aorta is a rare but devastating event that continues to haunt the vascular surgeon. Szilagyi et al,1 in a landmark 1978 publication, noted only a 0.25% incidence of paraplegia or pariparesis after 3614 abdominal aortic operations. Classically, the symptoms of spinal cord ischemia of the lower extremities include sensory and motor deficits associated with bladder/rectal incontinence with conserva391
392 Rosenthal
tion of vibratory and proprioceptive sensation, thus reflecting damage to the anterior horn of the spinal cord. The causative factors associated with spinal cord ischemia include interruption of the greater radicular artery (Adamkiewicz’s artery), prolonged aortic occlusion, intraoperative hypotension, atheromatous embolization, and interruption of the internal iliac artery circulation.2-8 The purpose of this report, the largest series to date, was to evaluate the possible causes of spinal cord ischemia after abdominal aortic operations and to determine whether it is preventable. MATERIALS AND METHODS A survey was performed among members of the Southern Association for Vascular Surgery, and 16 surgeons reported 18 patients who had spinal cord ischemia develop manifested by paraparesis or paraplegia after abdominal aortic operations. Of these 18 patients, 15 underwent preoperative computed tomographic (CT) scanning of the abdomen, 17 underwent aortography, and one underwent magnetic resonance (MR) scanning. Eleven patients underwent resection of infrarenal abdominal aortic aneurysms (AAAs). Seven of these underwent tube graft placement, three underwent aortobifemoral grafting, and one underwent placement of an aortobiiliac graft. One patient underwent suprarenal AAA resection with an interposition graft to a previous aortobiiliac graft. Six other patients underwent operation for aortoiliac occlusive disease: five had aortobifemoral graft placement and one patient had aortobiiliac graft placement. Hospital charts, operative notes, anesthesia records, radiologic reports, and follow-up data from all patients were retrospectively reviewed in an attempt to identify the factors that might have contributed to the occurrence of spinal cord ischemia. These possible factors included: type of operation, aortic crossclamp location and duration, presence of hypotension, use of heparin therapy, operative duration, maintenance of internal or external iliac artery perfusion, and intraoperative embolization. The postoperative neurologic examinations, electromyography, CT scanning, or MR imaging studies results, and the clinical course were reviewed to determine the nature and extent of the neurologic injury. RESULTS Fourteen of the 18 patients had paraplegia develop associated with bowel and bladder dysfunction, and four patients had paraparesis manifested by unilateral limb motor and sensory deficits. All the symptoms were noted within 20 hours after operation.
JOURNAL OF VASCULAR SURGERY September 1999
The intraoperative factors that may have contributed to spinal cord ischemia were evaluated (Table I). The aortic cross clamp was placed infrarenally in 16 of 18 patients, with a mean cross-clamp time of 48 minutes (range: 24 to 97 minutes). Seventeen proximal aortic anastomoses were performed end to end, and one aortobifemoral graft had an end-to-side aortic anastomosis. The mean intraoperative blood loss was 1250 mL (range, 300 to 9000 mL), and a cell saver was used during all the operations. Hypotension, defined as a systolic blood pressure less than 100 mm Hg, occurred in eight patients. The average minimum systolic blood pressure during aortic cross clamping in these eight patients was 96 mm Hg (range, 80 to 130 mm Hg). Intravenous heparin therapy was administered during 17 operations: one patient who had a ruptured AAA did not undergo heparin therapy. The mean duration of operation was 3 hours and 39 minutes (range, 2 hours and 45 minutes to 6 hours and 30 minutes). In 11 patients, preoperative arteriographic results identified patent external and internal iliac arteries, and six patients had either unilateral or bilateral occlusive disease (>50% reduction in diameter) of the internal iliac vessels. On the basis of the preoperative aortographic findings and the type of operation, internal iliac perfusion was maintained with either prograde flow (10 patients) or retrograde flow (eight patients). Two patients who had undergone aortobifemoral graft placement had inferior mesenteric artery reimplantation to protect the pelvic circulation: one of these patients had severe bilateral internal iliac artery occlusive disease and the other had an internal iliac artery aneurysm oversewn. No cases of massive embolization occurred, although one aortobifemoral graft limb necessitated thrombectomy. When paraplegia was suspected after operation (16 to 20 hours after surgery), five patients underwent urgent lumbar drainage. However, no clinical improvement was noted. Eleven patients underwent CT scanning, and nine underwent nuclear MR scanning of the spinal cord. One MR scan showed a zone of high-signal intensity consistent with edema after cord infarction. Eight patients died within 30 days of operation. Death was attributed to cardiac causes in five patients and to visceral or somatic ischemia that led to multi-system organ failure in three patients. Ten patients survived: four patients remained paraplegic, three patients with postoperative paraparesis were able to ambulate with assistance (ie, walker, leg brace) at discharge, and three patients were lost to follow-up.
JOURNAL OF VASCULAR SURGERY Volume 30, Number 3
DISCUSSION Ischemic injury to the spinal cord after abdominal aortic reconstruction is fortunately rare. Since the initial clinical report by McCune9 in 1956, we are aware of only 111 reported English-language cases.10-36 Of these 111 patients, 70 (63%) underwent operations for AAAs (39 ruptured, 31 elective) and 41 (37%) underwent operations for aortoiliac occlusive disease. The early, elegant work of Adams and Geertruyden9 indicated that spinal cord ischemia most frequently occurred after operation for ruptured AAA because these patients had significant periods of hypotension, often underwent suprarenal or supraceliac cross clamping, and usually did not undergo anticoagulation therapy. Of interest, however, was the fact that nearly two thirds (64%; 72 of 111) of the reported patients who had spinal cord ischemia develop had undergone elective abdominal aortic operations. This finding is similar in our series, in which 17 of 18 patients underwent elective operations, and indicates that the risks associated with emergent operations (ie, hypotension, suprarenal or celiac cross clamping) may not be such causative factors of spinal cord ischemia as previously believed. The location and duration of the aortic cross clamp have been suggested as possible causes of spinal cord ischemia. In this review, 16 aortic cross clamps were placed infrarenally. All the patients, except one (with a ruptured AAA), underwent anticoagulation therapy, and the mean aortic crossclamp time was 48 minutes. Forty-five minutes of cross-clamp time has been shown to be a safe limit, although no distinct limit has been defined.37 The location and duration of the aortic cross clamp in our series could not, therefore, be indicted as probable causes of spinal cord ischemia. Hypotension plays an important role in the cause of spinal cord ischemia. Several postmortem study results have shown that ischemia may occur without occlusion of the anterior spinal artery.38 If the distal spinal cord blood supply is compromised, severe hypotension alone may surpass the safe limits of cord perfusion. The average minimum blood pressure in our study among eight patients who were hypotensive was 96 mm Hg (range, 80 to 100 mm Hg). Lynch et al,39 however, found an increased risk of central nervous system ischemia when the mean arterial pressure fell below 55 mm Hg, a number that correlated well with the loss of brain and spinal cord autoregulation observed experimentally at 50 mm Hg. Certainly, hypotension should be avoided during the perioperative period, especially in patients with severe arterial occlusive disease, but hypoten-
Rosenthal
393
Fig 1. Intrinsic and extrinsic blood supply to spinal cord. Unshaded area is perfused by anterior spinal artery, shaded area by posterior spinal arteries. (Plecha EJ, Seabrook GR, Freischlag JA, et al. Neurologic complications of reoperative and emergent abdominal aortic reconstruction. Ann Vasc Surg 1995;9:95-101. Reprinted with permission).
sion alone in our patients was not believed to be a cause of spinal cord ischemia. The etiology of spinal cord ischemia after abdominal aortic operation is likely multifactorial, but the fundamental cause is always an alteration in blood supply to the spinal cord. The blood supply to the cord consists of an intrinsic and extrinsic network of arteries.4,7,36 The intrinsic blood supply is composed of three longitudinal systems: one anterior and two posterior spinal arteries (Fig 1). Intramedullary anastomoses between the anterior and posterior spinal systems are nonexistent except along the circumference, where the arterial vasocoronae coalesce. The anterior spinal artery extends the full 40-cm to 45cm length of the spinal cord, but it can be so narrow at the distal thoracic level that it cannot compensate for an interrupted radicular artery, thus increasing the risk of infarction after loss of a single radicular artery.37 The anterior spinal artery is, therefore, the major independent source of spinal cord perfusion, and inadvertent interruption of the largest of its segmental radicular arteries (the greater radicular artery or the artery of Adamkiewicz) has been implicated as the principal cause of spinal cord ischemia.4
394 Rosenthal
Fig 2. Anatomy of blood supply to distal cord and lumbosacral plexus. (Gloviczki P, Cross SA, Stanson AW, et al. Ischemic injury to the spinal cord or lumbosacral plexus after aorto-iliac reconstruction. Am J Surg 1991;162:1316. Reprinted with permission).
The extrinsic blood supply to the spinal cord is made up of segmental medullary feeder arteries that supply the anterior median spinal artery and the spinal arteries, which are present at each vertebral level (Fig 2). The anterior spinal artery is fed by these segmental medullary arteries, which originate from the subclavian, intercostal, and lumbar arteries. Perfusion of the distal spinal cord is derived from the lumbar, iliolumbar, and lateral sacral arteries that anastomose with the intrinsic spinal arteries at the level of the conus medullaris. The greater radicular artery (Adamkiewicz’s artery) is located between T9 and T12 in 75% of the cases, but, if it originates abnormally high or is chronically compromised,
JOURNAL OF VASCULAR SURGERY September 1999
these lower lumbar arteries that anastomose with the intrinsic network may contribute significantly to the blood supply of the terminal spinal cord.7 In general, when the greater radicular artery is open and of normal size, the pelvic blood supply is of minor importance. When the greater radicular artery is compromised, however, the pelvic blood supply may be critically important. Indeed, Picone et al4 identified several factors that may lead to spinal cord ischemia as the result of interference of pelvic blood flow and therefore should be avoided: (1) internal iliac devascularization with resultant severe pelvic oligemia, (2) significant hypotension, and (3) massive embolization to the pelvis. These authors suggested that intraoperative technical factors that may lead to these problems be avoided by revascularization of the internal iliac artery with a separate graft, by the avoidance of hypotension, and by antegrade flushing and irrigation of the graft. The authors also recommended that if an internal iliac artery aneurysm must be oversewn, it should be done so from within the aneurysm to preserve the collateral branch flow to the lateral sacral and ilial lumbar arteries. It was of interest to note that among the 10 patients in our series who underwent aortobifemoral (n = 8) or aortobiiliac (n = 2) grafting, only one proximal aortic anastomosis was end to side, which theoretically might preserve pelvic flow. Of these 10 patients, however, only four had internal iliac occlusive disease that was identified before surgery (Table I). In the performance of an aortic operation with an endto-end proximal anastomosis in the presence of significant iliac artery occlusive disease, the better part of surgical wisdom is to ensure internal iliac artery perfusion. If iliac perfusion is uncertain or seems inadequate, a separate internal iliac artery graft likely is appropriate. On the basis of the anatomy of the blood supply to the distal spinal cord, nerve roots, and lumbosacral plexus, correlated with the observed neurologic deficit, Gloviczki et al7 identified six types of ischemic neurologic injury. Type I cord injuries are associated with global ischemia to the distal cord and conus: these injuries are characterized by flaccid paraplegia and poor long-term recovery. A type II injury, or anterior spinal artery syndrome, is characterized by flaccid paraplegia with loss of pain and temperature sensation, with preservation of proprioception and vibratory sensation as the result of the posterior spinal artery circulation. A type II injury is frequently seen after thoracic or thoracoabdominal reconstruction, and the prognosis is guarded, with minimal long-term
JOURNAL OF VASCULAR SURGERY Volume 30, Number 3
Rosenthal
395
Table I. Risk factors for spinal cord ischemia after abdominal aortic operations Operation AAA
Location/duration Hypotension (systolic, Heparin of aortic cross clamp <100 mm Hg) (IU) Infrarenal/30 minutes
Operation time
Internal iliac artery status
None
5000
3 hours, 10 minutes
Patent bilaterally
AAA/ABF Infrarenal/95 minutes and AR Suprarenal AAA SupraSMA/59 minutes bilateral renal reimplant AAA/ABI Infrarenal/48 minutes
90 mm Hg × 40 min None
7500
6 hours, 30 minutes
5000
3 hours, 10 minutes
Bilateral 50% stenosis 50% unilateral stenosis
None
5000
3 hours, 39 minutes
Patent bilaterally
AAA/ABF/ IMA graft
None
5000
3 hours
85 mm Hg × 35 minutes
9000
6 hours, 24 minutes
Patent bilaterally (internal iliac aneurysm oversewn) Unilateral occlusion
5000
3 hours
Patent bilaterally
5000
3 hours, 30 minutes
Patent bilaterally
No
3 hours
Unknown
4000
2 hours, 45 minutes
Patent bilaterally
Infrarenal/60 minutes
ABF with IMA, Infrarenal/57 minutes SMA thrombectomy, bowel resection AAA Infrarenal/63 minutes AAA
Infrarenal/30 minutes
Ruptured AAA
Infrarenal/45 minutes
ABF
Infrarenal/25 minutes
90 mm Hg × 6 minutes 80 mm Hg × 15 minutes 90 mm Hg × 6 minutes None
ABF/IMA reimplant
Infrarenal/31 minutes
None
5000
3 hours, 30 minutes
AAA/ABF
Infrarenal/97 minutes
5000
5 hours, 40 minutes
ABF/FP
Infrarenal/38 minutes
80 mm Hg × 14 minutes None
5000
4 hours, 20 minutes
Severe bilateral iliac occlusive disease Occluded bilaterally Patent bilaterally
AAA
Infrarenal/48 minutes
None
6000
4 hours, 30 minutes
Patent bilaterally
AAA/ABF
Infrarenal/24 minutes
90 mm Hg × 8 minutes None None
6000
3 hours, 30 minutes
5000 5000
4 hours, 10 minutes Unknown
50% stenosis bilaterally Patent bilaterally Patent bilaterally
5000
3 hours, 15 minutes
Unknown
ABF Infrarenal/40 minutes ABF/bilateral Infrarenal/49 minutes RA grafts/FP AAA/ABI Suprarenal/unknown
<80 mm Hg × 12 minutes
Neurologic deficit Paraplegia with bowel/ bladder dysfunction Paraparesis Paraplegia with bowel/ bladder dysfunction Paraplegia with bowel/ bladder dysfunction Paraplegia with bowel/ bladder dysfunction Paraplegia with bowel/ bladder dysfunction Paraplegia with bowel/ bladder dysfunction Paraplegia with bowel/ bladder dysfunction Paraplegia intact sensation Paraplegia with bowel/ bladder dysfunction Paraparesis Paraparesis Paraplegia with bowel/ bladder dysfunction Paraplegia with bowel/ bladder dysfunction Paraparesis Paraparesis Paraplegia with bowel/ bladder dysfunction Paraplegia with bowel/ bladder dysfunction
AAA, Abominal aortic aneurysm repair; ABF, aortobifemoral graft placement; AR, aorto renal; ABI, aortobiiliac graft placement; IMA, inferior mesenteric artery repair; SMA, superior mesenteric artery repair; FP, femoropopliteal bypass grafting; RA, renal artery repair.
improvement. Type III and IV injuries involve the nerve roots and lumbosacral plexus. Most cord functions are therefore preserved but with neurologic asymmetric motor and sensory deficits with or without bladder and bowel incontinence. Because the injury is at the axonal level, patients with type III and IV injuries have favorable prognoses. A type V injury is caused by segmental spinal cord infarction and may be seen after embolization to the cord at a specific level or hypoperfusion in a watershed region, and type VI injuries occur as a result of interruption of the posterior spinal arteries blood supply to the posterior columns. These injuries are rare and are characterized
by loss of proprioception and vibratory sensation alone. The prognoses for type V and VI injuries are favorable. The neurologic deficits in our patients appear to have been type II deficits (anterior spinal artery syndrome) in 13 patients. Type III or IV injuries occurred in four patients, and type V occurred in one patient. The type III, IV, and V injuries occurred in patients who had undergone aortobifemoral bypass grafting with end-to-end proximal anastomoses and may have been the result of inadequate retrograde pelvic arterial flow. In an editorial in 1993, Szilagyi40 reported that the approximate incidence of paraplegia after AAA
396 Rosenthal
was 1 in 400 and after aortoiliac reconstruction for occlusive disease was 1 in 5000. Szilagyi stated that the greater radicular artery may be “harmlessly occluded much more frequently than it is occluded with serious consequences”40 and attributed this to ample pelvic collateral flow, which prevents damaging ischemia to the spinal cord. Spinal cord ischemia after abdominal aortic operations is likely multifactorial and may involve interference with pelvic blood supply, prolonged aortic cross clamping, intraoperative hypotension, and aortic embolization. As stated earlier, however, none of these factors could be indicted in our study as a sole cause of cord ischemia. It cannot, however, be refuted that the revascularization of one internal iliac artery, the avoidance of profound hypotension and its associated low-flow state to the spinal cord, the avoidance of prolonged supraceliac cross clamping, the intraoperative use of heparin therapy, and the gentle surgical techniques to avoid embolization are all sound operative principles that minimize the potential for spinal cord injury. If the surgeon must perform an end-to-end proximal aortic anastomosis in the presence of severe iliac occlusive disease, awareness of these principles will hopefully decrease the incidence of spinal cord ischemia. Nevertheless, spinal cord ischemia after abdominal aortic operations remains a tragically unpredictable, random, and unpreventable event. REFERENCES 1. Szilagyi DE, Hageman JH, Smith RF, Illiott JP. Spinal cord damage in surgery of the abdominal aorta. Surgery 1978; 83:38-56. 2. Zuber WF, Gaspar MR, Rothschild PD. The anterior spinal artery syndrome—a complication of abdominal surgery: report of five cases and review of the literature. Am Surg 1970;172:909-15. 3. Ferguson LR, Bergan JJ, Conn J Jr, Yao JST. Spinal ischemia following abdominal aortic surgery. Ann Surg 1975;181: 2677-82. 4. Picone AL, Green RM, Ricotta JR, May AG, DeWeese JA. Spinal cord ischemia following operations on the abdominal aorta. J Vasc Surg 1986;3:94-103. 5. Iliopoulos JI, Howanitz PE, Pierce GE, Kueshkerian SM, Thomas JH, Hermreck AS. The critical hypogastric circulation. Am J Surg 1987;154:671-5. 6. Dimakakos P, Arapoglou B, Katsenis K, et al. Ischemia of the spinal cord following elective operative procedures of the infrarenal aorta. J Cardiovasc Surg 1996;37:243-7. 7. Gloviczki P, Cross SA, Stanson AW, et al. Ischemic injury to the spinal cord or lumbosacral plexus after aorto-iliac reconstruction. Am J Surg 1991;162:131-6. 8. Senapati A, Browsf NL. Gluteal necrosis and paraplegia following postoperative bilateral internal iliac artery occlusion. J Cardiovasc Surg 1990;31:194-6. 9. McCune WS. Discussion. In: Adams HD, Geetruyden HH, editors. Neurologic complications of aortic surgery. Ann Surg 1956;144:574:610-6.
JOURNAL OF VASCULAR SURGERY September 1999
10. Bates T. Paraplegia following resection of abdominal aneurysm. Br J Surg 1971;58:913-6. 11. Coupland GAE, Reeve TS. Paraplegia: a complication of excision of abdominal aortic aneurysm. Surgery 1964; 878:19-23. 12. Di Chiro G, Doppman JL. Paraplegia after resection of aneurysm. N Engl J Med 1969;281:79-83. 13. Edmondson HT, Gindin RA. Paraplegia as a complication of abdominal aortic resection. Am Surg 1970;36:383-7. 14. Golden GT, Sears HF, Wellons HA Jr, Muller W Jr. Paraplegia complication resection of aneurysms of the infrarenal abdominal aorta. Surgery 1973;73:91-5. 15. Gump FE. Paraplegia after resection of aneurysm. N Engl J Med 1969;281:798-801. 16. Hara M, Lipin RJ. Spinal cord injury following resection of abdominal aortic aneurysm. Arch Surg 1960;80:419-22. 17. Hogan EL, Romanul FCA. Spinal cord infarction occurring during insertion of aortic graft. Neurology 1966;16:67-74. 18. Jellinger K. Spinal cord arteriosclerosis and progressive vascular myelopathy. J Neurol Neurosurg Psychiatry 1967; 30:195-8. 19. Lake PA. Paraplegia after resection of aneurysm. N Engl J Med 1969;281:798-801. 20. Mehrez IO, Nasbeth DC, Hogan EL, Deterling RA. Paraplegia following resection of abdominal aortic aneurysm. Ann Surg 1962;156:890-8. 21. Pasternak BM, Boyd DP, Ellis FH Jr. Spinal cord injury after procedures on the aorta. Surg Gynecol Obstet 1972;135: 29-34. 22. Reich MP. Paraplegia following resection of an abdominal aortic aneurysm: report of a case of atheromatous embolization to the anterior spinal artery. Vasc Surg 1968;2:230-4. 23. Sher MH, Healy EH. Paraplegia following infrarenal aneurysmorrhaphy. Vasc Surg 1971;5:171-4. 24. Skillman JJ, Zervas NT, Weintraub RM, Mayman C. Paraplegia after resection of aneurysms of the abdominal aorta. N Engl J Med 1969;281:422-4. 25. Lynch C, Weingarden SI. Paraplegia following aortic surgery. Paraplegia 1982;20:196-200. 26. Mills RP, Robbs JV, Duncan HJ. Paraplegia following abdominal aortic surgery. S Afr Med J 1983;63:930-2. 27. Friedman SG, Moccio CG. Spinal cord ischemia following elective aortoiliac reconstruction. Ann Vasc Surg 1988;2:295-7. 28. Michaels L. Case report: spinal cord ischemia associated with repair of a ruptured abdominal aortic aneurysm. Stroke 1972;3:238-42. 29. Connolly JE. Prevention of paraplegia secondary to operation on the aorta. J Cardiovasc Surg 1986;27:410-7. 30. Nielson BP. Ischemic injury to the spinal cord as a complication to abdominal aortic surgery. Acta Chir Scand 1985; 151:433-5. 31. Perry MO. Spinal cord injury following aortorenal bypass. Cardiovascular Surg 1979;20:261-4. 32. Salam AA, Sholkamy SM, Chaikof EL. Spinal cord ischemia after abdominal aortic procedures: is previous colectomy a risk factor? J Vasc Surg 1993;17:1108-10. 33. Graci RR, Matton KL. Anterior spinal artery syndrome following abdominal aortic aneurysmectomy. Arch Surg 1977; 112:813-5. 34. Gunderren J, Liungqvist V, Bergentz SE. Paraplegia after resection of a ruptured aortic aneurysm. Vasa 1980;9:304-5. 35. Jacoss PB, Dicke HW. Ischemic damage to the spinal cord following surgery of the abdominal aorta. Neth J Surg 1984; 36:1-5.
JOURNAL OF VASCULAR SURGERY Volume 30, Number 3
36. Plecha EJ, Seabrook GR, Freischlag JA. Neurologic complications of reoperative and emergent abdominal aortic reconstruction. Ann Vasc Surg 1995;9:95-101. 37. Defraigne JO, Otto B, Sakalihasan N, Limet R. Spinal ischemia after surgery for abdominal infrarenal aortic aneurysm: diagnosis with nuclear magnetic resonance. Acta Chir Belg 1997; 97:250-6. 38. Singh V, Silver JR, Weply NC. Hypotensive infarction of the spinal cord. Paraplegia 1994;32:314-22.
DISCUSSION Dr Elliot L. Chaikof (Atlanta, Ga). In a sobering presentation, Dr Rosenthal and colleagues have highlighted that spinal cord ischemia after abdominal aortic reconstruction is somewhat akin to playing Russian roulette: the gun is always loaded and ready to fire, and little or no forewarning is provided. In medicine, we take comfort from the time-honored practice of analyzing our outcomes, particularly those that are less than ideal. Implicit to this approach is the assumption that in experience there are always lessons to be learned and more appropriate paths that can be pursued the next time around. Unfortunately, the patterns of life’s experiences are occasionally not very neat, and, in this regard, spinal cord ischemia after abdominal aortic surgery appears to be excluded from this paradigm. Second guessing is common after these rare but catastrophic events. However, most recommendations, including the avoidance of systemic hypotension or prolonged suprarenal clamping, are so general and inherent to the standard practice of aortic surgery as to have little practical value. I have always held to the belief that the preservation of pelvic circulation is likely the most important factor in avoiding neurologic complications after abdominal aortic reconstruction. Although no one would deny the potential significance of pelvic ischemia, this could not be identified as a dominant factor in this collected series. In fact, in at least seven of these 18 patients, the operation was routine in every possible way, without systemic hypotension or significant iliac artery disease and with infrarenal clamping of short duration. An additional six patients had relatively mild hypotensive episodes during surgery but, in all cases, for periods of less than 15 minutes. This report may find its greatest impact in the medicolegal arena, and I believe it is important that events as rare and unpredictable as these not dictate established practice patterns. Nonetheless, the underlying message in this review is that bad things sometimes happen to good people during surgery, despite everyone’s best efforts. This knowledge held in the back recesses of our consciousness should give us some pause and require that recommendations for surgical intervention always be provided in a most considered manner. As a final comment, I believe it is worth noting that this report is an important outgrowth of Dr Clagett’s rec-
Rosenthal
397
39. Lynch DR, Dawson TM, Raps EC, Galetta SL. Risk factors for the neurologic complications associated with aortic aneurysms. Arch Neurol 1992;49:284-8. 40. Szilagyi DE. A second look at the etiology of spinal cord damage in surgery of the abdominal aorta. J Vasc Surg 1993; 17:1111-3.
Submitted Feb 1, 1999; accepted Jun 17, 1999.
ommendations to our Association during his 1998 presidential address. The collective experience of a society whose members practice in both community and academic settings can indeed be harnessed to address some very significant clinical problems. In sharing their experience, Dr Rosenthal and the participating members of the Southern Assocation for Vascular Surgery have done us all an important service. I have three questions for Dr Rosenthal. First, eight of the 18 patients died within 30 days of operation from somatic or visceral ischemia. Can this be taken as an indication that, in this series, atheroembolization is the likely cause of most cases of spinal ischemia? Second, anesthesiologists, on occasion, remind us that an epidural hematoma can complicate catheter placement in the patient who undergoes aortic surgery. Were you able to exclude epidural hematoma as a causative factor of spinal ischemia in this series? Finally, how much information should we provide to our patients on a routine basis regarding the risks of rare and random catastrophic events? I would like to thank the Association for the privilege of discussing this report. Dr David Rosenthal. The first question was in reference to atheromatous emboli, a common cause of spinal cord ischemia. In actuality, of the eight deaths, five were caused by cardiovascular sequelae and three, as Dr Chaikof mentioned, were related to somatic or visceral ischemia. These, however, were not the massive, catastrophic embolization problems that other have reported in which profound ischemia to the pelvis and lower extremities occurs. I believe three of the deaths may have been embolic related, but there is no way of proving this. The second question was in reference to epidural hematoma, which is an appropriate question. As I believe, all the patients for elective surgery had epidural catheters. All of the patients in the study underwent radiologic diagnostic studies when spinal cord ischemia was suspected. Eleven had computed tomographic scans and nine had magentic resonance scans, but there was no evidence of epidural hematoma in any of these patients. Interestingly, only one patient, on a magentic resonance scan, had edema of the cord that indicated ischemia. Certainly, if a patient awakens from an aortic operation and has lumbar pain and a persistent motor blockade
398 Rosenthal
when there is recovery from sensory anesthesia, we should be alerted that an epidural hematoma may be present. Because early intervention is the key to success in the management of an epidural hematoma, urgent neurologic assessment with computed tomographic or magentic resonance scanning is mandatory, followed by urgent operation to evacuate the hematoma. The last question related to the concern of the spinal cord ischemia being part of our informed consent, and, as usual, Dr Chaikof has hit to the heart of the matter. I personally have never mentioned this as a possible risk to a patient who was about to undergo aortic operation, but at my hospital in Georgia the standard hospital consent form lists paraplegia as an operative risk. As Dr Chaikof also mentioned, this raises a possible medicolegal problem for all of us, which possibly should be addressed by our National Vascular Societies Ethics Council to offer all of us structured guidelines. Dr G. Melville Williams (Baltimore, Md). Dr Rosenthal, I really appreciated your in-depth analysis, and I just stand to request more information on the diagnostic evaluation of these patients after the event. It has been interesting to me, in doing thoracoabdominal procedures, that we have evolved from the dense paraplegia that happened all too frequently, so that the patient would awake paralyzed, to a more subtle form in which the paraplegia occurs days later and may even be a monoplegia or some other weird syndrome. I have tried to figure this out with magnetic resonance imaging, magnetic resonance angiography and all this business of the spinal cord blood supply, and what I plead for is a thorough and intense neurologic evaluation by every method we now know to try to characterize exactly what goes wrong in these patients. I am persuaded that a lot happens with nerve root ischemia or injury as opposed to the spinal cord itself. I do not know how else you can explain the paraparalysis limited to one extremity. And even when we do magnetic resonance imaging studies, there is controversy in the interpretation. There can be a little streak that someone sees and says, “Ah ha, that is indicative of an ischemic insult.” So, I think one of the things as a society we ought to demand, and I think it is going to help us in every medicolegal adventure that we get into with these kind of patients, is a complete neurologic evaluation with every known method to try to investigate just what is ischemic and what went wrong. Dr Rosenthal. Dr Williams, I could not agree more. In this retrospective study, you saw how variable the studies were: some patients had electromyographys, some nerve conduction velocity studies, some the computed tomographic or magnetic resonance scans. From what I learned in my readings, it seems that if you are going to obtain only one study, a magnetic resonance scan is the most appropriate, but to gain more knowledge on the pathophysiology of what has occurred, a complete battery of studies should be analyzed. Dr Roger T. Gregory (Norfolk, Va). Dr Rosenthal, I
JOURNAL OF VASCULAR SURGERY September 1999
thoroughly enjoyed that presentation. Fortunately, we were not part of this group. Paraplegia has not been seen by our group in Norfolk in more than 30 years. We have, however, seen three patients with lumbosacral plexopathy after ruptured aneurysms, which we reported a few years ago. These were type III and type IV aneurysms or a type IV plexopathy. All of these were in ruptured aneurysms. Aggressive therapy and evaluation, as Dr Williams has just suggested, however, resulted in all three patients in an ambulatory status a few months later. Therefore, I think it is incumbent on all of us, and I would certainly stress your point, to evaluate these patients thoroughly, because those with plexopathy have actually a very reasonable prognosis. Dr Rosenthal. It is interesting, Dr Gregory, that we did a very extensive literature review, which we mention in the manuscript, and as far as I could tell, in the English language, there have been 107 case reports of paraplegia or paraparesis in the literature dating back to 1956, when the first patient was identified. All of the early reports, in the first decade, were on ruptured abdominal aortic aneurysms. But the thing that struck me when I was doing the literature search and review was that in the last several decades, more than two thirds of all these reports are in patients for elective surgery and paraphrase what Dr Chaikof said. I think that when you get that paraparesis, it is because of one of two factors: it is either microembolization to the conus region, or sometimes when you get these patients who have a massive hematoma tracking down the gutter towards the pelvis, you can get literally ischemia of the spinal roots as they are coming out of the pelvis itself. But those are the patients that get reclamation of function. Dr Hazim J. Safi (Houston, Tex). I enjoyed your presentation, and I think this is going to be a landmark paper. We have seen two kinds of postoperative paraplegia or neurological deficit—immediate and delayed—and we found that delayed paraplegia responds to cerebrospinal fluid drainage. I would like to know if you looked into the phenomenon of immediate delayed paraplegia. Dr Rosenthal. All of the patients were suspected of having spinal cord ischemia within 20 hours of operation. The delayed cord ischemia symptoms that you and Dr Williams see with the thoracoabdominal operations as the result of a possible hyperemia response may well respond to cerebrospinal fluid drainage, as others have recommended. Six of the 18 patients, I believe, had cerebrospinal fluid drainage as soon as the neurologic event was suspected, but tragically their clinical outcome did not improve. Thank you. Dr Gary M. Gross (Huntsville, Ala). This complication is a rare event. A survey of the Society is a way, perhaps the only way, of finding enough of these events to be worth studying. I am glad that the Lifeline Foundation is considering supporting some of these survey studies. My question is about an end-to-side case. Presumably, the existing pelvic circulation and the collaterals were preserved. What was the suspected cause in that case?
JOURNAL OF VASCULAR SURGERY Volume 30, Number 3
Dr Rosenthal. I do not know. That patient had undergone an elective aortobifemoral graft placement, with one internal iliac artery patent on a preoperative arteriography, yet the patient awoke from anesthesia paraplegic. You would have to think this was caused by either a low anterior spinal artery takeoff or pelvic oligemia. It is interesting that when we were gathering data for the study, I literally took 10 sheets of paper and hung them on a basement wall to make a flow chart in the hope we could finally find an answer to this devastating problem. We have 18 patients, this is the largest series reported, and despite all our efforts, many of the same questions remain unanswered. Dr S. Edwin Duncan (Tyler, Tex). Dr Rosenthal, I enjoyed the paper. The question I had was that some of the old literature talked about the significance of large, maybe not paired, lumbar vessels that were a possible etiology. Did you notice such a lumbar artery, and would you consider reimplantation of some significant honker lumbar? And I just wondered, in any of your review of these things, was that mentioned as something that was a consideration? Is that something we need to be aware of? Is that another risk factor to pay attention to? Thank you. Dr Rosenthal. In the literature that I reviewed there is anecdotal information on the large lumbar artery that does not bleed back, similar to what we see during thoracoabdominal aneurysm repair. Obviously, most of the lumbars that we are dealing with in the infrarenal aorta we routinely oversew because of vigorous back bleeding. In our data retrieval form, and in the operative notes I reviewed, I specifically looked for mention of lumbar vessels but could not find anything consistent. Certainly, after having been part of this study, if I come across an inordinately large patent lumbar artery during an aneurysm
Rosenthal
399
operation that is not back bleeding, I would strongly consider reimplantation. Dr Jose Alvarez (Miami, Fla). Did you notice how many of these patients had epidural anesthesia, because most of them underwent elective aneurysmectomies, and was there any role for this intervention in the course of their paraplegia now that they underwent heparin therapy and we know about that problem? Dr Rosenthal. This is similar to Dr Chaikof’s question and epidural hematoma. As I mentioned, each patient underwent a radiologic study, either computed tomographic or magnetic resonance scanning, but no epidural hematoma was identified.
APPENDIX Contributing members Colby P. Atkins, MD, Lexington, Ky Robert A. Brigham, MD, West Reading, Pa T.J. Bunt, MD, Bullhead City, Ariz Matthew P. Campbell, MD, Houston, Tex G. Patrick Clagett, MD, Dallas, Tex Steven J. Eskind, MD, Nashville, Tenn Robert W. Feldtman, MD, Houston, Tex D. Nelson Gwaltney, MD, Bristol, Tenn William D. Jordan, Jr, MD, Birmingham, Ala Herman W. Kaebnick, MD, Louisville, Ky Irving L. Kron, MD, Charlottesville, Va Samuel R. Money, MD, New Orleans, La Andrew J. Olinde, MD, Baton Rouge, La J. Mark Rheudasil, MD, Atlanta, Ga Edward E. Rigdon, MD, Jackson, Miss Thomas H. Schwarcz, MD, Lexington, Ky