Spinal cord injury has been a devastating complication

Selective Perfusion of Segmental Arteries in Patients Undergoing Thoracoabdominal Aortic Surgery Toshihiko Ueda, MD, Hideyuki Shimizu, MD, Atsuo Mori, MD, Ichiro Kashima, MD, Katsumi Moro, MD, and Shiaki Kawada, MD Department of Cardiovascular Surgery, School of Medicine, Keio University, Tokyo, Japan Background. Reattachment of segmental arteries is one method used to prevent paraplegia associated with thoracoabdominal aortic repair. Nevertheless, even when important segmental arteries are reattached, ischemia causing spinal injury may occur during anastomosis. Methods. In 27 patients undergoing thoracoabdominal aortic repair, we attempted to perfuse the segmental arteries to be reattached with catheters connected to the distal bypass circuit. To identify perioperative risk factors for spinal ischemia, we examined changes in spinal somatosensory evoked potentials. Results. A median value of four segmental arteries were perfused in 20 (74%) of the 27 patients. Changes in somatosensory evoked potential indicative of spinal ischemia were observed in 13 patients (48%). The only risk factor associated with changes in evoked potentials revealed by a multivariate analysis was prolonged aortic cross-clamp time (> 120 minutes). Of the 2 patients who suffered paraplegia, one had the longest clamp time and the other showed spinal cord necrosis due to embolic shower. Conclusions. Despite selective perfusion of segmental arteries, spinal ischemia associated with aortic crossclamping may occur when clamping is prolonged over 120 minutes. Most of the changes appear to be reversible, however. (Ann Thorac Surg 2000;70:38 43) 2000 by The Society of Thoracic Surgeons Spinal cord injury has been a devastating complication associated with surgical repair of the thoracoabdominal aorta. Although the mechanisms underlying paraplegia in this situation are complicated, perioperative spinal ischemia plays a fundamental role [1 3]. Distal aortic perfusion [4 7] and reimplantation of the important segmental arteries [4 6,8,9] are widely adopted surgical procedures undertaken to protect the spinal cord. Although Griepp and associates have concluded that the presence of so-called critical segmental artery was unlikely [7], the high incidence of spinal injury resulting from extended repairs that they reported in their own study suggests the necessity of the segmental artery reimplantation procedure. Whether or not the important radicular artery has been preserved cannot be determined without a special attempt to identify the vessel [10 12]. It is preferable to attach a number of pairs of segmental arteries located between vertebrae T8 and L1, because the arteria radicularis magna, also known as the artery of Adamkiewicz, usually arises from this segment [13]. Even when the important segmental artery is successfully preserved, transient ischemia may occur during reimplantation and thus cause spinal cord injury [14,15]. Accepted for publication March 27, 2000. Address reprint requests to Dr Ueda, Department of Cardiovascular Surgery, School of Medicine Keio University, 35 Shinanomachi, Shinjuku-ku 160-8582 Tokyo, Japan; e-mail: ueda@med.keio.ac.jp. Maintaining blood pressure within the aortic segment isolated by double cross-clamping has been shown experimentally to improve spinal cord blood flow [16,17]. Thus when reattachment of the segmental arteries is necessary, perfusion of these vessels during anastomosis will reduce spinal cord ischemia. Somatosensory evoked potential (SEP) monitoring has been used to detect spinal cord ischemia during aortic operations [18]. Although Crawford and associates denied the reliability of SEP monitoring [19], recent reports have revealed that the procedure is highly accurate and can produce significant benefits [7,20 23]. To determine any problems associated with our spinal protection technique, we examined the perioperative factors that could cause SEP changes in patients undergoing thoracoabdominal aortic repair. Material and Methods Between September 1993 and August 1997, 27 consecutive patients underwent replacement of the thoracoabdominal aorta at our hospital. During reimplantation, we attempted to perfuse the segmental arteries in all patients using our newly developed catheters (Sumibe Medical Co, Tokyo, Japan) (Fig 1) [24]. Patient Characteristics Patients ranged in age from 28 to 72 years (median, 62 y) (Table 1). Seven (26%) were female. Eight (30%) had 2000 by The Society of Thoracic Surgeons 0003-4975/00/$20.00 Published by Elsevier Science Inc PII S0003 4975(00)01488 0

Ann Thorac Surg UEDA ET AL 2000;70;38 43 SELECTIVE PERFUSION OF SEGMENTAL ARTERIES 39 Fig 1. Catheters used in this study are made from polyurethane, have a total length of 30 cm, and are available in two sizes: 1-mm and 1.25-mm (outer diameter of head portion). Perfusion flow rates through 1-mm and 1.25-mm catheters measured by mock circulation with glycerin under 60 mm Hg perfusion pressure are approximately 10 and 50 ml/min, respectively. undergone prior aortic surgery. The aortic pathology in 16 of the 27 patients (59%) was dissection. The preoperative circulatory status for all patients was stable. Surgical Technique Arterial pressure was monitored at the right radial artery and at the dorsalis pedis artery opposite the femoral artery that was to be cannulated for the bypass. After induction of general anesthesia, patients were placed in the left lateral decubitus position with hips swiveled posteriorly. A double-lumen endotracheal tube was used in all patients. A standard left thoracoabdominal retroperitoneal incision was made, and the left thorax was entered through the sixth rib bed or the seventh intercostal space. The costal arch was excised, and the diaphragm was incised circumferentially to the aortic hiatus. The aortic segment to be replaced was entirely exposed. The femoral venoarterial bypass circuit used for the procedure included a centrifugal pump (HPM-15, Nikkiso, Tokyo, Japan) and a membrane oxygenator with heat exchanger (Sarns Turbo 440, Sarns, Ann Arbor, MI) (Fig 2). The circuit was branched for selective perfusion of the segmental arteries and major abdominal vessels. Activated clotting time was maintained at around 200 seconds through the administration of low-dose heparin. When massive bleeding was anticipated, a cardiotomy reservoir and suction was used with full-dose heparin. The perfusion flow rate, initially 40 ml kg 1 min 1, was adjusted to keep the proximal systolic pressure at more than 90 mm Hg and the mean distal perfusion pressure at more than 60 mm Hg. Patients were kept normothermic by a heat exchanger. A protein-impregnated Dacron polyester fabric graft (Hemashield, Meadox Medicals, Oakland, NJ) was used. Proximal aortic anastomosis was performed under segmental double cross-clamping when feasible. A number of pairs of segmental arteries located between the T8 and L1 vertebrae were selected for reimplantation. As many of the selected arteries as possible were perfused by the catheters connected to the bypass circuit. To prevent backbleeding from the vessels in cases where placement of indwelling catheters was not possible, balloon-tipped occluders were used. The remaining segmental arteries were oversewn. The celiac artery, superior mesenteric artery, and bilateral renal arteries were also selectively perfused by balloon-tipped 8F catheters (Sumibe Medical Co, Tokyo, Japan) connected to the circuit. The segmental arteries to be reattached were trimmed even with the aortic wall and then anastomosed to the side of the Dacron graft. In some cases, reimplantation of the segmental arteries was included in a beveled anastomosis of the distal aorta. Major abdominal branches were reattached by the button technique or with interposition of a small Dacron graft. The four principal procedures were carried out in the following order: proximal aortic anastomosis, reattachment of the segmental arteries, reattachment of the abdominal branches, and distal aortic anastomosis. Spinal Evoked Potential Monitoring of SEPs was performed in all patients throughout the operation with a Neuropack MEB 7120 (Nihon Koden Co, Tokyo, Japan). Stimuli were applied at an ankle overlying the posterior tibial nerve opposite to the femoral artery to be cannulated. Potentials were recorded at multiple loci on the scalp by means of skin needles. According to Laschinger and associates, a 10% or greater prolongation of the latent period, a 40% or greater decrease in the amplitude in comparison to the control waveform, or both was considered to be abnormal [18]. Abnormal SEPs associated with clamping of the iliac artery were excluded. Statistical Analysis Risk factors were screened for possible association with the changes in SEPs by a univariate analysis with 2 tests. In terms of continuous data, the patients were divided into two groups according to the cut-off value at which the p value was the lowest. A p value of less than 0.05 was considered significant. Any factor that had a p value of

40 UEDA ET AL Ann Thorac Surg SELECTIVE PERFUSION OF SEGMENTAL ARTERIES 2000;70;38 43 Table 1. Risk Factors for Intraoperative Changes in Somatosensory Evoked Potentials Variable Range (median) No. SEP Change p Value ( 2 ) p Value (MLR) Total patient group 27 13 (48%) Age (y) 28 72 (62) 0.34 55 13 7 55 14 6 Sex 0.32 Male 20 8 Female 7 5 Pathology 0.028 0.6402 Dissection 16 11 True aneurysm 11 2 Crawford type 0.061 0.3827 II 7 6 I/III/IV 20 7 Prior aortic operation 0.58 Yes 8 5 No 19 8 Proximal clamp site 0.53 Proximal to T8 16 9 Distal to T8 11 4 Distal clamp site 0.16 Proximal to RA 11 3 Distal to RA 16 10 Involved pairs of segmental arteries 3 14 (10) 0.13 7 18 11 7 9 2 Involved pairs of patent segmental arteries (no) 1 14 (7) 0.065 0.4388 6 17 6 6 10 2 Reattached pairs of segmental arteries 1 6 (4) 0.57 4 13 7 4 14 6 Heparin dosage 0.64 Full-dose 6 4 Low-dose 21 9 Perfused segmental arteries 0 8 (4) 0.10 0.6409 0 7 1 1 20 12 Proximal hypotension (mm Hg) a 0.035 0.1195 70 5 5 70 22 8 Distal hypotension (mm Hg) a 0.69 50 10 6 50 17 7 Aortic cross-clamp time (min) 69 271 (157) 0.0029 0.0045 120 16 12 120 11 1 Ischemic time for reattached segmental arteries (min) 46 215 (72) 0.026 0.2482 90 8 7 90 19 6 a Persisting 30 min. MLR multiple logical regression; RA renal artery; SEP somatosensory evoked potential. less than 0.10 was entered into multivariate logistic regression analysis. Statistical analysis was performed with the SAS statistical software program (SAS Institute, Cary, NC). Results The extents of the replaced aortic segments according to Crawford s classification were type I in 8 patients (30%),

Ann Thorac Surg UEDA ET AL 2000;70;38 43 SELECTIVE PERFUSION OF SEGMENTAL ARTERIES 41 Table 2. Summary of Changes in SEP a Pattern of Change No. (%) b Disappeared during proximal aortic anastomosis 1 (8) and never recovered (case 1 c ) Changed while SAs to be reattached were placed 12 (92) between double aortic clamps Disappeared Recovered following initiation of selective 1 (8) perfusion Recovered after unclamping of reattached SAs 2 (15) Never recovered (case 2 c,d ) 1 (8) Not disappeared Recovered after unclamping of reattached SAs 6 (46) Incompletely recovered after unclamping of 2 e (16) reattached SAs a Of the 27 study patients, 13 (48%) evinced changes in SEP during thoracoabdominal aortic surgery and 14 (52%) did not. b Percentage of 13 patients with changes. c Suffered paraplegia. d SEP began to change during reimplantation of abdominal branches while reattached segments were left clamped. e Postoperative spinal function could not be examined in 1 patient. SA segmental artery; SEP somatosensory evoked potential. Fig 2. Illustration of the partial cardiopulmonary bypass circuit. (IVC inferior vena cava; FA femoral artery.) type II in 7 (26%), type III in 9 (33%), and type IV in 3 (11%). A proximal aortic clamp was placed between the left common carotid artery and the left subclavian artery in 8 patients (30%), between the left subclavian artery and T8 vertebra in 9 (33%), and below T8 in 10 (37%). During distal aortic anastomosis, a distal clamp was placed between the celiac artery and the renal artery in 11 patients (41%), and below the renal artery in 16 (59%). A minimum of 3 to a maximum of 14 pairs (median,10 pairs) of segmental arteries were involved in the replaced aorta. In each case, at least 1 patent vessel was found in 1 to 14 artery pairs (median, 7 pairs). One to 6 pairs (median, 4 pairs) were reattached in each patient. In 7 patients (26%), the reattached segmental arteries were too small to be cannulated; in the remaining 20 patients, 1 to 8 vessels (median. 4 vessels) could be perfused selectively. Full-dose heparin was administered in 6 patients (22%). Proximal hypotension persisted for more than 30 minutes at less than 80 mm Hg in 8 patients (33%), at less than 70 mm Hg in 5 (19%), and at less than 60 mm Hg in 1 (4%). In contrast, 30-minute or longer episodes of distal perfusion pressure of less than 60 mm Hg occurred in 17 patients (63%) and similar episodes of less than 50 mm Hg in 10 (37%). The aortic cross-clamp time ranged from 69 to 271 minutes (median, 157 min). The ischemic time of the reattached segmental arteries, defined as the interval during which they were placed between two cross-clamps, ranged from 46 to 215 minutes (median, 72 minutes). Three of the 27 patients died during hospitalization (mortality rate, 11%). One patient (case 1) died within 30 days of the operation because of extended bowel infarction due to intraoperative multiple thrombosis. Another patient, who had suffered brain damage associated with severe hypotension caused by postoperative bleeding, died of multiple organ failure. The third patient died of panperitonitis due to pancreatic necrosis. Changes in SEP Abnormal SEPs were observed in 13 patients (48%) (Table 2). In one patient (case 1), the SEPs disappeared during proximal aortic anastomosis and were not restored at any point during the operation. In the remaining 12 patients with abnormal SEPs (92%), changes in the measurement were observed while the segmental arteries were placed between double aortic cross-clamps. In 1 of those patients, the waveform disappeared soon after applying the double cross-clamps but was later restored following initiation of selective perfusion of six segmental arteries. The SEPs that had faded out in 2 other patients were restored soon after unclamping of the reattached segmental arteries. In 1 patient (case 2), SEPs began to change and later disappeared during reimplantation of the abdominal branches. The potentials remained lost at the end of the operation. The SEPs changed but did not disappear in the remaining 8 patients. The waveforms also recovered after unclamping of the reattached segmental arteries, but the recovery was incomplete in 2 patients. Postoperative Spinal Function Spinal function was examined after the operation in all patients except for 1 who, as a result of brain damage, never awoke. The 2 patients (8%) whose SEPs remained lost at the end of the operation suffered paraplegia. Both patients (case 1 and 2) had dissection and type II aneu-

42 UEDA ET AL Ann Thorac Surg SELECTIVE PERFUSION OF SEGMENTAL ARTERIES 2000;70;38 43 rysm. No spinal dysfunction occurred in the remaining patients. In case 1, a patient with mural thrombi, SEPs disappeared soon after applying the cross-clamps at vertebrae T3 and T5 and never reappeared. Only six patent segmental arteries were observed in this patient. All were reattached, and four of them were selectively perfused. An autopsy revealed massive infarctions in multiple abdominal organs and spinal cord necrosis below the T9 vertebra. In case 2, four pairs of segmental arteries were reattached and five of these vessels were selectively perfused. Following completion of the anastomosis, the catheters were withdrawn, but the reattached segmental arteries were left clamped for technical reasons until all the abdominal branches had been reimplanted. The clamp time of the reattached segmental arteries was 215 minutes. Postoperative magnetic resonance imaging revealed a spinal cord necrosis below T9. Statistical Data An aortic cross-clamp time of longer than 120 minutes, ischemia lasting longer than 90 minutes in the reattached segmental arteries, dissection, and proximal hypotension below 70 mm Hg were the significant risk factors for changes in SEP, according to univariate analysis (Table 1). However, multivariate logistic regression analysis yielded only one significant factor, an aortic cross-clamp time of longer than 120 minutes. Comment The accuracy of SEP monitoring can be improved by direct stimulation of the spinal cord [21] or by application of multilevel stimulation [22,23]. According to the above studies, global brain hypoperfusion caused by extreme hypotension, brain ischemia due to left carotid artery cross-clamping, and posterior tibial nerve ischemia were the factors responsible for the false-positive changes in SEPs. Thus spinal cord ischemia can be detected accurately by ordinary SEP monitoring if those factors are ruled out. In the majority of patients in the present study, altered SEPs returned to normal after the reattached segmental arteries were unclamped. Postoperative paraplegia occurred in only those patients whose SEPs never reappeared. Thus we believe that most of the SEP changes observed in our patients were associated with spinal cord ischemia. Rarely did SEPs change in the patients with small segmental arteries in which indwelling catheters could not be placed. We concluded that the anterior spinal artery in those patients had good continuity, and thus reimplantation of the segmental arteries might be not necessary [7]. Although it is impossible to perfuse a small segmental artery even if the artery is important, communication among the segmental arteries and spinal branches is reportedly not rare [11,12]. It is possible that blood from the vessels being perfused reaches the spinal cord through such collateral circulation. The observed changes in SEPs indicate that selective perfusion of the segmental arteries was not sufficient to maintain adequate spinal perfusion [16,17]. In one patient, however, initiation of selective perfusion led to restoration of the absent SEPs, and most of the SEP changes in the other patients were also reversible. The paraplegia that occurred in case 2 might have been prevented if the reattached segmental arteries had been unclamped at completion of the anastomosis. The spinal cord ischemia in case 1, on the other hand, was most likely caused by emboli scattered by retrograde perfusion. As we previously reported, this disastrous complication could have been prevented if the distal aortic anastomosis had been performed prior to institution of the bypass [25]. One important problem was that we were not able to determine the rate of flow through perfusion catheters, because only one pressure pump was used for the circuit. To attain a stable flow, the segmental arteries should be perfused by a separate pump. Insertion of catheters was often time-consuming and thus prolonged the aortic cross-clamp time. Improvements should be made to the catheter itself to help shorten insertion times. Nevertheless, the incidence of postoperative paraplegia in our study was not different from that of other recent studies [4 9]. Aside from distal perfusion, no other methods for spinal protection, such as hypothermia [4,6,7,9] or cerebrospinal fluid drainage, were used [7,8]. Using selective perfusion of the segmental arteries, SEP changes indicating spinal ischemia continued to occur in patients in whom aortic cross-clamping was prolonged for more than 120 minutes. However, most of the changes were reversible. The attempt to place numerous catheters into small segmental arteries should not prolong crossclamping time considerably, as this is still the main risk factor for paraplegia. In cases where the procedure is time-consuming, additional protective measures such as hypothermia would be required. References 1. Svensson LG, Crawford ES, Hess KR, Coselli JS, Safi HJ. Experience with 1509 patients undergoing thoracoabdominal aortic operations. J Vasc Surg 1993;17:357 70. 2. Adams HD, van Geertruyden HH. Neurologic complications of aortic surgery. Ann Surg 1956;144:574 610. 3. Mauney MC, Blackbourne LH, Langenburg SE, Buchanan SA, Kron IL, Tribble CG. Prevention of spinal cord injury after repair of the thoracic or thoracoabdominal aorta. Ann Thorac Surg 1995;59:245 52. 4. Fehrenbacher JW, McCready RA, Hormuth DA, et al. Onestage segmental resection of extensive thoracoabdominal aneurysms with left-sided heart bypass. J Vasc Surg 1993;18: 366 71. 5. Svensson LG, Kenneth RH, Coselli JS, Safi HJ. Influence of segmental arteries, extent, and atriofemoral bypass on postoperative paraplegia after thoracoabdominal aortic operations. J Vasc Surg 1994;20:256 62. 6. Frank SM, Parker SD, Rock P, Gorman RB, Kelly S, Beattie C, Williams GM. Moderate hypothermia, with partial bypass and segmental sequential repair of thoracoabdominal aortic aneurysm. J Vasc Surg 1994;19:687 97. 7. Griepp RB, Ergin MA, Galla JD, et al. Looking for the artery of Adamkiewicz: a quest to minimize paraplegia after oper-

Ann Thorac Surg UEDA ET AL 2000;70;38 43 SELECTIVE PERFUSION OF SEGMENTAL ARTERIES 43 ations for aneurysms of the descending thoracic and thoracoabdominal aorta. J Thorac Cardiovasc Surg 1996;112: 1202 15. 8. Hollier LH, Money SR, Naslund TC, et al. Risk of spinal cord dysfunction in patients undergoing thoracoabdominal aortic replacement. Am J Surg 1992;164:210 4. 9. Cambria RP, Davison JK, Zannetti S, et al. Clinical experience with epidural cooling for spinal cord protection during thoracic and thoracoabdominal aneurysm repair. J Vasc Surg 1997;25:234 43. 10. Kiefer E, Richard T, Chiras J, Godet G, Cormier E. Preoperative spinal cord arteriography in aneurysmal disease of the descending thoracic and thoracoabdominal aorta: preliminary results in 45 patients. Ann Vasc Surg 1989;3:34 46 11. Svensson LG, Patel V, Robinson MF, Ueda T, Roehm JOJ, Crawford ES. Influence of preservation or perfusion of intraoperatively identified spinal cord blood supply on spinal motor evoked potentials and paraplegia after aortic surgery. J Vasc Surg 1991;13:355 65. 12. Heinemann MK, Brassel F, Herzog T, Dresler C, Becker H, Borst HG. The role of spinal angiography in operations of the thoracic aorta: Myth or reality. Ann Thorac Surg 1998;65: 346 51. 13. Dommisse GF. The blood supply of the spinal cord: a critical vascular zone in spinal surgery. J Bone Joint Surg 1974;56: 225 35. 14. Mori A, Ueda T, Nakamichi T, et al. Detrimental effects of exogenous glutamate on spinal cord neurons during brief ischemia in vivo. Ann Thorac Surg 1997;63:1057 62. 15. Sakurai M, Hayashi T, Abe K, Sadahiro M, Tabayashi K. Delayed and selective motor neuron death after transient spoinal cord ischemia: a role of apoptosis? J Thorac Cardiovasc Surg 1998;115:1310 5. 16. Wadouh F, Arndt CF, Oppermann E, Borst HG, Wadouh R. The mechanism of spinal cord injury after simple and double aortic cross-clamping. J Thorac Cardiovasc Surg 1986;92:121 7. 17. Daput OE, Midulla PS, Sadeghi AM, et al. Pathogenesis of spinal cord injury during simulated aneurysm repair in a chronic animal model. Ann Thorac Surg 1994;58:689 97. 18. Laschinger JC, Cunningham JN, Nathan IM, Knopp EA, Cooper MM, Spencer FC. Experimental and clinical assessment of the adequacy of partial bypass in maintenance of spinal cord blood flow during operations on the thoracic aorta. Ann Thorac Surg 1983;36:417 26. 19. Crawford ES, Mizrahi EM, Hess KR, Cosseli JS, Safi HJ, Patel VM. The impact of distal aortic perfusion and somatosensory evoked potential monitoring on prevention of paraplegia after aortic aneurysm operation. J Thorac Cardiovasc Surg 1988;95:357 67. 20. Suzuki T, Kawada K, Soma Y, Imamura H, Taguchi S, Inoue T. Usefulness and related problems of somatosensory evoked potential monitoring for prevention of spinal cord injury secondary to operation of the aorta. Jpn J Cardiovasc Surg 1989;18:497 505. 21. Shiiya N, Yasuda K, Matui Y, Sakuma M, Sasaki S. Spinal cord protection during thoracoabdominal aortic aneurysm repair: results of selective reconstruction of the critical segmental arteries guided by evoked spinal cord potential monitoring. J Vasc Surg 1995;21:970 5. 22. Shahin GMM, Hamerlijnck RPHM, Schepens MAAM, Ter Beek HTM, Vermeulen FEE, Boezeman EHJF. Upper and lower extremity somatosensory evoked potential recording during surgery for aneurysms of the descending thoracic aorta. Eur J Cardiothorac Surg 1996;10:299 304. 23. Guérit JM, Verheist R, Rubay J, Khoury G, Matta A, Dion R. Mutilevel somatosensory evoked potentials SEPs for spinal cord monitoring in descending thoracic and thoracoabdominal aortic surgery. Eur J Cardiothorac Surg 1996;10: 93 104. 24. Ueda T, Nakamichi T, Yasudo M, et al. Selective perfusion of the intercostal arteries during thoracoabdominal aortic surgery. J Jpn Coll Angiol 1996;36:949 51. 25. lnoue T, Kawada S, Kawada K, Soma Y, Yozu R, Kudo M. Extended reconstruction for dissecting aneurysm of the aorta. Jpn J Thorac Surg 1992;45:387 94.