Assessment of Spinal Cord Circulation and Function in Endovascular Treatment of Thoracic Aortic Aneurysms

Assessment of Spinal Cord Circulation and Function in Endovascular Treatment of Thoracic Aortic Aneurysms Geert Willem H. Schurink, MD, PhD, Robbert J. Nijenhuis, MD, Walter H. Backes, PhD, Werner Mess, MD, PhD, Michiel W. de Haan, MD, PhD, Bas Mochtar, MD, PhD, and Michael J. Jacobs, MD, PhD Departments of Vascular Surgery, Radiology, Neurophysiology, and Cardiothoracic Surgery, University Hospital Maastricht, Maastricht, Netherlands Background. In thoracic stent graft repair, the importance of segmental artery (SA) occlusion and the role of blood pressure management during the intraoperative and directly postoperative period are not clear. To study these aspects in relation to spinal cord ischemia, our protocol in the endovascular treatment of descending thoracic aneurysms covering segmental arteries T8 and lower includes preoperative assessment of the spinal cord circulation using magnetic resonance angiography, intraoperative cerebrospinal fluid drainage, and spinal cord function monitoring using motor evoked potentials (MEPs). Methods. Thirteen patients with thoracic aortic aneurysms and dissections needing stent graft coverage of T8 and lower were included. In 9 patients, spinal cord circulation was evaluated preoperatively by magnetic resonance angiography. In 12 patients, MEPs were recorded during the endovascular procedure. A combination of both techniques was used in 8 patients. Results. The distal stent graft landing zone covered the intercostal arteries up to T10 in 4 patients, up to T11 in 7 patients, up to T12 in 1 patient, and all SAs to the aortic bifurcation in 1 patient. In 6 patients, the SA feeding the Adamkiewicz artery was covered by the stent graft. In three patients, intersegmental collaterals were present to the SA feeding the Adamkiewicz artery. The MEPs decreased to 50% and 30% in 2 patients, recovering to levels above 50% by elevation of the mean arterial pressure. Postoperatively, no signs of paraplegia were present. Conclusions. We believe that the presence of intersegmental collaterals decreases the risk of spinal cord ischemia during endovascular thoracic aortic aneurysm repair. Monitoring of MEPs during endovascular thoracic procedures shows no decrease in most cases. However, if a decrease of MEPs occurs, this can be reversed by elevation of the mean arterial pressure. (Ann Thorac Surg 2007;83:S877 81) 2007 by The Society of Thoracic Surgeons Presented at Aortic Surgery Symposium X, New York, NY, April 27 28, 2006. Address correspondence to Dr Schurink, Department of Surgery, P Debyelaan 25, University Hospital Maastricht, Maastricht 6202 AZ, Netherlands; e-mail: gwh.schurink@surgery.azm.nl. Endovascular interventions for thoracic aortic aneurysms (TAA) and thoracoabdominal aortic aneurysms (TAAA) potentially carry the same risk for developing spinal cord ischemia and paraplegia as the open approach. As the most obvious reason for developing paraplegia in open TAA and TAAA repair is a period of spinal cord ischemia, both the spinal cord blood supply and function are of interest during intervention. The focus in preoperative imaging is the Adamkiewicz artery (AKA) and its segmental supply because it is considered to be the most important provider of blood to the thoracolumbar spinal cord. Owing to atherosclerosis of the aortic wall, many segmental artery (SA) orifices are occluded. Blood supply to the AKA and spinal cord may therefore strongly depend upon collateral circulation. Recently, less invasive techniques such as magnetic resonance angiography (MRA) and computed tomographic angiography have been improved to image the spinal vasculature [1 4]. Since these new techniques are safe and have a high sensitivity, there is now renewed interest in preoperative imaging of the spinal cord blood supply. In addition to preoperative imaging, vascular surgeons can have access to intraoperative neuronal information on spinal cord function, for example, provided by transcranial motor evoked potentials (MEPs). In open repair, the importance of this technique has been demonstrated by adjusting hemodynamic and surgical strategies, including increasing distal aortic pressure and reattaching critical SAs for spinal cord perfusion [5, 6]. In endovascular repair of thoracic aortic aneurysms and dissections, the importance of segmental artery occlusion and the role of blood pressure management during the intraoperative and directly postoperative period is not clear. To study these aspects and their relationship to spinal cord ischemia, our protocol in the endovascular treatment of descending thoracic aneurysms covering segmental arteries T8 and lower includes preoperative assessment of the spinal cord circulation using MRA, intraoperative 2007 by The Society of Thoracic Surgeons 0003-4975/07/$32.00 Published by Elsevier Inc doi:10.1016/j.athoracsur.2006.11.028

S878 AORTIC SURGERY SYMPOSIUM X SCHURINK ET AL Ann Thorac Surg ENDOVASCULAR TAA TREATMENT IMPACT ON SPINAL CORD 2007;83:S877 81 cerebrospinal fluid drainage, and spinal cord function monitoring using MEPs. Patients and Methods Patients In our hospital, TAAs and TAAAs have been treated since June 2000. Until June 2005, 112 patients were operated on for type I and II TAAAs [5]. During the same period, 63 patients underwent 65 endovascular stent graft procedures for descending thoracic pathology. Emergency procedures were performed in 35 patients. Thirty elective procedures were performed for degenerative thoracic aortic aneurysms (n 27), and aneurysms in chronic aortic dissections (n 3). In 18 elective cases, the distal landing zone of the stent graft was caudal to the level of T8. In 9 patients, spinal cord circulation was evaluated preoperatively by MRA. In 10 patients, as well as in 2 who had acute type B dissections with distal thoracic aortic rupture, MEPs were recorded during the endovascular procedure. In 1 patient, MEPs were not measured owing to logistic problems. A combination of both techniques was used in 8 patients. This study was approved by the Ethics Committee of the University Hospital Maastricht, and informed consent was obtained from all patients. Technique of MRA Contrast-enhanced MRA was performed on a commercially available 1.5 Tesla scanner (Philips Intera; Philips Medical Systems, Best, Netherlands) to localize the AKA and the SAs supplying [1]. The MRA examination consisted of two dynamic phases: each phase took 40 s to enable differentiation of the AKA from the great anterior radiculomedullary vein based on temporal changes of contrast enhancement. The field of view of the MRA pulse covered T3 to S1 (45 to 50 cm). After acquisition, the level and side of the AKA was determined using curved multiplanar reformatted images targeted to the anterior surface of the spinal cord. In addition, targeted maximum intensity projections were created in the sagittal view to depict any intersegmental collaterals. Technique of MEPs The function of the spinal cord was monitored intraoperatively by measuring MEPs [6]. The brain was stimulated electrically (Digitimer D-185; Digitimer, Herfordshire, United Kingdom) by a train of five stimuli of 500 V and about 1 to 1.5 A each, with an interstimulus interval of 2 ms. The resulting MEPs were recorded with surface electrodes from the abductor pollicis brevis and the anterior tibial muscle on both sides, and the amplitude was measured between the maximal negative and positive deflection. The degree of muscle relaxation was adjusted from a measurement of the compound muscle action potential (CMAP) of the abductor digiti quinti muscle after a single supramaximal stimulation of the ulnar nerve at the wrist. During the endovascular procedure, we strove to achieve a value (T1%) of about 20% compared with the CMAP before induction of muscle relaxation. We used vecuronium administered through an infusion pump, the velocity of which was adjusted manually according to the CMAP values. All MEP amplitude values, blood pressure data, and the degree of muscle relaxation were transferred to an external database that allowed for graphic displays of trends in time and the calculation of the ratio between the amplitude of each anterior tibial MEP and the mean of both abductor pollicis MEPs. These ratios were the mainstay for qualifying whether amplitude changes of the anterior tibial anterior MEPs were critical. Table 1. Information Concerning Stent-Graft Extent, Spinal Cord Circulation on Magnetic Resonance Angiography (MRA), and Spinal Cord Function Monitoring MRA MEPs Patient Extent Stent-Graft Level AKA Side AKA Collaterals Registration Action 1 LSA T10 (no MRA) No changes 2 T7 T11 (no MRA) Decrease to 30% MAP 90 mm Hg 3 T6 T10 (no MRA) No changes 4 T4 T10 (no MRA) Decrease to 50% MAP 80 mm Hg 5 LSA T11 T12 Left No changes 6 T6 T10 T10 Left Collaterals from T11 No changes 7 LSA T11 T11 Left No collaterals No changes 8 LSA T11 T10 Right No collaterals No changes 9 Elephant trunk T11 T12 Left No changes 10 T3 T10 T12 Right No changes 11 LSA aortic bifurcation T10 Left No collaterals No changes 12 T3 T11 T11 Right Collaterals from T12 No changes 13 T4 T12 T11 Left Collaterals from T12 (no MEPs) AKA Adamokiewicz artery; LSA left subclavian artery; MAP mean arterial pressure; MEPs motor evoked potentials.

Ann Thorac Surg AORTIC SURGERY SYMPOSIUM X SCHURINK ET AL 2007;83:S877 81 ENDOVASCULAR TAA TREATMENT IMPACT ON SPINAL CORD S879 Fig 1. (A) Sagittal maximum intensity projection image of a 68- year-old man showing the aorta, segmental arteries (SAs), and intersegmental collaterals (arrowheads). Note that not all SAs have an open connection with the aorta. (B) In the coronal oblique multiplanar reformatted image, the spinal cord blood supply pathway to the Adamkiewicz artery (AKA [arrow]) is depicted. Segmental artery thoracic 10 (T10), which has a direct connection with the AKA, is occluded in the aortic wall (asterisk). The SA T10 and thereby the AKA are supplied by SA thoracic 11 (T11) through an intersegmental collateral (arrowhead). The distal landing zone of the thoracic stent graft did not cover SA T11, thereby ensuring blood supply to the AKA. Because spontaneous fluctuations of MEPs are common, changes of the ratio as high as 50% were encountered that were clearly not related to any intervention in most patients. So, only consistent decreases of MEP ratios (ie, reproducible during three consecutive stimulations) of greater than 50% were considered significant and relevant. Results In the 13 patients, the proximal landing zone of the stent graft was just distal to the left subclavian artery in 7 patients, covered level T4 in 2 patients, covered level T6 in 2 patients, and in an elephant trunk in 1 patient (Table 1). The distal landing zone covered the intercostal arteries up to T10 in 4 patients, up to T11 in 7 patients, and up to T12 in 1 patient. In 1 patient, all SAs from the subclavian artery to the aortic bifurcation were excluded from the circulation after debranching of the visceral arteries. Results of MRA Preoperative imaging of the spinal cord blood supply by MRA clearly identified the AKA in all 9 patients. The AKA arose between T10 and T12. In 6 patients (67%), the supplying SA of the AKA derived from the left, and in 3 (33%) from the right. The SA supplying the AKA was in 3 of the 9 patients not covered by the stent graft (Table 1). In 6 patients, the SA at the level of the AKA origin was covered by the stent graft. In 3 of these patients, the origin of the SA at the level of the AKA was already occluded in the aortic wall. The AKA was in these 3 cases supplied by a SA and an intersegmental collateral originating one level above or below the partially occluded SA (Fig 1). In 1 of these patients, the SA supplying the intersegmental collateral was also covered by the stent graft. In the 3 remaining patients, the SA at the level of the AKA was open in the aortic wall and thus was the SA supplying the AKA. In those 3 patients, the SA supplying the AKA was covered, and no intersegmental collateral could be imaged preoperatively. Results of MEPs In 12 of the 13 patients, MEPs were monitored during the endovascular procedure. No changes in MEPS were seen in 10 patients. In the other 2 patients, MEPs of the lower legs decreased after stent graft deployment to 50% and 30% of the MEPs registered at the arms. Stable MEPs above 50% were achieved by elevation of the mean arterial pressure up to 80 mm Hg and 90 mm Hg, respectively (Fig 2). These mean arterial pressures were maintained for 3 days. Postoperatively, no signs of paraplegia were present. No late paraplegia occurred. Comment Fortunately, in this group of patients, no paraplegia was encountered. Decrease in MEPs occurred twice, but was never below 25%, which is considered a sign of severe spinal cord ischemia. As in open TAAA repair, the first strategy after decrease of MEPs is elevation of the mean arterial pressure. In both patients with decrease of MEPs, this strategy led to restoration of the MEPs above 50%, which is considered to be a safe range. The monitoring of the MEPs during the endovascular intervention provided the postoperative blood pressure level. Chiesa and colleagues [7] suggest that a postoperative mean arterial blood pressure above 90 mm Hg is safe to prevent paraplegia. However, that would imply that all patients need the same mean arterial pressure to preserve adequate spinal cord perfusion. Probably the MEP-derived mean arterial pressure permits a more individualized approach. In the beginning of our endovascular thoracic experience, the patient was prepared for a conversion to open repair in case of absence of recovery of the MEPs. In case of spinal cord ischemia, little is known about the time available for reattaching SA without permanent spinal cord damage. Ishimaru and associates [8] reported the

S880 AORTIC SURGERY SYMPOSIUM X SCHURINK ET AL Ann Thorac Surg ENDOVASCULAR TAA TREATMENT IMPACT ON SPINAL CORD 2007;83:S877 81 Fig 2. Motor evoked potential (MEP) registration during endovascular exclusion of a distal descending thoracic aneurysm. The MEP registration of the left anterior tibial muscle (TAL [dashed line]) and the right anterior tibial muscle (TAR [solid line]) and the mean arterial pressure (MAP [broken line]) are reflected. The MEPs decrease to 30% of the initial value (MEPs to the arms stay unchanged). After elevation of the MAP, the MEPs increase to above 50%. use of a retrievable stent graft, which could be removed in case of decrease in evoked potentials. In 5 of the 8 patients who underwent both MRA and MEPs, the absence of a decrease in MEPs could be explained by preserving the feeding SA to the AKA, or by preserved SAs supplying intersegmental collaterals to the AKA. However, in the other 3 patients, the stent graft occluded the SA supplying the AKA, without the preoperatively demonstrated presence of an intersegmental collateral circulation on MRA. In these patients, it is unknown how their spinal cord circulation is maintained. Owing to the use of stainless steel containing stent grafts, postoperative MRA which could show the change in arterial supply to the spinal cord circulation is not advisable. Compared with open repair of thoracic aortic aneurysms and dissections, the low incidence of paraplegia in endovascular treatment is remarkable. In several publications of more than 100 patients, the paraplegia rate in endovascular treatment varies between 0% and 6% [7, 9 12]. Several possible explanations for the absence of spinal cord complications can be devised. First, less hemodynamic instability during intervention is probably an important phenomenon. A second important difference is that the aorta is not opened. Opening the aorta during open repair will lead to back-bleeding from the SAs, and depressurizing of the SA network, including the AKA. Further, the lack of aortic clamping and the need for distal aortic perfusion will guarantee uninterrupted pulsatile perfusion of the SAs. All the reasons lead to the absence of a spinal cord reperfusion syndrome, which may also be responsible for a part of the spinal cord complications. With the introduction of fenestrated and branched stent grafts, the endovascular treatment of thoracic and thoracoabdominal aneurysms is gaining ground [13]. More and more, SAs between T8 and L1 will be occluded by stent grafts. Techniques able to provide information about spinal cord circulation and function can probably help to select patients who will not suffer from SA coverage, and can provide important guidelines for postoperative management. Series with much larger experience with respect to intercostal artery management during open thoracoabdominal aneurysm repair, reported uncomplicated occlusion of many segmental arteries and concluded that routine surgical implantation of segmental vessels is not indicated, and that with evolving understanding of spinal cord perfusion, endovascular repair of the entire thoracic aorta should ultimately be possible without spinal cord injury [14]. These conclusions match with our much smaller experience.

Ann Thorac Surg AORTIC SURGERY SYMPOSIUM X SCHURINK ET AL 2007;83:S877 81 ENDOVASCULAR TAA TREATMENT IMPACT ON SPINAL CORD S881 In conclusion, we cannot draw firm conclusions from the small number of patients we treated in our protocol with preoperative MRA and intraoperative MEPs. However, we believe that intersegmental collaterals indicate the presence of a collateral circulation to the spinal cord. That decreases the risk of spinal cord ischemia during an endovascular thoracic aortic aneurysm repair. Motor evoked potentials during endovascular thoracic procedures do not decrease in most cases. However, if a decrease of MEPs occurs, this can be reversed by elevation of the mean arterial pressure in the majority of cases. References 1. Nijenhuis RJ, Gerretsen S, Leiner T, et al. Comparison of 0.5-M Gd-DTPA with 1.0-M gadobutrol for magnetic resonance angiography of the supplying arteries of the spinal cord in thoracoabdominal aortic aneurysm patients. J Magn Reson Imaging 2005;22:136 44. 2. Hyodoh H, Kawaharada N, Akiba H, et al. Usefulness of preoperative detection of artery of Adamkiewicz with dynamic contrast-enhanced MR angiography. Radiology 2005; 236:1004 9. 3. Kawaharada N, Morishita K, Fukada J, et al. Thoracoabdominal or descending aortic aneurysm repair after preoperative demonstration of the Adamkiewicz artery by magnetic resonance angiography. Eur J Cardiothorac Surg 2002;21:970 4. 4. Yamada N, Okita Y, Minatoya K, et al. Preoperative demonstration of the Adamkiewicz artery by magnetic resonance angiography in patients with descending or thoracoabdominal aortic aneurysms. Eur J Cardiothorac Surg 2000;18:104 11. 5. Jacobs MJ, Mess W, Mochtar B, et al. The value of motor evoked potentials in reducing paraplegia during thoracoabdominal aneurysm repair. J Vasc Surg 2006;43:239 46. 6. Jacobs MJ, Mess WH. The role of evoked potential monitoring in operative management of type I and type II thoracoabdominal aortic aneurysms. Semin Thorac Cardiovasc Surg 2003;15:353 64. 7. Chiesa R, Melissano G, Marrocco-Trischitta MM, et al. Spinal cord ischemia after elective stent-graft repair of the thoracic aorta. J Vasc Surg 2005;42:11 7. 8. Ishimaru S, Kawaguchi S, Koizumi N, et al. Preliminary report on prediction of spinal cord ischemia in endovascular stent graft repair of thoracic aortic aneurysm by retrievable stent graft. J Thorac Cardiovasc Surg 1998;115:811 8. 9. Greenberg RK, O Neill S, Walker E, et al. Endovascular repair of thoracic aortic lesions with the Zenith TX1 and TX2 thoracic grafts: intermediate-term results. J Vasc Surg 2005; 41:589 96. 10. Bortone AS, De Cillis E, D Agostino D, de Luca Tupputi Schinosa L. Endovascular treatment of thoracic aortic disease: four years of experience. Circulation 2004;110(Suppl 1): II262 7. 11. Leurs LJ, Bell R, Degrieck Y, et al. Endovascular treatment of thoracic aortic diseases: combined experience from the EUROSTAR and United Kingdom Thoracic Endograft registries. J Vasc Surg 2004;40:670 80. 12. Mitchell RS, Miller DC, Dake MD, et al. Thoracic aortic aneurysm repair with an endovascular stent graft: the first generation. Ann Thorac Surg 1999;67:1971 80. 13. Anderson JL, Adam DJ, Berce M, Hartley DE. Repair of thoracoabdominal aortic aneurysms with fenestrated and branched endovascular stent grafts. J Vasc Surg 2005;42: 600 7. 14. Etz CD, Halstead JC, Spielvogel D, et al. Thoracic and thoracoabdominal aneurysm repair: is reimplantation of spinal cord arteries a waste of time? Ann Thorac Surg 2006;82:1670 7.