Spinal Cord Protection During Open Repair of Thoracic and Thoracoabdominal Aortic Aneurysms Using Profound Hypothermia and Circulatory Arrest

Spinal Cord Protection During Open Repair of Thoracic and Thoracoabdominal Aortic Aneurysms Using Profound Hypothermia and Circulatory Arrest Thoralf M Sundt, MD, FACS, Mark D Flemming, MD, Gustavo S Oderich, MD, FACS, Norman E Torres, MD, Zhuo Li, MS, Judy Lenoch, BA, Manju Kalra, MBBS BACKGROUND: STUDY DESIGN: RESULTS: CONCLUSIONS: Reduced risk of paraplegia is argued as an advantage of endovascular repair of descending thoracic aortic aneurysms (DTA) and thoracoabdominal aortic aneurysms (TAAA); however, paraplegia rates with open repair vary widely. We identified consecutive patients undergoing open repair of TAAA or DTA with or without arch replacement using profound hypothermia and circulatory arrest as a spinal cord protection strategy on a single surgical service between June 1, 2001 and September 20, 2010. Ninety-nine procedures were performed in 94 patients with a mean age of 59 years (range 19 to 84 years), 56 of whom were male (60%). The extent of repair was TAAA in 37 (Crawford extent I in 6, extent II in 28, and extent III in 3), DTA in 37, and DTA plus arch in 25. Surgery was urgent or emergent in 25 patients (25%). Operative mortality (30-day) was 10% (10 of 99), including a mortality of 12% for arch DTA (3 of 26), 11% for TAAA (4 of 25), and 5% for isolated DTA (2 of 37). There were 11 (11%) strokes and 11 patients experienced renal failure (7 with dialysis). There were 15 late deaths and survival at 5 years was 74% (95% CI, 62.4 88.2%). No patients experienced paraplegia, although one had delayed paraparesis thought to be secondary to refractory hypotension postoperatively. Although the mortality and stroke risks for patients undergoing repair of DTA or TAAA using profound hypothermia and circulatory arrest are substantial, the risk for paraplegia is low. In appropriately selected patients, profound hypothermia and circulatory arrest should be the preferred technique for spinal cord protection for DTA and TAAA. (J Am Coll Surg 2011;212: 678 685. 2011 by the American College of Surgeons) The highly lethal nature of aneurysmal disease of the thoracic aorta has long been recognized 1 ; its operative repair has proven to be a formidable surgical challenge. In particular, finding strategies to minimize or prevent ischemic end-organ complications, particularly and most vexingly the spinal cord, has been an area of ongoing investigation for more than half a century. 2 Accordingly, development of endovascular stent-graft technology as an option for their treatment has generated great interest. 3 The success achieved with thoracic endografts and the recent development Disclosure Information: Nothing to disclose. Presented at Southern Surgical Association 122nd Annual Meeting, Palm Beach, FL, December 2010. Received December 15, 2010; Accepted December 15, 2010. From the Divisions of Cardiovascular Surgery (Sundt, Lenoch), Vascular Surgery (Flemming, Oderich, Kalra), Anesthesiology (Torres), and Biostatistics (Li), Mayo Clinic Rochester, Rochester, MN. Correspondence address: Thoralf M Sundt, MD, Division of Cardiac Surgery, Massachusetts General Hospital, 55 Fruit St, Boston, MA 02114. Email: tsundt@partners.org of fenestrations and branches to incorporate the visceral arteries has inspired the application of the endovascular approach to a broader population, including those with lower than expected operative risk and more extensive disease. 4 With this, one of the limitations of endovascular repair is the risk for paraplegia associated with extensive coverage of intercostal arteries. Determining the appropriate place for each of these approaches in our clinical armamentarium demands an understanding of the true contemporary risk of open repair. These data cannot be obtained from large multicenter databases because the risk of these complex procedures is dependent on both individual surgeon experience and hospital volume. 5 In addition, considerable progress has been made in reducing the risk of surgery through the appropriate use of adjunctive techniques. Techniques to provide distal aortic perfusion have been applied almost as early as the introduction of the heart-lung machine into clinical practice. 6 Drainage of cerebrospinal fluid from the spinal canal has been added to left-heart bypass in large series 7 and 2011 by the American College of Surgeons ISSN 1072-7515/11/$36.00 Published by Elsevier Inc. 678 doi:10.1016/j.jamcollsurg.2010.12.022

Vol. 212, No. 4, April 2011 Sundt et al Profound Hypothermia Spinal Cord Protection 679 has been argued to have virtually neutralized the impact of aortic clamp time. 8 Neuroprotective effects of hypothermia have also been advocated from the advent of efforts to repair such aneurysms. 2 This approach has been taken to its logical extent by Kulik and colleagues, in using profound hypothermia as a primary protective strategy. 9 This same approach has been adopted by others as well, 10,11 with good results and, in particular, paraplegia rates that are far lower than those often cited in the literature in support of more widespread endografting. During the past decade in our practice, we have increasingly used profound hypothermia and circulatory arrest as a spinal cord protective strategy for treatment of aneurysmal disease involving the thoracic and thoracoabdominal aorta, particularly as we have increasingly used endovascular stent grafts, 12 viewing these approaches as complementary rather than competitive. In the interest of better defining the place of each, we examined our results with profound hypothermia and circulatory arrest in the repair of aneurysmal disease of the thoracic aorta with or without concomitant replacement of the aortic arch proximally or abdominal aorta distally. METHODS Approval was obtained from the IRB with waiver of studyspecific consent. Only patients providing prospective consent for inclusion in research studies were included. Operative cases in which all or part of the descending thoracic aorta was replaced by open surgical repair were identified by a search of the computerized database of the Division of Cardiovascular Surgery at Mayo Clinic Rochester. As paraplegia has been reported as an important risk in the repair of various segments of the descending thoracic aorta, we included all cases involving replacement of the descending thoracic aorta, including those with concomitant replacement of the proximal (arch with or without ascending) or distal (abdominal) segments. To ensure some degree of uniformity in perfusion technique, only those cases performed on a single cardiovascular surgical service (TMS) were considered. Cases involving the thoracoabdominal aorta are routinely approached in a collaborative manner with members of the Division of Vascular Surgery and various vascular surgeons were involved in those cases. operative technique Substantial aortic valvular regurgitation has been considered a contraindication to the use of this technique, and an indication for the use of left-heart partial bypass or, more commonly, femoro-femoral partial distal bypass. Over time, we have moved from femoral arterial cannulation toward axillary artery cannulation. In rare cases, we have passed a cannula across the aortic valve via the left ventricle apex. If left ventricular dilation is appreciated, the left ventricle is vented via the apex. The proximal anastomosis is performed at the initiation of circulatory arrest and perfusion is reinstituted to minimize cerebral ischemic times. In the case of arch replacement, brachiocephalic reconstructions are performed first for the same reason. In the case of thoracoabdominal aneurysms, rewarming is performed after the distal anastomosis and spinal reimplantation have been completed and during visceral artery reconstruction. It is our practice to reimplant intercostal vessels between T9 and L1 during thoracoabdominal aortic aneurysm (TAAA) repair. When resecting the supradiaphragmatic descending thoracic aorta we bevel the distal anastomosis during descending aneurysm repair. Perfusion technique varied somewhat according to the extent of repair; however, duration of cooling was in, all instances, a minimum of 30 minutes with a target temperature of 18 C nasopharyngeal. Perfusate temperature is not permitted to be 13 C during cooling or 37.5 C during warming. The temperature gradient is maintained at 10 C between heat exchanger water temperature and venous return temperature. During cooling, ph management is performed by the alpha-stat method during cooling to 32 C and converted to ph-stat below that point. During warming alpha-stat is used. Hemodilution is to 7 to 8 g/dl during cooling and restored to 9.5 g/dl before terminating bypass. Immediately before arrest, we administer 20 to 40 meq KCl intravenously to induce electromechanical arrest. We do not administer barbiturates or steroids. The target systemic mean blood pressure on bypass is 55 to 65 mmhg and 85 to 90 mmhg post-bypass and during postoperative recovery. statistical methods Descriptive statistics for categorical variables are reported as frequency and percentage, and continuous variables are reported as mean (SD) or median (range) as appropriate. Logistic regression models were used to find the univariate predictors of postoperative outcomes. Kaplan-Meier method was used to draw the survival curve and calculate 5-year survival statistics. Log-rank test and Cox regression models were used to find the univariate predictors for longterm mortality. All statistical tests were 2-sided, with the level set at 0.05 for statistical significance. RESULTS There were 99 procedures performed in 94 patients during the study interval (Table 1). During this same interval, there were 3 TAAAs repaired with hypothermic arrest by

680 Sundt et al Profound Hypothermia Spinal Cord Protection J Am Coll Surg Table 1. Demographic Characteristics of the Study Population Descending* arch (n 25) Descending* (n 37) TAAA* (n 37) Total patients (n 94) n % n % n % n % Age, y, mean SD (median) 61 16 (65) 58 19 (61) 60 19 (69) 59 18 (65) Male 14 56 21 57 23 62 56 60 Hypertension 20 80 25 68 29 78 73 78 Diabetes 2 8 1 3 3 8 6 6 Renal failure 2 8 1 3 1 3 4 4 Chronic lung disease 5 20 7 19 12 32 24 26 Earlier CV surgery 14 58 17 46 22 60 49 52 CAD 12 48 12 32 8 22 31 33 CV disease 10 42 5 14 3 8 17 18 Marfan 5 20 4 11 6 16 15 15 Dissection Acute 2 8 6 16 1 3 9 9 Chronic 11 44 14 38 14 38 37 37 *Five patients had 2 operations, so there were 99 operations among 94 patients. CAD, coronary artery disease; CV, cardiovascular; TAAA, thoracoabdominal aortic aneurysm. another cardiovascular surgeon and 72 thoracoabdominal aneurysms repaired using other techniques, including leftheart bypass or partial femoro-femoral bypass. Hypertension was common in all regardless of the surgical procedure performed. More than half had undergone earlier cardiovascular surgery. Fifteen patients had a history of Marfan syndrome and 45 had acute (n 9) or chronic (n 36) dissection. Of these, 25 procedures involved replacement of the entire aortic arch well as the upper descending thoracic aorta and the remainder were split evenly between descending and thoracoabdominal replacement. Extent of descending thoracic replacement was Safi extent A in 16 (64%) and C in the remainder 13 (Table 2). The extent of resection was split roughly evenly between Safi A and C among those undergoing replacement of the descending thoracic aorta only. In those patients undergoing thoracoabdominal aortic replacement, most (76%) were extent II. Approximately one-fourth of the procedures were performed urgently Table 2. Details of Operative Procedures Descending arch (n 25) Descending (n 37) TAAA (n 37) Total (n 99) Urgent/emergent, n (%) 8 (32) 12 (32) 5 (14) 25 (25) Concomitant procedure, n (%) AVR 5 (20) 0 (0) 0 (0) 5 (5) Root replacement 3 (12) 1 (3) 0 (0) 4 () CABG 6 (24) 4 (11) 1 (3) 11 (11) Other 3 (12) 9 (24) 29 (78) 41 (41) Extent Safi, n (%) A 16 (64) 17 (46) NA 33 (33) B 0 (0) 1 (3) NA 1 (1) C 9 (36) 19 (51) NA 28 (28) Crawford, n (%) I NA NA 6 (16) 6 (6) II NA NA 28 (76) 28 (28) III NA NA 3 (8) 3 (3) CPB time (min), mean SD (median) 233 40 (224) 193 52 (189) 236 43 (227) 219 49 (214) Cerebral time (min), mean SD (median) 29 15 (34) 29 14 (26) 21 8 (20) 26 13 (24) Femoral cannulation, n (%) 2 (8) 9 (24) 7 (19) 18 (18) AVR, aortic valve replacement; CABG, coronary artery bypass graft; CPB, cardiopulmonary bypass.

Vol. 212, No. 4, April 2011 Sundt et al Profound Hypothermia Spinal Cord Protection 681 Table 3. Perioperative (30-day and In-hospital) Outcomes Descending arch (n 25) Descending (n 37) TAAA (n 37) Total (n 99) n % n % n % n % Paraplegia 0 0 1 3 0 0 1 1 Stroke 3 12 6 16 2 5 11 11 Renal failure 3 12 3 8 5 14 11 11 Death 3 12 3 8 4 11 10 10 TAAA, thoracoabdominal aortic aneurysm. or emergently. Concomitant procedures were most common among those having arch surgery. As shown in Table 3, only 1 patient experienced a spinal cord ischemic complication. He awakened and moved his lower extremities, but during the early postoperative period experienced refractory hypotension with a systolic pressure of 70 to 80 mmhg secondary to ventricular irritability and persistent pacing via his internal cardioverter defibrillator/ pacer device. He has fortunately experienced considerable recovery of function since surgery. Stroke occurred most commonly among patients undergoing repair of a descending thoracic aneurysm, and in most cases was thought to be secondary to femoral arterial cannulation for arterial inflow. We have migrated away from femoral cannulation because of this observation. Renal failure, defined as serum creatinine 2.0 g/dl, a 2-fold increase in serum creatinine, or new need for dialysis, was experienced in 11% of patients overall. Seven patients required use of dialysis on at least a temporary basis. Operative mortality was 10%. There were too few events to permit multivariate analysis of predictors of outcomes. Univariate analysis demonstrated age to be a predictor of perioperative death, stroke, and renal failure (Table 4). It is notable that 8 of the 10 mortalities occurred in patients older than 70 years of age. Presence of coronary artery disease was a strong predictor of both renal failure and death. Late survival, as shown in Figure 1, was impacted by age, hypertension, renal dysfunction, and chronic lung disease. DISCUSSION Results of this study confirm the effectiveness of profound hypothermia as a protective strategy against spinal cord ischemia during surgery of the descending thoracic and thoracoabdominal aorta. As such, other surgical approaches to this challenging anatomic territory, both surgical and endovascular, should be considered with this modality as an option. Still, the mortality and morbidity associated with this approach is considerable and should not be dismissed. The principal determinants of risk with open repair using profound hypothermia and circulatory arrest in our experience are age and comorbidities, including presence of coronary artery disease. Assuming a differential impact of these comorbidities on periprocedural risk Table 4. Univariate Predictors of Adverse Outcomes Outcomes Risk factor OR (95% CI) p Value Early outcomes Paraplegia NA NA NA Stroke Age 1.1 (1.0 1.2) 0.022 Renal failure Age 1.1 (1.0 1.2) 0.025 CAD 4.3 (1.2 16.1) 0.030 Earlier CV surgery 0.1 (0 0.6) 0.015 CPB time (per 10 min) 1.2 (1.0 1.4) 0.017 Death Age 1.1 (1.0 1.2) 0.037 CAD 5.8 (1.4 24.5) 0.016 Late outcomes HR (95% CI) Death Age 1.1 (1.0 1.1) 0.0200 Hypertension NA* 0.0474* Renal failure 5.8 (1.6 21.6) 0.0080 Chronic lung disease 3.2 (1.1 9.5) 0.0035 *Among the patients without hypertension, there is no death event so HR and 95% CI cannot be estimated. However, from log-rank test, the p 0.0474, indicating the difference between the survival of the 2 groups is statistically different. CAD, coronary artery disease; CPB, cardiopulmonary bypass; CV, cardiovascular; HR, hazard ratio; OR, odds ratio.

682 Sundt et al Profound Hypothermia Spinal Cord Protection J Am Coll Surg Figure 1. Late survival after repair of thoracic aortic aneurysmal disease (operative survivors). for endovascular therapy, our data would support a selective, patient-specific approach to the application of these complementary therapies. Results we report here confirm those reported previously by Kulik and colleagues 9 and Fehrenbacher and colleagues. 10 In 2010, Kouchoulkos detailed his experience from 1986 to 2008 with the repair of TAAAs using profound hypothermia and circulatory arrest. Among 218 patients, including 57 with Crawford extent I disease, 91 with Crawford extent II disease, and 70 with Crawford extent II disease, paraplegia was observed in only 10 patients (4.6%). The operative mortality rate was somewhat lower than ours at 7.3%, as was that of stroke at 3.7%. Operations in this series were undertaken emergently somewhat less frequently (8%). Fehrenbacher and colleagues reported paraplegia in 4 patients among 173 undergoing repair of descending thoracic or thoracoabdominal aneurysm (extent I in 27, extent II in 49, extent III in 20, and extent IV in 14) using profound hypothermia and circulatory arrest, for a rate of 2.4%. Their operative mortality rate was also lower than ours at 4% and a stroke rate of 4.1%. In their series, 10% of cases were urgent or emergent. Our results compare favorably with those reported using spinal cord cooling both anecdotally in our own practice and as reported in the literature. Among 576 patients reported from Massachusetts General Hospital between 1987 to 2005, the paraplegia rate was 6.6% among 105 patients having thoracic endovascular repair and 12% among 93 undergoing open surgical repair of descending thoracic aortic pathology. 14 Among TAAA repairs, the paraplegia rate was 24% among 58 with extent I disease, 20% among 65 with extent II disease, 13% among 156 with type III disease, and 2% among 99 with type IV disease. The mortalty rate in this series was 8%. The more widely used technique of left-heart bypass was introduced by Cooley and colleagues almost at the advent of extracorporeal circulation, 6 and has been most strongly advocated by the other pioneers in the field from Houston, including Crawford 15 and, more recently, his successors. Coselli and colleagues have reported a 3.8% paraplegia rate and a remarkable operative mortality rate of 5% among 2,286 patients undergoing repair of thoracoabdominal aneurysms between 1986 and 2006. 7 Looking specifically at those with Crawford extent II aneurysms, which were the highest risk, the mortality rate was 6% and paraplegia rate was 6.3%. Safi and associates have argued convincingly on the basis of their similarly large exerience that the use of 3 adjuncts, specifically permissive hypothermia to 34 C, leftheart bypass for distal perfusion, and spinal cord drainage, effectively neutralizes the impact of aortic occlusion time during repair of TAAAs. 8 In their series, when these adjuncts were used in Crawford extent II aneurysm, the paraplegia rate was only 3.3%. Schepens and colleagues have reported a reduction in the paraplegia rate from 8.3% to 4.4% with the addition of intraoperative motor-evoked potentials to guide care in 571 patients undergoing thoracoabdominal aortic repair between 1981 and 2008. 16 Their 30-day mortality rate was 8.9%. Endovascular repair of thoracoabdominal aneurysms has gained acceptance in several centers using fenestrated and branched technology or debranching procedures. 17,18 Although the main advantage of an endovascular approach includes shorter operative time, less blood loss, and, potentially, less morbidity and mortality, a limitation continues to be the increased risk of spinal cord injury associated with extensive coverage of collateral network to the spinal cord. 19 This is especially the case in extent I and II TAAAs. In the largest comparative report from the Cleveland Clinic, the risk of spinal cord injury was 19% for type II extent TAAA treated endovascularly and 20% for those treated open, although the latter were largely done with heart bypass and not with deep hypothermia. Therefore, in young patients who are good surgical candidates with extensive aortic disease, particularly those with chronic dissections or connective tissue disorders, open repair using the strategy of deep hypothermia and circulatory arrest might offer the best risk-to-benefit ratio and the lowest reported rates of spinal cord injury. The limitations of our study include its relatively small size and heterogeneous population of patients. Although our study is far smaller than those reported from Houston or a handful of other very large centers in Europe, which limits our ability to make definitive comments about the very predictors of outcomes that are required to best choose a particular therapy for a particular patient, our experience is likely more representative of that in most centers in the United States. Although the large number of cases reported from the centers in Houston permit the kind of statistical analysis we seek, repair of thoracic and thora-

Vol. 212, No. 4, April 2011 Sundt et al Discussion 683 coabdominal aortic aneurysms is highly complex and the results achieved are dependent on interdisciplinary teamwork and the associated learning curves of all members of that team. Few centers can, therefore, reproduce the truly remarkable results reported from these virtuoso centers. Our study is also limited by the potential for selection bias, as not all thoracic and thoracoabdominal aortic aneurysms are repaired in this manner. Apart from the patient-related exclusionary criteria, including substantial aortic valvular regurgitation, during the study interval not all vascular or cardiac surgeons preferred this approach. Accordingly, many patients underwent repair using left-heart partial bypass or other techniques, with circulatory arrest sought only in some circumstances. It is difficult to account for the impact this might have had on either increasing or decreasing the observed operative risk of death. Regardless, this should not have impacted the risk of paraplegia among our patients, which was the principle focus of our analysis. CONCLUSIONS We conclude that profound hypothermia with circulatory arrest is a technique providing outstanding spinal cord protection. Even as our practice of endovascular surgery has expanded to include fenestrated thoracoabdominal repairs in addition to descending thoracic repair, we have, as an institution, migrated increasingly to circulatory arrest when open repair is chosen. It is our practice to use open repair in younger patients who have fewer comorbidities to provide the most durable repair with the lowest risk of paraplegia. In higher-risk elderly patients, we might be willing to accept a higher risk of spinal cord complications in return for a lower risk of mortality when making our recommendations for care to our patients. The variety of techniques now available to cardiovascular surgeons for the repair of thoracic aortic disease should be considered complementary. Author Contributions Study conception and design: Sundt, Flemming, Oderich, Kalra Acquisition of data: Sundt, Flemming, Oderich, Li, Lenoch, Kalra Analysis and interpretation of data: Sundt, Flemming, Oderich, Torres, Li, Lenoch, Kalra Drafting of manuscript: Sundt, Li, Lenoch Critical revision: Sundt, Flemming, Oderich, Torres, Li, Lenoch, Kalra REFERENCES 1. Joyce JW, Fairbairn JF 2nd, Kincaid OW, Juergen JL. Aneurysms of the thoracic aorta. A clinical study with special reference to prognosis. Circulation 1964;29:176 181. 2. De Bakey ME, Cooley DA. Surgical considerations of acquired diseases of the aorta. Ann Surg 1954;139:763 775. 3. Makaroun MS, Dillavou ED, Kee ST, et al. Endovascular treatment of thoracic aortic aneurysms: results of the phase II multicenter trial of the GORE TAG thoracic endoprosthesis. J Vasc Surg 2005;41:1 9. 4. Bavaria JE, Appoo JJ, Makaroun MS, et al. Endovascular stent grafting versus open surgical repair of descending thoracic aortic aneurysms in low-risk patients: a multicenter comparative trial. J Thorac Cardiovasc Surg 2007;133:369 377. 5. Cowan JA Jr, Dimick JB, Henke PK, et al. Surgical treatment of intact thoracoabdominal aortic aneurysms in the United States: hospital and surgeon volume-related outcomes. J Vasc Surg 2003;37:1169 1174. 6. Cooley DA, Belmonte BA, De Bakey ME, Latson JR. Temporary extracorporeal circulation in the surgical treatment of cardiac and aortic disease; report of 98 cases. Ann Surg 1957;145:898 912. 7. Coselli JS, Bozinovski J, LeMaire SA. Open surgical repair of 2286 thoracoabdominal aortic aneurysms. Ann Thorac Surg 2007;83:S862 S864. 8. Safi HJ, Estrera AL, Miller CC, et al. Evolution of risk for neurologic deficit after descending and thoracoabdominal aortic repair. Ann Thorac Surg 2005;80:2173 2179. 9. Kulik A, Castner CF, Kouchoukos NT. Outcomes after thoracoabdominal aortic aneurysm repair with hypothermic circulatory arrest. J Thorac Cardiovasc Surg 2010 Jul 9. [Epub ahead of print]. 10. Fehrenbacher JW, Hart DW, Huddleston E, et al. Optimal endorgan protection for thoracic and thoracoabdominal aortic aneurysm repair using deep hypothermic circulatory arrest. Ann Thorac Surg 2007;83:1041 1046. 11. Coselli JS, Bozinovski J, Cheung C. Hypothermic circulatory arrest: safety and efficacy in the operative treatment of descending and thoracoabdominal aortic aneurysms. Ann Thorac Surg 2008;85:956 963. 12. Oderich GS, Ricotta JJ 2 nd. Modified fenestrated stent grafts: device design, modifications, implantation, and current applications. Perspect Vasc Surg Endovasc Ther 2009;21:157 167. 13. Estrera AL, Miller CC 3rd, Chen EP, et al. Descending thoracic aortic aneurysm repair: 12-year experience using distal aortic perfusion and cerebrospinal fluid drainage. Ann Thorac Surg 2005;80:1290 1296; discussion 1296. 14. Conrad MF, Ye JY, Chung TK, et al. Spinal cord complications after thoracic aortic surgery: long-term survival and functional status varies with deficit severity. J Vasc Surg 2008;48:47 53. 15. Crawford ES. Thoraco-abdominal and abdominal aortic aneurysms involving renal, superior mesenteric, celiac arteries. Ann Surg 1974;179:763 772. 16. Schepens MA, Heijmen RH, Ranschaert W, et al. Thoracoabdominal aortic aneurysm repair: results of conventional open surgery. Eur J Vasc Endovasc Surg 2009;37:640 645. 17. Greenberg RK, Lu Q, Roselli EE, et al. Contemporary analysis of descending thoracic and thoracoabdominal aneurysm repair: a comparison of endovascular and open techniques. Circulation 2008;118:808 817. 18. Chuter TA, Rapp JH, Hiramoto JS, et al. Endovascular treatment of thoracoabdominal aortic aneurysms. J Vasc Surg 2008; 47:6 16. 19. Etz CD, Zoli S, Bischoff MS, et al. Measuring the collateral network pressure to minimize paraplegia risk in thoracoabdominal aneurysm resection. J Thorac Cardiovasc Surg 2010; 140[Suppl]:S125 S130; discussion S142 S146.

684 Sundt et al Discussion J Am Coll Surg Discussion DR LYNN HARRISON (Miami, FL): If one were to superimpose a frequency distribution plot for this type of surgery over a map of the United States, we would see that both K2 and Mount Everest reside in Houston, Annapurna is in Cleveland, and New York, St Louis, and more recently, Rochester, MN, have formidable peaks of their own. In comparison with the numbers that Dr Coselli shared with us earlier this morning, this may seem like a relatively small series. But let me point out, this is a single-surgeon experience, which exceeds in numbers 95% of the institutional experiences reported annually to the Society of Thoracic Surgeons database. So this is a prodigious accomplishment. And it certainly sets the benchmark for spinal cord protection during these difficult surgical procedures. As I looked through the manuscript, which Dr Sundt thoughtfully provided before the meeting, I found myself searching for some pearl, some technical key, or some selection criterion that would have enabled these results to occur. But the only thing that I was able to identify was the routine reimplantation of all intercostals, T9 through L1. I don t know of other series that do that routinely in all cases, but that is really the only thing that I could identify that might explain these extraordinary results. If it were brilliant technical expertise in the operating room or exceptional postoperative care, then I would have expected the incidence of death, renal failure, and stroke also to be quite low. But in fact, they are in about the middle, the respectable middle, of a number of other series that are reported. So, Thor, my first question to you is, do you have some insight as to how you achieved these excellent numbers? You indicated that you switched from retrograde femoral artery perfusion to axillary artery perfusion, and had the impression that that decreased your incidence of stroke. I assume, then, that you use right axillary perfusion for ascending lesions that you approached through median sternotomy and left axillary perfusion on patients whom you explored by a left thoracotomy. But have you considered or would you consider the right axillary in your left thoracotomy patients in that it affords you the opportunity to clamp the base of the brachiocephalic artery and then provide continuous antegrade perfusion during the circulatory arrest period, which I am convinced contributes to cerebral preservation. Finally, is this the swan song or the last hurrah of the open approach? As you know, in May, in Toronto, Dr Coselli and his colleagues presented an extensive meta-analysis of nearly 12,000 patients who were operated on or who underwent the endovascular approach across the country. And interestingly, the mortality rate was exactly the same. The endovascular group went home a day and a half earlier, which is probably not terribly important. But they were 4 times more likely than the operated group to go home without the need for ongoing care. Now, this was despite the fact that the endovascular group was fully a decade older than the surgical group and had more comorbidities, which were more severe. So 5 or 10 years hence, will open repair be an interesting historical footnote? What do you think the future holds? This was a terrific paper. Because it would have been a welcome addition to the program of any of the cardiovascular subspecialty societies, I m particularly pleased you presented it here. DR JAY ZWISCHENBERGER (Lexington, KY): I want to thank the Southern Surgical Association for the privilege of commenting on 2 aortic papers at the same meeting. It s rare that one is able to comment on 2 of the greatest aortic surgeons in the world (Joe Coselli and Thor Sundt) sequentially. Thor, I am very pleased to see that you have built on the pioneering work of your mentor, Nic Kouchoukos. As we learned from Joe Coselli earlier today, circulatory arrest has not been adopted widely, likely in part out of concern for the bleeding complications encountered using this technique. Do you have a sense for the comparative use of blood products between cases in which you used circulatory arrest versus left heart bypass? Second, you have published on, and Joe earlier demonstrated, the occurrence of delayed paraplegia. You mentioned that you didn t have enough numbers to comment on, but certainly you have a sense for measures to avoid this dreaded complication. Third, how many cases were done during this same time using other techniques? Last, how does your group address patient selection using, or not, as Joe implied, a multidisciplinary approach? DR THOR SUNDT (Rochester, MN): Thank you for reminding me to make note of the coauthors. Importantly, the coauthors noted here are vascular surgeons. Drs Fleming, Oderich, and Kalra are vascular surgeons, as well as Dr Norman Torres, who is an outstanding cardiovascular anesthesiologist, and that leads into the answer to Dr Zwischenberger s question about a multidisciplinary approach. The way we approach thoracoabdominal aneurysms routinely at Mayo is in a combined and collaborative manner with our vascular surgeons. And so these cases are all done with them. I was the only cardiac surgeon involved, in part because I have the greatest interest in this condition. But these are vascular surgeons, so I appreciate your bringing that to light. The first question was about insights into the reason we are able to keep the rate of paraplegias so low. I think that hypothermia provides the best possible neurologic protection, and so we are able to avoid intraoperative injury. To be honest with you, Hazzim Safi has argued that there s a tradeoff in choosing between left heart bypass and circulatory arrest that amounts to a tradeoff between paraplegia and death. I suspect Joe Coselli would say the same thing. Our mortality rate is a bit higher than Hazzim or Joe has reported. I am not as pleased with our mortality rate; it s significant, although, as I said, the deaths were in the older patients. But I think that there probably is a bit of a trade-off there between paraplegia and death. I think the quality of intraoperative preservation of the spinal cord is unparalleled with this approach. We are aggressive about reimplantation of the intercostal vessels too. This is because of our experience in St Louis with delayed paraplegia, as previously reported. In St Louis, we tended to do more with patients with left heart bypass. With careful attention to intraoperative strategies, our incidence of delayed paraplegia actually exceeded that of immediate paraplegias, just as Joe has described in his experience. We believed that the best protection against this was reimplanting intercostal vessels, and we became enthusiastic proponents of this. Lynn, you asked about axillary artery for inflow. We have done that more and more in all of our aortic surgery. I haven t used the

Vol. 212, No. 4, April 2011 Sundt et al Discussion 685 right axillary when we re going through the left chest because it is so far away, although I take your point about being able to control the base of the innominate artery. I have used the left ventricular apex for inflow on a couple of cases when there was some mild to moderate aortic regurgitation, and found that passing a femoral cannula through the apex and up through the aortic valve provided antegrade perfusion and minimized the aortic regurgitation. And on occasion, we have actually cannulated the ascending aorta from the left chest just by opening the pericardium and reflecting the pulmonary artery down. You ask if this is the swan song for open surgery given advances in endovascular surgery. Surely, if one takes the long view, 100 years from now no one will believe that we did the kinds of things that we re doing today. So yes, in a way I acknowledge that it s not a matter of if, it s a matter of when. I certainly don t want to appear a Luddite. I welcome these new technologies, and they can benefit our patients, which, after all, is the most important thing. And I think that what we need to do is tailor the therapy to the patients. So, for example, today at Mayo when we have a descending thoracic or thoracoabdominal aneurysm, we ll very often have a conversation between the vascular surgeons and myself regarding what we think is the best treatment for this particular patient. If it s a young person with Marfan syndrome, we re very likely to go with an open operation. If it s a very elderly person with many comorbidities, then we re more likely to take the risk of paraplegia associated with extensive coverage of intercostal vessels in an endovascular approach. We believe that the risk of paraplegia with extensive coverage of the thoracoabdominal aorta is greater. If it s a very elderly person who we think is a high operative risk, we re more likely to go that route. Zwisch, you asked about bleeding. It s my sense that the bleeding is equivalent between the two. There certainly can be a lot of bleeding with the left heart approach as well. And we are quite meticulous with our hemostasis using this approach. And of course, the whole time that you are on pump, you are able to recirculate that blood through the pump suckers. So I actually think it may be a wash in terms of bleeding complications. You asked about measures to prevent delayed paraplegia. I think the principal one is reimplantation of the vessels, although I have also been quite adamant with our ICU about maintaining the mean blood pressure postoperatively above 90 mmhg or so. Much to my chagrin, when I first moved to Mayo, I discovered that the lore was to maintain the blood pressure low to prevent bleeding. I would rather deal with bleeding than paraplegia. So I found that my mission for several years was to convince people that we needed to keep the mean blood pressure of 85 to 90 mmhg, and we actually post that in the rooms of the patients to prevent delayed paraplegia. How many other cases were performed in other manners? During this same interval, 3 other cases were done with circulatory arrest with other cardiac surgeons involved. About 75 cases were done with left heart bypass in a variety of combinations of surgical groups or surgeons involved. I don t have the results of those at hand to do any kind of a comparison. I m not sure that that would be very valid in any event. And again, I think it s important that we have a multidisciplinary approach. It s not quite like your tumor board, but I can tell you that the conversations in the hallway in our institution tend to function like that.