Hypothermic Bypass and Circulatory Arrest for Operations on the Descending Thoracic and Thoracoabdominal Aorta

Transcription

1 Hypothermic Bypass and Circulatory Arrest for Operations on the Descending Thoracic and Thoracoabdominal Aorta Nicholas T. Kouchoukos, MD, Bill B. Daily, MD, PhD, Chris K. Rokkas, MD, Suzan F. Murphy, RN, Spomenko Bauer, MD, and Nabil Abboud, MD Division of Cardiothoracic Surgery, Department of Surgery, Washington University School of Medicine, St. Louis, Missouri Background. Hypothermic cardiopulmonary bypass with intervals of circulatory arrest is a useful adjunct during operations on the descending thoracic aorta and distal aortic arch when severe aortic disease precludes placement of clamps on the aorta. Hypothermia also has a marked protective effect on spinal cord function during periods of aortic occlusion. Methods. Fifty-one patients (age range, 22 to 79 years) with descending thoracic or thoracoabdominal aortic disease had resection and graft replacement of the diseased aortic segments using hypothermic cardiopulmonary bypass and intervals of circulatory arrest in situations where the location, extent, or severity of disease precluded placement of clamps on the proximal aorta (8 patients) or (in 43 patients) when extensive thoracic (11) or thoracoabdominal (32) aortic disease was present and the risk for development of spinal cord ischemic injury and renal failure was judged to be increased. Patent intercostal (below T-6) and upper lumbar arteries were attached to the graft whenever possible. Results. Thirty-day mortality was 9.8% (5 patients). Paraplegia occurred in 2 and paraparesis in 1 of the 46 3-day survivors (6.5%). Among the 27 operative survivors with thoracoabdominal aneurysms, paraplegia occurred in I of 12 with Crawford type I (8%), of 1 with type II, and 1 of 5 with type III aneurysms (2%). Paraplegia occurred in none of the 12 patients with aortic dissection and in 2 of the 15 patients with degenerative aneurysms. Renal failure requiring dialysis occurred in 1 (2.2%) of the 46 3-day survivors. Conclusions. Hypothermic circulatory arrest is a valuable adjunct for the treatment of complex aortic disease involving the aortic arch and thoracoabdominal aorta. In patients with thoracoabdominal aneurysms, its use has been associated with a low incidence of renal failure and an incidence of paraplegia/paraparesis in traditionally high-risk subsets (type I and II aneurysms, aortic dissection), which may be less than that observed with other surgical techniques. (Ann Thorac Surg 1995;6:67-77) p rofound hypothermic circulatory arrest with an anterior approach is currently a widely used technique for protection of the central nervous system during operations for treatment of congenital and acquired cardiac lesions and for aneurysms and dissections of the ascending aorta and aortic arch. The technique has been used less frequently for treatment of aortic disease that necessitates posterolateral exposure. However, it has proved to be of considerable value for the treatment of disease of the aortic arch and descending thoracic aorta when placement of aortic clamps is not possible or is considered hazardous [1-8]. Hypothermia has been shown experimentally to have a marked protective effect on spinal cord function during periods of aortic occlusion [9-12] and in studies employing hypothermic cardiopulmonary bypass and circulatory arrest in large animals where resection and graft replacement of the thoracoabdominal aorta were simu- Presented at the Thirty-first Annual Meeting of The Society of Thoracic Surgeons, Palm Springs, CA, Jan 29-Feb 1, Address reprint requests to Dr Kouchoukos, Division of Cardiothoracic Surgery, Washington University School of Medicine, Jewish Hospital, 216 S Kingshighway, St. Louis, MO lated [13, 14]. We evaluated the role of elective hypothermic cardiopulmonary bypass and circulatory arrest in patients with aortic disease involving the distal aortic arch, the descending thoracic aorta, and the thoracoabdominal aorta where the location, extent, or severity of disease precluded placement of clamps on the aortic arch or the proximal descending thoracic aorta, and when extensive thoracic and thoracoabdominal aortic disease was present and the risk of spinal cord ischemic injury was judged to be increased. Material and Methods Patient Selection Between January 1986 and October 1994, 51 patients with aortic disease involving the distal aortic arch, the descending thoracic aorta, or the thoracoabdominal aorta underwent resection and graft replacement of the diseased aortic segments using elective hypothermic cardiopulmonary bypass usually in combination with a period of circulatory arrest. The patients ranged in age from 22 to 79 years (mean age, 57 years) and 27 (53%) were male. Seven patients had the clinical stigmata of Marfan's syndrome. Thirty-five patients had symptoms associated 1995 by The Society of Thoracic Surgeons /95/$ (95)353-M

2 68 KOUCHOUKOS ET AL Ann Thorac Surg HCA FOR THORACOABDOMINAL AORTA OPERATIONS 1995;6:67-77 Table 1. Extent of Aortic Disease and Aortic Replacement Extent Aortic Disease. of Patients Dissection Other Distal aortic arch and proximal descending thoracic aorta Arch and entire DTA Most or all of DTA Thoracoabdominal aorta Crawford type I Crawford type II Crawford type III Total DTA = descending thoracic aorta. with their aortic disease. The remaining patients had aneurysms that were more than twice the size of the adjacent normal aorta or had evidence, by serial computed tomographic scans, of progressive enlargement of the aorta. Seven of 51 patients underwent emergent operation for rupture of the aneurysm or because of the presence of acute dissection. Thirty patients had previous operations on the intrathoracic or thoracoabdominal aorta. Eighteen patients had previous aortic valve replacement (11 as a component of a composite graft, 6 separate from graft replacement of the aorta, and I as an aortic root allograft). Seven patients had previous coronary artery bypass grafting procedures and 2 had mitral valve replacement. Four patients had undergone resection and graft replacement of the infrarenal abdominal aorta. The clinical characteristics of the patients are described in Appendix 1. The etiology of the aortic disease was as follows: degenerative, 24; aortic dissection, 21 (acute, 3; chronic, 18); coarctation, 4; (primary repair with aneurysm, 1, false aneurysm, 2, recurrent [postrepair] stenosis, 1); false aneurysm, 1; and malfunction of ringed intraluminal prosthesis, 1. Degenerative disease with aneurysm formation was the most common indication for operation. The extent of aortic disease is shown in Table 1. The decision to employ hypothermic cardiopulmonary bypass and circulatory arrest was made in 8 patients because the location, extent, and severity of the aortic disease precluded the safe placement of clamps on the aortic arch or the proximal descending thoracic aorta. In the remaining 43 patients, it was used because the aortic disease requiring graft replacement involved most or all of the entire descending thoracic aorta (11 patients) or the thoracoabdominal aorta (32 patients), with or without involvement of the distal aortic arch, and the risk of spinal cord ischemic injury was judged to be increased. Using the classification of Crawford and associates [15] 13 of the 32 thoracoabdominal aneurysms were type I, 13 were type II, and 6 were type III (see Table 1). During the study interval, patients with aortic disease confined to localized segments of the descending thoracic aorta (usually in the upper or middle one third) and those with Crawford type IV aneurysms (those confined to the infradiaphragmatic aorta) were treated using alternative techniques. Computed tomographic scanning and aortography were performed preoperatively in all patients. During angiography, no attempt was made to identify intercostal or lumbar arteries that supplied the spinal cord. If noninvasive studies suggested moderate or severe myocardial ischemia, coronary arteriography also was performed preoperatively. Two patients had prophylactic coronary artery bypass grafting and I patient had percutaneous transluminal coronary angioplasty performed in the 3 months prior to the procedure on the thoracic aorta. Operative Technique The general technique employed has been described previously [16]. With increasing experience, a number of modifications have been introduced. Anesthesia was induced with midazolam and fentanyl and maintained with fentanyl and isoflurane. High-dose aprotinin was administered according to a standard protocol in an attempt to reduce blood loss in 8 patients [17]. In 13 patients, a thermistor probe (Model 511; Yellow Springs Instrument Co, Inc, Yellow Springs, OH) was introduced into the intrathecal space through the L3-L4 interspace for continuous measurement of temperature in the spinal canal. After placement of a right radial artery cannula, and pulmonary artery catheter, and a doublequmen endotracheal tube, the descending thoracic aorta was exposed through a posterolateral thoracotomy incision through the bed of the resected fourth or fifth rib. When necessary, the incision was extended obliquely across the costal margin to the midline of the abdomen below the umbilicus. The diaphragm was incised radially or circumferentially. To reduce operative time, cardiopulmonary bypass was established immediately after the chest was entered. Methylprednisolone (7 mg/kg) was administered during this time. The left common femoral artery and vein were exposed through a vertical incision and a 28F or 32F long cannula was inserted and positioned in the right atrium. When a cannula of this size can be inserted into the atrium, flows of greater than 2. to 2.2 L min ~ m 2 can be achieved and cannulation of the pulmonary artery is not necessary. The femoral artery was cannulated with a 2F or 22F cannula. During the period of cooling, the abdominal portion of the incision was completed and the abdominal organs and the kidney were retracted medially after the peritoneum was incised in the left gutter. Thiopental (1 to 15 mg/kg) was given in divided doses during cooling. The left lung was collapsed and gently retracted anteriorly to minimize manipulation and injury. After the heart fibrillated, a small incision was made in the pericardium and a sump-tipped venting catheter was placed in the left ventricle for decompression. The ascending aorta is no longer clamped, and cardioplegic solution was not administered to any patient. The aorta distal to the diseased segment was isolated circumferentiauy. The remainder of the aorta was not dissected to minimize bleeding. Cooling was continued until the nasopharyn-

3 Ann Thorac Surg KOUCHOUKOS ET AL ;6:67-77 HCA FOR THORACOABDOMINAL AORTA OPERATIONS geal temperature reached 12 to 14 C and the bladder or rectal temperature reached 15 to 19 C. There was good correlation between the spinal cord temperature and the nasopharyngeal temperature during cooling and during the period of circulatory arrest. Electroencephalographic monitoring was used and electrical quiescence always was achieved at these temperatures. Monitoring of somatosensory or motor evoked potentials was not used. When required, circulatory arrest was established after the patient was placed in the head-down position, and 1, to 1,5 ml of blood was drained from the patient into the reservoir of the pump-oxygenator system. The venting catheter in the left ventricle was occluded during this time to prevent suction of air into the heart. For operations confined to the distal aortic arch and proximal descending thoracic aorta, resection and graft replacement were performed during a single period of hypothermic circulatory arrest. Clamps no longer are placed on the aortic arch or proximal aorta. The aorta distal to the involved segment was clamped or occluded intraluminally with a balloon catheter to minimize blood loss. As the final anastomosis was being completed, arterial perfusion was reestablished through the femoral artery. This maneuver assisted in the filling of the opened aorta and enhanced the evacuation of air. After completion of the final anastomosis, air was evacuated from the graft with an 18-gauge needle and cardiopulmonary bypass and rewarming were initiated. Cardiopulmonary bypass was discontinued when the rectal/bladder temperature reached 35 C. For procedures that required resection of all or the distal two thirds of the descending thoracic aorta and the abdominal aorta, circulatory arrest was established, the distal aorta was occluded, and the proximal aorta was transected at the appropriate level. After completion of the anastomosis of the Dacron graft to the proximal aorta, a metal-tipped aortic perfusion cannula or an 8-mm collagen-impregnated Dacron graft (Hemashield, Meadox Medicals, Inc, Oakland, NJ) connected to a second arterial line, oxygenator, and roller pump from the extracorporeal circuit was attached to the aortic graft. With the head of the patient in the dependent position, blood was infused into this line to remove air from the aortic arch. Flow through the lower perfusion circuit also was initiated to assist in the removal of air. The aortic graft then was occluded with a clamp distal to the proximal arterial line and flow into the upper aorta was established. Thirty-five percent of the total arterial flow was directed through the proximal arterial line and 65% through the distal line. The temperature of the perfusate was adjusted to maintain the rectal/bladder temperature less than 2 C (range, 15 to 19 C) and total flow was maintained at 1. to 1.5 L- min 1. m 2. During the period of hypothermic low flow, the anastomoses between the aortic graft and aortic tissue surrounding the lower intercostal, lumbar, visceral, and renal arteries were completed. An attempt was made to attach all patent intercostal and lumbar arteries that were below the level of the sixth intercostal space to the aortic graft. Intercostal and lumbar arteries were attached to the graft in 36 of the 43 Table 2. Cardiopulmonary Bypass Data Variable Mean Range Time (min) Cardiopulmonary bypass Cooling Circulatory arrest a Low-flow hypothermic bypass b Hypothermic ventricular fibrillation Rewarming Temperature ( C) Lowest nasopharyngeal Lowest rectal/bladder Lowest intrathecal space (n = 13) a Circulatory arrest not used in 4 patients, bypass not used in 18 patients. b Low-flow hypothermic patients who had replacement of the distal descending thoracic or thoracoabdominal aorta. Whenever possible, the proximal aortic clamp was repositioned below the intercostal artery-to-graft anastomoses before the anastomoses were performed to the visceral and renal arteries and to the distal aorta, and rewarming was initiated at that time to minimize the duration of cardiopulmonary bypass. If this was not possible, rewarming to a rectal/ bladder temperature of 35 C was not initiated until all anastomoses had been completed. The extent of aortic replacement in the 51 patients is shown in Table 1. Variables related to cardiopulmonary bypass are shown in Table 2. During the period of cooling, the heart spontaneously fibrillated when the nasopharyngeal temperature reached 28 to 3 C and became quiescent with further cooling. During rewarming, spontaneous defibrillation occurred in most patients when the nasopharyngeal temperature reached 26 to 28 C. The left ventricular venting catheter then was removed, cardiopulmonary bypass was discontinued, and the cannulas were removed. In the last 28 patients collagen-impregnated woven Dacron grafts (Hemashield) were used to minimize intraoperative blood loss. Statistical Methods The variables examined as potential risk factors for various events are presented in Appendix 1. Bivariable contingency tables of discrete variables, logistic transformation and trend analysis of ordinal variables, and t testing of continuous variables were used to explore the relation of an organized set of potential risk factors to each binary (no/yes) outcome variable. Thereafter, groups of related variables were entered sequentially into multivariable logistic regression analyses, beginning with patient variables, then procedural, support, and experience variables [18]. A p value of.1 was required for retention of a variable in the final model. Analysis of the continuous variables for blood loss and blood product use employed multiple linear regression on the logarithmic transformed variable. This was necessary because of the strong rightward skewness of the data.

4 7 KOUCHOUKOS ET AL Ann Thorac Surg HCA FOR THORACOABDOMINAL AORTA OPERATIONS 1995;6:67-77 Table 3. 3-Day Mortality. of 3-Day Mortality Extent of Disease Patients. % CI Arch and descending 19 % %-1% thoracic aorta Thoracoabdominal Crawford I % 1%-24% Crawford II % 1%-41% Crawford III % 2%-46% Total % 6%-16% p Value (2).17 ci confidence interval. The distribution of the interval between operation and death was estimated independently by a nonparametric life table method [19] and by a parametric method [2]. Risk factors for death were identified in the parametric domain by the hazard function regression methodology. For all analyses, a p value of.5 or less was considered significant. Results Mortality The 3-day mortality was 9.8% (5 patients) (Table 3). Three patients died intraoperatively of myocardial failure. A 77-year-old man with an acute type B (Crawford type II) dissection had severe coronary artery disease in the anterior descending and circumflex systems, which was documented by preoperative coronary angiography. Intractable hypotension and ventricular tachycardia developed. In a 76-year-old man with a ruptured degenerative Crawford type II thoracoabdominal aneurysm, hypotension and bradycardia developed after discontinuation of cardiopulmonary bypass. He had undergone coronary artery bypass grafting 13 years previously. A 47-year-old man with Marfan's syndrome died of intractable ventricular arrhythmias after resection of a Crawford type III aneurysm associated with chronic aortic dissection. He had undergone two previous operations on the ascending aorta and aortic arch, one on the proximal descending thoracic aorta and one on the infrarenal abdominal aorta. A 62-year-old man with a ruptured degenerative Crawford type II aneurysm died on the first postoperative day of low cardiac output associated with a diffuse intravascular coagulopathy. He had undergone previous resection of the ascending aorta and aortic arch and of the infrarenal abdominal aorta. He also had a previous left pleurodesis for repeated pneumothoraces. He received aprotinin intraoperatively. The fifth patient, a 75-year-old woman with a Crawford type II degenerative aneurysm, died 3 days postoperatively of low cardiac output associated with diffuse intravascular coagulopathy. This patient also received aprotinin intraoperatively. Postmortem examination demonstrated multiple foci of acute myocardial infarction and renal infarction associated with extensive microvascular thrombosis [21]. Using multivariable analysis, none of the variables examined was an incremental risk factor for 3-day mortality. The p value (Fisher) for Crawford type II and III aneurysms versus all other patients was.6. The hospital mortality was 11.8% (6 patients). In the sixth patient, a 74-year-old woman with a Crawford type II degenerative aneurysm, renal and respiratory failure developed postoperatively. She received aprotinin intraoperatively, and postmortem examination demonstrated microvascular thrombosis in the heart and kidneys [18]. There have been eight late deaths. patient died of aortic disease. The actuarial survival rates at 6 months and 2 years were 79% and 76%, respectively. When the hazard function was separated into early and constant phases, aortic rupture or an acutely dissected aneurysm and the presence of a Crawford type II or III aneurysm were significant incremental risk factors for death in the early phase, and older age at operation and the presence of a Crawford type II or III aneurysm were significant risk factors in the constant phase. Morbidity BLEEDING AND TRANSFUSION REQUIREMENTS. Reoperation for bleeding was required in two patients. The volumes of red blood cells, fresh frozen plasma and platelets transfused intraoperatively and postoperatively are shown in Table 4. Three patients received cryoprecipitate. Four patients (7.8%) received no blood products intraoperatively and 17 (35%) of the 48 patients who survived the operation required no transfusions postoperatively. Using multiple linear regression analysis, the presence of a Crawford type II thoracoabdominal aortic aneurysm and a longer elapsed time of cardiopulmonary bypass were significant (p ~.5) incremental risk factors for the intraoperative and early postoperative use of platelets and fresh frozen plasma. NEUROLOGIC COMPLICATIONS. Stroke with resulting hemiparesis occurred in 2 patients (4.3% of the 46 3-day survivors). The aorta proximal to the thoracic aortic disease was clamped in both patients during the operative procedure, and both had severe atherosclerotic dis- Table 4. Blood Component Transfusion Requirements Blood Component. of Units Percentile. of Patients Mean Min 25% 5% 75% Max Intraoperative Red blood cells Fresh-frozen plasma Platelets Postoperative Red blood cells Fresh-frozen plasma Platelets Max = maximum; Min - minimum.

5 Ann Thorac Surg KOUCHOUKOS ET AL ;6:67-77 HCA FOR THORACOABDOMINAL AORTA OPERATIONS ease in the aortic segment that was clamped. Paraplegia occurred in 2 and paraparesis in 1 of the 46 3-day survivors (6.5%). The neurologic deficit was apparent immediately upon recovery from anesthesia. Paraparesis occurred in a patient who had resection of the distal aortic arch and the proximal two-thirds of the descending thoracic aorta. The intercostal arteries between the T-4 and T-8 levels were occluded by the atherosclerotic process. The pair of intercostal arteries at the T-8 level and the left intercostal artery at the T-9 level were sacrificed. The aorta surrounding these arteries was severely atherosclerotic and could not be sutured to the aortic graft. All intercostal arteries distal to this level were preserved. This patient is ambu!atory with residual weakness of the flexor muscles of the hip. Among the 27 operative survivors with thoracoabdominal aortic disease, paraplegia occurred in I (8.3%) of the 12 with Crawford type I, in none of the 1 with Crawford type IL and in 1 (2%) of the 5 with Crawford type III aneurysms. Paraplegia occurred in none of the 13 patients with aortic dissection and in 2 of the 14 patients with degenerative aneurysms. In I patient with a Crawford type I aneurysm in whom paraplegia developed, two pairs of intercostal arteries above T-6 were sacrificed. The left intercostal artery at the T-9 level was implanted into the graft and the small right intercostal artery at this level, which was surrounded by severe aortic atherosclerosis, was sacrificed. other intercostal or lumbar arteries were patent. In the other patient, who had a Crawford type III aneurysm, a single patent intercostal artery at the T-1 level was implanted into the graft. A single patent lumbar artery located several centimeters below the renal arteries was sacrificed. other intercostal or lumbar arteries were patent. In a multivariable analysis, no incremental risk factors for the development of spinal cord ischemic injury were identified. The relation between the duration of spinal cord ischemia, defined as the time between onset of circulatory arrest or aortic clamping and the establishment of flow to the intercostal arteries below the T-6 interspace, and the incidence of paraplegia or paraparesis is shown in Figure 1. RENAL DYSFUNCTION. Renal failure requiring dialysis occurred in I (2.2%) of the 46 3-day survivors. This patient received aprotinin. Renal dysfunction, defined as an elevation of the serum creatinine level in the first postoperative week to a level of 1.5 times the preoperative level or higher, occurred in 5 additional patients. Two of these patients received aprotinin. PULMONARY DYSFUNCTION. Thirty-two of the operative survivors were extubated within 24 hours and 2 within 48 hours of operation. Prolonged (>48 hours) mechanical ventilation was required in 14 (29%) of the operative survivors. Five patients (1%) required a tracheostomy. Three patients received antibiotic therapy for pneumonitis. In a multivariable analysis employing all variables, older age at operation and use of aprotinin were signifi- 5 1 "~) ::~ I5.Q 1 5 O. Duration % (Min) % / 45 7.% 6 6.6% % % "-.. P=.9.. -" Duration (Min) Of Spinal Cord Ischemia Fig 1. Risk of paraplegia or paraparesis according to the duration of spinal cord ischemia. The dashed lines represent the 7% confidence limits. The p value relates to the relation between ischemic time and the probability of paraplegia or paraparesis. cant incremental risk factors for prolonged mechanical ventilation. CARDIAC COMPLICATIONS. Of the 46 3-day survivors, 2 patients required inotropic support for more than 48 hours for low cardiac output. The median length of stay in the intensive care unit was 4 days (range, 1 to 48 days). Fourteen patients remained in the intensive care unit for less than 48 hours. The median postoperative length of stay was 12 days (range, 6 to 69 days). Nineteen patients were discharged by the 1th postoperative day. Comment When used in conjunction with hypothermic cardiopulmonary bypass for management of disease of the descending thoracic aorta, circulatory arrest offers certain advantages over other techniques such as simple aortic clamping or use of distal perfusion with atriofemoral or femoral-femoral bypass. These include minimal dissection of the aorta, elimination of the need for proximal aortic clamping, access to the proximal aortic arch and ascending aorta, and a bloodless field. The profound hypothermia provides effective protection of the brain, spinal cord, kidneys, and the abdominal viscera. In certain circumstances, this technique may be the only suitable option for management of aortic disease that involves the distal aortic arch and the adjacent descending thoracic aorta [1-8]. In the 8 patients in our series in this category, there was no mortality or serious morbidity. The rationale for use of hypothermic cardiopulmonary bypass with or without a period of circulatory arrest for extensive disease involving the descending thoracic and thoracoabdominal aorta is to increase the tolerable duration of spinal cord ischemia so that attachment of intercostal and lumbar arteries and reestablishment of aortic flow can be completed before the onset of permanent cord injury. In studies in baboons and pigs, we

6 72 KOUCHOUKOS ET AL Ann Thorac Surg HCA FOR THORACOABDOMINAL AORTA OPERATIONS 1995;6:67-77 demonstrated that hypothermia (15 C) induced and maintained by cardiopulmonary bypass in a model where resection of the thoracoabdominal aorta was simulated and the aorta was occluded proximally and distally for 6 minutes provided superior spinal cord protection when compared with normothermic bypass [13, 14]. Before the technique is applied widely, however, the safety as well as the effectiveness of hypothermic cardiopulmonary bypass should be demonstrated. Comparison of mortality and morbidity rates with those of other currently used techniques is also necessary. In our study, no incremental risk factors for 3-day mortality were identified. Aprotinin was used in 3 of the 6 patients who died in the hospital. The deaths of 2 of these patients could be attributed directly to use of this drug, and we no longer use it in patients who require hypothermic cardiopulmonary bypass and circulatory arrest [17, 21]. When all deaths were analyzed, acute dissection or rupture of the aneurysm, the presence of a Crawford type II or type III aneurysm and increasing age were incremental risk factors for death. These variables have been found to be significant incremental risk factors for early and late death in patients with thoracic and thoracoabdominal aortic disease in whom other techniques (simple proximal aortic clamping and distal perfusion) have been employed [15, 22, 23]. When compared with series of patients having resection of extensive thoracic or thoracoabdominal aneurysrns in which alternative techniques were employed (simple aortic clamping, atriofemoral bypass, or femoralfemoral bypass) the frequency of reoperation for bleeding and the perioperative transfusion requirements in our patients did not differ from those reported with these techniques [15, 22, 24-29]. The ability to return virtually all shed blood into the pump-oxygenator system with full cardiopulmonary bypass and heparinization reduces the need for salvage of red cells from cell-saving devices. Use of membrane oxygenators and impermeable grafts may have reduced the requirements for platelets, fresh-frozen plasma, and cryoprecipitate. In clinical studies, the largest experience with extensive resections of the descending thoracic and thoracoabdominal aorta has involved the use of simple aortic clamping without distal perfusion [15, 22-25]. These studies all have demonstrated a clear correlation between the extent of aorta that is resected or the time required to effect this resection and the incidence of spinal cord ischemic injury. For aneurysms involving most or all of the descending thoracic aorta, the incidence of paraplegia or paraparesis has ranged from 6% to 13% [23, 25]. In subgroups of patients with prolonged aortic clamp times (~3 minutes), the incidence was 15% or greater [23]. For aneurysms of the thoracoabdominal aorta, the incidence of ischemic injury has ranged from 5% to 21% [15, 22, 25, 3-34]. The highest incidence occurred with type II aneurysms [15, 32-34], and the incidence was substantially higher for patients with aortic dissection than for those with degenerative aneurysms [15, 33]. As an example, for patients in the study by Crawford and associates [15] with type II aneurysm and dissection, the incidence of paraplegia or paraparesis was 4% (23 of 57 patients). In our series, the extent of aorta replaced or the duration of spinal cord ischemia did not correlate with the incidence of spinal cord ischemic injury (see Fig 1, Appendix 1). Furthermore, paraplegia or paraparesis developed in only I (4.5%) of the 22 3-day survivors with Crawford type I or type II thoracoabdominal aortic aneurysms. Although the number of cases is small, these findings are in marked contrast to those observed with simple aortic clamping. In a comparison using Fisher's exact test of our series with the results of Crawford and associates [15], Golden and co-workers [32], and Cox and colleagues [33] the incidence of paraplegia or paraparesis was significantly lower in our series for patients with Crawford type I! aneurysm (p =.2,.2, and.5, respectively). ne of the 6 patients in our series with extensive thoracic aneurysms and dissection and none of the 12 patients with Crawford type I, II, or III thoracoabdominal aneurysms and dissection had development of paraplegia or paraparesis. In two other series employing hypothermic circulatory arrest for extensive thoracic and thoracoabdominal aneurysms (a total of 29 patients), where the type of aortic disease was documented, paraplegia and paraparesis occurred in only 2 patients who had atherosclerotic aneurysms [4, 35]. spinal cord ischemic injury occurred in any of the patients with aortic dissection. Patent intercostal and lumbar arteries were reattached in these patients whenever possible. Taken together, these findings suggest that the technique of profound hypothermic cardiopulmonary bypass may offer substantial protection of the spinal cord during these extensive procedures, particularly in patients with aortic dissection, and that it can safely extend the duration of spinal cord ischemia for up to 9 to 1 minutes. The incidence of renal failure requiring dialysis in our series is lower than that reported with simple aortic clamping for extensive descending thoracic and thoracoabdominal aneurysms [15, 3-33] and with techniques employing normothermic distal perfusion for thoracoabdominal aneurysms [31, 36]. Use of distal perfusion with moderate hypothermia also has been associated with a low incidence of renal failure [27, 28]. These observations suggest that hypothermia has an important protective effect on renal function during these extensive procedures. Although a standard methodology has not been used to report pulmonary complications after extensive resections of the thoracic aorta, the incidence of major pulmonary complications (prolonged mechanical ventilation, need for tracheostomy) in our series did not differ appreciably from that reported in patients in whom simple aortic clamping was used [22, 25, 3, 31, 33]. Cardiac complications occurred infrequently in our series and in the reports by Caramutti and colleagues [3] and Kieffer and co-workers [4[, where elective hypothermic cardiopulmonary bypass was used, indicating that hypothermic fibrillation and venting of the heart, when appropriate, provide adequate myocardial protection. We believe that

7 Ann Thorac Surg KOUCHOUKOS ET AL ;6:67-77 HCA FOR THORACOABDOMINAL AORTA OPERATIONS internal or external occlusion of the ascending aorta and infusion of cardioplegic solution is not necessary for myocardial protection during these procedures, and that occlusion of the ascending aorta increases the potential for embolization of atheromatous debris into the brachiocephalic arteries. Our experience with hypothermic cardiopulmonary bypass and circulatory arrest confirms the safety and efficacy of the technique for operations that required resection of the aortic arch through a lateral approach. Our results suggest that for patients with extensive descending thoracic and Crawford type II thoracoabdominal aneurysms, the incidence of spinal cord ischemic injury may be less than that observed with simple aortic clamping techniques. The technique is associated with an incidence of renal failure that appears to be lower than that observed with simple aortic clamping and with normothermic distal perfusion techniques. The incidence of other major complications does not appear to be higher than those observed with other techniques. The optimal level of hypothermia that will provide maximal protection of the spinal cord and kidneys is undetermined. It is possible that a lesser degree of hypothermia may be equally protective and will be associated with less mortality and morbidity than the technique we have reported. However, hypothermia and circulatory arrest will be required for aortic disease that necessitates resection of the aortic arch. Techniques involving the use of distal perfusion with normothermia or mild hypothermia also have been associated with a lower incidence of spinal cord ischernic injury than the simple aortic clamping technique [26-28, 36]. However, these techniques require the use of sequential aortic clamping to minimize the duration of spinal cord ischemia, and this will not be possible in all patients [26]. Because the etiology of spinal cord ischemic injury associated with operations on the descending thoracic and thoracoabdominal aorta is probably multifactorial, it is unlikely that any single intervention such as the use of profound hypothermia will eliminate the paraplegia and paraparesis that occurs after these operations. However, hypothermia appears to increase substantially the tolerable duration of spinal cord ischemia. We believe that continued evaluation and use of the technique is indicated. Addendum Since submission of the manuscript for publication, 7 additional patients (4 with Crawford type II thoracoabdominal aneurysms and 3 with extensive descending thoracic aneurysms) have had resection of the aneurysms using the described technique with no early mortality, paraplegia or paraparesis, or renal failure. Supported by a grant from the Shoenberg Foundation. We are grateful to Dr Eugene H. Blackstone for assistance with the statistical analyses, to Dr Thomas H. Wareing who operated on 3 of the patients, and to Ms Mary Sue Williams for preparation of the manuscript. William Klausing, JoAnn Sabelli, Timothy Burns, Patrick O'Donnell, and Deanne Barlas-Collins provided superb cardiopulmonary perfusion. References 1. Borst HG, Schaudig A, Rudolph W. Arteriovenous fistula of the aortic arch: repair during deep hypothermia and circulatory arrest. J Thorac Cardiovasc Surg 1964;48: Crawford ES, Coselli JS, Sail JH. Partial cardiopulmonary bypass, hypothermic circulatory arrest, and posterolateral exposure for thoracic aortic aneurysm operation. J Thorac Cardiovasc Surg 1987;94: Caramutti VM, Dantur JR, Favaloro MR, et al. Deep hypothermia and circulatory arrest as an elective technique in the treatment of type B dissecting aneurysm of the aorta. J Card Surg 1989;4: Kieffer E, Koskas F, Walden R, et al. Hypothermic circulatory arrest for thoracic aneurysmectomy through left-sided thoractomy. J Vasc Surg 1994;19: Mahfood S, Qazi A, Garcia J, et al. Management of aortic arch aneurysm using profound hypothermia and circulatory arrest. Ann Thorac Surg 1985;39: Westaby S. Hypothermic thoracic and thoracoabdominal aneurysm operation: a central cannulation technique. Ann Thorac Surg 1992;54: Szentpetery S, Crisler C, Grinnan GLB. Deep hypothermic arrest and left thoracotomy for repair of difficult thoracic aneurysms. Ann Thorac Surg 1993;55: Massimo CG, Presenti LF, Favi PP, et al. Simultaneous total aortic replacement from valve to bifurcation: experience with 21 cases. Ann Thorac Surg 1993;56: Pontius RG, Brockman L, Hardy EG, et al. The use of hypothermia in the prevention of paraplegia following temporary aortic occlusion: experimental observations. Surgery 1954;36: Owens JC, Prevedel AE, Swan H. Prolonged experimental occlusion of thoracic aorta during hypothermia. Arch Surg 1955;7: Robertson CS, Foltz R, Grossman RG, et al. Protection against experimental ischemic spinal cord injury. J Neurosurg 1986;64: Colon R, Frazier OH, Cooley DA, et al. Hypothermic regional perfusion for protection of the spinal cord during periods of ischemia. Ann Thorac Surg 1987;43: Rokkas CK, Sundaresan S, Shuman TA, et al. Profound systemic hypothermia protects the spinal cord in a primate model of spinal cord ischemia. J Thorac Cardiovasc Surg 1993;16: Rokkas CK, Cronin CS, Nitta T, et al. Profound systemic hypothermia inhibits the release of neurotransmitter amino acids during spinal cord ischemia. J Thorac Cardiovasc Surg (in press). 15. Crawford ES, Crawford JL, Sail, HJ, et al. Thoracoabdominal aortic aneurysms: preoperative and intraoperative factors determining immediate and long-term results of operations in 65 patients. J Vasc Surg 1986;3: Kouchoukos NT, Wareing TH, Izumoto H, et al. Elective hypothermic cardiopulmonary bypass and circulatory arrest for spinal cord protection during operations on the thoracoabdominal aorta. J Thorac Cardiovasc Surg 199;99: Sundt TM, Kouchoukos NT, Saffitz JE, et al. Renal dysfunction and intravascular coagulation with aprotinin and hypothermic circulatory arrest. Ann Thorac Surg 1993;55: Walker SH, Duncan DB. Estimation of the probability of an event as a function of several independent variables. Biometrika 1967;54: Kaplan EL, Meier P. nparametric estimation from incomplete observations. J Am Stat Assoc 1958;53: Blackstone EH, Naftel DC, Turner ME Jr. The decomposition of time-varying hazard into phases, each incorporating a separate stream of concomitant information. J Am Stat Assoc 1986;81: Saffitz JE, Stahl DJ, Sundt TM, et al. Disseminated intravascular coagulation after administration of aprotinin in combination with deep hypothermic circulatory arrest. Am J Cardiol 1993;72:18-82.

8 74 KOUCHOUKOS ET AL Ann Thorac Surg HCA FOR THORACOABDOMINAL AORTA OPERATIONS 1995;6: Schepens MAAM, Defauw JJAM, Hammerlijnck RPHM, DeGeest R, Vermeulen FEE. Surgical treatment of thoracoabdominal aortic aneurysms by simple crossclamping, J Thorac Cardiovasc Surg 1994;17: Lawrie GM, Earle N, DeBakey ME. Evolution of surgical techniques for aneurysms of the descending thoracic aorta: twenty-nine years" experience with 659 patients. J Card Surg 1994; 9: Svensson LG, Crawford ES, Hess KR, et al. Experience with 159 patients undergoing thoracoabdominal aortic operations. J Vasc Surg 1993;17: Svensson LG, Crawford ES, Hess KR, et al. Variables predictive of outcome in 832 patients undergoing repairs of the descending thoracic aorta. Chest 1993;14: Coselli JS. Thoracoabdominal aortic aneurysms: experience with 372 patients. J Card Surg 1994;9: 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: Frank SM, Parker SD, Rock P, et al. Moderate hypothermia with partial bypass and segmental sequential repair for thoracoabdominal aortic aneurysm. J Vasc Surg 1994;19: Von Segesser LK, Killer I, Jenni R, et al. Improved distal circulatory support for repair of descending thoracic aortic aneurysms. Ann Thorac Surg 1993;56: Hollier LH, Symmonds JB, Pairolero PC, et al. Thoracoabdominal aortic aneurysm repair. Arch Surg 1988;123: Cambria RP, Brewster DC, Moncure AC, et al. Recent experience with thoracoabdominal aneurysm repair. Arch Surg 1989;124: Golden MA, Donaldson MC, Whittemore AD, et al. Evolving experience with thoracoabdominal aortic aneurysm repair at a single institution. J Vasc Surg 1991;13: Cox GS, O'Hara PJ, Hertzer NR, et al. Thoracoabdominal aneurysm repair: a representative experience. J Vasc Surg 1992;15: Archer CW, Wynn MM, Hoch JR, et al. Combined use of cerebral spinal fluid drainage and naloxone reduces the risk of paraplegia in thoracoabdominal aneurysm repair. J Vasc Surg 1994;19: Grabenw/Sger, Ehrlich M, Simon P, et al. Thoracoabdominal aneurysm repair: spinal cord protection using profound hypothermia and circulatory arrest. J Card Surg 1994;9: Sail HJ, Bartoli S, Hess KR, et al. Neurologic deficit in patients at high risk with thoracoabdominal aortic aneurysms: the role of cerebral spinal fluid drainage and distal aortic perfusion. J Vasc Surg 1994;2: Appendix 1. Univariate Analysis of Patient-Related Variables 3-Day. of. of Mortality 3-Day Variable Patients. % p Value Survivors Age (y) Sex.2 Male Female Aortic disease.7 Degenerative Dissection Other 6 Marfan's syndrome Dissection.3 Acute Chronic ne Extent of aortic disease.4 Distal arch and proximal DTA 8 Arch and entire DTA 7 Most or all of DTA 4 Thoracoabdominal aorta Crawford I Crawford II Crawford III Asymptomatic Paraparesis/ Paraplegia. % p Value

9 Ann Thorac Surg KOUCHOUKOS ET AL ;6:67-77 HCA FOR THORACOABDOMINAL AORTA OPERATIONS Appendix 1. Continued Variable. of Patients 3-Day Mortality. % p Value. of Paraparesis/ 3-Day Paraplegia Survivors. % p Value Rupture or acute dissection Hypertension Chronic pulmonary disease Renal dysfunction (creatinine mg/dl) Atherosclerotic heart disease Diabetes Transient ischemic attack or stroke Previous proximal aortic operation Previous distal aortic operation Previous aortic valve replacement Previous CABG or PTCA Year of operation Duration of cardiopulmonary bypass (min) >21 Duration of circulatory arrest (min) ne used Duration of spinal cord ischemia (min) < >

10 76 KOUCHOUKOS ET AL Ann Thorac Surg HCA FOR THORACOABDOMINAL AORTA OPERATIONS 1995;6:67-77 Appendix 1. Continued Paraparesis/ 3-Day. of Paraplegia. of Mortality 3-Day Variable Patients. % p Value Survivors. % p Value Implantation of intercostal arteries below.6.4 T-6 ~ Use of aprotinin Reoperation for bleeding Postoperative stroke Postoperative paraplegia/paraparesis (min) b Postoperative cardiac complications Mechanical ventilation >48 hours Renal failure Postoperative creatinine elevated (->1.5.7 baseline) b Gastrointestinal bleeding Sepsis Major wound complication Forty-three patients with descending thoracic and thoracoabdominal disease, b Thirty-day survivors only. CABG = coronary artery bypass grafting; DTA = descending thoracic aorta; PTCA = percutaneous transluminal coronary angioplasty. DISCUSSION DR JOSEPH S. COSELLI (Houston, TX): Once again, Dr Kouchoukos' presentations and publications remain an ideal we all strive to achieve. I thank Dr Kouchoukos for the opportunity to review the manuscript before the meeting. Doctor Kouchoukos and associates report on an experience with truly excellent results in a selective cohort of patients treated with an evolving strategy over a 9-year period. In this group of patients, however, I believe we need to make two clear distinctions, which Dr Kouchoukos has already pointed out. One is the use of deep hypothermia and circulatory arrest in the left chest for aneurysms where, for technical or anatomic reasons, it is simply not possible to place a clamp for proximal control. These would include large aneurysms of the proximal descending thoracic aorta where no appropriate neck for proximal clamping is anatomically available. The second is the extension of this modality for the purpose of spinal cord and visceral protection, particularly renal, in patients in whom cross-clamping proximally is possible. In my experience, it has been used relatively infrequently, for a total of 28

11 Ann Thorac Surg KOUCHOUKOS ET AL ;6:67-77 HCA FOR THORACOABDOMINAL AORTA OPERATIONS patients in whom circulatory arrest was used through the left chest out of 717 patients with either descending thoracic or thoracoabdominal aortic aneurysms. In these 28 patients, the development of paraplegia or a paraparesis was 7.14% and the early mortality was 3 patients (1.7%), comparing favorably with other techniques such as atriofemoral bypass or simple crossclamping. In 17 patients treated with atriofemoral bypass, the overall neurologic deficit rate, including paraplegia and paraparesis was 3.74% and mortality was 4.67%; for simple crossclamping the rates were 5.42% and 4.55%, respectively. Although the mortality is increased numerically over the latter two techniques, no statistical difference regarding the incidence of neurologic deficits has been identified. A problem still remains regarding the postoperative development of neurologic deficits in the Crawford type II patients (thoracoabdominal aortic aneurysms). In 137 Crawford type II patients out of 498 consecutive patients in a personal series, the mortality was 5.8% and the paraplegia and paraparesis rate combined was 1.9%. It makes me wonder whether in Kouchoukos and associates" series we are not trading off a certain amount of mortality for a decrease in the incidence of neurologic deficits. Doctor Kouchoukos, why did you use patients from other institutions for comparison with your own data, rather than using patients from your own institution? Additionally, were the 51 patients in your series consecutive; if not, how many other operations in patients with descending and thoracoabdominal aneurysms were carried out during the same time period of the study? The results in those patients, I think, would be informative. DR ALLAN WOLPOWITZ (Hollywood, FL): I have two technical questions. If you have to reperfuse the descending thoracic aorta at the same time that you are perfusing the femoral, the two perfused systems require different techniques, necessitating two oxygenators and two arterial heads. Is this a standard set-up? With all patients with elevated creatinine levels we routinely use continuous arteriovenous hemofiltration, but it appears as if we dialyze aprotinin out. This might be the reason that there is disseminated intravascular coagulation postoperatively despite use of aprotinin. DR TIRONE E. DAVID (Toronto, Ont, Canada): I enjoyed your presentation very much, Dr Kouchoukos. I have been very reluctant to place patients on cardiopulmonary bypass through femoral cannulation if they have extensive atherosclerotic aneurysm because of the possibility of embolization to the brain once the flow in the thoracic and abdominal aorta is reversed. Would you care to comment about that? DR KOUCHOUKOS: Doctor Coselli, the issue of trading a potentially higher early mortality rate for a lower incidence of spinal cord ischemic injury is a very important one. We made the decision to use hypothermic cardiopulmonary bypass and circulatory arrest in patients in whom the risk of spinal cord ischemic injury was judged to be approximately 1% or greater, and this included patients with aneurysms involving most or all of the descending thoracic aorta and patients with Crawford type I, II, and III thoracoabdominal aneurysms. The 3-day mortality rate for the 43 patients in these categories was 11.6%, and the rate of paraplegia or paraparesis was 7%. During the study interval, alternative operative techniques were not used for patients in these categories. For patients with aneurysms confined to the upper or the mid-descending thoracic aorta we used distal perfusion with femoral vein to femoral artery bypass, and for patients with Crawford type 1V thoracoabdominal aneurysms we used simple aortic clamping. Doctor Wolpowitz, we used separate pump-oxygenator systems to perfuse the upper and lower parts of the body. Separate systems are necessary to regulate blood flow precisely to these two areas. We did not routinely use a hemoconcentrator, and thus I cannot comment on whether or not aprotinin was removed from the circulation. Doctor David, I share your concern about the potential for embolization of atheromatous material to the brain if femoral artery perfusion is used in patients with arteriosclerotic aneurysms. The 2 patients in our series in whom embolic strokes developed had atherosclerotic disease in the aorta proximal to the aneurysm and had clamps placed on this segment. We believe this is a more likely cause for embolic stroke and we no longer place clamps on the proximal aorta. Since we abandoned this maneuver we have seen no embolic strokes. The optimal level of hypothermia that will provide maximal protection of the spinal cord and kidneys is undetermined, and it is possible that a lesser degree of hypothermia may be equally protective and will be associated with less mortality and morbidity than the technique we have reported. However, hypothermia and circulatory arrest will be required for aortic disease that necessitates resection of the aortic arch. Techniques involving the use of distal perfusion with normothermia or mild hypothermia, which have been used by Dr Coselli and others, also have been associated with a lower incidence of spinal cord ischemia injury than the simple aortic clamping technique. However, these perfusion techniques require the use of sequential aortic clamping to minimize the duration of spinal cord ischemia, and this will not be possible in all cases. Thus, in situations where aortic clamping is not possible, use of hypothermic circulatory arrest should be considered.