Retrograde Cerebral and Distal Aortic Perfusion During Ascending and Thoracoabdominal Aortic Operations

Transcription

1 Retrograde Cerebral and Distal Aortic Perfusion During Ascending and Thoracoabdominal Aortic Operations Joseph E. Bavaria, MD, Y. Joseph Woo, MD, R. Alan Hall, MD, Jeffrey P. Carpenter, MD, and Timothy J. Gardner, MD Department of Surgery, Hospital of the Universi~ of Pennsylvania, Philadelphia, Pennsylvania Background. Several alternative circulatory management techniques during thoracic aortic reconstruction have been implemented at this institution. This study was performed to assess whether retrograde cerebral perfusion during proximal aortic operations and distal aortic perfusion during thoracoabdominal aortic operations have improved outcomes. Methods. A retrospective review of 156 patients undergoing elective and emergent operations of the thoracic aorta over the past 7 years was performed. Seventy-five patients underwent proximal aortic procedures: 22 with ascending aneurysms, 45 with type A dissections, and 8 with arch reconstructions. Eighty-one patients underwent descending thoracic or thoracoabdominal procedures: 26 with Crawford type I aneurysms, 18 with type II, 8 with type Ill, 8 with type IV, 11 with traumatic transections, and 10 with type B dissections. Outcomes measured were neurologic injury, renal failure, and mortality. Results. For proximal aortic procedures, the stroke rate was 12% using cardiopulmonary bypass and 48% using hypothermic circulatory arrest. The addition of retrograde cerebral perfusion decreased the stroke rate to 0% (p < 0.01) and the mortality rate to 7.1% compared with 37% for hypothermic circulatory arrest (p < 0.05). For thoracic and thoracoabdominal aortic operations, straight cross-clamping resulted in a 27% rate of spinal cord injury and a 24% rate of renal failure, whereas the addition of distal aortic bypass resulted in a statistically significant reduction (p < 0.01) in neurologic injury to 7% and a notable, but not statistically significant, decrease in renal failure to 13%. Distal aortic bypass also reduced the mortality rate from 22% to 7% (p < 0.05). Conclusions. Retrograde cerebral perfusion decreases the stroke rate and mortality rate in proximal aortic operations and distal aortic perfusion decreases the rates of neurologic injury, renal failure, and mortality in thoracoabdominal aortic operations. (Ann Thorac Surg 1995;60:345-53) O ver the past decade, many innovations in circulation management technique during thoracic aortic reconstruction have been developed. For the management of proximal aortic disease and acute type A dissection, the classic technique has entailed aortic cross-clamping at the level of the innominate artery with subsequent repair of the ascending aortic pathology. More recently, authors have advocated an "open" distal anastomosis for reconstruction of acute type A dissection with simultaneous use of hypothermic circulatory arrest (HCA). Additionally, HCA has been used for aneurysmal disease involving the aortic arch as well as extension of aortic dissection tears into the aortic arch [1-7]. Others have advocated the use of selective antegrade cerebral perfusion during open arch operations [8]. Retrograde cerebral perfusion (RCP) originally was described for the management of inadvertent massive air embolism during cardiac operations [9]. It first was used for cerebral protection during aortic arch operations by Ueda and associates Presented at the Forty-first Annual Meeting of the Southern Thoracic Surgical Association, Marco Island, FL, Nov 10-12, Address reprint requests to Dr Bavaria, Department of Surgery, Hospital of the Universi~, of Pennsylvania, 3400 Spruce St, Philadelphia, PA [10]. Retrograde cerebral perfusion currently is being widely investigated in animal models and clinically [11-15]. In fact, some have advocated total body retrograde perfusion during proximal aortic operations [16, 17]. Advances in circulation management for thoracic and thoracoabdominal aortic reconstruction likewise have provided the surgeon many options. These have included straight aortic cross-clamping, cerebrospinal fluid (CSF) drainage, and various pharmacologic manipulations to decrease the incidence of spinal cord paraplegia by increasing spinal cord perfusion pressure and altering the reperfusion injury [18]. More recently many authors have advocated the routine use of distal aortic perfusion via an ~,xtracorporeal circuit to provide additional protection against ischemia in all organ systems distal to the aortic cross-clamp [19-21]. Even more recently, the use of hypothermic circulation management techniques during descending thoracic aortic operation has been advocated [221. This study is a retrospective review comparing the efficacy of several major circulation management strategies used during proximal and distal aortic reconstruction. Techniques used during proximal aortic repair in by The Society of Thoracic Surgeons /95/$ (95)00447-S

2 346 BAVARIA ET AL Ann Thorac Surg CIRCULATION MANAGEMENT IN AORTIC OPERATIONS 1995;60: Table 1. Proximal Aortic Reconstntctive Procedures Procedure No. Ascending aortic aneurysm (total) 22 Straight graft 8 Straight graft with aortic valve 3 resuspension Wheat 4 Composite 7 Emergency (%) 13.7 :, Type A dissection (total) 45 Straight graft 15 Straight graft with aortic valve 17 resuspension Wheat 6 Composite 7 Emergency (%) 100% Arch reconstruction (total) 8 Emergency (%) 37.5% Total 75 Total emergency (%) 68% clude (1) straight cross-clamping of the distal ascending aorta, (2) HCA, and (3) HCA with RCP. Techniques used during distal aortic repair include (1) straight aortic cross-clamping and (2) distal aortic perfusion. Material and Methods Patients A retrospective analysis was undertaken of 156 consecutive patients who underwent elective and emergent operations of the proximal and distal thoracic/thoracoabdominal aorta at the Hospital of the University of Pennsylvania on the Cardiothoracic Surgery Service during the period beginning June 1, 1987, and ending September 30, Patients who died intraoperatively before the interruption of aortic circulation were excluded. Furthermore, patients who underwent simultaneous thoracic procedures such as lung transplantation were excluded as were 4 patients who had combined distal aortic arch and Crawford type I or II aneurysmal repair via "open" proximal anastomosis under HCA and staged reconstruction of the distal aortic segment. Proximal Aortic Reconstruction The proximal aortic group comprised 75 patients: 22 with ascending aneurysms, 45 with type A dissections, and 8 with arch aneurysms (Table 1). Patients were divided further based on the specific operation performed. The proportion of emergent cases, as defined by operation within 12 hours of urgent nonscheduled admission, also was determined. Outcomes were evaluated in terms of clinically evident cerebrovascular accident and 60-day mortality in the proximal aortic group. Data were evaluated by standard 9( 2 statistical analysis. Thoracic/Thoracoabdominal Aortic Reconstruction The distal aortic group consisted of 81 patients: 26 with Crawford type I thoracoabdominal aneurysms, 18 with type II, 8 with type III, 8 with type IV, 11 with traumatic transections, and 10 with type B aortic dissections (Table 2). Outcomes were expressed in rates of spinal neurologic injury (paraplegia/paraparesis), renal failure as defined by an increase in serum creatinine level 2.0 mg]dl above preoperative levels or initiation of dialysis, and 60-day mortality. Techniques CIRCULATION MANAGEMENT: PROXIMAL AORTIC RECON- STRUCTION. By standard median sternotomy incision, the heart and great vessels were exposed. For subsequent cannulation, the left femoral artery was exposed. Three basic types of circulation management strategy were used: (1) Aortic cross-clamping at a level just proximal to the innominate artery with standard cardiopulmonary bypass and moderate hypothermia (28 to 32 C) was used in both emergent and elective operations with isolated aortic root pathology and a normal aorta at the level of the innominate artery, and for the great majority of type A aortic dissections before (2) Standard cardiopulmonary bypass with HCA was used in all patients with type A dissections with aortic arch involvement before 1992 and all primary aortic arch reconstructions before Deep hypothermia during HCA averaged 16 C. (3) Circulation management using circulatory arrest with RCP was used in most patients with type A aortic dissection or aortic arch reconstructions since January The retrograde cerebral perfusion circuit (Fig 1) consisted of standard bicaval cannulation (Research Medical, Midvale, UT) as well as femoral artery cannulation with 17F to 21F (Biomedicus, Minneapolis, MN) wire-wrapped cannulas. A 30F superior vena caval cannula was used intentionally in all patients to allow greater cephalad flow during the retrograde cerebral perfusion phase of the circulatory management. All hearts were vented via the right superior pulmonary vein. Patients were cooled for a Table 2. Distal (Descending) Aortic Reconstructive Procedures Emergency Type No. (%) Thoracic/Crawford type I % aortic aneurysms Crawford type II % thoracoabdominal aneurysms Crawford type III % thoracoabdominal aneurysms Thoracic aortic traumatic % transection Type B dissection 10 20% Crawford type IV thoracoabdominal aneurysms 8 25% Total 81 Total emergency (%) 30.9%

3 Ann Thorac Surg BAVARIA ET AL ;60: CIRCULATION MANAGEMENT IN AORTIC OPERATIONS ~,,~C antegrade cardiopulmonary bypass was started via the side-arm graft. The patient then was rewarmed and the proximal aortic reconstruction was performed during rewarming. This would include placement of the proximal anastomosis, resuspension of an aortic valve during type A dissection, placement of an aortic valve for a "Wheat" procedure, or placement of a composite graft with coronary reimplantation. H Venous Arterial Fig I. Circuit used for retrograde cerebral perfusion. Clamps on the "bridge" are controlled on the table by the surgeon. During retrograde cerebral perfusion clamps are placed on the proximal venous line (A), the distal arterial line (B), and the inferior vena caval cannula (C). Note: Recently, ante~rade perfusion is transferred to a lo-mm side arm after completion of distal anastomosis. minimum of 30 minutes to a nasopharyngeal temperature of 15 C. As soon as total body and cerebral temperatures were of appropriate level, the patient was exsanguinated and the cardiopulmonary bypass circuitry was converted into the RCP system. This consisted of taking the two clamps off the "bridge" between the venous and arterial lines and placing these clamps on the proximal venous line, the distal arterial line, and the inferior vena cava cannula. This allowed for direct blood flow from the arterial line through the "bridge" into the superior vena caval venous line retrograde and directly into the superior vena cava. Circuit clamping was controlled on the table by the surgeon. Central venous pressure (CVP) monitoring at the level of the jugular vein was used and RCP pressures were maintained at 20 mm Hg. Average flow was 400 ml/min with a range from 300 to 750 ml/min to target a CVP of 20 mm Hg. Using the retrograde cerebral perfusion technique, the aortic arch was opened and inspected, and dark blood was verified emanating from the brachiocephalic orifices. Temperature of inflow blood during retrograde cerebral perfusion was 10 C. The patient was placed into a slight Trendelenburg position during RCP. When all arch anastomoses were completed, the patient then was "deaired" via the retrograde perfusion circuit and blood was allowed to occupy fully the native aorta and graft. At this point, a cross-clamp was placed onto the graft, RCP was terminated, and antegrade perfusion using standard cardiopulmonary bypass via the femoral artery was initiated. More recently, a 10-mm side-arm graft has been placed on the Hemashield graft, and after completion of the aortic arch anastomosis, true P U m p CIRCULATION MANAGEMENT: THORACIC/THORACOABDOMI- NAL AORTIC RECONSTRUCTION. Circulation management for thoracic and thoracoabdominal reconstruction consisted of two fundamental strategies: (1) straight crossclamp and no distal perfusion and (2) distal perfusion, which in the majority of cases entailed a left atrial to distal aortic circuit. Since July 1, 1992, our protocol for thoracic and thoracoabdominal reconstruction also has included an initial CSF drainage catheter placed to right atrial pressure with the goal of maintaining CSF pressure less than 12 mm Hg. For more extensive thoracoabdominal operations, a staged cross-clamp technique was used with segmental reconstruction of the aorta as necessary. For typical Crawford type II thoracoabdominal reconstruction, the conduct of the operation included the following segmental anastomoses in sequence: (1) the proximal anastomosis with intercostal reimplantation and appropriate "bevel" of the anastomosis, (2) reimplantation of the left renal artery to a 6-mm graft, (3) creation of the mesenteric patch establishing flow to the celiac, superior mesenteric, and right renal arteries, and (4) completion of the distal anastomosis or aorto-bifemoral graft as necessary (Fig 2). Moderate core cooling was used during distal perfusion to approximately 32 C. Upon completion of the mesenteric anastomosis, the patient was rewarmed to 37 C. Five thousand units of intravenous heparin was used routinely and the activated clotting time was titrated between 180 and 200 seconds. At the completion of the mesenteric anastomosis, the distal aortic perfusion apparatus was transferred to a preconstructed 8-mm side-arm graft off of the thoracic aortic portion of the graft for rewarming. This in essence unloaded the left ventricle during rewarming. Target goals for distal perfusion included a mean proximal arterial pressure of approximately 80 to 90 mm Hg and a mean distal arterial pressure of greater than 60 mm Hg. Intercostal arteries routinely were implanted or included in either the proximal or distal beveled portion of the anastomosis. Reconstruction of Crawford type I and Crawford type III thoracoabdominal aneurysms required minor modifications in the above technique. For Crawford type I thoracoabdominal aneurysms there was no distal aortic reconstruction necessary. The great majority of these operations were completed using a beveled distal anastomosis into the mesenteric patch after placement of the separate left renal artery graft. Likewise, Crawford type IlI thoracoabdominal reconstruction used the same circulation management but there usually was no intercostal reimplantation. These patients usually had a signifi-

4 348 BAVARIA ET AL Ann Thorac Surg CIRCULATION MANAGEMENT IN AORTIC OPERATIONS 1995;60: ~c~ TM - "N.. i "~ ~" ~'-~= o~ 30- i ~ 25- [,. ; ; To SICU i ~ 35.5 oc....,. Transfer Perfusion to 8ram Sidearm I II I ;,t I Proximal and 88 I ~- Anastomoseslnterc stal 99. Meseniaric~[l~._ ABG _~ ~1 ~l-aanastomosis. Anastomosis "~---Dissection~ ~ Left Renal~ Anastomosis "~-- Closure Time (minutes) Fig 2. Conduct of operation for typical Craw.t~rd ~, pe I1 and ~pe II1 thoracoabdominal aortic aneurysms using staged segmental reconstruction. Type 1 rq~airs would preclude the aorto-b!h'moral graft (ABG). (CSF ~ cerebrospinal flui& LA - left atrium; Rap = right atn'al pressure; SICU = suryjcal intensive care unit.) cant proximal bevel, which included the important distal thoracic intercostals. Otherwise the same circulation management protocol for type li thoracoabdominal aneurysms was used. Crawford type 1V thoracoabdominal aneurysms were treated with straight aortic crossclamping without the benefit of distal perfusion, and a beveled proximal anastomosis into the mesenteric portion of the aorta was used. The only exception to this case was 1 patient who had a previous abdominal aortic aneurysm repair and a right lower pole renal artery that came off the abdominal aortic aneurysm graft and whose baseline creatinine level was 2.5 mg/dl. In this particular case, we used a distal aortic perfusion circuit for management of a type IV thoracoabdominal aneurysm to maintain flow through this important accessory right renal artery. Pharmacologic adjunctive management of these patients included administration of magnesium as well as methylprednisolone (125 nag) before cross-clamping. Mannitol was given just before removal of the crossclamp. Results Proximal Aortic Reconstruction For all proximal aortic procedures, the neurologic coinplication rate was 20% and the mortality rate was 21% (Table 3). Patients who died intraoperatively were excluded from calculation of stroke rates because of obvious inability to assess neurologic outcome. Stroke rates by circulatory techniques were as follows: straight cardiopulmonary bypass with cross-clamp, 12%; HCA, 48%; and RCP, 0%. Mortality rates by circulatory techniques were as follows: straight cardiopulmonary bypass with cross-clamp, 16%; HCA, 37%; hypothermia with RCP, 7%. Compared with HCA, RCP significantly lowered stroke rates (p < 0.01) and mortality (p < 0.05) (Table 4). The only RCP death occurred on postoperative day 10 while the patient was recovering well on the regular inpatient floor. Unstable rapid atrial fibrillation developed acutely, which immediately degenerated into ventricular fibrillation from which the patient was unable to be resuscitated. Additionally, when looking at type A dissection alone, hypothermia with RCP, which was used in 9 patients, had a 0% stroke rate compared with a 50% stroke rate with HCA and a 20% stroke rate with "classic" distal ascending aortic cross-clamp without an open distal anastomosis. Although the numbers were small, an additional 8 isolated aortic arch reconstructive procedures were performed. Five of these used HCA alone and the most recent three procedures used RCP. One HCA patient

5 Ann Thorac Surg BAVARIA ET AL ;60: CIRCULATION MANAGEMENT IN AORTIC OPERATIONS died intraoperatively and neurologic deficits developed in 2 of the other 4. The 3 RCP patients had no neurologic deficits. Emergent operations, as would be expected, had a higher mortality rate than elective operations. Sixty-eight percent of all proximal aortic procedures were emergent, usually type A dissection (see Table 1). The overall mortality rate for emergent proximal aortic reconstruction was 25.4% compared with 9.5% for elective cases (see Table 3). Thoracic and Thoracoabdominal Reconstructions For all thoracic and thoracoabdominal aortic operations (n = 81), the incidence of neurologic injury was 14%, renal failure 18%, and 60-day mortality 14% (Table 5), The percentage of emergency cases in this total group was 30.9% (see Table 2). "Classic" aortic cross-clamping resulted in a 27% incidence of spinal cord injury and a 24% renal failure rate. Circulation management using distal aortic perfusion resulted in a statistically significant reduction in neurologic deficit and 60-day mortality. Spinal cord paraplegia/paraparesis rates decreased from 27.2% to 6.7% (p < 0.01). Mortality decreased from 22.2% to 6.7% (p < 0.05). Distal aortic perfusion also resulted in a notable, but statistically not significant, decrease in renal failure from 24.2% to 13.3% (Table 6). Further analysis revealed operations on the thoracic Table 3. Proximal Aortic Reconstructive Procedures Procedure and Stroke 60-Day Circulation Management No. Rate ~ Mortality Ascending aortic aneurysm repair % 13.6% Distal ascending aortic 15 0% 13.3% clamp Open anastomosis: 5 40% 20% HCA alone Open anastomosis: 2 0% 0% HCA and RCP Type A dissection % 26.7% Distal ascending aortic clamp 22 20% 18.2% Open anastomosis: HCA alone 14 50% 50% Open anastomosis: 9 0% 11% HCA and RCP Aortic arch ".",, 12.5% reconstruction HCA alone 5 50% 20% HCA and RCP 3 0% 0% Proximal aortic % 21.3% reconstruction (totals) Emergent (%) 25.4% Elective (%) 9.5% a Patients who died intraoperatively were not included in determination of stroke rates. HCA hypothermic circulatory, arrest; RCP retrograde cerebral perfusion. Table 4. Proximal Aortic Reconstruction: Results of Circulation Management Strategy 60-Day Stroke Mortality Strategy No. (%) (%) Distal ascending % 16.2% aorta clamp Open anastomosis/ % 37.5% HCA alone Open anastomosis/ 14 0% 7.1% HCA and RCP (p < 0.01) (p < 0.05) HCA hypothermic circulatory arrest; RCP retrograde cerebral perfusion. and thoracoabdominal aorta that did not require a mandatory renal ischemic time, secondary to necessity of replacement of the mesenteric portion of the aorta, yielded a three-fourths reduction in renal failure. When the circulation management allowed for continuous perfusion of the renal arteries throughout the entire aortic reconstruction, renal failure rates decreased from 21.1% using standard straight aortic cross-clamping to 5.6% using distal aortic perfusion (Table 7). Comment There have been many reports documenting the serious nature and high incidence of major morbidity and mortality associated with reconstruction of the proximal aorta for both aneurysm and dissection. Reports from the 1970s and 1980s have shown mortality rates of 6% to 10% and stroke rates of 3% to 8% [5, 23-25]. Operative mortality is notably higher for repair of acute type A dissection. The incidence of transient neurologic symptoms has been reported to be as high as 18% in HCA patients. Additionally, until relatively recently, most patients with aortic arch aneurysms and other complex aortic pathology in tl~e.arch often were not offered definitive repair. This was secondary to the extreme technical difficulty of these operations and associated mortality and major morbidity. The first adjunct in circulation management of these patients consisted of the introduction of HCA [11. Using HCA, many authors reported that patients with complex proximal aortic pathology could be repaired with relative safety; indeed, HCA was mandatory for arch repair. Despite this advance, mortality and major morbidity remained challenging. There also have been selected reports using antegrade perfusion strategies and separate cannulation of the carotid arterial systems for management of the cerebral circulation during aneurysm and dissection repair [8]. These circulation management schemes are quite effective but are technically cumbersome and possibly no more effective than HCA alone. In 1988, RCP was introduced clinically for reconstruction of proximal aortic pathology by Ueda and associates [10]. This method has been used by many others with success, and the optimum technical strategy for retrograde cerebral perfusion presently is evolving. There are some reports revealing that increased flow at high CVP is

6 350 BAVARIA ET AL Ann Thorac Surg CIRCULATION MANAGEMENT IN AORT1C OPERATIONS 1995;60: Table 5. Neurologic Injury, Renal Failure, and Mortality Rates for Distal Aortic Reconstructions by Circulatory Management Neurologic Management Strategy No. Injuw a Renal Failure d 60-Day Mortality Thoracic aorta/crawford type I aneurysms 26 12% 8% 7.7% Straight X-clamp % 9.1% 16.7% Distal perfusion circuit % 0% Crawford type II thoracoabdominal aneurysms % 23.5% 22.2% Straight X-clamp 6 40% 20% 33.3% Distal perfusion circuit % 25% 16.7% Crawford type IIl thoracoabdominal 8 0% 12.5% 0% aneurysms Straight X-clamp 1 0% 0% 0% Distal perfusion circuit 7 0% 14.3% 0% Crawford type IV thoracoabdominal aneurysm % 2.5% 12.5% Straight X-clamp % 25% 12.5% Thoracic aortic traumatic transection 11 20% 20% 18.1% Straight cross-clamp 3 100% 50% 66.7% Distal perfusion 8 0% 12.5% 0% Type B dissection 10 10% 20% 20% Straight cross-clamp % 33% 16.7% Distal perfusion 4 0% 0% 25% Total % 17.9% 13.6% Emergency 28% Elective 7.1% a Patients who died intraoperatively were not included in determination of neurologic and renal dysfunction rates. X-clamp - cross-clamp. detrimental and other reports revealing that low flow with a low CVP or inadequate delivery of blood retrograde to the brain is likewise detrimental. Nojima and associates [26] suggested that RCP at 20 mm Hg is more effective than 10 mm Hg or 30 mm Hg. As a result of multiple clinical reports over the past few years, this technique has been used by aortic surgeons with greater frequency. In our report we not only used RCP for three aortic arch reconstructions but also used this technique during repair of nine acute aortic dissections allowing for complete inspection of the aortic arch and open distal anastomosis with either simple anastomosis or "herniarch" reconstruction. Our results were quite satisfactory, showing no strokes or transient neurologic deficits and only one death in this group, which occurred on postoperative day 10 of sudden ventricular fibrillation in an otherwise completely intact and alert 79-year-old woman. Our data corroborate the experi- Table 6. Overall Neurologic and Renal Dysfunction and Mortality Rates of Circulatory Management Techniques in Distal Aortic Reconstructions (n = 81) Neurologic Renal 60-Day Management Strategy No. Injury Failure MortaliW Straight cross-clamp % 24.2% 22.2% (no distal perfusion) Distal aortic perfusion % 13.3% 6.7% p Value ~0.01 NS ~0.05 mental data of Nojima and associates [26] showing RCP at a flow achieving a CVP of approximately 20 mm Hg in a slight Trendelenburg position is a safe circulation management technique during reconstruction of type A aortic dissection and arch aneurysm. This report also lends credence to the idea established more than a decade ago and adopted by more and more aortic surgeons that an open distal anastomosis is a preferred technique for repair of all acute type A aortic dissections [4]. As with proximal aortic surgery, the circulation management strategies during reconstruction of the thoracic and thoracoabdominal aorta also have evolved over the past decade. Distal aortic perfusion of various types is not Table 7. Effect of Circulation Management Strategy on Renal Function No Renal Renal Management Strategy Failure Failure Total Crawford type II and type III aneurysms X-clamp 5 1 Distal perfusion 15 4 Crawford type I and thoracic aortic procedures X-clamp 14 3 Distal perfusion 17 1 X-clamp cross-clamp. Renal Failure Rate % % % %

7 Ann Thorac Surg BAVARIA ET AL ;60: CIRCULATION MANAGEMENT 1N AORTIC OPERATIONS a new idea. Arterial to arterial shunts, including the Gott shunt, have been recommended and used by many groups [19]. Although these have limitations, there also have been reports of excellent results with these measures. Likewise, left atrial to femoral artery bypass as well as femoral venous to femoral arterial bypass have been used successfully during reconstruction of the distal thoracic and thoracoabdominal aorta. However, Hollier and his group [27] have had excellent success using straight cross-clamping of the aorta modified with CSF drainage and pharmacologic manipulation of the reperfusion injury. Despite these data, more and more authors have been advocating the routine use of distal aortic perfusion methods during reconstruction of the thoracic and thoracoabdominal aorta. Borst and Cunninghamm have advocated the use of these techniques maintaining distal aortic perfusion pressures of greater than 60 mm Hg [28]. Safi and Bartoli [29] have shown that the combination of distal aortic perfusion with routine CSF drainage has resulted in a substantial improvement in paraplegia rates compared with controls from an earlier period when neither were used. With more complex thoracoabdominal aortic reconstruction, Fehrenbacher and colleagues [30] showed that "staged" segmental reconstruction of the aorta using left heart bypass resulted in verv acceptable morbidity and mortality. Our data support the hypothesis that the distal aortic perfusion circuit maintaining a distal mean arterial pressure greater than 60 mm Hg improves paraplegia rates compared with straight aortic cross-clamping alone. Examination of our data from July 1, 1992, to present reveals 50 consecutive thoracic and thoracoabdominal aortic reconstructions with a 2% paraplegia rate. With this cohort of patients, routine adherence to the protocol in Figure 2 was performed, with distal aortic perfusion pressures of 60 mm Hg or greater, CSF drainage, moderate core cooling, intercostal reimplantation, and pharmacologic manipulation with magnesium, mannitol, and methylprednisolone. The only paraplegia in this group of patients was in a complete and total reconstruction of a Crawford type II thoracoabdominal aneurysm from the level of the subclavian artery into both common iliac arteries. In summary, based upon this retrospective analysis of 156 patients we conclude that RCP is an effective circulation management strategy during the reconstruction of type A aortic dissection and aneurysmal disease involving the aortic arch. We believe that RCP flows should be maintained at a CVP of approximately 20 mm Hg. Broadening the use of RCP to type A dissections allows for a safe and effective open distal aortic anastomosis or hemiarch procedure as necessary with full inspection of the aortic arch. We also conclude that for thoracic and thoracoabdominal aortic reconstruction, distal aortic perfusion in conjunction with the above adjunctive measures has produced excellent results. Another subset of patients that would benefit from this circulation management would be those with transections of the thoracic aorta secondary to trauma where left atrial to femoral artery bypass without heparin has been our recent standard. There has been no incidence of paraplegia using this system, whereas straight aortic cross-clamping has produced paraplegia at our institution. Retrograde cerebral perfusion and distal aortic perfusion in proximal and distal aortic reconstructive surgery provide protection against neurologic injury and reduce mortality. These circulatory management techniques are under current investigation and warrant extensive further laboratory investigation and clinical application. References 1. Griepp RB, Stinson EB, Hollingsworth JF, Buehler D. Prosthetic replacement of the aortic arch. J Thorac Cardiovasc Surg 1975;70: Cooley DA, Ott DA, Frazier OH, Walker WE. Surgical treatment of the transverse aortic arch: experience with 25 patients using hypothermic techniques. Ann Thorac Surg 1981;32: Crawford ES, Saleh SA. Transverse aortic arch aneurysm: improved results of treatment employing new modifications of aortic reconstruction and hypothermic cerebral circulatory arrest. Ann Surg 1981;194: Livesay JJ, Cooley DA, Duncan JM, Ott DA, Walker WE, Reul GJ. Open aortic anastomosis: improved results in the treatment of aneurysms of the aortic arch. Circulation 1982; 66(Suppl 1): Crawford ES, Svensson LG, Coselli JS, Sail HJ, Hess KR. Surgical treatment of aneurysm and/or dissection of the ascending aorta, transverse aortic arch, and ascending aorta and transverse aortic arch. Factors influencing survival in 717 patients. J Thorac Cardiovasc Surg 1989;98: Davis EA, Gillinov AM, Cameron DE, Reitz BA. Hypothermic circulatory, arrest as a surgical adjunct: a 5-year experience with 60 adult patients. Ann Thorac Surg 1992;53: Svensson LG, Crawford ES, Hess KR, et al, Deep hypothermia with circulatory arrest. Determinants of stroke and early mortality in 656 patients. J Tborac Cardiovasc Surg 1993;106: Kazui T, Inoue N, Yamada O, Komatsu S. Selective cerebral perfusion during operation for aneurysms of the aortic arch: a reassessment. Ann Thorac Surg 1992;53: Mill NL, Ochsner JL. Massive air embolism during cardiopulmonary bypass. J Thorac Cardiovasc Surg 1980;80: Ueda Y, Miki S, Kusuhara K, et al. Surgical treatment of the aneurysm of dissection involving the ascending aorta and aortic arch utilizing circulatory arrest and retrograde cerebral perfusion. J Jpn Assoc Thorac Surg 1988;36: Ueda Y, Miki S, Kusuhara K, Okita Y, Tahata T, Yamanaka K. Deep hypothermic systemic circulatory arrest and continuous retrograde cerebral perfusion for surgery of aortic arch aneurysm. Eur J Cardiothorac Surg 1992;6: Usui A, Hotta T, Hiroura M, et al. Retrograde cerebral perfusion through a superior vena caval cannula protects the brain. Ann Thorac Surg 1992;53: Sail HJ, Brien HW, Winter JN, et al. Brain protection via cerebral retrograde perfusion during aortic arch aneurysm repair. Ann Thorac Surg 1993;56: Usui A, Oohara K, Liu TL, et al. Determination of optimum retrograde cerebral perfusion conditions. J Thorac Cardiovasc Surg 1994;107: Usui A, Oohara K, Liu T, et al. Comparative experimental study between retrograde cerebral perfusion and circulatory arrest. J Thorac Cardiovasc Surg 1994;107: Yasuura K, Ogawa Y, Okamoto H, et al. Clinical application of total body retrograde perfusion to operation for aortic disse~tion. Ann Thorac Surg 1992;53: Yasuura K, Okamoto H, Ogawa Y, et al. Resection of aortic aneurisms without aortic cross-clamp technique with the aid 6f'hypothermic total body retrograde perfusion. J Thorac Cardiovasc Surg 1994;107: Wadouh F. New surgical techniques for the prevention of

8 352 BAVARIA ET AL Ann Thorac Surg CIRCULATION MANAGEMENT 1N AORTIC OPERATIONS 1995;60: paraplegia during aortic surgery. Thorac Cardiovasc Surg 1992;40: Verdant A. Descending thoracic aortic aneurysms: surgical treatment with the Gott shunt. Can J Surg 1992;35: Higgins RS, Sanchez JA, DeGuidis L, et al. Mechanical circulatory support decreases neurologic complications in the treatment of traumatic injuries of the thoracic aorta. Arch Surg 1992;127: Borst HG, Jurmann M, Buhner B, Laas J. Risk of replacement of descending aorta with a standardized left heart bypass technique. J Thorac Cardiovasc Surg 1994;107: Massimo CG, Presenti LF, Favi PP, Crisci C, Guadron EAC. Simultaneous total aortic replacement from valve to bifurcation: experience with 21 cases. Ann Thorac Surg 1993;56: Cabrol CE, Gandjbakeh I, Pavie A, Bors V. Total replacement of the ascending aorta: La Pitie experience. Semin Thorac Cardiovasc Surg 1990;3: Crawford ES, Coselli JS. Replacement of the aortic arch. Semin Thorac Cardiovasc Surg 1990;3: Griepp RB, Ergin MA, Lansman SL, Galla JD, Pogo G. The physiology of hypothermic circulatory arrest. Semin Thorac Cardiovasc Surg 1990;3: Nojima T, Mori A, Watarida M, et al. Experimental studies of retrograde cerebral perfusion [Abstract]. Aortic Surgery Symposium IV, New York, NY, Hollier LH, Money SR, Nasland TC, et al. Risk of spinal cord dysfunction in patients undergoing thoracoabdominal aortic replacement. Am J Surg 1992;164: Borst HG, Moderator. Descending and thoracoabdominal aorta: panel discussion. Semin Thorac Cardiovasc Surg 1991; 3: Sail HJ, Bartoli KR. Neurologic deficit in high risk patients with thoracoabdominal aortic aneurysms: the role of cerebral spinal fluid drainage. Aortic Surgery Symposium IV, New York, NY, 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: DISCUSSION DR SAFUH ATTAR (Baltimore, MD): 1 congratulate Dr Bavaria and his associates on their solution to this very complicated problem. In our experience, cerebral perfusion, whether antegrade or retrograde, is indicated only when the transverse arch is involved by aneurysmal dissection. There is no need for cerebral perfusion in ascending aortic aneurysms, whether it is dissecting or not. As far as the descending thoracic aortic aneurysms, our results compare with those of Dr Crawford and are about the same as yours. We have had 87 patients undergoing resection of the descending thoracic aortic aneurysms; 61 had acute traumatic aortic ruptures and 26 chronic aortic aneurysms. Among the acute traumatic ruptures, 22 were resected without shunt with a cross-clamp time of 41 minutes; 31 patients were shunted with a cross-clamp time of 74 minutes. In the chronic aneurysms, 13 had adjunctive support with a cross-clamp time of 22 minutes (these include thoracic and thoracoabdominal aortic aneurysms), and 13 patients had shunt and bypass with a cross-clamp time of 45 minutes. The paraplegia without shunt occurred in 4 of 22 with traumatic aortic rupture, an incidence of 17%, with the shunt it occurred in 6 of 39, which is 15%; in other words, no difference. In patients with chronic aortic aneurysms it occurred in I of 13, that is 7%, with or without shunt; the rates were equal. The incidence of renal failure was 5 of 39, that is 12%, in the shunted acute ruptures and 0% in the nonshunted group. In patients with chronic aortic aneurysms, the rate was 1 of 13, again, 7%, without shunt and 15% with the shunt; that is 2 of 13. The survival rate was 72% in acute traumatic aortic ruptures without the shunts and 59% with the shunts. In the chronic descending thoracic aneurysms, we had 11 of 13 survivors (84%) without shunt and 7 of 13 (54%) in the shunted group. I do not think there is an indication for the use of retrograde cerebral perfusion in ascending aortic aneurysms, and the only reason we currently are using the Bio-Medicus pump with the Carmeda system is for medicolegal reasons. I do not believe any method will prevent the incidence of paraplegia completely, because the etiology is multifactorial. DR BAVARIA: Regarding Dr Attar's question concerning the indication for retrograde cerebral perfusion in ascending aortic aneurysms, I believe he is fundamentally correct. Retrograde cerebral perfusion is only used during the repair of ascending aortic aneurysms when the fusiform aneurysmal component of the aneurysm extends into the proximal aortic arch. In this particular case, the use of retrograde cerebral perfusion is akin to the indication for retrograde cerebral perfusion in an uncomplicated acute type A aortic dissection. In these cases, the quality of the distal anastomosis is much improved using an "open anastomotic" technique. Although the open anastomotic technique for either acute type A dissection or ascending/proximal arch aneurysm only requires between 15 and 30 minutes of retrograde cerebral perfusion, I continue to believe that retrograde cerebral perfusion adds protection even during these short periods of hyperthermic circulatory arrest. I arrive at this conclusion based on an observation we have made regarding postoperative neurologic function in patients undergoing pulmonary thromboendarterectomy. Our standard technique for these patients is to perform the pulmonary thromboendarterectomy during 10- to 12-minute periods of hypothermic circulatory arrest followed by a period of reperfusion. Sometimes it will take between three and five periods of hypothermic circulatory arrest to complete both the left and right endarterectomy. It is my observation that these patients often will have short-term transient nonfocal neurologic damage in spite of the relatively short periods of hypothermic circulatory arrest. For this reason, as well as deairing and embolic reasons, I believe retrograde cerebral protection is appropriate even for relatively short periods of circulatory arrest. An open anastomosis is important in the reconstruction of ascending aortic pathology that extends to the level of the innominate artery because, as Dr Miller of Stanford has said, "There is nothing quite so terrifying as to take off all your clamps and see your distal anastomosis start coming apart after acute type A dissection." In response to Dr Attar's questions regarding traumatic aortic transection, there has been an excellent study by Swenson, after reviewing Crawford's data, showing that when thoracic aortic cross-clamp time exceeded 40 minutes, there was a significant paraplegia rate. Although we usually can repair a thoracic aortic transection faster than 40 minutes, this is not always the case. Additionally, a recent paper by Von Opel revealed that the best circulation management strategy for thoracic aortic transection was distal aortic perfusion without the use of heparin. Our data also support this as we have no paraplegia rate at all in traumatic aortic transection using distal aortic perfusion.

9 Ann Thorac Surg BAVARIA ET AL ;60: CIRCULATION MANAGEMENT IN AORTIC OPERATIONS DR HAZIM J. SAFI (Houston, TX): We used hypothermic circulatory arrest and retrograde cerebral perfusion for 29 patients during ascending and arch repair, and no strokes developed. Using conventional methods for 27 patients, stroke developed in 3. We kept the electroencephalogram isoelectric for both groups. With the 29 patients, upon initiation of retrograde cerebral perfusion, the electroencephalogram came back in 2 patients with 1 remaining for a full 55 minutes. Have you had this sort of experience? DR BAVARIA: As noted, I showed one slide during the talk that was a recording of continuous electroencephalogram during a "hemi-arch" aortic reconstruction. As was shown on the slide, we obtained some activity on the electroencephalogram as soon as we started retrograde cerebral perfusion. So I agree with your observation completely. DR SAFh Do you also use cerebrospinal fluid drainage for your patients? DR BAVARIA: I routinely include cerebrospinal fluid drainage for all distal aortic cases that are not emergent. The only time 1 would not use cerebrospinal fluid drainage is if I decide to go on full cardiopulmonary bypass with 25,000 units of heparin as opposed to just 5,000 units. In those patients I will not use cerebrospinal fluid drainage. DR SAFI: I would like to make an appeal to the community of vascular surgeons at large: when we report the surgical results of thoracoabdominal aortic aneurysm repair, we should make sure to distinguish the separate aneurysm types and more or less concentrate on the higher risk type I and type II aneurysms, because for type IV and type III, the incidence of paraplegia is less than 5% and you need 500 cases to prove your point. DR BAVARIA: I completely agree. For the statistical analysis of paraplegia in this article, I removed Crawford type 1V thoracoabdominal aneurysms from consideration because there is no place for "circulation management" in type W repairs. However, during the routine reconstruction of Crawford type III thoracoabdominal aneurysms, we do put an aortic cross-clamp directly superior to the celiac artery and perform the proximal anastomosis using distal perfusion and then perform the mesenteric anastomosis in a sequential fashion. Notice Regarding "Our Surgical Heritage" Section Readers are solicited to contribute recollections, vignettes, interviews, and incidents involving surgical pioneers. The items need not be lengthy, but they should have involved the author directly. Humorous incidents will not be excluded, and the style of writing can be such as to provide a few minutes of entertainment. Full articles of historic nature will, of course, continue to be welcome. Anthony R. C. Dobell, MD Associate Editor Our Surgical Heritage Section