Ascending and Transverse Aortic Arch Repair The Impact of Retrograde Cerebral Perfusion

Transcription

1 Ascending and Transverse Aortic Arch Repair The Impact of Retrograde Cerebral Perfusion Anthony L. Estrera, MD; Charles C. Miller, III, PhD; Taek-Yeon Lee, MD; Pallav Shah, MD; Hazim J. Safi, MD Background The benefit of retrograde cerebral perfusion (RCP) with profound hypothermic circulatory arrest has been subject to much debate. We examined our experience with ascending and transverse arch repairs to determine the impact of retrograde cerebral perfusion on stroke and mortality. Methods and Results Between August 1991 and June 2007, we performed 1107 repairs of the ascending and transverse aortic arch. RCP was used in 82% of cases (907 of 1107). Sixty-two percent were men (682 of 1107); median age was 64 years (range, 16 to 93 years). Perioperative variables were evaluated using univariate and multivariable analysis for mortality and stroke. Thiry-day mortality was 10.4% (115 of 1107). Stroke occurred in 2.8% (31 of 1107) of patients. Univariate risk factors for mortality were increasing age (P ), history of coronary artery disease (P 0.02), previous coronary artery bypass (P 0.02), emergency status (P ), acute dissection (P 0.02), rupture (P ), preoperative glomerular filtration rate, bypass time (P ), crossclamp time (P 0.007), RCP time (P ), and packed red blood cell transfusions (P ). Univariate risk factors for stroke included emergency status (P 0.02), cerebrovascular disease (P 0.02), and crossclamp time (P 0.04). Independent risk factors for mortality were glomerular filtration rate 90 ml/min (P ), emergency status (P 0.006), rupture (P 0.004), cardiopulmonary bypass time 120 minutes (P 0.04), and packed red blood cell transfusions (P ). Risk factors for stroke were emergency status (P 0.009) and hypertension (P 0.05). RCP was protective against mortality and stroke. Conclusions The use of RCP with profound hypothermic circulatory arrest was associated with a reduction in mortality and stroke. The use of RCP remains warranted during repairs of the ascending and transverse aortic arch. (Circulation. 2008;118[suppl 1]:S160 S166.) Key Words: aortic arch cerebral protection circulatory arrest perfusion surgery The optimal approach to cerebral protection during ascending and transverse aortic arch repairs remains undetermined. Since Griepp originally reported the use of profound hypothermic circulatory arrest (PHCA) for transverse arch repairs, 1 differing approaches have been used in combination with PHCA to reduce neurological complications. 2 5 Neurological complications of stroke and temporary neurological deficit (TND) or encephalopathy have been reported to occur in as many as 5% to 40% of these repairs Moreover, it was also identified that greater lengths of circulatory arrest were associated with greater risk for neurological injury. 11,12 For these reasons, the strategies of retrograde and antegrade cerebral perfusion were devised as adjuncts to PHCA for repairs of the transverse aortic arch. Ueda first reported using retrograde cerebral perfusion (RCP) in conjunction with PHCA during ascending and transverse arch repairs. 2 Reported advantages of RCP were flushing of atheromatous debris, uniform cerebral cooling, ease of use, and potential nutritive support. 2,6,13,14 In contrast, advocates of antegrade cerebral perfusion suggested decreased neurological complications due to improved cerebral perfusion. Despite studies comparing the different adjuncts, it has been difficult to demonstrate a significant advantage of one approach over the other. Animal and clinical studies led us to adopt RCP as the main method of cerebral protection since ,16 Since then, we have reported satisfactory results regarding neurological outcome during ascending and arch repairs. 6,17 The purpose of this study was to examine our experience with ascending and transverse arch repairs and determine the impact of RCP on stroke and mortality. Methods and Materials The Committee for the Protection of Human Subjects at the University of Texas Houston Medical School approved review of the data collected for this study. The study design was a retrospective observational study. Patient Population Between August 1991 and June 2007, we performed 1107 cases of ascending and transverse arch repairs. Sixty-two percent were men (682 From the Department of Cardiothoracic and Vascular Surgery, The University of Texas at Houston Medical School, Memorial Hermann Heart and Vascular Institute, Houston, Texas. Presented at the American Heart Association Scientific Sessions, November 4 7, 2007, Orlando, Fla. Correspondence to Anthony L. Estrera, MD, Department of Cardiothoracic and Vascular Surgery, The University of Texas Houston Medical School, 6410 Fannin, Suite 425, Houston, TX anthony.l.estrera@uth.tmc.edu 2008 American Heart Association, Inc. Circulation is available at DOI: /CIRCULATIONAHA S160

2 Estrera et al Ascending and Transverse Aortic Arch Repair S161 of 1107); median age was 64 years (range, 16 to 93 years). Of these patients, RCP was used in 82% of cases (907 of 1107). Antegrade cerebral perfusion (ACP), combined ACP and RCP, or PHCA alone generated the remaining cases. Retrograde cerebral perfusion was not used in the following cases: when ACP was used alone, when the superior vena cava could not be cannulated, when false proximal aneurysms precluded safe median sternotomy, and cases were performed before our conventional use of RCP. Coronary artery bypass was performed in 13% (145 of 1107) of cases, aortic root replacement in 25% (281 of 1107), with aortic valve replacement in 31% (345 of 1107). Among all patients, 20% (218 of 1107) underwent stage 1 elephant trunk procedure. Preoperative and operative variables are listed on Table 1. Technique Our approach for transverse arch repairs has been reported previously. 18,19 Basic features of our current technique for acute dissection or aneurysm of the distal ascending or transverse aortic arch included cardiopulmonary bypass, profound hypothermia, circulatory arrest, and RCP. Cerebral monitoring with near-infrared spectroscopy or power mode transcranial Doppler confirmed adequate cerebral circulation during perfusion. A 10-lead electroencephalogram monitored cerebral function. Once the electroencephalogram was isoelectric, which coincided with a nasopharyngeal temperature of 15 C to 20 C, cardiopulmonary bypass was discontinued and circulation was arrested. RCP was begun through the superior vena cava cannula. RCP flow rate will vary depending on the information obtained from bilateral power mode transcranial Doppler ultrasound and bilateral near-infrared spectroscopy (cerebral oximetry) When power mode transcranial Doppler ultrasound was used, adequacy of RCP flow was determined by the presence of reversed blood flow in the middle cerebral arteries. Although a higher opening pressure is required, the maintenance flow rate is often below 500 ml/min, maintaining the pressure in the superior vena cava line pressure below 25 mm Hg. 20 The information obtained from power mode transcranial Doppler ultrasound was correlated with information obtained with near-infrared spectroscopy (cerebral oximetry). When total arch repair was performed, island patch reattachment of the innominate, left common carotid, and left subclavian arteries was performed unless the patient had Marfan syndrome. In these cases, a premanufactured multibranched graft was used to bypass the great vessels. Preoperative factors analyzed included age, gender, emergency status, hypertension, renal insufficiency as determined by glomerular filtration rate (GFR), chronic obstructive pulmonary disease, peripheral arterial disease, cerebrovascular disease, and smoking. Operative factors analyzed were RCP times, aortic crossclamp time, cardiopulmonary bypass time, and intraoperative transfusions required. Acute dissection referred to dissection occurring within 2 weeks based on initial onset of pain. Chronic obstructive pulmonary disease was defined by a history of chronic bronchitis and emphysema, or, while on bronchodilators, less than 60% of predicted forced expired volume in 1 second. A serum creatinine level of 2.0 mg/dl or the need for dialysis defined renal dysfunction. GFR was estimated using the Cockcroft-Gault equation. 21 Study end points included 30-day mortality, stroke, and TND. Thirty-day mortality refers to deaths that occurred within 30 days of surgery. Stroke was defined as any gross focal neurological brain injury, either temporary or permanent, as identified on neurological examination by a neurology consultant and confirmed with CT or MRI. Temporary confusion, delirium, agitation, disorientation, or altered mental status denoted TND on all patients who survived operation as defined by Ergin et al. 7 We modified this classification by waiting 24 hours after complete reversal of anesthesia before the pronouncement of TND. Acute stroke was excluded by CT scan or MRI of the head in all patients suspected of having TND. All patients with TND were followed for the entire hospitalization for ultimate resolution of dysfunction. Statistical Analysis Treatment technique was not randomized but was determined by best medical judgment for each individual case. Analysis was retrospective. Data were collected from chart reviews by a trained nurse abstractor and were entered into a dedicated Microsoft Access database. Data were exported to SAS for analysis, and all computations were performed using SAS version running under Windows XP. Univariate analyses were performed using unpaired t tests and 2 or Fisher exact tests as appropriate. Multivariable analyses were conducted by multiple logistic regression. The null hypothesis was rejected at P Statement of Responsibility The authors had full access to the data and take responsibility for its integrity. All authors have read and agree to the manuscript as written. Results For the entire cohort, 30-day mortality was 10.4% (115 of 1107). Stroke occurred in 2.8% (31 of 1107) of patients. TND occurred in 15.5% (172 of 1107) of patients. Median cardiopulmonary bypass time was 144 minutes (range, 11 to 535 minutes). Median aortic crossclamp time was 83 minutes (range, 6 to 306 minutes). Median RCP time was 26 minutes (range, 3 to 89 minutes). Median transfusions of packed red blood cells, fresh-frozen plasma, and platelets were 4, 6, and 11 U, respectively. Univariate risk factors for 30-day mortality were increasing age (P ), history of coronary artery disease (P 0.02), previous coronary artery bypass (P 0.02), emergency status (P ), acute dissection (P 0.02), rupture (P ), preoperative GFR, bypass time (P ), aortic crossclamp time (P 0.007), retrograde cerebral perfusion time (P ), and packed red blood cell transfusions (P ; Table 1). Univariate risk factors for stroke included emergency status (P 0.02), history of cerebrovascular disease (P 0.02), and crossclamp time (P 0.04). By multivariable analysis, independent risk factors for mortality were GFR 90 ml/min (P ), emergency status (P 0.006), rupture (P 0.004), cardiopulmonary bypass time 120 minutes (P 0.04), and packed red blood cell transfusions (P ; Table 2) By multivariable analysis, independent risk factors for stroke were emergency status (P 0.009) and hypertension (P 0.05; Table 3). RCP was protective against mortality and stroke on univariate and multivariable analysis (P and P 0.02, respectively). By multivariable analysis, independent risk factors for TND were GFR 90 ml/min (P 0.04), previous stroke (P ), previous transient ischemic attack (P 0.03), and packed red blood cell transfusions (P ). Ascending and only proximal transverse arch repair was associated with decreased TND (Table 4). Analysis of the cause of death revealed that most deaths were related to cardiac etiology (23%), multiple organ failure (20%), and intraoperative causes (17%), which were either cardiac failure or bleeding (Table 5). Discussion The optimal strategy for cerebral perfusion during ascending and transverse aortic arch repairs has yet to be determined. Many variables have been associated with poor neurological

3 S162 Circulation September 30, 2008 Table 1. Risk Factors for 30-Day Mortality Variable No. of Patients (%) No. 30-Day Deaths (%) OR 95% CI P Overall 1107 (100.0) 115 (10.4) Age, years (24.8) 21 (7.7) (0.0009) (23.3) 12 (4.7) (25.8) 33 (11.5) (26.1) 49 (17.0) Female 425 (38.4) 49 (11.5) Male 682 (61.6) 66 (9.7) Hypertension 761 (68.7) 78 (10.3) No hypertension 346 (31.3) 37 (10.7) Proximal Not total arch 680 (61.4) 66 (9.7) Total arch 427 (38.6) 49 (11.5) Root 281 (25.4) 34 (12.1) Without root 826 (74.6) 81 (9.8) Coronary artery disease 252 (22.8) 36 (14.3) No coronary artery disease 855 (77.2) 79 (9.2) Prior coronary artery bypass graft 133 (12.0) 22 (16.5) No history of coronary artery bypass graft 974 (88.0) 93 (9.6) Congestive heart failure 23 (2.1) 3 (13.0) No congestive heart failure 1084 (97.9) 112 (10.3) Chronic obstructive pulmonary disease 223 (20.1) 23 (10.3) No chronic obstructive pulmonary disease 884 (79.9) 92 (10.4) Marfan 32 (2.9) 4 (12.5) Not Marfan 1075 (97.1) 111 (10.3) Emergent 273 (24.6) 46 (16.9) Elective 834 (75.4) 69 (8.3) Dissection None 624 (56.4) 56 (8.9) 1 Acute 279 (25.2) 41 (14.7) Chronic 204 (18.4) 18 (8.8) Rupture 54 (4.9) 18 (33.3) Intact 1053 (95.1) 97 (9.2) 1 GFR (24.4) 48 (19.1) (0.0001) (24.7) 23 (9.1) (23.6) 19 (7.8) (27.2) 13 (4.6) ET 218 (19.7) 19 (8.7) Conventional 889 (80.3) 96 (10.8) Coronary artery bypass graft 145 (13.3) 17 (11.7) No coronary artery bypass graft 962 (86.9) 98 (10.2) Aortic valve Normal 452 (40.8) 51 (11.2) 1 Repaired 310 (28.0) 28 (9.0) Replaced 345 (31.2) 36 (10.4) RCP 907 (81.9) 86 (9.5) No RCP 200 (18.1) 29 (14.5) 1 (Continued)

4 Estrera et al Ascending and Transverse Aortic Arch Repair S163 Table 1. Continued Variable No. of Patients (%) No. 30-Day Deaths (%) OR 95% CI P RCP time (RCP group only) Missing 33 (3.6) 4 (12.2) (29.2) 11 (4.2) (0.04) (22.9) 13 (6.3) (21.9) 24 (12.1) (0.04) (22.2) 34 (16.8) Clamp time (25.6) 22 (8.2) (0.0001) (24.3) 18 (7.0) (25.2) 27 (10.2) (0.0001) (24.9) 41 (15.7) Pump time (24.7) 19 (7.1) (0.0001) (25.4) 18 (6.5) (24.2) 19 (7.2) (25.6) 55 (19.8) Packed red blood cell units (27.4) 25 (8.2) (0.0001) (29.5) 17 (5.2) (20.3) 19 (8.4) (0.0001) (22.8) 54 (21.4) Fresh-frozen plasma units (28.2) 28 (8.9) (0.04) (25.7) 25 (8.8) (26.2) 29 (10.0) (0.04) (19.9) 33 (15.0) Platelet units (28.2) 33 (10.6) (0.36) (22.9) 22 (8.7) (0.36) (30.0) 30 (9.0) (18.9) 30 (14.4) Postoperative stroke 31 (2.8) 6 (19.4) No stroke 1076 (97.2) 109 (10.1) 109 (10.1) Prolonged Ventilation 243 (21.9) 40 (16.5) Short ventilation 864 (78.1) 75 (8.9) Renal failure 61 (5.5) 19 (31.2) No failure 1046 (94.5) 96 (9.2) *For dichotomous variables, the OR represents a test against a reference category whose referent OR is equal to 1. For continuous data, the OR refers to the increase in odds associated with a 1-unit increase in the variable value. Although continuous data are presented in quartiles, the ORs are against the continuous variable. This reflects the units against which its companion OR is computed. CIs are test-based. Type I statistical error (common P value). Values without parentheses are Pearson 2 probabilities. Probability values in parentheses are univariate logistic regression likelihood ratio P values. outcomes, including emergency status, aortic dissection, degree of atheromatous disease, diabetes, history of cerebrovascular accidents, and renal insufficiency. Operative factors affecting neurological outcome have also included cannulation strategy, acid/base management, duration of PHCA, 11 temperature management, and cerebral perfusion technique (ACP or RCP). Because many variables influence neurological outcome, it is not surprising that it has been difficult to demonstrate superiority of one technique over another. In previous animal and clinical studies, we demonstrated an advantage using RCP during arch repairs with regard to neurological outcomes. 6,16 Currently, the incidence of stroke of 2.8% compared favorably with other contemporary series reporting stroke incidences from 3% to 12% regardless of

5 S164 Circulation September 30, 2008 Table 2. Risk Factors for 30-Day Mortality: Multiple Logistic Regression Variable Parameter Estimate PR 95% CI P Intercept GFR Emergency Rupture Pump Packed red blood cells RCP technique of cerebral perfusion. 3,22,23 In addition, our incidence of TND at 15% is comparable to other reported studies. 7 Some have suggested RCP as being inferior to ACP in providing adequate cerebral protection. 12,24 These concerns have been based on observations of inadequate cerebral perfusion as detected by specific cerebral monitoring modalities (transcranial Doppler) or direct perfusion studies. 19,25 Yet in many of these reports, the stroke incidence was still significant, ranging from 3% to 12%. This reiterates the point that postoperative stroke after ascending and transverse arch repairs remains multifactorial in etiology and that providing adequate cerebral perfusion, whether antegrade or retrograde, does not ensure prevention of neurological complications. In fact, in a recent report, it was observed that up to two thirds of patients with stroke after ascending and arch repair were of embolic origin as compared with only one third attributed to hypoperfusion. 26 This suggests that providing ACP may be in part beneficial but not the most important variable when protecting against stroke. Considering this, it is probable that the primary effect of RCP in preventing stroke was related to flushing of atheromatous emboli. It has been our practice to provide at a minimum RCP during any period of PHCA unless not feasible. Some have suggested that RCP may even be detrimental with concern for increased cerebral edema and hyperemia, 27 although others have not. 28,29 On the contrary, our results of the RCP group were comparable to recent results reporting outcomes with either RCP or ACP alone and we have not observed increased cerebral edema. 30,31 Since adopting RCP, we have also modified our technique for administering RCP to improve its efficacy. We determined that the adequacy of RCP was dependent on modifying pressure and flow as guided by cerebral monitoring and that without cerebral monitoring, inadequate flow was observed in over 78% of cases. 30,32 Because cerebral venous capacitance vessels may require a Table 3. Risk Factors for Stroke: Multiple Logistic Regression Variable Parameter Estimate OR 95% CI P Intercept Emergency Hypertension RCP Table 4. Risk Factors for Transient Neurologic Deficit: Multiple Logistic Regression Variable Parameter Estimate OR 95% CI P Intercept GFR Prior stroke Prior transient ischemic attack Proximal only Packed red blood cells higher pressure to establish retrograde cerebral blood flow, nonguided RCP may not achieve adequate cerebral protection. This may be an explanation for inferior results using nonguided RCP demonstrated in previous studies. Although the correlation of RCP with improved neurological outcomes was not surprising, the beneficial impact of RCP on mortality was unexpected. It remained unclear, however, why RCP, which theoretically should only affect neurological outcomes, would also provide a survival advantage. One possibility is related to prevention of stroke. Because it is acknowledged that postoperative stroke after cardiovascular surgery predicts poor early and late survival, 33 if modalities such as RCP prevented stroke, improved early survival could be expected. Interestingly, analysis of causes of death revealed stroke or neurological injury to occur in only 6% of the deaths, whereas cardiac and multiorgan failure was the most frequent causes of death. Other significant risk factors included preoperative renal dysfunction as determined by GFR, prolonged cardiopulmonary bypass time ( 120 minutes), emergency procedures, and rupture. We have recently reported the significance of using GFR as a predictor of early mortality and the fact that using serum creatinine alone was relatively insensitive at detecting subclinical renal insufficiency. 34 Prolonged bypass time was a risk factor as expected, which likely represents the complexity of the procedures performed. Rupture and emergency procedures would have been expected to influence early mortality. Of interest, although age was a univariate risk factor for early death, age was not an independent risk factor for early mortality. Although controversial, this also Table 5. Causes of 30-Day Mortality During Ascending and Transverse Arch Repair Cause of Death No. (%; n 115) Cardiac 27 (23%) Multiorgan failure 24 (21%) Intraoperative 17 (15%) Respiratory 11 (10%) Sepsis 8 (7%) Aortic rupture 9 (8%) Stroke 7 (6%) Bowel necrosis 5 (4%) Pulmonary embolism 2 (2%) Unknown 5 (5%)

6 Estrera et al Ascending and Transverse Aortic Arch Repair S165 suggested that age should not be a contraindication for surgical repair. This study should be viewed with certain limitations. The study was retrospective in nature and inherent biases likely existed. Although RCP in conjunction with PHCA was preferred, RCP was not used in all cases. Those cases that did not involve RCP alone included cases of ACP, combined ACP and RCP, or PHCA alone. No method for patient selection for perfusion type existed, because surgeon preference determined the specific approach for cerebral perfusion used, and thus the cohort was subject to bias. Second, in evaluating the causes of early deaths, a significant number of deaths were related to early cardiac failure, rupture, or intraoperative bleeding where early neurological evaluation may not have been feasible. It was therefore possible that some of these early deaths could have sustained neurological injuries but were undetected. We hope that our large cohort and relatively low event rate would compensate for this. Third, because this study occurred over a long period of time, it is possible that changes in surgical technique, cannulation strategy, cerebral monitoring, anesthetic management, and intensive care unit care could have had an impact. Last, it must be recognized that the mean RCP time was 26 minutes and that only 22% of patients underwent a circulatory arrest time beyond 40 minutes. It has been previously established that longer circulatory arrest times are associated with a greater risk of morbidity and mortality, 11 and it could be argued that our relatively short circulatory arrest times might not have allowed RCP to have any appreciable effect. Although we acknowledge that adequate nutritive flow may not be provided by RCP, an important benefit of flushing atheromatous debris likely accounts for its beneficial effect. In conclusion, the use of RCP with profound hypothermic circulatory arrest is associated with a reduction in mortality and stroke. The use of RCP remains warranted during repairs of the ascending and transverse aortic arch. Acknowledgments We thank Ken Goodrick for his editorial assistance. Source of Funding The research reported here was supported by grant 5 P50 HL (TAAD-SCCOR) from the National Heart, Lung, and Blood Institute. None. Disclosures References 1. Griepp RB, Stinson EB, Hollingsworth JF, Buehler D. Prosthetic replacement of the aortic arch. J Thorac Cardiovasc Surg. 1975;70: Ueda Y, Miki S, Kusuhara K, Okita Y, Tahata T, Yamanaka K. Surgical treatment of aneurysm or dissection involving the ascending aorta and aortic arch, utilizing circulatory arrest and retrograde cerebral perfusion. J Cardiovasc Surg. 1990;31: Kazui T, Washiyama N, Muhammad BA, Terada H, Yamashita K, Takinami M, Tamiya Y. Total arch replacement using aortic arch branched grafts with the aid of antegrade selective cerebral perfusion. Ann Thorac Surg. 2000;70: Bachet J, Guilmet D, Goudot B, Dreyfus GD, Delentdecker P, Brodaty D, Dubois C. Antegrade cerebral perfusion with cold blood: a 13-year experience. Ann Thorac Surg. 1999;67: Spielvogel D, Halstead JC, Meier M, Kadir I, Lansman SL, Shahani R, Griepp RB. Aortic arch replacement using a trifurcated graft: simple, versatile, and safe. Ann Thorac Surg. 2005;80: Safi HJ, Letsou GV, Iliopoulos DC, Subramaniam MH, Miller CC III, Hassoun H, Asimacopoulos PJ, Baldwin JC. Impact of retrograde cerebral perfusion on ascending aortic and arch aneurysm repair. Ann Thorac Surg. 1997;63: Ergin MA, Uysal S, Reich DL, Apaydin A, Lansman SL, McCullough JN, Griepp RB. Temporary neurological dysfunction after deep hypothermic circulatory arrest: a clinical marker of long-term functional deficit. Ann Thorac Surg. 1999;67: Hagl C, Ergin MA, Galla JD, Lansman SL, McCullough JN, Spielvogel D, Sfeir P, Bodian CA, Griepp RB. Neurologic outcome after ascending aorta-aortic arch operations: effect of brain protection technique in high-risk patients. J Thorac Cardiovasc Surg. 2001;121: Bavaria JE, Pochettino A, Brinster DR, Gorman RC, McGarvey ML, Gorman JH, Escherich A, Gardner TJ. New paradigms and improved results for the surgical treatment of acute type A dissection. Ann Surg. 2001;234: Sinatra R, Melina G, Pulitani I, Fiorani B, Ruvolo G, Marino B. Emergency operation for acute type A aortic dissection: neurologic complications and early mortality. Ann Thorac Surg. 2001;71: Svensson LG, Crawford ES, Hess KR, Coselli JS, Raskin S, Shenaq SA, Safi HJ. Deep hypothermia with circulatory arrest. Determinants of stroke and early mortality in 656 patients. J Thorac Cardiovasc Surg. 1993;106: Griepp RB. Cerebral protection during aortic arch surgery. J Thorac Cardiovasc Surg. 2001;121: Deeb GM, Jenkins E, Bolling SF, Brunsting LA, Williams DM, Quint LE, Deeb ND. Retrograde cerebral perfusion during hypothermic circulatory arrest reduces neurologic morbidity. J Thorac Cardiovasc Surg. 1995; 109: Ehrlich MP, Fang WC, Grabenwoger M, Kocher A, Ankersmit J, Laufer G, Grubhofer G, Havel M, Wolner E. Impact of retrograde cerebral perfusion on aortic arch aneurysm repair. J Thorac Cardiovasc Surg. 1999;118: Safi HJ, Brien HW, Winter JN, Thomas AC, Maulsby RL, Doerr HK, Svensson LG. Brain protection via cerebral retrograde perfusion during aortic arch aneurysm repair. Ann Thorac Surg. 1993;56: Safi HJ, Iliopoulos DC, Gopinath SP, Hess KR, Asimacopoulos PJ, Bartoli S, Raskin SA, Shaibani AT, Leveque CM, Yawn DH. Retrograde cerebral perfusion during profound hypothermia and circulatory arrest in pigs. Ann Thorac Surg. 1995;59: Estrera AL, Miller CC III, Villa MA, Lee TY, Meada R, Irani A, Azizzadeh A, Coogan S, Safi HJ. Proximal reoperations after repaired acute type A aortic dissection. Ann Thorac Surg. 2007;83: Svensson LG, Nadolny EM, Penney DL, Jacobson J, Kimmel WA, Entrup MH, D Agostino RS. Prospective randomized neurocognitive and S-100 study of hypothermic circulatory arrest, retrograde brain perfusion, and antegrade brain perfusion for aortic arch operations. Ann Thorac Surg. 2001;71: Duebener LF, Hagino I, Schmitt K, Sakamoto T, Stamm C, Zurakowski D, Schafers HJ, Jonas RA. Direct visualization of minimal cerebral capillary flow during retrograde cerebral perfusion: an intravital fluorescence microscopy study in pigs. Ann Thorac Surg. 2003;75: Katz MG, Khazin V, Steinmetz A, Sverdlov M, Rabin A, Chamovitz D, Schachner A, Cohen AJ. Distribution of cerebral flow using retrograde versus antegrade cerebral perfusion. Ann Thorac Surg. 1999;67: Cockcroft DW, Gault MH. Prediction of creatinine clearance from serum creatinine. Nephron. 1976;16: Moon MR, Sundt TM III. Influence of retrograde cerebral perfusion during aortic arch procedures. Ann Thorac Surg. 2002;74: Matalanis G, Hata M, Buxton BF. A retrospective comparative study of deep hypothermic circulatory arrest, retrograde, and antegrade cerebral perfusion in aortic arch surgery. Ann Thorac Cardiovasc Surg. 2003;9: Bonser RS, Wong CH, Harrington D, Pagano D, Wilkes M, Clutton-Brock T, Faroqui M. Failure of retrograde cerebral perfusion to attenuate metabolic changes associated with hypothermic circulatory arrest. J Thorac Cardiovasc Surg. 2002;123: Tanoue Y, Tominaga R, Ochiai Y, Fukae K, Morita S, Kawachi Y, Yasui H. Comparative study of retrograde and selective cerebral perfusion with transcranial Doppler. Ann Thorac Surg. 1999;67:

7 S166 Circulation September 30, Gega A, Rizzo JA, Johnson MH, Tranquilli M, Farkas EA, Elefteriades JA. Straight deep hypothermic arrest: experience in 394 patients supports its effectiveness as a sole means of brain preservation. Ann Thorac Surg. 2007;84: Juvonen T, Zhang N, Wolfe D, Weisz DJ, Bodian CA, Shiang HH, McCullough JN, Griepp RB. Retrograde cerebral perfusion enhances cerebral protection during prolonged hypothermic circulatory arrest: a study in a chronic porcine model. Ann Thorac Surg. 1998;66: Li Z, Yang L, Jackson M, Summers R, Donnelly M, Deslauriers R, Ye J. Increased pressure during retrograde cerebral perfusion in an acute porcine model improves brain tissue perfusion without increase in tissue edema. Ann Thorac Surg. 2002;73: Kawata M, Sekino M, Takamoto S, Ueno S, Yamaguchi S, Kitahori K, Tsukihara H, Suematsu Y, Ono M, Motomura N, Morota T, Murakami A. Retrograde cerebral perfusion with intermittent pressure augmentation provides adequate neuroprotection: diffusion- and perfusion-weighted magnetic resonance imaging study in an experimental canine model. J Thorac Cardiovasc Surg. 2006;132: Estrera AL, Garami Z, Miller CC III, Sheinbaum R, Huynh TT, Porat EE, Winnerkvist A, Safi HJ. Determination of cerebral blood flow dynamics during retrograde cerebral perfusion using power M-mode transcranial Doppler. Ann Thorac Surg. 2003;76: Estrera AL, Garami Z, Miller CC III, Sheinbaum R, Huynh TT, Porat EE, Allen BS, Safi HJ. Cerebral monitoring with transcranial Doppler ultrasonography improves neurologic outcome during repairs of acute type A aortic dissection. J Thorac Cardiovasc Surg. 2005;129: Estrera A, Safi H. Repair of the transverse arch using retrograde cerebral perfusion during repair of acute type A aortic dissection. Operative Techniques in Thoracic and Cardiovascular Surgery, a Comparative Atlas. 2005;10: Anyanwu AC, Filsoufi F, Salzberg SP, Bronster DJ, Adams DH. Epidemiology of stroke after cardiac surgery in the current era. J Thorac Cardiovasc Surg. 2007;134: Estrera AL, Miller CM, Madisetty J, Bourgeois S, Azizzadeh A, Villa MA, Safi HJ. Ascending and transverse aortic arch repair: the impact of glomerular filtration rate on mortality. Ann Surg. 2008;247: