Improving Results of Open Arch Replacement

Improving Results of Open Arch Replacement Thoralf M. Sundt III, MD, Thomas A. Orszulak, MD, David J. Cook, MD, and Hartzell V. Schaff, MD Divisions of Cardiovascular Surgery and Anesthesiology, Mayo Clinic Rochester, Rochester, Minnesota ADULT CARDIAC Background. The rapid evolution of endovascular approaches to arch reconstruction such as brachiocephalic cerebral perfusion and to 6.0% (3 of 50) with selective and circulatory arrest to 21.1% (4 of 19) with retrograde debranching and endovascular stent grafting renders an antegrade cerebral perfusion p ( < 0.01), and for elective accurate understanding of contemporary outcomes of cases from 30.0% (6 of 20) with profound hypothermia conventional open arch surgery particularly relevant. and circulatory arrest to 14.3% (2 of 14) with retrograde Methods. Cases of arch reconstruction were identified cerebral perfusion, and 2.7% (1 of 37) with selective by search of the computerized cardiovascular surgical antegrade cerebral perfusion p ( < 0.01). The corresponding stroke rates were 19.2% (5 of 26) with profound database. Perioperative (30-day) outcomes as per The Society of Thoracic Surgeons database were evaluated. hypothermia and circulatory arrest, 5.3% (1 of 19) with Results. Between January 1, 1993, and June 30, 2007, 347retrograde cerebral perfusion, and 6.0% (3 of 50) with patients (195 male, 152 female; median age, 69 years; selective antegrade cerebral perfusion p ( 0.18) overall range, 21 to 88 years) underwent aortic arch replacement. and 15.0% (3 of 20) with profound hypothermia and Procedures were elective in 232 cases. Total arch replace-circulatory arrest, 7.1% (1 of 14) with retrograde cerebral ment was performed in 95, including 15 with concomi-perfusiontant replacement of the descending thoracic aorta by cerebral perfusion (p 0.46) for elective cases. and 5.4% (2 of 37) with selective antegrade means of bilateral thoracosternotomy. Modalities adjunctive to profound hypothermia and circulatory arrest can be accomplished under elective circumstances with Conclusions. Currently, open aortic arch replacement for cerebral protection have been introduced, including low operative mortality, particularly with adjunctive retrograde cerebral perfusion and, more recently, selective antegrade cerebral perfusion. The overall mortality cular therapies should be measured against contempo- measures for cerebral protection. The results of endovas- rate was 8.9% (elective procedures, 6.0%) and stroke raterary surgical series. was 8.4% (elective procedures, 6.9%). The mortality rate for total arch replacement has declined with adjuncts overall from 34.6% (9 of 26) with profound hypothermia (Ann Thorac Surg 2008;86:787 96) 2008 by The Society of Thoracic Surgeons Aortic arch replacement is a technically challenging mains the cornerstone for protection of the central nervous procedure, and one that has been associated in thesystem and other end organs. Adjunctive techniques have past with significant mortality and morbidity. Accordingly, been introduced with the aim of addressing the shortcomings of PHCA [4], as well as the need for conduct of more as success has been achieved with endovascular approaches to descending thoracic aortic aneurysmal disease, complex procedures to correct extensive aortic diseases [5]. attention has turned to the application of these technologiesretrograde cerebral perfusion (RCP) was introduced with to diseases of the aortic arch. These include debranchingpromising early results [6], and continues to be embraced procedures [1] and use of fenestrated or branched endovascular stent grafts [2]. To evaluate the appropriate applica- surgical community, however, whether substantial cerebral by many [7 9]. Controversy has built within the cardiac tion of these less invasive technologies, however, one needsflow is provided by RCP as evaluated in the laboratory [10], a clear understanding of the results that can be expectedor reflected in substantive clinical impact [11]. More recently, Kazui and associates [12] and Bachet and coworkers from open approaches using current techniques. Significant advances in open arch surgery have been [13] have championed selective antegrade perfusion made during the past several decades. Since the earliest(sacp) with excellent outcomes in their hands and those of series of arch replacements reported by Griepp and col-otherleagues [3], profound hypothermia and circulatory arrest der the protective affects of profound hypothermia in favor [14, 15]. Some surgeons have been willing to surren- (PHCA) has served our patients remarkably well and re-of only moderate systemic cooling in the interest of reducing perfusion times and coagulopathy associated with prolonged cardiopulmonary bypass [14]. In addition, use of the Accepted for publication May 5, 2008. Presented at the Forty-fourth Annual Meeting of The Society of Thoracic Surgeons, Fort Lauderdale, FL, Jan 28 30, 2008. Address correspondence to Dr Sundt, Division of Cardiovascular Surgery, Mayo Clinic, 200 First St, SW, Rochester, MN 55905; e-mail: with Boston Scientific. Dr Sundt discloses that he has a financial relationship sundt.thoralf@mayo.edu. 2008 by The Society of Thoracic Surgeons 0003-4975/08/$34.00 Published by Elsevier Inc doi:10.1016/j.athoracsur.2008.05.011

ADULT CARDIAC 788 SUNDT ET AL Ann Thorac Surg IMPROVING RESULTS OF OPEN ARCH REPLACEMENT 2008;86:787 96 Table 1. Patient Characteristics (n 347) Variable Total (n 347) PHCA (n 220) RCP (n 53) SACP (n 74) n (%) n (%) n (%) n (%) p Value Age, years (median) 69 (21 88) 65 (13) 71 (8) 64 (16) 0.05 Sex: female 152 (44) 89 (40) 27 (51) 36 (49) 0.25 Etiology Degenerative 195 (56) 134 (61) 28 (53) 33 (45) 0.04 Acute dissection 60 (17) 35 (16) 12 (23) 13 (18) 0.51 Chronic dissection 62 (18) 36 (16) 8 (15) 18 (24) 0.26 Marfan syndrome 12 (4) 6 (3) 0 (0) 6 (8) 0.03 Other 18 (5) 9 (4) 5 (9) 4 (5) 0.29 Renal failure 17 (5) 11 (5) 3 (6) 3 (4) 0.91 Hypertension 262 (76) 163 (74) 42 (79) 57 (77) 0.73 Hypercholesterolemia 178 (52) 114 (52) 25 (47) 39 (53) 0.79 Diabetes mellitus 25 (7) 17 (8) 1 (2) 7 (9) 0.23 Prior CVA 31 (9) 21 (10) 2 (4) 8 (11) 0.34 Prior MI 25 (7) 19 (9) 5 (9) 1 (1) 0.09 Prior cardiac surgery 105 (30) 67 (30) 12 (23) 26 (35) 0.32 EF 0.40 14 (5) 10 (6) 2 (5) 2 (4) 0.74 CVA cerebrovascular accident; EF ejection fraction; MI myocardial infarction; PHCA profound hypothermia and circulatory arrest; RCP retrograde cerebral perfusion; SACP selective antegrade cerebral perfusion. right axillary artery for arterial inflow has been introduced and also widely accepted [16, 17]. The two techniques are complementary, as axillary perfusion makes SACP quite straightforward technically. The summary result has been a remarkable reduction in mortality and morbidity for arch replacement [18]. With a view toward evaluating our current results with open aortic arch repair as a foundation for better determining which patients should receive traditional reconstruction versus stent graft repair, we reviewed our recent experience with arch replacement, focusing principally on the end points of mortality and stroke. During that interval our group has used PHCA alone, RCP, and SACP. Accordingly, the impact of these cerebral protective strategies as well as that of axillary inflow on these outcomes was examined. Material and Methods Cases performed by the Division of Cardiovascular Surgery at Mayo Clinic-Rochester have been cataloged Table 2. Operative Procedures (n 347) Variable Total (n 347) PHCA (n 220) RCP (n 53) SACP (n 74) n (%) n (%) n (%) n (%) p Value Elective status 232 (67) 155 (70) 32 (60) 45 (61) 0.17 Hemiarch 247 (71) 191 (87) 32 (60) 24 (32) 0.01 Total arch 95 (27) 26 (12) 19 (36) 50 (68) 0.01 Total arch descending 54 (16) 15 (7) 7 (13) 32 (43) 0.01 Cannulation Central 137 (40) 106 (48) 25 (47) 6 (8) 0.01 Femoral 116 (33) 84 (38) 28 (53) 4 (5) 0.01 Axillary 94 (27) 30 (14) 0 (0) 64 (86) 0.01 Circulatory arrest time (min) 29 ( 19) 27 ( 14) 33 ( 13) 41 ( 28) 0.01 CPB time (min) 157 ( 61) 141 ( 54) 176 ( 65) 188 ( 62) 0.01 Concomitant procedures Root replacement 86 (25) 54 (25) 15 (28) 17 (23) 0.78 AVR 36 (10) 25 (11) 2 (4) 9 (12) 0.23 MVR 2 (1) 1 (0) 0 (0) 1 (1) 0.57 CABG 92 (27) 62 (28) 12 (23) 18 (24) 0.64 Other 103 (30) 69 (31) 13 (25) 21 (28) 0.60 AVR aortic valve replacement; CABG coronary artery bypass grafting; CPB cardiopulmonary bypass; MVR mitral valve replacement; PHCA profound hypothermia and circulatory arrest; RCP retrograde cerebral perfusion; SACP selective antegrade cerebral perfusion.

Ann Thorac Surg SUNDT ET AL 2008;86:787 96 IMPROVING RESULTS OF OPEN ARCH REPLACEMENT 789 prospectively according to The Society for Thoracic Surgeons database criteria since January 1, 1993. After obtaining approval by the Mayo Clinic Rochester Institutional Review Board, this computerized database was searched retrospectively for patients undergoing partial or total arch replacement. A waiver of individual consent for this study was granted by the institutional review board on November 21, 2006, in accordance with 45 CFR 46.116 (d) as justified by the investigator, and waiver of HIPAA authorization in accordance with applicable HIPAA regulations. The committee determined that this constitutes a minimal risk collection of data or specimens that have already been collected for nonresearch purposes, and therefore was eligible for expedited review in accordance with 45 CFR 46.110 (b) (1) and 63 FR 60364, item 5. As of the time of this study, cases had been abstracted up to June 30, 2007. Patient demographics and comorbidities were derived from The Society of Thoracic Surgeons database. Thirty-day outcomes including mortality and stroke were the primary outcomes of interest. Operative notes were screened by one of the authors (T.M.S.) to assure consistency of recording circulatory arrest episodes and technical factors. Both cerebral circulatory arrest and corporeal circulatory arrest were considered in statistical analysis. ADULT CARDIAC Fig 1. (A) During the course of the study interval femoral cannulation (white bars) has been used less frequently in favor of axillary cannulation (gray bars). Central cannulation (black bars) continues to be used in a significant percentage of cases, particularly when atherosclerosis is thought to be unlikely and selective antegrade cerebral perfusion is not planned, as it is simpler and obviates the risks of brachial plexus injury with axillary cannulation. (B) Coincident with the rise in axillary cannulation, we have increasingly used selective antegrade cerebral perfusion (gray bars). (black bars profound hypothermia and circulatory arrest; white bars retrograde cerebral perfusion.) Surgical Technique There is no standardized approach to thoracic aortic disease within our institution. These procedures were performed by 12 surgeons, although greater than 90% were performed by the authors and one previous member of the division (K.J. Zehr). There are some general statements that can be made, however, which are applicable in most instances. The site of arterial cannulation is determined at the discretion of the operating surgeon. In general, axillary artery inflow is currently used in the repair of degenerative aneurysms in the interest of reducing the risk of atheroembolism from a central cannulation site. Axillary inflow provides for flow directed away from the cerebral circulation, particularly across anastomoses, at all phases of the operation. Axillary inflow is also preferred when total arch replacement is anticipated as it simplifies the institution of SACP. We routinely conduct axillary perfusion using an 8-mm sidearm graft rather than through direct arterial cannulation as this optimizes flows and reduces the risks of cannulation-induced vascular complications. Consonant with our institutional posture toward simplifying procedures whenever possible, however, we use central cannulation in the setting of connective tissue disease disorders such as Marfan s or in the setting of bicuspid aortic valve when hemiarch replacement is anticipated, as there is less concern regarding the atherosclerotic burden of the ascending aorta itself. This obviates the need for axillary dissection with the concomitant risks of brachial plexus injury or vascular complications, both of which are uncommon but can be catastrophic. The cornerstone of cerebral protection remains profound hypothermia. Cardiopulmonary bypass is routinely instituted at 2.2 to 2.5 L min 1 m 2. Cooling is conducted for a minimum of 30 minutes with a maximum temperature gradient of 10 C between the water entering the heat exchanger and the patient s nasopharyngeal temperature. Cooling is generally to a nasopharyngeal temperature of 16 to 18 C or an isoelectric processed electroencephalogram, and bladder temperature less than 23 C. More recently, particularly with the introduction of SACP, we have been willing to accept somewhat higher core temperatures (25 C) when hemiarch replacement is planned, particularly among patients with high body mass, although thus far we have been reluctant to reduce the minimum cooling time. During systemic cooling we use the ph-stat strategy for acid-base management until circulatory arrest. Our target hematocrit is 21% to 27% during cooling between 37 and 32 C, and 18% to 23% before circulatory arrest. During the arrest episode the head is

ADULT CARDIAC 790 SUNDT ET AL Ann Thorac Surg IMPROVING RESULTS OF OPEN ARCH REPLACEMENT 2008;86:787 96 Table 3. Mortality and Morbidity (n 347) Total (n 347) PHCA (n 220) RCP (n 53) SACP (n 74) Operative mortality 31 (9) 16 (7) 9 (17) 6 (8) 0.08 CVA 29 (8) 20 (9) 5 (9) 4 (5) 0.58 Renal failure 29 (8) 16 (7) 7 (13) 6 (8) 0.37 Reoperation for bleeding 32 (9) 13 (6) 7 (13) 12 (16) 0.02 Prolonged ventilation 113 (33) 66 (30) 19 (36) 28 (38) 0.40 Mortality and Morbidity for Elective Cases Only (n 232) Total (n 232) PHCA (n 155) RCP (n 32) SACP (n 45) Operative mortality 14 (6.0) 9 (5.8) 3 (9.4) 2 (4.4) 0.65 CVA 16 (6.9) 12 (7.7) 2 (6.3) 2 (4.4) 0.74 Renal failure 8 (3) 6 (4) 1 (3) 1 (2) 0.86 Reoperation for bleeding 22 (9) 10 (6) 5 (16) 7 (16) 0.08 Prolonged ventilation 58 (25) 36 (23) 13 (41) 9 (20) 0.08 Mortality and Morbidity for Emergency Cases Only (n 115) Total (n 115) PHCA (n 65) RCP (n 21) SACP (n 29) Operative mortality 17 (14.8) 7 (10.8) 6 (28.6) 4 (13.8) 0.13 CVA 13 (11.3) 8 (12.3) 3 (14.3) 2 (6.9) 0.67 Renal failure 21 (18) 10 (15) 6 (29) 5 (17) 0.39 Reoperation for bleeding 10 (9) 3 (5) 2 (10) 5 (17) 0.13 Prolonged ventilation a 55 (48) 30 (46) 6 (29) 19 (66) 0.03 a Prolonged ventilation defined as greater than 48 hours. CVA cerebrovascular accident; PHCA profound hypothermia and circulatory arrest; RCP retrograde cerebral perfusion; SACP selective antegrade cerebral perfusion. packed in ice to prevent ambient warming. Once rewarming has begun, alpha-stat management is initiated with a target hematocrit of 18% to 21% below 25 C and 24% to 30% above 32 C. During rewarming the maximum temperature gradient is again 10 C, and arterial inflow temperature does not exceed 37.5 C. Application of adjunctive cerebral protection measures such as RCP or SACP is at the discretion of the operating surgeon, although there is a general preference for the use of such adjuncts principally in the setting of total arch replacement as compared with hemiarch replacements. Retrograde cerebral perfusion was accomplished either by reversing flow in a venous return cannula placed in the superior vena cava or through a retrograde cardioplegia cannula placed in the superior vena cava. This technique has been essentially abandoned in our current practice, however, in favor of SACP. Selective antegrade perfusion is most often instituted by means of right axillary artery inflow as this obviates the need to cannulate the brachiocephalic trunk, which may clutter the operative field and introduces the possibility of vascular injury or atheroembolism with instrumentation. We have not perfused the left carotid artery routinely, as we believe our primary protection remains hypothermia and we have regarded SACP as an adjunct rather than substitute for the same. It is also our practice to reconstruct the arch vessels first, using either a trifurcated graft or a branched arch graft, such that once these anastomoses have been performed, antegrade flow in the left carotid and left subclavian is restored through the common lumen. During branch reconstruction or hemiarch replacement with SACP, the stumps of all three vessels are controlled with vascular clamps or, in selected cases in which stump pressures are to be measured, with balloon catheters. During SACP the perfusate is maintained at 13 C and flow through the axillary artery is maintained between 10 and 15 ml kg 1 min 1 based on current recommendations from the literature. Among those in whom prolonged episodes of SACP are anticipated, the flow rate has most recently been guided by pressures measured in the ipsilateral brachiocephalic artery and left carotid stumps with a goal to maintain an ipsilateral mean arterial pressure between 50 and 60 mm Hg and contralateral mean arterial pressure greater than 30 mm Hg. In 7 of the most recent total arch cases, brachiocephalic and left carotid stump pressures were measured using retrograde cardioplegia balloon catheters and recorded at 5, 10, and 15 ml kg 1 min 1.We do not routinely use cerebral oximetry devices as the

Ann Thorac Surg SUNDT ET AL 2008;86:787 96 IMPROVING RESULTS OF OPEN ARCH REPLACEMENT Table 4. Mortality and Morbidity for Total Arch Replacement (n 95) Total (n 95) PHCA (n 26) RCP (n 19) SACP (n 50) 791 ADULT CARDIAC Operative mortality 16 (16.8) 9 (34.6) 4 (21.1) 3 (6.0) 0.01 CVA 9 (9.5) 5 (19.2) 1 (5.3) 3 (6.0) 0.18 Mortality and Morbidity for Elective Total Arch Replacement (n 71) Total (n 71) PHCA (n 20) RCP (n 14) SACP (n 37) Operative mortality 9 (12.7) 6 (30.0) 2 (14.3) 1 (2.7) 0.01 CVA 6 (8.5) 3 (15.0) 1 (7.1) 2 (5.4) 0.46 Mortality and Morbidity for Emergency Total Arch Replacement (n 24) Total (n 24) PHCA (n 6) RCP (n 5) SACP (n 13) Operative mortality 7 (29.2) 3 (50.0) 2 (40) 2 (15.4) 0.30 CVA 3 (12.5) 2 (33.3) 0 (0) 1 (7.7) 0.24 CVA cerebrovascular accident; PHCA profound hypothermia and circulatory arrest; RCP retrograde cerebral perfusion; SACP selective antegrade cerebral perfusion. degree to which they meaningfully measure intracranial oxygen saturation is unproven in our opinion. Statistical Methods Descriptive statistics for categorical variables are reported as frequency and percentage whereas continuous variables are reported as mean (standard deviation) or median (range) as appropriate. Categorical variables were compared between perfusion groups using 2 test and continuous variables were compared using analysis of variance or Kruskal-Wallis rank-sum test as appropriate. Logistic regression models were used to find the univariate and multivariate predictors of operative mortality and postoperative stroke. The multivariable model considered univariately significant variables (p 0.05) with model selection using the stepwise method. All statistical tests were two-sided with the alpha level set at 0.05 for statistical significance. Results During the study interval, 347 patients underwent partial or complete arch replacement (Table 1). As expected the majority of aneurysms (56%) were attributable to degenerative disease, including bicuspid aortic valve disease. A slight majority (56%) of patients were male, with hypertension present in 76% and hypercholesterolemia in 52%. Nine percent of patients had a history of prior stroke. A higher percentage of cases using RCP were for acute dissection (23% versus 16% for PHCA and 18% for SACP), although this difference did not reach statistical significance. Two thirds of the procedures (67%) were performed electively (Table 2). Hemiarch replacement was preferred when technically feasible, being performed in 71% of cases. The remainder underwent total arch replacement with or without an elephant trunk. In only 15 patients was a transverse thoracosternotomy ( clam-shell ) incision required for replacement of a portion of the descending thoracic aorta. Concomitant aortic root replacement was performed in 25% of patients and coronary artery bypass grafting in a similar number (27%) of cases. During the course of the study as a whole, the site of arterial inflow has been relatively evenly distributed among central, femoral, and axillary sites; however, the trend with time has been away from femoral cannulation (Fig 1A). We continue to use central cannulation in a significant percentage of cases because of its simplicity as noted above. Like the site of cannulation, the modality of cerebral protection was at the discretion of the operating surgeon. Coincident with the shift toward axillary cannulation during the course of our experience, there has been a shift toward the use of SACP (Fig 1B). During the last 5 years, approximately 1 of 3 arch cases were performed using this technique; however, usage has been increasing during this interval and it is currently used in an even higher percentage of cases. As shown in Table 2, the patients undergoing SACP less often underwent hemiarch replacement and far more often had axillary cannulation (Table 2). When SACP was used, circulatory arrest times (to the corpus) were significantly longer (41 28 minutes versus 27 14 minutes for PHCA) as were cardiopulmonary bypass times (188 62 minutes versus 141 54 minutes for PHCA). The upper descending thoracic aorta was more often addressed in some manner using SACP, but there were no other significant difference in concomitant procedures. Operative outcomes are shown in Table 3. When considered according to perfusion strategy, the mortality rate overall was similar whether PHCA or SACP was used; however, it was significantly higher with RCP. The

ADULT CARDIAC 792 SUNDT ET AL Ann Thorac Surg IMPROVING RESULTS OF OPEN ARCH REPLACEMENT 2008;86:787 96 Fig 2. (A) Mortality ( ) and stroke ( ) by circulatory arrest time for profound hypothermia and circulatory arrest (A), retrograde cerebral perfusion (B), and selective antegrade cerebral perfusion (C) patients are shown. The risk of operative mortality and stroke rise with arrest time when profound hypothermia and circulatory arrest alone is used. This effect is less apparent when retrograde cerebral perfusion is used. Selective antegrade cerebral perfusion appears to neutralize the impact of arrest time on mortality and stroke. difference between RCP and PHCA groups likely in part reflects the slight overrepresentation of acute dissections in the former group (Table 1), although the same trend toward higher mortality was present among elective cases as well. Total arch replacement was also more commonly performed in the RCP than PHCA group (36% versus 12%) as shown in Table 2. In contrast, the similarity in mortality rate between SACP and PHCA is particularly notable in this regard as total arch replacement was performed in 68% of SACP patients (Table 2). There was a slight trend overall for a somewhat lower stroke rate with SACP (Table 3), although it must be acknowledged that there is always the possibility that stroke rate is underestimated owing to unrecognized strokes among those patients with perioperative mortalities who never awakened. Reoperation for bleeding was more common in the SACP group, perhaps because of the higher frequency of total arch replacement. When stratified according to operative status (Table 3), operative mortality was more than double under nonelective circumstances regardless of perfusion strategy, and postoperative stroke was somewhat higher. This is likely a reflection of the significant representation of acute dissection in nonelective cases. Under elective circumstances, there were no statistically significant differences among perfusion strategies, likely because of small numbers, despite the aforementioned differences in extent of procedure. There was a trend toward lower stroke rate in the SACP group (4.4% versus 6.3% with RCP and 7.7% with PHCA). Among emergent cases both mortality and stroke were double that of elective cases, but again there were no statistically significant differences according to perfusion strategy. Because procedure types differed among perfusion strategy groups with hemiarch replacement most often conducted in our practice with PHCA alone, we focused on total arch replacement as one would anticipate arrest times to be longest here and the impact of perfusion strategy to matter most (Table 4). Overall, operative mortality was statistically significantly lower when SACP was used despite small numbers of cases performed. The difference in stroke was not statistically significant, although there was a trend for lower stroke rates with adjuncts. Among elective cases alone, again SACP was associated with lower mortality rate (2.7% versus 30.0% for PHCA and 14.4% for RCP; p 0.01) Again, numbers are small and statistical significance was not achieved; however, the stroke rate also trended lower with SACP (5.4%) versus PHCA (15.0%) and RCP (7.1%). Under emergent circumstances the mortality rate with SACP was high (15.4%), but lower than for PHCA (50%) or RCP (40%). Considered in another way, if one examines mortality rate and stroke rate stratified by perfusion strategy and arrest times, there is a gradual increase in both adverse outcomes with arrest time when PHCA alone is used (Fig 2A). This trend is less apparent with RCP (Fig 2B) and appears to resolve with SACP (Fig 2C). Although again numbers are small, this supports a protective effect of

Ann Thorac Surg SUNDT ET AL 2008;86:787 96 IMPROVING RESULTS OF OPEN ARCH REPLACEMENT Table 5. Univariate Analysis for Operative Mortality Variable OR Lower 95% CL Upper 95% CL p Value Female 1.22 0.59 2.56 0.59 Hypertension 2.3 0.78 6.77 0.13 Emergency 2.7 1.28 5.7 0.01 SACP a 1.12 0.42 2.99 0.81 RCP a 2.61 1.08 6.28 0.03 Acute 1.77 0.75 4.16 0.19 CVA 0.68 0.15 3.00 0.61 Previous surgery 0.53 0.21 1.32 0.17 Axillary cannulation 0.62 0.25 1.57 0.31 Total arch b 3.13 1.48 6.63 0.01 Age 1.02 0.99 1.06 0.15 Perfusion time 1.01 1.00 1.02 0.01 Circulatory arrest time 1.02 1.01 1.04 0.01 Multivariate Analysis for Operative Mortality Emergency 2.81 1.26 6.28 0.01 Total arch b 2.85 1.25 6.51 0.01 Perfusion time 1.01 1.00 1.01 0.03 793 RCP was a significant predictor of death versus PHCA. By multivariate analysis, emergent status, total arch versus hemiarch, and perfusion time remained significant predictors. For stroke, only age was predictive by univariate analysis. Notably, there were no strokes among patients with a prior history of cerebrovascular events. The results of measurements of brachiocephalic and left carotid artery stump pressures during selective antegrade perfusion through axillary cannulation are shown in Figure 3. There was significant heterogeneity both in the ipsilateral stump pressures and in the contralateral pressures. At a given flow rate, the ipsilateral brachiocephalic pressure could vary by as much as a factor of 2, suggesting that reliance on flow rates alone may be inadequate, risking either hypoperfusion or hyperperfusion. Furthermore, the correlation between ipsilateral brachiocephalic and contralateral carotid pressures was poor, suggesting that unilateral perfusion may be inadequate if one is depending on SACP entirely for cerebral protection, as may be the case with the increasing popularity of only modest hypothermic corporeal cooling. ADULT CARDIAC Univariate Analysis for Postoperative Stroke Female 1.92 0.89 4.15 0.10 Hypertension 2.11 0.71 6.25 0.18 Emergency 1.72 0.8 3.71 0.17 SACP a 0.57 0.19 1.73 0.32 RCP a 1.04 0.37 2.92 0.94 Acute 1.59 0.65 3.91 0.31 Previous surgery 0.87 0.37 2.03 0.74 Axillary cannulation 0.84 0.35 2.05 0.71 Total arch b 1.19 0.52 2.71 0.68 Age 1.05 1.01 1.09 0.02 Perfusion time 1.00 0.99 1.00 0.40 Circulatory arrest time 1.01 1.00 1.03 0.12 a Versus PHCA. b Versus hemiarch. CI confidence interval; CVA cerebrovascular accident; OR odds ratio; PHCA profound hypothermia and circulatory arrest; RCP retrograde cerebral perfusion; SACP selective antegrade cerebral perfusion. SACP permitting safe conduct of complex operations requiring prolonged corporeal arrest times. Only a small number of patients (n 15) required repair of the ascending aorta, aortic arch, and descending thoracic aorta through transverse thoracosternotomy. Equal numbers were performed under elective and nonelective circumstances. Operative mortality was 1 of 15 (6.7%). The mean time on the ventilator was 61.9 hours and 5 of 15 (33.3% remained mechanically ventilated for greater than 72 hours. Predictors of operative mortality by univariate analysis (Table 5) included emergent status (p 0.01), total arch replacement (p 0.01 versus hemiarch replacement), and perfusion time as well as arrest time. Use of Comment The results of this study demonstrate that open aortic arch replacement can be undertaken under elective circumstances using contemporary techniques with low operative mortality and morbidity. Furthermore, although simple PHCA suffices in many cases, there appears to be benefit to the application of SACP among those patients in whom prolonged episodes of circulatory arrest are required to perform complex repairs. Indeed this technique for cerebral protection permits corporeal arrest episodes well in excess of the traditional safe episode, enabling careful elephant trunk repairs or simultaneous descending thoracic aortic replacements with cerebral arrest times as low as the single digits. In addition, although the impact of SACP during procedures with relatively low circulatory arrest times such as hemiarch replacement appears negligible for mortality, there is a trend for reduction in stroke. This may be attributable to axillary cannulation, which is, in our practice, most often used to accomplish SACP. Finally, given the current interest in the substitution of SACP for profound systemic cooling as a primary means of cerebral protection, our data suggest that perfusion strategies guided only by flow rate may risk significant hypoperfusion or hyperperfusion, and unilateral perfusion techniques may inadequately supply the contralateral cerebral hemisphere. These data suggest that direct measurement of cerebral perfusion pressures may be important as the safety margin of hypothermia is surrendered and higher perfusion temperatures are contemplated. Antegrade perfusion of the cerebral vasculature during arch replacement is not a novel concept, having been used by the pioneers of arch surgery [19, 20]. Indeed, given the burgeoning interest in SACP with

ADULT CARDIAC 794 SUNDT ET AL Ann Thorac Surg IMPROVING RESULTS OF OPEN ARCH REPLACEMENT 2008;86:787 96 Fig 3. In 7 patients undergoing selective antegrade cerebral perfusion, stump pressures in the brachiocephalic trunk and left carotid artery were measured. With increasing flow rates, ipsilateral pressures rise; however, there is wide variation in the associated stump pressure at any given flow, raising concerns as to the optimal rate. Furthermore, there is a significant gradient between the brachiocephalic and left carotid stump pressures, raising concerns about the adequacy of left hemispheric perfusion with unilateral inflow if higher temperatures are adopted. only mild hypothermic systemic perfusion and cannulation of individual brachiocephalic vessels, it is humbling to recognize that PHCA was heralded as an advance principally because it simplified the operation, freeing the operative field from multiple cannulas [3]. This advantage to PHCA remains as true today as before, and is the principal advantage cited by those who continue to advocate it as the preferred method for cerebral protection. The Yale group argues cogently that in cases in which the arrest episode can be expected to be brief and for those surgeons who perform arch surgery infrequently, PHCA is the simplest and most straightforward approach [21]. The vast majority of acute aortic dissections treated in the United States with hemiarch replacement likely fall into this group, and we would agree that in this setting simplicity is of particular value. We also note that we are reluctant to perform total arch replacement in the setting of acute dissection, a position supported by the data presented here. Despite the advantage of technical simplicity for the surgeon, the limitations of PCHA alone as a strategy for cerebral protection have become increasingly apparent as operative mortality has declined and attention has turned to preservation of neurologic function. In the aforementioned series from Yale, when arrest episodes exceed 40 minutes, the stroke rate exceeded 13%. Arrest times longer than 40 minutes also approached statistical significance (p 0.06) for mortality (odds raio, 1.56) [21]. The incidence of transient neurologic dysfunction, itself a marker for subtle neurologic damage, has also been shown to correlate with duration of circulatory arrest [4]. Furthermore, surgeons are increasingly challenged to repair more complex diseases which demand longer periods of corporeal arrest [5]. In this regard SACP has emerged as an enabling technology; the recent literature is replete with reports, some quite large, of remarkably low operative mortality rates and stroke rates despite complex aortic arch reconstructions [15]. Our practice has been influenced by these trends, particularly since the popularization of axillary artery inflow [16, 22], which obviates the need for multiple cannulas in the operative field. We have thus far continued to cool systemically for a minimum of 30 minutes and to a target temperature of 18 to 20 C nasopharyngeal despite the suggestion by others that mild hypothermia will suffice [23]. Inour experience the oft-mentioned adverse effects of profound hypothermia have been less troublesome than others have suggested, with the principal disadvantage being the inconvenience of prolonged rewarming time. Accordingly we have treated SACP as an adjunct rather than a substitute for what has proven to be quite satisfactory neurologic protection. We have been reluctant to reduce cooling out of concern for the adequacy of contralateral cerebral perfusion given our current technique providing only unilateral inflow. The validity of this concern is born out by the clinical results reported with this approach in 1999 by Wozniak and colleagues [23]. They reported 2 cases of temporary left-sided hemiparesis among 21 patients treated with moderate hypothermic perfusion (28 C) and unilateral perfusion. Although the authors thought this was an adequate outcome, the rate of neurologic event (9.5%) is twice what we have observed with SACP and profound hypothermia. Our unease about unilateral perfusion stimulated us to measure left carotid stump pressures, and indeed the results support those concerns. The obvious alternative of simultaneous left carotid cannulation and perfusion remains an option as well, although this again clutters the operative field. These same measurements also demonstrated concerning variation between arterial inflow rates and stump pressures. Unfortunately there is a paucity of data in the literature to support recommended flow rates during SACP. Even more uncertainty arises out of the loss of cerebral autoregulation at low temperatures, making the risks of hyperperfusion as real as those of hypoperfusion. These findings support routine moni-

Ann Thorac Surg SUNDT ET AL 2008;86:787 96 IMPROVING RESULTS OF OPEN ARCH REPLACEMENT toring of cerebrovascular pressures as well as flow rates. The current study has limitations shared by all retrospective studies. Although it would be optimal to test strategies prospectively in a randomized manner, arch replacement in our institution remains relatively uncommon and such a study would be impractical. Although observational studies such as this may be subject to selection or treatment bias, if anything, one might anticipate SACP to be applied in higher risk cases. This should bias the results against SACP. In addition, other elements in our perioperative management strategy have changed with time. Finally, the study itself is based on the experience of a number of surgeons, which may be perceived as a strength or weakness. Although this introduces some inconsistency in approach, it makes it perhaps somewhat more representative of the generalizability of the approach. In summary, our current practice is to use axillary perfusion when undertaking repair of atherosclerotic aneurysms. Selective antegrade perfusion is used when total arch replacement is planned or a prolonged episode of circulatory arrest is otherwise anticipated. The operative risk is low with PHCA alone when circulatory arrest times are brief, as may be anticipated with hemiarch replacement. If in doubt, however, axillary perfusion permits rapid change in strategy to use SACP without complicating matters excessively. We have concerns regarding the adequacy of unilateral SACP alone as a neuroprotective strategy in the setting of only moderate hypothermia. The authors wish to acknowledge the tireless efforts of Judy Lenoch for her supervision of the database and work to extract the data presented here, and the efforts of Zhuo Li for her statistical analysis of the findings. We also acknowledge the significant clinical contributions of Kenton Zehr, MD, who performed many of the operative procedures described here while he was at Mayo Clinic Rochester. References 795 1. Czerny M, Gottardi R, Zimpfer D, et al. Mid-term results of supraaortic transpositions for extended endovascular repair of aortic arch pathologies. Eur J Cardiothorac Surg 2007;31: 623 7. 2. Saito N, Kimura T, Odashiro K, et al. Feasibility of the Inoue single-branched stent-graft implantation for thoracic aortic aneurysm or dissection involving the left subclavian artery: short- to medium-term results in 17 patients. J Vasc Surg 2005;41:206 12. 3. Griepp RB, Stinson EB, Hollingsworth JF, Buehler D. Prosthetic replacement of the aortic arch. J Thorac Cardiovasc Surg 1975;70:1051 63. 4. Ergin MA, Galla JD, Lansman L, et al. Hypothermic circulatory arrest in operations on the thoracic aorta. Determinants of operative mortality and neurologic outcome. J Thorac Cardiovasc Surg 1994;107:788 99. 5. Kouchoukos NT, Mauney MC, Masetti P, Castner CF. Optimization of aortic arch replacement with a one-stage approach. Ann Thorac Surg 2007;83(Suppl):S811 4; discussion S824 31. 6. Ueda Y, Miki S, Kusuhara K, et al. Deep hypothermic systemic circulatory arrest and continuous retrograde cerebral perfusion for surgery of aortic arch aneurysm. Eur J Cardiothorac Surg 1992;6:36 42. 7. Estrera AL, Garami Z, Miller CC 3rd, et al. Determination of cerebral blood flow dynamics during retrograde cerebral perfusion using power M-mode transcranial Doppler. Ann Thorac Surg 2003;76:704 10. 8. Appoo JJ, Augoustides JG, Pochettino A, et al. Perioperative outcome in adults undergoing elective deep hypothermic circulatory arrest with retrograde cerebral perfusion in proximal aortic arch repair: evaluation of protocol-based care. J Cardiothorac Vasc Anesth 2006;20:3 7. 9. Geirsson A, Bavaria JE, Swarr D, et al. Fate of the residual distal and proximal aorta after acute type a dissection repair using a contemporary surgical reconstruction algorithm. Ann Thorac Surg 2007;84:1955 64. 10. Ehrlich MP, Hagl C, McCullough JN, et al. Retrograde cerebral perfusion provides negligible flow through brain capillaries in the pig. J Thorac Cardiovasc Surg 2001;122: 331 8. 11. Moon MR, Sundt TM 3rd. Influence of retrograde cerebral perfusion during aortic arch procedures. Ann Thorac Surg 2002;74:426 31. 12. Kazui T, Washiyama N, Muhammad BA, et al. Improved results of atherosclerotic arch aneurysm operations with a refined technique. J Thorac Cardiovasc Surg 2001;121:491 9. 13. Bachet J, Guilmet D, Goudot B, et al. Cold cerebroplegia. A new technique of cerebral protection during operations on the transverse aortic arch. J Thorac Cardiovasc Surg 1991; 102:85 94. 14. Kamiya H, Hagl C, Kropivnitskaya I, et al. The safety of moderate hypothermic lower body circulatory arrest with selective cerebral perfusion: a propensity score analysis. J Thorac Cardiovasc Surg 2007;133:501 9. 15. Sasaki H, Ogino H, Matsuda H, et al. Integrated total arch replacement using selective cerebral perfusion: a 6-year experience. Ann Thorac Surg 2007;83(Suppl):S805 10; discussion S824 31. 16. Sabik JF, Nemeh H, Lytle BW, et al. Cannulation of the axillary artery with a side graft reduces morbidity. Ann Thorac Surg 2004;77:1315 20. 17. Strauch JT, Spielvogel D, Lauten A, et al. Axillary artery cannulation: routine use in ascending aorta and aortic arch replacement. Ann Thorac Surg 2004;78:103 8. 18. Spielvogel D, Etz CD, Silovitz D, et al. Aortic arch replacement with a trifurcated graft. Ann Thorac Surg 2007; 83(Suppl):S791 5; discussion S824 31. 19. Pearce CW, Weichert RF 3rd, del Real RE. Aneurysms of aortic arch. Simplified technique for excision and prosthetic replacement. J Thorac Cardiovasc Surg 1969;58:886 90. 20. De Bakey ME, Crawford ES, Cooley DA, Morris GC Jr. Successful resection of fusiform aneurysm of aortic arch with replacement by homograft. Surg Gynecol Obstet 1957;105: 657 64. 21. Gega A, Rizzo JA, Johnson MH, et al. Straight deep hypothermic arrest: experience in 394 patients supports its effectiveness as a sole means of brain preservation. Ann Thorac Surg 2007;84:759 67. 22. Svensson LG, Blackstone EH, Rajeswaran J, et al. Does the arterial cannulation site for circulatory arrest influence stroke risk? Ann Thorac Surg 2004;78:1274 84. 23. Wozniak G, Dapper F, Zickmann B, et al. Selective cerebral perfusion via innominate artery in aortic arch replacement without deep hypothermic circulatory arrest. Int J Angiol 1999;8:50 6. ADULT CARDIAC

ADULT CARDIAC 796 SUNDT ET AL Ann Thorac Surg IMPROVING RESULTS OF OPEN ARCH REPLACEMENT 2008;86:787 96 DISCUSSION DR Y. JOSEPH WOO (Philadelphia, PA): Did you have a larger number of type A dissections in the hemiarch RCP (retrograde cerebral perfusion) group? That is your highest mortality group (17%). DR SUNDT: I don t think I ve looked at it that way. So you re asking whether the distribution of acute dissections was higher in the RCP group or not. I ll have to look back at that to be honest with you. Well, yes, I guess I didn t obsess over it too much because we have pretty well abandoned RCP. It s of historical interest in my book. But I ll look back and look at that. It s a good suggestion. DR THOMAS G. GLEASON (Pittsburgh, PA): Most groups come to the same conclusion that your group has, that clearly antegrade cerebral perfusion offers a lower stroke rate and a better survival. The question you raise about its safety as [we] push [to] doing these cases at warmer temperatures (ie, moderate hypothermia) is quite a relevant one. One of my specific concerns, even with bilateral carotid perfusion, is as we warm the temperature, are we putting the spinal cord at greater risk? What are your thoughts about spinal cord protection and its adequacy with the use of antegrade cerebral perfusion and only moderate hypothermia? DR SUNDT: Yes, I agree with your concerns. And I m a bit of a little old lady about those things. It is taking me a while to get the confidence to warm the perfusate up. If it s a complex repair, something that we re doing through a clamshell or something like that, I like to be cold. That s the way Dr Kouchoukos taught me to work and I feel comfortable that way. I think the patients where I m tempted to go warmer and warmer are the hemiarch replacements, when you really have to ask yourself, Why did I cool so long when all I m doing is a hemiarch replacement with a very brief arrest episode? And so those are the ones that I m going warmer and warmer with. But I worry about the left side of the brain and especially subtle changes on the left side that may not manifest as a stroke but sort of a unilateral version of transient neurologic dysfunction. That can t be good for your head. DR T. BRUCE FERGUSON (Greenville, NC): So, Lord Kelvin, please, how warm is warm? DR SUNDT: How warm is warm? Well, I am not sure, but I think I know how cold cold is. And cold is 18 nasopharyngeal, so that s the target that I cool to. And I was taught to cool for at least 30 minutes. And I just am having a hard time backing off of that. That s too warm or too cold? DR FERGUSON: That s 291.15 in Kelvin s scale. DR JEFFREY H. SHUHAIBER (Cambridge, UK): As you are well aware, there is a school of thought that you could use retrograde cerebral perfusion right at the beginning and at the end of your arch repair. What do you think of that procedure? Doctor Sundt, I enjoyed your presentation. DR SUNDT: To flush debris? DR SHUHAIBER: Yes. In combination with selective antegrade cerebral perfusion. DR SUNDT: I think that that s a reasonable approach, but if you do selective antegrade from the axillary, you get the same effect. It blows blood and debris away from the head vessels. So that s another reason why I actually like the selective antegrade. DR ANTONIO LAUDITO (Turin, Italy): Regarding the surgical strategy for the open hemiarch anastomosis: would you be comfortable to keep the temperature not lower than 26 C perfusing the left, the right carotid and having a catheter going down in the descending aorta, like a simple Foley, that will perfuse the lower part of the body? DR SUNDT: There are others who feel that way, who are interested in perfusing both carotids and the descending with separate cannulae. Actually, that is why I put up that quote from Dr Ecker. Isn t it remarkable to remember our history, our surgical history, our surgical heritage as we look at what we do today. The approach with multiple perfusion catheters is the way Dr Cooley and Dr Debakey did the first aneurysms! And when Dr Griepp introduced profound hypothermia, it was welcomed as a way to simplify the operation. So to some degree we re going backwards repeating the past. In the end I think you have to balance what s optimal for you and for your patient s circumstances. For me, a hemiarch replacement with a lot of catheters in the lumen is a pain in the neck. So if you re going to be able to do the anastomosis relatively quickly, it s probably simpler to just cool a bit more and do it without any kind of perfusion. Especially if the patient is not too big and rewarming will not take too long. But there are lots of ways to skin a cat.