with aneurysms of other causes; AXC was associated aorta/root repairs (1995 to 2005), principally for atherosclerotic

ORIGINAL ARTICLES: SURGERY: The Annals of Thoracic Surgery CME Program is located online http://cme.ctsnetjournals.org. at To take the CME activity related to this article, you must have either an STS member or an individual non-member subscription to the journal. Axillary Cannulation Significantly Improves Survival and Neurologic Outcome After Atherosclerotic Aneurysm Repair of the Aortic Root and Ascending Aorta Christian D. Etz, MD, Konstadinos A. Plestis, MD, Fabian A. Kari, MS, Daniel Silovitz, MS, Carol A. Bodian, DrPh, David Spielvogel, MD, and Randall B. Griepp, MD Mount Sinai School of Medicine, New York, New York Background. The impact of axillary artery cannulation (AXC) on survival and neurologic outcome after operation for ascending aortic disease was retrospectively evaluated. Methods. We reviewed 869 patients with ascending with aneurysms of other causes; AXC was associated aorta/root repairs (1995 to 2005), principally for atherosclerotic and degenerative aneurysms and chronic and patients taken together. with a significantly better outcome p ( 0.05) in the 869 acute type A dissections. Arterial cannulation was Conclusions. AXC resulted in superior survival and through the ascending aorta (AAC) in 157 patients, theneurologic outcome in patients with atherosclerotic an- and a marginally better outcome than with femoral artery (FAC) in 261, and the right axillary arteryeurysms (AXC) in 451. Patients cannulated at different sites werecannulation at other sites during proximal aortic procedures for all causes. This study supports AXC in patients compared for preoperative comorbidities and outcomes (mortality and stroke) for each cause. with atherosclerotic disease who require complex cardiothoracic operations and in patients requiring proximal Results. Of the 122 patients with atherosclerotic aneurysms, 66 with right AXC had significantly better out-aortic intervention regardless of cause. comes (p 0.02): 64 of 66 survived vs 24 of 26 with FAC and 27 with 30 of AAC; no strokes occurred (vs 2 of 26 with FAC and 4 of 30 with AAC). No significant advantage for AXC was found with ascending aortic operation in 495 degenerative aneurysms, 106 chronic, or 65 acute type A dissections, 41 patients with endocarditis, or in 18 (Ann Thorac Surg 2008;86:441 7) 2008 by The Society of Thoracic Surgeons I njury to the central nervous system remains one of therotic aneurysm disease, central cannulation enhances the major causes of morbidity and mortality after proximal aortic and arch operations, affecting not only qualityresulting in focal lesions producing neurologic injury. potential for embolization into the cerebral circulation, of life postoperatively but also resulting in prolonged When the ascending aorta is unsuitable, cannulation of hospitalization and increased cost of treatment. The the femoral artery (FAC) is a commonly used alternative. cause of most major cerebral insults after ascending/ But retrograde flow in a severely atherosclerotic and aortic root reconstruction is stroke. Two influences on thediseased aorta poses major risks, including dislodgement incidence of stroke after replacement of the aortic archof plaques and aortic dissection, both of which may lead and ascending aorta are the site of cannulation forto cerebral as well as peripheral injury [1 3]. cardiopulmonary bypass (CPB) and perfusion technique. For these reasons, cannulation of the axillary artery The preferred site of cannulation for CPB is usually the (AXC) has become increasingly widespread and is more ascending aorta, but in patients with severe atheroscle- frequently being used for ascending aorta/root repair 4]. [ Accepted for publication Feb 27, 2008. Axillary artery cannulation preserves antegrade flow in the descending aorta while eliminating some of the risks Presented at the Fifty-fourth Annual Meeting of the Southern Thoracic Surgical Association, Bonita Springs, FL, Nov 7 10, 2007. associated with direct cannulation of the ascending aorta (AAC). It lowers the potential for embolization into Address correspondence to Dr Etz, Department of Cardiothoracic Surgery, Mount Sinai School of Medicine, One Gustave L. Levy Pl, New York, right-sided cerebral vessels by perfusing them with flow NY 10029; e-mail: christian.etz@mountsinai.org. that has not traversed the arch. Axillary artery cannula- 2008 by The Society of Thoracic Surgeons 0003-4975/08/$34.00 Published by Elsevier Inc doi:10.1016/j.athoracsur.2008.02.083

442 ETZ ET AL Ann Thorac Surg AXILLARY CANNULATION EFFECTS ON OUTCOME 2008;86:441 7 Table 1. Clinical Profile of Patients Demographics All Aorta Femoral Axillary p Value Patients, No. 869 157 (18.0) 261 (30.0) 451 (51.9) Age, mean SD years 58.1 17.5 55.7 19.3 56.1 17.6 60.0 16.5 0.003 Age 60 years, No. (%) 449 (51.7) 73 (46.5) 125 (47.9) 251 (55.7) 0.05 Male sex, No. (%) 639 (73.5) 105 (66.9) 199 (76.3) 335 (74.3) 0.1 Previous cardioaortic procedures, No. (%) 184 (21.2) 33 (21.0) 63 (24.1) 88 (19.5) 0.35 Timing of operation, No. (%) 0.24 Elective 714 (82.2) 131 (83.4) 196 (75.1) 387 (85.8) Urgent 75 (8.6) 17 (10.8) 28 (10.7) 30 (6.7) Emergency 80 (9.2) 9 (5.7) 37 (14.2) 34 (7.5) Risk factors, No. (%) History of neurologic dysfunction 48 (5.57) 9 (5.8) 15 (5.8) 24 (5.3) 0.96 History of hypertension 444 (51.2) 74 (47.4) 129 (49.6) 241 (53.4) 0.36 Coronary artery disease 200 (23.1) 36 (24.4) 45 (17.3) 117 (25.9) 0.03 COPD 27 (3.1) 3 (1.9) 2 (0.8) 22 (4.9) 0.006 Diabetes 47 (5.47) 11 (7.1) 13 (5.0) 23 (5.1) 0.61 Clot or atheroma 43 (5.07) 8 (5.1) 16 (6.2) 19 (4.2) 0.51 COPD chronic obstructive pulmonary disease. tion also eliminates the sandblasting effect of turbulent flow from a catheter tip close to atherosclerotic lesions in the ascending aorta or aortic arch, and thus also reduces the risk of embolization into left-sided cerebral vessels. Arterial inflow through the axillary artery also increases the ease of using selective cerebral perfusion during arch repairs, which allows the surgeon to construct open distal anastomoses while the lower body is kept hypothermic during circulatory arrest. Not surprisingly, AXC has been advocated as particularly suitable for CPB in patients with severe atherosclerotic disease prohibiting femoral or direct aortic cannulation [1, 5]. Indirect AXC through a side graft has been demonstrated to reduce perfusion-related morbidity and stroke risk after complex cardioaortic operations that necessitate circulatory arrest [6]. We undertook this retrospective evaluation of the influence of direct AXC and its effect on survival and neurologic outcome after aortic root/ascending aortic repairs for ascending aortic disease of differing causes in 869 patients requiring proximal aortic repair. Patients and Methods From March 1990 to May 2005, 869 patients who underwent aortic root/ascending aorta repairs were identified Table 2. Cause Cause No. Aorta Femoral Axillary Degenerative 495 17% 27% 56% Atherosclerosis 122 25% 21% 54% Chronic dissection 106 8% 44% 47% Acute dissection 65 5% 48% 48% Endocarditis 41 39% 32% 29% Other 40 40% 23% 38% from a prospectively compiled database. All patients who underwent this operation in whom the site of cannulation could be ascertained from the database were included. This study was approved by the Institutional Review Board of the Mount Sinai Hospital and included a waiver of informed consent for chart review and follow-up. An urgent or emergency proximal aortic repair was done in 80 patients. The overall mean age was 58 18 years; 589 patients were men (67.7%), and 407 (46.8%) were aged 60 years or older. Of these, 184 patients (21%) had undergone previous cardioaortic procedures. Table 1 reports the clinical profiles of the patients in detail, as well as potential risk factors for death and stroke, allowing comparison of those who underwent AXC with those who had AAC or FAC for CPB. The definition of the type of aneurysm was based on the surgeon s classification at the time of operation. A diagnosis of atherosclerotic aneurysm was generally reached as a result of identification of atherosclerotic plaques in the aorta intraoperatively from observation in the open aortic arch or by means of echocardiography, computed tomography scan, or magnetic resonance imaging. Other factors leading to a diagnosis of atherosclerotic aneurysms included evidence of atherosclerotic disease elsewhere, such as in the coronary, carotid or ileofemoral arteries, and confirmation by surgical pathology. Aneurysms not classified as atherosclerotic or attributable to another specific cause, and which were not associated with either acute or chronic dissection, were classified as degenerative. Indications The principal indication for aortic root/ascending aorta reconstruction in 495 patients was a degenerative aortic aneurysm (Table 2). An atherosclerotic aortic aneurysm was present in 122 patients, chronic dissections in 106, and acute dissections in 65. Aortic valve endocarditis

Ann Thorac Surg ETZ ET AL 2008;86:441 7 AXILLARY CANNULATION EFFECTS ON OUTCOME 443 Fig 1. Cannulation of the right axillary artery. affected the root in 14 patients, and 40 patients had other pathologic conditions. Cannulation and Myocardial Protection Through 1998, cannulation for CPB was through the femoral artery or ascending aorta in most patients. Beginning in 1999, right AXC was used with increasing frequency (Fig 1). Venous cannulation was usually through a two-stage catheter in the right atrium, but in some patients in whom the heart was not fully exposed, the right atrium was accessed through a wire-directed catheter placed through the femoral vein. Antegrade crystalloid or blood cardioplegia was infused directly into the coronary ostia; topical cooling was also used, and in patients with severe coronary disease, retrograde blood cardioplegia was used. Cardioplegia was administered every 20 to 30 minutes during periods of myocardial ischemia. A vent was placed in the left ventricle to allow decompression of the heart. Hypothermic Circulatory Arrest Hypothermic circulatory arrest (HCA) was routinely used to permit open distal anastomosis and complete ascending aorta or hemiarch resection. HCA was effected by surface and perfusion cooling. Adequate cerebral cooling was ensured by cooling to an esophageal temperature of 12 to 15 C and maintaining a jugular venous saturation greater than 95%. The mean HCA interval was 21 minutes (range, 12 to 28 minutes). The head was packed circumferentially in ice. Supplemental antegrade and retrograde cerebral perfusion were not used in this patient group. Perfusion warming was performed at the end of the procedure with the gradient between the esophageal and blood temperature maintained at less than 10 C. Warming was maintained until the esophageal temperature reached 35 C and bladder temperature exceeded 32 C. Technique for Axillary Artery Cannulation Our technique for direct AXC has previously been described in detail [7]. Briefly, arterial pressure is routinely measured with a left radial artery cannula. Before the median sternotomy, a 6-cm transverse skin incision is made approximately 1 cm below the middle and lateral part of the right clavicle (deltopectoral groove). Following the direction of its fibers, the pectoralis major muscle is separated and the underlying pectoralis minor muscle retracted laterally. The axillary artery is identified by palpation and then gently mobilized by sharp dissection without touching the medial and lateral brachial plexus cords posterior to the artery. The artery is then controlled with loops of silicone elastomer tape. After the administration of heparin, the axillary artery is occluded distally with a silicone elastomer vessel loop, and a transverse or longitudinal arteriotomy is done. The axillary artery is then cannulated directly using a 20F to 26F wire-reinforced right-angled flexible cannula (axillary access arterial cannula, Edwards Lifescience LLC, Irvine, CA). After proximal clamp removal, the cannula is advanced 3 cm into the artery, and the snare on the vessel tape is tightened. The cannula is held in place by a ligature on the snare and a skin stitch at the lateral end of the incision. Free backflow of blood is assured before perfusion is initiated. At the end of the procedure, the artery is repaired with a 6-0 polypropylene continuous suture. Aortic Root Reconstruction A button Bentall operation was done in 91% of patients in this study, with a modification of the original technique described by Kouchoukos and coworkers [8] in 1991. The Cabrol technique was used in 4%, and 5% of patients had a classic Bentall procedure. Almost all operations were performed with an open distal anastomosis, often with hemiarch replacement. Our current technique and its rationale for using either a mechanical or biologic conduit [9] have been described in detail elsewhere [10]. Statistical Methods Data were entered in an Excel spreadsheet (Microsoft Corp, Redmond, WA) and transferred to SAS software (SAS Institute, Cary, NC) for data description and analysis. Characteristics and risk factors in this sample of patients are described as percentages or as means and standard deviations and compared among the cannulation sites by 2 tests or analysis of variance, respectively. Patients cannulated at different sites were compared for the hospital outcomes of death and stroke for each of the aforementioned indications. The 2 tests for trend within etiologic groups were used to compare aortic, femoral, and axillary cannulation sites for death (with or without stroke), stroke survival, and uneventful recovery. Signif-

444 ETZ ET AL Ann Thorac Surg AXILLARY CANNULATION EFFECTS ON OUTCOME 2008;86:441 7 Table 3. Neurologic Outcome and Mortality by Cause and Cannulation Site Outcome All Aorta Femoral Axillary p Value Degenerative aneurysms, No. 495 83 135 277 Stroke, % 1 1 1 1 Death, % 1... 2 1 Adverse outcome, % 2 1 3 2 0.65 Atherosclerotic aneurysms, No. 122 30 26 66 Stroke, % 5 13 8... Death, % 6 10 8 3 Adverse outcome, % 9 20 12 3 0.02 a Chronic dissection. No. 106 9 47 50 Stroke, % 3 11 4 0 Death, % 10 11 9 12 Adverse outcome, % 13 22 13 13 0.87 Acute dissection, No. 65 3 31 31 Stroke, % 3... 3 3 Death, % 12... 19 6 Adverse outcome, % 14... 19 10 0.46 Acute root endocarditis, No. 41 16 13 12 Stroke, %............ Death, % 10 13 8 8 Adverse outcome, % 10 13 8 8 0.89 Other c root pathologies, No. 40 16 9 15 Stroke, %............ Death, % 8 13 11... Adverse outcome, % 8 13 11... 0.11 All causes, No. 869 157 261 451 Stroke, % 2.0 3.8 2.7 0.9 Death, % 4.6 5.1 6.5 3.3 Adverse outcome, % 6.0 8.3 8.1 4.0 0.05 a p 0.03 when data are stratified by years 1990 1998, 1999 2002, and 2003 2007. Adverse outcome percentage is less than the sum of stroke percentage and death percentage when some patients had a stroke and then died. Significance levels pertain to 2 tests for trend among the outcomes (death with or without stroke, stroke survivors, stroke free survival). icant findings were checked by stratifying the data on groups of calendar years representing different periods of clinical practice 1990 to 1998, 1999 to 2002, and 2003 to 2007 and by comparing axillary and nonaxillary cannulation for the frequency of the adverse outcome of stroke or death. Results Arterial cannulation was accomplished directly using AAC in 157 patients, FAC in 261, and right AXC in 451. The AXC patients were older and more likely to have coronary artery disease or chronic obstructive pulmonary disease, or both, compared with the AAC and FAC patients. Outcome Overall hospital mortality defined as death in the hospital or within 30 days postoperatively was 4.6%. Adverse outcome was defined as postoperative death or stroke within 30 days after operation, with postoperative deaths occurring before discharge from the hospital included as adverse outcome even if the hospital stay extended beyond 30 days. An adverse outcome occurred in 6% of patients, and 2% had permanent strokes. In 122 patients with atherosclerotic aneurysms, AXC in 66 was associated with a significantly better outcome (p 0.02, Table 3). Of patients with AXC, 64 of 66 (97%) survived vs 24 of 26 (92%) with FAC and 27 of 30 (90%) with AAC. None of the patients with AXC had a stroke, in contrast to 2 of 26 (8%) with FAC and 4 of 30 (13%) with AAC. Overall, adverse outcome was seen in 2 of 66 patients who had AXC for atherosclerotic disease vs 9 of 56 with non-axc (p 0.01). Although some demographic variability in possible risk factors for death or stroke are apparent among patients who underwent cannulation at different sites (Table 1), no consistent differences suggesting decreased risk in the AXC group could be demonstrated. In contrast with the results in patients with atherosclerotic aneurysms, no significant advantage for AXC was found with ascending aortic procedures for degenerative aneurysm in 495, chronic type A dissection in 106, acute type A dissection in 65, endocarditis in 41, or aortic root/ascending aorta operations in 40 patients

Ann Thorac Surg ETZ ET AL 2008;86:441 7 AXILLARY CANNULATION EFFECTS ON OUTCOME 445 with other etiologies. When the 869 patients were analyzed together, however, AXC conferred a significant advantage in averting an adverse outcome compared with cannulation at other sites (p 0.05). The overall results, as well as neurologic outcome and mortality for each indication according to cannulation site, are summarized in Table 3. Comment This retrospective review of patients who underwent operations on the aortic root or ascending aorta using different sites for cannulation for CPB demonstrated a significantly better outcome in patients with atherosclerotic aneurysms with AXC than with either FAC or AAC. The discovery that AXC is of significant benefit in the patients with atherosclerotic aneurysms, but is not as clearly linked with a more favorable outcome in the patients with other aortic pathologies, is not altogether surprising, because the risk of stroke is much higher in the patients with atheromatous disease. The results suggest that AXC should perhaps be recommended for use in other complex cardiac operations in patients with atherosclerotic disease. Although no clear advantage of AXC for groups without atherosclerosis was shown when they were analyzed separately, AXC was seen to be of significant benefit in avoiding the adverse outcome of stroke or death in the group as a whole. This was true even though patients undergoing AXC where older and more likely to have coronary artery disease or chronic obstructive pulmonary disease, or both. A 50% decrease in adverse outcome was seen in the patients with AXC for acute dissection, and a larger patient sample might well have shown a statistically significant advantage in this group. The axillary artery is now our cannulation site of choice in patients with both atherosclerotic aneurysms and dissections, and even a conservative interpretation of the results suggests that use of AXC was not associated with any increase in adverse outcome in patients with degenerative aneurysms, acute or chronic dissections, or proximal aortic operations for miscellaneous other aortic pathologies. The utility of AXC in facilitating access to CPB in patients with previous cardioaortic intervention has previously been recognized. We have shown, in another study, that AXC is an independent factor favoring long-term survival after aortic root/ascending aorta reoperations, and there is some indication that AXC may provide some advantage for 30-day survival in aortic reoperations as a whole. Axillary artery cannulation also provides an excellent route for selective cerebral perfusion during aortic arch procedures, providing optimal protection during circulatory arrest and thus reducing neurologic injury and early mortality [7, 11, 12]. In proximal aortic operations, however, the circulatory arrest time is usually shorter, and therefore, optimal cerebral protection is arguably less critical. This may perhaps contribute to the difficulty in demonstrating any superiority of this approach in the current study in patients whose underlying disease did not predispose them to embolization. The question of whether one should initiate direct AXC or use a graft is somewhat controversial. We have not encountered problems with direct cannulation and therefore see the use of a graft as an unnecessary additional step. But we cannot reiterate frequently enough the importance of keeping the interval of AXC as brief as possible to avoid complications secondary to distal ischemia. It should be noted that use of a graft does not eliminate complications: both dissection and hyperperfusion of the arm have been described [13]. Because the choice of cannulation site varied with surgeon preference and with date of operation, this study is subject to the usual limitations of retrospective reviews. Therefore, notwithstanding our observation that patient characteristics did not appear to differ over time or between patients in the various cannulation groups, our analysis must be regarded as suggestive rather than conclusive. The results of the current review suggest that more widespread adoption of AXC for complex aortic operations, particularly in atherosclerotic aneurysms and possibly also its use in patients with atheromatous lesions undergoing coronary or valve operations should be encouraged. References 1. Sabik JF, Lytle BW, McCarthy PM, Cosgrove DM. Axillary artery: an alternative site of arterial cannulation for patients with extensive aortic and peripheral vascular disease. J Thorac Cardiovasc Surg 1995;109:885 90; discussion 890 1. 2. Bichell DP, Balaguer JM, Aranki SF, et al. Axilloaxillary cardiopulmonary bypass: a practical alternative to femorofemoral bypass. Ann Thorac Surg 1997;64:702 5. 3. Schachner T, Vertacnik K, Laufer G, Bonatti J. Axillary artery cannulation in surgery of the ascending aorta and the aortic arch. Eur J Cardiothorac Surg 2002;22:445 7. 4. Strauch JT, Spielvogel D, Lauten A, et al. Technical advances in total aortic arch replacement. Ann Thorac Surg 2004;77: 581 9; discussion 589 90. 5. Borst HG. Axillary artery for extracorporeal circulation. J Thorac Cardiovasc Surg 1995;110:1775. 6. 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; discussion 1274 84. 7. 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; discussion 103 8. 8. Kouchoukos NT, Wareing TH, Murphy SF, Perrillo JB. Sixteen-year experience with aortic root replacement. Results of 172 operations. Ann Surg 1991;214:308 18; discussion 318 20. 9. Etz CD, Homann TM, Rane N. Aortic root reconstruction with a bioprosthetic valved conduit: a consecutive series of 275 procedures. J Thorac Cardiovasc Surg 2007;133: 1455 63. 10. Ergin MA, Griepp EB, Lansman SL, Galla JD, Levy M, Griepp RB. Hypothermic circulatory arrest and other methods of cerebral protection during operations on the thoracic aorta. J Card Surg 1994;9:525 37. 11. Strauch JT, Spielvogel D, Lansman SL, Lauten AL, Bodian C, Griepp RB. Long-term integrity of teflon felt-supported suture lines in aortic surgery. Ann Thorac Surg 2005; 79:796 800.

446 ETZ ET AL Ann Thorac Surg AXILLARY CANNULATION EFFECTS ON OUTCOME 2008;86:441 7 12. Spielvogel D, Etz CD, Silovitz D, Lansman SL, Griepp RB. Aortic arch replacement with a trifurcated graft. Ann Thorac Surg 2007;83:S791 5; discussion S824 31. 13. 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. DISCUSSION DR TOMAS D. MARTIN (Gainesville, FL): If I might ask just a question, could you define for us what degenerative and atherosclerotic aneurysms were in your group? Were those defined by CT scan, were they defined by echo? DR ETZ: No. Aneurysms were classified according to a pathological definition. DR MARTIN: So you had pathology specimens on all these? DR ETZ: On all cases. DR MARTIN: Did you look at atheromatous disease? Was that included in that? DR ETZ: Atheromatous disease was included, but not all of the 122 patients had what we label clot or atheroma. Visible atheromata are not required for a classification for atherosclerotic disease. The aneurysm was classified as atherosclerotic if there were signs of systemic atherosclerosis and if the appearance of the aneurysm in the operating room and the specimen pathology were consistent with this diagnosis. In the severe cases, visible and sometimes mobile atheromata were present. DR MARTIN: If I might address the panelists, I don t know how you define it, but I am a little confused at the difference between those two groups and what we see generally. DR HAZIM J. SAFI (Houston, TX): Well, I am confused too. You said that you chose axillary cannulation because you knew that atherosclerosis was present. We know that aneurysm is always a medial disease and that atheromatous plaque is an intimal disease, and that sometimes they are superimposed. We use transesophageal echo to discern these conditions. If there is atheromatous plaque of grade 1 or 2, we will go ahead and cannulate whatever is convenient. However if it is atheromatous, with an appearance like cauliflower, then this will affect our choice of cannulation. Was that the way you approached it? DR ETZ: It is evident that I should clarify my statements. The site of cannulation was not chosen because we were able to identify those patients at unusually high risk of embolism beforehand. We began using the axillary cannulation technique thinking that it would be universally useful. Our classification of aneurysms is a somewhat retrospective definition using the clinical picture at the time of surgery, and the pathology. We used axillary cannulation increasingly starting in 1999 2000, regardless of what was known preoperatively about the etiology of the aneurysm. DR NICHOLAS T. KOUCHOUKOS (St. Louis, MO): Can you explain the difference between the atherosclerotic group and the other groups in terms of the morbid events? DR ETZ: We came up with two or three possible explanations. Obviously the flow direction is different if you cannulate using the axillary artery: you have retrograde flow in the innominate artery, which possibly protects the cerebral circulation from clots that you may dislodge when you manipulate the aortic root. This difference may be more significant when you compare axillary to femoral cannulation, which results in retrograde flow in the arch, and can flush distal atheromic debris into the cerebral circulation. Secondly, direct aortic cannulation is known for a sandblast effect that increases local pressures and creates turbulent flow, potentially directly dislodging plaques in the arch. We think one of the most important reasons for the success of axillary cannulation is that it provides a gentler form of antegrade flow. DR JOHN S. IKONOMIDIS (Charleston, SC): I have a question for the speaker and a question for the panel. The question for the speaker is, these patients included patients that had ascending replacements and in those that extended into the hemiarch. So did you separate your analysis on neurologic outcome by those that just had an ascending replacement where you presumably put the cross-clamp on the ascending aorta vs those that had hemiarch replacements where the clamp was probably more likely on the innominate artery? DR ETZ: We did not distinguish ascending aorta from hemiarch replacements. They were all done with an open distal anastomosis, so we thought this distinction was not really important. DR IKONOMIDIS: And the question I have for the panel is, advantages and disadvantage of direct cannulation of the axillary artery versus placement of a side graft. DR MARTIN: Before we answer that, John, could I make one comment before we get on that specific thing? First, I would like to compliment you on your presentation and on the choice of a very timely topic. We recently looked at our arch experience. In the last 6 years we have performed 660 arch aneurysm operations requiring circulatory arrest with a similar experience in terms of mortality and a similar experience in terms of stroke rate. From your paper, I would hesitate to recommend to the audience that every aneurysm that is atherosclerotic that you should use an axillary cannula. I am personally not an axillary cannulation fan and have actually told my residents the axillary artery in many cases is not your friend. So I would just like to make a comment, and I would be very, very interested in the rest of the esteemed panel s thoughts on this, that I would not at the moment say that axillary cannulation is the cannulation of choice for ascending and arch aneurysms. DR JOSEPH S. COSELLI (Houston, TX): We are completely on the other side of that issue. We over the last few years have defaulted to the axillary artery as the site of perfusion in these cases and have probably in less than 1% of cases actually cannulated the femoral artery. Our cannulation is a little different. We don t cannulate with a plastic cannula but sew on a Dacron [DuPont, Wilmington, DE] side graft. That probably eases into some of the other aspects of that question in that it allows us with a right radial artery line to monitor the perfusion pressure as well as the flows. When I am unable to do that with a Dacron graft, you can just simply put a little catheter, stick it in the side of the graft and monitor the pressure in the graft when

Ann Thorac Surg ETZ ET AL 2008;86:441 7 AXILLARY CANNULATION EFFECTS ON OUTCOME 447 the right radial art line is not available to you for whatever reason. I would like to know from your group what circumstances in your current practice would you use the femoral artery, and could you expand upon your concepts of the proper flows and monitoring of pressures and cerebral perfusion? DR ETZ: The pressure is only monitored in the left radial artery. Femoral artery cannulation is almost never used anymore; its use has been gradually decreasing over the last 10 years. Axillary artery cannulation is almost always utilized. In 2% to 3% of patients, however, the artery is small or friable, or we do not observe torrential backflow through the cannula. If there is not unequivocal free flow, we go to an alternate site. We have published a report of complications of direct axillary cannulation in the Annals of Thoracic Surgery. DR MARC R. MOON (St. Louis, MO): Can we get Dr Kouchoukos s thoughts on axillary cannulation versus femoral? DR KOUCHOUKOS (St. Louis, MO): We do what Dr Coselli has described, and that is, use a side arm exclusively. We have never actually directly cannulated. And I would like to ask you if you have any complications related to direct cannulation of the axillary artery? DR ETZ: Interestingly, in this series, there were no complications with malperfusion of the arm, but it may be worthwhile noting that we try to keep the time of the cannulation really short. The cannula is put in right before CPB, and it is taken out as soon as possible. One localized dissection of the arch required repair 4 years later. In four or five cases, owing to friability, the artery was ligated rather than repaired without ischemic sequelae. DR KOUCHOUKOS: Do you have some bias anastomosing a graft to the artery? DR ETZ: I discussed this with Dr Griepp before I came here to get his opinion on this question. He thinks that indirect cannulation is an option, but he has not done it because he has encountered few problems using direct cannulation and worries that there may be a problem, with a side graft, of hyperperfusion of the arm; he also doubts that use of a graft is any less likely to cause arterial injury or dissection. But he does not necessarily feel that a graft is an inferior option. DR SAFI: We rarely use the axillary artery, but when we do, often in some cases at the insistence of my associate, Dr Tony Estrera, we use a Dacron graft because of the relative ease of use. Otherwise we use the femoral artery and the ascending aorta, and we have very good results with stroke. The important issue is not the mode of cannulation, but the question of how well you can perfuse the brain. We feel the use of transcranial Doppler is essential to assess degree of perfusion. Perfusion can be estimated using radial artery and pressure measures. We find that when we have a good echo, the transcranial Doppler, we are able to determine flow to the brain. This requires continuous monitoring. I think there is a need to train a lot of people to use this monitoring technique to reduce the guesswork that sometimes occurs.