When Should the Aortic Arch Be Replaced in Marfan Patients?

Transcription

1 When Should the Aortic Arch Be Replaced in Marfan Patients? Jean Bachet, MD, Fabrice Larrazet, MD, Bertrand Goudot, MD, Gilles Dreyfus, MD, Thierry Folliguet, MD, François Laborde, MD, and Daniel Guilmet, MD Département de Pathologie Cardiaque, Institut Mutualiste Montsouris, Paris; and Service de Chirurgie Cardio-Vasculaire. Hôpital Foch, Suresnes, France Background. The purpose of this study was to assess the prevalence, indications, and results of aortic arch replacement in Marfan patients with and without acute dissection. Methods. Between January 1993 and December 2005, our group performed 76 aortic replacements in 54 Marfan patients (mean age, 38.3 years), of whom 20 had already undergone one or two replacements of the thoracic aorta, and 3 required one late procedure each in other institutions. So, the 54 patients underwent a total of 100 aortic operations. Indication for initial surgery was elective aortic root replacement in 25 patients (46%), acute type A dissection in 19 (35%), acute type B dissection in 2 (4%), and chronic type B dissection in 8 (15%). Indication for reoperation was residual chronic dissection in the proximal aorta in 14 patients (36%), in the distal aorta in 22 (56%), and acute retrograde type A dissection in 3 (8%). Results. At initial operation, the aortic arch was not involved in the 25 patients with aneurysm of the aortic root and was replaced in only 1 of the 19 patients with acute type A dissection (1/44 patients, 2.3%). At the second or third operation, the arch had to be replaced in 4 (16%) of 25 patients initially operated on for aortic root aneurysm, in 14 (73%) of 19 patients operated on for acute type A dissection, and in 3 (30%) of 10 patients with previous acute or chronic type B dissection. The difference between patients with initial elective aortic root replacement and patients with acute dissection was highly significant (p < 0.001). Overall in-hospital mortality was 13%. The risk of death was 9.6% per procedure. Conclusions. Aortic arch replacement in Marfan patients is not indicated during elective aortic root replacement. In contrast, the significant rate of aneurysmal dilatation of the aortic arch after surgery for acute type A dissection may be an incentive for a more aggressive approach toward the aortic arch during initial surgery. (Ann Thorac Surg ) 2007 by The Society of Thoracic Surgeons Patients with Marfan syndrome are generally affected by the cardiovascular features of this hereditary disease, and most are prone to develop aortic aneurysms. In most instances, these lesions first affect the aortic root or the ascending aorta, or both. When these lesions are not diagnosed or the patient is not operated on in a timely fashion, acute type A dissection requiring emergent surgery may occur. Some patients, however, experience type B dissection as the first cardiovascular complication of their defect, which generally does not require immediate surgery. Marfan patients are therefore mostly referred to cardiac surgery either for elective replacement of the aortic root and ascending aorta or emergent surgery for acute type A dissection. However, because of the nature of the disease, the chronic evolution of a persistent aortic dissection, or improper surgical treatment, a fair number of patients need one or more subsequent surgical procedures. It is currently widely demonstrated that the immediate and long-term results of elective surgery limited to the Presented at Aortic Surgery Symposium X, New York, NY, April 27 28, Address correspondence to Dr Bachet, Institut Mutualiste Montsouris, Cardiovascular Surgery, 42 Boulevard Jourdan, Paris, France; jean.bachet@imm.fr. proximal aorta are quite satisfactory, provided the aortic root is completely replaced [1 4]. It has also been observed that reoperations after surgery for acute type A dissection are frequent [5]. In contrast, few reports have addressed the issue of the extent of the aortic replacement and, specifically, whether the aortic arch should be totally replaced during elective surgery for root aneurysm or emergent surgery for acute type A dissection [6 9]. Relying on experience with initial and subsequent operations in patients with Marfan syndrome, the present retrospective observational study is aimed at addressing two incompletely resolved questions: when is the aortic arch involved, and when should it be replaced? The study was approved by the Ethics Committee of Institut Mutualiste Montsouris, Paris, but individual patient consent was waived because the patients are not identified. Patients and Methods Between January 1993 and December 2005, 54 patients with Marfan syndrome (30 men [60%]), with a mean age of years (range, 17 to 67 years), were referred to our group for initial surgery or reoperation required by aortic aneurysm, dissection, or one of their complica by The Society of Thoracic Surgeons /07/$32.00 Published by Elsevier Inc doi: /j.athoracsur

2 Ann Thorac Surg AORTIC SURGERY SYMPOSIUM X BACHET ET AL S775 Table 1. Procedures Performed at Initial Operation OUR Group Other Centers Total Aortic root aneurysm Bentall Valve sparing procedure 8 8 Cabrol 1 1 Ascending aorta AVR 4 4 Ascending aorta AV repair 1 1 Acute type A dissection Bentall Bentall arch replacement 1 1 Valve sparing procedure 2 2 Ascending aorta AVR 4 4 Ascending aorta AV repair 4 4 Type B dissection Descending aorta replacement Thoracoabdominal replacement 5 5 Total AV aortic valve; AVR aortic valve replacement. tions. Because of the time span of the present study, the diagnosis of Marfan syndrome had been established in all patients by the presence of two major criteria, or one major and two minor criteria of the Berlin recommendation [10], which are less stringent than the Ghent criteria [11]. In 3 patients in whom the clinical features could be questioned, the diagnosis was confirmed by a genealogic and genetic study within a specialized multidisciplinary outpatient clinic. Thirty-four patients were referred to us for initial surgery, 20 patients had already undergone one (n 19) or2(n 1) surgical procedures in other centers, and 3 underwent an additional procedure each in other institutions after our group had operated on them. So, the 54 patients underwent a total of 100 operations for a mean rate of 1.9 procedures per patient: 22 (40%) had one operation, 22 (40%) had two, 7 (13%) had three, 2 (4%) had four, and 1 (2%) required five procedures. The cause and type of the initial procedures are indicated in Table 1. The mean interval was years (range, 0.25 to 18 years) between the initial procedure and the second operation, years (range, 0.16 to 14.5 years) between the second and the third operations, and years (range, 0.25 to 6.9 years) between the third and the fourth operations. The median intervals were 3.6 years between the first and second procedures and 1.3 years between the second and third procedures. All patients primarily referred to our group for uncomplicated aneurysm or acute type A dissection had a complete replacement of the aortic root using either the modified Bentall procedure [12] or the Yacoub [13] or the David [14] valve-sparing procedure. In contrast, only 3 (37%) of 8 patients with aneurysm of the aortic root and 3 (30%) of 10 patients with acute type A dissection first operated on in other centers had complete replacement of the aortic root. It is noteworthy that at the initial surgery, the aortic arch was never involved in patients with uncomplicated aneurysm of the aortic root and that it was replaced only once in patients with acute type A dissection. The rate of aortic arch replacement was therefore 2% (1/54). Follow-Up The cumulative follow-up of the whole cohort amounts to 414 patient-years, with a mean follow-up of years (range, 5 months to 22.3 years). Follow-up was calculated from the time of the first procedure undergone by the patients. Therefore, although this study concerns patients referred to us between 1993 and 2005, the follow-up takes into account the interval between initial surgery and the date of referral to our group in all patients operated on before this period of time. Four patients (8%) were lost to follow-up. Statistical Analysis Hospital mortality was defined as the mortality rate within 30 postoperative days or before hospital discharge. Categoric variables are expressed as percentages, and continuous data are expressed as mean values SD. Comparisons of characteristics were calculated by using the 2 test. Freedom from reoperation and subsequent arch replacement was estimated by the Kaplan-Meier method. Curves of freedom from reoperation were compared between groups using the log-rank test. Patients who were operated on several times were included only once. Table 2. Procedures Performed at Reoperation Initial Presentation Aortic root aneurysm Acute type A dissection Procedure Reoperation 1st 2nd 3rd 4th AVR 2 MVR 1 Bentall 5 False aneurysm 1 Aortic arch 2 1 Thoracoabdominal 2 1 AVR arch 1 Bentall arch 7 2 Arch 2 1 Arch thoracoabdominal 1 Descending aorta 2 1 a Thoracoabdominal Abdominal aorta 1 Type B Bentall 1 dissection Bentall arch 2 Ascending Aorta arch 1 Descending Aorta 1 Total a Descending aortic stent graft. AVR aortic valve repair; MVR mitral valve repair.

3 S776 AORTIC SURGERY SYMPOSIUM X BACHET ET AL Ann Thorac Surg The linear rate of any reoperation was 7.4% patient/y in all patients with uncomplicated aortic root aneurysm. It is, however, noteworthy that it was 4.6% patient/year for patients who had complete aortic root replacement, and 18% patient/y (p 0.01) in patients who had no root replacement during the initial operation. The linear rate of any reoperation was, respectively, 17% patient/year and 4.2% patient/year in patients presenting initially with acute type A or type B dissection (either chronic or acute; Fig 2). Fig 1. Actuarial (Kaplan-Meier) representation of freedom from reoperation in all patients. Results Reoperations Thirty-five (55%) of the 53 patients who survived the initial procedure underwent 45 reoperations. The cause of reoperations varied, but can be classified into the four main categories of (1) valvular dysfunction and aortic root enlargement, (2) evolving dissecting aneurysm distal to or within a previous repair, (3) recurring acute type A dissection in patients with previous replacement of the thoracic or thoracoabdominal aorta, and (4) a few examples of false aneurysms or malperfusion. The type and chronology of the various reoperations are indicated in Table 2. Figure 1 indicates the freedom from reoperation in all patients. Among the 20 patients presenting initially with uncomplicated aneurysm that had a complete replacement of the aortic root, 3 patients (15%) required reoperation on the proximal aorta, consisting of aortic valve replacements in 2 and false aneurysm in 1, and all 5 patients who had not undergone complete root replacement required a subsequent Bentall procedure (p 0.01). Similarly, among the 19 patients presenting initially with acute type A dissection, a repeat Bentall procedure was required in 3 (27%) of the 11 who had had complete replacement of the root and in 7 (87%) of the 8 patients who had not undergone total root replacement (p 0.05). All those procedures were associated with transverse arch replacement (Table 2). In 3 (12%) of 25 patients with initial aortic root aneurysm, thoracic descending and thoracoabdominal replacement was required in 2 patients at a second operation and at a third operation in 1 patient. Those distal replacements were mandated by the occurrence and subsequent evolution of a type B dissection during follow-up. In 10 (53%) the 19 patients with initial type A dissection, replacement of the thoracic descending or thoracoabdominal aorta was required in 4 and 6 patients at second and third or fourth operations, respectively. Those operations were mandated by the aneurysmal evolution of a persistent distal dissection in 9 patients, and the presence of a false aneurysm in 1. Replacement of the Aortic Arch As already indicated, 1 patient with acute type A dissection underwent total arch replacement associated with a modified Bentall procedure during the initial operation. A total of 21 transverse arch replacements had to be performed during a second, third, or fourth operation. Among the 25 patients first presenting with an uncomplicated aneurysm of the aortic root, replacement of the transverse aortic arch was required in 4 (16%), 3 at the second operation and 1 at the third. In the 19 patients with initial acute type A dissection, 14 arch replacements (73%) were done (11 during a second operation and 3 during a third or fourth operation). Those replacements were isolated in 3 patients, associated with a modified Bentall procedure in 10, and with a thoracoabdominal replacement in 1. Of the 10 patients initially operated on for type B dissection, 9 survived and 3 (33%) required a subsequent total arch replacement for occurrence of retrograde type A acute dissection. The arch replacement in these patients was associated with a Bentall procedure in 2 or replacement of the ascending aorta in 1. The yearly linear rates per patient of aortic arch replacement were 1.7% for patients with uncomplicated aneurysm of the aortic root, 11%, for acute type A dissection, and 4.4% for type B dissection. The relative risk for patients presenting initially with acute type A dissection was 4.6, compared with patients with an uncomplicated aneurysm of the aortic root. Fig 2. Actuarial (Kaplan-Meier) representation of freedom from reoperation in patients initially presenting with aortic root aneurysm or chronic dissection (continuous line) vs patients presenting with acute type A aortic dissection (dotted line). The difference between the groups is highly significant (log-rank test, p 0.001).

4 Ann Thorac Surg AORTIC SURGERY SYMPOSIUM X BACHET ET AL S777 Table 3. Causes of Hospital Death Procedure Cause of Procedure Type of Procedure Death (n) Cause of Death First Type B diss Thoracoabdominal 1 Hemorrhage MOF Second Dilated diss. aneurysm Arch desc 1 Stroke Bentall arch 1 Hemorrhage Type B diss Thoracoabdominal 1 Hemorrhage Third Type B diss Thoracoabdominal 3 Coma MOF MOF MOF multi-organ failure. The surgical technique used during the procedure of aortic arch replacement included femoral artery cannulation in all patients but 2, operated on recently, who underwent right axillary artery cannulation and selective antegrade cerebral perfusion with moderate core hypothermia according to the technique described by our group [15] until 2002, and according to the technique described by Kazui and colleagues [16] thereafter. The supraaortic vessels were reimplanted en bloc into the arch prosthesis in all but 3 patients, in whom the left subclavian artery (1 case) or the innominate artery (2 cases) had to be reimplanted separately. Hospital Mortality There were no deaths among the 17 patients with uncomplicated aneurysm of the aortic root and, more surprisingly, in the 8 patients with acute type A dissection referred to us for the initial operation. After reoperation, 7 patients (13%) died within 30 days or before discharge. Two patients (2/22) died after aortic arch replacement for a mortality rate of 9%, and 5 of 17 died after thoracoabdominal replacement for a mortality rate of 29% (NS). The causes of death are listed in Table 3. If we consider that, by definition, the patients initially operated on in other centers and referred to us for reoperation survived the first procedure, the hazard rate of hospital mortality was 7.2 % for 97 surgical procedures and 9.3% for the 76 operations performed by our group. The mortality rates varied with the number and type of procedures (Table 4). It is of interest that it also varied with the number of procedures that we performed in each patient, regardless whether it was a first procedure or a reoperation. So, the mortality rates were 3% (1/33) after one operation, 10% (3/10) after two procedures, and 33% (3/9) after three procedures. No patients died in whom we performed four procedures. Table 4. Risk of Mortality According to the Number of Procedures Procedures (n) Patients (n) Deaths (n) Risk (%) Neurologic Complications After the 76 operations performed by our group, seven major neurologic complications were observed: five strokes (one death), consisting of one hemiplegia, one cerebellar infarction, three cases of oculomotor nerve palsy in patients who underwent arch replacement, and two cases of paraplegia (one death) in patients who underwent thoracic or thoracoabdominal replacement. The overall rate of neurologic complication was 9.2%. The rate of cerebral injury in arch replacement was 22% (5/22), and the rate of paraplegia during thoracoabdominal replacement was 11%; however, the neurologic status of the patients who died intraoperatively remained unknown. All neurologic complications were observed after the second or third procedures. No complication occurred at initial surgery. Comment The present study, although retrospective and strictly observational, tends to confirm some data and considerations that were observed a long time ago and appear to be the matter of a wide consensus concerning the surgical management of patients with Marfan syndrome [17]; however, some questions remain unresolved or incompletely resolved. In particular, it is yet unclear whether the technique and the extent of aortic replacement have any influence on the risk and rate of further reoperation. One point highlighted by the present study is that Marfan patients who are first operated on for acute type A dissection have a risk of being reoperated on that is about four times higher than that of Marfan patients first operated on for uncomplicated dilatation of the aortic root. Thus, there is no doubt that elective surgery should be considered before the occurrence of acute dissection or rupture. Although classically an aortic diameter of 55 mm is considered the upper limit before surgery is indicated, the present report, as well as our whole experience of complete replacement of the aortic root in non-marfan patients [18 20], has progressively led us to consider surgery in Marfan patients with smaller diameters. We think presently that a maximum diameter of 50 mm should be accepted and that patients with a family

5 S778 AORTIC SURGERY SYMPOSIUM X BACHET ET AL Ann Thorac Surg history of acute dissection or sudden death, or in whom the aorta diameter increases more than 2 mm in 1 year, should be operated on as soon as the aortic root reaches 45 mm. Despite the lack of statistics or data supporting such an opinion, it is reinforced by the experience reported by Ergin and colleagues [21], in which 73% of patients (Marfan and non-marfan) with acute type A dissection had an aortic root diameter of less than 48 mm. Another point is certainly that in those patients, the aortic root should be totally replaced during the initial operation performed either electively for dilatation of the aortic root or emergently for acute type A dissection. In this regard, the Bentall procedure and, in particular, its modified version, the button technique as popularized by Kouchoukos and colleagues [12], has represented a major breakthrough and is undoubtedly a key factor in the dramatic improvement in the life expectancy of Marfan patients. Many groups have reported large experiences with very low hospital mortality and morbidity and quite stable long-term results [1 3, 12, 17]. Modern techniques of valve-sparing procedures may represent a valuable option in some patients, but despite the excellent results reported recently, some degree of uncertainty remains concerning their very long-term durability, in particular in Marfan patients. Indeed, in the present experience, 3 of 10 patients with such a repair had to undergo a second operation for aortic valve replacement within 6 to 19 months after the first procedure. In addition, those techniques require a fair degree of expertise in aortic root surgery and may prove somewhat difficult especially in an emergent procedure. Nevertheless, it is rather surprising to observe that some Marfan patients still undergo limited resection of the ascending aorta, as it was the case in 13 (65%) of the 20 patients in the present study who were first operated on in other centers. This is particularly true during surgery for acute type A dissection. In patients with elective initial surgery for aneurysm of the aortic root, such unsatisfactory techniques have resulted in a risk of reoperation that is fourfold higher than in patients with complete root replacement and comparable with that of patients first presenting with acute type A dissection. But the most unresolved key question remains whether the aortic arch should be replaced systematically during elective surgery for aneurysm of the aortic root or during surgery for acute type A dissection to reduce the rate of reoperation and the number of subsequent arch replacements. Our experience strongly suggests that the answer to the first question is negative. Indeed, no primary aneurysm of the transverse arch or root dilatation extending into the transverse arch was observed during our reported experience. Only a few (2 [8%]) arch replacements were required after elective complete aortic root replacement. These were mandated at second and third reoperation in 1 patient by the presence of a false aneurysm, and in 1 patient by the occurrence, 9 years after initial elective surgery, of an acute type B dissection treated conservatively. The two other arch replacements performed during reoperation were done in patients with an unsatisfactory initial repair who required a subsequent Bentall procedure. We therefore concur with the general opinion that probably prophylactic replacement of the aortic arch is not indicated in the setting of patients operated on electively for aortic root dilatation. Concerning the patients who first present with acute type A dissection, the answer may be quite different. Indeed, in the present experience, the transverse aortic arch was replaced only once in 19 patients (5.2%) during the initial operation. In contrast, a subsequent arch replacement was required in most patients (73%) because of further evolution of a persistent false lumen. Indeed, the arch is almost always involved during the process of acute type A dissection. Because of the basic weakness of the aortic wall and the abnormal condition of flow in the dissected aorta, the false lumen is prone to dilate. Hence, systematic arch replacement during initial surgery appears logical. Several recent publications have reported excellent results with a low risk of mortality and adverse neurologic events [22]. Conversely, it can be argued that our experience extended over a long period of time and that many patients were operated on with techniques that are not currently considered optimal. Indeed, it is likely that the systematic use of right subclavian artery cannulation, antegrade perfusion of the aorta [23] with selective antegrade cerebral perfusion during moderate hypothermia [15, 16], and open distal anastomosis would result in less late aneurysmal dilatation of the distal aorta and thus in a reduced number of subsequent arch replacements. One could also argue that systematic total replacement of the aortic arch might be delicate during emergencies and could entail a certain degree of morbidity and mortality, trading off its advantages against its drawbacks. Only a randomized controlled study could help resolving this dilemma, but such a study is quite improbable [24]. In addition, it should be observed that the risk of arch replacement during reoperation is associated with an acceptable risk [25]. Notwithstanding the significant rate of reoperation required in these patients, the almost systematic necessity of replacing the jeopardized transverse arch during these reoperations is certainly a strong incentive for a more aggressive approach toward the aortic arch during initial surgery. Concerning patients who are operated on initially for type B dissection, the number of patients with immediate or late requirement for arch replacement is too small to allow any significance other than anecdotal. It seems, however, that the difficulty of arch replacement through a left thoracotomy and the necessity of resorting to deep hypothermia associated with circulatory arrest in most instances preclude systematic replacement. In addition, in our experience, subsequent arch replacement in these patients could be done routinely through a median sternotomy and was associated with no hospital mortality and only one neurologic complication. Because of its retrospective, observational nature, the present study fails to definitively answer the question raised in the title, but it certainly confirms ideas, statements, and strategies concerning the modern manage-

6 Ann Thorac Surg AORTIC SURGERY SYMPOSIUM X BACHET ET AL S779 ment of Marfan patients who require surgery. There is, indeed, a growing consensus that patients must be operated on sooner than classically admitted with use of techniques intended to eliminate the impaired aortic segments during initial surgery. In this regard, complete aortic root replacement, with or without valve preservation, is absolutely mandatory during surgery on the aortic root, whereas prophylactic replacement of the transverse arch does not seem indicated. In contrast, because of the major risk of reoperation, systematic, potentially curative total replacement of the aortic arch appears to be a logical option in patients presenting with acute type A dissection to reduce the number of reoperative procedures in these generally young patients who already pay a high toll to their disease. This article was published with the support of a grant from L Association pour le Développement des Techniques de Chirurgie Cardiovasculaire (ADETEC) Suresnes, France. References 1. Gott VL, Greene PS, Alejo DE, et al. Replacement of the aortic root in patients with Marfan s syndrome. N Engl J Med 1999;340: Finkbohner R, Johnston D, Crawford ES, Coselli J, Milewicz DM. Marfan syndrome. Long-term survival and complications after aortic aneurysm repair. Circulation 1995;91: Alexiou C, Langley SM, Charlesworth P, Haw MP, Livesey SA, Monro JL. Aortic root replacement in patients with Marfan s syndrome: the Southampton experience. Ann Thorac Surg 2001;72:1502 7; discussion David TE. Aortic surgery in the Marfan syndrome. Adv Card Surg 2001;13: Bachet J, Termignon JL, Dreyfus G, et al. Aortic dissection. Prevalence, cause, and results of late reoperations. J Thorac Cardiovasc Surg 1994;108: ; discussion Crawford ES, Kirklin JW, Naftel DC, Svensson LG, Coselli JS, Safi HJ. Surgery for acute dissection of the ascending aorta: should the arch be involved. J Thorac Cardiovasc Surg 1992;104: Pugliese P, Pessotto R, Santini F, Montalbano G, Luciani GB, Mazzucco A. Risk of late reoperations in patients with acute type A aortic dissection: impact of a more radical surgical approach. Eur J Cardiothorac Surg 1998;13: Carrel T, Beyeler L, Schnyder A, et al. Reoperation and late adverse outcome in Marfan patients following cardiovascular surgery. Eur J Cardiothorac Surg 2004;25: Tagusari O, Ogino H, Kobayashi J, et al. Should the transverse aortic arch be replaced simultaneously with aortic root replacement for annuloaortic ectasia in Marfan syndrome? J Thorac Cardiovasc Surg 2004;127: ; discussion Beighton P, De Paepe A, Danks D, et al. International nosology of inheritable disorders of connective tissue. Am J Med Genet 1986;29: De Paepe A, Devereux RB, Dietz HC, Hennekam RC, Pyeritz RE. Revised diagnostic criteria for the Marfan syndrome. Am J Med Genet 1996;62: Kouchoukos NT, Wareing TH, Murphy SF, Perrillo JB. Sixteen year experience with aortic root replacement: results of 172 operations. Ann Surg 1991;214: Yacoub MH, Gehle P, Chandrasekaran V, Birks EJ, Child A, Radley-Smith R. Late results of a valve-preserving operation in patients with aneurysms of the ascending aorta and root. J Thorac Cardiovasc Surg 1998;115: David TE, Feindel CM. An aortic valve-sparing operation for patients with aortic incompetence and aneurysm of the ascending aorta. J Thorac Cardiovasc Surg 1992;103:617 21; discussion 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 93; discussion Kazui T, Inoue N, Yamada O, Komatsu S. Selective cerebral perfusion during operation for aneurysms of the aortic arch: a reassessment. Ann Thorac Surg 1992;53: Baumgartner WA, Cameron DE, Redmond JM, Greene PS, Gott VL. Operative management of Marfan syndrome: the Johns Hopkins experience. Ann Thorac Surg 1999;67: Bachet J, Termignon JL, Goudot B, et al. Aortic root replacement with a composite graft. Factors influencing immediate and long-term results. Eur J Cardiothorac Surg 1996;10: Bachet J, Goudot B, Dreyfus G, et al. Current practice in Marfan s syndrome and annulo-aortic ectasia: aortic root replacement with a composite graft over a twenty-year period. J Card Surg 1997;12(2 Suppl): Guilmet D, Bonnet N, Saal JP, Le Houerou D, Ghorayeb G. Long term survival with the Bentall button operation in 150 patients. Arch Mal Coeur Vaiss 2004;97: Ergin A, Spielvogel D, Apaydin A, et al. Surgical treatment of the dilated ascending aorta: when, and how? Ann Thorac Surg 1999;67: Kazui T, Yamashita K, Washiyama N, et al. Impact of an aggressive surgical approach on surgical outcome in type A aortic dissection. Ann Thorac Surg 2002;74:S ; discussion S 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 Myrmel T, Lai DT, Miller DC. Can the principles of evidence-based medicine be applied to the treatment of aortic dissections? Eur J Cardiothorac Surg 2004;25:236 42; discussion Tan ME, Dossche KM, Morshuis WJ, Kelder JC, Waanders FG, Schepens MA. Is extended arch replacement for acute type A aortic dissection an additional risk factor for mortality? Ann Thorac Surg 2003;76: