Predicting outcome after reoperative procedures ... - Semantic Scholar

European Journal of Cardio-thoracic Surgery 17 (2000) 602±607 www.elsevier.com/locate/ejcts

Predicting outcome after reoperative procedures on the aortic root and ascending aorta q Giovanni Battista Luciani*, Gianluca Casali, Giuseppe Faggian, Alessandro Mazzucco Division of Cardiac Surgery, University of Verona, O.C.M. Piazzale Stefani 1, Verona 37126, Italy Received 6 September 1999; received in revised form 24 January 2000; accepted 8 February 2000

Abstract Objective: Reoperations on the ascending aorta after prior aortic procedures are formidable challenges. In order to identify factors predictive of clinical outcome using a uniform surgical approach, results of a 15-year experience were reviewed. Methods: Between 1983 and 1998, 78 reoperations on the ascending aorta were performed in 71 consecutive patients. There were 56 males and 15 females, aged 54 ^ 13 years (10±73 years), with a mean interval to ®rst reoperation of 60 ^ 76 months (5±223 months). The original operation was replacement of ascending aorta (23), aortic valve (25), aortic root (7), ascending aorta with valve preservation (9), ascending aorta and aortic valve (7). Surgical approach included femoral vessels dissection and repeat sternotomy, with femoro-femoral bypass limited to cases of traumatic reentry. Reoperation consisted in replacement of the aortic root (48), ascending aorta (15), ascending aorta and aortic valve (6), aortic root with ascending aorta and arch (6), ascending aorta and aortic arch (3). Average aortic crossclamp and cardiopulmonary bypass times were 122 ^ 86 and 188 ^ 60 min, respectively. Results: Early deaths were ®ve (7%), due to low output syndrome (3), hemorrhage (1) and sepsis (1). Mortality for emergent reoperation was signi®cantly higher (38 vs. 3%, P ˆ 0:001). A total of 39 early complications were observed in 78 reinterventions (50%), including: traumatic reentry requiring emergent femoro-femoral bypass (4), reexploration for bleeding (4), respiratory failure (12), sepsis (5), transient neurologic dysfunction (4), renal failure (3), myocardial infarction (3), circulatory insuf®ciency requiring mechanical life support (2), and wound infection (2). Average intensive care unit stay was 4:5 ^ 9:7 days (0.5±40 days). Survival was 92 ^ 4%, 78 ^ 10% and 78 ^ 10% at 1, 5, and 10 years, respectively. At follow-up (mean 34 ^ 36 months, 1±170), survivors were in satisfactory clinical conditions (1:6 ^ 0:8 mean NYHA class, 1±3) with no evidence of renal, respiratory or neurologic dysfunction. Multivariable analysis showed emergent reoperation (P ˆ 0:001), prior aortic valve replacement (P ˆ 0:005) and need for arch replacement (P ˆ 0:03) to be predictive of higher operative mortality. Longer duration of bypass (P ˆ 0:01) and aortic arch replacement (P ˆ 0:04) were predictive of higher prevalence of postoperative complications. Conclusions: Reoperations on the ascending aorta via repeat sternotomy without preventive femoral bypass are associated with low operative risk and high prevalence early complications. Emergent reintervention due to aortic dissection, particularly in patients with prior aortic valve replacement, and need for arch repair are predictive of poorer perioperative outcome. Long-term outlook of hospital survivors is satisfactory. q 2000 Elsevier Science B.V. All rights reserved. Keywords: Aortic aneurysm; Aortic dissection; Aortic root replacement; Endocarditis; Ascending aorta replacement; Reoperation

1. Introduction Primary replacement of the ascending aorta and aortic root currently represent low-risk surgical procedures, particularly when performed electively [1±7]. Progress in perioperative care of patients with aortic disease and in operative techniques, including myocardial and cerebral protection, must be credited for this achievement. The direct q Presented at the 13th Annual Meeting of the European Association for Cardio-thoracic Surgery, Glasgow, Scotland, UK, September 5±8, 1999. * Corresponding author. Tel.: 139-45-807-2485; fax: 139-45-8073308. E-mail address: [email protected] (G.B. Luciani)

consequence has been extension of aortic surgical procedures to patients at the extremes of the age spectrum and to patients with mildly symptomatic or asymptomatic disease. In addition, the rising popularity of biological conduits (autografts, homografts, xenografts), as substitutes of the aortic valve or root [8], and of aortic root remodeling operations with preservation of the native valve [9] have all translated into a substantial increase in the population at risk for reoperative procedures on the proximal thoracic aorta. Since reinterventions on the ascending aorta and aortic root pose a unique surgical challenge, it has become urgent to de®ne the risk and outcome connected with these procedures. In an attempt to predict the results after redo opera-

1010-7940/00/$ - see front matter q 2000 Elsevier Science B.V. All rights reserved. PII: S10 10-7940(00)0038 7-0

G.B. Luciani et al. / European Journal of Cardio-thoracic Surgery 17 (2000) 602±607

tions on the proximal thoracic aorta, review of a 15-year experience at our Institution was undertaken. 2. Materials and methods 2.1. Patients Between January 1983 and December 1998, 71 consecutive patients underwent one or more reoperations on the ascending aorta or aortic root. Due to the different implications in terms of risk of aortic trauma upon reentry and of operative strategy, a series of patients was excluded from the present review, including: (1) patients with isolated myocardial revascularization as original operation; (2) patients with prior replacement of the extrapericardial thoracic aorta (i. e. aortic arch, descending aorta); (3) patients requiring isolated arch replacement as reoperative procedure; (4) patients requiring isolated reoperative aortic valve surgery; (5) patients needing reoperation within 1 month of original operation. The clinical pro®le of the patient population is outlined in Table 1. Mean interval to ®rst reoperation averaged 5 years, with most procedures performed electively or urgently (within 24 h of admission) and only 12 (15%) of the operations emergently (immediately after diagnosis). Causes of emergent reintervention were acute dissection with aortic rupture or hemodynamic instability (8/12 procedures) and prosthetic valve endocarditis with hemodynamic instability (4/12 procedures). Five patients had two or more reoperations during the study period (two patients had three, and three had two). Indications to reoperation was generally due to presence of a new aortic aneurysm, true or false, or occurrence of dissection in a different aortic segment, as reported in Table 2. Prosthetic valve malfunction and endocarditis with aortic root involvement represented a less common indication. Failure of the native aortic valve after root replacement or remodeling accounted for the remaining cases. Preoperative diagnostic workup included routine cardiac catheterization except for cases of acute dissection with emergent indication where the Table 1 Clinical pro®le of the population No. Age (years) Male (%) Interval to ®rst redo (months) No. of reoperations Emergent reoperation (%) Moderate/severe aortic regurgitation (%) Original procedure (n ˆ 71) Aortic valve replacement Aortic valve repair Aortic root replacement Aortic root remodeling with valve preservation Ascending aorta replacement Ascending aorta 1 aortic valve replacement Ascending aorta 1 aortic valve resuspension

71 54 ^ 13, 10±73 56/71 (79%) 60 ^ 76, 5±223 78 12/78 (15%) 39/78 (50%) 14 11 7 5 23 7 4

603

Table 2 Indications to reoperation (n ˆ 78) Aortic aneurysm Aortic pseudoaneurysm Acute dissection Chronic dissection Prosthetic valve endocarditis Prosthetic valve malfunction Aortic valve malfunction after root operation

19 (24%) 10 (13%) 10 (13%) 15 (19%) 7 (9%) 8 (10%) 9 (12%)

diagnosis relied exclusively on transesophageal echocardiographic examination. 2.2. Operative technique The surgical approach included repeat median sternotomy with preventive exposure of the common femoral artery and vein. Femoral artery cannulation was used only in cases of acute dissection or traumatic sternal reentry. When cardiopulmonary bypass needed to be established prior to sternal division, a femoral venous cannula was also inserted via the common femoral vein and connected to a `y' circuit so as to allow subsequent cannulation of the superior vena cava or right atrium, as well. In case of coexisting severe aortic regurgitation, attention was paid not to cool the perfusate prior to venting the left heart via the right superior pulmonary vein or the apex of the left ventricle. After sternal division, an attempt at circumferential isolation of the ascending aorta was always made. Myocardial protection was obtained using antegrade cold crystalloid cardioplegia with maintenance doses every 30 min, before 1994. Thereafter, cold blood cardioplegia induction and maintenance with warm reperfusion became routine practice. Route of administration of blood cardioplegia included antegrade induction in the aortic root or retrograde in the coronary sinus, depending on the absence or presence of aortic insuf®ciency. Maintenance and reperfusion doses were injected alternatively in the coronary sinus or coronary ostia. Operations were performed under moderately hypothermic cardiopulmonary bypass (288C rectal temperature), except for cases where associated arch replacement was necessary (acute dissection, chronic dissection or aneurysm of the arch). In such instances profound hypothermia (188C rectal) was induced and aortic arch replacement was completed during a period of circulatory arrest, associated with selective antegrade cerebral perfusion when complete arch replacement was required. The procedure on the aortic valve or root was usually performed during cooling, when concomitant arch repair was anticipated. Exception to this strategy was the need to induce hypothermia and low-¯ow or circulatory arrest for unexpected intraoperative hemorrhage. Both the replacement of the aortic root and of the ascending aorta were performed avoiding any inclusion of the graft within the diseased aortic wall, which was always thoroughly resected. Routine technique for coronary artery anastomosis in root replacement was the

604

G.B. Luciani et al. / European Journal of Cardio-thoracic Surgery 17 (2000) 602±607

button technique [3], regardless of the root substitute utilized (composite, autograft, homograft, xenograft). When heavy calci®cation or scarring of the aortic root was present, thorough removal of calci®ed tissue was undertaken including the coronary ostia so as to allow complete mobilization. Any de®ciency of coronary button tissue resulting from debridement of calcium was repaired by autologous or bovine pericardial patches tailored in a ringlike fashion. In the rare case of insuf®cient coronary ostia mobilization (1/54 root replacement operations), interposition of a short segment of autologous saphenous vein between the coronary and the graft was used. The average aortic crossclamp time for the 78 procedures was 122 ^ 86 min, while bypass time averaged 188 ^ 60 min. A total of 29 (37%) procedures required a period of deep hypothermic circulatory arrest (mean arrest time 34 ^ 28 min), including nine complete and 20 partial arch replacements. 2.3. Operations performed The types of reoperative procedure performed on the 71 patients are listed in Table 3. The majority of patients having aortic root replacement received a composite graft (38/54, 70%), while the remaining 16 had a biological conduit (autograft in seven patients, homograft in seven, xenograft in two). All seven cases of prosthetic valve endocarditis were treated with homograft root replacement, with one recurrence of infection where a second homograft root was used. Tube graft replacement of the ascending aorta was associated with aortic valve replacement in one fourth (6/ 24) of the cases. Complete aortic arch replacement was needed in six cases of root replacement and in three of simple replacement of the ascending aorta. Myocardial revascularization and mitral valve surgery comprised the other commonly associated procedures. 2.4. Follow-up and statistical analysis Follow-up was gathered by telephone interview and/or clinical examination. The database was last updated between January and February 1999. All 66 operative survivors were reached. Data were expressed as mean values ^ SD or as percentages. Survival was calculated using the Kaplan±Meier product limit estimate. Patients with several operations were not censored at the next operation. The Fisher's exact test was used to compare categorical variTable 3 Type of reoperative procedure (n ˆ 78) Primary procedure Aortic root replacement Ascending aorta replacement Associated procedure CABG Complete aortic arch replacement Aortic valve replacement Mitral valve repair/replacement

54 (69%) 24 (31%) 12 9 6 4

(15%) (12%) (8%) (5%)

ables, the Mann±Whitney test for continuous variables. These tests were used for analysis of early events. Stepwise forward logistic regression analysis was employed to identify independent predictors of early mortality and early morbidity, number of early morbid events. The Cox proportional hazard method was used for predictors of late mortality. The study end-points were as follows: early mortality (prior to discharge or within 30 days of last procedure), early morbidity (i.e. traumatic reentry, myocardial infarction, circulatory failure, respiratory failure, renal failure, hepatic failure, neurologic dysfunction or damage, sepsis, wound infection), number of early complications, late mortality. The variables entered in the multivariate analysis were: age, sex, interval to ®rst redo, number of prior operations, severe preoperative aortic regurgitation, emergent reoperation, type of original procedure, prosthetic valve endocarditis, dissection, type of reoperative procedure, type of associated procedure, duration of bypass, duration of aortic crossclamp, need for hypothermic circulatory arrest, type of cardioplegia. 3. Results 3.1. Early mortality There were ®ve (7%, CL 2±16%) early casualties, two in patients undergoing elective and three in patients undergoing emergent reoperation (2=63 ˆ 3% vs. 3=8 ˆ 38%, P ˆ 0:001). Cause of death were perioperative low output syndrome in three patients who had been admitted in NYHA functional class IV. Massive perioperative hemorrhage in a patient with type A acute dissection ruptured into the left pleural cavity and septic shock in a patient needing a second root replacement for recurrent prosthetic valve endocarditis accounted for the other two early deaths. Multivariate analysis disclosed need for emergent reoperation (P ˆ 0:001), prior aortic valve replacement (P ˆ 0:005) and need for aortic arch replacement (P ˆ 0:03), as independent predictors of early mortality. 3.2. Early morbidity Overall prevalence of major non-lethal perioperative adverse events ranked high (39 events/78 procedures, 50%). However, a trend toward occurrence of multiple complications in the same complex patients was noticed. Intensive care unit stay for the entire population averaged 4:5 ^ 9:7 days, ranging from 12 h to 40 days. Four procedures (5%), where femoro-femoral bypass had not been preventively planned, required its institution due to trauma to the ascending aorta (3) or right ventricle (1). In one, bypass could be discontinued after repair of the aortic tear, while in the remaining three deep hypothermic circulatory arrest was required instead. Four cases (5%) required sternal reexploration for postoperative bleeding. Respiratory failure de®ned as need for prolonged (48 h or longer)

G.B. Luciani et al. / European Journal of Cardio-thoracic Surgery 17 (2000) 602±607

assisted mechanical ventilation complicated 12 operations (15%), while sepsis ®ve (6%). Transient neurologic dysfunction was recorded in four cases (5%), all showing complete clinical resolution prior to discharge. Renal insuf®ciency requiring postoperative dialysis was observed in three (4%) cases. Three episodes (4%) of acute myocardial infarction were identi®ed, two of which causing circulatory failure which required mechanical life support (IABP, in one patient, ECMO later followed by RVAD 1 IABP, in the other). Lastly, two instances (3%) of sternal wound infection requiring surgical debridement were recorded. Multivariate analysis failed to identify any independent predictor of a speci®c postoperative morbid event. Nevertheless, when the number (i.e. one complication vs. two or more) of early complications occurring in the same patient were used as end-point for the analysis, the duration of cardiopulmonary bypass (P ˆ 0:01) and the need for aortic arch replacement (P ˆ 0:04) were isolated as predictors of greater number of adverse events. 3.3. Late survival and functional status Follow-up of survivors ranged 1±170 months (mean 34 ^ 36 months). Two late deaths were recorded during the study period, due to lung cancer in one patient and to myocardial infarction in the other. Late survival was 92 ^ 4% at 1, 78 ^ 10% at 5 and 10 years, respectively (Fig. 1). Multivariate analysis failed to identify any independent predictor of late mortality. Assessment of follow-up clinical conditions of long-term survivors (64 patients) revealed a population in satisfactory functional status (mean NYHA class 1:6 ^ 0:8, range 1±3), with no cases of persistent neurologic, respiratory or renal dysfunction. Only two (3%) patients, one child with severe aortic regurgitation after autograft root replacement and one adult with aortic root pseudoaneurysm after homograft root replacement are currently awaiting a new aortic root reoperation. 4. Discussion The present experience demonstrates that replacement of

Fig. 1. Survival of 71 patients undergoing 78 reoperative procedures on the aortic root or ascending aorta at the University of Verona between January 1983 and December 1998. Patients at risk are reported in parenthesis.

605

the aortic root or ascending aorta after prior aortic operations can be performed with an overall low (7%) operative risk. Furthermore, the risk can be lowered to estimates typical of primary aortic procedures (3%) if reoperation is carried out on an elective basis. Previous works have shown that early mortality rates after reoperation on the proximal thoracic aorta can vary greatly, ranging between 6 and 22% [3,4,10±13]. Several factors account for this apparent wide variability, including the different prevalence of dissection as cause of reoperation, the techniques used for reoperation and the time interval considered for analysis. In detail, when reintervention is needed for recurrent or new aortic dissection, mortality ranks high, as shown by Bachet et al. [12]. The latter work further demonstrates that emergent reoperation is associated with increased operative risk [12]. Although our analysis failed to isolate aortic dissection as predictor of early mortality possibly due to limitations in sampling, the present series supports the thesis of Bachet et al. as emergent reoperation on the proximal thoracic aorta was associated with a 10-fold increase in operative risk. Multivariate analysis con®rmed emergent reintervention to be an incremental risk factor for operative mortality. Direct comparison of the results herein with the only other existing analysis focused on risk and outcome after reoperative procedures on ascending aorta and aortic root by Dougenis et al. [13] is unsound, as need for emergent reoperation was not entered in their multivariate analysis. Nevertheless, the observation that poor preoperative clinical conditions (NYHA class III or IV), which possibly included patients with active prosthetic valve endocarditis and aortic dissection, proved to be associated with increased operative mortality in the latter study may reconcile our ®ndings with those of Dougenis et al. [13]. The present analysis further identi®ed aortic valve replacement as original procedure as a risk factor for early mortality. This result is in apparent contrast with growing scienti®c evidence which shows that reoperative procedures on the aortic valve can be performed with limited surgical risk [14]. Based on these ®ndings, it may seem counterintuitive that replacement of the aortic root or ascending aorta after prior aortic valve operation should be associated with greater operative risk. Nonetheless, it has since been demonstrated how aortic valve replacement may be complicated by acute aortic dissection, late after the original intervention [15,16]. The independent progression of pathologic changes of the ascending aortic wall either secondary to the valve lesion or simply associated with it [17] has been held responsible for the above sudden event. Accordingly, the occurrence of acute aortic dissection, as well as, the appearance of endocarditis with hemodynamic compromise in patients previously subjected to aortic valve surgery adequately explain why both emergent reoperation and prior aortic valve replacement proved to be incremental risk factors for early mortality in our series. Lastly, need for complete aortic arch replacement was associated with greater operative risk in the present study.

606

G.B. Luciani et al. / European Journal of Cardio-thoracic Surgery 17 (2000) 602±607

This ®nding is not surprising as it identi®es the most lengthy and complex surgical procedures. Failure to isolate arch replacement as predictor of early mortality in previously published work [13] may be due to the distinction between partial and complete replacement, which was applied in our analysis. Whereas factors in¯uencing early mortality after reoperative aortic procedures have received great attention in prior studies [11,13], de®nition and prediction of early morbidity has less commonly been afforded by existing analyses. It is evident that primary replacement of the proximal thoracic aorta may associated with morbidity as low as 7% [1], which is signi®cantly lower than the estimates, ranging between 20 and 65%, reported for reoperative procedures [13,18,19]. The present work con®rms that multiple operations on the ascending aorta or aortic root carry a high prevalence of major postoperative complications. This translates into prolonged intensive care unit stay and increased resource utilization, as a large proportion of patients may require arti®cial respiratory, circulatory or renal support, ultimately resulting in greater hospitalization costs, albeit not calculated in our analysis. While multivariate analysis was unable to isolate predictors of speci®c postoperative complications, occurrence of two or more morbid events in the same patients was associated with more complex operations involving arch replacement and prolonged cardiopulmonary bypass times. Explanation for this ®nding is once again intuitive as use of deep hypothermic circulatory arrest in and of itself is known to profoundly affect postoperative neurologic, respiratory and renal recovery. In addition, complex reoperative root replacement, such as required in prosthetic valve endocarditis, may also result in prolonged cardiopulmonary bypass time and is certainly associated with greater prevalence of postoperative circulatory and infectious complications, as suggested by other studies [10,18]. It must be emphasized how progress in the ®eld of perioperative intensive care has substantially improved the prognosis after reoperative aortic procedures, as attested by several clinical series where early mortality was very low in the presence of high early morbidity [18,19]. Accordingly, despite the number and severity of complications recorded in our experience all patients were eventually discharged and remained free of respiratory, neurologic, renal or infectious sequelae at follow-up. Late survival after reoperative procedures on the proximal thoracic aorta is satisfactory and comparable to previous ®ndings [13]. No predictors of late mortality could be isolated in our analysis possibly due to the low number of events. Contrary to what observed for early mortality, however, it is conceivable that late mortality may be more deeply in¯uenced by the underlying disease process than by the demographic or operative variables, as shown by others [13]. Due to the low risk of primary aortic operations [1±7], to the growing enthusiasm for pulmonary autografts and aortic homografts [8], and to the expanding interest in aortic valve

preserving techniques [9], reoperations on the ascending aorta are destined to increase steadily. This has proved true of our own experience with aortic surgery during the past 8 years of practice, as well [20,21]. As evident from Fig. 1, the vast majority (71/78) of reinterventions have been performed during the 1990s and half of them in the last 3 years of activity. Nonetheless, a series of technical measures can be adopted in order to limit the prevalence and the impact (emergent vs. elective presentation) of reoperations, as previously advocated by us and others [10± 13,16,21,22]. These include: (1) radical resection of all diseased aortic tissue in acute as well as chronic aortic pathology and strict avoidance of any inclusion of the graft; (2) uniform use of coronary button technique for root replacement; (3) extensive repair in acute type A dissection, from remodeling of the sinotubular junction with valve resuspension to `open' aortic arch replacement; (4) elective replacement of the ascending aorta during valve procedures for diameters of 45 to 55 mm in the presence of bicuspid aortic valve or thinned aortic wall. In conclusion, reoperative replacement of the ascending aorta or the aortic root can be performed with low operative mortality, but high early morbidity. Strategies aimed at decreasing the need for emergent reintervention, particularly in patients with prior aortic valve replacement, may further improve early outcome.

References [1] Cohn LH, Rizzo RJ, Adams DH, Aranki SF, Couper GS, Beckel N, Collins Jr JJ. Reduced mortality and morbidity for ascending aortic aneurysm resection regardless of cause. Ann Thorac Surg 1996;62:463±468. [2] Crawford ES, Svensson LG, Coselli JS, Sa® HJ, Hess KR. Surgical treatment of aneurysm and/or dissection of the ascending aorta, transverse aortic arch, and ascending aorta and transverse aortic arch: factors in¯uencing survival in 717 patients. J Thorac Cardiovasc Surg 1989;98:659±673. [3] Kouchoukos NT, Wareing TH, Murphy SF, Perrillo JB. Sixteen-year experience with aortic root replacement. Ann Surg 1991;214:308± 320. [4] Svensson LG, Crawford ES, Hess KR, Coselli JS, Sa® HJ. Composite valve graft replacement of the proximal aorta: comparison of techniques in 348 patients. Ann Thorac Surg 1992;54:427±439. [5] Lewis CT, Cooley DA, Murphy MC, Talledo O, Vega D. Surgical repair of aortic root aneurysm in 280 patients. Ann Thorac Surg 1992;53:38±46. [6] Gott VL, Gillinov AM, Pyeritz RE, Cameron DE, Reitz BA, Greene PS, Stone CD, Ferris RL, Alejo DE, McKusick VA. Aortic root replacement: risk factor analysis of a seventeen year experience with 270 patients. J Thorac Cardiovasc Surg 1995;109:536±545. [7] Luciani GB, Casali G, Pugliese P, Mazzucco A. Impact of surgical technique on long-term outcome after aortic root replacement, Proceedings of the 2nd International Congress of Thorax Surgery, June 24±27, Bologna, Italy: Editrice Compositori, 1998. pp. 499±502. [8] Elkins RC, Lane MM, McCue C. Pulmonary autograft reoperation: incidence and management. Ann Thorac Surg 1996;62:450±455. [9] David TE. Aortic root aneurysms: remodeling or composite replacement? Ann Thorac Surg 1997;64:1564±1568. [10] Crawford ES, Crawford JL, Sa® HJ, Coselli JS. Redo operations for

G.B. Luciani et al. / European Journal of Cardio-thoracic Surgery 17 (2000) 602±607

[11] [12]

[13] [14] [15] [16] [17] [18]

[19] [20] [21]

[22]

recurrent aneurysmal disease of the ascending aorta and transverse aortic arch. Ann Thorac Surg 1985;40:439±455. Carrel T, Pasic M, Jenni R, Tkebuchava T, Turina MI. Reoperations after operation on the thoracic aorta: etiology, surgical techniques, and prevention. Ann Thorac Surg 1993;56:259±268. Bachet JE, Termignon JL, Dreyfus G, Goudot B, Martinelli L, Piquois A, Brodaty D, Dubois C, Delentdecker P, Guilmet D. Aortic dissection: prevalence, cause, and results of late reoperations. J Thorac Cardiovasc Surg 1994;108:199±205. Dougenis D, Daily BB, Kouchoukos NT. Reoperations on the aortic root and ascending aorta. Ann Thorac Surg 1997;64:986±992. Cohn LH, Aranki SF, Rizzo RJ, Adams DH, Cogswell KA, Kinchla NM, Couper GS, Collins Jr JJ. Decrease in operative risk of reoperative valve surgery. Ann Thorac Surg 1993;56:15±20. Pieters FA, Widdershoven JW, Gerardy AC, Geskes G, Cheriex EC, Wellens HJ. Risk of aortic dissection after aortic valve replacement. Am J Cardiol 1993;72:1043±1047. Prenger K, Pieters F, Cheriex E. Aortic dissection after aortic valve replacement: incidence and consequences for strategy. J Card Surg 1994;9:495±498. Nistri S, Sorbo MD, Marin M, Palisi M, Scognamiglio R, Thiene G. Aortic root dilatation in young men with normally functioning bicuspid aortic valves. Heart 1999;82:19±22. Sundt 3rd TM, Rasmi N, Wong K, Radley-Smith R, Khaghani A, Yacoub MH. Reoperative aortic valve operation after homograft root replacement: surgical options and results. Ann Thorac Surg 1995;60:S95±S99. Hahn C, Tam SK, Vlahakes GJ, Hilgenberg AD, Akins CW, Buckley MJ. Repeat aortic root replacement. Ann Thorac Surg 1998;66:88±91. Luciani GB, Casali G, Tomezzoli A, Mazzucco A. Recurrence of aortic insuf®ciency after aortic root remodeling with valve preservation. Ann Thorac Surg 1999;67:1849±1852. Pessotto R, Santini F, Pugliese P, Montalbano G, Luciani GB, Faggian G, Bertolini P, Mazzucco A. Preservation of the aortic valve in acute type A aortic dissection complicated by aortic regurgitation. Ann Thorac Surg 1999;67:2010±2013. 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;13:576±580.

Appendix A. Conference discussion Dr M. Irarrazaval (Santiago, Chile): Could you tell us a little bit about the incision. Do you have special tricks on very extended arches? Do you sometimes do something like T incisions, or are all of them approached through a median sternotomy? Also, do you always use the femoral cannulation or sometimes you use subclavians, to have a better perfusion system? Dr Luciani: Your question is relative to the cases which had associated complete arch replacement? Dr Irarrazaval: Yes. Dr Luciani: These were nine in our series, so it's a limited number. We approached those through a median sternotomy. We did not attempt any of these combined incisions or hemiclamshell. In those cases, perfusion through the groin was established; and in most of those cases, also, selected cannulation of the brachiocephalic vessels was done generally, through separate incisions on the neck, but less commonly with direct cannulation like the Kazui technique. Dr N. Kouchoukos (St. Louis, Missouri, USA): I noted in your indications for operation that you had no patients who had had previous coronary

607

bypass grafting. In our practice, we are seeing more patients present for reoperation with dissections and aneurysms involving the ascending aorta and the root in patients who previously had coronary bypass grafting. I wonder if you have any experience with such patients. Our experience, to some degree, reinforces your statement on your last slide that in patients who have moderate dilatation of the ascending aorta, and who may have coexisting coronary disease, consideration should be given to replacing those aortas at the time of the coronary bypass operation. Dr Luciani: Yes, I have to admit we had in comparison a more limited experience with those patients. The other reason why we felt it important to exclude those patients was that sometimes the operation was required based on the longevity of the bypass graft, and this was a different indication compared to the majority of patients. We somewhat feel that most of the indications we have presented in our series have many things in common. Indeed, many of the subsets of patients with aortic valve pathology present with intrinsic root pathology. And this has been increasingly shown by reports on bicuspid aortic valve. And certainly when you have acute dissection, no matter how you repair it, you may be exposed to root pathology, residual or recurrent, and to arch pathology. Although realizing that our series includes a very heterogeneous group of patients, we believe these patients have more in common than they have with patients requiring coronary bypass grafting. I'm aware of a landmark publication out of the Cleveland Clinic which has examined the effect of acute dissection on patients with prior bypass grafting. The authors came to conclusions similar to ours. But again, our experience with that subset was limited and we arbitrarily chose to exclude those patients from our series. I de®nitely would agree empirically with your idea of being more aggressive in replacing segments of the ascending aorta even in patients needing coronary bypass operations. Dr M. Turina (Zurich, Switzerland): You made a suprising statement, that you always use the coronary button technique. Could you give us more details; particularly, how do you perform the mobilization of the coronary buttons? I ®nd this extremely dif®cult on a second or third redo, when the coronary ostia are usually a mass of calcium and foreign material. And if the aneurysm is 10 or 12 cm in diameter, the left coronary artery will have to be mobilized for about 2 or 3 cm. In those cases, I'm using the Cabrol type of anastomosis of a small graft which is placed between the left and right coronary artery and anastomosed to the aortic graft. You seem not to use this technique. How do you mobilize the coronary arteries for this length to perform a safe anastomosis? Dr Luciani: It's obviously a matter of the patient subsets you're looking at. We have been possibly lucky not to experience many of these extensive root calci®cations in large aneurysms. We have operated mostly on aneurysms that were either due to dissection or degenerative disease. And in those cases, it's not dif®cult even in reoperative procedures to mobilize the coronaries. Sometimes we have resorted to the utilization of a segment of saphenous vein graft interposed between the coronary and the ascending aortic wall. But again, this has been possible due to the type of patients we operated on. In our experience it has not been necessary to resort to the Cabrol technique and only occasionally to the interposition of short segments of saphenous vein grafts. Dr C. Minale (Genova, Italy): Have you had in your set of patients any of them in which, the aneurysm was extended down to the descending aorta? And what have you done in such cases? Dr Luciani: The question is whether we had associated descending aortic pathology? Dr Minale: Yes, the aneurysm that extended down to the descending aorta. Dr Luciani: In some cases requiring arch replacement, primarily for dissection and not for sclerotic aneurysms, we have used the elephant trunk technique, but we have not done simultaneous replacement of the entire intrathoracic aorta.