Composite Valve Graft Replacement of the Proximal Aorta: Comparison of Techniques in 348 Patients Lars G. Svensson, MD, PhD, E. Stanley Crawford, MD, Kenneth R. Hess, MS, Joseph S. Coselli, MD, and Hazim J. Safi, MD Department of Surgery, Baylor College of Medicine, Houston, Texas
Composite valve graft replacement of the ascending aorta is being increasingly used, although it is not clear which technique, the Bentall, Cabrol, or button, is the best method for coronary artery ostial reattachment. We retrospectively analyzed our results with respect to these three techniques in 348 consecutive patients operated on between September 17, 1979, and January 29, 1991. Variables included aortic arch replacement in 88 patients (25%), need for deep hypothermia and circulatory arrest in 119 (34%),aortic dissection in 131 (38%),acute dissection in 34 (9.8%), reoperation in 79 (23%), and insertion of St. Jude prostheses in 270 (78%).The 30-day survival rate was 91% (316/348), the in-hospital survival rate was 90% (312/348), and the 30-day incidence of postoperative new transient (n = 6) and permanent (n = 6) stroke was 3% (12/348). The 30-day survival rates for each method were as follows: Cabrol, 92% (144/157); button, 91% (39/43); and Bentall, 91% (125/137). On stepwise multivariate logistic regression analysis with control for operative date and independent prognostic factors, operative
M
any techniques have been proposed and successfully used to insert composite valve grafts and reattach the coronary artery ostia for repair of the ascending aorta [l-241. The Bentall, aortic button, and Cabrol techniques are the three most commonly used methods [4, 6, 9-17, 22-24]. Good results have been reported by Gott and colleagues [23] with the Bentall technique for patients undergoing elective repair for Marfan’s syndrome, although more recently this group has adopted the button technique without wrapping of the aortic prosthesis. In previous reports, we [ll-131 have documented that with the Bentall technique of direct reanastomosis of the coronary artery ostia, false aneurysms requiring later reoperation can form at the coronary ostial anastomoses, particularly at the left main coronary artery ostium in patients with Marfan’s syndrome [13] or in patients with aortic dissection [12]. Kouchoukos and colleagues [6] and other surgeons [17] Presented at the Twenty-eighth Annual Meeting of The Society of Thoracic Surgeons, Orlando, FL, Feb 3-5, 1992. Address reprint requests to Dr Svensson, Department of Surgery, Baylor College of Medicine, One Baylor Plaza, Houston, TX 77025.
0 1992 by The Society of Thoracic Surgeons
technique was not an independent determinant of early mortality or stroke. On late follow-up, the Kaplan-Meier 5-year survival rate was 71% with no significant difference between the groups (3-year survival: Cabrol, 76%; Bentall, 79%; and button, 81%; p = 0.28). The 3-year freedom from reoperation was 95% (Cabrol, 97%; Bentall, 91%; and button, 100%;p = 0.17). We conclude that for patients undergoing reoperation or complicated repairs or when tension on the ostial anastomoses may occur, the Cabrol technique is preferable. If feasible, however, the button technique has better long-term results for both survival and rate of reoperation. An alternative technique is to use an interposition graft to reattach the left coronary artery and excise an aortic button for the right coronary artery reattachment. This has the advantages of technical ease in reattaching the left coronary artery, good results for reattachment of the right coronary artery, minimal tension on the anastomoses, and visualization of all anastomoses. (Ann Thorac Surg 1992;54:427-39)
have stressed the importance of not wrapping the aneurysm wall around the aortic graft because of the increased incidence of false aneurysms associated with wrapping. The cause of the false aneurysms appears to be the tension exerted on the anastomoses when the gap between the aneurysm wall and the tubular graft is excessive, especially when the accumulation of blood within the aneurysm wrap results in increased tension on the anastomoses. Furthermore, tamponade of the aortic prosthesis by the accumulated blood within the wrap can lead to an iatrogenic ascending aortic coarctation. Moreover, visualization of bleeding from the anastomoses can be difficult. To obviate this problem, some surgeons [5, 7, 8, 121have advocated excision of the coronary ostial buttons, mobilization of the proximal coronary arteries, and then reattachment of the ostia to the composite valve graft. This allows an increased length of artery that should bridge the gap between the aneurysm wall and the graft with minimal tension on the anastomoses. The drawbacks of this technique have been the time required to dissect out the ostia, particularly the ostium of the left coronary artery, and the risk of damage, or occlusion by tension, to the left main coronary artery, the circumflex artery, or the 0003-4975/92/$5.OO
428
SVENSSON ET AL COMPOSITE VALVE GRAFTS
first septa1perforator. Furthermore, in patients with aortic dissection, mobilization of the coronary ostia may not be feasible. If bleeding should occur at the anastomosis between the left coronary ostium and the graft, as can also happen with the Bentall technique, obtaining hemostasis by adding sutures to the anastomosis can be very difficult because it is on the posterior aspect of the composite graft. Thus, this technique has the potential for a higher operative mortality rate and increases the risk of uncontrolled bleeding at the anastomoses. Nevertheless, the long-term results in patients who safely undergo this type of repair, whether for Marfan’s syndrome, aortic dissection, or medial degenerative disease, have been excellent [11-13], and no patient in our experience has required reoperation [11-13]. In an attempt to make insertion of a composite graft easier, Cabrol and colleagues [3, 41 developed a technique whereby a 10-mm interposition tube graft was looped between the two coronary ostia, and then the loop was attached to the ascending aortic composite valve graft by a side-bo-side anastomosis. This approach also allowed easier access to the anastomoses for the purpose of obtaining hemostasis at the suture lines, including the aortic valve annulus and the coronary ostial anastomoses. The problem with this technique is that if the tube graft is placed to the left of the ascending aortic graft, the limb to the left coronary artery tends to kink. Similarly, with placement of the tube graft to the right of the ascending aortic graft, our preferred method [9], the side-to-side anastomosis has to be placed high on the aortic graft and in a left anterolateral position so that the graft to the right coronary artery does not kink. The left limb, however, can kink because of excessive tension over its long course at the tethered side-to-side anastomosis. Thus, the right limb can still be dysfunctional, or it can occlude because of both the angle at which the tube graft to the right coronary artery ostium has to be placed and the fact that the ostium has to be rotated through 90 degrees at the anastomosis. We [13] have also noted on long-term follow-up that the limb to the right coronary artery can occlude. Because all the techniques just discussed have some technical drawbacks with the possible risks of reoperation or graft occlusion, we retrospectively reviewed our experience with these three techniques in an attempt to offer some guidelines as to their use. We also examined our results (according to whether the composite graft was wrapped or not with the aneurysm sac when the Bentall technique was used. Kouchoukos and co-workers [6] found that the results are improved if the graft is not wrapped.
Ann Thorac Surg 1992;54:427-39
Material and Methods
whom had previously undergone composite valve graft insertion elsewhere by the Bentall technique. They were excluded because the old graft had not been excised in all of them, as the false aneurysms could sometimes be repaired, and because such patients are particularly prone to requiring reoperation a third time as we [12] have documented previously (p = 0.0004). The other patient who was excluded had undergone a Cabrol repair elsewhere and had occlusion of the graft to the right coronary artery ostium [13]. Patients who did not respond to a two-page questionnaire concerning their health and further operations were directly contacted by telephone, or their physicians were contacted for follow-up. All patients had at least a 30-day or in-hospital follow-up. A clinical profile of the 348 patients is provided in Table 1. Briefly, 29% (101/348)were female; 16%(57/348)were in aneurysm symptom grade I11 or IV (namely, severe unrelenting pain [grade 1111 or acute dissection, rupture, shock, or organ failure [grade IV]) [ll-131; aortic dissection was present in 38% (131/348) and was acute in 9.8% (34/348); 33% (113/348) had Marfan’s syndrome; 27% (94/ 348) had concurrent aortic aneurysms at distal sites; 23% (79/348) were undergoing reoperation for a procedure done elsewhere (excluding the patients already indicated who had reoperation for composite grafts done elsewhere); 25% (88/348) required aortic arch repair; and 34% (119/348)underwent repair using deep hypothermia with circulatory arrest usually for aortic arch disease or acute dissection. The method of coronary artery reattachment was a modified Cabrol technique in 45% (157/348), the Bentall technique in 39% (137/348), the button technique in 12% (43/348), and other techniques in 3% (11/348). The last category included methods for complicated problems requiring pulmonary autografts, homografts, and oversewing of the coronary ostia with placement of distal coronary artery bypass grafts either for dissection extending into the coronary arteries or for failure of a Cabrol repair because of occlusion of the 10-mm interposition graft (1 patient). Five patients who underwent a Cabrol repair required reversed saphenous vein grafts to the distal right coronary artery intraoperatively because of inadequate flow through the right Cabrol limb to the right coronary ostium with resulting right ventricular dysfunction. St. Jude valve prostheses were used in 78% (270/348)of the patients and Bjork-Shiley prostheses in the remainder (22%, 78/348). After 1985, only St. Jude prostheses were used because of their good flow characteristics, performance, and long-term durability [ll]. Another advantage was the broad sewing ring, which increased the contact surface area with the aortic annulus, thus reducing the risk of bleeding from the annular anastomosis. No composite valve grafts with biological valves were inserted.
Of 1,108 patients who underwent operation on the ascending aorta, aortic arch, or both, 348 had insertion of a composite valve graft. The records of all 348 patients, seen between September 17, 1979, and January 29, 1991 (11.4 years), were reviewed, and long-term follow-up was obtained. Excluded from the study were 31 patients, 30 of
The operative techniques for insertion of composite valve grafts have been well described [l-241, including the modified Cabrol technique that we used in these patients [9]. Deep hypothermia and circulatory arrest as described
Operative Technique
SVENSSON ET AL COMPOSITE VALVE GRAFTS
Ann Thorac Surg 1992:5442739
429
Table 1. Univariate Analysis of 30-Day Deaths Variable Age (Y) 1136 3746 47-61 62-78 Sex Female Male NYHA dyspnea class I
I1 111 IV NYHA angina class I I1 IIWV
No. of Patients
No. of Deaths
("/.I
Odds Ratio (95% CI)
p Value
1 0.82 (0.25, 2.68) 1.13 (0.39, 3.25) 2.29 (0.86, 6.15)
0.17
1 2.36 (0.88, 6.30)
0.080
190 106 39 13
1 0.77 (0.31, 1.93) 1.60 (0.55, 4.66) 4.83 (1.34, 17.5)
0.031
299 40 9
1 1.50 (0.54, 4.16) 1.31 (0.16, 10.9)
0.72
1 3.75 (1.28, 10.9) 12.5 (2.54, 61.9) 11.8 (3.90, 35.6)
Composite valve graft replacement of the ascending aorta is being increasingly used, although it is not clear which technique, the Bentall, Cabrol, or button, is the best method for coronary artery ostial reattachment. We retrospectively analyzed our results with respect to these three techniques in 348 consecutive patients operated on between September 17, 1979, and January 29, 1991. Variables included aortic arch replacement in 88 patients (25%), need for deep hypothermia and circulatory arrest in 119 (34%),aortic dissection in 131 (38%),acute dissection in 34 (9.8%), reoperation in 79 (23%), and insertion of St. Jude prostheses in 270 (78%).The 30-day survival rate was 91% (316/348), the in-hospital survival rate was 90% (312/348), and the 30-day incidence of postoperative new transient (n = 6) and permanent (n = 6) stroke was 3% (12/348). The 30-day survival rates for each method were as follows: Cabrol, 92% (144/157); button, 91% (39/43); and Bentall, 91% (125/137). On stepwise multivariate logistic regression analysis with control for operative date and independent prognostic factors, operative
M
any techniques have been proposed and successfully used to insert composite valve grafts and reattach the coronary artery ostia for repair of the ascending aorta [l-241. The Bentall, aortic button, and Cabrol techniques are the three most commonly used methods [4, 6, 9-17, 22-24]. Good results have been reported by Gott and colleagues [23] with the Bentall technique for patients undergoing elective repair for Marfan’s syndrome, although more recently this group has adopted the button technique without wrapping of the aortic prosthesis. In previous reports, we [ll-131 have documented that with the Bentall technique of direct reanastomosis of the coronary artery ostia, false aneurysms requiring later reoperation can form at the coronary ostial anastomoses, particularly at the left main coronary artery ostium in patients with Marfan’s syndrome [13] or in patients with aortic dissection [12]. Kouchoukos and colleagues [6] and other surgeons [17] Presented at the Twenty-eighth Annual Meeting of The Society of Thoracic Surgeons, Orlando, FL, Feb 3-5, 1992. Address reprint requests to Dr Svensson, Department of Surgery, Baylor College of Medicine, One Baylor Plaza, Houston, TX 77025.
0 1992 by The Society of Thoracic Surgeons
technique was not an independent determinant of early mortality or stroke. On late follow-up, the Kaplan-Meier 5-year survival rate was 71% with no significant difference between the groups (3-year survival: Cabrol, 76%; Bentall, 79%; and button, 81%; p = 0.28). The 3-year freedom from reoperation was 95% (Cabrol, 97%; Bentall, 91%; and button, 100%;p = 0.17). We conclude that for patients undergoing reoperation or complicated repairs or when tension on the ostial anastomoses may occur, the Cabrol technique is preferable. If feasible, however, the button technique has better long-term results for both survival and rate of reoperation. An alternative technique is to use an interposition graft to reattach the left coronary artery and excise an aortic button for the right coronary artery reattachment. This has the advantages of technical ease in reattaching the left coronary artery, good results for reattachment of the right coronary artery, minimal tension on the anastomoses, and visualization of all anastomoses. (Ann Thorac Surg 1992;54:427-39)
have stressed the importance of not wrapping the aneurysm wall around the aortic graft because of the increased incidence of false aneurysms associated with wrapping. The cause of the false aneurysms appears to be the tension exerted on the anastomoses when the gap between the aneurysm wall and the tubular graft is excessive, especially when the accumulation of blood within the aneurysm wrap results in increased tension on the anastomoses. Furthermore, tamponade of the aortic prosthesis by the accumulated blood within the wrap can lead to an iatrogenic ascending aortic coarctation. Moreover, visualization of bleeding from the anastomoses can be difficult. To obviate this problem, some surgeons [5, 7, 8, 121have advocated excision of the coronary ostial buttons, mobilization of the proximal coronary arteries, and then reattachment of the ostia to the composite valve graft. This allows an increased length of artery that should bridge the gap between the aneurysm wall and the graft with minimal tension on the anastomoses. The drawbacks of this technique have been the time required to dissect out the ostia, particularly the ostium of the left coronary artery, and the risk of damage, or occlusion by tension, to the left main coronary artery, the circumflex artery, or the 0003-4975/92/$5.OO
428
SVENSSON ET AL COMPOSITE VALVE GRAFTS
first septa1perforator. Furthermore, in patients with aortic dissection, mobilization of the coronary ostia may not be feasible. If bleeding should occur at the anastomosis between the left coronary ostium and the graft, as can also happen with the Bentall technique, obtaining hemostasis by adding sutures to the anastomosis can be very difficult because it is on the posterior aspect of the composite graft. Thus, this technique has the potential for a higher operative mortality rate and increases the risk of uncontrolled bleeding at the anastomoses. Nevertheless, the long-term results in patients who safely undergo this type of repair, whether for Marfan’s syndrome, aortic dissection, or medial degenerative disease, have been excellent [11-13], and no patient in our experience has required reoperation [11-13]. In an attempt to make insertion of a composite graft easier, Cabrol and colleagues [3, 41 developed a technique whereby a 10-mm interposition tube graft was looped between the two coronary ostia, and then the loop was attached to the ascending aortic composite valve graft by a side-bo-side anastomosis. This approach also allowed easier access to the anastomoses for the purpose of obtaining hemostasis at the suture lines, including the aortic valve annulus and the coronary ostial anastomoses. The problem with this technique is that if the tube graft is placed to the left of the ascending aortic graft, the limb to the left coronary artery tends to kink. Similarly, with placement of the tube graft to the right of the ascending aortic graft, our preferred method [9], the side-to-side anastomosis has to be placed high on the aortic graft and in a left anterolateral position so that the graft to the right coronary artery does not kink. The left limb, however, can kink because of excessive tension over its long course at the tethered side-to-side anastomosis. Thus, the right limb can still be dysfunctional, or it can occlude because of both the angle at which the tube graft to the right coronary artery ostium has to be placed and the fact that the ostium has to be rotated through 90 degrees at the anastomosis. We [13] have also noted on long-term follow-up that the limb to the right coronary artery can occlude. Because all the techniques just discussed have some technical drawbacks with the possible risks of reoperation or graft occlusion, we retrospectively reviewed our experience with these three techniques in an attempt to offer some guidelines as to their use. We also examined our results (according to whether the composite graft was wrapped or not with the aneurysm sac when the Bentall technique was used. Kouchoukos and co-workers [6] found that the results are improved if the graft is not wrapped.
Ann Thorac Surg 1992;54:427-39
Material and Methods
whom had previously undergone composite valve graft insertion elsewhere by the Bentall technique. They were excluded because the old graft had not been excised in all of them, as the false aneurysms could sometimes be repaired, and because such patients are particularly prone to requiring reoperation a third time as we [12] have documented previously (p = 0.0004). The other patient who was excluded had undergone a Cabrol repair elsewhere and had occlusion of the graft to the right coronary artery ostium [13]. Patients who did not respond to a two-page questionnaire concerning their health and further operations were directly contacted by telephone, or their physicians were contacted for follow-up. All patients had at least a 30-day or in-hospital follow-up. A clinical profile of the 348 patients is provided in Table 1. Briefly, 29% (101/348)were female; 16%(57/348)were in aneurysm symptom grade I11 or IV (namely, severe unrelenting pain [grade 1111 or acute dissection, rupture, shock, or organ failure [grade IV]) [ll-131; aortic dissection was present in 38% (131/348) and was acute in 9.8% (34/348); 33% (113/348) had Marfan’s syndrome; 27% (94/ 348) had concurrent aortic aneurysms at distal sites; 23% (79/348) were undergoing reoperation for a procedure done elsewhere (excluding the patients already indicated who had reoperation for composite grafts done elsewhere); 25% (88/348) required aortic arch repair; and 34% (119/348)underwent repair using deep hypothermia with circulatory arrest usually for aortic arch disease or acute dissection. The method of coronary artery reattachment was a modified Cabrol technique in 45% (157/348), the Bentall technique in 39% (137/348), the button technique in 12% (43/348), and other techniques in 3% (11/348). The last category included methods for complicated problems requiring pulmonary autografts, homografts, and oversewing of the coronary ostia with placement of distal coronary artery bypass grafts either for dissection extending into the coronary arteries or for failure of a Cabrol repair because of occlusion of the 10-mm interposition graft (1 patient). Five patients who underwent a Cabrol repair required reversed saphenous vein grafts to the distal right coronary artery intraoperatively because of inadequate flow through the right Cabrol limb to the right coronary ostium with resulting right ventricular dysfunction. St. Jude valve prostheses were used in 78% (270/348)of the patients and Bjork-Shiley prostheses in the remainder (22%, 78/348). After 1985, only St. Jude prostheses were used because of their good flow characteristics, performance, and long-term durability [ll]. Another advantage was the broad sewing ring, which increased the contact surface area with the aortic annulus, thus reducing the risk of bleeding from the annular anastomosis. No composite valve grafts with biological valves were inserted.
Of 1,108 patients who underwent operation on the ascending aorta, aortic arch, or both, 348 had insertion of a composite valve graft. The records of all 348 patients, seen between September 17, 1979, and January 29, 1991 (11.4 years), were reviewed, and long-term follow-up was obtained. Excluded from the study were 31 patients, 30 of
The operative techniques for insertion of composite valve grafts have been well described [l-241, including the modified Cabrol technique that we used in these patients [9]. Deep hypothermia and circulatory arrest as described
Operative Technique
SVENSSON ET AL COMPOSITE VALVE GRAFTS
Ann Thorac Surg 1992:5442739
429
Table 1. Univariate Analysis of 30-Day Deaths Variable Age (Y) 1136 3746 47-61 62-78 Sex Female Male NYHA dyspnea class I
I1 111 IV NYHA angina class I I1 IIWV
No. of Patients
No. of Deaths
("/.I
Odds Ratio (95% CI)
p Value
1 0.82 (0.25, 2.68) 1.13 (0.39, 3.25) 2.29 (0.86, 6.15)
0.17
1 2.36 (0.88, 6.30)
0.080
190 106 39 13
1 0.77 (0.31, 1.93) 1.60 (0.55, 4.66) 4.83 (1.34, 17.5)
0.031
299 40 9
1 1.50 (0.54, 4.16) 1.31 (0.16, 10.9)
0.72
1 3.75 (1.28, 10.9) 12.5 (2.54, 61.9) 11.8 (3.90, 35.6)
