Concomitant resection of ascending aortic aneurysm and replacement of the aortic valve

J

THORAC CARDIOVASC SURG

79:388-401, 1980

Concomitant resection of ascending aortic aneurysm and replacement of the aortic valve Operative and long-term results with "conventional" techniques in ninety patients We reviewed a consecutive series of 90 patients undergoing concomitant resection of ascending aortic aneurysm and aortic valve replacement (AVR) utilizing noncomposite "conventional" techniques in order to assess the early and late results, to define limitations of this operative approach, and thereby to clarify the indications for composite reconstruction of the aortic root. Mean age was 55 years. Twenty percent had Marfan's syndrome, and 13 % had aortic dissections. The cause of the aneurysm was dissection in 13% of cases, syphilis in 11%, atherosclerosis in 9%, and degeneration (with or without cystic medionecrosis) in 67%. Follow-up averaged 3.8 years and extended to 11.5 years maximum. AVR and complete excision of the aneurysm (preserving small tongues of aortic wall circumscribing the coronary artery ostia) coupled with tubular graft replacement of the ascending aorta were performed. Nineteen percent of patients required individual technical modifications relating to the coronary arteries. Operative mortality rate was 13%, with the majority of deaths being due to cardiac causes. Contemporary (1975 to 1978) operative mortality rate was 4.3%. Seven percent required re-explorationfor hemorrhage and 2.4% had perioperative myocardial infarctions. Late functional results were generally good (average N.Y.H.A. Class 1.4). Late thromboembolism, angina, myocardial infarction, and congestive heart failure occurred at linearized rates of 3.4% per patient-year, 4.9% per patient-year, 1.1 % per patient-year, and 5.2% per patient-year, respectively. No prosthetic valve endocarditis, graft infection, or recurrent aneurysms of the aortic root were observed. Late reoperation was necessary in eight patients (3% per patient-year), but reoperation for disease confined to the ascending aorta accounted for only three of these cases ( 1.1 % per patient-year). Overall actuarial survival rates were 67% :!: 5% at 5 years and 50% :!: 9% at 10 years; survival rates for the 78 operative survivors were 77% :!: 5% and 57% :!: 10% at the same time intervals, respectively. Only one late death could be attributed to complications arising in the reconstructed aortic root. These results confirm that such simple, noncomposite techniques are safe, portend minimal risk of late complications and the attendant necessity for reoperation, and provide satisfactory long-term survival. We believe that composite techniques should be primarily reserved for selected cases of advanced necrotizing prosthetic or natural endocarditis.

D. Craig Miller, M.D., Edward B. Stinson, M.D., Philip E. Oyer, M.D., Ricardo J. Moreno-Cabral, M.D., Bruce A. Reitz, M.D., Stephen J. Rossiter, M.D., and Norman E. Shumway, M.D., Stanford, Calif.

T

he evolution of surgical treatment of aneurysms of the thoracic aorta has been punctuated by substantial

From the Department of Cardiovascular Surgery, Stanford University School of Medicine, Stanford, Calif. Read at the Fifth Annual Meeting of The Samson Thoracic Surgical Society, Sun Valley, Idaho, June 5 to 8, 1979. Address for reprints: D. C. Miller, M.D., Department of Cardiovascular Surgery, Stanford University Medical Center, Stanford, California 94305.

388

advances since 1952, when Cooley and De Bakey' stated: "Aneurysm of the thoracic aorta is a fatal disorder, comparable in this respect to cancer, emphasizing the need for an aggressive attitude with emphasis on removal of the diseased tissue and reconstruction of the vessel where possible ... The gravity of the disease fully justifies exploratory thoracotomy, not only to establish the diagnosis, but also to determine the applicability of appropriate surgical treatment. " However, surgical treatment of ascending aortic aneurysms asso-

0022-5223/80/030388+14$01.40/0 © 1980 The C. V. Mosby Co.

Volume 79

Resection of ascending aortic aneurysm and AVR

Number 3

389

March,1980

Table I. Preoperative characteristics AVR + AscAo resection (N = 90) Mean age (±SEM) (yr) Age range (yr) Male: female ratio Marfan's syndrome* (%) Arterial hypertensiont (%) Angina pectorist (%) Myocardial infarction (%) Congestive heart failure (%) Aortic dissection * (%) Angiographic CAD Angiographic LV dyssynergy (%) AS* (%) Mixed AS/AR (%) AR*(%) Prosthetic dysfunction (%) AverageN.Y.H.A. Class (±SEM)

55.3 ± 1.5 20-78 3.1: 1

20 32 33

14 62 =

56.5 ± 0.4 8-84 3.0: 1 0.4 18 48 13

13

9/31

AVR only (N = 1,349)

29%

36 13

17 67 3 2.7 ± 0.07

70 0.1 272/615 = 44% 44 58 13

22 6 2.8 ± 0.02

Legend: AscAo, Ascending aortic aneurysm. AS, Aortic stenosis. AR, Aortic regurgitation. AVR, Aortic valve replacement. CAD, Coronary artery disease. LV, Left ventricular. N.Y.H.A., New York Heart Association. SEM, Standard error of the mean.

*p < 0.01. tp < 0.05.

ciated with aortic valvular disease has remained a formidable technical challenge. In an attempt to reduce operative morbidity and mortality, Bentall and DeBono, 2 Singh and Bentall," and Edwards and Kerr' described techniques for complete replacement of the ascending aorta which employed a composite tubular graft containing a prosthetic valve and reimplantation of the coronary artery ostia. Additional experience':" demonstrated that application of such techniques was associated with operative mortality rates ranging between 4% and 17% and late attrition rates ranging up to 32%. Also, complications unique to this method, occasionally mandating urgent reoperation, have been described.P- 8-13 Despite these drawbacks, broader and enthusiastic application of this technique has been advocated.v ": 11, 12, 14-16 This present study was therefore carried out to assess our experience with these combined entities. We utilized a more conventional technique in order to define the necessity and indications for the composite technique. Patients and methods

Between July, 1965, and June, 1968, a consecutive series of 90 patients underwent concomitant resection of ascending aortic aneurysm and aortic valve replacement (AVR) on the University Cardiovascular Surgical Services at Stanford University Medical Center. Patients undergoing multiple valve replacements were excluded. Age at operation ranged from 20 to 78 years. A concurrent comparison population was provided by

1,349 patients undergoing isolated AVR between 1963 and 1978. These latter patients do not represent a control group, owing to variance in extent and type of disease, but they do add perspective pertaining to operative mortality and long-term survival. Table I summarizes selected preoperative characteristics of these two groups and Fig. 1 illustrates the hemodynamic profile of those patients in both groups who underwent preoperative catheterization. The causes of the ascending aortic aneurysms were as follows: aortic dissection in 12 cases (acute in five), degeneration (idiopathic and/or cystic medionecrosis) in 60 cases (sinus of Valsalva aneurysm in three), syphilis in 10 cases, and atherosclerosis in eight cases. A piece of aortic wall was not examined histologically in all cases, but cystic medionecrosis was present in 22 specimens. Fourteen of the 90 patients (16%) had previously undergone a cardiovascular surgical procedure an average of 5.5 years prior to concomitant replacement of the ascending aorta and AYR. These operations included seven AVRs, two aortic valvuloplasties, one septal myomectomy for idiopathic hypertrophic subaortic stenosis, one coarctectomy, and three abdominal aortic aneurysmectomies. Methods of patient follow-up have been described previously in detail. 17 Closing intervals varied between 1975 (allograft valves) and 1978 (xenograft valves). A total of 268 patient-years of follow-up for the 90 patients undergoing concomitant resection of ascending

The Journal of

390 Miller et al.

Thoracic and Cardiovascular Surgery

:=E w 5(1) +1

!2AVR + ASCENDING AORTIC REPLACEMENT l'f]AVR ONLY

N

OPERATIVE RANGE: 1963 - 1978

4..§

c:

'E 3:J w ~ 20 ~

W

II: Q..

10

Eil_

n- 50 967

RA

I_I~ JI_I

54 1,007

49 966

------ -----PA Sy st

PA

p = .01

p = ,03

49901 PCW

38 878

LVED

u

«

C

II:

47 971

« U

C.1.

Fig. 1. Comparative preoperative hemodynamics for the studypopulation of 90 patientsundergoing simultaneous ascending aortic aneurysm resection and aortic valve replacement and 1,349 patients concurrently undergoing isolated aortic valve replacement. SEM, Standard error of the mean. AVR, Aortic valve replacement. RA, Mean right atrial. PAsysf, Pulmonary artery systolic. PA, Mean pulmonary artery. pew, Mean pulmonary capillary wedge. LVED, Left ventricular end-diastolic. C./., Cardiac index. aortic aneurysm and AVR was available for analysis; two were lost to follow-up (2%). Average follow-up span was 3.8 years and maximum was 11.5 years. Thirty-two patients were followed for 4 years or longer. For perspective, a total of 4,510 patient-years of follow-up was accumulated for the 1,349 patients undergoing AVR alone; 21 patients were lost to follow-up (1.6%). Average follow-up interval was 3.8 years and extended to 11.3 years. Four hundred sixty-three patients in this group were followed for 4 years or longer. Standard nonparametric actuarial statistical methods were employed to analyze time-dependent discrete events (e.g., death and reoperation). Noncontinuous variables were compared by means of chi square contingency analysis or the Fisher exact test. Continuous variables were compared by Student's t test. Linearized rates were compared by Cox's F-test or the maximal likelihood ratio hypothesis. Two-tailed significance levels were assigned to the p values thereby obtained. These methods have been cited individually in a previous publication;"

Operative technique Low-flow, low-pressure cardiopulmonary bypass with moderate (30° to 32° C) systemic hypothermia was employed. Left ventricular venting was not generally used. Profound topical hypothermia (occasionally with both a pericardial and an endocardial continuous flush line of 4° C) saline provided myocardial protection. Re-

cently, cold cardioplegia has been used to complement the topical cooling in certain cases. The average (±SEM) aortic cross-clamp time was 79 ± 2 minutes, and the average cardiopulmonary bypass time was 125 ± 4 minutes. Both of these times were significantly longer (p < 0.001) than those for the group of 1,349 patients undergoing AVR with or without concomitant coronary artery bypass grafting (65 ± 1 and 100 ± 1 minutes, respectively). A single crossclamp period was used for valve replacement, aneurysm resection, and construction of all distal coronary artery bypass graft anastomoses (16% underwent concomitant coronary artery bypass grafting). A Starr-Edwards prosthesis was used in 59% of the cases, an allograft valve substitute was used in 4%, and a porcine xenograft bioprosthesis* in 37%. A similar distribution of types of prostheses was evident in the larger group undergoing isolated AYR. The technique utilized was basically that initially described by Wheat and associates" in 1964. The aneurysm was opened and the aortic valve completely excised. Valve replacement was then carried out in a routine manner, with the use of interrupted mattress sutures of braided polyester (Fig. 2). The proximal aortic root was then completely transected just above the bioprosthetic sewing ring, with only small (3 to 5 mm) tongues of aortic wall left surrounding the coro*Hancock Laboratories, Anaheim, Calif.

Volume 79

Resection of ascending aortic aneurysm and A VR

Number 3 March,198O

39 1

Fig. 2. Operative technique (A). Aortic valve replacement.

nary artery ostia. Thereafter, the aneurysm was completely excised distally. A woven or low-porosity woven Dacron tubular graft of appropriate size was then sewn to the small rim of remaining aortic root (Fig. 3). The posterior proximal suture line was constructed first, and accurate full-thickness suture placement was assured, for this area later becomes relatively inaccessible. Finally, the distal anastomosis was constructed in a similar fashion with the use of continuous polyester or polypropylene (Prolene) suture (Fig. 4). The patient was placed in a steep Trendelenburg position, the heart was allowed to fill with blood, air was evacuated, and the cross-clamp was released. In cases of aortic dissection or cystic medionecrosis in which the structural integrity of the aortic tissue was markedly reduced, circumferential buttressing strips of Teflon felt were incorporated into the aortic suture lines (usually just on the adventitial surface). Furthermore, the false lumen in cases of aortic dissection was obliterated both proximally and distally by continuous

polypropylene suture prior to performance of the graft anastomoses. In 17 cases (19%) various individual technical modifications were necessary to revascularize the coronary arteries. Fourteen patients underwent concomitant coronary artery bypass grafting with reversed saphenous vein conduits. This procedure was necessary to bypass intrinsic coronary atherosclerotic stenoses in 10 cases, to bypass dissected right coronary artery ostia in two cases, and to bypass an iatrogenically damaged right coronary ostium in one case. Triple bypass grafts were constructed in the fourteenth case (extra-anatomic supracoronary AVR with ligation of the native coronary ostia) because severe, necrotizing, active prosthetic endocarditis precluded inserting a new bioprosthesis in the region of the destroyed natural aortic anulus. In two cases, either the left or the right coronary ostium was reimplanted into the prosthetic graft as a Carrel patch (Fig. 5). This was necessary in cases of annulo-aortic ectasia when the coronary ostium was

The Journal of

392 Miller et al.

Thoracic and Cardiovascular Surgery

Fig. 3. Operative technique (B). Excision of the ascending aortic aneurysm, with preservation of a small rim of aorta circumscribing the coronary ostia to which the proximal graft is anastomosed.

markedly displaced superiorly away from the aortic anulus, as previously described by Najafi," ROSS,20 Symbas," and their colleagues. A similar technique involving use of a short saphenous vein interposition graft (anastomosed end to end to the main right coronary artery) was employed in another case of annuloaortic ectasia to ensure that no tension existed at the coronary artery-graft anastomosis. This modification has also been described previously by Najafi and coworkers'? and Zubiate and Kay. 15 Anastomosis of a coronary ostium from within the wall of the aneurysm "side to side" to the aortic graft (as carried out in the composite techniques of Bentall and DeBon0 2 and Edwards and Kerr") was not performed in this series. To minimize bleeding through the interstices of the graft, we preclotted regular woven Dacron grafts, occasionally using topical thrombin. The recent introduc-

tion of low-porosity woven Dacron grafts* has eliminated the necessity for preclotting the graft prior to heparinization and the threat of hemorrhage through the graft interstices. However, these low-porosity grafts are less compliant and more difficult to suture; for these reasons they are not used in all cases.

Results Operative deaths. Twelve of the 90 patients undergoing simultaneous resection of an ascending aortic aneurysm and AVR died during the initial hospitalization (irrespective of time after operation) or within 30 days, for an operative mortality rate of 13%. Within selected subsets the operative mortality rate was higher: 29% (4/14) of those patients previously operated upon (p = *Meadox Medicals, Oakland, N. J.

Volume 79 Number 3 March, 1980

0.08) and 42% (5/12) of patients with aortic dissections (p < 0.01). The most common cause of operative death was myocardial failure (7/12 = 58%). Five of the seven (71%) could not be weaned from cardiopulmonary bypass and died in the operating room. Two of these five intraoperative deaths occurred in patients who were taken to operation on an emergency basis after sustaining cardiac arrest. One patient died of hemorrhage owing to a consumptive coagulopathy. Two operative deaths were due to technical complications. In one case a suture ligature "rolled-off" the divided innominate vein and cardiac arrest resulted. This patient was returned to the operating room and cardiopulmonary bypass was reinstituted, but he could not be weaned from bypass. The other case involved disruption of the distal aortic suture line. The initial aortic graft was too short and was successfully replaced. However, myocardial function was compromised and the patient later died. One death was attributed to cerebral hypoxia: This patient (with cystic medionecrosis) underwent isolated AVR with a Starr-Edwards prosthesis in 1965; postoperative hemorrhage emanating from the aortotomy mandated four re-explorations the same day. Eventually, cardiopulmonary bypass was reinstituted and the ascending aorta was replaced with a tubular woven graft. However, the patient never regained consciousness and died 2 weeks later. Because AVR and replacement of the ascending aorta were performed the same day, this patient was included in this present series of combined procedures, although the initial operation consisted of isolated AYR. The other operative death was due to pulmonary insufficiency and sepsis. The overall operative mortality rate of 13% contrasted significantly (p < 0.05) with the 6.9% operative mortality rate for the 1,349 patients concurrently undergoing isolated AYR. However, between 1975 and 1978, 23 patients have undergone comcomitant resection of an ascending aortic aneurysm and AVR with one operative death, for a contemporary operative mortality rate of 4.3%. Operative morbidity. Excluding the five intraoperative deaths, 85 patients were at risk to develop postoperative complications. Six patients (7%) required re-exploration for hemorrhage. Two of these six patients died; in both cases technical errors were identified (vide supra). Fifteen patients (18%) required prolonged myocardial inotropic support, two (2.4%) had perioperative myocardial infarctions, and four sustained episodes of ventricular tachyarrhythmias. Permanent complete heart block occurred in five patients. Seven patients (8%) had some degree of postoperative

393

Resection of ascending aortic aneurysm and AVR

Fig. 4. Operative technique (C). Construction of the distal graft anastomosis and the completed procedure.

Table II. Late results expressed as linearized functions of elapsed time Thromboembolic events (%/pt-yr) Anticoagulant-related hemorrhage (%/pt-yr) Angina pectoris (%/pt-yr) Myocardial infarction (%/pt-yr) Congestive heart failure (%/pt-yr) Late endocarditis (%/pt-yr) Recurrent aneurysm (%/pt-yr)

3.4 4.5 4.9 1.1 5.2 0 0

Legend: pt-yr, Patient-year.

renal insufficiency, and three required temporary hemodialysis. Six patients (7%) required intensive treatment for pulmonary insufficiency, including tracheostomies in four. Two patients had focal neurologic deficits, which resolved in both. No cases of early prosthetic valve endocarditis or graft infection occurred. Late results. Average New York Heart Association

394 Miller et al .

The Journal of Thoracic and Cardiovascular Surgery

Fig. 5. Operativetechnique (D) . If coronaryartery reimplantation is necessary, a rubber 8 Fr. catheter is placed in the coronary ostium to facilitate dissection of a full-thickness button. One or both of the coronary ostia are then reimplanted into the graft as a Carrel patch. As with our standard technique, the aneurysmal sac is completely excised. functional classification of the 56 current survivors at an average of 3.8 years postoperatively was 1.4 (two patients being lost to follow -up) . All patients except one were in N .Y.H .A. Clas s I or II. The rates of postoperative thromboembolism, anticoagulant-related hemorrhage, angina pectoris, myocardial infarction, and congestive heart failure were calculated in a linearized manner as a function of cumulative follow-up time and are summarized in Table II. No cases of late prosthetic valve endocarditis or graft infection occurred. No recurrent aneurysms' of the aortic root are known to have developed , and no patient currently has residual dysfunction of an aortic prosthesis or bioprosthesis (including peri prosthetic leaks) . Thirty of the 56

current traceable survivors (54 %) are still receiving anticoagulants orally. Late reoperation. A total of eight patients required late reoperation between I and 5 years postoperatively, for a linearized rate for reoperation of 3% per patientyear . Fig. 6 depicts the actuarial rate of reoperation. Actuarial probability of being free from reoperation was 83% ± 6% at 5 years, with no additional reoperations occurring within the I 1.5 year maximum followup interval. The average interval between initial and subsequent operations was 2 year s. Three patients (l % per patient-year, or 3.8% of the 78 discharged patients) developed peri prosthetic leaks which were repaired ; one of these patients died following reoperation .

Volume 79

Resection of ascending aortic aneurysm and AVR

Number 3 March,1980

n = 90 OPERATIVE RANGE 1963 - 1978 FOLLOW-UP: AVERAGE - 3.1 YR MAXIMUM - 11.5 YR CUMULATIVE - 288 PT-YR

90

~-~XlCX)()()OOOOOClOOOO

80

(33) 8 REOPERATIONS

70

60

395

(22)

3 AORTIC PERIVALVULAR LEAKS (1.1%/PT-YR) 2 DESCENDING THORACIC AORTIC RESECTIONS (O.7%/PT-YR) 2 AAA OR AFB (O.7%/PT-YR) 1 COARCTATION (O.4%/PT-YR)

I

1

2

3 4 5 6 YEARS POSTOPERATIVE

;

I

12

Fig. 6. Actuarial probability of late reoperation. Only three of the eight late reoperations involved the ascending aorta or aortic valve. Numbers in parentheses indicate number of patients at risk at each time interval. PT-YR. Patient-years. AAA. Abdominal aortic aneurysmectomy. AFR. Aortofemoral bypass.

Neither coronary artery ostial obstruction nor recurrent aneurysm of the aortic root has been observed. Five patients required additional operative procedures for problems remote from the ascending aorta: two resections of the descending thoracic aorta (one acute and one chronic dissection), one abdominal aortic aneurysmectomy, one aortofemoral bypass, and one coarctation resection. All of these patients survived reoperation. Late survival. A total of 20 late deaths occurred among the 78 patients surviving operation. The overall survival rate calculated by the actuarial method was 67% ± 5% at 5 years and 50% ± 9% at 10 years (Fig. 7, A). The survival rate for those 78 patients surviving operation was 77% ± 5% at 5 years and 57% ± 10% at 10 years (Fig. 7, B). The 5 year overall survival rate for patients with nondissecting aneurysms was 68% ± 6%, significantly (p = 0.04) superior to the 50% ± 14% survival rate for the 12 patients with aortic dissections. This difference reflected primarily the higher operative mortality rate associated with aortic dissections since calculation of survival rates excluding operative deaths showed no significant difference (p = 0.76) between these two subsets. As implied by the parallel slopes of the survival curves in Fig. 7, A, there was no significant difference in late survival rates between those patients undergoing combined operation and those undergoing only AVR when operative deaths were excluded (p = 0.23) (Fig.

7, B). This constitutes indirect evidence that the noncomposite technique of simultaneous resection of ascending aortic aneurysm and AVR is durable and is not associated with an inordinately high risk of late serious complication or death. Autopsy diagnoses were available in 12 of the 20 late deaths. Cardiac causes were judged to be responsible for 50% (10/20) of the late deaths: congestive heart failure in five cases, acute myocardial infarction in two cases, and sudden, unexpected deaths in three cases. No autopsy revealed abnormalities involving the prosthetic valve or ascending aortic graft. Three late deaths (15%) were due to cerebrovascular accidents, and two (10%) were pulmonary in nature. One death each was caused by neoplasm, renal insufficiency, hepatitis, and suicide. The remaining late death was caused by rupture of an aneurysm of the descending thoracic aorta (atherosclerotic). Therefore, as judged by clinical, radiologic, and necropsy findings, only one late death could be attributed to complications arising in the reconstructed aortic root.

Discussion The formidable technical challenge associated with simultaneous resection of the ascending aorta and AVR has been clearly appreciated since Groves," Wheat.!" Bloodwell," and their associates first reported simultaneous surgical management of both entities. An alternative, composite technique which included reimplan-

The Journal of Thoracic and Cardiovascular Surgery

396 Miller et at.

~100

p

:::E

w

en

= 0.007

(591)

+1

{o--<>

AVR ONLV In = 1,349) ........ AVR + ASCENDING AORTIC REPLACEMENT [n = 90) OPERATIVE RANGE: 1963 - 1978

...J

ct ~ 60 > a:

::l

en ~

Z w

40

o

ffi 0.

20

--

C>--OAVERAGE FOLLOW-UP = 3.1 VR ....... AVERAGE FOLLOW-UP = 3.8 VR

°o~-~----'~-~-""""--~-~~-~-~

16

CD ~ w

100

• SIMULTANEOUS AVR + ASCENDING AORTIC REPLACEMENT (n=78)

en +\ ....I

«

> > a: ::::l

en

~ 40 w o a: ~

20

(2)

2

4

6 8 10 12 YEARS POSTOPERATIVE

14

16

Fig. 7. A, Overall actuarial survival including operative deaths. B, Actuarial survival excluding operative mortality. The previously apparent superior survival of patients undergoing aortic valve replacement alone is no longer appreciable (p = 0.23) tation of the coronary ostia was introduced by Bentall and Debono," Singh and Bentall," and Edwards and Kerr' in an effort to reduce the incidence of intraoperative bleeding complications. These problems have been of particular concern in patients with Marfan's syndrome and those with cystic medionecrosis. 13, 23-26 However, patients with atherosclerotic and syphilitic aneurysms associated with aortic valve disease also have presented a technical challenge.": 28 The composite method has been thought to have additional theoretical merit in that, since no remnant of aortic root remains, the chance of late recurrent aneu-

rysm formation or redissection is minimized. However, the evidence supporting the validity of this hypothesis is scarce. Weldon29 commented upon two cases of recurrent aneurysm causing aortic insufficiency; Anagnostopoulos"? alluded to its occurrence but did not provide substantiating data. Both referred to patients undergoing either aortic reanastomosis (for dissection) or supracoronary aortic replacement without AYR. In both of these circumstances it seems obligatory to retain all or nearly all of the aortic root in order to maintain adequate commissural support of the native aortic valve. We contend that this situation is not analogous to

Volume 79 Number 3

Resection of ascending aortic aneurysm and A VR

397

March, 1980

that presented by patients who undergo simultaneous replacement of both the ascending aorta and the aortic valve; in this situation only small tongues of aortic root are preserved. Symbas and colleagues" reported the case of one patient with Marfan's syndrome who developed a dissecting aneurysm in the remaining aortic root 2 years following (supracoronary) ascending aortic aneurysm resection and AYR. The initial tubular graft had been anastomosed to a circumferential aortic cuff 1 em distal to the coronary artery ostia. This recurrent aneurysm was successfully managed by the composite technique, but re-exploration was necessary to control hemorrhage emanating from disruption of the left coronary artery anastomosis. McCready and Pluth" reported five cases of recurrent aortic root aneurysm formation among 31 patients who were evaluated an average of 6 years postoperatively. In these cases, as well as that of Symbas and associates," relatively large, circumferential portions of aortic root had been preserved at the time of initial supracoronary ascending aortic replacement. Again, this method is not directly analogous to the technique we have used, in which only small tongues of aortic wall circumscribing the coronary ostia are preserved. Wheat and co-workers.s" in summarizing their experience with 13 patients treated with "conventional" noncomposite techniques in 1971, reported an operative mortality rate of 15% and a 17% incidence of reexploration for hemorrhage. Seven of the 11 hospital survivors (64%) died between 1 month and 5 years later. No recurrent aneurysms or redissections of the aortic root were noted, but two patients (18%) required reoperation. Liddicoat and associates;" using similar techniques in 63 patients, reported in 1975 an operative mortality rate of 11%, with 3% of patients requiring re-exploration for hemorrhage. Late attrition (average follow-up 4.7 years) was also high in this report, with 16 of 53 (30%) traceable hospital survivors dying. Although the aorta became dilated proximal to the graft in three patients no reoperations were reported. In a series of 30 patients with Marfan' s syndrome treated by various noncomposite methods, Nasrallah, Cooley, and colleagues" reported a 20% operative mortality rate. Late attrition was 45% (10/22 traceable hospital survivors) over a 3.5 year average follow-up interval. Four late deaths were due to "additional dissection of the aorta," but the site of dissection was not mentioned. The most recent sizeable report describing noncomposite techniques is from the Mayo Clinic." Twenty-five patients were operated upon between 1972 and 1977, with an operative mortality rate of 8% and a 12% incidence of re-exploration for hemorrhage. Follow-up of

this group of patients (over an average of 2.9 years) revealed a late attrition rate of 22%. Two of the five late deaths occurred following reoperation for malfunctioning Braunwald-Cutter valves. Five patients (22%) required reoperation (approximately 7% per patientyear), but four procedures were necessary because of Braunwald-Cutter valve failure. One reoperation was performed to correct a periprosthetic leak. No recurrent aortic root aneurysms were observed, and no late deaths were attributed to problems with the basic technique of reconstruction. These same authors" reported the cases of 40 additional patients who were operated upon with similar techniques during an earlier (1965 to 1972) period. For this earlier population the operative and late mortality rates were 23% (9/40) and 26% (8/ 31), respectively; however, five patients (approximately 2.7% per patient-year) required late reoperation to correct recurrent aneurysms of the aortic root over a substantially longer (6 year average) follow-up interval. In comparison, the operative mortality rate in this present series of 90 patients operated upon over a 13 year span was 13%. Recent operative mortality rate was 4.3%. Seven percent required re-exploration for hemorrhage. No patient had early complications involving the valve or aortic graft. Over the 3.8 year average follow-up interval, eight patients (10% of the 78 initial survivors) required reoperation, a linearized rate of 3% per patient-year. Only three of these eight late reoperations, however, concerned lesions related to the aortic prosthesis or aortic root (3.8% of discharged patients, or 1.1 % per patient-year). Twenty late deaths occurred, for an overall late mortality rate of 26%. The cumulative total (268 patient years) and maximum duration (11.5 years) of follow-up in this present study, however, substantially exceed those in most other reports. This makes comparison of crude late survival rates rather difficult, if not specious. Computed on the more meaningful actuarial basis, the 5 year survival rate of discharged patients was 77% ± 5% and the 10 year survival rate was 57% ± 10%. Unfortunately, actuarial life-table methods have not been utilized in previous reports. Reported series of patients undergoing concomitant resection of ascending aortic aneurysm and A VR by use of composite techniques are smaller in size and generally limited in follow-up duration. 3-9, 11, 14-16 Bentall,? in recently summarizing his experience with 23 patients at the Hammersmith Hospital, reported an operative mortality rate of 17% and a late mortality rate of 32%. Average follow-up was 6.5 years. No reoperations were mentioned. Kouchoukos" from the University of Alabama reported an operative mortality rate of

398

Miller et al.

4% among 25 patients. Two patients (8%) required early re-exploration, one for hemorrhage and one for acute prosthetic insufficiency. The late mortality rate was 13%, but follow-up periods extended only to a maximum of2.3 years. Two patients (8%) required late reoperation. The University of Minnesota group recently reported a series of 16 patients who underwent composite aneurysm and valve replacement. 6 The operative mortality rate was 6%, no patient required reexploration for bleeding, and the late attrition rate was 13% (average follow-up 2.2 years). However, four of the 13 (31%) patients who underwent postoperative arteriography were found to have either anastomotic pseudoaneurysms or periprosthetic leaks. Furthermore, three patients (20% of hospital survivors) had either undergone or had been scheduled for reoperative procedures. These authors concluded that significant late problems may develop following composite replacement of the ascending aorta and aortic valve and advised that routine postoperative aortography be performed. The recent comprehensive report by McCready and Pluth" from the Mayo Clinic also included a series of 15 patients operated upon with the use of composite techniques. One operative death (7%) occurred, and one re-exploration (7%) was performed to control postoperative hemorrhage. Over an average follow-up interval of 2.2 years, two late deaths (14%) occurred, and one late reoperation (7%) was necessary. Pseudoaneurysms, as defined by leakage of contrast material around the coronary anastomoses, were apparent in both of the two patients who underwent late aortography. However, the problem which mandated reoperation in one patient was an aneurysm of the descending thoracic aorta. One observation relevant to the composite technique has been the occurrence of unique early and late complications. Crosby and co-workers 11 reported one case in which both early and late supravalvular extrinsic aortic obstruction developed owing to a hematoma between the tubular graft and the surrounding aneurysm wall. Two reoperations were necessary to correct this problem in this one patient. Klopp's group'? and McCready and Pluth" each reported a similar complication. One of the patients in Kouchoukos '8 series developed a hematoma which impinged on the prosthetic valve disc and produced acute prosthetic insufficiency. Cabrol and associates" have even suggested that a fistula be surgically created between the aneurysmal sac and the right atrial appendage to decompress this perigraft space. Serious bleeding from the coronary arterial anastomoses, both early and late postoperatively, 10. 11. 13

The Journal of Thoracic and Cardiovascular Surgery

and coronary anastomotic pseudoaneurysms": 6 have been reported. To avoid these problems, several authors'': 13. 15, 30 have advocated reimplanting the coronary ostia end to side into the graft as full-thickness Carrel buttons, with or without interposition saphenous vein grafts. This procedure appears to have reduced the incidence of bleeding and pseudoaneurysm formation associated with the side-to-side anastomotic technique performed from within the aneurysmal sac. A potential problem, which currently has not been clearly defined, concerns re-replacement of these composite grafts if prosthetic dysfunction or prosthetic endocarditis should occur. Kouchoukos" reported one late death occurring at the time of reoperation for endocarditis, but gave no details. Prolonged cardiopulmonary bypass support increases the risk of hemorrhagic complications, because extracorporeal perfusion damages platelet membranes and denatures plasma proteins. Proponents of the composite technique have stated that one of its merits is reduction of operating time." The average duration of cardiopulmonary bypass ranged from 102 to 232 minutes in reports of composite ascending aneurysm and valve replacement." 6. 8, 9. 16 Although this is partially a function of the mode of myocardial protection employed, the average cardiopulmonary bypass time was 125 minutes in the present series, in which noncomposite techniques were employed. The average aortic cross-clamp time was 79 minutes. We have found that the introduction of low-porosity woven grafts and the availability of individual coagulation factors appears to have reduced the frequency of postoperative bleeding. Our approach has been not to rely on wrapping the graft with surrounding aneurysmal sac to help promote hemostasis. Indeed, this maneuver has evinced further complications.P- 6, 8, 10, II, 13, 31 Currently, there appears to be no real difference between the two techniques in cardiopulmonary bypass time and the rate of hemorrhagic complications. In conclusion, we believe that the early results of simultaneous resection of ascending aortic aneurysm and AVR utilizing the noncomposite technique described in this report are satisfactory and compare favorably with those of composite methods. Furthermore, long-term follow-up has shown that such noncomposite techniques are durable, impose minimal risk of late recurrent aneurysm formation, and are associated with a low risk of perioprosthetic leak, prosthetic dysfunction, coronary artery complications, and reoperation. Five and 10 year survival rates, moreover, have been satisfactory. In this perspective, we believe the composite technique of Bentall and Edwards has

Volume 79 Number 3

Resection of ascending aortic aneurysm and A VR

3 99

March,1980

relatively limited applicability and should be reserved for selected cases of advanced necrotizing prosthetic or natural endocarditis. When the composite technique is indicated, it would seem prudent to reimplant the coronary arteries (with or without interposition saphenous vein grafts) end to side as full-thickness Carrel patches, rather than side to side from within the aneurysmal sac. REFERENCES Cooley DA, De Bakey ME: Surgical considerations of intrathoracic aneurysms of the aorta and great vessels. Ann Surg 135:660-680. 1952 2 Bentall HH, DeBono A: A technique for complete replacement of the ascending aorta. Thorax 23:338-339, 1968 3 Singh MP, Bentall HH: Complete replacement of the ascending aorta and the aortic valve for the treatment of aortic aneurysm. J THoRAc CARDIOVASC SURG 63:218225. 1972 4 Edwards WS, Kerr AR: A safer technique for replacement of the entire ascending aorta and aortic valve. J THoRAc CARDIOVASC SURG 59:837-839, 1970 5 McCready RA, Pluth JR: Surgical treatment of ascending aortic aneurysms associated with aortic valve insufficiency. Ann Thorac Surg 28:307-316,1979 6 Mayer JE, Lindsay WG, Wang Y, Jorgensen CR, Nicoloff DM: Composite replacement of the aortic valve and ascending aorta. J THoRAc CARDIOVASC SURG 76: 816-823, 1978 7 Bentall HH: Discussion of Mayer et al" 8 Kouchoukos NT, Karp RB, Lell WA: Replacement of the ascending aorta and aortic valve with a composite graft. Results in 25 patients. Ann Thorac Surg 24:140-148, 1977 9 Hashimoto A, Kitamura N, Koyanagi H, Konno F: Surgical treatment of annulo-aortic ectasia. J Cardiovasc Surg (Torino) 17:240-247, 1976 10 Klopp E, Hauer J, Zinner M, Brawley R: Acute supravalvular stenosis following replacement of the aortic valve and ascending aorta in a patient with Marfan's syndrome. Report of a case. Surgery 84:292-294, 1978 I I Crosby IK, Ashcraft WC, Reed WA: Surgery of proximal aorta in Marfan's syndrome. J THORAC CARDIOVASC SURG 66:75-81, 1973 12 Campbell CD, Hardesty RL, Siewers RD, Larberg DB, Peel RL, Bahnson HT: Selected therapy for ascending aortic aneurysms. Arch Surg 113:1324-1330, 1978 13 Symbas TN, Raizner AE, Tyras DH, Hatcher CR Jr, Ingelesby TV, Baldwin DJ: Aneurysms of all sinuses of Valsalva in patients with Marfan's syndrome. Ann Surg 174:902-907, 1971 14 Koizumi S, Mohri H, Kagawa Y, Saji K, Haneda K, Kahata 0, Itoh T, Yokoyama A, Ohmi M, Horiuchi T: Surgical treatment of annulo-aortic ectasia. Experience in seven consecutive patients. Ann Thorac Surg 25:425430. 1978

15 Zubiate P, Kay JH: Surgical treatment of aneurysm of the ascending aorta with aortic insufficiency and marked displacement of the coronary ostia. J THoRAc CARDIOVASC SURG71:415-421,1976 16 Helseth HK, Haglin 11, Stenlund RR, Peterson CR: Evaluation of composite graft replacement of the aortic root and ascending aorta. Ann Thorac Surg 18:138-141, 1974 17 Miller DC, Stinson EB, Oyer PE, Rossiter SJ, Reitz BA. Shumway NE: Surgical implications and results of combined aortic valve replacement and myocardial revascularization. Am J Cardiol 43:494-50 I, 1979 18 Wheat MW Jr, Wilson JR. Bartley TD: Successful replacement of the entire ascending aorta and aortic valve. JAMA 188:717-719, 1964 19 Najafi H, Dye WS, Javid H, Hunter JA, Goldin MD. Serry C: Aortic insufficiency secondary to aortic root aneurysm or dissection. Arch Surg 110: 1401-1407, 1975 20 Ross DN, Frazier IT, Gonzalez-Lavin L: Surgery of Marfan's syndrome and related conditions of the aortic root (annulo-aortic ectasia). Thorax 27:52-57, 1972 21 Groves LK, Effler DB, Hawk WA, Gulati K: Aortic insufficiency secondary to aneurysmal changes in the ascending aorta. Surgical management. J THoRAc CARDIOVASC SURG 48:362-379, 1964 22 Bloodwell RD, Hallman GL, Cooley DA.: Aneurysm of the ascending aorta with aortic valvular insufficiency. Arch Surg 92:588-599, 1965 23 Davis L, Pluth JR, Giuliani ER: The Marfan syndrome and cardiac surgery. J THoRAc CARDIOVASC SURG 75:505-509, 1978 24 Nasrallah AT, Cooley DA, Goussous Y, Hallman GL, Lufschanowski R, Leachman RD: Surgical experience in patients with Marfan's syndrome, ascending aortic aneurysm and aortic regurgitation. Am J Cardiol 36:338-341, 1975 25 Ferlic RM, Goott B, Edwards JE, Liliehei CW: Aortic valvular insufficiency associated with cystic medial necrosis. Ann Surg 165:1-9, 1967 26 Wheat MW Jr, Boruchow B, Ramsey HW: Surgical treatment of aneurysms of the aortic root. Ann Thorac Surg 12:593-607, 1971 27 Liddicoat JE, Bekassy SM, Rubio PA, Noon GP, and De Bakey ME: Ascending aortic aneurysms. Review of 100 consecutive cases. Circulation 51,52:Suppl 1:202-209, 1975 28 Blank RH, Pupello DF, Connar RG, Bessone LN: Resection of the ascending aorta using profound topical hypothermia for myocardial protection. Ann Surg 181: 742-746, 1975 29 Weldon CS: Discussion of Mayer et al" 30 Anagnostopoulos CE: Discussion of Kouchoukos et al" 31 Cabrol C, Gandjbakhch I, Cham B: Anevrismes de I'aorte ascendante. Remplacement total avec reimplantation des arteres coronaires. Nouv Presse Med 7:363-365, 1978

The Journal of

400 Miller et al.

Discussion DR. DEY R. MAN HAS Seattle. Wash.

I must congratulate Dr. Miller for an excellent presentation. The Stanford group, as usual, has presented a large series of patients with a relatively uncommon but a challenging problem with low operative mortality rate and superb long-term follow-up. Annulo-aortic ectasia, a term first used by Ellis and Cooley to describe a pathological complex of ascending aortic aneurysms associated with aortic insufficiency, can be managed by two techniques at the present time. The first is a noncomposite technique as shown by Dr. Miller. The second is a composite technique first described by Bentall and DeBono in 1968. The main criticisms of the noncomposite technique have been excessive postoperative bleeding, leading to a high rate of re-exploration, and also recurrent aneurysmal dilatation of the segment of the aorta left between the graft and the valve. Both these criticisms have been refuted by the results reported by Dr. Miller. There was no instance of recurrent aneurysmal dilatation after a long-term follow-up of about 10 years in the authors' presentation. In my opinion, this is one of the most important messages of this paper. We have used noncomposite techniques only in those situations in which for technical reasons, it is not feasible to implant the coronary ostia. One such example is the patient in whom the coronary ostia are very close to the aortic anulus. In annulo-aortic ectasia the ostia are displaced several centimeters away from the anulus, so that it is easy to implant them with the graft. Another situation is acute dissection of the ascending aorta, for implantation of the coronary arteries may be difficult because of extreme friability of the tissues. [Slide] One example is the case of a patient with acute dissection of the ascending aorta, which leaked into the pericardium and caused a big clot to develop. On opening the aneurysm, we found the dissection extending around the right coronary artery. This coronary artery was partially tom, so that implanting it in the graft would have been difficult because of extreme friability of the tissue. We ligated the coronary artery close to the ostium, reinforced the aortic wall by putting Teflon felt in the false channel both proximally and distally, and oversewed the free margin. We then replaced the dissection with a woven Dacron graft and implanted a saphenous vein coronary artery bypass graft to the distal right coronary artery. [Slide] In patients with classical annulo-aortic ectasia, who have a big, pear-shaped aneurysm of the ascending aorta with aortic insufficiency, we prefer to use the composite technique. This composite technique is similar to that described by Bentall in 1968, except for a few modifications. At the present time there are composite grafts available commercially, but one can make a composite graft in the operating room by sewing an aortic valve of one's choice to a woven Dacron tube graft. The important point is that the

Thoracic and Cardiovascular Surgery

Dacron tube graft should be 3 to 5 mm bigger than the aortic valve. The aortic valve is sewn to the aortic anulus with horizontal mattress sutures, as Dr. Miller mentioned. In implanting the coronary artery, the surgeon should make a big hole in the graft directly opposite the coronary ostia after fully stretching the graft. The suture should be placed, not in the wall of the ostia, but in the surrounding aortic wall. After releasing the cross-clamp, we carefully inspect all the suture lines for any bleeding. If there is excessive bleeding from the coronary suture lines or from the posterior portion of the proximal or the distal aortic suture lines, we try to control it from inside the graft. We cross-clamp the aorta again and open the graft transversely. With the heart cooled by cardioplegia, we reinforce the leaking suture line from inside. Attempting to control the suture line from outside, in our opinion, not only is difficult from the standpoint of exposure but also risks disrupting the coronary anastomosis while the graft is being retracted away from the aortic wall in order to gain the exposure. After controlling the bleeding, we remove the cross-clamp and wrap the aneurysmal wall around the fully distended graft, so that the graft is not constricted. We approximate this aneurysmal wall loosely, so that if there is any bleeding, blood escapes into the pericardial cavity. Formation of a big hematoma between the graft and the aneurysmal wall, which can lead to valve dysfunction or further disruption of the coronary anastomosis, is thereby avoided. I would like to conclude by asking Dr. Miller a question. In view of the excellent long-term results with noncomposite techniques, why do you recommend the composite technique for annulo-aortic ectasia? DR. MILLE R (Closing) I would like to thank Dr. Manhas for his remarks. Focusing on the aortic dissections first, we individualize what has to be done to the coronary arteries. If they are involved and irreversibly compromised by the dissecting process, we usually use conventional coronary artery bypass grafting or, in certain cases, transect the proximal right coronary artery and perform an end-to-end anastomosis to a saphenous vein graft. Rarely, coronary ostial reimplantation (such as a Carrel patch) is necessary. One must individualize each of these cases of aortic dissections; this is equally true for patients with annulo-aortic ectasia. In patients with acute dissections (when concomitant aortic valve replacement is not necessary), I believe it is important not to leave a large amount of aorta beyond the remaining sinotubular ridge. It is in this situation that recurrent aneurysms of the aortic root have been reported: that is, supracoronary aortic graft interposition in which the sinuses of Yalsalva and a variable amount of proximal native aorta are left. This situation is not analogous to the technique just described, in which AYR is carried out simultaneously. I think the modified version of the composite technique

Volume 79 Number 3

Resection of ascending aortic aneurysm and AVR

40 1

March. 1980

employed by Dr. Manhas and Dr. Thomas in Seattle avoids two specific problem areas: namely, the technique of the coronary artery anastomosis and the method of securing hemostasis. Dr. Manhas stated explicitly that he uses full-thickness (or through-and-through) bites of the aorta in constructing the coronary anastomoses. If one is going to use a composite technique, I think this approach is wise in order to avoid the all too frequent postoperative complication of false aneurysm of the coronary "side-to-side" anastomosis. Dr. Manhas also mentioned that meticulous hemostasis is surgically accomplished prior to closing the aneurysmal sac around the graft. Minimizing attention given to this step has led to many problems with the composite technique, including perigraft hematoma, iatrogenic supravalvular aortic stenosis, coronary artery false aneurysms, and impingment of an extravascular hematoma upon the disc of a Bjork-Shiley valve causing massive aortic regurgitation. All of these problems have been reported to result in emergency early and late reoperations. To blindly close the aneurysmal sac around the graft and expect hemostasis is not judicious, in my opinion.

Pertaining to our indications for application of the composite technique, our current philosophy mandates use of such a technique only in cases of advanced necrotizing prosthetic or natural endocarditis. In such cases, extensive periannular abscess formation and necrosis preclude implanting a bioprosthesis in the region of the natural annulus. We oversew the native coronary ostia and anastomose a xenograft-valved conduit to the proximal (relatively normal) aorta. We then re-establish coronary flow by means of saphenous vein grafts. Fortunately, such severe cases of endocarditis are relatively rare. Additionally, there are cases of severe annulo-aortic ectasia in which the coronary ostium is so markedly displaced cephalad that our basic technique is not applicable. In such cases we recommend reanastomosing the left main and right coronary artery ostia end-to-side as fullthickness Carrel patches into the graft. A similar approach may be useful in some cases of chronic aortic dissection as well. In this context, I believe Dr. Manhas and I concur and there is no significant discord.