- Email: [email protected]
Surgical treatment of the ascending aorta
J THORAC
CARDIOVASC SURG
1989;98:675-82
Surgical treatment of the ascending aorta Fourteen years' experience with 83 patients Between December 1972 and December 1986, 83 patients with aneurysmal disease (n = 37) or disseclion (n = 46) imolving tbe ascending aorta underwent a variety of operations, including composite valve-graft repairs (n = 39~ separated replacements of the aortic valve and ascending aorta (n = 18~ resuspension and graft replacement of tbe ascending aorta (n = 9~ graft replacement of the ascending aorta only (n = 8), bomograft root replacement (n = 3), aortic valve replacement with aortorrbapby in = 3), aortorrbapby alone (n = 2), and use of a sutureless intraluminal prosthesis (n = 1). The inclusion method was used in nine patients. The bospital mortality rate was 10% for patients with annuloaortic ectasia, 21 % (70% confidence interval 13% to 30%) for acute dissection, and 18% (70% confidence interval 14 % to 22 %) for the entire group. Logistic regression analysis showed age and cumulative bypass time to be significant for bospital death. The estimated 5-year survival rates are 69.5% ± 7.2% and 67.0% ± 9.0% and IO-year estimates are 34.6% ± 10.6% and 61.4% ± 9.8% lor dissection and aneurysm, respectively. Patient survival was related to differing pathology and type ef operation, and log-rank testing sbowed no differences at the 5% level. Attrition (17 late deaths) was mostly due to left ventricular dysfunction, myocardial infarction, or aneurysmal disease in ungrafted aorta, Actuarial freedom from thromboembolism in patients with prosthetic valves is 92.0 % ± 4.0 % IlDd 83.5% ± 6.8% at 5 and 10 years. Freedom from aU late graft and cardiac complications is 72.5% ± 9.1 % and 48.8% ± 13.1 % at 5 and 10 years for aneurysmal disease and 79.1 % ± 7.3% IlDd 67.3% ± 9.9% at 5 and 10 years for dissection. Reoperation in nine patients was required for pseudoaneurysms (n = 3), other aortic aneurysms (n = 3), persistent aortic regurgitation (n = 1), and ebsolescent valve prosthesis (n = 2). Thus hospital mortality does not seem to be significantly related to the type of operation used for pathologic conditions of the ascending aorta unless cumulative bypass time exceeds about 2 bours. Many nonfatal late complications are associated with a prosthetic valve, but late death is due primarily to cardiac causes and residual disease in other parts of tbe aorta.
P.1. Raudkivi, FRACS,* J. D. Williams, MSc,a J. L. Monro, FRCS, and 1. K. Ross, MS, FRCS, Southampton, England
Operations for aneurysma'l dilatation of the aortic root and ascending aorta or dissections involving the ascending aorta include the composite valve-graft method, I re-
the Wessex Cardiothoracic Centre, Southampton General Hospital, and Medical Statistics and Computing, University of Southampton," Southampton, England.
FroIO
Received for publication Nov. 19, 1987. Accepted for publication March 13, 1989. Address for reprints: J. K. Ross, MS, FRCS, Wessex Cardiothoracic Centre, General Hospital, Southampton S09 4XY, England. 'Present address: Cardiothoracic Surgical Unit, Green Lane Hospital, Auckland, New Zealand. 12/1/13089
suspension of the aortic valve and Dacron graft replacement of the ascending aorta, and separated replacement of the valve and ascending aorta. Wheat, Wilson, and Bartley? reported a procedure in which the aneurysm was resected down to the anulus, the coronary ostia were left in situ, and the prosthesis was scalloped around them. This operation differs from the supracoronary method reported by Groves and colleagues.' In the inclusion method the native aorta is closed completely and tightly over the graft." The supracoronary method has been associated with a high incidence of recurrent aneurysms, 5 as well as pseudoaneurysms and other technical problems." This study evaluates the use of various operations for the ascending aorta in a heterogeneous group of patients who have had surgical treatment of aneurysmal disease or 675
676
The Journal of Thoracic and Cardiovascular Surgery
Raudkivi et al.
Table III. Hospital mortality related to disease
Table I. Patient characteristics Variable n
Age (yr) Median Range Male:female Marfan's disease
70~{
Aneurysm
Dissection 46
37
59.0 26-70 35:11 8
54.0 20-74 27:10 4
Table II. Operation Type of disease
Disease
II
liM
cillA1
Acute dissection Subacute disseetion Acute plus subacute dissection Chronie dissection Annuloaorticectasia Degenerative aneurysms Marfan's disease Syphilis Endocarditis
28
(,
5
]
33
7
21.4 20.0 21.2
]3 20
2 2 0
(,
12 4 3
I
2
(
CL )
]3-30 14-29
]5,4
]0.0 0 8.3 50.0 33
HM. Hospital mortality: CL confidence limits.
Aneurysmal disease
Type of operation
Dissection
Graft replacement of ascending aorta Resuspension and graft Composite valve-graft Separated replacements Other
7
9 15 12 3
0 24 6 6
Total
46
37
dissection. The four principal types of operation used have been graft replacement of the ascending aorta, valve resuspension and graft replacement, composite valve-graft replacements, and separated replacements.
Patients and methods Between June 1972 and September 1986, 83 patients had operations for dissection involving the ascending aorta (n = 46) or for aneurysmal disease of the ascending aorta and/or anulus (n = 37). All of these patients were operated on by three surgeons at the Wessex Cardiothoracic Center. Basic patient characteristics are shown in Table I. The median age for patients with Marfan's disease at the time of operation was 46.5 years. Indications for operation included DeBakey type I or type II dissections (with or without aortic regurgitation) 7, annuloaortic ectasia, degenerative aneurysmal disease, syphilis, and endocarditis with annular destruction and aneurysmal dilatation of the sinuses of Valsalva. Dissection involved approximately the full length of the aorta (type I) in 32 patients (70%) and was confined to the ascending aorta (type II) in 14 (30%). The site of the tear was in the proximal ascending aorta in 20 (43%), in the distal ascending aorta in II (24%), in the arch in three patients, was not identified in six patients, and was not recorded in six. Patients with dissection included two with rupture into the pulmonary artery, one with rupture into the right atrium, and one with an iatrogenic cause. Definitions. Annuloaortic ectasia was defined as dilatation of the sinuses of Valsalva, associated aortic incompetence, and cephalad displacement of the coronary ostia with varying degrees of dilatation of the ascending aorta.f Degenerative aneurysmal disease was defined as an aneurysm showing generalized features of atherosclerosis in a patient
without the stigmata of Marfan's disease or histopathologic evidence of aortitis or cystic medial necrosis. Patients with dissection have been considered in terms of the interval from the moment of dissection to the time of operation. Acute dissection was defined as presentation within 24 hours of acute chest pain. subacute dissection as presentation any time between 24 hours and 7 days. and chronic dissection as disease presumed to have been present for more than a week. Hospital mortality was defined as death within 30 days of operation. Standardized definitions were used for hospital mortality, thromboembolism, anticoagulant-related bleeding, and valve-related morbidity." Prosthetic valve failure in this study refers to the need for elective replacement of Braunwald-Cutter valves. Surgical methods. The commonest procedure (Ta ble II) was composite valve-graft replacement; a 27 mm Bjork-Shiley valve was attached to a 30 mm low-porosity woven Dacron graft and sutured to the anulus with continuous or interrupted pledgetsupported sutures. Coronary arteries were anastomosed to the graft and the distal anastomosis to the ascending aorta with either a single or double layer ofTefton felt. The inclusion method was used in four patients with dissection and live with aneurysmal disease. The graft was not routinely preclotted. Additional procedures in the composite valve-graft group included coronary grafting (n = 2), reimplantation of the innominate artery (n = I), mitral valve replacement (n = I). and mitral valve replacement with double coronary artery bypass grafting (n = I). Reimplantation of the coronary arteries with an intermediary Dacron graft 10 was done in one patient. Short interposition vein or prosthetic conduits between the graft and coronary ostia and establishment of a fistula between the right atrium and periprosthetic space were not used. The aortic valve was resuspended by approximating the layers of the ascending aorta and the valve commissures with a single or double felt sandwich. and the ascending aorta was replaced with a low-porosity graft without interfering with the coronary arteries. Separate aortic valve replacement was usually done with a Bjork-Shiley prosthesis (also three porcine. three Starr-Edwards, and three Braunwald-Cutter prostheses). and then a Dacron graft was placed in the ascending aorta above the coronary ostia. Coronary artery bypass grafting was done in one of these patients who had surgical therapy for dissection. In patients with aneurysms. other procedures included aortic valve replacement with aortorrhaphy (n = 3) and homograft aortic
Volume 98
Surgical treatment of ascending aorta
Number 5, Part 1 November 1989
677
Table IV. Logistic regression for hospital mortality Factur CBT (min)
Age (yr) Sex (female) Dissection (compared with aneurysm) Year (post-l 977) NYHA class (relative to class I) class II class III or IV Interval (acutecompared with chronic)' Etiology (annuloaortic ectasia compared with others)"
[J
Value
0.005 0.024 0.136 0.884 0.949 0.668
0.265 0.455
Relative risk
95% C1for relative risk (%)
1.02 1.1I 0.51 0.95 0.98
1.01, 1.04 1.01, 1.22 0.20, 1.31 0.45,2.00 0.45,2.11
1.22 1.28 1.83 0.17
0.41, 3.69 0.51,3.23 0.61,5.46 0.001, 3.88
CBT. Cumulative bypass time: CI. confidence interval. "Dissection only (n = 44). rAncurvsm only (n = 30).
root replacement (n = 3). In patients with dissection, other procedures included use of a sutureless intraluminal prosthesis (n = I) and aortorrhaphy (n = 2). Femoral artery cannulation and hypothermic heart-lung bypass were used in all patients. Coronary perfusion was used between 1972 and 1977, but after 1977 cardioplegic arrest was achievedwith St. Thomas' Hospital solution given at 30-minute intervals together with topical myocardial cooling with saline. The median cumulative bypass time was 139 (range 55 to 345) minutes for the dissection group (n = 44) and 106 (range 60 to 271) minutes for those with aneurysmal disease (n = 37). The median crossclamp time was 109 (range 40 to 190) minutes for thosewith dissection (n = 40) and 90 (range 60 to 210) minutes for those with aneurysms (n = 33). Follow-up. Information was obtained by telephone consultation with the family practitioner or referring physician. The median duration of follow-up for hospitalll survivors was 70 months for dissection and 59 months for aneurysmal disease. At theclosingmonth of the study (September 1986) there had been 17 late deaths, a total of 48 survivors, and three patients were lost to follow-up and excluded from statistical analysis. Statistical methods. Life tables for patient survival were estimated by the method of Kaplan and Meier, 11 and heterogeneity between groups of patients was assessed by the log-rank test. Estimated survival proportions 5 and 10 years postoperatively are reported ± I standard error of the estimates. Logistic regression analysis was used to identify factors associated specifically with hospital mortality, and Cox's proportional hazards technique'? was used to identify factors associated with late death for hospital survivors. An estimate of the relative risk (with 95% confidence interval [CI]) associated with each prognostic factor was also calculated. These analyses were performed by BMDP programs PI L, P2L, and PLR.13
Early results Mortality. The hospital mortality rate was 18% (70% CI 14%to 22%) for the entire group of83 patients (Table III), Among the incremental risk factors for hospital death (80 cases analyzed), etiology is not significant and the interval between the moment of dissection and the time of
operation does not appear to influence the early outcome of the procedure (there were 13 patients with chronic dissection, for a hospital mortality rate of 15.4% [95% CI 0% to 64%]). The significant factors at the 5% level are cumulative bypass time and age (Table IV). The estimated probability of death within 30 days exceeds 10% if the cumulative bypass time is more than 2 hours (Fig. I), The largest proportion of hospital deaths was due to low cardiac output (Table V). Deaths resulting from acute cardiac failure included preoperative myocardial infarction (n = I), intractable arrhythmias (n = 2), inability to be weaned from bypass (n = 1), and asystole (n = 1). Uncontrollable bleeding accounted for two of four deaths in the operating room. Morbidity. Neurologic events and arrhythmias formed the greatest proportion of hospital complications (Table VI). Late results Mortality. There were 17 late deaths (20%), with two deaths directly related to the use of a prosthetic valve (anticoagulant-related bleeding from a gastric ulcer in one and stroke in the other) (Table VII). One death was due to bleeding from a pseudoaneurysm discovered at electiveoperation for replacement of a Braunwald-Cutter valve. Additional deaths resulting from aneurysms in other parts of the aorta included rupture of the descending thoracic aorta (n = 3), thoracoabdominal aneurysm (n = I), and abdominal aneurysm (n = 1). Cardiac failure accounted for four deaths, and the cause of death was not ascertained in four patients. The significant factor associated with late death (at the 5% level) is cumulative bypass time (Table VIII). In particular, however, the type of operation used, irrespective of the type of disease, has not been shown to be a signif-
The Journal of Thoracic and Cardiovascular Surqery ,
6 7 8 Raudkivi et al.
!
Table VI. Hospital complications in survivors
10
s
~
(}8
.~
c~ -c ....c ~ ~
-'-<:l iil ....C
C4
Minor neurologic Stroke Reoperation (bleeding) Sternal dehiscence Deep venous thrombosis Renal failure Pneumonia Supraventricular tachycardia Heart block
(}6 0·4 02 0
Dissection In)
Event
I
Aneurysm In)
3
3
o
3 I I I I I 6
I
o I
o
O·
10 I
o
o cul1lUlative bypass time (min)
Fig. 1. Estimated probability of hospital death related to cumulative bypass time from logistic regression model including cumulative bypass time and age. Dotted lines enclose 70% confidence interval.
Table V. Causes of hospital death Mode
Hemorrhage Stroke Acute cardiac failure Mesenteric infarction Renal failure Respiratory failure
Aneurysm (n)
Dissection (n)
1*
2* 2
0
3t
5t
0
I
I I
0 0
Table VII. Late complications Event
Embolic transient ischemic attack Frank stroke Anticoagulant-related bleeding Pseudoaneurysm Other aortic aneurysm Aortic regurgitation Paravalvular leak Prosthetic valve failure Left ventricular failure Myocardial infarction Cancer
Aneurysmal disease In)
Dissection in]
2'
o
2* (1) 1* (I)
2* 1*
I 2 (2)
2 (1) 3 (2)
2t 1*
3 1* 1* (I)
1 (I)
4 (4) I (I)
o o 1
o
'Death on table (n = I). tDeath on table death (n = I) (preoperative myocardial infarction).
Figures in parentheses indicate late death. 'Patient has a prosthetic valve. tHomograft root replacement.
icant variable affecting the late outcome of surgical therapy in this series. Fig. 2 compares cumulative survival in patients with aneurysmal disease and in those who have had a dissection. Both groups include patients with Marfan's disease. The estimated 5-yearsurvival rates are 67.0% ± 9.0% for aneurysmal disease and 69.5% ± 7.2% for dissections. At 10 years the estimates are 61.4% ± 9.8% and 34.6% ± 10.6% for aneurysm and dissection, respectively. Log-rank comparison shows no significant (x 2 = 0.57, df = 1, p = 0.452) overall difference. Patient survival for differing pathology shows that the cumulative proportion alive at 5 years is 65.4% ± 8.2% and at 10 years 26.6% ± 13.8% for dissection in patients without Marfan's disease. For dissection in patients with Marfan's disease, the estimated survival rate at 5 years is 87.5% ± 11.7%. For patients with degenerative aneurysms, the 5-year survival rate is 55.6% ± 24.9%, and for
those with annuloaortic ectasia, 75.0% ± 11.0% (logrank test, x 2 = 1.16, df= 3, p = 0.764). Patient survival was analyzed with respect to type of operation. Five-year survival rates are as follows: separated replacements, 77.4% ± 10.0%; aortic graft only, 60.0% ± 18.7%; composite graft-valve replacement, 74.0% ± 7.6%; and resuspension and grafting, 50.0% ± 18.6% (log-rank test, Xl = 4.04, df = 3, p = 0.257). Patient survival was also analyzed for operative method and etiologic group. For patients with dissection, 5-year survival rates are 92.3% ± 7.4%, 80.8% ± 12.3%, and 50.0% ± 18.6% for composite valve-graft operations, separated valve replacement, and resuspension with aortic grafting, respectively. Patients who have had composite valve-graft operations for arteurysmal disease (i.e., degenerativeorannuloaorticectasia)havea63.8% ± 12.5% 5-year survival rate, and there is no attrition at 10 years.
Volume 98 Number 5, Part 1
Surgical treatment of ascending aorta
November 1989
1.0
L
C>
.;
's
.80
("OII"~"'''''''!o'!l§"""",~
~
(fl
c
o 1:
o a. o
~"'OO~&ns--1III!
=,
.60
c:
'"
.~
iii
679
• = Aneurysm
.40
0= Dissection
:; E
p = 0.45
:J
o
.20
• i• •
Number at risk at start of year Aneurysm 35 Dissection 45
0.0
I
-
---
-I
0
24 33
20 29
19 27
I
2
3
17 23 4
14 22 5
12 19
10 15
9 12
10
I
I
I
I
6
7
8
8
9
8
5 10
11
12
I
13
14
Years
Fi!:. 2. Survival curves for patients with aneurysmal disease or dissection involving the ascending aorta.
Table VIII. Proportional hazards model for late death Faclor
n Value
Relative risk
95% C/ for relative risk
CBT (min)
0.D3~
0.9~
age (yr ) Sex (female Dissection (compared with aneurysm)
0.529
1.02 0.57 2.23
0.96. 1.00 0.96. 1.0~ 0.15. I.~O 0.73.6.85
1.79 3.25 1.53 2.36
0.48.6.61 O.~~. 12.00 0.31.7.60 0.22, 25.42
"'iYHA class (compared with class I) Class II Class III or IV Interval (acute compared with chronic)' Etiology (annuloaonic ectasia compared with othcrs)f "Di-scction only (n = tAncur yvm only (n
0.30~
0.162 0.190
0.603 0.478
3~).
= 29).
These differences are not significant by log-rank test (XC = 5.84, df = 3, p = 0.120). Morbidity. Late complications are summarized in Table VII. All but one of the patients with a cerebrovascular accident had a prosthetic valve. Aortic regurgitation has been discovered in five patients, two of whom have had homograft root replacement. Of the remaining three, one patient (who had resuspension and grafting) has had aortic valve replacement, one patient had a redissection after aortic grafting and died later of cardiac failure without surgical treatment, and the third has mild aortic regurgitation and has not been offered surgical therapy. Aneurysmal recurrence related to the operation has been identified in three patients. One aneurysm was at the lower end of the graft and was discovered during an elec-
tive operation for an obsolescent Braunwald-Cutter valve. The patient did not survive the procedure because of complications related to bleeding. In another patient, the initial procedure was separated replacement. Nearly 9 years later, an aneurysm ruptured into pulmonary artery. Composite valve-graft replacement was successful, and the patient is free of complications 54 months postoperatively. Another patient was recognized to have had a proximal redissection early after replacement without a specific operation for the valve. Porcine aortic valve replacement was undertaken uneventfully 3 months later. Actuarial freedom from thromboembolism in patients with prosthetic valves is 92.0% ± 4.0% and 83.5% ± 6.8% at 5 and 10 years. When all late graft and cardiac complications are considered together, freedom
The Journal of Thoracic and Cardiovascular Surgery
6 8 0 Raudkivi et al.
1.0
Cl
c:
's
R1•••••• I••lilIIII08,.';:rg, "0"'0'0.""",,)
.80
•
.~
::::l
en
0
• 8:v
,
L.J--------------•
c:
o
~c-
•
.60
•1---------
o
c:
Cl>
.e C; "3
.40
E
::::l
• =Aneurysm o = Dissection p =0.52
U
.20
Number at risk at start of year Ar.eurysm 36 Dissection 45
22 27
19 24
18 21
I
2
15 18 i
4
13 17 I
5 Years
10 14 i
6
9 10 I
7
7 7 I
8
5 7 i
9
5 5 i
10
I
11
I
12
Fig. 3. Freedom from any graft or cardiovascular complication.
from any of these events is 72.5% ± 9.1% and 48.8% ± 13.1% at 5 and 10 years, respectively, among patients with aneurysmal disease and 79.1% ± 7.3% and 67.3% ± 9.9% at 5 and 10 years, respectively, among those with dissections (Fig. 3) (log-rank test, x2 = 0.41, df= I, P = 0.523). Reoperation has been required for residual problems in the root, ascending aorta, or more distal aorta in 11% (9/ 83), including four with Marfan's disease. BraunwaldCutter valve replacements prompted two reoperations, but remaining indications were persistent aortic regurgitation (n = 1), dissection (n = 3), and aneurysm (n = 3). Three patients have had an operation for descending thoracic, thoracoabdominal, and abdominal aortic aneurysm. Discussion
Early results. Recent experience shows that a hospital mortality rate of about 20% or even less than 10% may be accomplished.v 8 10 13-15 For dissection, results may reflect variable case selection for in earlier years only patients with serious complications of aortic regurgitation, tamponade, renal artery involvement, or peripheral arterial occlusion tended to be selected for surgical treatment. Selection has probably been less marked for the aneurysm group. Although 37% of patients with dissection and 24% with aneurysmal disease were considered to have New York Heart Association class III or IV disease, preoperative status has not appeared to be a significant incremental risk factor. Multivariate testing further reveals that neither the operative
method nor the year of operation significantly affects early results, and this view is shared by others." Age and cumulative bypass time emerge as significant predictors of early death. The hospital mortality rate was 29% for patients in the seventh decade of life who had dissection and 50% for patients over 70 years of age with aneurysmal disease. Long cumulative bypass times reflect additional coronary grafting and/or valve replacement, difficulty with bleeding, or difficulty with weaning from bypass because of left ventricular failure. Mean cumulative bypass times reported here (141 and 122 minutes) are similar to the range of 105 to 145 minutes reported elsewhere. 15-17 Our experience is similar to that reported by others, 15. 16 where low cardiac output predominated over hemorrhage as the leading cause of hospital death. It is interesting that pathology is not a predictor of hospital deaths. If we presume that operative difficulties are particularly related to the quality of the aortic wall, however, this may reflect the histopathologic similarities that occur in patients with dissection and aneurysmal
disease.?: 8 Operative considerations and pseudoaneurysms. For
aortic dissection we have adhered to the principle that the native valve should be preserved if the anulus is not dilated. For composite valve-graft operations we have seldom used the inclusion method." Complete transection of the aorta allows considerably improved access to both the upper and lower suture lines and facilitates the coronary anastomoses. With accurate anastomoses and improved visibility for correction of subsequent anastomotic
Volume 98 Number 5, Part 1 November 1989
leakage, we have not found hemorrhage to be the leading cause of either hospital death or morbidity. Recently, it has been suggested that the inclusion method is associated with pseudoaneurysms.", In nine patients so treated these have not been identified at late death or during follow-up, whichaverages 81 months. These patients, however, have not had aortography or computed tomography. With respect to separated replacements, the proximal suture line has been just above the level of the aortic commissures. We have not resected the sinuses of Valsalva, as described by Wheat, Wilson, and Bartley? and more recently reemphasized by Miller and associates'? and Borst." In contrast to those patients with composite valve-graft operations, two patients who had separated replacement (without inclusion) have required reoperation for pseudoaneurysms involving the proximal suture line. A further 16 patients had aortic grafting without valve replacement, and no pseudoaneurysms have been identified. We have not had significant technical problems with coronary anastomotic dehiscence and therefore have not used intermediary 8 mm Dacron grafts'? or short interpositionvein or prosthetic conduits between the graft and coronary orifice. Although the importance of preclotting has been emphasized most often" we do not preclot the graft, because we have found that bleeding through low-porosity grafts usually ceases rapidly and that major bleeding is usually related to the suture line. In this context, also we have not used biologic glues.'? Homograft root replacement'? has been reserved specifically for infective destruction of the aortic root. We now consider that other plastic procedures to the ascending aorta have no role in patients with either dissection or aneurysmal disease. Late results. Long-term results for surgically treated aneurysmal disease of the ascending aorta are sparse. The 5-year survival rates for annuloaortic ectasia were 80%6 and 77%.20 These figures compare with our 5- and IO-year survival rates of 7~% ± II % and 64.3% ± 12.4%, respectively (18 patients). For aneurysmal disease without subdivision for pathology, we found 5- and 10-year survival rates of 67% ± 9% and 61.4% ± 9.8%, respectively. We found late survival rates after surgical therapy for acute dissection in the ascending aorta to be 69.5% ± 7.2% at 5 years and 34.6% ± 10.6% at 10 years. These figures compare with 5-year survival rates of 51% ± 9% to 79% ± 4%15.20.21 and 10-yearsurvivalratesof49%to 53%.15.21 Thus, as reported elsewhere," late survival by nonparametric comparison is similar for annuloaortic ectasia and dissection. Multivariate testing does not
Surgical treatment of ascending aorta
681
reveal pathology to be an independent predictor of late death. Thirty percent of late complications were related to a prosthetic valve. Freedom from thromboembolic disease (in patients with prosthetic valves) is 83.5% ± 6.8% at 10 years, which is similar to other data for composite valve grafts" and for isolated Bjork-Shiley aortic valve replacement.22.23 Of all thromboembolic, neurologic, and hemorrhagic events, only one (a stroke) occurred in a patient with a native valve. Other prosthetic valve-related complications (anticoagulant-related hemorrhage and paravalvular leak) were rare, and prosthetic valve endocarditis was not diagnosed. This information is offset against the 5-year survival figures of 50% ± 18.6% for resuspension and grafting and 60% ± 18.2 % for aortic grafting alone, compared with 74.0% ± 7.6% for composite valve graft operations and 77% ± 10% for separated replacements. Although there are data showing that the late mortality rate for aortic valve replacements is not excessiveI 6. 17 (confirmed by multivariate analysis of our own data), there is an increased incidence of nonfatal complications, for which reason we recommend preservation of the native valve in patients without Marfan's disease or annular dilatation. Aortic regurgitation occurred in two of 16 patients whose native valve was preserved. The failure mode is benign and encourages us to persevere with resuspension. Much late attrition is due to left ventricular dysfunction, myocardial infarction, and aneurysmal disease in ungrafted aortic segments. Reaperation has been required for residual problems in the root or ascending aorta or more distal aorta in 11% (9/83). This group of nine patients includes four with Marfan's disease. We have identified two other patients with persistence of the false lumen in the descending thoracic aorta, one of whom died 3 years postoperatively of cardiac failure. The other has had an additional operation. This experience highlights the need for indefinite surveillance and regular assessment of the thoracic aorta by computed tomography in all patients requiring surgical treatment of the ascending aorta. REFERENCES 1. Bentall H, De Bono A. A technique for complete replacement of the ascending aorta. Thorax 1967;23:338-9. 2. Wheat MW Jr, Wilson JR, Bartley TO. Successful replacement of the entire ascending aorta and aortic valve. JAMA 1964;188:717-9. 3. Groves LK, Effler DB, Hawk W A, Gulati K. Aortic insufficiency secondary to aneurysmal changes in the ascending aorta: surgical management. J THORAC CARDIOVASC SURG 1964;48:362-79. 4. Edwards WS, Kerr AR. A safer technique for replacement
The Journal of Thoracic and Cardiovascular Surgery
6 8 2 Raudkivi et al.
5.
6.
7.
8.
9.
10.
II.
12. 13. 14.
of the entire ascending aorta and aortic valve. J THORAC CARDIOVASC SURG 1970;59:837-9. McCready RA, Pluth JR. Surgical treatment of ascending aortic aneurysms associated with aortic valve insufficiency. Ann Thorac Surg 1979;28:307-16. Kouchoukos NT, Marshall WG, Wedige-Stechner TA. Eleven-year experience with composite graft replacement of the ascending aorta and aortic valve. J THORAC CAROlOVASC SURG 1986;92:691-705. DeBakey ME, McCollum CH, Crawford ES, et al. Dissection and dissecting aneurysms of the aorta: twenty-year follow-up of five hundred twenty-seven patients treated surgically. Surgery 1982;92: 1118-34. Savunen T. Annulo-aortic ectasia: a clinical, structural, and biochemical study. Scand J Thorac Cardiovasc Surg 1986;suppl 37:7-45. Edmunds LH, Clark RE, Cohn LH, Miller DC, Weisel RD. Guidelines for reporting morbidity and mortality after cardiac valvular operations. Ann Thorac Surg 1988;46: 257-9. Cabral C. Pavie A, Mersinildrey P, et al. Long-term results with total replacement of the aortic valve and aortic root with implantation of the coronary arteries. J THORAC CARDIOVASC SURG 1986;91:17-25. Kaplan EL, Meier P. Non-parametric estimation from incomplete observations. Am Stat Assoc J 1958;53:457-81. Cox DR. Regression models and life-tables. J R Stat Soc (B) 1972;34:187-200. Dixon W J. BMDP statistical software 1985; University of California Press. Gott VL, Pyertiz RE, Magovern GJ, Cameron DE, MeKusick VA. Surgical treatment of aneurysms of the ascending aorta in the Marfan syndrome: results of compos-
15.
16.
17.
18. 19.
20.
21.
22.
23.
ite-graft repair in 50 patients. N EnglJ Med 1986;314: 10704. Cachera J-P, Vouhe PR, Loisance DY, et al. Surgical management of acute dissections involving the ascending aorta: early and late results with 38 patients. J THORAC CARDIOVASC SURG 1981;82:576-84. Jex RK, SchaffHV, Piehler JM, et al. Repair of ascending aortic dissection: influence of associated aortic valve insufficiency on early and late results. J THORAC CARDIOVASC SURG1987;93:375-84. . Miller DC, Stinson EB, Oyer PE, et al. Operative treatment of aortic dissections: experience with 125 patients over a 16 year period. J THORAC SURG 1979;78:365-82. Borst HG. Replacement of the ascending aorta and aortic valve. Ann Thorac Surg 1981;32:613-4. Bachet J, Gigou F, Laurian C, Bical 0, Goudot B, Guilmer D. Four-year clinical experience with the gelatin-resorcineformol biologic glue in acute aortic dissection. J THORAC CARDIOVASC SURG 1982;83:212-7. Donaldson RM, Ross ON. Replacement of the ascending aorta and aortic valve with a composite graft: results and follow-up in 71 patients. Thorac Cardiovasc Surg 1981; 29:195-9. Kirklin JW, Barratt-Boyes BG. Cardiac surgery: morphology, diagnostic criteria, natural history, techniques, results and indications. 1st ed. New York: John Wiley, 1986:1488. Karp RB, Cyrus RJ, Blackstone EH, Kirklin JW, Kouchoukos NT, Pacifico AD. The Bjork-Shiley valve. J THORAC CARDIOVASC SURG 1981;81:602-14. Cohn LH, Allred EN, DiSesa VJ, Sawtelle K, Shemin RJ. Collins JJ. Early and late risk or aortic valve replacement. J THORAC CARDIOVASC SURG 1984;88:695-705.
Bound volumes available to subscribers Bound volumes of THE JOURNAL OFTHORACIC AND CARDIOVASCULAR SURGERY are available to subscribers (only) for the 1989 issues from the Publisher, at a cost of $55.00 ($72.00 international) for Vol. 97 (January-June) and Vol. 98 (July-December). Shipping charges are included. Each bound volume contains a subject and author index and all advertising is removed. Copies are shipped within 60 days after publication of the last issue of the volume. The binding is durable buckram with the JOURNAL name. volume number. and year stamped in gold on the spine. Payment must accompany all orders. Contact The C.V. Mosby Company. Circulation Department. 11830 Westline Industrial Drive. St. Louis. Missouri 63146-3318. USA; phone (800) 325-4177. ext. 351. Subscriptions must be in force to qualify. Boond volumes are not available in place of a regular JOURNAL subscription.
CARDIOVASC SURG
1989;98:675-82
Surgical treatment of the ascending aorta Fourteen years' experience with 83 patients Between December 1972 and December 1986, 83 patients with aneurysmal disease (n = 37) or disseclion (n = 46) imolving tbe ascending aorta underwent a variety of operations, including composite valve-graft repairs (n = 39~ separated replacements of the aortic valve and ascending aorta (n = 18~ resuspension and graft replacement of tbe ascending aorta (n = 9~ graft replacement of the ascending aorta only (n = 8), bomograft root replacement (n = 3), aortic valve replacement with aortorrbapby in = 3), aortorrbapby alone (n = 2), and use of a sutureless intraluminal prosthesis (n = 1). The inclusion method was used in nine patients. The bospital mortality rate was 10% for patients with annuloaortic ectasia, 21 % (70% confidence interval 13% to 30%) for acute dissection, and 18% (70% confidence interval 14 % to 22 %) for the entire group. Logistic regression analysis showed age and cumulative bypass time to be significant for bospital death. The estimated 5-year survival rates are 69.5% ± 7.2% and 67.0% ± 9.0% and IO-year estimates are 34.6% ± 10.6% and 61.4% ± 9.8% lor dissection and aneurysm, respectively. Patient survival was related to differing pathology and type ef operation, and log-rank testing sbowed no differences at the 5% level. Attrition (17 late deaths) was mostly due to left ventricular dysfunction, myocardial infarction, or aneurysmal disease in ungrafted aorta, Actuarial freedom from thromboembolism in patients with prosthetic valves is 92.0 % ± 4.0 % IlDd 83.5% ± 6.8% at 5 and 10 years. Freedom from aU late graft and cardiac complications is 72.5% ± 9.1 % and 48.8% ± 13.1 % at 5 and 10 years for aneurysmal disease and 79.1 % ± 7.3% IlDd 67.3% ± 9.9% at 5 and 10 years for dissection. Reoperation in nine patients was required for pseudoaneurysms (n = 3), other aortic aneurysms (n = 3), persistent aortic regurgitation (n = 1), and ebsolescent valve prosthesis (n = 2). Thus hospital mortality does not seem to be significantly related to the type of operation used for pathologic conditions of the ascending aorta unless cumulative bypass time exceeds about 2 bours. Many nonfatal late complications are associated with a prosthetic valve, but late death is due primarily to cardiac causes and residual disease in other parts of tbe aorta.
P.1. Raudkivi, FRACS,* J. D. Williams, MSc,a J. L. Monro, FRCS, and 1. K. Ross, MS, FRCS, Southampton, England
Operations for aneurysma'l dilatation of the aortic root and ascending aorta or dissections involving the ascending aorta include the composite valve-graft method, I re-
the Wessex Cardiothoracic Centre, Southampton General Hospital, and Medical Statistics and Computing, University of Southampton," Southampton, England.
FroIO
Received for publication Nov. 19, 1987. Accepted for publication March 13, 1989. Address for reprints: J. K. Ross, MS, FRCS, Wessex Cardiothoracic Centre, General Hospital, Southampton S09 4XY, England. 'Present address: Cardiothoracic Surgical Unit, Green Lane Hospital, Auckland, New Zealand. 12/1/13089
suspension of the aortic valve and Dacron graft replacement of the ascending aorta, and separated replacement of the valve and ascending aorta. Wheat, Wilson, and Bartley? reported a procedure in which the aneurysm was resected down to the anulus, the coronary ostia were left in situ, and the prosthesis was scalloped around them. This operation differs from the supracoronary method reported by Groves and colleagues.' In the inclusion method the native aorta is closed completely and tightly over the graft." The supracoronary method has been associated with a high incidence of recurrent aneurysms, 5 as well as pseudoaneurysms and other technical problems." This study evaluates the use of various operations for the ascending aorta in a heterogeneous group of patients who have had surgical treatment of aneurysmal disease or 675
676
The Journal of Thoracic and Cardiovascular Surgery
Raudkivi et al.
Table III. Hospital mortality related to disease
Table I. Patient characteristics Variable n
Age (yr) Median Range Male:female Marfan's disease
70~{
Aneurysm
Dissection 46
37
59.0 26-70 35:11 8
54.0 20-74 27:10 4
Table II. Operation Type of disease
Disease
II
liM
cillA1
Acute dissection Subacute disseetion Acute plus subacute dissection Chronie dissection Annuloaorticectasia Degenerative aneurysms Marfan's disease Syphilis Endocarditis
28
(,
5
]
33
7
21.4 20.0 21.2
]3 20
2 2 0
(,
12 4 3
I
2
(
CL )
]3-30 14-29
]5,4
]0.0 0 8.3 50.0 33
HM. Hospital mortality: CL confidence limits.
Aneurysmal disease
Type of operation
Dissection
Graft replacement of ascending aorta Resuspension and graft Composite valve-graft Separated replacements Other
7
9 15 12 3
0 24 6 6
Total
46
37
dissection. The four principal types of operation used have been graft replacement of the ascending aorta, valve resuspension and graft replacement, composite valve-graft replacements, and separated replacements.
Patients and methods Between June 1972 and September 1986, 83 patients had operations for dissection involving the ascending aorta (n = 46) or for aneurysmal disease of the ascending aorta and/or anulus (n = 37). All of these patients were operated on by three surgeons at the Wessex Cardiothoracic Center. Basic patient characteristics are shown in Table I. The median age for patients with Marfan's disease at the time of operation was 46.5 years. Indications for operation included DeBakey type I or type II dissections (with or without aortic regurgitation) 7, annuloaortic ectasia, degenerative aneurysmal disease, syphilis, and endocarditis with annular destruction and aneurysmal dilatation of the sinuses of Valsalva. Dissection involved approximately the full length of the aorta (type I) in 32 patients (70%) and was confined to the ascending aorta (type II) in 14 (30%). The site of the tear was in the proximal ascending aorta in 20 (43%), in the distal ascending aorta in II (24%), in the arch in three patients, was not identified in six patients, and was not recorded in six. Patients with dissection included two with rupture into the pulmonary artery, one with rupture into the right atrium, and one with an iatrogenic cause. Definitions. Annuloaortic ectasia was defined as dilatation of the sinuses of Valsalva, associated aortic incompetence, and cephalad displacement of the coronary ostia with varying degrees of dilatation of the ascending aorta.f Degenerative aneurysmal disease was defined as an aneurysm showing generalized features of atherosclerosis in a patient
without the stigmata of Marfan's disease or histopathologic evidence of aortitis or cystic medial necrosis. Patients with dissection have been considered in terms of the interval from the moment of dissection to the time of operation. Acute dissection was defined as presentation within 24 hours of acute chest pain. subacute dissection as presentation any time between 24 hours and 7 days. and chronic dissection as disease presumed to have been present for more than a week. Hospital mortality was defined as death within 30 days of operation. Standardized definitions were used for hospital mortality, thromboembolism, anticoagulant-related bleeding, and valve-related morbidity." Prosthetic valve failure in this study refers to the need for elective replacement of Braunwald-Cutter valves. Surgical methods. The commonest procedure (Ta ble II) was composite valve-graft replacement; a 27 mm Bjork-Shiley valve was attached to a 30 mm low-porosity woven Dacron graft and sutured to the anulus with continuous or interrupted pledgetsupported sutures. Coronary arteries were anastomosed to the graft and the distal anastomosis to the ascending aorta with either a single or double layer ofTefton felt. The inclusion method was used in four patients with dissection and live with aneurysmal disease. The graft was not routinely preclotted. Additional procedures in the composite valve-graft group included coronary grafting (n = 2), reimplantation of the innominate artery (n = I), mitral valve replacement (n = I). and mitral valve replacement with double coronary artery bypass grafting (n = I). Reimplantation of the coronary arteries with an intermediary Dacron graft 10 was done in one patient. Short interposition vein or prosthetic conduits between the graft and coronary ostia and establishment of a fistula between the right atrium and periprosthetic space were not used. The aortic valve was resuspended by approximating the layers of the ascending aorta and the valve commissures with a single or double felt sandwich. and the ascending aorta was replaced with a low-porosity graft without interfering with the coronary arteries. Separate aortic valve replacement was usually done with a Bjork-Shiley prosthesis (also three porcine. three Starr-Edwards, and three Braunwald-Cutter prostheses). and then a Dacron graft was placed in the ascending aorta above the coronary ostia. Coronary artery bypass grafting was done in one of these patients who had surgical therapy for dissection. In patients with aneurysms. other procedures included aortic valve replacement with aortorrhaphy (n = 3) and homograft aortic
Volume 98
Surgical treatment of ascending aorta
Number 5, Part 1 November 1989
677
Table IV. Logistic regression for hospital mortality Factur CBT (min)
Age (yr) Sex (female) Dissection (compared with aneurysm) Year (post-l 977) NYHA class (relative to class I) class II class III or IV Interval (acutecompared with chronic)' Etiology (annuloaortic ectasia compared with others)"
[J
Value
0.005 0.024 0.136 0.884 0.949 0.668
0.265 0.455
Relative risk
95% C1for relative risk (%)
1.02 1.1I 0.51 0.95 0.98
1.01, 1.04 1.01, 1.22 0.20, 1.31 0.45,2.00 0.45,2.11
1.22 1.28 1.83 0.17
0.41, 3.69 0.51,3.23 0.61,5.46 0.001, 3.88
CBT. Cumulative bypass time: CI. confidence interval. "Dissection only (n = 44). rAncurvsm only (n = 30).
root replacement (n = 3). In patients with dissection, other procedures included use of a sutureless intraluminal prosthesis (n = I) and aortorrhaphy (n = 2). Femoral artery cannulation and hypothermic heart-lung bypass were used in all patients. Coronary perfusion was used between 1972 and 1977, but after 1977 cardioplegic arrest was achievedwith St. Thomas' Hospital solution given at 30-minute intervals together with topical myocardial cooling with saline. The median cumulative bypass time was 139 (range 55 to 345) minutes for the dissection group (n = 44) and 106 (range 60 to 271) minutes for those with aneurysmal disease (n = 37). The median crossclamp time was 109 (range 40 to 190) minutes for thosewith dissection (n = 40) and 90 (range 60 to 210) minutes for those with aneurysms (n = 33). Follow-up. Information was obtained by telephone consultation with the family practitioner or referring physician. The median duration of follow-up for hospitalll survivors was 70 months for dissection and 59 months for aneurysmal disease. At theclosingmonth of the study (September 1986) there had been 17 late deaths, a total of 48 survivors, and three patients were lost to follow-up and excluded from statistical analysis. Statistical methods. Life tables for patient survival were estimated by the method of Kaplan and Meier, 11 and heterogeneity between groups of patients was assessed by the log-rank test. Estimated survival proportions 5 and 10 years postoperatively are reported ± I standard error of the estimates. Logistic regression analysis was used to identify factors associated specifically with hospital mortality, and Cox's proportional hazards technique'? was used to identify factors associated with late death for hospital survivors. An estimate of the relative risk (with 95% confidence interval [CI]) associated with each prognostic factor was also calculated. These analyses were performed by BMDP programs PI L, P2L, and PLR.13
Early results Mortality. The hospital mortality rate was 18% (70% CI 14%to 22%) for the entire group of83 patients (Table III), Among the incremental risk factors for hospital death (80 cases analyzed), etiology is not significant and the interval between the moment of dissection and the time of
operation does not appear to influence the early outcome of the procedure (there were 13 patients with chronic dissection, for a hospital mortality rate of 15.4% [95% CI 0% to 64%]). The significant factors at the 5% level are cumulative bypass time and age (Table IV). The estimated probability of death within 30 days exceeds 10% if the cumulative bypass time is more than 2 hours (Fig. I), The largest proportion of hospital deaths was due to low cardiac output (Table V). Deaths resulting from acute cardiac failure included preoperative myocardial infarction (n = I), intractable arrhythmias (n = 2), inability to be weaned from bypass (n = 1), and asystole (n = 1). Uncontrollable bleeding accounted for two of four deaths in the operating room. Morbidity. Neurologic events and arrhythmias formed the greatest proportion of hospital complications (Table VI). Late results Mortality. There were 17 late deaths (20%), with two deaths directly related to the use of a prosthetic valve (anticoagulant-related bleeding from a gastric ulcer in one and stroke in the other) (Table VII). One death was due to bleeding from a pseudoaneurysm discovered at electiveoperation for replacement of a Braunwald-Cutter valve. Additional deaths resulting from aneurysms in other parts of the aorta included rupture of the descending thoracic aorta (n = 3), thoracoabdominal aneurysm (n = I), and abdominal aneurysm (n = 1). Cardiac failure accounted for four deaths, and the cause of death was not ascertained in four patients. The significant factor associated with late death (at the 5% level) is cumulative bypass time (Table VIII). In particular, however, the type of operation used, irrespective of the type of disease, has not been shown to be a signif-
The Journal of Thoracic and Cardiovascular Surqery ,
6 7 8 Raudkivi et al.
!
Table VI. Hospital complications in survivors
10
s
~
(}8
.~
c~ -c ....c ~ ~
-'-<:l iil ....C
C4
Minor neurologic Stroke Reoperation (bleeding) Sternal dehiscence Deep venous thrombosis Renal failure Pneumonia Supraventricular tachycardia Heart block
(}6 0·4 02 0
Dissection In)
Event
I
Aneurysm In)
3
3
o
3 I I I I I 6
I
o I
o
O·
10 I
o
o cul1lUlative bypass time (min)
Fig. 1. Estimated probability of hospital death related to cumulative bypass time from logistic regression model including cumulative bypass time and age. Dotted lines enclose 70% confidence interval.
Table V. Causes of hospital death Mode
Hemorrhage Stroke Acute cardiac failure Mesenteric infarction Renal failure Respiratory failure
Aneurysm (n)
Dissection (n)
1*
2* 2
0
3t
5t
0
I
I I
0 0
Table VII. Late complications Event
Embolic transient ischemic attack Frank stroke Anticoagulant-related bleeding Pseudoaneurysm Other aortic aneurysm Aortic regurgitation Paravalvular leak Prosthetic valve failure Left ventricular failure Myocardial infarction Cancer
Aneurysmal disease In)
Dissection in]
2'
o
2* (1) 1* (I)
2* 1*
I 2 (2)
2 (1) 3 (2)
2t 1*
3 1* 1* (I)
1 (I)
4 (4) I (I)
o o 1
o
'Death on table (n = I). tDeath on table death (n = I) (preoperative myocardial infarction).
Figures in parentheses indicate late death. 'Patient has a prosthetic valve. tHomograft root replacement.
icant variable affecting the late outcome of surgical therapy in this series. Fig. 2 compares cumulative survival in patients with aneurysmal disease and in those who have had a dissection. Both groups include patients with Marfan's disease. The estimated 5-yearsurvival rates are 67.0% ± 9.0% for aneurysmal disease and 69.5% ± 7.2% for dissections. At 10 years the estimates are 61.4% ± 9.8% and 34.6% ± 10.6% for aneurysm and dissection, respectively. Log-rank comparison shows no significant (x 2 = 0.57, df = 1, p = 0.452) overall difference. Patient survival for differing pathology shows that the cumulative proportion alive at 5 years is 65.4% ± 8.2% and at 10 years 26.6% ± 13.8% for dissection in patients without Marfan's disease. For dissection in patients with Marfan's disease, the estimated survival rate at 5 years is 87.5% ± 11.7%. For patients with degenerative aneurysms, the 5-year survival rate is 55.6% ± 24.9%, and for
those with annuloaortic ectasia, 75.0% ± 11.0% (logrank test, x 2 = 1.16, df= 3, p = 0.764). Patient survival was analyzed with respect to type of operation. Five-year survival rates are as follows: separated replacements, 77.4% ± 10.0%; aortic graft only, 60.0% ± 18.7%; composite graft-valve replacement, 74.0% ± 7.6%; and resuspension and grafting, 50.0% ± 18.6% (log-rank test, Xl = 4.04, df = 3, p = 0.257). Patient survival was also analyzed for operative method and etiologic group. For patients with dissection, 5-year survival rates are 92.3% ± 7.4%, 80.8% ± 12.3%, and 50.0% ± 18.6% for composite valve-graft operations, separated valve replacement, and resuspension with aortic grafting, respectively. Patients who have had composite valve-graft operations for arteurysmal disease (i.e., degenerativeorannuloaorticectasia)havea63.8% ± 12.5% 5-year survival rate, and there is no attrition at 10 years.
Volume 98 Number 5, Part 1
Surgical treatment of ascending aorta
November 1989
1.0
L
C>
.;
's
.80
("OII"~"'''''''!o'!l§"""",~
~
(fl
c
o 1:
o a. o
~"'OO~&ns--1III!
=,
.60
c:
'"
.~
iii
679
• = Aneurysm
.40
0= Dissection
:; E
p = 0.45
:J
o
.20
• i• •
Number at risk at start of year Aneurysm 35 Dissection 45
0.0
I
-
---
-I
0
24 33
20 29
19 27
I
2
3
17 23 4
14 22 5
12 19
10 15
9 12
10
I
I
I
I
6
7
8
8
9
8
5 10
11
12
I
13
14
Years
Fi!:. 2. Survival curves for patients with aneurysmal disease or dissection involving the ascending aorta.
Table VIII. Proportional hazards model for late death Faclor
n Value
Relative risk
95% C/ for relative risk
CBT (min)
0.D3~
0.9~
age (yr ) Sex (female Dissection (compared with aneurysm)
0.529
1.02 0.57 2.23
0.96. 1.00 0.96. 1.0~ 0.15. I.~O 0.73.6.85
1.79 3.25 1.53 2.36
0.48.6.61 O.~~. 12.00 0.31.7.60 0.22, 25.42
"'iYHA class (compared with class I) Class II Class III or IV Interval (acute compared with chronic)' Etiology (annuloaonic ectasia compared with othcrs)f "Di-scction only (n = tAncur yvm only (n
0.30~
0.162 0.190
0.603 0.478
3~).
= 29).
These differences are not significant by log-rank test (XC = 5.84, df = 3, p = 0.120). Morbidity. Late complications are summarized in Table VII. All but one of the patients with a cerebrovascular accident had a prosthetic valve. Aortic regurgitation has been discovered in five patients, two of whom have had homograft root replacement. Of the remaining three, one patient (who had resuspension and grafting) has had aortic valve replacement, one patient had a redissection after aortic grafting and died later of cardiac failure without surgical treatment, and the third has mild aortic regurgitation and has not been offered surgical therapy. Aneurysmal recurrence related to the operation has been identified in three patients. One aneurysm was at the lower end of the graft and was discovered during an elec-
tive operation for an obsolescent Braunwald-Cutter valve. The patient did not survive the procedure because of complications related to bleeding. In another patient, the initial procedure was separated replacement. Nearly 9 years later, an aneurysm ruptured into pulmonary artery. Composite valve-graft replacement was successful, and the patient is free of complications 54 months postoperatively. Another patient was recognized to have had a proximal redissection early after replacement without a specific operation for the valve. Porcine aortic valve replacement was undertaken uneventfully 3 months later. Actuarial freedom from thromboembolism in patients with prosthetic valves is 92.0% ± 4.0% and 83.5% ± 6.8% at 5 and 10 years. When all late graft and cardiac complications are considered together, freedom
The Journal of Thoracic and Cardiovascular Surgery
6 8 0 Raudkivi et al.
1.0
Cl
c:
's
R1•••••• I••lilIIII08,.';:rg, "0"'0'0.""",,)
.80
•
.~
::::l
en
0
• 8:v
,
L.J--------------•
c:
o
~c-
•
.60
•1---------
o
c:
Cl>
.e C; "3
.40
E
::::l
• =Aneurysm o = Dissection p =0.52
U
.20
Number at risk at start of year Ar.eurysm 36 Dissection 45
22 27
19 24
18 21
I
2
15 18 i
4
13 17 I
5 Years
10 14 i
6
9 10 I
7
7 7 I
8
5 7 i
9
5 5 i
10
I
11
I
12
Fig. 3. Freedom from any graft or cardiovascular complication.
from any of these events is 72.5% ± 9.1% and 48.8% ± 13.1% at 5 and 10 years, respectively, among patients with aneurysmal disease and 79.1% ± 7.3% and 67.3% ± 9.9% at 5 and 10 years, respectively, among those with dissections (Fig. 3) (log-rank test, x2 = 0.41, df= I, P = 0.523). Reoperation has been required for residual problems in the root, ascending aorta, or more distal aorta in 11% (9/ 83), including four with Marfan's disease. BraunwaldCutter valve replacements prompted two reoperations, but remaining indications were persistent aortic regurgitation (n = 1), dissection (n = 3), and aneurysm (n = 3). Three patients have had an operation for descending thoracic, thoracoabdominal, and abdominal aortic aneurysm. Discussion
Early results. Recent experience shows that a hospital mortality rate of about 20% or even less than 10% may be accomplished.v 8 10 13-15 For dissection, results may reflect variable case selection for in earlier years only patients with serious complications of aortic regurgitation, tamponade, renal artery involvement, or peripheral arterial occlusion tended to be selected for surgical treatment. Selection has probably been less marked for the aneurysm group. Although 37% of patients with dissection and 24% with aneurysmal disease were considered to have New York Heart Association class III or IV disease, preoperative status has not appeared to be a significant incremental risk factor. Multivariate testing further reveals that neither the operative
method nor the year of operation significantly affects early results, and this view is shared by others." Age and cumulative bypass time emerge as significant predictors of early death. The hospital mortality rate was 29% for patients in the seventh decade of life who had dissection and 50% for patients over 70 years of age with aneurysmal disease. Long cumulative bypass times reflect additional coronary grafting and/or valve replacement, difficulty with bleeding, or difficulty with weaning from bypass because of left ventricular failure. Mean cumulative bypass times reported here (141 and 122 minutes) are similar to the range of 105 to 145 minutes reported elsewhere. 15-17 Our experience is similar to that reported by others, 15. 16 where low cardiac output predominated over hemorrhage as the leading cause of hospital death. It is interesting that pathology is not a predictor of hospital deaths. If we presume that operative difficulties are particularly related to the quality of the aortic wall, however, this may reflect the histopathologic similarities that occur in patients with dissection and aneurysmal
disease.?: 8 Operative considerations and pseudoaneurysms. For
aortic dissection we have adhered to the principle that the native valve should be preserved if the anulus is not dilated. For composite valve-graft operations we have seldom used the inclusion method." Complete transection of the aorta allows considerably improved access to both the upper and lower suture lines and facilitates the coronary anastomoses. With accurate anastomoses and improved visibility for correction of subsequent anastomotic
Volume 98 Number 5, Part 1 November 1989
leakage, we have not found hemorrhage to be the leading cause of either hospital death or morbidity. Recently, it has been suggested that the inclusion method is associated with pseudoaneurysms.", In nine patients so treated these have not been identified at late death or during follow-up, whichaverages 81 months. These patients, however, have not had aortography or computed tomography. With respect to separated replacements, the proximal suture line has been just above the level of the aortic commissures. We have not resected the sinuses of Valsalva, as described by Wheat, Wilson, and Bartley? and more recently reemphasized by Miller and associates'? and Borst." In contrast to those patients with composite valve-graft operations, two patients who had separated replacement (without inclusion) have required reoperation for pseudoaneurysms involving the proximal suture line. A further 16 patients had aortic grafting without valve replacement, and no pseudoaneurysms have been identified. We have not had significant technical problems with coronary anastomotic dehiscence and therefore have not used intermediary 8 mm Dacron grafts'? or short interpositionvein or prosthetic conduits between the graft and coronary orifice. Although the importance of preclotting has been emphasized most often" we do not preclot the graft, because we have found that bleeding through low-porosity grafts usually ceases rapidly and that major bleeding is usually related to the suture line. In this context, also we have not used biologic glues.'? Homograft root replacement'? has been reserved specifically for infective destruction of the aortic root. We now consider that other plastic procedures to the ascending aorta have no role in patients with either dissection or aneurysmal disease. Late results. Long-term results for surgically treated aneurysmal disease of the ascending aorta are sparse. The 5-year survival rates for annuloaortic ectasia were 80%6 and 77%.20 These figures compare with our 5- and IO-year survival rates of 7~% ± II % and 64.3% ± 12.4%, respectively (18 patients). For aneurysmal disease without subdivision for pathology, we found 5- and 10-year survival rates of 67% ± 9% and 61.4% ± 9.8%, respectively. We found late survival rates after surgical therapy for acute dissection in the ascending aorta to be 69.5% ± 7.2% at 5 years and 34.6% ± 10.6% at 10 years. These figures compare with 5-year survival rates of 51% ± 9% to 79% ± 4%15.20.21 and 10-yearsurvivalratesof49%to 53%.15.21 Thus, as reported elsewhere," late survival by nonparametric comparison is similar for annuloaortic ectasia and dissection. Multivariate testing does not
Surgical treatment of ascending aorta
681
reveal pathology to be an independent predictor of late death. Thirty percent of late complications were related to a prosthetic valve. Freedom from thromboembolic disease (in patients with prosthetic valves) is 83.5% ± 6.8% at 10 years, which is similar to other data for composite valve grafts" and for isolated Bjork-Shiley aortic valve replacement.22.23 Of all thromboembolic, neurologic, and hemorrhagic events, only one (a stroke) occurred in a patient with a native valve. Other prosthetic valve-related complications (anticoagulant-related hemorrhage and paravalvular leak) were rare, and prosthetic valve endocarditis was not diagnosed. This information is offset against the 5-year survival figures of 50% ± 18.6% for resuspension and grafting and 60% ± 18.2 % for aortic grafting alone, compared with 74.0% ± 7.6% for composite valve graft operations and 77% ± 10% for separated replacements. Although there are data showing that the late mortality rate for aortic valve replacements is not excessiveI 6. 17 (confirmed by multivariate analysis of our own data), there is an increased incidence of nonfatal complications, for which reason we recommend preservation of the native valve in patients without Marfan's disease or annular dilatation. Aortic regurgitation occurred in two of 16 patients whose native valve was preserved. The failure mode is benign and encourages us to persevere with resuspension. Much late attrition is due to left ventricular dysfunction, myocardial infarction, and aneurysmal disease in ungrafted aortic segments. Reaperation has been required for residual problems in the root or ascending aorta or more distal aorta in 11% (9/83). This group of nine patients includes four with Marfan's disease. We have identified two other patients with persistence of the false lumen in the descending thoracic aorta, one of whom died 3 years postoperatively of cardiac failure. The other has had an additional operation. This experience highlights the need for indefinite surveillance and regular assessment of the thoracic aorta by computed tomography in all patients requiring surgical treatment of the ascending aorta. REFERENCES 1. Bentall H, De Bono A. A technique for complete replacement of the ascending aorta. Thorax 1967;23:338-9. 2. Wheat MW Jr, Wilson JR, Bartley TO. Successful replacement of the entire ascending aorta and aortic valve. JAMA 1964;188:717-9. 3. Groves LK, Effler DB, Hawk W A, Gulati K. Aortic insufficiency secondary to aneurysmal changes in the ascending aorta: surgical management. J THORAC CARDIOVASC SURG 1964;48:362-79. 4. Edwards WS, Kerr AR. A safer technique for replacement
The Journal of Thoracic and Cardiovascular Surgery
6 8 2 Raudkivi et al.
5.
6.
7.
8.
9.
10.
II.
12. 13. 14.
of the entire ascending aorta and aortic valve. J THORAC CARDIOVASC SURG 1970;59:837-9. McCready RA, Pluth JR. Surgical treatment of ascending aortic aneurysms associated with aortic valve insufficiency. Ann Thorac Surg 1979;28:307-16. Kouchoukos NT, Marshall WG, Wedige-Stechner TA. Eleven-year experience with composite graft replacement of the ascending aorta and aortic valve. J THORAC CAROlOVASC SURG 1986;92:691-705. DeBakey ME, McCollum CH, Crawford ES, et al. Dissection and dissecting aneurysms of the aorta: twenty-year follow-up of five hundred twenty-seven patients treated surgically. Surgery 1982;92: 1118-34. Savunen T. Annulo-aortic ectasia: a clinical, structural, and biochemical study. Scand J Thorac Cardiovasc Surg 1986;suppl 37:7-45. Edmunds LH, Clark RE, Cohn LH, Miller DC, Weisel RD. Guidelines for reporting morbidity and mortality after cardiac valvular operations. Ann Thorac Surg 1988;46: 257-9. Cabral C. Pavie A, Mersinildrey P, et al. Long-term results with total replacement of the aortic valve and aortic root with implantation of the coronary arteries. J THORAC CARDIOVASC SURG 1986;91:17-25. Kaplan EL, Meier P. Non-parametric estimation from incomplete observations. Am Stat Assoc J 1958;53:457-81. Cox DR. Regression models and life-tables. J R Stat Soc (B) 1972;34:187-200. Dixon W J. BMDP statistical software 1985; University of California Press. Gott VL, Pyertiz RE, Magovern GJ, Cameron DE, MeKusick VA. Surgical treatment of aneurysms of the ascending aorta in the Marfan syndrome: results of compos-
15.
16.
17.
18. 19.
20.
21.
22.
23.
ite-graft repair in 50 patients. N EnglJ Med 1986;314: 10704. Cachera J-P, Vouhe PR, Loisance DY, et al. Surgical management of acute dissections involving the ascending aorta: early and late results with 38 patients. J THORAC CARDIOVASC SURG 1981;82:576-84. Jex RK, SchaffHV, Piehler JM, et al. Repair of ascending aortic dissection: influence of associated aortic valve insufficiency on early and late results. J THORAC CARDIOVASC SURG1987;93:375-84. . Miller DC, Stinson EB, Oyer PE, et al. Operative treatment of aortic dissections: experience with 125 patients over a 16 year period. J THORAC SURG 1979;78:365-82. Borst HG. Replacement of the ascending aorta and aortic valve. Ann Thorac Surg 1981;32:613-4. Bachet J, Gigou F, Laurian C, Bical 0, Goudot B, Guilmer D. Four-year clinical experience with the gelatin-resorcineformol biologic glue in acute aortic dissection. J THORAC CARDIOVASC SURG 1982;83:212-7. Donaldson RM, Ross ON. Replacement of the ascending aorta and aortic valve with a composite graft: results and follow-up in 71 patients. Thorac Cardiovasc Surg 1981; 29:195-9. Kirklin JW, Barratt-Boyes BG. Cardiac surgery: morphology, diagnostic criteria, natural history, techniques, results and indications. 1st ed. New York: John Wiley, 1986:1488. Karp RB, Cyrus RJ, Blackstone EH, Kirklin JW, Kouchoukos NT, Pacifico AD. The Bjork-Shiley valve. J THORAC CARDIOVASC SURG 1981;81:602-14. Cohn LH, Allred EN, DiSesa VJ, Sawtelle K, Shemin RJ. Collins JJ. Early and late risk or aortic valve replacement. J THORAC CARDIOVASC SURG 1984;88:695-705.
Bound volumes available to subscribers Bound volumes of THE JOURNAL OFTHORACIC AND CARDIOVASCULAR SURGERY are available to subscribers (only) for the 1989 issues from the Publisher, at a cost of $55.00 ($72.00 international) for Vol. 97 (January-June) and Vol. 98 (July-December). Shipping charges are included. Each bound volume contains a subject and author index and all advertising is removed. Copies are shipped within 60 days after publication of the last issue of the volume. The binding is durable buckram with the JOURNAL name. volume number. and year stamped in gold on the spine. Payment must accompany all orders. Contact The C.V. Mosby Company. Circulation Department. 11830 Westline Industrial Drive. St. Louis. Missouri 63146-3318. USA; phone (800) 325-4177. ext. 351. Subscriptions must be in force to qualify. Boond volumes are not available in place of a regular JOURNAL subscription.