280 © 2013 Wiley Periodicals, Inc.

DISEASES OF THE AORTA

ORIGINAL ARTICLE______________________________________________________________

Long-Term Mortality and Morbidity

after Button Bentall Operation
Tae Sik Kim, M.D.,* Chan-Young Na, M.D.,y Sam Sae Oh, M.D.,*
and Jae Hyun Kim, M.D.*

*Department of Thoracic and Cardiovascular Surgery, Sejong General Hospital, Sejong Heart
Institute, Bucheon, Republic of Korea; and yDepartment of Thoracic and Cardiovascular
Surgery, Keimyung University Dongsan Medical Center, Daegu, Republic of Korea

ABSTRACT Background and aim of the study: The purpose of this study is to evaluate the long-term outcomes
of the button Bentall procedure for the correction of aortic root disease. Methods: A total of 195 patients who
underwent the button Bentall procedure between 1997 and 2010 were studied. The main pathology was
annuloaortic ectasia. A mechanical valve was used in 163 patients (83.6%). The median duration of follow-up
was 64 months (14133.0 patient-years). Results: There were five operative deaths (2.6%). Late overall
mortality was 7.9%. The actuarial overall survival rate was 95.8 W 1.5% at 5 years, 89.6 W 3.4% at 10 years,
and 75.9 W 7.3% at 15 years. Multivariate logistic regression analysis identified preoperative poor mobility,
cardiopulmonary bypass time, deep hypothermic circulatory arrest (DHCA) use, embolism, and bleeding
event as significant independent risk factors for the late overall mortality. At 5, 10, and 15 years, actuarial
composite valve graft-related event-free survival was 85.8 W 2.8%, 75.5 W 4.4%, and 69.3 W 7.3%,
respectively. Hypertension and concomitant coronary artery bypass graft (CABG) were independent
predictors of composite valve graft-related events. Age, concomitant CABG, and DHCA use were associated
with bleeding. Conclusions: Valve-related morbidities, such as embolism and bleeding, determine the long-
term overall mortality in aortic root replacement with button Bentall operation, similar to that in routine
valve surgery. doi: 10.1111/jocs.12085 (J Card Surg 2013;28:280–284)

Aortic root replacement procedures, which include
the modifications of the original Bentall operation, the
Carrel button technique to reattach the coronary ostia,
and the open technique, are referred to as the button
Bentall procedure. Only a few papers regarding the
button Bentall procedure include mean follow-up
duration of more than five years.1,2 We evaluated the
long-term clinical outcomes and sought to determine
the risk factors of long-term mortality and morbidity for
the button Bentall procedure in 195 patients.
button Bentall technique. During the study period,
valve-sparing procedures were performed in 19 pa-
tients (reimplantation in 6 patients, remodeling in 13
patients). These patients were excluded from this
study. Preoperative variables are depicted in Table 1.
EuroSCORE II was available in 149 patients (76.4%).3
Marfan syndrome was diagnosed using the Ghent
criteria.4 The study was reviewed and approved by the
Institutional Review Board of our center. Individual
patient consent was waived.

MATERIALS AND METHODS Surgical technique

Patients’ characteristics
Our study includes patients who were operated on
from 1997 to 2010 in a single center. A total of 195
patients underwent aortic root replacement, using the
Conflict of interest: The authors acknowledge no conflict of interest
in the submission.
Address for correspondence: Chan-Young Na, M.D., Department of
Thoracic and Cardiovascular Surgery, Keimyung University Dongsan
Medical Center, 56 Dalseong-ro, Jung-gu, Daegu 700-712, Republic of
Korea. Fax: 82-53-250-7795; e-mail: koreaheartsurgeon@hotmail.com
The button Bentall operation was performed via a
median sternotomy. A repeat median sternotomy was
required in 17 patients. Cardiopulmonary bypass (CPB)
was instituted with arterial cannulation via the ascend-
ing aorta, femoral artery, or right axillary artery, and with
venous cannulation via bicaval, right atrium, or femoral
vein as necessary. Myocardial protection was provided
with cold crystalloid cardioplegia (HTK solution; n ¼ 16)
or cold blood cardioplegia (1:4; n ¼ 179). The heart was
vented via the right superior pulmonary vein. Deep
hypothermic circulatory arrest (DHCA) with antegrade
selective (n ¼ 12), retrograde cerebral perfusion
J CARD SURG KIM, ET AL. 281
2013;28:280–284 LONG-TERM MORBIDITY BUTTON BENTALL

TABLE 1 TABLE 2
Preoperative Characteristics (N = 195) Operative Data (N = 195)

Variables Number of Patients (%) Variables Number of Patients (%)

Mean age 50.48  15.16* (51, 22)** Bicuspid aortic valve 47 (24.1)
Female 57 (29.2) Implanted valve type
NYHA class 2.52  0.68* (3, 1)** Mechanical valve 163 (83.6)
Hypertension 82 (42.1) Bioprosthetic valve 32 (16.4)
Diabetes mellitus 14 (7.2) Arch replacement
Coronary artery disease 20 (10.3) Hemiarch 10 (5.1)
CVA 16 (8.2) Total arch with elephant trunk 2 (1.0)
Previous cardiac operation 17 (8.7) Concomitant cardiac procedures
Marfan syndrome 47 (24.1) Mitral valve surgery 21 (10.8)
Atrial fibrillation 27 (13.8) Tricuspid valve surgery 7 (3.6)
EuroSCORE II 4.21  4.99* (3, 2.49)** Coronary artery bypass graft 22 (11.3)
Echocardiographic findings Maze operation 16 (8.2)
Ejection fraction (%) 55.67  12.63* (55, 19)** Others 6 (3.1)
Grade of aortic regurgitation 3.18  1.19* (4, 1)** Cardiopulmonary bypass
Main aortic pathology CPB time (min) 189.76  71.83* (168, 62)**
Annuloaortic ectasia 106 (54.4) ACC time (min) 142.15  53.16* (127, 53)**
Ascending aortic aneurysm 51 (26.2) DHCA use 33 (16.9)
Acute type A dissection 24 (12.3) Secondary CPB 8 (4.1)
Chronic type A dissection 5 (2.6) Timing of operation 26 (13.3)
Previous valve failure*** 5 (2.6) Emergency 14 (7.2)
Severe aortic stenosis 3 (1.5) Urgency 12 (6.2)
Infective endocarditis 1 (0.5)
CPB, cardiopulmonary bypass; ACC, aortic cross-clamp;
NYHA, New York Heart Association; CVA, cerebrovascular DHCA, deep hypothermic circulatory arrest.*Mean  standard
accident; EuroSCORE II was available in 149 patients (76.4%) deviation;**Median, interquatile range.
because of incomplete data collection.*Mean  standard
deviation;**Median, interquatile range;***Prosthetic (N ¼ 2),
homograft (N ¼ 3).

(n ¼ 17), or DHCA only (n ¼ 4) was used in 33 patients. Anticoagulation

The aorta was completely transected near the sino-
tubular junction, and distally for 2 cm after the cross
clamping. The coronary buttons were excised with a
1.5 cm diameter cuff of aortic wall and mobilized to
facilitate reimplantation. Proximal anastomosis of the
graft was accomplished with interrupted pledgeted
sutures. We used the St. Jude mechanical valve graft
prosthesis (St. Jude Medical, St. Paul, MN, USA) in 155
patients. When other valves (8 mechanical valves and
32 biological valves) were used, we made a conduit by
joining the sewing ring of the valve prosthesis to a
Hemashield graft (MAQUET, Rastatt, Germany) or
Gelweave graft (Vascutek, Scotland, UK) with a running
suture. The tube graft was oversized 3 to 7 mm larger
than the diameter of the prosthetic valve. Patients older
than 65 years of age and those with contraindications to
anticoagulation were generally advised to choose a
biologic valve. Patients younger than 65 years of age
were usually recommended a mechanical valve. Coro-
nary button anastomoses were performed with contin-
uous running sutures, reinforced with small strips of
Teflon felt. Distal anastomosis of the graft to the aortic
wall was performed with continuous running sutures or
interrupted pledgeted sutures. In the case of continu-
ous running suture, a narrow band of Teflon felt was
placed outside of the aortic wall. Concomitant proce-
dures were performed in 62 patients (31.8%); mitral
valve surgery in 21, tricuspid valve surgery in 7,
coronary artery bypass graft (CABG) in 22, hemiarch
replacement in 10, total arch replacement with elephant
trunk in 2, maze operation in 16, and others in 6.
Operative data are described in Table 2.
After the operation using a mechanical valve, warfarin
was started with a target prothrombin time international
normalized ratio (INR) of 2.0 to 3.0. In patients with atrial
fibrillation and previous thromboembolism, INR with
warfarin was maintained between 2.5 and 3.5. Patients
with biological valves were advised to take warfarin for
the first three months. In the case of concomitant
CABG with any valve type, 100 mg of aspirin per day
was added.
Follow-up
Clinical follow-up data were collected from the
database and retrospective review of medical records.
Follow-up information was also obtained by means of
telephone interview of the patient or the patient’s
relatives. The date of last inquiry was between March
and June 2012. The median duration of the follow-up of
190 hospital survivors was 64 months (interquatile
range, 74 months [14133.0 patient-years]), and 90.5%
(n ¼ 172) of hospital survivors had completed follow-up.
The definition of complications followed the “Guide-
lines for reporting morbidity and mortality after cardiac
valvular operations.”5 The postoperative morbidity also
included pseudoaneurysm formation at the anastomo-
sis site. Operative mortality was defined as death
occurring during the hospitalization or within 30 days of
the operation. Follow-up duration for the overall survival
was measured from the date of the surgery to the
date of death, or of last contact alive, and for the event-
free survival, from the date of the surgery to the date of
282 KIM, ET AL.
LONG-TERM MORBIDITY BUTTON BENTALL
the first event, or last contact alive. From one month
after the operation, chest computed tomography (CT)
was checked in 113 patients (59.5%; of 190 hospital
survivors) for the detection of pseudoaneurysm forma-
tion. Asymptomatic routine follow-up chest CT was not
mandatory. The median follow-up duration of chest CT
was 45 months (interquatile range, 61 months).
Statistical analysis
Descriptive statistics were presented as the mean 
standard deviation (SD) or median [interquatile range
(IQR)] for continuous variables and number and
percentage for categorical variables. Differences be-
tween continuous variables were tested using Stu-
dent’s t-test. Differences between categorical variables
were tested using Fisher’s exact test. The results with
p values of less than 0.05 were considered statistically
significant. Subsequent stepwise logistic regression
analysis (backward methods) was performed to find the
potential risk factors of mortality and morbidity. Long-
term survival and event-free survival were analyzed
with the Kaplan–Meier method, and compared with log-
rank test. The SPSS Statistical software 18.0 program
(SPSS, Chicago, IL, USA) was used for statistical
analyses.
RESULTS
Early outcomes
The operative mortality was 2.6% (5 of 195 patients).
Causes of early death were heart failure (n ¼ 1), acute
myocardial infarction (n ¼ 1), respiratory failure (n ¼ 1),
bowel ischemia (n ¼ 1), and diffuse hypoxic brain
damage (n ¼ 1).
Early complications included postoperative rester-
notomy for bleeding (n ¼ 11), left ventricular dysfunc-
tion (ejection fraction lower than 30%; n ¼ 12), atrial
fibrillation (n ¼ 22), atrial flutter (n ¼ 2), complete
atrioventricular block (n ¼ 2), mediastinitis (n ¼ 3),
pneumonia (n ¼ 4), prolonged mechanical ventilation
(more than five days; n ¼ 20), cerebral infarction
(n ¼ 5), renal failure needed hemodialysis (n ¼ 2),
pericardial effusion needed drainage (n ¼ 5), gastroin-
testinal bleeding (n ¼ 1), bowel ischemia (n ¼ 1),
brachial plexus injury (n ¼ 1), deep vein thrombosis
(n ¼ 1), and diaphragm palsy (n ¼ 3).
Long-term results
There have been 15 late deaths (7.9% of 190 hospital
survivors) during the follow-up period. There were
seven cardiac and eight noncardiac deaths. Causes of
cardiac death included heart failure (n ¼ 2), cerebral
hemorrhage (n ¼ 2), cerebral infarction (n ¼ 2), and
unknown (n ¼ 1). There were eight noncardiac deaths:
cancer (n ¼ 3), trauma (n ¼ 2), pneumonia (n ¼ 1),
sepsis (n ¼ 1), and aplastic anemia (n ¼ 1).
The actuarial rate of the overall survival at 5, 10,
and 15 years was 95.8  1.5%, 89.6  3.4%, and
75.9  7.3%, respectively (Fig. 1). Multivariate logistic
regression analysis identified preoperative poor mobili-
J CARD SURG
2013;28:280–284
Figure 1. Kaplan–Meier curve of overall survival of 190
hospital survivors.
ty (OR, 30.831; 95% CI, 1.673 to 568.108; p ¼ 0.0211),
CPB time (OR, 1.009; 95% CI, 1.001 to 1.016;
p ¼ 0.00180), DHCA use (OR, 3.361; 95% CI, 0.888
to 12.715; p ¼ 0.0742), embolism (OR, 12.43; 95% CI,
1.540 to 100.227; p ¼ 0.0180), and bleeding event (OR,
3.40; 95% CI, 0.814 to 14.210; p ¼ 0.0935) as
significant independent risk factors for the late overall
death (Table 3).
The actuarial composite valve graft-related event-free
survival at 5, 10, and 15 years was 85.8  2.8%,
75.5  4.4%, and 69.3  7.3%, respectively. Multivar-
iate logistic regression analysis identified the following
factors to be independent predictors of composite valve
graft-related events: hypertension (OR, 2.408; 95% CI,
1.061 to 5.464; p ¼ 0.0356) and concomitant CABG
(OR, 5.538; 95% CI, 2.027 to 15.127; p ¼ 0.0008;
Table 4).
Bleeding events occurred in 22 patients (11.6%).
Prothrombin time INR at the time of the bleeding event
was 3.98  3.18 (mean  SD; Table 5). Of these 22
patients, two (INR in one, 1.99; other, unknown) had a
bioprosthetic valve. Multivariate logistic regression
analysis revealed that age (OR, 1.045; 95% CI, 1.005
to 1.086; p ¼ 0.0255), concomitant CABG (OR, 4.18;
95% CI, 1.328 to 13.16; p ¼ 0.0145), and DHCA use
(OR, 4.217; 95% CI, 1.419 to 12.53; p ¼ 0.0096) were
associated with bleeding events (Table 4).
Embolism occurred in seven patients (3.7%), with
neurologic events, including five temporary and two
permanent. Of these, one patient experienced two
temporary neurologic events. INR at the time of
embolism was 1.34  0.22 (mean  SD; Table 5). Of
the seven patients, three patients (INR in one, 1.65;
others, unknown) had bioprosthetic valves.
Structural valve deterioration and valve thrombosis
were not noted in any patients. One patient (0.5%)
experienced nonstructural valve dysfunction. This
patient, with suspected Behcet’s disease, had
J CARD SURG KIM, ET AL. 283
2013;28:280–284 LONG-TERM MORBIDITY BUTTON BENTALL

TABLE 3
Independent Predictors of Late Overall Mortality (N = 15) in the Multivariate Logistic Regression Analysis

Variables OR SE 95% CI P-Value

Poor mobility 30.831 1.4867 1.673–568.108 0.0211

CPB time 1.009 0.00359 1.001–1.016 0.0180
DHCA use 3.361 0.6789 0.888–12.715 0.0742
Embolism 12.425 1.0652 1.54–100.227 0.0180
Bleeding event 3.4 0.7297 0.814–14.21 0.0935
OR, odds ratio; SE, standard error; 95% CI, 95% confidence interval; CPB, cardiopulmonary bypass; DHCA, deep hypothermic
circulatory arrest.

TABLE 4
Independent Predictors of Composite Valve Graft-Related Events (N = 31) and Bleeding Event (N = 22) in
the Multivariate Logistic Regression Analysis

Variables OR SE 95% CI P-Value

Graft-related events
Hypertension 2.408 0.4181 1.061–5.464 0.0356
Concomitant CABG 5.538 0.5127 2.027–15.127 0.0008
Bleeding event
Age 1.045 0.0197 1.005–1.086 0.0255
Concomitant CABG 4.18 0.5851 1.328–13.160 0.0145
DHCA use 4.217 0.5556 1.419–12.530 0.0096
OR, odds ratio; SE, standard error; 95% CI, 95% confidence interval; CABG, coronary artery bypass graft; DHCA, deep hypothermic
circulatory arrest.

infective endocarditis and ascending aorta graft re-

TABLE 5 placement due to pseudoaneurysm at the distal
Prothrombin Time International Normalized Ratio anastomosis site.
(INR) at the Time of Bleeding Event and Embolism

Bleeding Event Embolism DISCUSSION

Variables (N = 18)* (N = 4)**
Mean, SD 3.981, 3.184 1.353, 0.224
Median, 2.705, 3.098 1.300, 0.418
IQR
Range 1.18–13.6 1.12–1.65
SD, standard deviation; IQR, interquatile range.*Of total 22
bleeding events, INR in 18 events was available;**Of total
seven embolisms, INR in four events was available.
paravalvular leakage with pseudoaneurysm formation
six months after the operation. He died due to heart
failure before reoperation. Valve endocarditis occurred
in three patients (1.6%); two of them were treated with
antibiotics, and the other underwent aortic root re-
replacement with an aortic homograft.
Three pseudoaneurysm formations (1.6%) were
detected on follow-up CT scanning, two at the proximal
and one at the distal graft anastomosis sites. No
pseudoaneurysms occurred at the coronary button
sites. One patient with a pseudoaneurysm in the
proximal anastomosis site died due to heart failure
before the reoperation. The other two patients under-
went reoperations (1.0% of 190 hospital survivors):
redo root replacement with aortic homograft due to
The median duration of follow-up in our study was
64 months (5.3 years). There are only two papers
written about the button Bentall procedure which
include a mean or median follow-up duration of more
than five years: the mean 97 months (8.1 years) by
Savunen et al.1 and median 6.6 years by Etz and et al.2
(Table 6).
In our series, concomitant CABG was a significant
predictor of composite valve graft-related events (p ¼
0.0008) along with bleeding events (p ¼ 0.0145). In
addition, of the 22 patients who experienced the
bleeding event, only two had bioprosthetic valves.
We speculate that concomitant CABG is a risk factor for
bleeding events due to the combined therapy of
warfarin and aspirin. In the meta-analysis by Cappelleri
et al.,6 the risk of bleeding events after mechanical
heart valve replacement was increased by the combi-
nation of a vitamin K antagonist and an antiplatelet
agent. Laffort et al.7 reported that the combination of
vitamin K antagonist and aspirin had a higher rate of
major bleeding than that with vitamin K antagonists
alone. Their study included patients with St. Jude
Medical valve prostheses, despite relatively higher
dosage (200 mg daily) of aspirin and higher INR ranges
(2.5 to 3.5) than our own patients.
284 KIM, ET AL.
LONG-TERM MORBIDITY BUTTON BENTALL
TABLE 6
Previous Reports about Button Bentall Operation
Operative
Refs. Follow-Up Mortality (%)
Biglioli et al.8 Mean, 39 months 14.9
Lepore et al.9 Mean, 4.7 years 18
Hilgenberg et al.10 — 7.3
Savunen et al.1 Mean, 97 months 3 —
Westaby et al.11 Mean, 33.8 months 5.7
Hagl et al.12 Median 3.5 years 0 95% (five years),
Etz et al.2 Median, 6.6 years 3.9
—, not reported.
Since the late 1990s, pseudoaneurysm formation
after the button Bentall operation have rarely been
reported (Table 6).1,2,8–12 In our study, three pseudoa-
neurysm formations (1.6%) were detected at the graft
anastomosis sites, not in the coronary button sites.
Unfortunately, because chest CT was performed in only
59.5% of hospital survivors, our incidence of pseudoa-
neurysm formation might be underestimated.
Our study has some limitations that are inherent in a
retrospective review. Observational data do not provide
causal evidence. Owing to incomplete data collection,
several important variables, such as EuroSCORE II,
aortic root diameter, and medical therapy, were omitted
in the statistical analysis. The valve-related morbidity
cannot be accurately estimated in this series because
18 patients (9.5%) of 190 hospital survivors were lost to
follow-up. The estimation of odds ratio, with regard to
operative death (n ¼ 5) and late cardiac death (n ¼ 7),
could not be possible in logistic regression analysis due
to the low number of events.
In this study, we identified that valve-related morbid-
ities, such as embolism and bleeding events, are
significant determinants of long-term mortality in aortic
root replacement with button Bentall operation, similar
to that in routine valve surgery. Furthermore, concomi-
tant CABG is a risk factor for bleeding events in the
button Bentall procedure.
Acknowledgments: We thank Hyonggin Ann, Ph.D., our
consultant statistician, for biostatistical advice of the study.
REFERENCES
1. Savunen T, Inberg M, Niinikoski J, et al: Composite graft in
annulo-aortic ectasia: Nineteen years’ experience without
graft inclusion. Eur J Cardiothorac Surg 1996;10:428–
432.
J CARD SURG
2013;28:280–284
Bleeding
Survival Rate Event Pseudoaneurysm
78% (one years), — —
67% (five years)
67% (five years), 7.8% 5.8%
48% (ten years)
78% (five years), — 0%
70% (ten years)
— 0%
94% (one year), 0.9% 0%
79% (five years)
7% —
93% (eight years)
— — —
2. Etz CD, Bischoff MS, Bodian C, et al: The Bentall
procedure: Is it the gold standard? A series of 597
consecutive cases. J Thorac Cardiovasc Surg 2010;140:
S64–S70.
3. Nashef SA, Roques F, Sharples LD, et al: EuroSCORE II.
Eur J Cardiothorac Surg 2012;41:734–744.
4. De Paepe A, Devereux RB, Dietz HC, et al: Revised
diagnostic criteria for the Marfan syndrome. Am J Med
Genet 1996;62:417–426.
5. Akins CW, Miller DC, Turina MI, et al: Guidelines for
reporting mortality and morbidity after cardiac valve
interventions. Eur J Cardiothorac Surg 2008;33:523–528.
6. Cappelleri JC, Fiore LD, Brophy MT, et al: Efficacy and
safety of combined anticoagulant and antiplatelet therapy
versus anticoagulant monotherapy after mechanical heart
valve replacement: A meta-analysis. Am Heart J 1995;
130:547–552.
7. Laffort P, Roudaut R, Roques X, et al: Early and long-term
(one-year) effects of the association of aspirin and oral
anticoagulant on thrombi and morbidity after replacement
of the mitral valve with the St. Jude medical prosthesis.
J Am Coll Cardiol 2000;35:739–746.
8. Biglioli P, Sala A, Spirito R, et al: Composite valve graft
replacement of the ascending aorta and the aortic valve by
a modified button technique: The influence of aortic
pathology on early mortality and late survival. Eur J
Cardiothorac Surg 1995;9:483–490.
9. Lepore V, Larsson S, Bugge M, et al: Replacement of the
ascending aorta with composite valve grafts: Long term
results. J Heart Valve Dis 1996;5:240–246.
10. Hilgenberg AD, Akins CW, Logan DL, et al: Composite
aortic root replacement with direct coronary artery
implantation. Ann Thorac Surg 1996;62:1090–1095.
11. Westaby S, Katsumata T, Vaccari G: Aortic root replace-
ment with coronary button re-implantation: Low risk and
predictable outcome. Eur J Cardiothorac Surg 2000;17:
259–265.
12. Hagl C, Strauch JT, Spielvogel D, et al: Is the Bentall
procedure for ascending aorta or aortic valve replacement
the best approach for long-term event free survival? Ann
Thorac Surg 2003;76:698–703.