Keywords
aortopathy - valve-sparing root replacement - durability - survival
Introduction
Bicuspid aortic valve (BAV) is commonly associated with proximal aortic aneurysm,
which presents as either ascending aorta or root phenotypes.[1] Patients with root phenotype have regurgitant aortic valves and dilated aortic anuli.[2] Preservation or repair of native BAV in this relatively young population by valve-sparing
root replacement seems an attractive alternative to a Bentall procedure with its anticoagulation
requirement, related thromboembolism, and bleeding complications.
Root reimplantation and root remodeling are currently the two main valve-sparing root
replacement approaches. Apart from their widespread application and excellent outcome
in the setting of tricuspid aortic valves (TAVs),[3] these approaches to BAV root aneurysms have limited clinical experience, with fewer
than 100 cases in most single-center reports.[4] Beckmann and colleagues[5] reported a 20-year survival of 84% and freedom from reoperation of 74% among 50
patients with BAV who underwent root reimplantation. We recently found that the durability
of root reimplantation in the BAV setting may be inferior to TAV.[3] Root remodeling is another promising valve-sparing root replacement approach for
BAV root aneurysms. Schneider and colleagues[6] reported a 15-year survival of 81% and freedom from reoperation of 78% among 357
patients with BAV root remodeling.
Due to the limited clinical experience so far, there is little evidence regarding
whether root reimplantation or remodeling is better for the BAV root phenotype. The
choice between the two valve-sparing root replacement approaches now mainly depends
on a surgeon's experience and preference. Therefore, in this study, we compared outcomes
of patients with BAV root phenotype who underwent root reimplantation versus root
remodeling, aiming to provide information to support evidence-based surgical decision-making.
Patients and Procedures
Patients
From January 2000 to January 2022, 214 patients underwent valve-sparing root replacement
to treat BAV root aneurysm at Cleveland Clinic. After excluding emergency surgery,
type A aortic dissection, endocarditis, and aortic valve reoperation, 206 patients
were included in the study ([Fig. 1]), with 11% female and mean age 47 ± 12 years, among whom 32 underwent root remodeling
and 174 root reimplantation ([Fig. 1]). Although the number of root remodeling procedures was nearly constant over time,
the number of root reimplantation gradually increased after 2003 ([Fig. 2]).
Fig. 1 CONSORT-style diagram of patients undergoing bicuspid aortic valve root remodeling
or reimplantation. AV, aortic valve.
Fig. 2 Number of patients undergoing bicuspid aortic valve (BAV) root remodeling (blue line
and symbols) or reimplantation (red line and symbols) over the study period. Symbols
are a yearly number of cases in each group.
Compared with the root remodeling group, patients who underwent root reimplantation
had more aortic regurgitation (severe: 61/174 [35%] vs. 3/32 [9.4%]), larger left
ventricular end-systolic volume index (26 ± 11 vs. 21 ± 9.1 mL/m2), and smaller aortic roots (sinus diameter: 4.3 ± 0.56 vs. 4.6 ± 0.47 cm; [Table 1]).
Table 1
Baseline characteristics of patients undergoing bicuspid aortic valve root remodeling
or reimplantation
|
Characteristics
|
BAV root remodeling (n = 32)
|
BAV root reimplantation (n = 174)
|
Standardized mean difference (%)
(95% PR[b])
|
|
n
[a]
|
Number (%) or mean ± SD
|
n
[a]
|
Number (%) or mean ± SD
|
|
Demographics:
|
|
Age (years)
|
32
|
45 ± 12
|
174
|
47 ± 12
|
−15 (−38, 37)
|
|
Female
|
32
|
1 (3.1)
|
174
|
22 (13)
|
−36 (−36, 36)
|
|
Body surface area (m2)
|
32
|
2.1 ± 0.18
|
174
|
2.1 ± 0.25
|
9.3 (−39, 39)
|
|
Aortic regurgitation grade
|
32
|
–
|
174
|
–
|
−63 (−36, 36)
|
|
None/Trace
|
|
15(47)
|
|
49(28)
|
|
|
Mild
|
8(25)
|
29(17)
|
|
Moderate
|
6(19)
|
35(20)
|
|
Severe
|
3(9.4)
|
61(35)
|
|
Aortic root size (diameter):
|
|
Aortic valve anulus (cm)
|
8
|
2.7 ± 0.35
|
63
|
2.6 ± 0.51
|
27 (−69, 68)
|
|
Aortic sinus (cm)
|
28
|
4.6 ± 0.47
|
159
|
4.3 ± 0.56
|
55 (−38, 38)
|
|
Sinutubular junction (cm)
|
22
|
4.2 ± 0.54
|
109
|
3.9 ± 0.59
|
51 (−52, 49)
|
|
Middle ascending aorta (cm)
|
30
|
4.5 ± 0.63
|
157
|
4.5 ± 0.74
|
3.3 (−39, 39)
|
|
LV morphology and function:
|
|
Ejection fraction (%)
|
32
|
57 ± 7.7
|
174
|
60 ± 5.8
|
44 (−37, 39)
|
|
End-diastolic volume index (mL/m2)
|
31
|
58 ± 22
|
157
|
70 ± 24
|
−52 (−40, 40)
|
|
End-systolic volume index (mL/m2)
|
31
|
21 ± 9.1
|
156
|
26 ± 11
|
−47 (−39, 43)
|
|
Mass index (g/m2)
|
30
|
104 ± 35
|
156
|
115 ± 39
|
−24 (−40, 38)
|
|
Relative wall thickness (cm)
|
30
|
0.44 ± 0.13
|
157
|
0.40 ± 0.10
|
44 (−41, 41)
|
|
Other cardiovascular comorbidities:
|
|
Atrial fibrillation or flutter
|
30
|
3 (10)
|
171
|
15 (8.8)
|
4.2 (−49, 37)
|
|
Heart failure
|
32
|
1 (3.1)
|
173
|
20 (12)
|
−33 (−52, 39)
|
|
Prior cardiac surgery
|
32
|
4 (13)
|
174
|
15 (8.6)
|
13 (−50, 34)
|
|
Noncardiac comorbidities:
|
|
Pharmacologically treated diabetes
|
32
|
1 (3.1)
|
174
|
3 (1.7)
|
9.1 (−22, 27)
|
|
COPD
|
32
|
0 (0)
|
174
|
6 (3.4)
|
−27 (−27, 34)
|
|
Peripheral artery disease
|
32
|
1 (3.1)
|
174
|
6 (3.4)
|
−1.8 (−29, 31)
|
|
Hypertension
|
32
|
20 (63)
|
174
|
104 (60)
|
5.6 (−40, 38)
|
|
History of smoking
|
31
|
13 (42)
|
174
|
50 (29)
|
28 (−43, 40)
|
|
Dyslipidemia
|
30
|
12 (40)
|
141
|
55 (39)
|
2.0 (−41, 40)
|
|
Creatinine (mg/dL)
|
32
|
1.01 ± 0.14
|
174
|
0.99 ± 0.37
|
6.7 (−29, 29)
|
|
Hematocrit (%)
|
32
|
43 ± 3.1
|
174
|
44 ± 3.6
|
−31 (−40, 39)
|
|
Concomitant procedures:
|
|
Aortic arch repair
|
32
|
5 (16)
|
174
|
20 (11)
|
12 (−39, 32)
|
|
Coronary artery bypass grafting
|
32
|
1 (3.1)
|
174
|
7 (4.0)
|
−4.8 (−31, 27)
|
|
Mitral valve repair
|
32
|
0 (0)
|
174
|
7 (4.0)
|
−29 (−29, 31)
|
|
Procedure for atrial fibrillation
|
32
|
1 (3.1)
|
174
|
26 (15)
|
−42 (−42, 37)
|
|
Aortic repair techniques:
|
|
Cabrol suture
|
32
|
22 (69)
|
174
|
25 (14)
|
132 (−43, 39)
|
|
Cusp plication
|
32
|
11 (34)
|
174
|
91 (52)
|
−37 (−37, 39)
|
|
Figure-of-8 hitch-up stitch
|
32
|
14 (44)
|
174
|
10 (5.7)
|
98 (−37, −34)
|
Abbreviations: BAV, bicuspid aortic valve; COPD, chronic obstructive pulmonary disease;
LV, left ventricular; PR, plausible range; SD, standardized deviation.
a Patients with data available.
b Plausible range under the null hypothesis that the population standardized difference
is zero based on 1,000 permutations.
Operative Techniques
For root remodeling, we prefer to use the inclusive technique for left and right coronary
reimplanting to achieve better hemostasis.[7] We routinely perform root reimplantation using Svensson's modified approach,[8] placing pledgeted stitches through the aortoventricular junction and reducing the
anular size by tying these stitches around a Hegar dilator of customized size (which
is typically larger [23–25 mm] as compared to those with a tricuspid valve).
To achieve aortic valve competence, we usually use additional repair techniques. Compared
with root remodeling, reimplantation was accompanied more by cusp plication (91/174
[52%] vs. 11/32 [34%]) and supracommissural sutures (47/174 [27%] vs. 1/32 [3.1%])
and less by figure-of-8 hitch-up stitches[9] (10/174 [5.7%] vs. 14/32 [44%]) and Cabrol sutures (25/174 [14%] vs. 22/32 [69%];
[Table 2]). Concomitant cardiac procedures included coronary artery bypass grafting, mitral
valve surgery, aortic arch repair, and surgical ablation for atrial fibrillation.
Table 2
Operative details and in-hospital outcomes of patients undergoing bicuspid aortic
valve root remodeling or reimplantation
|
Characteristics
|
BAV root remodeling (n = 32)
|
BAV root reimplantation (n = 174)
|
Standardized mean difference (%)
(95% PR[b])
|
|
n
[a]
|
Number (%) or median [15th, 85th percentile]
|
n
[a]
|
Number (%) or Median [15th, 85th percentile]
|
|
|
Operative details:
|
|
Aortic repair techniques:
|
|
Cabrol suture
|
32
|
22 (69)
|
174
|
25 (14)
|
132 (−43, 39)
|
|
Cusp plication
|
32
|
11 (34)
|
174
|
91 (52)
|
−37 (−37, 39)
|
|
Figure-of-8 hitch-up stitch
|
32
|
14 (44)
|
174
|
10 (5.7)
|
98 (−37, −34)
|
|
Concomitant procedures:
|
|
Aortic arch repair
|
32
|
5 (16)
|
174
|
20 (11)
|
12 (−39, 32)
|
|
Coronary artery bypass grafting
|
32
|
1 (3.1)
|
174
|
7 (4.0)
|
−4.8 (−31, 27)
|
|
Mitral valve repair
|
32
|
0 (0)
|
174
|
7 (4.0)
|
−29 (−29, 31)
|
|
Procedure for atrial fibrillation
|
32
|
1 (3.1)
|
174
|
26 (15)
|
−42 (−42, 37)
|
|
In-hospital outcomes:
|
p-Value
|
|
Operative death
|
32
|
1 (3.1)
|
174
|
1 (0.57)
|
0.29
|
|
Permanent stroke
|
32
|
1 (3.1)
|
174
|
0 (0)
|
0.16
|
|
Reoperation for valve dysfunction
|
32
|
2 (6.3)
|
174
|
0 (0)
|
0.02
|
|
Reoperation for bleeding or tamponade
|
32
|
0 (0)
|
174
|
5 (2.9)
|
>0.99
|
|
Any blood product transfusion
|
32
|
14 (44)
|
173
|
78 (45)
|
0.89
|
|
New requirement for dialysis
|
27
|
0 (0)
|
171
|
2 (1.2)
|
>0.99
|
|
Prolonged ventilation
|
29
|
0 (0)
|
174
|
2 (1.1)
|
>0.99
|
|
New postoperative atrial fibrillation
|
27
|
4 (15)
|
156
|
34 (22)
|
0.61
|
|
AR severity at discharge
|
21
|
|
142
|
|
0.04
|
|
None
|
|
17(81)
|
|
130 (92)
|
|
|
Mild
|
2 (9.5)
|
11 (7.8)
|
|
Moderate
|
1 (4.8)
|
1 (0.70)
|
|
Severe
|
1 (4.8)
|
0 (0)
|
|
ICU length of stay (hours)
|
32
|
32 [22, 73]
|
172
|
39 [21, 70]
|
0.69
|
|
Operative length of stay (days)
|
32
|
6 [5, 6]
|
174
|
6 [5, 8]
|
0.56
|
Abbreviations: AR, aortic regurgitation; BAV, bicuspid aortic valve; ICU, intensive
care unit; PR, plausible range.
a Patients with data available.
b Plausible range under the null hypothesis that the population standardized difference
is zero based on 1,000 permutations.
Data
Preoperative patient characteristics and operative details were collected prospectively
for quality reporting by independent registry nurses and entered into the cardiovascular
information registry. Transthoracic echocardiographic data were measured and entered
into the echocardiography database by clinical echosonographers. Other Cleveland Clinic
electronic medical record databases were also queried. All data used for this study
were approved for use in research by the Cleveland Clinic Institutional Review Board,
with patient consent waived (IRB #22-671, approved July 11, 2022).
Endpoints
Operative Mortality and Morbidity
Operative mortality included in-hospital deaths and any death occurring after hospital
discharge but within 30 days after surgery. In-hospital morbidity was defined according
to the Society of Thoracic Surgeons National Database.
Longitudinal Echocardiographic Outcomes
Aortic valve regurgitation grade (none, mild, moderate, or severe), mean gradient,
and left ventricular mass index were assessed on serial postoperative echocardiograms.
A total of 123 echocardiograms were available for 75% of patients (24/32) in the remodeling
group, and 531 echocardiograms were available for 94% of patients (164/174) in the
reimplantation group ([Supplementary Fig. S1] [available in the online version]). All longitudinal measurements were censored
at reoperation.
Time-related Aortic Valve Reoperation and Mortality
To assess reoperations on the aortic valve and vital status, patients underwent follow-up
systematically at 2 and 5 years and 5-year intervals thereafter, via mailed questionnaires
or telephone contact with the patient or family member. Systematic follow-up for vital
status was supplemented with Social Security Death Master File data (to 2011) and
Ohio State Death Registry data. The median follow-up duration for reoperation on the
aortic valve in the remodeling group was 2.0 years, with 25% of patients undergoing
follow-up of more than 5.3 years and 10% for more than 15 years. In the reimplantation
group, the median follow-up duration was 2.0 years, with 25% of patients undergoing
follow-up for more than 5.0 years and 10% for more than 15 years. The median follow-up
duration for vital status in the remodeling group was 5.1 years, with 25% undergoing
follow-up more than 13 years and 10% more than 18 years. In the reimplantation group,
the median follow-up duration was 3.0 years, with 25% followed more than 5.5 years
and 10% more than 8.3 years.
Statistical Analysis
Statistical analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC)
and R software version 4.0.3 (R Foundation for Statistical Computing, Vienna, Austria).
Continuous variables are summarized as mean ± standard deviation or as equivalent
15th, 50th (median), and 85th percentiles when the distribution of values was skewed.
Categorical data are summarized by frequencies and percentages. Continuous outcomes
were compared using the Wilcoxon rank-sum test and categorical outcomes using the
chi-square test or Fisher's exact test, if appropriate. Confidence intervals for longitudinal
estimates used a bootstrap percentile method to obtain 68% confidence bands (equivalent
to ±1 standard error) and the delta method for time-related events. A type I error
of 0.05 was used to assess statistical significance. For baseline characteristics,
standardized mean differences (SMDs) were used for comparison. An absolute value of
SMD equal to or less than 10% is usually interpreted as an acceptable difference.[10] However, when the sample size is smaller in the overall cohort or one group, as
in this study, there may be larger variability in SMD that could yield estimated SMD
values greater than 10%. To assess if the true underlying SMD is zero, we estimated
a 95% plausible range using the empirical distribution of SMD from 1,000 permutations
under the null hypothesis that population SMD is zero.[11] Any estimated SMD not falling within the interval was considered imbalanced.
Echocardiographic Longitudinal Analysis
To assess the temporal trend of individual grades of postoperative aortic regurgitation
(ordinal longitudinal data), follow-up transthoracic echocardiograms were analyzed
for the pattern of change across time using a nonlinear, multiphase, mixed-effects,
cumulative logistic regression model.[12] The model resolved the number of time phases to form a temporal decomposition model
and estimate its shaping parameters. The prevalence of each aortic regurgitation grade
over time was estimated by averaging the patient-specific profiles. A multiphase,
nonlinear, mixed-effects regression model for continuous longitudinal data was used
to estimate the temporal trend of postoperative mean gradient and left ventricular
mass index.[13]
Time-related Analysis
Survival was estimated nonparametrically by the Kaplan–Meier method. Freedom from
aortic valve reoperation was similarly estimated but with patients censored for the
competing risk of death.
Results
Operative Mortality and Morbidity
Aortic clamp time was longer for root reimplantation compared with root remodeling
(median 136 [15th and 85th percentiles of 4 and 162, respectively] vs. 76 [15th and
85th percentiles of 56 and 106, respectively] minutes, p < 0.001), as was cardiopulmonary bypass time (median 153 [15th and 85th percentiles
of 121 and 192, respectively] vs. 97 [15th and 85th percentiles of 67 and 136, respectively]
minutes, p < 0.001). One operative death occurred in each group ([Table 2]). The remodeling group had more in-hospital reoperations due to valve dysfunction
(2/32) than the reimplantation group (0/174). Occurrence of other in-hospital outcomes
was similar between groups ([Table 2]), including blood transfusions, reoperation for bleeding, and intensive care unit
(median 39 [15th and 85th percentiles of 21 and 70, respectively] vs. 32 [15th and
85th percentiles of 22 and 73, respectively] hours, p = 0.69) and postoperative (median 6 [15th and 85th percentiles of 5 and 8, respectively]
vs. 6 [15th and 85th percentiles of 5 and 6, respectively] days, p = 0.56) lengths of stay. Prevalence of moderate or more aortic valve regurgitation
at discharge was 2/21 (9.5%) in the remodeling group and 1/142 (0.70%) in the reimplantation
group (p = 0.04; [Table 2]).
Longitudinal Echocardiographic Trends
At 1 year postoperatively, prevalence of severe aortic regurgitation was 6.7% in the
remodeling group versus 1.8% in the reimplantation group. At 5 years, these proportions
increased to 16% versus 5.0%, respectively (p = 0.06; [Fig. 3]). No significant difference was found in aortic valve mean gradient between the
groups after index surgery (remodeling vs. reimplantation: 8.8 vs. 9.7 mm Hg at 1
year, and 11 vs. 10 mm Hg at 5 years, p = 0.12; [Fig. 4]). At 1 year postoperatively, mean left ventricular mass index was higher in the
remodeling group versus the reimplantation group (115 vs. 101 g/m2, p = 0.05). However, no significant difference was found at 5 years postoperatively
(105 vs. 100 g/m2, p = 0.52; [Supplementary Fig. S2] [available in the online version]).
Fig. 3 Postoperative prevalence of severe aortic regurgitation (AR) in patients undergoing
bicuspid aortic valve (BAV) root remodeling (blue lines and symbols) or reimplantation
(red lines and symbols). Solid lines represent longitudinal trends enclosed within
a dashed 68% confidence band, and symbols represent data grouped (without regard to
repeated measurements) within time frames to provide crude verification of model fit.
Fig. 4 Postoperative trend of the mean gradient in patients undergoing bicuspid aortic valve
(BAV) root remodeling (blue lines and symbols) or reimplantation (red lines and symbols).
The format is as in Fig. 3.
Time-related Reoperation and Mortality
There were 6 aortic valve reoperations after root remodeling and 10 after root reimplantation.
Freedom from aortic valve reoperation was lower after remodeling than reimplantation
(93 vs. 99% at 30 days, 93 vs. 97% at 1 year, and 77 vs. 94% at 5 years [p = 0.14]; [Fig. 5]). Reasons for reoperation after root remodeling included cusp prolapse, natural
progression, and tearing ([Supplementary Table S1] [available in the online version]). After root reimplantation, these included cusp
perforation, infection, disease progression, and tearing.
Fig. 5 Freedom from aortic valve (AV) reoperation after bicuspid aortic valve (BAV) root
remodeling (blue lines and symbols) or reimplantation (red lines and symbols). Each
symbol represents a Kaplan–Meier estimate of the event, and vertical bars are 68%
confidence limits equivalent to ±1 standard error. Numbers below the horizontal axis
are patients remaining at risk.
Nine deaths occurred during follow-up: three in the remodeling group and six in the
reimplantation group. No significant difference was found in risk of death between
the two groups (p = 0.71; [Fig. 6]). Survival estimates after root remodeling versus reimplantation were 96.8 versus
98.8% at 1 year, 96.8 versus 95.3% at 5 years, and 96.8 versus 93.5% at 8 years (p = 0.71; [Fig. 6]).
Fig. 6 Survival after bicuspid aortic valve (BAV) root remodeling (blue lines and symbols)
or reimplantation (red lines and symbols). The format is as in [Fig. 5].
Discussion
Principal Findings
This comparative study showed that root reimplantation and root remodeling could be
performed safely and had good outcomes in patients with BAV and a root aneurysm. However,
root implantation has better durability, including fewer aortic valve reoperations
and less severe aortic valve regurgitation.
Findings in Context
Aortic root anatomy and repair logic differ between bicuspid and tricuspid valve pathology.
A regurgitant TAV root reimplantation effectively downsizes the aortic anulus, which
not only increases coaptation height but also prevents later dilatation of the aortic
anulus by securing the aortoventricular junction. However, the anatomy of a BAV root
has more variations, including cusp fusion, commissural orientation, and location
of coronary ostia, so we believe that valve-sparing aortic root replacement is more
challenging in a BAV root than a TAV. It has also been pointed out that root reimplantation
in a BAV root may not be as effective as in a TAV root because reimplantation of a
BAV into a tube graft often reduces intercommissural distance and induces cusp prolapse.[7] By contrast, a large tube graft for root remodeling puts tension on the cusps longitudinally,
which helps secure the competence of bicuspid aortic cusps. Concerns about root remodeling
include bleeding risk from the long anastomosis line and lack of support for the aortic
anulus.[7]
In this study, we showed that both root remodeling and reimplantation can be safely
performed to treat root aneurysms and preserve or repair regurgitant BAVs. Likely
because it did not need deep dissection around the aortic root or placement of aortoventricular
junction stitches, root remodeling required less aortic clamp and on-pump time. In
addition, root remodeling did not have a higher risk of intraoperative and postoperative
bleeding complications compared with root reimplantation, which may be attributed
to Svensson's inclusive technique. Nevertheless, more patients in the remodeling group
required additional commissural elevation and anuloplasty by figure-of-8 hitch-up
stitches and Cabrol sutures, and still, there were two early repair failures (2/32)
after root remodeling. Conversely, we did not see any in-hospital recurrence of moderate
or severe aortic regurgitation after root reimplantation.
Despite the similarly good operative results in both groups, root remodeling may not
be as durable as reimplantation. During follow-up, we found that patients who underwent
root remodeling were more likely to have a recurrence of clinically important aortic
regurgitation or receive an aortic valve reoperation than those who underwent root
reimplantation despite having less severe aortic regurgitation preoperatively. As
our preoperative patient characteristics show, a majority of patients in this cohort
had dilated aortic anuli with a mean diameter exceeding 2.6 cm. We assume that dilated
aortic anuli may be associated with inferior long-term durability since root remodeling
was not able to effectively secure the aortic anulus even with Cabrol sutures. This
finding reinforces the belief that a complete anuloplasty provided by the reimplantation
strategy is preferred. Despite consistent anulus reduction, no significant increase
in aortic valve gradient was found within 5 years after remodeling or reimplantation.
Longer follow-up is needed to determine the long-term progression of BAV pathology
after valve-sparing root replacement and to guide refinements in patient selection
and techniques.
Limitations
This is an observational clinical study at a single institution. Due to the limited
sample size in the remodeling group, we could not perform the propensity-score matching
to minimize confounding bias. In addition, our echocardiographic follow-up data were
available only for patients routinely followed at our institution. Longer follow-up
is needed to assess longer-term outcomes.
Conclusion
Both root remodeling and reimplantation can be safely performed in patients with BAV
and root aneurysms with similarly good midterm outcomes. Although root remodeling
is a shorter operation than reimplantation, less late aortic valve regurgitation and
fewer valve reoperations lead us to recommend reimplantation. A better understanding
of BAV root anatomy and more technological innovation is needed to improve its long-term
outcome.
