Keywords
coronary artery bypass grafts surgery - CABG - heart valve surgery - heart valve -
transapical - percutaneous (TAVI) - cardiac catheterization/intervention (incl. PCI)
Introduction
The prevalence of a relevant coronary artery disease (CAD) in patients with severe
aortic stenosis (AS) varies widely between studies depending on definitions and study
design. The German Aortic Valve Registry (GARY), including 140,000 patients who underwent
aortic valve replacement (AVR) or transcatheter aortic valve implantation (TAVI) for
AS, reported a prevalence of CAD of 24.4 and 62.6%, respectively.[1]
[2] This finding relates to similar pathogenesis with degenerative and/or arteriosclerotic
processes on the basis of similar risk factors.[3] Patients with severe AS and relevant CAD can be treated surgically with simultaneous
AVR and coronary artery bypass graft (CABG) (class I; level of evidence: C).[4]
[5] Nevertheless, a combined surgical procedure is associated with higher perioperative
risk compared with an isolated AVR or an isolated CABG due to extended surgical trauma
and prolonged operative and ischemia time. The in-hospital mortality was 2.1% after
AVR and 4.5% after AVR + CABG for patients undergoing conventional surgery in the
GARY registry in 2017.[6] Nonetheless, omission of revascularization during AVR was shown to be the strongest
independent predictor of early death in patients with concomitant relevant CAD.[1]
[7]
A catheter-based approach is an established treatment of severe AS in elderly and
high-risk patients, with a massively expanding spectrum. Furthermore, catheter-based
approach for complex CAD has shown to be feasible and safe in selected high-risk or
inoperable patients.[4]
[8] A complete interventional approach combining percutaneous coronary intervention
(PCI) and TAVI or hybrid procedures could provide an alternative to CABG + AVR. The
available evidence allows a European Society of Cardiology/European Association for
Cardiothoracic Surgery (ESC/EACTS) recommendation of class IIa for PCI + TAVI.[5]
In our analysis, we compared the early outcome of patients treated surgically with
that of patients treated with completely catheter-based approach with severe AS and
concomitant CAD. The aim of our work is to highlight the decision-making process and
management of these patients by the heart team.
Methods
This is a monocentric observational analysis of consecutive patients with AS and CAD.
For the analysis, all patients were identified who were treated for the above two
conditions at our hospital since 2017. We included patients who had undergone concomitant
CABG + AVR for severe symptomatic AS or combined stenosis and insufficiency and relevant
coronary stenosis. For the comparison group, we included all AS + CAD patients who
received a staged interventional therapy with PCI + TAVI, with a maximal time interval
of 6 months between PCI and TAVI. All included patients were discussed at the weekly
interdisciplinary conference at our institution. The clinical condition and comorbidities
of each patient were evaluated with appropriate assessment of the operative risk.
The most appropriate therapeutic procedure for each patient was consensually chosen
by the heart team. An agreement of the ethical committee was exempted due to the retrospective
character of this study. Preoperative data on patient demographics, comorbidities,
and previous cardiac interventions were extracted out of digital patients records
and analyzed. Early patient outcome was examined to identify in-hospital mortality
and complications.
We excluded all patients who had one of the following diagnoses as an indication for
surgery: aortic valve insufficiency without stenosis, endocarditis of the aortic valve,
combined procedures including valve intervention/surgery other than aortic valve (in
mitral, tricuspid, or pulmonary position).
The end points were in-hospital mortality, periprocedural new onset of atrioventricular
block (AVB) II–III°, paravalvular leak (PVL), myocardial infarction (MI), cerebrovascular
stroke, the length of stay in the intensive care unit (ICU), and the total in-hospital
stay.
Severe AS was defined according to the current echocardiographic guidelines.[5] CAD was defined according to the current angiographic guidelines.[9] CAD severity was classified into 1-, 2-, and 3-vessel disease (3-VD). The European
System for Cardiac Operative Risk Evaluation (EuroSCORE) II was used to assess the
operative risk. The corresponding definitions of the Valve Academic Research Consortium
(VARC) criteria were applied to define the above-mentioned end points.[10]
[11] The 30-day mortality included all-cause death during the hospital stay or within
30 days postprocedural. PVL presence and severity was detected with echocardiography
and classified after the European Association of Cardiovascular Imaging (EACVI) guideline
into “none,” “mild,” “moderate,” and “severe.”[11] MI was defined as type 5 MI associated with surgical revascularization, AVR, or
TAVI.[12] Cerebrovascular events (CVEs) were considered to be any new-onset neurologic deficit
of cerebral origin, assessed by a neurologist, in association with signs of hemorrhage
or ischemia on computed tomography/magnetic resonance imaging (CT/MRI) of the brain
that occurred during the primary hospital stay.
All CABG and AVR were performed as combined procedures, whereas TAVI were staged within
6 months after PCI. Surgical management was performed under general anesthesia via
median sternotomy, with CPB, mild hypothermia, and cardioplegic arrest. In this study,
the following conventional prothesis were implanted: Perimount (Carpentier-Edwards
Lifesciences Corp., Irvine, California, United States), Magna Ease (Carpentier-Edwards
Lifesciences Corp., Irvine, California, United States), and Trifecta (Abbott, Valve
with Glide Technology (GT), Allschwil, Switzerland). Rapid deployment prostheses such
as Perceval (LivaNova, London, United Kingdom) and INTUITY (Edwards Lifesciences Corp.,
Irvine, California, United States) were used. Stentless valves were not used in this
work. In young patients (<60 years), mechanical valves comprising Carbomedics (LivaNova,
Saluggia (VC), Italy) and On-X (Life Technologies, Kennesaw, Georgia, United States)
were used.
TAVI procedures were performed in the hybrid operating room in our institution by
the interdisciplinary heart team involving an interventional cardiologist, a cardiac
surgeon, and a cardio-anesthetist. Transfemoral (TF), transapical (TA), or transaxillary
(TAx) access routes were used in this study. The procedure was usually performed under
general anesthesia for TA-TAVI and TAx-TAVI, whereas conscious sedation was usually
enough for TF-TAVI. The predefined access route for the TAVI sheath was then prepared
accordingly either percutaneously or surgically. Balloon valvuloplasty before, through,
and/or after valve delivery was decided according to extension of annulus calcification,
prothesis type, and/or a resulting PVL. In this study, balloon-expandable SAPIEN prostheses
(SAPIEN; SAPIEN XT, SAPIEN 3, Edwards Lifesciences Corporation, Irvine, California,
United States) were used. The following self-expanding prostheses were implanted:
Core Valve prostheses (COREVALVE, COREVALVE Evolut R, and COREVALVE Evolut pro, Medtronic,
Minneapolis, Minnesota, United States); ACURATE and ACURATE neo (formerly Symetis
Inc., Ecublens, Switzerland and now Boston Scientific, Marlborough, Massachusetts,
United States); JENAVALVE (JenaValve Technology Inc., Munich, Germany); and LOTUS
(Boston Scientific, Marlborough, MA, USA).
Statistical Analysis
All data were statistically analyzed using SPSS Statistics version 27.0 (IBM Corporation,
Armonk. New York, United States) and R Statistical Software (V.3.3.2; R Foundation
for Statistical Computing, Vienna. Austria). For nominal variables, the absolute number
(n) was calculated with percentages. For metric variables, the mean with the respective
standard deviation or the median with the 25th and 75th percentiles were calculated
for their description, depending on the variable distribution. Group comparisons were
performed using the Student's t-test or Wilcoxon–Mann–Whitney test for continuous variables and the chi-square test
or Fisher's exact test for categorical variables. Univariate normality assumptions
were verified using the Shapiro–Wilk test. The significance level was set at α = 0.05 and a p-value < 0.05 was thus considered statistically significant.
Propensity Score Matching
Propensity score matching (PSM) was used to minimize differences in demographic characteristics
and potential bias. This was done using logistic regression analysis with inclusion
of the following variables: age, gender, body mass index (BMI), body surface area,
CAD 3-VD, left main stem involvement, diabetes mellitus, hypertension, previous MI,
chronic obstructive pulmonary disease (COPD), calculated creatinine clearance, peripheral
vascular disease, atrial fibrillation (AF), left ventricular ejection fraction (<30%),
and EuroSCORE II. A one-to-one matching algorithm was performed identifying 101 PS-matched
pairs. The logrank test was used to analyze survival rates and graphically presented
in Kaplan–Meier curves.
Results
Of all patients treated since 2017 by TAVI, we identified 242 (17.7%) patients who
underwent PCI within 6 months (median, 43 days; interquartile range, 1–171 days) before
TAVI. After adjusting the exclusion criteria, a total of 237 patients were included
in the PCI + TAVI group. In the surgical group, 517 patients underwent CABG + AVR
during the same period. After adjustment of the exclusion criteria, 241 patients were
included in the CABG + AVR group. All patients received the appropriate revascularization
method deemed feasible by the heart team at the time of TAVI or at the time of CABG + AVR,
with no further planned coronary interventions.
The median age of patients in the CABG + AVR group was 71.9 years (67.7–77.8) at the
time of the operation, while the median age in the PCI + TAVI group was 81.4 (78.7–85.8)
years (p < 0.001). Also, 23.2% in the CABG + AVR group and 40.9% in the PCI + TAVI group were
females. The mean BMI was comparable between the two groups: 28.2 kg/m2 in CABG +AVR and 26.4 kg/m2 in PCI + TAVI (p = 0.100) ([Table 1]).
Table 1
Demographics of both CABG + AVR and PCI + TAVI groups before PSM
|
Preoperative variables
|
CABG + AVR
|
PCI + TAVI
|
p-Value
|
|
Unmatched cohort
|
|
n = 241
|
n = 237
|
|
Age (y)
|
71.9 (67.7–77.8)*
|
81.4 (78.7–85.8)
|
<0.001
|
|
Gender (male), n (%)
|
185 (76.8)
|
140 (59.1)
|
0.198
|
|
BMI (kg/m2)
|
28.2 (24.9–30.5)*
|
26.4 (23.8–28.3)
|
0.100
|
|
BSA (m2)
|
1.9 (±0.18)
|
1.8 (±0.19)
|
0.101
|
|
NYHA, n (%)
|
|
I
|
17 (3.1)
|
33 (13.8)
|
0.118
|
|
II
|
25 (10.3)
|
52 (21.8)
|
|
III
|
158 (66.0)
|
99 (41.4)
|
|
IV
|
41 (17)
|
53 (22.7)
|
|
CCS (III–IV), n (%)
|
28 (11.6)
|
2 (0.8)
|
0.010
|
|
Decompensation
|
22 (9.1)
|
4 (1.7)
|
<0.001
|
|
CAD (number of vessel disease treated)
|
|
1, n (%)
|
48 (19.9)
|
144 (60.7)
|
<0.001
|
|
2, n (%)
|
55 (22.8)
|
30 (12.6)
|
|
3, n (%)
|
131(54.3)
|
52 (21.9)
|
|
Left main disease
|
59 (24.4)
|
34 (14.6)
|
0.05
|
|
ACS, n (%) (unstable AP/STEMI/NSTEMI within 2 wk before index procedure)
|
19 (7.8)
|
0
|
<0.001
|
|
Previous PCI (older than 6 mo before index procedure)
|
37 (15.4)
|
132 (57.6)
|
<0.001
|
|
PCI/DES, n (%) (within 6 mo before index-procedure)
|
47 (19.5)
|
237 (100)
|
<0.001
|
|
Previous cardiac surgery, n (%)
|
0
|
65 (27.2)
|
<0.001
|
|
CABG
|
|
29 (12.2)
|
|
|
AVR
|
|
6 (2.5)
|
|
|
Others
|
|
30 (12.6)
|
|
|
CKD, n (%) (creatinine clearance <50 mL/min)
|
48 (19.9)
|
41 (17.2)
|
0.105
|
|
Peripheral vascular disease, n (%)
|
22 (9.1)
|
53 (22.3)
|
<0.001
|
|
COPD, n (%)
|
50 (21.0)
|
64 (27.0)
|
0.134
|
|
Arterial hypertension, n (%)
|
228 (94.6)
|
220 (92.8)
|
0.456
|
|
Pulmonary hypertension, n (%)
|
6 (2.5)
|
149 (62.8)
|
<0.001
|
|
Diabetes mellitus, n (%)
|
95 (39.4)
|
76 (32.0)
|
0.105
|
|
Atrial fibrillation, n (%)
|
34 (14.1)
|
103 (43.4)
|
<0.001
|
|
LVEF < 30%, n (%)
|
6 (2.4)
|
36 (15.1)
|
<0.001
|
|
EuroSCORE II (%)
|
5.6 (±3.5)
|
7.9 (±5.4)
|
0.059
|
Abbreviations: ACS, acute coronary syndrome; AVR, aortic valve replacement; AP, Angina
Pectoris; BMI, body mass index; BSA, body surface area; CABG, coronary artery bypass
grafting; CAD, coronary artery disease; CCS, Canadian Cardiovascular Society classification
for angina pectoris; CKD, chronic kidney disease; COPD, chronic obstructive pulmonary
disease; DES, drug-eluting stent; LVEF, left ventricular ejection fraction.; NYHA,
New York Heart Association; NSTEMI, Non-ST-elevation myocardil infarction; PCI, percutaneous
coronary intervention; STEMI, ST-elevation myocardial infarction; TAVI, transcatheter
aortic valve implantation.
Note: Depending on variable distribution, metric variables are calculated either as
mean with respective standard deviation (±) or median with 25th and 75th percentiles
([*]). For nominal variables, the absolute number (n) is calculated with percentage (%).
In the CABG + AVR group, 54.3% of patients had 3-VD versus 21.9% in the PCI + TAVI
group. 1-VD was noted in 19.9 and 60.7% in the CABG + AVR and PCI + TAVI groups, respectively
(p < 0.001). Left main stem involvement was documented in 59 (24.4%) versus 34 (14.6%)
in the CABG + AVR group versus PCI + TAVI (p = 0.05). There were no patients with previous cardiac surgery in the CABG + AVR group.
In the PCI + TAVI group, 65 (27.2%) patients had a prior cardiac surgery. Pulmonary
vascular disease (PVD), severely impaired left ventricular function, renal insufficiency,
AF, and pulmonary hypertension (PHT) were more prevalent in the PCI + TAVI group.
The EuroSCORE II was 5.6% (±3.5) in the CABG + AVR group versus 7.9% (±5.4) in the
PCI + TAVI group (p = 0.059).
All TAVI patients in this analysis were elective, whereas 7 (2.9%) patients underwent
emergency and 12 (4.9%) patients underwent urgent surgery for acute coronary syndrome
(ACS). Seventeen (7.05%) patients were hemodynamically unstable with high catecholamine
requirement due to preoperative cardiogenic shock ([Table 2]). Of these, three (1.2%) patients were on intra-aortic balloon pump support and
three (1.2%) patients were on extracorporeal membrane oxygenation support.
Table 2
Procedural data of both groups CABG + AVR and PCI + TAVI before PSM
|
Procedural data
|
CABG + AVR
|
PCI + TAVI
|
p-Value
|
|
Unmatched population
|
|
n = 241
|
n = 237
|
|
Urgency, n (%)
|
|
Elective
|
222 (92.1)
|
237 (100)
|
<0.001
|
|
Urgent
|
12 (4.9)
|
0
|
|
Emergency
|
7 (2.9)
|
0
|
|
Perioperative instability, n (%) (high vasopressors or assist device)
|
17 (7.1)
|
6 (2.5)
|
0.045
|
|
Access route
|
|
Median sternotomy
|
241 (100)
|
–
|
|
|
Transfemoral
|
–
|
214 (90.2)
|
|
|
Transapical
|
–
|
20 (8.4)
|
|
|
Transaxillary
|
–
|
3 (1.3)
|
|
|
Duration (h)
|
3.8 (3.12–7.2)*
|
1.1 (1.0–1.3)
|
<0.001
|
|
Valve prothesis size (mm)
|
23.0 (23–25)*
|
26.0 (24–27)
|
<0.001
|
Abbreviations: AVR, aortic valve replacement; CABG, coronary artery bypass grafting;
PCI, percutaneous coronary intervention; TAVI, transcatheter aortic valve implantation.
Note: Metric variables are calculated as median with 25th and 75th percentiles ([*]). For nominal variables, absolute number (n) is calculated with percentage (%). Assist device: intra-aortic balloon pump support
and/or extracorporeal membrane oxygenation and/or Impella.
Rethoracotomy for relevant hemorrhage or tamponade was necessary in 16 (6.6%) patients
in the CABG + AVR group. In the PCI + TAVI group, thoracotomy was necessary in nine
(3.7%) patients (p < 0.010). Postoperatively, new AVB requiring a permanent pacemaker therapy was significantly
higher after TAVI with 10.5 versus 4.1% after CABG + AVR. A moderate-to-severe PVL
was detected in 16 (6.7%) patients after TAVI (p < 0.001). No PVL was noted in the CABG + AVR group. Patients stayed a median of 2
days longer in the ICU after surgical management (p = 0.010). CVEs were documented in 1.6% after CABG + AVR and 2.9% after TAVI (p = 0.070). All-cause mortality was 2.1% after TAVI versus 7.8% after CABG + AVR (p = 0.012) ([Table 3]).
Table 3
Postoperative outcome of both groups CABG + AVR and PCI + TAVI before PSM
|
Postoperative variables
|
CABG + AVR
|
PCI + TAVI
|
p-Value
|
|
Unmatched population
|
|
n = 241
|
n = 237
|
|
Rethoracotomy or conversion, n (%)
|
16 (6.6)
|
9 (3.7)
|
0.010
|
|
Packed RBCs, unit
|
1.2 (0–4)*
|
0.74 (0–2)
|
<0.001
|
|
Creatinine (mg/dL)
|
1.7 (±0.3)
|
1.3 (±0. 3)
|
<0.525
|
|
CK (U/L)
|
1112.3 (±205.1)
|
151.1 (±23.4)
|
<0.001
|
|
CK-MB (U/L)
|
71.1 (±20.3)
|
19.0 (±2.3)
|
<0.001
|
|
Troponin (ug/L)
|
2.11 (±0.7)
|
0.23 (±0.1)
|
<0.001
|
|
New AVB II–III°, n (%)
|
10 (4.1)
|
25 (10.5)
|
<0.010
|
|
Wound infection, n (%)
|
14 (5.8)
|
0 (0)
|
<0.001
|
|
Major peripheral vascular complications, n (%)
|
7 (2.9)
|
3 (1.3)
|
0.068
|
|
CVEs, n (%)
|
4 (1.6)
|
7 (2.9)
|
0.070
|
|
PVL II–III, n (%)
|
0
|
16 (6.7)
|
<0.001
|
|
ICU stay (d)
|
3 (1–6)*
|
1 (1–4)
|
0.010
|
|
In-hospital stay (d)
|
13 (11–17)*
|
8 (5–23)
|
0.375
|
|
30-d mortality, n (%)
|
19 (7.8)
|
5 (2.1)
|
0.012
|
Abbreviations: AVB, atrioventricular block; AVR, aortic valve replacement; CABG, coronary
artery bypass grafting; CK, creatinine kinase; CVE, cerebrovascular event; ICU, intensive
care unit; PCI, percutaneous coronary intervention; PVL, paravalvular leak; TAVI,
transcatheter aortic valve implantation.
Note: Metric variables are calculated either as mean with respective standard deviation
(±) or median with 25th and 75th percentiles ([*]), depending on variable distribution. For nominal variables, the absolute number
(n) is calculated with percentage (%).
Comparison of Patient Population by PSM
After PSM, 101 matched pairs were identified. The baseline variants were well balanced
between groups ([Table 4]). The prevalence of AF, arterial hypertension, and diabetes mellitus was comparable
in both groups. Nonetheless, significantly higher prevalence of PVD, chronic kidney
disease, COPD, and PHT was noticed in the PCI + TAVI group. The mean operative risk
assessed with EuroSCORE II was 6.2% (±2.7) versus 7.6% (±5.1) for CABG + AVR versus
PCI + TAVI, respectively (p = 0.279). Emergency surgery patients were excluded. However, 9 (8.9%) patients underwent
CABG + AVR urgently ([Table 5]).
Table 4
Preoperative data of CABG + AVR and PCI + TAVI after PSM
|
Preoperative variables
|
CABG + AVR
|
PCI + TAVI
|
p-Value
|
|
Matched population
|
|
n = 101
|
n = 101
|
|
Age (y)
|
77.2 (74.8–80.2)*
|
78.5 (74.5–82.0)
|
0.141
|
|
Gender (male), n (%)
|
69 (68.3)
|
69 (68.3)
|
1.000
|
|
BMI (kg/m2)
|
27.1 (23.8–29.4)*
|
26.7 (24.2–29.7)
|
0.867
|
|
BSA (m2)
|
1.9 (±0.18)
|
1.8 (±0.19)
|
0.318
|
|
NYHA, n (%)
|
|
I
|
2 (2.0)
|
14 (13.8)
|
0.295
|
|
II
|
10 (9.9)
|
22 (21.8)
|
|
III
|
68 (67.3)
|
42 (41.4)
|
|
IV
|
17 (16.8)
|
23 (22.7)
|
|
CCS (III–IV), n (%)
|
9 (8.9)
|
7 (6.9)
|
0.093
|
|
Decompensation, n (%)
|
9 (8.9)
|
11 (10.8)
|
0.785
|
|
CAD (number of vessel disease treated)
|
|
1, n (%)
|
23 (22.8)
|
48 (47.5)
|
1.000
|
|
2, n (%)
|
35 (34.7)
|
10 (9.9)
|
|
3, n (%)
|
39 (38.6)
|
39 (38.6)
|
|
Left main disease, n (%)
|
37 (36.6)
|
29 (28.7)
|
0.187
|
|
ACS, n (%) (unstable AP/STEMI/NSTEMI within 2 wk before the index procedure)
|
13 (12.8)
|
0
|
<0.001
|
|
Recent PCI, n (%)
(within 6 mo before the index procedure)
|
20 (19.8)
|
101 (100)
|
<0.001
|
|
Previous cardiac surgery, n (%)
CABG
AVR
Other procedures
|
0
|
40 (39.6)
40 (100)
0 (0)
11 (27.5)
|
<0.001
|
|
CKD, n (%) (creatinine clearance < 50 mL/min)
|
17 (16.8)
|
24 (23.8)
|
<0.040
|
|
Peripheral vascular disease, n (%)
|
22 (9.1)
|
53 (22.3)
|
<0.001
|
|
COPD, n (%)
|
17 (16.8)
|
30 (29.7)
|
0.045
|
|
Arterial hypertension, n (%)
|
94 (93.1)
|
95 (94.1)
|
1.000
|
|
Pulmonary hypertension, n (%)
|
1 (0.9)
|
57 (56.4)
|
<0.001
|
|
Diabetes mellitus, n (%)
|
37 (36.6)
|
33 (32.7)
|
0.658
|
|
Atrial fibrillation, n (%)
|
28 (27.7)
|
26 (25.7)
|
0.874
|
|
LVEF < 30%, n (%)
|
4 (3.9)
|
7 (6.9)
|
0.537
|
|
EuroSCORE II (%)
|
6.2 (±2.7)
|
7.6 (±5.1)
|
0.279
|
Abbreviations: ACS, acute coronary syndrome; AVR, surgical aortic valve replacement;
BMI, body mass index; BSA, body surface area. NYHA, New York Heart Association; CABG,
aortocoronary bypass surgery; CCS, Canadian Cardiovascular Society classification
for angina pectoris; CKD, chronic kidney disease; COPD, chronic obstructive pulmonary
disease; DM, diabetes mellitus; LVEF, left ventricular ejection fraction; PCI, percutaneous
coronary intervention; PHT, pulmonary hypertension; TAVI, transcatheter aortic valve
implantation; VHF, atrial fibrillation.
Note: Metric variables are calculated either as mean with respective standard deviation
(±) or median with 25th and 75th percentiles ([*]), depending on variable distribution. For nominal variables, the absolute number
(n) is calculated with percentage (%).
Table 5
Procedural data of both groups CABG + AVR und PCI + TAVI after PSM
|
Procedural data
|
CABG + AVR
|
PCI + TAVI
|
p-Value
|
|
Matched population
|
|
n = 101
|
n = 101
|
|
Urgency, n (%)
|
|
Elective
|
91 (90.1)
|
101 (100)
|
<0.055
|
|
Urgent
|
9 (8.9)
|
0
|
|
Emergency
|
0 (0.0)
|
0
|
|
Access route, n (%)
|
|
Median sternotomy
|
101 (100)
|
–
|
|
|
Transfemoral
|
–
|
88 (87.1)
|
|
|
Transapical
|
–
|
11 (10.9)
|
|
|
Transaxillary
|
–
|
2 (2.0)
|
|
|
Duration (h)
|
3.8 (3.1–4.1)*
|
1.1 (1.0–1.3)
|
<0.001
|
|
Size of valve prothesis (mm)
|
23.0 (23–25)*
|
26.0 (24–27.5)
|
<0.001
|
|
Valve prothesis type, n (%)
|
|
Conventional bioprothesis
|
62 (61.3)
|
–
|
|
|
Rapid deployment valves
|
36 (35.6)
|
–
|
|
|
Balloon expandable
|
–
|
74 (73.3)
|
|
|
Self-expandable
|
–
|
27 (26.7)
|
|
AVR, surgical aortic valve replacement; CABG, aortocoronary bypass surgery; N/A, not
applicable; PCI, percutaneous coronary intervention; TAVI, transcatheter aortic valve
implantation.
Note: Metric variables are calculated as median with 25th and 75th percentiles ([*]). For nominal variables, absolute number (n) is calculated with percentage (%).
In the CABG + AVR group, 36 rapid-deployment valves (16 Edwards Intuity and 20 Sorin
Perceval), 55 Perimount CE Pericardial prostheses, 4 Magna Ease, 3 Trifecta, 2 Carbomedics,
and 1 On-X aortic valve prosthesis were implanted. In the matched TAVI cohort, 74
balloon-expandable valves (65 S3 and 9 Edwards XT) and 27 self-expandable prostheses
(19 Symetis, 7 Core Valve, and 1 Lotus) were implanted.
The 30-day all-cause mortality was 5 (4.9%) after CABG + AVR and 1 (1.0%) after PCI + TAVI
(p = 0.099) and after Fischer's exact test (p = 0.212). [Figs. 1] and [2] show the 30-day survival before and after PSM in Kaplan–Meier curves. Two male patients
in the CABG + AVR died of cardiogenic shock, one patient died of hemorrhagic shock,
one patient died of sepsis and multiple organ failure, and one female patient died
due to massive cerebral ischemia. Four of the five deceased patients in the CABG + AVR
were over 80 years old.
Fig. 1 Kaplan–Meier curve for 30-day survival before matching.
Fig. 2 Kaplan–Meier curve of overall survival after CABG + AVR vs PCI + TAVI after PSM.
Rethoracotomy due to massive bleeding or tamponade was performed in eight (7.9%) cases
in the CABG + AVR group. In the PCI + TAVI group, an unplanned sternotomy with CPB
was emergently performed in two (2.0%) cases because of pericardial tamponade due
to left ventricular perforation and valve dislocation. One patient survived the emergency
procedure and could be discharged home.
Cerebral events were comparable in both groups with 2.0% after CABG + AVR and 3.0%
after PCI + TAVI (p = 1.00). Cardiac biomarkers were significantly higher after CABG + AVR. Postprocedural,
new permanent pacemaker implantation was significantly higher after PCI + TAVI with
6.9% than after CABG + AVR with 4.1% (p = 0.010). A relevant PVL II–III° was diagnosed in five (4.9%) patients after TAVI
(p < 0.027). There was no detected PVL after CABG + AVR. Although TAVI patients received
significantly larger valves, there was no difference in PPM after matching (p < 0.872). Other postoperative complications such as major peripheral vascular complications,
periprocedural acute renal failure, and dialysis showed no significant statistical
difference between both groups. Complications related to sternotomy including wound
dehiscence or mediastinitis were noted in four (3.9%) patients in the CABG + AVR group.
The length of ICU stay after PSM was comparable between groups (p = 0.066). However, the overall length of hospital stay remained longer after CABG + AVR
(p = 0.001). Postoperative outcomes after PSM are summarized in [Table 6].
Table 6
Postoperative outcome for CABG + AVR vs PCI + TAVI after PSM
|
Postoperative variables
|
CABG + AVR
|
PCI + TAVI
|
p-Value
|
|
Matched population
|
|
n = 101
|
n = 101
|
|
Rethoracotomy or conversion, n (%)
|
8 (7.9)
|
2 (2.0)
|
0.035
|
|
PRBCs, unit
|
1.3 (0–4)*
|
0.72 (0–2)
|
0.026
|
|
Creatinine clearance < 50 mL/min, n (%)
|
42 (41.5)
|
31 (30.6)
|
0.084
|
|
Dialysis postoperative, n (%)
|
1 (1.0)
|
3 (3.0)
|
0.100
|
|
CK post-op (U/L) (normal <190 U/L)
|
977.4 (±96.7)
|
186.5 (±15.4)
|
<0.001
|
|
CK-MB post-op (U/L) (normal < 25 U/L)
|
73.7 (±11.2)
|
21.7 (±7.5)
|
<0.001
|
|
Troponin post-op (ug/L) (normal < 0.014 ug/L)
|
1.72 (±0.3)
|
0.25 (±0.2)
|
<0.001
|
|
New AVB II–III°, n (%)
|
4 (4.1)
|
7 (6.9)
|
0.010
|
|
Wound infections, n (%)
|
4 (4.0)
|
0 (0.0)
|
0.021
|
|
Major peripheral vascular complications, n (%)
|
1 (1.0)
|
3 (3.0)
|
0.100
|
|
CVEs, n (%)
|
2 (2.0)
|
3 (3.0)
|
1.000
|
|
PVL II–III, n (%)
|
0
|
5 (4.9)
|
0.027
|
|
Moderate PPM (iAVA < 0.85 cm2/m2), n (%)
|
31 (31)
|
30 (29)
|
1.000
|
|
Severe PPM (iAVA < 0.65 cm2/m2), n (%)
|
10 (9.9)
|
9 (8.9)
|
0.872
|
|
ICU stay (d)
|
3.7 (1–6)*
|
3.1 (1–4)
|
0.066
|
|
In-hospital stay (d)
|
14.4 (11–17)*
|
9.4 (5–23)
|
<0.001
|
|
30-d mortality, n (%) (with Fischer's exact test)
|
5 (4.9)
|
1 (1.0)
|
0.099
(0.212)
|
Abbreviations: AVR: surgical aortic valve replacement. PCI: percutaneous coronary
intervention. TAVI: transcatheter aortic valve implantation. CK: creatinine kinase.
CVE: cerebrovascular event. PVL: paravalvular leak. AVB: AV block. PPM: patient-prosthesis
mismatch. ICU: intensive care unit.
Metric variables are calculated either as mean with respective standard deviation
(±) or median with 25th and 75th percentiles ([*]). depending on variable distribution. For nominal variables, the absolute number
(n) is calculated with percentage (%). CABG: aortocoronary bypass surgery.
Discussion
Variable treatment options are available for patients presenting with severe AS and
relevant CAD, including open surgical, minimally invasive, interventional, and hybrid
procedures. Individual operative risk assessment and anatomical conditions guide decision
making.[13]
[14] Combined surgical procedure comprising of AVR + CABG is associated with higher perioperative
risk compared with an isolated AVR or an isolated CABG, with an in-hospital mortality
of 4.5 versus 2.1% respectively.[6] With the steadily increasing accessibility of catheter-based techniques, a staged
interventional approach in terms of PCI and TAVI could minimize the operative/interventional
burden.[15] Nonetheless, a meta-analysis of patients who underwent a concomitant PCI + TAVI
reported a significantly higher 30-day mortality rate compared with patients who underwent
an isolated TAVI.[16] Moreover, the currently available evidence to support a staged interventional treatment
strategy relies on two meta-analyses by Kotronias et al and Tarus et al, which were
mainly based on retrospective and monocentric studies that reported unadjusted and
inconsistent results.[16]
[17] The SURTAVI trial is the only available randomized controlled trial (RCT) for patients
with AS + CAD. The authors reported no difference in the end point outcome after CABG + AVR
or PCI + TAVI at 2 years.[18] However, the included patients did not have a complex CAD with a mean SYNTAX score
of 8.1 and 8.5 in both groups, respectively. Thus, a generalization for more complex
CAD patients is not justifiable. Therefore, the role of the heart team in patient/procedure
coupling is pertinent.[5] An individual evaluation for each patient considering underlying anatomical features
of the coronaries and the aortic valve annulus as well as other bystander cardiac
or systemic conditions is recommended.[5] Moreover, outcome aspects such as completeness of revascularization, durability
of the valve prosthesis, and the expected complication spectrum of each procedure
have to be taken into consideration in the decision-making process.
In the present work, the PCI + TAVI cohort was significantly older, with higher operative
risk and frequent previous cardiac surgery. Similar distribution of demographics and
risk profiles for surgically or interventionally treated patients with AS and CAD
was reported in numerous studies.[16]
[17]
[19] A prior cardiac surgery per se is not a contraindication to surgical treatment and
was not considered as an exclusion criterion in our analysis. Nevertheless, those
patients were more likely to undergo TAVI in consensus of the heart team due to the
considerably higher operative risk.
We refer the inverse proportion of 1- or 3-VD in both groups before PSM to the heart
team decision made by the time of treatment assignment. Obviously, surgical revascularization
was more often recommended in cases of 3-VD (54.3%) or left main stem involvement
(24.4%). On the contrary, a percutaneous approach was performed more often in less
complex CAD, with 60.7% of patients having had 1-VD . This variation was balanced
after PSM with 38.6% of patients having had 3-VD in both groups (p = 1.00). Left main involvement was adjusted as well, after matching, with 36.6 and
28.7% in the CABG + AVR and PCI + TAVI group, respectively (p = 0.187). Further, SYNTAX score–based or functional-based assessment of CAD severity
was not conducted in this analysis.
In our analysis, the sequence of PCI and TAVI was discussed individually depending
on the patient's clinical condition, severity of CAD, and risk of myocardial ischemia.
Available data seem to indicate that PCI is feasible both before and together with
TAVI without impact on early survival. However, PCI weeks to months before TAVI is
currently the most established strategy with dividing of the interventional burden
and easier cannulation of coronary ostia before TAVI. Nevertheless, the ACTIVATION
study showed that a routine PCI strategy in patients with obstructive CAD before TAVI
is not clinically beneficial.[20] Furthermore, it is associated with higher bleeding rates with dual antiplatelet
therapy and a propensity for more frequent acute renal failure.[20] Regardless, a recent study reported technical success in 46 cases of coronary angiography,
including 26 PCI, performed immediately after implantation of self-expandable CoreValve
Medtronic prostheses within the same operative session without additional procedural
complexity.[21]
In our analysis, patients in the CABG + AVR had more frequently ACS preoperatively.
As in the PCI + TAVI group, PCI was already performed before the index intervention
(TAVI). Most patients had, by definition, no longer ACS at the time of the index intervention,
even if ACS was diagnosed at the time of PCI.
In our analysis, 30-day all-cause mortality after PSM was 4.9% in the CABG + AVR group
and 1.0% in the PCI + TAVI group (p = 0.212) ([Fig. 2]). Our results are comparable with the national average of 30-day mortality risk
after CABG + AVR of 4.5%; however, in the GARY register, patients were younger with
a mean age of 73 years.[6] In an analysis from the U.S. National Registry for aortic valve procedures with
a mean age of 73 years, the authors reported 5.1% incidence of in-hospital mortality
after CABG + AVR. This incidence increased with age, with 7.4% in patients older than
85 years.[22] On the other hand, the GARY registry showed in 2015 an in-hospital mortality of
2.3% after TF-TAVI and 3.7% after TA-TAVI in a high-risk cohort with an average age
of 80 years.[6] A more recent evaluation in 2020 showed a significant reduction in TAVI mortality
at average of all access routes of 1.6%.[2] Furthermore, a recent meta-analysis showed an advantage in favor of PCI + TAVI regarding
the 30-day mortality incidence with 5.2 versus 7.5% after combined surgical approach,
without statistical significance.[16] Nonetheless, the initial advantage of PCI + TAVI group equaled out in the 2-year
mortality analysis compared with CABG + AVR. The small number of included studies
limited this meta-analysis.
In the CABG + AVR group, 7.9% of patients underwent rethoracotomy due to major hemorrhage
or pericardial tamponade with subsequently significant more blood transfusion than
in the PCI + TAVI group. A single-center analysis by Baumbach et al showed a comparable
incidence of rethoracotomy of 7.8% after CABG + AVR.[3] The incidence of surgical conversion during TAVI procedures was 0.7%, as published
in the National German Register.[2]
[6] Despite the fact that serious complications during PCI or TAVI procedures are rare
with an overall incidence of 0.4 to 1.3%, they are associated with a higher mortality
of more than 50%.[23] Our experience supports the current ESC/EACTS guideline for valve disease insisting
on the presence of a cardiac surgery department “on-site” as a prerequisite for performing
TAVI (class I; evidence level: C).[5]
The rate of CVEs at 30 days was comparable between groups in our analysis, with 2%
in the CABG + AVR group and 3% in the PCI + TAVI group (p = 1.00). In an analysis from the U.S. National Registry in 2015, a CVE incidence
of 2.9% was reported after CABG + AVR perioperatively.[22] The risk increased with age, with an incidence of 1.3% in patients younger than70
years and 3.2% in those older than 85 years. The incidence of CVEs after TAVI ranged
from 2.7 to 5.5% at 30 days in most TAVI RCTs.[13]
[14]
[15]
In our study, the absolute values of cardiac biomarkers were significantly higher
postoperatively in the CABG + AVR group than in the PCI + TAVI. However, after applying
the definition of type 5 MI, there was no clear evidence of more frequent MI postoperatively
in the CABG + AVR compared with PCI + TAVI.[12] There were no coronary obstructions or repeat revascularization in the matched cohort
at 30 days. In a recent multicenter analysis of AS patients with complex CAD (156
pairs after PSM), Alperi et al reported a comparable outcome of major adverse cardiac
and cerebrovascular events (MACCE) at 3 years.[19] However, the authors further reported a significantly higher incidence of repeat
revascularization in the PCI + TAVI group.
We observed moderate-to-severe PVL in 4.9% of cases in the PCI + TAVI group. No PVL
was reported in the matched CABG + AVR cohort. Moderate-to-severe PVL occurs in 0.9%
of conventional AVR and in approximately 6.4% of TAVI procedures.[23] Furthermore, the 1-year report of the PERSIST trial showed a similar low incidence
of PVL using perceval valves compared with conventional prostheses in AVR.[24] Several studies report that moderate and severe PVL are associated with 2- to 12-fold
increase in 1-year mortality.[25] The PARTNER II study reported a significantly higher mortality among patients with
moderate or severe PVL (64.8%) than among patients with mild PVL (48.7%) or minor
PVL (41.1%) after TAVI with balloon expandable valves.[26] Despite reduction of its incidence over the past decade, PVL still represents one
of the major limitations of the TAVI procedure.
Conduction disturbances with subsequent pacemaker dependency is the most procedure-associated
complication after both AVR and TAVI with a considerably higher risk after TAVI.[23] Moreover, new pacemaker implantation was independently associated with 1-year mortality
after TAVI. Furthermore, the development of a pacemaker dependency postinterventional
is associated with longer hospitalization of patients and higher costs as well as
a relative reduction in the quality of life.[27] In our analysis, we noted a pacemaker dependency of 4.0% after CABG + AVR and 6.9%
after PCI + TAVI at 30 days (p = 0.010). Taking into consideration that more than one-third of patients in the CABG + AVR
group received a rapid-deployment prosthesis, this may be considered a very good outcome.
A retrospective analysis showed a higher risk for AVB postoperatively after using
rapid deployment prostheses with 8.8% compared to conventional bioprosthesis with
3.7%, without survival benefit.[28] Furthermore, the 1-year report of the PERSIST trial mentioned a higher incidence
of new pacemaker implantation after perceval of 11.1% versus 3.6% compared with conventional
bioprothesis.[24] On the other hand, the rate of AVB after TAVI varies between 3.5 and 28.6% for currently
available prosthesis models (Sapien 3: 12.4%; Evolut R: 17.5% Acurate neo: 9.9%).[23] The GARY registry showed a rate of AVB across all TAVI prosthesis between 2011 and
2015 of 16.6%.[6] This was significantly higher compared with conventional bioprothesis with 3.5%
over the same period. Nevertheless, with establishment of new generations TAVI prothesis
and a better recognition of the predictors of AVB, the risk is supposedly going to
continue to decrease.
Patients who underwent CABG + AVR stayed longer in hospital with a mean of 14.4 (11–17)
versus 9.4 (5–23) days after PCI + TAVI (p < 0.001). However, the last refers to the length of stay for TAVI without adding
that for the corresponding previously performed PCI. A recent meta-analysis showed
a mean length of stay of 12.1 (±6.7) and 8.35 (±5.95) in CABG + AVR and PCI + TAVI,
respectively.[17]
Limitations
The present study is limited due to its retrospective design. The collected data originate
from an unselected patient collective from a single center. This may have resulted
in a referral or selection bias.
Furthermore, the collective size was limited for the evaluation of some relatively
rare outcomes such as CVEs or MI. Therefore, a type II statistical error cannot be
excluded in this regard. A treatment by means of hybrid approach was excluded in this
analysis.
Based on the profound differences of both groups, even matching for COPD and PHT did
not result in an equal distribution among treatment groups. This imbalance was finally
accepted balancing the number of patients to be analyzed. Post hoc case review of
all included individuals did not show a relevant influence of COPD and PHT on the
final conclusion.
Conclusion
Our analysis showed a short-term survival advantage in favor of PCI + TAVI compared
with CABG + AVR for management of CAD and AS. However, PCI + TAVI provide a significantly
higher risk of AVB and PVL. Long-term results including data on revascularization
incidence, durability of TAVI prostheses, and impact of PVL on long-term survival
are necessary to guide decision making by the heart team.
