The Impact of Calcium Volume and Distribution in Aortic Root Injury Related to Balloon-Expandable Transcatheter Aortic Valve Replacement

Page 1

Q6

Q2

ww.sciencedirect.com

123456789

1011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465

J o u rn a l o f C a r d i o v a s c u l a r C om p u t e d T omog r a p h y x x ( 2 0 1 5 ) 1e1 1

JCCT792_proof ■ 12 May 2015 ■ 1/11

Available online at w

ScienceDirect

journal homepage: www.JournalofCardiovascularCT.com

66676869707172737475

Original Research Article

76777879808182

The impact of calcium volume and distributionin aortic root injury related to balloon-expandabletranscatheter aortic valve replacement

8384858687888990919293949596979899

100101102103104105106107108109110111112113114115116117118119120121

Nicolaj C. Hansson MDa,*, Bjarne L. Nørgaard MD, PhDa,Marco Barbanti MDb,c, Niels Erik Nielsen MDd, Tae-Hyun Yang MDb,e,Corrado Tamburino MDc, Danny Dvir MDb, Hasan Jilaihawi MDf,Phillip Blanke MDg, Raj R. Makkar MDf, Azeem Latib MDh,Antonio Colombo MDh, Giuseppe Tarantini MDi, Rekha Raju MDb,David Wood MDb, Henning R. Andersen MD, DMScia,Henrique B. Ribeiro MDj, Samir Kapadia MDk, James Min MDf,Gudrun Feuchtner MDl, Ronen Gurvitch MDm, Faisal Alqoofi MDn,Marc Pelletier MDo, Gian Paolo Ussia MDp, Massimo Napodano MDh,Fabio Sandoli de Brito Jr. MDq, Susheel Kodali MDr, Gregor Pache MDg,Sergio J. Canovas MDs, Adam Berger MDb, Darra Murphy MDb,Lars G. Svensson MDk, Josep Rodes-Cabau MDj, Martin B. Leon MDr,John G. Webb MDb, Jonathon Leipsic MDb

aDepartment of Cardiology, Aarhus University Hospital, Skejby, Brendstrupgaardsvej 100, DK-8200 Aarhus N,

Aarhus, Denmarkb St Paul’s Hospital, Vancouver, Canadac Ferrarotto Hospital, University of Catania, Italyd Linkoping University Hospital, Linkoping, SwedeneBusan Paik Hospital, Inje University, South KoreafCedars-Sinai Heart Institute, Los Angeles, CA, USAgUniversity Hospital Freiburg, GermanyhEMO-GVM Centro Cuore Columbus and San Raffaele Scientific Institute, Milan, ItalyiUniversity of Padua, Padua, ItalyjQuebec Heart and Lung Institute, Laval University, Quebec City, CanadakCleveland Clinic, Cleveland, OH, USAl Innsbruck Medical University, Innsbruck, AustriamRoyal Melbourne Hospital, Melbourne, AustralianUniversity of Calgary, Calgary, CanadaoNew Brunswick Heart Centre, Saint John, CanadapDepartment of Cardiovascular Disease, Tor Vergata University of Rome, Italy

Conflict of interest: David Wood, Hasan Jilaihawi, Susheel Kodali, Josep Rodes-Cabau, Martin B. Leon, John G. Webb, and JonathonLeipsic are consultants to Edwards Lifesciences. Gian Paolo Ussia and Azeem Latib are consultants to Medtronic Inc. Nicolaj C. Hanssonand Bjarne L. Nørgaard have received unrestricted grant support from Edwards Lifesciences. Nicolaj C. Hansson has also receivedunrestricted grant support from the Danish Heart Foundation. The other authors report that they have no conflicts of interest.* Corresponding author.E-mail address: [email protected] (N.C. Hansson).

122123124125126127128129130

1934-5925/$ e see front matter ª 2015 Society of Cardiovascular Computed Tomography. All rights reserved.http://dx.doi.org/10.1016/j.jcct.2015.04.002

Page 2

Q3

J o u r n a l o f C a r d i o v a s c u l a r Com p u t e d T omog r a p h y x x ( 2 0 1 5 ) 1e1 12

131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195

196197

JCCT792_proof ■ 12 May 2015 ■ 2/11

qHospital Israelita Albert Einstein, Sao Paulo, BrazilrColumbia University Medical Center/New York-Presbyterian Hospital, New York, NY, USAsArrixaca University Hospital, Murcia, Spain

198199

200201202203204205206207208209210211212213214215216217

a r t i c l e i n f o

Article history:

Received 17 November 2014

Received in revised form

24 February 2015

Accepted 4 April 2015

Available online xxx

Keywords:

Aortic root calcification

Aortic root injury

Multidetector computed

tomography

Transcatheter aortic valve

replacement

218219220221222223224225226227228229230231

a b s t r a c t

Background: A detailed assessment of calcium within the aortic root may provide important

additional information regarding the risk of aortic root injury during transcatheter heart

valve replacement (TAVR).

Objective: We sought to delineate the effect of calcium volume and distribution on aortic

root injury during TAVR.

Methods: Thirty-three patients experiencing aortic root injury during TAVR with a balloon-

expandable valve were compared with a control group of consecutive 153 TAVR patients

without aortic root injury (as assessed by post-TAVR multidetector CT). Using a commer-

cial software product to analyze contrast-enhanced pre-TAVR CT scans, calcium volume

was determined in 3 regions: (1) the overall left ventricular outflow tract (LVOT), extending

10 mm down from the aortic annulus plane; (2) the upper LVOT, extending 2 mm down

from the annulus plane; and (3) the aortic valve region.

Results: Calcium volumes in the upper LVOT (median, 29 vs 0 mm3; P < .0001) and overall

LVOT (median, 74 vs 3 mm3; P ¼ .0001) were higher in patients who experienced aortic root

injury compared with the control group. Calcium in the aortic valve region did not differ

between groups. Upper LVOT calcium volume was more predictive of aortic root injury

than overall LVOT calcium volume (area under receiver operating curve [AUC], 0.78; 95%

confidence interval, 0.69e0.86 vs AUC, 0.71; 95% confidence interval, 0.62e0.82; P ¼ .010).

Upper LVOT calcium below the noncoronary cusp was significantly more predictive of

aortic root injury compared to calcium underneath the right coronary cusp or the left

coronary cusp (AUC, 0.81 vs 0.68 vs 0.64). Prosthesis oversizing >20% (likelihood ratio test,

P ¼ .028) and redilatation (likelihood ratio test, P ¼ .015) improved prediction of aortic root

injury by upper LVOT calcium volume.

Conclusion: Calcification of the LVOT, especially in the upper LVOT, located below the

noncoronary cusp and extending from the annular region, is predictive of aortic root injury

during TAVR with a balloon-expandable valve.

ª 2015 Society of Cardiovascular Computed Tomography. All rights reserved.

232

233234 1. Introduction ventricular outflow tract (LVOT) calcification, respectively, has 235236237238239240241242243244245246247248249250251252253254255256257258259260

Transcatheter aortic valve replacement (TAVR) or trans-

catheter aortic valve implantation has become a well-

established treatment for symptomatic severe aortic

stenosis in patients with high or prohibitive surgical risk.1,2

Contrast-enhanced multidetector CT (MDCT) imaging has

emerged to play a key role in pre-TAVR screening. MDCT

evaluation of annular dimensions and other aortic root fea-

tures such as the presence of calcification, sinus of Valsalva

diameter, and coronary ostia height has been shown to have

important implications for transcatheter heart valve (THV)

size selection.3 Device selection has been reliant on the need

for THV oversizing in a controlled and predictable fashion in

order to reduce the risk of paravalvular regurgitation without

increasing the risk of aortic root injury.4e7

Potential future broadening of TAVR indications calls for

an increasingly safe procedure. Although rare, aortic root

injury related to balloon-expandable TAVR remains a feared

complication with poor clinical outcome.7 By evaluating pre-

TAVR MDCT scans in patients experiencing aortic root

injury, an association between the risk of aortic root injury in

balloon-expandable TAVR and THV oversizing >20% relative

to the MDCT annular area and moderate or severe left

been demonstrated.7 To date, the evaluation of LVOT calcifi-

cation has been limited to largely qualitative assessment

limiting the discriminatory ability of preprocedural MDCT. It

has been suggested that a detailed assessment of the distri-

bution of calcium within the aortic root and LVOT may pro-

vide important additional information regarding the risk of

aortic root injury. However, historically, such analysis has

been challenging because of the lack of uniform calcium

quantification methods and the complex 3-dimensional

nature of the aortic root.

The aim of this study was to build on the existing knowl-

edge regarding the role of LVOT calcification in aortic root

injury and to evaluate whether volume and particularly loca-

tion of calcium play a role, beyond qualitative assessment, for

the risk of aortic root injury during balloon-expandable TAVR.

2. Methods

2.1. Study population

Thirty-three patients who underwent pre-TAVR MDCT and

experienced contained or uncontained aortic root injury

Page 3

web4C=FPO

J o u rn a l o f C a r d i o v a s c u l a r C om p u t e d T omog r a p h y x x ( 2 0 1 5 ) 1e1 1 3

261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325

326327328329330331332333334335336337338339340341342343344345

JCCT792_proof ■ 12 May 2015 ■ 3/11

related to TAVR between May 2010 and February 2013 were

collected from 16 centers worldwide. This cohort of patients

has previously been described in detail.7

The control group consisted of 153 consecutive TAVR

patients without aortic root injury who underwent pre-TAVR

MDCT and post-TAVR MDCT (confirming the absence of

aortic root injury) at 1 center (Aarhus University Hospital,

Skejby, Denmark) between June 2011 and December 2013.

In all patients, TAVR was performed with the balloon-

expandable SAPIEN or SAPIEN XT prosthesis (Edwards Life-

sciences; Irvine, CA) using standard implantation techniques.7

2.2. MDCT image acquisition

All patients underwent ECG-gated, contrast-enhanced MDCT

before TAVR. MDCT protocols were applied according to each

center’s standard practice. MDCT examinations were per-

formed on either a 64-slice Discovery HD 750 high-definition

or volume CT scanner (GE Healthcare; Milwaukee, WI), a

Fig. 1 e Multidetector CT aortic root dimensions. (A) Stretched M

Short-axis view of the sinotubular junction level. (C) Short-axis v

aortic annulus level. (E) Short-axis view of the LVOT level 5 mm

distance from the aortic annulus to the right coronary ostium.

annulus to the left coronary ostium. LVOT, left ventricular outfl

Siemens first-generation or second-generation dual-source

scanner (Siemens Healthcare; Erlangen, Germany), or a

Toshiba Aquilion ONE 320 row scanner (Toshiba Medical Sys-

tems; Tokyo, Japan). Patients in the control group additionally

had contrast-enhanced MDCT performed 1 to 3 months

post-TAVR as part of a standard institutional protocol.8

2.3. MDCT aortic root dimensions

Images were postprocessed offline at a workstation using

dedicated TAVR planning software (3mensio Structural Heart;

3mensio Medical Imaging BV, Bilthoven, the Netherlands).

High reproducibility and reliability of this software have been

demonstrated.9,10 After automatic reconstruction and seg-

mentation of the aortic root, the aortic annulus was manually

identified as the plane immediately below the nadir of the

aortic cusps. The software then provided perpendicular

short-axis views along a centerline from the LVOT into the

ascending aorta (Fig. 1). The centerline was manually

PR view of the aortic root demonstrating analysis levels. (B)

iew of the sinus of Valsalva level. (D) Short-axis view of the

below the aortic annulus. (F) Stretched MPR displaying the

(G) Stretched MPR displaying the distance from the aortic

ow tract; MPR, multiplanar reformation.

346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390

Page 4

eb4C=FPO

J o u r n a l o f C a r d i o v a s c u l a r Com p u t e d T omog r a p h y x x ( 2 0 1 5 ) 1e1 14

391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455

456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486

JCCT792_proof ■ 12 May 2015 ■ 4/11

adjusted, if necessary. LVOT dimensions were measured in a

short-axis view perpendicular to the centerline 5 mm below

the aortic annulus. This distance was chosen based on pre-

viously reported distances between the aortic annulus and the

lower rim of the Edwards THV.10e12 Sinus of Valsalva dimen-

sionsweremeasured in a short-axis view perpendicular to the

centerline at the level of the largest sinus of Valsalva area. The

sinotubular junction dimensions were measured in a short-

axis view perpendicular to the centerline at the distal end of

the sinus of Valsalva. LVOT, annular, and sinotubular junction

dimensions were assessed using the short- and long-axis

diameters as well as area by planimetry (Fig. 1B,D,E). For

the sinus of Valsalva, area by planimetry and the 3 cusp

commissure diameters were measured (Fig. 1C). Sinotubular

junction height (distance from the aortic annulus to the

sinotubular junction) and coronary ostia heights were deter-

mined on stretched multiplanar reformation views (Fig. 1F,G).

The degree of THV oversizing (positive percentage) or

undersizing (negative percentage) was calculated as (THV

nominal area/MDCT annular area � 1) � 100, where THV

nominal areas are 4.15 cm2 (23-mm THV), 5.31 cm2 (26-mm

THV), and 6.61 cm2 (29-mm THV), and MDCT annular area

was measured by planimetry. Mean diameter was calculated

as (short-axis diameter þ long-axis diameter)/2. Eccentricity

indexes for the LVOT, aortic annulus, and sinotubular junc-

tion were calculated as 1 � (short-axis diameter/long-axis

diameter). To evaluate the mismatch of dimensions within

the aortic root, ratios between LVOT, annulus, sinus of Val-

salva, and sinotubular junction areas were determined.

w

Fig. 2 e A representative example of an image histogram of

the normal blood pool region. (A) A 200 m3 VOI is placed in

the uniform normal blood pool region at the level of the left

coronary ostium. (B) A Gaussian curve ( green) is fitted to

the image histogram (brown) of the VOI, and the patient-

specific calcium detection threshold is calculated based on

meanfit and SDfit (see Methods section). HU, Hounsfield

units; VOI, volume of interest. (For interpretation of the

references to color in this figure legend, the reader is

referred to the Web version of this article.)

487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520

2.4. MDCT calcium analysis

The applied software automatically quantifies calcium vol-

ume based on a user-defined calcium detection threshold and

a manually identified volume of interest (VOI). Owing to the

variability of acquisition-related and patient-related factors,

mean contrast attenuation of the aortic lumen varies signifi-

cantly (from 158 to 826 HU in the present study cohort).

Therefore, patient-specific calcium detection thresholds were

determined in a fashion similar to that previously applied to

coronary calcium analysis.13,14 Accordingly, a 200-mm3 VOI

was placed in the uniform normal blood pool region at the

level of the left main stem. The image histogram of the VOI

was then obtained, and a Gaussian curve fitted to the histo-

gram (Fig. 2).13 On the basis of the mean value (meanfit) and

standard deviation (SDfit) of the fitted Gaussian curve, the

patient-specific calciumdetection thresholdwas calculated as

follows:

Calcium detection threshold ¼ meanfit þ 4SDfit

Varying values of the constant multiplying the SDfit were

tested in MDCT scans before transcatheter aortic valve

implantation from a subgroup of 40 patients. The optimal

visual discrimination between calcium and contrast was

found by using 4SDfit.

Calcium was quantified in 3 specific regions (Fig. 3): (1) the

overall LVOT (from the aortic annulus plane and 10 mm into

the left ventricle); (2) the upper LVOT (from the aortic annulus

plane and 2mm into the left ventricle); and (3) the aortic valve

region (from the aortic annulus plane to the left coronary

ostia). Each region was further subdivided according to the 3

aortic cups. The distal boundary of the upper LVOT region (ie,

2 mm into the left ventricle) was chosen based on previously

reported distances from the lower rim of the THV to the aortic

annulus in order to include the calcium that would be in direct

contact with the THV device.11,12 Furthermore, the presence of

continuous calcium extending through all the 3 regions was

recorded.

2.5. Statistical analysis

The distribution of continuous variables was tested for

normality by Q-Q plots and histograms. Continuous para-

metric variables are presented as mean � SD and compared

using the Student t test. Continuous nonparametric variables

are presented as median (interquartile range) and compared

Page 5

web4C=FPO

Fig. 3 e Quantitative analysis of calcium distribution. (A) Stretched MPR view of the aortic root. Three regions are defined:

the overall LVOT (from the aortic annulus and 10 mm into the left ventricle), the upper LVOT (from the aortic annulus and

2mm into the left ventricle), and the aortic valve region (from the aortic annulus to the left coronary ostia). (B1, B2) Long- and

short-axis views of the aortic valve region demonstrating calcium in relation to the 3 cusps. (C1, C2) Long- and short-axis

views of the upper LVOT demonstrating calcium below the noncoronary cusp. (D1, D2) Long- and short-axis views of the

overall LVOT demonstrating calcium below the NCs and LCs. LVOT, left ventricular outflow tract; LC, left coronary cusp;

MPR, multiplanar reformation; NC, noncoronary cusp; RC, right coronary cusp.

J o u rn a l o f C a r d i o v a s c u l a r C om p u t e d T omog r a p h y x x ( 2 0 1 5 ) 1e1 1 5

521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565566567568569570571572573574575576577578579580581582583584585

586587588589590591592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635636637

JCCT792_proof ■ 12 May 2015 ■ 5/11

using the Mann-Whitney U test. Categorical variables are

presented as frequencies (percentages) and compared using

the Fisher exact test or chi-square test as appropriate.

Receiver operating curve analysis was performed to test the

discriminatory power of calcium volumes in relation to

localization and aortic root mismatch parameters for predic-

tion of aortic root injury. Areas under the curves (AUCs) were

compared using themethod of DeLong et al15 with adjustment

for multiple comparisons. The regional calcium volume

parameter most predictive of aortic root injury was entered

into amultivariate logistic regressionmodel together with the

previously identified predictors, THV oversizing >20% and

redilatation.7 Likelihood ratio (LR) test was performed to

determine if the inclusion of the latter variables improved

the prediction of aortic root injury. A 2-tailed P value <.05

was considered statistically significant. All statistical

analyses were performed using Stata 12 (StataCorp LP; College

Station, TX).

638639640641642643644645646647648649650

3. Results

3.1. Study population

Baseline clinical and echocardiographic characteristics are

listed in Table 1. Patients with aortic root injury had a higher

median Society of Thoracic Surgeons (STS) risk score when

compared with the control group patients (7.3 [4.8e9.3] vs 5.1

[3.6e6.8]; P ¼ .004). There was no significant difference in

aortic valve area (0.63 � 0.20 vs 0.68 � 0.16 cm2; P ¼ .16)

between groups. Aortic root injury occurred in the LVOT in 4

patients (12.1%), at the level of the annulus in 22 patients

(66.7%), in the sinus of Valsalva in 5 patients (15.2%), and at the

sinotubular junction in 2 patients (6.1%). In patients with

aortic root injury, THV oversizing >20% (60.6% vs 37.9%; P ¼.020) and redilatation (24.2% vs 7.8%; P ¼ .011) were more

frequent than in the control group patients (Table 2).

3.2. Aortic root dimensions

Annular, LVOT, and sinus of Valsalva dimensions were

consistently smaller in patients with aortic root injury

(Table 3). There were no differences between groups with

regard to eccentricity at any level of the aortic root. The

annulus/LVOT area ratio (AUC, 0.55; 95% confidence interval

[CI], 0.45e0.65), sinus of Valsalva/annulus area ratio (AUC,

0.60; 95% CI, 0.48e0.71), or sinotubular junction/annulus area

ratio (AUC, 0.59; 95% CI, 0.48e0.70) did not predict aortic root

injury.

3.3. Calcium analysis

Examples of MDCT scans in patients experiencing aortic root

injury are depicted in Figure 4. Individual patient calcium

volumes in relation to the anatomic region in the 2 groups are

shown in Figure 5, and calcium volumes are summarized in

Table 4. There were no differences in median contrast atten-

uation ormedian calciumdetection threshold between groups

Page 6

Q7Table 1 e Baseline clinical and echocardiographic characteristics.

Characteristic All patients(n ¼ 186)

Patients with aorticroot injury (n ¼ 33)

Control group(n ¼ 153)

P value

Clinical characteristics

Age (y), median (interquartile range) 83 (78e86) 83 (79e87) 83 (78e86) .547

Female sex, n (%) 110 (59.1) 23 (69.7) 87 (56.9) .241

COPD, n (%) 68 (36.6) 8 (24.2) 60 (39.5) .114

Porcelain aorta, n (%) 23 (12.4) 4 (12.1) 19 (12.4) 1.000

Previous cardiac surgery, n (%) 41 (22.0) 4 (12.5) 37 (26.6) .110

STS risk score, median (interquartile range) 5.3 (3.7e7.2) 7.3 (4.8e9.3) 5.1 (3.6e6.8) .004

Echocardiographic characteristics

Ejection fraction <35%, n (%) 26 (14.2) 2 (6.5) 13 (8.5) 1.000

Mean transaortic gradient (mmHg), mean � SD 44.8 � 17.4 54.0 � 14.8 40.1 � 18.6 <.001

AVA (cm2), mean � SD 0 67 � 0.17 0.63 � 0.20 0.68 � 0.16 .097

AVA, aortic valve area; COPD, chronic obstructive pulmonary disease; SD, standard deviation; STS, Society of Thoracic Surgeons.

J o u r n a l o f C a r d i o v a s c u l a r Com p u t e d T omog r a p h y x x ( 2 0 1 5 ) 1e1 16

651652653654655656657658659660661662663664665666667668669670671672673674675676677678679680681682683684685686687688689690691692693694695696697698699700701702703704705706707708709710711712713714715

716717718719720721722723724725726727728729730731732733734735736737738739740741742743744745746747748749750751752753754755756757758759

JCCT792_proof ■ 12 May 2015 ■ 6/11

(Table 4). Patients with aortic root injury had significantly

more overall LVOT calcium; median volumes were 74 (5e326)

vs 3 (0e63) mm3 (P ¼ .0001). Median upper LVOT calcium

volumewas higher in the aortic root injury group compared to

the control group, 29 (3e66) vs 0 (0e9) mm3 (P< .0001). Median

aortic valve region calcium volume was not statistically

significantly different between the 2 groups, 848 (390e1138)

mm3 vs 580 (296e991) mm3 (P ¼ .115).

Receiver operating curve analysis confirmed that overall

LVOT (AUC, 0.71; 95% CI, 0.62e0.82) and upper LVOT calcium

volume (AUC, 0.78; 95% CI, 0.69e0.86), as opposed to the aortic

valve region calcium volume (AUC, 0.58; 95% CI, 0.47e0.69),

predicted aortic root injury (Fig. 6 and Table 5). The predictive

value of upper LVOT calcium volume for aortic root injury was

significantly higher than that of overall LVOT calcium

(P ¼ .010). Within the upper LVOT, calcium below the non-

coronary cusp was more predictive of aortic root injury (AUC,

0.81; 95% CI, 0.71e0.89) compared to calcium below the right

coronary cusp (AUC, 0.68; 95% CI, 0.59e0.77; P ¼ .033) or

calcium below the left coronary cusp (AUC, 0.64; 95% CI,

0.54e0.74; P ¼ .022). Continuous calcium extending

throughout all the 3 anatomic regions was more frequent in

patients experiencing aortic root injury compared with pa-

tients in the control group, 67% vs 37% (P ¼ .002).

Table 2 e Procedural characteristics.

Characteristic All patients(n ¼ 186)

Patroo

Access route, n (%)

Femoral 109 (58.6)

Apical 73 (39.3)

Aortic 3 (1.6)

Subclavian 1 (0.5)

THV size, n (%)

23 40 (21.5)

26 88 (47.3)

29 58 (31.2)

Redilatation 20 (10.8)

THV oversizing (%), mean � SD 19.6 � 12.7

THV oversizing >20%, n (%) 78 (41.9)

SD, standard deviation; THV, transcatheter heart valve.

3.4. Multivariate analysis of predictors of aortic rootinjury

In multivariable analysis, upper LVOT calcium volume (odds

ratio [OR], 1.34 per 10 mm3; 95% CI, 1.16e1.54; P < .0001), THV

oversizing >20% (OR, 2.69; 95% CI, 1.10e6.58; P ¼ .030), and

redilatation (OR, 4.61; 95% CI, 1.43e14.83; P ¼ .010) were

independently associated with aortic root injury. Addition of

THV oversizing >20% (LR test; P ¼ .028) and redilatation (LR

test; P ¼ .015), respectively, significantly improved the pre-

diction value of aortic root injury by upper LVOT calcium

volume.

4. Discussion

The present study demonstrates that calcium volume and

distribution may play an important role in predicting patients

at risk of aortic root injury during balloon-expandable TAVR.

Consistent with prior findings, overall LVOT calcium as

opposed to aortic valve region calciumwas predictive of aortic

root injury.7 However, by applying a novel method of detailed

3-dimensional regional calcium quantification, the present

study elaborates further on these findings and identifies upper

ients with aortict injury (n ¼ 33)

Control group(n ¼ 153)

P value

.009

25 (75.8) 84 (54.9)

6 (18.2) 67 (43.8)

1 (3.0) 2 (1.3)

1 (3.0) 0

.021

11 (33.3) 29 (19.0)

18 (54.6) 70 (45.8)

4 (12.1) 54 (35.3)

8 (24.2) 12 (7.8) .011

26.6 � 15.2 18.1 � 11.6 <.001

20 (60.6) 58 (37.9) .020

760761762763764765766767768769770771772773774775776777778779780

Page 7

Table 3 e Multidetector CT aortic root and LVOT dimensions.

Dimension All patients (n ¼ 186),mean � SD

Patients with aorticroot injury (n ¼ 33),

mean � SD

Control group (n ¼ 153),mean � SD

P value

Aortic annulus

Short-axis diameter 21.9 � 2.6 20.3 � 2.7 22.2 � 2.5 <.001

Long-axis diameter 27.4 � 2.7 26.1 � 3.1 27.7 � 2.6 .002

Mean diameter 24.7 � 2.5 23.2 � 2.7 25.0 � 2.3 <.001

Eccentricity index (%) 20.1 � 7.0 22.1 � 6.9 19.7 � 6.9 .084

Area (cm2) 4.62 � 0.90 4.10 � 0.95 4.73 � 0.84 <.001

LVOT

Short-axis diameter 20.3 � 3.4 18.7 � 3.1 20.6 � 3.3 .002

Long-axis diameter 28.9 � 4.0 27.1 � 3.5 29.3 � 4.0 .004

Mean diameter 24.6 � 3.4 22.9 � 2.9 25.0 � 3.3 .001

Eccentricity (%) 29.6 � 8.7 30.9 � 9.2 29.4 � 8.6 .366

Area (cm2) 4.44 � 1.24 3.80 � 1.05 4.57 � 1.24 .001

Sinus of Valsalva

Noncoronary cusp commissure diameter 32.4 � 3.5 30.7 � 3.9 32.7 � 3.3 .002

Right coronary cusp commissure diameter 31.4 � 3.6 29.5 � 4.1 31.9 � 3.4 <.001

Left coronary cusp commissure diameter 33.3 � 3.7 31.5 � 4.1 33.7 � 3.5 .002

Mean diameter 32.4 � 3.5 30.5 � 3.9 32.8 � 3.3 <.001

Area (cm2) 8.58 � 1.86 7.88 � 1.98 8.73 � 1.81 .017

Sinotubular junction

Short-axis diameter 28.2 � 3.4 27.3 � 4.4 28.4 � 3.2 .096

Long-axis diameter 29.8 � 3.7 28.8 � 4.6 30.0 � 3.4 .087

Mean diameter 29.0 � 3.5 28.1 � 4.5 29.2 � 3.2 .085

Eccentricity (%) 5.1 � 4.1 5.0 � 4.2 5.1 � 4.1 .867

Area (cm2) 6.61 � 1.66 6.21 � 2.07 6.70 � 1.55 .136

Height from aortic annulus 23.4 � 2.5 21.5 � 2.3 23.8 � 2.4 <.0001

Coronary ostia

Height left coronary ostia 14.6 � 3.2 14.2 � 2.4 14.7 � 3.3 .417

Height right coronary ostia 17.2 � 2.5 16.6 � 2.7 17.3 � 2.5 .149

Other

Annulus/LVOT area ratio 1.07 � 0.15 1.09 � 0.14 1.07 � 0.16 .361

Sinus of Valsalva/Annulus area ratio 1.87 � 0.29 1.94 � 0.34 1.85 � 0.27 .106

Sinotubular junction/Annulus area ratio 1.44 � 0.28 1.51 � 0.35 1.42 � 0.26 .079

LVOT, left ventricular outflow tract; SD, standard deviation.

Fig. 4 e Multidetector CT (MDCT) scans in 3 patients (A, B, and C) experiencing aortic root injury. (A1, A2) Double-oblique

transverse and sagittal oblique views displaying calcium in the upper LVOT. (B1, B2) Double-oblique transverse and coronal

oblique views displaying calcium extending through the overall left ventricular outflow tract. (C1, C2) Double-oblique

transverse and sagittal oblique views displaying left ventricular outflow tract calcium below the noncoronary cusp. (C3, C4)

Post-TAVR MDCT double-oblique transverse and coronal oblique views demonstrating contained aortic root injury (white

arrows) in relation to the left sinus of Valsalva in the patient depicted in panels C1 and C2.

J o u rn a l o f C a r d i o v a s c u l a r C om p u t e d T omog r a p h y x x ( 2 0 1 5 ) 1e1 1 7

781782783784785786787788789790791792793794795796797798799800801802803804805806807808809810811812813814815816817818819820821822823824825826827828829830831832833834835836837838839840841842843844845

846847848849850851852853854855856857858859860861862863864865866867868869870871872873874875876877878879880881882883884885886887888889890891892893894895896897898899900901902903904905906907908909910

JCCT792_proof ■ 12 May 2015 ■ 7/11

Page 8

web4C=FPO

Fig. 5 e Individual patient calcium volumes in the control and aortic root injury groups in the upper LVOT (A), the overall

LVOT (B), and aortic valve regions (C). Orange lines indicate median values (see also Table 4). *P < .0001 vs control group.

LVOT, left ventricular outflow tract. (For interpretation of the references to color in this figure legend, the reader is referred to

the Web version of this article.)

Q4

J o u r n a l o f C a r d i o v a s c u l a r Com p u t e d T omog r a p h y x x ( 2 0 1 5 ) 1e1 18

911912913914915916917918919920921922923924925926927928929930931932933934935936937938939940941942943944945946947948949950951952953954955956957958959960961962963964965966967968969970971972973974975

976977978979980981982983984985986987988989990991992993994995996997998999

1000100110021003100410051006100710081009101010111012101310141015101610171018

JCCT792_proof ■ 12 May 2015 ■ 8/11

LVOT calcium (ie, subannular calcium within 2 mm of the

aortic annulus plane) as themost important predictor of aortic

root injury. Furthermore, previously identified predictors of

aortic root injury, THV oversizing >20% and redilatation,

significantly improve the prediction of aortic root injury by

upper LVOT calcium volume.

Various quantitative and semiquantitative methods have

been applied using either noncontrast or contrast-enhanced

MDCT scans to assess the influence of aortic root calcifica-

tion on TAVR outcomes, in particular in relation to para-

valvular regurgitation.16e22 Traditional coronary calcium

quantification is performed on noncontrast CT examinations

by using the Agatston method involving a fixed calcium

detection threshold of 130 HU.23 When performing pre-TAVR

MDCT scans, however, intravenous contrast administration

is essential to gain detailed information regarding the aortic

annulus and peripheral vessels.24 Contrast enhancement

complicates calcium quantification as a fixed calcium

Table 4 e Multidetector CT calcium volumes.

All patients (n ¼ 186),median (interquartile

range) mm3

Patieinju

(inter

Aortic valve region calcium 621.4 (301.2e1023) 8

Noncoronary cusp 257.9 (135.6e444.5) 2

Right coronary cusp 171.9 (58.6e298.8) 2

Left coronary cusp 153.7 (71.3e326.1) 2

Upper LVOT calcium 0.8 (0e16.7)

Below the noncoronary cusp 0 (0e1.3)

Below the right coronary cusp 0 (0e0)

Below left coronary cusp 0 (0e4.8)

Overall LVOT Calcium 10.2 (0e87.8)

Below the noncoronary cusp 0 (0e14.3)

Below the right coronary cusp 0 (0e0.1)

Below the left coronary cusp 0.3 (0e27.5)

LVOT, left ventricular outflow tract.

detection threshold does not take into account the variation in

contrast attenuation. In accordance with previous findings,

mean contrast attenuation in the aortic root varied consider-

ably in this study.14,18 Furthermore, the complex anatomic

nature of the aortic root makes it challenging to accurately

perform 3-dimensional regional calcium quantification on

most standard CT workstations. Indeed, traditional axial

Agatston calcium scoring does not allow for detailed

3-dimensional aortic root assessment. Moreover, volumetric

quantification of calcium has been shown to improve inter-

scan reproducibility compared to the Agatston method owing

to less influence by volume averaging and slice thickness.25e27

In this study, we applied a new and more comprehensive

method of quantitative calcium analysis incorporating

patient-specific calcium detection thresholds and detailed

3-dimensional regional analysis on contrast-enhanced MDCT

scans. The software used ensures working in perpendicular

views along a centerline and has dedicated 3-dimensional VOI

nts with aortic rootry (n ¼ 33), medianquartile range) mm3

Control group (n ¼ 153),median (interquartile

range) mm3

P value

47.7 (389.7e1138) 580.0 (296.0e991.1) .115

78.5 (207.9e582.3) 237.4 (129.2e420.5) .080

14.9 (69.2e353.2) 151.4 (58.6e295.1) .324

30.5 (75.9e344.4) 149.6 (67.6e319.4) .430

28.7 (3.1e65.9) 0 (0e8.7) <.0001

5.5 (0.3e36.7) 0 (0e0.4) <.0001

0 (0e1.8) 0 (0e0) <.0001

0.7 (0e16.7) 0 (0e3.7) .005

74.3 (4.6e326.3) 3.1 (0e69.3) .0001

26.3 (0.4e125.1) 0 (0e5.3) <.0001

0.4 (0e4.25) 0 (0e0) <.0001

11.6 (0.1e74.3) 0 (0e22.9) .009

1019102010211022102310241025102610271028102910301031103210331034103510361037103810391040

Page 9

web4C=FPO

Fig. 6 e Receiver operating curves assessing the predictive

value of calcium volumes for aortic root injury. (A) Overall

LVOT, upper LVOT, and aortic valve region calcium and (B)

upper LVOT calcium below the noncoronary cusps, right

coronary cusps, and left coronary cusps, respectively (see

also Table 4). LVOT, left ventricular outflow tract.

Table 5 e Receiver operating analysis for prediction ofaortic root injury by calcium volume.

AUC 95% CI

Aortic valve region calcium volume 0.58 0.47e0.69

Noncoronary cusp 0.60 0.49e0.70

Right coronary cusp 0.55 0.44e0.67

Left coronary cusp 0.55 0.44e0.67

Upper LVOT calcium volume 0.78* 0.69e0.86

Below the noncoronary cusp 0.81y 0.71e0.89

Below the right coronary cusp 0.68 0.59e0.77

Below the left coronary cusp 0.64 0.54e0.74

Overall LVOT calcium volume 0.71z 0.62e0.82

Below the noncoronary cusp 0.74 0.65e0.84

Below the right coronary cusp 0.70 0.61e0.80

Below the left coronary cusp 0.63 0.54e0.74

AUC, area under curves; CI, confidence interval; LVOT, left

ventricular outflow tract.

* P value<.01 compared to AUC for aortic valve calcium and overall

LVOT calcium volumes, respectively.

y P value <.05 compared to AUC for proximal LVOT calcium

volume below the right and left coronary cusp, respectively.

z P value <.05 compared to AUC for aortic valve calcium volume.

J o u rn a l o f C a r d i o v a s c u l a r C om p u t e d T omog r a p h y x x ( 2 0 1 5 ) 1e1 1 9

10411042104310441045104610471048104910501051105210531054105510561057105810591060106110621063106410651066106710681069107010711072107310741075107610771078107910801081108210831084108510861087108810891090109110921093109410951096109710981099110011011102110311041105

11061107110811091110111111121113111411151116111711181119112011211122112311241125112611271128112911301131113211331134113511361137113811391140114111421143114411451146114711481149115011511152115311541155115611571158115911601161116211631164116511661167116811691170

JCCT792_proof ■ 12 May 2015 ■ 9/11

definition capabilities, which allowed for precise regional

calcium quantification.

It is somewhat intuitive that upper LVOT calcium is more

predictive of aortic root injury given its contact with the

deployed THV and exposure to the force created during

balloon expansion. Interestingly though, our results suggest

that within the upper LVOT region, it is calcium located below

the noncoronary cusp that is most predictive of aortic root

injury. The reason for this finding is not entirely clear. Previ-

ous case reports suggest that the left aortic sinus may be the

most vulnerable area with regard to aortic root injury possibly

because of the lack of supporting cardiac structures in this

area.28,29 However, it may be speculated whether the culprit

site is located at the site of calcification or at the aortic wall

opposite to the calcification because of THV migration away

from hard calcified areas during balloon expansion. If the

latter were to hold true, this could potentially explain our

findings.

Previous studies have demonstrated lack of association

between annular eccentricity and paravalvular regurgitation

as well as postimplant THV geometry.5,11,30 Correspondingly,

in this study, eccentricity at different levels of the aortic root

was not associated with aortic root injury, hence emphasizing

the compliant nature of the aortic root. Of importance, aortic

root injury was observed at all levels of the aortic root.7 As

selection of THV size is based on annular dimensions, it

therefore may be speculated whether mismatch between

different portions of the aortic root could play a role in the

mechanism of aortic root injury. In this study, there was no

evidence of such an association; however, it cannot be

excluded thatmismatch in combinationwith calcium or other

procedural or anatomic characteristics constitutes an

elevated risk of aortic root injury.

4.1. Limitations

Different MDCT scanners and protocols were used, and this

could potentially influence calcium volume measurements.

To accommodate this, we, in contrast to previous studies,

used a patient-specific calcium detection threshold and

volumetric assessment as this has proven more reproduci-

ble between scans compared to the traditional Agatston

score.25e27 MDCT analysesweremade using specific dedicated

software, and therefore, specific values for calcium volumes

may not be immediately applicable to other workstation

platforms. That being said, the primary finding of our analysis

was to determine the distribution and location of calcium that

are associated with aortic root rupture rather than specific

thresholds for calcium volumes.

Page 10

5

J o u r n a l o f C a r d i o v a s c u l a r Com p u t e d T omog r a p h y x x ( 2 0 1 5 ) 1e1 110

11711172117311741175117611771178117911801181118211831184118511861187118811891190119111921193119411951196119711981199120012011202120312041205120612071208120912101211121212131214121512161217121812191220122112221223122412251226122712281229123012311232123312341235

12361237123812391240124112421243124412451246124712481249125012511252125312541255125612571258125912601261126212631264

JCCT792_proof ■ 12 May 2015 ■ 10/11

The unmatched control group comprised consecutive

patients from one of the participating centers. We acknowl-

edge that this design may have led to confounding. Of note,

patients experiencing aortic root injury had higher STS risk

scores and smaller aortic root dimensions compared to the

control group patients. STS risk score may be associated with

aortic root injury; however, the difference in STS risk score

between the 2 groups in this study may also in part reflect

differences in institutional practice. Furthermore, the STS risk

score of patients with aortic root injury was not particularly

high when comparing with data from large TAVR regis-

tries.31,32 The differences in aortic root dimensions between

groups may be a result of differences in geographic origin

between study groups; thus, whether a causal relationship

between aortic root dimensions and aortic root injury exists

cannot be concluded from this study. Of importance, the

control group appears appropriate with respect to confirming

both oversizing >20% and redilatation as independent pre-

dictors of aortic root injury. Although this study is based on

the largest cohort of patient with aortic root injury yet

collected, the relatively low number of cases did not allow for

more elaborate adjustments for confounders. Larger studies

are warranted to further study the interactions between pre-

dictors of aortic root injury. Finally, the data from this study

concern balloon-expandable Edwards SAPIEN and SAPIEN

XT and therefore may not be directly applied to other types

of THVs.

126512661267126812691270127112721273127412751276127712781279

5. Conclusion

Through detailed quantitative calcium analysis, this study

builds on our previous findings by demonstrating increased

predictive value of aortic root injury of upper LVOT calcium

compared with overall LVOT calcium and interestingly, a

heightened risk of rupture in the setting of calcification below

the noncoronary cusp. The effect of upper LVOT calcium on

the risk of aortic root injury is augmented in the setting of the

THV oversizing >20% and balloon redilatation. These findings

may help predefine patients at increased risk of aortic root

injury more precisely.

12801281128212831284128512861287128812891290129112921293129412951296129712981299

r e f e r e n c e s

1. Leon MB, Smith CR, Mack M, et al. Transcatheter aortic-valveimplantation for aortic stenosis in patients who cannotundergo surgery. N Engl J Med. 2010;363:1597e1607.

2. Smith CR, Leon MB, Mack MJ, et al. Transcatheter versussurgical aortic valve replacement in high-risk patients. N EnglJ Med. 2011;364:2187e2198.

3. Blanke P, Schoepf UJ, Leipsic JA. CT in transcatheter aorticvalve replacement. Radiology. 2013;269:650e669.

4. Binder RK, Webb JG, Willson AB, et al. The impact ofintegration of a multidetector computed tomography annulusarea sizing algorithm on outcomes of transcatheter aorticvalve replacement: a prospective, multicenter, controlledtrial. J Am Coll Cardiol. 2013;62:431e438.

5. Hansson NC, Thuesen L, Hjortdal VE, et al. Three-dimensionalmultidetector computed tomography versus conventional2-dimensional transesophageal echocardiography for

annular sizing in transcatheter aortic valve replacement:influence on postprocedural paravalvular aortic regurgitation.Catheter Cardiovasc Interv. 2013;82:977e986.

6. Jilaihawi H, Kashif M, Fontana G, et al. Cross-sectionalcomputed tomographic assessment improves accuracy ofaortic annular sizing for transcatheter aortic valvereplacement and reduces the incidence of paravalvular aorticregurgitation. J Am Coll Cardiol. 2012;59:1275e1286.

7. Barbanti M, Yang TH, Rodes Cabau J, et al. Anatomical andprocedural features associated with aortic root rupture duringballoon-expandable transcatheter aortic valve replacement.Circulation. 2013;128:244e253.

8. Leetmaa TH, Hansson N, Leipsic J, et al. Early aortictranscatheter heart valve thrombosis: diagnostic value ofcontrast-enhanced multidetector computed tomography. CircCardiovasc Interv. 2015;8. Q

9. Delgado V, Ng ACT, Schuijf JD, et al. Automated assessmentof the aortic root dimensions with multidetector rowcomputed tomography. Ann Thorac Surg. 2011;91:716e723.

10. Buellesfeld L, Stortecky S, Kalesan B, et al. Aortic rootdimensions among patients with severe aortic stenosisundergoing transcatheter aortic valve replacement. JACCCardiovasc Interv. 2013;6:72e83.

11. Binder RK, Webb JG, Toggweiler S, et al. Impact of post-implant SAPIEN XT geometry and position on conductiondisturbances, hemodynamic performance, and paravalvularregurgitation. JACC Cardiovasc Interv. 2013;6:462e468.

12. Willson AB, Webb JG, Gurvitch R, et al. Structural integrity ofballoon-expandable stents after transcatheter aortic valvereplacement: assessment by multidetector computedtomography. JACC Cardiovasc Interv. 2012;5:525e532.

13. Dey D, Cheng VY, Slomka PJ, et al. Automated 3-dimensionalquantification of noncalcified and calcified coronary plaquefrom coronary ct angiography. J Cardiovasc Comput Tomogr.2009;3:372e382.

14. Mylonas I, Alam M, Amily N, et al. Quantifying coronaryartery calcification from a contrast-enhanced cardiaccomputed tomography angiography study. Eur Heart JCardiovasc Imaging. 2014;15:210e215.

15. DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing theareas under two or more correlated receiver operatingcharacteristic curves: a nonparametric approach. Biometrics.1988;44:837e845.

16. Feuchtner G, Plank F, Bartel T, et al. Prediction of paravalvularregurgitation after transcatheter aortic valve implantation bycomputed tomography: value of aortic valve and annularcalcification. Ann Thorac Surg. 2013;96:1574e1580.

17. Leber AW, Kasel M, Ischinger T, et al. Aortic valve calciumscore as a predictor for outcome after TAVI using theCoreValve revalving system. Int J Cardiol. 2013;166:652e657.

18. Ewe SH, Ng AC, Schuijf JD, et al. Location and severity ofaortic valve calcium and implications for aortic regurgitationafter transcatheter aortic valve implantation. Am J Cardiol.2011;108:1470e1477.

19. Schultz C, Rossi A, Van Mieghem NM, et al. Aortic annulusdimensions and leaflet calcification from contrast MSCTpredict the need for balloon post-dilatation after TAVI withthe Medtronic CoreValve prosthesis. EuroIntervention.2011;7:564e572.

20. Colli A, Gallo M, Bernabeu E, D’Onofrio A, Tarzia V, Gerosa G.Aortic valve calcium scoring is a predictor of paravalvularaortic regurgitation after transcatheter aortic valveimplantation. Ann Thorac Surg. 2012;1:156e159.

21. John D, Buellesfeld L, Yuecel S, et al. Correlation of Devicelanding zone calcification and acute procedural success inpatients undergoing transcatheter aortic valve implantationswith the self-expanding CoreValve prosthesis. JACCCardiovasc Interv. 2010;3:233e243.

1300

Page 11

J o u rn a l o f C a r d i o v a s c u l a r C om p u t e d T omog r a p h y x x ( 2 0 1 5 ) 1e1 1 11

130113021303130413051306130713081309131013111312131313141315131613171318131913201321132213231324132513261327132813291330133113321333133413351336133713381339134013411342134313441345134613471348134913501351135213531354135513561357135813591360136113621363136413651366136713681369

13701371137213731374137513761377137813791380138113821383138413851386138713881389139013911392

JCCT792_proof ■ 12 May 2015 ■ 11/11

22. Staubach S, Franke J, Gerckens U, et al. German transcatheteraortic valve implantation-registry I. Impact of aortic valvecalcification on the outcome of transcatheter aortic valveimplantation: results from the prospectivemulticenter GermanTAVI registry. Catheter Cardiovasc Interv. 2013;81:348e355.

23. Agatston AS, Janowitz WR, Hildner FJ, Zusmer NR,Viamonte Jr M, Detrano R. Quantification of coronary arterycalcium using ultrafast computed tomography. J Am CollCardiol. 1990;15:827e832.

24. Achenbach S, Delgado V, Hausleiter J, Schoenhagen P, Min JK,Leipsic JA. SCCT expert consensus document on computedtomography imaging before transcatheter aortic valveimplantation (TAVI)/transcatheter aortic valve replacement(TAVR). J Cardiovasc Comput Tomogr. 2012;6:366e380.

25. Callister TQ, Cooil B, Raya SP, Lippolis NJ, Russo DJ, Raggi P.Coronary artery disease: improved reproducibility of calciumscoring with an electron-beam CT volumetric method.Radiology. 1998;208:807e814.

26. Morgan-Hughes GJ, Owens PE, Roobottom CA, Marshall AJ.Three dimensional volume quantification of aortic valvecalcification using multislice computed tomography. Heart.2003;89:1191e1194.

27. Ohnesorge B, Flohr T, Fischbach R, et al. Reproducibility ofcoronary calcium quantification in repeat examinations with

retrospectively ECG-gated multisection spiral CT. Eur Radiol.2002;12:1532e1540.

28. Blanke P, Reinohl J, Schlensak C, et al. Prosthesis oversizing inballoon-expandable transcatheter aortic valve implantationis associated with contained rupture of the aortic root. CircCardiovasc Interv. 2012;5:540e548.

29. Hayashida K, Bouvier E, Lefevre T, et al. Potentialmechanism of annulus rupture during transcatheter aorticvalve implantation. Catheter Cardiovasc Interv.2013;82:E742eE746.

30. Willson AB, Webb JG, LaBounty TM, et al. 3-dimensionalaortic annular assessment by multidetector computedtomography predicts moderate or severe paravalvularregurgitation after transcatheter aortic valve replacement: amulticenter retrospective analysis. J Am Coll Cardiol.2012;59:1287e1294.

31. Gilard M, Eltchaninoff H, Iung B, et al. Registry oftranscatheter aortic-valve implantation in high-risk patients.N Engl J Med. 2012;366:1705e1715.

32. Genereux P, Head SJ, Van Mieghem NM, et al. Clinicaloutcomes after transcatheter aortic valve replacement usingvalve academic research consortium definitions: a weightedmeta-analysis of 3,519 patients from 16 studies. J Am CollCardiol. 2012;59:2317e2326.

139313941395139613971398139914001401140214031404140514061407140814091410141114121413141414151416141714181419142014211422142314241425142614271428142914301431143214331434143514361437