Embed Size (px)
Citation preview
TIMJDU
(ac(marci
ptIwcpi
4
A(svwtbctmuc
vaidct
A
ASH
©P
AD
ULT
CA
RD
IAC
ransapical Versus Transfemoral Aortic Valvemplantation: A Comparison of Survival and Safetyalin Johansson, MD, Shahab Nozohoor, MD, PhD, Per Ola Kimblad, MD, PhD,
an Harnek, MD, PhD, Göran K. Olivecrona, MD, PhD, and Johan Sjögren, MD, PhDepartments of Cardiothoracic Surgery, Cardiothoracic Anesthesia and Intensive Care, and Coronary Heart Disease, Skane
niversity Hospital and Lund University, Lund, SwedenoavtctAt
unfcpb
Background. Transcatheter aortic valve implantationTAVI) is a therapeutic option for high-risk patients withortic stenosis. Procedural mortality remains high inomparison with conventional aortic valve replacementAVR) because patients determined for TAVI are com-
only denied conventional surgery. We aimed to evalu-te access-related complications between the transfemo-al (TF) and the transapical (TA) approach and toompare survival between TAVI and conventional AVRn propensity-score-matched patients.
Methods. Between January 2008 and November 2009, 40atients underwent TAVI (TF, n � 10; TA, n � 30) with
he Edwards Sapien bioprosthesis (Edwards Lifesciences,rvine, CA). Survival and postoperative complicationsere evaluated between the TF and the TA approach. A
omparison of survival was made between the TAVIatients and propensity-score-matched patients undergo-
ng conventional AVR.Results. Successful implantation rate was 92.5% (37 of
0). Thirty-day mortality was 5.0% (2 of 40), and the
ddathau
(aTnar
cas
urgery, Cardiothoracic Anesthesia and Intensive Care, Skane Universityospital, Lund, SE-221 85, Sweden; e-mail: [email protected].
2011 by The Society of Thoracic Surgeonsublished by Elsevier Inc
verall in-hospital mortality was 10.0% (4 of 40). Survivalfter TAVI was 77% at both 6 months and 1 year. Majorascular complications occurred in 3 of 10 patients (all inhe TF group), and 3 of 40 patients (7.5%) sufferederebrovascular events. A comparison of survival be-ween TAVI and propensity score-matched conventionalVR patients showed no significant difference in either
he TA group (p � 0.73) or the TF group (p � 0.59).Conclusions. The vascular complications occurring when
sing the TF approach were probably related to a combi-ation of a wide introducer sheath and heavily calcifiedemoral arteries in a high-risk population. No seriousomplications were encountered when using the TA ap-roach. After propensity-score matching, survival withoth the TA and TF approaches is similar to that after AVR.
(Ann Thorac Surg 2011;91:57–63)
© 2011 by The Society of Thoracic Surgeonsortic stenosis is the most common heart valve dis-ease in adults [1], and aortic valve replacement
AVR) is the treatment of choice for symptomatic aortictenosis [2]. However, according to the Euro Heart Sur-ey on Valvular Heart Disease, one third of all patientsith severe symptomatic aortic stenosis who were more
han 75 years of age were not offered surgery [1], mainlyecause of high age, left ventricular dysfunction, or otheromorbidity. Since the introduction of transcatheter aor-ic valve implantation (TAVI), these high-risk patients
ay now be offered an alternative therapeutic optionsing a less invasive technique without the need forardiopulmonary bypass [3].
Patient selection for TAVI is still subject to debate, andarious risk stratification models have been used to obtainn objective risk assessment [4, 5]. However, these tools aremprecise and may exclude significant comorbidity oftenemonstrated in the TAVI population. Furthermore, re-ently published data indicate that comparisons betweenhe operative methods of AVR and TAVI should include the
ccepted for publication July 21, 2010.
ddress correspondence to Dr Sjögren, Department of Cardiothoracic
istribution of coronary artery disease [6]. In addition, it isifficult to compare survival and morbidity between TAVInd AVR owing to obvious differences in baseline charac-eristics, as patients determined for TAVI are consideredigh risk and commonly denied conventional AVR. Toddress this concern, propensity score matching can besed to balance measured baseline covariates [7, 8].Several centers specialize in either the transfemoral
TF) or the transapical (TA) approach, but we havedopted both techniques in one multidisciplinary team.his strategy gives us the opportunity to use the tech-ique considered most suitable for the individual patientnd also provides an opportunity to compare access-elated complications.
The aim of the present study was to evaluate the compli-ations occurring when using the TF and the TA techniquesnd to compare clinical outcome of TAVI to a propensitycore matched group undergoing conventional AVR.
Dr Olivecrona discloses that he has a financial rela-
tionship with Edwards Lifesciences.0003-4975/$36.00doi:10.1016/j.athoracsur.2010.07.072
M
PBw(SITmdwimtappwidvv
pbuRTccptiraowms2wSSnbU
CBuapsd
OBh
swwiahttwpitiuAtettwsaFdlfiwcaa
SCccecmpam3(apcpPpuniubtncv
58 JOHANSSON ET AL Ann Thorac SurgTAVI AND SURVIVAL 2011;91:57–63A
DU
LTC
AR
DIA
C
aterial and Methods
atient Population and Study Designetween January 2008 and November 2009, 40 patientsith severe symptomatic aortic stenosis underwent TAVI
41 procedures) with the 23-mm or 26-mm EdwardsAPIEN transcatheter heart valve (Edwards Lifesciences,rvine, CA). This prosthesis is approved for both TF andA access and is suitable for native annulus sizes of 18m to 25 mm. The TF and TA techniques have been
escribed in detail previously [9–11]. Both proceduresere performed by a team of cardiac surgeons and
nterventional cardiologists/radiologists. The initial treat-ent strategy (January 2008 to January 2009) was to use
he TF approach as the first option, reserving the TApproach for patients who were declined the TF ap-roach. However, owing to access-related vascular com-lications during the initial procedures, the TA approachas adopted as the primary option. The study population
nitially included 11 patients selected for the TF proce-ure, and 29 for the TA procedure; however, owing toascular complications, 1 TF patient was converted to TAalve implantation.The criteria for inclusion and exclusion used in the
resent study have been previously described [12]. Theaseline risk of the patient population was estimatedsing the logistic European System for Cardiac Operativeisk Evaluation (EuroSCORE) [5] and The Society ofhoracic Surgeons score, which was used together withlinical judgment. This approach allowed risk factors notovered in the risk score models, such as malignancy,orcelain aorta, and previous radiation therapy to be
aken into account (Table 1). Preprocedural screeningncluded transthoracic and transesophageal echocardiog-aphy, computed tomography, coronary angiography,ortography, and peripheral vascular angiography. Pre-perative, perioperative, and postoperative variablesere prospectively collected and entered into the depart-ent’s computerized cardiac surgical database for retro-
pective analysis. Follow-up was performed in January010 and was 100% complete. The mean follow-up timeas 10 � 8 months (median 7.5; interquartile range, 16).urvival data and cause of death were obtained from thewedish National Board of Health and Welfare or, ifecessary, from patient records. The study was approvedy the Ethics Committee for Clinical Research at Lundniversity, Sweden.
onventional AVR Populationetween January 1999 and April 2009, 2,262 patientsnderwent conventional AVR with or without coronaryrtery bypass surgery at our department. Using thisopulation, matching (relation 1:1) based on the propen-ity score was used to compare patients undergoing theifferent procedures.
perative Management and Definitionsefore the TAVI procedure, 40% (16 of 40) of the patients
ad significant coronary artery stenosis and underwent (uccessful percutaneous coronary intervention 1 to 2eeks before the TAVI procedure. All TAVI proceduresere performed under general anesthesia in a catheter-
zation laboratory under surveillance using fluoroscopynd transesophageal echocardiography. Intravenouseparin was administered at 80 IU/kg body weight with
he aim of achieving an activated clotting time greaterhan 200 s to 250 s. Normal clotting was reestablishedith protamine at the end of the procedure. A transientacemaker wire was inserted transvenously before valve
mplantation. An introducer sheath (6F) was inserted intohe most suitable femoral artery (based on the prescreen-ng examination) for the introduction of a pigtail cathetersed to visualize the aortic root using contrast injections.ll TF patients were treated with the SAPIEN valve using
he Retroflex (Edwards Lifesciences) transfemoral deliv-ry system (22F for the 23-mm prosthesis and 24F size forhe 26-mm prosthesis), and all the TA patients werereated with the SAPIEN valve using the Ascendra (Ed-ards Lifesciences) transapical delivery system. Prosthe-
is function and placement was assessed by angiographynd intraoperative transesophageal echocardiography.or postoperative anticoagulation, 75 mg clopidogrelaily was administered for 1 month, together with a
ifetime dose of 75 mg daily aspirin. Patients with atrialbrillation or other indications for warfarin receivedarfarin and aspirin without clopidogrel. To facilitate
omparisons to previous studies, measures of outcomend complications were defined according to Wendlernd associates [12].
tatistical Analysisontinuous variables are presented as means � SD, and
ategorical variables are presented as relative frequen-ies. Proportions were compared using the �2 or Fisher’sxact test (when frequencies were less than 5), andontinuous variables using Student’s t test. Because treat-ent assignment was not based on random allocation,
ropensity score adjustment was used to reduce imbal-nces in covariates at baseline [8]. A logistic regressionodel was fitted (Hosmer-Lemeshow goodness-of-fit: �2
.4, p � 0.91; and c-statistic: 0.923), where treatmentTAVI versus AVR) was the outcome, and baseline char-cteristics from the EuroSCORE model in addition toresence of coronary artery disease (defined as previousoronary artery bypass graft or PCI before the TAVIrocedure) were the covariates in a bivariate analysis.ropensity scores were generated for the AVR and TAVIatients using an SPSS macro (SPSS, Chicago, IL), andsed to match patients from the two groups in a nearest-eighbor fashion. The propensity score adjusted sample
ncluded 40 patients who underwent AVR and 40 whonderwent TAVI (TA, n � 30; TF, n � 10). The covariatealance achieved by matching was assessed by checking
hat the variables included in the propensity score wereo longer significant in the matched sample as well asalculating the absolute standardized differences in co-ariates between patients undergoing AVR and TAVI
Table 2). An absolute standardized difference of less
taetwtes
R
OT4g
Trdgipudsdpsiq
TR
V
FHR
DCNRPRPPANLLPCPO
ASLCBA
L
V
CO n;a neous
59Ann Thorac Surg JOHANSSON ET AL2011;91:57–63 TAVI AND SURVIVAL
AD
ULT
CA
RD
IAC
han 10% for the measured covariate suggests appropri-te balance between the patients undergoing the differ-nt treatment modalities. The survival function was illus-rated by Kaplan-Meier curves, and survival distributionsere compared with the log rank test. A value of p less
han 0.05 was considered a statistically significant differ-nce. Statistical analysis was performed using SPSS ver-ion 17 software (SPSS Inc, Chicago, IL).
esults
perative Datahe overall procedural success of TAVI was 92.5% (37 of0). In the TA group, it was 93% (28 of 30), and in the TF
able 1. Preoperative Characteristics of Patients Undergoing Teplacement (AVR)
ariable
TF Approach
n � 10 %
emale 5 50ypertension 3 30edo surgeryPrevious CABG 4 40Previous AVR 0 0iabetes mellitus 1 10OPD 1 10eurologic dysfunction 1 10enal failure 1 10reoperative dialysis 1 10ecent myocardial infarction 1 10ulmonary hypertension 0 0eripheral vascular disease 5 50trial fibrillation 2 20YHA functional class IV 1 10VEF 30 to 50 1 10VEF �30 2 20CI before TAVI 5 50ancer 1 10orcelain aorta 0 0ther severe comorbidity 6 60
Mean SD
ge, years 83 6tandard EuroSCORE, points 11 2ogistic EuroSCORE, % 25.6 15reatinine 117 61ody mass index, kg/m2 28 5ortic gradient, mm HgPeak 89 38Mean 54 24
VOT, mm 22 2
alues given are number and percentage of patients, or mean � SD.
ABG � coronary artery bypass surgery; COPD � chronic obstruperative Risk Evaluation; LVEF � left ventricular ejection fractio
pplicable; NYHA � New York Heart Association; PCI � percuta
roup, it was 91% (10 of 11; 1 TF patient was converted to o
A-TAVI). The TF procedure was converted owing to aupture of the femoral artery after insertion of the intro-ucer sheath. A Fluency Plus self-expanding coveredraft (Bard Peripheral Vascular, Tempe, AZ) was inserted
nto the left femoral artery, covering the rupture. Theatient was uneventfully discharged after 5 days andnderwent TA-TAVI successfully 5 months later. Proce-ural failure occurred in 2 patients in the TA group: openurgery had to be performed in 1 patient owing toislocation of the valve in the left ventricle; and onerocedure was discontinued because of transient ob-truction of the left main coronary artery during thenitial balloon valvuloplasty. One patient (TA-TAVI) re-uired valve-in-valve treatment (SAPIEN-in-SAPIEN)
catheter Aortic Valve Implantation (TAVI) and Aortic Valve
TA Approach AVR
p Value� 30 % n � 40 %
15 50 22 55 0.9116 53 14 35 0.23
10 25 0.455 171 39 30 4 10 0.074
12 40 12 30 0.204 13 4 10 0.901 3 3 33 0.531 3 0 0 0.184 13 4 10 0.904 13 1 3 0.12
14 47 17 43 0.8912 40 9 23 0.2911 37 2 5 0.002
8 27 10 25 0.543 10 5 13 0.71
11 37 NA 0.483 10 NA 1.008 27 NA 0.17
21 70 NA 0.70
Mean SD Mean SD
80 6 81 5 0.3610 3 11 3 0.8823.5 17 22.7 16 0.73
104 50 99 29 0.5027 4 26 5 0.50
80 24 87 23 0.4945 15 NA 0.1722 2 NA 0.54
pulmonary disease; EuroSCORE � European System for CardiacLVOT � left ventricle outflow tract; NA � data not available orcoronary intervention; TA � transapical; TF � transfemoral.
rans
n
ctive
wing to significant transprosthetic regurgitation.
PPppra(tlfhswbeccsptoC1vhaharfipt
ETmwgttwgp3fdv
1wpmoppfttsp
C
C
Ta
V
FRCNRRPPLLC
ASLP
V
A ctiveO
60 JOHANSSON ET AL Ann Thorac SurgTAVI AND SURVIVAL 2011;91:57–63A
DU
LTC
AR
DIA
C
ostoperative Complicationsostoperative outcome is summarized in Table 3. Majorostoperative vascular complications occurred in 3 of 10atients (30%) in the TF group. One patient underwenteoperation because of severe bleeding from the femoralccess point caused by a dislocated wound closure deviceProstar XL, Abbott Vascular, Abbott Park, IL). The pa-ient was taken to the operating room, an explorativeaparotomy was performed, and the insertion point in theemoral artery was surgically closed. The second patientad a severe hemorrhage from the right femoral arteryhortly after arrival in the intensive care unit and under-ent emergent surgery with a repair of a rupturedranch of the femoral artery. Thereafter, an arterialmbolectomy was successfully performed because oflinical signs of limb ischemia. The third patient hadritical ischemia and was embolectomized at the sameide as the femoral access in the lower limb on the 10thostoperative day. As the ischemia progressed slowly,
he patient underwent a transtibial amputation, but diedf multiple organ failure on the 35th postoperative day.erebrovascular events occurred in 3 patients (2 TF andTA). The 2 TF patients were diagnosed with cerebro-
ascular ischemia due to embolism, and the TA patientad a traumatically incurred subarachnoid hemorrhagefter falling in the ward. Four patients (3 TA and 1 TF)ad renal failure (creatinine increase to �200 �mol/L ornuria). Two of the 3 TA patients with renal failureequired hemodialysis. New-onset postoperative atrialbrillation occurred in 3 patients, all in the TA group. Noatient required pacemaker implantation for postopera-
able 2. Propensity Score Adjustment Between Transcatheternd Aortic Valve Replacement (AVR) With or Without Conco
ariable
TAVI
n � 40 %
emale sex 20 50edo surgery 10 25OPD 13 32.5eurologic dysfunction 5 12.5enal failure 2 5ecent myocardial infarction 5 12.5ulmonary hypertension 4 10eripheral vascular disease 19 47.5VEF 30–50 9 22.5VEF �30 5 12.5oronary artery disease 25 62.5
Mean SD
ge, years 81 6tandard EuroSCORE, points 11 3ogistic EuroSCORE, % 24 17ropensity score 0.19 0.21
alues given are number and percentage of patients, or mean � SD.
bSD � absolute standardized difference; COPD � chronic obstruperative Risk Evaluation; LVEF � left ventricular ejection fraction.
ive atrioventricular block. a
arly and Late Outcomehere was no intraoperative mortality. The overall 30-dayortality was 5.0% (2 of 40). The in-hospital mortalityas 10.0% (4 of 40): TA group 6.7% (2 of 30) versus TFroup 20% (2 of 10). In the TA group, 1 patient died ofraumatic subarachnoid hemorrhage 8 days postopera-ively due to accidentally falling in the ward while onarfarin treatment. The second patient died of multior-an failure 17 days postoperatively. In the TF group, 1atient died of intestinal carcinoma with liver metastases1 days postoperatively. One patient died of multiorganailure related to critical ischemia in the lower limb 35ays postoperatively. Autopsies demonstrated functionalalve prosthesis in all 4 patients.Late survival after TAVI was 77% both at 6 months andyear. One patient in the TF group died of heart failureith pulmonary edema 96 days postoperatively. Fouratients in the TA group died during follow-up, 3 ofyocardial infarction at 133, 153, and 481 days, and 1 died
f heart failure 123 days postoperatively. The results afterropensity score adjustment are given in Table 2. Pro-ensity score adjusted Kaplan-Meier survival estimates
or the TF and TA treatment groups are compared withhose for conventional AVR in Figure 1, demonstratinghat conventional AVR was not associated with a higherurvival rate than the TAVI procedure (TA, p � 0.73; TF,� 0.59).
omment
onventional AVR is the treatment of choice for calcific
c Valve Implantation (TAVI) With Coronary Artery Diseasent Coronary Artery Bypass Graft Surgery at Baseline
AVR
p Value AbSD (%)n � 40 %
22 55 0.65 �1010 25 1.0 �012 30 0.81 5.4
4 10 1.0 7.91 2.5 1.0 13.14 10 1.0 7.91 2.5 0.36 31
17 43 0.65 9.110 25 0.79 �5.95 12.5 1.0 �0
23 57.5 0.65 �10.2
Mean SD
81 5 0.73 �5.611 3 0.8 �4.823 16 0.71 6.8
0.16 0.17 0.50
pulmonary disease; EuroSCORE � European System for Cardiac
Aortimita
ortic stenosis, and the results are favorable even in the
ewmttctvTopfi
hcTaavSAlse
pttttasrt
is(dtiblevrpc
T
V
P
HPNLND
CCUVPIL
TCH
Vh
A d;t
61Ann Thorac Surg JOHANSSON ET AL2011;91:57–63 TAVI AND SURVIVAL
AD
ULT
CA
RD
IAC
lderly [13]. However, a considerable number of patientsith aortic stenosis have significant comorbidity [1]. Thatay result in fewer patients being referred for conven-
ional AVR, despite their symptomatic status [14], sincehe comorbidity may pose relative contraindications toardiopulmonary bypass surgery. One feasible alterna-ive to conventional AVR may be the emerging method ofalve implantation with the transcatheter technique [15].he TAVI procedure has been used as a therapeuticption at our department since January 2008, and in theresent study, we evaluated our clinical experience of therst 40 consecutively treated patients.All the patients included in the present study were at
igh surgical risk or presented technical challenges toonventional AVR. Initially, the preferred strategy forAVI at our department was to employ the transfemoralpproach. However, during the initial procedures (Janu-ry 2008 to January 2009), we experienced some severeascular complications in the iliac and femoral arteries.imilar complications have previously been described byl-Attar and colleagues [16], and are probably due to the
arge diameter and stiffness of the introducer deliveryystem, in combination with vascular calcification in this
able 3. Intraoperative and Postoperative Data for Transcathe
ariable
TF Approach
n � 10
rosthesis size23 mm 426 mm 6emodialysis 0rolonged ventilator time 1ew-onset postoperative AF 0evosimendan infusion 0orepinephrine �48 hours 2obutamine �48 hours 2
Mean
ontrast medium, mL 288reatinine peak, �mol/L 179rine output 12 hours, mL 1,300entilator time, hours 16erioperative bank blood, units
CU bank blood, units1.60.8
ength of stayICU, hours, median (IQR) 34 2Total, days, median (IQR) 13 1
nT peak, �g/L 0.2K-MB peak, �g/L 7.7emoglobin, g/LPreoperatively 118At ICU discharge 112
alues given are number and percentage of patients, mean � SD, or medpostoperatively.
F � atrial fibrillation; CK-MB � creatine kinase–myocardial banransapical; TF � transfemoral; TnT � troponin T.
lderly population. A device with a smaller diameter may c
robably reduce the risk of vascular access complica-ions. As a consequence of these vascular complications,he treatment algorithm was changed to the antegrade,ransapical technique as our default strategy. As a result,he TA approach has been used exclusively since Febru-ry 2009, and to date, we have not observed any accessite complications such as wound infection, apex aneu-ysm, heart tamponade, or any bleedings requiringransfusion.
Our findings did not demonstrate any clinically signif-cant differences in postoperative outcome in the inten-ive care unit between the TF and the TA proceduresTable 3), apart from the complications at the access siteescribed above. The amount of contrast medium used in
he TF group during TAVI was significantly higher thann the TA group (p � 0.03). One explanation of this maye that the TA procedure was mainly performed in the
atter stage of the study, and was therefore less influ-nced by a learning curve effect. The increased contrastolume in the TF group was not correlated to impairedenal function, as assessed by urine output or creatinineeak level in the intensive care unit. The higher levels ofardiac enzymes in the TA group (troponin T � 0.5,
ortic Valve Implantation (TAVI) Patients
TA Approach
p Valuen � 30 %
13 4516 552 6.7 1.001 3.3 0.443 10 0.561 3.31 3.3 0.151 3.3 0.15
Mean SD
209 100 0.03152 107 0.52
1,200 550 0.638 10 0.25
.1
.10.90.6
1.41.1
0.250.23
–19) 40 61 (23–5) 0.7523) 7 3 (6–3) 0.11.2 0.5 0.3 0.01.2 20.1 7.8 �0.001
129 14 0.038110 12 0.65
nd interquartile range (IQR). Prolonged ventilator time � ventilator �48
ICU � intensive care unit; IQR � interquartile range; TA �
ter A
%
40600
1000
2020
SD
47143450
2021
6 (231 (7–
03
1613
ian a
reatine kinase–myocardial band � 20.1) compared with
tmsstrIa2M
mfmTsraslctrm3vwtsA
ptuvtasTdvacia
eitaut
scpiATtdb
R
Fw[api
62 JOHANSSON ET AL Ann Thorac SurgTAVI AND SURVIVAL 2011;91:57–63A
DU
LTC
AR
DIA
C
he TF group (troponin T � 0.2, creatine kinase–yocardial band � 7.7) are probably related to the
urgical manipulation of the apex and had no clinicalignificance. Postoperative atrioventricular block withhe need for permanent pacemaker implantation was notequired in any patient in the current study population.n contrast, other studies have presented this complications the most frequent, with an incidence ranging from 4% to9% [17), with the majority of patients receiving theedtronic CoreValve (Medtronic, Minneapolis, MN) [18].Previous studies have demonstrated that EuroSCOREay overestimate the risk of mortality in AVR, especially
or high-risk patients [19, 20]. The predicted 30-dayortality using the logistic EuroSCORE was 25.6% in the
F group and 23.5% in the TA group. These values areimilar to those presented in previous studies, and probablyeaffirm the recent statement by Grossi and coworkers [19]nd Bleiziffer and colleagues [17] that the EuroSCORE is notignificantly correlated with the mortality in a highly se-ected cohort of aged patients with cardiac and noncardiacomorbidity. The actual mortality in the present popula-ion was favorable compared to the predicted mortalityates, with a 30-day mortality of 5.0%, and an in-hospitalortality of 10% in all TAVI patients. In comparison, the
0-day mortality in the propensity score matched con-entional AVR patient population was 10.0% comparedith 2.8% for all 2,262 patients who underwent conven-
ional AVR during the period 1999 to 2009. After propen-ity score matching, the difference between TAVI and
ig 1. Comparisons between 2-year survival estimates for patientsho underwent conventional aortic valve replacement (AVR
straight line]), transapical valve implantation (TA [broken line]),nd transfemoral valve implantation (TF [broken dotted line]) inropensity-matched groups. (TAVI � transcatheter aortic valvemplantation.)
VR was no longer statistically significant (Fig 1).
However, one should bear in mind that the majority ofatients undergoing TAVI have been denied conven-
ional AVR. Forty percent of the current TAVI patientsnderwent percutaneous coronary intervention beforealve implantation, and 10% had previous coronary ar-ery bypass graft surgery. Recently, Dewey and associ-tes [6] demonstrated that coronary artery disease is aignificant risk factor for mortality among patients havingAVI. Patients with coronary artery disease remain can-idates for TAVI, although risk prediction should beiewed with recognition of the influence of coronaryrtery disease on procedural outcome. One of the mainhallenges in the future will be the determination of clearndications for surgical and interventional treatment ofortic stenosis.One limitation of the present study was that our early
xperience of TAVI may have been influenced by learn-ng curve effects, and the potential benefits of TAVI may,herefore, have been underestimated. Propensity scorenalyses can not necessarily account for bias due tonmeasured covariates, and the method may only par-
ially compensate for baseline differences.In conclusion, our results suggest that TAVI can be
afely performed in selected high-risk patients. The vas-ular complications occurring when using the TF ap-roach were probably related to a combination of a wide
ntroducer sheath and heavily calcified femoral arteries.fter propensity score matching, survival with both theA and TF approaches was similar to that after conven-
ional AVR. However, owing to the lack of long-termata, the relationship between TAVI and AVR appears toe complementary rather than substitutive.
eferences
1. Iung B, Baron G, Butchart EG, et al. A prospective survey ofpatients with valvular heart disease in Europe: the EuroHeart Survey on valvular heart disease. Eur Heart J 2003;24:1231–43.
2. Bonow RO, Carabello BA, Kanu C, et al. ACC/AHA 2006guidelines for the management of patients with valvularheart disease. Circulation 2006;114:e84–231.
3. Cribier A, Eltchaninoff H, Tron C, et al. Early experience withpercutaneous transcatheter implantation of heart valve pros-thesis for the treatment of end-stage inoperable patients withcalcific aortic stenosis. J Am Coll Cardiol 2004;43:698–703.
4. Hannan EL, Wu C, Bennett EV, et al. Risk index for predict-ing in-hospital mortality for cardiac valve surgery. AnnThorac Surg 2007;83:921–9.
5. Nashef SA, Roques F, Michel P, Gauducheau E, Lemeshow S,Salamon R. European system for cardiac operative risk evalu-ation (EuroSCORE). Eur J Cardiothorac Surg 1999;16:9–13.
6. Dewey TM, Brown DL, Herbert MA, et al. Effect of concom-itant coronary artery disease on procedural and late out-comes of transcatheter aortic valve implantation. Ann Tho-rac Surg 2010;89:758–67.
7. Rubin DB. On principles for modeling propensity scores inmedical research. Pharmacoepidemiol Drug Saf 2004;13:855–7.
8. D’Agostino RB. Propensity scores in cardiovascular re-search. Circulation 2007;115:2340–3.
9. Lichtenstein SV, Cheung A, Ye J, et al. Transapical trans-catheter aortic valve implantation in humans: initial clinical
experience. Circulation 2006;114:591–6.
1
1
1
1
1
1
1
1
1
1
2
N
TMetdbw
c(
63Ann Thorac Surg JOHANSSON ET AL2011;91:57–63 TAVI AND SURVIVAL
©P
AD
ULT
CA
RD
IAC
0. Walther T, Simon P, Dewey T, et al. Transapical minimallyinvasive aortic valve implantation: multicenter experience.Circulation 2007;116:I240–5.
1. Cribier A, Eltchaninoff H, Tron C, et al. Treatment of calcificaortic stenosis with the percutaneous heart valve: mid-termfollow-up from the initial feasibility studies: the Frenchexperience. J Am Coll Cardiol 2006;47:1214–23.
2. Wendler O, Walther T, Nataf P, et al. Trans-apical aorticvalve implantation: univariate and multivariate analyses ofthe early results from the SOURCE registry. Eur J Cardio-thorac Surg 2010;38:119–27.
3. Sundt TM, Bailey MS, Moon MR, et al. Quality of life afteraortic valve replacement at the age of �80 years. Circulation2000;102:III70–4.
4. Iung B, Cachier A, Baron G, et al. Decision-making in elderlypatients with severe aortic stenosis: why are so many deniedsurgery? Eur Heart J 2005;26:2714–20.
5. Rodes-Cabau J, Webb JG, Cheung A, et al. Transcatheteraortic valve implantation for the treatment of severe symp-
otice From the American Boa
written) examination.
ei
BC2
2011 by The Society of Thoracic Surgeonsublished by Elsevier Inc
surgical risk: acute and late outcomes of the multicenterCanadian experience. J Am Coll Cardiol 2010;55:1080–90.
6. Al-Attar N, Himbert D, Descoutures F, et al. Transcatheteraortic valve implantation: selection strategy is crucial foroutcome. Ann Thorac Surg 2009;87:1757–62.
7. Bleiziffer S, Ruge H, Mazzitelli D, et al. Results of percuta-neous and transapical transcatheter aortic valve implanta-tion performed by a surgical team. Eur J Cardiothorac Surg2009;35:615–21.
8. Tchetche D, Dumonteil N, Sauguet A, et al. Thirty-dayoutcome and vascular complications after transarterial aorticvalve implantation using both Edwards Sapien andMedtronic CoreValve bioprostheses in a mixed population.EuroIntervention 2010;5:659–65.
9. Grossi EA, Schwartz CF, Yu PJ, et al. High-risk aortic valvereplacement: are the outcomes as bad as predicted? AnnThorac Surg 2008;85:102–6.
0. Osswald BR, Gegouskov V, Badowski-Zyla D, et al. Overes-timation of aortic valve replacement risk by EuroSCORE:implications for percutaneous valve replacement. Eur
tomatic aortic stenosis in patients at very high or prohibitive Heart J 2009;30:74–80.
rd of Thoracic Surgery
he 2011 Part I (written) examination will be held ononday, November 21, 2011. It is planned that the
xamination will be given at multiple sites throughouthe United States using an electronic format. The closingate for registration is August 15, 2011. Those wishing toe considered for examination must apply online atww.abts.org.To be admissible to the Part II (oral) examination, a
andidate must have successfully completed the Part I
A candidate applying for admission to the certifyingxamination must fulfill all the requirements of the Boardn force at the time the application is received.
Please address all communications to the Americanoard of Thoracic Surgery, 633 N St. Clair St, Suite 2320,hicago, IL 60611; telephone: (312) 202-5900; fax: (312)02-5960; e-mail: [email protected].
Ann Thorac Surg 2011;91:63 • 0003-4975/$36.00
