original article
The
new engl a nd jour nal
o f
medicine
n engl j med 366;18 nejm.org may 3, 2012
1686
Two-Year Outcomes after Transcatheter
or Surgical Aortic-Valve Replacement
Susheel K. Kodali, M.D., Mathew R. Williams, M.D., Craig R. Smith, M.D.,
Lars G. Svensson, M.D., Ph.D., John G. Webb, M.D., Raj R. Makkar, M.D.,
Gregory P. Fontana, M.D., Todd M. Dewey, M.D., Vinod H. Thourani, M.D.,
Augusto D. Pichard, M.D., Michael Fischbein, M.D., Ph.D., Wilson Y. Szeto, M.D.,
Scott Lim, M.D., Kevin L. Greason, M.D., Paul S. Teirstein, M.D.,
S. Chris Malaisrie, M.D., Pamela S. Douglas, M.D., Rebecca T. Hahn, M.D.,
Brian Whisenant, M.D., Alan Zajarias, M.D., Duolao Wang, Ph.D.,
Jodi J. Akin, M.S., William N. Anderson, Ph.D., and Martin B. Leon, M.D.,
for the PARTNER Trial Investigators*
From Columbia University Medical Cen-
ter and New York Presbyterian Hospital
(S.K.K., M.R.W., C.R.S., R.T.H., M.B.L.)
and Lenox Hill Hospital (G.P.F.) — both
in New York; Cleveland Clinic Foundation,
Cleveland (L.G.S.); University of British
Columbia and St. Paul’s Hospital, Vancou-
ver, Canada (J.G.W.); Cedars–Sinai Medi-
cal Center, Los Angeles (R.R.M.); Medi-
cal City Dallas, Dallas (T.M.D.); Emory
University School of Medicine, Atlanta
(V.H.T.); Washington Hospital Center,
Washington, DC (A.D.P.); Stanford Uni-
versity Medical School, Palo Alto (M.F.),
Scripps Clinic, La Jolla (P.S.T.), and Ed-
wards Lifesciences, Irvine (J.J.A., W.N.A.)
— all in California; Hospital of the Uni-
versity of Pennsylvania, Philadelphia
(W.Y.S.); University of Virginia, Charlottes-
ville (S.L.); Mayo Clinic, Rochester, MN
(K.L.G.); Northwestern University, Chi-
cago (S.C.M.); Duke University Medical
Center, Durham, NC (P.S.D.); Intermoun-
tain Medical Center, Salt Lake City (B.W.);
Washington University Medical School
and Barnes-Jewish Hospital, St. Louis
(A.Z.); and London School of Hygiene
and Tropical Medicine, London (D.W.).
Address reprint requests to Dr. Kodali at
Columbia University Medical Center/
New York Presbyterian Hospital, 161 Fort
Washington Ave., 6th Fl., New York, NY
10032, or at sk2427@columbia.edu.
* The investigators, institutions, and re-
search organizations participating in the
Placement of Aortic Transcatheter Valves
(PARTNER) trial are listed in the Supple-
mentary Appendix, available at NEJM.org.
This article (10.1056/NEJMoa1200384) was
published on March 26, 2012, at NEJM.org.
N Engl J Med 2012;366:1686-95.
Copyright © 2012 Massachusetts Medical Society.
ABSTR ACT
Background
The Placement of Aortic Transcatheter Valves (PARTNER) trial showed that among
high-risk patients with aortic stenosis, the 1-year survival rates are similar with trans-
catheter aortic-valve replacement (TAVR) and surgical replacement. However, longer-
term follow-up is necessary to determine whether TAVR has prolonged benefits.
Methods
At 25 centers, we randomly assigned 699 high-risk patients with severe aortic stenosis
to undergo either surgical aortic-valve replacement or TAVR. All patients were fol-
lowed for at least 2 years, with assessment of clinical outcomes and echocardiographic
evaluation.
Results
The rates of death from any cause were similar in the TAVR and surgery groups
(hazard ratio with TAVR, 0.90; 95% confidence interval [CI], 0.71 to 1.15; P = 0.41) and
at 2 years (Kaplan–Meier analysis) were 33.9% in the TAVR group and 35.0% in the
surgery group (P = 0.78). The frequency of all strokes during follow-up did not differ
significantly between the two groups (hazard ratio, 1.22; 95% CI, 0.67 to 2.23;
P = 0.52). At 30 days, strokes were more frequent with TAVR than with surgical re-
placement (4.6% vs. 2.4%, P = 0.12); subsequently, there were 8 additional strokes in
the TAVR group and 12 in the surgery group. Improvement in valve areas was similar
with TAVR and surgical replacement and was maintained for 2 years. Paravalvular
regurgitation was more frequent after TAVR (P<0.001), and even mild paravalvular
regurgitation was associated with increased late mortality (P<0.001).
Conclusions
A 2-year follow-up of patients in the PARTNER trial supports TAVR as an alternative
to surgery in high-risk patients. The two treatments were similar with respect to
mortality, reduction in symptoms, and improved valve hemodynamics, but paraval-
vular regurgitation was more frequent after TAVR and was associated with in-
creased late mortality. (Funded by Edwards Lifesciences; ClinicalTrials.gov number,
NCT00530894.)
The New England Journal of Medicine
Downloaded from nejm.org at LONDON SCH HYGIENE & TROPICAL MED on January 21, 2014. For personal use only. No other uses without permission.
Copyright © 2012 Massachusetts Medical Society. All rights reserved.
Transcatheter vs. Surgical Aortic-Valve Replacement
n engl j med 366;18 nejm.org may 3, 2012
1687
A
ortic stenosis is associated with
high mortality after the appearance of car-
diac symptoms.
1
Nevertheless, many pa-
tients do not undergo surgical aortic-valve replace-
ment owing to real or perceived increased risks
associated with surgery.
2-5
Transcatheter aortic-
valve replacement (TAVR) has emerged as an al-
ternative therapy in high-risk patients with aortic
stenosis.
6-10
Observational registries from vari-
ous countries have reported 1-month and 1-year
outcomes after TAVR,
11-14
but there are limited
long-term follow-up data.
15
The Placement of Aortic Transcatheter Valves
(PARTNER) trial was a randomized trial compar-
ing TAVR with standard-of-care therapies in high-
risk patients with aortic stenosis. One-year mor-
tality outcomes from PARTNER showed that TAVR
was superior to standard therapy in patients who
could not undergo surgery
16
and was noninferior
to surgical replacement in high-risk patients who
could undergo surgery.
17
However, longer-term
data are required to assess valve durability and
to monitor late clinical complications, before
TAVR is used more widely in clinical practice.
This report describes the 2-year (and longer) clini-
cal outcomes and echocardiographic findings after
TAVR or surgical aortic-valve replacement in the
high-risk patients in the PARTNER trial who could
undergo surgery.
Methods
Patients
Patient selection for this cohort of the PARTNER
trial has been described previously.
17
Inclusion cri-
teria were severe symptomatic aortic stenosis (an
aortic-valve area ≤0.8 cm
2
plus a peak velocity ≥4 m
per second or a mean valve gradient ≥40 mm Hg)
and high-risk status for surgical aortic-valve re-
placement, as determined by experienced surgeons.
Patients were considered to be at high surgical
risk if they had coexisting conditions that were
associated with a risk of death of at least 15% by
30 days after the operation.
Study Device and Procedure
The SAPIEN heart-valve system (Edwards Life-
sciences) and the TAVR procedure have been de-
scribed previously.
16,17
Most procedures were per-
formed in a hybrid operating room with a fixed
fluoroscopic imaging system, while the patient was
under general anesthesia, and with transesopha-
geal echocardiography. Transapical TAVR was per-
formed through a small intercostal incision over
the left ventricular apex with the use of a dedicat-
ed delivery catheter and the same SAPIEN valve.
Heparin was administered during the proce-
dure, and dual antiplatelet therapy (aspirin and
clopidogrel) was recommended for 6 months af-
terward. The outpatient regimen was frequently
modified by the treating physicians because of
increased bleeding risks.
Study Design and Oversight
The study design and data-management practices
have been described previously.
16,17
A total of 699
patients from 25 sites were randomly assigned to
TAVR or surgical replacement. Patients assigned
to TAVR were treated by either the transfemoral or
transapical approach on the basis of whether pe-
ripheral arteries could accommodate the large
sheaths required (22 French for the 23-mm valve
and 24 French for the 26-mm valve). Patients who
were randomly assigned to surgical replacement
were stratified according to whether a transfemo-
ral or transapical approach would have been used.
The study was designed and monitored by the
sponsor, Edwards Lifesciences, and the executive
committee, which included four interventional
cardiologists and four cardiac surgeons. The spon-
sor funded the study and participated in the se-
lection and management of the sites, the collec-
tion of the data, and data monitoring. The first
author and members of the executive committee
had unrestricted access to the data after the data-
base had been locked and prepared all drafts of
the manuscript; they attest to the completeness
and accuracy of the reported data and to the ad-
herence of the study to the protocol (available with
the full text of this article at NEJM.org). The trial
was approved by the institutional review board
at each site. Written informed consent was ob-
tained from all patients.
Study End Points
The prespecified primary end point of the PART-
NER trial was all-cause mortality at 1 year for the
pooled cohort. Prespecified secondary end points
included cardiovascular mortality, stroke, repeat
hospitalization, acute kidney injury, vascular com-
plications, bleeding events, and New York Heart
Association (NYHA) functional class. All patients
were followed for at least 2 years and had annual
clinical visits and echocardiographic evaluations.
Crossovers between the two treatment groups were
not permitted. A clinical-events committee was
The New England Journal of Medicine
Downloaded from nejm.org at LONDON SCH HYGIENE & TROPICAL MED on January 21, 2014. For personal use only. No other uses without permission.
Copyright © 2012 Massachusetts Medical Society. All rights reserved.
The
ne w engl and jour nal
o f
medicine
n engl j med 366;18 nejm.org may 3, 2012
1688
responsible for adjudicating all end points. Defi-
nitions of the end points are identical to those in
the original trial and have been reported else-
where.
16,17
Statistical Analysis
For data analyses, the intention-to-treat analysis
started at the time of randomization, and the as-
treated analysis started at the time of induction
of anesthesia in the procedure room. All clinical
outcomes were primarily analyzed with the use
of an intention-to-treat analysis, but the results
of as-treated analyses are also presented for com-
parison. All echocardiographic analyses were per-
formed with the use of the as-treated data. Cat-
egorical variables were compared with the use of
Fisher’s exact test. Continuous variables, presented
as means ±SD, were compared with the use of
Student’s t-test. Survival curves for time-to-event
variables were constructed with the use of Kap-
lan–Meier estimates based on all available data
and were compared with the use of the log-rank
test. To study the effect of risk factors on mortal-
ity, Cox proportional-hazards regression was per-
formed. For the multivariable analyses, multiple
imputations were used to accommodate missing
baseline variables. The multivariable models in-
cluded covariates with a P value of less than 0.20
in univariate analyses. An additional time-depen-
dent covariate analysis was performed to test the
association of complications during TAVR or sur-
gical replacement with subsequent mortality. All
statistical analyses were performed with the use
of SAS software, version 9.2.
Results
Patients
In the randomized TAVR group, 244 patients had
acceptable vascular access and were treated by
means of the transfemoral approach, and the re-
maining 104 patients were treated by means of
the transapical approach. Surgical replacement
was performed in 351 patients. Figure 1 in the
Supplementary Appendix, available at NEJM.org,
shows the study-group assignments and follow-
up. All patients were followed for at least 2 years
(median, 727 days; maximum, 1490 days). The
overall study population was elderly (mean age,
84.1±6.6 years), had severe cardiac symptoms
(94.1% had NYHA class III or IV status), and had
frequent coexisting conditions (75.5% had a his-
tory of coronary artery disease, 43.0% had a his-
tory of coronary-artery bypass surgery, 42.4% had
peripheral vascular disease, 43.3% had pulmo-
nary disease, and 41.3% had diabetes). The TAVR
and surgery groups were generally well matched
with regard to baseline characteristics (
Table 1
in the Supplementary Appendix), except for a
slightly higher incidence of renal dysfunction in
the TAVR group (creatinine level >2 mg per deci-
liter [177 µmol per liter]: 10.8%, as compared
with 6.4% in the surgery group; P = 0.04). The
mean Society of Thoracic Surgeons predicted risk
of death at 30 days was 11.8±3.4%.
Of the 699 study patients, 42 did not receive
the assigned therapy: 4 in the TAVR group and 38
in the surgery group.
17
The main reasons for non-
treatment were withdrawal from the study and
the patient’s decision not to undergo surgery (28
patients).
Mortality and Stroke
Outcomes at 30 days and 1 year have been de-
scribed previously.
17
For the duration of the trial,
there were no significant differences in survival
between the TAVR and surgery groups in either
the intention-to-treat analysis (hazard ratio with
TAVR, 0.90; 95% confidence interval [CI], 0.71 to
1.15; P = 0.41) or the as-treated analysis (hazard
ratio, 0.98; 95% CI, 0.76 to 1.25; P = 0.85) (Fig. 1).
Between 1 and 2 years, there were 32 additional
deaths in the TAVR group and 25 in the surgery
group. At 2 years, there were no significant dif-
ferences in mortality from any cause between the
TAVR group (33.9%; 95% CI, 28.9 to 39.0) and
the surgery group (35.0%; 95% CI, 29.8 to 40.2;
P = 0.78) (
Table 1
). Cardiovascular mortality at
2 years was also similar in the TAVR and surgery
groups (21.4% [95% CI, 16.8 to 26.0] and 20.5%
[95% CI, 15.8 to 25.3], respectively; P = 0.80). Sim-
ilarly, in the as-treated analysis, the TAVR and sur-
gery groups did not differ significantly with re-
spect to all-cause mortality (33.9% and 32.7%,
respectively; P = 0.75) or cardiovascular mortality
(20.8% and 18.5%, respectively; P = 0.50) (
Table 2
in the Supplementary Appendix).
Between 1 and 2 years, eight strokes occurred
(four in the TAVR group and four in the surgery
group) and three transient ischemic attacks (two
in the TAVR group and one in the surgery group).
The frequency of all neurologic events (strokes and
transient ischemic attacks) at 2 years was higher
with TAVR than with surgical replacement (11.2%
The New England Journal of Medicine
Downloaded from nejm.org at LONDON SCH HYGIENE & TROPICAL MED on January 21, 2014. For personal use only. No other uses without permission.
Copyright © 2012 Massachusetts Medical Society. All rights reserved.
Transcatheter vs. Surgical Aortic-Valve Replacement
n engl j med 366;18 nejm.org may 3, 2012
1689
vs. 6.5%, P = 0.05). However, there was no signifi-
cant difference in the number of overall strokes
between the TAVR and surgery groups (hazard
ratio, 1.22; 95% CI, 0.67 to 2.23; P = 0.52) (Fig. 1).
After the early increased hazard of stroke in the
first 30 days associated with TAVR (4.6% with
TAVR vs. 2.4% with surgical replacement, P = 0.12),
there were 8 additional strokes in the TAVR group
and 12 in the surgery group, such that the total
number of strokes over the follow-up period (36
months) was 24 in the TAVR group and 20 in the
surgery group. The composite of the rate of death
from any cause or stroke did not differ signifi-
cantly between the two treatment groups (hazard
ratio, 0.93; 95% CI, 0.73 to 1.18; P = 0.55) (Fig. 1);
at 2 years, the rate was 37.1% in the TAVR group
and 36.4% in the surgery group (P = 0.85).
Other Clinical Outcomes
Other clinical events are summarized in
Table 1
.
Major vascular complications and major bleeding
events were frequent procedure-related complica-
tions in the TAVR and surgery groups, respec-
tively, but after 1 year, these events were uncom-
mon and did not differ significantly between the
groups. No patients were treated with balloon aor-
tic valvuloplasty or repeat TAVR between 1 and
2 years. Endocarditis was rare and occurred at a
similar rate in the two groups (1.5% in the TAVR
group and 1.0% in the surgery group, P = 0.61).
No patients in either group had structural valve
deterioration requiring surgical replacement dur-
ing follow-up.
At 2 years, there was no significant difference
in the rate of repeat hospitalization between the
Death from Any Cause (%)
60
40
30
10
50
20
0
0 6 12 18 24 30 36
Month
A Death from Any Cause, Intention-to-Treat Population
Hazard ratio, 0.90 (95% CI, 0.71–1.15)
P=0.41
No. at Risk
TAVR
Surgery
348
351
298
252
234
217
260
236
172
165
70
65
31
32
Death from Any Cause (%)
60
40
30
10
50
20
0
0 6 12 18 24 30 36
Month
B Death from Any Cause, As-Treated Population
Hazard ratio, 0.98 (95% CI, 0.76–1.25)
P=0.85
No. at Risk
TAVR
Surgery
344
313
291
243
232
211
259
229
155
143
70
63
29
28
Event Rate (%)
60
40
30
10
50
20
0
0 6 12 18 24 30 36
Month
C Stroke, Intention-to-Treat Population
Hazard ratio, 1.22 (95% CI, 0.67–2.23)
P=0.52
No. at Risk
TAVR
Surgery
348
351
287
246
224
211
249
230
162
160
65
62
28
31
Event Rate (%)
60
40
30
10
50
20
0
0 6 12 18 24 30 36
Month
D Death from Any Cause or Stroke, Intention-to-Treat Population
Hazard ratio, 0.93 (95% CI, 0.73–1.18)
P=0.55
No. at Risk
TAVR
Surgery
348
351
291
247
230
213
254
232
168
162
68
63
29
31
TAVR
TAVR
Surgery
Surgery
Surgery
Surgery
TAVR
TAVR
Figure 1. Time-to-Event Curves for the Primary and Other Selected End Points.
Events were calculated with the use of Kaplan–Meier methods and compared with the use of a log-rank test. TAVR denotes transcatheter
aortic-valve replacement.
The New England Journal of Medicine
Downloaded from nejm.org at LONDON SCH HYGIENE & TROPICAL MED on January 21, 2014. For personal use only. No other uses without permission.
Copyright © 2012 Massachusetts Medical Society. All rights reserved.
The
ne w engl and jour nal
o f
medicine
n engl j med 366;18 nejm.org may 3, 2012
1690
TAVR and surgery groups (24.7% and 21.7%, re-
spectively; P = 0.41). Among survivors at 2 years,
the mean NYHA class was similar in the TAVR
and surgery groups (1.72 and 1.70, P = 0.87), and
the majority of patients in both groups had NYHA
class I or II status (83.9% in the TAVR group and
85.2% in the surgery group). Clinical outcomes in
the subgroups of patients in whom a transfemo-
ral or transapical approach was used are shown
in Tables 3 through 6 in the Supplementary Ap-
pendix.
Predictors of Mortality
Predictors of mortality for the overall trial cohort
as well as for each of the randomized groups are
presented in
Table 2
. Treatment assignment was
not a significant predictor of the risk of death.
The time-dependent effect of treatment com-
plications on subsequent mortality was also eval-
uated. Strokes greatly increased the hazard of
death in both groups (TAVR: hazard ratio, 2.47;
95% CI, 1.42 to 4.30; P<0.001; surgery: hazard
ratio, 5.20; 95% CI, 3.07 to 8.80; P<0.001). Major
Table 1. Clinical Outcomes at 1 Year and 2 Years with TAVR or Surgery (Intention-to-Treat Population).*
Outcome 1 Year 2 Years
Surgery
(N = 351)
TAVR
(N = 348) P Value†
Surgery
(N = 351)
TAVR
(N = 348) P Value†
no. of patients (%) no. of patients (%)
Death
From any cause 89 (26.8) 84 (24.3) 0.45 114 (35.0) 116 (33.9) 0.78
From cardiovascular causes 40 (13.0) 47 (14.3) 0.63 59 (20.5) 67 (21.4) 0.80
Repeat hospitalization‡ 51 (17.7) 59 (18.6) 0.78 60 (21.7) 74 (24.7) 0.41
Death from any cause or repeat
hospitalization‡
125 (37.7) 121 (34.9) 0.45 152 (46.5) 159 (46.6) 0.99
Stroke or TIA§
All 13 (4.3) 28 (8.7) 0.03 18 (6.5) 34 (11.2) 0.05
Stroke 10 (3.2) 20 (6.0) 0.08 14 (4.9) 24 (7.7) 0.17
TIA 4 (1.5) 8 (2.6) 0.32 5 (2.0) 10 (3.6) 0.26
Death from any cause or stroke 95 (28.6) 95 (27.4) 0.74 119 (36.4) 127 (37.1) 0.85
Myocardial infarction 2 (0.6) 0 0.16 4 (1.5) 0 0.05
Major vascular complication¶ 13 (3.8) 39 (11.3) <0.001 13 (3.8) 40 (11.6) <0.001
Major bleeding‖ 88 (26.7) 52 (15.7) <0.001 95 (29.5) 60 (19.0) 0.002
Endocarditis 3 (1.0) 2 (0.6) 0.63 3 (1.0) 4 (1.5) 0.61
Renal failure** 20 (6.5) 18 (5.4) 0.57 21 (6.9) 20 (6.2) 0.75
New pacemaker 16 (5.0) 21 (6.4) 0.44 19 (6.4) 23 (7.2) 0.69
SVD requiring surgical replacement 0 0 0 0
* All percentages are Kaplan–Meier estimates at the specific time point and thus do not equal the number of patients
divided by the total number in the study group. SVD denotes structural valve deterioration, TAVR transcatheter aortic-
valve replacement, and TIA transient ischemic attack.
† P values are for between-group comparisons of the frequency of the event at each time point.
‡ Repeat hospitalizations were included in the analysis if they were for symptoms of heart failure, angina, or syncope
due to aortic-valve disease that required aortic-valve intervention or intensified medical management.
§ Stroke was defined as a neurologic deficit lasting more than 24 hours or lasting less than 24 hours with a brain-imag-
ing study showing infarction.
¶ Major vascular complications were defined as thoracic aortic dissection; access-site or access-related vascular injury
leading to death, the need for substantial blood transfusion (>3 units), or percutaneous or surgical intervention; and
distal embolization (noncerebral) from a vascular source requiring surgery or amputation or resulting in irreversible
end-organ damage.
‖ Major bleeding was defined as any episode of major internal or external bleeding that caused death, hospitalization,
or permanent injury or that necessitated the transfusion of at least 3 units of packed red cells or a pericardiocentesis
procedure.
** Renal failure was defined as any condition requiring the initiation of any dialysis.
The New England Journal of Medicine
Downloaded from nejm.org at LONDON SCH HYGIENE & TROPICAL MED on January 21, 2014. For personal use only. No other uses without permission.
Copyright © 2012 Massachusetts Medical Society. All rights reserved.
