Myocardial Viability and Survival
in Ischemic Left Ventricular Dysfunction
Robert O. Bonow, MD
On behalf of the STICH Trial Investigators
STICH Financial Disclosures
Original Recipient Institution
Principal Investigator
Activity
Duke University Medical Center
Robert H. Jones
Clinical Coordinating Ctr
Duke University Medical Center
Kerry L. Lee
Statistical and Data CC
Northwestern University
Robert O. Bonow
Radionuclide Core Lab
Washington Hospital Center
Julio A. Panza
Dobutamine Echo Core
Mayo Clinic
Jae K. Oh
Echo Core Laboratory
Univ of Alabama-Birmingham
Gerald M. Pohost
CMR Core Laboratory
University of Pittsburgh
Arthur M. Feldman
NCG Core Laboratory
Baylor University Medical Ctr
Paul Grayburn
MR TEE Substudy
Duke University Medical Center
Daniel B. Mark
EQOL Core Laboratory
Funding Sources:
National Heart, Lung and Blood Institute 98%
Abbott Laboratories 2%
Background
• LV dysfunction in patients with CAD is not
always an irreversible process, as LV function
may improve substantially after CABG
• Assessment of myocardial viability is often
used to predict improvement in LV function after
CABG and thus select patients for CABG
• Numerous studies have suggested that
identification of viable myocardium also predicts
improved survival after CABG
Limitations of Cohort Studies
• Decision for CABG may have been influenced
by viability status
• No (or inadequate) adjustment for key baseline
variables (age, comorbidities)
• Cohort studies carried out before modern
aggressive medical therapy
STICH Revascularization Hypothesis
• The first prospective randomized trial testing the
hypothesis that CABG improves survival in
patients with ischemic LV dysfunction compared
to outcome with aggressive medical therapy
• Provides the first opportunity to assess the
interaction between myocardial viability and
survival in randomized patients who were all
eligible for medical management alone and
also eligible for CABG.
STICH Viability Hypothesis
In this prospective substudy, we tested the
hypothesis that assessment of myocardial
viability identifies patients with CAD and LV
dysfunction who have the greatest survival
benefit with CABG compared to aggressive
medical therapy
STICH Viability Hypothesis
• All randomized patients were eligible for
viability testing with SPECT myocardial
perfusion imaging or dobutamine echo.
• Viability testing was optional at enrolling
sites and was not a prerequisite for
enrollment.
STICH Viability Hypothesis
SPECT protocols:
• Thallium-201 stress-redistribution-reinjection
• Thallium-201 rest-redistribution
• Nitrate-enhanced Tc-99m perfusion imaging
Dobutamine echo protocols:
• Staged increase in dobutamine starting at
5 μg/kg/min
STICH Viability Hypothesis
Criteria for myocardial viability were prospective
and pre-specified
SPECT:
• 17 segment model
• ≥11 segments manifesting viability based on
relative tracer activity
Dobutamine echo:
• 16 segment model
• ≥5 segments with dysfunction at rest
manifesting contractile reserve with dobutamine
STICH Viability Hypothesis
Primary endpoint:
▪ All-cause mortality
Secondary endpoints:
▪ Mortality plus cardiovascular hospitalization
▪ Cardiovascular mortality
Intention-to-treat analysis
Patients randomized in STICH
Revascularization Hypothesis
1212
Patients with
myocardial
viability test
618
611
Patients with no
usable myocardial
viability test
Unusable test
• Timing
• Poor quality
17
Patients with
usable myocardial
viability test
594
Patients with no
myocardial
viability test
601
Patients randomized in STICH
Revascularization Hypothesis
1212
SPECT
n=471
Dobutamine echo
n=280
321 150 130
611
Patients with
usable myocardial
viability test
601
114
487
Viable
Nonviable
Patients with no
usable myocardial
viability test
Baseline Characteristics
Patients With and Without Myocardial Viability
Variable
Viable
(n=487)
Non-Viable
(n=114)
P value
Age
61 ± 10
61 ± 9
NS
Multivessel CAD
73%
73%
NS
Proximal LAD stenosis
64%
70%
NS
Risk score *
12.4 ± 8.7
12.9 ± 9.3
NS
Previous MI
76.6%
94.7%
<0.001
LV ejection fraction (percent)
28 ± 8
23 ± 9
<0.001
LV end-diastolic volume index (ml/m2)
117 ± 37
147 ± 53
<0.001
LV end-systolic volume index (ml/m2)
86 ± 33
116 ± 50
<0.001
* Significant covariates in risk model: Age, renal function, heart failure,
ejection fraction, CAD index, mitral regurgitation, stroke
Myocardial Viability and Mortality
1.0
Without viability
Variables associated with mortality
With viability
Mortality Rate
0.8
HR
0.64
95% CI
P
0.48,0.86 0.003
0.6
Risk score
LV ejection fraction
LV EDVI
LV ESVI
Myocardial viability
Chi-square
p
33.26
24.80
35.36
33.90
8.54
<0.001
<0.001
<0.001
<0.001
0.003
50%
0.4
33%
0.2
0.0
Without viability
With viability
0
1
114
487
99
432
2
3
4
Years from Randomization
85
409
80
371
63
294
5
6
36
188
16
102
Myocardial Viability and Mortality
Variable
No.
Univariate
Chi-square
p value
Multivariable
Chi-square
p value
SPECT and/or DE 601
8.54
0.003
1.57
0.210
SPECT alone
471
7.35
0.007
0.58
0.444
DE alone
280
1.18
0.277
0.42
0.518
Myocardial Viability and Cardiovascular Mortality
1.0
Without viability
Cardiovascular Mortality Rate
With viability
Univariate
0.8
HR
0.61
95% CI
P
0.44,0.84 0.003
Chi-square p value
8.81
0.003
Multivariable
Chi-square
p value
0.91
0.339
0.6
43%
0.4
29%
0.2
0.0
Without viability
With viability
0
1
114
487
99
432
2
3
4
Years from Randomization
85
409
80
371
63
294
5
6
36
188
16
102
Myocardial Viability and Mortality + CV Hospitalization
Mortality and CV Hospitalization Rate
1.0
Without viability
82%
With viability
0.8
HR
0.59
95% CI
P
0.47,0.74 0.001
0.6
63%
0.4
Univariate
Multivariable
HR
95% CI
P
Chi-square p value
Chi-square<0.001
p value
0.59 0.47,0.44
0.2
20.27
<0.001
8.60
0.003
5
6
14
94
5
41
0.0
Without viability
With viability
0
1
114
487
56
327
2
3
4
Years from Randomization
41
284
34
238
22
166
Patients with
viability tests
601
Patients with
myocardial viability
487
Patients without
myocardial viability
114
243
244
60
54
MED
49.9%
CABG
50.1%
MED
52.6%
CABG
47.4%
Baseline Characteristics
Viable (n=487)
Non-Viable (n=114)
Variable
MED
(n=243)
CABG
Variable P value
(n=244)
MED
(n=60)
CABG
(n=54)
P value
Age
60 ± 10
62 ± Age
9
62 ± 9
60 ± 9
NS
NS
Gender (% male)
84%
86%Gender (% NS
male)
92%
93%
NS
Previous MI
78%
75%Previous MINS
93%
96%
NS
Multivessel CAD
72%
73%MultivesselNS
CAD
68%
78%
NS
Proximal LAD
65%
63%Proximal LAD
NS
70%
70%
NS
13.7 ± 9.8
12.9 ± 9.3
NS
Risk score *
11.9 ± 8.4
12.8 ± Risk
903 score NS
*
LV EF (percent)
28 ± 8
27± LV
8 EF (percent)
NS
23 ± 9
23 ± 9
NS
LV EDVI (ml/m2)
118 ± 38
116 ±LV
35 EDVI (ml/m
NS 2)
151 ± 51
140 ± 54
NS
LV ESVI (ml/m2)
86 ± 34
86 ±LV
32ESVI (ml/m
NS 2)
121 ± 50
111 ± 51
NS
* Significant covariates in risk model: Age, renal function,
heart failure, ejection fraction, CAD index, MR, stroke
Myocardial Viability and Mortality
Without Viability
1.0
Mortality Rate
0.8
With Viability
MED (33 deaths)
MED (95 deaths)
CABG (25 deaths)
CABG (83 deaths)
56%
0.6
35%
0.4
42%
31%
0.2
0.0
0
1
2
3
4
5
Years from Randomization
MED
60
51
44
39
29
CABG
54
48
41
41
34
6
0
1
2
3
4
5
Years from Randomization
6
14
4
243
219
206
179
146
94
51
22
12
244
213
203
192
148
94
51
Myocardial Viability and Mortality
Without Viability
1.0
Mortality Rate
0.8
With Viability
MED (33 deaths)
MED (95 deaths)
CABG (25 deaths)
CABG (83 deaths)
56%
0.6
35%
0.4
42%
31%
0.2
0.0
0
1
2
3
4
5
Years from Randomization
6
0
1
2
3
4
5
Years from Randomization
Subgroup
N
Deaths
HR
95% CI
Interaction
P value
Without viability
114
58
0.70
0.41, 1.18
0.528
With viability
487
178
0.86
0.64, 1.16
0.25
0.5
CABG
better
1
2
MED
better
6
Interaction of Viability and Treatment on CV Outcomes
Endpoint
Mortality
Mortality or CV
hospitalization
CV mortality
Events Treatment
236
422
187
p value
As randomized
0.528
As treated
0.962
As randomized
0.390
As treated
0.975
As randomized
0.697
As treated
0.261
STICH Viability Hypothesis
Limitations:
• Lack of viability data in all patients; patients
represent a subpopulation of STICH
• Analysis limited to SPECT and dobutamine
echo, not PET or cardiac MRI
STICH Viability Hypothesis
• STICH represents the largest report to date
relating myocardial viability to clinical outcomes
of patients with CAD and LV dysfunction
• … and is the first to assess these relationships
prospectively among patients who were all
eligible for CABG as well as optimal medical
management alone
STICH Viability Hypothesis
STICH results:
…demonstrate a significant association between
myocardial viability and outcome, but this association
is rendered non-significant when subjected to a
multivariable analysis that includes other prognostic
variables.
…fail to demonstrate a significant interaction between
myocardial viability and medical versus surgical
treatment with respect to mortality, whether assessed
according to treatment assigned (intention to treat) or
to the treatment actually received.
STICH Viability Hypothesis
Implications:
In patients with CAD and LV dysfunction,
assessment of myocardial viability does not
identify patients who will have the greatest
survival benefit from adding CABG to
aggressive medical therapy
Full report available at www.NEJM.org