why clinical trials of hfpef have failed : is it due to...
TRANSCRIPT
Why Clinical Trials of HFpEF have
Failed Is it due to Comorbidities
Grant Support andor Consulting from
NIHNHLBI Cardiorentis Amgen Roche Diagnostics Medtronic Otsuka Singulex
Christopher M OrsquoConnor MD
Stack Distinguished Professor of Medicine
Chief of Cardiology
Director Duke Heart Center
Normal
Systolic
Heart
Failure
Diastolic
Heart Failure
Aurigemma Zile Gaasch
Circulation 2005
0
1
2
3
4
5
6
7
8
9
10
pre
val
ence
Prevalence of Heart Failure
age range
mean age
66-103
78
75
75
gt 50
-
gt 40
60
55-95
65
75-86
-
70-84
76
75
51
England
(Poole)
64
45
Den
(Copen)
49
29
Spain
(Asturias)
82
42
Finland
(Helsinki)
Portugal
(EPICA)
17
42
gt25
68
2
1
15
Nether
(Rotter)
67
Sweden
(Vasteras)
31
USA
(CHS)
88
48
Proportion with decreased
LV systolic function
Proportion with preserved
LV systolic function
Why Target
Co-Morbidities
Heterogeneity of population problematic
No therapeutic advances in broad approach
Comorbidities associated with high risk
May be differential association in HFpEF and
HFrEF
European Heart Failure Pilot Survey
3226 outpatients with chronic HF
74 had a least one co-morbidity
CKD (41) anemia (29) and diabetes (29)
Co-morbidities were independently associated
with higher age higher NYHA class ischemic
etiology of HF higher heart rate HTN and AF
Only diabetes CKD and anemia were
independently associated with a higher risk of
mortality andor HF hospitalization
There were marked regional differences in
prevalence and prognostic implications of co-
morbidities
van Deursen VM Eur J Heart Fail 2014 Jan16(1)103-11
Mentz RJ OrsquoConnor CM et al JACC 2014
HFpEF vs HFrEF
ADHERE Registry(2) GWTG Registry(8)
Reduced Preserved EFlt40 EFge50
Age (y) 70 plusmn 14 74 plusmn 13 70 (58-80) 78 (67-85)
Female 40 62 36 63
African American 22 17 25 16
Medical History
COPD or asthma 27 31 27 33
CRI 26 26 48 52
Anemia - - 14 22
Diabetes mellitus 40 45 22 oral
18 insulin
24 oral
22 insulin
Mentz RJ OrsquoConnor CM et al JACC 2014
Bidirectional Impact
Mentz RJ OrsquoConnor CM et al JACC 2014
HFpEF and Comorbidities
Paulus WJ et al JACC 2013
PATHOPHYSIOLOGIC TARGETS
Diastolic dysfunction
Ventriculo-arterial disociation
Pulmonary Hypertension
Chronotropic Incompetance
Systemic Hypertension
HFpEF ndash Clinical Trials to Date
HFpEF ndash Landmark Trials
No proven therapy
All treatment recommendations ndash Class C evidence
Completed
Dig (Mortalityharr HF Hospitalizationsdarr)
CHARM-Preserved (trend towards darr HF hosp CV mortality)
PEP-CHF (trend towards darr mortality hf hospitalization frac14 of pts withdrew to go on open label ACEI at 1 year)
SENIORS - nebivolol (darr time to deathhf hospitalization but few LVEFgt50)
I-PRESERVE (ARB) did not improve outcomes
Recent Trials
TOPCAT (Spironolactone)
RELAX (sildenafil)
Paramount
Have HFpEF Trials Failed
Cleland JG Pellicori P Dierckx R
Heart Fail Clin 2014 Jul1
Reasons for Failure
Should the Trialists be on Trial
Was it Heart Failure
Different Demographics
Too many Co-Morbidities
Trial Design
HFpEF
Neither clinical history nor echo is a
reliable diagnostic method in patients
with HFpEF
Prior Hospitalization
Natriuretic Peptide Level
Natriuretic Peptide Enhances Risk
Cleland EHJ2006
Increase Mortality after Hospitalization
Both HF and non -HF
Carson JACC-HF
Risk of Events in HFpEF The I-Preserve Model
Previous hospitalization
Natriuretic Peptide
Age
Diabetes
Renal Disease
COPD
CAD
KomajdaCirc HF 2011
Duke Heart Center
A high degree of disease heterogeneity exists within heart failure patients and there is a need
for improved phenotyping of the syndrome
A new taxonomy of heart failure based on both clinical and molecular measures may provide a
more accurate classification of disease and ultimately enhance diagnosis and treatment
Duke Heart Center
Cluster Analysis
J Am Coll Cardiol 2014641765-74
Cluster analysis is an unsupervised learning task of grouping a set of
objects in such a way that objects in the same group are more similar
to each other than to those in other groups
Duke Heart Center
Cluster Analysis of Heart Failure to Uncover
Distinct Phenotypes
Circulation 2014 Nov 14
Duke Heart Center
Objective Measures of Heart Failure
Duke Heart Center
Clinical Implications on Outcomes
Duke Heart Center
bull Caucasian
bull Female
bull NICM
bull Few Co-Morbidities
Clinical Implications
Some Clusters Respond to Exercise
Implications of Co-Morbidities
Increase heterogeneity
Complicates management(Beta agonistsNSAID)
Associated with worse outcomes
Increase in non-cardiac outcomes
Mentz RJ and Felker GM Heart Fail Clin 2013
Duke Heart Center
Risk of Co-Morbidity and Death
in HFpEF
bull COPD
bull Rheumatologic
Disorders
bull More dangerous in
HFpEF than HFrEF
AtherJACC2012
Total Hospitalizations I-Preserve
Many Co-Morbid Hospitalizations
CarsonJACC-HF in press
Mode of Death I-Preserve
60 cardiovascular death
40 non-cardiovascular death or unknown
May require a doubling of sample size for mortality component
Similar Issue with HFH
ZileCirc HF 2010
Mode of Non- Cardiovascular Death I-Preserve
ZileCirc HF 2010
Regional Differences in Therapy
OConnorJACC2012
Spironolactone
Placebo
HR = 089 (077 ndash 104)
p=0138
3511723 (204)
3201722 (186)
1degOutcome (CV Death HF Hosp or Resuscitated Cardiac Arrest)
Placebo Rates Primary Outcome by Region
US Canada
Argentina Brazil
Russia Rep Georgia
126 per 100 pt-
yr
23 per 100 pt-yr
Placebo
280881 (318)
Placebo
71842 (84)
HR=082 (069-098)
HR=110 (079-151)
Interaction p=0122
US Canada
Argentina Brazil
Russia Rep Georgia
Placebo
280881 (318)
Placebo
71842 (84)
Exploratory (post-hoc)Placebo vs Spiro
by Region
Of 22 pre-specified only 1 - Stratum - showed
a significant interaction with treatment
Enrolled
by Spiro Placebo
Hazard Ratio
(95 CI)
P-value
Natriuretic
peptide
78490
(159)
116491
(236)
065 (049-087)
0003
Heart Failure
Hosp
2421232
(196)
2351232
(191)
101 (084-121)
0923
P=0013 for interaction
Subgroups
ACEIARB Meta-Analysis Favors
Treatment(HR=088 p=005)
ShahJCF 2010
Hope for HFpEF
Duke Heart Center
LCZ696 Favorable on the most likely
Surrogate
bull Reduced NT-proBNP
bull Reduced LA size
bull Improved NYHA
Class
bull PARAGON
OUTCOME Trial
SolomonLancet2012
Have we failed in HFpEF
As clinical trialists we have sometimes failed(unknowingly) in the
design and conduct of HFpEF studies
Most traditional therapies(ACEIARBMRABB) have not failed(modest
reduction in recurrent HFH) but they have not passed regulatory
standards ndashWhat about guidelines
Must address the moderate number of Co- Morbidities differently
New therapies will have the advantage of our informed journey and will
likely yield positive results
HFpEF Trials
Half Empty Half Full hellip Yet No Banners
Normal
Systolic
Heart
Failure
Diastolic
Heart Failure
Aurigemma Zile Gaasch
Circulation 2005
0
1
2
3
4
5
6
7
8
9
10
pre
val
ence
Prevalence of Heart Failure
age range
mean age
66-103
78
75
75
gt 50
-
gt 40
60
55-95
65
75-86
-
70-84
76
75
51
England
(Poole)
64
45
Den
(Copen)
49
29
Spain
(Asturias)
82
42
Finland
(Helsinki)
Portugal
(EPICA)
17
42
gt25
68
2
1
15
Nether
(Rotter)
67
Sweden
(Vasteras)
31
USA
(CHS)
88
48
Proportion with decreased
LV systolic function
Proportion with preserved
LV systolic function
Why Target
Co-Morbidities
Heterogeneity of population problematic
No therapeutic advances in broad approach
Comorbidities associated with high risk
May be differential association in HFpEF and
HFrEF
European Heart Failure Pilot Survey
3226 outpatients with chronic HF
74 had a least one co-morbidity
CKD (41) anemia (29) and diabetes (29)
Co-morbidities were independently associated
with higher age higher NYHA class ischemic
etiology of HF higher heart rate HTN and AF
Only diabetes CKD and anemia were
independently associated with a higher risk of
mortality andor HF hospitalization
There were marked regional differences in
prevalence and prognostic implications of co-
morbidities
van Deursen VM Eur J Heart Fail 2014 Jan16(1)103-11
Mentz RJ OrsquoConnor CM et al JACC 2014
HFpEF vs HFrEF
ADHERE Registry(2) GWTG Registry(8)
Reduced Preserved EFlt40 EFge50
Age (y) 70 plusmn 14 74 plusmn 13 70 (58-80) 78 (67-85)
Female 40 62 36 63
African American 22 17 25 16
Medical History
COPD or asthma 27 31 27 33
CRI 26 26 48 52
Anemia - - 14 22
Diabetes mellitus 40 45 22 oral
18 insulin
24 oral
22 insulin
Mentz RJ OrsquoConnor CM et al JACC 2014
Bidirectional Impact
Mentz RJ OrsquoConnor CM et al JACC 2014
HFpEF and Comorbidities
Paulus WJ et al JACC 2013
PATHOPHYSIOLOGIC TARGETS
Diastolic dysfunction
Ventriculo-arterial disociation
Pulmonary Hypertension
Chronotropic Incompetance
Systemic Hypertension
HFpEF ndash Clinical Trials to Date
HFpEF ndash Landmark Trials
No proven therapy
All treatment recommendations ndash Class C evidence
Completed
Dig (Mortalityharr HF Hospitalizationsdarr)
CHARM-Preserved (trend towards darr HF hosp CV mortality)
PEP-CHF (trend towards darr mortality hf hospitalization frac14 of pts withdrew to go on open label ACEI at 1 year)
SENIORS - nebivolol (darr time to deathhf hospitalization but few LVEFgt50)
I-PRESERVE (ARB) did not improve outcomes
Recent Trials
TOPCAT (Spironolactone)
RELAX (sildenafil)
Paramount
Have HFpEF Trials Failed
Cleland JG Pellicori P Dierckx R
Heart Fail Clin 2014 Jul1
Reasons for Failure
Should the Trialists be on Trial
Was it Heart Failure
Different Demographics
Too many Co-Morbidities
Trial Design
HFpEF
Neither clinical history nor echo is a
reliable diagnostic method in patients
with HFpEF
Prior Hospitalization
Natriuretic Peptide Level
Natriuretic Peptide Enhances Risk
Cleland EHJ2006
Increase Mortality after Hospitalization
Both HF and non -HF
Carson JACC-HF
Risk of Events in HFpEF The I-Preserve Model
Previous hospitalization
Natriuretic Peptide
Age
Diabetes
Renal Disease
COPD
CAD
KomajdaCirc HF 2011
Duke Heart Center
A high degree of disease heterogeneity exists within heart failure patients and there is a need
for improved phenotyping of the syndrome
A new taxonomy of heart failure based on both clinical and molecular measures may provide a
more accurate classification of disease and ultimately enhance diagnosis and treatment
Duke Heart Center
Cluster Analysis
J Am Coll Cardiol 2014641765-74
Cluster analysis is an unsupervised learning task of grouping a set of
objects in such a way that objects in the same group are more similar
to each other than to those in other groups
Duke Heart Center
Cluster Analysis of Heart Failure to Uncover
Distinct Phenotypes
Circulation 2014 Nov 14
Duke Heart Center
Objective Measures of Heart Failure
Duke Heart Center
Clinical Implications on Outcomes
Duke Heart Center
bull Caucasian
bull Female
bull NICM
bull Few Co-Morbidities
Clinical Implications
Some Clusters Respond to Exercise
Implications of Co-Morbidities
Increase heterogeneity
Complicates management(Beta agonistsNSAID)
Associated with worse outcomes
Increase in non-cardiac outcomes
Mentz RJ and Felker GM Heart Fail Clin 2013
Duke Heart Center
Risk of Co-Morbidity and Death
in HFpEF
bull COPD
bull Rheumatologic
Disorders
bull More dangerous in
HFpEF than HFrEF
AtherJACC2012
Total Hospitalizations I-Preserve
Many Co-Morbid Hospitalizations
CarsonJACC-HF in press
Mode of Death I-Preserve
60 cardiovascular death
40 non-cardiovascular death or unknown
May require a doubling of sample size for mortality component
Similar Issue with HFH
ZileCirc HF 2010
Mode of Non- Cardiovascular Death I-Preserve
ZileCirc HF 2010
Regional Differences in Therapy
OConnorJACC2012
Spironolactone
Placebo
HR = 089 (077 ndash 104)
p=0138
3511723 (204)
3201722 (186)
1degOutcome (CV Death HF Hosp or Resuscitated Cardiac Arrest)
Placebo Rates Primary Outcome by Region
US Canada
Argentina Brazil
Russia Rep Georgia
126 per 100 pt-
yr
23 per 100 pt-yr
Placebo
280881 (318)
Placebo
71842 (84)
HR=082 (069-098)
HR=110 (079-151)
Interaction p=0122
US Canada
Argentina Brazil
Russia Rep Georgia
Placebo
280881 (318)
Placebo
71842 (84)
Exploratory (post-hoc)Placebo vs Spiro
by Region
Of 22 pre-specified only 1 - Stratum - showed
a significant interaction with treatment
Enrolled
by Spiro Placebo
Hazard Ratio
(95 CI)
P-value
Natriuretic
peptide
78490
(159)
116491
(236)
065 (049-087)
0003
Heart Failure
Hosp
2421232
(196)
2351232
(191)
101 (084-121)
0923
P=0013 for interaction
Subgroups
ACEIARB Meta-Analysis Favors
Treatment(HR=088 p=005)
ShahJCF 2010
Hope for HFpEF
Duke Heart Center
LCZ696 Favorable on the most likely
Surrogate
bull Reduced NT-proBNP
bull Reduced LA size
bull Improved NYHA
Class
bull PARAGON
OUTCOME Trial
SolomonLancet2012
Have we failed in HFpEF
As clinical trialists we have sometimes failed(unknowingly) in the
design and conduct of HFpEF studies
Most traditional therapies(ACEIARBMRABB) have not failed(modest
reduction in recurrent HFH) but they have not passed regulatory
standards ndashWhat about guidelines
Must address the moderate number of Co- Morbidities differently
New therapies will have the advantage of our informed journey and will
likely yield positive results
HFpEF Trials
Half Empty Half Full hellip Yet No Banners
0
1
2
3
4
5
6
7
8
9
10
pre
val
ence
Prevalence of Heart Failure
age range
mean age
66-103
78
75
75
gt 50
-
gt 40
60
55-95
65
75-86
-
70-84
76
75
51
England
(Poole)
64
45
Den
(Copen)
49
29
Spain
(Asturias)
82
42
Finland
(Helsinki)
Portugal
(EPICA)
17
42
gt25
68
2
1
15
Nether
(Rotter)
67
Sweden
(Vasteras)
31
USA
(CHS)
88
48
Proportion with decreased
LV systolic function
Proportion with preserved
LV systolic function
Why Target
Co-Morbidities
Heterogeneity of population problematic
No therapeutic advances in broad approach
Comorbidities associated with high risk
May be differential association in HFpEF and
HFrEF
European Heart Failure Pilot Survey
3226 outpatients with chronic HF
74 had a least one co-morbidity
CKD (41) anemia (29) and diabetes (29)
Co-morbidities were independently associated
with higher age higher NYHA class ischemic
etiology of HF higher heart rate HTN and AF
Only diabetes CKD and anemia were
independently associated with a higher risk of
mortality andor HF hospitalization
There were marked regional differences in
prevalence and prognostic implications of co-
morbidities
van Deursen VM Eur J Heart Fail 2014 Jan16(1)103-11
Mentz RJ OrsquoConnor CM et al JACC 2014
HFpEF vs HFrEF
ADHERE Registry(2) GWTG Registry(8)
Reduced Preserved EFlt40 EFge50
Age (y) 70 plusmn 14 74 plusmn 13 70 (58-80) 78 (67-85)
Female 40 62 36 63
African American 22 17 25 16
Medical History
COPD or asthma 27 31 27 33
CRI 26 26 48 52
Anemia - - 14 22
Diabetes mellitus 40 45 22 oral
18 insulin
24 oral
22 insulin
Mentz RJ OrsquoConnor CM et al JACC 2014
Bidirectional Impact
Mentz RJ OrsquoConnor CM et al JACC 2014
HFpEF and Comorbidities
Paulus WJ et al JACC 2013
PATHOPHYSIOLOGIC TARGETS
Diastolic dysfunction
Ventriculo-arterial disociation
Pulmonary Hypertension
Chronotropic Incompetance
Systemic Hypertension
HFpEF ndash Clinical Trials to Date
HFpEF ndash Landmark Trials
No proven therapy
All treatment recommendations ndash Class C evidence
Completed
Dig (Mortalityharr HF Hospitalizationsdarr)
CHARM-Preserved (trend towards darr HF hosp CV mortality)
PEP-CHF (trend towards darr mortality hf hospitalization frac14 of pts withdrew to go on open label ACEI at 1 year)
SENIORS - nebivolol (darr time to deathhf hospitalization but few LVEFgt50)
I-PRESERVE (ARB) did not improve outcomes
Recent Trials
TOPCAT (Spironolactone)
RELAX (sildenafil)
Paramount
Have HFpEF Trials Failed
Cleland JG Pellicori P Dierckx R
Heart Fail Clin 2014 Jul1
Reasons for Failure
Should the Trialists be on Trial
Was it Heart Failure
Different Demographics
Too many Co-Morbidities
Trial Design
HFpEF
Neither clinical history nor echo is a
reliable diagnostic method in patients
with HFpEF
Prior Hospitalization
Natriuretic Peptide Level
Natriuretic Peptide Enhances Risk
Cleland EHJ2006
Increase Mortality after Hospitalization
Both HF and non -HF
Carson JACC-HF
Risk of Events in HFpEF The I-Preserve Model
Previous hospitalization
Natriuretic Peptide
Age
Diabetes
Renal Disease
COPD
CAD
KomajdaCirc HF 2011
Duke Heart Center
A high degree of disease heterogeneity exists within heart failure patients and there is a need
for improved phenotyping of the syndrome
A new taxonomy of heart failure based on both clinical and molecular measures may provide a
more accurate classification of disease and ultimately enhance diagnosis and treatment
Duke Heart Center
Cluster Analysis
J Am Coll Cardiol 2014641765-74
Cluster analysis is an unsupervised learning task of grouping a set of
objects in such a way that objects in the same group are more similar
to each other than to those in other groups
Duke Heart Center
Cluster Analysis of Heart Failure to Uncover
Distinct Phenotypes
Circulation 2014 Nov 14
Duke Heart Center
Objective Measures of Heart Failure
Duke Heart Center
Clinical Implications on Outcomes
Duke Heart Center
bull Caucasian
bull Female
bull NICM
bull Few Co-Morbidities
Clinical Implications
Some Clusters Respond to Exercise
Implications of Co-Morbidities
Increase heterogeneity
Complicates management(Beta agonistsNSAID)
Associated with worse outcomes
Increase in non-cardiac outcomes
Mentz RJ and Felker GM Heart Fail Clin 2013
Duke Heart Center
Risk of Co-Morbidity and Death
in HFpEF
bull COPD
bull Rheumatologic
Disorders
bull More dangerous in
HFpEF than HFrEF
AtherJACC2012
Total Hospitalizations I-Preserve
Many Co-Morbid Hospitalizations
CarsonJACC-HF in press
Mode of Death I-Preserve
60 cardiovascular death
40 non-cardiovascular death or unknown
May require a doubling of sample size for mortality component
Similar Issue with HFH
ZileCirc HF 2010
Mode of Non- Cardiovascular Death I-Preserve
ZileCirc HF 2010
Regional Differences in Therapy
OConnorJACC2012
Spironolactone
Placebo
HR = 089 (077 ndash 104)
p=0138
3511723 (204)
3201722 (186)
1degOutcome (CV Death HF Hosp or Resuscitated Cardiac Arrest)
Placebo Rates Primary Outcome by Region
US Canada
Argentina Brazil
Russia Rep Georgia
126 per 100 pt-
yr
23 per 100 pt-yr
Placebo
280881 (318)
Placebo
71842 (84)
HR=082 (069-098)
HR=110 (079-151)
Interaction p=0122
US Canada
Argentina Brazil
Russia Rep Georgia
Placebo
280881 (318)
Placebo
71842 (84)
Exploratory (post-hoc)Placebo vs Spiro
by Region
Of 22 pre-specified only 1 - Stratum - showed
a significant interaction with treatment
Enrolled
by Spiro Placebo
Hazard Ratio
(95 CI)
P-value
Natriuretic
peptide
78490
(159)
116491
(236)
065 (049-087)
0003
Heart Failure
Hosp
2421232
(196)
2351232
(191)
101 (084-121)
0923
P=0013 for interaction
Subgroups
ACEIARB Meta-Analysis Favors
Treatment(HR=088 p=005)
ShahJCF 2010
Hope for HFpEF
Duke Heart Center
LCZ696 Favorable on the most likely
Surrogate
bull Reduced NT-proBNP
bull Reduced LA size
bull Improved NYHA
Class
bull PARAGON
OUTCOME Trial
SolomonLancet2012
Have we failed in HFpEF
As clinical trialists we have sometimes failed(unknowingly) in the
design and conduct of HFpEF studies
Most traditional therapies(ACEIARBMRABB) have not failed(modest
reduction in recurrent HFH) but they have not passed regulatory
standards ndashWhat about guidelines
Must address the moderate number of Co- Morbidities differently
New therapies will have the advantage of our informed journey and will
likely yield positive results
HFpEF Trials
Half Empty Half Full hellip Yet No Banners
Why Target
Co-Morbidities
Heterogeneity of population problematic
No therapeutic advances in broad approach
Comorbidities associated with high risk
May be differential association in HFpEF and
HFrEF
European Heart Failure Pilot Survey
3226 outpatients with chronic HF
74 had a least one co-morbidity
CKD (41) anemia (29) and diabetes (29)
Co-morbidities were independently associated
with higher age higher NYHA class ischemic
etiology of HF higher heart rate HTN and AF
Only diabetes CKD and anemia were
independently associated with a higher risk of
mortality andor HF hospitalization
There were marked regional differences in
prevalence and prognostic implications of co-
morbidities
van Deursen VM Eur J Heart Fail 2014 Jan16(1)103-11
Mentz RJ OrsquoConnor CM et al JACC 2014
HFpEF vs HFrEF
ADHERE Registry(2) GWTG Registry(8)
Reduced Preserved EFlt40 EFge50
Age (y) 70 plusmn 14 74 plusmn 13 70 (58-80) 78 (67-85)
Female 40 62 36 63
African American 22 17 25 16
Medical History
COPD or asthma 27 31 27 33
CRI 26 26 48 52
Anemia - - 14 22
Diabetes mellitus 40 45 22 oral
18 insulin
24 oral
22 insulin
Mentz RJ OrsquoConnor CM et al JACC 2014
Bidirectional Impact
Mentz RJ OrsquoConnor CM et al JACC 2014
HFpEF and Comorbidities
Paulus WJ et al JACC 2013
PATHOPHYSIOLOGIC TARGETS
Diastolic dysfunction
Ventriculo-arterial disociation
Pulmonary Hypertension
Chronotropic Incompetance
Systemic Hypertension
HFpEF ndash Clinical Trials to Date
HFpEF ndash Landmark Trials
No proven therapy
All treatment recommendations ndash Class C evidence
Completed
Dig (Mortalityharr HF Hospitalizationsdarr)
CHARM-Preserved (trend towards darr HF hosp CV mortality)
PEP-CHF (trend towards darr mortality hf hospitalization frac14 of pts withdrew to go on open label ACEI at 1 year)
SENIORS - nebivolol (darr time to deathhf hospitalization but few LVEFgt50)
I-PRESERVE (ARB) did not improve outcomes
Recent Trials
TOPCAT (Spironolactone)
RELAX (sildenafil)
Paramount
Have HFpEF Trials Failed
Cleland JG Pellicori P Dierckx R
Heart Fail Clin 2014 Jul1
Reasons for Failure
Should the Trialists be on Trial
Was it Heart Failure
Different Demographics
Too many Co-Morbidities
Trial Design
HFpEF
Neither clinical history nor echo is a
reliable diagnostic method in patients
with HFpEF
Prior Hospitalization
Natriuretic Peptide Level
Natriuretic Peptide Enhances Risk
Cleland EHJ2006
Increase Mortality after Hospitalization
Both HF and non -HF
Carson JACC-HF
Risk of Events in HFpEF The I-Preserve Model
Previous hospitalization
Natriuretic Peptide
Age
Diabetes
Renal Disease
COPD
CAD
KomajdaCirc HF 2011
Duke Heart Center
A high degree of disease heterogeneity exists within heart failure patients and there is a need
for improved phenotyping of the syndrome
A new taxonomy of heart failure based on both clinical and molecular measures may provide a
more accurate classification of disease and ultimately enhance diagnosis and treatment
Duke Heart Center
Cluster Analysis
J Am Coll Cardiol 2014641765-74
Cluster analysis is an unsupervised learning task of grouping a set of
objects in such a way that objects in the same group are more similar
to each other than to those in other groups
Duke Heart Center
Cluster Analysis of Heart Failure to Uncover
Distinct Phenotypes
Circulation 2014 Nov 14
Duke Heart Center
Objective Measures of Heart Failure
Duke Heart Center
Clinical Implications on Outcomes
Duke Heart Center
bull Caucasian
bull Female
bull NICM
bull Few Co-Morbidities
Clinical Implications
Some Clusters Respond to Exercise
Implications of Co-Morbidities
Increase heterogeneity
Complicates management(Beta agonistsNSAID)
Associated with worse outcomes
Increase in non-cardiac outcomes
Mentz RJ and Felker GM Heart Fail Clin 2013
Duke Heart Center
Risk of Co-Morbidity and Death
in HFpEF
bull COPD
bull Rheumatologic
Disorders
bull More dangerous in
HFpEF than HFrEF
AtherJACC2012
Total Hospitalizations I-Preserve
Many Co-Morbid Hospitalizations
CarsonJACC-HF in press
Mode of Death I-Preserve
60 cardiovascular death
40 non-cardiovascular death or unknown
May require a doubling of sample size for mortality component
Similar Issue with HFH
ZileCirc HF 2010
Mode of Non- Cardiovascular Death I-Preserve
ZileCirc HF 2010
Regional Differences in Therapy
OConnorJACC2012
Spironolactone
Placebo
HR = 089 (077 ndash 104)
p=0138
3511723 (204)
3201722 (186)
1degOutcome (CV Death HF Hosp or Resuscitated Cardiac Arrest)
Placebo Rates Primary Outcome by Region
US Canada
Argentina Brazil
Russia Rep Georgia
126 per 100 pt-
yr
23 per 100 pt-yr
Placebo
280881 (318)
Placebo
71842 (84)
HR=082 (069-098)
HR=110 (079-151)
Interaction p=0122
US Canada
Argentina Brazil
Russia Rep Georgia
Placebo
280881 (318)
Placebo
71842 (84)
Exploratory (post-hoc)Placebo vs Spiro
by Region
Of 22 pre-specified only 1 - Stratum - showed
a significant interaction with treatment
Enrolled
by Spiro Placebo
Hazard Ratio
(95 CI)
P-value
Natriuretic
peptide
78490
(159)
116491
(236)
065 (049-087)
0003
Heart Failure
Hosp
2421232
(196)
2351232
(191)
101 (084-121)
0923
P=0013 for interaction
Subgroups
ACEIARB Meta-Analysis Favors
Treatment(HR=088 p=005)
ShahJCF 2010
Hope for HFpEF
Duke Heart Center
LCZ696 Favorable on the most likely
Surrogate
bull Reduced NT-proBNP
bull Reduced LA size
bull Improved NYHA
Class
bull PARAGON
OUTCOME Trial
SolomonLancet2012
Have we failed in HFpEF
As clinical trialists we have sometimes failed(unknowingly) in the
design and conduct of HFpEF studies
Most traditional therapies(ACEIARBMRABB) have not failed(modest
reduction in recurrent HFH) but they have not passed regulatory
standards ndashWhat about guidelines
Must address the moderate number of Co- Morbidities differently
New therapies will have the advantage of our informed journey and will
likely yield positive results
HFpEF Trials
Half Empty Half Full hellip Yet No Banners
European Heart Failure Pilot Survey
3226 outpatients with chronic HF
74 had a least one co-morbidity
CKD (41) anemia (29) and diabetes (29)
Co-morbidities were independently associated
with higher age higher NYHA class ischemic
etiology of HF higher heart rate HTN and AF
Only diabetes CKD and anemia were
independently associated with a higher risk of
mortality andor HF hospitalization
There were marked regional differences in
prevalence and prognostic implications of co-
morbidities
van Deursen VM Eur J Heart Fail 2014 Jan16(1)103-11
Mentz RJ OrsquoConnor CM et al JACC 2014
HFpEF vs HFrEF
ADHERE Registry(2) GWTG Registry(8)
Reduced Preserved EFlt40 EFge50
Age (y) 70 plusmn 14 74 plusmn 13 70 (58-80) 78 (67-85)
Female 40 62 36 63
African American 22 17 25 16
Medical History
COPD or asthma 27 31 27 33
CRI 26 26 48 52
Anemia - - 14 22
Diabetes mellitus 40 45 22 oral
18 insulin
24 oral
22 insulin
Mentz RJ OrsquoConnor CM et al JACC 2014
Bidirectional Impact
Mentz RJ OrsquoConnor CM et al JACC 2014
HFpEF and Comorbidities
Paulus WJ et al JACC 2013
PATHOPHYSIOLOGIC TARGETS
Diastolic dysfunction
Ventriculo-arterial disociation
Pulmonary Hypertension
Chronotropic Incompetance
Systemic Hypertension
HFpEF ndash Clinical Trials to Date
HFpEF ndash Landmark Trials
No proven therapy
All treatment recommendations ndash Class C evidence
Completed
Dig (Mortalityharr HF Hospitalizationsdarr)
CHARM-Preserved (trend towards darr HF hosp CV mortality)
PEP-CHF (trend towards darr mortality hf hospitalization frac14 of pts withdrew to go on open label ACEI at 1 year)
SENIORS - nebivolol (darr time to deathhf hospitalization but few LVEFgt50)
I-PRESERVE (ARB) did not improve outcomes
Recent Trials
TOPCAT (Spironolactone)
RELAX (sildenafil)
Paramount
Have HFpEF Trials Failed
Cleland JG Pellicori P Dierckx R
Heart Fail Clin 2014 Jul1
Reasons for Failure
Should the Trialists be on Trial
Was it Heart Failure
Different Demographics
Too many Co-Morbidities
Trial Design
HFpEF
Neither clinical history nor echo is a
reliable diagnostic method in patients
with HFpEF
Prior Hospitalization
Natriuretic Peptide Level
Natriuretic Peptide Enhances Risk
Cleland EHJ2006
Increase Mortality after Hospitalization
Both HF and non -HF
Carson JACC-HF
Risk of Events in HFpEF The I-Preserve Model
Previous hospitalization
Natriuretic Peptide
Age
Diabetes
Renal Disease
COPD
CAD
KomajdaCirc HF 2011
Duke Heart Center
A high degree of disease heterogeneity exists within heart failure patients and there is a need
for improved phenotyping of the syndrome
A new taxonomy of heart failure based on both clinical and molecular measures may provide a
more accurate classification of disease and ultimately enhance diagnosis and treatment
Duke Heart Center
Cluster Analysis
J Am Coll Cardiol 2014641765-74
Cluster analysis is an unsupervised learning task of grouping a set of
objects in such a way that objects in the same group are more similar
to each other than to those in other groups
Duke Heart Center
Cluster Analysis of Heart Failure to Uncover
Distinct Phenotypes
Circulation 2014 Nov 14
Duke Heart Center
Objective Measures of Heart Failure
Duke Heart Center
Clinical Implications on Outcomes
Duke Heart Center
bull Caucasian
bull Female
bull NICM
bull Few Co-Morbidities
Clinical Implications
Some Clusters Respond to Exercise
Implications of Co-Morbidities
Increase heterogeneity
Complicates management(Beta agonistsNSAID)
Associated with worse outcomes
Increase in non-cardiac outcomes
Mentz RJ and Felker GM Heart Fail Clin 2013
Duke Heart Center
Risk of Co-Morbidity and Death
in HFpEF
bull COPD
bull Rheumatologic
Disorders
bull More dangerous in
HFpEF than HFrEF
AtherJACC2012
Total Hospitalizations I-Preserve
Many Co-Morbid Hospitalizations
CarsonJACC-HF in press
Mode of Death I-Preserve
60 cardiovascular death
40 non-cardiovascular death or unknown
May require a doubling of sample size for mortality component
Similar Issue with HFH
ZileCirc HF 2010
Mode of Non- Cardiovascular Death I-Preserve
ZileCirc HF 2010
Regional Differences in Therapy
OConnorJACC2012
Spironolactone
Placebo
HR = 089 (077 ndash 104)
p=0138
3511723 (204)
3201722 (186)
1degOutcome (CV Death HF Hosp or Resuscitated Cardiac Arrest)
Placebo Rates Primary Outcome by Region
US Canada
Argentina Brazil
Russia Rep Georgia
126 per 100 pt-
yr
23 per 100 pt-yr
Placebo
280881 (318)
Placebo
71842 (84)
HR=082 (069-098)
HR=110 (079-151)
Interaction p=0122
US Canada
Argentina Brazil
Russia Rep Georgia
Placebo
280881 (318)
Placebo
71842 (84)
Exploratory (post-hoc)Placebo vs Spiro
by Region
Of 22 pre-specified only 1 - Stratum - showed
a significant interaction with treatment
Enrolled
by Spiro Placebo
Hazard Ratio
(95 CI)
P-value
Natriuretic
peptide
78490
(159)
116491
(236)
065 (049-087)
0003
Heart Failure
Hosp
2421232
(196)
2351232
(191)
101 (084-121)
0923
P=0013 for interaction
Subgroups
ACEIARB Meta-Analysis Favors
Treatment(HR=088 p=005)
ShahJCF 2010
Hope for HFpEF
Duke Heart Center
LCZ696 Favorable on the most likely
Surrogate
bull Reduced NT-proBNP
bull Reduced LA size
bull Improved NYHA
Class
bull PARAGON
OUTCOME Trial
SolomonLancet2012
Have we failed in HFpEF
As clinical trialists we have sometimes failed(unknowingly) in the
design and conduct of HFpEF studies
Most traditional therapies(ACEIARBMRABB) have not failed(modest
reduction in recurrent HFH) but they have not passed regulatory
standards ndashWhat about guidelines
Must address the moderate number of Co- Morbidities differently
New therapies will have the advantage of our informed journey and will
likely yield positive results
HFpEF Trials
Half Empty Half Full hellip Yet No Banners
Mentz RJ OrsquoConnor CM et al JACC 2014
HFpEF vs HFrEF
ADHERE Registry(2) GWTG Registry(8)
Reduced Preserved EFlt40 EFge50
Age (y) 70 plusmn 14 74 plusmn 13 70 (58-80) 78 (67-85)
Female 40 62 36 63
African American 22 17 25 16
Medical History
COPD or asthma 27 31 27 33
CRI 26 26 48 52
Anemia - - 14 22
Diabetes mellitus 40 45 22 oral
18 insulin
24 oral
22 insulin
Mentz RJ OrsquoConnor CM et al JACC 2014
Bidirectional Impact
Mentz RJ OrsquoConnor CM et al JACC 2014
HFpEF and Comorbidities
Paulus WJ et al JACC 2013
PATHOPHYSIOLOGIC TARGETS
Diastolic dysfunction
Ventriculo-arterial disociation
Pulmonary Hypertension
Chronotropic Incompetance
Systemic Hypertension
HFpEF ndash Clinical Trials to Date
HFpEF ndash Landmark Trials
No proven therapy
All treatment recommendations ndash Class C evidence
Completed
Dig (Mortalityharr HF Hospitalizationsdarr)
CHARM-Preserved (trend towards darr HF hosp CV mortality)
PEP-CHF (trend towards darr mortality hf hospitalization frac14 of pts withdrew to go on open label ACEI at 1 year)
SENIORS - nebivolol (darr time to deathhf hospitalization but few LVEFgt50)
I-PRESERVE (ARB) did not improve outcomes
Recent Trials
TOPCAT (Spironolactone)
RELAX (sildenafil)
Paramount
Have HFpEF Trials Failed
Cleland JG Pellicori P Dierckx R
Heart Fail Clin 2014 Jul1
Reasons for Failure
Should the Trialists be on Trial
Was it Heart Failure
Different Demographics
Too many Co-Morbidities
Trial Design
HFpEF
Neither clinical history nor echo is a
reliable diagnostic method in patients
with HFpEF
Prior Hospitalization
Natriuretic Peptide Level
Natriuretic Peptide Enhances Risk
Cleland EHJ2006
Increase Mortality after Hospitalization
Both HF and non -HF
Carson JACC-HF
Risk of Events in HFpEF The I-Preserve Model
Previous hospitalization
Natriuretic Peptide
Age
Diabetes
Renal Disease
COPD
CAD
KomajdaCirc HF 2011
Duke Heart Center
A high degree of disease heterogeneity exists within heart failure patients and there is a need
for improved phenotyping of the syndrome
A new taxonomy of heart failure based on both clinical and molecular measures may provide a
more accurate classification of disease and ultimately enhance diagnosis and treatment
Duke Heart Center
Cluster Analysis
J Am Coll Cardiol 2014641765-74
Cluster analysis is an unsupervised learning task of grouping a set of
objects in such a way that objects in the same group are more similar
to each other than to those in other groups
Duke Heart Center
Cluster Analysis of Heart Failure to Uncover
Distinct Phenotypes
Circulation 2014 Nov 14
Duke Heart Center
Objective Measures of Heart Failure
Duke Heart Center
Clinical Implications on Outcomes
Duke Heart Center
bull Caucasian
bull Female
bull NICM
bull Few Co-Morbidities
Clinical Implications
Some Clusters Respond to Exercise
Implications of Co-Morbidities
Increase heterogeneity
Complicates management(Beta agonistsNSAID)
Associated with worse outcomes
Increase in non-cardiac outcomes
Mentz RJ and Felker GM Heart Fail Clin 2013
Duke Heart Center
Risk of Co-Morbidity and Death
in HFpEF
bull COPD
bull Rheumatologic
Disorders
bull More dangerous in
HFpEF than HFrEF
AtherJACC2012
Total Hospitalizations I-Preserve
Many Co-Morbid Hospitalizations
CarsonJACC-HF in press
Mode of Death I-Preserve
60 cardiovascular death
40 non-cardiovascular death or unknown
May require a doubling of sample size for mortality component
Similar Issue with HFH
ZileCirc HF 2010
Mode of Non- Cardiovascular Death I-Preserve
ZileCirc HF 2010
Regional Differences in Therapy
OConnorJACC2012
Spironolactone
Placebo
HR = 089 (077 ndash 104)
p=0138
3511723 (204)
3201722 (186)
1degOutcome (CV Death HF Hosp or Resuscitated Cardiac Arrest)
Placebo Rates Primary Outcome by Region
US Canada
Argentina Brazil
Russia Rep Georgia
126 per 100 pt-
yr
23 per 100 pt-yr
Placebo
280881 (318)
Placebo
71842 (84)
HR=082 (069-098)
HR=110 (079-151)
Interaction p=0122
US Canada
Argentina Brazil
Russia Rep Georgia
Placebo
280881 (318)
Placebo
71842 (84)
Exploratory (post-hoc)Placebo vs Spiro
by Region
Of 22 pre-specified only 1 - Stratum - showed
a significant interaction with treatment
Enrolled
by Spiro Placebo
Hazard Ratio
(95 CI)
P-value
Natriuretic
peptide
78490
(159)
116491
(236)
065 (049-087)
0003
Heart Failure
Hosp
2421232
(196)
2351232
(191)
101 (084-121)
0923
P=0013 for interaction
Subgroups
ACEIARB Meta-Analysis Favors
Treatment(HR=088 p=005)
ShahJCF 2010
Hope for HFpEF
Duke Heart Center
LCZ696 Favorable on the most likely
Surrogate
bull Reduced NT-proBNP
bull Reduced LA size
bull Improved NYHA
Class
bull PARAGON
OUTCOME Trial
SolomonLancet2012
Have we failed in HFpEF
As clinical trialists we have sometimes failed(unknowingly) in the
design and conduct of HFpEF studies
Most traditional therapies(ACEIARBMRABB) have not failed(modest
reduction in recurrent HFH) but they have not passed regulatory
standards ndashWhat about guidelines
Must address the moderate number of Co- Morbidities differently
New therapies will have the advantage of our informed journey and will
likely yield positive results
HFpEF Trials
Half Empty Half Full hellip Yet No Banners
HFpEF vs HFrEF
ADHERE Registry(2) GWTG Registry(8)
Reduced Preserved EFlt40 EFge50
Age (y) 70 plusmn 14 74 plusmn 13 70 (58-80) 78 (67-85)
Female 40 62 36 63
African American 22 17 25 16
Medical History
COPD or asthma 27 31 27 33
CRI 26 26 48 52
Anemia - - 14 22
Diabetes mellitus 40 45 22 oral
18 insulin
24 oral
22 insulin
Mentz RJ OrsquoConnor CM et al JACC 2014
Bidirectional Impact
Mentz RJ OrsquoConnor CM et al JACC 2014
HFpEF and Comorbidities
Paulus WJ et al JACC 2013
PATHOPHYSIOLOGIC TARGETS
Diastolic dysfunction
Ventriculo-arterial disociation
Pulmonary Hypertension
Chronotropic Incompetance
Systemic Hypertension
HFpEF ndash Clinical Trials to Date
HFpEF ndash Landmark Trials
No proven therapy
All treatment recommendations ndash Class C evidence
Completed
Dig (Mortalityharr HF Hospitalizationsdarr)
CHARM-Preserved (trend towards darr HF hosp CV mortality)
PEP-CHF (trend towards darr mortality hf hospitalization frac14 of pts withdrew to go on open label ACEI at 1 year)
SENIORS - nebivolol (darr time to deathhf hospitalization but few LVEFgt50)
I-PRESERVE (ARB) did not improve outcomes
Recent Trials
TOPCAT (Spironolactone)
RELAX (sildenafil)
Paramount
Have HFpEF Trials Failed
Cleland JG Pellicori P Dierckx R
Heart Fail Clin 2014 Jul1
Reasons for Failure
Should the Trialists be on Trial
Was it Heart Failure
Different Demographics
Too many Co-Morbidities
Trial Design
HFpEF
Neither clinical history nor echo is a
reliable diagnostic method in patients
with HFpEF
Prior Hospitalization
Natriuretic Peptide Level
Natriuretic Peptide Enhances Risk
Cleland EHJ2006
Increase Mortality after Hospitalization
Both HF and non -HF
Carson JACC-HF
Risk of Events in HFpEF The I-Preserve Model
Previous hospitalization
Natriuretic Peptide
Age
Diabetes
Renal Disease
COPD
CAD
KomajdaCirc HF 2011
Duke Heart Center
A high degree of disease heterogeneity exists within heart failure patients and there is a need
for improved phenotyping of the syndrome
A new taxonomy of heart failure based on both clinical and molecular measures may provide a
more accurate classification of disease and ultimately enhance diagnosis and treatment
Duke Heart Center
Cluster Analysis
J Am Coll Cardiol 2014641765-74
Cluster analysis is an unsupervised learning task of grouping a set of
objects in such a way that objects in the same group are more similar
to each other than to those in other groups
Duke Heart Center
Cluster Analysis of Heart Failure to Uncover
Distinct Phenotypes
Circulation 2014 Nov 14
Duke Heart Center
Objective Measures of Heart Failure
Duke Heart Center
Clinical Implications on Outcomes
Duke Heart Center
bull Caucasian
bull Female
bull NICM
bull Few Co-Morbidities
Clinical Implications
Some Clusters Respond to Exercise
Implications of Co-Morbidities
Increase heterogeneity
Complicates management(Beta agonistsNSAID)
Associated with worse outcomes
Increase in non-cardiac outcomes
Mentz RJ and Felker GM Heart Fail Clin 2013
Duke Heart Center
Risk of Co-Morbidity and Death
in HFpEF
bull COPD
bull Rheumatologic
Disorders
bull More dangerous in
HFpEF than HFrEF
AtherJACC2012
Total Hospitalizations I-Preserve
Many Co-Morbid Hospitalizations
CarsonJACC-HF in press
Mode of Death I-Preserve
60 cardiovascular death
40 non-cardiovascular death or unknown
May require a doubling of sample size for mortality component
Similar Issue with HFH
ZileCirc HF 2010
Mode of Non- Cardiovascular Death I-Preserve
ZileCirc HF 2010
Regional Differences in Therapy
OConnorJACC2012
Spironolactone
Placebo
HR = 089 (077 ndash 104)
p=0138
3511723 (204)
3201722 (186)
1degOutcome (CV Death HF Hosp or Resuscitated Cardiac Arrest)
Placebo Rates Primary Outcome by Region
US Canada
Argentina Brazil
Russia Rep Georgia
126 per 100 pt-
yr
23 per 100 pt-yr
Placebo
280881 (318)
Placebo
71842 (84)
HR=082 (069-098)
HR=110 (079-151)
Interaction p=0122
US Canada
Argentina Brazil
Russia Rep Georgia
Placebo
280881 (318)
Placebo
71842 (84)
Exploratory (post-hoc)Placebo vs Spiro
by Region
Of 22 pre-specified only 1 - Stratum - showed
a significant interaction with treatment
Enrolled
by Spiro Placebo
Hazard Ratio
(95 CI)
P-value
Natriuretic
peptide
78490
(159)
116491
(236)
065 (049-087)
0003
Heart Failure
Hosp
2421232
(196)
2351232
(191)
101 (084-121)
0923
P=0013 for interaction
Subgroups
ACEIARB Meta-Analysis Favors
Treatment(HR=088 p=005)
ShahJCF 2010
Hope for HFpEF
Duke Heart Center
LCZ696 Favorable on the most likely
Surrogate
bull Reduced NT-proBNP
bull Reduced LA size
bull Improved NYHA
Class
bull PARAGON
OUTCOME Trial
SolomonLancet2012
Have we failed in HFpEF
As clinical trialists we have sometimes failed(unknowingly) in the
design and conduct of HFpEF studies
Most traditional therapies(ACEIARBMRABB) have not failed(modest
reduction in recurrent HFH) but they have not passed regulatory
standards ndashWhat about guidelines
Must address the moderate number of Co- Morbidities differently
New therapies will have the advantage of our informed journey and will
likely yield positive results
HFpEF Trials
Half Empty Half Full hellip Yet No Banners
Bidirectional Impact
Mentz RJ OrsquoConnor CM et al JACC 2014
HFpEF and Comorbidities
Paulus WJ et al JACC 2013
PATHOPHYSIOLOGIC TARGETS
Diastolic dysfunction
Ventriculo-arterial disociation
Pulmonary Hypertension
Chronotropic Incompetance
Systemic Hypertension
HFpEF ndash Clinical Trials to Date
HFpEF ndash Landmark Trials
No proven therapy
All treatment recommendations ndash Class C evidence
Completed
Dig (Mortalityharr HF Hospitalizationsdarr)
CHARM-Preserved (trend towards darr HF hosp CV mortality)
PEP-CHF (trend towards darr mortality hf hospitalization frac14 of pts withdrew to go on open label ACEI at 1 year)
SENIORS - nebivolol (darr time to deathhf hospitalization but few LVEFgt50)
I-PRESERVE (ARB) did not improve outcomes
Recent Trials
TOPCAT (Spironolactone)
RELAX (sildenafil)
Paramount
Have HFpEF Trials Failed
Cleland JG Pellicori P Dierckx R
Heart Fail Clin 2014 Jul1
Reasons for Failure
Should the Trialists be on Trial
Was it Heart Failure
Different Demographics
Too many Co-Morbidities
Trial Design
HFpEF
Neither clinical history nor echo is a
reliable diagnostic method in patients
with HFpEF
Prior Hospitalization
Natriuretic Peptide Level
Natriuretic Peptide Enhances Risk
Cleland EHJ2006
Increase Mortality after Hospitalization
Both HF and non -HF
Carson JACC-HF
Risk of Events in HFpEF The I-Preserve Model
Previous hospitalization
Natriuretic Peptide
Age
Diabetes
Renal Disease
COPD
CAD
KomajdaCirc HF 2011
Duke Heart Center
A high degree of disease heterogeneity exists within heart failure patients and there is a need
for improved phenotyping of the syndrome
A new taxonomy of heart failure based on both clinical and molecular measures may provide a
more accurate classification of disease and ultimately enhance diagnosis and treatment
Duke Heart Center
Cluster Analysis
J Am Coll Cardiol 2014641765-74
Cluster analysis is an unsupervised learning task of grouping a set of
objects in such a way that objects in the same group are more similar
to each other than to those in other groups
Duke Heart Center
Cluster Analysis of Heart Failure to Uncover
Distinct Phenotypes
Circulation 2014 Nov 14
Duke Heart Center
Objective Measures of Heart Failure
Duke Heart Center
Clinical Implications on Outcomes
Duke Heart Center
bull Caucasian
bull Female
bull NICM
bull Few Co-Morbidities
Clinical Implications
Some Clusters Respond to Exercise
Implications of Co-Morbidities
Increase heterogeneity
Complicates management(Beta agonistsNSAID)
Associated with worse outcomes
Increase in non-cardiac outcomes
Mentz RJ and Felker GM Heart Fail Clin 2013
Duke Heart Center
Risk of Co-Morbidity and Death
in HFpEF
bull COPD
bull Rheumatologic
Disorders
bull More dangerous in
HFpEF than HFrEF
AtherJACC2012
Total Hospitalizations I-Preserve
Many Co-Morbid Hospitalizations
CarsonJACC-HF in press
Mode of Death I-Preserve
60 cardiovascular death
40 non-cardiovascular death or unknown
May require a doubling of sample size for mortality component
Similar Issue with HFH
ZileCirc HF 2010
Mode of Non- Cardiovascular Death I-Preserve
ZileCirc HF 2010
Regional Differences in Therapy
OConnorJACC2012
Spironolactone
Placebo
HR = 089 (077 ndash 104)
p=0138
3511723 (204)
3201722 (186)
1degOutcome (CV Death HF Hosp or Resuscitated Cardiac Arrest)
Placebo Rates Primary Outcome by Region
US Canada
Argentina Brazil
Russia Rep Georgia
126 per 100 pt-
yr
23 per 100 pt-yr
Placebo
280881 (318)
Placebo
71842 (84)
HR=082 (069-098)
HR=110 (079-151)
Interaction p=0122
US Canada
Argentina Brazil
Russia Rep Georgia
Placebo
280881 (318)
Placebo
71842 (84)
Exploratory (post-hoc)Placebo vs Spiro
by Region
Of 22 pre-specified only 1 - Stratum - showed
a significant interaction with treatment
Enrolled
by Spiro Placebo
Hazard Ratio
(95 CI)
P-value
Natriuretic
peptide
78490
(159)
116491
(236)
065 (049-087)
0003
Heart Failure
Hosp
2421232
(196)
2351232
(191)
101 (084-121)
0923
P=0013 for interaction
Subgroups
ACEIARB Meta-Analysis Favors
Treatment(HR=088 p=005)
ShahJCF 2010
Hope for HFpEF
Duke Heart Center
LCZ696 Favorable on the most likely
Surrogate
bull Reduced NT-proBNP
bull Reduced LA size
bull Improved NYHA
Class
bull PARAGON
OUTCOME Trial
SolomonLancet2012
Have we failed in HFpEF
As clinical trialists we have sometimes failed(unknowingly) in the
design and conduct of HFpEF studies
Most traditional therapies(ACEIARBMRABB) have not failed(modest
reduction in recurrent HFH) but they have not passed regulatory
standards ndashWhat about guidelines
Must address the moderate number of Co- Morbidities differently
New therapies will have the advantage of our informed journey and will
likely yield positive results
HFpEF Trials
Half Empty Half Full hellip Yet No Banners
HFpEF and Comorbidities
Paulus WJ et al JACC 2013
PATHOPHYSIOLOGIC TARGETS
Diastolic dysfunction
Ventriculo-arterial disociation
Pulmonary Hypertension
Chronotropic Incompetance
Systemic Hypertension
HFpEF ndash Clinical Trials to Date
HFpEF ndash Landmark Trials
No proven therapy
All treatment recommendations ndash Class C evidence
Completed
Dig (Mortalityharr HF Hospitalizationsdarr)
CHARM-Preserved (trend towards darr HF hosp CV mortality)
PEP-CHF (trend towards darr mortality hf hospitalization frac14 of pts withdrew to go on open label ACEI at 1 year)
SENIORS - nebivolol (darr time to deathhf hospitalization but few LVEFgt50)
I-PRESERVE (ARB) did not improve outcomes
Recent Trials
TOPCAT (Spironolactone)
RELAX (sildenafil)
Paramount
Have HFpEF Trials Failed
Cleland JG Pellicori P Dierckx R
Heart Fail Clin 2014 Jul1
Reasons for Failure
Should the Trialists be on Trial
Was it Heart Failure
Different Demographics
Too many Co-Morbidities
Trial Design
HFpEF
Neither clinical history nor echo is a
reliable diagnostic method in patients
with HFpEF
Prior Hospitalization
Natriuretic Peptide Level
Natriuretic Peptide Enhances Risk
Cleland EHJ2006
Increase Mortality after Hospitalization
Both HF and non -HF
Carson JACC-HF
Risk of Events in HFpEF The I-Preserve Model
Previous hospitalization
Natriuretic Peptide
Age
Diabetes
Renal Disease
COPD
CAD
KomajdaCirc HF 2011
Duke Heart Center
A high degree of disease heterogeneity exists within heart failure patients and there is a need
for improved phenotyping of the syndrome
A new taxonomy of heart failure based on both clinical and molecular measures may provide a
more accurate classification of disease and ultimately enhance diagnosis and treatment
Duke Heart Center
Cluster Analysis
J Am Coll Cardiol 2014641765-74
Cluster analysis is an unsupervised learning task of grouping a set of
objects in such a way that objects in the same group are more similar
to each other than to those in other groups
Duke Heart Center
Cluster Analysis of Heart Failure to Uncover
Distinct Phenotypes
Circulation 2014 Nov 14
Duke Heart Center
Objective Measures of Heart Failure
Duke Heart Center
Clinical Implications on Outcomes
Duke Heart Center
bull Caucasian
bull Female
bull NICM
bull Few Co-Morbidities
Clinical Implications
Some Clusters Respond to Exercise
Implications of Co-Morbidities
Increase heterogeneity
Complicates management(Beta agonistsNSAID)
Associated with worse outcomes
Increase in non-cardiac outcomes
Mentz RJ and Felker GM Heart Fail Clin 2013
Duke Heart Center
Risk of Co-Morbidity and Death
in HFpEF
bull COPD
bull Rheumatologic
Disorders
bull More dangerous in
HFpEF than HFrEF
AtherJACC2012
Total Hospitalizations I-Preserve
Many Co-Morbid Hospitalizations
CarsonJACC-HF in press
Mode of Death I-Preserve
60 cardiovascular death
40 non-cardiovascular death or unknown
May require a doubling of sample size for mortality component
Similar Issue with HFH
ZileCirc HF 2010
Mode of Non- Cardiovascular Death I-Preserve
ZileCirc HF 2010
Regional Differences in Therapy
OConnorJACC2012
Spironolactone
Placebo
HR = 089 (077 ndash 104)
p=0138
3511723 (204)
3201722 (186)
1degOutcome (CV Death HF Hosp or Resuscitated Cardiac Arrest)
Placebo Rates Primary Outcome by Region
US Canada
Argentina Brazil
Russia Rep Georgia
126 per 100 pt-
yr
23 per 100 pt-yr
Placebo
280881 (318)
Placebo
71842 (84)
HR=082 (069-098)
HR=110 (079-151)
Interaction p=0122
US Canada
Argentina Brazil
Russia Rep Georgia
Placebo
280881 (318)
Placebo
71842 (84)
Exploratory (post-hoc)Placebo vs Spiro
by Region
Of 22 pre-specified only 1 - Stratum - showed
a significant interaction with treatment
Enrolled
by Spiro Placebo
Hazard Ratio
(95 CI)
P-value
Natriuretic
peptide
78490
(159)
116491
(236)
065 (049-087)
0003
Heart Failure
Hosp
2421232
(196)
2351232
(191)
101 (084-121)
0923
P=0013 for interaction
Subgroups
ACEIARB Meta-Analysis Favors
Treatment(HR=088 p=005)
ShahJCF 2010
Hope for HFpEF
Duke Heart Center
LCZ696 Favorable on the most likely
Surrogate
bull Reduced NT-proBNP
bull Reduced LA size
bull Improved NYHA
Class
bull PARAGON
OUTCOME Trial
SolomonLancet2012
Have we failed in HFpEF
As clinical trialists we have sometimes failed(unknowingly) in the
design and conduct of HFpEF studies
Most traditional therapies(ACEIARBMRABB) have not failed(modest
reduction in recurrent HFH) but they have not passed regulatory
standards ndashWhat about guidelines
Must address the moderate number of Co- Morbidities differently
New therapies will have the advantage of our informed journey and will
likely yield positive results
HFpEF Trials
Half Empty Half Full hellip Yet No Banners
PATHOPHYSIOLOGIC TARGETS
Diastolic dysfunction
Ventriculo-arterial disociation
Pulmonary Hypertension
Chronotropic Incompetance
Systemic Hypertension
HFpEF ndash Clinical Trials to Date
HFpEF ndash Landmark Trials
No proven therapy
All treatment recommendations ndash Class C evidence
Completed
Dig (Mortalityharr HF Hospitalizationsdarr)
CHARM-Preserved (trend towards darr HF hosp CV mortality)
PEP-CHF (trend towards darr mortality hf hospitalization frac14 of pts withdrew to go on open label ACEI at 1 year)
SENIORS - nebivolol (darr time to deathhf hospitalization but few LVEFgt50)
I-PRESERVE (ARB) did not improve outcomes
Recent Trials
TOPCAT (Spironolactone)
RELAX (sildenafil)
Paramount
Have HFpEF Trials Failed
Cleland JG Pellicori P Dierckx R
Heart Fail Clin 2014 Jul1
Reasons for Failure
Should the Trialists be on Trial
Was it Heart Failure
Different Demographics
Too many Co-Morbidities
Trial Design
HFpEF
Neither clinical history nor echo is a
reliable diagnostic method in patients
with HFpEF
Prior Hospitalization
Natriuretic Peptide Level
Natriuretic Peptide Enhances Risk
Cleland EHJ2006
Increase Mortality after Hospitalization
Both HF and non -HF
Carson JACC-HF
Risk of Events in HFpEF The I-Preserve Model
Previous hospitalization
Natriuretic Peptide
Age
Diabetes
Renal Disease
COPD
CAD
KomajdaCirc HF 2011
Duke Heart Center
A high degree of disease heterogeneity exists within heart failure patients and there is a need
for improved phenotyping of the syndrome
A new taxonomy of heart failure based on both clinical and molecular measures may provide a
more accurate classification of disease and ultimately enhance diagnosis and treatment
Duke Heart Center
Cluster Analysis
J Am Coll Cardiol 2014641765-74
Cluster analysis is an unsupervised learning task of grouping a set of
objects in such a way that objects in the same group are more similar
to each other than to those in other groups
Duke Heart Center
Cluster Analysis of Heart Failure to Uncover
Distinct Phenotypes
Circulation 2014 Nov 14
Duke Heart Center
Objective Measures of Heart Failure
Duke Heart Center
Clinical Implications on Outcomes
Duke Heart Center
bull Caucasian
bull Female
bull NICM
bull Few Co-Morbidities
Clinical Implications
Some Clusters Respond to Exercise
Implications of Co-Morbidities
Increase heterogeneity
Complicates management(Beta agonistsNSAID)
Associated with worse outcomes
Increase in non-cardiac outcomes
Mentz RJ and Felker GM Heart Fail Clin 2013
Duke Heart Center
Risk of Co-Morbidity and Death
in HFpEF
bull COPD
bull Rheumatologic
Disorders
bull More dangerous in
HFpEF than HFrEF
AtherJACC2012
Total Hospitalizations I-Preserve
Many Co-Morbid Hospitalizations
CarsonJACC-HF in press
Mode of Death I-Preserve
60 cardiovascular death
40 non-cardiovascular death or unknown
May require a doubling of sample size for mortality component
Similar Issue with HFH
ZileCirc HF 2010
Mode of Non- Cardiovascular Death I-Preserve
ZileCirc HF 2010
Regional Differences in Therapy
OConnorJACC2012
Spironolactone
Placebo
HR = 089 (077 ndash 104)
p=0138
3511723 (204)
3201722 (186)
1degOutcome (CV Death HF Hosp or Resuscitated Cardiac Arrest)
Placebo Rates Primary Outcome by Region
US Canada
Argentina Brazil
Russia Rep Georgia
126 per 100 pt-
yr
23 per 100 pt-yr
Placebo
280881 (318)
Placebo
71842 (84)
HR=082 (069-098)
HR=110 (079-151)
Interaction p=0122
US Canada
Argentina Brazil
Russia Rep Georgia
Placebo
280881 (318)
Placebo
71842 (84)
Exploratory (post-hoc)Placebo vs Spiro
by Region
Of 22 pre-specified only 1 - Stratum - showed
a significant interaction with treatment
Enrolled
by Spiro Placebo
Hazard Ratio
(95 CI)
P-value
Natriuretic
peptide
78490
(159)
116491
(236)
065 (049-087)
0003
Heart Failure
Hosp
2421232
(196)
2351232
(191)
101 (084-121)
0923
P=0013 for interaction
Subgroups
ACEIARB Meta-Analysis Favors
Treatment(HR=088 p=005)
ShahJCF 2010
Hope for HFpEF
Duke Heart Center
LCZ696 Favorable on the most likely
Surrogate
bull Reduced NT-proBNP
bull Reduced LA size
bull Improved NYHA
Class
bull PARAGON
OUTCOME Trial
SolomonLancet2012
Have we failed in HFpEF
As clinical trialists we have sometimes failed(unknowingly) in the
design and conduct of HFpEF studies
Most traditional therapies(ACEIARBMRABB) have not failed(modest
reduction in recurrent HFH) but they have not passed regulatory
standards ndashWhat about guidelines
Must address the moderate number of Co- Morbidities differently
New therapies will have the advantage of our informed journey and will
likely yield positive results
HFpEF Trials
Half Empty Half Full hellip Yet No Banners
HFpEF ndash Clinical Trials to Date
HFpEF ndash Landmark Trials
No proven therapy
All treatment recommendations ndash Class C evidence
Completed
Dig (Mortalityharr HF Hospitalizationsdarr)
CHARM-Preserved (trend towards darr HF hosp CV mortality)
PEP-CHF (trend towards darr mortality hf hospitalization frac14 of pts withdrew to go on open label ACEI at 1 year)
SENIORS - nebivolol (darr time to deathhf hospitalization but few LVEFgt50)
I-PRESERVE (ARB) did not improve outcomes
Recent Trials
TOPCAT (Spironolactone)
RELAX (sildenafil)
Paramount
Have HFpEF Trials Failed
Cleland JG Pellicori P Dierckx R
Heart Fail Clin 2014 Jul1
Reasons for Failure
Should the Trialists be on Trial
Was it Heart Failure
Different Demographics
Too many Co-Morbidities
Trial Design
HFpEF
Neither clinical history nor echo is a
reliable diagnostic method in patients
with HFpEF
Prior Hospitalization
Natriuretic Peptide Level
Natriuretic Peptide Enhances Risk
Cleland EHJ2006
Increase Mortality after Hospitalization
Both HF and non -HF
Carson JACC-HF
Risk of Events in HFpEF The I-Preserve Model
Previous hospitalization
Natriuretic Peptide
Age
Diabetes
Renal Disease
COPD
CAD
KomajdaCirc HF 2011
Duke Heart Center
A high degree of disease heterogeneity exists within heart failure patients and there is a need
for improved phenotyping of the syndrome
A new taxonomy of heart failure based on both clinical and molecular measures may provide a
more accurate classification of disease and ultimately enhance diagnosis and treatment
Duke Heart Center
Cluster Analysis
J Am Coll Cardiol 2014641765-74
Cluster analysis is an unsupervised learning task of grouping a set of
objects in such a way that objects in the same group are more similar
to each other than to those in other groups
Duke Heart Center
Cluster Analysis of Heart Failure to Uncover
Distinct Phenotypes
Circulation 2014 Nov 14
Duke Heart Center
Objective Measures of Heart Failure
Duke Heart Center
Clinical Implications on Outcomes
Duke Heart Center
bull Caucasian
bull Female
bull NICM
bull Few Co-Morbidities
Clinical Implications
Some Clusters Respond to Exercise
Implications of Co-Morbidities
Increase heterogeneity
Complicates management(Beta agonistsNSAID)
Associated with worse outcomes
Increase in non-cardiac outcomes
Mentz RJ and Felker GM Heart Fail Clin 2013
Duke Heart Center
Risk of Co-Morbidity and Death
in HFpEF
bull COPD
bull Rheumatologic
Disorders
bull More dangerous in
HFpEF than HFrEF
AtherJACC2012
Total Hospitalizations I-Preserve
Many Co-Morbid Hospitalizations
CarsonJACC-HF in press
Mode of Death I-Preserve
60 cardiovascular death
40 non-cardiovascular death or unknown
May require a doubling of sample size for mortality component
Similar Issue with HFH
ZileCirc HF 2010
Mode of Non- Cardiovascular Death I-Preserve
ZileCirc HF 2010
Regional Differences in Therapy
OConnorJACC2012
Spironolactone
Placebo
HR = 089 (077 ndash 104)
p=0138
3511723 (204)
3201722 (186)
1degOutcome (CV Death HF Hosp or Resuscitated Cardiac Arrest)
Placebo Rates Primary Outcome by Region
US Canada
Argentina Brazil
Russia Rep Georgia
126 per 100 pt-
yr
23 per 100 pt-yr
Placebo
280881 (318)
Placebo
71842 (84)
HR=082 (069-098)
HR=110 (079-151)
Interaction p=0122
US Canada
Argentina Brazil
Russia Rep Georgia
Placebo
280881 (318)
Placebo
71842 (84)
Exploratory (post-hoc)Placebo vs Spiro
by Region
Of 22 pre-specified only 1 - Stratum - showed
a significant interaction with treatment
Enrolled
by Spiro Placebo
Hazard Ratio
(95 CI)
P-value
Natriuretic
peptide
78490
(159)
116491
(236)
065 (049-087)
0003
Heart Failure
Hosp
2421232
(196)
2351232
(191)
101 (084-121)
0923
P=0013 for interaction
Subgroups
ACEIARB Meta-Analysis Favors
Treatment(HR=088 p=005)
ShahJCF 2010
Hope for HFpEF
Duke Heart Center
LCZ696 Favorable on the most likely
Surrogate
bull Reduced NT-proBNP
bull Reduced LA size
bull Improved NYHA
Class
bull PARAGON
OUTCOME Trial
SolomonLancet2012
Have we failed in HFpEF
As clinical trialists we have sometimes failed(unknowingly) in the
design and conduct of HFpEF studies
Most traditional therapies(ACEIARBMRABB) have not failed(modest
reduction in recurrent HFH) but they have not passed regulatory
standards ndashWhat about guidelines
Must address the moderate number of Co- Morbidities differently
New therapies will have the advantage of our informed journey and will
likely yield positive results
HFpEF Trials
Half Empty Half Full hellip Yet No Banners
HFpEF ndash Landmark Trials
No proven therapy
All treatment recommendations ndash Class C evidence
Completed
Dig (Mortalityharr HF Hospitalizationsdarr)
CHARM-Preserved (trend towards darr HF hosp CV mortality)
PEP-CHF (trend towards darr mortality hf hospitalization frac14 of pts withdrew to go on open label ACEI at 1 year)
SENIORS - nebivolol (darr time to deathhf hospitalization but few LVEFgt50)
I-PRESERVE (ARB) did not improve outcomes
Recent Trials
TOPCAT (Spironolactone)
RELAX (sildenafil)
Paramount
Have HFpEF Trials Failed
Cleland JG Pellicori P Dierckx R
Heart Fail Clin 2014 Jul1
Reasons for Failure
Should the Trialists be on Trial
Was it Heart Failure
Different Demographics
Too many Co-Morbidities
Trial Design
HFpEF
Neither clinical history nor echo is a
reliable diagnostic method in patients
with HFpEF
Prior Hospitalization
Natriuretic Peptide Level
Natriuretic Peptide Enhances Risk
Cleland EHJ2006
Increase Mortality after Hospitalization
Both HF and non -HF
Carson JACC-HF
Risk of Events in HFpEF The I-Preserve Model
Previous hospitalization
Natriuretic Peptide
Age
Diabetes
Renal Disease
COPD
CAD
KomajdaCirc HF 2011
Duke Heart Center
A high degree of disease heterogeneity exists within heart failure patients and there is a need
for improved phenotyping of the syndrome
A new taxonomy of heart failure based on both clinical and molecular measures may provide a
more accurate classification of disease and ultimately enhance diagnosis and treatment
Duke Heart Center
Cluster Analysis
J Am Coll Cardiol 2014641765-74
Cluster analysis is an unsupervised learning task of grouping a set of
objects in such a way that objects in the same group are more similar
to each other than to those in other groups
Duke Heart Center
Cluster Analysis of Heart Failure to Uncover
Distinct Phenotypes
Circulation 2014 Nov 14
Duke Heart Center
Objective Measures of Heart Failure
Duke Heart Center
Clinical Implications on Outcomes
Duke Heart Center
bull Caucasian
bull Female
bull NICM
bull Few Co-Morbidities
Clinical Implications
Some Clusters Respond to Exercise
Implications of Co-Morbidities
Increase heterogeneity
Complicates management(Beta agonistsNSAID)
Associated with worse outcomes
Increase in non-cardiac outcomes
Mentz RJ and Felker GM Heart Fail Clin 2013
Duke Heart Center
Risk of Co-Morbidity and Death
in HFpEF
bull COPD
bull Rheumatologic
Disorders
bull More dangerous in
HFpEF than HFrEF
AtherJACC2012
Total Hospitalizations I-Preserve
Many Co-Morbid Hospitalizations
CarsonJACC-HF in press
Mode of Death I-Preserve
60 cardiovascular death
40 non-cardiovascular death or unknown
May require a doubling of sample size for mortality component
Similar Issue with HFH
ZileCirc HF 2010
Mode of Non- Cardiovascular Death I-Preserve
ZileCirc HF 2010
Regional Differences in Therapy
OConnorJACC2012
Spironolactone
Placebo
HR = 089 (077 ndash 104)
p=0138
3511723 (204)
3201722 (186)
1degOutcome (CV Death HF Hosp or Resuscitated Cardiac Arrest)
Placebo Rates Primary Outcome by Region
US Canada
Argentina Brazil
Russia Rep Georgia
126 per 100 pt-
yr
23 per 100 pt-yr
Placebo
280881 (318)
Placebo
71842 (84)
HR=082 (069-098)
HR=110 (079-151)
Interaction p=0122
US Canada
Argentina Brazil
Russia Rep Georgia
Placebo
280881 (318)
Placebo
71842 (84)
Exploratory (post-hoc)Placebo vs Spiro
by Region
Of 22 pre-specified only 1 - Stratum - showed
a significant interaction with treatment
Enrolled
by Spiro Placebo
Hazard Ratio
(95 CI)
P-value
Natriuretic
peptide
78490
(159)
116491
(236)
065 (049-087)
0003
Heart Failure
Hosp
2421232
(196)
2351232
(191)
101 (084-121)
0923
P=0013 for interaction
Subgroups
ACEIARB Meta-Analysis Favors
Treatment(HR=088 p=005)
ShahJCF 2010
Hope for HFpEF
Duke Heart Center
LCZ696 Favorable on the most likely
Surrogate
bull Reduced NT-proBNP
bull Reduced LA size
bull Improved NYHA
Class
bull PARAGON
OUTCOME Trial
SolomonLancet2012
Have we failed in HFpEF
As clinical trialists we have sometimes failed(unknowingly) in the
design and conduct of HFpEF studies
Most traditional therapies(ACEIARBMRABB) have not failed(modest
reduction in recurrent HFH) but they have not passed regulatory
standards ndashWhat about guidelines
Must address the moderate number of Co- Morbidities differently
New therapies will have the advantage of our informed journey and will
likely yield positive results
HFpEF Trials
Half Empty Half Full hellip Yet No Banners
Have HFpEF Trials Failed
Cleland JG Pellicori P Dierckx R
Heart Fail Clin 2014 Jul1
Reasons for Failure
Should the Trialists be on Trial
Was it Heart Failure
Different Demographics
Too many Co-Morbidities
Trial Design
HFpEF
Neither clinical history nor echo is a
reliable diagnostic method in patients
with HFpEF
Prior Hospitalization
Natriuretic Peptide Level
Natriuretic Peptide Enhances Risk
Cleland EHJ2006
Increase Mortality after Hospitalization
Both HF and non -HF
Carson JACC-HF
Risk of Events in HFpEF The I-Preserve Model
Previous hospitalization
Natriuretic Peptide
Age
Diabetes
Renal Disease
COPD
CAD
KomajdaCirc HF 2011
Duke Heart Center
A high degree of disease heterogeneity exists within heart failure patients and there is a need
for improved phenotyping of the syndrome
A new taxonomy of heart failure based on both clinical and molecular measures may provide a
more accurate classification of disease and ultimately enhance diagnosis and treatment
Duke Heart Center
Cluster Analysis
J Am Coll Cardiol 2014641765-74
Cluster analysis is an unsupervised learning task of grouping a set of
objects in such a way that objects in the same group are more similar
to each other than to those in other groups
Duke Heart Center
Cluster Analysis of Heart Failure to Uncover
Distinct Phenotypes
Circulation 2014 Nov 14
Duke Heart Center
Objective Measures of Heart Failure
Duke Heart Center
Clinical Implications on Outcomes
Duke Heart Center
bull Caucasian
bull Female
bull NICM
bull Few Co-Morbidities
Clinical Implications
Some Clusters Respond to Exercise
Implications of Co-Morbidities
Increase heterogeneity
Complicates management(Beta agonistsNSAID)
Associated with worse outcomes
Increase in non-cardiac outcomes
Mentz RJ and Felker GM Heart Fail Clin 2013
Duke Heart Center
Risk of Co-Morbidity and Death
in HFpEF
bull COPD
bull Rheumatologic
Disorders
bull More dangerous in
HFpEF than HFrEF
AtherJACC2012
Total Hospitalizations I-Preserve
Many Co-Morbid Hospitalizations
CarsonJACC-HF in press
Mode of Death I-Preserve
60 cardiovascular death
40 non-cardiovascular death or unknown
May require a doubling of sample size for mortality component
Similar Issue with HFH
ZileCirc HF 2010
Mode of Non- Cardiovascular Death I-Preserve
ZileCirc HF 2010
Regional Differences in Therapy
OConnorJACC2012
Spironolactone
Placebo
HR = 089 (077 ndash 104)
p=0138
3511723 (204)
3201722 (186)
1degOutcome (CV Death HF Hosp or Resuscitated Cardiac Arrest)
Placebo Rates Primary Outcome by Region
US Canada
Argentina Brazil
Russia Rep Georgia
126 per 100 pt-
yr
23 per 100 pt-yr
Placebo
280881 (318)
Placebo
71842 (84)
HR=082 (069-098)
HR=110 (079-151)
Interaction p=0122
US Canada
Argentina Brazil
Russia Rep Georgia
Placebo
280881 (318)
Placebo
71842 (84)
Exploratory (post-hoc)Placebo vs Spiro
by Region
Of 22 pre-specified only 1 - Stratum - showed
a significant interaction with treatment
Enrolled
by Spiro Placebo
Hazard Ratio
(95 CI)
P-value
Natriuretic
peptide
78490
(159)
116491
(236)
065 (049-087)
0003
Heart Failure
Hosp
2421232
(196)
2351232
(191)
101 (084-121)
0923
P=0013 for interaction
Subgroups
ACEIARB Meta-Analysis Favors
Treatment(HR=088 p=005)
ShahJCF 2010
Hope for HFpEF
Duke Heart Center
LCZ696 Favorable on the most likely
Surrogate
bull Reduced NT-proBNP
bull Reduced LA size
bull Improved NYHA
Class
bull PARAGON
OUTCOME Trial
SolomonLancet2012
Have we failed in HFpEF
As clinical trialists we have sometimes failed(unknowingly) in the
design and conduct of HFpEF studies
Most traditional therapies(ACEIARBMRABB) have not failed(modest
reduction in recurrent HFH) but they have not passed regulatory
standards ndashWhat about guidelines
Must address the moderate number of Co- Morbidities differently
New therapies will have the advantage of our informed journey and will
likely yield positive results
HFpEF Trials
Half Empty Half Full hellip Yet No Banners
Reasons for Failure
Should the Trialists be on Trial
Was it Heart Failure
Different Demographics
Too many Co-Morbidities
Trial Design
HFpEF
Neither clinical history nor echo is a
reliable diagnostic method in patients
with HFpEF
Prior Hospitalization
Natriuretic Peptide Level
Natriuretic Peptide Enhances Risk
Cleland EHJ2006
Increase Mortality after Hospitalization
Both HF and non -HF
Carson JACC-HF
Risk of Events in HFpEF The I-Preserve Model
Previous hospitalization
Natriuretic Peptide
Age
Diabetes
Renal Disease
COPD
CAD
KomajdaCirc HF 2011
Duke Heart Center
A high degree of disease heterogeneity exists within heart failure patients and there is a need
for improved phenotyping of the syndrome
A new taxonomy of heart failure based on both clinical and molecular measures may provide a
more accurate classification of disease and ultimately enhance diagnosis and treatment
Duke Heart Center
Cluster Analysis
J Am Coll Cardiol 2014641765-74
Cluster analysis is an unsupervised learning task of grouping a set of
objects in such a way that objects in the same group are more similar
to each other than to those in other groups
Duke Heart Center
Cluster Analysis of Heart Failure to Uncover
Distinct Phenotypes
Circulation 2014 Nov 14
Duke Heart Center
Objective Measures of Heart Failure
Duke Heart Center
Clinical Implications on Outcomes
Duke Heart Center
bull Caucasian
bull Female
bull NICM
bull Few Co-Morbidities
Clinical Implications
Some Clusters Respond to Exercise
Implications of Co-Morbidities
Increase heterogeneity
Complicates management(Beta agonistsNSAID)
Associated with worse outcomes
Increase in non-cardiac outcomes
Mentz RJ and Felker GM Heart Fail Clin 2013
Duke Heart Center
Risk of Co-Morbidity and Death
in HFpEF
bull COPD
bull Rheumatologic
Disorders
bull More dangerous in
HFpEF than HFrEF
AtherJACC2012
Total Hospitalizations I-Preserve
Many Co-Morbid Hospitalizations
CarsonJACC-HF in press
Mode of Death I-Preserve
60 cardiovascular death
40 non-cardiovascular death or unknown
May require a doubling of sample size for mortality component
Similar Issue with HFH
ZileCirc HF 2010
Mode of Non- Cardiovascular Death I-Preserve
ZileCirc HF 2010
Regional Differences in Therapy
OConnorJACC2012
Spironolactone
Placebo
HR = 089 (077 ndash 104)
p=0138
3511723 (204)
3201722 (186)
1degOutcome (CV Death HF Hosp or Resuscitated Cardiac Arrest)
Placebo Rates Primary Outcome by Region
US Canada
Argentina Brazil
Russia Rep Georgia
126 per 100 pt-
yr
23 per 100 pt-yr
Placebo
280881 (318)
Placebo
71842 (84)
HR=082 (069-098)
HR=110 (079-151)
Interaction p=0122
US Canada
Argentina Brazil
Russia Rep Georgia
Placebo
280881 (318)
Placebo
71842 (84)
Exploratory (post-hoc)Placebo vs Spiro
by Region
Of 22 pre-specified only 1 - Stratum - showed
a significant interaction with treatment
Enrolled
by Spiro Placebo
Hazard Ratio
(95 CI)
P-value
Natriuretic
peptide
78490
(159)
116491
(236)
065 (049-087)
0003
Heart Failure
Hosp
2421232
(196)
2351232
(191)
101 (084-121)
0923
P=0013 for interaction
Subgroups
ACEIARB Meta-Analysis Favors
Treatment(HR=088 p=005)
ShahJCF 2010
Hope for HFpEF
Duke Heart Center
LCZ696 Favorable on the most likely
Surrogate
bull Reduced NT-proBNP
bull Reduced LA size
bull Improved NYHA
Class
bull PARAGON
OUTCOME Trial
SolomonLancet2012
Have we failed in HFpEF
As clinical trialists we have sometimes failed(unknowingly) in the
design and conduct of HFpEF studies
Most traditional therapies(ACEIARBMRABB) have not failed(modest
reduction in recurrent HFH) but they have not passed regulatory
standards ndashWhat about guidelines
Must address the moderate number of Co- Morbidities differently
New therapies will have the advantage of our informed journey and will
likely yield positive results
HFpEF Trials
Half Empty Half Full hellip Yet No Banners
HFpEF
Neither clinical history nor echo is a
reliable diagnostic method in patients
with HFpEF
Prior Hospitalization
Natriuretic Peptide Level
Natriuretic Peptide Enhances Risk
Cleland EHJ2006
Increase Mortality after Hospitalization
Both HF and non -HF
Carson JACC-HF
Risk of Events in HFpEF The I-Preserve Model
Previous hospitalization
Natriuretic Peptide
Age
Diabetes
Renal Disease
COPD
CAD
KomajdaCirc HF 2011
Duke Heart Center
A high degree of disease heterogeneity exists within heart failure patients and there is a need
for improved phenotyping of the syndrome
A new taxonomy of heart failure based on both clinical and molecular measures may provide a
more accurate classification of disease and ultimately enhance diagnosis and treatment
Duke Heart Center
Cluster Analysis
J Am Coll Cardiol 2014641765-74
Cluster analysis is an unsupervised learning task of grouping a set of
objects in such a way that objects in the same group are more similar
to each other than to those in other groups
Duke Heart Center
Cluster Analysis of Heart Failure to Uncover
Distinct Phenotypes
Circulation 2014 Nov 14
Duke Heart Center
Objective Measures of Heart Failure
Duke Heart Center
Clinical Implications on Outcomes
Duke Heart Center
bull Caucasian
bull Female
bull NICM
bull Few Co-Morbidities
Clinical Implications
Some Clusters Respond to Exercise
Implications of Co-Morbidities
Increase heterogeneity
Complicates management(Beta agonistsNSAID)
Associated with worse outcomes
Increase in non-cardiac outcomes
Mentz RJ and Felker GM Heart Fail Clin 2013
Duke Heart Center
Risk of Co-Morbidity and Death
in HFpEF
bull COPD
bull Rheumatologic
Disorders
bull More dangerous in
HFpEF than HFrEF
AtherJACC2012
Total Hospitalizations I-Preserve
Many Co-Morbid Hospitalizations
CarsonJACC-HF in press
Mode of Death I-Preserve
60 cardiovascular death
40 non-cardiovascular death or unknown
May require a doubling of sample size for mortality component
Similar Issue with HFH
ZileCirc HF 2010
Mode of Non- Cardiovascular Death I-Preserve
ZileCirc HF 2010
Regional Differences in Therapy
OConnorJACC2012
Spironolactone
Placebo
HR = 089 (077 ndash 104)
p=0138
3511723 (204)
3201722 (186)
1degOutcome (CV Death HF Hosp or Resuscitated Cardiac Arrest)
Placebo Rates Primary Outcome by Region
US Canada
Argentina Brazil
Russia Rep Georgia
126 per 100 pt-
yr
23 per 100 pt-yr
Placebo
280881 (318)
Placebo
71842 (84)
HR=082 (069-098)
HR=110 (079-151)
Interaction p=0122
US Canada
Argentina Brazil
Russia Rep Georgia
Placebo
280881 (318)
Placebo
71842 (84)
Exploratory (post-hoc)Placebo vs Spiro
by Region
Of 22 pre-specified only 1 - Stratum - showed
a significant interaction with treatment
Enrolled
by Spiro Placebo
Hazard Ratio
(95 CI)
P-value
Natriuretic
peptide
78490
(159)
116491
(236)
065 (049-087)
0003
Heart Failure
Hosp
2421232
(196)
2351232
(191)
101 (084-121)
0923
P=0013 for interaction
Subgroups
ACEIARB Meta-Analysis Favors
Treatment(HR=088 p=005)
ShahJCF 2010
Hope for HFpEF
Duke Heart Center
LCZ696 Favorable on the most likely
Surrogate
bull Reduced NT-proBNP
bull Reduced LA size
bull Improved NYHA
Class
bull PARAGON
OUTCOME Trial
SolomonLancet2012
Have we failed in HFpEF
As clinical trialists we have sometimes failed(unknowingly) in the
design and conduct of HFpEF studies
Most traditional therapies(ACEIARBMRABB) have not failed(modest
reduction in recurrent HFH) but they have not passed regulatory
standards ndashWhat about guidelines
Must address the moderate number of Co- Morbidities differently
New therapies will have the advantage of our informed journey and will
likely yield positive results
HFpEF Trials
Half Empty Half Full hellip Yet No Banners
Natriuretic Peptide Enhances Risk
Cleland EHJ2006
Increase Mortality after Hospitalization
Both HF and non -HF
Carson JACC-HF
Risk of Events in HFpEF The I-Preserve Model
Previous hospitalization
Natriuretic Peptide
Age
Diabetes
Renal Disease
COPD
CAD
KomajdaCirc HF 2011
Duke Heart Center
A high degree of disease heterogeneity exists within heart failure patients and there is a need
for improved phenotyping of the syndrome
A new taxonomy of heart failure based on both clinical and molecular measures may provide a
more accurate classification of disease and ultimately enhance diagnosis and treatment
Duke Heart Center
Cluster Analysis
J Am Coll Cardiol 2014641765-74
Cluster analysis is an unsupervised learning task of grouping a set of
objects in such a way that objects in the same group are more similar
to each other than to those in other groups
Duke Heart Center
Cluster Analysis of Heart Failure to Uncover
Distinct Phenotypes
Circulation 2014 Nov 14
Duke Heart Center
Objective Measures of Heart Failure
Duke Heart Center
Clinical Implications on Outcomes
Duke Heart Center
bull Caucasian
bull Female
bull NICM
bull Few Co-Morbidities
Clinical Implications
Some Clusters Respond to Exercise
Implications of Co-Morbidities
Increase heterogeneity
Complicates management(Beta agonistsNSAID)
Associated with worse outcomes
Increase in non-cardiac outcomes
Mentz RJ and Felker GM Heart Fail Clin 2013
Duke Heart Center
Risk of Co-Morbidity and Death
in HFpEF
bull COPD
bull Rheumatologic
Disorders
bull More dangerous in
HFpEF than HFrEF
AtherJACC2012
Total Hospitalizations I-Preserve
Many Co-Morbid Hospitalizations
CarsonJACC-HF in press
Mode of Death I-Preserve
60 cardiovascular death
40 non-cardiovascular death or unknown
May require a doubling of sample size for mortality component
Similar Issue with HFH
ZileCirc HF 2010
Mode of Non- Cardiovascular Death I-Preserve
ZileCirc HF 2010
Regional Differences in Therapy
OConnorJACC2012
Spironolactone
Placebo
HR = 089 (077 ndash 104)
p=0138
3511723 (204)
3201722 (186)
1degOutcome (CV Death HF Hosp or Resuscitated Cardiac Arrest)
Placebo Rates Primary Outcome by Region
US Canada
Argentina Brazil
Russia Rep Georgia
126 per 100 pt-
yr
23 per 100 pt-yr
Placebo
280881 (318)
Placebo
71842 (84)
HR=082 (069-098)
HR=110 (079-151)
Interaction p=0122
US Canada
Argentina Brazil
Russia Rep Georgia
Placebo
280881 (318)
Placebo
71842 (84)
Exploratory (post-hoc)Placebo vs Spiro
by Region
Of 22 pre-specified only 1 - Stratum - showed
a significant interaction with treatment
Enrolled
by Spiro Placebo
Hazard Ratio
(95 CI)
P-value
Natriuretic
peptide
78490
(159)
116491
(236)
065 (049-087)
0003
Heart Failure
Hosp
2421232
(196)
2351232
(191)
101 (084-121)
0923
P=0013 for interaction
Subgroups
ACEIARB Meta-Analysis Favors
Treatment(HR=088 p=005)
ShahJCF 2010
Hope for HFpEF
Duke Heart Center
LCZ696 Favorable on the most likely
Surrogate
bull Reduced NT-proBNP
bull Reduced LA size
bull Improved NYHA
Class
bull PARAGON
OUTCOME Trial
SolomonLancet2012
Have we failed in HFpEF
As clinical trialists we have sometimes failed(unknowingly) in the
design and conduct of HFpEF studies
Most traditional therapies(ACEIARBMRABB) have not failed(modest
reduction in recurrent HFH) but they have not passed regulatory
standards ndashWhat about guidelines
Must address the moderate number of Co- Morbidities differently
New therapies will have the advantage of our informed journey and will
likely yield positive results
HFpEF Trials
Half Empty Half Full hellip Yet No Banners
Increase Mortality after Hospitalization
Both HF and non -HF
Carson JACC-HF
Risk of Events in HFpEF The I-Preserve Model
Previous hospitalization
Natriuretic Peptide
Age
Diabetes
Renal Disease
COPD
CAD
KomajdaCirc HF 2011
Duke Heart Center
A high degree of disease heterogeneity exists within heart failure patients and there is a need
for improved phenotyping of the syndrome
A new taxonomy of heart failure based on both clinical and molecular measures may provide a
more accurate classification of disease and ultimately enhance diagnosis and treatment
Duke Heart Center
Cluster Analysis
J Am Coll Cardiol 2014641765-74
Cluster analysis is an unsupervised learning task of grouping a set of
objects in such a way that objects in the same group are more similar
to each other than to those in other groups
Duke Heart Center
Cluster Analysis of Heart Failure to Uncover
Distinct Phenotypes
Circulation 2014 Nov 14
Duke Heart Center
Objective Measures of Heart Failure
Duke Heart Center
Clinical Implications on Outcomes
Duke Heart Center
bull Caucasian
bull Female
bull NICM
bull Few Co-Morbidities
Clinical Implications
Some Clusters Respond to Exercise
Implications of Co-Morbidities
Increase heterogeneity
Complicates management(Beta agonistsNSAID)
Associated with worse outcomes
Increase in non-cardiac outcomes
Mentz RJ and Felker GM Heart Fail Clin 2013
Duke Heart Center
Risk of Co-Morbidity and Death
in HFpEF
bull COPD
bull Rheumatologic
Disorders
bull More dangerous in
HFpEF than HFrEF
AtherJACC2012
Total Hospitalizations I-Preserve
Many Co-Morbid Hospitalizations
CarsonJACC-HF in press
Mode of Death I-Preserve
60 cardiovascular death
40 non-cardiovascular death or unknown
May require a doubling of sample size for mortality component
Similar Issue with HFH
ZileCirc HF 2010
Mode of Non- Cardiovascular Death I-Preserve
ZileCirc HF 2010
Regional Differences in Therapy
OConnorJACC2012
Spironolactone
Placebo
HR = 089 (077 ndash 104)
p=0138
3511723 (204)
3201722 (186)
1degOutcome (CV Death HF Hosp or Resuscitated Cardiac Arrest)
Placebo Rates Primary Outcome by Region
US Canada
Argentina Brazil
Russia Rep Georgia
126 per 100 pt-
yr
23 per 100 pt-yr
Placebo
280881 (318)
Placebo
71842 (84)
HR=082 (069-098)
HR=110 (079-151)
Interaction p=0122
US Canada
Argentina Brazil
Russia Rep Georgia
Placebo
280881 (318)
Placebo
71842 (84)
Exploratory (post-hoc)Placebo vs Spiro
by Region
Of 22 pre-specified only 1 - Stratum - showed
a significant interaction with treatment
Enrolled
by Spiro Placebo
Hazard Ratio
(95 CI)
P-value
Natriuretic
peptide
78490
(159)
116491
(236)
065 (049-087)
0003
Heart Failure
Hosp
2421232
(196)
2351232
(191)
101 (084-121)
0923
P=0013 for interaction
Subgroups
ACEIARB Meta-Analysis Favors
Treatment(HR=088 p=005)
ShahJCF 2010
Hope for HFpEF
Duke Heart Center
LCZ696 Favorable on the most likely
Surrogate
bull Reduced NT-proBNP
bull Reduced LA size
bull Improved NYHA
Class
bull PARAGON
OUTCOME Trial
SolomonLancet2012
Have we failed in HFpEF
As clinical trialists we have sometimes failed(unknowingly) in the
design and conduct of HFpEF studies
Most traditional therapies(ACEIARBMRABB) have not failed(modest
reduction in recurrent HFH) but they have not passed regulatory
standards ndashWhat about guidelines
Must address the moderate number of Co- Morbidities differently
New therapies will have the advantage of our informed journey and will
likely yield positive results
HFpEF Trials
Half Empty Half Full hellip Yet No Banners
Risk of Events in HFpEF The I-Preserve Model
Previous hospitalization
Natriuretic Peptide
Age
Diabetes
Renal Disease
COPD
CAD
KomajdaCirc HF 2011
Duke Heart Center
A high degree of disease heterogeneity exists within heart failure patients and there is a need
for improved phenotyping of the syndrome
A new taxonomy of heart failure based on both clinical and molecular measures may provide a
more accurate classification of disease and ultimately enhance diagnosis and treatment
Duke Heart Center
Cluster Analysis
J Am Coll Cardiol 2014641765-74
Cluster analysis is an unsupervised learning task of grouping a set of
objects in such a way that objects in the same group are more similar
to each other than to those in other groups
Duke Heart Center
Cluster Analysis of Heart Failure to Uncover
Distinct Phenotypes
Circulation 2014 Nov 14
Duke Heart Center
Objective Measures of Heart Failure
Duke Heart Center
Clinical Implications on Outcomes
Duke Heart Center
bull Caucasian
bull Female
bull NICM
bull Few Co-Morbidities
Clinical Implications
Some Clusters Respond to Exercise
Implications of Co-Morbidities
Increase heterogeneity
Complicates management(Beta agonistsNSAID)
Associated with worse outcomes
Increase in non-cardiac outcomes
Mentz RJ and Felker GM Heart Fail Clin 2013
Duke Heart Center
Risk of Co-Morbidity and Death
in HFpEF
bull COPD
bull Rheumatologic
Disorders
bull More dangerous in
HFpEF than HFrEF
AtherJACC2012
Total Hospitalizations I-Preserve
Many Co-Morbid Hospitalizations
CarsonJACC-HF in press
Mode of Death I-Preserve
60 cardiovascular death
40 non-cardiovascular death or unknown
May require a doubling of sample size for mortality component
Similar Issue with HFH
ZileCirc HF 2010
Mode of Non- Cardiovascular Death I-Preserve
ZileCirc HF 2010
Regional Differences in Therapy
OConnorJACC2012
Spironolactone
Placebo
HR = 089 (077 ndash 104)
p=0138
3511723 (204)
3201722 (186)
1degOutcome (CV Death HF Hosp or Resuscitated Cardiac Arrest)
Placebo Rates Primary Outcome by Region
US Canada
Argentina Brazil
Russia Rep Georgia
126 per 100 pt-
yr
23 per 100 pt-yr
Placebo
280881 (318)
Placebo
71842 (84)
HR=082 (069-098)
HR=110 (079-151)
Interaction p=0122
US Canada
Argentina Brazil
Russia Rep Georgia
Placebo
280881 (318)
Placebo
71842 (84)
Exploratory (post-hoc)Placebo vs Spiro
by Region
Of 22 pre-specified only 1 - Stratum - showed
a significant interaction with treatment
Enrolled
by Spiro Placebo
Hazard Ratio
(95 CI)
P-value
Natriuretic
peptide
78490
(159)
116491
(236)
065 (049-087)
0003
Heart Failure
Hosp
2421232
(196)
2351232
(191)
101 (084-121)
0923
P=0013 for interaction
Subgroups
ACEIARB Meta-Analysis Favors
Treatment(HR=088 p=005)
ShahJCF 2010
Hope for HFpEF
Duke Heart Center
LCZ696 Favorable on the most likely
Surrogate
bull Reduced NT-proBNP
bull Reduced LA size
bull Improved NYHA
Class
bull PARAGON
OUTCOME Trial
SolomonLancet2012
Have we failed in HFpEF
As clinical trialists we have sometimes failed(unknowingly) in the
design and conduct of HFpEF studies
Most traditional therapies(ACEIARBMRABB) have not failed(modest
reduction in recurrent HFH) but they have not passed regulatory
standards ndashWhat about guidelines
Must address the moderate number of Co- Morbidities differently
New therapies will have the advantage of our informed journey and will
likely yield positive results
HFpEF Trials
Half Empty Half Full hellip Yet No Banners
Duke Heart Center
A high degree of disease heterogeneity exists within heart failure patients and there is a need
for improved phenotyping of the syndrome
A new taxonomy of heart failure based on both clinical and molecular measures may provide a
more accurate classification of disease and ultimately enhance diagnosis and treatment
Duke Heart Center
Cluster Analysis
J Am Coll Cardiol 2014641765-74
Cluster analysis is an unsupervised learning task of grouping a set of
objects in such a way that objects in the same group are more similar
to each other than to those in other groups
Duke Heart Center
Cluster Analysis of Heart Failure to Uncover
Distinct Phenotypes
Circulation 2014 Nov 14
Duke Heart Center
Objective Measures of Heart Failure
Duke Heart Center
Clinical Implications on Outcomes
Duke Heart Center
bull Caucasian
bull Female
bull NICM
bull Few Co-Morbidities
Clinical Implications
Some Clusters Respond to Exercise
Implications of Co-Morbidities
Increase heterogeneity
Complicates management(Beta agonistsNSAID)
Associated with worse outcomes
Increase in non-cardiac outcomes
Mentz RJ and Felker GM Heart Fail Clin 2013
Duke Heart Center
Risk of Co-Morbidity and Death
in HFpEF
bull COPD
bull Rheumatologic
Disorders
bull More dangerous in
HFpEF than HFrEF
AtherJACC2012
Total Hospitalizations I-Preserve
Many Co-Morbid Hospitalizations
CarsonJACC-HF in press
Mode of Death I-Preserve
60 cardiovascular death
40 non-cardiovascular death or unknown
May require a doubling of sample size for mortality component
Similar Issue with HFH
ZileCirc HF 2010
Mode of Non- Cardiovascular Death I-Preserve
ZileCirc HF 2010
Regional Differences in Therapy
OConnorJACC2012
Spironolactone
Placebo
HR = 089 (077 ndash 104)
p=0138
3511723 (204)
3201722 (186)
1degOutcome (CV Death HF Hosp or Resuscitated Cardiac Arrest)
Placebo Rates Primary Outcome by Region
US Canada
Argentina Brazil
Russia Rep Georgia
126 per 100 pt-
yr
23 per 100 pt-yr
Placebo
280881 (318)
Placebo
71842 (84)
HR=082 (069-098)
HR=110 (079-151)
Interaction p=0122
US Canada
Argentina Brazil
Russia Rep Georgia
Placebo
280881 (318)
Placebo
71842 (84)
Exploratory (post-hoc)Placebo vs Spiro
by Region
Of 22 pre-specified only 1 - Stratum - showed
a significant interaction with treatment
Enrolled
by Spiro Placebo
Hazard Ratio
(95 CI)
P-value
Natriuretic
peptide
78490
(159)
116491
(236)
065 (049-087)
0003
Heart Failure
Hosp
2421232
(196)
2351232
(191)
101 (084-121)
0923
P=0013 for interaction
Subgroups
ACEIARB Meta-Analysis Favors
Treatment(HR=088 p=005)
ShahJCF 2010
Hope for HFpEF
Duke Heart Center
LCZ696 Favorable on the most likely
Surrogate
bull Reduced NT-proBNP
bull Reduced LA size
bull Improved NYHA
Class
bull PARAGON
OUTCOME Trial
SolomonLancet2012
Have we failed in HFpEF
As clinical trialists we have sometimes failed(unknowingly) in the
design and conduct of HFpEF studies
Most traditional therapies(ACEIARBMRABB) have not failed(modest
reduction in recurrent HFH) but they have not passed regulatory
standards ndashWhat about guidelines
Must address the moderate number of Co- Morbidities differently
New therapies will have the advantage of our informed journey and will
likely yield positive results
HFpEF Trials
Half Empty Half Full hellip Yet No Banners
Duke Heart Center
Cluster Analysis
J Am Coll Cardiol 2014641765-74
Cluster analysis is an unsupervised learning task of grouping a set of
objects in such a way that objects in the same group are more similar
to each other than to those in other groups
Duke Heart Center
Cluster Analysis of Heart Failure to Uncover
Distinct Phenotypes
Circulation 2014 Nov 14
Duke Heart Center
Objective Measures of Heart Failure
Duke Heart Center
Clinical Implications on Outcomes
Duke Heart Center
bull Caucasian
bull Female
bull NICM
bull Few Co-Morbidities
Clinical Implications
Some Clusters Respond to Exercise
Implications of Co-Morbidities
Increase heterogeneity
Complicates management(Beta agonistsNSAID)
Associated with worse outcomes
Increase in non-cardiac outcomes
Mentz RJ and Felker GM Heart Fail Clin 2013
Duke Heart Center
Risk of Co-Morbidity and Death
in HFpEF
bull COPD
bull Rheumatologic
Disorders
bull More dangerous in
HFpEF than HFrEF
AtherJACC2012
Total Hospitalizations I-Preserve
Many Co-Morbid Hospitalizations
CarsonJACC-HF in press
Mode of Death I-Preserve
60 cardiovascular death
40 non-cardiovascular death or unknown
May require a doubling of sample size for mortality component
Similar Issue with HFH
ZileCirc HF 2010
Mode of Non- Cardiovascular Death I-Preserve
ZileCirc HF 2010
Regional Differences in Therapy
OConnorJACC2012
Spironolactone
Placebo
HR = 089 (077 ndash 104)
p=0138
3511723 (204)
3201722 (186)
1degOutcome (CV Death HF Hosp or Resuscitated Cardiac Arrest)
Placebo Rates Primary Outcome by Region
US Canada
Argentina Brazil
Russia Rep Georgia
126 per 100 pt-
yr
23 per 100 pt-yr
Placebo
280881 (318)
Placebo
71842 (84)
HR=082 (069-098)
HR=110 (079-151)
Interaction p=0122
US Canada
Argentina Brazil
Russia Rep Georgia
Placebo
280881 (318)
Placebo
71842 (84)
Exploratory (post-hoc)Placebo vs Spiro
by Region
Of 22 pre-specified only 1 - Stratum - showed
a significant interaction with treatment
Enrolled
by Spiro Placebo
Hazard Ratio
(95 CI)
P-value
Natriuretic
peptide
78490
(159)
116491
(236)
065 (049-087)
0003
Heart Failure
Hosp
2421232
(196)
2351232
(191)
101 (084-121)
0923
P=0013 for interaction
Subgroups
ACEIARB Meta-Analysis Favors
Treatment(HR=088 p=005)
ShahJCF 2010
Hope for HFpEF
Duke Heart Center
LCZ696 Favorable on the most likely
Surrogate
bull Reduced NT-proBNP
bull Reduced LA size
bull Improved NYHA
Class
bull PARAGON
OUTCOME Trial
SolomonLancet2012
Have we failed in HFpEF
As clinical trialists we have sometimes failed(unknowingly) in the
design and conduct of HFpEF studies
Most traditional therapies(ACEIARBMRABB) have not failed(modest
reduction in recurrent HFH) but they have not passed regulatory
standards ndashWhat about guidelines
Must address the moderate number of Co- Morbidities differently
New therapies will have the advantage of our informed journey and will
likely yield positive results
HFpEF Trials
Half Empty Half Full hellip Yet No Banners
Duke Heart Center
Cluster Analysis of Heart Failure to Uncover
Distinct Phenotypes
Circulation 2014 Nov 14
Duke Heart Center
Objective Measures of Heart Failure
Duke Heart Center
Clinical Implications on Outcomes
Duke Heart Center
bull Caucasian
bull Female
bull NICM
bull Few Co-Morbidities
Clinical Implications
Some Clusters Respond to Exercise
Implications of Co-Morbidities
Increase heterogeneity
Complicates management(Beta agonistsNSAID)
Associated with worse outcomes
Increase in non-cardiac outcomes
Mentz RJ and Felker GM Heart Fail Clin 2013
Duke Heart Center
Risk of Co-Morbidity and Death
in HFpEF
bull COPD
bull Rheumatologic
Disorders
bull More dangerous in
HFpEF than HFrEF
AtherJACC2012
Total Hospitalizations I-Preserve
Many Co-Morbid Hospitalizations
CarsonJACC-HF in press
Mode of Death I-Preserve
60 cardiovascular death
40 non-cardiovascular death or unknown
May require a doubling of sample size for mortality component
Similar Issue with HFH
ZileCirc HF 2010
Mode of Non- Cardiovascular Death I-Preserve
ZileCirc HF 2010
Regional Differences in Therapy
OConnorJACC2012
Spironolactone
Placebo
HR = 089 (077 ndash 104)
p=0138
3511723 (204)
3201722 (186)
1degOutcome (CV Death HF Hosp or Resuscitated Cardiac Arrest)
Placebo Rates Primary Outcome by Region
US Canada
Argentina Brazil
Russia Rep Georgia
126 per 100 pt-
yr
23 per 100 pt-yr
Placebo
280881 (318)
Placebo
71842 (84)
HR=082 (069-098)
HR=110 (079-151)
Interaction p=0122
US Canada
Argentina Brazil
Russia Rep Georgia
Placebo
280881 (318)
Placebo
71842 (84)
Exploratory (post-hoc)Placebo vs Spiro
by Region
Of 22 pre-specified only 1 - Stratum - showed
a significant interaction with treatment
Enrolled
by Spiro Placebo
Hazard Ratio
(95 CI)
P-value
Natriuretic
peptide
78490
(159)
116491
(236)
065 (049-087)
0003
Heart Failure
Hosp
2421232
(196)
2351232
(191)
101 (084-121)
0923
P=0013 for interaction
Subgroups
ACEIARB Meta-Analysis Favors
Treatment(HR=088 p=005)
ShahJCF 2010
Hope for HFpEF
Duke Heart Center
LCZ696 Favorable on the most likely
Surrogate
bull Reduced NT-proBNP
bull Reduced LA size
bull Improved NYHA
Class
bull PARAGON
OUTCOME Trial
SolomonLancet2012
Have we failed in HFpEF
As clinical trialists we have sometimes failed(unknowingly) in the
design and conduct of HFpEF studies
Most traditional therapies(ACEIARBMRABB) have not failed(modest
reduction in recurrent HFH) but they have not passed regulatory
standards ndashWhat about guidelines
Must address the moderate number of Co- Morbidities differently
New therapies will have the advantage of our informed journey and will
likely yield positive results
HFpEF Trials
Half Empty Half Full hellip Yet No Banners
Duke Heart Center
Objective Measures of Heart Failure
Duke Heart Center
Clinical Implications on Outcomes
Duke Heart Center
bull Caucasian
bull Female
bull NICM
bull Few Co-Morbidities
Clinical Implications
Some Clusters Respond to Exercise
Implications of Co-Morbidities
Increase heterogeneity
Complicates management(Beta agonistsNSAID)
Associated with worse outcomes
Increase in non-cardiac outcomes
Mentz RJ and Felker GM Heart Fail Clin 2013
Duke Heart Center
Risk of Co-Morbidity and Death
in HFpEF
bull COPD
bull Rheumatologic
Disorders
bull More dangerous in
HFpEF than HFrEF
AtherJACC2012
Total Hospitalizations I-Preserve
Many Co-Morbid Hospitalizations
CarsonJACC-HF in press
Mode of Death I-Preserve
60 cardiovascular death
40 non-cardiovascular death or unknown
May require a doubling of sample size for mortality component
Similar Issue with HFH
ZileCirc HF 2010
Mode of Non- Cardiovascular Death I-Preserve
ZileCirc HF 2010
Regional Differences in Therapy
OConnorJACC2012
Spironolactone
Placebo
HR = 089 (077 ndash 104)
p=0138
3511723 (204)
3201722 (186)
1degOutcome (CV Death HF Hosp or Resuscitated Cardiac Arrest)
Placebo Rates Primary Outcome by Region
US Canada
Argentina Brazil
Russia Rep Georgia
126 per 100 pt-
yr
23 per 100 pt-yr
Placebo
280881 (318)
Placebo
71842 (84)
HR=082 (069-098)
HR=110 (079-151)
Interaction p=0122
US Canada
Argentina Brazil
Russia Rep Georgia
Placebo
280881 (318)
Placebo
71842 (84)
Exploratory (post-hoc)Placebo vs Spiro
by Region
Of 22 pre-specified only 1 - Stratum - showed
a significant interaction with treatment
Enrolled
by Spiro Placebo
Hazard Ratio
(95 CI)
P-value
Natriuretic
peptide
78490
(159)
116491
(236)
065 (049-087)
0003
Heart Failure
Hosp
2421232
(196)
2351232
(191)
101 (084-121)
0923
P=0013 for interaction
Subgroups
ACEIARB Meta-Analysis Favors
Treatment(HR=088 p=005)
ShahJCF 2010
Hope for HFpEF
Duke Heart Center
LCZ696 Favorable on the most likely
Surrogate
bull Reduced NT-proBNP
bull Reduced LA size
bull Improved NYHA
Class
bull PARAGON
OUTCOME Trial
SolomonLancet2012
Have we failed in HFpEF
As clinical trialists we have sometimes failed(unknowingly) in the
design and conduct of HFpEF studies
Most traditional therapies(ACEIARBMRABB) have not failed(modest
reduction in recurrent HFH) but they have not passed regulatory
standards ndashWhat about guidelines
Must address the moderate number of Co- Morbidities differently
New therapies will have the advantage of our informed journey and will
likely yield positive results
HFpEF Trials
Half Empty Half Full hellip Yet No Banners
Duke Heart Center
Clinical Implications on Outcomes
Duke Heart Center
bull Caucasian
bull Female
bull NICM
bull Few Co-Morbidities
Clinical Implications
Some Clusters Respond to Exercise
Implications of Co-Morbidities
Increase heterogeneity
Complicates management(Beta agonistsNSAID)
Associated with worse outcomes
Increase in non-cardiac outcomes
Mentz RJ and Felker GM Heart Fail Clin 2013
Duke Heart Center
Risk of Co-Morbidity and Death
in HFpEF
bull COPD
bull Rheumatologic
Disorders
bull More dangerous in
HFpEF than HFrEF
AtherJACC2012
Total Hospitalizations I-Preserve
Many Co-Morbid Hospitalizations
CarsonJACC-HF in press
Mode of Death I-Preserve
60 cardiovascular death
40 non-cardiovascular death or unknown
May require a doubling of sample size for mortality component
Similar Issue with HFH
ZileCirc HF 2010
Mode of Non- Cardiovascular Death I-Preserve
ZileCirc HF 2010
Regional Differences in Therapy
OConnorJACC2012
Spironolactone
Placebo
HR = 089 (077 ndash 104)
p=0138
3511723 (204)
3201722 (186)
1degOutcome (CV Death HF Hosp or Resuscitated Cardiac Arrest)
Placebo Rates Primary Outcome by Region
US Canada
Argentina Brazil
Russia Rep Georgia
126 per 100 pt-
yr
23 per 100 pt-yr
Placebo
280881 (318)
Placebo
71842 (84)
HR=082 (069-098)
HR=110 (079-151)
Interaction p=0122
US Canada
Argentina Brazil
Russia Rep Georgia
Placebo
280881 (318)
Placebo
71842 (84)
Exploratory (post-hoc)Placebo vs Spiro
by Region
Of 22 pre-specified only 1 - Stratum - showed
a significant interaction with treatment
Enrolled
by Spiro Placebo
Hazard Ratio
(95 CI)
P-value
Natriuretic
peptide
78490
(159)
116491
(236)
065 (049-087)
0003
Heart Failure
Hosp
2421232
(196)
2351232
(191)
101 (084-121)
0923
P=0013 for interaction
Subgroups
ACEIARB Meta-Analysis Favors
Treatment(HR=088 p=005)
ShahJCF 2010
Hope for HFpEF
Duke Heart Center
LCZ696 Favorable on the most likely
Surrogate
bull Reduced NT-proBNP
bull Reduced LA size
bull Improved NYHA
Class
bull PARAGON
OUTCOME Trial
SolomonLancet2012
Have we failed in HFpEF
As clinical trialists we have sometimes failed(unknowingly) in the
design and conduct of HFpEF studies
Most traditional therapies(ACEIARBMRABB) have not failed(modest
reduction in recurrent HFH) but they have not passed regulatory
standards ndashWhat about guidelines
Must address the moderate number of Co- Morbidities differently
New therapies will have the advantage of our informed journey and will
likely yield positive results
HFpEF Trials
Half Empty Half Full hellip Yet No Banners
Duke Heart Center
bull Caucasian
bull Female
bull NICM
bull Few Co-Morbidities
Clinical Implications
Some Clusters Respond to Exercise
Implications of Co-Morbidities
Increase heterogeneity
Complicates management(Beta agonistsNSAID)
Associated with worse outcomes
Increase in non-cardiac outcomes
Mentz RJ and Felker GM Heart Fail Clin 2013
Duke Heart Center
Risk of Co-Morbidity and Death
in HFpEF
bull COPD
bull Rheumatologic
Disorders
bull More dangerous in
HFpEF than HFrEF
AtherJACC2012
Total Hospitalizations I-Preserve
Many Co-Morbid Hospitalizations
CarsonJACC-HF in press
Mode of Death I-Preserve
60 cardiovascular death
40 non-cardiovascular death or unknown
May require a doubling of sample size for mortality component
Similar Issue with HFH
ZileCirc HF 2010
Mode of Non- Cardiovascular Death I-Preserve
ZileCirc HF 2010
Regional Differences in Therapy
OConnorJACC2012
Spironolactone
Placebo
HR = 089 (077 ndash 104)
p=0138
3511723 (204)
3201722 (186)
1degOutcome (CV Death HF Hosp or Resuscitated Cardiac Arrest)
Placebo Rates Primary Outcome by Region
US Canada
Argentina Brazil
Russia Rep Georgia
126 per 100 pt-
yr
23 per 100 pt-yr
Placebo
280881 (318)
Placebo
71842 (84)
HR=082 (069-098)
HR=110 (079-151)
Interaction p=0122
US Canada
Argentina Brazil
Russia Rep Georgia
Placebo
280881 (318)
Placebo
71842 (84)
Exploratory (post-hoc)Placebo vs Spiro
by Region
Of 22 pre-specified only 1 - Stratum - showed
a significant interaction with treatment
Enrolled
by Spiro Placebo
Hazard Ratio
(95 CI)
P-value
Natriuretic
peptide
78490
(159)
116491
(236)
065 (049-087)
0003
Heart Failure
Hosp
2421232
(196)
2351232
(191)
101 (084-121)
0923
P=0013 for interaction
Subgroups
ACEIARB Meta-Analysis Favors
Treatment(HR=088 p=005)
ShahJCF 2010
Hope for HFpEF
Duke Heart Center
LCZ696 Favorable on the most likely
Surrogate
bull Reduced NT-proBNP
bull Reduced LA size
bull Improved NYHA
Class
bull PARAGON
OUTCOME Trial
SolomonLancet2012
Have we failed in HFpEF
As clinical trialists we have sometimes failed(unknowingly) in the
design and conduct of HFpEF studies
Most traditional therapies(ACEIARBMRABB) have not failed(modest
reduction in recurrent HFH) but they have not passed regulatory
standards ndashWhat about guidelines
Must address the moderate number of Co- Morbidities differently
New therapies will have the advantage of our informed journey and will
likely yield positive results
HFpEF Trials
Half Empty Half Full hellip Yet No Banners
Implications of Co-Morbidities
Increase heterogeneity
Complicates management(Beta agonistsNSAID)
Associated with worse outcomes
Increase in non-cardiac outcomes
Mentz RJ and Felker GM Heart Fail Clin 2013
Duke Heart Center
Risk of Co-Morbidity and Death
in HFpEF
bull COPD
bull Rheumatologic
Disorders
bull More dangerous in
HFpEF than HFrEF
AtherJACC2012
Total Hospitalizations I-Preserve
Many Co-Morbid Hospitalizations
CarsonJACC-HF in press
Mode of Death I-Preserve
60 cardiovascular death
40 non-cardiovascular death or unknown
May require a doubling of sample size for mortality component
Similar Issue with HFH
ZileCirc HF 2010
Mode of Non- Cardiovascular Death I-Preserve
ZileCirc HF 2010
Regional Differences in Therapy
OConnorJACC2012
Spironolactone
Placebo
HR = 089 (077 ndash 104)
p=0138
3511723 (204)
3201722 (186)
1degOutcome (CV Death HF Hosp or Resuscitated Cardiac Arrest)
Placebo Rates Primary Outcome by Region
US Canada
Argentina Brazil
Russia Rep Georgia
126 per 100 pt-
yr
23 per 100 pt-yr
Placebo
280881 (318)
Placebo
71842 (84)
HR=082 (069-098)
HR=110 (079-151)
Interaction p=0122
US Canada
Argentina Brazil
Russia Rep Georgia
Placebo
280881 (318)
Placebo
71842 (84)
Exploratory (post-hoc)Placebo vs Spiro
by Region
Of 22 pre-specified only 1 - Stratum - showed
a significant interaction with treatment
Enrolled
by Spiro Placebo
Hazard Ratio
(95 CI)
P-value
Natriuretic
peptide
78490
(159)
116491
(236)
065 (049-087)
0003
Heart Failure
Hosp
2421232
(196)
2351232
(191)
101 (084-121)
0923
P=0013 for interaction
Subgroups
ACEIARB Meta-Analysis Favors
Treatment(HR=088 p=005)
ShahJCF 2010
Hope for HFpEF
Duke Heart Center
LCZ696 Favorable on the most likely
Surrogate
bull Reduced NT-proBNP
bull Reduced LA size
bull Improved NYHA
Class
bull PARAGON
OUTCOME Trial
SolomonLancet2012
Have we failed in HFpEF
As clinical trialists we have sometimes failed(unknowingly) in the
design and conduct of HFpEF studies
Most traditional therapies(ACEIARBMRABB) have not failed(modest
reduction in recurrent HFH) but they have not passed regulatory
standards ndashWhat about guidelines
Must address the moderate number of Co- Morbidities differently
New therapies will have the advantage of our informed journey and will
likely yield positive results
HFpEF Trials
Half Empty Half Full hellip Yet No Banners
Duke Heart Center
Risk of Co-Morbidity and Death
in HFpEF
bull COPD
bull Rheumatologic
Disorders
bull More dangerous in
HFpEF than HFrEF
AtherJACC2012
Total Hospitalizations I-Preserve
Many Co-Morbid Hospitalizations
CarsonJACC-HF in press
Mode of Death I-Preserve
60 cardiovascular death
40 non-cardiovascular death or unknown
May require a doubling of sample size for mortality component
Similar Issue with HFH
ZileCirc HF 2010
Mode of Non- Cardiovascular Death I-Preserve
ZileCirc HF 2010
Regional Differences in Therapy
OConnorJACC2012
Spironolactone
Placebo
HR = 089 (077 ndash 104)
p=0138
3511723 (204)
3201722 (186)
1degOutcome (CV Death HF Hosp or Resuscitated Cardiac Arrest)
Placebo Rates Primary Outcome by Region
US Canada
Argentina Brazil
Russia Rep Georgia
126 per 100 pt-
yr
23 per 100 pt-yr
Placebo
280881 (318)
Placebo
71842 (84)
HR=082 (069-098)
HR=110 (079-151)
Interaction p=0122
US Canada
Argentina Brazil
Russia Rep Georgia
Placebo
280881 (318)
Placebo
71842 (84)
Exploratory (post-hoc)Placebo vs Spiro
by Region
Of 22 pre-specified only 1 - Stratum - showed
a significant interaction with treatment
Enrolled
by Spiro Placebo
Hazard Ratio
(95 CI)
P-value
Natriuretic
peptide
78490
(159)
116491
(236)
065 (049-087)
0003
Heart Failure
Hosp
2421232
(196)
2351232
(191)
101 (084-121)
0923
P=0013 for interaction
Subgroups
ACEIARB Meta-Analysis Favors
Treatment(HR=088 p=005)
ShahJCF 2010
Hope for HFpEF
Duke Heart Center
LCZ696 Favorable on the most likely
Surrogate
bull Reduced NT-proBNP
bull Reduced LA size
bull Improved NYHA
Class
bull PARAGON
OUTCOME Trial
SolomonLancet2012
Have we failed in HFpEF
As clinical trialists we have sometimes failed(unknowingly) in the
design and conduct of HFpEF studies
Most traditional therapies(ACEIARBMRABB) have not failed(modest
reduction in recurrent HFH) but they have not passed regulatory
standards ndashWhat about guidelines
Must address the moderate number of Co- Morbidities differently
New therapies will have the advantage of our informed journey and will
likely yield positive results
HFpEF Trials
Half Empty Half Full hellip Yet No Banners
Total Hospitalizations I-Preserve
Many Co-Morbid Hospitalizations
CarsonJACC-HF in press
Mode of Death I-Preserve
60 cardiovascular death
40 non-cardiovascular death or unknown
May require a doubling of sample size for mortality component
Similar Issue with HFH
ZileCirc HF 2010
Mode of Non- Cardiovascular Death I-Preserve
ZileCirc HF 2010
Regional Differences in Therapy
OConnorJACC2012
Spironolactone
Placebo
HR = 089 (077 ndash 104)
p=0138
3511723 (204)
3201722 (186)
1degOutcome (CV Death HF Hosp or Resuscitated Cardiac Arrest)
Placebo Rates Primary Outcome by Region
US Canada
Argentina Brazil
Russia Rep Georgia
126 per 100 pt-
yr
23 per 100 pt-yr
Placebo
280881 (318)
Placebo
71842 (84)
HR=082 (069-098)
HR=110 (079-151)
Interaction p=0122
US Canada
Argentina Brazil
Russia Rep Georgia
Placebo
280881 (318)
Placebo
71842 (84)
Exploratory (post-hoc)Placebo vs Spiro
by Region
Of 22 pre-specified only 1 - Stratum - showed
a significant interaction with treatment
Enrolled
by Spiro Placebo
Hazard Ratio
(95 CI)
P-value
Natriuretic
peptide
78490
(159)
116491
(236)
065 (049-087)
0003
Heart Failure
Hosp
2421232
(196)
2351232
(191)
101 (084-121)
0923
P=0013 for interaction
Subgroups
ACEIARB Meta-Analysis Favors
Treatment(HR=088 p=005)
ShahJCF 2010
Hope for HFpEF
Duke Heart Center
LCZ696 Favorable on the most likely
Surrogate
bull Reduced NT-proBNP
bull Reduced LA size
bull Improved NYHA
Class
bull PARAGON
OUTCOME Trial
SolomonLancet2012
Have we failed in HFpEF
As clinical trialists we have sometimes failed(unknowingly) in the
design and conduct of HFpEF studies
Most traditional therapies(ACEIARBMRABB) have not failed(modest
reduction in recurrent HFH) but they have not passed regulatory
standards ndashWhat about guidelines
Must address the moderate number of Co- Morbidities differently
New therapies will have the advantage of our informed journey and will
likely yield positive results
HFpEF Trials
Half Empty Half Full hellip Yet No Banners
Mode of Death I-Preserve
60 cardiovascular death
40 non-cardiovascular death or unknown
May require a doubling of sample size for mortality component
Similar Issue with HFH
ZileCirc HF 2010
Mode of Non- Cardiovascular Death I-Preserve
ZileCirc HF 2010
Regional Differences in Therapy
OConnorJACC2012
Spironolactone
Placebo
HR = 089 (077 ndash 104)
p=0138
3511723 (204)
3201722 (186)
1degOutcome (CV Death HF Hosp or Resuscitated Cardiac Arrest)
Placebo Rates Primary Outcome by Region
US Canada
Argentina Brazil
Russia Rep Georgia
126 per 100 pt-
yr
23 per 100 pt-yr
Placebo
280881 (318)
Placebo
71842 (84)
HR=082 (069-098)
HR=110 (079-151)
Interaction p=0122
US Canada
Argentina Brazil
Russia Rep Georgia
Placebo
280881 (318)
Placebo
71842 (84)
Exploratory (post-hoc)Placebo vs Spiro
by Region
Of 22 pre-specified only 1 - Stratum - showed
a significant interaction with treatment
Enrolled
by Spiro Placebo
Hazard Ratio
(95 CI)
P-value
Natriuretic
peptide
78490
(159)
116491
(236)
065 (049-087)
0003
Heart Failure
Hosp
2421232
(196)
2351232
(191)
101 (084-121)
0923
P=0013 for interaction
Subgroups
ACEIARB Meta-Analysis Favors
Treatment(HR=088 p=005)
ShahJCF 2010
Hope for HFpEF
Duke Heart Center
LCZ696 Favorable on the most likely
Surrogate
bull Reduced NT-proBNP
bull Reduced LA size
bull Improved NYHA
Class
bull PARAGON
OUTCOME Trial
SolomonLancet2012
Have we failed in HFpEF
As clinical trialists we have sometimes failed(unknowingly) in the
design and conduct of HFpEF studies
Most traditional therapies(ACEIARBMRABB) have not failed(modest
reduction in recurrent HFH) but they have not passed regulatory
standards ndashWhat about guidelines
Must address the moderate number of Co- Morbidities differently
New therapies will have the advantage of our informed journey and will
likely yield positive results
HFpEF Trials
Half Empty Half Full hellip Yet No Banners
Mode of Non- Cardiovascular Death I-Preserve
ZileCirc HF 2010
Regional Differences in Therapy
OConnorJACC2012
Spironolactone
Placebo
HR = 089 (077 ndash 104)
p=0138
3511723 (204)
3201722 (186)
1degOutcome (CV Death HF Hosp or Resuscitated Cardiac Arrest)
Placebo Rates Primary Outcome by Region
US Canada
Argentina Brazil
Russia Rep Georgia
126 per 100 pt-
yr
23 per 100 pt-yr
Placebo
280881 (318)
Placebo
71842 (84)
HR=082 (069-098)
HR=110 (079-151)
Interaction p=0122
US Canada
Argentina Brazil
Russia Rep Georgia
Placebo
280881 (318)
Placebo
71842 (84)
Exploratory (post-hoc)Placebo vs Spiro
by Region
Of 22 pre-specified only 1 - Stratum - showed
a significant interaction with treatment
Enrolled
by Spiro Placebo
Hazard Ratio
(95 CI)
P-value
Natriuretic
peptide
78490
(159)
116491
(236)
065 (049-087)
0003
Heart Failure
Hosp
2421232
(196)
2351232
(191)
101 (084-121)
0923
P=0013 for interaction
Subgroups
ACEIARB Meta-Analysis Favors
Treatment(HR=088 p=005)
ShahJCF 2010
Hope for HFpEF
Duke Heart Center
LCZ696 Favorable on the most likely
Surrogate
bull Reduced NT-proBNP
bull Reduced LA size
bull Improved NYHA
Class
bull PARAGON
OUTCOME Trial
SolomonLancet2012
Have we failed in HFpEF
As clinical trialists we have sometimes failed(unknowingly) in the
design and conduct of HFpEF studies
Most traditional therapies(ACEIARBMRABB) have not failed(modest
reduction in recurrent HFH) but they have not passed regulatory
standards ndashWhat about guidelines
Must address the moderate number of Co- Morbidities differently
New therapies will have the advantage of our informed journey and will
likely yield positive results
HFpEF Trials
Half Empty Half Full hellip Yet No Banners
Regional Differences in Therapy
OConnorJACC2012
Spironolactone
Placebo
HR = 089 (077 ndash 104)
p=0138
3511723 (204)
3201722 (186)
1degOutcome (CV Death HF Hosp or Resuscitated Cardiac Arrest)
Placebo Rates Primary Outcome by Region
US Canada
Argentina Brazil
Russia Rep Georgia
126 per 100 pt-
yr
23 per 100 pt-yr
Placebo
280881 (318)
Placebo
71842 (84)
HR=082 (069-098)
HR=110 (079-151)
Interaction p=0122
US Canada
Argentina Brazil
Russia Rep Georgia
Placebo
280881 (318)
Placebo
71842 (84)
Exploratory (post-hoc)Placebo vs Spiro
by Region
Of 22 pre-specified only 1 - Stratum - showed
a significant interaction with treatment
Enrolled
by Spiro Placebo
Hazard Ratio
(95 CI)
P-value
Natriuretic
peptide
78490
(159)
116491
(236)
065 (049-087)
0003
Heart Failure
Hosp
2421232
(196)
2351232
(191)
101 (084-121)
0923
P=0013 for interaction
Subgroups
ACEIARB Meta-Analysis Favors
Treatment(HR=088 p=005)
ShahJCF 2010
Hope for HFpEF
Duke Heart Center
LCZ696 Favorable on the most likely
Surrogate
bull Reduced NT-proBNP
bull Reduced LA size
bull Improved NYHA
Class
bull PARAGON
OUTCOME Trial
SolomonLancet2012
Have we failed in HFpEF
As clinical trialists we have sometimes failed(unknowingly) in the
design and conduct of HFpEF studies
Most traditional therapies(ACEIARBMRABB) have not failed(modest
reduction in recurrent HFH) but they have not passed regulatory
standards ndashWhat about guidelines
Must address the moderate number of Co- Morbidities differently
New therapies will have the advantage of our informed journey and will
likely yield positive results
HFpEF Trials
Half Empty Half Full hellip Yet No Banners
Spironolactone
Placebo
HR = 089 (077 ndash 104)
p=0138
3511723 (204)
3201722 (186)
1degOutcome (CV Death HF Hosp or Resuscitated Cardiac Arrest)
Placebo Rates Primary Outcome by Region
US Canada
Argentina Brazil
Russia Rep Georgia
126 per 100 pt-
yr
23 per 100 pt-yr
Placebo
280881 (318)
Placebo
71842 (84)
HR=082 (069-098)
HR=110 (079-151)
Interaction p=0122
US Canada
Argentina Brazil
Russia Rep Georgia
Placebo
280881 (318)
Placebo
71842 (84)
Exploratory (post-hoc)Placebo vs Spiro
by Region
Of 22 pre-specified only 1 - Stratum - showed
a significant interaction with treatment
Enrolled
by Spiro Placebo
Hazard Ratio
(95 CI)
P-value
Natriuretic
peptide
78490
(159)
116491
(236)
065 (049-087)
0003
Heart Failure
Hosp
2421232
(196)
2351232
(191)
101 (084-121)
0923
P=0013 for interaction
Subgroups
ACEIARB Meta-Analysis Favors
Treatment(HR=088 p=005)
ShahJCF 2010
Hope for HFpEF
Duke Heart Center
LCZ696 Favorable on the most likely
Surrogate
bull Reduced NT-proBNP
bull Reduced LA size
bull Improved NYHA
Class
bull PARAGON
OUTCOME Trial
SolomonLancet2012
Have we failed in HFpEF
As clinical trialists we have sometimes failed(unknowingly) in the
design and conduct of HFpEF studies
Most traditional therapies(ACEIARBMRABB) have not failed(modest
reduction in recurrent HFH) but they have not passed regulatory
standards ndashWhat about guidelines
Must address the moderate number of Co- Morbidities differently
New therapies will have the advantage of our informed journey and will
likely yield positive results
HFpEF Trials
Half Empty Half Full hellip Yet No Banners
Placebo Rates Primary Outcome by Region
US Canada
Argentina Brazil
Russia Rep Georgia
126 per 100 pt-
yr
23 per 100 pt-yr
Placebo
280881 (318)
Placebo
71842 (84)
HR=082 (069-098)
HR=110 (079-151)
Interaction p=0122
US Canada
Argentina Brazil
Russia Rep Georgia
Placebo
280881 (318)
Placebo
71842 (84)
Exploratory (post-hoc)Placebo vs Spiro
by Region
Of 22 pre-specified only 1 - Stratum - showed
a significant interaction with treatment
Enrolled
by Spiro Placebo
Hazard Ratio
(95 CI)
P-value
Natriuretic
peptide
78490
(159)
116491
(236)
065 (049-087)
0003
Heart Failure
Hosp
2421232
(196)
2351232
(191)
101 (084-121)
0923
P=0013 for interaction
Subgroups
ACEIARB Meta-Analysis Favors
Treatment(HR=088 p=005)
ShahJCF 2010
Hope for HFpEF
Duke Heart Center
LCZ696 Favorable on the most likely
Surrogate
bull Reduced NT-proBNP
bull Reduced LA size
bull Improved NYHA
Class
bull PARAGON
OUTCOME Trial
SolomonLancet2012
Have we failed in HFpEF
As clinical trialists we have sometimes failed(unknowingly) in the
design and conduct of HFpEF studies
Most traditional therapies(ACEIARBMRABB) have not failed(modest
reduction in recurrent HFH) but they have not passed regulatory
standards ndashWhat about guidelines
Must address the moderate number of Co- Morbidities differently
New therapies will have the advantage of our informed journey and will
likely yield positive results
HFpEF Trials
Half Empty Half Full hellip Yet No Banners
HR=082 (069-098)
HR=110 (079-151)
Interaction p=0122
US Canada
Argentina Brazil
Russia Rep Georgia
Placebo
280881 (318)
Placebo
71842 (84)
Exploratory (post-hoc)Placebo vs Spiro
by Region
Of 22 pre-specified only 1 - Stratum - showed
a significant interaction with treatment
Enrolled
by Spiro Placebo
Hazard Ratio
(95 CI)
P-value
Natriuretic
peptide
78490
(159)
116491
(236)
065 (049-087)
0003
Heart Failure
Hosp
2421232
(196)
2351232
(191)
101 (084-121)
0923
P=0013 for interaction
Subgroups
ACEIARB Meta-Analysis Favors
Treatment(HR=088 p=005)
ShahJCF 2010
Hope for HFpEF
Duke Heart Center
LCZ696 Favorable on the most likely
Surrogate
bull Reduced NT-proBNP
bull Reduced LA size
bull Improved NYHA
Class
bull PARAGON
OUTCOME Trial
SolomonLancet2012
Have we failed in HFpEF
As clinical trialists we have sometimes failed(unknowingly) in the
design and conduct of HFpEF studies
Most traditional therapies(ACEIARBMRABB) have not failed(modest
reduction in recurrent HFH) but they have not passed regulatory
standards ndashWhat about guidelines
Must address the moderate number of Co- Morbidities differently
New therapies will have the advantage of our informed journey and will
likely yield positive results
HFpEF Trials
Half Empty Half Full hellip Yet No Banners
Of 22 pre-specified only 1 - Stratum - showed
a significant interaction with treatment
Enrolled
by Spiro Placebo
Hazard Ratio
(95 CI)
P-value
Natriuretic
peptide
78490
(159)
116491
(236)
065 (049-087)
0003
Heart Failure
Hosp
2421232
(196)
2351232
(191)
101 (084-121)
0923
P=0013 for interaction
Subgroups
ACEIARB Meta-Analysis Favors
Treatment(HR=088 p=005)
ShahJCF 2010
Hope for HFpEF
Duke Heart Center
LCZ696 Favorable on the most likely
Surrogate
bull Reduced NT-proBNP
bull Reduced LA size
bull Improved NYHA
Class
bull PARAGON
OUTCOME Trial
SolomonLancet2012
Have we failed in HFpEF
As clinical trialists we have sometimes failed(unknowingly) in the
design and conduct of HFpEF studies
Most traditional therapies(ACEIARBMRABB) have not failed(modest
reduction in recurrent HFH) but they have not passed regulatory
standards ndashWhat about guidelines
Must address the moderate number of Co- Morbidities differently
New therapies will have the advantage of our informed journey and will
likely yield positive results
HFpEF Trials
Half Empty Half Full hellip Yet No Banners
ACEIARB Meta-Analysis Favors
Treatment(HR=088 p=005)
ShahJCF 2010
Hope for HFpEF
Duke Heart Center
LCZ696 Favorable on the most likely
Surrogate
bull Reduced NT-proBNP
bull Reduced LA size
bull Improved NYHA
Class
bull PARAGON
OUTCOME Trial
SolomonLancet2012
Have we failed in HFpEF
As clinical trialists we have sometimes failed(unknowingly) in the
design and conduct of HFpEF studies
Most traditional therapies(ACEIARBMRABB) have not failed(modest
reduction in recurrent HFH) but they have not passed regulatory
standards ndashWhat about guidelines
Must address the moderate number of Co- Morbidities differently
New therapies will have the advantage of our informed journey and will
likely yield positive results
HFpEF Trials
Half Empty Half Full hellip Yet No Banners
Hope for HFpEF
Duke Heart Center
LCZ696 Favorable on the most likely
Surrogate
bull Reduced NT-proBNP
bull Reduced LA size
bull Improved NYHA
Class
bull PARAGON
OUTCOME Trial
SolomonLancet2012
Have we failed in HFpEF
As clinical trialists we have sometimes failed(unknowingly) in the
design and conduct of HFpEF studies
Most traditional therapies(ACEIARBMRABB) have not failed(modest
reduction in recurrent HFH) but they have not passed regulatory
standards ndashWhat about guidelines
Must address the moderate number of Co- Morbidities differently
New therapies will have the advantage of our informed journey and will
likely yield positive results
HFpEF Trials
Half Empty Half Full hellip Yet No Banners
Duke Heart Center
LCZ696 Favorable on the most likely
Surrogate
bull Reduced NT-proBNP
bull Reduced LA size
bull Improved NYHA
Class
bull PARAGON
OUTCOME Trial
SolomonLancet2012
Have we failed in HFpEF
As clinical trialists we have sometimes failed(unknowingly) in the
design and conduct of HFpEF studies
Most traditional therapies(ACEIARBMRABB) have not failed(modest
reduction in recurrent HFH) but they have not passed regulatory
standards ndashWhat about guidelines
Must address the moderate number of Co- Morbidities differently
New therapies will have the advantage of our informed journey and will
likely yield positive results
HFpEF Trials
Half Empty Half Full hellip Yet No Banners
Have we failed in HFpEF
As clinical trialists we have sometimes failed(unknowingly) in the
design and conduct of HFpEF studies
Most traditional therapies(ACEIARBMRABB) have not failed(modest
reduction in recurrent HFH) but they have not passed regulatory
standards ndashWhat about guidelines
Must address the moderate number of Co- Morbidities differently
New therapies will have the advantage of our informed journey and will
likely yield positive results
HFpEF Trials
Half Empty Half Full hellip Yet No Banners
HFpEF Trials
Half Empty Half Full hellip Yet No Banners