1. Effect of Phosphodiesterase-5 Inhibition on Exercise
Capacity and Clinical Status in Heart Failure with
Preserved Ejection Fraction (RELAX):
A Randomized Clinical Trial
Margaret M Redfield, MD
on behalf of the
NHLBI Heart Failure Clinical Research Network

2. Background ● Phosphodiesterase type-5 (PDE-5) metabolizes nitric
oxide (NO) and natriuretic peptide (NP) generated
cGMP
● If PDE-5 is activated in HF; may limit beneficial NO
and NP actions in the heart, vasculature and kidney
● PDE-5 Inhibitor therapy approved for
• Erectile dysfunction
• Group I pulmonary arterial hypertension (PAH)
● Role in heart failure (HF) with reduced (HFrEF) or
preserved (HFpEF) ejection fraction unclear

3. Background ● Experimental HF: PDE-5 inhibition
• Reversed cardiac remodeling and dysfunction
• Improved vascular and renal function
● Small Clinical Studies: PDE-5 inhibition (sildenafil)
• HFrEF
Improved maximal exercise capacity
• HFpEF + PAH + RV dysfunction
Improved hemodynamics, lung function, RV function
and LV remodeling

4. Hypothesis In comparison to placebo, chronic (24 weeks)
therapy with the PDE-5 inhibitor sildenafil will
improve exercise capacity (peak VO2) and
clinical status in HFpEF.

5. Study population ≥ 400 pg/ml or
● NYHA class II-IV HF symptoms
● EF ≥ 50%
● Objective evidence of HF (at least one)
• HF hospitalization or ED visit + iv diuretic
• Elevated PCWP at catheterization for dyspnea
• Left atrial enlargement + chronic diuretic for HF
● At study entry (both)
• Peak VO2 < 60% age/sex nl value + RER ≥ 1.0
• NT-proBNP
≥ 400 pg/ml or
< 400 pg/ml with documented ↑ PCWP ≤ two weeks
of NT-proBNP < 400 

6. Study Design Placebo 20 mg TID Sildenafil 20 mg TID
Baseline CPXT, 6MWT, Echo-Doppler, MLHFQ,
Biomarkers, and CMR (sinus rhythm)
Double-blind; 1 to 1 randomization stratified by site and rhythm (AF)
Placebo 20 mg TID
Sildenafil 20 mg TID
12 week CPXT, 6MWT, MLHFQ
Placebo 60 mg TID
Sildenafil 60 mg TID
24 week CPXT, 6MWT, Echo-Doppler, MLHFQ,
Biomarkers and CMR

7. Study Endpoints ● Secondary Endpoints ● Other Endpoints
● Primary Endpoint
• Change in peak VO2 after 24 weeks of therapy
● Secondary Endpoints
• Change in 6MWD after 24 weeks of therapy
• Hierarchical composite clinical rank score
● Other Endpoints
• Change in CV structure and function (24 weeks)
Echo-Doppler
Cardiac magnetic resonance imaging (CMR)
• Change in biomarkers (24 weeks)

8. Hierarchical composite clinical rank score
At 24 weeks, all patients ranked 1 X
FIRST Death
LAST Death

9. Hierarchical composite clinical rank score
At 24 weeks, all patients ranked 1 X
X+1 Y
FIRST Death
LAST Death
Alive with FIRST CV or cardiorenal hsp
Alive with LAST CV or cardiorenal hsp

10. Hierarchical composite clinical rank score
At 24 weeks, all patients ranked 1 X
X+1 Y
Y+1 Z
FIRST Death
LAST Death
Alive with FIRST CV or renal hospitalization
Alive with LAST CV or renal hospitalization
LEAST favorable change in MLHFQ
MOST favorable change in MLHFQ
Mean rank score (lower worse) compared between treatment groups
Anchor value (no treatment effect) = Z / 2

11. HF hospitalization in past year Hx hypertension
Baseline Features
Characteristic
Age (years)
Female
White race
BMI (kg/m2)
NYHA class II/III
HF hospitalization in past year
Hx hypertension
Hx of coronary artery disease
Diabetes
Hx of atrial fibrillation
Median values or % shown
Placebo
(N = 103) 69
53%
92% 33
45% / 55%
39%
90%
36%
44%
50%
Sildenafil
(N = 113) 68
43%
90% 33
49% / 51%
35%
80%*
42%
52%
*p-value < 0.05

12. Peak VO2 (ml/kg/min) (% predicted) Chronotropic incompetence present
Baseline Features
Characteristic
Ejection fraction (%)
NT-proBNP (pg/ml)
Peak VO2 (ml/kg/min) (% predicted)
Chronotropic incompetence present
6MWD (m) (% predicted)
Cardiac index (L/min/m2) - (normal > 2.5)
Relative Wall Thickness ≥ 0.42
E/e’ - (normal ≤ 8)
LA volume index (ml/m2) - (normal < 29)
PASP (mmHg) - (normal < 30)
Median values or % shown
Placebo
(N = 103) 60
648
11.9 (41%)
78%
305 (68%)
2.48
44% 17 43 41
Sildenafil
(N = 113) 60
757
11.7 (41%)
76%
308 (70%)
2.47
48% 15 44 41
All p > 0.05

13. Results:

14. Results: Primary Endpoint
Change in Peak VO2
1.5
1.0
0.5
0.0
-0.5
-1.0
-1.5
ml/kg/min
Placebo
n = 94
(91%)
Sildenafil
n = 91
(80%)
Withdrew consent (n=14), death (n=3), unwilling (n=3) or unable (n=9)
to complete CPXT, inadequate peak VO2 data (n=2)

15. Results: Primary Endpoint
Change in Peak VO2
p = 0.90
1.5
1.0
0.5
0.0
-0.5
-1.0
-1.5
ml/kg/min
Placebo
n = 94
(91%)
Sildenafil
n = 91
(80%)
Sensitivity analyses for missing data
Multiple imputation: p = 0.98; LOCF: p = 0.98
Data are median and IQR

16. Results: Secondary Endpoints
Change in 6MWD
p = 0.92 60 40 20
-20
-40
meters
Placebo
n = 95
Sildenafil
n = 90
Data are median and IQR

17. Results: Secondary Endpoints
Mean Clinical Rank Score
105
100
p = 0.85
Anchor Value 95 90 85
Placebo
n = 94
Sildenafil
n=95

18. Results: Safety Sildenafil 0% 3% 13% 13% 76% 80% 16% 22% 8% 16%
Characteristic
Death (%)
CV or cardiorenal hospitalization (%)
Adverse events (%)
Serious adverse events (%)
Withdrew or Unwilling or Unable to
complete 24 week CPXT
Placebo 0%
13%
76%
16% 8%
Sildenafil 3%
13%
80%
22%
16%
All p > 0.05

19. Results: Other endpoints
Placebo
0.6
-1.6 -2
0.01
-23
-0.01
Sildenafil
-1.5
0.2 2
0.05*
15*
0.38*
0.30*
*p-value < 0.05
Characteristic
Change in LV mass by CMR (g)
Change in E/e’
Change in PASP (mmHg)
Change in creatinine (mg/dl)
Change in cystatin C (mg/L)
Change in NT-proBNP (pg/ml)
Change in endothelin-1 (pg/ml)
Change in uric acid (mg/dl)
Median values shown

20. Conclusions ● Chronic therapy with the PDE-5 inhibitor sildenafil
was not associated with clinical benefit in HFpEF
● Continued efforts to identify key pathophysiologic
perturbations and novel therapeutic targets in HFpEF
are needed

21. Montreal Heart Institute BIOMARKER CORE University of Vermont
Mayo Clinic
ECHO CORE
DCRI
BIOMARKER CORE
University of Vermont
Harvard University
CPXT CORE
NHLBI
CMR CORE
Duke University
Minnesota Heart
Failure Consortium
University of Utah
Morehouse School
of Medicine
Baylor College
of Medicine

22. JAMAARTICLE FIGURE