relax: inhibidor de la fosfodiesterasa-5. insuficiencia cardiaca con función ventricular preservada

7/28/2019 RELAX: Inhibidor de la fosfodiesterasa-5. Insuficiencia cardiaca con funcin ventricular preservada

1/10

ORIGINAL CONTRIBUTION

ONLINE FIRST

Effect of Phosphodiesterase-5 Inhibitionon Exercise Capacity and Clinical Statusin Heart Failure With Preserved Ejection FractionA Randomized Clinical Trial

Margaret M. Redfield, MD

Horng H. Chen, MD

Barry A. Borlaug, MD

Marc J. Semigran, MDKerry L. Lee, PhD

Gregory Lewis, MD

Martin M. LeWinter, MD

Jean L. Rouleau, MD

David A. Bull, MD

Douglas L. Mann, MD

Anita Deswal, MD

Lynne W. Stevenson, MD

Michael M. Givertz, MD

Elizabeth O. Ofili, MD

Christopher M. OConnor, MDG. Michael Felker, MD

Steven R. Goldsmith, MD

Bradley A. Bart, MD

Steven E. McNulty, MS

Jenny C. Ibarra, MSN

Grace Lin, MD

Jae K. Oh, MD

Manesh R. Patel, MD

Raymond J. Kim, MD

Russell P. Tracy, PhD

Eric J. Velazquez, MD

Kevin J. Anstrom, PhD

Adrian F. Hernandez, MD

Alice M. Mascette, MD

Eugene Braunwald, MD

for the RELAX TrialAuthor Affiliations are listed at the end of thisarticle.Corresponding Author: Margaret M. Redfield, MD,

Division of Cardiovascular Diseases, MayoClinic, 200First St SW, Guggenheim 9, Rochester, MN 55905([email protected]).

Importance Studies in experimental and human heart failure suggest that phos-phodiesterase-5 inhibitors may enhance cardiovascular function and thus exercise ca-pacity in heart failure with preserved ejection fraction (HFPEF).

Objective To determine the effect of the phosphodiesterase-5 inhibitor sildenafil com-pared with placebo on exercise capacity and clinical status in HFPEF.

Design Multicenter,double-blind, placebo-controlled, parallel-group, randomized clini-cal trial of 216 stable outpatients with HF, ejection fraction 50%, elevated N-terminalbrain-typenatriureticpeptideor elevatedinvasivelymeasured fillingpressures,and reducedexercise capacity. ParticipantswererandomizedfromOctober2008throughFebruary2012at 26 centers in North America. Follow-up was through August 30, 2012.

Interventions Sildenafil (n=113)or placebo (n=103)administered orally at 20 mg,3 times daily for 12 weeks, followed by 60 mg, 3 times daily for 12 weeks.

Main Outcome Measures Primary end point was change in peak oxygen consump-tion after 24 weeks of therapy. Secondary end points included change in 6-minute walkdistance and a hierarchical composite clinical status score (range, 1-n, a higher value in-dicates better status; expected value with no treatmenteffect, 95)basedon time to death,time to cardiovascular or cardiorenal hospitalization, and change in quality of life for par-ticipants without cardiovascular or cardiorenal hospitalization at 24 weeks.

Results Median age was 69 years, and 48% of patients were women. At baseline,median peak oxygen consumption (11.7 mL/kg/min) and 6-minute walk distance (308m) were reduced. The median E/e' (16), left atrial volume index (44 mL/m2), and pul-monary artery systolic pressure (41 mm Hg) were consistent with chronically elevatedleft ventricular filling pressures. At 24 weeks, median (IQR) changes in peak oxygen con-sumption (mL/kg/min) in patients who received placebo (0.20 [IQR,0.70 to 1.00])or sildenafil (0.20 [IQR, 1.70 to 1.11]) were not significantly different (P=.90) inanalyses in which patients with missing week-24 data were excluded, and in sensitivityanalysis based on intention to treat with multiple imputation for missing values (meanbetween-group difference, 0.01 mL/kg/min, [95% CI,0.60 to 0.61]). Themean clini-cal status rank score was not significantly different at 24 weeks between placebo (95.8)and sildenafil (94.2) (P=.85). Changes in 6-minute walk distance at 24 weeks in pa-tients who received placebo (15.0 m [IQR, 26.0 to 45.0]) or sildenafil (5.0 m [IQR,37.0 to 55.0]; P=.92) were also not significantly different. Adverse events occurred in78 placebo patients (76%) and 90 sildenafil patients (80%). Serious adverse events oc-curred in 16 placebo patients (16%) and 25 sildenafil patients (22%).

Conclusion and Relevance Among patients with HFPEF, phosphodiesterase-5 in-hibition with administration of sildenafil for 24 weeks, compared with placebo, didnot result in significant improvement in exercise capacity or clinical status.

Trial Registration clinicaltrials.gov Identifier: NCT00763867

JAMA. 2013;309(12):doi:10.1001/jama.2013.2024 www.jama.com

2013 American Medical Association. All rights reserved. JAMA, Published online March 11, 2013 E1

wnloaded From: http://jama.jamanetwork.com/ on 03/11/2013


2/10

HEART FAILURE WITH PRE-served ejection fraction(HFPEF) or diastolic heartfailure is a common and

highly morbidcondition.1 Clinicaltrialsof renin-angiotensin system antago-

nists have not demonstrated improve-ment in outcomes or clinical status inHFPEF, and effective therapies areneeded.2 Phosphodiesterase-5 (PDE-5)metabolizes the nitric oxide and natri-uretic peptide systems second messen-ger cyclicguanosinemonophosphateandthusmay limitbeneficial nitricoxideandnatriuretic peptide actions in the heart,vasculature, and kidneys.

Preclinical studies suggest thatinhibition of PDE-5 reverses adversecardiacstructural andfunctional remod-eling and enhances vascular, neuroen-

docrine, and renal function.3 In clini-cal studies, PDE-5 inhibitor therapyimproved exercise tolerance and clini-calstatusinpatientswithidiopathicpul-monary arterial hypertension and inpatients with heart failure andreducedejection fraction.4-7 A small, single-center study of HFPEF observedimprovedhemodynamics, leftventricu-lar diastolic function, right ventricularsystolic function,leftventricularhyper-trophy, and lung function with 6 to 12months of therapywith a PDE-5 inhibi-

tor compared with placebo.8

In aggre-gate, these studies suggest the poten-tial for PDE-5 inhibition to ameliorateseveral key pathophysiological pertur-bations in HFPEF, and thus improveexercise capacity and clinical status.

Accordingly, the Phosphodiester-ase-5 Inhibitionto Improve Clinical Sta-tus and Exercise Capacity in Heart Fail-ure with Preserved Ejection Fraction(RELAX) trial was designed to test thehypothesis that, compared with pla-cebo, therapy with the PDE-5 inhibi-

tor sildenafil would improve exercisecapacity in HFPEF after 24 weeks oftherapy, assessed by the change in peakoxygen consumption.

METHODS

Study Oversight

All study participants provided writ-ten informed consent prior to screen-

ing. The National Heart, Lung, andBlood Institutesponsored Heart Fail-ure Clinical Research Network con-ceived, designed, and conducted theRELAX trial. The trial protocol was ap-proved by a National Heart, Lung, and

Blood Instituteappointed protocol re-view committee and data and safetymonitoring board, and by the institu-tional review board at each participat-ingsite. TheDuke ClinicalResearch In-stitute served as the data coordinatingcenter.

Study Design

The rationale for, and design of, theRELAX trial have been described.3 Pa-tients with normal (50%) ejectionfraction and heart failure with NewYork Heart Association functional class

II through IV whose symptoms arestable while receiving medical therapywere eligible to participate if they hadobjective evidence of heart failure (pre-vious heart failure hospitalization, oracute heart failure therapy with intra-venous diuretic therapy, or chronicloopdiuretic therapy for heart failure withleft atrial enlargement, or invasivelydocumented elevation in left ventricu-lar filling pressures).

A peak oxygen consumption of 60%or less of the age- and sexspecific nor-

mal value9

with a respiratory ex-change ratio of 1.0 or more at the car-diopulmonary exercise test screening,and either elevated (400 pg/mL) N-terminal fragment of the precursor tobrain-type natriuretic peptide (NT-proBNP) level or elevation in left ven-tricular filling pressures at the time ofan NT-proBNP level less than 400pg/mL were required for study entry.A complete list of thetrial inclusion andexclusion criteria is provided in the on-line supplement (eTable 1, available at

http://www.jama.com). As required inf e d e r a l l y f u n d e d t r i a l s , a s e l f -identification of investigator-definedrace or ethnicity option was recorded.

Participants who met screening cri-teria underwent baseline studies (his-tory and physical examination, cardio-pulmonary exercisetest,6-minutewalkdistance, Minnesota Living with Heart

Failure Questionnaire, echocardiogra-phy, cardiac magnetic resonanceimaging [CMRI, if sinus rhythm], andphlebotomy for biomarkers), and werethen randomly assigned, in a 1:1 ratio,to receive either sildenafil or placebo

with the use of an automated web-based system. A permuted block ran-domization scheme was used withstratification according to clinical siteand the presence of atrial fibrillation.

Study drug was administered orallyat 20 mg, 3 times daily for 12 weeks.At the end of 12 weeks, each patientshistory and physical examination, car-diopulmonary exercise test, 6-minutewalk distance, and heart failure ques-tionnaire were repeated. Phlebotomyforsildenafil levels was performed 2 hoursafter a scheduled dose of the study drug

were obtained. The dose was then in-creased to 60 mg, 3 times daily for 12weeks,after which baseline studieswererepeated, including phlebotomy forsildenafil levels2 hours after study drug.If adverse effects developed, study staffcould recommend discontinuation orreturn to a lower or previously toler-ated dose of the study drug.

Blinded core laboratories assessedbiomarkers (University of Vermont),cardiopulmonary exercise tests (Mas-sachusetts General Hospital, Harvard

University), CMRIs (Duke Univer-sity), and echocardiograms (MayoClinic).

Study End Points

The primary end point was the changein peak oxygen consumption after 24weeks of therapy. A number of sub-group analyses were prespecified. Sec-ondary end points included a compos-ite hierarchical-rank clinical score inwhich patients were ranked (range, 1-nwith data) based on time to death

(tier 1), time to hospitalization forcardiovascular or cardiorenal causes(tier 2),andchangein heart failure ques-tionnaire from baseline (tier 3) for pa-tients without cardiovascular or cardio-renal hospitalization after 24 weeks oftherapy.10 Because189patientshaddatafor this end point, the anchor value(mean value in each group indicatingno

PHOSPHODIESTERASE-5 INHIBITION IN HEART FAILURE

E2 JAMA, Published online March 11, 2013 2013 American Medical Association. All rights reserved.



3/10

treatment effect) was 95.Other second-ary end points included change in6-minute walk distance at 24 weeksandchange in peakoxygenconsumptionand6-minute walk distance after 12 weeksof therapy. Peak sildenafil levels at 12

and24 weeks and coinciding plasma cy-clicguanosine monophosphate levels at24 weekswere assessed.Using other pre-specifiedendpoints, we assessed the ef-fect of PDE-5 inhibition on left ventricu-lar structure and vascular function byCMRI, Doppler-estimateddiastolic func-tion parameters and pulmonary arterysystolic pressure, and biomarkers thatreflect renal and neuroendocrine func-tion, oxidative stress, and collagenmetabolism.

The percent-predicted peak oxygenconsumption, 6-minute walk dis-

tance, and the presence of chrono-tropic incompetence and left ventricu-lar hypertrophy were calculated usingpublished criteria (eMethods).11-14

Statistical Analysis

Powercalculationswerebasedonthestan-darddeviation forchangeinpeakoxygenconsumption and the magnitude of itschange when associated with improve-ments in other markers ofclinical status(New York Heart Association class,6-minutewalkdistance,andqualityoflife

scores) in heart failure trials.5,14,15

Basedon these studies, a difference betweentreatmentgroupsof1.2mL/kg/mininthechangeinpeakoxygenconsumptionwasconsidered clinically significant.

We estimated a 20% rate of incom-plete primary end point data due todeath, withdrawal, or incidence of newfactorslimiting abilityto exercise. Usinga 2-sample t test and a 2-sided valueof .05, a sample size of 190 patients had85% power to detect a difference of 1.2mL/kg/min in change in peak oxygen

consumption, assuming 20% missingdata and a standard deviation of changein peak oxygen consumption of 2.5 mL/kg/min. As an early-blinded interimanalysis of aggregated primary endpoint completeness indicated that themissingness rate approached 20%, theblinded investigators recommended in-creasing thesample size to 215 patients.

The primary analyses were 2-sided,and patientswithout24-week datawereexcluded.Sensitivity analysis for the pri-mary end point was based on the inten-tion to treat principle and utilized mul-tiple imputations with 100imputed data

sets to account for missing 24-weekdata.In addition, a prespecified last obser-vation carried forward sensitivity analy-sisused carry-forwardof 12-week dataif 24-week data were missing.

Data are presented as median (inter-quartile range [IQR]). For the com-parison of treatment groups in theprimary analysis, a multivariable linear-regression model wasusedadjusting forbaseline peak oxygen consumption. Asimilar approach was used for the sec-ondary end point of change in 6-min-

ute walk distance, adjusting for base-line 6-minute walk distance. For thecomposite hierarchical-rank clinicalscore, patients were ranked indepen-dent of treatment assignment from 1(worst outcomethe earliest death) ton (best outcomesurvival with no car-diovascular or cardiorenal hospitaliza-tion, and the most favorable improve-

ment in heart failure questionnaire) andtreatments compared using the Wil-coxon rank sum test. For primary andsecondary end points, a P value lessthan .05 was considered significant. Forsubgroup analyses, a treatment by sub-

group interaction P less than .001 wasconsidered significant.

All analyseswere conducted with theuseofSASstatisticalsoftware,version9.2.

RESULTS

Patient Population

A total of 216 patients were enrolled inthetrial from October 13,2008, throughFebruary 21, 2012, at 26 sites in theUnitedStatesandCanada(FIGURE). Thebaseline characteristics were not sig-nificantly different between treatment

groups except for a lower prevalence ofhypertension in the sildenafil group(TABLE 1). Patients in this study wereolder, 48% women, andobese with con-trolled blood pressureand multiple co-morbidities including hypertension, is-chemic heart disease, atrial fibrillation,diabetes, anemia, and chronic kidneydisease.

Figure. Flow of Patients Through the Trial

94 Included in the primary analysis

9 Excluded

5 Withdrew consent

2 Unable to perform exercise test

1 Unwilling to perform exercise test

1 Had inadequate peak oxygen consumption data

91 Included in the primary analysis

22 Excluded

9 Withdrew consent

7 Unable to perform exercise test

2 Unwilling to perform exercise test

1 Had inadequate peak oxygen consumption data

3 Died

98 Completed 24-week follow-up visit

82 Completed study per protocol

16 Did not complete study per protocol

3 Discontinued study drugb13 Reduced dosage of study drug

101 Completed 24-week follow-up visit

74 Completed study per protocol

27 Did not complete study per protocol

8 Discontinued study drugc19 Reduced dosage of study drug

5 Did not complete 24-week follow-up visit(withdrew consent)b

12 Did not complete 24-week follow-up visit9 Withdrew consentc3 Died

103 Randomized to receive placebo

103 Received placebo as assigned

113 Randomized to receive sildenafil

113 Received sildenafil as assigned

216 Outpatients with heartfailure randomizeda

Some patients discontinued study drug for a specific reason (1 may apply) and later withdrew.aData regarding the number of patients screened for eligibility was not collected.b Study drug discontinued dueto adverse event (1), physiciandecision (1), other (2), or forwithdrawal of con-sent (4).c Study drug discontinued due to adverse event (9), physician decision (1), other (5), or for withdrawal of con-sent (5).





4/10

Table 1. Baseline Characteristics of the Patients

No. (%) of Patients

All (N = 216) Placebo (n = 103) Sildenafil (n = 113)

Age, median (IQR), y 69 (62-77) 69 (62-77) 68 (62-77)

Women 104 (48) 55 (53) 49 (43)

Self-reported white race 197 (91) 95 (92) 102 (90)

BMI, median (IQR) 32.9 (28.3-39.1) 32.8 (28.6-38.6) 33.3 (28.2-40.0)

NYHA functional classificationII 101 (47) 46 (45) 55 (49)

III 115 (53) 57 (55) 58 (51)

MLHFQ total score, median (IQR) 43 (30-62) 43 (29-58) 44 (30-63)

Systolic blood pressure, median (IQR), mm Hg 126 (113-138) 127 (113-138) 124 (113-138)

Heart rate, median (IQR), beats/min 69 (61-78) 68 (62-77) 70 (61-80)

Jugular venous pressure 8 cm a 95/209 (45) 48/99 (48) 47/110 (43)

Edema 125 (58) 61 (59) 64 (57)

1 Heart failure hospitalizations in previous year 79 (37) 40 (39) 39 (35)

Qualifying ejection fraction, median (IQR), % 60 (55-66) 60 (55-65) 60 (55-66)

Medical historyHypertension 183 (85) 93 (90) 90 (80)

Ischemic heart disease 84 (39) 37 (36) 47 (42)

Atrial fibrillation or flutter 111 (51) 52 (50) 59 (52)

Diabetes mellitus 93 (43) 45 (44) 48 (42)

COPD 42 (19) 20 (19) 22 (19)

Anemiab 76 (35) 34 (33) 42 (38)

Local laboratory creatinine, median (IQR), mg/dL 1.2 (0.9-1.5) 1.1 (0.9-1.5) 1.3 (1.0-1.5)

Glomerular filtration rate, median (IQR), mL/min/1.73 m2 57 (43-75) 59 (43-76) 57 (43-70)

Stage 3/4 chronic kidney disease 119 (55) 56 (54) 63 (56)

Medications at enrollmentLoop diuretic 166 (77) 79 (77) 87 (77)

Any diuretic 186 (86) 86 (83) 100 (88)

ACE inhibitor or angiotensin II receptor blocker 152 (70) 78 (76) 74 (65)

-Blocker 164 (76) 77 (75) 87 (77)

Aldosterone antagonist 23 (11) 9 (9) 14 (12)

Calcium channel blocker 66 (31) 35 (34) 31 (27)

Statin 138 (64) 67 (65) 71 (63)

Core laboratory biomarker dataCystatin C, median (IQR), mg/Lc 1.31 (1.07-1.74) 1.34 (1.08-1.76) 1.29 (1.04,1.70)

NT-proBNP, median (IQR), pg/mLd 700 (283-1553) 648 (247-1679) 757 (318-1465)

Aldosterone, median (IQR), ng/dL c 18.9 (11.8-28.3) 18.0 (11.4-29.4) 20.7 (13.0-27.5)

Endothelin-1, median (IQR), pg/mL c 2.36 (1.95-3.20) 2.37 (2.03-3.14) 2.35 (1.85-3.22)

NT-procollagen III, median (IQR), g/Lc 7.7 (6.1-10.0) 7.9 (6.4-11.2) 7.5 (5.5-9.3)

Uric acid, median (IQR), mg/dL e 7.4 (5.8-8.5) 7.1 (5.8-8.5) 7.5 (5.8-8.5)

Peak functional statusOxygen consumption, median (IQR), mL/kg/min f 11.7 (10.2-14.4) 11.9 (10.1-14.4) 11.7 (10.4-14.5)

Predicted peak oxygen consumption, median (IQR), % 41 (35-49) 41 (36-49) 41 (34-48)

Respiratory exchange ratio, median (IQR) f 1.09 (1.02-1.15) 1.10 (1.03-1.15) 1.09 (1.02-1.16)

Systolic blood pressure, median (IQR), mm Hg g 156 (132-170) 154 (134-168) 156 (130-170)

Chronotropic incompetence f 164 (77) 79 (78) 85 (76)

6-Minute walk distance, median (IQR), m 308 (229-383) 305 (213-372) 308 (241-392)

Predicted 6-minute walk distance, median (IQR), % 69 (51-83) 68 (48-83) 70 (54-83)

Abbreviations: ACE, angiotensin-converting enzyme; BMI, body mass index, calculated as weight in kilograms divided by height in meters squared; COPD, chronic obstructivepulmonary disease; IQR, interquartile range; MLHFQ, Minnesota Living with Heart Failure Questionnaire; NT-proBNP, N-terminal fragment of the precursor to brain-type natri-uretic peptide; NYHA, New York Heart Association.

SI conversion factors: To convert aldosterone from ng/dL to pmol/L, multiply by 27.74; creatinine from mg/dL to mol/L, multiply by 88.4; and uric acid from mg/dL to mol/L,multiply by 59.485.

a Indicates the number of patients/number of patients with nonmissing data for the variable (percentage).bAnemia is defined as hemoglobin13 g/dL for men, and 12 g/dL for women.cAvailable in 214 (102 placebo and 112 sildenafil-treated) patients.dAvailable in 213 (101 placebo and 112 sildenafil-treated) patients.eAvailable in 212 (101 placebo and 111 sildenafil-treated) patients.fAvailable in 215 (103 placebo and 112 sildenafil-treated) patients due to technical limitation precluding core laboratory measurement of peak oxygen consumption.gAvailable in 209 (99 placebo and 110 sildenafil-treated) patients.





5/10

The heart failure questionnaire scorewas consistent with the distribution ofNew York Heart Association functionalclass observed in the study population.Evidence of volume overload (elevatedjugular venous pressure or edema) and

hospitalization forheart failurein thepre-vious year were common. The majorityof patients were taking diuretics, renin-angiotensinsystemantagonists,-block-ers, and statins. There was evidence ofneuroendocrine activation and alteredcollagen metabolism as levels of NT-proBNP,aldosterone, endothelin-1, andNT-procollagen III exceeded referenceranges(eTable 2, availableat http://www.jama.com).Both peakoxygenconsump-tionand the6-minute walk distance werereduced and chronotropic incompe-tence was common.

Baseline echocardiographic andCMRI characteristics of the patientswere not significantly different be-tween treatment groups (TABLE 2). Atechocardiography,ejectionfractionandleft ventricular diastolic dimension(5.6 cm) were within normal limits,while cardiacindex wasreduced. Nearly50% of patients had left ventricular hy-

pertrophy or evidence of concentricre-modeling or hypertrophy (relative wallthickness 0.42).16 There was Dop-pler evidence of diastolic dysfunctionand elevatedleftventricular filling pres-sures with reduced early diastolic me-

dial mitral annular tissue velocity (e'

)and elevated ratio of early to late trans-mitral ventricular filling velocity (E/A),ratio of early transmitral filling veloc-ity to early diastolic medial mitral an-nular tissue velocity (E/e'), left atrial vol-ume, and pulmonary artery systolicpressure.

Overall, 132 patients (61%) were eli-gible for CMRI, and 117 of these un-derwent CMRI (89%). Ejection frac-tion and left ventricular end-diastolicvolume index were within normal lim-its while cardiac index was reduced. In

the CMRI cohort, 25% of patients hadleft ventricular hypertrophy. Aortic dis-tensibility was reduced compared withpublished values in older participantswithout heart failure.17

Primary End Point

At 24 weeks, themedian changein peakoxygen consumption frombaselinewas

not significantly different in patientstreated with placebo (0.20 [IQR,0.70 to 1.00]) and patients treatedwith sildenafil (0.20 [IQR,1.70 to1.11]), (P=.90)(TABLE3). Using mul-tiple imputation to account for miss-

ing 24-week data, the mean differencebetween sildenafil and placebo was 0.01mL/kg/min (favoring sildenafil, [95%CI, 0.60 to 0.61], P=.98). Carryingforward 12-week peak oxygen con-sumption when 24-week data weremissing (placebo [n=5] and sildenafil[n=9]), median change in peak oxy-gen consumption (mL/kg/min) frombaseline was 0.20 (IQR, 0.83 to1.10) in patients treated with placeboand0.13 (IQR,1.50 to 1.16) in pa-tients treated with sildenafil (P=.98).In subgroup analyses (eTable 3), the

change in peak oxygen consumptionwas not significantly different be-tween treatment groups when analy-sis was restricted to those patients stilltaking thestudy drug at week 24, in pa-tients with or without left ventricularhypertrophy by CMRI, with lower orhigher pulmonary artery systolicpressure, with lower or higher NT-

Table 2. Baseline Core Laboratory Echocardiographic and Cardiac Magnetic Resonance Imaging Characteristics of the Patients

Placebo SildenafilAll Patients,

Median (IQR)(N=216)

No. of Patients( n = 103) Median ( IQR)

No. of Patients( n = 113) Median (IQR)

Doppler echocardiographic dataEjection fraction, % 103 60 (56-65) 110 60 (55-65) 60 (56-65)

Cardiac index, L/min/m2 89 2.48 (2.06-2.86) 86 2.47 (2.09-2.92) 2.47 (2.07-2.92)

Left ventricular end-diastolic dimension, cm 76 4.6 (4.3-5.1) 88 4.6 (4.2-5.2) 4.6 (4.2-5.2)

Left ventricular mass/BSA, g/m2 73 77 (63-90) 85 78 (61-97) 77 (62-96)

Left ventricular hypertrophy, No. (%) 73 36 (49) 85 40 (47) 76 (48)

Relative wall thickness 0.42, No. (%) 73 32 (44) 85 43 (51) 75 (48)

E/A ratio 70 1.3 (1.0-2.0) 72 1.6 (1.0-2.1) 1.5 (1.0-2.0)

Deceleration time, ms 95 190 (156-228) 98 182 (153-212) 185 (155-218)

Medial e', m/s 92 0.06 (0.05-0.07) 105 0.06 (0.05-0.08) 0.06 (0.05-0.08)

Medial E /e' 90 17 (12-24) 98 15 (10-23) 16 (11-23)

Left atrial volume index, mL/m2 74 43 (38-59) 75 44 (33-59) 44 (35-59)

Pulmonary artery systolic pressure, mm Hg 75 41 (34-54) 63 41 (32-51) 41 (33-51)

CMRI data, median (IQR)Ejection fraction, % 56 66 (59-71) 61 66 (55-70) 66 (58-70)

Cardiac index, L/min/m2 56 2.38 (1.94-2.80) 59 2.32 (2.11-2.74) 2.35 (1.98-2.76)

Left ventricular end-diastolic volume/BSA, mL/m2 56 58 (49-68) 61 53 (46-66) 56 (47-67)

Left ventricular mass/BSA, g/m2 56 61 (53-73) 61 65 (54-78) 63 (54-77)

Left ventricular hypertrophy, No. (%) 56 14 (25) 61 16 (26) 30 (26)

Aortic distensibility, median (IQR), 103 mm Hg-1 41 1.08 (0.63-1.64) 45 1.18 (0.74-2.07) 1.17 (0.67-1.76)

Abbreviations: BSA, body surface area; CMRI, cardiac magnetic resonance imaging; e ', early diastolic medial mitral annular tissue velocity; E/A, elevated ratio of early to late trans-mitral ventricular filling velocity; E/e', ratio of early transmitral filling velocity to early diastolic medial mitral annular tissue velocity; IQR, interquartile range.





6/10

proBNP, with or without atrial fibril-lation or treated or not with renin-angiotensin system antagonists,-blockers, or statins.

Secondary End Points

and Safety Data

There were no significant differences inthe clinical rankscore, change in 6-min-ute walk distance at 24 weeks, orchangein peak oxygenconsumption or6-minute walk distance at 12 weeks be-tween treatment groups (Table 3).There were no significant differences inthe components of the clinical rankscore at 24 weeks or in the overall in-cidence of adverse, or serious adverse,events in the treatment groups.

Adverse events occurred in 78 pa-tients (76%) who received placebo and

90 patients (80%) who receivedsildenafil. Serious adverse events oc-curred in 16 patients (16%) who re-ceived placebo and 25 patients (22%)who received sildenafil. Adverse eventsoccurring in 5% or more of either studygroup are listedin eTable 4 (available athttp://www.jama.com). Patients treatedwith sildenafil hada higher incidenceofvascular adverse events, which in-cluded (but were not limited to) head-ache, flushing, and hypotension, al-though the change in mean arterial

pressure from baseline to 24 weeks wasnot significantly different in patients

treated withsildenafil (1 [95% CI,8to 6]) and patients treated with placebo(2 [95% CI, 10 to 7]), (P=.45). Allserious adverse eventsexclusiveof deathor cardiovascular or cardiorenal hospi-talization (Table 3) are listed in eTable

5. There were no other notable differ-ences in the incidence of specific seri-ousadverse eventsbetween study groups.

Study Drug and Cyclic Guanosine

Monophosphate Levels

Median sildenafil concentrations mea-sured approximately2 hours after thelastdose were 78 (IQR, 35-130 ng/mL) at 12weeks and 200 (IQR, 92-330 ng/mL) at24 weeks. At week 24, there was a weakcorrelation between sildenafil dose andsildenafil level (r=0.29, P=.008). Inpaired analysis, plasma cyclic guano-

sine monophosphatelevels increased sig-nificantly from baseline to 24 weeks inpatients randomized to receive sildena-fil (mean increase, 8.72pmol/mL, [95%CI, 2.56-14.87],P=.006), but not in pa-tients randomized to placebo (mean in-crease, 1.28 pmol/mL, [95% CI,6.27to 8.83],P=.74), although thechange incyclic guanosine monophosphate wasnot significantly different between groups(P=.11).

Additional End Points

At CMRI, there was no difference inchange in left ventricular mass or left

ventricular end-diastolic volume indexbetween treatment groups (TABLE 4).There was also no difference in changein Doppler-assessedleft ventricular dia-stolic function parameters or pulmo-nary artery systolic pressure between

treatmentgroups. By CMRI,arterial elas-tance decreasedmore andsystemic vas-cular resistancetended to decrease morein patients treated with sildenafil(P=.09). However, the change in meanarterial pressure in the entire studypopulation was not significantly differ-ent between groups, as noted above.More patients had missing data for aor-ticdistensibility at 24 weeksthan at base-line, but there was no difference inchangein distensibility between groups.Patients treated with sildenafil had agreater increase in creatinine, cystatin C,

NT-proBNP, uric acid, and endothe-lin-1 than patients treated with pla-cebo, whereas changes in aldosteroneand NT-procollagen III werenot signifi-cantly different between groups.

COMMENT

To our knowledge, the RELAX trial isthe first multicenter study to investi-gate the effect of PDE-5 inhibition inHFPEF. Contrary to our hypothesis,long-term PDE-5 inhibition in HFPEFhad no effect on maximal or submaxi-

mal exercise capacity, clinical status,quality of life, left ventricular remod-

Table 3. Primary, Secondary, and Safety End Points

Placebo Sildenafil

P

ValueNo. of

Patients VariableNo. of

Patients Variable

Primary end pointChange in peak oxygen consumption at 24 wk, median (IQR), mL/kg/min 94 0.20 (0.70 to 1.00) 91 0.2 (1.70 to 1.11) .90

Secondary end pointsClinical rank score, meana 94 95.8 95 94.2 .85

Change in 6- minute walk distance at 24 wk, median (IQR), m 95 15.0 (26.0 to 45.0) 90 5.0 (37.0 to 55.0) .92

Change in peak oxygen consumption at 12 wk, median (IQR), mL/kg/min 96 0.03 (1.10 to 0.67) 97 0.01 (1.35 to 1.25) .98

Change in 6- minute walk distance at 12 wk, median (IQR), m 96 18.0 (14.5 to 48.0) 99 10.0 (25.0 to 36.0) .13

Components of clinical rank score at 24 wkDeath, No. (%)b 103 0 113 3 (3) .25

Hospitalization for cardiovascular or renal cause, No. (%) 103 13 (13) 113 15 (13) .89

Change in MLHFQ, median (IQR) 91 8 (21 to 5) 91 8 (19 to 0) .44

Safety end points, No. (%)Adverse events 103 78 (76) 113 90 (80) .49

Serious adverse events 103 16 (16) 113 25 (22) .22

Abbreviations: IQR, interquartile range; MLHFQ, Minnesota Living with Heart Failure Questionnaire.aA mean value of 95 in each group is expected under the null hypothesis of no treatment effect.b Site investigator identified causes of death were sudden death (n=1), progressive cardiorenal failure (n=1), and noncardiovascular (n=1).





7/10

eling, diastolic function parameters, orpulmonary artery systolic pressure. Re-nal function worsened more and NT-proBNP, endothelin-1, and uric acidlevels increased more in patientstreatedwithsildenafil. Furthermore, there were

more (but not significantly more) pa-tients in the sildenafil group who with-drew consent, died, or were too ill toperform the cardiopulmonary exer-cise test, and patients treated withsildenafil had a higher incidence of vas-cular adverseevents. These findings donotsuggest that therapy with thePDE-5inhibitor sildenafil provides clinicalbenefit in the general HFPEF popula-tion.

Given the strong rationale for test-ing PDE-5 inhibitionin HFPEF3 and thelack of benefit observed, it is impor-

tant to consider whether the studypopulation, specifics of the therapeu-tic intervention, and end points wereappropriate.

The clinical characteristics of pa-tients with HFPEF enrolled in major on-going or completed clinical trials havebeen summarized.18 The RELAX studypopulation wassimilarto othersin terms

of age, sex distribution, and body size.Severityof heartfailure in our study wasalso similar to or greater than otherHFPEF trials, althoughcomorbiditybur-den (diabetes, atrial fibrillation, kidneydisease) may have been greater in this

study population. Concentric remodel-ing and hypertrophy were common butnot severe, and Doppler evidence of el-evated filling pressures and pulmonaryhypertension were present as was neu-roendocrine activation consistent withthe heart failure state.

However, the characteristics of thestudy population were notably differ-ent from the only other study to ourknowledge that evaluated the effect ofPDE-5 inhibition in HFPEF. In thestudy by Guazzi and coauthors,8

sildenafil had a number of beneficial ef-

fects as outlined above, although ef-fect on exercise capacity wasnot tested.Importantly, in thatstudy, patientswithHFPEF had fewer comorbidities andsignificantly higher blood pressure, leftventricular mass, and pulmonary ar-tery systolic pressure than the pa-tients in the RELAX study and cath-eterization-documented pulmonary

arterial hypertension, profound rightventricular systolic dysfunction, andright ventricular failure were present.This profile is somewhat atypical forHFPEF cohorts.19

It may be that the primary therapeu-

tic effects of PDE-5 inhibitors in heartfailure involve the drugs ability to di-late the pulmonary vascular bed, en-hance right ventricular contractility,and reduce ventricular interdepen-dence,4,19-22 and that pulmonary arte-rial hypertension and right ventricu-lar failure must be significant in orderto observe clinical benefit in HFPEF.Our subgroup analysis didnotshow anytrends toward improvement in peakoxygen consumption in patients withhigher pulmonary artery systolic pres-sure, but the presence of pulmonary ar-

terial hypertension or right ventricu-lar dysfunction was not assessed in thisstudy.

Although left ventricular hypertro-phy wascommon in participants in thisstudy, it was far less severe than amongparticipants in the study by Guazzi etal.8 In murine-pressure overload stud-ies, PDE-5 inhibition did not have an-

Table 4. Additional Prespecified End Points

Placebo Sildenafil

P

ValueNo. of

Patients Median (IQR)No. of

Patients Median ( IQR)

Change in left ventricular structure by CMRI at 24 wkLeft ventricular mass by CMRI, g 47 0.6 (5.7 to 7.9) 49 1.5 (5.9 to 7.1) .93

Left ventricular end-diastol ic volume by CMRI, mL 47 4.3 (15.5 to 8.1) 49 3.7 (4.9 to 14.5) .13

Change in diastolic function parameters at 24 wkMedial e', m/s 83 0.00 (0.01 to 0.01) 77 0.00 (0.01 to 0.01) .88

E/e ' 80 1.6 (4.7 to 2.2) 75 0.2 (2.4 to 3.1) .16

PA systolic pressure, mm Hg 58 2 (8 to 8) 45 2 (5 to 7) .94

Change in vascular function by CMRI at 24 wkArterial elastance, mm Hg/mL 45 0.03 (0.14 to 0.23) 47 0.08 (0.41 to 0.12) .02

Systemic vascular resistance, wood units 45 0.13 (0.15 to 0.43) 47 0.04 (0.38 to 0.21) .09

Aortic distensibility, 103 mm Hg-1 31 0.13 (0.17 to 0.42) 29 0.18 (0.12 to 0.67) .38

Change in core laboratory biomarkers at 24 wkCreatinine, mg/dL 94 0.01 (0.10 to 0.09) 94 0.05 (0.04 to 0.15) .047

Cystatin C, mg/L 95 0.01 (0.08 to 0.11) 95 0.05 (0.04 to 0.16) .01

NT-proBNP, pg/mL 94 23 (198 to 139) 95 15 (90 to 372) .03

Endothelin-1, pg/mL 95 0.01 (0.48 to 0.47) 95 0.38 (0.10 to 0.97) .046

Aldosterone, ng/dL 95 0 (7.0 to 4.8) 95 1.1 (7.7 to 3.0) .85

NT-procollagen III, g/L 93 0.03 (1.49 to 1.54) 95 0.07 (1.17 to 1.42) .77

Uric acid, mg/dL 94 0.1 (0.7 to 0.7) 94 0.3 (0.4 to 1.4) .02

Abbreviations: CMRI, cardiac magnetic resonance imaging; e', early diastolic medial mitral annular tissue velocity; E/e ', ratio of early transmitral filling velocity to early diastolic medialmitral annular tissue velocity; IQR, interquartile range; NT-proBNP, N-terminal fragment of the precursor to brain-type natriuretic peptide; PA, pulmonary artery.

SI conversions: To convert aldosterone from ng/dL to pmol/L, multiply by 27.74; creatinine from mg/dL to mol/L, multiply by 88.4; and uric acid from mg/dL to mol/L, multiplyby 59.485.





8/10

tihypertrophic effects in mice with lesssevere pressure overload and compen-sated left ventricular hypertrophy withrelatively preserved ejection fraction,whereas dramatic antiremodeling ben-efits were observed in mice with se-

vere pressure overload, eccentric leftventricular hypertrophy, reduced ejec-tion fraction, and pulmonary conges-tion.23,24 Conceivably, activation ofPDE-5 or of cyclic guanosine mono-phosphatesensitive downstreampath-ways in the left ventricle or other or-gans may occur only in heart failureassociated with advanced left ventricu-lar remodeling.

In a randomized clinical trial ofPDE-5 inhibition in pulmonary arte-rial hypertension, the effect of sildena-fil on exercise capacity was not dose re-

lated as improvement in 6-minute walkdistance was seen with 20 mg 3 timesdaily after just 4 weeks of therapy withno further improvement with higherdoses or longer duration of therapy.4

Sildenafil levels at 12 and 24 weeks werevariable, but on average, similar to ran-dom levels observed with similar dosesamong participants with heart failureandreduced ejection fraction for whomimprovement in exercise capacity withlong-term sildenafil use was observed(Gregory Lewis, MD, Massachusetts

General Hospital, written communica-tion, February 2, 2013).5

For patientswith heart failure andre-duced ejection fraction, the benefit ofPDE-5 inhibition on exercise capacityhas been demonstrated acutely after asingle 50-mg dose,25,26 and with 75 mg,3 times daily for 12 weeks (dose upti-trated over the course of 6 weeks).5 Al-though only 73% of patients in thesildenafil group attending the 24-week follow-up were taking the per-protocol dose, 92% of patientswere tak-

ing at least 20 mg 3 times daily. Whilestudies in pulmonary arterial hyper-tension and heart failure and reducedejection fraction have observed effectson exercise capacity with similar dosesand duration of therapy, we cannot ex-clude the possibility that inadequatedose or duration of PDE-5 inhibitioncontributed to our findings.

Therapeutic sildenafil levels wereassociated with minimal increases inplasma cyclic guanosine monophos-phate. Heart failure with preservedejection fraction is characterized byendothelial dysfunction27 and by

lower natriuretic peptide levels thanobserved in heart failure and reducedejection fraction,28 which may suggestlimited nitric oxide and natriureticpeptide activity in HFPEF. Inability toenhance cyclic guanosine monophos-phate w i th PDE- 5 i nhi b i ti o n i nHFPEF may have contributed to ourfindings.

As previously described,3 change inpeak oxygen consumption was cho-sen as the primary end point in theRELAX trial based on previous pre-clinical and clinical studies and be-

cause noncardiovascular comorbidi-ties and motivational factors caninfluence measures of submaximal ex-ercise performance in HFPEF.Our trialwas powered to detect a clinically sig-nificantdifference in thechangein peakoxygen consumption between groupsand the estimate of variability (SD, 2.5mL/kg/min) in change in peak oxygenconsumption used in the power calcu-lations was consistent with the stan-dard deviation of change in peak oxy-gen consumption observed in the

patients treated with placebo (2.0 mL/kg/min). Thelack of treatmenteffectonsubmaximal exercise, clinical status,andphysiologic end points supportsthevalidity of the observed lack of treat-ment effect on maximal exercise ca-pacity.

The high prevalence of chrono-tropicincompetence in the study popu-lation is noteworthy. Chronotropic in-competence may contribute to exerciselimitation in HFPEF, and may not beimproved by PDE-5 inhibition.

Although numerous studies in ani-mal models of renal dysfunction sug-gest that PDE-5 inhibition amelio-rates progression of renal dysfunctionof various etiologies,3,29-31 in this trial,modest but statisticallysignificant wors-ening of renal function was observedin patients treated with sildenafil andwas associated with concordant in-

creases in NT-proBNP, uric acid, andendothelin-1, suggesting that the de-cline in renal function was physiologi-cally significant. Studies in pulmo-nary arterial hypertension and erectiledysfunction have not reported wors-

ening of renal function with PDE-5 in-hibitor therapy, but little is known ofthe effect of PDE-5 inhibition on renalfunction in HFPEF.

There were more (but not signifi-cantly more) patients who withdrewconsent, died, or were too ill to per-form the cardiopulmonary exercise testin the sildenafil treatment group, po-tentially accentuating the lack of ben-efit observed, particularly if those whowithdrew did so due to adverse effectsor poor clinical status.

A modest decrease in arterial elas-

tance wasnoted in patients treated withsildenafil in the CMRI cohort. This mayhave been related to an effect ofsildenafil on resistance that tended todecrease more in patients treated withsildenafil, butin theentirecohort, therewere no differences in change in meanarterial pressure between treatmentgroups.

The findings ofthisstudy must be in-terpreted in the context of other po-tential limitations. Multicenter trialsusing peak oxygen consumption as a

primary end point are challenging, butrigorous methodologies were used inthe design and execution of the exer-cise test study protocol.3 Patients wereselected who could perform the test,and who had significant reduction inpeak oxygen consumptionthese en-try criteria may have selected for auniqueHFPEF phenotype.Thetrial wasnot powered to address differences inclinical outcomes.

CONCLUSIONS

Among patientswithHFPEF,PDE-5 in-hibition with administration ofsildenafil for 24 weeks, compared withplacebo, did not resultin significantim-provement in exercise capacity or clini-cal status. Continued efforts to iden-tify key pathophysiologic perturbationsandnovel therapeutic targets in HFPEFare needed.





9/10

Published Online: March 11, 2013. doi:10.1001/jama.2013.2024Author Affiliations: Department of Medicine, MayoClinic, Rochester, Minnesota (Drs Redfield, Chen,Borlaug,Lin, andOh); Department of Medicine, Mas-sachusetts GeneralHospital, Boston(Drs Semigran andLewis); Departments of Biostatisticsand Bioinformat-ics(Drs Leeand Anstrom) andMedicine (DrsVelazquezand Hernandez), Statistics (Mr McNulty), and Proj-

ect Management (Ms Ibarra), Duke Clinical ResearchInstitute, Durham, North Carolina; Departments ofMedicine (Dr LeWinter) and Pathologyand Biochem-istry (DrTracy), Universityof Vermont, Burlington; De-partment of Medicine, University of Montreal andMontreal Heart Institute, Montreal, Canada (Dr Rou-leau); Department of Surgery,University of Utah,SaltLake City (Dr Bull); Department of Medicine, BarnesJewish Hospital, St Louis, Missouri (Dr Mann); De-partment of Medicine,Michael E. DeBakey VA Medi-cal Center and Baylor College of Medicine, Houston,Texas (Dr Deswal); Department of Medicine, Brighamand Womens Hospital, Boston, Massachusetts (DrsStevenson, Givertz, and Braunwald); Department ofMedicine, Morehouse School of Medicine, Atlanta,Georgia (Dr Ofili); Departments of Medicine (DrsOConnor, Felker, Patel, and Kim) and Duke Cardio-vascular Magnetic Resonance Center (Dr Kim), DukeUniversityMedicalCenter andDuke HeartCenter,Dur-

ham, North Carolina; Department of Medicine, Hen-nepin County Medical Center, Minneapolis, Minne-sota (Drs Goldsmith and Bart); and Division ofCardiovascular Sciences, National Heart, Lung, andBlood Institute, Bethesda, Maryland (Dr Mascette).For a complete listing of the RELAX site representa-tives see the online supplement at jama.com.Author Contributions: Drs Lee and Anstrom had fullaccess to all the data in the study and take responsi-bility for the integrity of the data and the accuracy ofthe data analysis.Study concep t and desig n: Chen, Semigran, Lee,LeWinter, Rouleau,Bull, Deswal, Stevenson, OConnor,Felker, Goldsmith, Bart, McNulty, Velazquez,Hernandez, Braunwald, Redfield.Acquisition of data: Chen, Borlaug, Semigran, Lee,Lewis, LeWinter, Rouleau, Bull, Deswal, Stevenson,Givertz,Ofili, OConnor, Felker, Bart, McNulty,Ibarra,Lin, Patel,Kim, Tracy,Velazquez, Hernandez, Redfield.

Analysis and interpretation of data:Borlaug, Semigran,Lee, Lewis, LeWinter, Bull, Mann, Deswal, Givertz,Felker, McNulty, Ibarra, Lin, Oh, Patel, Kim, Tracy,Velazquez,Anstrom,Hernandez, Mascette,Braunwald,Redfield.Drafting of the manuscript: Chen,Borlaug,LeWinter,Bull, OConnor, McNulty, Velazquez, Redfield.Critical revision of the manuscript for important in-tellectual content: All authors.Statistical analysis: Lee, Bull, McNulty, Velazquez,Anstrom.Obtainedfunding: Chen, Semigran, Lee, Bull, Deswal,Ofili, Felker, McNulty, Velazquez, Hernandez,Mascette, Redfield.Administrative, technical,or material support:Chen,Borlaug, Lee,Rouleau,Bull, OConnor, McNulty, Ibarra,Oh, Kim, Tracy, Velazquez, Hernandez, Mascette,Braunwald.Study supervision: Borlaug, Lewis, Deswal, Givertz,

OConnor, Felker, Goldsmith, Bart, McNulty, Kim,Velazquez, Mascette, Braunwald, Redfield.Conflictof InterestDisclosures: All authorshave com-pleted and submitted the ICMJE Form for Disclosureof Potential Conflicts of Interest. Dr Redfield re-ported receiving financial support from the NationalInstitutes of Health (NIH) and royalties from An-nexon. Dr Chen reported receiving funding for pat-ents that hisinstitutionhas licensed to NileTherapeu-tics and Annexon with other patents pending at theUS Patent andTrademark Office.He also reported thathe has royaltieswith Nile Therapeutics,Annexon, and

Up To Date. Dr Borlaug reported receiving consult-ing fees from Medscape, Amgen,and GlaxoSmithKlineandreceiving financial support fromAtcorMedical andGilead. Dr Semigran reported receiving financial sup-port from the NIH for participation in the Heart Fail-ureNetwork andfrom Bayer. Dr Leereportedreceiv-ing financial support from the NIH for participationin the Heart Failure Network. Dr LeWinter reportedreceiving financial support from the NIH for partici-

pation in the Heart Failure Network. Dr Bull reportedreceiving financial support from the NIH for partici-pation in the Heart Failure Network. Dr Deswal re-ported receiving financial support from the NIH forparticipation in the HeartFailure Network. Dr Steven-son reported receiving financial support from the NIHfor participation in the Heart Failure Network. Dr Gi-vertz reportedreceiving financial support fromthe NIHforparticipationin theHeart Failure Network. Dr Ofilireported receivingfinancialsupport from the NIH, re-ceiving consulting feesfrom Merck, Novartis, and Ar-borPharmaceuticals andlecturefees fromArbor Phar-maceuticals. She is currently employed by theMorehouseSchool of Medicine andhas grantspend-ing from the National Heart, Lung, and Blood Insti-tute and National Institute on Minority Health andHealth Disparities. Dr OConnor reportedreceivingcon-sulting fees from Roche Diagnostics, Amgen, Novar-tis, Ikaria, Acetlion Pharma, Heartware, ResMed, Pozen,

GE Healthcare, Johnson& Johnson,Gilead,Critical Di-agnostics, BG Medicine, Otsuka,Astellas, Novella,Cy-tokinetics, and Capricor; holding stock options withNeurotronik/Interventional Autonomics Corpora-tion;editing AmericanCollegeof Cardiology; and co-owning Cardiology Consulting Associates. Dr Felkerreported receiving financial support fromthe NIH;con-sulting fees from Novartis, Amgen, Trevena, RocheDiagnostics, Merck, Singulex, and BG Medicine; andfunding from Amgen, Otsuka, and Roche Diagnos-tics. Dr Bart reported receivingfinancial support fromthe NIH for participation in the Heart Failure Net-work. Ms Ibarra reported receiving financial supportfromthe NIH forparticipation in theHeart FailureNet-work. Dr Oh received financial support from the NIHfor the Echo Core Lab. Dr Patel has received consult-ingfees fromGenzyme,Baxter, Jensen, andBayer, andhis financial institutionreceives financial support fromAstraZeneca, Johnson& Johnson, andMaquet. Dr Kim

reported receiving financial support from the NIH asthe Heart Failure Network Data Coordinating Centerand MRI Core Lab, and funding for a patent withNorthwestern University. His institution has receiveda grant from Siemens. Dr Tracy reported receiving fi-nancial support from the NIH for participation in theHeartFailure Network, andconsultingfees fromMerck,TibotecJohnson& Johnson, andAbbott. He hasalsogiven experttestimony about biomarkersin blood clot-ting and is the owner of Haematologic Technologies.Dr Velazquez reported receiving financial support fromthe NIH for participation in the Heart Failure Net-work. Dr Anstrom reported receiving financial sup-port from the NIH for participation in the Heart Fail-ure Network. Dr Hernandezreportedreceivingfinancialsupport from the NIH for participation in the HeartFailure Network and consulting fees from AstraZen-eca, Corthera, Janseen, and Bristol-Myers Squibb. Hisinstitution has received funding from Amylin, John-

son & Johnson, and Bristol-Myers Squibb. Dr Braun-wald reported receiving consulting fees from Merck,Daiichi Sankyo, Genzyme, Amorcyte, Medicines Co,MC Communications, Ikaria, CardioRentis, sanofi-aventis; funding (or pending funding) fromAstraZen-eca, Johnson & Johnson, Merck, sanofi-aventis, Dai-ichi Sankyo, GlaxoSmithKline, Bristol-Myers Squibb,BeckmanCoulter, Roche Diagnostics, andPfizer; com-pensation for lectures for Eli Lilly, Merck, CVRx (nocompensation), CV Therapeutics (now Gilead), Dai-ichi Sankyo, MC Communications, Manarini Interna-tional,Medscape, andBayer; andpaymentfor devel-

opment of educational presentations from MCCommunications. No other disclosures were re-ported.Funding/Support:This work wassupportedby grantsU10HL084904(for the datacoordinating center),andU10HL084861, U10HL084875, U10HL084877,U10HL084889, U10HL084890, U10HL084891,U10HL084899, U10HL084907, and U10HL084931(forthe clinicalcenters) fromthe National Heart, Lung,

andBlood Institute.Thisworkis also supportedby grantUL1TR000454 from the National Centerfor Advanc-ingTranslational Sciences andgrant8 U54MD007588from the National Institute on Minority Health andHealthDisparities.Pfizer provided studydrug (sildenafiland matched placebo).Role of the Sponsor: Pfizer had no role in the designand conduct of the study, collection, management,analysis, or interpretation of the data,or preparation,review, or approval of the manuscript. National Heart,Lung, andBloodInstitute representativesprovided ad-vice on the trial design and appointed the protocolreview committee and data and safety monitoringboard.Disclaimer: Theauthors aresolelyresponsiblefor thecontent of this article, which does not necessarily rep-resent the official views of the National Heart, Lung,andBloodInstitute or theNational Institutes of Health.Online-Only Material: The eMethods, eReferences,eTables 1 through 5, and eAppendices are availableat http://www.jama.com.AdditionalContributions: Wethankthepatientswhoparticipated in this study, the Heart Failure Networksite investigators and coordinators, the members ofthe Heart Failure Network data and safety monitor-ingboardand protocol review committee, andthe Na-tional Heart, Lung, and Blood Institute representa-tives. For a complete listing of the Heart FailureNetwork members, see Appendix C available athttp://www.jama.com.

REFERENCES

1. Owan TE, Redfield MM. Epidemiology of dia-stolic heart failure. Prog Cardiovasc Dis. 2005;47(5):320-332.2. Shah RV,Desai AS,Givertz MM.The effectof renin-angiotensin system inhibitors on mortality and

heartfailure hospitalization in patients with heartfail-ure and preserved ejection fraction: a systematic re-view and meta-analysis. J Card Fail. 2010;16(3):260-267.3. Redfield MM, Borlaug BA, Lewis GD, et al; HeartFailureClinical Research Network. PhosphdiesteRasE-5inhibition to improve clinical status and exercise ca-pacityin diastolic heart failure (RELAX)trial:rationaleand design. Circ Heart Fail. 2012;5(5):653-659.4. Gali N, Ghofrani HA, Torbicki A, et al; SildenafilUse in Pulmonary ArterialHypertension(SUPER) StudyGroup. Sildenafil citrate therapy for pulmonary arte-rial hypertension. N Engl J Med. 2005;353(20):2148-2157.5. Lewis GD, Shah R, Shahzad K, et al. Sildenafil im-provesexercise capacity andquality of lifein patientswith systolic heart failure and secondary pulmonaryhypertension. Circulation. 2007;116(14):1555-1562.

6. Guazzi M, Vicenzi M, Arena R. Phosphodiester-ase 5 inhibition with sildenafil reverses exercise oscil-latory breathing in chronic heart failure: a long-termcardiopulmonary exercise testing placebo-controlledstudy. Eur J Heart Fail. 2012;14(1):82-90.7. Guazzi M, Vicenzi M, Arena R, Guazzi MD. PDE5inhibition with sildenafil improves left ventriculardia-stolic function, cardiac geometry, and clinical statusin patients with stable systolic heart failure: resultsof a 1-year, prospective, randomized, placebo-controlled study. Circ Heart Fail. 2011;4(1):8-17.





10/10

8. Guazzi M, Vicenzi M, Arena R, Guazzi MD. Pul-monary hypertension in heart failure with preservedejection fraction: a target of phosphodiesterase-5 in-hibition in a 1-year study. Circulation. 2011;124(2):164-174.9. Fletcher GF, Balady G, Froelicher VF, Hartley LH,Haskell WL, PollockML; Writing Group. Exercisestan-dards: a statement for healthcare professionals fromthe American Heart Association. Circulation. 1995;

91(2):580-615.10. Packer M. Proposal for a new clinical end pointto evaluate the efficacy of drugs and devices in thetreatment of chronic heart failure. J Card Fail. 2001;7(2):176-182.11. Enright PL, Sherrill DL. Reference equations forthe six-minutewalk in healthyadults. AmJ Respir CritCare Med. 1998;158(5 Pt 1):1384-1387.12. KhanMN, Pothier CE,LauerMS. Chronotropic in-competenceas a predictorof deathamongpatients withnormal electrogramstakingbeta blockers(metoprololoratenolol). Am J Cardiol. 2005;96(9):1328-1333.13. Chirinos JA, Segers P, De Buyzere ML, et al. Leftventricular mass: allometric scaling, normative val-ues, effect of obesity, and prognostic performance.Hypertension. 2010;56(1):91-98.14. Feldman AM,Silver MA,FrancisGS, et al;PEECHInvestigators. Enhanced external counterpulsation im-proves exercise tolerance in patientswith chronicheartfailure. J Am Coll Cardiol. 2006;48(6):1198-1205.15. Abraham WT, Fisher WG, Smith AL, et al;MIRACLE StudyGroup, MulticenterInSync Random-ized Clinical Evaluation. Cardiac resynchronization inchronic heart failure. N Engl J Med. 2002;346(24):1845-1853.16. Lang RM, Bierig M, Devereux RB, et al; Cham-ber Quantification Writing Group; American Societyof Echocardiographys Guidelines and StandardsCommittee; EuropeanAssociation of Echocardiography.Recommendations for chamber quantification: a re-

port from the American Society of Echocardio-graphys Guidelines andStandardsCommitteeand theChamber QuantificationWritingGroup, developed inconjunction with the European Association of Echo-cardiography, a branch of the European Society ofCardiology. J Am Soc Echocardiogr. 2005;18(12):1440-1463.17. Hundley WG, Kitzman DW, Morgan TM, et al.Cardiac cycle-dependent changes in aortic area and

distensibility are reduced in older patients with iso-lated diastolic heart failure and correlate with exer-cise intolerance. J Am Coll Cardiol. 2001;38(3):796-802.18. Shah SJ, Heitner JF, Sweitzer NK, et al. Baselinecharacteristics of patients in the Treatment of Pre-served Cardiac Function Heart Failure with an Aldo-sterone Antagonist (TOPCAT) trial. Circ Heart Fail.2012.19. Forfia PR, Borlaug BA. Letter by Forfia and Bor-laug regarding article, Pulmonary hypertension inheart failure with preserved ejection fraction: a tar-get of phosphodiesterase-5 inhibition in a 1-yearstudy. Circulation. 2012;125(8):e408-e410.doi:10.1161/CIRCULATIONAHA.111.064584.20. Nagendran J, Archer SL, Soliman D, et al. Phos-phodiesterase type 5 is highly expressed in the hy-pertrophied human right ventricle, and acute inhibi-tionof phosphodiesterase type5 improves contractility.Circulation. 2007;116(3):238-248.21. Xie YP, Chen B, Sanders P, et al. Sildenafil pre-ventsand reverses transverse-tubuleremodelingandCa(2) handling dysfunction in right ventricle failurein d u ce d b y p u lmo n a ry a rte ry h y p e rte n sio n .Hypertension. 2012;59(2):355-362.22. Borgdorff MA, Bartelds B, Dickinson MG, et al.Sildenafil enhances systolic adaptation, but does notprevent diastolic dysfunction, in the pressure-loadedright ventricle. Eur J Heart Fail. 2012;14(9):1067-1074.

23. Nagayama T, Hsu S, Zhang M, et al. Pressure-overload magnitude-dependence of the anti-hypertrophic efficacy of PDE5A inhibition.J Mol CellCardiol. 2009;46(4):560-567.24. Takimoto E, Champion HC, Li M, et al. Chronicinhibition of cyclic GMP phosphodiesterase 5A pre-vents and reverses cardiac hypertrophy. Nat Med.2005;11(2):214-222.25. Lewis GD, Lachmann J, Camuso J, et al. Sildena-

fil improves exercise hemodynamics and oxygen up-takein patientswith systolic heartfailure. Circulation.2007;115(1):59-66.26. Guazzi M, Tumminello G, Di Marco F, FiorentiniC, Guazzi MD. Theeffects of phosphodiesterase-5 in-hibition with sildenafil on pulmonary hemodynamicsand diffusion capacity, exercise ventilatory effi-ciency, and oxygen uptake kinetics in chronicheart failure. J Am Coll Cardiol. 2004;44(12):2339-2348.27. Borlaug BA, Olson TP, Lam CS, et al. Global car-diovascular reserve dysfunction in heart failure withpreserved ejection fraction. J Am Coll Cardiol. 2010;56(11):845-854.28. Bishu K, Deswal A, Chen HH, et al. Biomarkersin acutely decompensated heart failure with pre-served or reducedejectionfraction. AmHeart J. 2012;164(5):763-770 e763.29. Tapia E,Sanchez-LozadaLG, Soto V, etal. Sildenafiltreatment prevents glomerular hypertension and hy-perfiltration in rats with renal ablation. Kidney BloodPress Res. 2012;35(4):273-280.30. Medeiros PJ, Villarim Neto A, Lima FP, AzevedoIM, Leao LR, Medeiros AC. Effect of sildenafil in re-nal ischemia/reperfusion injury in rats. Acta Cir Bras.2010;25(6):490-495.31. Choi DE, Jeong JY, Lim BJ, et al. Pretreatment ofsildenafil attenuates ischemia-reperfusion renal in-jur y in rat s. Am J Physiol Renal Physiol. 2009;297(2):F362-F370.



relax: inhibidor de la fosfodiesterasa-5. insuficiencia cardiaca con función ventricular preservada

Documents