cttcpawHaeatwbwfae

DCivLr

t0

0d

Comparison of Patients With Heart Failure and Preserved LeftVentricular Ejection Fraction Among Those With Versus Without

Diabetes Mellitus

David Aguilar, MDa,b,*, Anita Deswal, MD, MPHa,b,c, Kumudha Ramasubbu, MDa,b,Douglas L. Mann, MDd, and Biykem Bozkurt, MDa,b

Heart failure (HF) with preserved left ventricular ejection fraction (LVEF) and diabetescommonly coexist, but the impact of diabetes on HF outcomes in patients with HF andpreserved LVEF has not been well studied. We assessed the risk of HF death or hospital-ization for worsening HF associated with diabetes by studying 987 patients with HF andpreserved LVEF enrolled in the Digitalis Investigation Group (DIG) ancillary study.Diabetics (n � 285, 28.9%) were younger, had a larger body mass index, faster heart rate,and higher pulse pressure than nondiabetics. Diabetics were also more likely to be women,have a history of hypertension, ischemic cause for HF, and were more likely to be treatedwith diuretics. During the mean follow-up of 37 months, 88 (30.9%) diabetics and 133(19.0%) nondiabetics developed the primary outcome of HF hospitalization or HF death.After adjustments for baseline differences, diabetes was associated with a 68% increasedrisk of HF hospitalization or HF death (adjusted hazard ratio 1.68, 95% confidence interval1.26 to 2.25, p <0.001). In conclusion, in patients with HF and preserved LVEF, diabetesis associated with significantly increased risk of developing adverse HF outcomes. © 2010

Elsevier Inc. All rights reserved. (Am J Cardiol 2010;105:373–377)

M

tIaBidsgcr�twtwappHl

shwdeb

w

Diabetes mellitus is an important co-morbidity likelyontributing to adverse heart failure (HF) outcomes in pa-ients with HF and preserved left ventricular ejection frac-ion (LVEF). The prevalence of diabetes continues to in-rease in the general population1 and in community-basedatients with HF.2 Although diabetes has been shown to bepredictor of adverse cardiovascular outcomes in patientsith decreased LVEF,3 the impact of diabetes on adverseF outcomes has not been well studied in patients with HF

nd preserved LVEF. Given potential diabetes-associatedffects on vascular compliance4 and LV systolic stiffnessnd potential effects on LV diastolic function,5 processeshat have been implicated in the pathophysiology of HFith preserved LVEF, we hypothesized that diabetes woulde associated with increased rates of hospitalization fororsening HF or HF death in this group of patients. There-

ore, we examined the prognostic impact of diabetes ondverse HF outcomes in patients with preserved LVEFnrolled in the Digitalis Investigation Group (DIG) study.

aWinters Center for Heart Failure Research and Section of Cardiology,epartment of Medicine, Baylor College of Medicine, and bSection ofardiology and cHouston Center for Quality of Care and Utilization Stud-

es, Michael E. DeBakey VA Medical Center, Houston, Texas; and dDi-ision of Cardiology, Washington University School of Medicine, St.ouis, Missouri. Manuscript received July 18, 2009; revised manuscript

eceived and accepted September 8, 2009.This study was supported in part by a National Institutes of Health (Be-

hesda, Maryland), Mentored Career Development Award (5K01-HL092585-2) to Dr. Aguilar.

*Corresponding author: Tel: 713-798-0287; fax: 713-798-0270.

iE-mail address: daguilar@bcm.edu (D. Aguilar).

002-9149/10/$ – see front matter © 2010 Elsevier Inc. All rights reserved.oi:10.1016/j.amjcard.2009.09.041

ethods

The database used for this study was a public-use copy ofhe DIG study from the National Heart, Lung, and Bloodnstitute (Bethesda, Maryland). The design6 and primarynalyses7 of the DIG study have been described in detail.riefly, the DIG study tested the effects of effects of digoxin

n mortality and hospitalization in patients with HF. Theiagnosis of HF was based on current or previous clinicalymptoms, signs, or radiologic evidence of pulmonary con-estion. Important exclusion criteria included recent myo-ardial infarction or unstable angina (�4 weeks from en-ollment) and renal insufficiency (defined as creatinine level3.0 mg/dl). Patients were recruited from February 1991

hrough August 1993.7 Although the main results of the trialere published in the 6,800 patients with an LVEF �45%,7

he DIG study also included an ancillary trial of 988 patientsith HF and an LVEF �45%. The results of digoxin ther-

py on cardiovascular outcomes in the patients with HF andreserved LVEF of the DIG ancillary trial were recentlyublished.8 Our analyses are limited to these patients withF and LVEF �45% who were enrolled in the DIG ancil-

ary trial.Baseline demographics were obtained at entry into the

tudy. Classification of diabetes was based on a reportedistory of diabetes at study entry. Data on baseline diabetesere available in 987 participants; 1 patient did not haveata available for the classification of diabetes and wasxcluded from the study. Diabetic treatment or classificationetween type 1 and type 2 diabetes was not available.

The primary outcome was time to hospitalization fororsening HF or HF death. Secondary outcomes included

ndividual outcomes of times to hospitalization for wors-

www.AJConline.org

evciufcvaipovt

Fhospitalization for worsening HF by diabetic status (p �0.001, log-rank test).

TB

V

AWNBL

HS

D

B

S

CE

PIN

PCHPM

D

a

TC

V

H

H

H

T

C

C

dtt

374 The American Journal of Cardiology (www.AJConline.org)

ning HF, HF mortality, all-cause mortality and cardio-ascular mortality, and time to combined outcome ofardiovascular mortality and hospitalization for worsen-ng HF. Differences in baseline variables were comparedsing chi-square tests for categorical variables and t testsor continuous variables. Two-sided p values �0.05 wereonsidered statistically significant. Univariate and multi-ariable Cox proportional hazards models were used tossess the relation between diabetes and outcomes ofnterest. Covariates included variables identified as im-ortant predictors of the combined outcome of HF deathr hospitalization for worsening HF, and variables with palues �0.10 by univariate analyses were entered into

able 1aseline characteristics by diabetic status

ariable Diabetes Mellitus pValue

Yes(n � 285)

No(n � 702)

ge (years) 64.8 � 9.8 67.6 � 10.4 �0.001omen 48.8% 38.0% 0.002onwhite 16.5% 12.8% 0.13ody mass index (kg/m2) 30.8 � 7.3 27.7 � 5.5 �0.001eft ventricular ejection

fraction (%)55.1 � 7.9 55.6 � 8.2 0.37

eart rate (beats/min) 77.7 � 12.4 75.0 � 11.7 0.002ystolic blood pressure

(mm Hg)141.4 � 21.8 136.0 � 20.9 �0.001

iastolic blood pressure(mm Hg)

76.5 � 12.0 77.0 � 11.1 0.58

rachial pulse pressure(mm Hg)

64.8 � 18.5 59.0 � 18.1 �0.001

igns and/or symptoms ofheart failure

Rales 74.7% 73.2% 0.62Increased jugular venous

pressure50.5% 46.3% 0.23

Peripheral edema 69.5% 55.0% �0.001S3 31.9% 33.5% 0.64Radiologic evidence of

pulmonary congestion68.4% 60.0% 0.01

ardiothoracic ratio 0.52 � 0.08 0.52 � 0.08 0.50stimated glomerular filtration

rate (ml/min/1.73 m2)60.0 � 20.8 62.4 � 20.4 0.09

otassium (mmol/L) 4.4 � 0.4 4.3 � 0.5 0.007schemic heart failure cause 64.6% 53.1% 0.001ew York Heart Association

class III–IV24.2% 21.1% 0.29

revious myocardial infarction 54.4% 47.6% 0.05urrent angina 33.7% 28.2% 0.09ypertension 70.9% 55.1% �0.001revious digoxin use 32.0% 36.6% 0.17edication

Nonpotassium-sparingdiuretic

81.8% 73.7% 0.007

Potassium-sparing diuretic 7.4% 8.3% 0.64Angiotensin-converting

enzyme inhibitors88.8% 85.2% 0.14

Nitrates 44.6% 37.3% 0.04igoxin 47.4% 50.7% 0.34

Continuous variables are presented as means � SDs, categorical vari-bles as percentages.

he multivariable model. The final model included age, u

igure 1. Kaplan-Meier event curves for combined outcome of HF death or

able 2ardiovascular outcomes by diabetic status

ariable Diabetes Mellitus p Value

Yes No(n � 285) (n � 702)

eart failure death orhospitalization forworsening heart failure

88 (30.9%) 133 (19.0%)

Unadjusted hazard ratio(95% CI)

1.79 (1.37–2.35) reference �0.001

Adjusted hazard ratio(95% CI)*

1.68 (1.26–2.25) reference �0.001

ospitalization forworsening heart failure

81 (28.4%) 116 (16.5%)

Unadjusted hazard ratio(95% CI)

1.88 (1.42–2.50) reference �0.001

Adjusted hazard ratio(95% CI)

1.76 (1.30–2.39) reference �0.001

eart failure death 21 (7.4%) 43 (6.1%)Unadjusted hazard ratio

(95% CI)1.26 (0.75–2.12) reference 0.39

Adjusted hazard ratio(95% CI)

1.40 (0.78–2.50) reference 0.26

otal mortality 81 (28.4%) 150 (21.4%)Unadjusted hazard ratio

(95% CI)1.38 (1.06–1.81) reference 0.02

Adjusted hazard ratio(95% CI)

1.48 (1.10–1.99) reference 0.009

ardiovascular mortality 58 (20.3%) 104 (14.8%)Unadjusted hazard ratio

(95% CI)1.43 (1.03–1.97) reference 0.03

Adjusted hazard ratio(95% CI)

1.54 (1.08–2.18) reference 0.02

ardiovascular mortality orhospitalization forworsening heart failure

116 (40.7%) 180 (25.6%)

Unadjusted hazard ratio(95% CI)

1.76 (1.39–2.22) reference �0.001

Adjusted hazard ratio(95% CI)

1.69 (1.32–2.17) reference �0.001

* Model includes age, gender, LVEF, cardiothoracic ratio, heart rate,iastolic blood pressure, brachial pulse pressure, number of signs or symp-oms of heart failure, NYHA classification, history of myocardial infarc-ion, ischemic HF cause, previous digoxin use, nonpotassium-sparing di-

retic use, GFR, and BMI.

gbo(isba(watcpahifm9

R

imidwfnralp

ttwc2gbsdnjot

icmecrb0

i

tIcn

ictip5mo2ddhplaa(tyc

D

fciostptrumtwemfatrbdieaRwah

mo

375Heart Failure/Diabetes and HF With Preserved LVEF

ender, LVEF, cardiothoracic ratio, heart rate, diastoliclood pressure, brachial pulse pressure, number of signsr symptoms of HF, New York Heart AssociationNYHA) classification, history of myocardial infarction,schemic HF cause, previous digoxin use, nonpotassium-paring diuretic use, glomerular filtration rate (GFR), andody mass index (BMI). GFR was calculated using thebbreviated Modification of Diet in Renal DiseaseMDRD) Study Group equation.9 Brachial pulse pressureas calculated as the difference between brachial systolic

nd diastolic blood pressures. Tests for interaction be-ween diabetes and subgroups (gender, age, ischemicause of HF, obesity, and NYHA classification) wereerformed by adding a cross-product term of these vari-bles on the combined primary outcome of HF death orospitalization for worsening HF; the test for a statisticalnteraction between diabetes and subgroup were per-ormed before other variables being entered into theodel. Statistical analysis was performed using STATA

.2 (STATA Corp., College Station, Texas).

esults

Of the total 987 patients with HF and an LVEF �45%ncluded in this analysis, 285 patients (28.9%) had a docu-ented history of diabetes. Baseline characteristics accord-

ng to diabetic status are listed in Table 1. At baseline,iabetic patients were younger and were more likely to beomen, have a history of hypertension, have ischemic cause

or HF, and were more likely to be treated with diuretics anditrates. Diabetic patients also had a larger BMI, faster heartate, had a higher systolic blood pressure and pulse pressure,nd were more likely to have peripheral edema and radio-ogic evidence of pulmonary congestion than nondiabeticatients at baseline.

During the mean follow-up of 37 months, diabetic pa-ients were more likely than nondiabetic patients to develophe primary outcome of HF mortality or hospitalization fororsening HF (unadjusted hazard ratio [HR] 1.79, 95%

onfidence interval 1.37 to 2.35, p �0.001; Figure 1, Table). After adjusting for potential confounders, including age,ender, LVEF, cardiothoracic ratio, heart rate, diastoliclood pressure, brachial pulse pressure, number of signs orymptoms of HF, NYHA classification, history of myocar-ial infarction, ischemic HF cause, previous digoxin use,onpotassium-sparing diuretic use, GFR, and BMI, the ad-usted HR for the combined primary outcome of HF deathr hospitalization for worsening HF was 1.68 (95% CI 1.26o 2.25, p �0.001).

Other cardiovascular end points are listed in Table 2.To assess the potential confounder of interim myocardial

nfarction on HF outcomes, analyses were performed ex-luding patients (n � 54) who were hospitalized for acuteyocardial infarction during the follow-up period. After

xclusion of these patients, the hazard of the primary out-ome of HF mortality or hospitalization for worsening HFemained higher in diabetic patients compared to nondia-etic patients (adjusted HR 1.63, 95% CI 1.21 to 2.21, p �.002).

Hospitalization for worsening HF was also markedly

ncreased in diabetic patients compared to nondiabetic pa- c

ients (adjusted HR 1.76, 95% CI 1.30 to 2.39, p �0.001).n addition, hospitalization admissions for cardiovascularauses occurred in 155 diabetic patients (54.4%) and 311ondiabetic patients (44.3%, p �0.01).

Further exploratory analyses were performed examin-ng the potential interaction of diabetes and gender on theombined primary outcome of HF death or hospitaliza-ion for worsening HF. Prevalence of diabetes was highern women than in men (34.2% vs 25.1%, respectively,

�0.001). Of the 406 women enrolled in the study,6 diabetic women (40.3%) and 64 nondiabetic wo-en (24.0%) developed the primary outcome of HF death

r HF hospitalization (adjusted HR 1.94, 95% CI 1.31 to.87, p �0.01). Similarly, in the 581 men in the study, 32iabetic men (21.9%) and 69 nondiabetic men (15.9%)eveloped the primary outcome of HF death or HFospitalization (adjusted 1.37, 95% CI 0.87 to 2.15,� 0.18). Although the HR associated with diabetes was

arger in women than in men, the interaction of gendernd diabetes was not statically significant (p for inter-ction � 0.33). Additional exploratory subgroup analysesFigure 2) did not demonstrate a statistical interac-ion between diabetes and subgroups of age �60ears, ischemic HF cause, obesity, and NYHA classifi-ation.

iscussion

Previous studies have established diabetes as a risk factoror development of adverse events in patients with de-reased LVEF3; however, only 1 previous study has exam-ned the prognostic impact of diabetes on cardiovascularutcomes in patients with HF and preserved LVEF.10 In thistudy of patients with HF and preserved LVEF enrolled inhe DIG ancillary study, we demonstrate that diabetes wasresent in approximately 30% of patients. At baseline, pa-ients with diabetes had increased rates of peripheral edema,adiologic evidence of pulmonary congestion, and diureticse supporting the hypothesis that diabetic patients may beore prone to problems of congestion and volume reten-

ion. Importantly, the presence of diabetes was associatedith significantly increased risk of hospitalization for wors-

ning HF or death attributed to HF. Increased morbidity andortality persisted even after adjustments for potential con-

ounders such as hypertension, obesity, renal insufficiency,nd ischemic heart disease. Although a previous analysis ofhis cohort has demonstrated that diabetes is an independentisk factor for mortality,11 our study adds to the existingody of literature regarding HF with preserved LVEF byemonstrating that diabetes is associated with a nearly 70%ncreased hazard of HF death or hospitalization for wors-ning HF. This hazard is similar in magnitude to a recentnalysis of the Candesartan in Heart Failure–Assessment ofeduction in Mortality and Morbidity (CHARM) program,hich demonstrated that presence of diabetes was associ-

ted with a doubling in risk of cardiovascular death or HFospitalization.10

There are several mechanisms through which diabetesay be contributing to the pathophysiology and adverse

utcomes in patients with HF and preserved LVEF. Asso-

iated co-morbid conditions and complications such as hy-

pctaniqlcvmdsTdslHvyi

a(apssatcpruaLr

nftaowiicpLchTiu

me of H

376 The American Journal of Cardiology (www.AJConline.org)

ertension, accelerated atherosclerosis, and renal insuffi-iency are likely contributing to the adverse outcomes, buthe increased hazard seen with diabetes persists despitedjusting for these co-morbidities. One important mecha-ism linking diabetes and HF with preserved LVEF may bencreased arterial load in patients with diabetes and subse-uent matched increases in LV systolic stiffness (ventricu-ar–arterial coupling).12–16 Although ventricular–arterialoupling appears beneficial in maintaining optimal cardio-ascular performance,13,14 marked increases of arterial loadatched with increased LV systolic stiffness may have

etrimental cardiovascular effects, including increased sen-itivity to volume shifts and decreased exercise capacity.12

his process may be particularly relevant to patients withiabetes, because studies have suggested that diabetes re-ults in increased arterial stiffness,4 a process likely re-ated to effects of advanced glycation end products.igher brachial pulse pressure, an indirect surrogate ofascular compliance, in diabetic participants of this anal-sis is consistent with a potential pathogenic role ofncreased vascular stiffness.

In addition to abnormalities of LV systolic stiffness andrterial stiffness, abnormalities of LV diastolic functionimpaired LV relaxation and increased passive LV stiffness)re likely contributing to the greater HF burden in diabeticatients with HF and preserved LVEF. Previous noninva-ive studies of diabetic patients without HF have demon-trated that abnormalities of diastolic function are commonnd are independent of LV mass, systolic function, hyper-ension, or other co-morbidities.5 Potential mechanismsontributing to increased diastolic abnormalities in diabeticatients include abnormalities of myocardial fibrosis,17 di-ect and indirect effects of advanced glycation end prod-cts,18 abnormal myocardial metabolism,19 and increasedrterial stiffness.20 Importantly, increased abnormalities ofV diastolic function have been shown to play an important

Figure 2. Effect of diabetes on adjusted HRs for outco

ole in the pathophysiology of HF with preserved LVEF.16,21,22

Several limitations of this exploratory study should beoted. Participants for the DIG ancillary trial were recruitedrom 1991 to 1993. Although the prevalence of diabetes inhis clinical trial was approximately 29% and was similar torecent analysis from the CHARM study,10 the prevalencef diabetes has continued to increase in community patientsith HF over time,2 and prevalence estimates from a clin-

cal trial may underestimate the true prevalence of diabetesn patients with HF and preserved LVEF in the generalommunity. Indeed, in a contemporary registry of patientsresenting with acute decompensated HF and preservedVEF, prevalence of diabetes was 45%.23 In addition, gly-ometabolic measurements, such as glucose or glycosylatedemoglobin, or specific diabetic therapy were not available.hus, degree of glycemic control in the diabetic population

n the DIG trial could not be assessed. Rates of �-blockerse were not available in this study.

1. Mokdad AH, Ford ES, Bowman BA, Dietz WH, Vinicor F, Bales VS,Marks JS. Prevalence of obesity, diabetes, and obesity-related healthrisk factors, 2001. JAMA 2003;289:76–79.

2. From AM, Leibson CL, Bursi F, Redfield MM, Weston SA, Jacob-sen SJ, Rodeheffer RJ, Roger VL. Diabetes in heart failure: prev-alence and impact on outcome in the population. Am J Med 2006;119:591–599.

3. De Groote P, Lamblin N, Mouquet F, Plichon D, McFadden E, VanBelle E, Bauters C. Impact of diabetes mellitus on long-term survivalin patients with congestive heart failure. Eur Heart J 2004;25:656–662.

4. Cruickshank K, Riste L, Anderson SG, Wright JS, Dunn G, GoslingRG. Aortic pulse-wave velocity and its relationship to mortality indiabetes and glucose intolerance: an integrated index of vascular func-tion? Circulation 2002;106:2085–2090.

5. Liu JE, Palmieri V, Roman MJ, Bella JN, Fabsitz R, Howard BV,Welty TK, Lee ET, Devereux RB. The impact of diabetes on leftventricular filling pattern in normotensive and hypertensive adults: theStrong Heart Study. J Am Coll Cardiol 2001;37:1943–1949.

6. Rationale, design, implementation, and baseline characteristics of pa-tients in the DIG trial: a large, simple, long-term trial to evaluate the

F death or HF hospitalization in selected subgroups.

effect of digitalis on mortality in heart failure. Control Clin Trials1996;17:77–97.

1

1

1

1

1

1

1

1

1

1

2

2

2

2

377Heart Failure/Diabetes and HF With Preserved LVEF

7. The Digitalis Investigation Group. The effect of digoxin on mortalityand morbidity in patients with heart failure. N Engl J Med 1997;336:525–533.

8. Ahmed A, Rich MW, Fleg JL, Zile MR, Young JB, Kitzman DW, LoveTE, Aronow WS, Adams KF Jr, Gheorghiade M. Effects of digoxin onmorbidity and mortality in diastolic heart failure: the ancillary digitalisinvestigation group trial. Circulation 2006;114:397–403.

9. Levey AS, Coresh J, Greene T, Stevens LA, Zhang YL, Hendriksen S,Kusek JW, Van Lente F. Using standardized serum creatinine valuesin the modification of diet in renal disease study equation for estimat-ing glomerular filtration rate. Ann Intern Med 2006;145:247–254.

0. Macdonald MR, Petrie MC, Varyani F, Ostergren J, Michelson EL,Young JB, Solomon SD, Granger CB, Swedberg K, Yusuf S, PfefferMA, McMurray JJ. Impact of diabetes on outcomes in patients withlow and preserved ejection fraction heart failure: an analysis of theCandesartan in Heart failure: Assessment of Reduction in Mortalityand Morbidity (CHARM) programme. Eur Heart J 2008;29:1377–1385.

1. Jones RC, Francis GS, Lauer MS. Predictors of mortality in patientswith heart failure and preserved systolic function in the DigitalisInvestigation Group trial. J Am Coll Cardiol 2004;44:1025–1029.

2. Kawaguchi M, Hay I, Fetics B, Kass DA. Combined ventricularsystolic and arterial stiffening in patients with heart failure and pre-served ejection fraction: implications for systolic and diastolic reservelimitations. Circulation 2003;107:714–720.

3. Chantler PD, Lakatta EG, Najjar SS. Arterial-ventricular coupling:mechanistic insights into cardiovascular performance at rest and dur-ing exercise. J Appl Physiol 2008;105:1342–1351.

4. De Tombe PP, Jones S, Burkhoff D, Hunter WC, Kass DA. Ventric-ular stroke work and efficiency both remain nearly optimal despitealtered vascular loading. Am J Physiol Heart Circ Physiol 1993;264:H1817–H1824.

5. Borlaug BA, Lam CS, Roger VL, Rodeheffer RJ, Redfield MM.

Contractility and ventricular systolic stiffening in hypertensive heart

disease insights into the pathogenesis of heart failure with preservedejection fraction. J Am Coll Cardiol 2009;54:410–418.

6. Lam CS, Roger VL, Rodeheffer RJ, Bursi F, Borlaug BA, Ommen SR,Kass DA, Redfield MM. Cardiac structure and ventricular-vascularfunction in persons with heart failure and preserved ejection fractionfrom Olmsted County, Minnesota. Circulation 2007;115:1982–1990.

7. Asbun J, Villarreal FJ. The pathogenesis of myocardial fibrosis in thesetting of diabetic cardiomyopathy. J Am Coll Cardiol 2006;47:693–700.

8. Goldin A, Beckman JA, Schmidt AM, Creager MA. Advanced glyca-tion end products: sparking the development of diabetic vascularinjury. Circulation 2006;114:597–605.

9. Diamant M, Lamb HJ, Groeneveld Y, Endert EL, Smit JW, Bax JJ,Romijn JA, de Roos A, Radder JK. Diastolic dysfunction is associatedwith altered myocardial metabolism in asymptomatic normotensivepatients with well-controlled type 2 diabetes mellitus. J Am CollCardiol 2003;42:328–335.

0. Sharman JE, Haluska BA, Fang ZY, Prins JB, Marwick TH. Associ-ation of arterial wave properties and diastolic dysfunction in patientswith type 2 diabetes mellitus. Am J Cardiol 2007;99:844–888.

1. Zile MR, Baicu CF, Gaasch WH. Diastolic heart failure—abnormali-ties in active relaxation and passive stiffness of the left ventricle.N Engl J Med 2004;350:1953–1959.

2. Westermann D, Kasner M, Steendijk P, Spillmann F, Riad A, Weit-mann K, Hoffmann W, Poller W, Pauschinger M, Schultheiss HP,Tschope C. Role of left ventricular stiffness in heart failure withnormal ejection fraction. Circulation 2008;117:2051–2060.

3. Yancy CW, Lopatin M, Stevenson LW, De Marco T, Fonarow GC.Clinical presentation, management, and in-hospital outcomes of pa-tients admitted with acute decompensated heart failure with preservedsystolic function: a report from the Acute Decompensated Heart Fail-ure National Registry (ADHERE) Database. J Am Coll Cardiol 2006;

47:76–84.