J A C C : C A R D I O V A S C U L A R I M A G I N G V O L . 1 0 , N O . 1 0 , 2 0 1 7

ª 2 0 1 7 B Y T H E AM E R I C A N C O L L E G E O F C A R D I O L O G Y F O U N D A T I O N

P U B L I S H E D B Y E L S E V I E R

I S S N 1 9 3 6 - 8 7 8 X / $ 3 6 . 0 0

h t t p : / / d x . d o i . o r g / 1 0 . 1 0 1 6 / j . j c m g . 2 0 1 7 . 0 6 . 0 1 3

The Prevalence, Correlates, andImpact on Cardiac Mortality ofRight Ventricular Dysfunction inNonischemic Cardiomyopathy

Andreas Pueschner,a Pairoj Chattranukulchai, MD, MSC,a,b John F. Heitner, MD,c Dipan J. Shah, MD,d

Brenda Hayes, BS,a Wolfgang Rehwald, PHD,a Michele A. Parker, MS,a Han W. Kim, MD,a,e Robert M. Judd, PHD,a,f

Raymond J. Kim, MD,a,e,f Igor Klem, MDa,e

ABSTRACT

Fro

ofcD

De

Ca

sup

gra

Ma

OBJECTIVES This study sought to determine the prevalence, correlates, and impact on cardiac mortality of right

ventricular (RV) dysfunction in nonischemic cardiomyopathy.

BACKGROUND Current heart failure guidelines place little emphasis on RV assessment due to limited available data on

determinants of RV function, mechanisms leading to its failure, and relation to outcomes.

METHODS We prospectively studied 423 patients with cardiac magnetic resonance (CMR). The pre-specified study

endpoint was cardiac mortality. In 100 patients, right heart catheterization was performed as clinically indicated.

RESULTS During a median follow-up time of 6.2 years (interquartile range: 2.9 to 7.6 years), 101 patients (24%) died of

cardiac causes. CMR right ventricular ejection fraction (RVEF) was a strong independent predictor of cardiac mortality

after adjustment for age, heart failure–functional class, blood pressure, heart rate, serum sodium, serum creatinine,

myocardial scar, and left ventricular ejection fraction (LVEF). Patients with the lowest quintile of RVEF had a nearly 5-fold

higher cardiac mortality risk than did patients with the highest quintile (hazard ratio: 4.68; 95% confidence interval

[CI]: 2.43 to 9.02; p < 0.0001). RVEF was positively correlated with LVEF (r ¼ 0.60; p < 0.0001), and inversely

correlated with right atrial pressure (r ¼ �0.32; p ¼ 0.001), pulmonary artery pressure (r ¼ �0.34; p ¼ 0.0005),

transpulmonary gradient (r ¼ �0.28; p ¼ 0.006) but not with pulmonary wedge pressure (r ¼ �0.15; p ¼ 0.13). In

multivariable logistic regression analysis of CMR, clinical, and hemodynamic data the strongest predictors of right

ventricular dysfunction were LVEF (odds ratio [OR]: 0.85; 95% CI: 0.78 to 0.92; p < 0.0001), transpulmonary gradient

(OR: 1.20; 95% CI: 1.09 to 1.32; p ¼ 0.0003), and systolic blood pressure (OR: 0.97; 95% CI: 0.94 to 0.99; p ¼ 0.02).

CONCLUSIONS CMR assessment of RVEF provides important prognostic information independent of established risk

factors and LVEF in heart failure patients with nonischemic cardiomyopathy. Right ventricular dysfunction is strongly asso-

ciated with both indices of intrinsic myocardial contractility and increased afterload from pulmonary vascular dysfunction.

(J Am Coll Cardiol Img 2017;10:1225–36) © 2017 by the American College of Cardiology Foundation.

H eart failure (HF) is an important cause ofmorbidity and mortality in developedcountries, with an estimated 5.7 million

people affected in the United States and 5-year

m the aDuke Cardiovascular Magnetic Resonance Center, Duke University

Cardiology, Department of Medicine, Chulalongkorn University, King Ch

epartment of Cardiology, New York Methodist Hospital, Brooklyn, N

Bakey Heart and Vascular Center, Houston, Texas; eDivision of Cardiolog

rolina; and the fDepartment of Radiology, Duke University Medical

ported in part by Medtronic Inc. (to Dr. Klem). Dr. Rehwald is an employee of

nt fromMedtronic. All other authors have reported that they have no relations

nuscript received April 5, 2017; revised manuscript received June 19, 201

mortality reaching 50% (1). The focus of HF manage-ment has been on left ventricular (LV) systolic func-tion, which is the principal criterion forclassification of HF into preserved and reduced

Medical Center, Durham, North Carolina; bDivision

ulalongkorn Memorial Hospital, Bangkok, Thailand;

ew York; dDepartment of Cardiology, Methodist

y, Duke University Medical Center, Durham, North

Center, Durham, North Carolina. This research is

Siemens Healthineers. Dr. Klem has received a research

hips relevant to the contents of this paper to disclose.

7, accepted June 26, 2017.

ABBR EV I A T I ON S

AND ACRONYMS

CI = confidence interval(s)

CMR = cardiac magnetic

resonance

DE-CMR = delayed-

enhancement cardiac magnetic

resonance

HF = heart failure

HR = hazard ratio(s)

IQR = interquartile range

LV = left ventricle/ventricular

LVEF = left ventricular ejection

fraction

NICM = nonischemic

cardiomyopathy

NYHA = New York Heart

Association

OR = odds ratio(s)

PAP = pulmonary artery

pressure

PH = pulmonary hypertension

PCWP = pulmonary capillary

wedge pressure

RAP = right atrial pressure

RHC = right heart

catheterization

RV = right ventricle/ventricular

RVD = right ventricular

dysfunction

RVEF = right ventricular

ejection fraction

TPG = transpulmonary gradient

Pueschner et al. J A C C : C A R D I O V A S C U L A R I M A G I N G , V O L . 1 0 , N O . 1 0 , 2 0 1 7

RVD in NICM O C T O B E R 2 0 1 7 : 1 2 2 5 – 3 6

1226

ejection fraction in recent guidelines (2).There are scant diagnostic or therapeuticrecommendations based on right ventricular(RV) assessment, mainly due to limited dataon determinants of RV function, mechanismsleading to its failure, and relation to outcomesas ascertained by a recent National Heart,Lung, and Blood Institute task force report(2,3). Prior studies have suggested that RVdysfunction (RVD) may be an important prog-nostic marker in patients with left-sided HF(4–11). Most of these studies though had arelatively small sample size and includedheterogeneous patient populations in termsof etiology of cardiomyopathy and left ven-tricular ejection fraction (LVEF), or invasivemethods for RV assessment were used,limiting the clinical applicability (4,5,12–15).

Several factors may affect RV systolicfunction. RV is very sensitive to afterload,with RV stroke volume decreasing rapidlywhen pulmonary artery pressure is increasedin experimental studies (16). In pulmonaryhypertension (PH) due to left heart disease,RV afterload may be elevated either due toelevated pulmonary capillary wedge pressure(isolated post-capillary PH), or progression toincreased transpulmonary gradient indi-cating a pulmonary vascular process (com-bined post-capillary and pre-capillary PH)(17,18). Within physiologic limits, increase inventricular pre-load can improve RV systolicfunction based on the Frank-Starling mech-

anism (19). Conversely, intrinsic myocardial contrac-tility may be impaired from various cardiomyopathicprocesses and affect RV systolic function indepen-dent of the loading conditions.

SEE PAGE 1237

This study had 3 specific aims. First, we wanted toexplore the determinants of RVD, specifically whetherright ventricular ejection fraction (RVEF) is morereflective of abnormal loading conditions or indices ofintrinsic myocardial contractility (e.g., LVEF). RV andLV systolic function were assessed using cardiacmagnetic resonance (CMR), which allows accurate andreproducible measurements of both ventricles (20).Second, we wanted to investigate the prevalence of RVDin HF patients with reduced LVEF and, third, to assessthe impact of RVD on long-term cardiac mortality.

METHODS

POPULATION. We prospectively enrolled 423consecutive patients evaluated by the HF service

from January 1, 2003 to June 30, 2008 who haddilated cardiomyopathy based on American HeartAssociation definitions (2) and LVEF <50%. Weexcluded patients with ischemic heart disease, whichwas defined as stenosis $50% of left main or $70% ofmajor coronary artery, or insignificant coronary arterydisease with definite evidence of myocardial infarc-tion: for example, ruptured plaque or acute thrombuson coronary angiography, if no angiography wasperformed, the absence of definite, enzymaticallyproven type 1 myocardial infarction (21,22); primaryvalvular or congenital heart disease; constrictivepericarditis; known hypertrophic or arrhythmogenicRV cardiomyopathy; acute inflammatory or infiltra-tive heart disease; and those without a reasonableexpectation of survival for at least 1 year or who werescheduled for major cardiothoracic surgery. Thestudy protocol was approved by the Duke Institu-tional Review Board. All patients gave writteninformed consent.

PROTOCOL. A comprehensive medical historyincluding comorbidities, New York Heart Association(NYHA) functional class, and medications at the timeof enrolment was obtained. Additionally, 12-leadelectrocardiography was performed a median of 1.0days (interquartile range [IQR]: 0 to 4 days) from CMRand was interpreted blinded to clinical and CMR datafor QRS and QTc intervals and bundle branch blockwith Minnesota code (23).

Cardiac magnetic resonance. All patients underwentCMR imaging on a 1.5-T or 3.0-T CMR scanner(Siemens, Erlangen, Germany) as part of clinical HFmanagement for assessment of LVEF and myocardialscar as previously described (24). Briefly, steady-state-free precession cine images were acquired inmultiple short-axis and 3 long-axis views. Delayed-enhancement CMR (DE-CMR) was performed usinga segmented inversion-recovery gradient-echotechnique in the identical views as cine-CMR 10 minafter contrast administration (gadoversetamide,0.15 mmol/kg).

Cine and DE-CMR images were evaluated separately,masked to all patient information. The analysis of RVvolumes was performed by determining end-diastolicand end-systolic frames on short-axis cine views, andsemiautomated thresholding was used to delineateendocardial borders. The systolic descent of thetricuspid valve plane was tracked on long-axis views tocorrectly identify the ventricular volume in the basalshort-axis slice. Papillary muscles and trabeculationswere excluded from the RV volumes. Reproducibility ofRV measurements was tested in 20 randomly selectedpatients. To test the intraobserver variability, the same

J A C C : C A R D I O V A S C U L A R I M A G I N G , V O L . 1 0 , N O . 1 0 , 2 0 1 7 Pueschner et al.O C T O B E R 2 0 1 7 : 1 2 2 5 – 3 6 RVD in NICM

1227

observer measured the same group of patients with3-month time separation. For assessment of interob-server variability, a second reader, blinded to the resultsof the first reader, performed RV measurements in thesame patients. LV volumes, LV mass, and LVEF werequantitatively measured from the stack of short-axiscine images using standard techniques (24). LV and RVvolumes were indexed to body surface area. Thepresence of hyperenhanced tissue on DE-CMR,which was interpreted as representing myocardialreplacement fibrosis (24), was determined by visualinspection using the American Heart Association17-segment model (25). The extent of scar as percentageof LV mass was estimated based on a 5-point scale aspreviously described (26).

Right heart catheterization. In 100 patients (24%) rightheart catheterization (RHC) was performed as part ofclinical HF management (for assessment of volumestatus in 62 patients, in 17 refractory to HF therapy, in8 due to hypotension, and in 13 for heart transplantevaluation) within a median of 0 days (IQR: �2 to þ1days) of CMR according to standard protocols. Pa-tients were in chronic stable condition at the time ofCMR and RHC, without intervening events betweenthe 2 tests. During the procedure mean right atrialpressure (RAP), pulmonary artery pressure (PAP),mean pulmonary capillary wedge pressure (PCWP),and systemic arterial blood pressure were measured(27). Cardiac index, total pulmonary resistance, sys-temic vascular resistance, and pulmonary vascularresistance were calculated. A mean-PAP >25 mm Hgwas used to define PH. The transpulmonary gradient(TPG), which is calculated as mean PAP � meanPCWP, was used to differentiate increased RV after-load due to pulmonary vascular disease (TPG >12mm Hg) from increased downstream left heart pres-sures (TPG #12 mm Hg) (17).

Follow-up. Information concerning mortality statuswas obtained at regular intervals via: 1) telephoneinterview with the patient, or if deceased, with familymembers; 2) contact with the patient’s physician(s);and 3) hospital records. Mortality status was verifiedindependently through the Social Security DeathIndex and death certificates. Cause of death wasdetermined based on a review of death certificates,post-mortem exam reports when available, medicalrecords for patients who died while hospitalized, andcontact of patient’s physician(s), and it was classifiedas cardiac or noncardiac death (28). All event infor-mation was obtained and classified without knowl-edge of clinical, hemodynamic, or CMR findings.

STATISTICAL ANALYSIS. Continuous data areexpressed as mean � SD or median (IQR) as

appropriate. Comparisons between groups were madeusing 2-sample Student t or Wilcoxon rank sum testsfor continuous data, and chi-square tests for discretedata.

Pearson correlation analysis was used to examinerelationships between hemodynamic parameters andRVEF. Logistic regression analysis was used to iden-tify determinants of RVD, and clinically relevantvariables were selected a priori for model entry (9,29).To assess the importance of nonhemodynamic vari-ables separated from the loading condition, a secondanalysis adjusted for mean PAP was performed. Threemultivariable models were developed: first, todifferentiate myocardial contractility from hemody-namic loading, LVEF, mean PAP, and mean RAP wereconsidered simultaneously; second, mean PAP wassubstituted by TPG and mean PCWP in the samemodel; and finally, the best model was identifiedamong all univariable predictors by a stepwise se-lection process.

The pre-specified endpoint was cardiac death.Patients who underwent heart transplantation duringfollow-up were censored on the date of surgery. Cu-mulative event rates were calculated according to theKaplan-Meier method. Differences in event rates be-tween groups were assessed with the log-rank test,with Bonferroni adjustment for multiple pairwisecomparisons. Using Cox proportional hazards regres-sion analysis, 2 multivariable models were developed.First, RVEF and established predictors of mortality inHF patients, reported in at least 2 studies cited in the2013 Heart Failure Guideline Data Supplements, DataSupplement 3, were included (2). Myocardial fibrosis,which was shown to be an important predictor ofmortality in nonischemic cardiomyopathy (NICM) af-ter the guideline document publication date was alsoincluded (11). Second, the 10 most significant pre-dictors by univariable analysis were included ascandidate variables in a stepwise selection process. Toavoid the potential for overfitting, no more than 10variables were included in any multivariable model.Relative risks were expressed as hazard ratios (HRs)with associated 95% confidence intervals (CIs).Reclassification of patient risk was determined usingnet reclassification improvement analyses. Statisticalanalysis was performed using SAS version 9.3 (SASInstitute, Cary, North Carolina).

RESULTS

BASELINE CHARACTERISTICS. The baseline charac-teristics of all 423 patients are listed in Table 1. Themedian age was 54 years, about one-half were maleand of non-white ethnicity. HF symptoms with NYHA

TABLE 1 Baseline Patient Characteristics

All Patients(N ¼ 423)

Cardiac Death(n ¼ 101)

No Cardiac Death(n ¼ 322) p Value

Age, yrs 54.4 (45.8–65.8) 58.0 (47.6–72.9) 53.9 (45.5–64.4) 0.02

Male 222 (52) 63 (62) 159 (49) 0.02

Non-white race 225 (53) 59 (58) 166 (52) 0.23

Clinical history

Diabetes mellitus 106 (25) 34 (34) 72 (22) 0.02

Hypertension 239 (57) 63 (62) 176 (55) 0.17

Cigarette smoker 93 (22) 29 (29) 64 (20) 0.06

Hyperlipidemia 190 (45) 44 (44) 146 (45) 0.75

Family history of CAD 96 (23) 25 (25) 71 (22) 0.57

NYHA functional class* 0.01

I 120 (28) 20 (20) 100 (31)

II 109 (26) 29 (29) 80 (25)

III 162 (38) 45 (45) 117 (36)

IV 13 (3) 6 (6) 7 (2)

Mean NYHA functional class 2.1 � 1.0 2.3 � 0.9 2.0 � 1.0 0.001

Duration of NICM, months 24 � 40 35 � 58 20 � 41 0.02

Atrial fibrillation or flutter 73 (17) 25 (25) 48 (15) 0.02

Nonsustained ventricular tachycardia 57 (13) 13 (13) 44 (14) 0.84

Syncope 33 (8) 6 (6) 27 (8) 0.42

Body mass index 28.7 � 6.8 28.3 � 6.8 28.8 � 6.7 0.51

Systolic blood pressure, mm Hg 126.0 � 23.4 121.2 � 23.1 127.5 � 23.4 0.02

Diastolic blood pressure, mm Hg 75.8 � 15.7 75.0 � 15.4 76.1 � 15.8 0.55

Heart rate, beats/min 78 (68–90) 80 (66–93) 78 (69–90) 0.69

Laboratory results

Serum sodium, mmol/l† 139 (137–141) 139 (136–141) 139 (137–141) 0.22

Serum creatinine, mg/dl† 1.1 (0.9–1.5) 1.4 (1.1–1.9) 1.1 (0.9–1.4) <0.0001

Serum proBNP, pg/ml‡ 2,154 (699–5,772) 4,748 (1,859–9,703) 1,374 (525–5,112) <0.0001

Medications

ACE inhibitor 293 (69) 68 (67) 225 (70) 0.63

ARB 41 (10) 11 (11) 30 (9) 0.64

Antiarrhythmic drug 41 (10) 15 (15) 26 (8) 0.04

Aspirin 263 (62) 66 (65) 197 (61) 0.45

Beta-blocker 332 (78) 79 (78) 253 (79) 0.94

Calcium-channel blocker 54 (13) 15 (15) 39 (12) 0.47

Digitalis 76 (18) 29 (29) 47 (15) 0.001

Diuretics§ 290 (69) 80 (79) 210 (65) 0.008

Insulin 56 (13) 20 (20) 36 (11) 0.03

Oral antidiabetic 50 (12) 12 (12) 38 (12) 0.98

Nitrates 52 (12) 16 (16) 36 (11) 0.21

Statin 144 (34) 36 (36) 108 (34) 0.70

Warfarin 100 (24) 35 (35) 65 (20) 0.003

Continued on the next page

Pueschner et al. J A C C : C A R D I O V A S C U L A R I M A G I N G , V O L . 1 0 , N O . 1 0 , 2 0 1 7

RVD in NICM O C T O B E R 2 0 1 7 : 1 2 2 5 – 3 6

1228

functional class II or higher were present in 284patients (67%). Over 75% of patients were treatedwith beta-blockers and ACE-angiotensin-convertingenzyme inhibitors/angiotensin-receptor blockers,among those with LVEF #40%, 98% were on bothmedications. LVEF was significantly reduced with33 � 12%, the mean RVEF was 47 � 13%. The meandifference (and 95% limits of agreement) for theintraobserver and interobserver reproducibilityanalysis was 0.2% (�3.8 to 3.4), and 0.3% (�7.6 to8.1), respectively.

The 100 patients who underwent RHC were me-dian age 51.1 years (IQR: 40.5 to 59.5 years), 54 (54%)were male, had a mean NYHA functional class of2.2 � 1.0, and mean LVEF of 29 � 12%. The meanRVEF was 41 � 14% compared with 49 � 12% in thegroup not referred to RHC (p < 0.0001). The hemo-dynamic results are shown in Table 2. The medianPCWP was 20 mm Hg (IQR: 12 to 27 mm Hg), andmedian RAP 10 mm Hg (IQR: 5 to 14 mm Hg). Themedian PAP was 33 mm Hg (IQR: 25 to 40 mm Hg).PH was present in 72 patients, 36 (50%) had isolated

TABLE 1 Continued

All Patients(N ¼ 423)

Cardiac Death(n ¼ 101)

No Cardiac Death(n ¼ 322) p Value

ElectrocardiogramkHeart rate, beats/min 79 (68–89) 84 (68–97) 79 (68–89) 0.29

QRS, ms 107.8 � 29.1 111.7 � 30.5 106.6 � 28.6 0.13

QTc, ms 460.5 � 45.1 468.4 � 46.0 458.0 � 44.6 0.04

Left bundle branch block¶ 76 (18) 23 (23) 53 (17) 0.16

Right bundle branch block# 21 (5) 4 (4) 17 (5) 0.58

CMR

RVEF, % 47.2 � 12.9 42.4 � 13.9 48.7 � 12.2 <0.0001

RVEDV index, ml/m2 77.3 � 29.7 87.6 � 34.9 74.1 � 27.2 0.0005

RVESV index, ml/m2 43.1 � 25.2 53.7 � 31.4 39.7 � 22.0 <0.0001

LVEF, % 32.9 � 11.6 29.0 � 12.1 34.2 � 11.2 <0.0001

LVEDV index, ml/m2 114.8 � 45.0 127.3 � 49.8 110.9 � 42.7 0.003

LVESV index, ml/m2 80.4 � 43.1 94.5 � 48.4 76.0 � 40.3 0.0007

LV mass, g 204.4 � 71.2 223.4 � 73.6 198.4 � 69.5 0.002

Any scar on DE-CMR 215 (51) 68 (67) 147 (46) 0.0001

Scar extent, %LV mass 3.7 � 6.5 5.4 � 8.4 3.2 � 5.7 0.02

Values are median (interquartile range), n (%), or mean � SD. Bold values are statistically significant. *NYHA functional class was documented at the time of study enrollment.In 19 patients (5%) no heart failure symptoms were present. The p value pertains to the comparison between the groups with and without events in the distribution of patientsaccording to NYHA functional class. †Available in 416 patients. ‡Available in 273 patients. §Including aldosterone blockers. kAvailable in 412 patients. ¶Minnesota codes 7-1-1and 7-1-2. #Minnesota codes 7-2-1 and 7-2-2.

ACE ¼ angiotensin-converting enzyme; ARB ¼ angiotensin-receptor blocker; CAD ¼ coronary artery disease; CMR ¼ cardiac magnetic resonance; DE-CMR ¼ delayed-enhancement cardiac magnetic resonance; EDV ¼ end-diastolic volume; EF ¼ ejection fraction; ESV ¼ end-systolic volume; LV ¼ left ventricular; NICM ¼ nonischemiccardiomyopathy; NYHA ¼ New York Heart Association; proBNP ¼ pro–B-type natriuretic peptide; RV ¼ right ventricular.

J A C C : C A R D I O V A S C U L A R I M A G I N G , V O L . 1 0 , N O . 1 0 , 2 0 1 7 Pueschner et al.O C T O B E R 2 0 1 7 : 1 2 2 5 – 3 6 RVD in NICM

1229

post-capillary and 36 (50%) had combined post-capillary/pre-capillary PH.FOLLOW-UP RESULTS. During a median follow-uptime of 6.2 years (IQR: 2.9 to 7.6 years) 170 patients(40%) died; of those deaths, 101 (24%) were cardiacdeaths. Fifteen patients (4%) underwent hearttransplantation after study enrollment. No patientwas lost to follow-up. In the subgroup that underwent

TABLE 2 Hemodynamic Results

All Patients*(N ¼ 100)

C

Heart rate, beats/min 88 (76–100)

Cardiac index, l/min/m2 2.3 (1.8–2.7) 2

RAP mean, mm Hg 10 (5–14)

PAP systolic, mm Hg 47 (34–55)

PAP diastolic, mm Hg 23 (15–28)

PAP mean, mm Hg 33 (25–40)

PCWP mean, mm Hg 20 (12–27)

TPG, mm Hg 11 (7–16)

TPR, Wood units 7.6 (4.8–10.4) 7

PVR, Wood units 2.2 (1.5–4.0) 2

SVR, Wood units 19.1 (15.5–23.8) 16

BP systolic, mm Hg 120 (105–140) 1

BP diastolic, mm Hg 76 (66–85)

Values are median (IQR). Bold values are statistically significant. *RHC performed in 100

BP ¼ blood pressure; CMR ¼ cardiac magnetic resonance; IQR interquartile range; PApulmonary vascular resistance; RAP ¼ right atrial pressure; RHC ¼ right heart catheterizapulmonary resistance.

RHC, the cardiac mortality rate was higher than forthe patients who did not undergo RHC (28% vs. 23%),but this difference was not statistically significant(p ¼ 0.27). A total of 135 patients (32%) underwentimplantable cardioverter-defibrillator implantation,and 52 (12%) underwent implantable cardioverterdefibrillator plus cardiac resynchronization therapydevice implantation during follow-up.

ardiac Death(n ¼ 28)

No Cardiac Death(n ¼ 72) p Value

87 (73–100) 88 (78–100) 0.65

.3 (1.7–2.7) 2.2 (1.9–2.7) 0.97

13 (7–19) 8 (5–14) 0.02

57 (35–63) 44 (33–51) 0.02

28 (16–32) 21 (15–27) 0.04

38 (25–45) 32 (25–38) 0.04

21 (13–29) 20 (12–26) 0.43

12 (7–19) 11 (7–15) 0.34

.7 (4.9–13.3) 7.6 (4.8–9.6) 0.31

.5 (1.4–6.1) 2.2 (1.4–3.6) 0.50

.9 (13.0–21.5) 19.6 (16.3–24.5) 0.02

11 (103–122) 124 (105–150) 0.06

73 (65–80) 76 (68–85) 0.10

patients within a median of 0 days (IQR: �2 to þ1 days) from CMR.

P ¼ pulmonary artery pressure; PCWP ¼ pulmonary capillary wedge pressure; PVR ¼tion; SVR ¼ systemic vascular resistance; TPG ¼ transpulmonary gradient; TPR ¼ total

FIGURE 1 Cardiac Mortality Risk According to RVEF

0≤25 25 - ≤35 35 - ≤45 45 - ≤55

*

*

>55

1

2

5

4

3

Haza

rd R

atio

RVEF

The hazard ratios for cardiac mortality are shown according to different levels of right

ventricular ejection fraction (RVEF). The yellow dashed line displays the hazard ratio of

1.00 of the reference group with RVEF >55%. Patients with RVEF #35% had a signifi-

cantly higher event risk than did the reference group. Patients with RVEF >35

and #45%, as well as those >45% and #55% had similar risk to the group with RVEF

>55%. *p < 0.001.

FIGURE 2 Kaplan-

A

33984

100

90

80

70

60

50

40

030

Prob

abili

ty o

f Sur

viva

l (%

)

Numberat Risk

(A) The difference in

(B) The survival estim

Pueschner et al. J A C C : C A R D I O V A S C U L A R I M A G I N G , V O L . 1 0 , N O . 1 0 , 2 0 1 7

RVD in NICM O C T O B E R 2 0 1 7 : 1 2 2 5 – 3 6

1230

ASSOCIATION OF RVD AND CARDIAC MORTALITY.

We sought to determine the optimal threshold fordefining RVD based on cardiac mortality risk. Figure 1details the relationship between RVEF categories andHR for cardiac death. Patients with an RVEF #25%(n ¼ 27) had the highest cardiac death risk compared

Meier Estimates of Survival According to RVEF

RVEF ≤35%

RVEF >35%

108

292 253 206 83 1555 43 31 7 1

642Year

B100

80

60

40

20

0

Prob

abili

ty o

f Sur

viva

l (%

)

cardiac survival between patients with right ventricular ejection fraction (RVEF

ates in subgroups by deciles in RVEF are shown (see text for details).

with that of the reference group with RVEF >55%(HR: 4.68; 95% CI: 2.43 to 9.02; p < 0.0001). Patientswith RVEF >25% and #35% (n ¼ 57) had also highercardiac mortality than did those with RVEF >55%(HR: 2.84; 95% CI: 1.56 to 5.19; p ¼ 0.0007). An RVEF>35% and #45% (n ¼ 84) was associated with a small,but statistically insignificant risk increase (HR: 1.45;95% CI: 0.78 to 2.70; p ¼ 0.24), whereas patients withan RVEF >45% and #55% (n ¼ 126) had nearly iden-tical cardiac mortality risk (HR: 1.13; 95% CI: 0.64 to2.00; p ¼ 0.68) compared with the reference group(n ¼ 129). Similar relationships between the RVEFcategories and cardiac mortality risk were observedwhen adjusted for LVEF in all patients, and afterload(mean PAP) in the subgroup with RHC. Consequently,RVEF #35% was used to define RVD.

Figure 2A shows the survival curves for the entirestudy population stratified based on RVEF #35%versus >35%. Among the 84 patients with RVD, 36(43%) died of cardiac causes during follow-up,whereas among the 339 patients without RVD, 65(19%) reached the study endpoint (HR: 3.00; 95% CI:1.99 to 4.51; p < 0.0001). The Kaplan-Meier estimatesof cardiac death-free survival stratified by RVEF areshown in Figure 2B. Both subgroups with aRVEF #25% and RVEF >25% and #35% had signifi-cantly lower probability of survival than did thosewith RVEF >45% (p < 0.001 for both), whereas thosewith RVEF >35% and #45% had a similar probabilityof survival than did those with RVEF >45% (p ¼ 0.09).

RVEF ≤25%

RVEF >35 - ≤45%

RVEF >25 - ≤35%

RVEF >45%

1086420Year

)#35% and>35% was significant (p< 0.0001, by the log rank test).

J A C C : C A R D I O V A S C U L A R I M A G I N G , V O L . 1 0 , N O . 1 0 , 2 0 1 7 Pueschner et al.O C T O B E R 2 0 1 7 : 1 2 2 5 – 3 6 RVD in NICM

1231

PREDICTORS OF CARDIAC MORTALITY. HR associ-ated with all clinical, laboratory, electrocardio-graphic, and CMR parameters are shown in Table 3.RVEF was a strong, independent predictor for cardiacmortality (HR: 0.96; 95% CI: 0.94 to 0.98; p ¼ 0.0003)after adjustment for established risk markers ofmortality in HF patients including age, NYHA func-tional class, systolic blood pressure, heart rate, serumsodium, renal dysfunction, LVEF, and myocardialscar (model 1). On stepwise multivariable analysesincluding the 10 strongest univariable predictors(excluding RVEF), age (HR: 1.03; 95% CI: 1.01 to 1.04;p ¼ 0.0006), serum creatinine (HR: 1.14; 95% CI: 1.04to 1.25; p ¼ 0.004), myocardial scar (HR: 1.94; 95% CI:1.28 to 2.96; p ¼ 0.002), and LVEF (HR: 0.96; 95% CI:0.95 to 0.98; p < 0.0001) were independent predictorsof cardiac mortality. When RVEF was added as cova-riable (model 2 in Table 3, including RVEF), it wasamong the strongest risk markers of cardiac death(HR: 0.96; 95% CI: 0.94 to 0.98; p ¼ 0.0001), andLVEF was no longer a significant predictor. In addi-tion, the global chi-square of the latter modelincluding RVEF was significantly improved over themodel excluding RVEF (58.00 vs. 43.08; p ¼ 0.0001),substantiating the incremental prognostic value ofRVEF. In addition, the predicted cardiac mortalityrisk was determined on the basis of the best clinicalmodel, and the relative improvement in patientreclassification associated with RVEF was assessedusing thresholds of 0% to 15%, 16% to 30%, and>30%. The addition of RVEF resulted in 50 correct(down) reclassification and 36 incorrect (up) reclas-sifications in the 316 survivors. Additionally, 12 cor-rect (up) classifications and 11 incorrect (down)reclassifications occurred in subjects who died ofcardiac causes. Overall, 3.6% of patients werecorrectly reclassified by the addition of RVEF toestablished clinical parameters (net reclassificationimprovement: 3.6%; p < 0.0001).

In patients with LVEF #35% (n ¼ 222), RVD waspresent in 76 (34%). This subgroup with biventriculardysfunction (e.g., LVEF #35% and RVEF #35%) had amore than 2-fold higher cardiac mortality risk thandid those with LV dysfunction alone (e.g., LVEF #35%and RVEF >35%) (HR: 2.30; 95% CI: 1.42 to 3.73;p ¼ 0.0008). In patients with LVEF >35% (n ¼ 201),only 8 patients (4%) had RVD, which was also asso-ciated with higher cardiac mortality compared withthose with preserved RVEF (HR: 5.50; 95% CI: 1.93 to15.66; p ¼ 0.001). On multivariable Cox regressionanalyses, performed separately in both subgroups,RVEF was an independent predictor of cardiac deathafter adjustment for age, serum creatinine, and scar.

DETERMINANTS OF RVD. RVD was present in 20% ofthe entire study population, and in 43% among thosewho underwent RHC. The relationships among RVEFand afterload (mean PAP), pre-load (mean RAP), andLVEF are shown in Figure 3. There was a moderate,positive correlation between RVEF and LVEF(r ¼ 0.60; p < 0.0001). We observed only a modest,negative correlation with RAP (r ¼ �0.32; p ¼ 0.001),indicating that pre-load may have a narrow physio-logic window of improving RV systolic functionbefore volume overload effects impair RVEF. Therewas also a modest, inverse relation between RVEFand mean PAP (r ¼ �0.34; p ¼ 0.0005). Interestingly,when considering the components of afterload—TPGand PCWP—we found RVEF was weakly correlatedwith the former (r ¼ �0.28; p ¼ 0.006) but not withthe latter (r ¼ �0.15; p ¼ 0.13).

The strongest predictors of RVD were reducedLVEF, lower systemic systolic blood pressure, malesex, lower cardiac index, higher heart rate, andincreased mean PAP (Table 4). Both pre-load (RAP)and the pulmonary vascular component of afterload(TPG and pulmonary vascular resistance), but notPCWP were predictive of RVD. When adjusting forafterload (mean PAP), RAP was no longer a significantpredictor of RVD, but age, male sex, LVEF, cardiacindex, heart rate, and systolic blood pressureremained associated with RVD. A history of atrialfibrillation was not related to RVD.

In multivariable analysis, lower LVEF (odds ratio[OR]: 0.86; 95% CI: 0.81 to 0.92; p < 0.0001) andelevated mean PAP (OR: 1.01; 95% CI: 1.01 to 1.15;p ¼ 0.04) both were independently associated withRVD, but not mean RAP (OR: 1.02; 95% CI: 0.92 to 1.13;p ¼ 0.73). When mean PAP was substituted by itscomponents, TPG (OR: 1.15; 95% CI: 1.05 to 1.26;p ¼ 0.003) in addition to LVEF were independentpredictors of RVD, whereas mean RAP and meanPCWP were not predictive of RVD. Stepwise selectionof candidate variables revealed that LVEF (OR: 0.85;95% CI: 0.78 to 0.92; p < 0.0001), TPG (OR: 1.20; 95%CI: 1.09 to 1.32; p ¼ 0.0003), and systolic blood pres-sure (OR: 0.97; 95% CI: 0.94 to 0.99; p ¼ 0.02) werethe strongest independent predictors of RVD.

DISCUSSION

In this study of patients with NICM, we found thatRVEF determined by CMR strongly predicted cardiacmortality independent of established risk factors inHF patients. Furthermore, the grade of RVD associ-ated with a significant cardiac mortality risk(RVEF #35%) was present in 20% of patients.

TABLE 3 Predictors of Cardiac Mortality

Univariable Analysis Multivariable Analysis

HR (95% CI) p Value

Model 1 Model 2

HR (95% CI) p Value HR (95% CI) p Value

Age, yrs 1.02 (1.01–1.03) 0.005 1.03 (1.02–1.05) <0.0001 1.03 (1.02–1.05) <0.0001

Male 1.63 (1.09–2.44) 0.02

Non-white race 1.37 (0.92–2.04) 0.12

Diabetes mellitus 1.72 (1.14–2.60) 0.01

Hypertension 1.33 (0.89–1.99) 0.17

Cigarette smoker 1.60 (1.04–2.46) 0.03

Hyperlipidemia 0.97 (0.65–1.44) 0.88

Family history of CAD 1.03 (0.65–1.61) 0.91

NYHA functional class 1.48 (1.20–1.84) 0.0003 1.16 (0.91–1.49) 0.22

Atrial fibrillation or flutter 1.68 (1.07–2.65) 0.02

NSVT 0.89 (0.50–1.59) 0.68

Syncope 0.68 (0.30–1.54) 0.35

Body mass index 0.99 (0.96–1.02) 0.33

Systolic blood pressure, per 1 mm Hg 0.99 (0.98–0.99) 0.02 0.99 (0.98–0.99) 0.04

Diastolic blood pressure, per 1 mm Hg 0.99 (0.98–1.00) 0.46

Heart rate, beats/min 0.99 (0.99–1.01) 0.75 0.99 (0.98–1.01) 0.25

Serum sodium, per mmol/l 0.97 (0.91–1.02) 0.22 1.02 (0.96–1.09) 0.47

Serum creatinine, per mg/dl 1.10 (1.01–1.19) 0.02 1.18 (1.06–1.30) 0.002 1.15 (1.05–1.26) 0.003

Serum proBNP, per 1,000 pg/ml 1.01 (1.00–1.02) 0.045

ACE inhibitor 0.88 (0.58–1.33) 0.54

ARB 1.09 (0.58–2.03) 0.79

Antiarrhythmic drug 1.71 (0.99–2.95) 0.06

Aspirin 1.21 (0.80–1.82) 0.36

Beta-blocker 0.97 (0.60–1.55) 0.89

Calcium-channel blocker 1.24 (0.71–2.14) 0.45

Digitalis 1.87 (1.21–2.87) 0.005

Diuretics 1.98 (1.22–3.20) 0.005

Insulin 1.87 (1.15–3.06) 0.01

Oral antidiabetic 1.11 (0.61–2.03) 0.74

Nitrates 1.42 (0.83–2.43) 0.20

Statin 1.15 (0.76–1.73) 0.51

Warfarin 1.90 (1.26–2.86) 0.002

Heart rate on ECG, beats/min 1.01 (0.99–1.02) 0.18

QRS on ECG, ms 1.00 (0.99–1.01) 0.18

QTc on ECG, ms 1.00 (1.00–1.01) 0.03

Left bundle branch block on ECG 1.35 (0.85–2.16) 0.20

Right bundle branch block on ECG 0.87 (0.32–2.35) 0.78

RVEF, % 0.96 (0.95–0.98) <0.0001 0.96 (0.94–0.98) 0.0003 0.96 (0.94–0.98) 0.0001

RVEDV index, ml/m2 1.01 (1.00–1.02) <0.0001

RVESV index, ml/m2 1.02 (1.01–1.02) <0.0001

LVEF, % 0.96 (0.95–0.98) <0.0001 0.99 (0.97–1.02) 0.66 0.99 (0.97–1.01) 0.25

LVEDV index, ml/m2 1.00 (1.00–1.01) 0.001

LVESV index, ml/m2 1.00 (1.00–1.01) <0.0001

LV mass, g 1.00 (1.00–1.00) 0.0009

Any scar on DE-CMR 2.12 (1.40–3.22) 0.0004 1.86 (1.21–2.87) 0.005 1.90 (1.25–2.89) 0.003

Scar extent, %LV mass 1.04 (1.02–1.06) 0.001

ICD 0.72 (0.47–1.12) 0.14

ICDþCRT 1.05 (0.60–1.85) 0.87

Bold values are statistically significant.

CI ¼ confidence interval; CRT ¼ cardiac resynchronization therapy; ECG ¼ electrocardiogram; HR ¼ hazard ratio; ICD ¼ implantable cardioverter-defibrillator; NSVT ¼nonsustained ventricular tachycardia; other abbreviations as in Table 1.

Pueschner et al. J A C C : C A R D I O V A S C U L A R I M A G I N G , V O L . 1 0 , N O . 1 0 , 2 0 1 7

RVD in NICM O C T O B E R 2 0 1 7 : 1 2 2 5 – 3 6

1232

FIGURE 3 Correlation Between RVEF and RV Pre-Load, Afterload, and LVEF

CA

7060

80706050

203040

504030

r = –0.34p = 0.0005

10 2000

10RV E

ject

ion

Frac

tion

(%)

Mean PA Pressure (mm Hg)403025 35

80706050

203040

2015

r = 0.60p < 0.0001

5 1000

10

LV Ejection Fraction (%)

B

1512

80706050

203040

96

r = –0.32p = 0.001

300

10

Mean RA Pressure (mm Hg)

D

50

80706050

203040

4030

r = –0.28p = 0.006

10 2000

10RV E

ject

ion

Frac

tion

(%)

Transpulmonary Gradient (mm Hg)

E

453025 4035

80706050

203040

2015

r = –0.15p = 0.13

5 1000

10

Mean PCWP (mm Hg)

The graphs show correlations with regression lines between right ventricular ejection fraction (RVEF) and afterload (A), pre-load (B), and left ventricular ejection

fraction (LVEF) (C), and correlations with the components of RV afterload, that is, transpulmonary gradient (TPG) (D) and mean pulmonary capillary wedge pressure

(PCWP) (E). PA ¼ pulmonary arterial; RA ¼ right atrial.

J A C C : C A R D I O V A S C U L A R I M A G I N G , V O L . 1 0 , N O . 1 0 , 2 0 1 7 Pueschner et al.O C T O B E R 2 0 1 7 : 1 2 2 5 – 3 6 RVD in NICM

1233

We sought to study determinants of RVD andobserved strong, independent associations both withreduced LVEF and increased RV afterload.

RV function may be compromised in the setting ofreduced LVEF due to several factors. First, the sameprocess resulting in LV cardiomyopathy may impairRV myocardial contractility. Second, LV dysfunctionwith impaired contraction of septal myocardium andmutually encircling epicardial muscle-fiber layers maybe responsible for up to 40% reduction in RV systolicpressure and volume outflow due to systolic ventric-ular interdependence (30). Third, an increase in LVvolume or LV pressure overload may shift upward theRV diastolic pressure–volume relationship due to thepericardial constraints and result in RV impairmentthrough diastolic ventricular interdependence (31).Fourth, LV failure may result in increased RV afterloadby a rise in LV end-diastolic, pulmonary venous, and

ultimately pulmonary arterial pressures (group 2World Health Organization PH) (32,33).

We found a strong relation between RVD andafterload (mean PAP), interestingly though this wasspecifically related to the pulmonary vascularcomponent (TPG), whereas elevated PCWP was notassociated with RVD. This suggests that as PH due toleft heart disease progresses from passive backwardtransmission of an elevated PCWP in isolated post-capillary PH to impaired vascular reactivity andendothelial dysfunction and perturbed pulmonarycapillary and arteriolar structure with elevated TPG incombined post-capillary/pre-capillary PH (18), RVDmay ensue at some point along this process. Thissupports the notion that drug therapies targeting thepulmonary vasculature to lower the PAP in left-sidedHF may effectively improve RVEF. Going a stepfurther, because RVD is strongly associated with

TABLE 4 Predictors of RV Systolic Dysfunction

OR (95% CI)* p Value

RV Afterload-Adjusted†

OR (95% CI) p Value

Age, yrs 0.97 (0.94–1.00) 0.03 0.96 (0.93–0.99) 0.02

Male 4.63 (1.94–11.03) 0.0005 4.08 (1.67–9.99) 0.002

Duration of NICM, months 0.98 (0.95–0.99) 0.04 0.97 (0.95–0.99) 0.03

History of atrial fibrillation 0.99 (0.21–4.69) 0.99 1.07 (0.21–5.45) 0.94

LV ejection fraction 0.88 (0.83–0.92) <0.0001 0.86 (0.81–0.92) <0.0001

PAP mean 1.06 (1.02–1.10) 0.005 — —

PCWP mean 1.03 (0.98–1.08) 0.23 —‡ —

RAP mean 1.08 (1.01–1.15) 0.02 1.03 (0.95–1.11) 0.54

Cardiac index 0.27 (0.12–0.59) 0.001 0.32 (0.14–0.71) 0.005

Systolic blood pressure 0.96 (0.94–0.98) <0.0001 0.95 (0.93–0.97) <0.0001

Heart rate 1.05 (1.02–1.07) 0.002 1.04 (1.01–1.07) 0.02

PVR 1.33 (1.09–1.63) 0.006 —‡ —

TPG 1.07 (1.01–1.12) 0.02 —‡ —

Bold values are statistically significant. *OR for continuous data are expressed for 1 unit change. †RV afterload: PAP mean. ‡These parameters were not adjusted for because ofcollinearity with PAP mean.

OR ¼ odds ratio; other abbreviations as in Tables 1 to 3.

Pueschner et al. J A C C : C A R D I O V A S C U L A R I M A G I N G , V O L . 1 0 , N O . 1 0 , 2 0 1 7

RVD in NICM O C T O B E R 2 0 1 7 : 1 2 2 5 – 3 6

1234

cardiac mortality, such intervention might alsoimprove outcomes of HF patients with secondary PH.In this context, we note that the selective inhibitor oftype 5-phosphodiesterase sildenafil was shown in asmall, randomized clinical trial in 34 HF patients withelevated PA pressures to improve exercise capacityand quality of life (27). Interestingly, a higherimprovement in the primary endpoint (peakoxygen consumption) was observed in patients withbelow-median RVEF. Although currently pulmonaryarterial hypertension drug therapies are not recom-mended in PH due to left-heart disease (17), the use oftype 5-phosphodiesterase inhibition (NCT01616381,NCT01910389) or other vasoactive compounds(NCT02070991) are studied in ongoing clinical trials.Due to its robust relation with cardiac mortality,CMR-RVEF could be a surrogate endpoint in efficacystudies on drug treatment of PH due to left-heartdisease.

Data on the prevalence of PH due to left-heartdisease and subtypes are scarce and subject to vari-ability in definitions (34). In this study, PH was foundin 72% of patients, and one-half of those had com-bined post-capillary/pre-capillary PH. Our findingsare similar to data from a large PH center, whichshowed that 45% of PH due to left-heart disease wascombined PH (35). The high prevalence corroboratesthe importance of exploring novel drug therapies forPH due to left-heart disease.

We have measured RVEF using CMR, which allowsan accurate and reproducible assessment of the

complex anatomical structure of the RV (19).Furthermore, we stratified patients into RVEF sub-groups on the basis of the associated cardiac mortalityrisk to substantiate grades of severity of RVD assuggested by the American Society of Echocardiog-raphy (36). To date, limited data is available ondefining RVD based on CMR-RVEF and how to gradeseverity of RVD based on associated cardiac mortalityrisk. The current practice is to assign grades of mild,moderate, and severe RVD based on a set number ofSDs below the reference limit of a normal population(37,38).

STUDY LIMITATIONS. A limitation of this study isthat only 24% of all patients were referred for RHC,and thus selection bias is present for the evaluation ofdeterminants of RVD.

CONCLUSIONS

We have demonstrated that CMR-RVEF is a strongindependent and incremental risk factor for cardiacmortality in patients with NICM. RVD was related toboth LVEF and increased afterload, and treatmentstargeting the pulmonary vascular remodeling processmay improve RV systolic function and cardiacoutcomes.

ADDRESS FOR CORRESPONDENCE: Dr. Igor Klem,Duke Cardiovascular MRI Center, Cardiology,DUMC-3934, Durham, North Carolina 27710. E-mail:igorklem@duke.edu.

PERSPECTIVES

COMPETENCY INPATIENTCAREANDPROCEDURAL

SKILLS: This study suggests that it may be important to

assess RVEF in addition to LV parameters during a CMR

exam in HF patients with NICM. More importantly, our

results substantiate grades of severity for the interpreta-

tion of RVD based on the associated cardiac mortality risk,

which was not well established to date for CMR-RVEF.

TRANSLATIONAL OUTLOOK: Future research

should be directed on how medical or device

interventions may improve RV function and—given the

strong relationship with cardiac mortality—potentially

also prognosis.

J A C C : C A R D I O V A S C U L A R I M A G I N G , V O L . 1 0 , N O . 1 0 , 2 0 1 7 Pueschner et al.O C T O B E R 2 0 1 7 : 1 2 2 5 – 3 6 RVD in NICM

1235

RE F E RENCE S

1. Mozaffarian D, Benjamin EJ, Go AS, et al. HeartDisease and Stroke Statistics—2016 update: areport from the American Heart Association.Circulation 2016;133:e38–360.

2. Yancy CW, Jessup M, Bozkurt B, et al. 2013ACCF/AHA guideline for the management of heartfailure: a report of the American College of Car-diology Foundation/American Heart AssociationTask Force on Practice Guidelines. J Am Coll Car-diol 2013;62:e147–239.

3. Voelkel NF, Quaife RA, Leinwand LA, et al. Rightventricular function and failure: report of aNational Heart, Lung, and Blood Institute WorkingGroup on cellular and molecular mechanisms ofright heart failure. Circulation 2006;114:1883–91.

4. Di Salvo TG, Mathier M, Semigran MJ, Dec GW.Preserved right ventricular ejection fraction pre-dicts exercise capacity and survival in advancedheart failure. J Am Coll Cardiol 1995;25:1143–53.

5. Ghio S, Gavazzi A, Campana C, et al. Indepen-dent and additive prognostic value of right ven-tricular systolic function and pulmonary arterypressure in patients with chronic heart failure.J Am Coll Cardiol 2001;37:183–8.

6. Meluzin J, Spinarova L, Hude P, et al. Prog-nostic importance of various echocardiographicright ventricular functional parameters in patientswith symptomatic heart failure. J Am Soc Echo-cardiogr 2005;18:435–44.

7. Kjaergaard J, Akkan D, Iversen KK, Kober L,Torp-Pedersen C, Hassager C. Right ventriculardysfunction as an independent predictor of short-and long-term mortality in patients with heartfailure. Eur J Heart Fail 2007;9:610–6.

8. Gavazzi A, Ghio S, Scelsi L, et al. Response ofthe right ventricle to acute pulmonary vasodilationpredicts the outcome in patients with advancedheart failure and pulmonary hypertension. AmHeart J 2003;145:310–6.

9. Melenovsky V, Hwang SJ, Lin G, Redfield MM,Borlaug BA. Right heart dysfunction in heart fail-ure with preserved ejection fraction. Eur Heart J2014;35:3452–62.

10. Mohammed SF, Hussain I, AbouEzzeddine OF,et al. Right ventricular function in heart failurewith preserved ejection fraction: a community-based study. Circulation 2014;130:2310–20.

11. Gulati A, Ismail TF, Jabbour A, et al. Theprevalence and prognostic significance of rightventricular systolic dysfunction in nonischemicdilated cardiomyopathy. Circulation 2013;128:1623–33.

12. de Groote P, Millaire A, Foucher-Hossein C,et al. Right ventricular ejection fraction is an in-dependent predictor of survival in patients withmoderate heart failure. J Am Coll Cardiol 1998;32:948–54.

13. Gavazzi A, Berzuini C, Campana C, et al. Valueof right ventricular ejection fraction in predictingshort-term prognosis of patients with severechronic heart failure. J Heart Lung Transplant1997;16:774–85.

14. Juilliere Y, Barbier G, Feldmann L,Grentzinger A, Danchin N, Cherrier F. Additionalpredictive value of both left and right ventricularejection fractions on long-term survival in idio-pathic dilated cardiomyopathy. Eur Heart J 1997;18:276–80.

15. La Vecchia L, Paccanaro M, Bonanno C,Varotto L, Ometto R, Vincenzi M. Left ventricularversus biventricular dysfunction in idiopathicdilated cardiomyopathy. Am J Cardiol 1999;83:120–2, A9.

16. MacNee W. Pathophysiology of cor pulmonalein chronic obstructive pulmonary disease. PartOne. Am J Respir Crit Care Med 1994;150:833–52.

17. Galie N, Humbert M, Vachiery JL, et al. 2015ESC/ERS guidelines for the diagnosis and treat-ment of pulmonary hypertension: the Joint TaskForce for the Diagnosis and Treatment of Pulmo-nary Hypertension of the European Society ofCardiology and the European Respiratory Society.Eur Heart J 2016;37:67–119.

18. Guazzi M, Borlaug BA. Pulmonary hyperten-sion due to left heart disease. Circulation 2012;126:975–90.

19. Haddad F, Hunt SA, Rosenthal DN, Murphy DJ.Right ventricular function in cardiovascular dis-ease, part I: anatomy, physiology, aging, andfunctional assessment of the right ventricle. Cir-culation 2008;117:1436–48.

20. Beygui F, Furber A, Delepine S, et al. Routinebreath-hold gradient echo MRI-derived rightventricular mass, volumes and function: accuracy,

reproducibility and coherence study. Int J Car-diovasc Imaging 2004;20:509–16.

21. Bardy GH, Lee KL, Mark DB, et al. Amiodaroneor an implantable cardioverter-defibrillator forcongestive heart failure. N Engl J Med 2005;352:225–37.

22. Thygesen K, Alpert JS, White HD, et al. Uni-versal definition of myocardial infarction. Circula-tion 2007;116:2634–53.

23. Prineas RJ, Crow RS, Blackburn H. The Min-nesota Code: Manual of ElectrocardiographicFindings. Standards and Procedures for Measure-ment and Classification. Littleton, MA: Wright;1982.

24. Klem I, Weinsaft JW, Bahnson TD, et al.Assessment of myocardial scarring improves riskstratification in patients evaluated for cardiacdefibrillator implantation. J Am Coll Cardiol 2012;60:408–20.

25. Cerqueira MD, Weissman NJ, Dilsizian V, et al.Standardized myocardial segmentation andnomenclature for tomographic imaging of theheart. A statement for healthcare professionalsfrom the Cardiac Imaging Committee of theCouncil on Clinical Cardiology of the AmericanHeart Association. Circulation 2002;105:539–42.

26. Kim RJ, Albert TS, Wible JH, et al. Perfor-mance of delayed-enhancement magnetic reso-nance imaging with gadoversetamide contrast forthe detection and assessment of myocardialinfarction: an international, multicenter, double-blinded, randomized trial. Circulation 2008;117:629–37.

27. Lewis GD, Shah R, Shahzad K, et al. Sildenafilimproves exercise capacity and quality of life inpatients with systolic heart failure and secondarypulmonary hypertension. Circulation 2007;116:1555–62.

28. Epstein AE, Carlson MD, Fogoros RN,Higgins SL, Venditti FJ Jr. Classification of death inantiarrhythmia trials. J Am Coll Cardiol 1996;27:433–42.

29. Haddad F, Doyle R, Murphy DJ, Hunt SA. Rightventricular function in cardiovascular disease, partII: pathophysiology, clinical importance, andmanagement of right ventricular failure. Circula-tion 2008;117:1717–31.

Pueschner et al. J A C C : C A R D I O V A S C U L A R I M A G I N G , V O L . 1 0 , N O . 1 0 , 2 0 1 7

RVD in NICM O C T O B E R 2 0 1 7 : 1 2 2 5 – 3 6

1236

30. Santamore WP, Dell’Italia LJ. Ventricularinterdependence: significant left ventricular con-tributions to right ventricular systolic function.Prog Cardiovasc Dis 1998;40:289–308.

31. Taylor RR, Covell JW, Sonnenblick EH,Ross J Jr. Dependence of ventricular distensibilityon filling of the opposite ventricle. Am J Physiol1967;213:711–8.

32. Setaro JF, Cleman MW, Remetz MS. The rightventricle in disorders causing pulmonary venoushypertension. Cardiol Clin 1992;10:165–83.

33. McLaughlin VV, Archer SL, Badesch DB, et al.ACCF/AHA 2009 expert consensus document onpulmonary hypertension: a report of the AmericanCollege of Cardiology Foundation Task Force onExpert Consensus Documents and the American

Heart Association. J Am Coll Cardiol 2009;53:1573–619.

34. Cheli M, Vachiery JL. Controversies in pulmo-nary hypertension due to left heart disease.F1000Prime Rep 2015;7:07.

35. Gerges C, Gerges M, Lang MB, et al. Diastolicpulmonary vascular pressure gradient: a predictorof prognosis in “out-of-proportion” pulmonaryhypertension. Chest 2013;143:758–66.

36. Lang RM, Bierig M, Devereux RB, et al.Recommendations for chamber quantification: areport from the American Society ofEchocardiography’s Guidelines and StandardsCommittee and the Chamber QuantificationWriting Group. J Am Soc Echocardiogr 2005;18:1440–63.

37. Marcus FI, McKenna WJ, Sherrill D, et al.Diagnosis of arrhythmogenic right ventricularcardiomyopathy/dysplasia: proposed modificationof the task force criteria. Circulation 2010;121:1533–41.

38. Maceira AM, Prasad SK, Khan M, Pennell DJ.Reference right ventricular systolic and diastolicfunction normalized to age, gender and bodysurface area from steady-state free precessioncardiovascular magnetic resonance. Eur Heart J2006;27:2879–88.

KEY WORDS heart failure, nonischemiccardiomyopathy, right ventriculardysfunction