Embed Size (px)
Citation preview
International Journal of Cardiology 177 (2014) 429–435
Contents lists available at ScienceDirect
International Journal of Cardiology
j ourna l homepage: www.e lsev ie r .com/ locate / i j ca rd
Right ventricular dysfunction, late gadolinium enhancement, and femalegender predict poor outcome in patients with dilated cardiomyopathy
Christina Doesch a,c,⁎, Désirée-Marie Dierks a, Dariusch Haghi a,c, Rainer Schimpf a,c, Jürgen Kuschyk a,c,Tim Suselbeck a,c, Stefan O. Schoenberg b,c, Martin Borggrefe a,c, Theano Papavassiliu a,c
a 1st Department of Medicine, Cardiology, University Medical Centre Mannheim, Medical Faculty Mannheim, University of Heidelberg, Germanyb Institute of Clinical Radiology and Nuclear Medicine, University Medical Center Mannheim, Medical Faculty Mannheim, University of Heidelberg, Germanyc DZHK (German Centre for Cardiovascular Research) partner site, Mannheim, Germany
⁎ Corresponding author at: 1st Department of MediciDZHK (German Centre for Cardiovascular ResarcUniversity Medical Center Mannheim, Medical FacuHeidelberg, Theodor-Kutzer-Ufer 1-3, 68167 Mannh383 2204; fax: +49 621 383 3821.
E-mail address: [email protected] (C. Doesch
http://dx.doi.org/10.1016/j.ijcard.2014.09.0040167-5273/© 2014 Elsevier Ireland Ltd. All rights reserved
a b s t r a c t
a r t i c l e i n f oArticle history:
Received 21 December 2013Received in revised form 31 July 2014Accepted 15 September 2014Available online 7 October 2014Keywords:Dilated cardiomyopathyCardiovascular magnetic resonance imagingRight ventricular ejection fractionRisk stratification
Aims: Dilated cardiomyopathy (DCM) shows a variable disease course and is associated with significantmorbidity and mortality. So far, left ventricular function (LVF) is the major determinant for risk stratification.However, since it has shown to be a poor guide to individual outcome, we studied the prognostic value ofcardiovascular magnetic resonance imaging (CMR) parameters, late gadolinium enhancement (LGE) andepicardial adipose tissue (EAT).Methods and results: 140 patients with DCM underwent late gadolinium enhancement (LGE) CMR. During amedian follow-up of 3 years, 22 patients (16%) died and another 51 (36%) were hospitalized due to congestiveheart failure (CHF). Female gender and right ventricular ejection fraction (RV-EF) below themedian of 38%wereindependent predictors of all-causemortality inmultivariable analysis. In patients whowere hospitalized due toCHF, RV-EF below the median of 38% was the only independent predictor in multivariable analysis. When
patients where further stratified according to systolic LV-EF, the prognostic value of RV-EF to predict mortalityand cardiac morbidity remained unchanged. Looking at DCM patients who died during follow-up compared tothose who were hospitalized due to CHF, the former presented with a higher prevalence of LGE as well asreduced indexed EAT.Conclusion: Female gender, RV-EF and the presence of LGE are of prognostic importance in patients with DCM.Therefore, the present study underlines the role of CMR as an important tool for risk stratification in patientswith DCM.© 2014 Elsevier Ireland Ltd. All rights reserved.
1. Introduction
Dilated cardiomyopathy (DCM) shows a variable disease course andis associated with significant morbidity and mortality [1–3]. So far, leftventricular ejection fraction (LV-EF) is the major determinant for riskstratification and the current guidelines recommend implantation of adefibrillator for primary prevention in symptomatic patients(New York Heart Association functional class II/III) with a LV-EF lessthan 35% [4,5]. However, LV-EF alone has shown to be a poor guide tooutcome.
Cardiovascular magnetic resonance imaging (CMR) is the goldstandard for non-invasive, accurate, and reproducible assessmentof left and right ventricular function, cardiac mass and morphology
ne, Cardiology affiliated at theh) partner site Mannheim,lty Mannheim, University ofeim, Germany. Tel.: +49 621
).
.
[6]. Due to the use of late gadolinium enhancement (LGE) technique,CMR allows in vivo quantification of regions of replacement fibrosisin patients with DCM that has shown a good correlation withhistological data [7]. In prior studies [7–10], the presence of LGE inpatients with DCM was also associated with an unfavorable progno-sis. Additionally, CMR allows the quantification of epicardial adiposetissue (EAT) that has been shown to be reduced in patients with DCM[11–13] and became also suspect to be associated with a poorprognosis [13].
Since one single parameter does not seem to be sufficient to predictthe clinical outcome of patients with NICM, we studied the prognosticvalue of CMR parameters, LGE and EAT.
2. Methods
2.1. Study population
Between February 2003 and February 2011, 150 consecutive patients with DCM thatunderwent late gadolinium enhancement CMR to quantify left ventricular (LV) functionand myocardial scarring as part of their routine clinical work-up were enrolled at ourtertiary referral hospital.
430 C. Doesch et al. / International Journal of Cardiology 177 (2014) 429–435
The diagnosis of DCMwas based on the 1995WHO/International Society and Federa-tion of Cardiology criteria [14]. All patients had undergone coronary angiography andwere classified as non-ischemic if they had no history ofmyocardial infarction or revascu-larization and no evidence of coronary artery stenoses N50% of 2 ormore epicardial vesselsor left main or proximal anterior descending coronary artery N50% [15]. Patients with anormal CMR-derived left ventricular ejection fraction (LV-EF N 55%) were not includedin the study. Other exclusion criteria were standard contraindications to CMRexamination.
CMR examinationwas not possible in 5 (3.3%) patients due to claustrophobia and in 3(2.0%) patients due to severe obesity. 2 (1.3%) patients were lost during follow-up so thatthe final study population consisted of 140 patients. Of these 140 patients, 19 patientswere included in an earlier study [13] and are now reported with an extended follow-up.
The study was approved by the local ethics committee and informed consent was ob-tained from all patients.
2.2. Image acquisition
All studies were performed using 1.5 Tesla whole-body imaging systems (MagnetomSonata and Avanto, SiemensHealthcare Sector, Erlangen, Germany). To evaluate function-al parameters, electrocardiogram-gated cine images were acquired using a segmentedsteady-state free precession [fast imaging with steady-state precession (true-FISP)] se-quence (time to echo/time of repetition 1.6/3.2 ms, temporal resolution 35 ms, in-planespatial resolution 1.4 × 1.8mm, slice thickness 8mm, interslice gap 2mm). For the assess-ment of the epicardial adipose tissue, we used a dark blood prepared T1-weighted multi-slice turbo spin-echo pulse sequence with a water suppression prepulse (time ofrepetition = 800 ms, time to echo = 24 ms, slice thickness = 6 mm, interslice gap =2mm, and field of view= 30 to 34 cm) in the same orientations as the short-axis images.
2.3. LGE
Ten minutes after contrast agent injection (BW gadoterate meglumine, Dotarem,Guerbet, France), late gadoliniumenhancement (LGE) imageswere acquired. An inversiontime (TI) scoutwasperformed to choose the optimal TIs between200 and 360ms. LGE im-ages were assessed using either an inversion recovery Turbo FLASH 2D sequence: field ofview 300–340 mm, TR 9.56 ms, TE 4.38 ms, flip angle 25°; matrix 166 × 256 and slicethickness 6 mm or a phase-sensitive inversion recovery TrueFISP sequence [16]: field ofview 290 mm × 260 mm, TR 2.2 ms, TE 1.1 ms, flip angle 50°; matrix, 140 × 192 andslice thickness 6 mm, in-plane resolution 1.4 × 1.9 × 6 mm. LGE was only considered tobe present if it was also present in the same slice after swapping phase encoding, thusexcluding artifacts.
2.4. Image analysis
Image analysis and quantitative analysis were performed off-line using dedicatedsoftware (ARGUS, Siemens, Germany). The readers were blinded to patient data and out-come. On the four-chamber view, the tricuspid annular plane systolic excursion (TAPSE)was calculated as previously described [17]. The amount of EATwas determined accordingto the method described by Fluechter S et al. [18].
Table 1Baseline demographic and clinical characteristics.
All DCM patients n = 140
Male n (%) 108 (77.1)Age (yrs) 59.3 ± 13.8NYHA functional class• I 7 (5.0)• II 26 (18.6)• II 64 (45.7)• IV 43 (30.7)
2.5. Extent of LGE
The extent of LGEwas assessed visually by two independent experienced readers. LGEwas only considered to be present if it was also present in the same slice after swappingphase encoding, thus excluding artifacts. The pattern of LGE was characterized as mid-wall, patchy foci, epicardial, or diffuse [9,10]. For quantification of fibrosis, LGEwas definedas areas with a signal intensity N 2 standard deviation (SD) abovemean signal intensity ofremote myocardium in the same short-axis slice [8]. Areas were measured by planimetryand expressed as percentage of the myocardial area using the VPT tool (SiemensHealthcare Systems Erlangen, Germany).
Atrial fibrillation n (%) 54 (38.6)Family history of DCM n (%) 8 (5.7)Hypertension n (%) 46 (32.9)Smoking n (%) 20 (14.3)Hyperlipidemia n (%) 41 (29.3)Diabetes n (%) 31 (22.1)Medication n (%)• Beta-blocker 118 (84.3)• ACEI 113 (80.7)• ARB 21 (15.0)• Spironolactone 34 (24.3)• Diuretics 100 (71.4)• Digoxin 38 (27.1)• Amiodarone 12 (8.6)
Abbreviations: ACEI: angiotensin-converting-enzyme inhibitor, ARB: angiotensin II recep-tor blockers, CHF: congestive heart failure, n: number, DCM: dilated cardiomyopathy,NYHA: New York Heart association functional class, yrs: years
2.6. Follow-up data and definition of study endpoints
The long-term follow-up was performed by patient interview at our outpatientclinic and by telephone contact. The observers were unaware of the CMR resultsand collected data with a standardized questionnaire. Reported clinical events wereconfirmed by review of the corresponding medical records in our electronic HospitalInformation System, contact with the general practitioner, referring cardiologist, orthe treating hospital. The definition of cardiac event required the documentation ofsignificant ventricular arrhythmia or cardiac arrest or death attributable to congestiveheart failure or myocardial infarction in the absence of any other precipitating factor. Incase of out-of-hospital death not followed by autopsy, sudden unexpected death wasclassified as cardiac death. The primary study endpoint was a combined endpoint of all-cause mortality including non cardiac and cardiac death as well as heart transplantation(HTX). The secondary endpoint was hospitalization due to worsening of CHF. Patientswhowerehospitalized and died over the course of follow-upwere only counted regardingthe primary endpoint and not the secondary endpoint.
2.7. Statistical analysis
Since we aimed to study to what extent CMR results, age and gender were associatedwith events, Cox proportional hazard regressionmodels were constructed for age, genderand CMR parameters. To check the proportional hazard assumption for different catego-ries, they were plotted against time to ensure that the curves were reasonably parallel.Those variables which appeared to be associated with events at a value of p ≤ 0.1 levelin univariable analysis were eligible for multivariable analysis to predict hospitalizationfor CHF. Forward stepwise logrank regression (p b 0.2 for entry, p N 0.1 for removal) wasused. Due to the limitednumber of all-causemortality events, thenumber of candidate pa-rameters for multivariable analysis was limited to two (the clinical and CMR parameterwith the best performance in univariable analysis) to avoid model overfitting. Resultsare presented as hazard ratioswith 95% confidence intervals (CIs). The follow-up durationwas measured from the CMR study date.
The performance of the final models was assessed with respect to discrimination.Discrimination is the model's ability to separate patients with different outcomes. Toquantify the discrimination, we used the c-statistic (Harrell's C) [19]. The maximumvalue of the c-statistic is 1.0; indicating a perfect predictionmodel. A value of 0.5 indicatesthat patients are correctly classified in 50% of the cases, e.g. as good as chance. The perfor-mance of a prediction model is generally worse in new patients than initially expected.This “optimism” can be studied with internal validation techniques [20]. Internal validityof our models was assessed with standard bootstrapping techniques [20]. The c-statisticof the final multivariable model, corrected for optimism, was reported.
Kaplan–Meier analysis was performed for the independent predictors of all-cause-mortality. For this analysis the study population was divided into two groups accordingto the median value of the entire study population. Difference in survival over time wasevaluated by a log-rank test.
Analysiswas performed using SPSS statistical software (version 14.0, SPSS Inc., Chicago,Illinois), Stata version 11 (StataCorp), and R software (version 2.8.1, R foundation forstatistical computing, Vienna, Austria).
3. Results
140 patients with DCM (77% men, mean age 59.2 ± 13.9) were in-cluded in the study. The baseline clinical characteristics are presentedin Table 1. Most patients presented with symptomatic heart failure(NYHA N I). Median follow-up was 3 years (interquartile range 0.5–5.0 years).
3.1. Outcome
During the follow-up, 22 (16%) patients died. Thereof 15 (11%)patients suffered cardiac death, 1 (0.7%) patient underwent HTX andnon cardiac death was reported in 6 (4%) patients. 51 (36%) werehospitalized due to congestive heart failure (CHF). 67 (48%) showedan event-free survival.
Table 3Cox proportional hazard analysis for the time to the occurrence of all-cause mortality.
Univariable analysis Multivariable analysis
Hazard ratio p value Hazard ratio p value
431C. Doesch et al. / International Journal of Cardiology 177 (2014) 429–435
3.2. Demographic, clinical and CMR parameters of all patients with DCM
Tables 1 and 2 show the demographic, clinical and CMR parametersin all patients with DCM.
(95% CI) (95% CI)
Female gender 3.6 (1.2–10.7) 0.02 3.4 (1.4–8.3) 0.006Age ≥ 65 yrs 2.4 (1.0–5.7) 0.05LV-EF b 35% 1.5 (0.5–4.4) 0.5LV-EDVI ≥ 137 ml/m2 1.3 (0.6–3.0) 0.5LV-ESVI ≥ 99 ml/m2 1.9 (0.8–4.5) 0.1RV-EF b 38% 3.7 (1.5–9.2) 0.004 4.7 (1.8–12.0) 0.001TAPSE b 1.7 3.4 (1.5–8.0) 0.005RV-EDVI ≥ 92 ml/m2 2.5 (1.0–5.9) 0.04RV-ESVI ≥ 54 ml/m2 3.3 (1.3–8.1) 0.01LGEpos 2.3 (1.0–5.2) 0.06Indexed EAT b 22 g/m2 1.3 (0.5–3.2) 0.6
Abbreviations: EAT: epicardial adipose tissue, g = gram, LGE: late gadolinium enhance-
3.3. Predictors of all-cause mortality compared to event-free survival
By univariable analysis, significant associations were observedbetween age, female gender, RV-EF, TAPSE, RV-EDVI, RV-ESVI and thepresence of LGE (Table 3). Due to the limited number of events onlyfemale gender and RV-EF as representative of right ventricular functionand volumes were used in multivariable analysis (Table 3). Thec-statistic of the final multivariable model including female genderand RV-EF was 0.73 (95% CI 0.72–0.92), while the bootstrapping proce-dure resulted in an optimism corrected c-statistic of 0.71.
ment, LV: left ventricular, LV-EDVI: left ventricular end diastolic volume index, LV-EF:left ventricular ejection fraction, LV-ESVI: left ventricular end-systolic volume index, m2:square meter, ml: milliliter, n: number, RVEDD: right ventricular end diastolic diameter,RV-EDVI: right ventricular end diastolic volume index RV-EF: right ventricular ejectionfraction, RV-ESVI: right ventricular end-systolic volume index, TAPSE: tricuspid annularsystolic excursion
3.4. Survival analysis for all-cause mortality
Fig. 1 A–D illustrates the Kaplan–Meier curves for the independentpredictors of all-cause mortality. After 2.5 years, female gender was as-sociated with a significantly higher all-cause mortality rate (p = 0.02,Fig. 1 A). LGEpos DCM patients showed a higher rate of all-causemortality that became particularly pronounced after 6 years (p= 0.047,Fig. 1 B). TAPSE below the median of 1.7 cm also was associatedwith a markedly higher all-cause mortality rate during the follow-up (p = 0.0003, Fig. 1 C). During the entire follow-up period, RV-EF below the median 38% correlated significantly with a higher rateof all-cause mortality (p = 0.002, Fig. 1 D).
When patients where further stratified according to systolic LV-EF,the prognostic value of RV-EF remained unchanged. While LV-EFalone was not a significant discriminator Fig. 2 A), patients with aRV-EF above the median 38% showed a better survival irrespectiveof LV-EF (Fig. 2 B).
Table 2CMR characteristics.
All DCM patients (n = 140)
LV-EF (%) 27.8 ± 10.6EDM/BSA (g/m2) 105.5 ± 30.1LV-EDVI (ml/m2) 146.2 ± 48.1LV-ESVI (ml/m2) 107.8 ± 46.6LVEDD (mm) 63.3 ± 15.8RVEDD (mm) 44.2 ± 7.8RAD (mm) 48.2 ± 8.1TAPSE (cm) 1.7 ± 0.8RV-EF (%) 36.9 ± 16.5RV-EDVI (ml/m2) 102.7 ± 39.1RV-ESVI (ml/m2) 66.9 ± 41.4LVRI (g/ml) 0.7 ± 0.2Indexed EAT mass (g/m2) 23.2 ± 6.6Presence of LGE n (%) 44 (31.4)LGE extent (%) 2.2 ± 4.2LGE pattern- midwall 21 (15.0)- patchy foci 13 (9.3)- epicardial 7 (5.0)- diffuse 3 (2.1)
Abbreviations: CHF: congestive heart failure, EAT: epicardial adipose tissue, g = gram,LGE: late gadolinium enhancement, LV: left ventricular, LV-EDD: left ventricular enddiastolic diameter, LV-EDVI: left ventricular end diastolic volume index, LV-EF: leftventricular ejection fraction, LV-ESVI: left ventricular end-systolic volume index, LVRI:left ventricular remodelling index, m2: square meter, ml: milliliter, n: number, RAD:right atrial diameter, RVEDD: right ventricular end diastolic diameter, RV-EDVI: right ven-tricular end diastolic volume index RV-EF: right ventricular ejection fraction, RV-ESVI:right ventricular end-systolic volume index, TAPSE: tricuspid annular systolic excursion
3.5. Predictors of hospitalization due to congestive heart failure comparedto event-free survival
CHF was associated with LVF, LV-EDVI, LVESVI, RV-EF and RV-ESVI by univariable analysis (Table 4). In multivariable analysisonly RV-EF proved to be independently associated with CHF(Table 4). The c-statistic for RV-EF to predict hospitalization due toCHF was 0.65 (95% CI: 0.55–0.75). The optimism corrected c-statisticswas 0.62.
3.6. Survival analysis for congestive heart failure
Kaplan–Meier curve for the only independent predictor of CHF isshown in Fig. 3. During the entire follow-up period, RV-EF below themedian 38% is associated with a markedly higher rate of hospitalizationdue to CHF (p = 0.001).
Also with regard to hospitalization due to CHF, the prognostic valueof RV-EF lasted even after stratification of patients according to systolicLV-EF. While LV-EF alone was not a significant discriminator Fig. 4 A),patients with a RV-EF above the median 38% showed a lower hospitali-zation rate irrespective of LV-EF (Fig. 4 B).
3.7. Comparison of patients who died and those hospitalized due tocongestive heart failure
Looking at the patients who died during the follow-up compared tothose whowere hospitalized due to CHF, the former presented with re-duced indexed EAT (19.6 ± 6.0 g/m2 vs 23.2 ± 5 g/m2) as well as ahigher prevalence of LGE (54.6% vs 27.5%). All other CMR characteristicswere comparable between the two groups.
4. Discussion
The first main result of the present study is that among the analyzedCMRparameters including LGE andEAT, RV-EF is the strongest indepen-dent predictor of all-cause mortality, and a modest predictor of cardiacmorbidity due to CHF in patients with DCM. Secondly, even when pa-tients where further stratified according to systolic LV-EF, the prognos-tic value of RV-EF to predict all-cause mortality and cardiac morbidityremained unchanged. Thirdly, looking at the DCM patients who diedduring the follow-up compared to those who were hospitalized due toCHF, the former presentedwith a higher prevalence of LGE aswell as re-duced indexed EAT.
Fig. 1. Kaplan–Meier curves for all-cause mortality. Kaplan–Meier curves for all-cause mortality according to gender (A), presence of LGE (B), TAPSE (C) and RV-EF (D).
432 C. Doesch et al. / International Journal of Cardiology 177 (2014) 429–435
4.1. Parameters to predict all-cause mortality in DCM
In univariable and multivariable analysis female gender was predic-tive of all-causemortality. Female gender had also shown to be an inde-pendent predictor of cardiovascular death in patients with CHF due toICM or DCM in a prior study by Bistola V et al. [21]. Besides, in an Italianmulticenter study [22] of patients with idiopathic dilated cardiomyopa-thy, women presented with more advanced disease with regard tosymptoms and LV dimensions as well as a trend toward poorer progno-sis. However, a limiting factor for evaluating gender-related aspects isthe small number of women usually included in CHF studies. In ourstudy population only 22.9% of patients included were female. Further-more, the number of womenwith severely reduced LV-EF is significant-ly smaller than among men which certainly constitutes a certaininclusion bias in most studies selecting the patients according to LV-EF.
Fig. 2. Event-free survival according to LV-EF and RV-EF. Kaplan–Meier analysis shows thatsignificantly better prognosis irrespective of LV-EF (B).
RV-EF had an important prognostic impact on survival in univariableand multivariable analysis. RV function and volumes are parametersthat are rarely taken under consideration when analyzing prognosis inpatients with DCM. However, in patients with DCM, RV function hasnot only shown to be impaired as consequence of LV dysfunction [23,24] but histopathologically also a direct right ventricular involvementcould be proved in patients with DCM [25–27]. In line with our results,a previous echocardiography study by Juillière Y et al. [28] also foundthat in addition to LV-EF, RV-EF appeared to be a complementary pre-dictor of survival in 62 patients with idiopathic DCM. Another echocar-diography study by Meluzin J et al. also observed a poor prognosis in177 patients with either ischemic or idiopathic DCM and RV systolicand diastolic dysfunction.
TAPSE represents a quick semi-quantitative approach to assess in-formation about the RV-EF [29], was found to be correlated to markers
LV-EF alone is not a significant discriminator (A) but patients with a RV-EF N38% had a
Table 4Cox proportional hazard analysis for the time to the occurrence of hospitalization due tocongestive heart failure.
Univariable analysis Multivariable analysis
Hazard ratio(95% CI)
p value Hazard ratio(95% CI)
p value
Female gender 0.8 (0.4–1.6) 0.6Age ≥ 65 yrs 1.0 (0.6–1.8) 0.9LV-EF b 35% 1.9 (0.9–4.3) 0.1 0.6LV-EDVI ≥ 137 ml/m2 1.7 (1.0–3.1) 0.04 0.2LV-ESVI ≥ 99 ml/m2 2.0 (1.2–3.6) 0.02 0.2RV-EF b 38% 2.6 (1.4–4.6) 0.001 2.6 (1.4–4.6) 0.001RV-ESVI ≥ 54 ml/m2 2.0 (1.1–3.5) 0.02 0.9LGEpos 0.9 (0.5–1.7) 0.7Indexed EAT b 22 g/m2 1.2 (0.7–2.2) 0.5
Abbreviations: LV-EDVI: left ventricular end diastolic volume index, LV-EF: left ventricularejection fraction, LV-ESVI: left ventricular end-systolic volume index, m2: square meter,ml: milliliter, RV-EF: right ventricular ejection fraction, RV-ESVI: right ventricularend-systolic volume index
Fig. 3. Kaplan–Meier curve for hospitalization due to CHF. Kaplan–Meier curve for hospi-talization due to CHF dependent on RV-EF.
433C. Doesch et al. / International Journal of Cardiology 177 (2014) 429–435
of diastolic dysfunction and has been proven to be a valuable prognosticmarker in various cardiac diseases, including heart failure [30]. Our dataalso showed a univariable association between reduced TAPSE and a
Fig. 4. Survival free from hospitalization due to CHF according to LV-EF and RV-EF. Kaplan–Meiewhereas patients with a RV-EF N38% revealed a significantly higher probability of event-freegadolinium enhancement, LV-EF: left ventricular ejection fraction, RV-EF: right ventricular ejec
worse outcome. Kaplan–Meier curves significantly differed for patientswith TAPSE above the median (TAPSE≥ 1.7 cm) and below the median(TAPSE b 1.7 cm). In multivariable analysis TAPSE had no independentprognostic impact.
In line with a previous study by Hombach V et al. [31] LGE wasonly univariably associated with all-cause mortality but failed toplay an independent prognostic role. Kaplan–Meier survivalestimates for all-cause mortality showed a better probability of sur-vival for patients without LGE compared to those with LGE. Studiesthat identified LGE as the most important indicator of outcome inpatients with DCM did not include either RV-EF [7,9] or RV volumes[8,10]. Besides, the study by Wu et al. [9] only included DCMpatients with an LV-EF b 35% and therefore did not represent thewhole spectrum of patients with DCM studied by Hombach V et al.[31] and in the present study. Additionally, the present study alsotook into consideration the effect of a new parameter EAT that hasalso shown to have an impact on survival in a prior study [13] of pa-tients with heart failure due to ICM or DCM and severely reducedLV-EF.
The performance of the final multivariable model including femalegender and RV-EF was good with an optimism corrected c-statistic of0.80. Therefore, these parameters can be used to elaborate risk stratifi-cation in patients with DCM.
4.2. Parameters to predict hospitalization due to CHF in DCM
With regard to hospitalization due to CHF in patients with DCM,RV-EF was the only independent predictor in multivariable analysis.This result is in line with a study by Meluzin J. et al. [32] who alsofound a better event-free survival in patients with DCM and preservedRV-EF estimated by echocardiographic peak systolic tricuspid annularvelocity. RV and LV volumes above the median were both associatedwith hospitalization due to CHF in univariable analysis but did notprove to be independent predictors in multivariable analysis. Althoughleft ventricular dilatation is the hallmark of disease, there exists a hugevariability in the degree of right ventricular dilatation. Lewis JF et al. [33]also showed a poorer survival in patients with a more severe rightdilation compared to patients with a predominant left dilatation. Inthe study by Hombach V et al. [31], RV dilatation was also a risk factorto reach the composite endpoint of cardiac death and hospitalizationdue to CHF.
As expected, patients with a reduced LV-EF and RV-EF had theworst prognosis with regard to all cause mortality, HTX and hospital-ization due to CHF. Interestingly, patients with a LV-EF ≤35% and aRV-EF above the median 38% did better with regard to bothendpoints than those with a LV-EF N 35% and a RV-EF below 38%.The optimism corrected c-statistics of 0.62 showed that RVEF is a
r analysis only showed a trend toward a better prognosis in patients with LV-EF N35% (A),survival irrespective of LV-EF (B). Abbreviations: CHF: congestive heart failure, LGE: latetion fraction.
434 C. Doesch et al. / International Journal of Cardiology 177 (2014) 429–435
modest predictor of hospitalization due to CHF. These results onceagain highlight the prognostic importance to assess RV-EF in patientswith DCM.
Looking at the discriminators between those patientswith a severelyreduced LV-EF and RV-EF who died as opposed to those hospitalizeddue to CHF, a reduced EAT and the presence of LGE designated thosethat were more likely to die.
4.3. Limitations
One limitation of our study is the limited sample size and thesmall number of events. Another limitation is that the data present-ed were observed only at a single center. Therefore, larger multi-center trials are needed to confirm our results and to allowstudying further parameters that might be of interest in thesepatients.
5. Conclusion
Female gender, RV-EF and the presence of LGE are of prognosticimportance in patients with DCM. In the present study, RV-EF provedto be an important independent predictor of all-cause mortality anda modest predictor of hospitalization due to CHF in patients withDCM. Further stratification of patients according to systolic LV-EFshowed that the prognostic value of RV-EF to predict mortality andcardiac morbidity remained unchanged. Therefore, the presentstudy underlines the role of CMR as an important risk stratificationtool in patients with DCM.
Funding sources
This study was supported by grants from the DZHK (“DeutschesZentrum für Herz-Kreislauf-Forschung”—German Centre for Cardiovas-cular Research, HD/MA 6.1 IM Multimodal Image Fusion) and by theBMBF (German Ministry of Education and Research, HD/MA 6.1 IMMultimodal Image Fusion).
Disclosures
None.
Conflict of interest
The authors report no relationships that could be construed as aconflict of interest.
Acknowledgment
The authors of this manuscript have certified that they comply withthe Principles of Ethical Publishing in the International Journal ofCardiology.
References
[1] Coughlin SS, Neaton JD, SenguptaA, Kuller LH. Predictors ofmortality from idiopathicdilated cardiomyopathy in 356,222 men screened for the Multiple Risk FactorIntervention Trial. Am J Epidemiol 1994;139:166–72.
[2] Maron BJ, Towbin JA, Thiene G, Antzelevitch C, Corrado D, Arnett D, et al. Contempo-rary definitions and classification of the cardiomyopathies: an American HeartAssociation Scientific Statement from the Council on Clinical Cardiology, HeartFailure and Transplantation Committee; Quality of Care and Outcomes Researchand Functional Genomics and Translational Biology Interdisciplinary WorkingGroups; and Council on Epidemiology and Prevention. Circulation 2006;113:1807–16.
[3] Adams Jr KF, Dunlap SH, Sueta CA, Clarke SW, Patterson JH, Blauwet MB, et al.Relation between gender, etiology and survival in patients with symptomaticheart failure. J Am Coll Cardiol 1996;28:1781–8.
[4] Dickstein K, Cohen-Solal A, Filippatos G, McMurray JJ, Ponikowski P, Poole-WilsonPA, et al. ESC Guidelines for the diagnosis and treatment of acute and chronic
heart failure 2008: the Task Force for the Diagnosis and Treatment of Acute andChronic Heart Failure 2008 of the European Society of Cardiology. Developed incollaboration with the Heart Failure Association of the ESC (HFA) and endorsed bythe European Society of Intensive Care Medicine (ESICM). Eur Heart J 2008;29:2388–442.
[5] Epstein AE, DiMarco JP, Ellenbogen KA, Estes III NA, Freedman RA, Gettes LS, et al.2012 ACCF/AHA/HRS focused update incorporated into the ACCF/AHA/HRS 2008guidelines for device-based therapy of cardiac rhythm abnormalities: a report ofthe American College of Cardiology Foundation/American Heart Association TaskForce on Practice Guidelines and the Heart Rhythm Society. Circulation 2013;127:e283–352.
[6] Rajappan K, Bellenger NG, Anderson L, Pennell DJ. The role of cardiovascularmagnetic resonance in heart failure. Eur J Heart Fail 2000;2:241–52.
[7] Gulati A, Jabbour A, Ismail TF, Guha K, Khwaja J, Raza S, et al. Association of fibrosiswith mortality and sudden cardiac death in patients with nonischemic dilatedcardiomyopathy. JAMA 2013;309:896–908.
[8] Assomull RG, Prasad SK, Lyne J, Smith G, Burman ED, Khan M, et al. Cardiovascularmagnetic resonance, fibrosis, and prognosis in dilated cardiomyopathy. J Am CollCardiol 2006;48:1977–85.
[9] Wu KC, Weiss RG, Thiemann DR, Kitagawa K, Schmidt A, Dalal D, et al. Lategadolinium enhancement by cardiovascular magnetic resonance heralds anadverse prognosis in nonischemic cardiomyopathy. J Am Coll Cardiol 2008;51:2414–21.
[10] Lehrke S, Lossnitzer D, SchobM, Steen H, Merten C, Kemmling H, et al. Use of cardio-vascular magnetic resonance for risk stratification in chronic heart failure: prognos-tic value of late gadolinium enhancement in patients with non-ischaemic dilatedcardiomyopathy. Heart 2011;97:727–32.
[11] Doesch C, Haghi D, Fluchter S, Suselbeck T, Schoenberg SO, Michaely H, et al. Epi-cardial adipose tissue in patients with heart failure. J Cardiovasc Magn Reson2010;12:40.
[12] Doesch C, Streitner F, Bellm S, Suselbeck T, Haghi D, Heggemann F, et al. Epicardialapipose tissue assessed by cardiac magnetic resonance imaging in patients withheart failure due to dilated cardiomyopathy. Obesity (Silver Spring) 2013;21:E253–61.
[13] Doesch C, Suselbeck T, Leweling H, Fluechter S, Haghi D, Schoenberg SO, et al.Bioimpedance analysis parameters and epicardial adipose tissue assessed by cardiacmagnetic resonance imaging in patients with heart failure. Obesity (Silver Spring)2010;18:2326–32.
[14] Richardson P, McKennaW, BristowM, Maisch B, Mautner B, O'Connell J, et al. Reportof the 1995 World Health Organization/International Society and Federation ofCardiology Task Force on the Definition and Classification of cardiomyopathies.Circulation 1996;93:841–2.
[15] Felker GM, Shaw LK, O'Connor CM. A standardized definition of ischemic cardiomy-opathy for use in clinical research. J Am Coll Cardiol 2002;39:210–8.
[16] Huber A, Bauner K, Wintersperger BJ, Reeder SB, Stadie F, Mueller E, et al. Phase-sensitive inversion recovery (PSIR) single-shot TrueFISP for assessment of myocardialinfarction at 3 tesla. Invest Radiol 2006;41:148–53.
[17] Nijveldt R, Germans T, McCann GP, Beek AM, van Rossum AC. Semi-quantitativeassessment of right ventricular function in comparison to a 3D volumetricapproach: a cardiovascular magnetic resonance study. Eur Radiol 2008;18:2399–405.
[18] Fluchter S, Haghi D, Dinter D, HeberleinW, Kuhl HP, NeffW, et al. Volumetric assess-ment of epicardial adipose tissue with cardiovascular magnetic resonance imaging.Obesity (Silver Spring) 2007;15:870–8.
[19] Harrell F. Regression modelling strategies with applications to linear models, logisticregression and survival analysis. New York: Springer; 2001.
[20] Harrell Jr FE, Lee KL, Mark DB. Multivariable prognostic models: issues in developingmodels, evaluating assumptions and adequacy, and measuring and reducing errors.Stat Med 1996;15:361–87.
[21] Bistola V, Parissis JT, Paraskevaidis I, Panou F, Nikolaou M, Ikonomidis I, et al.Prognostic value of tissue Doppler right ventricular systolic and diastolic function in-dexes combined with plasma B-type natriuretic peptide in patients with advancedheart failure secondary to ischemic or idiopathic dilated cardiomyopathy. Am JCardiol 2010;105:249–54.
[22] De Maria R, Gavazzi A, Recalcati F, Baroldi G, De Vita C, Camerini F. Comparison ofclinical findings in idiopathic dilated cardiomyopathy in women versus men. TheItalian Multicenter Cardiomyopathy Study Group (SPIC). Am J Cardiol 1993;72:580–5.
[23] Juilliere Y, Anconina J, Buffet P, Everaere S, Perrin O, Berder V, et al. Relationshipbetween right ventricular ejection fraction and pulmonary pressure in man. ArchMal Coeur Vaiss 1992;85:1305–10.
[24] Sibbald WJ, Driedger AA. Right ventricular function in acute disease states: patho-physiologic considerations. Crit Care Med 1983;11:339–45.
[25] Juilliere Y, Buffet P, Marie PY, Berder V, Danchin N, Cherrier F. Comparison of rightventricular systolic function in idiopathic dilated cardiomyopathy and healedanterior wall myocardial infarction associated with atherosclerotic coronary arterydisease. Am J Cardiol 1994;73:588–90.
[26] Olsen EG. Special investigations of COCM: endomyocardial biopsies (morphologicalanalysis). Postgrad Med J 1978;54:486–93.
[27] Richardson PJ, Olsen EG, Jewitt DE, Oram S. Proceedings: percutaneous technique ofleft ventricular biopsy, and comparison between right and left ventricular myocardialsamples. Br Heart J 1975;37:556.
[28] Juilliere Y, Barbier G, Feldmann L, Grentzinger A, Danchin N, Cherrier F. Addi-tional predictive value of both left and right ventricular ejection fractions onlong-term survival in idiopathic dilated cardiomyopathy. Eur Heart J 1997;18:276–80.
435C. Doesch et al. / International Journal of Cardiology 177 (2014) 429–435
[29] Kaul S, Tei C, Hopkins JM, Shah PM. Assessment of right ventricular function usingtwo-dimensional echocardiography. Am Heart J 1984;107:526–31.
[30] Kjaergaard J, Iversen KK, Akkan D, Moller JE, Kober LV, Torp-Pedersen C, et al.Predictors of right ventricular function as measured by tricuspid annular planesystolic excursion in heart failure. Cardiovasc Ultrasound 2009;7:51.
[31] Hombach V, Merkle N, Torzewski J, Kraus JM, Kunze M, Zimmermann O, et al.Electrocardiographic and cardiacmagnetic resonance imagingparameters as predictorsof a worse outcome in patients with idiopathic dilated cardiomyopathy. Eur Heart J2009;30:2011–8.
[32] Meluzin J, Spinarova L, Dusek L, Toman J, Hude P, Krejci J. Prognostic importance ofthe right ventricular function assessed by Doppler tissue imaging. Eur J Echocardiogr2003;4:262–71.
[33] Lewis JF, Webber JD, Sutton LL, Chesoni S, Curry CL. Discordance in degree of rightand left ventricular dilation in patients with dilated cardiomyopathy: recognitionand clinical implications. J Am Coll Cardiol 1993;21:649–54.
