Author + information
- Received July 30, 2018
- Revision received November 5, 2018
- Accepted November 5, 2018
- Published online March 25, 2019.
- Fadi Shamoun, MDa,∗ (, )
- Teresa De Marco, MDb,
- David DeMets, PhDc,
- Chaoqun Mei, MSc,
- JoAnn Lindenfeld, MDd,
- Leslie A. Saxon, MDe,
- John P. Boehmer, MDf,
- Jill Leigh, BSg,
- Patrick Yong, MSEEh,
- Arthur M. Feldman, MD, PhDi and
- Michael R. Bristow, MD, PhDi
- aDepartment of Cardiovascular Diseases, Mayo Clinic Arizona, Scottsdale, Arizona
- bDepartment of Medicine, Division of Cardiology, University of California Medical Center, San Francisco, California
- cStatistical Data Analysis Center, University of Wisconsin Madison, Wisconsin
- dDepartment of Medicine, Division of Cardiology, Vanderbilt University, Nashville, Tennessee
- eDepartment of Medicine, Division of Cardiology, University of Southern California, Los Angeles, California
- fDepartment of Medicine, Division of Cardiology, Pennsylvania State University Medical Center, Hershey, Pennsylvania
- gBoston Scientific, St. Paul, Minnesota
- hDepartment of Medicine, Division of Cardiology, Temple University, Philadelphia, Pennsylvania
- iDepartment of Medicine, Division of Cardiology and Cardiovascular Institute, University of Colorado, Boulder and Aurora, Colorado
- ↵∗Address for correspondence:
Dr. Fadi Shamoun, Department of Cardiovascular Diseases, Mayo Clinic Arizona, 13400 East Shea Boulevard, Scottsdale, Arizona 85259.
Objectives This study tested the hypothesis that the extent of left ventricular (LV) eccentric structural remodeling in heart failure with reduced ejection fraction (HFrEF) is directly associated with clinical event responses to cardiac resynchronization therapy (CRT).
Background Whether the severity of LV structural remodeling influences CRT treatment effects is unknown.
Methods COMPANION (Comparison of Medical Therapy, Pacing and Defibrillation in Heart Failure) trial data were analyzed retrospectively. Left ventricular internal dimensions at end diastole indexed by body surface area (LVEDDI) were measured pre-randomization by 2-dimensional echocardiography. LVEDDI values were stratified around the median value of 35 mm/m2, and CRT (including CRT-P [CRT with only pacing capability] and/or CRT-D [CRT with an implantable defibrillator]) treatment effects were assessed and compared by LVEDDI group. Patients assigned to these treatments were compared to those undergoing optimal pharmacologic therapy (OPT) for the outcomes of all-cause mortality (ACM) or ACM and heart-failure hospitalization (ACM/HFH).
Results In the LVEDDI ≥35 mm/m2 group (n = 614), CRT vs. OPT was associated with a lower ACM/HFH hazard ratio (HR) (HR: 0.53; 95% confidence interval [CI]: 0.40 to 0.70; p <0.001), whereas in the LVEDDI <35 mm/m2 group, the CRT vs. OPT ACM/HFH hazard ratio was not statistically significant (HR: 0.80; 95% CI: 0.59 to 1.08; p = 0.15). For ACM alone, in the LVEDDI ≥35 mm/m2 group, the hazard ratio for CRT-P was 0.59 (95% CI: 0.39 to 0.90; p = 0.012) and for CRT-D 0.50 (95% CI: 0.32 to 0.77; p = 0.002). Neither of the CRT groups showed a statistically significant reduction in ACM in the LVEDDI <35 mm/m2 group.
Conclusions Larger versus smaller LVEDDIs are associated with a reduction in ACM with CRT-P or CRT-D treatment, and with a more effective reduction in ACM/HFH for the combined CRT treatment groups.
Heart failure with reduced left ventricular ejection fraction (HFrEF) is a complex syndrome characterized by high morbidity and mortality. Cardiac resynchronization therapy (CRT) is a generally effective treatment for HFrEF patients with intraventricular conduction delays (IVCDs), but in 20% to 30% of patients, CRT fails to reverse remodeling, to alter clinical outcomes, or to favorably impact cardiac dyssynchrony (1–7). CRT correction of the abnormality that results from delayed activation of the left ventricle (LV) coordinates septal and free wall contraction, resulting in improved chamber contractility and myocardial efficiency, as well as a reversal of remodeling (1,8–11). However, it is unclear if the benefit of CRT treatment is proportional to the degree of underlying pathologic eccentric remodeling (defined as increased end-diastolic LV internal dimension or volume with an LV ejection fraction [EF] of ≤0.35), reflecting more to gain from higher clinical event rates, or if patients with less advanced remodeling benefit more, as might be expected from a more easily reversible pathophysiologic process.
There are 2 types of CRT devices: CRT-P describes a device with biventricular pacing only, whereas CRT-D is a device with biventricular pacing and defibrillation capacity. In patients with advanced HFrEF, CRT-P reduces the combination of mortality and hospitalizations (4), and CRT-D adds the additional benefit of protecting against sudden cardiac death and is more effective in reducing all-cause mortality (ACM) (4). In less advanced cases of HFrEF, CRT-P reduces ACM (5). Despite wide adoption of CRT, a significant number of patients still lack apparent benefit (12,13), and consequently, there has been considerable investigation attempting to identify variables that can predict clinical response (13). Regarding these possible predictors, the relationship between the degree of LV eccentric remodeling and CRT response is unclear. Uncontrolled studies have reported that less baseline remodeling or a smaller LV size is associated with more favorable remodeling (14,15) or clinical (15) response, whereas other studies have reported no obvious relationship between LV diastolic dimension or volume and subsequent CRT-associated reverse remodeling (16,17) or clinical endpoint effects (18). In cases of advanced heart failure, quantitative information about the relationship between the degree of baseline remodeling and CRT clinical response has not been previously reported in a controlled study. Accordingly, using data from a pharmacologic therapy–controlled trial, the current investigation sought to determine whether the degree of eccentric structural remodeling could limit CRT clinical responses.
There were 1,520 patients enrolled in the COMPANION (Comparison of Medical Therapy, Pacing and Defibrillation in Heart Failure) trial, all of whom provided written informed consent, approved by the corresponding center’s institutional review board (4). The COMPANION trial included patients with the New York Heart Association (NYHA) functional classes III and IV, a QRS duration of more than 120 ms, and an EF of ≤35%. The trial was unblinded because there were 2 device arms that were compared to an optimal pharmacologic therapy (OPT) arm. Patients were randomized in a 1:2:2 ratio to receive OPT or CRT-P or CRT-D treatment. OPT, used in all groups, consisted of stable doses of diuretic drugs (unless the drugs were not needed), angiotensin-converting enzyme inhibitors or angiotensin-receptor blockers (unless they were not tolerated), beta-blockers (unless they were not tolerated or contraindicated), and spironolactone (unless it was not tolerated or contraindicated). Digoxin and other medications used to treat heart failure were given at the investigator's discretion (4).
For the purposes of the present investigation, eccentric remodeling was defined as the COMPANION trial HFrEF entry criteria consisting of an increased end-diastolic LV internal dimension of ≥60 mm or >30 mm/m2, as measured by echocardiography, coupled with a decreased LVEF of ≤0.35%. Because LV mass and wall thickness measurements were not collected in the COMPANION cardiac imaging case report forms, the operational term eccentric remodeling, often used to delineate increased LV volume and decreased LVEF in the absence of an increase in LV mass (19), is agnostic to the presence of hypertrophy and is used in a more general sense to define the pathologic changes in LV geometry and function (20) that are characteristic of advanced HFrEF. In the present post hoc analysis, the primary remodeling independent variable was LV end diastolic dimension indexed to body surface area (LVEDDI), a measurement of eccentric structural remodeling (defined as increased LV end-diastolic volume or internal dimension).
The primary clinical endpoint used to assess outcomes was the composite of time to ACM or first heart failure hospitalization (ACM/HFH) (4), a derivative of the main protocol primary endpoint that is more specific to heart failure. For this endpoint, the CRT-P and CRT-D arms were combined, as in the entire cohort (4) and in the current dataset the hazard ratios (HR) were similar for the 2 treatment groups. In the current analysis, ACM was considered a secondary clinical endpoint, analyzed within each CRT group separately because of differences in outcomes in the entire cohort (4).
In the COMPANION trial, it was necessary to collect the qualifying LVEF data within 3 months of randomization, and the current analysis used LVEDDI (required for patients qualifying by echocardiographic measurement of LVEF) and LVEF data collected by 2-dimensional (2D) echocardiography. Body surface area was calculated by using the formula: square root [(height in cm) × (weight in kg)]/3,600. Of the 1,520 patients with qualifying LVEF measurements, echocardiograms provided the information for 1,266 patients, and other imaging techniques were used for qualifying 254 patients. Data adequate for calculating LVEDD were available in 1,260 2D echocardiograms. Baseline characteristics in the LVEDDI cohort were very similar to those in the 260 patients who had no evaluable echocardiographic data (Table 1). Baseline characteristics for the 3 treatment arms of the 1,260 LVEDDI cohort were recorded for 253 patients in the OPT arm, 521 patients in the CRT-P arm, and 486 patients in the CRT-D arm (Online Table 1) and were not significantly different among the groups. A CONSORT (Consolidated Standards of Reporting Trials) flow diagram of the 1,520 patients in the COMPANION trial who were randomized to the 1,260 LVEDDI-patient cohort is shown in Figure 1.
To evaluate the association between the degree of structural remodeling and the CRT response, the LVEDDI cohort was divided into 2 groups based on the median measurement of 35 mm/m2. The groups were individually analyzed by using previously described methodology (4). The Wilcoxon rank-sum test was used for continuous variables, and the Pearson chi-square test was used for categorical data. Unless otherwise specified, all p values are 2-sided and nominal, with <0.05 considered statistically significant. HRs for the endpoints of ACM and ACM/HFH were based on time to a first event, with the time to an event plotted according to the Kaplan-Meier method and differences among groups determined by the log-rank statistic. CRT-P and CRT-D data were combined for the ACM/HFH endpoint, as were functional outcomes of the 6-min walk, quality of life survey scores, and NYHA functional classes (4). Functional outcomes were measured at 6 months, as previously described (4), and clinical event data were used in the full study follow-up period (4). Forest plots depicting HRs and their 95% confidence intervals (CIs) were plotted within each group.
For the analysis of LVEDDI as a continuous variable related to clinical outcomes in the CRT treatment groups, a Cox proportional hazards model was used to adjust for variables that were statistically significant between the 2 LVEDDI groups. Height and weight were not included in the model because they were incorporated in the LVEDDI calculation. Including the 2 CRT treatment arms, 15 covariates were evaluated in the model. R software (R Core Team, Vienna Austria) was used for all statistical analyses.
The baseline characteristics of the 2 LVEDDI groups are given in Table 2. In the LVEDDI <35 mm/m2 group, there were more males, higher weight, more NYHA functional class III patients, shorter duration of heart failure, higher proportions with ischemic cardiomyopathy, fewer patients with left bundle branch block, and more occurrences of diabetes.
The median follow-up was 15.7 months for the <35 mm/m2 group and 15.2 months for the ≥35 mm/m2 group. Event rates in the ≥35 mm/m2 LVEDDI group were higher than those in the <35 mm/m2 group, respectively (events/patient), for ACM/HFH, 0.58 in OPT or 0.43 in the entire group versus 0.39 in OPT or 0.20 in the entire group; and for ACM alone, 0.32 in OPT or 0.23 in the entire group versus 0.20 in OPT or 0.19 in the entire group. For the entire LVEDDI patient population HRs for ACM/HFH were similar to those for in the 2 CRT groups (CRT-P vs. OPT HR: 0.65, CRT-D vs. OPT HR: 0.64; CRT-D vs. CRT-P HR: 0.98; 95% CI: 0.80 to 1.20). CRT was associated with improved ACM/HFH in the entire LVEDDI cohort (HR: 0.65; 95% CI: 0.53 to 0.80; p < 0.001) (Figure 2). In the LVEDDI <35 mm/m2 group, CRT was associated with a statistically nonsignificant effect compared to OPT (HR: 0.80; 95% CI: 0.59 to 1.08; p = 0.15) (Figure 3A). However, in the LVEDDI ≥35 mm/m2 group, CRT was associated with a larger and statistically significant treatment effect compared to OPT (HR: 0.53; 95% CI: 0.40 to 0.70; p < 0.001) (Figure 3B). The interaction p value between the CRT treatment effects (HR) in the 2 LVEDDI groups was 0.054.
As previously reported (4), there was a preferential mortality benefit in the CRT-D treatment arm compared to the CRT-P and OPT treatments in the entire COMPANION cohort. For that reason, ACM was evaluated across the 3 arms in the 2 LVEDDI groups, as shown in Figure 4A for the LVEDDI <35 mm/m2 group and in Figure 4B for the LVEDDI ≥35 mm/m2 group. For the LVEDDI <35 mm/m2 group, CRT-P had a hazard ratio of 0.95 (95% CI: 0.60 to 1.51; p = 0.83) compared to 0.86 (95% CI: 0.53 to 1.40; p = 0.54) for CRT-D (Figure 4A). In contrast, the LVEDDI ≥ 35 mm/m2 group (Figure 4B) showed HRs of 0.59 (95% CI: 0.39 to 0.90; p = 0.012) for CRT-P and 0.51 (95% CI: 0.33 to 0.78; p = 0.002) for CRT-D. Because the CRT-P and CRT-D HRs were similar in the 2 LVEDDI groups and the number of events was fewer for ACM/HFH, in order to maximize power, results from the 2 CRT arms were combined for the interaction test that compared treatment effects. This combination resulted in an interaction p value of 0.085 between LVEDDI groups. Figure 5 summarizes these comparisons by using a forest plot for the 2 treatment groups and outcomes of ACM or ACM/HFH. Compared to the <35 mm/m2 patients, there were consistently lower point estimates for the CRT, CRT-P, and CRT-D treatment arm HRs in the LVEDDI ≥35 mm/m2 group.
A Cox proportional hazards model was used (Table 3) to evaluate the effect of LVEDDI as a continuous variable relative to the ACM/HFH and ACM endpoints, adjusting for all variables in Table 2 that had p values of <0.05, and CRT-P or CRT-D treatment effect. For ACM/HFH and ACM, the LVEDDI post-adjustment p values were 0.010 and 0.0005, respectively (Table 3). The Cox model coefficient was positive, meaning that larger LVEDDIs were associated with greater ACM or ACM/HFH event rates after adjustments were made for other variables that were different between the 2 LVEDDI groups. Other variables that were independently related to both ACM/HFH and ACM event rates were NYHA functional class IV, CRT-D treatment, and female sex. Multiple variables were related to ACM/HFH and not to ACM, likely due to greater statistical power for the ACM/HFH endpoint. These variables included negative associations of ACM/HFH with LVEF, CRT-P treatment, ischemic versus nonischemic cause, beta-blocker therapy, and positive relationships among with heart failure duration, QRS length, and history of atrial fibrillation or flutter. Left bundle branch block, a known univariate predictor of CRT response (4,19), was not significantly related to either of the clinical outcomes in the covariate adjusted analysis.
The respective statistically significant covariates shown in Table 3 were used in a Cox model to adjust the CRT/OPT HRs within the LVEDDI groups, shown in Figures 3 and 4. There were no appreciable changes compared to the unadjusted values, except for CRT-D vs. OPT and the ACM endpoint, where the interaction p value between LVEDDI groups achieved statistical significance (p = 0.034; LVEDDI <35 mm/m2 HR: 0.94; 95% CI: 0.58 to 1.52; vs. ≥35 mm/m2 HR: 0.47; 95% CI: 0.30 to 0.72).
6-MIN walk and NYHA outcomes
Functional outcomes (defined as results from the 6-min walk test, quality of life scores, and functional status) were evaluated in both LVEDDI groups. Although differences in improvements in these functional outcomes were not significant between LVEDDI groups, there was a trend toward more favorable effects for CRT in the LVEDDI >35 mm/m2 group (Table 4). Compared to placebo, the 6-min walk distances improved, on average, by 46 m in the CRT arm compared to those in the OPT arm within the LVEDDI ≥35 mm/m2 group and compared to a 36-m improvement in the <35 mm/m2 group. Additionally, 29% of subjects in the LVEDDI ≥35 mm/m2 group showed improvements in NYHA functional class compared with those receiving OPT and compared to 14% in the <35 mm/m2 patients (interaction p = 0.092).
Analysis of the COMPANION trial data demonstrates that LV structural eccentric remodeling at baseline, as assessed by echocardiographic measurements of LVEDDI, is a reliable marker for prediction of clinical response to CRT. A value more than or equal to the median LVEDDI value of 35 mm/m2, indicative of advanced LV eccentric structural remodeling, was compared to a less advanced remodeling value below the LVEDDI median. The LVEDDI cohort was derived from the COMPANION parent population consisting of patients with advanced NYHA functional class III/IV heart failure with LVEF ≤35%, LVEDDI ≥60 mm or >30 mm/m2, and QRS lengths of ≥120 ms, which defines a high-risk, pronounced eccentric remodeling patient population with IVCDs, whose annualized mortality in the OPT group was 18.5% (11). Echocardiographic data were available for 1,260 individuals who constituted the analysis cohort; however, baseline characteristics for the 260 individuals who were not included were very similar, indicating that the LVEDDI cohort was likely highly representative of the COMPANION parent population.
Compared to the smaller LVEDDI group, the group with the larger LV size showed that CRT treatment had a greater treatment effect on ACM/HFH (respective HRs of 0.80 [95% CI: 0.59 to 1.08] and 0.53 [95% CI: 0.40 to 0.70; interaction p = 0.054]). Differential effects on ACM were even more pronounced, with HRs of 0.59 (95% CI: 0.39 to 0.90) for CRT-P and 0.51 (95% CI: 0.33 to 0.78) for CRT-D in the ≥35 mm/m2 group and nonsignificant HRs in the smaller LVEDDI group (HR: 0.95; 95% CI: 0.59 to 1.51) for CRT-P, and HR of 0.86 (95% CI: 0.53 to 1.40) for CRT-D, with an interaction p value of 0.085 for the combined CRT groups.
Several baseline characteristics were different between the 2 LVEDDI groups. Factors that could have been associated with an enhanced treatment effect in the ≥35 mm/m2 group included less incidence of diabetes, a greater percentage of left bundle branch block, a wider QRS length, and less ischemic cardiomyopathy. On the other hand, the ≥35 mm/m2 group was also associated with multiple factors that might have been expected to blunt a therapeutic effect (older patient population, more females, more NYHA functional class IV patients, longer duration of HF, lower LVEF, and lower blood pressure). Because of the imbalance in these characteristics, a covariate adjusted model was used to analyze LVEDDI as a continuous variable for predicting ACM/HFH. After adjustments were made for significant other baseline characteristics, LVEDDI was revealed as a significant, positive predictor of both ACM/HFH and ACM. Other significant positive predictors of ACM/HFH combined and ACM alone were NYHA functional class IV, absence of CRT-D treatment, and male sex. In the covariate adjusted model, LVEF, which has some inherent advantages compared to end-diastolic volume as a remodeling index (20), was independently and negatively associated with the ACM/HFH endpoint but not with ACM. This likely indicates that the functional component of LVEF, that is, the stroke volume term in the numerator of the LVEF equation, has some impact in the ACM/HFH multivariate analysis beyond LV end-diastolic volume or its surrogate LVEDDI. Although LVEDDI was superior to LVEF as a predictor of ACM, that could be related to the simplicity of the former’s measurement, for example, 1 versus 2 internal diameter calculations. A more likely explanation is that eccentric structural remodeling per se, as opposed to a measurement of structural remodeling and function, is a key determinant of mortality risk in HFrEF patients eligible for CRT. Considered together, these data support the extent of structural remodeling and LVEDDI particularly, as an independent predictor of clinical outcomes in HFrEF patients eligible for CRT treatment.
The increased risk of ACM/HFH combined or ACM alone, conferred by larger LVEDDIs, is presumably the basis for the greater effectiveness of CRT in patients with greater degrees of structural remodeling. CRT has been shown to reverse the eccentric remodeling process (21–23), similar to beta-blocker therapy (11), and it is logical that such an effect would have the greatest therapeutic impact in patients at the highest clinical risk. In the extensively investigated cohorts of HFrEF patients with moderate to severe LV dysfunction and eccentric remodeling who meet criteria for CRT (21–23), there has been no evidence that patients with the most advanced degrees of structural remodeling do not experience therapeutic reverse remodeling.
The potential importance of the extent of eccentric structural remodeling in CRT treatment effects has not been previously detected in large trials using pharmacologic (5) or implantable cardioverter-defibrillator and pharmacologic controls (18,24). The obvious differences between the current study and the CARE-HF (Cardiac Resynchronization-Heart Failure) trial (5) or the MADIT-CRT (Multicenter Automatic Defibrillator Implantation Trial–Cardiac Resynchronization Therapy) trial (18,24) are the extent of remodeling and the severity of heart failure. In the COMPANION trial, the LVEDDI cohort's median LVEF was 0.20 compared to 0.25 in the CARE-HF trial (5) and a mean of 0.24 in the MADIT-CRT trial (18). The NYHA functional class distributions were 86% functional class III and 14% functional class IV in the COMPANION LVEDDI cohorts; 94% functional class III and 6% functional class IV in the CARE-HF cohort (5); and 85% functional class II with no functional class III or IV, in the MADIT-CRT cohort (18). In the LVEDDI cohorts, there were 651 patients (52% of the total) with an LVEF ≤0.20% compared to approximately 200 patients in the CARE-HF trial (25% of the total) with an LVEF ≤0.21% (5), illustrating the differences in proportions of patients with advanced eccentric remodeling in the 2 trials, despite their having identical LVEF entry criteria (≤0.35%). It is likely that the inclusion of relatively more patients with advanced eccentric structural remodeling in the COMPANION trial provided the statistical advantage that enabled detection of a direct relationship between remodeling and CRT therapeutic clinical outcomes.
This study was a retrospective analysis comparing CRT treatment with OPT treatment, and baseline LVEDDI was not a pre-specified covariate for predicting response. There were no follow-up echocardiograms obtained in the COMPANION trial, so no assessment of the effects of reverse remodeling on clinical endpoints was possible. Echocardiographic data were missing for 260 patients; however, their baseline characteristics did not differ in ways that were clinically important from those of the LVEDDI cohort. Another issue was, as expected, the event rate in the LVEDDI <35 mm/m2 group was lower than that in the ≥35 mm/m2 group. This difference resulted in slightly lower statistical power within the group with less remodeling, which could have contributed being unable to detect a treatment effect for either clinical endpoint in the <35 mm/m2 group.
In advanced HFrEF patients with IVCDs meriting CRT treatment, those with greater degrees of eccentric structural remodeling have an enhanced benefit for prevention of ACM/HFH and an even more pronounced effect for ACM reduction. CRT-P treatment, which was not associated with a statistically significant reduction in ACM in the COMPANION parent population (4), exhibited a substantial (41%; p = 0.012) reduction in ACM in the more pronounced remodeling group compared with no statistically significant benefit in the LVEDDI <35 mm/m2 group. The degree of LV remodeling should be considered among the standard available data that predict CRT clinical response.
COMPETENCY IN MEDICAL KNOWLEDGE: The clinical relevance of this study is that patients with HFrEF, QRS lengthening, and LV electrical or mechanical dyssynchrony who have advanced degrees of eccentric remodeling may respond favorably to CRT and perhaps even better than patients with less advanced remodeling.
TRANSLATIONAL OUTLOOK: The finding that CRT response appears to be greater with more pronounced degrees of remodeling may relate to the effects of CRT-associated improved pump function having a relatively greater effect on attenuating adverse neurohormonal signaling or wall stress-activated pathways when remodeling is advanced. This hypothesis could be further tested in HFrEF patients or could be “reverse translated” in animal models.
The authors thank Laura Hofstatter and Rachel Rosenberg for manuscript preparation and handling. The authors also thank Ed Gill, MD, for assistance with echocardiographic data and methods.
Supported by Boston Scientific and by the Statistical Data Analysis Center, University of Wisconsin, Madison, Wisconsin. Analysis of data was supported by Statistical Data Analysis Center, University of Wisconsin. Dr. De Marco has received consulting fees from Boston Scientific; and is a speaker for Novartis. Dr. Lindenfield has consultant relationships with Abbott Lab, Edwards Life Science, Novartis, Boston Scientific, VWave, Relypsa, ResMed, CVRX, and Cardiotronix. Dr. Saxon is Executive Director, University of Southern California Center for Body Computing; and has received fees for consulting for Abbott Lab. Dr. Boehmer has received fees for consulting from Boston Scientific and Medtronic; and has received research support from Boston Scientific and Abbott Lab. Ms. Leigh is an employee of Boston Scientific. Mr. Yong is an employee of Boston Scientific. Dr. Bristow is a compensated Director of ARCA biopharma; and holds equity in ARCA biopharma. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose. Barry Greenberg, MD, served as Guest Editor for this article.
- Abbreviations and Acronyms
- all-cause hospitalization
- all-cause mortality
- cardiac resynchronization therapy
- CRT with an implantable defibrillator
- CRT with only pacing capability
- heart failure
- heart failure hospitalization
- HF with reduced LV ejection fraction
- implantable cardioverter-defibrillator
- intraventricular conduction delay
- left ventricle
- left ventricular end diastolic dimension, determined by 2D echocardiography, indexed for body surface area
- left ventricular ejection fraction
- left ventricular end systolic volume
- New York Heart Association
- Received July 30, 2018.
- Revision received November 5, 2018.
- Accepted November 5, 2018.
- 2019 American College of Cardiology Foundation
- Auricchio A.,
- Stellbrink C.,
- Block M.,
- et al.
- Leclercq C.,
- Hare J.M.
- Kass D.A.
- Ermis C.,
- Seutter R.,
- Zhu A.X.,
- et al.
- Auricchio A.,
- Stellbrink C.,
- Sack S.,
- et al.
- Bristow M.R.,
- Saxon L.A.,
- Feldman A.M.,
- Mei C.,
- Anderson S.A.,
- DeMets D.L.
- Lindenfeld J.,
- Feldman A.M.,
- Saxon L.,
- et al.
- Poole J.E.,
- Singh J.P.,
- Birgersdotter-Green U.
- Rinkuniene D.,
- Bucyte S.,
- Ceseviciute K.,
- et al.
- Agacdiken Agir A.,
- Celikyurt U.,
- Sahin T.,
- et al.
- Gaasch W.H.,
- Zile M.R.
- Bristow M.R.,
- Kao D.P.,
- Breathett K.K.,
- et al.
- St John Sutton M.G.,
- Plappert T.,
- Abraham W.T.,
- et al.
- Solomon S.D.,
- Foster E.,
- Bourgoun M.,
- et al.