Author + information
- Received March 9, 2011
- Revision received May 17, 2011
- Accepted May 24, 2011
- Published online August 23, 2011.
- ↵⁎Reprint requests and correspondence:
Dr. Selcuk Adabag, Veterans Administration Medical Center, One Veterans Drive, Minneapolis, Minnesota 55417
Objectives The purpose of this study was to perform a systematic review and meta-analysis of prospective randomized clinical trials of cardiac resynchronization therapy (CRT) versus implantable cardioverter-defibrillator (ICD) in patients with reduced ejection fraction (EF), prolonged QRS interval, and New York Heart Association (NYHA) functional class I to II heart failure (HF).
Background In patients with advanced HF, CRT improves left ventricular (LV) function and reduces mortality and hospitalizations. Recent data suggest that patients with milder HF also benefit from CRT.
Methods A meta-analysis of 5 clinical trials including 4,317 patients with NYHA functional class I/II HF was performed.
Results Average age of patients was 65 years, and 80% were male. Frequency of all-cause mortality for CRT versus ICD was 8% versus 11.5% (risk ratio [RR]: 0.81; 95% confidence interval [CI]: 0.65 to 0.99, p = 0.04); for HF hospitalization, it was 11.6% versus 18.2% (RR: 0.68; 95% CI: 0.59 to 0.79, p < 0.001). Patients assigned to CRT had a significantly greater improvement in LVEF (+5.9% vs. +2.2%, p < 0.001) and LV volume than ICD patients. Among mildly symptomatic (NYHA functional class II) patients, CRT was associated with significantly lower mortality and HF hospitalization (RR: 0.73; 95% CI: 0.64 to 0.83), p < 0.001). In asymptomatic (NYHA functional class I) patients, HF hospitalization risk was lower (RR: 0.57; 95% CI: 0.34 to 0.97, p = 0.04) with CRT; however, there was no difference in mortality. Twelve asymptomatic HF patients needed to be treated with CRT to prevent 1 hospitalization.
Conclusions Cardiac resynchronization therapy decreases all-cause mortality, reduces HF hospitalizations, and improves LVEF in NYHA functional class I/II HF patients. Although there was a reduction in HF hospitalization with CRT for asymptomatic (NYHA functional class I) patients, risks versus benefits have to be carefully considered in this subgroup.
Heart failure (HF) is a major public health problem in the United States, with an estimated 6 million current patients, >500,000 new patients each year, and an expenditure of $40 billion spent annually for treatment (1). Sudden cardiac death and worsening HF are the 2 predominant causes of death in HF patients with left ventricular (LV) dysfunction (2). Over the last 30 years, there has been a remarkable improvement in the survival of HF patients due to medical therapies and the implantable cardioverter-defibrillator (ICD) (3).
Cardiac resynchronization therapy (CRT) has been noted to improve symptoms and LV systolic function and reduce mortality and hospitalizations among patients with moderate to severe HF symptoms, reduced LV ejection fraction (EF), and a wide QRS complex on electrocardiogram (4–9). Recent evidence from randomized clinical trials suggests that the salutary effects of CRT on LV morphology, HF symptoms, hospitalizations, and mortality also extend to patients with mild HF (10–14). However, pooled analysis of CRT effect on patients with New York Heart Association (NYHA) functional class I and II HF has not been done, and CRT effects on subgroups of asymptomatic HF patients has not been examined. Indeed, each individual trial included a small number of such patients. Thus, the purpose of the present investigation was to perform a systematic review and meta-analysis of prospective randomized clinical trials of CRT in patients with NYHA functional class I/II HF, EF ≤40%, and QRS ≥120 ms.
Data sources and study search strategy
We identified randomized clinical trials published in MEDLINE from 1960 through January 2010 using the following keywords: cardiac resynchronization therapy or biventricular pacing. We limited our search criteria to include studies published in the English language and those involving humans. We identified additional studies by searching the website www.clinicaltrials.gov and by hand searching the references cited in relevant publications. The methodology utilized in this study has been published previously (15).
We included prospective randomized clinical trials that: 1) enrolled asymptomatic or mildly symptomatic (NYHA functional class I/II) HF patients with EF ≤40% and QRS duration ≥120 ms; 2) compared CRT versus ICD alone; and 3) reported mortality, heart failure, hospitalization, and LV dimensions or volume. In studies that also included patients with more severe HF, the subset of patients with NYHA functional class I/II symptoms were included in this analysis. When data were missing or incomplete, we contacted the authors/sponsors of the original studies to provide the necessary information.
Our initial search yielded 3,494 citations (Fig. 1). Of these, 3,001 were excluded by title search because of irrelevant content, animal subjects, or publication in a language other than English. The abstracts of the remaining 493 articles were reviewed. Of these, 473 were excluded because the study was not a randomized clinical trial (n = 343), it was a duplicate publication (n = 22), a retrospective analysis (n = 15), the control group did not get ICDs (n = 34), the outcomes of interest in this meta-analysis were not reported (n = 53); and specific study population was included (n = 6). Of the remaining 20 publications, 15 included patients with NYHA functional class III/IV HF. The remaining 5 trials were included in the systematic review (10–14) (Fig. 1).
Data were abstracted by 2 reviewers (S.A. and H.R.) using standardized data extraction forms. Discrepancies between the reviewers were infrequent and resolved by discussion. Abstracted information included study design, methods, patient characteristics, relevant outcomes, and adverse events.
The primary outcome for this systematic review was all-cause mortality. Secondary outcomes were HF events requiring inpatient hospitalization or outpatient treatment with intravenous medications and parameters of LV function and size on echocardiography.
Randomized clinical trials included in the meta-analysis were evaluated according to the Delphi criteria (16) and scored on the basis of the Jadad scale (17). Factors considered in methodological quality included adequacy of randomization method and allocation concealment, similarity of study groups at baseline, specification of eligibility criteria, blinding of the patient, the clinician, and the outcome assessor, presentation of point estimates and measures of variability for the outcome measures, description of withdrawals and dropouts, and application of intention-to-treat analysis (16,17). Heterogeneity among studies was examined with the I2 test (18). We limited publication bias by using a comprehensive electronic and hand-search strategy as well as contacting authors of identified studies.
Data synthesis and statistical analysis
We calculated weighted mean difference and 95% confidence interval (CI) for continuous variables, and weighted risk ratios (RRs) and 95% CI for categorical variables. RRs were defined as the risk of an event for CRT versus ICD alone. Each RR was calculated according to the DerSimonian-Laird random effects model. Automatic “zero cell” correction was used for studies with no events for a particular outcome. The I2 statistic was used to quantify heterogeneity. A value >50% was considered to indicate substantial heterogeneity. Pre-specified subgroup analyses of patients with NYHA functional class II and I symptoms separately were performed. All comparisons were 2-sided, and a p value <0.05 was considered statistically significant. All analyses were performed using STATA software, version 10.1 (STATA Corporation, College Station, Texas).
Study and patient characteristics
The characteristics of the 5 randomized clinical trials that met eligibility criteria are displayed in Table 1. All studies included HF patients with reduced EF, wide QRS interval, and NYHA functional class I or II symptoms, although the entrance criteria for EF and QRS duration varied slightly between studies (Table 1). In 2 studies, which also enrolled patients with more severe HF symptoms (10,14), only the outcomes of participants with NYHA functional class I/II symptoms were analyzed in this meta-analysis. Follow-up ranged from 6 to 40 months. In 3 studies (12–14), all participants underwent LV lead implantation, whereas in 2 studies (10,11) only patients assigned to CRT had LV leads implanted. Four studies performed intention-to-treat analysis and reported adverse events (10–13) (Table 1). All studies reported drop-outs.
The 5 clinical trials included a total of 4,317 HF patients (range 186 to 1,820 patients) with NYHA functional class I/II symptoms, randomly assigned to CRT (n = 2,429) versus ICD therapy (n = 1,888). The pooled baseline characteristics are displayed in Table 2. Mean age of the study participants was 65 years. Overall, 80% were male, 60% had ischemic cardiomyopathy, and 9% were asymptomatic (NYHA functional class I). At baseline, 97% of patients were taking angiotensin-converting enzyme inhibitors or angiotensin-receptor blockers. However, beta-blockers were used in ≥90% of patients in the trials published during or after 2008 (10–12) versus 55% of those published before 2008 (13,14) (Table 2). The baseline characteristics of patients assigned to CRT versus ICD therapy were well balanced in each individual trial (data not shown).
Mortality and HF hospitalization
Mortality and HF events or hospitalization data in NYHA I/II functional class patients were present in all 5 studies (Table 3). Of the 4,317 patients with NYHA functional class I/II HF, 412 (9.5%) died during follow-up. Mortality was lower among patients assigned to CRT versus ICD (8.0% CRT vs. 11.5% ICD; RR: 0.81, 95% CI: 0.65 to 0.99; p = 0.04; I2 = 0). Heart failure events or hospitalizations were also lower among patients assigned to CRT (11.6% CRT vs. 18.2% ICD; RR: 0.68, 95% CI: 0.59 to 0.79; p < 0.001; I2 = 0). There was a 28% reduction in the risk of composite of these outcomes (17.5% CRT vs. 26.4% ICD; RR: 0.72, 95% CI: 0.65 to 0.81; p < 0.001). Overall, 29 patients needed to be treated to prevent 1 death, and 15 needed to be treated to prevent 1 HF hospitalization (Table 3).
LV structure and function
The changes in LV dimensions and EF with CRT versus ICD therapy are displayed in Table 4. Mean increase in EF from baseline was 5.9% (95% CI: 1.7% to 9.9%) in patients assigned to CRT versus 2.2% (95% CI: 1.6% to 2.8%) with ICD (difference CRT vs. ICD 4.1% [95% CI: 0.1% to 8.0%]; p < 0.001]. Reductions in LV end-diastolic dimensions and volume were significantly greater for patients assigned to CRT than for patients assigned to ICD (Table 4).
The 5 clinical trials included 3,915 patients with NYHA II symptoms. Of these, we were granted access to subgroup data for 3,412 patients in 4 studies (10,11,13,14). There was a 22% reduction in mortality (9.6% CRT vs. 13.1% ICD; RR: 0.78, 95% CI: 0.65 to 0.95; p = 0.01; I2 = 0) and a 33% reduction in HF events or hospitalization (14.6% CRT vs. 21.5% ICD; RR: 0.67, 95% CI: 0.57 to 0.79; p < 0.001; I2 = 0) (Table 3).
Asymptomatic (NYHA functional class I) HF patients (n = 402) were included in 3 trials (11,12,14). Of these, we were granted access to subgroup data for 295 patients in 2 studies (11,14). There was a 43% reduction in HF hospitalization between patients assigned to CRT versus ICD (11.9% CRT vs. 20.5% ICD; RR: 0.57, 95% CI: 0.34 to 0.97; p = 0.04). Overall, 12 patients needed to be treated to prevent 1 HF hospitalization (Table 3). However, there was no difference in mortality with CRT among asymptomatic HF patients (6% CRT vs. 20.5% ICD; RR: 0.85, 95% CI: 0.34 to 2.01; p = 0.71) (Table 3).
Adverse events were reported in 4 studies including 4,414 patients (10–13). Pooled results are displayed in Table 5. Overall, 562 (12.7%) patients had an adverse event related to device implantation. The majority of these were LV lead dislodgement and failure to implant the LV lead. Adverse events, such as failure to implant, pneumothorax, pocket hematoma, and infection, were significantly higher among patients assigned to CRT versus ICD based on data from the MADIT-CRT (Multicenter Automatic Defibrillator Implantation Trial with Cardiac Resynchronization Therapy) and RAFT (Resynchronization/Defibrillation for Ambulatory Heart Failure) trials in which the patients in the control group were not implanted with an LV lead (10,11).
In this systematic review of 5 randomized clinical trials involving >4,000 patients with asymptomatic or mildly symptomatic HF (NYHA functional class I/II), reduced EF, and wide QRS complex, CRT was associated with a 19% reduction in mortality and 32% reduction in HF events or hospitalization in comparison with ICD therapy alone. Further, CRT was associated with a significant improvement in LV dimensions, volume, and EF. These benefits were robust for patients with NYHA functional class II symptoms. We also found, for the first time, that there was a significant reduction in HF events or hospitalizations with CRT in asymptomatic HF patients (NYHA functional class I).
A widened QRS complex on 12-lead electrocardiography represents delayed ventricular depolarization, which, particularly in patients with left bundle branch block, results in LV dyssynchrony and increased myocardial strain (19). In patients with myocardial disease, these findings are associated with abnormal cardiac remodeling, leading to increased LV volume and reduced EF. By pacing the LV through an electrode positioned in a coronary vein, CRT creates a more synchronous contraction with partial restoration of the LV systolic wringing and twisting motion (20). This improvement in contractile efficiency with CRT reverses abnormal cardiac remodeling, with improvement in cardiac size and function, and is associated with a reduction in myocardial energy cost and oxygen consumption (21–23).
Previously, among severely symptomatic HF patients (NYHA functional class III/IV) with reduced EF and widened QRS interval, CRT was associated with reverse remodeling of the LV resulting in improved EF, reduced LV dimensions and volume, and reduced mitral regurgitation (7,23–25). These structural changes have, in turn, led to improvement in symptoms, quality of life, and exercise capacity (4–6,14,25,26). Finally, in large clinical trials, utilization of CRT and ICD in addition to optimal medical therapy has resulted in a 39% reduction in HF hospitalization and a 36% reduction in mortality in comparison with ICD alone (8,9).
The present investigation shows that these salutary effects of CRT also extend to patients with mild HF. Further, to our knowledge, the present investigation is the first to show that CRT, in asymptomatic HF patients (NYHA functional class I) with a low EF and wide QRS interval, was associated with a significant reduction the risk of hospitalizations for HF. Previously, in 2 nonrandomized studies, CRT had led to a degree of improvement in LV volume and EF in patients with mild HF similar to that in patients with severe HF, supporting these data (27,28). Further, reverse cardiac remodeling with improvement in the structure and function of the LV with angiotensin-converting enzyme inhibitors was also associated with a reduced mortality and hospitalization rate in asymptomatic HF patients (29,30). Thus, cumulatively, these data support the notion that early administration of CRT delays progression of HF by reversing the LV remodeling process and improves clinical outcomes. However, potential benefits have to be carefully weighed against the risk of adverse events associated with device implantation in these asymptomatic patients.
In this meta-analysis, the mortality benefit associated with CRT among HF patients with NYHA functional class II symptoms was largely driven by the results of the RAFT trial (10). Survival curves in the CRT versus ICD arms in the RAFT study started to diverge after ∼2.5 years of follow-up, underscoring the delayed effect of CRT on mortality. Indeed, the RAFT study had longer follow-up (average 40 months) than the other trials in this meta-analysis, including the MADIT-CRT study (average follow-up 29 months) (Table 1).
Nearly 10% of the patients who participated in the 5 clinical trials that were included in this meta-analysis had device-related adverse events within 30 days of implantation. The most common adverse events were failure to implant the LV lead and LV lead dislodgement. Other adverse events, such as pneumothorax, pocket hematoma, and infection, were more likely to occur in patients assigned to CRT versus ICD. Whether these adverse events were associated with rehospitalizations were unspecified. Adverse events related to device implantation should also be taken into consideration when judging the benefit and cost effectiveness of CRT (31).
The limitations of this systematic review and meta-analysis include the following: The eligibility criteria of each individual trial included in this meta-analysis were slightly different from each other (Table 1). However, the differences were not large enough to prevent pooling of these study populations. Further, whereas 4 CRT trials (10,12–14) reported HF hospitalization as an outcome, the MADIT-CRT trial (11) reported HF events and hospitalizations, which included outpatient IV treatment of HF. However, the results did not change when we analyzed the data including only the HF hospitalizations in the MADIT-CRT study. Also, we were not granted access to the data of NYHA functional class I HF patients who participated in the REVERSE (Resynchronization Reverses Remodeling in Systolic Left Ventricular Dysfunction) trial (12). This reduced the number of NYHA functional class I patients analyzed to 295 from a potential 402, limiting the power of analysis in this subgroup. Finally, we did not have the data to separately examine subgroups such as females, patients with QRS >150 ms, and patients with left bundle branch block who benefited more from CRT in individual trials.
This first meta-analysis of randomized clinical trials of CRT versus ICD demonstrated that CRT improves LV dimensions and reduces mortality and hospitalizations among asymptomatic or mildly symptomatic HF patients (NYHA functional class I/II) with reduced EF and widened QRS complex on electrocardiogram. These benefits were robust for patients with NYHA functional class II symptoms. We also found, for the first time, that there was a significant reduction in HF events and hospitalizations with CRT in asymptomatic HF patients (NYHA functional class I).
This project was funded by a physician-initiated research grant from Boston Scientific Corporation, Arden Hills, Minnesota. Dr. Moss reports receiving a research grant from Boston Scientific to the University of Rochester in support of the MADIT clinical trials. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- confidence interval
- cardiac resynchronization therapy
- ejection fraction
- heart failure
- implantable cardioverter-defibrillator
- left ventricular
- New York Heart Association
- risk ratio
- Received March 9, 2011.
- Revision received May 17, 2011.
- Accepted May 24, 2011.
- American College of Cardiology Foundation
- Roger V.L.,
- Go A.S.,
- Lloyd-Jones D.M.,
- et al.
- Hunt S.A.,
- Abraham W.T.,
- Chin M.H.,
- et al.
- Sutton M.G.,
- Plappert T.,
- Hilpisch K.E.,
- Abraham W.T.,
- Hayes D.L.,
- Chinchoy E.
- Linde C.,
- Abraham W.T.,
- Gold M.R.,
- et al.
- Abraham W.T.,
- Young J.B.,
- León A.R.,
- et al.
- Higgins S.L.,
- Hummel J.D.,
- Niazi I.K.,
- et al.
- Adabag A.S.,
- Ishani A.,
- Bloomfield H.E.,
- Ngo A.K.,
- Wilt T.J.
- Higgins J.P.,
- Thompson S.G.,
- Deeks J.J.,
- Altman D.G.
- Bertini M.,
- Marsan N.A.,
- Delgado V.,
- et al.
- Solomon S.D.,
- Foster E.,
- Bourgoun M.,
- et al.
- St. John Sutton M.G.,
- Plappert T.,
- Abraham W.T.,
- et al.
- Nelson G.S.,
- Berger R.D.,
- Fetics B.J.,
- et al.
- Ypenburg C.,
- Lancellotti P.,
- Tops L.F.,
- et al.
- Auricchio A.,
- Stellbrink C.,
- Sack S.,
- et al.
- Bleeker G.B.,
- Holman E.R.,
- Steendijk P.,
- et al.
- Pfeffer M.A.,
- Braunwald E.,
- Moyé L.A.,
- et al.
- Linde C.,
- Mealing S.,
- Hawkins N.,
- et al.