Author + information
- Received May 29, 2009
- Revision received July 8, 2009
- Accepted August 3, 2009
- Published online November 10, 2009.
- Claude Daubert, MD⁎,⁎ (, )
- Michael R. Gold, MD, PhD†,
- William T. Abraham, MD, PhD‡,
- Stefano Ghio, MD§,
- Christian Hassager, MD, PhD∥,
- Grahame Goode, MD¶,
- Tamás Szili-Török, MD#,
- Cecilia Linde, MD, PhD⁎⁎,
- REVERSE Study Group
- ↵⁎Reprint requests and correspondence:
Dr. Claude Daubert, Service de cardiologie et maladies vasculaires, Centre cardio-pneumologique, Hôpital Pontchaillou-CHU, Rennes 35033, France
Objectives The aim of this study was to determine the long-term effects of cardiac resynchronization therapy (CRT) in the European cohort of patients enrolled in the REVERSE (Resynchronization Reverses Remodeling in Systolic Left Ventricular Dysfunction) trial.
Background Previous data suggest that CRT slows disease progression and improves the outcomes of asymptomatic or mildly symptomatic patients with left ventricular (LV) dysfunction and a wide QRS complex.
Methods We randomly assigned 262 recipients of CRT pacemakers or defibrillators, with QRS ≥120 ms and LV ejection fraction ≤40% to active (CRT ON; n = 180) versus control (CRT OFF; n = 82) treatment, for 24 months. Mean baseline LV ejection fraction was 28.0%. All patients were in sinus rhythm and receiving optimal medical therapy. The primary study end point was the proportion worsened by the heart failure (HF) clinical composite response. The main secondary study end point was left ventricular end-systolic volume index (LVESVi).
Results In the CRT ON group, 19% of patients were worsened versus 34% in the CRT OFF group (p = 0.01). The LVESVi decreased by a mean of 27.5 ± 31.8 ml/m2in the CRT ON group versus 2.7 ± 25.8 ml/m2in the CRT OFF group (p < 0.0001). Time to first HF hospital stay or death (hazard ratio: 0.38; p = 0.003) was significantly delayed by CRT.
Conclusions After 24 months of CRT, and compared with those of control subjects, clinical outcomes and LV function were improved and LV dimensions were decreased in this patient population in New York Heart Association functional classes I or II. These observations suggest that CRT prevents the progression of disease in patients with asymptomatic or mildly symptomatic LV dysfunction. (REsynchronization reVErses Remodeling in Systolic Left vEntricular Dysfunction [REVERSE]; NCT00271154)
- biventricular stimulation
- cardiac resynchronization therapy
- heart failure
- randomized controlled trial
- reverse cardiac remodeling
Cardiac resynchronization therapy (CRT), alone or in combination with an implantable cardioverter-defibrillator (ICD), is recommended as a class I indication level of evidence A for patients presenting in New York Heart Association (NYHA) heart failure (HF) functional class III or IV, a wide QRS, and left ventricular (LV) dysfunction (1,2). In this population, CRT alleviates symptoms and lowers major HF morbidity, all-cause mortality (3–7), and the risk of sudden death (8). Cardiac resynchronization therapy also induces progressive changes in LV structure and function, consistent with reverse remodeling, most prominent in patients with nonischemic heart disease (9,10).
On the basis of these observations and the results of previous small-size studies of patients in NYHA functional classes I and II (11,12), we hypothesized that CRT is also beneficial in mildly symptomatic or asymptomatic patients with LV dysfunction and markers of cardiac dyssynchrony, primarily by preventing the progression of disease over the long term (10). The randomized, double-blind REVERSE (Resynchronization Reverses Remodeling in Systolic Left Ventricular Dysfunction) trial was designed to test this hypothesis (13). The primary end point of the study was the proportion of patients worsened according to the 12-month clinical composite response, which was not significantly different between the actively treated (CRT ON) group and the control (CRT OFF) group (p = 0.10). A key secondary observation made in the first 12 months was a significantly greater degree of reverse ventricular remodeling in the actively treated than in the control group (14). All patients enrolled in the trial participated in the first 12 months of the study, at which time the North American patients learned their randomization assignment. However, the European cohort was prospectively randomized for 24 months. We report these long-term follow-up results in this article.
The present report refers to the European cohort of patients enrolled in the REVERSE trial. Patients eligible for the trial presented in American College of Cardiology/American Heart Association stage C (2), NYHA functional class I (previously symptomatic, currently asymptomatic) or class II (mildly symptomatic), for at least 3 months before enrollment. They were in sinus rhythm, with a 120-ms or longer QRS duration, a left ventricular ejection fraction (LVEF) ≤40%, and a 55-mm or wider LV end-diastolic diameter, measured by echocardiography. All patients were being treated with optimal medical therapy for HF (1) including stable doses of an angiotensin-converting enzyme inhibitor or angiotensin II receptor blocker and a beta-adrenergic blocker for at least 3 months. Patients that were, within 3 months before enrollment, in NYHA functional class III or IV or were hospitalized for HF were excluded from the study. Patients in need of permanent cardiac pacing or presenting with permanent or persistent atrial tachyarrhythmias were also excluded. Other exclusion criteria were published previously (13). The ethics committee of each participating center approved the study protocol, and all patients granted their written informed consent.
Study design and procedures
Patients meeting the study inclusion criteria underwent a baseline evaluation, including NYHA functional classification, quality-of–life assessment by the Minnesota Living with Heart Failure questionnaire and the Kansas City Cardiomyopathy Questionnaire (15), a 6-min hall-walk test, a 12-lead electrocardiogram for measurement of QRS duration, and a 2-dimensional Doppler-flow echocardiogram to assess the ventricular structure and function and the degree of mitral regurgitation. After this evaluation, the patients underwent implantation of CRT pacemaker or CRT defibrillator systems, depending on ICD indication. Patients who had undergone successful implantation were randomly assigned in a 2-to-1 scheme to a CRT ON or CRT OFF group for 24 months. Randomization occurred in permuted blocks within each center. The device of patients assigned to CRT ON was programmed to pace both ventricles and inhibit atrial pacing, unless the intrinsic rate was 35 beats/min or less. The pulse generator of patients assigned to CRT OFF was programmed to inhibit atrial or ventricular pacing, unless the intrinsic rate was 35 beats/min or less. The atrioventricular delay was optimized by echocardiography in all patients, regardless of the randomization assignment, before the final programming (13). For patients assigned to CRT ON, the atrioventricular delay remained unchanged from the time of their discharge from the hospital until the end of the study, unless the LV lead needed to be revised. The patients were followed at 1, 3, 6, 12, 18, and 24 months.
The physicians caring for the patients and responsible for the management of HF when patients were hospitalized were unaware of the pacing mode. Quality-of-life questionnaires, patient global assessments, NYHA functional classification, and 6-min walk tests were performed by observers unaware of the random treatment assignment. Device interrogations and programming, 12-lead electrocardiograms, echocardiograms, and pacemaker follow-ups were performed by unblinded observers.
The investigators were instructed to report all adverse events, which were adjudicated by an independent Adverse Event Advisory/Endpoint Committee unaware of the treatment assignment. An unblinded, independent Data Monitoring Committee reviewed the cumulative adverse events, hospital stays, and deaths. Crossover from the randomly assigned to the alternate treatment was disallowed before the 24-month assessment, except if a patient in either group developed progression of HF to NYHA functional class III or IV. Crossovers required prior communications between the site investigator and a member of the Steering Committee and were only allowed as a last resort.
Study end points
The primary end point of the study was the HF clinical composite response, used to classify patients as worsened, unchanged, or improved (12,16). Patients were classified as worsened if they: 1) died or were hospitalized for worsening HF; 2) crossed over to the alternate treatment or permanently discontinued double-blind treatment due to worsening HF; or 3) had an increase in NYHA functional class or a moderate or marked worsening in their overall clinical status. Patients were classified as improved if they had not worsened and had a decrease in NYHA functional class or had a moderate or marked improvement in their overall clinical status or both. Patients who were neither worsened nor improved were classified as unchanged.
The prospectively powered secondary end point was left ventricular end-systolic volume index (LVESVi), measured as the absolute change between baseline and end follow-up and compared between the 2 study groups. All European echocardiographic measurements were made by a single core laboratory.
At each follow-up evaluation and up to the 24-month follow-up, all adverse events, whether related or unrelated to CRT, were recorded. Any event that led to death or to serious deterioration in the health status—meaning a life-threatening illness or injury resulting in a permanent impairment, required hospital stays or prolongation of hospital stay, or medical or surgical intervention—was classified as a serious adverse event (SAE).
The data were analyzed according to the intention-to-treat principle for all randomized patients. For the pre-specified analysis of the primary end point, patients who were improved or unchanged were grouped and considered to have suffered no progression of HF, whereas the proportion of patients who worsened in each study group was compared to ascertain the efficacy of CRT. For all time-to-event analyses, the Kaplan-Meier method was used, and groups were compared with log-rank statistics. Time 0 was the date of randomization, and patients who did not experience an event were censored at the 24-month follow-up or on the date of their exit from the study. The proportions of patients who suffered at least 1 SAE in each study group were compared with Fisher exact test. Rates of SAE are reported as the number of SAE divided by the study sample size. Other between-groups comparisons were made with Fisher exact test (proportions) or Student 2-sample ttest (means). The HF clinical composite response (percent of patients worsened) is reported as the observed odds ratio of CRT ON to CRT OFF, and 95% Wald confidence limits were used. Location data are reported as mean ± SD. All p values are nominal, and all statistical tests were 2-sided. The threshold of significance was set at 0.05.
A total of 287 patients were enrolled in the study at 35 European medical centers, between December 2004 and September 2006. The baseline evaluation was completed in 277 (Fig. 1).The CRT implants were attempted in 274 and successfully performed in 266 patients, representing a 97% overall implantation success rate. The CRT defibrillators were implanted in 68% and CRT pacemakers were implanted in 32% of patients. The mean LVEF was 28.0 ± 6.8% and mean left ventricular end-diastolic dimension was 70 ± 9 mm. The mean QRS duration was 156 ± 23 ms. At the time of study enrollment, an angiotensin-converting enzyme inhibitor or angiotensin receptor blocker was being administered to 99.6%, and a beta-adrenergic blocker was being administered to 93.5% of patients. Target doses of beta-adrenergic blockers were administered to 36% of patients, whereas 62% were receiving at least 50% of the target dose recommended by current HF guidelines (1). Ultimately, 262 patients were randomly assigned to CRT ON (n = 180) versus CRT OFF (n = 82). The baseline characteristics of the 2 study groups were similar (Table 1).
European sample versus remainder of the REVERSE trial population
In the REVERSE trial, several important baseline characteristics of the European subgroup differed from the remainder of the study population (i.e., non-European participants) (14) (Table 2).Overall, compared with the non-European sample, the European sample was, on average, significantly younger with fewer comorbidities. Thus European patients had a lower mean prevalence of ischemic heart disease (44% vs. 63%; p < 0.0001), less peripheral artery disease, and less history of prior myocardial infarction (34% vs. 55%; p < 0.0001) and of hypertension (34% vs. 66%; p < 0.0001). Moreover, Europeans had a smaller body mass index and a longer mean QRS duration (156 ± 23 ms vs. 151 ± 21 ms; p = 0.008), covered a longer mean 6-min walk distance (439 ± 103 m vs. 363 ± 134 m; p < 0.0001), and had a smaller proportion of CRT defibrillators implanted (68% vs. 95%; p < 0.0001) compared with the non-European patients.
Effect of CRT on primary and secondary end points
The difference between the 2 study groups became statistically significant at the 6-month evaluation and continued to differ significantly and with a similar amplitude of difference over the entire follow-up period in favor of the CRT ON assignment (Fig. 2).At 24 months of follow-up, a worsening of the HF clinical composite response was observed in 34 of the 180 patients (19%) assigned to CRT ON compared with 28 of the 82 patients (34%) assigned to CRT OFF (p = 0.01). Worsening was attributable to death or HF hospital stays in 68% of worsened patients in the CRT OFF group. Analyzing the HF clinical composite response conventionally by using the entire distribution of worsened, unchanged, and improved (16) yielded a p value of 0.0006 at 24 months in favor of CRT ON.
Paired measurements of LVESVi were available for 187 of the 262 patients (71%). The reasons for missing data are reported in Table 3.Intraobserver and interobserver variability for LVESVi in the echocardiographic core laboratory have been previously published (17). The LVESVi decreased by a mean of 27.5 ± 31.8 ml/m2in the CRT ON (n = 136) compared with 2.7 ± 25.8 ml/m2in the CRT OFF (n = 51) group (p < 0.0001) (Fig. 3).In the CRT ON group, the decrease in LVESVi was more than twice as great (36.9 ± 32.4 ml/m2) among patients with nonischemic heart disease as among patients with ischemic heart disease (16.3 ± 27.3 ml/m2). The measurements of left ventricular end-diastolic volume index and LVEF revealed similarly greater changes, consistent with reverse remodeling, in the CRT ON than in the CRT OFF group (Fig. 3). The main ventricular remodeling effect conferred by CRT occurred during the first 6 months, with further improvements developing over the following 12 months.
Subgroup analyses yielded concordant results (Fig. 4).The improvements in clinical status and LV function conferred by CRT were significant and similar in patients with ischemic and nonischemic heart disease. A trend was observed toward less clinical efficacy conferred by CRT among patients in NYHA functional class I than in class II, though the statistical comparison was unreliable because of an insufficient number of observations. Patients with a QRS duration ≥152 ms tended to derive more benefit from CRT than patients with a QRS shorter than 152 ms.
Other secondary end points
The distance covered during the baseline 6-min hall walk test was consistent with the clinical status of the study population. The mean changes in the distance walked at 24 months (+29.2 ± 87.3 m with CRT ON vs. +21.9 ± 90.7 m with CRT OFF; p = 0.57) and quality of life ascertained by the Minnesota questionnaire score (−8.2 ± 15.5 with CRT ON vs. −7.0 ± 14.6 with CRT OFF; p = 0.62) did not differ between groups.
Hospital stays and death
During the 24-month follow-up, 106 of the 262 patients underwent a total of 199 hospital stays. In the CRT ON group, 40.5% of patients were hospitalized for any cause versus 41.2% of patients in the CRT OFF group (hazard ratio [HR]: 0.98, p = 0.94). The time to first HF hospital stay was significantly longer in the CRT ON (7.8% hospitalized) than in the CRT OFF (18.4%) group (HR: 0.39, p = 0.01).The absolute rates of HF hospital stay were 14 of 82 (17.1%) CRT OFF patients and 13 of 180 (7.2%) CRT ON patients. Time to first HF hospital stay or death, time to first HF hospital stay, and time to death from any cause are displayed in Figure 5.
During the 24-month follow-up, 13 patients died. Deaths were classified as sudden cardiac in 3 patients and HF-related in 3 patients and due to noncardiovascular causes in 4 patients. No cause of death was identified in 3 patients. The 24-month death rate was 5.7% in the CRT ON versus 8.6% in the CRT OFF group (HR: 0.40, p = 0.09).
The perioperative (0 to 30 days) SAE rate was 21%, consisting mainly of dislodgments of 13 left ventricular, 6 right atrial, and 5 right ventricular leads. After implantation and during the 24-month follow-up, 26 of the 262 device recipients experienced a total of 30 device-related SAE. The percentage of patients with SAE in the CRT ON versus CRT OFF study groups were similar (p = 0.66). The rate of SAE between 1 and 12 months was 6%, the most common SAE being lead dislodgement, particularly of the LV lead (n = 9). The rate of SAE between 12 and 24 months was 3%, consisting mainly of right ventricular lead dysfunction. At the time of study closure, 29 of the 30 post-implant, device-related SAE were resolved.
Follow-up compliance and crossovers
In addition to the 13 deaths, there were 4 exits from the study before 24 months. The 24-month follow-up was completed by the other 245 patients. During the 24 months, 15 of the 262 patients (5.7%) permanently crossed over from their randomized assignment to the alternate treatment, including 2 patients from CRT ON to CRT OFF due to worsening HF and upon the patient's request, respectively, and 13 patients from CRT OFF to CRT ON, because of worsening HF in 10 and other reasons in 3 patients. Crossover due to worsening HF occurred at various times during follow-up from 44 to 603 days after randomization.
The 24-month results of the European cohort of the large multicenter REVERSE trial show for the first time that CRT improves the HF clinical composite response, a measure of clinical outcome (16) as well as LV structure and function, in patients with asymptomatic or mildly symptomatic LV dysfunction. The CRT ON was associated with progressive reverse LV remodeling over time, which was significant at 6 months of follow-up with a further improvement over time and was accompanied by a significant decrease in morbidity and mortality. Not surprisingly, no significant improvement in quality of life or increase in exercise capacity by CRT was observed, an expected outcome in a cohort of patients with little functional impairment at baseline.
The present results with a twice-as-long follow-up were superior to those observed in the first phase of the study, in which no significant difference was observed in the primary end point between the 2 study groups (14). This significant and sustained benefit conferred by CRT has several putative explanations. First, disease progression in patients presenting with NYHA functional class I or II is usually slow. Its prevention or reversal by any treatment modality might require several years to become apparent (18). In the present study, the longer follow-up might have contributed to the highly significant results, manifest as progressive, time-dependent, therapeutic effects conferred by CRT. It does not, however, explain how the primary objective was reached as early as 6 months in the European cohort, in contrast to the main, much larger, 12-month study (14). The effect of a different patient population might be an important factor. The European cohort of the REVERSE trial was representative of a typical European HF population with, in particular, a relatively low proportion of ischemic heart disease, as observed in previous European trials of CRT with favorable outcomes (7). Although the influence of underlying heart disease on the long-term outcomes of CRT systems recipients remains to be confirmed, its major effect on the rate and time course of reverse remodeling conferred by CRT has been clearly established (9,10). As a matter of fact, in the present study the extent of reverse remodeling among patients with nonischemic heart disease after CRT was more than 2 times greater than among patients with ischemic heart disease.
The extent of reverse remodeling observed in the present study was similar to that observed in previous studies of CRT in patients in NYHA HF functional classes III and IV (9,10). Importantly, a larger proportion of patients in the REVERSE trial were treated with beta-adrenergic blockers for at least 3 months before randomization than in these previous trials, and a larger proportion of patients received target doses, with no changes in doses allowed during the randomization phase, unless medically necessary. This suggests that CRT has prominent effects on ventricular remodeling, above and beyond the effects of drugs prescribed for the management of HF. This observation, along with the higher survival associated with reverse LV remodeling (but not symptomatic improvement) reported by another study over 2 years (19), is consistent with the improvement in clinical outcomes observed in this study.
The mortality and hospital stay rates in the asymptomatic or mildly symptomatic patient population of the REVERSE trial were low, as expected. On average, the patients were also younger, had a shorter mean QRS duration, and were more rigorously managed at baseline according to current HF guidelines (1) than in previous trials. A clearly decreased morbidity and a trend toward a decreased mortality, attributable to CRT, were nevertheless observed in this study. Although the REVERSE trial was not designed to examine morbidity or mortality, our observations suggest that, in the long term, they might both be lowered by CRT in patients presenting in NYHA functional class I or II. The long-term effects of CRT on mortality and HF-related clinical events are currently being studied by the RAFT (Resynchronization/Defibrillation in Ambulatory Heart Failure) trial (20) and MADIT CRT (Multicenter Automatic Defibrillator Implantation Trial–Cardiac Resynchronization Therapy) trial (21).
The benefits conferred by CRT in a patient population similar to that of the REVERSE trial must be balanced against its risks. Because most patients in the REVERSE trial were candidates for the primary preventive implantation of an ICD (22), the added risk associated with the addition of a CRT system was related to the LV lead, which was 11% at 24 months with most complications occurring during the 12 first months. Furthermore, the implant success rate was very high, reflecting the progress made in implantation techniques and lead design since earlier studies of CRT devices (3–5).
Complete blinding is not feasible in device trials. In the present double-blind study, all precautions were taken to limit potential bias due to blinding procedure. Most importantly, different staff members evaluated the device and patients outcome, including the electrocardiograms and echocardiograms. Still, we cannot exclude that some situations—like examining the electrocardiogram while evaluating the patient—would have disclosed whether patients were programmed to CRT ON or CRT OFF and might have created the potential for bias. The present analysis was conducted in the European cohort of the study. Because there were some important differences in baseline demographic data between European and non-European patients, we cannot exclude that our results are in part explained by a population effect. Thus, our conclusions are mainly appropriate for a European NYHA functional class I/II HF population. The REVERSE trial was planned to enroll mildly symptomatic or asymptomatic HF patients with LV dysfunction. However, relatively few NYHA functional class I patients (17% of the entire study population) were recruited. Accordingly, this article more accurately reports on the results of CRT in patients with NYHA functional class II HF. The relatively high proportion of missing echocardiographic data is a common observation in studies with centralized analysis (9–11). In the present study, there are 2 main reasons for the imbalance in paired data: deaths (7% CRT OFF vs. 3% CRT ON) and nonreadable 24-month echocardiograph (18% CRT OFF and 12% CRT ON). Missing echocardiographic data were more common in the CRT OFF group than in the CRT ON group, but a sensitivity analysis revealed that the results would be the same even if we assumed outcomes in favor of CRT OFF over CRT ON for the missing patients.
Over a 24-month follow-up, CRT improved the clinical status, reduced the risk of first HF hospital stay or death, and reversed LV remodeling, in a large European population of patients in NYHA functional classes I or II.
The authors wish to acknowledge the 73 centers that contributed to this study, the echocardiographic core laboratory, the Data Monitoring Committee and Adverse Event Advisory Committee, as well as the Swedish Heart and Lung Foundation (Online Appendix).
For the centers and committee members who contributed to this study, please see the online version of this article.
This study was supported by the Medtronic Bakken Research Center BV, Maastricht, the Netherlands, and Medtronic, Minneapolis, Minnesota. Dr. Daubert receives consulting fees from Medtronic and St. Jude Medical, and speaker's honoraria from Medtronic, Sorin Group, and St. Jude Medical. Dr. Gold reports consultant and clinical trials with Medtronic, Boston Scientific, and St. Jude Medical. Dr. Abraham receives research grants and consulting fees from Medtronic, St. Jude Medical, and Biotronik. Dr. Szili-Török received consulting fees from Medtronic. Dr. Linde receives research grants, speaker's honoraria, and consulting fees from Medtronicand speaker's honoraria and consulting fees from St. Jude Medical.
- Abbreviations and Acronyms
- cardiac resynchronization therapy
- heart failure
- implantable cardioverter-defibrillator
- left ventricular ejection fraction
- left ventricular end-systolic volume index
- New York Heart Association
- serious adverse event
- Received May 29, 2009.
- Revision received July 8, 2009.
- Accepted August 3, 2009.
- American College of Cardiology Foundation
- Dickstein K.,
- Cohen-Solal A.,
- Filippatos G.,
- et al.
- Epstein A.E.,
- DiMarco J.P.,
- Ellenbogen K.A.,
- et al.
- Cleland J.G.,
- Daubert J.C.,
- Erdmann E.,
- et al.
- St. John Sutton M.G.,
- Plappert T.,
- Abraham W.T.,
- et al.
- Higgins S.L.,
- Hummel J.D.,
- Niazi I.K.,
- et al.
- Abraham W.T.,
- Young J.B.,
- Leon A.R.,
- et al.
- Linde C.,
- Gold M.,
- Abraham W.T.,
- Daubert J.C.
- Linde C.,
- Gold M.R.,
- Abraham W.T.,
- Daubert J.C.
- Green C.P.,
- Porter C.B.,
- Bresnahan D.R.,
- Spertus J.A.
- Donal E.,
- Leclercq C.,
- Linde C.,
- Daubert J.C.
- Yu C.M.,
- Bleeker G.B.,
- Fung J.W.,
- et al.