Author + information
- Received February 2, 2017
- Revision received March 31, 2017
- Accepted April 2, 2017
- Published online June 12, 2017.
- Dipak Kotecha, MBBS, PhDa,b,
- Marcus D. Flather, MBBSc,
- Douglas G. Altman, DScd,
- Jane Holmes, PhDd,
- Giuseppe Rosano, MD, PhDe,f,
- John Wikstrand, PhDg,
- Milton Packer, MDh,
- Andrew J.S. Coats, DSci,j,
- Luis Manzano, MDk,
- Michael Böhm, MDl,
- Dirk J. van Veldhuisen, MDm,
- Bert Andersson, MD, PhDn,
- Hans Wedel, PhDo,
- Thomas G. von Lueder, PhDb,p,
- Alan S. Rigby, MScq,
- Åke Hjalmarson, MD, PhDn,
- John Kjekshus, MD, PhDr,
- John G.F. Cleland, MDs,∗ (, )
- Beta-Blockers in Heart Failure Collaborative Group
- aUniversity of Birmingham Institute of Cardiovascular Sciences, Birmingham, United Kingdom
- bCentre of Cardiovascular Research and Education in Therapeutics, Monash University, Melbourne, Victoria, Australia
- cNorwich Medical School, University of East Anglia, Norwich, United Kingdom
- dCentre for Statistics in Medicine, Nuffield Department of Orthopaedics, Rheumatology and Musculoskeletal Sciences, University of Oxford, Oxford, United Kingdom
- eDepartment of Medical Sciences, IRCCS San Raffaele Pisana, Rome, Italy
- fCardiovascular and Cell Science Institute, St. George’s University of London, London, United Kingdom
- gWallenberg Laboratory for Cardiovascular Research, Sahlgrenska Academy, Gothenburg University, Gothenburg, Sweden
- hBaylor Heart and Vascular Institute, Baylor University Medical Center, Dallas, Texas
- iMonash University, Melbourne, Victoria, Australia
- jUniversity of Warwick, Warwick, United Kingdom
- kInternal Medicine Department, Hospital Universitario Ramón y Cajal, Universidad de Alcalá, Madrid, Spain
- lUniversitätsklinikum des Saarlandes, Homburg/Saar, Germany
- mDepartment of Cardiology, University Medical Centre Groningen, University of Groningen, Groningen, the Netherlands
- nDepartment of Cardiology, Sahlgrenska University Hospital and Gothenburg University, Gothenburg, Sweden
- oHealth Metrics, Sahlgrenska Academy, University of Gothenburg, Gothenburg, Sweden
- pDepartment of Cardiology, Oslo University Hospital, Oslo, Norway
- qAcademic Cardiology, Castle Hill Hospital, Kingston upon Hull, United Kingdom
- rRikshospitalet University Hospital and Faculty of Medicine, University of Oslo, Oslo, Norway
- sRobertson Institute of Biostatistics and Clinical Trials Unit, University of Glasgow, Glasgow, United Kingdom
- ↵∗Address for correspondence:
Prof. John G.F. Cleland, Robertson Institute of Biostatistics and Clinical Trials Unit, University of Glasgow, University Avenue, Glasgow G12 8QQ, United Kingdom.
Background The relationship between mortality and heart rate remains unclear for patients with heart failure with reduced ejection fraction in either sinus rhythm or atrial fibrillation (AF).
Objectives This analysis explored the prognostic importance of heart rate in patients with heart failure with reduced ejection fraction in randomized controlled trials comparing beta-blockers and placebo.
Methods The Beta-Blockers in Heart Failure Collaborative Group performed a meta-analysis of harmonized individual patient data from 11 double-blind randomized controlled trials. The primary outcome was all-cause mortality, analyzed with Cox proportional hazard ratios (HR) modeling heart rate measured at baseline and approximately 6 months post-randomization.
Results A higher heart rate at baseline was associated with greater all-cause mortality for patients in sinus rhythm (n = 14,166; adjusted HR: 1.11 per 10 beats/min; 95% confidence interval [CI]: 1.07 to 1.15; p < 0.0001) but not in AF (n = 3,034; HR: 1.03 per 10 beats/min; 95% CI: 0.97 to 1.08; p = 0.38). Beta-blockers reduced ventricular rate by 12 beats/min in both sinus rhythm and AF. Mortality was lower for patients in sinus rhythm randomized to beta-blockers (HR: 0.73 vs. placebo; 95% CI: 0.67 to 0.79; p < 0.001), regardless of baseline heart rate (interaction p = 0.35). Beta-blockers had no effect on mortality in patients with AF (HR: 0.96, 95% CI: 0.81 to 1.12; p = 0.58) at any heart rate (interaction p = 0.48). A lower achieved resting heart rate, irrespective of treatment, was associated with better prognosis only for patients in sinus rhythm (HR: 1.16 per 10 beats/min increase, 95% CI: 1.11 to 1.22; p < 0.0001).
Conclusions Regardless of pre-treatment heart rate, beta-blockers reduce mortality in patients with heart failure with reduced ejection fraction in sinus rhythm. Achieving a lower heart rate is associated with better prognosis, but only for those in sinus rhythm.
Beta-blockers reduce morbidity and mortality in patients with heart failure with reduced left ventricular ejection fraction (HFrEF) in sinus rhythm (1,2). It is not clear whether the key mechanism underpinning their benefits is protection of adrenergic receptors from heightened sympathetic activity or reduction in heart rate. It is also uncertain whether the efficacy of beta-blockers is related to dose, reduction in heart rate, or achieved heart rate (3–10). These questions are conceptually important for how clinicians manage and follow-up patients with HFrEF. Furthermore, there may be a clinically important interaction with heart rhythm (11). Although beta-blockers reduce the incidence of new-onset atrial fibrillation (AF) in patients with HFrEF (1,12), they do not appear to reduce mortality for patients with established HFrEF and concomitant AF (1).
The Beta-Blockers in Heart Failure Collaborative Group pooled individual patient data (IPD) from major randomized controlled trials (RCTs) comparing beta-blockers with placebo in patients with heart failure to investigate further their efficacy and safety (13). With almost all the available IPD, this analysis permits a robust assessment of the associations between heart rate, heart rhythm, and mortality. Our aims were to answer 3 questions for patients with HFrEF according to their heart rhythms. 1) Does baseline heart rate predict mortality? 2) Does the effect of beta-blockers on mortality differ according to baseline heart rate? 3) What is the association between achieved heart rate, achieved dose, and mortality?
The Beta-Blockers in Heart Failure Collaborative Group includes leading investigators from relevant landmark trials, with the support of the pharmaceutical companies that conducted them (AstraZeneca, Cambridge, United Kingdom; GlaxoSmithKline, Isleworth, London, United Kingdom); Merck Serono, Darmstadt, Germany; and Menarini, Florence, Italy). This report was prepared according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses IPD guidance (14) and prospectively registered with ClinicalTrials.gov (NCT00832442) and the PROSPERO database of systematic reviews (CRD42014010012) (15). Detailed rationale and methods have been previously published (1,2,13). Each trial required appropriate ethical approval.
Eligibility and search strategy
A systematic search was performed of MEDLINE and Current Contents, scrutiny of reference lists of trials, trial registries, meeting abstracts, and review papers, as well as discussion with group members and pharmaceutical manufacturers (1,2,13). We included RCTs that reported mortality as a primary or part of a composite outcome comparing beta-blockers versus placebo with recruitment of >300 patients and planned follow-up of >6 months. Eleven studies were included that accounted for 95.7% of eligible participants recruited in RCTs on the basis of a systematic review of published reports: ANZ (Australia/New Zealand Heart Failure Study) (16), BEST (Beta-Blocker Evaluation Survival Trial) (17), CAPRICORN (Carvedilol Post-Infarct Survival Control in LV Dysfunction Study) (18), CHRISTMAS (Carvedilol Hibernating Reversible Ischaemia Trial: Marker of Success Study) (19), CIBIS-I (Cardiac Insufficiency Bisoprolol Study) (20), CIBIS-II (Cardiac Insufficiency Bisoprolol Study II) (21), COPERNICUS (Carvedilol Prospective Randomized Cumulative Survival Study) (22), MDC (Metoprolol in Idiopathic Dilated Cardiomyopathy Study) (23), MERIT-HF (Metoprolol CR/XL Randomised Intervention Trial in Congestive Heart Failure) (24), US-HF SENIORS (Study of the Effects of Nebivolol Intervention on Outcomes and Rehospitalisation in Seniors With Heart Failure) (25), and the U.S. Carvedilol Heart Failure Study (26). All included studies had low risk for bias (27).
Data collection and IPD integrity
Data were extracted from original source files provided by the pharmaceutical companies and lead investigators (13). All trials provided IPD, and databases were harmonized according to a standardized data request form to match patient characteristics and outcomes across all trials. Discrepancies, inconsistencies, and incomplete data were checked against original case report forms, trial documentation, and published reports to ensure IPD integrity. The clinically measured resting heart rate was used in analysis, as this was consistently recorded in all trials at each major study visit. Because of the small amount of missing data, imputation was not performed.
We included all patients with baseline electrocardiograms that showed either sinus rhythm or AF or atrial flutter. For the purposes of this report, reference to AF therefore includes atrial flutter (1). Patients missing baseline electrocardiograms or in paced rhythm were excluded. We also excluded all patients with documented heart block because second- or third-degree heart block was an exclusion criterion in some of the trials.
Outcomes and effect measures
The outcome for this analysis was all-cause mortality, including additional deaths on follow-up available from 7 studies (19–21,25,26,28,29). Our analysis used heart rate as a continuous variable and also categorized into pre-specified groups (<70, 70 to 90, and >90 beats/min). All trials excluded patients with lower heart rates, as defined in Figure 1.
A statistical analysis plan was generated and finalized by the collaborative group in advance of data analysis. Summary results are presented as percentages, or median and interquartile range (IQR; displayed as 25th to 75th quartiles). Estimated glomerular filtration rate was calculated using the MDRD (Modification of Diet in Renal Disease) formula, normalized to a body surface area of 1.73 m2.
All analyses followed the principle of intention-to-treat. Baseline heart rhythm groups (sinus rhythm or AF) were analyzed separately. Outcomes were analyzed using a Cox proportional hazards regression model (28), stratified by study. This is a 1-stage fixed-effects approach and assumes that all trials are estimating a common treatment effect with baseline hazards that vary across studies. The independent variable was continuously distributed heart rate. We assessed the relationship between continuous heart rate and mortality using fractional polynomials to find the best transformation (29), but a linear association was the best fit (with note taken of the scarce data below a heart rate of 60 beats/min due to trial exclusion criteria). Hazard ratios (HRs) and 95% confidence intervals (CIs) are presented, along with corresponding p values. We pre-specified adjustment in Cox models for age, sex, left ventricular ejection fraction (LVEF), systolic blood pressure, prior myocardial infarction, and baseline use of angiotensin-converting enzyme inhibitors or angiotensin receptor blockers and diuretic therapy. Treatment allocation and heart rate were also adjusted for, where appropriate. The goodness-of-fit C statistic for the main stratified Cox model was 0.66 for sinus rhythm and 0.64 for AF at 20 months. Kaplan-Meier plots were used to graph the pooled trial data. Few patients were followed for more than 3 years, and therefore data were censored at 1,200 days (3.3 years) from randomization. Heterogeneity was assessed using the chi-square test and I2 statistic, with the estimate of heterogeneity taken from the inverse-variance fixed-effects 2-stage model (30). Pre-defined sensitivity analyses were alternative censor points (1 and 2 years), alternative methodology (2-stage meta-analysis and fixed vs. random effects ), and restriction to a heart rate between 60 and 140 beats/min at baseline.
Analyses at the interim study time point (mean 184 days from randomization) excluded those who had died, had withdrawn consent, or were lost to follow-up. Not all patients attended interim visits or had heart rates recorded at this time, but the number of patients without interim data was similar across treatment groups for both sinus rhythm and AF. Furthermore, there was no significant difference in baseline heart rate compared with those with interim data or any difference in the observed hazards for either heart rate or beta-blocker efficacy. We performed 2 post hoc analyses not detailed in our pre-specified analysis plan: 1) assessment of mortality in patients on beta-blockers who attained heart rates <60 beats/min; and 2) assessment of mortality according to beta-blocker dose achieved at the interim visit. There were missing data on dose in all studies, and 2 studies provided no information (17,19). For consistency across the different beta-blockers and trials, dose achieved was expressed as the percentage of maximum target dose according to the particular beta-blocker and specific trial design.
There was no evidence of violation of the proportional hazards assumption in any multivariate model, as determined by Schoenfeld residuals (32). Effect modification was assessed using p values from interaction terms fitted in the multivariate models (29,33). A 2-tailed p value of 0.05 was considered to indicate statistical significance. Analyses were performed in Stata version 14.1 (StataCorp, College Station, Texas) and R version 3.2.1 (R Core Team, Vienna, Austria).
IPD were obtained for 18,637 patients. Patients were excluded because of missing baseline electrocardiograms (n = 118), heart block (n = 510), or paced rhythm (n = 616). A further 15 participants had missing baseline heart rates. The final cohort included 14,313 patients in sinus rhythm and 3,065 in AF (Figure 1). Three patients (1 in sinus rhythm and 2 in AF) had missing event dates and were excluded from outcome analyses.
The median age was 65 years (IQR: 55 to 72 years), 24% were women, and the median LVEF at baseline was 0.27% (IQR: 0.21% to 0.33%). The median baseline heart rate was 80 beats/min for those in sinus rhythm (IQR: 72 to 88 beats/min) and 81 beats/min for those in AF (IQR: 72 to 92 beats/min). Characteristics according to baseline heart rhythm are presented in Table 1. Regardless of heart rhythm, patients with higher heart rates were younger and more likely to be women, to have nonischemic cardiomyopathy, and to have lower LVEF and more severe symptoms. There were no differences in patient characteristics according to randomized treatment for either sinus rhythm (Online Table 1) or AF (Online Table 2) in any heart rate group.
Heart rate at baseline and mortality for patients in sinus rhythm or AF
For patients in sinus rhythm, there were 2,141 deaths among 14,166 patients (15.1%) over a mean follow-up period of 1.5 ± 1.1 years. Baseline heart rate was associated with all-cause mortality, with an HR of 1.11 per 10 beats/min (95% CI: 1.07 to 1.15; p < 0.0001), adjusted for baseline variables and treatment allocation. From the Kaplan-Meier analysis (Figure 2A), higher baseline heart rates were associated with higher mortality in patients assigned to either placebo or beta-blockers.
For patients in AF at baseline, there were 609 deaths among 3,034 patients (20.1%), but there was no association between baseline heart rate and mortality (adjusted HR: 1.03; 95% CI: 0.97 to 1.08; p = 0.38) (Figure 2B).
The Central Illustration displays the modeling of heart rate as a continuous variable and the HR for death, according to baseline heart rhythm. Contrary to results in sinus rhythm, there was no relationship between baseline heart rate and mortality for those in AF (p = 0.003 for interaction). Sensitivity analyses showed similar results to the main findings (Online Table 3).
Efficacy of beta-blockers according to baseline heart rate
Beta-blockers reduced heart rate by 11 to 12 beats/min in both sinus rhythm and AF (Online Table 4, Figure 1). The overall HR for mortality comparing beta-blockers with placebo for patients in sinus rhythm was 0.73 (95% CI: 0.67 to 0.79; p < 0.0001), with similar benefit for all 3 strata of baseline heart rate (Table 2, Figure 3). There was no interaction with baseline heart rate as a continuous variable (p = 0.35). In contrast, beta-blockers did not reduce mortality for patients in AF, either overall (HR: 0.96; 95% CI: 0.81 to 1.12; p = 0.58) or for any baseline heart rate stratum (interaction p = 0.48) (Table 2). Similar results were seen in sensitivity analyses (Online Table 3).
Achieved versus change in post-randomization heart rate and mortality
A landmark analysis was performed, starting at an interim visit after expected dose titration for each surviving participant (mean 184 ± 144 days from randomization) with a recorded interim heart rate (n = 12,441 in sinus rhythm and n = 2,566 in AF). Mean heart rate was similar at the interim and final visits for surviving patients in sinus rhythm or AF, suggesting stable beta-blockade had been reached (Online Figure 1).
For patients in sinus rhythm, the heart rate achieved at the interim visit was more strongly associated with mortality than the change in heart rate from baseline (HR per 10 beats/min: 1.16; 95% CI: 1.11 to 1.22) (Online Table 5). The lowest mortality in sinus rhythm was observed in patients who attained lower heart rates after beta-blocker therapy (Figure 4A). Conversely, in patients with AF, neither attained nor change in heart rate were associated with survival (Figure 4B, Online Table 5).
Analysis of post-randomization beta-blocker dosage in sinus rhythm
Separately fitted models in patients with sinus rhythm for those assigned to placebo or beta-blockers showed consistent findings for the association of interim heart rate and mortality. In patients randomized to beta-blockers (n = 6,327), the adjusted HR was 1.12 per 10 beats/min (95% CI: 1.05 to 1.19). In patients randomized to placebo, the dose of which does not affect heart rate (n = 6,114), the adjusted HR was 1.13 per 10 beats/min (95% CI: 1.08 to 1.19).
Analysis of dose achieved (Online Table 6) was complicated by susceptibility to bias due to nonrandom missing data. Achieving a higher dose was associated with lower mortality in both the placebo and beta-blocker arms (Online Figure 2).
Our analysis confirms a reduction in mortality with beta-blockers for patients with HFrEF in sinus rhythm, irrespective of pre-treatment heart rate within the studied range. For patients in sinus rhythm, resting heart rate is an important prognostic indicator, both before and after the initiation of beta-blockers; a lower achieved heart rate is associated with lower subsequent mortality and is more likely to occur in patients initiated on beta-blockers. In patients with concomitant AF, heart rate was not associated with mortality, and beta-blockers did not reduce mortality at any observed heart rate.
Insights on the mechanism of action of beta-blockers
Whether reduction in morbidity and mortality in patients with HFrEF in sinus rhythm is related to myocardial protection from heightened sympathetic activity or due to reductions in heart rate is uncertain. Chronic adrenergic overstimulation is thought to provoke myocyte dysfunction and arrhythmias (34), providing a theoretical rationale for prescribing beta-blockers for HFrEF. However, a large trial of moxonidine, which inhibits sympathetic activation, was stopped prematurely for harm, which casts doubt on this hypothesis (35). Heart rate reduction may also improve cardiac myocyte metabolism by conserving energy, improving calcium recycling, increasing diastolic blood flow, and protecting against ischemia. However, our finding that beta-blockers reduce mortality, regardless of pre-treatment heart rate within the studied range, suggests that the mechanism of action of beta-blockers is not due simply to lowering heart rate. Moreover, ivabradine, which decreases heart rate by If-channel blockade rather than by sympathetic inhibition, did not reduce mortality overall when added to beta-blockers, although it did reduce the composite of cardiovascular death or hospital admission for worsening heart failure (36).
Divergent responses in patients with AF
Observational studies suggest a relationship between resting heart rate and prognosis in patients with AF predominantly without heart failure (37) and those with HFrEF in sinus rhythm (5,38). However, ventricular rate appears to be a poor predictor of outcomes for patients with concomitant HFrEF and AF. Lower ventricular rates in AF may even be associated with an adverse prognosis (39), but why the relationship between heart rate and prognosis should differ by heart rhythm is uncertain. Perhaps heart rate is a good reflection of sympathetic activation only for patients in sinus rhythm. Vagal activity is also a major determinant of heart rate, which may be increased by beta-blockade and potentially be more important for patients in sinus rhythm compared with AF (40). Alternatively, the relationship could be confounded by an increase in risk associated with variable R-R intervals in AF or ventricular pauses (41). The RACE-II (Rate Control Efficacy in Permanent Atrial Fibrillation) study, an RCT of strict compared with lenient ventricular rate control of AF, failed to show a difference in outcome between these strategies, even among those patients with concomitant heart failure (42).
We previously identified a highly significant interaction between the effects of beta-blockers on mortality and heart rhythm (p = 0.002) (1), but why beta-blockers do not reduce mortality in patients with HFrEF with AF remains unclear (43,44). If their benefits are mediated by blocking adrenergic receptors on cardiac myocytes, then heart rhythm should be irrelevant. Similarly, if their benefits are mediated by reducing ventricular rate, then these should also be similar, regardless of heart rhythm. Further work on the effects of autonomic modulation in patients with heart failure and AF is clearly warranted.
Target heart rate versus target dose
Whether clinicians should strive to achieve a target heart rate or a target dose of beta-blocker remains unanswered, and the authors of this paper were unable to reach a consensus. In this analysis, beta-blockers reduced mortality, regardless of baseline heart rates, for patients with HFrEF in sinus rhythm. All trial protocols, which form the basis for current international guidelines, requested titration to a target dose of beta-blocker, provided they were tolerated and did not cause excessive bradycardia. Dose-dependent improvements in LVEF and survival were observed in the MOCHA (Multicenter Oral Carvedilol Heart Failure Assessment) trial (45), although that trial included only 345 patients. No large trial has randomized patients to higher versus lower doses, although post hoc analyses suggest greater benefit from higher doses (3,9,46). A trial-level meta-analysis of 7 dose-ranging studies of beta-blockers provided inconclusive evidence of a dose relationship with mortality (47); further prospective trials are required to clarify this issue.
Conversely, for those who believe that lowering heart rate is the key mediator of beta-blocker benefit for patients with HFrEF in sinus rhythm, our analysis supports the notion that achieving a lower rate (∼60 beats/min) is beneficial, perhaps because it is a physiological marker indicating that adequate beta-receptor blockade has been achieved. The advantage of an approach that titrates to a target heart rate is clinical simplicity that, serendipitously, may lead to increased use of the guideline-recommended target doses of beta-blockers, as well as being a measure of patient adherence to therapy.
Ultimately, heart rate and prescribed beta-blocker dose are intimately related; one is a surrogate for the other, although the relationship may be complicated by other factors, such as genetic variations in beta-blocker response and drug metabolism. Our observation of dose-related differences in mortality in patients assigned to placebo clearly demonstrates that it is unsafe to make a strong inference from any analysis of a post-randomization variable such as dose. Dose achieved is itself an outcome (48), affected by confounding patient factors, adherence, physician preferences, and bias, including the perceived risk for adverse outcomes.
This was a retrospective analysis, and background therapy, including devices, will have changed since these trials were conducted. Heart rate was not measured in a standardized fashion across trials and may have been less accurate in patients with AF. Although by using IPD we were able to adjust for many known confounders with sufficient power for statistical analysis, unmeasured variables may have affected heart rate or dose of beta-blocker. The trials had different patient study groups and used different beta-blockers; we have previously demonstrated that excluding individual trials had no impact on results (1), and the diversity of trial participants could be considered a strength. Our analysis plan specified that only mortality would be analyzed as an outcome. Although data on hospitalization were available, this outcome may be biased, as heart rate can influence the likelihood of a physician admitting a patient. The power to explore effects in the subgroup with AF is limited by its modest size (albeit large in comparison with many other reports) and the inclusion of a small number of patients with atrial flutter. Few patients with resting heart rates <65 beats/min were enrolled in these RCTs, and hence we are unable to comment on patients who had slower heart rates before receiving beta-blockers. There is uncertainty as to where the nadir of risk in the relationship between heart rate and risk lies, but there will be a rate below which mortality rises.
Beta-blockers reduce mortality at all studied heart rates in patients with HFrEF in sinus rhythm, and those who achieved lower resting heart rates in dose-titrated RCTs had lower mortality. This did not hold true for patients with concomitant AF, for whom there was no mortality benefit from beta-blockade, nor a relationship between heart rate and mortality.
COMPETENCY IN MEDICAL KNOWLEDGE: In patients with HFrEF and sinus rhythm, lower heart rates are associated with lower mortality, and beta-blockers improve survival regardless of heart rate. In those with AF, however, heart rate is unrelated to mortality, and beta-blockers do not lengthen survival.
TRANSLATIONAL OUTLOOK: Further studies are needed to understand the differential effect of beta-blockers on clinical outcomes in patients with HFrEF and AF and to explore the impact of other drugs that affect heart rate and rhythm in these patients.
The authors are indebted to the other members of the Beta-Blockers in Heart Failure Collaborative Group for database access and extraction support (for the full list, please see the design paper ), the steering committees of the included trials (in particular representatives of the MERIT-HF trial), as well as the late Philip Poole-Wilson (1943 to 2009; Imperial College, London, United Kingdom). This work is dedicated to the memory of Henry Krum (1958 to 2015; Monash University, Melbourne, Australia), one of the founding members of the collaborative group. This project was possible only with the support of the pharmaceutical companies that have marketed beta-blockers in heart failure, and the group wishes to extend their gratitude to AstraZeneca, GlaxoSmithKline, Menarini Farmaceutica, and Merck Serono for full access to trial data. The authors gratefully acknowledge incorporation of data from the BEST trial through the National Heart, Lung, and Blood Institute’s BioLINCC program.
For supplemental tables and figures, please see the online version of this article.
Menarini Farmaceutica Internazionale provided an unrestricted research grant for administrative costs. GlaxoSmithKline provided data extraction support and a collaborative research grant to IRCCS San Raffaele. None of the pharmaceutical groups had any role in data analysis or manuscript preparation. The Steering Committee lead (Dr. Kotecha) and the Centre for Statistics in Medicine (Drs. Altman and Holmes) had full access to all the data and had joint responsibility for the decision to submit for publication after discussion with all the named authors. Dr. Kotecha is funded by a National Institute for Health Research (NIHR) Career Development Fellowship (CDF-2015-08-074). The opinions expressed are those of the authors and do not represent the NIHR or the U.K. Department of Health. Dr. Kotecha has received grants from Menarini, during the conduct of the study; nonfinancial support from Daiichi-Sankyo; personal fees from AtriCure; and has served as chief investigator for the RATE-AF (Rate Control Therapy Evaluation in Atrial Fibrillation) trial. Dr. Flather has received personal fees from AstraZeneca; and grants from Novartis. Dr. Wikstrand has an appointment as study team physician for MERIT-HF at AstraZeneca. Dr. Packer has received personal fees from Amgen, Admittance Technologies, Bayer, Boehringer Ingelheim, BioControl, Celyad, Daiichi-Sankyo, AstraZeneca, Cardiorentis, CardioKinetix, Relypsa, Novartis, Sanofi, Takeda, and ZS Pharma. Dr. Böhm has received personal fees from Servier, Medtronic, Bayer, and Pfizer. Dr. Andersson has received personal fees from Servier. Dr. Wedel has received personal fees from AstraZeneca. Prof. Cleland has received grants from Amgen and Novartis; personal fees from Novartis and Servier; and nonfinancial support from GlaxoSmithKline. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- atrial fibrillation
- confidence interval
- heart failure with reduced left ventricular ejection fraction
- hazard ratio
- individual patient data
- interquartile range
- left ventricular ejection fraction
- randomized controlled trial
- Received February 2, 2017.
- Revision received March 31, 2017.
- Accepted April 2, 2017.
- 2017 American College of Cardiology Foundation
- Kotecha D.,
- Manzano L.,
- Krum H.,
- et al.,
- for the Beta-Blockers in Heart Failure Collaborative Group
- Gullestad L.,
- Wikstrand J.,
- Deedwania P.,
- et al.,
- for the MERIT-HF Study Group
- Böhm M.,
- Perez A.C.,
- Jhund P.S.,
- et al.,
- for the I-Preserve Committees and Investigators
- Vazir A.,
- Claggett B.,
- Jhund P.,
- et al.
- Kato N.,
- Kinugawa K.,
- Teruhiko I.,
- et al.
- Wikstrand J.,
- Hjalmarson A.,
- Waagstein F.,
- et al.,
- for the MERIT-HF Study Group
- Kirchhof P.,
- Benussi S.,
- Kotecha D.,
- et al.
- ↵Kotecha D, Manzano L, Krum H, Altman DG, Holmes J, Flather M. The Beta-Blockers in Heart Failure Collaborative Group: individual patient data meta-analysis. Available at: http://www.crd.york.ac.uk/PROSPERO/display_record.asp?ID=CRD42014010012. Accessed April 4, 2017.
- Cleland J.G.F.,
- Pennell D.J.,
- Ray S.G.,
- et al.,
- for the CHRISTMAS (Carvedilol Hibernating Reversible Ischaemia Trial: Marker of Success) Investigators
- CIBIS Investigators and Committees
- Flather M.D.,
- Shibata M.C.,
- Coats A.J.,
- et al.
- ↵Higgins JPT, Altman DG, Sterne JAC, for the Cochrane Statistical Methods Group and the Cochrane Bias Methods Group. Chapter 8: assessing risk of bias in included studies. In: Higgins JPT, Green S, editors. Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0: The Cochrane Collaboration 2011. Available at: http://handbook.cochrane.org/chapter_8/8_assessing_risk_of_bias_in_included_studies.htm. Accessed April 4, 2017.
- Koopman L.,
- van der Heijden G.J.M.G.,
- Hoes A.W.,
- Grobbee D.E.,
- Rovers M.M.
- Schoenfeld D.
- Floras J.S.,
- Ponikowski P.
- Steinberg B.A.,
- Kim S.,
- Thomas L.,
- et al.,
- for the Outcomes Registry for Better Informed Treatment of AF (ORBIT-AF) Investigators and Patients
- Li S.J.,
- Sartipy U.,
- Lund L.H.,
- et al.
- Cullington D.,
- Goode K.M.,
- Zhang J.,
- Cleland J.G.F.,
- Clark A.L.
- Kotecha D.,
- Piccini J.P.
- Bristow M.R.,
- Gilbert E.M.,
- Abraham W.T.,
- et al.
- Dobre D.,
- van Veldhuisen D.J.,
- Mordenti G.,
- et al.,
- for the SENIORS Investigators
- Cleland J.G.F.,
- Coletta A.P.,
- Freemantle N.,
- Ahmed D.,
- Rubis P.,
- Clark A.L.