Author + information
- Received August 30, 2019
- Revision received September 11, 2019
- Accepted September 16, 2019
- Published online December 2, 2019.
- Dipak Kotecha, MD, PhD, MSca,b,∗ (, )@ICVS_UoB,
- Simrat K. Gill, MDa,
- Marcus D. Flather, MDc,
- Jane Holmes, PhDd,
- Milton Packer, MDe,
- Giuseppe Rosano, MD, PhDf,g,
- Michael Böhm, MDh,
- John J.V. McMurray, MDi,
- John Wikstrand, PhDj,
- Stefan D. Anker, MD, PhDk,
- Dirk J. van Veldhuisen, MDl,
- Luis Manzano, MDm,
- Thomas G. von Lueder, MD, PhDb,n,
- Alan S. Rigby, MSco,
- Bert Andersson, MD, PhDp,
- John Kjekshus, MD, PhDq,
- Hans Wedel, PhDr,
- Frank Ruschitzka, MDs,
- John G.F. Cleland, MDt,
- Kevin Damman, MD, PhDl,
- Josep Redon, MDu,
- Andrew J.S. Coats, MD, DScg,
- on behalf of the Beta-Blockers in Heart Failure Collaborative Group
- aInstitute of Cardiovascular Sciences, University of Birmingham, Birmingham, United Kingdom
- bCentre of Cardiovascular Research and Education in Therapeutics, Monash University, Melbourne, Victoria, Australia
- cNorwich Medical School, Faculty of Medicine and Health Science, 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
- eBaylor Heart and Vascular Institute, Baylor University Medical Center, Dallas, Texas
- fCardiovascular and Cell Science Institute, St. George’s University of London, London, United Kingdom
- gDepartment of Medical Sciences, IRCCS San Raffaele Pisana, Roma, Italy
- hKardiologie, Angiologie und Internistische Intensivmedizin, Universitatsklinikum des Saarlandes, Homburg/Saar, Germany
- iRobertson Institute of Biostatistics and Clinical Trials Unit, University of Glasgow, Glasgow, United Kingdom
- jWallenberg Laboratory for Cardiovascular Research, Sahlgrenska Academy, Gothenburg University, Gothenburg, Sweden
- kDepartment of Cardiology, Charite Campus Virchow-Klinikum, Berlin, Germany
- lUniversity of Groningen, Department of Cardiology, University Medical Centre Groningen, RB Groningen, the Netherlands
- mInternal Medicine Department, Hospital Universitario Ramón y Cajal, Universidad de Alcalá (IRYCIS), Plaza de San Diego, Madrid, Spain
- nDepartment of Cardiology, Oslo University Hospital, Oslo, Norway
- oHull York Medical School, Faculty of Health Sciences, University of Hull, Kingston-upon-Hull, United Kingdom
- pDepartment of Cardiology, Sahlgrenska University Hospital and Gothenburg University, Gothenburg, Sweden
- qRikshospitalet University Hospital and Faculty of Medicine, University of Oslo, Oslo, Norway
- rHealth Metrics, Sahlgrenska Academy, University of Gothenburg, Gothenburg, Sweden
- sKlinik fur Kardiologie, UniversitätsSpital Zürich, Zürich, Switzerland
- tInstitute of Cardiovascular and Medical Sciences, University of Glasgow, Glasgow, United Kingdom
- uINCLIVA Biomedical Research Institute, Valencia, Spain
- ↵∗Address for correspondence:
Dr. Dipak Kotecha, University of Birmingham Institute of Cardiovascular Sciences, Medical School, Vincent Drive, Birmingham B15 2TT, United Kingdom.
Background Moderate and moderately severe renal impairment are common in patients with heart failure and reduced ejection fraction, but whether beta-blockers are effective is unclear, leading to underuse of life-saving therapy.
Objectives This study sought to investigate patient prognosis and the efficacy of beta-blockers according to renal function using estimated glomerular filtration rate (eGFR).
Methods Analysis of 16,740 individual patients with left ventricular ejection fraction <50% from 10 double-blind, placebo-controlled trials was performed. The authors report all-cause mortality on an intention-to-treat basis, adjusted for baseline covariates and stratified by heart rhythm.
Results Median eGFR at baseline was 63 (interquartile range: 50 to 77) ml/min/1.73 m2; 4,584 patients (27.4%) had eGFR 45 to 59 ml/min/1.73 m2, and 2,286 (13.7%) 30 to 44 ml/min/1.73 m2. Over a median follow-up of 1.3 years, eGFR was independently associated with mortality, with a 12% higher risk of death for every 10 ml/min/1.73 m2 lower eGFR (95% confidence interval [CI]: 10% to 15%; p < 0.001). In 13,861 patients in sinus rhythm, beta-blockers reduced mortality versus placebo; adjusted hazard ratio (HR): 0.73 for eGFR 45 to 59 ml/min/1.73 m2 (95% CI: 0.62 to 0.86; p < 0.001) and 0.71 for eGFR 30 to 44 ml/min/1.73 m2 (95% CI: 0.58 to 0.87; p = 0.001). The authors observed no deterioration in renal function over time in patients with moderate or moderately severe renal impairment, no difference in adverse events comparing beta-blockers with placebo, and higher mortality in patients with worsening renal function on follow-up. Due to exclusion criteria, there were insufficient patients with severe renal dysfunction (eGFR <30 ml/min/1.73 m2) to draw conclusions. In 2,879 patients with atrial fibrillation, there was no reduction in mortality with beta-blockers at any level of eGFR.
Conclusions Patients with heart failure, left ventricular ejection fraction <50% and sinus rhythm should receive beta-blocker therapy even with moderate or moderately severe renal dysfunction.
Heart failure (HF) is associated with numerous comorbidities, of which renal dysfunction is both common and of particular importance due to its impact on mortality as well as on the use of guideline-recommended therapies (1). Patients with HF have a higher incidence of renal dysfunction due to shared pathophysiological pathways and mutual risk factors. In the Swedish HF registry (2), 51% of 47,716 patients with unselected HF had an estimated glomerular filtration (eGFR) of <60 ml/min/1.73 m2, and in the CHARM (Candesartan in Heart Failure: Assessment of Reduction in Mortality and Morbidity) program, between 33% and 43% of patients had eGFR <60 ml/min/1.73 m2 depending on HF phenotype (3). Impaired renal function is independently associated with worse outcomes; in meta-analyses of 57 studies including trials and cohorts in HF, there was a 2-fold increase in the odds of death comparing patients with and without renal dysfunction (4).
Renal impairment in HF patients also affects the prescription, dosage, maintenance, and possibly effectiveness of therapies (5,6). Although common in clinical practice, patients with renal dysfunction have often been excluded from major clinical trials, creating an evidence gap for many HF patients and a discrepancy with clinical need. Those with moderate renal dysfunction (eGFR 45 to 59 ml/min/1.73 m2) and moderately severe renal dysfunction (eGFR 30 to 44 ml/min/1.73 m2) have a higher risk of adverse outcomes and potentially more absolute benefit from HF therapy, but in addition have multiple comorbidities that can have an impact on clinical management. Previous analyses of beta-blockers in patients with HF and renal dysfunction suggest that efficacy may be maintained at different levels of baseline eGFR (7–10). However, the number of patients and events in these studies were limited, particularly at the more severe end of renal impairment, and hence clinicians remain uniformed about any possible interaction of treatment effect.
We examined the effect of renal dysfunction on outcomes in patients with HF and reduced ejection fraction (HFrEF) using the totality of individual patient data (IPD) from the landmark, double-blind, randomized controlled trials (RCTs) comparing beta-blockers with placebo (11). The Beta-Blockers in Heart Failure Collaborative Group is a multinational project that has systematically harmonized clinical trial data to improve management and outcomes in patients with HF (12–15). In this study, we tested the hypothesis that compared with placebo, beta-blockers reduce mortality in patients with moderate and moderately severe renal dysfunction. Further, we looked at the prognostic impact of renal dysfunction and associated variables, and how change in renal function affects mortality.
The Beta-Blockers in Heart Failure Collaborative Group (Collaborative Systematic Overview of Randomised Controlled Trials of Beta-Blockers in the Treatment of Heart Failure [BB-meta-HF]) includes the lead investigators from the relevant trials, with support of the 4 pharmaceutical companies that sponsored them (AstraZeneca, GlaxoSmithKline, Merck Serono, and Menarini). This report was prepared according to the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) IPD guidance (16), and prospectively registered with ClinicalTrials.gov (NCT00832442) and the PROSPERO database of systematic reviews (CRD42014010012) (17).
Eligibility and search strategy
Detailed rationale and methods have previously been published (11–13). Placebo-controlled trials were eligible for inclusion if they recruited >300 patients, were not confounded by investigation of other treatments, had a planned follow-up of >6 months, and explicitly reported mortality as an endpoint.
Eleven trials were included that account for 95.7% of eligible participants recruited in RCTs based on a systematic review of published reports: ANZ (Australia/New Zealand Heart Failure Study) (18), BEST (Beta-Blocker Evaluation Survival Trial) (19), CAPRICORN (Carvedilol Post-Infarct Survival Control in LV Dysfunction Study) (20), CHRISTMAS (Carvedilol Hibernating Reversible Ischaemia Trial: Marker of Success Study) (21), CIBIS I (Cardiac Insufficiency Bisoprolol Study) (22), CIBIS-II (Cardiac Insufficiency Bisoprolol Study II) (23), COPERNICUS (Carvedilol Prospective Randomized Cumulative Survival Study) (24), MDC (Metoprolol in Idiopathic Dilated Cardiomyopathy Study) (25), MERIT-HF (Metoprolol CR/XL Randomised Intervention Trial in Congestive Heart Failure) (26), SENIORS (Study of the Effects of Nebivolol Intervention on Outcomes and Rehospitalisation in Seniors with Heart Failure) (27); and US-HF (U.S. Carvedilol Heart Failure Program) (28).
All included studies had appropriate ethical approval. Using the Cochrane Collaborations Risk of Bias Tool, we established that each trial had a low risk of bias (29).
Data collection and IPD integrity
A standardized data request form to obtain IPD from each trial has been published, along with search results and individual study demographics (11). IPD were obtained for all 11 trials identified in the systematic review, and data were extracted from original source files provided by the pharmaceutical companies and lead investigators. All data were cross-checked across different trial databases and compared with published reports. Discrepancies, inconsistencies, and incomplete data were checked against original case report forms and trial documentation to ensure IPD integrity. All 11 trial databases were then harmonized according to the standardized data request form to match patient characteristics and outcomes across all trials. Due to the small amount of missing data for relevant covariates, imputation was not performed.
For this analysis, we included patients with a baseline creatinine available and left ventricular ejection fraction (LVEF) <50% (the CIBIS I trial was excluded due to a lack of renal function data). Because we have previously identified a significant treatment interaction comparing sinus rhythm and atrial fibrillation/flutter (AF) (12), patients were stratified by heart rhythm on the baseline electrocardiogram for treatment estimates. Those with a missing electrocardiogram or paced rhythm were excluded.
Renal dysfunction and related variables
Creatinine values were obtained for each enrolled patient at baseline, the interim study visit and the final follow-up visit, where available. Renal function was analyzed using eGFR calculated with the Modification of Diet in Renal Disease (MDRD) equation for nonstandardized creatinine: 186 × (serum creatinine in mg/dl)−1.154 × (age)−0.203 × (0.742 if female) × (1.21 if African/African American) in ml/min/1.73 m2. eGFR was categorized according to the National Kidney Foundation staging: category 1, eGFR ≥90 ml/min/1.73 m2 (normal); category 2, eGFR 60 to 89 ml/min/1.73 m2 (mildly decreased); category 3a, eGFR 45 to 59 ml/min/1.73 m2 (mildly to moderately decreased); category 3b, eGFR 30 to 44 ml/min/1.73 m2 (moderately to severely decreased); and category 4 and 5 combined, eGFR <30 ml/min/1.73 m2 (severely decreased or kidney failure). We pre-defined patients with eGFR 30 to 59 ml/min/1.73 m2 as a group of clinical interest, and worsening renal function as 20% or greater reduction in eGFR between baseline and follow-up. When used as a continuous variable in interaction analyses, the lowest and highest 1% of eGFR values were excluded to avoid leverage of extreme results. Anemia was classified according to the World Health Organization definition (hemoglobin <13.0 g/dl in men and <12.0 g/dl in women), and proteinuria was defined as 1+ on a dipstick or ≥30 mg/dl.
The outcome for this analysis was all-cause mortality, which included additional deaths reported after the censor date for 7 studies (18–20,24,25,27,28). There were no patients with missing vital status.
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 (displayed as 25th to 75th quartiles). Group comparisons were made using the Kruskal Wallis nonparametric rank test. Fractional polynomials were used to find the best transformation of eGFR in adjusted analysis, including nonlinear relationships (for sinus rhythm, the best fit was the inverse square root, and for AF the inverse squared eGFR).
All analyses of beta-blockers versus placebo followed the principle of intention to treat. Outcomes were analyzed using a Cox proportional hazards regression model stratified by study and grouped by heart rhythm and eGFR category. 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. Hazard ratios (HRs) and 95% confidence intervals (CIs) are presented, along with corresponding p values. We pre-specified adjustment in Cox models for baseline age, sex, LVEF, history of myocardial infarction, systolic blood pressure, heart rate, use of angiotensin-converting enzyme inhibitors or angiotensin receptor blockers, and diuretic therapy. Only a minority of patients were followed for an extended period, and therefore data were censored at 1,200 days (3.3 years) from randomization.
Effect modification was assessed using p values from interaction terms fitted in the multivariable models. The interactions of continuous eGFR with mortality or beta-blocker efficacy were assessed using cubic splines in the Cox model and the Royston-Parmar flexible parametric survival model (30). There was no evidence of violation of the proportional hazards assumption in any multivariable model as determined by Schoenfeld residuals. Kaplan-Meier plots were used to graph the pooled, unadjusted data for eGFR/treatment groups, with log-rank p values for comparison. For worsening eGFR, analysis time began on the date of the final eGFR measurement (hence excluding any patients who had died, withdrawn consent, or were lost to follow-up before their repeat eGFR), and ends 2 years after this date.
Pre-defined sensitivity analyses in patients with eGFR 30 to 59 ml/min/1.73 m2 were: 1) additional multivariable adjustment for diabetes, body mass index, New York Heart Association (NYHA) functional class (I/II vs. III/IV), and use of digoxin or aldosterone antagonists; 2) effect estimate in patients with LVEF <35% compared with 35% to 49%; and 3) exclusion of CAPRICORN (the only post-infarct trial) and BEST (utilizing a pharmacologically distinct beta-blocker). We also performed sensitivity analyses using all available eGFR measurements for interaction analyses, rather than just the central 99%. Heterogeneity for pooled outcomes was assessed using the I2 statistic from a fixed-effects 2-stage model. We performed post hoc analyses: 1) according to dose achieved at interim follow-up; 2) to assess the relationship between proteinuria and worsening renal function; and 3) using the CKD-EPI (Chronic Kidney Disease Epidemiology Collaboration) eGFR calculator. A 2-tailed p value of 0.05 was considered statistically significant. Analyses were performed on Stata version 14.2 software (StataCorp, College Station, Texas) and R version 3.2.1 (R Core Team, Vienna, Austria).
A total of 16,740 HF patients were included from 10 RCTs (Online Figure 1). Median age was 65 years (25th to 75th centiles 55 to 72 years), 23% were women, and median LVEF was 27% (21% to 33%). Baseline median eGFR was 63 ml/min/1.73 m2. A total of 1,781 patients (10.6%) had an eGFR >90 ml/min/1.73 m2, 7,641 (45.6%) 60 to 89 ml/min/1.73 m2, 4,584 (27.4%) 45 to 59 ml/min/1.73 m2, and 2,286 (13.7%) 30 to 44 ml/min/1.73 m2. Only 448 patients (2.7%) had an eGFR <30 ml/min/1.73 m2, reflecting the exclusion criteria for several trials (Online Table 1). Patients in sinus rhythm (n = 13,861) had better renal function at baseline than those with AF (n = 2,879); 64 ml/min/1.73 m2 compared with 60 ml/min/1.73 m2, with 42.9% versus 48.9% of patients with eGFR <60 ml/min/1.73 m2, respectively. Other factors associated with more advanced renal dysfunction were older age and female sex, a longer duration of HF, an ischemic etiology, and concomitant diabetes (Table 1 shows data for sinus rhythm, and Online Table 2 for AF).
Impact of renal dysfunction on mortality in HFrEF
During a median follow-up of 1.3 years (0.8 to 1.9 years), eGFR was associated with all-cause mortality independent of other measured prognostic variables, with a 12% increase in the hazard of death for every 10 ml/min lower eGFR (95% CI: 10% to 15%; p < 0.001). Mortality was particularly high for patients with more severe renal dysfunction, and their cause of death was more often due to progressive heart failure (Figure 1, Online Table 3).
In the subset of patients with hemoglobin values, anemia was associated with higher mortality (HR: 1.35 compared with no anemia, 95% CI: 1.22 to 1.50; p < 0.001; n = 9,906). This was evident at all levels of cardiorenal dysfunction, except for patients with eGFR <30 ml/min/1.73 m2 (Online Figure 2). Proteinuria at baseline was also independently associated with higher mortality (HR: 1.32 compared with no proteinuria, 95% CI: 1.02 to 1.70; p = 0.034; n = 3,081). The largest prognostic impact of proteinuria was seen in patients with eGFR ≥90 ml/min/1.73 m2 (Online Figure 2); in this group, the death rate was 12.3% without proteinuria and 28.6% with proteinuria (adjusted p = 0.032).
Beta-blocker efficacy according to renal function at baseline
In 13,861 patients in sinus rhythm, beta-blockers reduced mortality across all patients (HR: 0.71; 95% CI: 0.66 to 0.78; p < 0.0001), including those with moderate and moderately severe renal dysfunction (Table 2). The adjusted HR for beta-blockers versus placebo in sinus rhythm was 0.73 for eGFR 45 to 59 ml/min/1.73 m2 (0.62 to 0.86; p < 0.001), and 0.71 for eGFR 30 to 44 ml/min/1.73 m2 (0.58 to 0.87; p = 0.001). Absolute risk reductions and number needed to treat (NNT) to prevent 1 death were 4.0% (NNT = 25) and 4.7% (NNT = 21), respectively. There were insufficient numbers of patients with eGFR <30 ml/min/1.73 m2 to be certain of any benefit or harm from beta-blocker therapy (95% CI: 0.87 to 1.91). We detected an interaction between beta-blocker efficacy and eGFR in sinus rhythm (p = 0.021), but with weak effect and only at the lowest end of the eGFR range (Central Illustration). In a sensitivity analysis that included all patients, including the extremes of eGFR, the interaction p value was 0.062.
The efficacy of beta-blockers in patients with sinus rhythm and moderate or moderately severe renal impairment was not affected by the presence of either anemia or proteinuria (interaction p = 0.69 and p = 0.24), or by LVEF or additional adjustment (Online Table 4). In 2,879 patients with AF at baseline, there was no significant reduction in mortality with beta-blockers in any category of eGFR, and no interaction of beta-blocker efficacy with continuous eGFR (p for interaction = 0.18) (Table 3, Figure 2).
Change in renal function over time, dose of therapy, and adverse events
Only a small drop in mean eGFR was noted overall from baseline to the last available measurement: 2.0 ml/min/1.73 m2 lower eGFR after a median of 1.2 years for 7,420 surviving patients in sinus rhythm (SD ±15.2). In 3,179 patients with either moderate or moderately severe renal impairment at baseline, there was an increase of 1.3 ml/min/1.73 m2 (SD ±13.1). We observed little difference between patients randomized to beta-blockers or placebo (Figure 3, Online Table 5).
Worsening renal function of 20% or greater during follow-up was observed in 1,342 (18.1%) of patients in sinus rhythm, and was associated with a 28% higher adjusted risk of death during the subsequent 2 years (95% CI: 9% to 49%; p = 0.002; n = 4,725). The corresponding mortality increase was 46% for the subset of patients with combined moderate or moderately severe renal impairment at baseline (95% CI: 14% to 87%; p = 0.002; n = 2,175) (Figure 3).
In patients with sinus rhythm and moderate or moderately severe renal impairment, 77% managed to reach one-half of the target dose of beta-blocker or greater (compared with 80% with eGFR 60 to 89 ml/min/1.73 m2 and 84% with eGFR ≥90 ml/min/1.73 m2) (Online Table 6). These patients had substantially better prognosis than those at lower dose levels (Online Figure 3); however, the same pattern was also seen in patients randomized to placebo. Across both sinus rhythm and AF, discontinuation of beta-blockers was similar to that seen with placebo in patients with moderate and moderately severe renal dysfunction. This was the case for all adverse events leading to therapy discontinuation, and also those specifically related to renal impairment (Online Table 7). Overall, discontinuation rates for both beta-blockers and placebo were higher in those with more advanced renal dysfunction.
Renal impairment is often a barrier in clinical practice for the commencement and up-titration of guideline-recommended HFrEF therapy (5,6). Most patients with HFrEF have some degree of renal impairment, yet many randomized trials have excluded those with significant renal dysfunction, leading to concerns by clinicians about efficacy and safety. Using robust and high-quality data from the landmark beta-blocker trials, we have demonstrated with sufficient sample size that beta-blockers are effective in reducing mortality in patients with HFrEF and sinus rhythm, even in those with moderately severe renal dysfunction (as low as an eGFR of 30 to 44 ml/min/1.73 m2). Despite higher rates of comorbidities, the absolute benefit in this group was similar to patients with eGFR >90 ml/min/1.73 m2. Discontinuation due to adverse events was the same for both beta-blockers and placebo in these double-blind trials and renal function did not appear to worsen even in those with kidney dysfunction at baseline. These results suggest that renal impairment should not obstruct the prescription of beta-blockers in patients with HFrEF.
Prognostic implications of renal dysfunction, anemia, and proteinuria
The prognosis for patients with HFrEF has many determinants, and renal dysfunction is a well-known contributor to adverse outcomes (4). Our data highlight the different pattern of patient prognosis according to the severity of renal dysfunction, with a complete reversal in sudden cardiac death and death due to progressive HF comparing preserved and severe renal dysfunction. Whereas anemia was associated with higher mortality across most patients (with the exception of eGFR <30 ml/min/1.73 m2, where renal replacement therapy and erythropoietin come into play ), proteinuria had the most marked impact in those with preserved renal function. Proteinuria at baseline was not associated with a higher chance of worsening renal function (post hoc p = 0.61) and so could be a useful independent marker of elevated risk (despite apparently normal renal function), or even of insufficient HF therapy (32).
HF treatment in patients with renal impairment
Our results highlight the importance of appropriate HFrEF therapy for all patients, especially those with renal insufficiency who could benefit the most. Subgroup analyses from both the Val-HeFT (Valsartan in Heart Failure Trial) (N = 2,346) and the RALES (Randomized Aldactone Evaluation Study; N = 792) trials showed that the benefit of valsartan and spironolactone were consistent in patients above and below an eGFR of 60 ml/min/1.73 m2 (33,34). In more recent trials, no interactions have been noted for renal impairment using this same eGFR cutoff for eplerenone (n = 912) or sacubitril-valsartan (n = 3,061) (35,36). However, because the median eGFR in clinical practice is often around 60 ml/min/1.73 m2, it could be argued that a more realistic cutpoint is required to reassure clinicians about the safety and efficacy of therapy. Unfortunately, data specifically in patients with moderately severe renal dysfunction are limited (37). Subgroup analysis from the MERIT-HF trial (8), as well as the CIBIS-II (10) and SENIORS (9) trials, suggested potential benefit from beta-blockers in those with eGFR <45 ml/min/1.73 m2, hence the need for this analysis that included all of these RCTs and more. HFrEF patients with severe renal dysfunction or kidney failure have largely been excluded from RCTs. One exception was a small trial in 114 dialysis patients, where carvedilol was found to improve clinical status (38). Although we have pooled data from 10 placebo-controlled trials of beta-blockers, only 448 patients (2.7%) had an eGFR <30 ml/min/1.73 m2. Event rates were high, but due to the restricted sample size, we are unable to comment on the true efficacy of beta-blockers in this patient group; new RCTs are clearly warranted to address this knowledge gap.
Worsening renal function
Clinicians are often concerned about the potential for worsening renal function during initiation or up-titration of HF therapy. We show that beta-blockers do not lead to any overall deterioration in renal function in those with existing impairment. The results we present on worsening renal function (not caused by initiation of renin-angiotensin-aldosterone system inhibitors) are similar to other studies showing that deterioration of function is associated with higher mortality (39). Our data suggest that preservation of renal function may be an important management goal. However, post-randomization variables such as repeated eGFR measurement should be judged carefully even in double-blind trials, as they are prone to the same biases as observational data. Achieving target dosage remains an essential task for HF teams. We demonstrate that this is achievable for the majority of patients with renal dysfunction, even those with moderately severe impairment. However, dose is a complicated variable also affected by physician- and patient-level biases, as highlighted by the marked difference in mortality according to the placebo dose attained.
Patients with HF and concomitant atrial fibrillation
We pre-specified stratification of analyses by heart rhythm due to significant interactions with beta-blocker efficacy (12) and a marked difference in the association of heart rate with mortality comparing sinus rhythm with AF (14). The lack of benefit from beta-blockers regardless of eGFR in patients with AF was therefore unsurprising. Similar to sinus rhythm, we show that renal dysfunction in patients with AF is associated with more high-risk features, but the majority of patients can reach appropriate dosage. Higher rates of renal dysfunction in AF and worse prognosis across all eGFR categories compared with sinus rhythm demonstrate the need for improved multidisciplinary management of concomitant AF and HF (40).
Study strengths and limitations
The current analysis uses IPD from 10 landmark, placebo-controlled trials. An extensive period of additional data collation was performed from original case report forms, including events not originally reported, with data cleaning and harmonization according to a published design (11). Each of the trials had different inclusion and exclusion criteria, although treatment effects were similar for eGFR 30 to 59 ml/min/1.73 m2 even when the 2 more unique trials were excluded. Exclusion of these 2 RCTs led to no significant heterogeneity in treatment effect in the remaining studies (Online Table 4). Due to the large sample size, we were able to test the interaction of beta-blocker efficacy across the range of continuous eGFR. We confirmed nonlinearity and used fractional polynomial transformations to obtain optimal model fitting in sinus rhythm and AF. However, despite using nonlinear approaches, splines, and flexible parametric models, interaction tests have relatively low power, and we may have missed a clinically significant interaction in small subgroups. The MDRD formula was used for primary estimation of kidney function as the CKD-EPI calculation requires standardized creatinine measurements, although we saw no difference in treatment effects in a post hoc analysis. Due to our prior findings (15), we included patients with both mid-range and reduced LVEF in the current study and demonstrated similar efficacy from beta-blockers regardless of baseline (reduced) LVEF. As discussed, any analysis of post-randomization variables (follow-up eGFR, dose, and so on) should be considered exploratory and may be affected by regression toward the mean, and selection or survivor bias. We have deliberately not addressed hospital admissions in this paper, due to concerns that kidney disease can itself influence the likelihood of a physician admitting a patient, lead to withdrawal of other heart failure therapy, and confound the association with adverse outcomes. Finally, this a retrospective analysis, and further new RCTs should be encouraged in view of the commonality of renal dysfunction in HF.
Combining double-blind, individual patient-level data has provided a sufficient sample size to confirm the efficacy of beta-blockers in heart failure patients with reduced ejection fraction, sinus rhythm and renal dysfunction, including those with eGFR 30 to 44 ml/min/1.73 m2, the lowest range of eGFR tested in large placebo-controlled trials.
COMPETENCY IN MEDICAL KNOWLEDGE: Beta-blocker therapy reduces mortality in patients with heart failure and reduced ejection fraction in sinus rhythm, including those with moderate (eGFR 45 to 59 ml/min/1.73 m2) and moderately severe (eGFR 30 to 44 ml/min/1.73 m2) renal dysfunction.
TRANSLATIONAL OUTLOOK: Further research is needed in patients with atrial fibrillation, in whom beta-blockers are not associated with lower mortality rates, and those with severe renal impairment (eGFR <30 ml/min/1.73 m2), about whom there are limited data.
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, and the patients who contributed to these trials. This work is dedicated to the memory of Douglas Altman (1948–2018; University of Oxford, United Kingdom), Henry Krum (1958–2015; Monash University, Melbourne, Australia), and Philip Poole-Wilson (1943–2009; Imperial College London, United Kingdom). This project was only possible 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 BEST research materials obtained through the National Heart Lung and Blood Institute Biologic Specimen and Data Repository Information Coordinating Center; the manuscript does not reflect the opinions or views of BEST or the NHLBI. The Steering Committee Lead (Dr. Kotecha) and the Centre for Statistics in Medicine, Oxford (Dr. 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.
Menarini Farmaceutica Internazionale provided an unrestricted research grant for administrative costs, GlaxoSmithKline provided data extraction support, IRCCS San Raffaele provided a collaborative research grant, and the British Heart Foundation supports the University of Birmingham Institute of Cardiovascular Sciences with an Accelerator Award (AA/18/2/34218). None of the pharmaceutical groups had any role in data analysis or manuscript preparation. 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 BHF, NIHR or the U.K. Department of Health; has received personal fees from Bayer and Atricure, outside the submitted work; has been chief investigator for the RAte control Therapy Evaluation in permanent Atrial Fibrillation trial (RATE-AF; NCT02391337); and has received funding through a British Heart Foundation Project Grant (PG/17/55/33087), and EU Innovative Medicines Initiative Collaboration Grant (BigData@Heart; 116074). Dr. Gill has received funding through the BigData@Heart Innovative Medicines Initiative grant no. 116074. Prof. Packer has received personal fees from Abbive, Akcea, Actavis, Amgen, AstraZeneca, Bayer, Boehringer Ingelheim, Cardiorentis, Daiichi-Sankyo, Gilead, Johnson & Johnson, NovoNordisk, Pfizer, Relypsa, Sanofi, Synthetic Biologics, and Theravance, outside the submitted work. Prof. Böhm has received personal fees from Abbott, Amgen, AstraZeneca, Bayer, Boehringer Ingelheim, Servier, Medtronic, Vifor, and Novartis, outside the submitted work. Prof. McMurray has received grants from the British Heart Foundation Centre of Research Excellence Grant RE/18/6/34217, during the conduct of the study; and nonfinancial support from Bayer, Cardiorentis, Amgen, Oxford University/Bayer, Theracos, Abbvie, DalCor, Pfizer, Merck, AstraZeneca, GlaxoSmithKline, Novartis, Bristol-Myers Squibb, Vifor-Fresenius Pharma, and Kidney Research UK, outside the submitted work. Prof. Anker has received grants and personal fees from Vifor Int, and Abbott Vascular; and has received personal fees from Bayer, Boehringer Ingelheim, Novartis, Servier, Impulse Dynamics, AstraZeneca, and St. Jude Medical, outside the submitted work. Dr. von Lueder has received fees from AstraZeneca, Vifor, Pharmacosmos, Novartis, Pfizer, and Boehringer Ingelheim. Prof. Andersson has received grants from AstraZeneca, and Federal research grants, during the conduct of the study. Prof. Wedel has received personal fees from AstraZeneca, during the conduct of the study. Prof. Ruschitzka has received grants and personal fees from St. Jude Medical/Abbott, Servier, Bayer, and Novartis; has received personal fees from Zoll, AstraZeneca, Sanofi, Amgen, BMS, Pfizer, Fresenius, Vifor, Roche, Cardiorentis, and Boehringer Ingelheim; has received other fees from HeartWare; and has received grants from Mars, all are outside the submitted work. Prof. Cleland has received grants and personal fees from Amgen, Novartis, Vifor, and Stealth Biotherapeutics; has served on Advisory Boards for Abbott and Vifor; and has received nonfinancial support from Medtronic and Boston Scientific, all outside the submitted work. Prof. Coats has received personal fees from AstraZeneca, Menarini, Novartis, Nutricia, Servier, Vifor, Actimed, Enopace, Faraday, WL Gore, Respicardia, Stealth Peptides, and V-Wave, all outside the submitted work. 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
- estimated glomerular filtration rate
- heart failure
- heart failure with reduced ejection fraction
- hazard ratio
- individual patient data
- left ventricular ejection fraction
- number needed to treat
- randomized controlled trials
- Received August 30, 2019.
- Revision received September 11, 2019.
- Accepted September 16, 2019.
- 2019 The Authors
- Lofman I.,
- Szummer K.,
- Hagerman I.,
- Dahlstrom U.,
- Lund L.H.,
- Jernberg T.
- Hillege H.L.,
- Nitsch D.,
- Pfeffer M.A.,
- et al.
- Yancy C.W.,
- Jessup M.,
- Bozkurt B.,
- et al.
- Wali R.K.,
- Iyengar M.,
- Beck G.J.,
- et al.
- Ghali J.K.,
- Wikstrand J.,
- Van Veldhuisen D.J.,
- et al.
- Kotecha D.,
- Manzano L.,
- Krum H.,
- et al.
- Kotecha D.,
- Flather M.D.,
- Altman D.G.,
- et al.
- Cleland J.G.F.,
- Bunting K.V.,
- Flather M.D.,
- et al.
- Kotecha D.,
- Manzano L.,
- Krum H.,
- Altman D.G.,
- Holmes J.,
- Flather M.
- CIBIS Investigators and Committees
- Higgins J.P.T.,
- Altman D.G.,
- Sterne J.A.C.
- Vardeny O.,
- Wu D.H.,
- Desai A.,
- et al.
- Damman K.,
- Tang W.H.,
- Felker G.M.,
- et al.
- Cice G.,
- Ferrara L.,
- D'Andrea A.,
- et al.