Nearly six million Americans have heart failure, a leading driver of health care costs in the United States.
The “stiff heart” heart failure variant accounts for about half of all cases and the vast majority of such patients take beta-blocker medications despite unclear benefit from their regular use.
A new publication in JAMA Network Open links use of beta-blockers to heart failure hospitalizations among those with this common “stiff heart” heart failure subtype.
Heart failure occurs when the heart cannot meet the body’s demands. About half of patients have heart failure characterized by a normal squeeze but impaired relaxation of the heart muscle from a “stiff heart”.
This is also known as heart failure with preserved ejection fraction.
The other half of cases are due to a “weak heart” with an abnormal squeeze, also known as heart failure with reduced ejection fraction.
Beta-blockers – medications that lower the heart rate and blood pressure – are strongly recommended in national guidelines for treatment of “weak heart” heart failure because of their clear benefit.
“A big problem with ‘stiff heart’ heart failure is that we don’t have effective medical therapies,” says coauthor Timothy Plante, M.D., an assistant professor of medicine at the Larner College of Medicine at the University of Vermont.
“So, instead, we use the same medications that work for ‘weak heart’ heart failure.
Because beta-blockers save lives in ‘weak heart’ heart failure, we assume they are also effective in ‘stiff heart’ heart failure patients – this assumption may be wrong.”
Plante, lead author Daniel Silverman, M.D., senior author Markus Meyer, M.D., and colleagues analyzed data from the National Institutes of Health-funded TOPCAT (Treatment of Preserved Cardiac Function Heart Failure with an Aldosterone Antagonist) study, a trial of the medication spironolactone in patients with “stiff heart” heart failure. About four out of five study participants were on beta-blockers.
The researchers found beta-blocker use to be a risk factor for hospitalizations for heart failure among these patients with “stiff heart” heart failure.
“Beta-blocker use was associated with a 74 percent higher risk of heart failure hospitalizations among participants with heart failure and a normal pump function,” says Meyer, an associate professor of medicine at the University of Minnesota Medical School.
Despite their common use, the authors note that beta-blocker use in “stiff heart” heart failure has not been sufficiently studied.
This publication extends their prior work, which found that halting beta-blockers markedly improves levels of the heart failure blood test known as BNP among patients with “stiff heart” heart failure.
“In ‘stiff heart’ heart failure, the heart is less able to relax and fill with blood. Beta-blockers appear to increase pressures inside the heart.
This may lead to symptoms like worsening shortness of breath and retention of fluid,” says Silverman, a cardiology fellow and clinical instructor in medicine at the University of Vermont Medical Center and Larner College of Medicine.
“Even people without heart failure will have more shortness of breath and less exercise capacity. This has been a known class side effect for decades,” says Meyer. “It is important to understand that our findings are not proof that beta-blockers are harmful among patients with ‘stiff heart’ heart failure—it is just a concerning signal.”
They believe their findings warrant a clinical trial to evaluate the safety and effects of beta-blockers in patients with “stiff heart” heart failure. “There are some big next steps, like reproducing this finding in other studies and testing if there is a benefit of stopping beta-blockers in patients with ‘stiff heart’ heart failure,” says Silverman.
Provided by University of Vermont
Chronic heart failure (CHF) remains a public health problem attaching growing attention since recent years, for its high prevalence (with an increase 46% from 2012 to 2030 in U.S.A.) and low 5-year survival rates (with estimating 50% in U.S.A.) [1, 2], which are much worse than some types of cancers [3].
Atrial fibrillation (AF) is one of the most common complications of CHF, with report of prevalence up to 50% in CHF patients [4, 5].
CHF and AF could coexist and interact with each other, promoting the development of cardiac dysfunction and increasing the risk of mortality [6, 7]. It has been reported that CHF with new-onset AF had greater mortality in 1 year [5].
β-blockers play an important role in treatment of heart failure with reduced ejection fraction (HFrEF). Large amounts of studies confirmed that it could decrease the rate of mortality and hospital readmission for HF of HFrEF in sinus rhythm (SR) [8–10] and β-blockers had been recommended as Class IA in guidelines in treating those CHF with SR patients by both U.S.A. and Europe [4, 11].
Also, β-blockers is the first-line rate control treatment in AF and has received a Class IB recommendation in treating patients with AF with left ventricular ejection fraction (LVEF) ≥ 40% or < 40% in the latest recommendation in Europe [12].
However, it still lacks direct and strong evidence from data of large randomized control trials (RCTs) which are originally designed for patients with CHF and AF and there has been a controversy about the benefit of β-blockers in treating CHF patients with AF. A meta-analysis with four RCTs indicated that β-blockers did not decrease the risk of all-cause mortality and HF hospitalization in CHF patients with AF [10].
Also, according to individual patient-level meta-analysis conducted by β-Blockers in Heart Failure Collaborative Group, which included data from RCTs, β-blockers were not associated with the with decreasing the risk of mortality, nor the hospital admission outcomes [9], regardless of LVEF [13] or heart rate (HR) [14].
However, registry study indicated an opposite view that β-blockers treatment could decrease ll-cause mortality in patients with both CHF and AF. Recently, the European Society of Cardiology-Heart Failure (ESC-HF) Registry demonstrated that β-blockers could reduce the all-cause mortality in CHF and AF patients [Hazard ratio: 0.52; 95% Confidence Interval (CI) 0.31–0.89; p = 0.02], especially in the group with patients’ HR between 80 and 109 b.p.m. (beats per minute) [15]. Similarly, findings from one cohort study also support the idea that β-blockers could show a reduction of all-cause mortality in AF and CHF groups (Hazard ratio: 0.63, 95% CI: 0.50–0.79) [16].
Due to these inconsistent results in RCTs and registry studies, we performed a meta-analysis stratified by study design; to examine the effect of β-blockers on outcomes (i.e. all-cause mortality, cardiovascular mortality and hospitalization for HF) in CHF combined AF patients.
Discussion
The main finding of the present meta-analysis indicates that β-blockers treatment could decrease the risk of all-cause mortality in patients with CHF and AF, but did not reduce the risk of cardiovascular mortality and HF hospitalization. We also found that β-blockers treatment only shows a reduction in observational study group, whereas the reduction in subgroup analysis of RCT group was not significant.
The effect of β-blockers treatment in patients with CHF and AF has been reported by several meta-analyses. One meta-analysis with four RCTs indicated that β-blockers did not decrease the risk of all-cause mortality and HF hospitalization in CHF patients with AF [10]. Another previous individual-patient-data meta-analysis performed by Kotecha et al., which included ten RCTs with 18,254 CHF patients (containing 3066 with AF), also showed that β-blockers treatment had no effect in reducing all-cause mortality in patients who had both AF and CHF [9]. It should be noted that RCTs included in these meta-analyses was originally designed for patients with CHF and patients with CHF and AF only comprised around 19% of the whole group. So there may be a less power to detect benefits of β-blockers treatment in patients with CHF and AF. In order to have a more comprehensive understanding, we not only included data from but also from observational studies in this meta-analysis. Besides, in our study, we used a wider search strategy with more search terms, including “atrial fibrillation”, “heart failure”, “cardiac dysfunction”, “heart dysfunction”, “cardiac failure”, “heart weakness”, “beta blockers”, “adrenergic beta antagonists”, “bisoprolol”, “nebivolol”, “carvedilol”, “bucindolol”, “metoprolol”, “atenolol” and “ metoprolol CR/XL” whereas only “atrial fibrillation”, “heart failure”, “beta blockade”, “beta-blocker therapy” and “medical therapy” were used to detect associated studies in the prior meta-analysis. We believe that our wider search strategy is important for meta-analysis to avoid missing potentially relevant studies.
Meta-analyses may be biased when the literature search fails to identify all relevant studies.
Several explanations may be contributed the difference in these studies. First, the dose of beta blockers may be different in subgroup analysis of RCTs and observational studies. Patients who receive higher doses of β-blockers may have more chance exposed to its side effects such as hypotension and bradyarrhythmias [32], which may counterbalance its benefits in improving conditions in AF and CHF patients.
As shown in post-hoc analysis of RCTs, β-blockers were required to titrated to target dose or the tolerance dose for patients in follow-up period which were 25 mg twice daily [23], 10 mg of nebivolol daily [26], 10 mg of bisoprolol daily [25], 154 mg metoprolol CR/XL daily [27], respectively.
The individual-patient-data meta-analysis by Kotecha et al. showed that 72.1% patients with AF and CHF received maximum does of β-blockers [9]. However, in observational studies, which were much more revealing the medication prescription for patients in real world, the prescribed dosage of β-blockers may be depended on conditions of each patient and be more individualized, so dosages of β-blockers for patients may not be as high as those in RCTs. In one registry, Li et, al pointed out that only 30% of patients in AF and CHF treating with β-blockers reached its target does [30].
The target does of β-blockers in treating patients with CHF may not benefit patients with CHF and AF. A previous study showed that 50% of the target does of β-blockers linked with a better prognosis in patients in AF and HFrEF [33].
Up to now, we still lack studies to learn which dose of β-blockers could give a better prognosis to patients with CHF and AF. Second, the baseline HR and achieved HR with β-blockers treatment may be different in post-hoc analysis of RCTs and observational studies. In RACE (Rate Control Efficacy in Permanent Atrial Fibrillation) II study, results showed that permanent AF patients in strict rate control group (resting HR ≤ 80 b.p.m., and a HR ≤110 b.p.m. during moderate exercise) was not better than those in lenient rate control (resting HR < 110 b.p.m.) in all-cause mortality [34].
An observational study pointed out that in patients with CHF and AF, each increase in resting HR of 10 b.p.m. at baseline was associated with a 7% decrease in mortality per year and HR < 73b.p.m. in patients with CHF and AF seemed to have a worse survival [35].
Likewise, a registry study showed that AF and CHF patients with HR control between 80b.p.m.-109b.p.m. had better prognosis compared with other two groups (< 80 b.p.m. and ≥ 110b.p.m.) [15]. Thus, different benefits reported in subgroup analysis of RCTs and observational studies may be due to the relatively low controlled heart rate in post-hoc analysis of RCTs, which would lead to side effects of β-blockers such as atrioventricular block.
It is suggested that patients with HF and AF should pay attention to the ventricular rate, and the optimal dosage of β-blockers should be the dose that helps patients achieve the target ventricular rate, but not the dose designed in RCT which neglect patients’ target ventricular rate. Third, we also found that except for β-blockers, other medicines treating for AF and CHF may be different in both groups. For example, patients in post-hoc analysis of RCT group seemed to more likely to receive digoxin (mean 65.4%) than those in registry study group (mean 29.5%).
It has been demonstrated that digoxin could increase the risk of all-cause mortality in AF patients with or without CHF [36]. As digoxin has negative effect on atrioventricular conduction [37], the therapy combining digoxin and β-blockers may exacerbate the possibility of atrioventricular block.
Thus, in treating patients with AF and HF, clinicians should be more cautious when using combined medications of digoxin and β-blockers and avoid excessive dosage of these two drugs.
According to our analysis, β-blockers could reduce the risk of all-cause mortality among AF-CHF patients and should be considered as the first-line therapy for controlling rate. However, we could not ignore that there was not a significantly reduction in cardiovascular mortality and HF hospitalization when using β-blockers treatment, which require us to further investigate an optimal treatment scheme for patients with AF and CHF [38–40].
As previous studies mentioned, rate control were equivalent to rhythm control in reducing rates of all-cause mortality, cardiovascular mortality, thromboembolism in patients with AF, as well as in patients with AF and HFrEF [41].
Also, a study indicated that side-effect of drugs in both rate-control therapy and rhythm-therapy may contribute to progression of diseases and worse prognosis in AF-CHF patients [42].
However, recent report from Catheter ablation versus standard conventional treatment in patients with left ventricular dysfunction and atrial fibrillation (CASTLE-AF) trial demonstrated that AF ablation could significantly reduce composite endpoints of all-cause mortality and hospitalization for worsening HF in 60 months of follow-up (hazard ratio 0.62; 95% CI 0.43–0.87 P = 0.006) in enrolled AF and CHF patients [43]. Besides, previous studies showed that AF ablation was superior to rate control using medications in improving left ventricular function [44].
So AF ablation may become first choice for treating patients with AF and CHF in future, but its benefits need further studies.
We should note that our meta-analysis include both subgroup analysis of RCTs and observational studies and the outcome in two groups were different when analysis in stratifying by study design. Although large RCTs are considered as the highest level of evidence in professional societies, potential bias, such as patient selection and enrollment cannot be completely avoided.
Furthermore, clinical heterogeneity (e.g different baseline characteristics, risk profile of patients and different pharmacological profiles) between studies included in individual patient data meta-analysis may also made the combined results misleading.
Despite their often stated limitations, large sample observational studies can provide valuable information, which are critical to posing relevant questions in real world practice and help to inform the planning and design of RCTs. Taken these evidence together, we considered that: (1) Beta-blockers are still the first line medications for heart rate control in patients with HF and AF; (2) The effect on mortality of beta-blockers in patients with HF and AF should be furthermore evaluated based on HR strata.
Limitations
There are several limitations of our meta-analysis which should be considered. First, as there was not RCTs were not initially designed for patients with AF and CHF, we could only include post-hoc analysis of RCTs which was originally designed for CHF patients and that may limit our understanding of the real effect of β-blockers on patients with CHF and AF.
Second, the LVEF of patients in most articles we included was < 40% and we lacked sufficient data on patients whose LVEF ≥40%, so it is hard for us to extrapolate the conclusion to the whole population of AF and CHF which contain both heart failure with perserved ejection fraction and HFrEF.
Third, since there was not sufficient data or individual data provided in the included studies, we were not able to perform more subgroup analysis except for study design, and some confounding factors maybe underestimated.
Forth, included articles did not report the whether patients were still on treatment during follow-up period, so it is hard for us to evaluate if the compliance would affect the results and how it could affect the result.
Fifth, although we adopt random model analysis, we were still unable to avoid the inherent heterogeneity from different articles due to their different follow-up durations, treatments and population, etc. Sixth, thought we did an entire search in published data, we could not include those unpublished articles and it is needed for us to give further analysis.
Conclusion
β-blockers treatment was associated with significantly decrease the risk of all-cause mortality in patients with AF and CHF but not in reducing rates of cardiovascular mortality and HF hospitalization. The association between β-blockers treatment and the reduction of all-cause mortality was only seen in observational study group, whereas it was not significant in RCT group. Further RCTs targeting AF and CHF patients with β-blockers therapy as well as studies of new therapy, such as AF ablation, are needed for our better understanding of the management of patients with AF and CHF.
