Heart Rate Control in Heart Failure with Reduced Ejection Fraction: The Bright and the Dark Side of the Moon

This article refers to ‘Relationship between heart rate and outcomes in patients in sinus rhythm or atrial fibrillation with heart failure and reduced ejection fraction’ by K.F. Docherty et al., published in this issue on pages 528–538. Resting heart rate predicts prognosis in patients with heart failure with reduced (HFrEF) and preserved ejection fraction (HFpEF); however, only in the subset of patients with sinus rhythm. In contrast, in heart failure patients with atrial fibrillation (AF), ventricular rate does not predict outcomes.1, 2 Clinical studies describing these relationships so far were limited by the fact that endpoints had often not been prospectively collected and adjudicated, and/or by the lack of adjustment for other powerful predictors of outcome, e.g. circulating natriuretic peptides.3 Against this background, in the current issue of the Journal, Docherty et al.4 present important data showing for the first time that even after adjustment for N-terminal pro-B-type natriuretic peptide (NT-proBNP) levels, an elevated heart rate was an independent predictor of adverse cardiovascular outcomes in HFrEF patients in sinus rhythm, but not in patients with AF. In patients with sinus rhythm, both when analysed as tertiles but also as continuous variable, heart rate at baseline – after adjustment for many clinical parameters including natriuretic peptide levels – was highly predictive for all-cause mortality, cardiovascular mortality, heart failure hospitalization, pump failure-related death and sudden cardiac death. Heart rate also correlated with symptoms and quality of life. These associations were not observed in patients with AF. The particular strengths of the current analysis is (i) the derivation of the data from two very large randomized trials in HFrEF with meticulously adjudicated clinical endpoints, (ii) the adjustment for many prospectively assessed clinical and laboratory variables including NT-proBNP as well as drug treatment, and (iii) the derivation of rhythm and heart rate from baseline electrocardiograms (ECGs) in all patients, thus avoiding potential underestimation of ventricular heart rate by pulse taking in patients with AF. Nevertheless, as in all retrospective analyses, despite widespread correction for potential bias, unknown confounders cannot be completely excluded.5, 6 The authors also analysed the prognostic association of an increase or a decrease in heart rate of >10 bpm in their cohort during 12 months by assessing follow up ECGs in all patients.4 Again, in patients in sinus rhythm, clinical endpoints and quality of life were highly significantly worse in the patients with a heart rate increase compared to baseline. Interestingly, in patients with AF experiencing an increase or decrease in heart rate of >10 bpm over time, similar trends were seen although most of these did not reach statistical significance. Why lower heart rates are associated with clear beneficial outcome only in patients in sinus rhythm but not in AF remains unclear.1 The strength of the current study to derive heart rate from the 12-lead ECG might also be a disadvantage for assessing heart rate in AF. The snapshot of a 12-lead ECG may be too short to reliably measure heart rate in AF. Old studies suggest a minimum of 60 s time frame.7 Possibly mean heart rate derived from 24 h recording might be an even better indicator of ‘true’ heart rate in AF. Where to go from here? What clinical consequences for the treatment of HFrEF patients have to be drawn from the results? Clearly, maintenance of sinus rhythm appears to be an important goal in patients with HFrEF, and the application of beta-blockers to achieve a heart rate of 50–60 bpm, with the addition of ivabradine if heart rate is persistently >70 bpm, lowers mortality and heart failure hospitalization, and improves left ventricular remodelling and quality of life (Figure 1). In contrast, how aggressive heart rate control as well as rhythm control should be pursued in HFrEF patients with AF is still highly controversial. Regarding control of ventricular heart rate in HFrEF patients with AF, the lack of evidence is reflected in the discrepancy between guideline recommendations; while the 2014 American AF guideline proposes a target resting ventricular rate < 80 bpm, the European guidelines differ: the 2016 European Society of Cardiology (ESC) guideline for AF is comfortable with a target of <110 bpm, the 2016 ESC guideline for heart failure mentions a target of 60–100 bpm whereas the 2019 Heart Failure Association clinical practice update suggests that a heart rate of 70–90 bpm is preferable8-10 (Figure 1). As only a minority of patients in the RACE II trial had HFrEF, a rate goal of <110 bpm in patients with HFrEF and AF does not appear to be based on hard evidence.11 However, while the benefit of beta-blocker therapy in HFrEF patients in sinus rhythm is out of doubt, in patients with AF, Kotecha et al.12 did neither observe any prognostic impact of baseline heart rate on mortality nor an association of heart rate reduction in response to beta-blocker therapy with improved prognosis in any of the baseline heart rate tertiles. Thus, up-titration of beta-blocker dose in HFrEF patients in AF is not advocated to improve outcomes, but may be indicated to improve symptoms, e.g. palpitations. Due to side effects, amiodarone is not indicated for heart rate control in HFrEF with AF. Digitalis glycosides are used quite often for this purpose; however, as the DIG trial did not include patients with AF, current evidence is scarce. The ongoing DIGIT-HF trial13 (www.digit-hf.de, EudraCT-No.: 2013–005326-38) as well as the DECISION trial (ClinicalTrials.gov Identifier: NCT03783429) will provide data in the future. In the meantime, digitalis glycosides may be used in selected patients, provided that the levels of the drugs are controlled regularly (target digoxin: 0.5–0.9 ng/mL; digitoxin: 8–18 ng/mL), especially in the case of digoxin that is cleared predominantly by the kidney.6 While in retrospective non-randomized analyses digoxin was supposed to be associated with worse outcomes, further analyses have shown that these assumptions result mainly from prescription bias and not from an adverse drug effect.5, 6 Recently in older patients hospitalized for heart failure, discontinuation of pre-admission digoxin therapy was even associated with poor outcomes, albeit this study was not randomized.14 Thus, in selected symptomatic patients with HFrEF and AF on maximally tolerated beta-blocker dose, a treatment approach with digitalis glycosides to control heart rate seems to be justified. While a rate control strategy for persistent AF avoids invasive procedures and potentially toxic drugs, in patients with HFrEF and symptomatic AF or AF-related tachycardiomyopathy, a rhythm control strategy may be advisable.10 Although the CASTLE-AF as well as the CABANA trial suffer from several shortcomings limiting their generalizability to the broad spectrum of HFrEF patients,10, 15 catheter ablation of AF of patients with heart failure and AF may be considered in symptomatic patients despite implementation of guideline-recommended pharmacological and device therapy. In experienced centres catheter ablation of AF is a safe procedure with low complication and mortality rates.16 Although catheter ablation of AF does not completely suppress AF episodes in many patients, it is clearly more effective than long-term treatment with amiodarone.16, 17 In HFrEF patients with persistent AF, before proceeding to catheter ablation, electrocardioversion with or without amiodarone treatment may be performed, to get clearer insight whether restoration and maintenance of sinus rhythm leads to improvement of symptoms. In selected patients with refractory high ventricular rate non-responsive to other therapeutic approaches, implantation of a cardiac resynchronization device followed by atrioventricular node ablation is reasonable.18 Docherty et al.4 have to be complimented for an important study in a well-characterized HFrEF cohort derived from two large trials with similar inclusion criteria, and excellent adjustments for all major clinical and laboratory variables including for the first time natriuretic peptide levels. Their results remove the last doubts that baseline heart rate, as well as heart rate control over time, are essential for outcomes and quality of life in HFrEF patients in sinus rhythm. However, they also remind us that we are still far from evidence-based treatments for the huge number of patients with AF and often advanced HFrEF. Further studies are needed to get better insights into predictive parameters and effective therapeutic approaches for these patients. J.B. received funding from the German Ministry of Research and Education (BMBF) for the conduct of the DIGIT-HF trial. Conflict of interest: J.B. received honoraria for lectures/consulting from Novartis, Vifor, Bayer, Servier, Abiomed, Boehringer Ingelheim, Daiichi Sankyo, AstraZeneca, CVRx, Orion, BMS, Pfizer, MSD, Abbott, Medtronic; not related to this editorial. C.V. received honoraria for lectures/consulting from Bayer, Boehringer Ingelheim, Daiichi Sankyo, AstraZeneca, CVRx, BMS, Pfizer, Abbott, Medtronic, Zoll, Boston Scientific, Biotronik; not related to this editorial.