The Pilot European Survey of Atrial Fibrillation: How to Look at Heart Failure Through a Keyhole

This article refers to ‘Heart failure in patients with atrial fibrillation in Europe: a report from the EURObservational Research Programme Pilot survey on Atrial Fibrillation’,† by G.Y. Lip et al. published in this issue on page 570–582. Atrial fibrillation (AF) and heart failure (HF) are highly prevalent and costly conditions, frequently diagnosed together, and sharing common risk factors. In this issue of the journal, Lip and colleagues analyse baseline characteristics of 2972 patients with AF according to the co-existence of HF [with either reduced (HFrEF) or preserved (HFpEF) LVEF].1 Their investigation deserves careful consideration because it makes it possible to observe HF from an unconventional point of view and to challenge several of the most relevant topics encountered during contemporary clinical practice. The incidence of AF rises with age, as does the number of patients developing HF. Although AF can occur without structural heart disease, several cardiac disorders represent predisposing factors both for AF and for HF.2 We know that patients with AF at the greatest risk of developing HF can be identified simply by searching for ECG LV hypertrophy, diabetes, significant murmur, and history of myocardial infarction (besides considering advanced age and a large body mass index).3 Consistent with previous epidemiological data, Lip et al. found the above-mentioned risk factors more frequently in AF associated with HF (AF–HF).1 Sharing risk factors is not the only link between AF and HF. Both conditions interact with each other, amplifying the negative effects they exert when occurring alone. In animal models, a combination of atrial pacing and induced HF (mimicking AF–HF) causes effects on ionic current densities different from what is encountered in each condition considered alone.4 Lip et al. found that as many as 1411/2972 patients (48%) with AF had a concomitant diagnosis of HF. The somewhat higher than expected prevalence of this association is not surprising, and may be explained in light of the broad definition of HF used in the Registry. Indeed, HF was only based on investigator judgement and not on pre-specified established criteria, which are usually more restrictive.5 The authors found that patients with AF–HF were more severely symptomatic and reported a worse quality of life. By increasing heart rate at rest and during exercise, AF shortens diastolic time and decreases cardiac output.6 Irregularity of ventricular contractions and loss of atrial systole further contribute to haemodynamic deterioration.7 If it is true that AF negatively affects the physiopathology of HF, the opposite is also true. In HF, atrial stretch due to increased filling pressures results in activation of ionic currents, facilitating AF.8 Up-regulation of the renin–angiotensin–aldosterone system causes salt and water retention, elevated atrial pressures, and structural remodelling with fibrosis of the atrial and ventricular myocardium, eventually favouring maintenance and progression of AF.6 Lip and colleagues recorded that 555/1070 of the patients with HF (52%) had HFrEF. Since a second echocardiogram was not available during follow-up, no comments can be made concerning the prevalence of tachycardia-induced cardiomyopathy in the study population. The true prevalence of tachycardia-induced cardiomyopathy in HF is unknown, but it is estimated to affect up to 50% of patients with AF–HF.9 The mechanisms of tachycardia-induced cardiomyopathy are not yet fully understood, but appear to be related to myocardial ischaemia, energy depletion, and calcium regulation abnormalities affecting cardiac muscle.10 The study findings indicate that 1-year mortality of patients with AF–HF was significantly higher as compared with those with AF only (11 vs. 3%), and remarkably so both in AF–HFrEF and in AF–HFpEF. Unfortunately, in the study by Lip et al. the follow-up was completed at 1 year in 2354/2972 patients (79%) and no firm conclusions can therefore be made concerning outcome. Furthermore, the subjective estimation of EF used by the investigators (rather than a quantitative assessment of ventricular systolic function) further enhances the uncertainties regarding the outcome of HFrEF vs. HFpEF. Whether AF is a marker of more severe HF, or whether it contributes negatively and independently to prognosis remains unclear.6 These doubts contribute to the controversy on whether the main goal in AF should be rhythm vs. rate control.7 In the study by Lip et al., a pure strategy of rhythm control was more common in AF–HF, and particularly in AF–HFrEF. Although it may be hypothesized that sinus rhythm, particularly in HFrEF, is preferable to AF, the potential benefits of sinus rhythm restoration need to be counterbalanced by the adverse effects of antiarrhythmic agents. Notably, rhythm control has never been demonstrated to be superior to rate control in chronic HF, and it should probably be reserved for AF due to a reversible cause, and/or to patients who cannot tolerate AF after optimization of rate control and HF therapy.5 Things may be different in acute HF, where emergency cardioversion may be required to correct a profound haemodynamic instability.5 In the study by Lip et al., 82% of patients with HFrEF were treated with a beta-blocker, 39% with digoxin, and only 4% with non-dihydropyridine calcium channel blockers (non-DHPCCBs). If rate control is pursued, beta-blockers in AF–HFrEF represent an option commonly considered.5 A recent large meta-analysis failed to detect a beneficial effect of beta-blockers in AF–HFrEF.11 On the other hand, neither excess mortality nor higher hospital admission rates has been associated with beta-blockers in AF–HFrEF.11 Taken together, these findings should (at least partially) reasure physicians who decide to use beta-blockers in AF–HFrEF while waiting for specifically designed prospective trials (hopefully not excluding AF–HFpEF).11 Although digoxin is often insufficient as a single agent for rate control of AF, it can represent a valuable option to be added to beta-blockers, particularly in chronic HFrEF.12 The same does not apply to acute HFrEF, where digoxin exerts a modest effect because of its slow onset of action. Non-DDPCCBs were prescribed to 9% of patients with AF–HFpEF as compared with only 4% of those with AF–HFrEF. Because of their negative inotropic effect, non-DHPCCBs may not be tolerated at the doses required in HFrEF, whereas they could represent a therapeutic option (besides beta-blockers) in HFpEF.13 Implantable cardioverter defibrillators (ICDs) represent a cornerstone for the primary and secondary prevention of sudden cardiac death in HFrEF.5 In the study of Lip et al., 47% of patients with HFrEF had an LVEF <35%. Nevertheless, only 14% of them had an ICD implanted. Since as many as 80% of patients enrolled in the survey were hospital inpatients, it is possible that the physicians may have planned a follow-up echocardiography later on to confirm an LVEF value triggering prophylactic ICD implantation after an adequate period of optimal medical treatment. Similarly, it is likely that at least one group of patients enrolled were not considered eligible for prophylactic ICD due to the presence of very advanced HF. Analysing the baseline characteristics of patients enrolled in the EORP-AF Pilot survey,14 almost 42% of patients with HFrEF were in NYHA class III/IV. The role of prophylactic ICD implantation in patients in NYHA class IV has not been established.15 Similarly, the presence of significant co-morbidities usually characterizing HFrEF (particularly when considered outside of clinical trials) may have further discouraged the use of ICDs.16 Cardiac resynchronization therapy (CRT) is an effective device-based therapy able to improve symptoms and mortality in HFrEF with ventricular dyssynchrony.17 Although CRT effectiveness seems reduced in AF, it still represents a therapeutic option in these patients.18 By favourably influencing haemodynamic and cardiac remodelling, CRT can reduce the incidence of AF. In the study of Lip et al., overall pacemaker implantation (probably including CRT) was recorded in 12% of patients with HFrEF, perhaps reflecting the perception of limited efficacy of CRT in patients with HF and AF.18 A catheter ablation of AF was planned or performed in 2–3% of patients with either HFrEF or HFpEF.1 Catheter-based ablation of AF has recently been demonstrated to be effective in the HFrEF population.6 A potential benefit from ablation could derive from the ability to restore and maintain sinus rhythm without the detrimental effects of antiarrhythmic drugs. Further studies are needed to address the role of AF ablation in HF,19 particularly in patients with a controlled heart rate, where the option to correct a component of tachycardia-induced cardiomyoapthy is unlikely.5 In AF patients with high ventricular response refractory to drugs, catheter-based ablation of the atrioventricular node combined with CRT remains an option in carefully selected patients,20 but unfortunately no data were reported in this Registry on this approach. Similarly, Lyp et al. did not analyse the contribution of different types of AF to the outcome of HF patients.1 In conclusion, Lip et al. should be congratulated for providing us with many findings useful not only to improve knowledge of currently adopted treatments for HF, but also to make us reflect on which one to embrace (and in whom) during our daily clinical practice. Conflict of interest: none declared.