Heart failure with improved ejection fraction: Is a newcomer in the family important?

In the 2013 American College of Cardiology/American Heart Association guidelines, an emerging heart failure (HF) phenotype was introduced, named heart failure with preserved ejection fraction (HFpEF) improved. It is defined as improvement in left ventricular ejection fraction (LVEF) to > 40% in patients with previous heart failure with reduced ejection fraction (HFrEF).1 Later, similar entities with different names were described: heart failure with improved ejection fraction (HFIEF), heart failure with recovered ejection fraction (HFrecEF), and better HFEF. Although not consistent in cut-off LVEF values/degree of improvement, these groups of patients are clearly distinct from those with persistent HFrEF or HFpEF.2–7 In the current issue of the European Journal of Preventive Cardiology, Jorgensen and colleagues8 discuss the new HF entity, HFIEF. They should be acknowledged both for getting into the focus this important clinical topic and for using meta-analysis to get meaningful insights regarding a large number of the patients. They carried out a systematic review and meta-analysis on 24 observational studies in a group of 2663 patients with HFIEF (defined as ≥ 5% LVEF improvement) compared with 8355 patients with persistently reduced ejection fraction (HFpREF). LVEF was assessed at baseline and reassessed after 19 ± 19 months, with follow-up of 39 ± 12 months. The primary endpoints were all-cause mortality and appropriate implantable cardioverter defibrillator (ICD) discharge. Among HFIEF patients, LVEF improved by an average 16.3%. Meta-analysis covered studies with heterogeneous aetiology of HF, various therapeutic approaches and different baseline LVEF cut-off values. The results revealed a significantly lower risk of all-cause mortality in 25% of patients with HFIEF and a significantly lower risk of appropriate Implantable cardioverter defibrillator (ICD) activations, in comparison to HFpREF patients. The largest therapeutic benefit was demonstrated in patients who had cardiac resynchronization therapy-defibrillator (CRT-D) implantation. The variables associated with higher probability of HFIEF included: female sex, higher systolic blood pressure, absence of diabetes, non-ischaemic origin for HF and lower left ventricular end-diastolic diameter LVEDD). Every meta-analysis has inherent methodological drawbacks, which may obscure the validity of the results and this study is not the exception to this rule. The meta-analysis of Jorgensen and colleagues was based on observational studies with some limitations such as: selection bias, different and unreported baseline variables, length of follow-up period and various definitions of LVEF improvement. The studies used in this meta-analysis came from a period with a long time span and, therefore, optimal HF medical therapy used was different or unreported. A major limitation was also a lead time bias, since HFIEF patients had to survive long enough for a second LVEF assessment. Furthermore, events occurring between the first and second LVEF measurements were not analysed, since patients later to be classified as HFpREF were more likely to die before reaching the second LVEF measurement. To make matters more complex, the precise time frame needed for ventricular remodelling to occur (spontaneously or under treatment) is unknown. The results obtained represent the patients who had both baseline and final data analysed (‘completers’), and it will be impossible to predict the effect of data from patients without full assessment (‘non-completers’) on the overall results. Furthermore, the methods for measurement of LVEF evolved significantly, from less precise to sophisticated methods, which may have altered the results. Finally, differentiation between long-term functional or structural improvement of LVEF, would need a ‘withdrawal’ study (LVEF reassessed after withdrawal of drug or device), which is obviously ethically challenging. Disease-modifying medical therapy, interventional procedures, revascularization and devices are not only improving symptoms and prolonging life, but are also enhancing LVEF.9–14 However it is essential to clarify in which cases the rise of LVEF is spontaneous and unrelated to therapy, and when it is the result of the treatment. Understanding this phenomenon will allow the cardiologist to tailor the treatment individually, or even to safely discontinue the medication. A spontaneous LVEF improvement, even without medical therapy, was described in several instances.15 One of the most prominent clinical scenarios is symptomatic (or asymptomatic) inflammatory cardiomyopathy, where left ventricular (LV) haemodynamic recovery may take place in up to six months after the acute phase. Tachycardia-induced cardiomyopathy, benign forms of peripartum cardiomyopathy16 or toxin-induced LV dysfunction are also illustrative examples of possible LV improvement. The incidence of LVEF enhancement, its relations to the symptoms, natural history and outcomes are unknown and need to be adressed.17–20 The proportion of patients that improve ejection fraction under medication varies in different publications. In the Valsartan Heart Failure Trial (Val-HeFT), baseline ejection fraction recovered in 9% of patients and up to > 40%, at 12 months follow-up.6 Similarly, in the trials of Basuray et al. and Wilcox et al., close to 10% of patients with a baseline ejection fraction of < 50% recovered their ejection fraction to ≥ 50%.7,21 In addition, in the registry of Punnoose et al. approximately 30% of patients with normal LVEF, previously had reduced LVEF.22 Greater improvement was obtained in a single-centre cohort, in whom 42.9% of patients demonstrated HFrecEF during the three-year follow-up (Kalogeropoulos et al.,), while in the CARDIOCHUS-CHOP registry 52% had improvement of LVEF > 40% at one-year follow-up.23 Even after the improvement of LVEF > 50%, this group of patients remains symptomatic and more vulnerable than individuals without HF. Despite improved LVEF, permanent neurohormonal activation, oxidative stress and myocardial injury can be detected with ongoing troponin release, elevation in B-type natriuretic peptide, ST2 and soluble Fms like tyrosine kinase-1 (sFlt-1), leading to higher cardiovascular hospitalization rate.6,7,21 In patients with HFrecEF, a better three-year lower mortality and fewer all-cause cardiovascular and HF-related hospitalizations, compared to HFpEF or/and HFrEF, was revealed.3 Although there is a limited knowledge of the determinants of recovery, Jorgensen and colleagues discussed these variables extensively. In the Val-HeFT, female sex, non-ischaemic aetiology, lower body mass index, higher diastolic blood pressure, LVEDD index, beta-blocker and valsartan therapy were identified as correlates of LVEF recovery.6 This is consistent with other HFIEF populations, who presented with younger age,7,22 lower prevalence of coronary artery disease,22 better renal function,7,22 and better biomarker and neurohormonal profiles.7 In addition, Kalogeropoulos et al., showed that HFrecEF was associated with lower prevalence of hypertension, diabetes, chronic lung disease and atrial fibrillation, compared with patients with HFpEF.3 In the CARDIOCHUS-CHOP registry, variables linked with recovery were lower New York Heart Association (NYHA) class, treatment with Angiotensin receptor converting enzyme inhibitors (ACEIs) and beta-blockers, non-ischaemic aetiology and lack of ICD implantation.23 Additional data regarding LV recovery can be derived from the LV myocardial response on left ventricular assist device (LVAD) implantation. Bruckner et al. revealed that higher degree of LV fibrosis at the time of LVAD insertion was an unfavourable factor for LV recovery during LVAD unloading.24 The above-mentioned results are consequences of the pathophysiological distinction in myocardial remodelling. Myocardial recovery is a part of biological continuum of HF progression, and it occurs when the positive effect of therapy or withdrawal of the negative stimulus is taking place. It can only happen when the remodelling is not advanced and structural changes are still reversible, during the natural history of HF. Patients with an ischaemic pathogenesis, more dilated LV and a detectable hs-troponin T probably have more extensive LV damage and less capacity for improvement. On the contrary, patients with post-inflammatory, non-ischaemic aetiology may have a higher capacity for myocardial reverse remodelling. Females do have a higher percentage of non-ischaemic heart failure and subsequently have a better chance for LV improvement.1,25 Lower LVEF recovery after ICD implantation may be associated with more myocardial damage and lower potential for recovery.23 Patients with higher blood pressure at baseline are more prone to be randomised to intensive disease-modifying therapy, therefore expected to have the better effect on LV function. Due to the complexity of biological evolution of myocardial remodelling, relationship of its cause and consequences needs to be elucidated. Most of the patients continue medical and device treatment besides the recovery of LVEF, and it is unknown for how long it should carried on. Clinical deterioration and recurrence of cardiac dysfunction after discontinuation of HF treatment is described.26,27 The sustainability of LVEF recovery is essential to indicate the need for a mechanical device (ICD, VADs), ICD deactivation or CRT/ICD battery replacement. In the Multicenter Automatic Defibrillator Implantation Trial With Cardiac Resynchronization Therapy (MADIT-CRT) trial,28 patients with improved LVEF to 36–50% had 60% lower risk of appropriate ICD activation, whereas patients with recovered ejection fraction to > 50% had 94% lower risk. Jorgensen and colleagues suggested 42% lower risk of ICD activation in HFIEF patients, but high residual risk would require randomised trials. HFIEF is here to stay, which was confirmed by the meta-analysis of Jorgensen and colleagues. It is a clinical reality with possible therapeutic and prognostic considerations. Moving into the future, it is most likely that the importance of HFIEF in clinical practice should be determined as a result of the large randomised trial of homogeneous patient groups, with sufficient follow-up and use of contemporary metabolomic and imaging modalities. The author(s) declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article. The author(s) received no financial support for the research, authorship and/or publication of this article.