LVEF continues to be a practical and essential parameter for the classification, prognosis, and management of heart failure in daily clinical practice despite its known limitations.
‘Il meglio è nemico del bene’ (‘The better is enemy of the good’) Proverbi Italiani e Latini (Orlando Pescetti, 1603), popularized as ‘Le mieux est l'ennemi du bien’ by Voltaire (Dictionnaire philosophique, 1700; La Béguele, 1772) For decades, heart failure (HF) has been classified based on left ventricular ejection fraction (LVEF), which is calculated as stroke volume (i.e. end-diastolic volume minus end-systolic volume) divided by end-diastolic volume. Since these parameters are rather easy to understand and to obtain using various cardiac imaging techniques, LVEF is generally known and accepted by all physicians attending HF patients. However, LVEF has many acknowledged limitations, including technical issues in obtaining the images, data variability and poor reproducibility, variations among different imaging techniques, dependency on heart rate such that atrial fibrillation and rapid ventricular response may lead to a falsely low LVEF, and reduced accuracy in patients with left bundle branch block or ventricular pacing due to pacemaker implantation. Additionally, LVEF only provides a partial picture of patients with HF. Indeed, LVEF often shows inadequate correlation with functional capacity, and it does not provide accurate information regarding diastolic function. Moreover, even when only addressing left ventricular systolic dysfunction, LVEF may be less accurate than other echocardiographic measurements, such as longitudinal and global strain or speckle tracking. However, despite these issues, from the clinicians' perspective (beyond aetiology, geometry, volumes, diastolic issues, etc.), LVEF remains the most convenient parameter and is a cornerstone in four areas of daily clinical work: (i) HF classification, (ii) HF prognosis assessment, (iii) HF treatment and management, and (iv) HF monitoring. Below, we will discuss each of these aspects. Classically, two types of HF have been acknowledged relative to LVEF (or left ventricular performance): HF with reduced ejection fraction (HFrEF) and HF with preserved ejection fraction (HFpEF). While normal LVEF is generally considered to be >50%,1 both conditions have been arbitrary defined with regard to the cut-off points used to delineate HFrEF and HFpEF. Most randomized controlled trials aiming to establish optimal medical treatment define reduced LVEF as ≤35–40%,2 while most studies investigating HFpEF enrol patients having HF with an LVEF of >40–45% and overall HFpEF is variably defined by an LVEF of >40%, >45%, >50%, or ≥55%.2 In the 2013 American College of Cardiology/American Heart Association guidelines,2 HFrEF is defined as a clinical diagnosis of HF with an LVEF of ≤40%, HFpEF is defined by an LVEF of ≥50%, while an LVEF of 41–49% is considered borderline. The 2016 European Society of Cardiology guidelines1 present a new HF classification including a third category—HF with mid-range ejection fraction (HFmrEF), defined as an LVEF of 40–49%—which replaces the previously designated ‘grey zone’ or borderline category. This third group was created to boost investigation of these patients, who were poorly represented in clinical trials of both HFrEF and HFpEF. The emergence of HFmrEF led to a plethora of reports providing insight into the clinical characteristics and outcomes among these patients, with controversial results. When investigating patients with HFmrEF, some studies describe clinical characteristics similar to those of patients with HFrEF,3 while other studies report clinical characteristics and prognosis similar to HFpEF.4 The SwedeHF registry is a real-world contemporary and well-characterized nationwide registry-based cohort of patients with HF.5 In a recent report, the authors analysed data from 42 061 patients, of whom 56% had HFrEF, 21% HFmrEF, and 23% HFpEF. They report the clinical characteristics of patients with HFmrEF, finding that HFmrEF resembles HFrEF in terms of six characteristics, resembles HFpEF with regard to seven characteristics, and shows no pattern of resembling either condition for 10 characteristics. Collectively, the investigators propose that HFmrEF is an intermediate phenotype between HFrEF and HFpEF, but with the important exception of coronary artery disease, which is distinctly more common in HFrEF and HFmrEF compared to HFpEF. Based on the available data, some authors suggest that HFmrEF may, in many cases, be a transitional situation between HFrEF and HFpEF6, 7 rather than a truly distinct pathophysiologic entity.6 Further analysis of the dynamic nature of HFmrEF may reveal different patient subsets, including patients with HFrEF who experience LVEF recovery,8 those with well-treated coronary artery disease and showing limited LVEF loss (mainly ST-elevation myocardial infarction patients included in stent-for-life initiatives), patients with HFpEF who show a progressively declining LVEF (probably less frequent), or even patients who have ‘healed’ after myocarditis or stress-induced cardiomyopathy/takotsubo syndrome. At present, patients with HFmrEF receive inconsistent treatment, with some clinicians using HFrEF therapy and others awaiting more evidence and guideline recommendations. From the clinicians' perspective, it may not be useful to use three categories. It may eventually be deemed more practical to define HFpEF as including an LVEF of ≥50% and HFrEF as having an LVEF of <50%, and accepting that some patients might be in a transitional stage. However, at this time, there remains a need for further research regarding the dynamic behaviour of LVEF to better understand whether such patients should be given the same therapy as HFrEF patients or a more tailored therapy, which would have to be prospectively tested in appropriate trials. Regarding the use of LVEF in HF prognosis assessment, LVEF taken out of context is not accurate enough for patient risk stratification. This is even more apparent when considering the full spectrum of diseases that include HFpEF. Outcomes are importantly impacted by other factors, such as HF aetiology, myocardial fibrosis, cardiac volumes, biomarkers, and co-morbidities. However, when focusing on the subgroup of patients with HFrEF, LVEF is an important prognostic marker, with lower LVEF associated with worse prognosis. Thus, from a clinician's perspective, LVEF remains an acceptable predictor of outcomes in HFrEF, especially when repetitive LVEF measurements are analysed. At this time, LVEF is a cornerstone of the treatment and management of HF patients. Both pharmacologic and non-pharmacologic treatments are selected based on symptoms and LVEF, and sometimes natriuretic peptides, for example, when deciding upon the introduction of mineralocorticoid receptor antagonist and angiotensin receptor-neprilysin inhibitor in HFrEF.1 Recent clinical trials, particularly the pivotal PARADIGM-HF trial,9 have established a triad of criteria for patient enrolment: symptoms, LVEF, and biomarkers. Of these, symptoms and LVEF being irreplaceable, while the most appropriate biomarker is yet to be fully disclosed. Other important aspects of the disease must also be noted, including aetiology, risk of sudden death, family history, presence of fibrosis, left bundle branch block on the elctrocardiogram, etc. With regard to HF monitoring, the dynamics of LVEF must be considered. The past three decades have seen developments in drug therapy (mainly β-blocker therapy, but also angiotensin-converting enzyme inhibitor/angiotensin receptor blocker therapy to some extent), devices (mainly cardiac resynchronization therapy), coronary revascularization (both percutaneous and surgical), and valvular repair (also surgical and percutaneous). These advanced methods can restore LVEF to near-normal or even normal values in a substantial number of patients. We recently reported that up to 25% of patients with HFrEF experience LVEF recovery during follow-up8 and that such recovery predicts a better outcome, including all-cause death, cardiovascular death, HF-related death, and sudden death, as well as HF-related hospitalizations. Others also report a decline in all-cause mortality among HF-recovered patients.10, 11 As in our series, Kalogeropoulos et al.11 report that HF-recovered patients showed lower mortality, less frequent hospitalizations, and fewer composite endpoints in a retrospective cohort. LVEF recovery is more frequent in cases of HF with non-ischaemic aetiology, in the absence of left bundle branch block, in patients with a short-lasting HF diagnosis, and in patients with lower ST2, and can be reasonably well predicted using recently reported clinical tools.12, 13 Clinicians presently tend to maintain medical treatment with neurohormonal blockade in HF-recovered patients; however, further studies are needed to address whether treatment discontinuation is a valid option. Mann et al.14 elegantly discern between ‘myocardial remission’ observed along varying degrees of reverse remodelling (with molecular, cellular, and anatomic changes) and true ‘myocardial recovery’ seen in hearts with reversible damage. Precision medicine6 may offer a promising means of identifying HF-recovered patients with true ‘myocardial recovery’ vs. ‘myocardial remission’. In summary, despite the overt and acknowledged technical and clinical limitations of LVEF, this parameter remains a cornerstone for classification, stratification, management, and surveillance during follow-up in cases of HF. LVEF is easy to obtain, non-invasive, well-known and understood by a majority of internists and general practitioners, and pivotal for HF management by HF specialists and general cardiologists. Thus, we anticipate that it will remain alive and kicking in clinical practice for many more years to come. Nevertheless, novel biomarkers15 and molecular cardiology techniques are being developed to attain a more refined understanding of the multi-faceted phenotypes present in the syndrome of HF. Conflict of interest: none declared.
Lupón et al. (Mon,) studied this question.