Piecing together the puzzle of sex‐specific differences in left ventricular ejection fraction

This article refers to ‘Ejection fraction and mortality: a nationwide register-based cohort study of 499 153 women and men’ by S. Stewart et al., published in this issue on pages 406–416. Assessment of the left ventricular ejection fraction (LVEF) is one of the cornerstones of the echocardiographic evaluation and is the principal echocardiographic measure used to guide treatment and clinical decision-making in hear failure (HF) patients.1 While reduced LVEF was once considered the hallmark of HF, more recently, the landscape of HF has reformed to also include HF with mid-range reduced LVEF and HF with preserved LVEF (HFpEF). Yet regardless of LVEF, HF patients have a dim prognosis.2 Patients within these different HF groups differ with respect to clinical characteristics, with one notable feature being the distribution of men and women.2, 3 Studies have not only reported a higher proportion of women in HFpEF, but the PARAGON-HF trial also suggested differential treatment effects with respect to sex, with women deriving greater benefit than men.4 Furthermore, gender-specific cardiac remodelling in the development of HFpEF is well established.5 The underlying reason for these gender-specific findings are unclear, but a better understanding of sex-specific characteristics could help guide management through a more detailed risk stratification. In this issue of the journal, Stewart et al.6 investigated the association between sex-specific characteristics of systolic function with outcome. Based on the National Echocardiography Database of Australia (NEDA) they investigated approximately half a million subjects referred for echocardiography, with the specific purpose of exploring the sex-specific association of different LVEF strata with all-cause mortality and cardiovascular-related death. They present several key findings. First, the authors showed that a stepwise decrease in LVEF below 60% was associated with an increased risk of both all-cause and cardiovascular-related death in both men and women. This is consistent with current knowledge from HF studies, including the CHARM programme.7 Secondly, the authors observed that the risk of cardiovascular-related mortality started to increase at high LVEF levels (60–65%) in women but not in men. Although the gender-specific finding is novel, it has previously been observed in the MESA study that a supranormal LVEF (>70%) carries an increased risk of HF.8 Even though it might be counterintuitive that a high LVEF can be an indicator of poor outcome, a supranormal LVEF can be observed secondary to pathological geometric changes as seen with amyloidosis or hypertrophic cardiomyopathy. A supranormal LVEF can also develop as a compensatory mechanism as seen with severe mitral regurgitation. Although the authors sought to account for hypertrophy and valve disease, the lacking granularity with respect to clinical information in the NEDA study still leaves some residual confounding which could cloud the findings. Moreover, these data were derived from patients referred for echocardiography due to suspicion of cardiovascular disease or with established cardiovascular disease, resulting in potential referral bias. Individuals in the NEDA study could have been referred for echocardiography for various reasons, but most likely they either had symptoms or risk factors that would likely increase their risk of outcomes. Hence, these individuals are neither comparable to a general population nor to population-based studies. One possible explanation for the findings is that those referred for echocardiography because of symptoms with normal or supranormal LVEF may be at increased risk for non-cardiac disease; indeed, the authors did observe that the primary cause of death was non-cardiac in those with normal or supranormal LVEF. The NEDA study found that the median value of LVEF was higher for women than men (65% vs. 61.1%), and that the adjusted risk of cardiovascular-related death increased incrementally when LVEF was below 65% and 60% for women and men, respectively. Notwithstanding the epidemiological limitations, these findings suggest that the normal range for LVEF to be higher for women than men. Indeed gender-specific differences in normal ranges have been noted in guidelines. The differences in lower limits of normality have, however, been minor (52% for men and 54% for women).9 As the risk of outcome started to increase at LVEF levels considered within the normal range, the NEDA study makes an interesting observation, based on the to date largest observational dataset, that may prompt us to reconsider how we should interpret normality and consequently when to consider patients at risk of outcomes. In this regard, it should also be noted that LVEF was quantified by various methods in the NEDA study, including the Teichholz method, which often overestimates LVEF. However, the authors addressed this issue by showing consistent findings regardless of quantification method. Gender-specific differences in cardiac structure and remodelling are well established, with women having a smaller left ventricular cavity and often respond with concentric remodelling in the presence of cardiovascular risk factors.10 Since LVEF increases disproportionally with a smaller left ventricular cavity, this type of remodelling can lead to seemingly supranormal LVEF,11 while the presence of cardiac remodelling in itself indicates adaptations to cardiovascular risk factors and a worse prognosis.12 With the introduction of myocardial speckle tracking to quantify myocardial deformation, it has been recognized that even patients with preserved LVEF can have impaired systolic function by global longitudinal strain and that this confers a higher risk of adverse outcomes.13 At a preclinical stage, meaning in the presence of cardiovascular risk factors, myocardial deformation also differs between gender. Specifically, findings from community-based cohort studies suggest that global longitudinal strain is significantly reduced in men as compared to women and only carries prognostic information in men.10, 14 With age, women on the other hand have a greater extent of increased left ventricular torsion which seems to preserve LVEF.6 What causes this difference in myocardial deformation in the presence of cardiac risk factors is unclear, but may rely on gender differences in cellular response with aging as well as differences in geometric adaptions, including the smaller left ventricular cavity size in women.10, 15, 16 Thus, in the setting of cardiovascular risk factors, torsional mechanics may be the first sign of progressive left ventricular dysfunction in women, which could translate into an increased risk of adverse outcome while LVEF is either preserved or supranormal. This could also in part explain some of the findings made by Stewart et al. Detailed three-dimensional studies relating torsional mechanics to outcome would be needed to substantiate this hypothesis. As the above-mentioned sex-specific left ventricular mechanical adaptions become more notable with age, this does however not explain the stronger sex-specific associations observed in the younger women. The lacking details on clinical characteristics only allow us to speculate as to the mechanisms (Figure 1). The data do, however, reveal a previously unrecognized sex-specific association between LVEF and outcomes, the reason for which needs to be explored in population-based studies not subjected to the epidemiological limitations of the NEDA registry. Recently though, a large scale study from the UK revealed a slower rate of improvement in survival for women than men over two decades.17 This likely reflects that men more often present with HF with reduced LVEF and women with HFpEF, the latter of which still lack treatment options. This underlines the need for a deeper investigation of sex-specific differences in HF, and a need for better risk stratification in women to target patients who may benefit from future treatment. The findings from the NEDA study could prove valuable in this regard. F.J.O. was financed by the Danish Heart Foundation (grant no.: 18-R125-A8534-22083), King Christian the 10th Foundation, Herlev & Gentofte Hospital, and Mrs. Asta Florida Boldings Memorial Grant during preparation of this manuscript. T.B.S. was financed by the Novo Nordisk Foundation, the Lundbeck foundation, and Herlev & Gentofte Hospital during the preparation of this manuscript. Conflict of interest: T.B.S.: Steering Committee member of the Amgen financed GALACTIC-HF trial; Advisory Board: Sanofi Pasteur, Amgen; speaker honorarium: Novartis, Sanofi Pasteur; research grant: GE Healthcare, Sanofi Pasteur. All other authors have nothing to disclose.