Aortic Valve Disease and Heart Failure with Mildly Reduced Ejection Fraction: New Insights and Pending Questions

This article refers to ‘Prevalence and prognosis of aortic valve diseases in patients hospitalized with heart failure with mildly reduced ejection fraction’ by T. Schupp et al., published in this issue on pages xxx. Although left ventricular (LV) remodelling differs between the different types of valvular diseases, consequences on LV ejection fraction (LVEF) represent a major prognostic factor, and therefore a potent driver for intervention, particularly in asymptomatic patients.1, 2 Impairment of LVEF is also a strong determinant of the occurrence of symptoms of heart failure in valvular disease. In patients with severe aortic stenosis (AS), it is well known for decades that the association of heart failure with LVEF <50% is associated with a particularly poor outcome with a life expectancy (50% mortality) as short as 1 year.3, 4 The strongest evidence concerning the prognostic impact of LV dysfunction was shown for LVEF <40%. In one of the largest series including 1402 patients who underwent surgical aortic valve replacement (SAVR), the 416 patients with LVEF ≤40% presented also with more frequent comorbidities, particularly ischaemic heart disease, than those with LVEF >40%.5 This raises the question of the respective contributions of LV remodeling in response to pressure overload due to AS and other determinants of LV systolic dysfunction. In patients with severe chronic aortic regurgitation (AR), the response to volume overload leads to LV enlargement resulting at a later stage in an impairment of LVEF. The relationship between decreasing LVEF and gradual overmortality has also been clearly demonstrated.6 In recent years, multimodality imaging has led to better insights in LV remodelling in patients with valvular disease. Echocardiographic analysis of LV strain is more sensitive than LVEF to detect subtle impairment of LV function which has a prognostic value.7 Tissue characterization using cardiac magnetic resonance enable in-depth analysis of LV remodelling. In patients with AS, the type and amount of fibrosis have been identified as prognostic factors of incremental value, particularly regarding irreversible replacement fibrosis, which differs from diffuse interstitial fibrosis which is reversible.8 The prognostic value of myocardial fibrosis has also been shown in chronic AR. Multimodality imaging is also useful to identify other causes of LV dysfunction, such as fibrotic sequalae of ischaemic heart disease or transthyretin amyloidosis.7 In this issue of the Journal, Schupp et al.9 report a retrospective single-centre analysis of 2106 patients hospitalized for heart failure with mildly reduced ejection fraction (HFmrEF, defined as a LVEF 41–49%). The originality of this paper is to focus on the recently characterized entity of HFmrEF in a large population, while older studies on LV dysfunction in patients with valvular disease were based on smaller populations and concerned patients with a wider range of impaired LVEF. Another strength is the analysis of the whole spectrum of aortic valve disease, from mild to severe. In the study by Schupp et al., patients were stratified according to the presence and severity of aortic valve disease. Older age and more frequent comorbidities in patients who had aortic valve disease in this cohort reflect the contemporary epidemiology of valvular heart disease. AS was moderate or severe in 9.8% of cases (moderate in 6.1% and severe in 3.7%) and AR was moderate or severe in 3.9% of cases (moderate in 3.7% and severe in 0.2%). Increasing severity of AS was associated with older age, more frequent comorbidities, and more severe symptoms. During a median follow-up of 30 months, there was a gradual increase in mortality and heart failure hospitalization with increasing severity of AS and AR. In a multivariable analysis adjusting for comorbidities, mild, moderate, and severe AS were all associated with higher all-cause mortality. Conversely, moderate or severe AR was not associated with increased mortality. The findings of Schupp et al. represent an important contribution to the risk stratification of HFmrEF according to aortic valvular heart disease. They also raise questions and point out the need for improving our knowledge of interactions between valvular disease and LV function to optimize patient management. The strong association between AS of any severity and all-cause mortality should not lead to overinterpretation in terms of the causal role of AS. There is a consensus on the need for intervention in patients with heart failure when AS is severe. In patients with moderate AS, recent data suggest a possible benefit in selected patients with heart failure.10 Conversely, mild AS does not have significant haemodynamic consequences. Mild AS therefore appears to be more a marker of risk rather than a cause of overmortality per se. This finding is consistent with the relationship between aortic sclerosis and increased cardiovascular mortality.11 This association makes sense since there are common pathways in the pathophysiology of AS and atherosclerosis.12 However, increased risk associated with mild AS was more pronounced in the study by Schupp et al. for all-cause mortality than for hospitalization for heart failure, thereby suggesting that the relationship between mild AS and poor outcome is mediated not only by cardiovascular diseases. Although the relationship between mild AS and overmortality remained significant in multivariable analysis in the paper by Schupp et al., the influence of unmeasured non-cardiac confounding factors cannot be excluded. At the other end of the spectrum of AS, an unexpected finding of the study by Schupp et al. is the low rate of interventions in patients with HFmrEF and severe AS. Transcatheter aortic valve implantation (TAVI) was performed in 27.5% of patients with severe AS and SAVR in 8.8%. Almost two thirds of patients were not treated for severe AS although their presentation corresponds to a class I indication in European as well as in American guidelines.1, 2 Although the appropriateness of the decision to intervene cannot be assessed for each patient in the present study, these findings raise concern regarding a possible underuse of interventions on the aortic valve. Comorbidities may be a valid reason to deny surgery but were less frequently mentioned than patient refusal. This is surprising since patient reluctance to interventions now occurs less frequently with the wide availability of TAVI. It should be stressed that failure to intervene in severe AS has been described as a major risk factor for short-term mortality.13 The findings of the study by Schupp et al. support close follow-up, aggressive management of cardiovascular risk factors and guideline-directed medical therapy of HFmrEF in patients with mild or moderate AS, but also increased awareness toward indications of TAVI or SAVR in those with severe AS. At the present time, intervention on the aortic valve should not be proposed in patients with moderate AS in the absence of randomized trials, such as the ongoing TAVR UNLOAD (Transcatheter Aortic Valve Replacement to Unload the Left Ventricle in Patients With Advanced Heart Failure, NCT02661451). Conversely to AS, the relationship between AR and increased mortality was no longer significant in multivariable analysis. This should be interpreted in the light of the very low percentage of severe AR (0.2%) and it is of importance to recall that LVEF <50% is a class I indication for intervention in patients with severe AR.1, 2 Despite higher operative mortality, SAVR is associated with improved survival in patients with severe AR and LV dysfunction as compared with medical therapy, so that those patients should not be denied surgery.14 On the other hand, patients with mild or moderate AR deserve medical therapy (Figure 1). In conclusion, the study by Schupp et al. provides valuable insights into the risk stratification of patients with HFmrEF in the context of aortic valve disease. Their findings have the merit to draw attention on the particularly poor outcome of patients with AS and heart failure, even when LVEF is moderately impaired, and on the need for reinforcing guideline implementation in patients with severe AS. Further research is needed in patients with mild and moderate aortic valve disease to elucidate the underlying mechanisms of the association between valvular disease with few of no haemodynamic consequences and impaired outcome in patients with HFmrEF. Any overinterpretation of this association is necessary to avoid unjustified indications of intervention on the aortic valve. Conflict of interest: none declared.