Heart failure with preserved ejection fraction: a stepchild no more!

In his monumental description of the circulation published in 1628, William Harvey wrote: ‘The blood in the animal body moves around in a circle continuously and the function of the heart is to accomplish this by pumping …The intrinsic motion of the heart is not the diastole but the systole; neither is it in the diastole that the heart grows firm and tense, but in the systole, for then only is it moved and made vigorous’.1 For more than the three following centuries, the focus of physiologists and clinicians alike was on systolic function and dysfunction. The three great cardiovascular physiologists of the 19th and 20th centuries—Otto Frank (Munich), Ernest Starling (London), and Carl Wiggers (Cleveland)—paid little attention to diastole and treated it as a stepchild. It was assumed that heart failure was caused by impaired systolic function. I embraced that mindset in my efforts in the early 1960s to translate the physiologists’ experimental findings to patients, by describing two clinical techniques for assessing systolic function—the left ventricle’s systolic dp/dt and its ejection fraction. However, physiologists did pay sufficient attention to their stepchild to divide diastole into three sequential components: (i) ventricular relaxation, (ii) passive filling of the relaxed ventricle, and (iii) atrial contraction. In the latter half of the 20th century, it became clear that abnormalities in any of these could interfere with ventricular filling, usually referred to as diastolic dysfunction. Slowed ventricular relaxation as reflected in a prolonged isovolumic pressure fall interferes with ventricular filling, especially at rapid heart rates and/or in myocardial ischaemia. Dysfunction of passive filling reduces ventricular distensibility, which raises ventricular pressure at any given volume.2 Clinical investigators also entered the picture, slowly at first. In 1957, Robin and Burwell3 published a remarkably prescient paper entitled Hemodynamic aspects of diffuse myocardial fibrosis, writing ‘The cardiac failure of myocardial fibrosis is a form of low-output failure … There is impressive evidence of limited diastolic filling … This fibrous infiltration is evident by gross and microscopic examination’. In 1961, Goodwin et al.4 discussed three types of cardiomyopathy leading to heart failure; the least frequent simulated constrictive pericarditis and was subsequently termed restrictive cardiomyopathy. In 1968, my colleagues and I described Impaired rate of left ventricular filling in idiopathic hypertrophic subaortic stenosis (now referred to as hypertrophic cardiomyopathy).5 We proposed that the diastolic dysfunction in this condition resulted primarily from the marked ventricular hypertrophy, as well as by increased stiffness of the individual myocytes. While diastolic dysfunction was increasingly recognized in the 1970s and 1980s, in 1993, Brutsaert et al.2 defined diastolic heart failure as a ‘condition resulting from an increased resistance to filling one or both ventricles, leading to symptoms of congestion due to an inappropriate upward shift of the diastolic pressure-volume relation’. The term diastolic heart failure was then replaced by heart failure with preserved ejection fraction (HFpEF). A minority of patients with HFpEF have a specific aetiology, such as an infiltrative disease like amyloidosis, or a form of endocardial fibroelastosis. However, most patients with HFpEF present with one or more antecedent comorbid conditions—including advanced age, hypertension, diabetes, obesity, metabolic syndrome, and chronic kidney disease. Paulus and Zile support the view that HFpEF is frequently characterized by systemic inflammation and myocardial fibrosis involving the extracellular matrix, resulting in abnormalities in myocardial structure and function.6 There has been considerable debate about whether this latter very large subgroup of patients with HFpEF should be lumped together into a single entity or split into several distinct disorders. Unsupervised machine learning of patients with HFpEF has identified a number of mutually exclusive phenogroups, based on the presence of individual or groups of antecedent comorbid conditions, and levels of natriuretic peptides, leading to varying outcomes.6,7 This is not simply an academic exercise. If the ‘lumpers’ are correct, a single-management plan may prove to be effective for the majority of patients. If the ‘splitters’ are correct (and I lean towards this group), HFpEF would be regarded as a heterogeneous syndrome, with optimal treatment dictated by the patient’s specific phenogroup. An analogy might be pneumonia, a syndrome that can have a variety of aetiologies; an effort is usually made to define the cause of the pulmonary process before selecting therapy. Through the 1960s, 1970s, and early 1980s, HFpEF was not recognized frequently. In 1984, Dougherty et al.8 reported the first large series; 36% of a group of 188 patients with heart failure had an ejection fraction ≥45%. Both the incidence and prevalence of HFpEF have expanded rapidly since then and now constitute at least one half of all patients with heart failure. The diagnosis of HFpEF in the large clinical trials was purposely broad and usually included age >50 years, a left ventricular ejection fraction exceeding 40% as well as recent hospitalization for heart failure. The diagnosis is aided by elevated circulating natriuretic peptides and echocardiographic findings of reduced ventricular distensibility. An integer score based on a combination of these measurements described by Pieske et al.9 is useful in this regard. For many decades, the management of HFpEF was non-specific, including diuretics and control of comorbid conditions, such as hypertension and obesity, but the results were discouraging and the prognosis was guarded. However, relatively recently the situation has improved and several drugs show promise. Although TOPCAT, an international trial of the mineralocorticoid receptor antagonist spironolactone, was neutral, a large subgroup of patients enrolled in the Americas showed a significant benefit.10 Another large trial (PARAGON-HF), which compared the angiotensin receptor-neprilysin inhibitor sacubitril/valsartan with valsartan alone, showed a reduction of the primary endpoint which just missed nominal statistical significance (P = 0.06). However, several important subgroups in this trial, including women as well as patients in the lower part of the range of ejection fraction, showed benefit.11 The sodium-glucose cotransporter 2 inhibitors were shown to be effective in the treatment of heart failure with reduced ejection fraction. In 2018, Pabel et al.12 reported that empagliflozin improved diastolic function in anaesthetized rats, and in isolated human ventricular trabeculae. This was followed by the observation that dapagliflozin improved ventricular diastolic dysfunction in patients with diabetes.13 In the SOLOIST trial conducted in diabetic patients with heart failure, the SGLT1 and 2 inhibitor sotagliflozin was associated with a significant reduction in cardiovascular death, heart failure hospitalization, and urgent heart failure visits in a subgroup of patients with HFpEF.14 On 27 August 2021 at the European Society of Cardiology meeting, Anker et al. presented the EMPEROR-Preserved trial, in which empagliflozin was compared to placebo in 5988 patients with HFpEF. The primary endpoint, a composite of cardiovascular death and hospitalization for heart failure was reduced significantly by 21%.15 It would appear that finally the ‘dam has been broken’ and that HFpEF is no longer a stepchild! The results of a second phase 3 trial, the DELIVER trial in HFpEF and mildly reduced ejection fraction, testing dapagliflozin will become available in the next few months and are anxiously awaited.16 Hopefully, we will be able to delete the question mark shown above. Conflict of interest: Research grant support through Brigham and Women’s Hospital from: AstraZeneca, Daiichi-Sankyo, Merck, and Novartis; consulting for: Amgen, Boehringer-Ingelheim/Lilly, Cardurion, MyoKardia, NovoNordisk, and Verve.