Update on heart failure with preserved ejection fraction

Abstract

Heart failure with preserved ejection fraction (HFPEF) is the most common form of heart failure (HF) in older adults, and is increasing in prevalence as the population ages. Morbidity and long-term mortality in HFPEF are substantial and can be similar to HF with reduced ejection fraction (HFREF), yet HFPEF therapy remains empirical and treatment guidelines are based primarily on expert consensus. Neurohormonal blockade has revolutionized the management of HFREF, but trials in HFPEF based on this strategy have been disappointing to date. However, many recent studies have increased knowledge about HFPEF. The concept of HFPEF has evolved from a ‘cardio-centric’ model to a syndrome that may involve multiple cardiovascular and non-cardiovascular mechanisms. Emerging data highlight the importance of non-pharmacological management strategies and assessment of non-cardiovascular comorbidities. Animal models, epidemiological cohorts, and small human studies suggest that oxidative stress and inflammation contribute to HFPEF, potentially leading to development of new therapeutic targets.

Introduction: prevalence, outcomes, and definition of HFPEF

Current estimates suggest that over five million Americans have heart failure (HF); of these, approximately 50% have heart failure with preserved ejection fraction (HFPEF). [1] HFPEF is the predominant form of HF in older adults, and accordingly is increasing in prevalence as the overall population ages. [2] Long-term mortality in HFPEF is similar to heart failure with reduced ejection fraction (HFREF), with less than 50% five-year survival in community HFPEF cohorts. [2, 3] Outcomes following hospitalization for decompensated HFPEF are quite poor, with over 1/3 of patients dead or rehospitalized within 60–90 days of discharge. [4]

The diagnosis of HFPEF remains challenging due to the advanced age and frequent multiple concomitant illnesses. In fact, multiple comorbidities are the rule in HFPEF rather than the exception, [5] and significantly influence cardiovascular structure and function as well as long-term prognosis. [6] Many of these conditions (e.g. advanced age, obesity, atrial fibrillation, anemia) [7, 8] can mimic HF signs and symptoms, and some have questioned the concept of HFPEF as a distinct disorder. [9] Non-cardiovascular hospital readmissions and mortality are more frequent in HFPEF than in HFREF, [11] and while cardiovascular mortality in HFPEF is also lower than in HFREF, it is still substantial, accounting for 50% or more of all deaths. [10,12]

The European Society of Cardiology, the Heart Failure Society of America, and the American Heart Association/American College of Cardiology guidelines agree that HFPEF patients should have symptoms and/or signs of HF and a left ventricular ejection fraction of ≥ 50%, with exclusion of other primary causes of the symptom pattern. Previous diagnostic algorithms mandated the presence of left ventricular diastolic dysfunction, which remains an important and validating criterion. However, recognizing that many other mechanisms may also contribute, current guidelines support the diagnosis of HFPEF if the clinical picture is consistent and diastolic dysfunction is indeterminate, but other evidence of HFPEF-associated adverse cardiovascular remodeling (e.g. left ventricular hypertrophy, left atrial enlargement, atrial fibrillation) is present. [13–15] Diagnostic criteria for HFPEF will likely continue to evolve along with our understanding of the disorder.

Mechanisms of HFPEF

Hemodynamic/cardiovascular mechanisms

The classic paradigm for HFPEF implicates impaired diastolic ventricular filling due to delayed active relaxation, intrinsic ventricular stiffness, or a combination of these factors. [16, 17] Worsening left ventricular diastolic dysfunction is an important risk factor for developing HFPEF, [18] and strongly predicts mortality in unselected community cohorts and in patients with prevalent HFPEF. [19, 20] In comparison to age- and gender-matched controls, HFPEF patients had increased baseline ventricular stiffness, lower stroke volume during rapid atrial pacing, and exaggerated rise in end-diastolic pressures during handgrip exercise. [21] Borlaug et al. recently studied diastolic function in HFPEF patients undergoing cycle ergometry and confirmed an upward and leftward shift of the end-diastolic pressure-volume relationship, attributing increased filling pressures to intrinsic ventricular stiffness and reduced diastolic filling time at higher heart rates. [22]

HFPEF patients display combined ventricular and arterial stiffness, which increases in stress-induced blood pressure, cardiac metabolic demand, and the energy cost for cardiac output. [23] Aortic distensibility [24] and carotid artery distensibility [25] are severely reduced in elderly HFPEF patients. While resting values of arterial and end-systolic ventricular elastance are similar to age-matched hypertensive controls, [18] chamber-level and myocardial contractility in HFPEF are decreased. [26, 27] Moderate-intensity cycle ergometry exercise in HFPEF patients shows a disproportionate impact of proximal arterial stiffness on ventricular afterload. [28] Part of the mechanism of exercise intolerance, particularly in hypertensive HFPEF, may be an inability to augment contractile function to match arterial load. [23, 29, 28, 25]

Peripheral vasodilation may also be abnormal in HFPEF, particularly during exercise. [29] However, in HFPEF compared to age-matched controls, flow-mediated arterial dilation, measured in the femoral artery by phase-contrast MRI [30] and in the brachial artery by high-resolution ultrasound, [31] was found to be preserved.

A subset of HFPEF patients have normal ventricular stiffness and increased ventricular capacitance, particularly in the setting of comorbidities that are associated with increased plasma volume such as obesity, anemia, and renal insufficiency. [32, 33] Patients with ‘pre-clinical’ HFPEF have impaired natriuretic function, [34] and even seemingly well-compensated HFPEF patients often have demonstrably elevated plasma volume. [33] Some view volume overload and congestion as key contributors to HF development and progression. [35] Chronotropic incompetence is frequently present in HFPEF and contributes significantly to exercise intolerance, the primary manifestation of HFPEF. [36, 29] Consequently, chronotropic incompetence, which can easily be measured, [37] should be considered before agents that slow the heart rate are used. Echocardiographic strain studies indicate that left atrial stiffness and contractile dysfunction, particularly evident during preload changes induced by leg lifts, can help identify HFPEF. [38, 39]

Noncardiovascular mechanisms of HFPEF

The Cardiovascular Health Study cohort showed that frailty, as evidenced by slow gait speed and muscular weakness, strongly predicts hospital admission in older adults newly diagnosed with HF. [40] Rather than being simply a result of deconditioning, recent data suggest that frailty and muscular abnormalities may directly contribute to the HFPEF syndrome, a finding similar to HFREF, where skeletal muscle abnormalities can be independent of physical activity and deconditioning. [41, 42] Exercise training improves physical functioning and peak oxygen consumption in HFPEF patients, [43, 44] but has largely neutral effects on cardiac filling pressures, ventricular diastolic function, conduit artery endothelial function, and large-arterial stiffness. [45, 44, 46] Improvements in peak oxygen consumption following training in HFPEF patients relate more strongly to increased peak arterial-venous oxygen difference rather than increased cardiac output. [47] Analysis of lean body mass with dual-energy X-ray absorptiometry shows that sarcopenia (degenerative skeletal muscle loss) is common in HFPEF patients. Moreover, the increase in peak oxygen consumption per unit of lean body mass in HFPEF is markedly reduced compared with sedentary age-matched controls. [48] Taken together, these data suggest that impaired oxygen utilization by skeletal muscle contributes to the severe exercise intolerance in HFPEF and may represent a potential novel therapeutic target.

Cellular/metabolic mechanisms of HFPEF

With the increasing prevalence of associated factors such as advanced age, hypertension, diabetes mellitus, obesity, and chronic kidney disease in the U.S. population, the incidence of HFPEF is expected to rise in the years ahead. [2] The specific mechanisms that promote the development of HFPEF in patients with these risk factors have previously been unclear. However, hints come from two large and well-characterized U.S. cohorts of community-dwelling older adults, in which markers of systemic inflammation strongly predicted incident HF (in particular HFPEF) even following extensive adjustment for other known risk factors. [49, 50]

Several small-animal models of HFPEF have been studied, among them the Dahl S (salt-sensitive) rat, the obese spontaneously hypertensive rat, and the deoxycorticosterone/salt uninephrectomized mouse, [51–53] and the dog cellophane renal wrap model has been proposed as a large-animal example of HFPEF. [54] These and other similar experimental models share the common pathways of increased oxidative stress and perivascular inflammation. These factors are important driving mechanisms for HFPEF, as antioxidant supplementation and modulation of immune cell function in these or similar models greatly diminish vascular and cardiorenal damage and dysfunction. [55–57, 51, 58]

One recent human study examined inflammation in HFPEF using endomyocardial biopsies showed activated macrophages, staining TGF-β, increased vascular adhesion markers when compared with controls. Primary fibroblasts cultured from the HFPEF samples and stimulated with TGF-β transdifferentiated into myofibroblasts,. [59] In the Dahl S rat model of HFPEF, neutralizing antibodies to IL-16 greatly reduce cardiac macrophage infiltration and TGF-β production, myocardial fibrosis, and lung weight. An observational study linked serum levels of IL-16 to ventricular diastolic dysfunction and left atrial enlargement in HFPEF patients. [60] A cross-sectional study showed that older HFPEF patients have increased biomoarkers for inflammation. [61]

Elevated oxidative stress is present in endomyocardial biopsies from HFPEF patients showing increased DHE-positive nuclei [59] and nitrotyrosine content. [62] A new paradigm for HFPEF has recently been proposed, [63] focusing on comorbidity-induced oxidative stress as a central causative mechanism. In this construct, low nitric oxide bioavailability and reactive oxygen species lead to decreased cyclic GMP and protein kinase G activity. In strong support of this model, in comparison to patients with aortic stenosis or HFREF, patients with HFPEF have markedly reduced myocardial protein kinase G activity and cyclic GMP levels that are inversely proportional to myocardial nitrotyrosine residues. [62] The adverse effects of oxidative stress in HFPEF patients may not be confined to the heart and vasculature. Preliminary findings from ³¹phosphate magnetic resonance spectroscopy suggest that HFPEF patients have impaired skeletal muscle oxidative metabolism independent of vascular function or oxygen delivery. [64]

Treatment of HFPEF

Neurohormonal antagonists

Several studies have investigated angiotensin-converting enzyme inhibitors (ACEI) and angiotensin-receptor blocker (ARB) therapy, [65–67] a concept strongly grounded in data from animal models [68–71] as well as human hypertensives without heart failure. However, a 12-month, randomized controlled trial of the ACEI enalapril in elderly patients with established HFPEF showed no improvement in exercise capacity or quality of life. [72]

Of the three large randomized trials of ACEI/ARB performed to date in HFPEF, only the Candesartan in Heart failure: Assessment of Reduction in Mortality and morbidity – Preserved (CHARM-Preserved) study found nominal benefit for candesartan in reducing HF hospitalizations [HR 0.86 (95% CI 0.74–1.0), p=.051] over three years of follow-up, [65] and none showed benefit for their primary endpoints. Recently, Lund and colleagues explored community ACEI/ARB use in 16,216 HFPEF patients. They found a modest reduction in one-year mortality for ACEI/ARB therapy [propensity score–adjusted HR 0.90 (95% CI, 0.85–0.96), p<.001] with evidence for increased benefit at higher doses. However, mortality reduction was mainly observed in patients with LVEF 40–49%. [73]

The Phase II Prospective comparison of Angiotensin Receptor-neprilysin inhibitor with ARB on Management Of heart failUre with preserved ejectioN fracTion (PARAMOUNT) study randomized 301 HFPEF patients to valsartan vs. LCZ696, a combination ARB/neprilysin inhibitor (intended to inhibit the breakdown of natriuretic peptides). Compared to valsartan alone, the LCZ696 group had significantly lower nt-pro BNP levels and at 36 weeks, decreased left atrial size and a trend towards improved functional class. [74]

The Randomized Aldosterone Antagonism in Heart Failure With Preserved Ejection Fraction (RAAM-PEF) trial randomized 44 HFPEF patients to six months of eplerenone vs. placebo, and showed reductions in circulating markers of collagen turnover and modest improvements in diastolic function. [75] The larger Aldosterone Receptor Blockade in Diastolic Heart Failure (Aldo-DHF) randomized HFPEF patients to spironolactone or placebo for 12 months. Spironolactone reduced left ventricular mass and the mitral E/e′ ratio, although these findings were partially attenuated by adjustment for blood pressure reduction. [76] Despite these favorable signals, neither study demonstrated improvement in its primary outcome of six-minute walk distance, and Aldo-DHF participants reported no improvement in quality of life. Moreover, in a propensity-matched analysis of hospitalized older HFPEF patients from the Organized Program to Initiate Lifesaving Treatment in Hospitalized Patients with Heart Failure (OPTIMIZE-HF) study, aldosterone antagonists had no effect on all-cause mortality or hospitalization. [77] The Treatment of Preserved Cardiac Function with an Aldosterone Antagonist (TOPCAT) study has randomized HFPEF patients to spironolactone or placebo [78] TOPCAT is powered for the composite outcome of cardiovascular mortality, aborted cardiac arrest, and/or HF hospitalization, and should further define the role of aldosterone blockade in HFPEF.

Other pharmacological interventions

Pulmonary hypertension is common in HFPEF and predicts a poor prognosis. [79] Phosphodiesterase-5 inhibitors vasodilate the pulmonary vascular bed and improve functional capacity in pulmonary arterial hypertension. One study randomized 44 HFPEF patients with documented pulmonary hypertension to 12 months of sildenafil vs. placebo. Sildenafil markedly reduced pulmonary vascular resistance while significantly improving quality of life. [80] However, in a recent PhosphodiesteRasE-5 Inhibition to Improve Clinical Status And Exercise Capacity in Diastolic Heart Failure (RELAX) study, sildenafil did not improve six-minute walk distance or quality of life, and was associated with modest worsening of renal function and increases in neurohormone levels. [81] In a seven-day study, 61 HFPEF patients were randomized to placebo or ivabradine, a selective sinus node I_f sodium channel inhibitor that reduces heart rate without affecting contractility or lusitropy. Patients in the ivabradine group increased peak oxygen consumption and reduced exercise E/e′ ratio, and in this short-term study ivabradine was well-tolerated. [82]

Anemia is highly prevalent in HFPEF and carries a poor prognosis; leading to the hypothesis that epoeitin-alfa would improve submaximal exercise capacity and ventricular remodeling. However, after 24 weeks of therapy there was no change in 6-minute walk distance or left ventricular end-diastolic volume. [83] The ongoing RAnoLazIne for the Treatment of Diastolic Heart Failure (RALI-DHF) study, based on pre-clinical models, randomized HFPEF patients with invasively confirmed diastolic dysfunction to intravenous followed by oral ranolazine vs. placebo. [84] In acutely decompensated HF patients, intravenous serelaxin was highly effective in relieving dyspnea and was associated with encouraging trends for end-organ function and 6-month mortality in both HFPEF and HFREF; however, rehospitalization rates were not affected. [85].

Nonpharamacological strategies

Given that symptoms in HFPEF are most prominent during physical activity, interest has recently focused on exercise training as a potential treatment modality. In a single-blind, single-center study, 53 HFPEF patients were randomized to moderate-intensity aerobic exercise training vs. intention control. The intervention group exercised in a medically supervised environment three times weekly for 16 weeks, and the intervention increased peak exercise oxygen uptake, 6-minute walk distance, and physical quality-of-life scores. [44] Similar results for these endpoints were seen in a multicenter study of 40 HFPEF patients randomized to a 32-session, 3-month exercise protocol including both aerobic and resistance exercise training. [43] To date, positive effects on exercise intolerance from exercise training has been reported in 5 studies involving over 200 HFPEF patients. [86] The effect of exercise training on survival in HFPEF is unknown, but will be examined in the Ejection-HF trial. [87]

Patients with HFPEF are often advised to limit dietary sodium intake, [13] and those who receive this recommendation at hospital discharge have a lower risk of hospital readmission. [7] In multiple ‘salt-sensitive’ experimental models of HFPEF, high sodium consumption exacerbates oxidative stress and adverse cardiovascular remodeling. [53, 52, 51]. The sodium-restricted Dietary Approaches to Stop Hypertension (DASH/SRD) reduces oxidative stress and in a cohort of postmenopausal women the incidence of HF was inversely proportional to DASH diet adherence. [88–89] In a recent proof-of-concept study, 13 hypertensive HFPEF patients consumed the sodium-restricted DASH diet for 21 days. Clinic and 24-hour ambulatory blood pressure were significantly reduced, and urinary F2-isoprostanes, a measure of systemic oxidative stress, declined by 31%. [90] In addition, relaxation- and stiffness based measures of diastolic function improved, arterial elastance decreased, and the ventricular-vascular coupling ratio improved. [91] These hypothesis-generating findings remain to be confirmed in a larger randomized study.

Conclusions

In summary, recent work illustrates that HFPEF patients frequently have functional abnormalities in multiple cardiovascular and noncardiovascular domains. In any given patient, one mechanism may predominate or several may contribute simultaneously to the HFPEF syndrome. Structural and functional phenotyping of HFPEF patients may have important implications for clinical trial patient selection and individualized treatment plans. Animal models, epidemiological cohorts, and small mechanism-focused studies in humans suggest that oxidative stress and chronic inflammation are important underlying contributors to the development and progression of HFPEF. Considering phenotypic heterogeneity and novel mechanisms may lead to new therapeutic and prevention-focused strategies for HFPEF.