The traditional classification of HFPEF as diastolic heart failure is based on the premise that a single pathophysiologic mechanism underlies the genesis of HFPEF. Yet as understanding progressed and recognition that HFPEF is a complex clinical syndrome with multiple operative pathophysiologic mechanisms including several different subgroups of patients, it became apparent that diastolic dysfunction was inadequate to explain all cases of HFPEF [19]. Rational sub grouping of patients into cohorts that have operational and targetable pathophysiologic mechanisms can be appreciated by assessment of the pressure-volume (PV) relations that characterize different populations with HFPEF [20, 21]. PV measurements provide a systematic means of characterizing overall, systolic and diastolic properties of the heart, as well as isovolumetric pressure volume area (PVA_iso), which when indexed to EDP provides an afterload independent measure of the pump function of the ventricle [22]. End-systolic PV relation (ESPVR) and end-diastolic PV relation (EDPVR) in particular reflect global ventricular chamber pump properties that effect overall cardiac function. These parameters are in turn determined by intrinsic myocardial properties (systolic and diastolic function), but also by muscle mass, chamber architecture (how myocardial fibers are assembled), chamber shape, and sequence of myocardial activation.

Data from PV analyses in both animal and human studies have revealed at least three different phenotypes of HFPEF that are depicted in Fig. 7.2. In the first, patients have a normal ESPVR and upward shifted EDPVR—(e.g. reduced chamber capacitance or diastolic dysfunction) with normal chamber contractility (e.g. preserved systolic function) (Fig. 7.2a).

Patients with this classic diastolic heart failure paradigm have intrinsic myocardial disease causing impairment in cardiac relaxation and passive ventricular filling. Such severe and isolated diastolic dysfunction is often found in hypertrophic cardiomyopathy and infiltrative diseases such as cardiac amyloidosis. These patients with isolated diastolic dysfunction can be identified clinically by the presence of heart failure and a normal ejection fraction without concomitant hypertension [23]. Physical exam findings are often notable for signs and symptoms of a stiff ventricle, including an S4 gallop and other classic but non-specific signs of heart failure such as pulmonary congestion and elevated jugular venous distension. In subjects with restrictive cardiomyopathies the jugular venous pressure often has a “double dip” which is indicative of the rapid “x” and “y” descent seen on the right atrial pressure tracing in patients with restriction. Ejection fraction is usually greater than 50 % in the early phases of these conditions, and on echocardiogram Doppler evidence of diastolic dysfunction ranging from impaired relaxation to a restrictive filling pattern can be observed—although Doppler will be notably abnormal and indistinguishable in each of the following phenotypes of HFPEF [24]. Finally, the PVA_iso will be lower than normal in these subjects indicative of left ventricular pump dysfunction as a primary pathophysiologic mechanism underlying the phenotype.

In the second paradigm of HFPEF, ESPVR and EDPVR are both shifted upward and leftward, with a resulting decrease in chamber capacitance but with concomitant enhanced chamber systolic function—as evidenced by the upward shift of the ESPVR (Fig. 7.2b). In such subjects, the area between the ESPVR and EDPVR will be unchanged and the PVA_iso will remain normal. Examples of clinical conditions that produce this phenotype include increased afterload from chronic hypertension (usually from central conduit artery stiffening) and severe aortic stenosis. These two clinical scenarios are the most common causes of this phenotype in older adults with HFPEF. With normal physiologic aging, increasing central conduit arterial stiffness from collagen crosslinking leads to an increased arterial elastance with concomitant hypertension. Consequent concentric hypertrophy of the left ventricle yields subsequent reliance on atrial filling with corresponding left atrial (LA) enlargement [7, 8, 25, 26] Similarly, concentric remodeling with a small ventricular chamber is seen in patients with aortic stenosis who are characterized by this phenotype according to this paradigm. Ejection fraction is often supranormal in these patients (>60 %) as a result of an increase in end systolic elastance with a similar echocardiographic phenotype characterized by a normal to small LV chamber size, concentric LV hypertrophy and LA enlargement [27].

In the third subtype of HFPEF, patients can display normal ESPVR and normal or slightly rightward and downward shifted EDPVR (e.g. normal or increased capacitance) with increased end diastolic pressures. The latter results from an increased central volume that could be caused by either an expanded blood volume or a shift of the blood volume toward the central circulation from veno-constriction. Indeed, most of the blood volume in mammalian species is in the venous bed, and small changes in venous tone can result in large shift of blood volume back to the central circulation resulting in left ventricular overfilling [28]. Pathophysiologically, this phenotype is characterized by increased preload without significant alteration in LV systolic or diastolic function (e.g. the ESPVR and EDPVR are not significantly different from normal) (Fig. 7.2c) [29]. Although these patients also have hypertension, the primary operative mechanism is increased preload due conditions that expand the blood volume such as renal dysfunction, obesity, anemia and others. Indeed, age-related changes in non-cardiac systems including renal, pulmonary, endocrine, and autonomic nervous systems have been documented to produce a state of salt sensitive hypertension characterized by excessive salt and water retention [30–32]. The diminished pulmonary vascular capacitance with pulmonary arterial hypertension and endothelial dysfunction coupled with the volume redistribution that accompany advancing age (decreased beta-2 adrenergic responsiveness of the veins reduces venous capacitance) predispose the aging adult to the development of symptomatic pulmonary edema in the setting of normal LV systolic function. Patients display signs of volume overload on physical exam, and can have an S3 gallop. Ejection fraction is commonly >50 % and echocardiography shows a mildly dilated LV that often is unrecognized [33]. LV pump function as indexed by the PVA_iso to EDP relation is supranormal in these subjects suggesting that the myocardium is performing more work at any given filling pressure.

Finally, subtle reductions in systolic function as evidenced by a reduction in LV chamber contractile function (either from a downward shifted ESPVR or increased V₀ [volume axis intercept of the ESPVR]) can cause neurohormonal activation that leads to salt and water retention (Fig. 7.2d) [34, 35]. In these subjects, the EF may be low normal or slightly reduced (e.g. ~40–50 %). The primary mechanism in this underappreciated subtype is systolic dysfunction and PVA_iso to EDP relation that demonstrates a downward shift compared consistent with reduced chamber contractile strength. These patients are often hypertensive, and on physical exam will appear most similarly to patients with a mild form of systolic heart failure as demonstrated by non-invasive PV analysis [36].

The pathophysiology of HFPEF is heterogeneous, involving multiple physiologic domains in both cardiac and non-cardiac systems—and manifesting in patients with multiple co-morbid conditions. Despite the multifactorial etiology of the syndrome, there are several clinical signs and symptoms consistently present in patients that characterize the phenotype(s) of HFPEF: (i) Labile blood pressure with high resting LVEDP, (ii) predisposition for acute pulmonary edema, and (iii) effort intolerance with diminished exercise capacity [37–39]. Appreciating the distinct pathophysiologic mechanisms that contribute to these phenotypes is essential in guiding therapeutic interventions and in advancing the understanding and management of this condition.

Diagnosing HFPEF presents a clinical challenge stemming from its diverse etiology. Making the diagnosis of HFPEF is particularly difficult because: (1) there is no specific measure (such as EF in HFREF) to define the syndrome, (2) other disorders common in older adults (such as obesity with confounding comorbidities) can mimic the clinical syndrome, and (3) precise criteria have not been widely adapted. The original criteria developed to distinguish HFPEF from HFREF called for the presence of a clinical heart failure syndrome, a normal EF, and evidence of diastolic dysfunction by either catheterization or echocardiography [40]. Since then, several guidelines have been published that broaden the requirement of diastolic dysfunction to surrogate markers of diastolic LV dysfunction such as LV hypertrophy, LA enlargement, atrial fibrillation, or elevated plasma natriuretic peptides (NP) levels (Fig. 7.3) [41–45].

In the most recent guideline published by the Heart Failure and Echocardiography Associations of the European Society of Cardiology (ESC), the diagnosis of HFPEF required signs or symptoms of heart failure, a LVEF >50%, a LVEDV Index <97 mL/m², and evidence of diastolic LV dysfunction as shown by LVEDP >16 mmHg, pulmonary capillary wedge pressure (PCWP) >12 mmHg or ratio of mitral peak velocity of early filling (E) to early diastolic mitral annular velocity (E/E’) >15 alone. Alternatively criteria include an elevated NP with E/E’>8, a mitral flow velocity Doppler signal showing a ratio of early to late flow (E/A) <0.5 with deceleration time (DT) >280 ms, a pulmonary vein flow velocity signal showing a Ard-Ad >30 ms (Ard = duration of reverse pulmonary vein atrial systole flow; Ad = duration of mitral valve atrial wave flow), a LA size >40 mL/m², or a LV mass >149 g/m² in men or >122 g/m² in women [43]. These guidelines represent consensus criteria with ongoing validation studies needed to verify their utility (in particularly for those that are non-invasive). Although invasive measurements with cardiac catheterization remain the gold standard for identifying HFPEF, the invasive nature of catheterization makes it impractical to obtain during routine evaluation for all patients with the HFPEF phenotype.

The diagnostic criteria above are largely based on the premise that elevated filling pressures remain the primary hemodynamic abnormality in HFPEF. Yet in a recent study assessing the diagnostic utility of these ESC guidelines, only 25 % of patients with HFPEF and unexplained dyspnea who underwent invasive diagnostic testing fulfilled ESC criteria for HFPEF with evidence of elevated filling pressures at rest [46]. Indeed, a significant proportion of the patients demonstrated left ventricular stiffness, dyssynchrony, and dynamic mitral regurgitation, confirming patients with HFPEF can experience heart failure symptoms without demonstrating significant elevated filling pressures at rest or during hand grip exercise (increased afterload), leg lift (increased preload), nitroprusside infusion (decreased afterload), or dobutamine infusion (increased contractility). The specificity of these guidelines was also challenged, with 20–40 % of health controls demonstrating borderline E/E’ despite no evidence of heart failure and normal filling pressures [46]. Similarly, in a recent study assessing hemodynamic response to volume challenge in healthy young volunteers, older women, and patients with HFPEF, filling pressures (assessed by PCWP and mean pulmonary artery pressures [MPAP]) rose significantly with volume loading in all of the groups [47]. This raises additional concern for the validity of the cut points utilized to define an elevated filling pressure characteristic of HFPEF, suggesting more testing is needed to determine the threshold at which to classify HFPEF.

Despite these findings, several studies have validated the diagnostic capacity of various non-invasive estimates of elevated filling pressures. Conventional and tissue Doppler echocardiographic indices have been measured against conductance catheter PV loop analyses in patients with diastolic dysfunction (confirmed by invasively measured indices for diastolic relaxation [tau], LVEDP and LVEDV), demonstrating the LV filling index E/E’ to have the highest ability to detect diastolic dysfunction (86 % sensitivity) [48]. This measure (E/E’) has subsequently been evaluated against LA volume index and LV mass index, demonstrating increased sensitivity of detecting HFPEF in patients with LA size >40 mL/m² and LV mass >149 g/m² in men or >122 g/m² in women [46].Yet when these non-invasive indices were compared with pulmonary capillary wedge pressure (PCWP) assessed by right heart catheterizations, the non-invasive E/E’ did not reliably track changes in left-sided filling pressures [49]. This disconnect between patients with HFPEF and non-elevated filling pressures at rest has been examined further by a study of euvolemic patients with exertional dyspnea, preserved EF, and normal cardiac filling pressures at rest: during exercise, patients demonstrated significantly abnormal hemodynamic responses including elevated PCWP [50]. While exercise hemodynamics may lack specificity as abnormal findings can be found in normal aging, it presents an attractive method to yield findings that may not be apparent at rest. The role of exercise hemodynamics in the diagnosis of HFPEF presents an emerging area of potentially improved diagnostic sensitivity. Although further validation of exercise hemodynamics in the diagnosis of HFPEF is necessary, it may be useful in the evaluation of patients who do not meet other established criteria for HFPEF.

Finally, the measurement of circulating natriuretic peptides (NP) has been suggested as an adjunctive diagnostic tool, but is also confounded by the influence of common co-morbidities present in patients with HFPEF on the presence of elevated NP [46]. Recent studies have demonstrated the association between NP and HFPEF, however validation of this marker in the diagnosis of HFPEF needs to be verified in ongoing clinical studies [51, 52].

Given the non-specific nature of the current non-invasive measures and their lack of correlation with invasive measures, the diagnostic utility of the ESC criteria have been called into question [35, 49]. Since many older adults have comorbid conditions that mimic the symptoms of dyspnea, fatigue, paroxysmal nocturnal dyspnea, orthopnea and leg swelling classically associated with heart failure, the specificity of the clinical diagnosis of HFPEF is reduced. Similarly, physical exam findings are both difficult to assess and notoriously non-specific for diagnostic criteria, and the diagnostic EF cutoff defining “preserved EF” has not been consistently used in diagnostic studies [3, 17, 21, 42, 52–54]. The nature of these diagnostic dilemmas has created a clinical challenge for the treating physician. Consequently, many clinicians have opted for an approach to diagnosis that employs a combination of the previous diagnostic strategies that is tailored to specific clinical settings and includes a clinical phenotype consistent with heart failure, a preserved EF, and echocardiographic evidence for LV hypertrophy and LA enlargement [51].

The rationale behind many of the large pharmaceutical trials is that drugs that reduce ventricular hypertrophy or improve myocardial relaxation should benefit patients with HFPEF. This premise is primarily based on improving diastolic function, which as previously mentioned is only one of several mechanistic contributors to HFPEF. The following drug classes have been studied in large randomized placebo-controlled clinical trials (Table 7.1): beta-blockers, angiotensin-converting enzyme inhibitors (ACEi), angiotensin receptor blockers (ARB), aldosterone antagonists, and digoxin.

Each of these trials is notable for including subjects with clinical HFPEF and, with the exception of RAAM-PEF, having greater than 100 and up to 4,128 patients (I-PRESERVE). The trial subjects had a mean age ranging from 63 years to 76 years, and EF cutoffs ranging from ≥35 to ≥50 %. None of these trials demonstrated mortality benefit, and most failed to show benefit in the endpoint of heart failure hospitalizations; only candesartan (CHARM-P) and perindopril (PEP-CHF) exhibited decreased hospitalizations. Finally, despite a lack of placebo-controlled trials demonstrating clinical benefit of diuretic use in HFPEF, in the Hong Kong diastolic heart failure study of 150 elderly patients with HFPEF, diuretics significantly reduced heart failure symptoms and improved quality of life, and neither ramipril nor irbesartan demonstrated additional effect [66] Diuretics remain essential therapy in the symptomatic control of volume overloaded states, although their use should be cautioned among preload dependent euvolemic patients. See Table 7.2 [67–77].