Heart failure (HF) is a progressive condition in which the heart muscle is not able to pump enough blood to meet the needs of the body. The prevalence of HF is in 1–2% of adults in industrialized countries and is increasing with population ageing. HF then represents one of the major public health problems [1]. Heart failure with preserved ejection fraction (HFpEF) is a clinical syndrome in which patients have clinical features of HF in the presence of normal or near-normal left ventricular ejection fraction (LVEF), with LV not filling adequately because of diastolic dysfunction. HFpEF accounts for more than half of HF cases [2,3]. HFpEF is currently identified by a LVEF ≥ 50%, although different definitions of “preserved” EF have been employed in previous studies, with LVEF cut-offs ranging from 40% to 55% [4]. HFpEF was initially considered as a cardiac disorder characterized by diastolic dysfunction, cardiomyocyte hypertrophy and myocardial fibrosis. However, extra-cardiac mechanisms also have a crucial role to play in the pathophysiology of HFpEF, leading it to be rebranded as a multisystem disorder [5,6]. Indeed, HFpEF is frequently associated with non-cardiovascular comorbidities (e.g., chronic kidney disease, anemia, chronic obstructive pulmonary disease). All these diseases, as well as advanced age, promote a mild chronic inflammatory state. Myocardial microvascular inflammation, mediated by pro-inflammatory cytokines, leads to activation of endothelial cells, which highly express adhesion molecules that trigger monocyte migration from the bloodstream into the myocardium and their differentiation into macrophages. This vicious circle leads to a state called “endothelial dysfunction”, which contributes to fibrosis and progressive diastolic dysfunction [7,8]. Conversely, HF with reduced ejection fraction (HFrEF; LVEF ≤ 40%) is mainly characterized by systolic dysfunction as a consequence of a direct heart damage, such as an acute coronary syndrome, a cardiomyopathy or a valve disease. Therefore, the pathophysiology of HFpEF is multifactorial, whereas HFrEF is mostly associated with a neuroendocrine-based dysregulation of cardiovascular systems [9].

The risk of all-cause death is comparable in HFpEF and HF with mildly reduced EF (HFmrEF; LVEF 41–49%) and lower than in HFrEF, while the risk of death or HF hospitalization is lower for HFpEF than HFrEF or HFmrEF [10,11]. While several pharmacological and non-pharmacological approaches have been demonstrated to improve survival in HFrEF, only few interventions have proven able to modify the clinical course of HFpEF, possibly due to a consistent phenotypic variability and to the enrolment of heterogeneous populations in large clinical trials [4,12]. However, the recent EMPEROR-Preserved (EMPagliflozin outcome trial in Patients With chronic heart Failure With Preserved Ejection Fraction) trial has demonstrated improved outcomes in patients with HF and LVEF > 40% with the sodium-glucose cotransporter 2 inhibitor (SGLT2i) empagliflozin as compared to the placebo. There was no signal of a differential effect in the subgroups with or without diabetes as well as in patients with LVEF below and above 50% [13,14]. The results of the DELIVER (Dapagliflozin Evaluation to Improve the LIVEs of Patients With Preserved Ejection Fraction Heart Failure) trial (NCT03619213), a phase III trial enrolling HFpEF and HFmrEF patients and testing dapagliflozin versus placebo, are expected in the near future [15].

B-type natriuretic peptide (BNP) and the N-terminal fragment of proBNP (NT-proBNP) are produced from the cleavage of their 108-aminoacid precursor proBNP by proprotein convertases, such as corin and furin. The biologically active BNP is degraded by several peptidases, such as dipeptidyl peptidase IV and neutral endopeptidases (NEP or neprylisin) [19,20,21]. BNP is produced by ventricular myocytes in response to any kind of damage to the cardiovascular system, including increased myocardial wall stress, and plays a major role in HF pathophysiology, by counteracting the detrimental effects of renin-angiotensin-aldosterone system (RAAS) and sympathetic nervous system activation through their diuretic, natriuretic, vasodilator and anti-hypertrophic properties [22]. BNP and NT-proBNP are crucial biomarkers for the diagnosis of HFrEF, with a less established role for risk stratification and management [4,23]. Nonetheless, their clinical value has been studied across the whole spectrum of LV systolic function. Circulating levels of NPs are increased in patients with HFpEF and mirror the severity of cardiac morphological and functional abnormalities, such as LV hypertrophy, fibrosis and diastolic dysfunction [24,25]. Therefore, their measurement is a central element in the diagnostic algorithm for HF [4,23]. NP levels are less elevated in HFpEF than HFrEF, but no single cut-off value has been shown to accurately differentiate the two conditions [26,27]. A meta-analysis of 51 studies reported that NPs have reasonable diagnostic performance in the detection of HFpEF in a chronic setting (area under the receiver operating characteristics curve [AUC] 0.80; 95% CI 0.73–0.87) [28]. Comorbidities influence NP circulating levels in both HFpEF and HFrEF, which is of great clinical importance, given the higher prevalence of non-cardiac conditions in HFpEF. Several conditions are associated with higher NPs, including chronic obstructive pulmonary disease (COPD), atrial fibrillation (AF), kidney disease, diabetic ketosis, while NP levels may be significantly reduced in obese patients [29,30,31]. Furthermore, increased age is associated with higher NP concentrations [29,32]. NP elevation has been used in many trials as an inclusion criterion to improve the diagnostic specificity and for risk enrichment [33,34]. BNP and NT-proBNP have also been tested as tools for risk stratification. The prognostic performance of BNP in HFpEF is similar to that in HFrEF, since the rates of death and HF related hospitalization are similar to those of patients with impairment of systolic function for any given level of BNP [27]. In the Irbesartan in Patients with Heart Failure and Preserved Systolic Function Study (I-PRESERVE), which enrolled 4128 patients with HF and an LVEF ≥ 45% for a mean follow-up of 49.5 months, NT-proBNP above the median value of 339 ng/L was independently associated with an increased risk of the primary composite endpoint of all-cause death and cardiovascular hospitalization in patients with LVEF ≥ 45% [35]. In a sub-analysis of the I-PRESERVE trial including 2162 patients, Jhund et al. have investigated the association between changes in NT-proBNP over a 6-month follow-up and clinical outcomes (cardiovascular death or HF hospitalization; all-cause death, HF death or HF hospitalization). Changes in NT-proBNP were associated with the risk of clinical outcomes, and particularly with HF-related outcomes. A 1000 ng/L elevation in NT-proBNP over 6 months was associated with a 2-fold higher risk of cardiovascular death or HF hospitalization (HR 2.01, 95% CI 1.50–2.61) [36]. The prognostic role of NT-proBNP has been also evaluated in a recent analysis of the EMPEROR-Preserved trial, which enrolled and randomized 5988 patients with a LVEF > 40% and NYHA class II-IV to receive empaglifozin or a placebo [37]. 5986 (99.9%) participants had available baseline NT-proBNP measurements, with an overall median baseline NT-proBNP level of 974 ng/L (Q1 and Q3 at 499 and 1731 ng/L, respectively). Patients with higher NT-proBNP concentrations were older and showed a more severe degree of HF, including lower LVEF, worse clinical manifestations and health status measured by KCCQ score. Moreover, an increase in baseline NT-proBNP across quartiles was reflected by an enhanced risk of cardiovascular (CV) death, >4-fold higher in the placebo group compared with the highest quartile, and HF hospitalization, 5-fold higher total number of hospitalizations in the placebo group compared with the highest quartile. The increase in NT-proBNP level from baseline to 12 week was associated with risk for CV death both in the placebo (HR: 1.88, 95% CI: 1.57–2.26) and empagliflozin (HR: 1.57, 95% CI: 1.30–1.90) group. Treatment with empagliflozin reduced clinical outcomes across NT-proBNP quartiles without interaction with baseline NT-proBNP and contributing to mildly reduce NT-proBNP levels [37]. Even in patients with acute HF and preserved EF, NT-proBNP was a strong predictor of all-cause mortality [38,39,40]. In a study conducted on 205 patients with HFpEF hospitalized for acute HF, after a mean follow up of 28 ± 10 months, discharge NT-proBNP ≥ 1500 ng/L (HR: 5.23, CI 95%: 2.87–17.8, p < 0.001) and ≥50% NT-proBNP reduction between admission and discharge (HR: 0.62, CI 95%: 0.25–0.79, p = 0.019) were independent predictors of death and rehospitalization for HF. Moreover, the combination of E/e’ and NT-proBNP values at discharge significantly improved the prognostic ability compared to each variable analyzed independently (AUC, NT-proBNP at discharge: 0.80; E/e’ at discharge: 0.77; E/e’ + NT-proBNP: 0.88; p < 0.01) [41]. Another study explored the prognostic significance of NT-proBNP levels in patients hospitalized for acute HF with preserved versus reduced EF. Notably, discharge NT-proBNP concentrations predicted clinical outcomes similarly in HFpEF and HFrEF. In a cohort of patients with HFpEF (n = 283) compared to those with HFrEF (n = 776) followed up for 6 months, multivariable adjusted Cox regression analysis reported that for any 2.7-factor increase in NT-proBNP levels, the HR for mortality was 2.14 for HFpEF (95% CI 1.48 to 3.09) and 1.96 for HFrEF (95% CI 1.60 to 2.40). Furthermore, prognostically relevant comorbidities were more often present in patients with HFpEF than patients with HFrEF, but only in low (≤3000 ng/L) and not in high (>3000 ng/L) NT-proBNP discharge categories [40]. Guideline-recommended HF therapies reduce NP levels [42]. The use of NPs to guide HF therapy is still controversial, even though some metanalyses showed that a NP-guided treatment is associated with lower rates of all-cause mortality and HF hospitalization [43,44,45], but some uncertainty is reported in other systemic reviews [46,47]. A few studies have explored the effectiveness of a NP-guided therapy in HFpEF. Most notably, patients with LVEF > 45% randomized to medical therapy titrated to reduce symptoms to NYHA ≤ II presented a better 18-month outcome compared to those whose treatment also pointed at a reduction in NT-proBNP below the inclusion thresholds (>400 ng/L or >800 ng/L according to the age) [48]. Overall, further evaluation is warranted to better understand the differences between a NP-guided therapy and a clinically-guided therapy, and for the potential use of NPs in the follow up of HF patients. As for other NPs, the measurement of circulating atrial NP (ANP) is complicated due to its short half-life (2–5 min) related to the rapid cleavage by neprilysin, insulin-degrading enzyme and natriuretic peptide receptor-C. ANP precursor (proANP), which is stoichiometrically equimolar to ANP, has a longer half-life and it is more easily measurable by searching its mid-regional portion (MR-proANP) [9,49]. MR-proANP has been firstly examined for diagnosis in the Biomarkers in the Acute Heart Failure trial (BACH), in which it revealed great diagnostic ability in acute decompensated HF (cut-off point of ≥120 pmol/L had a sensitivity of 97%, specificity of 60% with accuracy of 74%) [50]. The use of MR-proANP as a biomarker has not yet been extensively investigated in HFpEF, but recent studies highlight its prognostic value. In a study population of 806 subjects with type 2 diabetes (T2D) from the Tousand&2 Study including 141 (17.5%) patients with HFpEF, 646 controls without HF and 19 patients with HFrEF, authors evaluated the association between cardiovascular events and MR-proANP, during a median follow up of 4.8 years. MR-proANP level was associated with a higher risk of incident cardiovascular events (multivariable model HR: 2.56, 95% CI 1.64–4.00) in patients with HFpEF and high MR-proANP, while patients with HFpEF and a low MR-proANP did not show a different risk for incident cardiovascular events compared to patients without HF (multivariable model HR: 2.18, 95% CI 0.78–6.14) [51]. Similar results were obtained on a cohort of 143 patients, including 57 controls without HF, 43 patients with HFpEF and 43 with HFrEF. MR-proANP was associated with the endpoint of HF hospitalization or death in HFpEF (HR adjusted for age, sex, and body mass index [BMI] 1.61, 95% CI 1.07–2.32) [52]. Cardiac troponins are released from intracellular space to the bloodstream following alterations in membrane properties. The release of myocardial troponins may not require myocardial cell death, but the extrusion of proteins from reversibly injured cardiomyocytes may occur during transient increases in cell permeability due to cell wounds [53,54]. In patients with HF, increased levels of troponin generally correlate with HF severity, especially in the acute setting, but the elevation of circulating troponins has been reported also in chronic HF, possibly due to mechanisms such as inflammation, neurohormonal activation, myocardial stretch, hypoxia, cytotoxicity [55]. Santhanakrishnan et al. have shown that hs-TnT was higher in HFrEF than in HFpEF (p < 0.04), after adjustment for age, sex and other clinical covariates (e.g., estimated glomerular filtration rate (eGFR), diabetes, hypertension, coronary artery disease, AF, and HF therapies) in 50 patients with HFpEF, 51 with HFrEF and 50 controls without HF [56]. Other studies on larger cohorts confirmed the higher hs-TnT levels in HFrEF and HFmrEF patients compared to HFpEF patients [57,58]. For example, the Trial of Intensified vs. Standard Medical Therapy in Elderly Patients With Congestive Heart Failure (TIME-CHF) study confirmed that patients with LVEF ≥ 50% have lower hs-TnT than those with LVEF ≤ 40% (27.7 [16.8–48.0] vs. 32.4 [19.2–59.0] ng/L, p = 0.03) [57]. Both hs-cTnT and hs-cTnI levels are increased in chronic HFpEF and show a stronger association with poorer outcomes in men (HR 3.33; 95% CI 1.82–6.09) than in women (HR 1.35; 95% CI 0.94–1.93), while there were no significant differences in HFrEF [59]. The mechanism of this sex-related difference in the prognostic value of hs-cTn is unclear. In the Multi-Ethnic Study of Atherosclerosis (MESA), hs-TnT and NT-proBNP could identify the subset of patients with LV hypertrophy at a higher risk for incident HF, both with and without impaired systolic function [60,61]. In a retrospective analysis of the longitudinal The St Vincent’s Screening TO Prevent Heart Failure (STOP-HF) study, hs-TnI at baseline was a significant predictor of HFpEF development among individuals with risk factors for HF, whereas changes in plasma levels over time were not predictive [62]. The hs-Tn thresholds for risk stratification in the general population have been recently suggested: hs-TnI < 4 or <6 ng/L is indicative of low risk in women and men, respectively, and >10 or >12 ng/L is indicative of a higher risk [63,64,65]. Several studies investigated the prognostic value of troponin assay in patients hospitalized for HF and having a preserved EF. In a cohort of 500 patients with LVEF ≥ 40%, TnT was directly correlated with serum creatinine and symptom severity, and independently predicted all-cause death and HF rehospitalization [66]. In a recent study enrolling 847 HF patients (43% with HFpEF), the AUC of hs-TnT for the prediction of mortality at 30 days was significantly lower in patients with HFpEF (AUC 0.61) than in those with HFmrEF (AUC 0.80, p = 0.01) or HFrEF (AUC 0.74, p = 0.04). hs-TnT displayed no significant association with 30-day outcome in the HFpEF group (odds ratio [OR], 1.48; [95%-CI 0.89–2.46]; p = 0.13), as opposed to HFmrEF (OR 4.53 [95%-CI 1.85–11.1]; p < 0.001) and HFrEF (OR 2.58 [95%-CI 1.57–4.23]; p < 0.001), suggesting a lesser prognostic value of hs-TnT in HFpEF [67]. As for chronic HF, a study on 155 HF patients (41% with HFpEF) over a median follow up of 449 days revealed that patients with HFpEF who developed adverse events had higher hs-TnT concentrations compared to those who did not (36 (20–66) ng/L vs. 21 (15–32) ng/L, p = 0.003). The AUC for hs-TnT was higher than BNP (0.739 vs. 0.631), and the optimal hs-TnT cut-off for adverse events (all-cause mortality, non-fatal myocardial infarction, non-fatal stroke, HF hospitalization) was 26 ng/L [58]. The phase 2 Prospective comparison of ARNI (angiotensin receptor–neprilysin inhibitor) with ARB (angiotensin-receptor blockers) on Management Of heart failUre with preserved ejectioN fraction (PARAMOUNT) and the phase 3 Prospective Comparison of ARNI with ARB Global Outcomes in HF With Preserved Ejection Fraction (PARAGON-HF) trials randomized patients with HFpEF to sacubitril/valsartan or valsartan, showing that patients receiving sacubitril/valsartan had a greater reduction in hs-TnT compared to those assigned to valsartan (PARAMOUNT: 14%, p = 0.03; PARAGON-HF: 9%, p < 0.001) [68,69]. Moreover, in the PARAGON-HF trial, patients with hs-TnT levels reduced by 16 weeks to ≤17 ng/L (median value at baseline) had a lower risk of CV death or HF hospitalization compared with those with persistently elevated hsTnT (p = 0.046) [69]. Data regarding the prognostic ability of troponins also derive from a recent analysis of the EMPEROR-Preserved trial. Of 5988 study participants with LVEF > 40%, 5825 (97.3%) had available hs-TnT measurements, with an overall median baseline of 17.8 ng/L (Q1 and Q3 at 11.6 and 26.9 ng/L respectively) and 3767 (65.7%) patients showing hs-TnT > 14 ng/L [37]. Similar to the results for NT-proBNP concentrations, higher hs-TnT levels were associated with more severe HF and comorbidities. A significant increase in the rates of CV death and HF hospitalization was observed across quartiles, with a 5-fold higher number of events between quartiles 1 and 4 in patients randomized to placebo [37]. Similar results were obtained in another analysis of EMPEROR-Preserved trial, in which patients with both the lowest NT-proBNP and lowest hs-TnT had a primary event (CV death and HF hospitalization) rate of 2.2 per 100 patient-years compared to 19.2 per 100 patient-years in those with highest NT-proBNP and hs-TnT, with a rate ratio of 8.7 [70].