Heart failure with preserved ejection fraction (HFpEF) affects around half of all heart failure (HF) cases, and throughout the past 20 years, this percentage has been rising [1, 2]. HFpEF prevalence in the adult population is around 1–3%, although it is increasing in the general global population [3]. Additionally, the 1-year mortality is about 6.3%, reaching 75.7% at 5 years, which is comparable to heart failure with reduced ejection fraction (HFrEF). While cardiovascular death remains the primary outcome for HF, non-cardiovascular causes are increasingly significant in HFpEF [4, 5, 6].

Compared to HF with impaired systolic function, HFpEF is regarded as a distinct entity with specific molecular vascular myocardial cell adaptations and a detailed pathophysiological profile (HFrEF) [7]. This evaluation is based on certain presumptions regarding the anatomy of the left ventricle (LV), the composition and structure of the heart, the geometry of the intrinsic fiber myocytes, the thickness of the vessels, and the effects of therapy [8]. Consequently, it is a diverse syndrome with a high comorbidity burden characterized by several related diseases and different clinical phenotypes. The considerable variety of enrollment criteria in the larger multicenter trials showed a lack of diagnostic consistency, and criteria uniformity is required for both the time course evaluation and the adopted treatment [9]. Understanding the underlying pathophysiological processes that generate hemodynamic changes and cardiovascular dysfunctions is crucial for addressing the optimized therapy according to different phenotypes. For example, patients with HFpEF and a prevalent coronary microvascular impairment as the main pathophysiological mechanism may benefit more from drugs acting on microcirculation that reduce coronary resistance, improving coronary blood flow. Alternatively, patients with HFpEF and prevalent extracellular fibrosis may benefit more from drugs with anti-fibrotic properties, which reduce collagen extracellular storage. Patients with early supraventricular tachyarrhythmias (SVTs) may benefit from early ablation procedures because recurrent supraventricular tachyarrhythmia (SVT) may trigger electrical disorder persistence and HFpEF progression. In conclusion, the emerging role of genetics and epigenetics may open new perspectives in HFpEF classification and treatment. The contrasting results of novel and conventional treatments emphasize the need to search for a more suitable mechanistic strategy based on more customized therapy that incorporates data on lifestyle, genetic, biochemical, environmental, and cardiovascular (CV) risk factors [10]. This review aims to revise the knowledge concerning this complex syndrome, particularly regarding the different phenotypes and the current and potential therapeutic targets in HFpEF.

Many questions are currently being raised regarding the influence of common cardiovascular risk factors in the general population that can induce HFpEF, the role that comorbidities play in the development of this syndrome, and the precise correlation between this syndrome and underlying cardiac dysfunction. These questions result from the inconsistent findings of clinical trials and the generation of different hypotheses. More recently, some epidemiological studies have proposed that HFpEF represents a terminal phase of various diseases with diverse clinical courses, outcomes, and phenotypes. Distinct patterns and HFpEF subtypes have been revealed by certain cluster analyses of big randomized clinical trials (RCTs). For example, in the TOPCAT trial, latent class analysis identified three primary categories based on left ventricular geometry, clinical profile, risk factor burden, and vascular characteristics [20]. The authors identified three main phenogroups: Group 1 with low NP levels, normal LV geometry, low arterial stiffness, and cardiac events; Group 2 was characterized by high NP levels, increased vascular stiffness associated with LV concentric hypertrophy, LA dilatation and an elevated incidence of cardiovascular events; Group 3 is the typical metabolic pattern with an elevated level of rennin, an abnormal glycemic and lipid profile, an increased inflammatory pattern and a better response to spironolactone treatment [20, 21].

Five distinct groups were chosen by another analysis based on shared laboratory characteristics: Age, gender, and related disorders, which came from the SwedHF registry. Similarly to the previous analysis, Cluster 1 included younger patients with low comorbidities and NP values. Cluster 2 encompasses patients with a high prevalence of AF and diabetes associated with impaired renal function; Cluster 3 demonstrated a higher incidence of AF, older age, the prevalence of females, and those associated with a high NP level; Cluster 4 could be identified by the elevated prevalence of hypertension and diabetes with high body mass index (BMI); Cluster 5 identified older patients with cardio–renal disorders and more advanced New York Heart Association (NYHA) classes. The current classifications show a poorer outcome among groups with cardio–renal and hypertension–diabetic patterns [39]. Similarly, an analysis of the Alberta cohort recognized four phenotypes ranging from healthy subjects at risk of HFpEF without signs of HF up to congested, older patients with elevated comorbidity burdens [40]. The more recent Prospective Comparison of angiotensin receptor–neprilysin inhibitor (ARNI) with angiotensin receptor blocker (ARB) Global Outcomes in HFpEF (PARAGON-HF) included older individuals with a mean age of 72 years, a mean EF of 57%, a significant prevalence of diabetes (43%), and chronic kidney disease (47%), reflecting the real HFpEF patients [41]. Looking at more detailed morphological investigations, these patients had similar LA sizes but lower evidence of increased LV mass values and tricuspid regurgitation velocity compared to previous trials. Conversely, the E/e’ ratio in the PARAGON trial was more noticeable, emphasizing the variability in the cardiac phenotype associated with this disease [42]. The comparison of this research highlights the differences in patient characteristics, especially the observation that only half of the people enrolled in the most well-known trials met the HFpEF diagnostic criteria. Interestingly, 50% of patients exhibited increased LV mass, yet while advanced diastolic dysfunction was common, treatment with sacubitril/valsartan was the most beneficial for those with an EF below normal [43] (Fig. 1).

Current discrepancies demonstrated that different mechanisms may primarily be responsible for disease onset and progression depending on different baseline conditions, burden of risk factors, and cardiovascular adaptation to the underlying triggers [44]. A certain amount of overlap and negligence between phenomapping analyses may reflect different HFpEF phases and selection criteria. Comprehensive morphological analysis, combined with environmental lifestyle and CV risk factors, may result in a more homogeneous classification and improved risk stratification of HFpEF [45, 46].

Screening and treating underlying causes and comorbidities, both cardiovascular and non-cardiovascular, are advised for HFpEF patients.

When AF causes HF, the clinical course tends to be more favorable compared to other causes of HF (known as tachycardiomyopathy) [47]. Nevertheless, the presence of AF may attenuate the prognostic benefits associated with beta-blockers and impede the efficacy of ivabradine [48]. Certain HF treatments, such as angiotensin-converting enzyme inhibitors (ACE-I), may reduce the risk of developing AF.

To manage HFpEF, beta-blockers, long-acting nitrates, calcium channel blockers (CCBs), ivabradine, ranolazine, trimetazidine, nicorandil, and their associations should be contemplated for angina relief, notwithstanding the absence of demonstrable benefits on HF progression or coronary endpoints [49, 50, 51, 52, 53].

Hypertension has emerged as a key driver of HFpEF, posing a significant treatment challenge with uncertain optimal approaches.

ARBs, ACE-I, and CCBs cause more effective LVH regression than beta-blockers or diuretics. However, caution is warranted to prevent hypotension in HFpEF patients with LVH and limited preload reserves.

Further, obesity strongly correlates with HFpEF, revealing distinct pathophysiological pathways in obese patients [8]. Weight reduction and increased physical activity may enhance symptom management and exercise capacity, especially in suitable candidates [54].

Chronic kidney disease (CKD) and declining renal function are more prevalent in HFpEF patients compared to heart failure with mildly reduced ejection fraction (HFmrEF) and HFrEF; however, their impact on outcomes appears less severe in HFpEF relative to the other classifications [55].

Management of HF in chronic obstructive pulmonary disease (COPD) patients generally proceeds without major concerns. While beta-blockers may exacerbate pulmonary function in select cases, they are not contraindicated in COPD or asthma [56, 57].

Before the EMPagliflozin outcomE tRial in patients with chrOnic heaRt failure (EMPEROR-Preserved) and Dapagliflozin Evaluation to Improve the LIVEs of Patients with Preserved Ejection Fraction Heart Failure (DELIVER) studies, American and European guidelines lacked recommendations for the use of disease-modifying HFrEF therapies as clinical trials with ACE-I, ARB, mineralocorticoid receptor antagonists (MRA), while ARNIs failed to achieve their primary endpoints in HFpEF patients.

Notable trials include PEP-CHF (perindopril) [58], CHARM-Preserved (candesartan) [59], I-PRESERVE (irbesartan) [60], TOPCAT (spironolactone) [61], DIG-Preserved (digoxin) [62], and PARAGON-HF (sacubitril/valsartan) [63]. Candesartan and spironolactone showed reductions in hospitalizations for heart failure, with a trend toward reductions observed while using sacubitril/valsartan. However, since these trials were neutral for their primary endpoints, these findings are considered hypothesis-generating only.

In the SENIORS trial, although nebivolol significantly decreased the combined primary endpoints of all-cause mortality and CV hospital admission, it only included 15% of participants with an LVEF greater than 50% [64].

Despite no treatment showing conclusive evidence of reducing mortality and morbidity in HFpEF patients, most of them have underlying hypertension and/or coronary artery disease (CAD), and many have already been treated with ACE-I/ARB, beta-blockers, or MRAs.

In congested patients with HFpEF, diuretics are recommended to mitigate symptomatic distress and clinical manifestations. Loop diuretics are preferred, while thiazide diuretics may also prove efficacious in managing hypertension.

The action mechanisms of ARNIs in HFpEF involve a multifaceted approach targeting various pathological processes, including neurohormonal activation, inflammation, and fibrosis. ARNIs represent a relatively novel class of drugs that combine the actions of ARBs and neprilysin inhibitors. The primary components of ARNIs are sacubitril and valsartan, combined in a unique crystalline salt complex with high stability and good pharmacodynamics. Sacubitril inhibits neprilysin, an enzyme responsible for degrading NPs, which has beneficial effects on blood pressure regulation and sodium balance. Conversely, valsartan is an ARB that blocks the effects of angiotensin II, a vasoconstrictor and pro-fibrotic peptide. Their combination is crucial because increasing the concentration of circulating neprilysin and sacubitril causes a reflex renin-angiotensin-aldosterone (RAAS) activation that limits its benefit, while valsartan inhibits this reflex [65].

The potential beneficial mechanisms of ARNIs in HFpEF are multifaceted:

• NP enhancement: By inhibiting neprilysin, ARNIs increase the levels of NPs, such as atrial natriuretic peptide (ANP) and B-type natriuretic peptide (BNP). These peptides promote vasodilation, natriuresis, and diuresis, which help reduce fluid overload and improve renal function.

• Angiotensin II blockade: Valsartan blocks the effects of angiotensin II, contributing to vasoconstriction, sodium retention, and myocardial remodeling.

• Improvement in cardiac remodeling: In HF, cardiac remodeling is often caused by microvascular inflammation and is often mediated by a reduction of intracellular cyclic guanosine monophosphate (cGMP). Increasing cGMP neprilysin counteracts this action, reducing left ventricular remodeling, myocardial hypertrophy, and fibrosis, thus reducing cardiomyocyte stiffness and ultimately improving cardiac function and structure.

• Beneficial metabolic effects: ARNIs act in both lipid and carbohydrate metabolism, increasing lipolysis and lipid oxidation, insulin secretion, and insulin sensitivity through many pathways. This is crucial because metabolic dysregulation is an important risk factor for HFpEF, especially affecting cardiac diastolic function [66] (Fig. 2).

HFpEF management is still a topic of debate because of the lack of specific therapies, even if many drugs designed for HFrEF have also been investigated for HFpEF. According to both the 2021 ESC guidelines and the 2022 American College of Cardiology (ACC)/American Heart Association (AHA) guidelines on HF, ARNIs may be considered in HFpEF (Class IIB) [11, 67]. Unfortunately, the class of recommendation is still low because, to date, clinical trials have failed to show the benefit of these drugs in hard endpoints such as cardiovascular deaths or hospitalizations.

In 2012, the Prospective Comparison of ARNI with ARB on the Management of Heart Failure with Preserved Ejection Fraction (PARAMOUNT) was the first trial that clearly demonstrated the potential role of ARNIs in HFpEF. PARAMOUNT is a phase 2 randomized double-blind trial comparing sacubitril/valsartan to valsartan alone in 301 patients with HFpEF (EF $\ge$45%) for a 36-week follow-up. It demonstrated a reduction in NT-proBNP levels in the sacubitril/valsartan group after 12 weeks, with an improvement in NYHA class and a reverse LA remodeling at 36 weeks [68]. Despite these results, subsequent trials have been controversial. In 2019, three important trials were published: the PARAGON-HF, the Comparison of Sacubitril/Valsartan Versus Enalapril on Effect on NT-proBNP in Patients Stabilized from an Acute Heart Failure Episode (PIONEER-HF), and the Prospective Comparison of ARNI vs. Comorbidity-Associated Conventional Therapy on Quality of Life and Exercise Capacity (PARALLAX) trials [63, 69, 70].

Using the same study groups as the PARAMOUNT trial but with 4822 patients, who were observed for a median follow-up of 35 months, the PARAGON-HF trial did not show a benefit in deaths or hospitalizations among sacubitril/valsartan patients, despite an improvement in the NYHA class and the quality of life, measured using the Kansas City Cardiomyopathy Questionnaire (KCCQ). Nevertheless, a subgroup analysis suggested potential ARNI survival benefits in women and patients with lower EF (45–57%). Furthermore, the benefit of ARNIs was larger in patients recently hospitalized for HF [63].

The PARALLAX trial with 2572 patients, EF $>$40%, and a median follow-up of 24 weeks showed a reduction in NT-proBNP in patients treated with sacubitril/valsartan compared to patients treated with an ACE-I or ARB. However, there was no significant improvement in the NYHA class, KCCQ, or the 6-minute walk distance [69]. Compared to the PARAMOUNT trial, these findings suggest that NT-proBNP reduction is probably an early effect of ARNI treatment, while other benefits are shown later. Furthermore, the ARNI effects on NT-proBNP are not limited to a chronic setting but are also important in acute HF, as shown by the PIONEER-HF trial in HFrEF [70] and by the Prospective comparison of ARNI with ARB Given following the stabiLization In DEcompensated HFpEF (PARAGLIDE) trial. However, also in the PARAGLIDE trial, sacubitril/valsartan did not reduce mortality, even if the hierarchical outcome showed a positive but not significant trend (unmatched win ratio: 1.19; 95% CI: 0.93–1.52; p = 0.16) [71] (Table 2, Ref. [63, 68, 69, 71]).

In conclusion, the role of ARNIs in HFpEF is still debated; however, it seems to lower HF hospitalizations in HFpEF patients [72]. Furthermore, a recent pooled analysis of the PARAGON and PARAGLIDE trials showed a significant reduction in deaths or hospitalizations in patients treated with sacubitril/valsartan compared to valsartan alone, with a larger benefit in patients with LVEF $\le$60% [73].

MRAs, such as spironolactone and eplerenone, antagonize mineralocorticoid receptors, reducing aldosterone effects that can harm HFpEF, promoting sodium and water retention, and exacerbating cardiac fibrosis and inflammation. The role of MRAs in HFpEF should probably be discussed alongside HF pathophysiology. Indeed, MRAs are probably more effective in patients with hypertension as an etiology of HF [74]. Their potential beneficial mechanisms in HFpEF include:

• Sodium and water balance: MRAs counteract sodium and water retention caused by aldosterone, reducing edema and congestion.

• Reduction in myocardial fibrosis: Aldosterone is implicated in cardiac fibrosis; thus, MRAs may have a role in reducing myocardial remodeling.

• Anti-inflammatory effects: MRAs may have anti-inflammatory properties, which could help ameliorate the systemic inflammation often seen in HFpEF patients [75, 76] (Fig. 3).

As for ARNIs, MRAs could also be used in HFpEF according to the latest ESC guidelines and ACC/AHA guidelines on HF (Class IIB) [11, 67]. However, their effects on CV deaths and hospitalizations remain uncertain and under investigation.

In 2013, the Aldosterone Receptor Blockade in Diastolic Heart Failure (ALDO-HF) trial showed that E/e’ and reduced left ventricular mass in the treatment group (n = 213) compared to the placebo (n = 209), even if it did not benefit the maximal exercise capacity or the patient’s symptoms and quality of life [77]. In 2014, the Treatment of Preserved Cardiac Function Heart Failure with an Aldosterone Antagonist trial showed a decrease in hospitalizations but not in mortality in patients with HFpEF [61, 78], even though a subsequent sub-analysis seems to illustrate a survival benefit in the American cohort of the trial compared to the Russian cohort; differences in drug compliance have been hypothesized to clarify these results [79]. Furthermore, in 2016, the Spironolactone in Myocardial Dysfunction with Reduced Exercise Capacity (STRUCTURE) trial showed an improvement in exercise capacity using the cardiopulmonary test in patients with HFpEF and exertional dyspnea [80].

Currently, the only large study on MRAs with positive results on the survival rate is the Canrenone Effects on Cardiovascular Mortality in Patients with Congestive Heart Failure (COFFEE-IT) trial, a retrospective study that showed the survival benefit of canrenone on older people (68–83 years old) after 10 years of treatment [81]. Despite this, canrenone remains poorly used among MRAs.

Results from two large trials investigating the impact of spironolactone on cardiovascular deaths and hospitalizations should be published in the near future from the Spironolactone Initiation Registry Randomized Interventional Trial in Heart Failure with Preserved Ejection Fraction (SPIRRIT-HF) [82] and the Spironolactone In The Treatment of Heart Failure (SPIRIT-HF) trials [83]. Furthermore, compared to traditional MRAs, non-steroidal MRAs are emerging as drugs with higher selectivity for mineralocorticoid receptors and without sex–steroid-related side effects [84]. Among them, finerenone is the most investigated drug, and subgroup analyses of the FIDELIO-DKD [85] and FIGARO-DKD on patients with HF (symptomatic HFrEF was an exclusion criterion in both trials) suggest a potential role of this drug in HFpEF [86]. Indeed, a reduction in the composite outcome of cardiovascular death, non-fatal myocardial infarction, stroke, or hospitalization for heart failure was observed in a mean follow-up of 2.6 years. As a result, the Finerenone in Heart Failure Patients (FINEARTS-HF) trial is currently ongoing, investigating finerenone effects on morbidity in more than 6000 patients with HF and EF $\ge$40% [87] (Table 3, Ref. [77, 79, 80, 82, 83, 87]).

Recently, the ESC guidelines recommended SGLT2i therapy for treating HFpEF [11], despite the exact mechanisms yet being fully understood. However, several potential mechanisms of action have been proposed based on clinical and pre-clinical evidence [85]. SGLT2i plays the potential role of anti-HFpEF through the direct or indirect synergy of multiple targets and pathways [88].

SGLT2i inhibits the absorption of sodium and glucose in the proximal renal tubule, leading to natriuresis, glucosuria, and elevated urine output. Thus, it has been associated with a reduction in blood pressure. Lowering blood pressure can alleviate cardiac workload and improve overall cardiovascular function.

The vascular and metabolic effects of SGLT2i have always been held responsible for cardiovascular benefits [89]. In HFpEF studies, researchers propose potent reno-protective effects of SGLT2i, with their impact on intraglomerular pressure prevailing over other mechanisms. SGLT2i, by restoring afferent arteriole tone, synergizes with renin-angiotensin system inhibitors, reducing intraglomerular pressure and preventing renal complications. Notably, despite a decline in the estimated glomerular filtration rate (eGFR), SGLT2 inhibition influences primary endpoints [90].

In a post-hoc analysis of the EMPEROR-Preserved Trial, empagliflozin was linked to a slight increase in the risk of volume depletion for patients concurrently using diuretics [91]. However, it was also associated with a decreased probability of initiating or escalating diuretic doses and an increased likelihood of reducing or permanently discontinuing diuretics [91]. Natriuresis derived from empagliflozin is not associated with neurohormonal activation, potassium loss, or impaired renal function (favorable diuretic profile) [92].

SGLT2i provides metabolic advantages, promotes weight loss, and enhances insulin sensitivity for improved cardiovascular health. It also exhibits intriguing effects, which include inducing a transcriptional paradigm, nutrient deprivation, hypoxia, increasing ketosis, erythropoietin, and autophagic flux [93]. Consequently, inflammasome activation is reduced, mitigating cardiomyocyte dysfunction and coronary microvascular injury. Moreover, alterations in iron homeostasis contribute to enhanced cardiac energetics and function. Additionally, SGLT2i reduces epicardial adipose tissue and modifies adipokine signaling, potentially contributing to the observed reductions in inflammation and oxidative stress associated with their use [93].

Dapagliflozin has a significant impact on energy metabolism related to fatty acid intake and mitochondrial dysfunction in HFpEF; it is achieved by elevating $\beta$-hydroxybutyric acid ($\beta$-OHB) levels, activating citrate synthase, reducing acetyl coenzyme A (acetyl-CoA) pools, regulating adenosine 5’-triphosphate production, and enhancing the expression of mitochondrial oxidative phosphorylation system complexes I–V. SGLT2i is beneficial in preventing and treating cardiac remodeling and dysfunction in HFpEF models by mitigating cardiometabolic dysregulation [94].

Endothelial dysfunction is a pivotal mechanism in HFpEF, diabetes mellitus (DM), and frailty. Treatment with the SGLT2i empagliflozin altered certain microRNAs, counteracting the changes observed in HFpEF patients. This suggests a potential restoration of endothelial function through empagliflozin treatment [95].

Chronic activation of the SGLT2i pathway may contribute to maladaptive cardiac remodeling. Inhibiting SGLT2 receptors could potentially mitigate adverse remodeling processes, improving cardiac structure and function [96].

SGLT2i can affect the extracellular matrix, contributing to collagen turnover and mitigating fibrosis—a significant feature in HFpEF. Notably, these inhibitors have demonstrated a capacity to reduce myofilament stiffness and remodel the extracellular matrix in the heart. This action improves diastolic function, offering potential benefits in the context of HFpEF [93].

Canagliflozin (CANA) treatment reduces myocardial hypertrophy and fibrosis and improves left ventricular diastolic function and remodeling. These positive effects are attributed to CANA’s ability to upregulate apelin, activate angiotensin-converting enzyme 2 (ACE2), and increase ACE2/Ang (1–7)/mast cell receptor (MASR) axis levels [97]. Canagliflozin treatment was found to counteract ferroptosis, a recently identified mechanism of iron-dependent non-apoptotic cell death in HF. In a rat model of HF induced by a high-salt diet, the study observed iron overloading and lipid peroxidation, both of which had been alleviated by administering canagliflozin [98]. SGLT2i mitigates the risk of hospitalization for HF in individuals with HFpEF. However, the specific hemodynamic mechanisms responsible for these benefits are poorly understood. In the CAMEO-DAPA trial, dapagliflozin treatment in HFpEF patients was associated with reduced resting and exercise pulmonary capillary wedge pressure (PCWP) and positive effects on plasma volume and body weight [99].

Emerging evidence also indicates that these drugs impact cardiomyocyte ionic homeostasis. Empagliflozin was observed to diminish the activity of the cardiac Na${}^{+}$/H${}^{+}$ exchanger, potentially enhancing cardiac function. Subsequently, it was discovered that dapagliflozin and canagliflozin also inhibited Na${}^{+}$/H${}^{+}$ exchanger activity, leading to a decrease in cytosolic Na${}^{+}$. Moreover, empagliflozin reduced the activity of Ca${}^{2+}$/calmodulin-dependent kinase II (CaMKII) and CaMKII-dependent sarcoplasmic reticulum Ca${}^{2+}$ leakage [100].

The effects of empagliflozin on HFpEF are primarily mediated by inhibiting Na${}^{+}$/H${}^{+}$ exchanger 1 (NHE1), influencing cardiomyocyte oxidative stress modulation, cardiomyocyte stiffness, myocardial extracellular matrix remodeling, heart concentric hypertrophy, and systemic inflammation [101]. Empagliflozin enhances the nitric oxide (NO)—soluble guanylate cyclase (sGC)—cGMP cascade and protein kinases GI$\alpha$ (PKGI$\alpha$) activity by reducing PKGI$\alpha$ oxidation in HFpEF [102]. Additionally, dapagliflozin inhibits the inflammatory response and activates the NO–cGMP–protein Kinases G (PKG) pathway in animal models [103].

Prior evidence showed that metabolites produced by gut microbiota play a crucial role in heart failure development. SGLT2i have been discovered to impact the gut microbiota in rodent studies. The EMPAGUM trial seeks to validate these human changes and explores the role of gut microbiota and their metabolites in the HFpEF process [104]. It is important to note that ongoing research further elucidates the mechanisms of SGLT2i in HFpEF, and the field continues to evolve (Fig. 4).

The EMPEROR-Preserved study study was carried out on 5988 patients (median age: 72, 45% women, median LVEF: 54%); half of the patients had diabetes, and half had an eGFR below 60 mL/min/1.73 m${}^2$. The primary endpoint (cardiovascular events and HF hospitalization) was met (hazard ratio (HR) 0.73; 95% confidence interval (CI), 0.6–0.88; p 0.001), marking the first instance where drug therapy achieved this goal in an HFpEF study.

The trial validated a 21% risk reduction in cardiovascular death or HF hospitalization from using empagliflozin on HF patients with an LVEF $>$40% [105]. Subgroup analysis has shown the highest benefit in LVEF 50% (HR 0.71; 95% CI, 0.57–0.88), a lower benefit in LVEF 50 to 60% (HR 0.80; 95% CI, 0.64–0.99), and no apparent benefit in LVEF $\ge$60% (HR 0.87; 95% CI, 0.69–1.10). Pooled analysis across EMPEROR-Reduced and EMPEROR-Preserved trials have indicated consistent HF outcomes with empagliflozin in patients with LVEF 25% to 65%, with diminished effects in those with LVEF $\ge$65% [106]. A potential drawback in the EMPEROR-Preserved trial was raised from the under-representation of females and a predominantly low NYHA class within the studied population. Additionally, there was no observed benefit of empagliflozin on overall mortality, as indicated by a hazard ratio of 1 (95% CI, 0.87–1.15).

In the Dapagliflozin Evaluation to Improve the LIVEs of Patients with Preserved Ejection Fraction Heart Failure trial of 6263 heart failure patients (mean age: 71, 44% female, mean LVEF: 54%), dapagliflozin led to an 18% risk reduction in the primary combined endpoint (HR 0.82; 95% CI, 0.73–0.92) and a non-significant 12% reduction in cardiovascular mortality (HR 0.88; 95% CI, 0.74–1.05) [107] over 2.3 years. Dapagliflozin lowered the risk of HF worsening or cardiovascular death by 18%, regardless of LVEF. Unlike prior trials, DELIVER included all LVEF ranges, but it is unclear if the benefit in patients with LVEF $>$60% specifically relates to a reduction in total CV events, HF events, or both.

DELIVER permitted randomization during or soon after hospitalization for HF in clinically stable patients without intravenous HF therapies. Starting dapagliflozin during, or soon after, HF hospitalization in patients with mildly reduced or preserved LVEF has appeared safe and effective; moreover, dapagliflozin has reduced the risk of worsening HF or cardiovascular death similarly in patients with and without a history of recent HF hospitalization [108]. In patients recently hospitalized with HF, initiating dapagliflozin has had trivial effects on blood pressure and has not worsened renal status [109].

In a meta-analysis [110] performed to investigate SGLT2i effects on HFpEF or HFmrEF, by pooling data from all clinical RCTs available and thus increasing power to testify, the authors have demonstrated that in patients with LVEF $>$40%, SGLT2i significantly reduces the composite risk of cardiovascular death and hospitalization for heart failure, although the risk of cardiovascular death and all-cause death did not reduce.

Type 2 diabetes mellitus (T2DM) and obesity could be considered metabolic disorders characterized by high CV risk [111].

Recently, a multinational, cross-sectional study of cardiovascular disease prevalence and etiology in adults with T2DM across 13 countries (CAPTURE study) showed that, among 9823 T2DM patients, one-third have a CV disease [112]. Moreover, the overall mean BMI was 29.0 kg/m${}^2$, highlighting a pathophysiological link between T2DM and obesity [112]. Importantly, according to U.S. National Health Interview Survey data, overall mortality in T2DM patients has reduced from 11.3% during 1988–1994 to 5.9% during 2010–2015. However, despite a paradigm shift in T2DM treatment beyond the hypoglycemic effect, the overall prevalence of CV complication is 32.2%, consisting mainly of CAD [113, 114, 115].

Obesity and T2DM are very common comorbidities in HFpEF patients and are closely involved in their pathophysiology and prognosis [116, 117]. T2DM-related metabolic derangements such as hyperglycemia, lipotoxicity, and hyperinsulinemia, associated with coronary microvascular rarefaction and advanced glycation end-products deposition, favor the development of DM-related cardiomyopathy (DMC) with HFpEF phenotype and concentric LV remodeling with diastolic LV dysfunction. Importantly, this phenotype is more prevalent in obese patients with T2DM [118]. Recently, a new paradigm of HFpEF syndrome has been developed: Particularly, HFpEF comorbidities are believed to induce a chronic systemic inflammatory state. In obese subjects, macrophages infiltrate adipose tissue, releasing proinflammatory cytokines. This condition promotes coronary microvascular endothelial inflammation, leading to a reduction in NO bioavailability. Reductions in both cGMP and protein kinase G activity increase wall tension, myocardial stiffness, and interstitial fibrosis [119]. These pathophysiological aspects suggest why innovative drugs targeting endothelial dysfunction are crucial in treating T2DM and obese patients and could be useful in diabetic and obese HFpEF phenotypes.

Incretin hormones—glucose-dependent insulinotropic polypeptide (GIP) and glucagon-like peptide 1 (GLP1)—are released from gut endocrine cells and potentiate meal-stimulated insulin secretion [120]. Biologically active GLP1 refers to both GLP1 (7–36) amide and GLP1 (7–37), which act on a single identified GLP1 receptor (GLP1R) on pancreatic islet $\beta$-cells, $\delta$-cells, and $\alpha$-cells, respectively, to increase insulin and somatostatin secretion and decrease glucagon secretion.

Synthetic GLP1R agonists (GLP1RAs) mediate the same biological effects of endogenous GLP1, binding to the GLP1R and stimulating glucose-dependent insulin release from the pancreatic islets.

A single, canonical GLP1R, expressed at low levels in the human atria and ventricles (including in cardiomyocytes) and in blood vessels, mediates the major CV actions of GLP1RAs, reducing blood pressure and improving microvascular and coronary flow, counteracting atherosclerosis, and promoting plaque stability [121]. Specifically, GLP1RAs might reduce myocardial apoptosis and inflammation in the heart, inducing glucose metabolism [120]. In addition, GLP1 and GLP1RAs have multiple extra-pancreatic effects that might indirectly reduce CV morbidity, inducing weight loss and acting on postprandial lipemia and inflammation [122]. Enterocytes and hepatocytes do not express GLP1R, although studies on mouse liver showed that a subset of GLP1R+ endothelial and intrahepatic $\gamma$$\delta$ T cells mediates a component of the anti-inflammatory effect of GLP1RAs [123]. As proof of this, patients with either T2DM, obesity, or non-alcoholic steatohepatitis (NASH) treated with semaglutide had shown a decrease in circulating triacylglycerol, low density lipoprotein cholesterol (LDLc), and non-high density lipoprotein (HDL) cholesterol [124]. Similarly, several GLP1RAs were shown to reduce postprandial plasma levels of triacylglycerol with little effect on fasting plasma LDLc, even in patients treated with statins [125]. The effect of a reduction in the fasting plasma levels of cholesterol and triacylglycerol could indirectly reflect the extent of weight loss achieved in these patients [119]. Weight loss is determined by reducing hunger by stimulating the GLP1Rs expressed in the central nervous system and slowing down gastric emptying [120].

The activity of GLP-1 and GIP is limited by the dipeptidyl peptidase-4 (DPP-4) enzyme, which rapidly inactivates incretin hormones. This provides the rationale for developing DPP4 inhibitors, which limit the breakdown of endogenous GLP-1 by the DPP-4 enzyme. Concentrations of the active intact GLP-1 and GIP are therefore increased, leading to an increased and prolonged action of these hormones. The biological effects of GLP1RAs are summarized in Fig. 5.

GLP1RAs and DPP4 inhibitors were developed for T2DM treatment based on this evidence. However, these effects may be particularly relevant in the context of metabolic alterations and inflammation of HFpEF syndrome. As proof, GLP1RAs were shown to have protective effects in experimental HF models in several species, including mice, rats, pigs, and dogs, further suggesting a role in specific HFpEF phenotypes [119, 126]. The same molecule used in experimental animal models of T2DM cardiomyopathy improved diastolic dysfunction, as reflected by a decreased E/e’ ratio [127].

Recently, a meta-analysis of eight RCTs about CV events, mortality, and kidney outcomes with GLP1RAs in 60,080 T2DM patients showed a significant reduction (–14%) in major cardiovascular events (MACEs), all-cause mortality (–12%), and composite kidney outcome (–21%) with no increase in the risk of severe hypoglycemia, retinopathy, or pancreatic adverse events. Moreover, GLP1RAs significantly reduced hospitalizations for HF by 11% [128]. Effects on HF-related events were different depending on the patients treated. For example, a post-hoc analysis from Harmony Outcomes aiming to explore the effects of the GLP1RA albiglutide on HF outcomes in T2DM patients and cardiovascular disease, with and without HF history, showed that albiglutide had no effect in reducing HF-related events among those with a history of HF [129]. Similarly, in a post-hoc analysis from the REWIND trial, which included $>$5.4 years of follow-up, dulaglutide was not associated with a reduction in HF events in patients with T2DM regardless of baseline HF status [130].

Since the effects of GLP1RAs are mainly related to anti-inflammatory and lipolytic effects, their administration is associated with weight loss regardless of diabetic status [131]. Obesity and, more specifically, increased epicardial fat in HFpEF are associated with more symptoms and worse prognoses [132, 133, 134]. This has led to weight loss as a specific target for HFpEF treatment and the realted role of GLP1RAs in this context. However, the effect of weight loss in HF patients is still partially unsettled. Weight loss is a poor prognostic factor in patients with HFrEF [135, 136]. Conversely, evidence suggested that weight loss has beneficial effects in obese patients with HFpEF: Kitzman et al. [137] showed that among clinically stable obese older HFpEF patients, aerobic exercise training or caloric restriction increased peak oxygen consumption, and drug effects were additive [138]. Bariatric surgery was also shown to improve symptoms, as well as reduce HF rehospitalizations and reverse left ventricular remodeling in obese HFpEF patients [137, 139].

However, the same beneficial effects of weight loss have not been replicated in non-obese patients with HFpEF: the FLAGSHIP cohort study aimed to examine the association between weight loss and HFpEF prognosis in 573 hospitalized obese and non-obese patients [140]. In particular, in non-obese patients, weight loss was associated with higher all-cause mortality and rehospitalization rates than ones without weight loss. Moreover, 6 months after hospital discharge, a high proportion of non-obese patients with weight loss showed functional limitations and anorexia, suggesting an impairment of their physical function and poor nutritional status. Conversely, weight loss was not associated with adverse events in obese patients with HFpEF [141].

Based on these results, much evidence suggests that GLP1RAs, promoting weight loss by reducing the generation of reactive oxygen species and systemic inflammation, may be highly effective in the obese phenotype of HFpEF. Among GLP1RAs, semaglutide was shown to be more effective in weight loss compared with other agents: In the SUSTAIN trials, a slightly greater weight loss was observed with subcutaneous once-weekly semaglutide administration, compared with exenatide, dulaglutide, or liraglutide [142]. Similarly, a secondary analysis of PIONEER 4 showed a greater weight loss with once-daily oral semaglutide administration than with subcutaneous liraglutide [142, 143, 144, 145].

The STEP HFpEF trial has recently shown a significant improvement in symptoms, quality of life, and exercise tolerance, assessed by the 6-minute walking test (6MWT), along with weight loss, in 529 non-diabetic obese patients with HFpEF randomized to subcutaneous semaglutide (2.4 mg weekly) or placebo [140]. This trial also showed a significant reduction in C-reactive protein (CRP) in the semaglutide arm compared with the placebo. Importantly, a prespecified trial sub-analysis showed that semaglutide improved symptoms, physical limitations, and exercise function and reduced inflammation and body weight across obesity categories. Furthermore, the magnitude of benefit was directly related to the extent of weight loss (adjusting for age, sex, body weight at baseline, and other confounding variables) [146]. Semaglutide also greatly improved HF-related symptoms, physical limitations, exercise function, and NT-proBNP regardless of baseline health status [147]. In the STEP-HFpEF, semaglutide was also shown to reduce NPs. This evidence suggests a direct hemodynamic mechanism for semaglutide, regardless of caloric and metabolic improvement [148].

The STEP HFpEF DM trial (NCT04916470), which recently completed enrolment, will test the safety and benefit of semaglutide in obesity-related HFpEF patients with T2DM and explore the interaction with SGLT2i, as their use was rare in STEP-HFpEF but not in STEP HFpEF DM (32%) [140] (Table 4, Ref. [146, 147, 148, 149, 150]).

Given the results derived from RCTs, international scientific societies currently recommend using GLP1RAs as part of a comprehensive strategy to reduce the risk of CV events in patients with T2DM [21], although they have yet to be recommended to prevent HF [151, 152].

More recently, results of a SELECT study demonstrated the superiority of semaglutide when added to a typical standard of care, compared with placebo, in overweight or obese patients without T2DM, in reducing the incidence of death from cardiovascular causes, nonfatal myocardial infarction, or nonfatal stroke at a mean follow-up of 39.8 months [153]. Semaglutide also showed a non-significant reduction in the heart failure composite endpoint (HR 0.82 [0.71–0.96]).

New molecules are currently being developed and tested. For example, tirzepatide is a dual agonist of glucose-dependent insulinotropic polypeptide and GLP1R, thereby constituting a novel treatment option for T2DM. This agent exerts additional effects in addition to improvement in glycemic control, which can benefit individuals with T2DM, especially those at risk for or with established CV disease or HF. However, current evidence is limited, although it is suggestive of the cardiovascular safety of tirzepatide [150]. SUMMIT is an ongoing RCT that will assess the efficacy and safety of tirzepatide (LY3298176), compared with the placebo, in participants with the obesity phenotype of HFpEF [154, 155, 156, 157, 158, 159, 160, 161, 162].

Vericiguat is a drug that stimulates the cGMP pathway through direct and indirect stimulation of soluble guanylate cyclase. The current mechanism increases nitric oxide synthase (eNOS) at the endothelial level, directly affecting cardiac workload by reducing systemic resistance and improving vascular compliance [156, 157]. The NO availability enhancement leads to smooth muscle cell relaxation and reduces hypertrophy, inflammation, and fibrosis [158]. Two recent trials demonstrated a potential drug effect in HFpEF: The SOCRATES-PRESERVED study showed no significant effect on mortality and hospitalization in patients with worsening heart failure. However, using vericiguat was associated with better tolerance and quality of life after a 3-month follow-up period [159]. More recently, the VITALITY study did not confirm preliminary findings, showing no differences in quality of life and 6-minute walking distance score [160]. These contrasting data could depend on the high frailty burden of enrolled patients and the short observational period [161].

Even though recent post-hoc analyses revealed that all drugs endorsed for patients with reduced ejection fraction have positive effects in subjects with HFpEF up to the cutoff 60%, there are still doubts regarding the effective benefits of all agents [162]. Presently, only SGLT2i have demonstrated a comparable effect in HFrEF and HFpEF, but the remaining data are based on putative analyses, surrogate endpoints, and retrospective data. Since HFpEF encompasses several subtypes with different risk profiles, patients’ frailty, pathophysiological mechanisms, and hemodynamic cardiovascular disorders, a tailored management algorithm needs to be precise according to these features. In particular, there are two settings in which therapeutic advances demonstrated significant improvement: inflammatory and metabolic pathways can be regulated by treatment, resulting in improved functional status, reduced HF hospitalization, and lower CV mortality.

In the last 5 years, additional putative analyses and post-hoc investigations in certain HFpEF phenotypes have indicated that traditional treatment used for HFrEF can be extended to individuals affected by HFpEF. The efficacy of various treatments helped to understand the pathophysiological mechanisms behind the development and maintenance of the HFpEF condition, resulting in the identification of new therapeutic targets. However, more research is needed to understand how these agents influence the natural history of specific HFpEF phenotypes. This is extremely important given the high frequency and poor prognosis of patients with HFpEF.

AP: conception and design; PS: interpretation of data; AD: acquisition of data; VM: analysis and interpretation of data; LT: substantial contribution to the study design; MC: conception and design of the work; GD: acquisition and analysis; FF: investigation and analysis; SN: study conception and reviewing. All authors contributed to editorial changes in the manuscript. All authors read and approved the final manuscript. All authors have participated sufficiently in the work and agreed to be accountable for all aspects of the work.

Not applicable.

Not applicable.

This research received no external funding.

The authors declare no conflict of interest. Alberto Palazzuoli is serving as one of the Editorial Board members of this journal. We declare that Alberto Palazzuoli had no involvement in the peer review of this article and has no access to information regarding its peer review. Full responsibility for the editorial process for this article was delegated to Wengen Zhu.