William Harvey’s epochal discovery of blood circulation driven by a heart pump in 1628 provided a framework for the basic concept of heart failure (HF). Ancient Greek and Roman physicians attributed edema, anasarca, and dyspnea to a variety of causes other than heart disease; pleural effusions were thought to originate in the brain, and the heart was believed to heat and distribute the vital spirit. Linking cardiac pathology to hemodynamic and clinical disorders following Harvey’s discovery took time. G.M. Lancisi (1654-1720) observed that dilatation weakens the heart. J.N. Corvisart wrote in 1806 that marked cardiac dilatation in association with valvular regurgitation portends HF and a bad prognosis. In 1892, W. Osler viewed cardiac hypertrophy as a step towards HF (“broken compensation”). Those and many other physicians, whether famous or less known, elaborated the clinical presentation and course of HF (1–3).

Early pathophysiological concepts were simplistic, limited to cardiac weakness causing low cardiac output and extracellular fluid retention with pulmonary and systemic congestion. Treatment options were empirical and ineffective. Lifestyle changes provided some relief. Bloodletting and leeches were used for centuries. Southey tubes for edema drainage have been long forgotten. Among many herbal treatments that have been tried, only digitalis (foxglove) preparations proved to be a lucky hit. Since introduction to medical use by W. Withering in 1785, digitalis had been a pillar of HF treatment for about 200 years until recently (4–9).

The creation of effective diuretics in late the 1950s heralded the modern era of pharmacological HF management. Until 1957, the only diuretics used in HF were intravenous or intramuscular mercurial agents, which were difficult to use and fraught with toxicity. The serendipitous discovery of modern diuretics took place during the study of sulphonamide side-effects. In 1937/8, research on sulphonamides revealed their diuretic effects. In 1945, the development of the carbonic anhydrase inhibitor acetazolamide improved the understanding of diuretic mechanisms in renal tubules. Among many compounds synthesized in the search for potent carbonic anhydrase inhibitors, researchers stumbled on chlorothiazide, which inhibited not only carbonic anhydrase but also the sodium chloride cotransport system. Introduced in 1958, chlorothiazide was the first useful oral diuretic. It is still used widely, but much more as an antihypertensive agent than a HF drug (10–12).

Thiazides and thiazide-like diuretics may be useful in the management of mild HF, but more potent diuretics are needed in severe forms. Furosemide, with a brand-name Lasix (“lasts six hours”), came to the rescue. The introduction of the loop diuretics furosemide and ethacrynic acid in the early 1960s dramatically improved medical practice. Until then, HF had been considered a terminal condition, but new diuretics allowed amazing relief of fluid retention. It was obvious that they worked, and little in the way of clinical trials was needed to accept them in clinical practice. Furosemide, as an archetype of potent loop diuretics, was followed by now forgotten ethacrynic acid and bumetanide, in addition to its current rival torasemide, and has remained a cornerstone in the treatment of congestive HF (12–14).

Spironolactone, a nonselective steroid aldosterone receptor antagonist, appeared in 1957 (15). Its independent diuretic action is weak, but the synergism with loop diuretics is strong, in addition to potassium-sparing effects. The value of spironolactone became evident in the RALES trial (1999) which showed a reduction in morbidity and death among patients with severe HF (16). Those data suggested a paradigm-shift in the concept of HF, demonstrating the neurohormonal actions of spironolactone. The selective aldosterone receptor antagonist epleronone shares beneficial effects with spironolactone but is devoid of side-effects on sexual steroid hormones (17). Finally, the use of a novel nonsteroidal aldosterone receptor antagonist fineronone is advantageous in patients with chronic kidney disease and type 2 diabetes (18).

Diuretics were accepted as symptom-relieving and lifesaving drugs before the days of large clinical trials with mortality and survival endpoints. Even nowadays, the lack of alternative to loop diuretics makes it hard to envision how to design such trials (19, 20).

With digoxin, furosemide, and spironolactone available, the stage was set for a dynamic era of HF management. Past stagnation and incapacity gave way to active pursuit of advances in HF treatment that dominated in the decades to follow. This pursuit rekindled research on HF pathophysiology (21). Research requires concepts that representing the basic models of HF. Those have changed significantly since the middle of the 20^th century. Senior cardiologists may recall them from their own professional experience. The formulation of those models may vary somewhat, but some of them, e.g. the neurohormonal model, have been widely accepted. Milton Packer has proposed four concepts: 1) the cardiorenal model (1940s through the 1960s), 2) the cardiocirculatory model or hemodynamic hypothesis (1970s and 1980s), 3) the neurohormonal model (from 1990s up to recently), and 4) the recent cellular stress model (22–24). Such a conceptual framework is interesting as it may explain many advancements but also some side-tracks. We will try to explicate the concepts, adding some remarks.

The cardiorenal model regarded HF as an edematous disorder with salt and water retention. The pillars of treatment were digitalis and diuretics. Although the congestion seen to be due mostly to increased venous pressures (“backward HF”) dominated the concept, low cardiac output (“forward HF”) was duly appreciated. Moreover, renal hypoperfusion due to low cardiac output was deemed essential for extracellular fluid retention by activating renal (glomerular, tubular, and peritubular) mechanisms of volume conservation. These may be viewed as atavistic responses to hypovolemia due to fluid loss. In due course, the concept was expanded by adrenal and neurohormonal (e.g. antidiuretic hormone secretion) mechanisms of volume retention with dilutional hyponatremia as an epiphenomenon. Late additions were synthesis with the neurohormonal HF concept. The erstwhile cardiorenal concept views HF as a cardiac disorder with anticipatable renal response (without disease), at variance with the modern concept of cardiorenal syndrome where the disease of any of those two organs induces the disorder of the other (21, 22).

The central tenet of the erstwhile cardiorenal concept assumes the main role of the heart disorder in the pathogenesis of HF, with the kidneys playing supporting roles. The widespread use of digitalis preparations (mostly digoxin) was intended to alleviate cardiac disorder as the origin of HF. The evaluation of efficacy and clinical utility was based on clinical observations, experience, and judgement; only later were some respectable randomized trials conducted, with neutral results on mortality (25–29). The weak positive inotropy of digitalis, due to the inhibition of Na⁺/K⁺-ATPase in cardiac myocytes with a consequent increase in cytosolic Ca²⁺ content through Na⁺/Ca²⁺ exchanger, was overrated. Negative chronotropy, desirable in tachycardia due to atrial fibrillation, was limited because of narrow therapeutic width. Despite limitations, digitalis preparations had remained the mainstay of HF treatment for two centuries (6). Perhaps M. Packer, when giving the European Society of Cardiology (ESC) Rene Laennec Lecture on Clinical Cardiology at the ESC Congress 2023, dated the cardiorenal model to the 1748-1965 period correctly; digitalis use dates back that far (24).

Potent diuretics became available not earlier than in the 1960s, when this period was at an end. The diuretics revolutionized HF treatment and have remained its mainstay irrespectively of the overall concept. They are needed whenever an excess of extracellular fluid arises (12, 14).

Finally, sticking to the cardiorenal model impeded advances in HF treatment. Lowering arterial pressure was deemed risky because of renal and myocardial hypoperfusion. The use of vasodilators was avoided except in severe hypertension. Similarly, the use of beta-blockers (BB) was restrained (propranolol appeared in the 1960s as an antianginal drug), fearing not only arterial hypotension but also negative inotropy (22, 30, 31).

The cardiocirculatory model or the hemodynamic hypothesis (1965-1992), as it was formulated by M. Packer, followed the cardiorenal model. The quest for new treatment strategies spurred by the unmet expectations of digoxin and diuretics led to reappraisal of pathophysiological concepts. The ensuing paradigm shift viewed HF principally as a hemodynamic disorder involving the whole cardiovascular system, not only the heart but also the vessels. The focus shifted from the kidneys to peripheral vessels and redistribution of intravascular volumes (22).

In HF, adrenergically mediated arterial and venous vasoconstriction devastatingly impairs hemodynamics. The resulting increase in afterload exhausts the myocardium and impedes cardiac output. Reduction of the venous reservoir increases cardiac preload and aggravates congestion. The rationale for the use of vasodilating drugs was to relieve the failing heart of preload and afterload burden, expecting a recovery in cardiac function (21, 22). The whole array of vasodilators was tested intravenously or orally to relieve vasoconstriction and to achieve a breakthrough in HF management. Senior cardiologists may recall the fear of causing hypotension with those unorthodox innovations (30, 31). The use of some initially promising vasodilators, e.g. the adrenergic alpha-receptor blockers phentolamine and prazosin, remained a dead-end attempt (32–36). Hydralazine and isosorbide dinitrate held some promise, which was not fulfilled (36–38). Calcium channel blockers were expected to improve hemodynamics, but controlled clinical trials showed worsening of HF (39–43). At best, oral vasodilators provided some temporary relief without long-term benefits on outcomes. Intravenously administered vasoactive agents for acute HF fared better, especially if guided by hemodynamic monitoring. Intravenous nitroglycerine, acting mostly as a venous vasodilator, is still valuable in acute cardiogenic pulmonary edema, while combined arterial and venous vasodilator sodium nitroprusside is helpful in severe acute HF in a critical care setting (44–47).

Disappointment with oral vasodilators shifted the focus to the heart itself. As the impairment of myocardial contractility was deemed to be the primary problem, the whole array of inotropic agents was created. The basic approach was to increase contractility, stimulating the influx of calcium ions or maintaining higher calcium levels in the cytosol of cardiac myocytes throughout the action potential. Dozens of such drugs progressed to phase 3 clinical trials. Among them, dobutamine, amrinone, milrinone, enoximone, levosimendan, pimobendane, and xamoterol were the most promising. Dobutamine was the first of them, created in 1975 by modifying the chemical structure of isoproterenol. The rationale was to benefit from adrenergic stimulation but to avoid detrimental vasoconstriction. A series of phosphodiesterase 3 (PDE-3) inhibitors followed, amrinone first, followed by milrinone and enoximone. Calcium sensitizers were developed in the 1980s, with levosimendan as a prototype. It increases contractile apparatus sensitivity to calcium ions during systole, not interfering with their diastolic release. Levosimendan also inhibits PDE-3 and activates ATP sensitive K⁺ channels, causing strong vasodilation. Pimobendan shares calcium sensitizing and PDE-3 inhibiting effects. Xamoterol is a beta1 selective partial adrenergic agonist. The ideal was to develop a positive inotropic agent with vasodilator properties, like levosimendan, or a vasodilator with positive inotropic properties, like flosequinan. This drug has both vasodilating and inotropic properties that are not entirely understood but are believed to be distinct from β-adrenergic receptor agonists and PDE inhibitors (48).

Those and other positive inotropic agents, irrespective of mode of action, provided a transient hemodynamic and clinical improvement but failed to reduce mortality and morbidity. Moreover, they were associated with increased mortality, except the prognostically neutral digoxin (25, 48, 49). Many inotropic drugs were created, but only those designed for short-term intravenous use (e.g. 48 h) in severe acute HF remained in use. Dobutamine, milrinone, and levosimendan are still indispensable in our cardiac care units (50–53).

The failure of inotropic agents to convert short-term improvements into long-term benefits, with an excess of mortality, can be explained only speculatively. A mechanistic explanation suggests that overstrain of an injured organ shortens its lifespan. Overstimulating sick myocardium unduly depletes its meagre stores of energy. The persistent adrenergic overstimulation is detrimental, whereas BBs reduce mortality and morbidity. In addition, the chronic use of drugs acting via cAMP modulation, like PDE-3 inhibitors and adrenergic stimulants, disrupts calcium homeostasis with desensitization of the contractile apparatus to calcium through impairment of early diastolic relaxation and ventricular arrhythmias. Disruption of cardiac myocyte energetics and calcium ion homeostasis are detrimental for cardiac myocytes. Furthermore, HF is a heterogeneous syndrome comprising a diverse spectrum of diseases. The harmful effects of chronic treatment with milrinone were conspicuously more prevalent in ischemic than in non-ischemic HF. The doses of inotropic agents used were perhaps too high, adjusted to achieve maximal immediate inotropic effect, notwithstanding later consequences. The two-century-long history of digitalis parallels the positive inotropes of the 1980s in overuse and overdosage deviations. More judicious and selective use of those inotropes may have yielded different end-results (48, 49, 54).

Even if the attempts to treat HF with inotropes failed, the concept itself may not be doomed to failure. Strengthening the weak heart was the primordial aim which appeared self-evident for generations of physicians and was incorporated in wishful thinking on digitalis inotropy. Positive inotropes are still expected to have significant roles in chronic HF treatment, but as part of an elaborate framework of concepts and not as a blunt overstimulation of already exhausted cardiac myocytes. A hiatus in the development of positive inotropes was followed by new cellular targets in the 2000s. Pharmacological and gene therapy approaches were directed at a key enzyme responsible for myocardial calcium homeostasis that is downregulated in HF: sarcoplasmic reticulum Ca²⁺-ATPase (SERCA2a). Another concept is related to cardiac myosin activators, which are a new class of myotropes that improve myocardial function by directly augmenting cardiac sarcomere function. Omecamtiv mecarbil, the first of this class, augments cardiac contractility by selectively binding to cardiac myosin, thus increasing the number of myosin heads that can bind to the actin filament and initiate a power stroke at the start of systole (48, 54). The GALACTIC-HF trial (2021) showed that omecamtiv mecarbil reduced the incidence of a composite of a heart failure events and death from cardiovascular causes among patients with HF with reduced ejection fraction (HFrEF) (55, 56). However, FDA has declined to approve omecamtiv mecarbil, citing a lack of evidence on efficacy in 2023. No positive inotrope is currently approved for long-term use in HF.

M. Packer, who formulated the hemodynamic model of HF in his original article (1993), did not pay much attention to the recognition of diastolic left ventricular dysfunction as the cause of HF. Though the pressure-volume relations during the cardiac cycle had already been recognized in the early 20^th century and the term left ventricular lusitropy, denoting the rate of early diastolic relaxation, predated clinical concepts, clinical research did not recognize the concept of left ventricular diastolic dysfunction until the 1970s and clinical practice did not use it until the 1980s (57–66). This was a real paradigm shift, since until then only the impairment of left ventricular systolic function (i.e. contractility) was regarded as a cause of HF. It was only then that cardiologists realized that left ventricular systolic function was preserved in 40-50% of HF cases. It was assumed that diastolic dysfunction was the main culprit, accounting for at least 30% of all HF cases (62, 67–69).

This led to the confusing question of how to treat those patients. Use of positive inotropic agents seemed senseless and diuretics necessary; the excess of fluid should be removed in any case. It was advised to do this cautiously, since the Frank Starling curve was supposed to be steep and shifted to the right. Therefore, a sudden contraction of intravascular volume could precipitate a sudden drop in stroke volume and cardiac output. Many small clinical studies and later some landmark hypertension trials showed that renin-angiotensin-aldosterone system (RAAS) antagonists may improve left ventricular diastolic function in parallel with left ventricular hypertrophy regression better than BBs. The latter may yet provide beneficial effects on diastolic function by prolonging diastole. Calcium channel blockers proved to be controversial. The evidence on efficacy of those treatments in HF was only circumstantial, limited to surrogate hemodynamic data without the data on morbidity and mortality outcomes (66, 70, 71).

The uncertainties about the treatment of HF due to the left ventricular diastolic dysfunction have been never resolved by clinical trials since the hemodynamic hypothesis fell into disrepute. The assessment of left ventricular diastolic dysfunction in clinical practice is often indeterminate or elusive (72, 73). The conceptual advancements made clear that HF in patients with preserved systolic function cannot be simplistically reduced to diastolic dysfunction. The role of left ventricular diastolic dysfunction in the pathophysiology of HF has not been ignored but has instead been incorporated into broader and more comprehensive concepts of HF with preserved systolic function, normal ejection fraction (HFNEF), and preserved ejection fraction (HFpEF) (74–79). These concepts outgrew the hemodynamically frame (77–81).

The neurohormonal hypothesis (1992-2019), as it was named, formulated, and dated by M. Packer, presented a radical paradigm-shift and a conceptual breakthrough which viewed HF as a systemic disorder involving a complex neurohormonal response with renin-angiotensin-aldosterone (RAAS) and adrenergic systems in protagonist roles (21–23). Senior cardiologists may remember that this approach emerged gradually, not because of ingenious thinking which revised pathophysiological concepts but rather arising empirically through studies on vasodilators. The first angiotensin converting enzyme inhibitor (ACEi) captopril, first isolated from snake venom, introduced in 1981 and initially considered a plain vasodilator useful in arterial hypertension, proved to be superior to prazosin, hydralazine, and other pure vasodilators in the treatment of chronic HF. The advantage was explained by the blockade of detrimental neuroendocrine responses (34–36, 82). Following captopril, many other ACEis with specific qualities were developed. They provided a quantum leap in the treatments across the cardiovascular continuum. A series of landmark clinical trials evaluating the treatment of chronic HF with impaired left ventricular systolic function demonstrated that ACEis provided not only short-term hemodynamic improvements but also long-term benefits to morbidity and mortality. The most cited game-changing trials were CONSENSUS I (enalapril, 1987), VHeFT I (enalapril, 1991), SOLVD (enalapril, 1991), SAVE (captopril after myocardial infarction, 1992), AIRE (ramipril after myocardial infarction, 1993), and TRACE (trandolapril, after myocardial infarction, 1995) (83–89).

An unprecedented breakthrough in HF trials with ACEs reinvigorated experimental scientific research on RAAS to explain those results and chart further advances. It became clear that besides of the endocrine component of RAAS, there was also a widespread tissue RAAS, with the heart, vessels, nervous system, and the kidneys as the main players. Endocrine (in blood), paracrine (in the tissues), and intracrine (in the cells), signaling was identified in addition to autocrine and juxtacrine RAAS. RAAS was recognized as a ubiquitous system for homeostasis and pathologies, biologically fundamental, with deep evolutionary roots and composed of ancient molecules. Its pivotal molecule, angiotensin II, an octapeptide with strong vasoconstrictive properties, arose in the early Cambrian ~500 million years ago, primarily as an epigenetic regulator of protein synthesis and growth-promoting factor. It is a key molecule in the signaling pathway of pathological myocardial hypertrophy and a potent promotor of atherogenesis (90–94).

Until the 90s, positive inotropic agents (mainly digoxin) were deemed a mainstay of HF treatment. As the poor left ventricular systolic function was viewed as the main cause of HF, all drugs with negative inotropic effect were “absolutely contraindicated” according to the practice guidelines up to 1995 (95). The idea of using BBs as a primary therapy for congestive HF to improve symptoms and prognosis seemed paradoxical and dissenting. The cardiac community reacted with skepticism and disbelief when, in 1975, the pioneering report of Waagstein et al. gave an account on 7 cases of refractory HF in patients with dilated cardiomyopathy treated successfully by already forgotten BBs alprenolol and practolol (96, 97). This unorthodox approach may be viewed as a bailout in a desperate situation. Waagstein violated a taboo and stirred controversy. A change of opinion took time. Many small studies with surrogate endpoints only added to confusion. Decades were needed for scientific research to recognize the detrimental effects of maladaptive adrenergic response to declining systolic function and vindicate large controlled clinical trials (98). The methodology was honed in HF trials with ACEs. Landmark clinical trials convincingly demonstrated the efficacy of four BBs in improving morbidity and reducing mortality in the patients with “systolic” HF. Pioneering trials were: CIBIS I (bisoprolol, 1994), CIBIS II (bisoprolol 1999), US Carvedilol HF Trials Program (1996, carvedilol), and MERIT-HF (metoprolol, 2000) (99–103). Other important trials with bisoprolol, carvedilol, and metoprolol followed, while the SENIORS trial affirming nebivolol took place slightly later (2005) (104).

While the ACEIs were introduced as vasodilators in the treatment of HF to improve hemodynamics and were only later recognized as neurohormonal agents, BBs were introduced primarily as neurohormonal agents. When M. Packer proposed his neurohormonal model of HF pathophysiology in 1993, BBs were only envisaged, but neither evaluated in clinical trials nor approved for the treatment of HF (23). The trials opened the door to the extensive use of BBs in the treatment of “systolic” HF and firmly established the neurohormonal concept. ACEIs and BBs took pole position in the guidelines for HF treatment on both sides of the Atlantic, sharing the first two positions (46, 105).

The neurohormonal concept led to reappraisal of spironolactone. After 40 years spent in a modest role as an adjunct potassium sparing diuretic, following the RAALES trial (1999) spironolactone was revisited as an essential neurohormonal drug which reduces morbidity and mortality in patients with HF (16). Considering the key role of aldosterone receptors in the pathogenesis of cardiovascular and renal pathologies, the results of RAALES trial might have been expected, but the clear reduction in all-cause mortality sent a clear message. The EPHESUS trial (2003) showed that the selective aldosterone receptor blocker eplerenone reduced mortality and morbidity in patients with acute myocardial infarction complicated by HF (17). This evidence positioned mineralocorticoid receptor antagonists (MRAs) firmly in third place on the list of preferred HF drugs, after ACEIs and BBs, while digoxin was relegated to the low fourth position (106, 107).

In 1992, research on angiotensin II led to losartan, an angiotensin II type 1 receptor antagonist with antihypertensive properties (108). It was a forerunner of the whole class of angiotensin receptor blockers (ARBs) which have been used as an alternative to ACEs ever since. Candesartan, eprosartan, irbesartan, valsartan, telmisartan, and olmesartan followed. Affirmed as an antihypertensive drug with heart and kidney protecting properties (RENAAL 2001; LIFE 2002), losartan was compared to captopril in the treatment of systolic HF ((ELITE I 1997; ELITE II, 2000) and acute myocardial infarction (OPTIMAAL 2002), showing non-inferiority with better tolerability (70, 109–113). The trials with newer ARBs upheld the message: ARBs are a noninferior alternative to ACEs in left ventricular systolic failure. Only valsartan (2001), candesartan (CHARM-Alternative 2003, CHARM-Added, 2023), losartan again (HEAAL, 2012), and to a lesser extent telmisartan (1999, 2010) were evaluated in “systolic” HF trials, while candesartan (CHARM-Preserved, 2003) and irbesartan (2008) were appraised in HF with preserved systolic function trials (114–121). European HF guidelines (2021) do not claim any preference for ACEs or ARBs, recommending them both as an alternative (IA). American guidelines (2022) prefer ACEs as the first choice in RAAS-naive patients (46, 105).

The use of ARBs in the treatment of HF with reduced ejection fraction (HFrEF) has been upgraded recently by neprilysin inhibition which fits neatly with the neurohormonal concept of HF (122). In PARADIGM-HF Trial (2014) the dual-acting angiotensin receptor-neprilysin inhibitor (ARNI) sacubitril valsartan reduced the composite endpoint of HF hospitalization and death in comparison with standard enalapril treatment (123, 124). ARNI therapy now has a class I indication for the treatment of patients with HFrEF (46, 105, 125).

The evolutionarily highly conserved family of natriuretic peptides comprises the atrial, brain, and C-type peptides (ANP, BNP, and CNP). ANP and BNP, secreted by the atria and ventricles, operate via the natriuretic peptide receptors type A (NPR-A) and type B (NPR-B), which are coupled to guanyl cyclase, mediating biological effects. Those include vasodilatation, natriuresis, and diuresis, inhibition of the RAAS, endothelin, and vasopressin, along with lipid mobilization. ANP is degraded rapidly by endopeptidase neprilysin. The inhibition of neprilysin increases the levels of ANP in circulation, with beneficial effects in HF (122, 126, 127).

Mechanistic reasoning may turn out to be overly simplistic when faced with the complexity of clinical medicine. Promising concepts may not work in clinical practice. Clinical trials may dash hopes placed in promising treatments. Nesiritide, a human recombinant B-type natriuretic peptide (BNP) with vasodilatory properties, binds to receptors in the vasculature, kidney, and other organs to mimic the actions of endogenous natriuretic peptides. It was approved by the FDA in 2001 for use in patients with acute HF based on studies showing hemodynamic and symptomatic improvements. A few years after its approval (2005), nesiritide fell out of use because small studies seemed to indicate an increased risk of kidney problems and an increased death rate. The ASCEND-HF study (2011) showed no impact of nesiritide on death or HF hospitalization (128–132).

Elevated resting heart rate has been linked to poor outcomes in patients with chronic systolic HF. Ivabradine may be added to neurohormonal treatments as the adjunctive therapy for HF with reduced ejection fraction (HFrEF) to slow sinus rhythm. It inhibits “funny” pacemaker current (I_f) of the sinoatrial node, not affecting the AV node, inotropy, diastolic function, cardiac output, vascular resistance, or blood pressure (133, 134). Ivabradine should be considered (IIa indication) in the patients with LVEF ≤35% in sinus rhythm and with a resting heart rate ≥70 bpm who remain symptomatic despite optimally up-titrated BBs, ACEi/ARNI, and MRA based treatment (or in BB intolerant patients) to reduce the risk of HF hospitalization and cardiovascular death (46). Ivabradine is not a substitute for BBs (133).

Neurohormonal activation is the crucial mechanism underlying the progression of HF, and therapeutic antagonism of neurohormonal systems has become the cornerstone of pharmacotherapy for HF. The perception of HF has changed: it is no longer regarded as a terminal syndrome with a dismal prognosis but as a treatable disorder (92).

However, it was not possible to ignore the fact that neurohormonal inhibition did not work in patients with HFpEF, representing at least a half of all patients with HF. Their share has risen to >50%, owing mostly to the aging of the population. The typical phenotype are obese elderly women with a small, well-contracting left ventricle, diabetes, hypertension, and atrial fibrillation (77, 135). Diastolic left ventricular dysfunction is a risk factor but not the main cause of HF. Landmark HF studies with RAAS antagonists and BBs excluded patients with LV EF ≥40%. Trials with candesartan, irbesartan, and spironolactone, designed to explicitly address the efficacy of RAAS antagonists in patients with EF ≥40%, yielded disappointing results. The CHARM-Preserved trial with candesartan (2003) showed only a moderate reduction in hospital admissions among patients with HF with LVEF >40%. The I-PRESERVE trial (2008) with irbesartan was neutral about outcomes in patients with HF with LVEF ≥45%. The TOPCAT trial (2014) demonstrated that in patients with HF and LVEF ≥45%, the treatment with spironolactone did not significantly reduce the incidence of the primary composite outcome of cardiovascular death, aborted cardiac arrest, or hospitalisation (120, 121, 136).

Failure of neurohormonal inhibition to improve HFpEF took the cardiac community aback since the prognosis of HFpEF may be as grave as that of HFrEF (137). Cardiologists were at a loss since a huge population of patients with HFpEF was left without any treatment strategy. Judicious use of diuretics for decongestion was the only treatment remaining, along with the treatment of comorbidities (138, 139). HFpEF is a phenotypically and pathophysiologically heterogeneous disorder in which therapy should target the underlying phenotypes, etiologies, and comorbidities, but such an approach was frustratingly complex (140). Relying on speculative treatment of LV diastolic dysfunction based on hemodynamics did not help. New treatment strategies were needed.

The cut-off point for HFpEF of 40-45% was set to distinguish HFpEF from HFrEF of earlier HF trials. However, trials like CHARM-Preserved and TOPCAT indicated the presence of a transitional EF range from 40% to 49% where the beneficial effects of candesartan and spironolactone were better than with EF ≥50%, albeit not as good as with <40%. The term HF mid-range EF (40-49%) was introduced (≈2014) and soon (2021) renamed to HF with mildly reduced EF (41-49%) or HFmrEF. It is an intermediate category between HFrEF and HFpEF with an estimated prevalence of 10% to 20% among all patients with HF. The 2021 ESC guidelines on HF stated that “although no strong recommendations for this HF phenotypes can be made ACEIs, ARBs, MRAs, BBs, and ARNI can be considered to reduce the risk of death and HF hospitalization in HFmEF (IIb)”. In other words, pharmacological neurohormonal inhibition may be beneficial in patients with HFmEF, but not as much as in patients with HFrEF, since the evidence is far weaker (46, 139–143). According to the 2023 ESC HF guidelines update, neurohormonal inhibitors comprising ACEI/ARNI/ARBs, BBs, and MRAs may be considered (class IIb recommendation) in patients with HFmrEF, while diuretics for fluid retention and dapagliflozin/empagliflozin are indicated with class I recommendation (125).

Two peculiar new non-neurohormonal drugs have raised hopes for the treatment of patients with HFrEF in addition to the standard neurohormonal inhibition. Vericiguat works via stimulation of soluble guanylate cyclase (sGC) in the NO-sGC-cGMP pathway, with resulting improvements in myocardial function and vasodilation (144). The cardiac myosin activator omecamtiv mecarbil, which potentiates the effects of myosin on actin, exerts positive inotropic effects. Both drugs reduced the composite endpoint of cardiovascular death and hospitalization in HFrEF trials, vericiguat in VICTORIA and omecamtiv mecarbil in GALACTIC-HF (55, 56, 145–147).

The cellular stress hypothesis (2019-) is the final concept of HF pathophysiology proposed by M. Packer. It filled the substantial gaps in the understanding of HF left by the neurohormonal and the previous hypotheses. It has been inspired by the “game changer” role of SGLT2 inhibitors (SGLT2is). The essence is that cellular dysfunction perpetuates chronic HF. Once again, chance observations, clinical practice, and trials led the way.

Sodium-glucose cotransporters SGLT1 and SGLT2 are mediators of epithelial glucose transport. While SGLT1 accounts for most of the dietary glucose uptake in the intestine, SGLT2 is accountable for the majority of glucose reuptake in the tubular system of the kidney. The medications that inhibit SGLT2 suffix with flozins. The prototype phlorizin was identified in root bark from trees as early as in 1835. Although phlorizin did not show any obvious medicinal value, its blood glucose-lowering and glucosuric effects were described as early as 1886. Only recently (2012) was dapagliflozin introduced as antidiabetic drug. SGLT2is modulate the sodium-glucose cotransporter on the nephron, inhibiting glucose reuptake, inducing glucosuria, and lowering the serum glucose levels. In a way, SGLT2is act as diuretics stimulating osmotic diuresis along with glucosuria and natriuresis. Compared with placebo, SGLT2is reduce HbA1c levels by an average of 0.5-0.8% when used as monotherapy or add-on therapy (along with a modest weight loss). As antidiabetic agents, which was the primary indication, SGLT2is are modestly effective, but they proved to be exceptional in cardiac and renal protection (148–151).

The stir caused by rosiglitazone, which was found to increase cardiovascular mortality in post-marketing surveillance, required that new antidiabetic drugs undergo cardiovascular outcome trials prior to FDA approval. This led to the EMPA-REG OUTCOME empagliflozin trial (2015) which unexpectedly revealed a significant reduction in primary composite outcomes of cardiovascular death, nonfatal myocardial infarction, and stroke in the treatment group (152). The serendipitous discovery led to several landmark confirmatory trials on cardiovascular outcomes including CANVAS (canagliflozin, 2017) and DECLARE-TIMI (dapagliflozin, 2019), with positive results, indicating that the favorable outcomes were attributable to a class effect of SGLT2 receptor inhibition. The risk of HF hospitalization was reduced in patients with and without HF history. As the benefits were observed already within 2 months, they were not attributed to the regulation of diabetes but to independent effects of SGLT2i (153, 154). Moreover, marked beneficial effects on chronic kidney disease were also observed (155–157).

Four landmark trials (two for HFrEF and two for HFpEF) set the standard for HF treatment, establishing the strongest (IA) recommendations for SGLT2is use across the whole EF range (HFrEF, HFmrEF, HFpEF) in HF guidelines (2022) (105, 125). Those trials were: DAPA-HF (dapagliflozin 2019), EMPEROR REDUCED (empagliflozin 2020), EMPEROR PRESERVED (empagliflozin 2021), and DELIVER (dapagliflozin 2022). The results for dapagliflozin and empagliflozin were highly congruent: both reduced the composite endpoint of cardiovascular death and HF hospitalization by nearly 20% (158–161).

The chance discovery which promoted the modest class of antidiabetic agents to a game changer in HF treatment stirred excitement but also took the cardiac community aback. The mechanisms behind the phenomenon were shrouded in mystery. The upheaval among clinical cardiologists stimulated a huge research effort to provide scientific explanations for the mechanisms of action. Experimental research proposed many explanations, mostly related to cytoprotection. SGLT2is exert cytoprotective effects on the failing heart via SGLT2-independent pathways to increase nutrient-deprivation signaling and autophagic flux, thus reducing cellular stress, improving mitochondrial vitality, and suppressing inflammatory signaling and apoptosis (162).

In the surge of data on the cellular effects of SGLT2, it is hardly possible to concisely outline a unifying concept. Perhaps closest to it is the hypothesis that SGLT2s provide cardiac and renal protection by inducing a state of fasting mimicry through the activation of low-energy sensors, which is not mediated through the SGLT2 protein. This state activates SIRT1/AMPK and suppresses Akt/mTOR signaling, which lead to a reduction in oxidative stress, normalized mitochondrial structure and function, suppression of inflammation, minimization of coronary microvascular injury, enhanced contractile performance, and myocardial protection. SGLT2is promote autophagy, independent of their effects on glucose. SGLT2is might enhance ATP and hemoglobin production by expanding the pool of reactive cytosolic Fe²⁺. This is due to SGLT2i-induced decline in hepcidin and ferritin levels. which alleviates functional iron deficit mediated by the HF inflammatory milieu (156, 162, 163).

In the failing heart, glucose transporter type 1 (GLUT1) levels are upregulated, along with excessive glycolysis and defective glucose oxidation, causing cytosolic accumulation of injurious glucose intermediates that activate mTOR and suppress nutrient-deprivation signaling. The uptake of long-chain fatty acids rises, but their oxidation is defective, impairing ATP production and causing cytosolic accumulation of toxic lipid intermediates, worsened by mitochondrial and nutrient-deprivation signaling dysfunction. The ensuing cytosolic accumulation of amino acids activates mTOR. SGLT2is cure the abnormalities in glucose, long-chain fatty acid, and amino acid metabolism by inhibiting GLUT1, stimulating nutrient-deprivation signaling, and restoring mitochondrial vitality. This improves nutrient oxidation and oxidative phosphorylation, preventing cytosolic accumulation of harmful glucose and lipid by-products (162, 163).

With regard to HF, fluid accumulation pattern is critical. SGLT2is may differentially regulate the interstitial vs. intravascular compartment when compared with loop diuretics. In congestive HF, interstitial edema is the hallmark of disease. SGLT2 inhibitors may selectively reduce interstitial volume with minimal change in blood volume, whereas loop diuretics reduce both interstitial and intravascular volume. It has been assumed that this differential volume regulation by SGLT2is (interstitial > intravascular) may limit the aberrant reflex neurohormonal stimulation induced by intravascular volume depletion (153).

The crucial question arises how SGLT2is work in HFpEF. The answer is complex, reflecting the heterogeneity of HFpEF phenotypes and the intricacy of SGLT2i actions. Many pieces of this jigsaw puzzle have been put together, noting the considerable overlapping between HFpEF and HFrEF. SGLT2is have been shown to induce a nutrient-deprivation and hypoxic-like transcriptional paradigm, with increased ketosis, erythropoietin, and autophagic flux in addition to altering iron homeostasis, which may improve cardiac energetics and function. These agents also reduce epicardial adipose tissue and modify adipokine signaling attenuating inflammation and oxidative stress. SGLT2is may affect cardiomyocyte ionic homeostasis. Finally, they have been shown to reduce myofilament stiffness as well as extracellular matrix remodeling/fibrosis in the heart, improving diastolic function. The research on the salutary mechanisms of SGLT2is in HFpEF is expected to improve both the understanding of HFpEF and its treatment (164).

The key-message from SGLT2i trials is that cellular stress, unrecognized in previous concepts, is essential for HF pathophysiology. The quest for efficient treatments was the main driver behind this evolution of concepts. The new concept did not repeal but instead revised the previous ones, improving treatment strategies.

Modern HFrEF and partly HFmrEF treatments are based on four pillars (besides diuretics): 1) ACEIs or ARBs, preferably in ARNI combination, 2) BBs, 3) MRAs, and 4) SGLT2is (listed chronologically). With regard to HFpEF, besides omnipresent diuretics, there are SGLT2is which span the entire EF spectrum, rendering EF irrelevant. Quadruple therapy with ARNI, BB, MRAs, and SGLT2is has been established as first-line therapy for patients with HFrEF in current HF guidelines. There is increasing evidence that many patients with HF with an LVEF >40% may benefit from these medications. SGLT2is are beneficial regardless of ejection fraction. HFpEF treatment should be targeted according to the underlying phenotype, comorbidities ant etiology. The hypertensive phenotype often requires ACEIs, ARBs, BBs, and not rarely MRAs, irrespectively of preserved EF, blurring the distinction between HFrEF and HFpEF. Obese and diabetic phenotypes may benefit greatly from GLP1 receptor agonists, in addition to SGLT2i (140, 165).

Some inconsistencies may be observed in clinical practice with regard to the use of MRAs in patients with HFpEF. The guidelines neither recommend the use of MRAs in those patients, nor oppose it. The question has been addressed in the 2023 update, since many patients with HFpEF who benefited from SGLT2i received MRAs and other neurohormonal inhibitors. The TOPCAT trial with spironolactone found generally neutral results but with ambiguity in interpretation due to marked regional differences in patient selection. A small trial (RAAM-PEF, 2011) demonstrated that the use of epleronone in patients with HFpEF was associated with significant reduction in markers of collagen turnover and improvement in diastolic function. Thus, it may make sense to prescribe a MRA drug along with a loop diuretic in patients with HFpEF (125, 166, 167).

A new entity called HF with recovered (or improved) ejection fraction (HFrecEF) has been recently recognized, along with uncertainties about prognosis and treatment (168–171). The data from the DELIVER trial indicate that these patients benefit from SGLT2 inhibition (161).

The idea that the failing myocardium recapitulates the fetal signaling program has been generally accepted but has been recently revisited and reinvigorated. The reactivation of fetal beta-myosin heavy chain isoform replacing the mature α variant and the shift from fatty acids to glucose as the main “fuel” in the myocytes of the failing heart are well-known examples. Fetal reprogramming is adaptive in the short time but is deleterious if sustained for long periods (172).

The uridine diphosphate N-acetylglucosamine (UDP-GlcNAc) molecule has been identified recently as a hub of fetal reprogramming in the failing heart. Prolonged increases in glucose uptake in overburdened cardiac myocytes, especially when SIRT1 and AMPK signaling is suppressed, enhances the hexosamine biosynthesis pathway. Its final product, UDP-GlcNAc, acts as a critical nutrient surplus sensor. UDP-GlcNAc is the key step for O-GlcNAcylation, which in collusion with mTOR-mediated phosphorylation rapidly and reversibly modifies a multitude of intracellular proteins in the nucleus, cytoplasm, and mitochondria. The deleterious effects are impaired calcium kinetics with contractile dysfunction, arrhythmias related to activation of voltage-gated sodium channels and Ca²⁺/CaMKII, mitochondrial dysfunction, maladaptive hypertrophy, fibrosis, and HF. Damage can be prevented by muting of O-GlcNAcylation, which poses a challenge for innovations, including the use of m-RNA technologies (173).

It turns out that chronic HF is heterogeneous disorder defying unifying concepts and treatment approaches. Ejection fraction is only an imperfect tool to distinguish between the phenotypes with different treatment strategies (174). The concept of cellular stress identifies common denominators at a cellular level. This raises hopes that experimental research will create new drugs based on this concept.

Unfortunately, the experience of previous blind alleys in research paths teaches that disappointment is a companion of hope. Positive inotropic agents and nesiritide are the examples. Monoclonal antibodies directed against tumor necrosis factor α (TNFα) etanercept and infliximab raised high hopes in the treatment of HF, considering the role of detrimental inflammatory response in the process of adverse myocardial remodeling. However, the clinical trials RECOVER, RENNESAINCE, and ATTACH dashed these hopes in 2001 (175, 176).

All that remains is to briefly comment on the evolving understanding of HF from simplistic clinical concepts to the thesis of a complex multicausal and multiorgan disorder. Viewpoints shifted from a mechanistic approach all the way down to the cellular level. Osler viewed HF as a terminal decompensated stage of many cardiac diseases. The diagnosis was based on clinical symptoms and signs alone. The traditional approach prevailed up to the 70s, when cardiac imaging provided hemodynamic quantitation. At the end of century, the neurohormonal concept stimulated the progress of laboratory diagnostics, leading to the launch of the NT-proBNP assay which enabled early recognition and monitoring of HF. Traditionally, the severity of symptoms was graded by NYHA functional classes from I to IV. The A-B-C-D HF staging was introduced in the USA in 2001. It describes the development of HF ranging from risk factors (A), structural heart disease without (B) and with prior or current failure symptoms (C), to refractory HF (D). European guidelines do not use such symbolic numeration but also describe the course of HF comprehensively, ranging from etiological factors to the advanced stage. Such a view contrasts with the historical notion of HF as a terminal edematous state with an ominous prognosis and detrimental clinical course (24, 177, 178).

This review focused on the pathophysiological concepts of HF in line with pharmacological management. A comprehensive approach includes other aspects, starting with preventive lifestyle changes. Devices, like cardiac resynchronization, physiological pacing, and implantable cardiac defibrillator (ICD), have improved clinical outcomes in addition to the drugs. Older outcome trials with HF drugs, conducted before the ICDs became a standard of care, are difficult to compare with the newer ones. Conversely, new HF treatments appear to reduce the benefit of ICD by decreasing the risk of sudden cardiac death (179, 180). Surgical or alternatively percutaneous revascularization has showed hardly any benefits on outcomes in the treatment of HF (HFrEF) due to ischemic heart disease. It may (or should) yet be considered in patients carefully selected on an individual basis (181–183). Epigenetic modulation, including mRNA technologies, is offering new prospects for treatment (184, 185). Regenerative strategies, which caused enthusiasm in the cardiac community two decades ago but then stalled, have been recently revisited (186–190). The modern armamentarium for HF treatment includes a gamut of approaches with cellular drugs and heart transplant at opposite sides of the spectrum. However, heart transplant is a bail-out action, while the drugs are the mainstay.

The diversity of treatment approaches reflects the complexity of HF pathophysiology, unmet needs (HF is still a detrimental disorder), and insufficient understanding of the underlying mechanisms. Reducing HF pathophysiology to the malfunction of the heart pump, with failure of the kidneys to excrete the excess of retained extracellular fluid, proved to be a simplistic concept insufficient for efficient treatment strategies. Adding the hemodynamics of circulation to the concept did not help much either; the resulting pharmacological interventions may have brought some temporary relief but without any impact on the final outcomes. Only a paradigm shift, discovering that the hub of HF pathophysiology is a detrimental systemic neurohormonal response, led to a breakthrough with reduction in adverse cardiovascular outcomes (including mortality) by pharmacological neurohormonal inhibition. HFpEF, where neurohormonal inhibition did not work, exposed the gaps in understanding. Chance discovery of SGLT2is, which are also beneficial for HFpEF and cardiorenal syndrome, highlighted the cellular aspects of HF. Revisiting cellular pathophysiology of the failing heart and the related organs, especially of the kidneys, holds promise to find new avenues of HF treatment. From the viewpoint of experimental research in cardiology, modulation of mitochondrial function may be one of those avenues (191–194).