Diabetes Mellitus and Heart Failure: Epidemiology, Pathophysiologic Mechanisms, and the Role of SGLT2 Inhibitors

Abstract

Diabetes mellitus (DM) and heart failure (HF) are frequently encountered afflictions that are linked by a common pathophysiologic background. According to landmark studies, those conditions frequently coexist, and this interaction represents a poor prognostic indicator. Based on mechanistic studies, HF can be propagated by multiple pathophysiologic pathways, such as inflammation, oxidative stress, endothelial dysfunction, fibrosis, cardiac autonomic neuropathy, and alterations in substrate utilization. In this regard, DM may augment myocardial inflammation, fibrosis, autonomic dysfunction, and lipotoxicity. As the interaction between DM and HF appears critical, the new cornerstone in DM and HF treatment, sodium-glucose cotransporter-2 inhibitors (SGLT2i), may be able to revert the pathophysiology of those conditions and lead to beneficial HF outcomes. In this review, we aim to highlight the deleterious pathophysiologic interaction between DM and HF, as well as demonstrate the beneficial role of SGLT2i in this field.

1. Introduction

Diabetes mellitus (DM) is a developing pandemic, with an anticipated global prevalence of 9.3%, and it is estimated that 50% of persons with DM are currently undiagnosed [1]. Due to the nature of its vascular consequences, it is a public health concern precipitating significant morbidity and death [2]. Cardiovascular complications continue to be the greatest cause of morbidity and death among people with DM. Overt cardiovascular diseases, such as obstructive coronary artery disease (CAD), acute ischemic stroke, or critical limb ischemia, are the most prevalent concerns in diabetic individuals [3]. DM is also a major cause of chronic kidney disease, retinopathy, and neuropathy.

DM may be involved in the development and progression of heart failure (HF), which is frequently termed diabetic cardiomyopathy. HF, irrespective of the cause, is characterized by a poor short- and long-term prognosis, with significantly impaired quality of life for the affected individuals. As the connection between DM and HF still remains incompletely understood, in this review article, we aim to provide an overview of the most recent evidence in the epidemiology and pathophysiology as well as the role of sodium-glucose cotransporter-2 inhibitors (SGLT2i) in the management of this deleterious interaction.

2. Epidemiological Trends: Heart Failure and Diabetes Mellitus

Since DM is a major risk factor for HF, it is no surprise that the prevalence and incidence rates of HF in individuals with DM are high. The authors of a recent study of 3.25 million individuals found that the presence of DM was associated with two-fold higher HF incidence [4]. We should note that in recent years, the proportion of patients who have suffered from HF that are obese or have DM increased to a large extent, as pointed by the study of Ciardullo et al. [5]. The longer duration and poorer control of DM are factors that signify an augmented risk of HF [4]. Nevertheless, HF may also develop in young individuals with a recent onset of DM [6,7]. A recent diagnosis of type 2 DM or insulin resistance was also associated with a greater risk of incident HF among UKPDS participants [8]. These results confirm previous longitudinal studies supporting such associations [9,10,11]. The increased HF risk may be even greater among women and those with type 1 DM, as shown in a recent systematic review and meta-analysis [12]. It should also be stated that HF with reduced (HFrEF) and preserved (HFpEF) left ventricular ejection fractions (LVEF) are equally common in individuals with DM [13], but HFpEF may remain undetected in a considerable proportion of patients [14].

HF may be the first cardiovascular manifestation of DM, especially among black women, as shown by a recent study conducted by Sinha et al. incorporating 40,117 individuals [15]. The analysis of larger cohorts has also demonstrated that HF is a common initial presentation in patients with DM [16,17]. Several noninvasive risk scores have been proposed with the objective of the risk stratification of patients with DM concerning the risk of incident HF [18]. These include the TRS-HFDM, WATCH-DM, Qdiabetes, BRAVO risk engine, Health ABC HF Score, and RECODe, among others [18]. It is particularly important to note that the co-presentation of DM in patients with HF may predispose them to a worse prognosis. DM was an independent predictor of short-term readmission in a study of 1727 Chinese patients with HF [19]. DM has been previously established as a severe risk factor for mortality in patients with HF [20]. The measurement of N-terminal, pro-B natriuretic peptides may be of importance in the risk prognostication of patients with DM and HF [21], as also endorsed by the latest HF guidelines [22].

The relationship between these two entities is bidirectional, since the development of insulin resistance or DM is common among individuals with HF. However, the reported prevalence of DM among patients with HF varies across studies. Pathological glucose homeostasis prevails in over 20% of patients with HF, with higher rates in clinical trials of pharmacological interventions and acute HF [23]. A contemporary registry of 7488 HF patients reported a DM prevalence of 29% [24]. Ethnic disparities also persist, with lower rates of DM being reported among African HF patients and higher rates among Middle Eastern individuals [25]. In a multinational registry, DM prevalence was 31% [26].

It becomes apparent that this interplay between DM and HF is based on shared pathophysiologic mechanisms that influence the natural history of both conditions, which will be discussed below.

3. Major Mechanisms in Heart Failure Pathophysiology

3.1. Inflammation

Inflammation represents a critical component of the pathophysiology of cardiovascular diseases [27,28,29]. In the setting of HF and irrespective of etiology, the injured myocardium results in the activation of the innate immune system through the interaction of pathogen- or damage-associated molecular patterns with pattern recognition receptors (toll-like receptors (TLRs) and NOD-like receptors), which are located on the myocardium and immune cells and are overexpressed in HF [30]. The ensuing increased excretion of pro-inflammatory cytokines and chemokines leads to endothelial cell activation as well as the activation of adaptive immunity through B and T cells, while the mobilization of immune cells (neutrophils, monocytes, and macrophages) further augments cytokine production and phagocytosis [30]. The association of inflammation with HF has also been determined in observational studies, since increased levels of inflammatory markers were associated with the development of HF [31]. Moreover, high C-reactive protein (CRP) levels at discharge following an acute HF decompensation event may be related to adverse prognosis and mortality [32].

3.2. Oxidative Stress

Along with inflammation, oxidative stress is also implicated in the development and progression of HF. The imbalance of the myocardial redox state, including the overproduction of reactive oxygen species (ROS) and defective innate antioxidant mechanisms, can result in adverse cardiac remodeling. Several hazardous effects have been described as a result of excessive ROS production, including cardiomyocyte electrophysiological dysfunction, impaired myocardial contractility due to calcium overload, mitochondrial dysfunction, and fibrosis due to an imbalance between the expression of the tissue inhibitors of metalloproteinases and matrix metalloproteinases. Furthermore, the major antioxidant mechanisms are impaired among individuals with HF, including superoxide dismutase [33], glutathione [34], and nicotinamide adenine dinucleotide [35].

3.3. Endothelial Dysfunction

Endothelial dysfunction is an additional consideration in the pathophysiology of HF, as it is a sequela of oxidative stress and inflammation [36]. The imbalance between nitric oxide (NO) and superoxide as the principal agents of endothelium-dependent vasodilation is highly prevalent in numerous cardiovascular risk factors (DM, arterial hypertension, and chronic kidney disease) apart from overt CAD, which is a fact that can explain the presence of endothelial dysfunction irrespective of LVEF and HF etiology [37]. Markers of HF and endothelial dysfunction have been correlated in previous studies, indicating the interaction between those two conditions [38]. It should also be noted that endothelial dysfunction is an important prognostic factor in chronic HF, as it is associated with a higher incidence of hospitalization and mortality [39].

3.4. Fibrosis

The increased concentration of collagen in the setting of myocardial fibrosis is an essential factor implicated in adverse myocardial remodeling and, eventually, HF development. Apoptotic cardiomyocyte cell death, as seen in acute ischemic events, represents the main source of replacement fibrosis [40]. Initially, fibroblasts are activated directly through microRNAs and matrix metalloproteinases or indirectly through endothelial and inflammatory cells [41]. Following activation, the proliferation and differentiation of fibroblasts has been observed owing to the action of growth factors (transforming growth factor-beta) or inflammatory cytokines (tumor necrosis factor-alpha), which leads to the formation of myofibroblasts, which, in turn, are responsible for collagen deposition in the extracellular matrix [42]. In the absence of apoptosis, other stimuli such as obesity, smoking, arterial hypertension, dyslipidemia, and DM have been implicated in the fibrotic process [43]. Several markers of fibrosis have been associated with HF, including galectin-3 and serum carboxy-terminal propeptide of procollagen type I [44,45]. The presence of fibrosis may also add important prognostic information in HF regardless of LVEF [46]. Previous studies of fibrosis of a histological (collagen volume fraction) or imaging-based (extracellular volume fraction) nature have suggested a deleterious association of fibrosis with adverse HF outcomes [47,48].

3.5. Cardiac Autonomic Neuropathy

Cardiac autonomic neuropathy (CAN) is defined by an imbalance between cardiac sympathetic and parasympathetic activity [49]. Due to the increased production of cardiac catecholamines, the activation of adrenergic receptors, and the activation of the renin-angiotensin-aldosterone system (RAAS), sympathetic activity might aggravate cardiomyopathy in individuals with DM. The longest nerve is the vagus nerve (parasympathetic); it innervates the heart and contains 75% of all parasympathetic nerve fibers. Postganglionic fibers innervate the atria in cardiac fat pads through ganglia, and subsequent neurotransmission is regulated by a nicotinic receptor. CAN affects the distal segment of the vagus nerve first, similar to the length-dependent aspect of diabetic peripheral neuropathy. A decrease in parasympathetic autonomic tone is largely evident in early CAN, as evaluated by cardiac autonomic reflex tests, and is accompanied by a reduction in heart rate variability (HRV). As a result, the sympathetic nervous system predominates, as seen by a greater resting heart rate. Autonomic nervous system imbalance emerges early in CAN pathogenesis and is related to increased cardiovascular risk prior to the development of definitive CAN. Insulin resistance, a hallmark of type 2 DM, prediabetes, and metabolic syndrome, exacerbates sympathetic predominance. Furthermore, ambulatory blood pressure monitoring reveals a lack of circadian rhythm in CAN patients with significant nocturnal hypertension. Untreated, parasympathetic vagal dysfunctions worsen, leading to the development of resting tachycardia caused by an unopposed sympathetic tone. Palpitations may be a common symptom of patients in this state, although they may also be entirely asymptomatic. As a result of advanced CAN, autonomic neuropathy develops in the (shorter) sympathetic plexus, with diminished sympathetic postural reflexes leading to postural hypotension. In end-stage CAN, the heart is completely denervated, and the resting increased heart rate is stable.

3.6. Alteration in Substrate Utilization

The heart is the most metabolically demanding organ; thus, it operates by consuming whichever energy substrate it has direct access to in order to maintain its contractile performance [50]. The healthy adult heart receives 60–80% of its energy through mitochondrial β-oxidation of fatty acids, with glucose, ketone bodies, lactate, and amino acids also being used to a lesser degree [50]. It should be stressed that glucose and ketone bodies are more efficient substrates compared to free fatty acids (FFAs) and that a shift towards FFA oxidation, as observed in patients with DM or insulin resistance, reduces the cardiac contractile function as a result of lipotoxicity [51].

However, the healthy adult heart is very flexible in terms of its regulation of overall energy metabolism, in which it constantly switches between glucose for oxidative metabolism in response to increases in circulating insulin postprandially and fatty acids in response to elevated plasma FFAs during prolonged fasting or starvation.

FAs can passively transfer across the cardiomyocytic plasma membrane to be utilized for cardiac ATP production, but their absorption is aided by both the FA translocase (CD36) and the FA binding protein (FABP) [52]. Carnitine palmitoyltransferase I (CPT1) converts FAs in the cytosol to long-chain acylcarnitine before they enter the mitochondria, where they are reversed to acyl-CoA by CPT2 and undergo β-oxidation [52]. Cytosolic FAs, on the other hand, can be esterified to produce triglycerides for storage in the myocardial triglyceride (TG) pool [52]. Despite the heart’s reliance on FAs for ATP, overall TG storage capacity is limited, with only around 3 mg of TGs stored per gram of myocardial tissue [50]. This could be due to the fact that excessive lipid accumulation can exert deleterious effects on the heart.

HF is exacerbated by changes in cardiac energy metabolism. The metabolic alterations that occur in HF, on the other hand, are complicated and rely not only on the degree and type of HF but also on the presence of common comorbidities such as obesity and type 2 DM. The failing heart has an energy deficit, which is caused mostly by a decline in mitochondrial oxidative capability. This is offset in part by an increase in ATP synthesis from glycolysis. The proportional contribution of various fuels to mitochondrial ATP synthesis changes as well, with glucose and amino acid oxidation decreasing and ketone oxidation increasing. Depending on the kind of HF, the heart’s fatty acid oxidation augments or decreases. For example, concerning HFpEF caused by diabetes or obesity, myocardial fatty acid oxidation increases [53,54], but in heart failure caused by hypertension or ischemia, it decreases. These energy-metabolic alterations, when combined, cause the failing heart to become less efficient.

4. The Role of Diabetes Mellitus in Heart Failure Pathophysiology

HF among individuals with DM is induced by the dysregulated management of glucose and insulin. The ultimate observation is the impairment of systolic and diastolic cardiac function by the reduction in myocardial contractility and compliance.

4.1. Myocardial Inflammation in Diabetes Mellitus

The role of inflammation in DM-induced cardiac dysfunction has already been established. In a recently reported experimental study, investigators detected an increase in MyD88-TLR2-TLR4 expression of primary cardiomyocytes, which was paired with an overexpression of inflammatory genes such as IL-1b, IL-6, and Tnfa [55]. These findings indicate the role of nuclear factor-kappaB (NF-κB), a major mediator of inflammation, in diabetic cardiomyopathy. Moreover, the authors proceeded to conduct an in vivo experiment using db/db mice, thereby confirming the in vitro findings. Advanced glycation end-products (AGEs) are also critical to NF-κB activation and myocardial inflammation in DM [56]. These types of molecules form when proteins or fats in the body become glycated. Their formation or binding to the receptor for AGEs (RAGE) expressed on the myocardial cell surface can induce inflammatory signaling. Moreover, RAGE signaling in endothelial cells and monocytes can also indirectly and adversely affect the myocardium [57]. The importance of AGEs in HF pathophysiology was recently demonstrated by the reported outcomes of the Rotterdam study [58]. The investigators detected an association between AGEs, whose presence is assessed by skin autofluorescence, and the prevalence of HF, which was more potent among individuals with DM. Interestingly, increasing AGEs were associated with lower LVEF in subjects without an HF diagnosis.

Following NF-κB activation, deleterious inflammatory processes ensue in the myocardium of patients with DM, including cardiomyocyte hypertrophy, fibrosis, and apoptosis, as well as the impairment of myocardial energetics, calcium management, and contractility [59]. Several human studies have suggested that an augmented systemic inflammatory burden could be predictive of incident HF in individuals with DM. In the ADVANE trial of 3098 participants with type 2 DM, one standard deviation increase in interleukin-6 (IL-6) (hazard ratio: 1.48, 95% confidence interval: 1.27 to 1.72) or high sensitivity CRP (hazard ratio: 1.32, 95% confidence interval: 1.12 to 1.55) was associated with a higher risk of HF incidence or progression [60].

4.2. Myocardial Fibrosis in Diabetes Mellitus

Numerous factors, such as oxidative stress, pro-inflammatory states, growth factor secretion, neurohumoral activation, the deposition of AGEs, and the activation of the RAAS, may raise the likelihood of diffuse myocardial fibrosis among people with DM [61]. The activation of myofibroblasts, which generate fibrous tissue, by these effectors may alter the balance of fibrosis deposition throughout breakdown [40]. Studies, both preclinical and clinical, have shown that DM induces a pro-fibrotic state. Four-week-old male Sprague–Dawley rats with streptozocin-induced DM exhibited a higher collagen type I and III accumulation compared to the control group [62]. Critically, a recently published systematic review and meta-analysis of 32 studies by Salvador et al. highlighted a relationship between DM and a higher degree of myocardial fibrosis assessed by either the histological collagen volume fraction (mean difference 5.80; 95% CI: 2.00–9.59) or extracellular volume fraction (mean difference: 2.09; 95% CI: 0.92–3.27) [63]. The findings were more pronounced in individuals with worse glycemic control [63]. It should also be stated that increasing myocardial fibrosis assessed by cardiac magnetic resonance imaging could signify an augmented risk of major adverse cardiac, cerebrovascular, and heart failure events in individuals with DM [64,65].

4.3. Cardiac Autonomic Neuropathy in Diabetes Mellitus

The presentation of dysglycemia and insulin resistance in individuals with DM are plausible mediators of excessive sympathetic activity, which leads to myocardial hypertrophy, fibrosis, and dysfunction, thereby easing progression to HF [66]. Hyperglycemia is assumed to be the chief factor, as it triggers a series of complicated processes and pathways that cause oxidative stress and toxic glycosylation products, ultimately leading to neuronal malfunction and death [67]. Hyperglycemia causes an increase in the mitochondrial synthesis of free radicals, which causes oxidative damage to the microvasculature that supplies these peripheral nerves [68]. Inflammation may be an additional driver of CAN, as indicated by increased levels of inflammatory molecules in those with CAN [69]. Moreover, increased IL-1 levels could predict progressive changes in the resting heart rate of individuals with type 2 DM, which is suggestive of inflammation’s role in the development of CAN [70]. However, the exact pathophysiology of CAN is unknown, since the processes involved have only been examined in somatic models and extended to the autonomic nervous system.

The criteria for the diagnosis of CAN among DM patients has not been established, and usually requires one to two abnormal functional tests (excitation/inhibition ratio, the Valsalva maneuver, 30:15 ratio, blood pressure postural decrease, and HRV indices). Its prevalence ranges from 17% to 73% in individuals with DM [69], and may also be increased in those with prediabetes compared to normoglycemic subjects [71]. The severity of CAN is dependent on the duration of DM, the degree of glycemic control, the presence of comorbidities, and specific genetic polymorphisms [72,73,74]. Moreover, the role of CAN in the development of HF was recently documented in the ACCORD study [75]. Analyzing 7160 participants without HF at baseline over a 4.9-year follow-up, the investigators highlighted that lower HRV measured using standard deviation of all normal-to-normal intervals (SDNN) led to a greater incident HF risk, even after adjustment for confounders (adjusted hazard ratio for lowest vs. highest SDNN quartile—1.70; 95% confidence interval—1.14 to 2.54). Moreover, individuals with CAN, defined as the lowest SDNN quartile with the highest QT interval index and heart rate quartiles, had an augmented risk of developing HF (adjusted hazard ratio—2.65; 95% confidence interval—1.57 to 4.48).

4.4. Cardiac Lipotoxicity in Diabetes Mellitus

Due to a ‘lipid spillover’ from overloaded adipose tissue, circulating free FAs (FFAs) and triglyceride (TG)-rich very low-density lipoproteins (VLDL) are elevated in DM sufferers, thus enhancing their availability for absorption by the heart. Elevated plasma FFA concentrations result in considerably increased myocardial FFA absorption and consequent intramyocardial lipid deposition [76], indicating that serum FFAs are important regulators of myocardial lipid accumulation. Furthermore, in type 2 DM, lipoprotein lipase (LPL) expression on cardiomyocytes is increased, thus improving VLDL hydrolysis in the coronary circulation and thereby increasing lipid availability [77]. As a result, lipid absorption by cardiomyocytes increases in DM-affected individuals, possibly stimulating an increase in FA oxidation. This increase in FA oxidation is likely further provoked by decreased glucose uptake induced by insulin resistance, as well as the suppression of glucose oxidation by lipid catabolism intermediates such as acetyl-CoA and citrate [78].

Despite this increase in FA oxidation, research shows that it is frequently insufficient to avoid myocardial TG buildup in patients with HFpEF. HFpEF patients showed 2.3-fold greater myocardial lipid content than the control participants in research utilizing CMR to measure cardiac lipids, and this content was independently linked with diastolic strain rate [79]. Similar research has shown that HFpEF patients have considerably more intramyocardial fat than non-HF and HFrEF patients, which corresponds to both the E/e’ ratio and the left atrial volume index, while women with subclinical HFpEF had more intramyocardial fat than reference controls [80,81]. A very low-calorie diet causes an initial increase in myocardial TG deposition, which is related to immediate diastolic functional impairment as measured by a lower E/A ratio in type 2 DM patients [76]. This group of patients also had a 2.1-fold higher myocardial TG content than the normoglycemic controls and a worse early diastolic filling rate [82]. These findings suggest that intramyocardial fat accumulation is linked to decreased diastolic performance among people with diabetes.

5. A Novel Era in Heart Failure Pharmacotherapy: SGLT2 Inhibitors

As the interaction of DM with HF is of particular interest, the effect of novel antidiabetic agents has been examined in HF populations, and the foremost of such agents is SGLT2i. Below, we elaborate on their pleiotropic mechanisms of action that could be of importance in reverting HF pathophysiology, as shown in preclinical studies. Moreover, we provide the available information stemming from clinical trials and cohort studies that demonstrate the effect of SGLT2i on the prevention of adverse HF outcomes.

5.1. SGLT2i as Hypoglycemic Agents

The SGLT2 protein consists of a glucose ring and a proximal and distal benzene ring connected by a methylene bridge. As they are responsible for the reabsorption of the majority of filtered glucose by the glomeruli, SGLT1 and SGLT2, which are found on the proximal convoluted tubule of nephrons, are essential mediators of glucose homeostasis. The large capacity, low-affinity SGLT2 absorbs 90% of the glucose, with a sodium turnover ratio of 1:1 [83,84]. As a result, facilitative glucose transporters (GLUTs) situated in the epithelial lining of proximal tubules unleash glucose into the bloodstream [83]. The impact of SGLT inhibition on diabetes was first observed in 1987 with the use of phlorizin, a non-specific SLGT1 and SGLT2 inhibitor [85]. Its treatment resulted in reduced hyperglycemia and increased glycosuria in an animal model of type 2 DM, along with the maintenance of normal insulin sensitivity. However, specific SGLT2i types have also been produced. Dapagliflozin was the first high-potency SLGT2i to be discovered [86], followed by canagliflozin [87], empagliflozin [88], and ertugliflozin [89]. Other compounds of this family, such as ipragliflozin, tofogliflozin, and luseogliflozin, have also been investigated, but less thoroughly. In terms of hypoglycemic impact, all licensed medications have shown efficacy in terms of decreasing glucose levels and glycated hemoglobin (HbA1c) in randomized clinical studies.

5.2. Pleiotropic Mechanisms of SGLT2i

Several mechanisms are thought to be involved in the pleiotropic effects conferred by SGLT2 inhibition to HF (

Table 1. Preclinical evidence of the pleiotropic mechanisms of action of SGLT2 inhibitors (SGLT2i).

	Type	Experimental Model	Disease Type	SGLT2i	Finding	Target Mechanism
Ren et al. [94]	In vitro	Cardiomyocytes	HF	DAPA	↑ SIRT1 ↓ fibroblast transformation to myofibroblast ↓ fibroblast migration	Autophagy Fibrosis
Yu et al. [95]	In vitro + In vivo	Primary cardiomyocytes Male C57/BL6 mice	IRI	DAPA	↓ IL-1β ↓ NLRP3 inflammasome activation ↓ Atg-5, Beclin-1, LC3B-II and P62 ↑ NLRP3 phagocytosis by autophagosomes	Inflammation Autophagy
Quaqliariello et al. [96]	In vivo	C57Bl/6 mice	Cardiotoxicity	EMPA	↓ intracellular ROS, MDA, lipid peroxidation↓ IL-6, IL-8, NF-κB, NLRP3 ↓ pro-collagen 1α1, MMP-9 ↓ apoptotic nuclei, caspase-3	Oxidative stress Inflammation Fibrosis Apoptosis
Sukhanov et al. [97]	In vitro	Aortic SMC	Inflammation	EMPA	↓ NLRP3, IL-1β, IL-18, Caspase-1 ↓ Superoxide, hydrogen peroxide	Inflammation Oxidative stress
Zhang et al. [98]	In vivo	Female landrace pigs	HF	DAPA	↓ Collagen-1 and -3, TGF-β	Fibrosis
Kondo et al. [99]	Ex vivo + In vitro	Atrial tissue H9C2 and primary human cardiomyocytes	High glucose	CANA	↓ NADPH ↑ BH4, NOS coupling ↓ NF-κB, TNF-α, and apoptosis pathways	Oxidative stress Inflammation Apoptosis
Cappetta et al. [100]	In vivo + In vitro	Dahl salt-sensitive rats Ventricular cardiomyocytes	Hypertension	DAPA	↓ VCAM-1, E-Selectin ↓ NF-κB, IL-6, MCP-1 ↓ Collagen-1, TGF-β, MMP-2	Endothelial dysfunction Inflammation Fibrosis
Madonna et al. [101]	In vivo	Male C57BL/6 mice	DM	EMPA	↓ collagen content ↓ p38	Fibrosis Inflammation
Young et al. [102]	In vivo	Male C57BL/6J mice	HF	SOTA	↓ histological fibrosis	Fibrosis
Tian et al. [103]	In vivo + In vitro	Male Sprague Dawley rats HUVECs	DM	DAPA	↓ collagen deposition ↓ TGF-β expression ↓ EndMT ↓ fibroblast activation ↓ ROS and NADPH Oxidase 4	Fibrosis Oxidative stress
Zhang et al. [104]	In vivo	Sprague-Dawley rats	HF	DAPA	↓ Collagen-1 and -3, TGF-β	Fibrosis
Liu et al. [105]	In vivo	C57BL/6 mice	HF	EMPA	↓ histological fibrosis ↓ caspase-3, Bcl2	Fibrosis Apoptosis
Santos-Gallego et al. [106]	In vivo	Yorkshire pigs	HF	EMPA	↓ histological and imaging fibrosis ↑ NO bioavailability, cGMP, PKG	Fibrosis Endothelial dysfunction
Marfella et al. [107]	In vitro	Ventricular cardiomyocytes	DM	Any	↑ JunD/PPAR-γ and ceramide ↓ IRS1 and IRS2	Lipotoxicity
Sun et al. [108]	In vivo + In vitro	C57BL/6J mice HL-1 cells	DM	CANA	↓ IL-6 and TNF-α ↓ ROS ↓ mTOR or HIF-1α signaling	Inflammation Oxidative stress Lipotoxicity

HF: heart failure, DAPA: dapagliflozin, IRI: ischemia/reperfusion injury, IL: interleukin, EMPA: empagliflozin, ROS: reactive oxygen species, MDA: malondialdehyde, NF-κB: nuclear factor-kappaB, MMP: matrix metalloproteinase, SMC: smooth muscle cell, TGF: transforming growth factor, BH4: tetrahydrobiopterin, CANA: canagliflozin, NADPH: nicotinamide adenine dinucleotide phosphate, TNF: tumor necrosis factor, VCAM: vascular cell adhesion molecule, MCP: monocyte chemoattractant protein, DM: diabetes mellitus, SOTA: sotagliflozin, EndMT: endothelial-to-mesenchymal transition, NO: nitric oxide, PKG: protein kinase G, PPAR: peroxisome proliferator-activated receptor, IRS: insulin receptor substrate, mTOR: mammalian target of rapamycin, and HIF: hypoxia-inducible factor. ↑ indicates an increase, ↓ indicates a decrease.

). To begin with, among the earliest hypothesized processes to be identified were increased glycosuria and natriuresis, which led to osmotic diuresis and volume control. However, based on current SGLT2i HF studies, it is extremely unlikely that this theory captures the major driver of decreased HF morbidity. Although there may be a brief rise in diuresis [90], there is no effect on natriuretic peptides during the early phase [91]. This suggests that a greater proportion of patients with a moderate or significant decrease in natriuretic peptides during long-term follow-up may be undergoing a reversal of the unfavorable cardiac remodeling process [92]. As we recently showed through a systematic review and meta-analysis, the use of SGLT2i led to significant improvement of systolic and diastolic function-imaging indices [93], which is indicative of the amelioration of cardiac remodeling that was effected through this treatment.

At this point, it should be noted that SGLT2 expression is lacking in the human heart, but SGLT1 is plentiful, particularly in pathologic conditions [109]. Nonetheless, the cardioprotective effects seen with even highly selective SGLT2i (empagliflozin) may suggest that these benefits are primarily regulated by indirect, systemic actions [110]. The emergence of hemoglobin and hematocrit as the most crucial predictors of the improvements reported with SGLT2-is, according to statistical mediation studies of major clinical trials, has led to a reassessment of the probable mechanistic relationship [111,112]. The impact of SGLT2 inhibition on erythropoiesis was investigated further in a sub-study of the EMPA-HEART CardioLink-6 randomized controlled trial (RCT), which revealed an increase in erythropoietin, hemoglobin, and hematocrit six months after randomization [113]. Since an increase in hematocrit and hemoglobin alone did not result in cardiovascular benefits when erythropoietin-like agents were used, this effect could be driven by hypoxia-inducible factors (HIFs) [114], potentially as an additional result of the stimulation of nutrient deprivation signaling pathways [115]. It has long been recognized that DM and HF are conditions characterized by a nutrient surplus, which leads to the inhibition of the nutrient deprivation sirtuin-1 (SIRT1)/HIF/5′ adenosine monophosphate-activated protein kinase (AMPK) pathway and the stimulation of the Akt/mammalian target of rapamycin complex 1 (mTORC1) pathway, resulting in increased endoplasmic reticulum stress, oxidative stress, inflammation, and apoptosis (Figure 1). According to a novel concept, SGLT2 inhibition may cause caloric loss in the urine and a reduced tissue distribution of glucose, resulting in a global oxygen and nutrient deficiency and, as a result, affecting the balance between the pathways listed above [29]. Indeed, SGLT2 expression appears to be negatively associated with SIRT1 in the proximal renal tubular area [30]. In a recent in vitro study of angiotensin II-stimulated cardiomyocytes, dapagliflozin activated SIRT1 and attenuated HF development by inhibiting the transformation of fibroblasts into myofibroblasts as well as fibroblast migration [94]. The study of Yu et al. also suggested that SGLT2i could alleviate myocardial injury by restoring autophagy, but other recent studies provided conflicting results [116,117]. Therefore, the existing evidence is not straightforward, and it may be the mediation of the selective autophagic flux that is critical. Lastly, the latest report of the EMPEROR program analyzing circulating proteomics supports the validity of the widespread effect of empagliflozin on the amelioration of autophagic flux at the heart, kidneys, and endothelium [118].

SGLT2-Is may be able to improve oxidative stress and inflammation [119,120]. Antioxidant and anti-inflammatory effects of SGLT2 inhibition have been demonstrated in recent experimental studies at the level of the heart [96,97], implicating key inflammatory mediators such as the NLRP3 inflammasome and NF-κB. According to our systematic review and meta-analysis, SGLT2i can significantly decrease the levels of circulating inflammatory markers, thereby confirming this potential anti-inflammatory action [121]. Furthermore, in a nine-week investigation of pigs with HF and retained LVEF, Zhang et al. found that daily dapagliflozin treatment resulted in decreased expression of inflammatory markers in the aortic tissue [98]. After discovering high expression of the SGLT1 isoform in human cardiomyocytes (SGLT2 was undetectable), Kondo et al. performed SGLT inhibition with canagliflozin [99]. The researchers discovered antioxidant effects such as decreased nicotinamide adenine dinucleotide phosphate oxidase activity, increased tetrahydrobiopterin bioavailability, and enhanced NO synthase coupling via a mechanism involving SGLT1/AMPK/Rac1. Additional anti-inflammatory and anti-apoptotic effects were detected. However, there is a scarcity of evidence in the clinical context. Although other studies found a substantial reduction in oxidative stress (8-iso-prostaglandin F2a and 8-hydroxy-20-deoxyguanosine) and inflammatory markers (CRP, IL-6, and tumor necrosis factor-alpha (TNF-α)), their designs were not capable of demonstrating the correctness of those findings [122]. In a study of patients who underwent coronary artery bypass grafting, those who received SGLT2i had significantly reduced IL-1, IL-6, and TNF-α levels at the 5-year follow-up [123]. It should be noted that a few studies [124,125,126,127] have revealed non-significant changes in the aforementioned parameters. However, in the EMPA-TROPISM study, 6-month therapy with empagliflozin resulted in a substantial decrease in inflammatory biomarkers in patients without diabetes, with HF, and a lower LVEF [128]. A recent systematic review and meta-analysis of RCTs has also suggested a significant anti-inflammatory effect of SGLT2 inhibition [129]. Additionally, the proteomic analysis on empagliflozin trials has also shown a differential expression of proteins associated with inflammation and oxidative stress at the level of the heart [118].

Fibrosis may also be reduced after SGLT2 inhibition, as indicated by decreased collagen deposition in cardiac cells after therapy with dapagliflozin or empagliflozin [100,101]. Furthermore, treatment with sotagliflozin, a dual SGLT1/2 inhibitor, resulted in reduced histological fibrosis in a mouse model of transverse aortic constriction-induced pressure overload [102]. In the same study, dapagliflozin reduced left ventricular fibrosis in a rat model of mitral regurgitation-induced myocardial failure, which was caused by decreased apoptosis and endoplasmic reticulum stress [130]. Tian et al. observed a reduction in cardiac fibrosis after treatment of dapagliflozin in a mouse model of diabetic cardiomyopathy, as evidenced by decreased collagen deposition and levels of fibrosis biomarkers [103]. The reduced endothelial-to-mesenchymal transition and fibroblast activation were identified as the orchestrators of this positive impact via the AMPK/TGF-β/Smad signaling pathway [103]. This signaling pathway was also linked to dapagliflozin’s anti-fibrotic actions in a rat model of angiotensin-II-induced heart failure [104]. Empagliflozin therapy resulted in the AMPK-mediated alleviation of myocardial fibrosis in a post-myocardial infarction animal model [105]. In addition, in a multimodality investigation of pigs with proximal left anterior descending artery occlusion-induced HF, empagliflozin treatment was linked with improved diastolic function based on decreased fibrosis and endothelial dysfunction [106]. There is a paucity of clinical evidence on the effect of SGLT2 inhibition on fibrosis. Daily empagliflozin administration for six months showed no effect on fibrosis as measured by cardiac magnetic resonance among type 2 DM patients [131]. Patients with type 2 DM and CAD experienced a decrease in extracellular volume in the randomized controlled experiment EMPA-HEART CardioLink-6 [132]. Finally, compared to the control group, participants in the EMPA-TROPISM study showed a substantial reduction in interstitial myocardium fibrosis following chronic therapy with empagliflozin [128]. Proteins associated with fibrosis were among those with altered cardiac expression in a proteomic signature analysis reported by the EMPEROR program, which further supports the anti-fibrotic hypothesis of SGLT2i [118].

Other mechanisms involved in the deleterious effect of DM on the heart may be influenced by SGLT2 inhibition but have been studied to a lesser extent. SGLT2i may attenuate JunD/PPAR-γ-mediated cardiac lipotoxicity [107]. Canagliflozin also ameliorated cardiomyocyte lipotoxicity through the mTOR/HIF-1α pathway in an experimental study conducted by Sun et al. [108]. Another explanatory concept is the decrease in epicardial fat caused by leptin downregulation [128,133]. Epicardial adipose tissue growth is linked to poor left ventricular relaxation and diastolic filling, either directly through constriction or indirectly through the regulation of pro-inflammatory molecule release, resulting in inflammation, microcirculatory dysfunction, and fibrosis [134]. SGLT2 expression has been established in human epicardial adipose tissue, and the majority of the data on individuals with type 2 DM show a decrease in epicardial adipose tissue mass after SGLT2 inhibition [134,135]. The newly published randomized, placebo-controlled EMPA-TROPISM study found that empagliflozin decreased epicardial fat content in individuals without T2DM, HF, and a lower LVEF [128]. At the same time, another molecular theory has been questioned, namely, the impact of SGLT2 inhibitors on the cardiac sodium–hydrogen exchanger-1 as a means of minimizing myocardial damage and incident HF [136,137,138]. Enhanced ATP generation owing to ketone body overproduction through beta-hydroxybutirate metabolism was an intriguing idea in this regard, since ketones are considered to be a more efficient substrate. However, this was not confirmed in a recent experiment [139]. Moving on to the effects of SGLT2i on CAN, in a study of patients with type 2 DM and vasovagal syncope recurrence, SLGT2i use was associated with a significantly lower low frequency/high frequency ratio and norepinephrine serum levels and a higher Heart to Mediastinum ratio in 123I-metaiodobenzylguanidine (123I-mIBG) myocardial scintigraphy at the conclusion of the study [140]. These SGLT2i-mediated effects might indicate more balanced autonomic system activity and an induced improvement in CAN. Dapagliflozin also ameliorated heart rate variability and heart rate turbulence parameters in patients with type 2 DM and CAN [141]. These findings contradict a recent meta-analysis that documented a trivial effect of SGLT2 inhibition on indices of CAN [142]. It should be noted that the improvement in CAN with SGLT2i may be more pronounced in individuals with HF, as suggested by the study of Hamaoka et al. [143]. Finally (but still of great importance), an improvement in renal function observed with SGLT2i may be responsible for some of the cardiac benefits [144].

5.3. SGLT2 Inhibitors and Heart Failure Outcomes

The programs of SGLT2i clinical trials have met with overwhelming success concerning HF outcomes, either in trials of solely HF populations or in studies of patients with DM or kidney diseases (Figure 2). Following the initial positive findings in cardiovascular outcome trials, research into HF patients was conducted to explore the working hypothesis of cardiac protection further. The DAPA-HF study found a decreased risk of worsening HF or cardiovascular mortality in patients with HF and a lowered LVEF, which were independent of DM status [145]. Notably, there was no difference in outcomes based on HF etiology (ischemic vs. non-ischemic) [146], the simultaneous use of an angiotensin receptor neprilysin inhibitor (ARNI) [147], or baseline health condition [148]. Simultaneously, patients with longer HF duration [149] and LVEF of 35% saw more substantial benefits [150]. Following therapy with dapagliflozin, the participants’ quality of life improved significantly [148]. The subsequent EMPEROR-REDUCED study investigating a comparable patient population found decreased rates of cardiovascular mortality and HF hospitalization in the empagliflozin arm compared to the placebo, regardless of DM [151]. In terms of the co-administration of agents proven to be effective for the management of HF with reduced LVEF, the use of mineralocorticoid receptor antagonists (MRAs) did not change the initial findings, whereas treatment with empagliflozin resulted in lower MRA discontinuation rates and higher MRA initiation rates [111]. Regarding angiotensin receptor neprilysin inhibitors (ARNIs), their usage was related to an incremental decline in endpoint occurrence [152]. Furthermore, the outcomes were constant regardless of chronic kidney disease status or baseline HbA1c [153,154].

SGLT2 inhibition has also proven to be effective in terms of the difficult-to-treat HF population with maintained LVEF, a condition with no previously authorized therapeutic options other than symptomatic. In the SOLOIST-WHF study, starting sotagliflozin after an episode of worsening HF resulted in decreased risks of cardiovascular death, HF hospitalizations, or urgent visits. Despite the fact that the study was prematurely discontinued and did not specifically investigate the effect of sotagliflozin on HF with preserved LVEF, a 52% relative risk reduction in the primary endpoint was observed in patients with HF with LVEF 50% on sotagliflozin [155]. The EMPEROR-PRESERVED trial findings corroborated the efficacy of SGLT2 inhibition towards HF with LVEF >40%, indicating a 29% reduction in HF hospitalizations associated with an improvement in patients’ quality of life, regardless of the presence of DM [156]. Despite the fact that patients with modestly decreased LVEF were also included in the study, the results were fairly constant across the LVEF range, with the highest advantages shown in those with LVEF 50% or lower. In the DELIVER trial of dapagliflozin, individuals with HFmrEF or HFpEF receiving the intervention experienced lower endpoint rates compared to those receiving a placebo [157]. Dapagliflozin’s effect was evident as early as 2 weeks after initiation [158], was more pronounced in subjects with a worse baseline symptom burden [159], and was consistent irrespective of age, body mass index, frailty, baseline glycemic status, or MRA/ARNI use [160,161,162,163,164]. The expected addition of SGLT2i in the societal guidelines of HFpEF/HFmrEF management may result in a significant alleviation of HF burden, with a recent study indicating 250,000 fewer hospitalizations for worsening HF in this group of patients [165].

6. Conclusions

Diabetes mellitus represents a major risk factor for the development of heart failure since the two afflictions’ pathophysiological characteristics are closely related, as proven by multiple preclinical and clinical studies. The emergence of a diabetic pharmacotherapy, particularly the use of sodium-glucose cotransporter-2 inhibitors, as an effective approach in the management of heart failure across the spectrum of left ventricular ejection fraction highlights the need for their implementation in real-world clinical practice to reduce the burden of both diseases.