Redefining Diabetic Cardiomyopathy: Perturbations in Substrate Metabolism at the Heart of Its Pathology Free

Despite the complications of diabetes affecting many organs in the body, the leading cause of mortality in people with type 1 (T1D) and type 2 (T2D) diabetes is cardiovascular disease. Diabetes increases the incidence of myocardial infarction (MI) and heart failure (1), and individuals with diabetes who have not had a prior MI have cardiovascular mortality rates comparable with those of individuals without diabetes who have had a previous MI (2). In part, this can be attributed to the macrovascular dysfunction, endothelial dysfunction, accelerated atherosclerosis, and hypertension found in many individuals with diabetes. In addition to these vascular complications, people with diabetes develop an often asymptomatic and undiagnosed diastolic dysfunction (3). This diastolic dysfunction is a hallmark of diabetic cardiomyopathy (DbCM) and is also a characteristic of heart failure with preserved ejection fraction (HFpEF). HFpEF is more prevalent within the population with diabetes; however, we currently do not fully understand the mechanism(s) responsible for HFpEF and have limited treatment options available for these individuals. As such, it is imperative we address whether treating diastolic dysfunction and alleviating DbCM in the early stages of T2D can influence the progression of HFpEF, which would have major implications for the burden imposed on our health care systems.

Herein we will provide an overview of the pathology of DbCM, highlighting some of the key mediators, with major emphasis on the role of perturbations in cardiac substrate metabolism. Furthermore, we will interrogate whether correcting these perturbations in cardiac substrate metabolism may represent viable targets for the treatment of DbCM. We will also consider the alternative role for disrupted metabolites, as signaling molecules regulating cell function in the heart. While our focus will primarily be DbCM in the setting of T2D, we will also consider these aspects in the context of T1D.

The clinical phenotype of DbCM was first described by Rubler et al. (4) in the 1970s, from autopsy findings of four individuals with diabetes with no evidence of MI but signs of left ventricular (LV) hypertrophy, cardiomegaly, and congestive heart failure of unknown causes. This led to the definition of a DbCM: a condition in patients with diabetes characterized by the presence of ventricular dysfunction but in the absence of underlying coronary artery disease and/or hypertension. Our understanding of the clinical phenotype of DbCM has greatly improved in the 21st century (Fig. 1), aided by advances in the diagnostic capabilities of noninvasive imaging modalities (5,6). Through these diagnostic advancements it has been found that a decline in diastolic function is a key feature of DbCM, often in the absence of systolic dysfunction. Impairments in diastolic function have been identified early in the progression of diabetes and were previously overlooked, as this group of asymptomatic individuals would not be routinely investigated for cardiac dysfunction (3). The prevalence of diastolic dysfunction in T2D has been reported to range from as low as 20% to nearly 80%, depending on the diagnostic criteria used and patient group studied (7–10). In addition to LV hypertrophy, DbCM is also characterized by increased wall thickness, diffuse myocardial fibrosis, and intramyocyte lipid accumulation (11). What remains unclear, despite the advances in our understanding of DbCM, is the progression of cardiac complications in people with diabetes and how the different structural and functional changes progress longitudinally with diabetes duration. In addition, developments in the field have been hindered by the lack of a universally accepted and consistently applied definition of DbCM.

Individuals with HFpEF have symptoms of heart failure in the absence of systolic dysfunction (ejection fraction >50%) and often display evidence of LV diastolic dysfunction and increased LV filling pressures (12). Development of HFpEF is particularly prevalent in individuals with diabetes; in a recent study investigators reported that 45% of HFpEF cases were in patients with diabetes (13). Phenomapping of HFpEF patients has identified a large subgroup characterized by a “metabolic, obese” phenotype (14). Individuals with diabetes and HFpEF have a higher prevalence of hypertension, pulmonary diseases, and renal dysfunction, while also having more severe cardiac remodeling with higher filling pressure and diastolic dysfunction compared to individuals with HFpEF in the absence of diabetes (15,16). In addition, women are more susceptible to HFpEF and diastolic dysfunction than men (17). Despite an overlap in clinical features between DbCM and HFpEF (Fig. 1), there are symptomatic differences, especially in relation to the severe exercise intolerance and exertional dyspnea in the latter. Clinical studies directly comparing DbCM with HFpEF in individuals with diabetes are needed, and we currently lack longitudinal studies in the population to determine whether DbCM increases risk for HFpEF. Such studies may explain why the abovementioned measures are exacerbated in those with diabetes and HFpEF.

Our understanding of the principal mechanisms underlying DbCM has primarily been uncovered using animal models of obesity/insulin resistance. Several preclinical models are available for investigating DbCM, of which the strengths and weaknesses have been extensively reviewed (18). A number of dysfunctional cellular processes have been identified within the diabetic myocardium. These include abnormal substrate metabolism, mitochondrial dysfunction and oxidative stress, dyslipidemia and lipotoxicity, increased fibrosis, inflammation, microvascular/endothelial dysfunction, endoplasmic reticulum stress, abnormal calcium handling, and glucotoxicity (Fig. 2), many of which promote cardiomyocyte death (and have been extensively reviewed in the article by Ritchie and Abel [3]). These mechanisms operating in isolation is highly unlikely, and they possibly interact to propagate DbCM. As an example, impaired insulin signaling causes abnormal substrate metabolism, which can lead to mitochondrial dysfunction, driving oxidative damage and lipotoxicity (19,20). Identification of the cellular processes which are “drivers” and those that are “passengers” of DbCM progression is needed, as this would allow for better identification of druggable targets for pharmacotherapy. Almost universally, changes in cardiac substrate metabolism are observed in human and animal models of both T1D and T2D. The following sections of this article will thus focus on these metabolic changes within the diabetic myocardium and whether they represent targets for potential pharmacotherapy.

The healthy adult heart is a metabolically flexible organ capable of switching between substrates including fatty acids, carbohydrates, amino acids, and ketones according to the underlying physiological state (21). The majority of ATP generated from substrate oxidation comes from breakdown of fatty acids in the fasted state, whereas carbohydrates become a more dominant fuel source following feeding (22–24). In individuals with either T1D or T2D, profound changes in cardiac metabolism have been identified (Fig. 3).

In clinical studies of arterial and coronary sinus blood samples from people with T1D without coronary artery disease, investigators observed elevated myocardial fatty acid uptake compared with that in healthy individuals (25,26). Similarly, positron emission tomography imaging studies revealed increased fatty acid oxidation and utilization within the myocardium of individuals with T2D (27). In animal models of diabetes, increased myocardial fatty acid oxidation rates have been demonstrated with isolated heart perfusions using radioisotopes (28,29). This was accompanied by increased incorporation of fatty acid into the triacylglycerol (TAG) pool and other lipid intermediates, indicative of cardiac lipotoxicity (30,31). Increased myocardial TAG content has also been confirmed in individuals with T2D using magnetic resonance spectroscopy (MRS) and was found to be an independent predictor for diastolic dysfunction (32). Extensive research in the 21st century has led to significant advancements in our understanding of the molecular mechanisms that drive excess cardiac fatty acid oxidation in diabetes. An early event in the pathogenesis is increased fatty acid uptake across the sarcolemma, predominantly regulated by fatty acid translocase (FAT/CD36). FAT/CD36 translocates between intracellular vesicles and the membrane, and in diabetes it is permanently relocated to the membrane, thereby facilitating excessive fatty acid uptake to fuel oxidation and esterification (33).

Transcriptional upregulation of multiple genes involved in fatty acid uptake, oxidation, and storage is mediated by the transcription factor peroxisome proliferator–activated receptor-α (PPARα), resulting in a coordinated increase of fatty acid metabolism in diabetes (34). Lipid intermediates are endogenous agonists for PPARα, and thereby fatty acids transcriptionally regulate their own fate. Key studies by Finck et al. (34) demonstrated that cardiac-specific PPARα overexpression in mice recapitulated a DbCM phenotype, providing seminal evidence that the metabolic changes within the heart were sufficient on their own to induce cardiac dysfunction.

It should be noted that there is extensive intracellular cross talk between the oxidation of different fuels to limit substrate wasting and futile cycling. This cross talk forms the basis of the “glucose–fatty acid cycle” reported first by Shipp et al. (35) in the heart and then later linked with muscle and adipose tissue by Randle et al. (36) and referred to as the “Randle cycle.” Furthermore, increased levels of fatty acid intermediates can suppress the utilization of glucose directly by allosterically or covalently regulating enzymes of glucose metabolism.

Accordingly, in people with T1D who exhibited increased myocardial fatty acid utilization, there was also a simultaneous decrease in glucose utilization (25,26). Using noninvasive imaging modalities, decreased myocardial glucose metabolism was also found in people with T2D (37–40). Moreover, the limited glucose that was metabolized was preferentially converted to lactate rather than being oxidized in the mitochondria (28). This has been further confirmed using cutting-edge hyperpolarized MRS techniques, which demonstrated decreased flux of cytosolic pyruvate into mitochondrial acetyl CoA via pyruvate dehydrogenase (PDH) in people with T2D (41). In preclinical animal models, the changes in glucose metabolism closely recapitulate those seen in humans, whereby both glycolysis and glucose oxidation are downregulated in isolated perfused hearts from mice and rats with T2D (28,29,42,43). Complementing these ex vivo studies, in vivo measurements of pyruvate oxidation using hyperpolarized [1-¹³C]pyruvate also demonstrate decreased flux through PDH in both T1D and T2D rodent hearts (44,45).

Mechanistically, changes in sarcolemmal glucose transport regulate glucose entry to the heart and have been heavily implicated in the metabolic changes in diabetes. GLUT4 protein expression and sarcolemmal localization are both decreased in T1D and T2D animal models (46,47). A key node for metabolic control in the heart is mitochondrial PDH, which regulates the coupling between glycolysis and glucose oxidation. There is now unequivocal evidence that decreased myocardial PDH activity is a major factor in the robust impairment of glucose oxidation in DbCM. Posttranslational modifications (PTMs) play a major role in its regulation, with PDH kinase (PDK)-mediated phosphorylation suppressing PDH activity, whereas PDH phosphatase (PDP)-mediated dephosphorylation increases its activity. Increases in myocardial PPARα activity contribute not only to molecular increases in fatty acid oxidation but also to decreases in glucose oxidation via increased transcription of Pdk4 (34,48). Additionally, in mice with T2D, increases in myocardial PPARα activity may stimulate transcription of the mitochondrial calcium uniporter complex inhibitory subunit (MCUb), which impairs PDH activity by decreasing mitochondrial calcium levels (49). More recently, it was also demonstrated that the transcription factor forkhead box protein O1 (FoxO1) contributes to elevations in myocardial Pdk4 transcription and impaired PDH activity in obesity/T2D (50–52). Finally, increased acetyl CoA concentrations in the diabetic heart, generated from increased fatty acid oxidation, activate PDK4 and thereby suppress PDH activity, demonstrating one of the key nodes of regulation within the “Randle cycle.”

It is increasingly becoming recognized that cardiovascular pathologies are also associated with perturbations in branched chain amino acid and ketone metabolism (53), the former of which also contributes to the pathology of T2D. However, both have been comparatively understudied in the diabetic heart versus fatty acid and carbohydrate metabolism, while myocardial amino acid metabolism in general has been ignored. Nonetheless, it has been reported that branched chain α-ketoacid CoA dehydrogenase protein expression is decreased in hearts from genetically obese rats, suggestive of an impairment in myocardial branched chain amino acid metabolism (54).

With regard to myocardial ketone metabolism, contrasting findings have been reported. Studies in rodents and humans have demonstrated increased myocardial β-hydroxybutyrate levels and reduced expression and activity of the ketone oxidation enzyme β-hydroxybutyrate dehydrogenase 1 (BDH1) (55). Conversely, studies using hyperpolarized [3-¹³C]acetoacetate revealed increased myocardial ketone utilization in a genetic nonobese rat model of T2D (56). These discrepancies may be explained by the fact that acetoacetate oxidation is dependent not on BDH1 activity but, rather, on SCOT activity, which is elevated in T2D (56,57).

Phosphorus spectroscopy allows for measurements of cardiac energetics for understanding how changes in substrate flux ultimately impact myocardial ATP and phosphocreatine (PCr) concentrations. Studies in individuals with T2D have shown a decrease in the myocardial PCr-to-ATP ratio, which is further exacerbated on exercise (58,59). In animal models, decreased PCr-to-ATP ratios have been measured in vivo in mice with T2D, while decreases in both ATP and PCr concentrations have been reported in rats with T2D (60,61).

The decrease in myocardial energetics in diabetes is consistent with mitochondrial dysfunction, though the precise interplay between the aforementioned perturbations in substrate metabolism and mitochondrial dysfunction in DbCM remains unclear. Disturbances to the mitochondrial network have been identified, associated with imbalances in fission/fusion homeostasis, as mitochondrial content is increased but fragmented (3,62). In addition, there is increased mitophagy in response to high-fat feeding, which appears to be an adaptive response to ensure mitochondrial quality control (63). There are also increases in reactive oxygen species and oxidative stress, with strategies to prevent this including increasing catalase expression, inhibiting NADPH oxidases, or using antioxidants, all of which lead to improvements in mitochondrial and cardiac function in diabetes (64).

Increasing evidence suggests that in addition to their traditional roles as substrates for ATP production, certain metabolites function as substrates for and as modulators of nonmetabolic cellular processes. The relative excess or absence of certain metabolites in the diabetic heart dysregulates these metabolite-controlled processes, providing an alternative mechanism linking the metabolic perturbations to the functional changes within the diabetic myocardium. In the next section we describe selected examples of metabolites directly regulating epigenetics, PTMs, and transcription factors in DbCM (Fig. 4).

Epigenetic alterations refer to changes in chromosome structure or composition without changes to the nucleotide sequence, which can contribute to disease pathogenesis by changing gene expression through the reversible covalent modification of DNA or histones. DNA cytosine methylation is increased in DbCM in comparison with healthy cardiac tissue (65), and this has been linked to decreased α-ketoglutarate (αKG) synthesis in diabetes. αKG is a positive allosteric activator of the demethylation complex thymine DNA glycosylase–ten eleven translocation protein 1 (TDG-TET1), and decreased αKG production in diabetes limits DNA demethylation within the heart.

PTMs of histones have been implicated in changing the transcriptional landscape in diabetes by regulating chromatin remodeling. Histone acetylation is regulated by a family of acetyltransferase and deacetylase enzymes, which confer changes in transcription of target genes. Histone deacetylase 4 is regulated by increased glucose concentrations via flux through the hexosamine biosynthetic pathway, which can change gene expression in the diabetic heart (66). Histone lysine lactylation is a newly described epigenetic modification mediated by addition of a lactyl group derived from lactate (67), which has been shown to be elevated in skeletal muscle biopsies from insulin-resistant individuals. In these individuals, protein lactylation correlated with markers of anaerobic glycolysis and could be recapitulated through culturing of cells with elevated lactate or glucose (68). However, the specific myocardial proteins undergoing this modification in DbCM have yet to be determined.

As metabolites function as substrates for many protein PTMs, alterations in intermediary metabolism contribute to DbCM pathogenesis by inappropriately increasing or decreasing protein PTMs and detrimentally altering a protein’s structure, stability, localization, or intermolecular interactions. Elegant works by Bertrand and colleagues have identified posttranslational acetylation of proteins involved in protein trafficking as a driver for metabolic dysfunction in diabetes. Elevated acetyl CoA within the diabetic myocardium (47), derived from fatty acids, ketones, and ketogenic amino acids, drives acetylation of α-tubulin lysine resides, causing decreased GLUT4 translocation and myocardial glucose utilization (69,70). Given that the majority of intracellular acetyl CoA is located within the mitochondria, changes in mitochondrial protein acetylation driven by elevated fatty acid–derived acetyl CoA have also been reported in diabetes, modulating mitochondrial respiration and morphology (61,71).

Another example of metabolite-derived PTMs in diabetes is O-GlcNAcylation, which involves the addition of β-N-acetylglucosamine to proteins, synthesized via the hexosamine biosynthetic pathway from sequential reactions involving glucose, glutamine, and acetyl CoA. O-GlcNAcylation is increased in the rodent and human heart in diabetes and can be mimicked by increasing supply of glucose or glucosamine to cardiomyocytes (72,73). A multitude of proteins involved in diverse cellular processes undergo O-GlcNAcylation; thus, the current challenge remains to identify which of those modifications that are elevated in the diabetic heart are causative to pathology.

In addition to modifying a cell’s epigenome, metabolites influence gene expression by modulating several transcription factors. Stabilization of the hypoxia-inducible factor (HIF)-1α transcription factor has been shown to be suppressed in the ischemic diabetic myocardium by the presence of elevated fatty acid concentrations (31,47,74). Fatty acids indirectly regulate HIF-1α by suppressing production of the Krebs cycle intermediates succinate and fumarate during hypoxia, which are needed for HIF-1α stabilization. Transcription factors have also been identified as targets for O-GlcNAcylation, and one such target that is excessively O-GlcNAcylated in diabetes is the homeodomain transcription factor Nkx2.5, implicated in growth and repair (75). As mentioned previously, PPARs are master regulators of fatty acid metabolism that are activated by lipid intermediates, resulting in a vicious cycle of fatty acid influx driving increased capacity for fatty acid metabolism (76).

Numerous preclinical studies have demonstrated that correcting cardiac substrate metabolism perturbations can attenuate the progression of DbCM and improve outcomes. In this next section we provide examples of pharmacological approaches of directly manipulating these pathways (Fig. 3) for the treatment of DbCM in both animal and small-scale human studies.

Trimetazidine, an antianginal agent that decreases fatty acid oxidation secondary to an inhibition of 3-ketoacyl CoA thiolase, alleviates diastolic dysfunction in mice with high-fat diet–induced obesity (77). Similarly, clinical studies with trimetazidine treatment for 6 months in individuals with T2D and idiopathic dilated cardiomyopathy led to improvements in both systolic and diastolic function (78). Strategies to inhibit fatty acid uptake across the sarcolemma by blocking FAT/CD36 using sulfo-N-succinimidyl oleate decreased fatty acid metabolism in rodent models of T2D, upregulating glycolysis and improving postischemic cardiac function (47). Blocking fatty acid uptake with sulfo-N-succinimidyl oleate also restored HIF-1α activation and downstream hypoxic signaling in insulin-resistant cardiomyocytes (74). Demonstrating the integrated nature of metabolic control, HIF-1α activation with molidustat (a drug developed for the treatment of anemia) decreased myocardial fatty acid oxidation and TAG accumulation in T2D rats (31).

Targeting defects in myocardial glucose metabolism has also shown benefit in preclinical studies. Activating mitochondrial PDH using the pan-PDK inhibitor dichloroacetate increased in vivo myocardial glucose oxidation rates in T2D rats and alleviated diastolic dysfunction (44). Increased FoxO1 transcriptional activity also contributes to impaired myocardial PDH activity in T2D (50), as inhibition of FoxO1 with AS1842856 improved diastolic dysfunction in T2D mice (51). Of interest, AS1842856 treatment decreased Pdk4 expression and subsequent PDH phosphorylation, while these benefits were abolished in mice with a cardiac-specific PDH deficiency, emphasizing that increases in glucose oxidation were responsible for the improved cardiovascular outcomes. Similar observations have been reported in T1D-related cardiomyopathy, as AS1842856 treatment also improved cardiac function in streptozotocin-treated rats (79). Treatment with an aldose reductase inhibitor, AT-001 (currently in phase 3 clinical trials), improved diastolic dysfunction in mice with T2D, and while inhibiting aldose reductase prevents glucose metabolism to sorbitol, decreases in myocardial fatty acid oxidation were also observed (80). Of clinical relevance, glucagon-like peptide 1 receptor (GLP-1R) agonists improve cardiovascular outcomes in people with T2D, and increased myocardial glucose oxidation rates have been observed in mice with T2D following systemic treatment with liraglutide, along with improvements in diastolic function (42). Also, in a recent study treatment with the long-acting GLP-1R agonist semaglutide resulted in improvement in the Kansas City Cardiomyopathy Questionnaire Clinical Summary Score and 6-min walk distance in individuals with HFpEF but without diabetes (81). The question of whether the observed benefit is dependent on increases in myocardial glucose oxidation remains but is an intriguing avenue for further interrogation.

There have been limited studies investigating whether manipulating myocardial amino acid or ketone metabolism can improve outcomes in experimental DbCM. Dietary supplementation with lysine, leucine, and arginine decreased myocardial wall thickness, thereby attenuating cardiac hypertrophy in insulin-resistant rats (82). Ketone ester administration to T2D mice resulted in improved diastolic and systolic function, associated with improvements in mitochondrial respiration (83). Recently, there has been much interest in sodium–glucose cotransporter 2 (SGLT2) inhibitors as manipulators of ketone metabolism, with clinical data demonstrating that they also improve cardiovascular outcomes in people with T2D. Treatment with the SGLT2 inhibitor empagliflozin ameliorates diastolic dysfunction in mice subjected to high-fat diet–induced obesity (84); however, empagliflozin failed to increase myocardial ketone oxidation rates in mice with T2D (85). Thus, it remains unclear whether the beneficial effects of SGLT2 inhibition involve changes in ketone metabolism, or whether this has simply been a metabolic “red herring.”

It has often been posited that decreasing fatty acid oxidation or increasing glucose oxidation will improve the efficiency of contractile function, due to the lower O₂ cost of generating ATP from glucose versus fatty acids. Whether this is relevant in DbCM where some of the primary pathological features are diastolic dysfunction and cardiac fibrosis is difficult to ascertain. It would seem prudent to determine how decreasing or increasing fatty acid and glucose oxidation, respectively, impacts the molecular control of ventricular relaxation, which is a highly energy-dependent process (86). Furthermore, as correcting these metabolic perturbations can be associated with alleviation of cardiac fibrosis (51), it would also be relevant to determine whether changes in fatty acid and/or glucose oxidation in cardiomyocytes or fibroblasts directly impacts fibrogenesis. Another key area of consideration revolves around the previously described nonenergetic aspects of metabolic intermediates and PTMs, as they may influence signaling and biological processes. For example, titin is a major protein of the cardiac sarcomere susceptible to several PTMs that regulates diastolic function. Understanding of how dysregulated metabolism influences the cardiac sarcomere may pave the way for novel areas of investigation in diabetes.

As highlighted throughout this article, a plethora of preclinical and clinical studies support that DbCM is characterized by several perturbations in cardiac substrate metabolism and the presence of diastolic dysfunction. However, there have been limited studies comparing DbCM with HFpEF in diabetes. Research in this area is needed to address 1) whether DbCM is predictive of or a precursor to HFpEF in T2D, 2) whether effective management of DbCM at its onset will decrease the prevalence of HFpEF, and 3) whether DbCM is truly its own unique clinical entity. In recent studies investigators identified a stepwise decrease in diastolic function and PCr-to-ATP ratio when comparing control subjects, patients with T2D, and patients with HFpEF, but exercise only induced transient pulmonary congestion in the HFpEF patients (87). This would suggest that changes in energy metabolism are following a trajectory via DbCM to HFpEF; however, the number of HFpEF patients with diabetes in this study was limited.

What is also lacking is a clear and consistently applied definition of DbCM. It has been suggested that with an improved understanding of the pathology that characterizes DbCM, perhaps the condition should be retermed “diabetic heart disease” (3). A limitation of this approach is that the term diabetic heart disease could be falsely construed to include all cardiovascular diseases associated with diabetes (not just those affecting the myocardium but also the vascular diseases). Therefore, our recommendation is that the term DbCM simply be redefined: we propose the definition “diastolic dysfunction in the presence of altered myocardial metabolism in a person with diabetes but absence of other known causes of cardiomyopathy and/or hypertension.” While advancements in noninvasive imaging technologies such as ¹³C hyperpolarized MRS and positron emission tomography imaging allow for assessment of myocardial glucose oxidation and fatty acid oxidation, respectively, these are unlikely to see routine use in clinical management in the foreseeable future. Thus, validation of metabolic biomarkers reflective of the perturbations in cardiac energy metabolism may serve as a more feasible approach for diagnosing DbCM based on this definition. Additionally, further research to characterize the metabolic phenotype in patients with diabetes with HFpEF will assist in unraveling the relationship between DbCM and HFpEF. Nonetheless, even this new definition cannot be used to universally distinguish DbCM, as there will be individuals with diabetes and systolic dysfunction more reminiscent of a heart failure with reduced ejection fraction phenotype. Why one individual with DbCM may develop diastolic dysfunction, whereas another may develop systolic dysfunction, remains unknown and, until appropriately designed larger population studies are pursued, will remain an enigma in the field of cardiovascular endocrinology.

Taken together, it is increasingly recognized that DbCM is more prevalent in people with diabetes than previously accepted, and DbCM is often present but undiagnosed in people with prediabetes or early-stage T2D. Cardiovascular outcomes trials (CVOTs) for two glucose-lowering medications, the SGLT2 inhibitors and GLP-1R agonists, reported decreased cardiovascular events in people with T2D (88,89). Despite these promising observations, it is important to note that the majority of individuals with diabetes included in CVOTs have established macrovascular cardiovascular disease and have had diabetes for years. Conversely, people with prediabetes or the early stages of T2D are underrepresented in CVOTs, and this population will need to be considered in future studies.

Funding. This work was supported by funding from the British Heart Foundation to L.C.H. (FS/17/58/33072) and to N.S. [FS/4yPhD/F/21/34160, Oxford 1st intake – British Heart Foundation 4 Year PhD Studentship (5th) Scheme, the McKie McLean Class of 2021] as well as a Canadian Institutes of Health Research Project Grant (PJT-159648) and a Diabetes Canada End Diabetes Award (OG-3-22-5606-JU) to J.R.U. J.R.U. is a Tier 2 Canada Research Chair (Pharmacotherapy of Energy Metabolism in Obesity).

Duality of Interest. No potential conflicts of interest relevant to this article were reported.