Circulating Cardiac Biomarkers in Diabetes Mellitus: A New Dawn for Risk Stratification-A Narrative Review.

The aim of this narrative review is to update the current knowledge on the differential choice of circulating cardiac biomarkers in patients with prediabetes and established type 2 diabetes mellitus (T2DM). There are numerous circulating biomarkers with unconfirmed abilities to predict clinical outcomes in pre-DM and DM individuals; the prognostication ability of the cardiac biomarkers reported here has been established, and they are still being studied. The conventional cardiac biomarkers, such as natriuretic peptides (NPs), soluble suppressor tumorigenisity-2, high-sensitivity circulating cardiac troponins and galectin-3, were useful to ascertain cardiovascular (CV) risk. Each cardiac biomarker has its strengths and weaknesses that affect the price of usage, specificity, sensitivity, predictive value and superiority in face-to-face comparisons. Additionally, there have been confusing reports regarding their abilities to be predictably relevant among patients without known CV disease. The large spectrum of promising cardiac biomarkers (growth/differential factor-15, heart-type fatty acid-binding protein, cardiotrophin-1, carboxy-terminal telopeptide of collagen type 1, apelin and non-coding RNAs) is discussed in the context of predicting CV diseases and events in patients with known prediabetes and T2DM. Various reasons have been critically discussed related to the variable findings regarding biomarker-based prediction of CV risk among patients with metabolic disease. It was found that NPs and hs-cTnT are still the most important tools that have an affordable price as well as high sensitivity and specificity to predict clinical outcomes among patients with pre-DM and DM in routine clinical practice, but other circulating biomarkers need to be carefully investigated in large trials in the future.

Key Summary Points

Introduction

Diabetes mellitus (DM) is a common metabolic disorder worldwide and has reached epidemic levels [1]. Type 2 DM (T2DM) comprises about 90% of the cases of the disease and affects 8.3% of the entire adult population [2]. It has been suggested that by 2035 about 592 million people will die of DM, and it will be one of the leading causes of mortality in the general population [3]. Although there are several well-established methods for the prevention and treatment of DM, the disease has been considered a powerful risk factor for cardiovascular (CV) disease, but evidence for a close relationship between various types of prediabetes, such as metabolically healthy obesity and metabolic syndrome, is still conflicting [4-6]. However, T2DM is closely linked to a substantial increase in all-cause mortality and CV mortality in the general population [7]. Therefore, DM patients with known CV diseases, including atherosclerosis, stable coronary artery disease (CAD), acute coronary syndrome/myocardial infarction (MI), heart failure (HF), arrhythmia (atrial fibrillation and flutter) and cardiomyopathies, have a higher risk of death than non-diabetics with CV diseases [8-12].

In 2019, the European Society of Cardiology (ESC) announced a new clinical guideline on diabetes, pre-diabetes and CV diseases [13]. This recommendation contains a section on circulating cardiac biomarkers, which are promising prognostic indicators for CAD, HF and major adverse CV events (MACEs) in DM patients. Notably, T2DM and diabetes-induced target organ damage have been considered factors that hinder clinical interpretations of circulating biomarkers’ peak levels [14]. For instance, abdominal obesity in diabetics was associated with increased endogenous activity of neprilysin and thereby decreased the expected levels of circulating natriuretic peptides (NPs), which were previously proposed as powerful diagnostic and accurate prognostic biomarkers in HF [15]. Nowadays, there is an appropriate correction for the NP discriminative value in DM to calculate CV risk [16]. On the other hand, at the early stage of DM in elderly people without CV disease, the circulating levels of soluble suppressor tumorogenicity-2 were significantly increased compared with healthy elderly volunteers [17]. These data need to be evaluated carefully to prevent both over- and underestimation of CV disease risk and inadequate care [18]. Finally, multiple cardiac biomarker measuring strategies have been proposed as tools to improve the sensitivity and specificity of conventional diagnostic and predictive models, and serial measurements of circulating cardiac biomarkers have been used in clinical practice. The current recommendations of the ESC and American Heart Association seriously differentiate in this setting [19]. The aim of the narrative review is to update the current knowledge on the differential choice of circulating cardiac biomarkers in patients with prediabetes and established T2DM. This article is based on previously conducted studies and does not involve any studies with human participants or animals performed by any of the authors.

Type 2 Diabetes Mellitus and CV Risk

CV risk in T2DM patients is caused by several mechanisms that are associated with the development of micro- and macrovascular dysfunction, accelerating atherosclerosis, an impaired endogenous repair system, direct cell metabolism impairment, oxidative stress, vascular and systemic inflammation, inadequate immune response, cardiac biomechanical stress, fibrosis, necrosis and apoptosis as well as thrombophilia and aggregation of circulating blood cells [20-25]. In fact, from the early stages of prediabetes, traditional CV risk factors (hypertension, smoking, obesity and dyslipidemia) are present in T2DM patients and have an association with CV death in the diabetic population [24, 25]. Not all cases of HF resulting from the progression of T2DM are associated with primary diabetes-induced metabolic impairments and ventricular hypertrophy [26-28].

T2DM patients have increased mortality and CV complications with a > 4-fold risk of death and CV among patients without T2DM [29]. However, there are differences in the all-cause mortality and CV death rates between T2DM patients and non-T2DM patients, and preventive and treatment approaches have been implemented over the last decade [20]. Additionally, the all-cause mortality rate among diabetics, as compared with that in the general population, is influenced by older age, poor glycemic control and renal complications, but these factors are not independently related to death due to CV disease and HF [29]. In this context, biomarkers reflecting CV risk appear to be promising for risk stratification and probably improve point of care in diabetics with unknown CV disease.

Cardiac Biomarkers in T2DM

Evidence shows that conventional CV risk factors have a negative influence on the mortality rate and quality of life of T2DM patients, and it is suggested that cardiac biomarkers reflecting various pathophysiologic stages of cardiac remodeling, such as biomechanical stress, inflammation, necrosis/apoptosis, fibrosis, hypertrophy and extracellular matrix remodeling, would have an incremental add-on value for the prediction of clinical outcomes (death, MACEs, hospital admission, HF onset) in the patient population. Moreover, measurement of circulating levels of cardiac biomarkers can demonstrate new individual predictive information that could have great predictive power beyond conventional CV risk factors. However, each biomarker has strengths and weaknesses, which affect the cost, specificity, sensitivity, predictive value and superiority in a face-to-face comparison. Because there are numerous circulating biomarkers with unconfirmed abilities to predict clinical outcomes in pre-DM and DM individuals, we here report on cardiac biomarkers whose prognostication value has been established and are still being studied. The utility of circulating cardiac biomarkers in patients with T2DM is reported in Table 1.

Table 1

Utility of circulating cardiac biomarkers in patients with prediabetes and T2DM

Pathogenetic condition	Biomarkers	Relation to CV risk in patients with prediabetes and T2DM	References
Cardiac biomechanical stress	NPs (atrial NP, NT-proANP, brain NP, NT-proBNP)	Independent predictors of new-onset CAD, and MACE	[46, 48, 53, 63]
		Independent predictors of HF	[49, 51, 54]
		Independent predictors of AF and sudden death	[52]
		Micro- and macrovascular complications	[53, 57, 58, 63]
		Predictors of adverse cardiac remodeling	[55]
Cardiac myocyte necrosis	hs-Tn I/T	Independent predictors of MACE	[46, 74, 84, 87]
		Predictors of CV death and HF	[70, 73, 81]
		Predictors of T2DM-induced CMP	[81, 87]
		Predictors of renal outcomes	[85, 86]
	H-FABR	Predictors of long-term mortality and re-infarction	[95]
		Predictor of premature death	[99]
		Predictor of asymptomatic cardiac ischemia	[101–103]
Inflammation	sST2	Predictor of CV disease and CV mortality	[111]
		Predictor of HF and HF-related outcomes	[110]
		Predictor of CV risk and mortality	[113–117]
	GDF15	Prediction of CV risk and mortality	[124, 138]
		Prediction of new-onset T2DM	[125, 127]
		Prediction of MACEs in ACS/MI	[131, 132]
		Prediction of HF and HF-related outcomes	[133, 134, 137]
		Prediction of T2DM-induced CMP	[139]
Fibrosis	Galectin-3	Prediction of T2DM-induced CMP	[155]
		Prediction of MACEs and all-cause mortality	[156, 157]
Myocardial hypertrophy	Cardiotrophin-1	Prediction of T2DM-induced CMP	[173, 174]
Extracellular matrix remodeling	Extracellular matrix biomarkers	Prediction of HFpEF, HF-related outcomes, MACEs	[175–180]

NPs natriuretic peptides, sST2 soluble suppressor tumorogenicity-2, CAD coronary artery disease, GDF15 growth/differential factor-15, ACS acute coronary syndrome, MI myocardial infarction, MACE major adverse cardiac events, CMP cardiomyopathy

Natriuretic Peptides

Natriuretic peptides have immense systemic homeostatic effects, playing a pivotal role in the regulation of natriuresis, electrolyte and water retention, vascular permeability and vasodilation, cardiac contractility and blood pressure changes; consequently, NPs are physiologic antagonists of the renin-angiotensin-aldosterone and sympatho-adrenal systems [30]. Several types of NPs are released predominantly from the myocardium, atrial (ANP) and brain (BNP) NPs, and vessels, bone and brain (C-type NP) [31]. Physiologic effects of NPs cause binding to the extracellular domains of the appropriate receptors, NPR-A, NPR-B and NPR-C. NPR-A and NPR-C are widely expressed on the surfaces of target cells and involved in the regulation of NP bioavailability independently from circulating neprilysin activity [32]. Synthesis and secretion of NPs, predominantly ANP and BNP, are carried out in response to myocardial stretching and fluid overload, while several stimuli have direct and indirect impacts on NP production, accumulation and secretion, such as ischemia/hypoxia, inflammation, hormones (catecholamines, aldosterone, renin) and growth factors (transforming growth factor-beta, vascular endothelial growth factor) [33]. The C-type NP is a locally produced peptide, which acts as an autocrine regulator of vascular function, bone ossification and development of the nervous system.

Additionally, NPs suppress the lipolytic activity of adipocytes through attenuation of adipose tissue-expressed NPR-A and NPR-C [34]. Nevertheless, NPs increasing p38 MAP kinase in brown adipose tissue cells cause overexpression of “browning” genes ensuring upregulation of energy expenditure and adaptive thermogenesis [35]. NPs are involved in transcriptional regulation of genes, which are responsible for mitochondrial biogenesis, uncoupled respiration (PPARγ coactivator‐1α and uncoupling protein 1), lipid oxidation, GLUT-4 synthesis and insulin sensitivity in various human cells, including adipocytes, skeletal muscle cell, myocardium, vasculature smooth muscle cells, endothelial cells and hepatocytes [36, 37]. In fact, the NPR-A signaling pathway in skeletal muscle cells and hepatocytes is crucial for the metabolic memory phenomenon and the change of pre-diabetes to T2DM [38, 39].

Previous observational and clinical studies have yielded evidence of altered clearance of NPs and impaired activity of neprilysin in patients with abdominal obesity, metabolic syndrome and T2DM in connection with fasting glucose impairment and insulin resistance [40, 41]. However, there are controversial findings related to the ability of circulating insulin to reciprocally regulate NPR-C expression on the surface of adipose tissue cells in obese individuals [42-44]. Therefore, patients with diabetes-induced nephropathy had increased circulating levels of BNP and NT-pro-BNP compared with those who did not have diabetes renal disease [45]. Overall, the primary cause of the fluctuation of circulating NP levels among patents with metabolic disease is not clear.

Current clinical guidelines recommend measuring NP levels to diagnose HF, stratify patients at higher CV risk including HF onset risk and predict short-term re-admission to the hospital because of HF decompensation [46, 47]. Asymptomatic and symptomatic HF patients can be stratified as at risk of death from any cause and CV disease if they have high circulating levels of NT-proBNP > 125 pg/ml or > 300 pg/ml, respectively [47]. In fact, increased age was associated with diagnostic cutoff points for the above-mentioned NT-proBNP upper values [45, 48]. Surprisingly, HF patients without T2DM with higher levels of NPs have shown much more predictive accuracy for MACEs, CV mortality and HF manifestation than those who have T2DM [46, 49].

Interestingly, women with HF with a preserved ejection fraction (HFpEF) had higher levels of NT-proBNP and consequently CV mortality risk than males with HFpEF [50]. HFpEF patients with T2DM had more ventricular hypertrophy and adverse cardiac remodeling compared with non-T2DM patients, while systolic and diastolic myocardial function and serum levels of NT-proBNP did not differ [51]. There is evidence of NPs in HF in reduced ejection fraction (HFrEF) patients with prediabetes/T2DM independently predicting atherosclerosis, atrial fibrillation, pulmonary hypertension and sudden death [52-54]. Overall, the NT-proBNP level predicted cardiac abnormalities and CV events regardless of glucose status, and multiple biomarker models are required to improve the predictive accuracy for HF [53-57].

Controversies Related to the Predictive Value of NP in Pre-Diabetics and Diabetics

There are some controversies related to the variability of predictive values of NPs in patients with abdominal obesity, metabolic syndrome and T2DM. NPs are similar to cardiac troponins in the prediction of micro- and macrovascular complications in pre-diabetics without known CV disease [58]. Among dysmetabolic patients with HF, adipocytokines (adiponectin, resistin, chemerin, leptin, visfatin) have exhibited additive discriminative power to NT-proBNP for MACEs regardless of glucose status [59-61].

There is a close inverse association between the number of metabolic syndrome components and circulating levels of NPs in middle-aged and elderly individuals, but not among young people without CV disease [62]. Overall, NPs better predict CV outcomes in patients with known HF than among individuals without HF independently of prediabetes and T2DM [63].

NPs in HFrEF Patients with Prediabetes and DM

It had been noted that HFrEF patients with prediabetes and T2DM treated with glucagon-like peptide-1 [GLP-1] analog (liraglutide) [64] and sodium-glucose co-transporter-2 [SGLT2] inhibitor (empagliflozine, dapaglyflozine) [65, 66] with benefits in CV outcomes have demonstrated a decrease in serum levels of NT-proBNP. In contrast, among non-HF patients with prediabetes/T2DM, serum levels of NT-proBNP remain unaltered despite improved glucose homeostasis and decreased CV risk [67]. Interestingly, the change in NT-proBNP serum levels correlated negatively with baseline levels of NT-proBNP in T2DM patients [68]. Additionally, in the DEFINE-HF Trial, the SGLT2 inhibitor dapagliflozin did not affect the mean NT-proBNP serum levels, but increased the proportion of patients (diabetics and non-diabetics) experiencing clinically meaningful improvements in HFrEF-related clinical status [69]. These facts require elucidation in large clinical studies in the future.

Overall, NPs continue to be the most important tool for identifying pre-diabetics and diabetics at CV risk that are affordable. The highest sensitivity and specificity were established for patients with HF symptoms, but for asymptomatic individuals the predictive accuracy of NP levels is superior to traditional CV risk factors. Probably the use of serial measurements of circulating levels of NPs in DM patients with HF treated with SGLT2 inhibitors can be disputed as surrogate markers with possible predictive value.

Cardiac Troponins

Cardiac troponins are established biomarkers of myocardial injury and necrosis, and, even below the 99th percentile, they strongly predict adverse outcomes in prediabetes and T2DM patients. There is evidence showing that patients with prediabetes and T2DM may develop asymptomatic myocardial damage beyond obvious ischemic causes [70, 71]. The primary causes that lead to increased permeability of cell membranes, leakage of a cytoplasmic pool of cardiac troponins and onset of small-sized myocardial necrosis in prediabetes and diabetes patients are lipotoxicity and myocardial steatosis, which influence the biomechanical myocardial stress, low-grade inflammation, oxidative stress, endoplasmic reticulum stress and mitochondrial stress, altered reparation due to the metabolic memory phenomenon and impaired intracellular metabolism [71, 72].

Micro- and macro-vasculopathies in prediabetes and diabetes are associated with accelerating atherosclerosis, plaque shaping, development of endothelial dysfunction and microvascular obstruction, which induce diabetes-related cardiomyopathy [72, 73]. However, a T2DM-induced significant decrease in sirtuin-1 and hypoxia-inducible factor (HIF)-1α expression in the myocardium as well as declining circulating levels of orexin B may contribute to ischemia-reperfusion injury and exacerbate the cardiac and vascular dysfunctions [74].

Therefore, there are extra-cardiac causes that increase high-sensitivity circulating troponin (hs-cTn) T/I levels. For instance, the lowered estimated glomerular filtration rate (per each 15 ml/min/1.73 m2 lower) showed an independent association with a steeper hs-cTnT increase [75]. All of these conditions yield an increased risk of CV death and CV events including MACEs and HF [46, 76, 77].

Patients with metabolic syndrome have demonstrated higher hs-cTnI levels than those who do not have the condition, but there were no significant differences in BNP serum levels [78]. Additionally, the levels of hs-cTnI significantly correspond to the presence and components of metabolic syndrome [78]. No significant difference was found in changes in hs-cTnT/I between non-ST elevation myocardial infarction patients with and without T2DM [79, 80]. However, elevated hs-Tnl levels were found to be independent predictors of MACE in individuals with known CAD regardless of glucose metabolism status [46, 79, 81, 82] as well as among patients with prediabetes and T2DM without known CAD [83, 84]. Interestingly, in several large clinical trials, such as SAVOR-TIMI 53 and TECOS, elevated levels of hs-cTnT in patients with T2DM during treatment with antidiabetic drugs (inhibitors of dipeptidyl peptidase 4 saxagliptin and sitagliptin) were significantly associated with renal outcomes rather than CV events and MACEs [85, 86], but in diabetics with CV disease, i.e., myocardial infarction, the hs-cTnT levels corresponded positively to CV death, MACEs and HF [87].

Considering the high availability of troponin measurements, affordability of the test and relatively high sensitivity and specificity for prediction of CV events, hs-cTnT/Is are promising for risk stratification of patients with pre-DM and DM with known CV diseases.

Heart-Type Fatty Acid-Binding Protein

Heart-type fatty acid-binding protein (H-FABR) is a novel serum biomarker of early myocardial ischemia and injury that has been recently reported to be related to CV diseases [88], acute myocardial infarction (MI) [89, 90] and long-term post-MI prognosis [91]. H-FABR is rapidly released into the circulation from cardiac myocytes after non-selective increased cell membrane permeability and myocardial injury [89]. H-FABR is more sensitive than conventional biomarkers (myoglobin and hs-cTnT/I) to diagnose acute MI [90]. Circulating levels of H-FABP were also found to be higher in prediabetes [92-94] and T2DM patients [95, 96] than in those who did not have DM. Although H-FABP levels predicted subclinical myocardial injury or subclinical atherosclerosis in patients with prediabetes and T2DM [97, 98], its predictive value for CV risk in patients with impaired glucose metabolism without established CAD is unclear. It has been suggested that higher H-FABP levels in asymptomatic patients at the early stages of metabolic disorders may reflect silent myocardial damage and susceptibility to HF development and the risk of premature cardiac death [99]. This evidence reflects the fact that overexpressed H-FABP in the sub-intima can induce multiple pathways of inflammation, growth and migration of vascular smooth muscle cells and thereby influence in-stent restenosis in the culprit coronary artery after percutaneous coronary intervention (PCI) [100]. Unfortunately, circulating H-FABP levels were found to be similar in T2DM patients without CAD and non-T2DM individuals [101]. Moreover, H-FABP was not a better independent powerful diagnostic biomarker when used alone than traditional CV risk factors [101, 102]. In fact, the strength of the biomarker is its ability to diagnose the early period of asymptomatic cardiac ischemia in T2DM patients with diabetic ketoacidosis and diabetic ketosis, and it is also inexpensive [103]. Large clinical trials are required to identify the predictive ability of the biomarker in a face-to-face comparison with other biomarkers in prediabetes and DM patients.

Soluble Suppressor Tumorogenisity-2

The soluble form of suppressor tumorogenisity-2 (sST2) acts as a decoy receptor of interleukin (IL)-33, inhibiting the effects of IL-33/ST2 ligand signaling, and it is produced by endothelial and epithelial cells, fibroblasts and certain immune cells in response to biomechanical stress, ischemia/necrosis, hypoxia and inflammatory cytokines [104]. In fact, the IL-33/ST2/sST2 axis is a core component of the autocrine/paracrine mechanism acting to prevent tissue injury [105, 106].

Elevated levels of sST2 were not found to be a specific diagnostic biomarker for a single disorder in humans, but increased serum concentrations of sST2 were linked to the progression of atherosclerosis [107], myocardial dysfunction [108], fibrosis and adverse cardiac remodeling [109], poor clinical outcomes in CV diseases including HF and atrial fibrillation [110] and metabolic disorders including diabetes mellitus and metabolic syndrome [111]. In fact, serum sST2 was measured in higher concentrations in T2DM patients, and the presence of left ventricular hypertrophy and diastolic and systolic cardiac dysfunction was associated with even higher sST2 levels [108]. sST2 has been validated as a predictive biomarker for CV disease and CV events, including HF [48]. Therefore, sST2 independently predicted the no-reflow phenomenon in STEMI patients undergoing primary PCI [112].

Previous clinical studies have shown that sST2 levels are strongly associated with several markers of T2DM including glycosylated hemoglobin, triglyceride levels, fasting glucose, HOMA-IR, ectopic fat accumulation and the glomerular filtration rate, and the levels in women are lower than in men, but the sST2 concentration increases with age [113, 114]. Elevated levels of sST2 are not an independent predictor of mortality and MACE in diabetics with acute coronary syndrome (ACS) and MI [115], whereas in ACS/MI patients without T2DM elevated levels of sST2 are independently associated with a risk of early in-hospital death, 30-day death and HF onset [116, 117].

Thus, most investigations have shown that sST2 levels are higher in patients with CV disease who also have either prediabetes or T2DM and that this association has independent predictive value for prognosis. The weakness of this biomarker is that that associations between serum levels of sST2 and other CV biomarkers (including NPs) and CV-related events in dysmetabolic individuals without CV disease have not been carefully studied, and the predictive power of sST2 beyond conventional CV risk factors requires confirmation in the future.

Growth Differentiation Factor 15

Growth differentiation factor 15 (GDF15), also known as macrophage-inhibiting cytokine 1, belongs to the transforming growth factor-ß (TGF-ß) superfamily [118]. GDF15 is released from a wide range of cells, such as mononuclear cells, macrophages, cardiac myocytes and adipocytes, under inflammatory conditions and oxidative stress [119, 120]. The main biologic role of the GDF15 is to regulate the inflammatory response, growth, cell differentiation, energy homoeostasis and weight loss. In fact, GDF-15 protects target tissues (myocardium, kidney, adipose tissue and vasculature) by several intracellular molecular pathways, such as inhibiting c-Jun N-terminal kinase, Bcl-2-associated death promoter, epidermal growth factor receptor and activating SMAD, endothelial NO synthase and phosphoinositide 3-kinase/AKT signaling pathways [121]. There is evidence of a role of GDF15 in oxidative stress, protein glycation, inflammation, cellular senescence and hormonal deregulation in aging and age-depending diseases [122, 123].

Serum levels of GDF15 were found to be higher in prediabetes/T2DM patients than in healthy volunteers and independently associated with CV risk scores [124], body mass index, waist-to-hip ratio [125], insulin resistance [126], hs-CRP [127] and parameters of glucose metabolism (C-peptide, fasting pre-hepatic beta cell function, impaired fasting glucose) [128, 129], while successful glycemic control did not cause a decrease in GDF15 levels [130].

Previous clinical studies have shown elevated levels of GDF15 are independently related to adverse cardiac remodeling and poor prognosis in ACS/MI [131, 132], HF [133], atrial fibrillation [134], renal dysfunction [135] and cachexia [135, 136]. Moreover, serial measurements of GDF15 have shown that an increase in GDF-15 over 1 year was independently associated with higher risks of future CV mortality beyond the NYHA functional class, left ventricular (LV) ejection fraction (EF) and circulating levels of NT-proBNP [137]. Additionally, elevated levels of GDF15 were strongly associated with an increase of the all-cause mortality rate in patients with atherosclerosis, such as angiographically proven CAD and peripheral artery disease, regardless of T2DM presentation [138].

Interestingly, elevated levels of the GDF15 over the cutoff point of 3812 pg/ml predicted T2DM-induced cardiomyopathy in the absence of other CV risk factors, such as age, smoking, hypertension and known CV disease [139]. Moreover, some antidiabetic drugs, such as metformin, promote their cardioprotective effects through GDF15 expression in target tissue [140]. Having strong evidence that GDF15 expression in multiple tissues is higher in prediabetes and T2DM patients than in individuals without metabolic disorders [141, 142], it has been suggested that GDF-15 may be a promising biomarker for identification of people at risk of metabolic-induced events (T2DM-induced cardiomyopathy) and CV disease/events. It can be of great clinical importance because a highly anticipated new class of GFRAL (receptor for GDF15)/RET (receptor tyrosine kinase)-based drugs for the treatment of abdominal obesity and metabolic syndrome may mediate the endogenous effects of GDF15 and thereby improve CV risk in individuals with metabolic diseases [143, 144].

Galectin-3

Galectin-3 (Gal-3) is a versatile protein that belongs to the lectin family and has been implicated predominantly in cardiac, liver and kidney fibrosis and inflammation [145, 146]. Overexpression of Gal-3 is associated with accumulation of advanced glycation end products (AGE), oxidative stress products (3-nitrotyrosine protein, superoxide radicals) and activation of the pro-apoptotic c-Jun-N-terminal kinase 1/2 stress signaling pathway, which directly influences the development of endothelial dysfunction and altered vascular reparation [147, 148].

Mediating profibrotic pathways, Gal-3 predicts cardiac remodeling and CV events that are independently related to it, such as HF and atrial fibrillation [149, 150]. Among asymptomatic adults from the general population, the highest quartile of Gal-3 levels was closely associated with two-fold increased odds of myocardial dysfunction compared with the lowest quartile of the biomarker [151]. Additionally, elevated levels of GDF15 were related to methylated arginine and hs-CRP in patients with prediabetes and T2DM without known CAD [152].

Based on previous clinical studies, Gal-3 has served as a prognostic clinical biomarker in HF [48], but its role in the prediction of T2DM was uncertain until the end of the Dallas Heart Study [153]. Gal-3 levels in the trial were associated with the incidence of T2DM even after adjustment for conventional metabolic and CV risk factors (age, gender, race, body mass index and hypertension) and renal function. Therefore, there were correlations between Gal-3 levels and circulating levels of inflammatory biomarkers (hs-CRP, IL-18, monocyte chemoattractant protein 1, soluble tumor necrosis factor receptor 1-alpha and myeloperoxidase), insulin secretion biomarkers (C-peptide), the homeostatic model assessment for insulin resistance and subcutaneous adiposity. Interestingly, in patients with abdominal obesity and prediabetes without established CAD, increased levels of the GDF15 were correlated with LV diastolic dysfunction and elevated pulmonary artery systolic pressure, but not with LV mass [154]. Elevated Gal-3 levels were associated with diminished global longitudinal strain in diabetics [155]. There is confirmation of a close association between elevated levels of GDF15 in prediabetic individuals and patients with T2DM and a risk of vascular calcification, plaque formation and endothelial dysfunction, corresponding to a risk of MACEs and all-cause mortality [156-158].

There are data elucidating the possible interrelation between dynamic changes in circulating levels of GDF15 and CV risk in T2DM patients treated with antidiabetic drugs [159, 160]. Serum Gal-3 levels in T2DM patients showed a modest increase from the baseline with the SGLT2 inhibitor canagliflozin versus placebo, whereas the concentrations of both NT-proBNP and hs-cTnI for > 2 years have demonstrated a tendency to decrease [161]. Large clinical trials are required to elucidate a strategy for GDF15-guided therapy of T2DM without CV disease and with established CV disease including HF. Finally, Gal-3 is a promising biomarker with high sensitivity and specificity that identifies preclinical metabolic heart disease and stratifies patients at risk of CV death and CV events including HF. The economic aspect of using this biomarker requires further elucidated.

Cardiotrophin-1

Cardiotrophin-1 (CT-1) is an adipocytokine that belongs to the IL-6 family and realizes its biologic effect by binding to the gp130 receptor [162]. CT-1 and its receptors are expressed in many tissues including the myocardium, brain, kidney, skeletal muscles, adipose tissue, liver, lung and testes and acts as an endocrine and paracrine regulator of physiologic and pathophysiologic functions [163]. The main source of CT synthesis is adipocytes, and down- and upregulated CT-1 gene expressions were found in white and subcutaneous adipose tissue, respectively [164, 165]. CT-1 decreases fasting glucose in an insulin-independent manner, mediates increased insulin sensitivity through the AKT-dependent pathway in skeletal muscles, reduces food intake, stimulates lipolysis and enhances energy expenditure [166].

It has been suggested that in patients with abdominal obesity, metabolic syndrome and T2DM adipose-derived CT-1 ensure metabolic circadian rhythms and adipose core clock genes [167] and promote tissue-protective effects including increased resistance of cardiac myocytes to hypoxia/ischemia, growth and differentiation of progenitor cells of different origin, reducing [163, 168]. Indeed, lowered circulating levels of CT-1 were associated with decreased risk of metabolic syndrome and T2DM in overweight patients [169]. Moreover, the CT-1 level was inversely correlated with the severity of obesity in non-T2DM patients [170].

A large body of evidence shows that the changes in LV geometry and development of LV hypertrophy and systolic and diastolic HF are related to lowered expression of CT-1 receptors in the heart and increased circulating levels of CT-1 in the peripheral blood [171, 172]. Additionally, the circulating level of CT-1 predicts the risk of T2DM manifestation, T2DM-induced target organ damage and CV complications regardless of classic CV risk factors [173, 174]. Although CT-1 has possible predictive ability for LV hypertrophy, T2DM-induced cardiomyopathy and HF appear to be promising; large clinical trials are required to ascertain whether this biomarker has independent discriminative power and is affordable to use in a face-to-face comparison with other outcome indicators.

Future Biomarkers

Collagen Turnover Biomarkers

There is a wide spectrum of biomarkers, which reflects several stages of TDM pathogenesis and could predict CV risk. For instance, carboxy-terminal telopeptide of collagen type 1 was measured in elevated concentrations in peripheral blood among diabetics compared with healthy volunteers and individuals with metabolically healthy obesity [51]. This is a profibrotic biomarker with predictive value for CV events that also is found in elevated levels in patients with HFpEF [175].

Other collagen biomarkers, such as procollagen type III N-terminal propeptide and collagen type I carboxy-terminal telopeptide, also appeared to be related to incident HFpEF, but not HFrEF [176]. Whether these pro-peptides, which reflect collagen synthesis and degradation, can be used to prognosticate CV risk in patients with metabolic disease is not fully understood, but a growing body of evidence suggests that they may play a pivotal role in point of care in patients with HFpEF and possibly HFrEF regardless of glucose impairment [177].

Biomarkers of Extracellular Matrix Remodeling

Extracellular matrix remodeling biomarkers, such as matrix metalloproteinase (MMP)-1, MMP-6 and MMP-9, and their tissue inhibitors, osteoprotegerin and osteopontin, are promising indicators of adverse cardiac remodeling and MACEs in the post-MI period, but they are not specific to T2DM-induced cardiac damage and correspond well to conventional CV risk factors including chronic kidney disease, aging and hypertension [178-180].

Apelin

Another promising biomarker for LV hypertrophy and HF is apelin, which has demonstrated vasodilator, inotropic and aquaretic properties and is a physiologic antagonist of RAAS [181, 182]. It has been shown that the development of DM-induced cardiomyopathy is attenuated by preventing mitochondrial dysfunction through the Apelin/Sirt3 pathway [183]. The predictive role of apelin in the Apelin/Sirt3 pathway for HF in DM patients is under investigation.

Non-Coding RNAs

Non-coding RNAs, including microRNAs (miRNAs), as well-established powerful regulators of posttranscriptional gene expression, could be potential biomarkers for CV risk in patients with prediabetes and DM [184]. Although there are organ-specific miRNAs, expression of which are highly up- or downregulated in HF, there are no clear advantages to the signature of circulating cell-free miRNAs and microvesicle-derived miRNAs compared with traditional CV risk biomarkers and HF biomarkers [185].

Conclusions

Although conventional cardiac biomarkers, such as NPs, sST2, cardiac troponins and galectin-3, can be useful for ascertaining CV risk in patients with prediabetes and T2DM, there are confusing reports regarding their ability to be prognostically relevant among patients without known CV disease. There have been many cohort clinical studies with small sample sizes in which the higher predictive power of these biomarkers for CV death and MACEs in patients with metabolic diseases was determined, but large clinical trials have not demonstrated significant results in this context. This is perhaps related to mixed patient cohorts (having and not having CV diseases) enrolled in the studies for which positive predictive values of biomarkers were sufficiently distinguished. Another cause is the significant difference in the quality of the studies. Therefore, numerous results were obtained from post hoc data analyses, and they were not given during prospective observation even when a large sample size was presented. Most positive findings regarding independent prediction of CV death, MACEs and HF in prediabetes and T2DM populations were obtained for NPs and hs-cTn I/T, whereas other cardiac biomarkers did not frequently show independent power in individuals without known CV disease. These discrepancies urged investigations of new biomarkers, such as GDF15, H-FABR, CT-1, carboxy-terminal telopeptide of collagen type 1, apelin and miRNAs as well as multiple biomarker predictive scores. Finally, NPs and hs-cTnT continue to be the most important tools with an affordable price as well as high sensitivity and specificity to predict clinical outcomes among patients with pre-DM and DM in routine clinical practice, but other circulating biomarkers need to be carefully investigated in large trials in the future.

The role of diabetes mellitus as a cause of asymptomatic and symptomatic cardiac disease is progressively increasing.

Conventional biomarkers of cardiac biomechanical stress and myocardial injury have demonstrated limited predictive value for patients with prediabetes and diabetes mellitus.

Circulating biomarkers of fibrosis (soluble ST2) and inflammation (growth-differentiation factor-15, galectin-3, cardiotrophin-1) are promising predictors of cardiac injury at an early stage.

Multiple biomarker predictive scores could be useful in personalizing stratification and care.