Triglyceride–glucose index as a cardiovascular risk marker: updated evidence, limitations, and future perspectives

Abstract

Cardiovascular diseases (CVDs) remain the foremost cause of morbidity and mortality worldwide, accounting for a substantial burden on healthcare systems, economies, and society. Traditional risk factors do not fully explain residual cardiovascular risk, highlighting the role of insulin resistance (IR) as a key pathophysiological link between metabolic disorders and CVD. The triglyceride–glucose (TyG) index, a simple surrogate marker derived from fasting triglyceride and glucose levels, has emerged as a practical tool for assessing IR and cardiovascular risk. This review summarizes current evidence on the role of the TyG index in the development, progression, and prognosis of CVD. The TyG index reflects both hepatic and adipose tissue insulin resistance and has shown comparable or superior performance to traditional markers such as homeostatic model assessment of insulin resistance (HOMA-IR) but is easier to calculate and more feasible for routine clinical use. Large epidemiological and cohort studies have consistently demonstrated associations between elevated TyG values and increased risk of type 2 diabetes mellitus, coronary heart disease, myocardial infarction, stroke, peripheral arterial disease, and heart failure in both general and high-risk populations. Evidence suggests a dose–response relationship between TyG and cardiovascular outcomes, with higher TyG levels or adverse longitudinal trajectories conferring greater risk. Importantly, TyG has been linked to subclinical atherosclerosis, coronary microvascular dysfunction, and adverse cardiac remodeling, indicating its relevance in early disease stages. Sex-specific differences have been reported, with several studies showing stronger associations between TyG and cardiovascular risk in women. In addition, TyG has demonstrated prognostic value across diverse clinical settings, including acute and chronic coronary syndromes, myocardial infarction with nonobstructive coronary arteries, and heart failure. Despite these findings, the clinical implementation of the TyG index remains limited. Its incremental value beyond that of established risk models is inconsistent, and its predictive performance varies across populations, age groups, and ethnicities. Most available data are observational, with underrepresentation of non-Asian populations and younger individuals, and standardized clinical cutoff values have not been universally established. In conclusion, the TyG index is a simple, inexpensive, and widely accessible marker of insulin resistance with considerable potential for cardiovascular risk stratification. While current evidence supports its association with incident CVD, atherosclerotic burden, and adverse outcomes, further prospective, multiethnic studies are needed to define optimal thresholds, clarify its role alongside existing risk models, and determine whether TyG-guided strategies can improve clinical outcomes.

Introduction

Cardiovascular diseases (CVDs) continue to represent the leading cause of death and disability worldwide, placing a substantial burden on healthcare systems and society [1, 2]. Although traditional risk factors—including older age, male sex, obesity, hypertension, hypercholesterolemia, and diabetes mellitus (DM)—explain a large proportion of cardiovascular events, many individuals experience CVD despite the absence of these established predictors [3, 4]. Moreover, despite considerable advances in pharmacological treatment and the implementation of evidence-based primary and secondary prevention strategies, patients with established CVD remain at considerable risk of recurrent cardiovascular complications [5]. These observations highlight the need for improved approaches to identify individuals at elevated cardiovascular risk and to refine current risk stratification strategies [2, 6]. In this context, the triglyceride–glucose (TyG) index has gained increasing attention as a practical surrogate marker of insulin resistance (IR). Accumulating evidence indicates that elevated TyG values are associated with both the onset and clinical outcomes of CVD, suggesting that this index may provide additional value for cardiovascular risk assessment [2].

Methods

This review is an updated narrative review aimed at summarizing the current evidence regarding the use of the TyG index as a surrogate marker of insulin resistance and its role in CVD development, progression, prognosis, and risk prediction. Although this review was not designed as a formal systematic review or meta-analysis, a structured literature search and study selection process were implemented to ensure comprehensive and balanced coverage of the available evidence. A literature search was performed in PubMed/MEDLINE, Scopus, and Web of Science. The final search was completed in July 2026. Publications from 1 January 2016 to July 2026 were primarily considered to provide an updated overview of contemporary evidence. Earlier landmark studies were included when necessary to describe the development and validation of the TyG index or to provide an essential historical context. The search strategy combined Medical Subject Headings (MeSH), Emtree terms (where applicable), and free-text keywords using Boolean operators. The principal search terms included (“triglyceride-glucose index” OR “TyG index” OR “TyG”) AND (“cardiovascular disease” OR “atherosclerosis” OR “coronary artery disease” OR “acute coronary syndrome” OR “chronic coronary syndrome” OR “myocardial infarction” OR “stroke” OR “cerebrovascular disease” OR “peripheral artery disease” OR “heart failure” OR “cardiovascular mortality” OR “insulin resistance” OR “risk prediction” OR “risk stratification”). Reference lists of eligible articles, systematic reviews, and meta-analyses were also manually screened to identify additional relevant publications not retrieved during the electronic database search. Eligible publications included systematic reviews, meta-analyses, randomized controlled trials, prospective and retrospective cohort studies, case–control studies, registry-based investigations, and large observational studies published in English. Conference abstracts, editorials, narrative commentaries without original data, letters lacking primary results, animal studies, in vitro studies, duplicate publications, and studies with insufficient methodological descriptions or unavailable cardiovascular outcome data were excluded. Titles and abstracts were independently screened by two reviewers to assess potential eligibility. Full texts were subsequently evaluated for relevance according to the predefined eligibility criteria. Disagreements regarding study inclusion were resolved through discussion until consensus was achieved. Because this review is narrative in nature, no formal quantitative data synthesis or meta-analysis was performed. Particular emphasis was placed on studies evaluating the following: the association between the TyG index and incident cardiovascular disease; its prognostic value in patients with established cardiovascular disease, subclinical atherosclerosis and vascular remodeling; its incremental predictive performance beyond established cardiovascular risk models; and the clinical applicability and limitations of TyG in cardiovascular risk assessment. The evidence was synthesized narratively according to major clinical settings rather than pooled quantitatively. Greater weight was given to large prospective cohort studies, multicenter investigations, systematic reviews, and meta-analyses, whereas smaller exploratory studies were interpreted cautiously.

What is the TyG Index?

The biological mechanisms linking the TyG index to CVD have not yet been fully elucidated. Because the TyG index integrates fasting triglyceride and glucose concentrations, it reflects two major metabolic abnormalities associated with cardiovascular risk and serves as an indirect marker of insulin resistance (IR) [2]. Increasing evidence supports the use of the TyG index as a reliable surrogate for IR, providing a plausible explanation for its association with adverse cardiovascular outcomes [7–9]. Insulin resistance is recognized as an important contributor to both the development and progression of CVD. It not only promotes cardiovascular risk in the general population and individuals with DM but also adversely affects prognosis among patients with established cardiovascular disease [2, 8]. The principal mechanisms linking IR with cardiovascular pathology are summarized in Fig. 1 [2].

Fig. 1.

Potential molecular mechanisms that contribute to the predictive roleof the triglyceride-glucose (TyG) index in cardiovascular diseases (CVD)

A well-established physiological relationship exists between insulin sensitivity and pancreatic β-cell function. As insulin sensitivity decreases, compensatory β-cell insulin secretion increases in an attempt to preserve normal glucose homeostasis [9]. However, persistent insulin resistance ultimately contributes to the development of chronic metabolic disorders, including DM and CVD [10, 11]. Consequently, elevated TyG values may identify individuals in whom these pathological mechanisms are already active and therefore signal a greater likelihood of future cardiovascular events. Experimental and clinical studies have further demonstrated that fasting plasma glucose primarily reflects hepatic insulin resistance, whereas fasting triglyceride concentrations predominantly reflect insulin resistance within adipose tissue. By combining these two metabolic parameters, the TyG index appears to capture systemic insulin resistance more comprehensively than either fasting glucose or triglycerides alone do, which may explain its stronger association with cardiovascular outcomes [9]. bThe homeostatic model assessment of insulin resistance (HOMA-IR) has long served as the conventional surrogate marker for IR in epidemiological investigations and has consistently been associated with increased cardiovascular risk. Nevertheless, several limitations restrict its routine clinical application. Its accuracy is reduced in individuals with normal or low body mass index, and its implementation depends on the availability of standardized insulin assays, which are not universally accessible [12]. Furthermore, HOMA-IR performs less reliably in patients receiving exogenous insulin therapy or in those with impaired pancreatic β-cell function [2, 13]. Although direct methods such as the euglycemic–hyperinsulinemic clamp and intravenous glucose tolerance test remain the reference standards for assessing insulin resistance, they are labor intensive, invasive, and expensive, making them unsuitable for routine clinical practice [13].

To overcome these practical limitations, the TyG index was introduced as a simple, inexpensive surrogate marker of insulin resistance. Unlike HOMA-IR, it does not require insulin measurements and can be applied regardless of insulin treatment status. Several studies have demonstrated that the TyG index performs at least as well as, and in some populations better than, HOMA-IR for identifying insulin resistance in both diabetic and nondiabetic individuals [14]. Beyond its value as a marker of metabolic dysfunction, growing evidence indicates that the TyG index independently predicts adverse clinical outcomes in patients with and without diabetes as well as in those with established cardiovascular disease, supporting its potential utility for cardiovascular risk stratification in routine clinical practice [2].

The TyG index is calculated as follows:

TyG = ln [fasting triglycerides × fasting glucose/2],

The TyG index is a composite marker derived from fasting triglyceride and fasting glucose concentrations. It was originally introduced in 2008 as a simple surrogate measure of insulin resistance (IR). In the initial large cross-sectional study involving apparently healthy individuals, the TyG index outperformed HOMA-IR in identifying IR, demonstrating a sensitivity of 84.0% but a relatively low specificity of 45.0% [15]. Consequently, despite its acceptable sensitivity, the modest specificity and corresponding risk of false-positive results limit its suitability as a standalone screening tool for IR [15]. Additional validation was provided by Guerrero-Romero et al., who evaluated 99 individuals with different body mass indices and degrees of glucose tolerance in 2010. Compared with the euglycemic–hyperinsulinemic clamp, the reference standard for assessing insulin resistance, the TyG index demonstrated excellent diagnostic performance, with a sensitivity of 96.5% and a specificity of 85.0% [16]. Similar findings were reported in a cross-sectional study of 82 Brazilian participants with either diabetes mellitus (DM) or normal glucose tolerance, in which the TyG index showed slightly better discriminative ability than the HOMA-IR index for detecting IR (AUC 0.79 vs. 0.77, respectively) [17]. Nevertheless, because both validation studies included relatively few participants, their findings should be interpreted with caution and require confirmation in larger populations. Although both the TyG index and the homeostasis model assessment of insulin resistance (HOMA-IR) are widely accepted surrogate markers of IR, increasing evidence indicates that the TyG index may offer greater prognostic value for cardiovascular outcomes. Unlike HOMA-IR, which depends on insulin measurements and may be influenced by interassay variability, the TyG index is calculated exclusively from routinely measured fasting glucose and triglyceride concentrations, making it a simple, inexpensive, and practical tool for widespread clinical use. Subsequent large-scale cohort studies have further established the clinical utility of the TyG index for identifying insulin resistance in high-risk populations. Because IR is a central mechanism underlying impaired glucose metabolism and the development of DM, early identification of affected individuals has important clinical implications. In a prospective study involving 5,354 middle-aged Korean adults without diabetes, Lee et al. reported that participants in the highest TyG quartile had a more than fourfold greater risk of developing DM than those in the lowest quartile did (relative risk 4.095; 95% CI 2.701–6.207) during long-term follow-up [18]. Moreover, compared with the HOMA-IR index, the TyG index showed better predictive performance for insulin resistance, supporting its value as an early marker of metabolic deterioration [18]. Consistent with these findings, Navarro-González et al. reported in 2016 that among 4,820 participants, the TyG index predicted the presence of DM more accurately than fasting glucose or triglyceride concentrations alone did. The TyG index achieved an AUC of 0.75 (95% CI 0.70–0.81), compared with 0.66 (95% CI 0.60–0.72) for fasting glucose and 0.71 (95% CI 0.65–0.77) for triglycerides [19]. These observations suggest that the TyG index may represent a practical approach for identifying individuals at increased risk of future diabetes, thereby facilitating earlier preventive interventions.

Incremental predictive value of the TyG index beyond established cardiovascular risk models

Although numerous studies have demonstrated associations between the TyG index and cardiovascular outcomes, epidemiological associations should not be equated with incremental predictive value [20–26]. Whether TyG improves established cardiovascular risk prediction models represents a distinct clinical question. Current evidence addressing this issue remains limited and heterogeneous [20–26]. Importantly, studies evaluating the incremental value of TyG differ substantially in design and can be categorized into four groups: (1) studies that add TyG alone to an externally validated risk score [20–24]; (2) studies that add TyG together with other biomarkers [25]; (3) studies that develop a new multivariable model that includes TyG [26]; and (4) studies that compare a newly developed model with an established prediction score [26]. These approaches address different methodological questions and should not be interpreted as equivalent evidence that TyG independently improves an established cardiovascular risk model.

Addition of TyG to established cardiovascular risk scores

The Tehran Lipid and Glucose Study [20] illustrates the distinction between association and prediction. Although higher TyG values were associated with incident coronary heart disease and composite cardiovascular disease, adding TyG to the Framingham risk score did not markedly improve the predictive performance. Thus, an epidemiological association does not necessarily translate into improved discrimination beyond conventional cardiovascular risk factors.

More recent evidence has evaluated TyG in relation to SCORE2. Mavraganis et al. [25] derived a TyG threshold of 8.46 in the ATTICA cohort and externally evaluated it in an independent primary-prevention cohort. The addition of TyG to SCORE2 reportedly improved both Harrell’s C-index (0.054 and 0.044 in the derivation and validation cohorts, respectively) and the net reclassification indices (0.668 and 0.469, respectively). Elevated TyG was also associated with multisite vascular disease and the progression of subclinical carotid atherosclerosis. These findings suggest that TyG may function as a potential risk-enhancing marker in primary prevention; however, further external validation is needed because data-derived thresholds may not be transportable across populations with different baseline cardiovascular risk, ethnic composition, metabolic characteristics, and laboratory methods.

In a contemporary Chinese cohort, recalibrated models incorporating both TyG and high-sensitivity C-reactive protein demonstrated improvements in calibration and decision curve analysis but did not significantly improve Harrell’s C-index [26]. Because both biomarkers were introduced simultaneously, the observed improvements cannot be attributed specifically to TyG. Consequently, this study supports the complementary value of combining metabolic and inflammatory biomarkers but does not establish the isolated incremental contribution of TyG beyond the contributions of the PCE, China-PAR, or PREVENT [26].

Machine learning studies require similarly cautious interpretation. A recent investigation in patients with newly diagnosed type 2 diabetes developed a LightGBM model incorporating six predictors: age, hypertension, TyG, carotid plaque burden, maximum plaque thickness, and carotid intima-media thickness [26]. The area under the receiver operating characteristic curve was 0.845 in the training cohort and 0.772 in the validation cohort, whereas the Framingham risk scores were 0.672 and 0.608, respectively. However, these results reflect the performance of a multimodal clinical and imaging model rather than the independent incremental contribution of TyG. Furthermore, because the model was derived in hospitalized patients with newly diagnosed type 2 diabetes, its generalizability to community-based primary prevention populations remains uncertain [26]. Taken together, the current evidence indicates that TyG may provide additional prognostic information in selected populations, but its incremental predictive value beyond that of established cardiovascular risk models remains inconsistent. Some studies demonstrate improvements in discrimination, whereas others report changes primarily in calibration, reclassification, or decision curve analysis. Interpretation should therefore consider whether TyG was evaluated alone or together with additional biomarkers, whether validation was internal or external, and whether comparisons were made against an externally validated clinical score or an internally developed model. Future investigations should report discrimination, calibration, reclassification, and clinical net benefit separately and should determine whether improvements in predictive performance translate into meaningful changes in clinical decision-making or patient outcomes.

Association of the TyG index with cardiovascular disease

Multiple cohort studies have consistently demonstrated an association between elevated TyG values and incident cardiovascular disease [2, 6, 20–26]. In the Tehran Lipid and Glucose Study, Barzegar et al. reported increased risks of coronary heart disease and composite cardiovascular disease among individuals with higher TyG values [20]. Similarly, two large population-based studies by Liu et al. [21] and Hong et al. [22], involving 96,541 and 5,593,134 participants, respectively, reported positive associations between TyG and the incidence of myocardial infarction, supporting an association between TyG and cardiovascular risk in the general population [6]. Cui et al. further demonstrated that cumulative TyG exposure was more strongly associated with cardiovascular disease than a single baseline measurement, with a multivariable-adjusted hazard ratio of 1.39 (95% CI 1.21–1.61) across TyG quartiles [23]. These findings suggest that a longitudinal assessment of TyG may better reflect long-term metabolic risk than a single measurement.

Most available evidence comes from middle-aged and older populations, whereas data in younger individuals remain limited. Although Dikaiakou et al. demonstrated a correlation between TyG and insulin resistance in children and adolescents [27], evidence supporting the prediction of future cardiovascular disease in younger populations is currently insufficient. Sex differences also remain incompletely understood [2, 12, 28–30]. Although men generally exhibit a greater burden of traditional cardiovascular risk factors [2, 29], several studies suggest stronger associations between TyG and cardiovascular events in women [12, 30], highlighting the need for further sex-specific analyses. Dietary factors may also influence triglyceride concentrations, although most cohorts lack detailed nutritional information [6].

Overall, higher TyG values are consistently associated with increased risks of coronary heart disease, myocardial infarction, and composite cardiovascular disease, with evidence supporting a dose‒response relationship. One-unit increases in TyG have been associated with 35% and 23% increases in CHD and CVD risk, respectively [6]. However, these epidemiological associations should not be interpreted as evidence that lowering TyG itself reduces cardiovascular risk or that TyG independently improves clinical risk prediction.

Evidence from Iranian cohorts illustrates the importance of distinguishing associations from predictions and of considering individual cardiovascular outcomes separately [20, 31, 32]. While the Tehran Lipid and Glucose Study demonstrated associations with incident coronary heart disease and composite cardiovascular disease [20], the addition of the TyG score to the Framingham risk score did not improve the predictive performance. Findings from the Isfahan Cohort Study further demonstrated that associations varied according to specific endpoints and adjustment models [32]. In a separate analysis of 5,432 adults without previous cardiovascular disease, higher TyG tertiles were initially associated with cardiovascular and all-cause mortality, but both associations lost statistical significance after adjustment for diabetes mellitus. TyG was not associated with noncardiovascular mortality, suggesting that diabetes may substantially mediate the observed mortality associations [32]. These findings emphasize the importance of reporting coronary disease, stroke, cardiovascular mortality, and all-cause mortality separately rather than collectively.

Several cohort studies have also demonstrated associations between TyG and cardiovascular disease among individuals without diabetes [12, 33, 34], although a meta-analysis did not confirm a significant association with myocardial infarction in nondiabetic individuals [35]. Mechanistic studies support insulin resistance as a plausible biological link between TyG and atherosclerosis through pathways involving dyslipidemia, inflammation, endothelial dysfunction, and blood pressure dysregulation [10, 36–40]. Additional prospective cohorts from Brazil and Spain reported increased cardiovascular risk among individuals with TyG values of approximately 8.8–9.0 [34, 41]. Finally, in the CUN-VMC cohort, TyG independently predicted coronary heart disease, cerebrovascular disease, and peripheral artery disease [19].

Role of the TyG Index across Clinical Manifestations of Atherosclerotic Cardiovascular Disease

Acute coronary syndrome (ACS)

ACS, including unstable angina (UA), non-ST-elevation myocardial infarction (NSTEMI), and ST-elevation myocardial infarction (STEMI), represents the most severe manifestation of coronary heart disease. Despite guideline-recommended revascularization strategies and optimal medical therapy, many ACS patients remain at high risk of recurrent cardiovascular events [35]. Consequently, accurate risk stratification in patients with ACS is essential. Increasing evidence suggests that the TyG index may serve as a useful prognostic marker in ACS patients, irrespective of diabetes status [35]. A 2024 meta-analysis including 21 cohorts and 20,403 patients with acute myocardial infarction revealed that patients in the highest TyG category had increased risks of major adverse cardiovascular and cerebrovascular events and all-cause mortality [42]. Associations were also observed when TyG was analyzed as a continuous variable, and subgroup findings were broadly consistent according to diabetes status, myocardial infarction phenotype, and reperfusion treatment. Nevertheless, the included studies were observational and differed in their TyG categorization, follow-up duration, endpoint definitions, adjustment variables, and clinical characteristics. A more recent systematic review and meta-analysis of 10 observational cohorts comprising 32,751 patients with ACS reported a pooled adjusted hazard ratio of 1.45 per-unit increase in TyG for major adverse cardiovascular events [43]. Substantial heterogeneity was present, and the association appeared stronger in cohorts restricted to patients with type 2 diabetes. Although sensitivity and trim-and-fill analyses suggested that the association remained statistically significant, evidence of heterogeneity and small-study effects limits the certainty and transportability of the pooled estimate. Recent investigations in patients with acute myocardial infarction further support the prognostic value of TyG, which is significantly associated with long-term major adverse cardiovascular events and with intracoronary thrombus burden following primary percutaneous coronary intervention, highlighting its potential utility for cardiovascular risk stratification in contemporary clinical practice [44, 45].

In a retrospective cohort of 2,531 diabetic ACS patients followed for three years, higher TyG tertiles were associated with increased major adverse cardiovascular and cerebrovascular events [46]. The TyG index independently predicted major adverse cardiovascular events (MACE) (HR = 1.455, 95% CI 1.208–1.753), regardless of treatment strategy [46]. Subgroup analysis indicated a significant association primarily in patients with UA (adjusted HR = 1.604, 95% CI 1.270–2.027), possibly reflecting limited statistical power [46]. In line with these findings, Luo et al. reported that among 1,092 STEMI patients who underwent PCI, elevated TyG values were associated with increased MACCE risk within one year (HR = 1.529, 95% CI 1.001–2.061) [47]. Similarly, Mao et al. demonstrated that in 438 NSTE-ACS patients, TyG independently predicted both metabolic abnormalities and adverse outcomes, including high SYNTAX scores (OR = 6.055, 95% CI 2.915–12.579) and MACEs (HR = 1.878, 95% CI 1.130–3.121) [48]. While early studies focused primarily on patients with diabetes or glucose intolerance, emerging data suggest that TyG may also have prognostic value in ACS patients without established metabolic disorders. In a cohort of 1,655 nondiabetic ACS patients with LDL-C concentrations < 1.8 mmol/L, higher TyG values were associated with an increased incidence of acute myocardial infarction (21.2% vs. 15.2%), a larger infarct size, and higher revascularization rates [49]. These patients also exhibited a higher risk of developing diabetes during follow-up, suggesting a link between elevated TyG levels and more extensive coronary disease. Thus, TyG may enable early risk stratification even in ACS patients considered to be at relatively low risk. Beyond obstructive ACS, the TyG index has also been associated with adverse outcomes in patients with myocardial infarction with nonobstructive coronary arteries (MINOCA). In a study of 1,179 MINOCA patients followed for a median of 41.7 months, those in the highest TyG tertile had a significantly increased risk of MACEs (HR = 1.33, 95% CI 1.04–1.69), independent of diabetes status or LDL-C level [50]. These findings support the use of the TyG index as a robust marker for outcome prediction across diverse ACS phenotypes. These findings support the use of TyG as a potentially useful prognostic marker in ACS, but they do not establish TyG as an actionable biomarker for treatment guidance. At present, there is no evidence that knowledge of TyG should independently change the timing of invasive assessment, revascularization strategy, intensity of lipid-lowering therapy, antiplatelet selection or duration, or other established components of acute coronary care. Similarly, compared with guideline-directed assessment and secondary prevention, a TyG-guided strategy does not improve cardiovascular outcomes. TyG may nevertheless identify residual metabolic risk that is not fully represented by conventional glucose or lipid measurements. A high TyG value could reasonably prompt recognition of an adverse metabolic phenotype and consideration of a comprehensive metabolic assessment, including evaluation for diabetes, glycemic abnormalities, obesity, hypertriglyceridemia, and associated lifestyle factors. However, such assessments are already appropriate in many patients with ACS and should not be presented as evidence that TyG directly determines treatment. Therefore, TyG should currently be characterized as an investigational or potentially complementary prognostic marker rather than a validated therapeutic decision tool. Future studies should evaluate its incremental performance beyond validated ACS models, particularly GRACE or GRACE 2.0, and should report discrimination, calibration, reclassification, and decision curve measures. Ultimately, prospective studies are needed to determine whether TyG-informed clinical decisions improve meaningful patient outcomes.

Chronic coronary syndrome (CCS)

Coronary microvascular dysfunction (CMD) is an important determinant of prognosis in patients with CCS. Although the TyG index is positively associated with CVD incidence and outcomes, its relationship with CMD in CCS patients has only recently been explored. Zhang W. et al. evaluated the association between the presence of TyG and CMD and prognosis in 430 CCS patients, of whom 221 were diagnosed with CMD [51]. Patients with CMD had significantly higher TyG values than those without CMD did. During follow-up, 63 major adverse cardiovascular events (MACE) were recorded among CMD patients, with a higher incidence in the highest TyG tertile than in the lower tertiles (39.2% vs. 20.5% vs. 25.7%; P = 0.035). Multivariate analysis revealed that the TyG index was an independent predictor of CMD (OR = 1.436; 95% CI 1.014–2.034; P = 0.042). Moreover, compared with the lowest tertile, the highest TyG tertile was independently associated with increased MACE risk in CMD patients (HR = 2.132; 95% CI 1.066–4.261; P = 0.032) [51]. Similarly, Tao S. et al. (2023) reported that individuals with the highest TyG index had a 59.5% increased risk of MACEs during one-year follow-up (HR = 1.595; 95% CI 1.370–1.855) [52]. Each 1-unit increase in TyG was additionally associated with 35% and 23% higher risks of coronary artery disease (CAD) and CVD, respectively. Collectively, these findings support the use of the TyG index as an independent predictor of CMD and adverse cardiovascular outcomes, in addition to traditional clinical risk factors, in both CCS patients and the general population [52].

Subclinical atherosclerosis

The TyG index has been established as a reliable surrogate marker of IR, demonstrating high sensitivity (96.5%) and specificity (85.0%) in early validation studies [15–17, 53]. As such, the TyG index enables indirect assessment of IR and may predict CVD incidence. Insulin resistance is characterized by low-grade systemic inflammation, endothelial dysfunction, and dysregulation of insulin signaling, promoting atherosclerosis and plaque progression through metabolic abnormalities, chronic hyperglycemia, dyslipidemia, oxidative stress, and proinflammatory pathways [53, 54]. Accordingly, TyG has emerged as a prognostic marker for atherosclerosis, MI, and CHD in both nondiabetic individuals and the general population [55–62]. The Progression of Early Subclinical Atherosclerosis Study reported a high prevalence (~ 50%) of subclinical atherosclerosis among middle-aged individuals without traditional cardiovascular risk factors [62], underscoring the importance of early risk identification. In a retrospective study of 1,250 asymptomatic individuals without conventional risk factors, Park et al. demonstrated that higher TyG values were independently associated with increased CHD risk (OR = 1.473, 95% CI 1.026–2.166), particularly in those with noncalcified or mixed plaques [63]. These findings support the use of TyG as an independent marker of subclinical CHD across both general and high-risk populations. IR also plays a key role in hyperlipidemia. Elevated triglyceride levels increase free fatty acid flux from adipose tissue to nonadipose tissue, further aggravating IR and promoting atherogenic dyslipidemia [59, 62, 63]. Retention of triglyceride- and cholesterol-rich ApoB-containing lipoprotein remnants within the arterial wall contributes directly to atherosclerosis development [39], suggesting that lowering triglyceride levels may represent an important therapeutic target in high-risk patients. Activation of the renin–angiotensin system and impaired myocardial calcium handling may further exacerbate vascular injury [40]. Notably, several studies have shown that TyG outperforms HOMA-IR in predicting atherosclerosis and adverse vascular outcomes, including carotid atherosclerosis [2, 17, 64]. Consistently, higher TyG values have been associated with an increased incidence of ASCVD in individuals without baseline ASCVD [6, 12]. Sex-specific differences have also been reported. The I-Lan Longitudinal Aging Study demonstrated a significant association between elevated TyG and subclinical atherosclerosis in nondiabetic women (OR = 1.510, 95% CI 1.010–2.257) but not in men [29], highlighting the potential need for sex-specific prevention strategies. Despite heterogeneity in sex adjustment across studies, pooled analyses confirmed a robust association between TyG and CHD/CVD risk (CVD HR = 1.59, 95% CI 1.21–2.09) [6, 35]. Imaging studies further support the link between TyG expression and atherosclerotic burden. Carotid ultrasound assessments revealed more frequent plaques in higher TyG quartiles in both sexes, with increased intima–media thickness predominantly observed in women [29, 65]. In a large prospective cohort with long-term follow-up, elevated TyG independently predicted CVD events across weight categories, with stronger effects in women, and provided incremental predictive value when combined with diabetes and hypertension [65]. Overall, TyG represents a simple, accessible, and clinically relevant marker for identifying individuals at increased risk of atherosclerosis and future cardiovascular events among middle-aged individuals.

Cerebrovascular diseases

Emerging evidence indicates that the TyG index is also associated with age-related cerebrovascular disease and markers of subclinical cerebrovascular atherosclerosis. In a study of young adults, Lin et al. evaluated several surrogate markers of insulin resistance in relation to carotid intima–media thickness (cIMT), a recognized indicator of subclinical atherosclerosis, and reported that the TyG index demonstrated the strongest association with increased cIMT [66]. In the overall study population, triglyceride levels were significantly associated with cIMT (β = 0.057; P = 0.04). Subgroup analyses further revealed that among overweight or hyperlipidemic individuals (BMI ≥ 25 kg/m²), cIMT was independently associated with triglycerides (β = 0.127; P = 0.004), the TyG index (β = 0.119; P = 0.01), and the TG/HDL-C ratio (β = 0.081; P = 0.03), whereas no significant associations were observed among participants with normal body weight [66]. Similar findings were reported by Meng et al., who investigated the relationship between the TyG index and the development of carotid atherosclerosis [67]. During follow-up, 551 participants developed new carotid plaques, corresponding to an incidence of 34.5%. Although univariate analyses revealed a significant association between higher TyG values and plaque formation, multivariable-adjusted analyses revealed that the TyG index was an independent predictor of increased carotid intima–media thickness but not plaque development. Specifically, each one-unit increase in the TyG index was associated with a 49% increased risk of IMT progression (RR 1.49; 95% CI 1.05–2.13; P = 0.026). This relationship was particularly pronounced in men, individuals aged ≥ 60 years, and participants without hypertension [67]. Xie et al. further explored the association between the TyG index and intracranial as well as extracranial atherosclerotic disease using digital subtraction angiography (DSA) to characterize cerebrovascular stenosis [64]. Their study demonstrated that higher TyG values were significantly associated with a greater incidence of intracranial arterial stenosis (ICAS), combined intracranial and extracranial stenosis (ICAS/ECAS), anterior and posterior circulation involvement, simultaneous anterior and posterior circulation stenosis, severe arterial narrowing, and both symptomatic and asymptomatic cerebrovascular stenosis [64].

Peripheral arterial disease

Peripheral arterial disease (PAD) is a common manifestation of systemic atherosclerosis that is associated with substantial morbidity and adverse cardiovascular outcomes. Nevertheless, its underlying pathophysiological mechanisms remain incompletely understood. Compared with that of coronary heart disease (CHD), the contribution of insulin resistance (IR) to PAD has received considerably less attention, while research has focused primarily on other pathogenic pathways, particularly inflammation. Consequently, the identification of metabolic markers associated with PAD remains an area of active investigation. Evidence from large prospective cohorts suggests that the TyG index may serve as a useful marker for PAD risk assessment. Gao et al. analyzed data from 12,320 participants enrolled in the Atherosclerosis Risk in Communities (ARIC) study to investigate the relationship between the TyG index and incident PAD, as well as the effect of longitudinal changes in TyG over time [65]. During a median follow-up of 23 years, PAD developed in 1,300 participants. After adjustment for conventional PAD risk factors, each one-standard-deviation increase in the TyG index (0.58 units) was associated with an 11.9% higher risk of incident PAD (hazard ratio 1.119; 95% CI 1.049–1.195). Moreover, compared with individuals who maintained persistently low TyG values, participants with high and very high TyG trajectories had significantly greater risks of future PAD, with adjusted odds ratios of 1.404 (95% CI 1.132–1.740) and 1.742 (95% CI 1.294–2.344), respectively [65]. Similar observations have been reported in patients with diabetes mellitus. Ning et al. evaluated 1,040 individuals with diabetes, including 168 patients diagnosed with lower extremity arterial disease (LEAD), and reported significantly higher TyG values in patients with LEAD than in those without vascular disease (9.94 ± 0.78 vs. 9.36 ± 0.70; P < 0.001) [68]. Multivariable analysis further demonstrated that the TyG index was independently associated with the presence of LEAD (OR 3.92; 95% CI 2.92–5.26; P < 0.001), supporting its potential role as a marker of lower-extremity atherosclerosis in diabetic patients [68]. Additional evidence was provided by Caliskan et al., who evaluated the diagnostic performance of the TyG index for PAD. In their study, the TyG index demonstrated moderate predictive accuracy, with a sensitivity of 57.8%, a specificity of 70.0%, and an area under the receiver operating characteristic curve (AUC) of 0.689 [69]. Although these results indicate only moderate discrimination, they suggest that the TyG index may contribute to the identification of individuals at increased risk when interpreted alongside established clinical and laboratory parameters. At present, no single biomarker provides optimal risk stratification or diagnostic accuracy for PAD. Although the causal role of insulin resistance in PAD pathogenesis has not been definitively established, increasing evidence consistently demonstrates a significant association between metabolic dysfunction and peripheral atherosclerosis. Given that the TyG index is an inexpensive, readily available surrogate marker of insulin resistance, it represents a practical tool for identifying individuals at elevated risk of PAD and other cardiovascular diseases in which insulin resistance is believed to play an important pathogenic role [65, 68–70].

HF and cardiac remodeling

Recent studies have indicated that IR is a major driver of poor prognosis in HF patients, suggesting that surrogate IR markers are potentially important for HF prevention and management [71–79]. Guo et al. reported that the TyG index was associated with worse outcomes in chronic HF patients with diabetes, with higher TyG values linked to increased cardiovascular mortality and more frequent HF-related rehospitalization [71]. In addition to prognosis, TyG has been proposed as a biomarker of myocardial fibrosis in patients with HF, a key determinant of cardiovascular risk, as assessed by cardiovascular magnetic resonance (CMR) [72]. In 103 hospitalized HF patients, Yang et al. demonstrated that the myocardial fibrosis quantified by the extracellular volume (ECV) fraction on CMR was significantly associated with TyG (r_partial = 0.36), and during a median follow-up of 12.3 months, TyG independently predicted all-cause mortality and HF-related hospitalization (HR = 2.01, 95% CI 1.03–4.01), supporting its potential value for risk stratification in HF management [73]. Given the contribution of nondiabetic hyperglycemia and elevated triglycerides to CVD risk, the TyG index may also be clinically useful for screening individuals without overt metabolic disorders, although prospective evidence using CVD as a primary endpoint remains limited [12]. Population-level data further support TyG as a predictor of incident HF. Li X. et al., using observational data from two large Chinese cohorts and a two-sample Mendelian randomization analysis, reported that a high TyG index was an independent and causal risk factor for HF: the highest quartile carried a 22% higher risk than the lowest quartile did (95% CI 11%–34%, P < 0.0001), each 1-unit increase was associated with a 15% increase in HF risk (95% CI 10%–21%, P < 0.00001), and genetically determined TyG remained associated with HF after sensitivity analyses (OR 1.19, 95% CI 1.05–1.35, P = 0.01) [74]. The Kailuan and Hong Kong cohorts reported 2,726 and 1,709 incident HF cases, respectively, over mean follow-up periods of 12.3 ± 2.2 and 16.2 ± 4.3 years, with incidence rates of 2.3 and 5.5 per 1000 person-years, respectively [74]. In a high-risk clinical population with ischemic cardiomyopathy and DM, Abuduaini et al. reported a higher incidence of HF in the T2/T3 TyG subgroup than in the T1 subgroup over 55.4 months, with 86 (5.7%) patients developing HF [76]. TyG has also been linked to subclinical cardiac dysfunction and remodeling. Mei J. et al. (2023) reported that the TyG index is a predictor of BMI-linked subclinical left atrial dysfunction (AUC 0.630, p = 0.031), with an optimal cutoff of 9.33 [72]. Chiu T. et al. reported that higher TyG was associated with greater left atrial diameter (LAD), lower left ventricular ejection fraction (LVEF), and lower ankle–brachial index (ABI), whereas observed/predicted left ventricular mass and brachial–ankle pulse wave velocity were not significantly associated [77]. Similarly, Wang C. et al. reported positive correlations between TyG and multiple hemodynamic parameters (LVDd, LVSd, PWT, IVST, LVM, and aortic root diameter) and a negative correlation with LVEF (p < 0.05) [78]. In hypertensive patients, Liao et al. reported that compared with non-HFpEF patients, HFpEF patients had higher LAD, LVMI, and E/e′ values and lower E/A and average e′ values, indicating worse cardiac structure and diastolic function [79] (Table 1).

Table 1.

Basic information of the included studies for TyG index and CVD

Study	Design	Country	Participants characteristics	Total, n	Mean age, years	Male, n (%)	Diabetes, n (%)	Groups	Follow-up duration (years)	Main findings
Cardiovascular disease (CVD)
Tian S [34]	A large-scale, longitudinal stady	China	General population	7632	≥ 18 and ≤ 85	3501 (45.9)	1649 (21.6)	MHNO, MUNO MHO, and MUO	1	MHO subjects were nearly 2-fold more likely to have hypertension than their MHNO peers. The adjusted ORs for having hypertension in MHO subjects were 1.94 (95% CI 1.60–2.35) using the ATP-III criteria, 1.98 (95% CI 1.61–2.43) using the Wildman criteria, 2.37 (95% CI 1.88–2.99) using the Karelis criteria, 2.26 (95% CI 1.96–2.61) using the HOMA criteria, and 2.54 (95% CI 2.14–3.00) using the TyG criteria. Similarly, compared to the MHNO phenotype, subjects in the MUO and MUNO phenotypes demonstrated significantly higher ORs for hypertension; the corresponding ORs ranged from 2.75 to 7.02 and from 1.23 to 3.99, respectively.
Sánchez-Íñigo L [42]	Prospective cohort study	Spain	Free of CVD at baseline, history of type 1 diabetes or latent autoimmune diabetes in adults, cancer in the palliative phase, familial hypertriglyceridaemia, extreme BMI (> 45 kg/m2) or a hypercoagulable state	5014	54.41	3068 (61.19)	0	Q4:Q1	8.75	A higher level of TyG index was significantly associated with an increased risk of developing CVD independent of confounding factors with a value of 2·32 (95% CI: 1·65–3·26) for those in the highest quintile and 1·52 (95% CI: 1·07–2·16) for those in the fourth quintile. The areas under the curve (AUC) of the ROC plots were 0·708 (0·68–0·73) for the Framingham model and 0·719 (0·70–0·74) for the Framingham + TyG index model (P = 0·014).
Fernández-Friera L [65]	Prospective cohort study	Spain	Patients CVRF-free	1779	45.0 ± 4.1	884 (49.7)	0	No Atherosclerosis Atherosclerosis	6	Subclinical atherosclerosis (plaque or coronary artery calcification) was present in 49.7% of CVRF-free participants. Together with male sex and age, LDL-C was independently associated with atherosclerosis presence and extent, in both the CVRF-free and CVRF-optimal groups (odds ratio [×10 mg/dl]: 1.14 to 1.18; p < 0.01 for all)
Vega GL [60]	Retrospective cohort study	USA	General population	39,447	44.90	39,444 (100)	394 (1)	NA	14.75	CV mortality: highest vs. lowest 0.89 (0.77–1.03); All-cause mortality: highest vs. lowest 0.89 (0.82–0.97)
Acute coronary syndrome (ACS)
da Silva [63]	Prospective study	Brazil	Patients have at least one CVD in the last 10 years	2330	63.2 ± 8.1	1358 (58.3)	1024 (44.0)	tertile 1 tertile 2 tertile 3	10	Risk of CAD: Asymptomatic: 0.98 (0.78–1.17) Symptomatic: 1.16 (1.01–1.33) Treated: 1.03 (0.97–1.10)
Hong [22]	Retrospective study	Korea	Community population without ASCVDs	559,134	53.0	282,363 (50.5)	20,688 (3.7)	Q4:Q1	8.2	Composite ASCVDs (146,744), CAD (62,577), and stroke (89,120). in the highest TyG index quartile demonstrated that these patients were at higher risk for stroke (HR = 1.259; 95% confidence interval [CI] 1.233–1.286), for MI (HR = 1.313; 95% CI 1.28–1.346), and for both (HR = 1.282; 95% CI 1.261–1.303) compared with participants in the lowest TyG index quartile.
Barzegar N [20]	Prospective study	Iran	Community population without ASCVDs	7521	46.6 ± 12.1	3367 (44.8 )	988 (13.1)	Q4:Q1	16.1	Composite ASCVDs (1084), and CAD (924). A 1-SD increase in TyG-index was associated with significant risk of CVD/CHD in multivariate analysis [1.16 (1.07–1.25) and 1.19 (1.10–1.29), respectively]. The cut-off value of TyG-index for incident CVD was 9.03 (59.2% sensitivity and 63.2% specificity); the corresponding value of TyG-index for incident CHD was 9.03 (60.0% sensitivity and 62.8% specificity), respectively.
Liu, [21]	Prospective study	China	Free of CVD and cancer history, and obesity (BMI > 45 kg/m2)	96,541	70.45	51,244 (53.08)	-	Q4:Q1	5.52	CVD events: Q1, HR (95% CI) was 1.12 (95%, 1.03–1.21) in Q2, 1.28 (95%, 1.18–1.38) in Q3, and 1.34 (95%, 1.23–1.45) in Q4.
Wang L, [49]	Retrospective, observational study	China	Patients with T2DM and ACS	2531	66,3 ± 6.8	1415 (55.9)	2531 (100)	tertile 1 tertile 2 tertile 3	3	289 (11.4%) MACEs were recorded, including 142 (49.1%) all-cause death, 101 (34.9%) non-fatal MI and 46 (16.0%) non-fatal stroke.
Mirshafiei H [43]	Prospective study	Iran	Free of diabetes, hyperlipidemia, hypertensive and CVD history	9704	48.09	3894 (40.13)	0	Q4:Q1	6.00	risk of ACS and cardiac death 1.362 [1.013–1.831] and 2.3 [1.247–4.241] respectively; the chance of CVD by 1.634 [1.304–2.047] and 1.068 [1.031–1.105)], respectively
Ray K []		)						a
Tian X [61]	Prospective study	China	Community population without ASCVDs	98,849	51.8	78,881 (79.8)	3035 (3.07)	Q4:Q1	11.0	The risk of MI increasedwith quartiles of the baseline and updated mean TyG index, the HR in quartile 4 versus quartile 1 was 2.08 (95% CI,1.77–2.45) and 1.58 (1.18–2.12), respectively
Luo E [50]	Observational study	China	Patients with acute ST-elevation myocardial infarction (STEMI) who underwent PCI	1092	62,4 ± 12,5	864 (79.1)	270 (24.7)	Q4:Q1	1	Risk factors for MACCEs: in STEMI patients after PCI was 0.685 (95% CI 0.610–0.761; P = 0.001); in STEMI patients within 1 year after PCI, with a value of 1.529 (95% CI 1.001–2.061; P = 0.003) for those in the highest quartile.
Mao Q [3751]	Observational study	China	Patients with non-ST-segment elevation acute coronary syndrome (NSTE-ACS)	438	62.5 (53.0–68.0)	295 (67.4)	143 (32.6)	low TyG index high TyG index	1	Risk MACEs: 1.951 (1.416–2.688). Risk cardiac death: 2.285 (1.195–4.371). Risk Nonfatal MI: 1.399 (0.717–2.731). Risk target vessel revascularization (TVR): 2.258 (1.285–3.968). Risk congestive heart failure: 1.015 (0.374–2.753). Risk nonfatal stroke: 4.421 (1.559–12.536).
Zhu Y []										,
Chronic coronary syndrome (CCS)
Park B [36]	Prospective study	Korea	Free of IHD or ischemic stroke history, type 2 diabetes or a fasting plasma glucose level ≥ 126 mg/dL, current use of dyslipidaemia medication or aspirin, and hs-CRP levels ≥ 10 mg/L	16,455	46.1	8426 (51.21)	0	Q4:Q1	4.17	Ischemic heart disease (incidence) Q2:Q4 1.61 (1.05–2.48); 1.85 (1.21–2.83); 2.28 (1.48–3.51)
Guo W [74]	Retrospective study	China	Patients with CHF and type 2 diabetes	546	65.18	362 (66.3)	546 (100)	tertile 1 tertile 2 tertile 3	1	The incidence of cardiovascular death or rehospitalization due to HF in the T3 group was significantly higher than that in the T1 group (HR: 2.15, 95%CI: 1.37–3.35). The incidence of primary outcome events in the T2 and T3 groups was significantly higher than that in the T1 group (HR: 1.66, 95%CI: 1.02–2.70 for T2; HR: 2.46, 95%CI: 1.51–4.01 for T3).
Zhang, W [54]	A single-center retrospective observational study	China	Patients suspected or established CCS, diagnosed according to the 2019 ESC guidelines for the diagnosis and management of CCS and met the diagnostic criteria of CCS	596	63.54 ± 9.61	254 (59.1)	150 (34.9)	Patients with CMD: tertile 1 tertile 2 tertile 3 non-CMD	4	Sixty-three (28.5%) MACE were recorded during the follow-up duration among the CMD population, including 4 (1.8%) cardiovascular death, 4 (1.8%) nonfatal MI, 18 (8.1%) heart failure, 15 (6.9%) nonfatal stroke, and 22 (10.0%) ischemia-driven revascularizations. Patients with higher TyG index (T3 group) had a higher rate of MACE as compared to lower TyG index patients (T1 and T2 groups) (39.2% vs. 20.5% vs. 25.7%; p = 0.035).
Tao S [55]	A single-center, retrospective, observational cohort study	China	Patients with CCS patients undergoing PCI were collected	1353	64 (57, 71)	979 (72.36)	612 (45.23)	tertile 1 tertile 2 tertile 3	2	An association between the TyG index score and an elevated risk of one-year endpoints (HR 1.595, 95% CI 1.370 ~ 1.855, P < 0.001). Taking the T1 in as a reference, the risks of the primary endpoint were 2.522-fold higher (HR 2.522, 95% CI 2.032 ~ 3.128, P < 0.001) and 2.297-fold higher (HR 2.297, 95% CI 1.842 ~ 2.864, P < 0.001) in the T3 of Model 2 and Model 3, respectively.
Cerebrovascular diseases (CVD)
Wang A [35]	Prospective study	China	Patients without stroke	97,653	51.67 (43.53–58.97)	77,748 (79.62)	2860 (2.93)	Q4:Q1	11.02	Stroke: 1.08 (0.99–1.18), 1.22 (1.12–1.33), and 1.32 (1.21–1.44); ischemic stroke: 1.14 (1.03–1.26), 1.31 (1.19–1.44), and 1.45 (1.31–1.59)
Cui H [23]	Prospective study	China	Patients without stroke, myocardial infarction, and cancer in the four consecutive examinations.	44,064	54.30 ± 11.65	32,960 (74.8)	3320 (7.53)	Q4:Q1	6.52	Тhe HR for CVD events were 1.25 (1.08–1.44) in Q2, 1.22 (1.05–1.40) in Q3 and 1.39 (1.21–1.61) in Q4, compared to Q1 group.
Kim J []
Lu YW [29]	The present cross-sectional study	Taiwan	Patients without CAD, without lipid-lowering agents and less than 80 years old	1457	62.34 ± 8.00	690 (47.4)	202 (13.9)	TyG index < 8.55 TyG index ≥ 8.55	2	Subclinical atherosclerosis in non-diabetic women after adjusting for traditional risk factors [adjusted odds ratio (OR): 1.510, 95% CI 1.010–2.257, p = 0.045].
Meng J [68]	Prospective study	China	The target population includes individuals with low income, low education, middle-aged and older adults, and those with limited education.	2702	59.60 ± 8.47	1076 (39.8)	0	Q4:Q1	6	Age, TyG index, and educational level as independent predictors of IMT increase. The TyG index was also found to be an independent risk factor for IMT increase, with each unit increase in TyG associated with a 49% increased risk (RR: 1.49; 95% CI: 1.05–2.13; P = 0.026). Notably, the second quartile of TyG (8.3921 ≤ TyG < 8.7676) was associated with a 54% increased risk of IMT increase compared to the first quartile (TyG < 8.3921) (RR: 1.54; 95% CI: 1.01–2.34; P = 0.043), while no significant associations were observed in the third and fourth quartiles (P > 0.05). The risk of IMT increase was 4.28-fold, 6.49-fold, and 10.75-fold higher in individuals aged 45–59, 60–74, and ≥ 75 years, respectively, compared to those aged < 45 years
Xie Y [69]	Retrospective study	China	Patients who underwent digital subtraction angiography (DSA)	1129	62 (55, 68)	805 (71.30)	335 (29.67)	Q3:Q1	4,9	Elevated TyG index was associated with the presence of both asymptomatic stenosis (OR, 1.47; 95%CI, 1.07–2.03, p = 0.018) and symptomatic stenosis (OR, 1.46; 95%CI, 1.09–1.96, p = 0.011).
Peripheral arterial disease (PAD)
Gao J [70]	Prospective study	China	free of a history of PAD	12,320	54.3 ± 5.7	5693 (46.2)	1378 (11.2)	Q4:Q1	23	After adjustment for traditional PAD risk factors, each 1-SD (0.58) increase in TyG index was associated with an 11.9% higher risk of incident PAD [hazard ratio, 1.119 (95% CI, 1.049–1.195)]. Those participants with TyG index trajectories at high and very high levels had an even greater risk of future incident PAD [odds ratio (95%CI): 1.404 (1.132–1.740) and 1.742 (1.294–2.344), respectively] after multivariate adjustments for traditional PAD risk factors.
Ning P [71]	Retrospective study	China	Diabetic patients a ged 18 or older with no acute complication of diabetes with and without lower extremity arterial disease (LEAD)	1040	59.20 ± 12.92	589 (56.63)	1040 (100)	LEAD: No LEAD	1	Increases in TyG index values were associated with the risk of LEAD in patients with diabetes (OR 3.92, 95% CI: 2.92–5.26, p < 0.001). Patients with LEAD had signi cantly higher TyG levels, compared with no LEAD group (9.94 ± 0.78 vs. 9.36 ± 0.70; p < 0.001).
Caliskan S [72]	Retrospective study	Turkey	Patients diagnosed with PAD/healthy controls	440	PAD: 69.41 ± 10.61 Control: 51.79 ± 15.15	273 (62)	117 (26.6)	PAD Control	5	The TyG index levels were signi cantly higher in the peripheral artery disease group than in the control group (9.19 ± 0.57 vs. 8.80 ± 0.59; p < 0.001). TyG index (OR = 1.111, 95% CI = 1.083–1.139; p < 0.001) was independent predictor of peripheral artery disease. TyG index had predictive ability for PAD (sensitivity: 57.8%, speci city: 70%, AUC: 0.689).
Heart failure (HF)
Tian S [2634]	a large-scale, longitudinal stady	China	General population	7632	≥ 18 and ≤ 85	3501 (45.9)	1649 (21.6)	MHNO, MUNO MHO, and MUO	1	MHO subjects were nearly 2-fold more likely to have hypertension than their MHNO peers. The adjusted ORs for having hypertension in MHO subjects were 1.94 (95% CI 1.60–2.35) using the ATP-III criteria, 1.98 (95% CI 1.61–2.43) using the Wildman criteria, 2.37 (95% CI 1.88–2.99) using the Karelis criteria, 2.26 (95% CI 1.96–2.61) using the HOMA criteria, and 2.54 (95% CI 2.14–3.00) using the TyG criteria. Similarly, compared to the MHNO phenotype, subjects in the MUO and MUNO phenotypes demonstrated significantly higher ORs for hypertension; the corresponding ORs ranged from 2.75 to 7.02 and from 1.23 to 3.99, respectively.
Guo W [74]	Retrospective study	China	Patients with acute decompensated heart failure (ADHF)	932	69	579 (62.1)	306 (32.8)	tertile 1 tertile 2 tertile 3	1,3	all-cause death: the highest tertile (TyG index ≥ 9.32) versus the lowest tertile (TyG index < 8.83) were 2.09 (95% confidence interval [CI], 1.23–3.55; p = 0.006); CV death: 2.31 (95% CI, 1.26–4.24; p = 0.007); MACCEs: 1.83 (95% CI, 1.18–3.01; p = 0,006).
Yang S [76]	Retrospective study	China	Patients with HF	103	58.3 (8.9)	71 (68.9)	45 (43.7)	With events Without events	1.3	predictive role: TyG index (AUC = 0.709, 95% CI = 0.611–0.794), ECV fraction (AUC 0.715, 95% CI = 0.618–0.80), and NT-proBNP (AUC 0.741, 95% CI = 0.646–0.823).
Li X [77]	Prospective cohort - the Kailuan study Retrospective study - the Hong Kong cohort	China	the Kailuan study - free prevalent cancer and cardiovascular diseases, including HF, atrial fibrillation (AF), myocardial infarction, and ischemic stroke and missing baseline levels of triglyceride (TG) or fasting blood glucose (FBG). the Hong Kong cohort - free with a history of ischemic heart disease, stroke, HF, AF, or cancer, as well as those who were pregnant at the time of inclusion, and those with missing baseline low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C), and total cholesterol levels.	the Kailuan study − 95,996 the Hong Kong cohort − 19,345	the Kailuan study − 51.4 ± 12.5 the Hong Kong cohort − 60.1 ± 12.9	the Kailuan study − 76,364(79.6) the Hong Kong cohort − 7738 (40.0)	the Kailuan study − 8408(8.8) the Hong Kong cohort − 6524(33.7)	the Kailuan study: Q4:Q1 the Hong Kong cohort: Q4:Q1	the Kailuan study − 1 the Hong Kong cohort − 3	The risk of incident HF of subjects in the highest quartile of the TyG index was 22% higher (95% CI 11% − 34%, P < 0.0001) than those in the lowest quartile, with every unit increment of the TyG index being associated with a 15% increase in the risk of incident HF (95% CI 10% − 21%, P < 0.00001). Similarly, subjects in the highest quartile of the TyG index had a 25% (95% CI 13% − 37%) increase in the sub-hazard of incident HF.
Abuduaini [79]	Retrospective study	China	Patients with cardiomyopathy and T2DM	1514	65.26 ± 10.036	1119 (73.9)	1514 (100)	tertile 1 tertile 2 tertile 3	12	chest pain [9.056 (4.370 to 18.767), p < 0.001], acute myocardial infarction [4.437 (1.420 to 13.869), p = 0.010], heart failure [7.334 (3.424 to 15.708), p < 0.001], cardiogenic shock [3.707 (1.207 to 11.384), p = 0.022], malignant arrhythmia [5.309 (2.367 to 11.908), p < 0.001], cerebral infarction [3.127 (1.596 to 6.128), p < 0.001], all-cause death [4.502 (3.478 to 5.827), p < 0.001].
Cardiac remodeling
Chiu T [80]	Retrospective study	Taiwan	Without atrial fibrillation, significant ankle edema, significant aortic or mitral valve diseases and poor image quality	823	61.3 ± 13.2	483 (58.69)	222 (26.97)	Q4:Q1	4	After adjusting for age, sex, history of DM and hypertension, mean arterial pressure, hematocrit, eGFR and total cholesterol, multivariate stepwise linear regression analysis revealed that young age (p = 0.013), DM (p = 0.021), low mean arterial pressure (p < 0.001) and low eGFR (p < 0.001) were significantly associated with high observed/predicted LVM, but that TyG index was not (unstandardized coefficient β, 4.401; 95% CI, -0.540 to 9.342; p = 0.081). LAD, a high TyG index (unstandardized coefficient β, 4.401; 95% CI, -0.540 to 9.342; p = 0.081) were significantly correlated with low LVEF.
Mei J [75]	Retrospective study	China	without prevalent cardiovascular disease	403	55.71 ± 11.28	236 (58.6)	172 (42.8)	Non-obese Obese	8 months	ROC analysis exhibited that the TyG index can be used as a predictor of LA dysfunction (AUC = 0.63), and the optimal cut-off point for the TyG index is 9.33.
Wang C [57]	Retrospective study	China	asymptomatic T2DM	647	53 ± 12	348	446	T2DM healthy controls	8 months	With the increase of TyG, the levels of left ventricular mass index (LVMI), left ventricular mass (LVM), left ventricular end diastolic diameter (LVDd), posterior wall thickness (PWT), and interventricular septum thickness (IVST) were also increased in T2DM individuals.

MHNO Metabolically healthy nonobese, MUNO Metabolically unhealthy nonobese, MHO; and MUO Metabolically unhealthy obese, LDL-C Low-density lipoprotein cholesterol, HDL-C High-density lipoprotein cholesterol, TG Triglyceride, AF Atrial fibrillation, FBG Fasting blood glucose, LVMI Left ventricular mass index, LVM Left ventricular mass, LVDd Left ventricular end diastolic diameter, PWT Posterior wall thickness, IVST Interventricular septum thickness

Conclusions

The TyG index is a simple, inexpensive surrogate marker of insulin resistance that has been consistently demonstrated to be epidemiologically associated with incident cardiovascular disease, including coronary heart disease, myocardial infarction, stroke, heart failure, and subclinical atherosclerosis, across multiple observational studies. Overall, the evidence supports the use of TyG as a marker of adverse cardiometabolic risk rather than as a causal determinant of cardiovascular disease. Evidence supporting the predictive performance of TyG beyond established cardiovascular risk models remains limited and heterogeneous. Although several studies have reported improvements in discrimination, calibration, reclassification, or decision curve analyses, these findings are inconsistent across populations and often depend on study design, concomitant biomarkers, or newly developed prediction models. Consequently, current evidence does not establish that the TyG score consistently provides meaningful incremental prediction beyond validated clinical risk scores. Evidence regarding the calibration and clinical net benefit is also incomplete. Only a limited number of studies have formally evaluated these metrics, and improvements have not been consistently demonstrated across different populations or externally validated cohorts. Furthermore, no standardized TyG thresholds have been universally accepted for clinical risk classification. At present, there is insufficient evidence to recommend TyG as a treatment-guiding biomarker or as a determinant of clinical decision-making. Although TyG may help identify individuals with an adverse metabolic phenotype who warrant comprehensive cardiovascular risk assessment, no prospective studies have demonstrated that TyG-guided management improves cardiovascular outcomes beyond guideline-directed care. Several important limitations remain. Most available evidence is observational, substantial heterogeneity exists among study populations and outcome definitions, standardized cutoffs have not been established, and data from non-Asian populations remain relatively limited. Future research should prioritize prospective multicenter studies with external validation, standardized TyG thresholds, comprehensive evaluation of discrimination, calibration, reclassification, and clinical net benefit, and randomized or implementation studies to determine whether TyG-guided strategies improve clinically meaningful outcomes before routine clinical implementation can be recommended.

Authors’ contributions

Mariia Cherska - A.B.C.D.Khrystyna Kukharchuk - A.B.C.D.Alexandr Ceasovschih - C.DVictorita Sorodoc - C.D.Federica Fogacci - E.F.Tatyana Storozhenko - E.F.A. M. Abd El-Aty - A.B.All authors reviewed the manuscript.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Competing interests

The authors declare no competing interests.