Triglyceride-glucose index and coronary artery disease: a systematic review and meta-analysis of risk, severity, and prognosis

This is a registered meta-analysis on the International Prospective Register of Systematic Reviews (https://​www.​crd.​york.​ac.​uk/​prospero/​) with registration number CRD42023422917. The systematic review and meta-analysis included cross-sectional studies, retrospective or prospective cohort studies that recruited patients with CAD, regardless of their nationality, race, age, gender, or course of the disease.

Two authors (S. Liang and J. Zhang) independently searched the PubMed, EMBASE, The Cochrane Library, and Web of Science databases for articles published from inception until May 1, 2023, using the following heading terms: “coronary artery disease”, “coronary heart disease”, “atherosclerotic cardiovascular diseases”, “CAD”, “CHD”, “triglyceride-glucose index”, and “TyG index”. The search was carried out by combining subject words and free words. No language restrictions were used. Relevant literature references were also searched to identify more eligible studies. The literature titles and abstracts were screened for primary screening, and then full-text acquisition and reading of the literature for rescreening were conducted.

The TyG index is calculated as ln[fasting triglycerides (mg/dL) × fasting glucose (mg/dL)/2]. For the analysis of CAD risk and severity, coronary artery calcification (CAC) is defined as coronary artery calcification score (CACS) > 0, and coronary artery stenosis is defined as the maximum intraluminal stenosis in any of the segments of the major epicardial coronary arteries > 70% [8]. Coronary plaque is defined as structures ≥ 1 mm² within or adjacent to the coronary artery lumen, which is characterized from the vessel lumen or surrounding pericardial tissues [8], and plaque progression is defined as the difference of the baseline and follow-up CACS or plaque volume > 0. For the analysis of CAD prognosis, the primary outcome is the major adverse cardiovascular event (MACE), which is defined as the composite outcome of all-cause death, cardiac death, myocardial infarction, revascularization, stroke, and heart failure. The secondary outcomes are all-cause death, cardiac death, myocardial infarction, revascularization, and stroke.

Two independent readers (S. Liang and C. Wang) extracted the data. They were not blinded to the authors and institutions of included studies. Disagreements were resolved by a third reader (Z. Liu) while Y. He and H. Huang supervised the entire process. This meta-analysis followed the guidelines for the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) [9].

The two reviewers independently extracted the following information: the first author, published year, sample size, demographic information, and variables adjusted in multivariate analysis. The risk of bias was assessed using the Newcastle–Ottawa Scale (NOS). The NOS ranges from 0 (lowest) to 9 (highest), and studies with scores ≥ 6 are considered high quality [10].

The RevMan version 5.3.5 (The Cochrane Collaboration, Copenhagen, Denmark) was properly used in all statistical analyses. Results were compiled using PRISMA. The two authors collecting the data, S. Liang and C. Wang, were not aware of the authors and institutions of included studies.

When the TyG index was analyzed as a categorical variable, the odds ratios (ORs) or hazard ratios (HRs) of patients with the highest TyG index level compared to those with the lowest TyG index level were extracted. When the TyG index was analyzed as a continuous variable, the HRs of the outcome incidence per 1-unit or 1-standard deviation (SD) increment of the TyG index were extracted. Statistical heterogeneity was assessed using the I square test. Heterogeneity was interpreted as absent (I²: 0–25%), low (I²: 25.1–50%), moderate (I²: 50.1–75%), or high (I²: 75.1–100%). A random-effects model was considered when the number of studies was relatively small, and a random-effects model was applied to estimate the continuous outcome data if the P-value < 0.1 and an I² value > 50%, indicating statistical heterogeneity [11]. Otherwise, a fixed-effects model was used. A P < 0.05 was regarded as statistical significance for the pooled OR and HR.

The authors conducted an online search using databases such as PubMed, EMBASE, The Cochrane Library, and Web of Science. The search initially yielded 381 literature citations, which were reduced to 141 after removing duplicates and irrelevant studies. A review of titles and keywords resulted in the exclusion of 80 studies, leaving 61 abstracts for evaluation by two authors (S. Liang and C. Wang). Thirteen studies were excluded due to their focus on heart failure or arterial stiffness. After full-text evaluation, 41 studies were selected for inclusion in the analysis. The systematic literature search and study selection process is displayed in Additional file 1: Figure S1 using the PRISMA flow chart.

The authors assessed the quality of the included studies using the NOS which rates the quality of non-randomized studies based on three criteria: selection, comparability, and outcome. The NOS score of most of the included studies was above 6, indicating good quality. Additional file 1: Table S1 provides the details of the NOS scores for each study.

The authors identified five studies that evaluated the relationship between the TyG index and CAD risk in patients without CAD or with suspected CAD [12‐16]. Details of these studies are presented in Table 1. One study [13] found that the TyG index is an independent risk factor for sub-clinical CAD in asymptomatic patients (OR: 2.007, 95% CI 1.066–3.780, P = 0.031). The other four studies included postmenopausal women [14], non-alcoholic fatty liver disease (NAFLD) patients [15], hypertensive patients [16], or patients who had at least one CVD in the last 10 years [12]. The pooled results of four studies [12, 14‐16] showed that individuals with higher TyG index levels were significantly more likely to have CAD than those with lower TyG index levels (OR: 1.94, 95% CI 1.20–3.14, I² = 91%, P = 0.007, Fig. 1). This finding was consistent when the TyG index was analyzed as a continuous variable in one study [16] (OR per 1-SD increment of the TyG index: 1.49, 95% CI 1.30–1.74, P = 0.007).

Four studies evaluated the relationship between the TyG index and coronary artery stenosis and/or calcification [17‐20], four studies evaluated the relationship between the TyG index and coronary plaque [21‐24], four studies evaluated the relationship between the TyG index and multi-vessel CAD [19, 25‐27], whilst two studies evaluated the relationship between the TyG index and in-stent re-stenosis (ISR) after drug-eluting stent (DES) [28, 29]. Details of these studies are presented in Table 1. One study [18] found that the TyG index is associated with a higher incidence rate of CAC in asymptomatic adults (OR: 1.95, 95% CI 1.23–3.11, P = 0.01). Ding et al. [20] also showed that in asymptomatic, non-diabetic patients undergoing maintenance hemodialysis, the prevalence ratio of CAC was 1.281 (95% CI 1.121–1.465).

The pooled results of the studies showed that when the TyG index was analyzed as a continuous variable, higher levels were associated with an increased likelihood of stenotic coronary artery (two studies [17, 19], OR: 3.49, 95% CI 1.71–7.12, I² = 0%, P = 0.0006), progressed plaques (two studies [21, 22], OR: 1.67, 95% CI 1.28–2.19, I² = 0%, P = 0.002), and more vessels involved (two studies [17, 25], OR: 2.33, 95% CI 1.59–3.42, I² = 0%, P < 0.0001). Figure 2 provides details of these results.

In this meta-analysis, a total of fifteen studies evaluated the prognosis of patients with ACS [30‐44], seven studies evaluated the prognosis of patients with CCS or stable CAD [29, 45‐50], and one study evaluated the prognosis of patients with myocardial infarction with non-obstructive coronary arteries (MINOCA) [51]. The details of the studies are shown in Table 2 and Additional file 1: Table S2.

The meta-analysis of thirteen studies [30, 32‐42, 44] revealed that ACS patients with the highest TyG index level had a significantly increased incidence rate of MACE compared to those with the lowest level (HR: 2.09, 95% CI 1.68–2.62, I² = 87%, P < 0.00001). This association was also observed when the TyG index was analyzed as a continuous variable using five studies [35‐37, 39, 41], with an HR per 1-unit/1-SD increment of the TyG index of 2.28 (95% CI 1.44–3.63, I² = 95%, P = 0.0005). Figure 3 provides more details on these findings.

Five studies [45, 46, 48‐50] indicated that individuals with the highest TyG index level have a higher incidence rate of MACE compared to those with the lowest level among CCS or stable CAD patients, although the trend was not statistically significant (HR: 1.24, 95% CI 0.96–1.60, I² = 85%, P = 0.09). Four studies analyzed the TyG index as a continuous variable [45, 46, 49, 50], and found that the risk of MACE increased by 1.49 times per 1-unit/1-SD increment of the TyG index (95% CI 1.21–1.83, I² = 75%, P = 0.0001) (Fig. 4). Only one study [51] analyzed the TyG index as a continuous variable, and reported an HR of 1.85 per 1-unit increment of the TyG index (95% CI 1.17–2.93, P = 0.008). Additional file 1: Table S3 and Figure 5 present the results of the secondary outcomes.

Additional file 1: Figure S2 shows the funnel plots for the association between the TyG index and the incidence rate of MACEs in ACS patients. The plots appear symmetrical upon visual inspection, indicating a low risk of publication bias. However, due to a limited number of datasets, it was challenging to estimate publication bias for the meta-analysis of other outcomes and diseases such as CCS and MINOCA.

The TyG index is a composite index that combines both fasting triglyceride levels and fasting blood glucose levels. It is a highly sensitive and specific marker for identifying IR, which is a risk factor for cardiovascular events [52]. An increased baseline TyG index can help identify individuals who are at high risk for these events. This meta-analysis found that individuals with the highest TyG index have a greater risk of coronary artery disease, more severe coronary artery lesions, and a worse prognosis compared to those with the lowest TyG index (Fig. 6).

IR may precede the onset of diabetes and CVDs. Even apparently healthy adults may still have IR, which is a key factor in the development of atherosclerosis [55]. In healthy individuals, IR can lead to abnormal metabolism and promote the development of CAD. A previous meta-analysis suggested that a higher TyG index may serve as an independent predictive indicator for an increased risk of CAD incidence in individuals without pre-existing CAD [7]. IR may also serve as a latent risk factor for the interactions between glucose intolerance and CAD [53]. Under normal physiological conditions, insulin induces vasodilation actions by producing nitric oxide. However, dysregulated insulin signaling can harm bioavailable nitric oxide and contribute to vascular stiffening [54]. Additionally, in patients receiving coronary artery bypass graft (CABG) surgery, a higher TyG index is proven to be associated with symptomatic graft failure due to endothelial dysfunction caused by the proinflammatory and procoagulatory effect of IR [49]. Our study contributes additional evidence on the association between the TyG index and the risk of CAD, which indicated that individuals with a higher TyG index have an increased susceptibility to developing CAD.

Nonenzymatic glycosylation of lipids is one of the factors that contributes to atherosclerosis, along with hypertriglyceridemia and hyperglycemia. In particular, hypertriglyceridemia can lead to the formation of low-density lipoproteins, which can promote atherosclerosis and weaken the protective effect of high-density lipoproteins [56]. What is more, hyperglycemia can be directly and/or indirectly related to the acceleration of atherosclerosis [57]. The severity of CAD is strongly associated with the extent of arterial stenosis and calcification, the presence of coronary plaque, the occurrence of multi-vessel lesions, and the development of ISR following the placement of DES. In a study by Thai et al. [19], a TyG index above 10 was significantly associated with coronary stenosis above 70%, with a sensitivity of 57% and specificity of 75%. Research has shown that the TyG index, compared with HOMA-IR, is more independently associated with the presence of calcified coronary artery plaques in healthy Korean adults undergoing cardiac computed tomography [18], even in those without traditional CVD risk factors [23]. Higher TyG index was also found to be associated with an increased presence of non-calcified and mixed coronary artery laques [23]. Multiple vessel lesions are strongly associated with a poor prognosis in CAD and can increase the complexity of PCI procedures. Wang et al. has demonstrated a dose–response relationship between the TyG index and the risk of developing multiple vessel CAD, with this association being particularly significant in males and older individuals [26]. ISR poses a significant challenge following PCI, and there is an independent positive correlation between an elevated TyG index and DES-ISR. However, the additional predictive value provided by the TyG index is minimal [28]. Therefore, current guidelines recommend evaluating the 10-year CVD risk in apparently healthy adults [58, 59], and the TyG index can provide new insight for identifying the development of CVDs at an early stage.

Compared to CAD patients without metabolic syndrome (MeS), those with MeS tend to have more severe coronary lesions and a higher incidence rate of MACE [60]. Even in patients with MINOCA, the presence of MeS significantly increases the risk of MACE [61]. In a retrospective cohort study of critically ill patients with CAD, a higher TyG index was found to be associated with greater mortality, as well as increased length of stay in the intensive care unit (ICU) and hospital, which imposes a significant financial burden on both families and society [62]. Consistent with previous studies, our meta-analysis found that ACS, CCS, and MINOCA patients with higher TyG index have an increased incidence rate of MACE, indicating that higher TyG index is associated with poorer outcomes in CAD.

The relationship between the TyG index and poor prognosis in CAD patients may be due to the association between IR and CAD. The TyG index reflects IR from both fasting blood glucose, which indicates IR from the liver, and fasting triglycerides, which indicate IR from adipose cells [63]. Therefore, the TyG index provides a more comprehensive evaluation of IR. IR can lead to an increase in sympathetic nervous system activity and renal sodium retention, along with hyperinsulinemia, resulting in increased blood pressure and elevated heart burden. Additionally, the metabolic effects of IR, such as hyperglycemia and dyslipidemia, have a synergistic effect with elevated blood pressure, leading to damage to the vascular and kidney systems, which can easily cause damage to the renal and cardiovascular systems [64]. Moreover, IR and insufficient insulin signaling can cause inappropriate activation of the renin–angiotensin–aldosterone system, oxidative stress, inflammation, and dysfunctional immune modulation, all of which can compromise cardiac function [65, 66]. For patients without diabetes, IR has also been found to be associated with reduced cardiac autonomic function, particularly vagal activity [67]. Furthermore, Gao et al. [68] found that in patients with chronic total occlusion lesions, an increased TyG index was strongly associated with less developed collateralization, indicating that these patients may face a larger ischemic area.

The baseline TyG index can help identify individuals at a higher risk of cardiovascular events at an early stage. In comparison to other indices such as triglycerides, atherogenic index of plasma, triglycerides to high-density lipoprotein cholesterol ratio, and lipoprotein combine index-adjusted HR, the TyG index has shown the best ability to predict MACE [52]. In clinical practice, different scoring systems are commonly used to manage patients with CAD throughout their treatment. The TyG index has shown to have independent and commendable performance in various aspects related to CAD. Therefore, it is possible that combining the TyG index with these scoring systems may provide additional value in managing CAD patients [69, 70].

The Framingham risk score (FRS) is a tool used to assess the risk of CVD in individuals. It takes into account factors such as gender, age, total cholesterol, high-density lipoprotein, systolic blood pressure level, and smoking status to stratify individuals based on their risk level. Guo et al. [71] found that the TyG index is significantly correlated with intermediate- or high-risk patients for CVD, suggesting that the TyG index can be used as an additional factor in assessing CVD risk. Sánchez-Íñigo et al. [52] included the TyG index in the FRS, and found the predictive accuracy of FRS was improved, especially in patients with intermediate risk (10–20% risk). In a 16-year follow-up study by Barzegar et al. [72], the TyG-index was found to be significantly associated with the risk of CVD/CHD incidence, especially among individuals younger than 60 years old. However, adding the TyG-index to the FRS did not provide better predictive ability for CVD risk. The differences of the population may be one possible explanation for the variations in different research findings.

The TyG index can also be applied to other scoring systems. The Gensini score and the SYNTAX score reflect the burden of plaque, the type and complexity of plaques, respectively. In our included studies, Shen et al. [44] found that incorporating the TyG index into the prediction model, along with left ventricular ejection fraction (LVEF) and Gensini score, significantly improves the ability for predicting the risk of all-cause mortality. In addition, Xiong et al. [73] also found that adding the residual SYNTAX score and TyG index to the baseline risk model had an incremental impact on the prediction of MACE. The GRACE score is used to predict in-hospital mortality of patients with ACS as well as the 6-month all-cause mortality rate after discharge. In our meta-analysis, Qin et al. [42] and Pang et al. [43] found that combining TyG index and the GRACE score can provide a better predictive value for the clinical prognosis of ACS patients.

In patients with ACS and CCS without diabetes [38, 48], the TyG index has been identified as an independent factor for those with well-controlled cholesterol levels. This indicates that the TyG index can be used as a tool for risk stratification and prognosis assessment in clinical practice, allowing for a comprehensive evaluation of metabolic status and cardiovascular risk in CAD patients. This tool can be utilized for clinical management and intervention strategies for CAD patients. However, it is important to note that the application of the TyG index in CCS patients may differ slightly from that in ACS patients. The value of the TyG index in ACS patients may be influenced by stress hyperglycemia [74], which can exacerbate the index and not reflect the true metabolic status of the patients. Additionally, ACS patients are more likely to experience a poor prognosis, which can obscure the prognostic effect of the TyG index.

What is more, one concerns have been raised regarding the use of baseline fasting triglycerides and fasting blood glucose as predictors for the prognosis of CAD in most studies. This is because CAD is a dynamic and progressive disorder, which makes it uncertain whether the TyG index based on a specific situation can accurately predict its progression [75]. To address this issue, Wang et al. monitored changes in the TyG index during the follow-up period and found that individuals with larger fluctuations in the TyG index were more likely to have a higher cardiovascular risk [76]. Similarly, Cui et al. proposed the use of the cumulative TyG index, which is calculated by summing up the average TyG index for each pair of consecutive examinations multiplied by the time between these two visits in years. They found that the cumulative TyG index was a better predictor of cardiovascular prognosis than the TyG index at baseline [6]. Therefore, dynamic monitoring of the TyG index during the follow-up duration may be a more effective approach to achieving good whole-course management in CAD patients [77].

It is also worth noting that an elevated level of triglycerides is linked to CVD and is a crucial factor in residual risk after statin therapy [78]. While epidemiological studies have found a correlation between triglycerides and cardiovascular risk, almost all clinical studies that have intervened with triglycerides have not yielded positive results [78]. The TyG index obtained by combining fasting glucose and triglycerides levels can provide more information. However, further research is needed to investigate whether drug intervention with the TyG index can improve prognosis.