Background: The relationship between the triglyceride–glucose (TyG) index and the risk of cardiovascular disease (CVD) in the U.S. population under 65 years of age with diabetes or prediabetes is unknown. The purpose of this study was to investigate the relationship between baseline TyG index and CVD risk in U.S. patients under 65 years of age with diabetes or prediabetes. Methods: We used data from the 2003–2018 National Health and Nutrition Examination Survey (NHANES). Multivariate regression analysis models were constructed to explore the relationship between baseline TyG index and CVD risk. Nonlinear correlations were explored using restricted cubic splines. Subgroup analysis and interaction tests were also conducted. Results: The study enrolled a total of 4340 participants with diabetes or pre-diabetes, with a mean TyG index of 9.02 ± 0.02. The overall average prevalence of CVD was 10.38%. Participants in the higher TyG quartiles showed high rates of CVD (Quartile 1: 7.35%; Quartile 2: 10.04%; Quartile 3: 10.71%; Quartile 4: 13.65%). For CVD, a possible association between the TyG index and the risk of CVD was observed. Our findings suggested a linear association between the TyG index and the risk of CVD. The results revealed a U-shaped relationship between the TyG index and both the risk of CVD (P nonlinear = 0.02583) and CHF (P nonlinear = 0.0208) in individuals with diabetes. Subgroup analysis and the interaction term indicated that there was no significant difference among different stratifications. Our study also revealed a positive association between the TyG index and comorbid MetS in the U.S. population under 65 years of age with prediabetes or diabetes. Conclusions: A higher TyG index was linked to an increased likelihood of CVD in the U.S. population aged ≤ 65 years with prediabetes and diabetes. Besides, TyG index assessment will contribute to more convenient and effective screening of high-risk individuals in patients with MetS. Future studies should explore whether interventions targeting the TyG index may improve clinical outcomes in these patients.
Keywords: TyG index; CVD; Diabetes; Pre-diabetes
Supplementary Information The online version contains supplementary material available at https://doi.org/10.1186/s12933-024-02261-8.
Cardiovascular disease (CVD) continues to increase in morbidity and mortality This upward trend imposes a substantial burden on both the healthcare system and overall human well-being, thereby emerging as a major public health issue of global concern [[
Insulin resistance (IR) is a pathophysiological disorder characterized by defective insulin regulation of glucose metabolism in tissue cells, which is primarily denoted by a diminished sensitivity and responsiveness of the body to insulin, potentially resulting in metabolic irregularities such as hyperglycemia, hyperlipidemia, and obesity [[
The TyG index is a useful, low-cost predictive marker of the risk of cardiovascular and coronary heart disease in the non-diabetic population [[
The National Health and Nutrition Examination Survey (NHANES), an ongoing project employing a complex, multistage probability sampling design to evaluate the health and nutritional status of the ambulatory population in the U.S., received approval from the Institutional Review Board of the National Center for Health Statistics (NCHS), with informed consent obtained from all participants. NHANES gathers questionnaire data through interviews, performs health screenings at mobile examination centers (MECs), and collects samples for laboratory testing. A comprehensive overview of the NHANES study and its data is accessible online at https://
We utilized NHANES survey cycles from 2003 to 2018, as these surveys provided comprehensive data on the TyG index and various cardiovascular conditions, including congestive heart failure (CHF), coronary heart disease (CHD), atherosclerotic cardiovascular disease (ASCVD), heart attack, angina, and stroke. Initially, 80,312 participants were enrolled in the study. After excluding individuals aged > 65 years (N = 10,489), those without diabetes or pre-diabetes (N = 9970), and those with missing data on the TyG index (N = 49,939), diabetes or pre-diabetes (N = 412) and specific cardiovascular conditions (N [CVD] = 5162 [CHF: 5095; CHD: 21; ASCVD: 0; heart attack: 11; angina: 28; stroke: 7]), our final analysis included 4340 eligible participants (Fig. 1).
Graph: Fig. 1 Flowchart of the sample selection from National Health and Nutrition Examination Survey (NHANES) 2003–2018
TyG was utilized as an exposure variable, and we computed the TyG index using the formula: Ln [triglycerides (mg/dL) * fasting glucose (mg/dL)/2]. Both triglyceride and fasting glucose concentrations were determined through an enzymatic assay employing an automatic biochemistry analyzer. Serum triglyceride concentration was measured using the Roche Modular P and Roche Cobas 6000 chemistry analyzers while fasting plasma glucose was assessed through the hexokinase-mediated reaction using the Roche/Hitachi Cobas C 501 chemistry analyzer.
The medical conditions section, identified by the variable name prefix MCQ, encompasses self- and proxy-reported personal interview data that spans a diverse array of health conditions and medical history for both children and adults. This section incorporates questions such as'Has a doctor or other health professional ever told you/SP that you/him/her... had CHF, CHD, angina (also called angina pectoris), heart attack (also called myocardial infarction), stroke, etc.?' These questions were denoted as MCQ160B-F in the household questionnaires administered during home interviews. Participants responding'yes' to any of these questions were classified as having a history of CVD. We established a composite endpoint for CVD, encompassing CHD, ASCVD, angina, stroke, and CHF as primary outcomes. Additionally, events related to CHD, ASCVD, angina, stroke, and CHF were separately analyzed as secondary outcomes.
Diabetes was defined as either treatment or medical diagnosis of hyperglycemia with hemoglobin A1c ≥ 6.5%, fasting plasma glucose (FPG) ≥ 126 mg/dL, or a 2-h blood glucose ≥ 200 mg/dL [[
MetS was defined according to the National Cholesterol Education Program/Adult Treatment Panel III criteria (NCEP-ATP III) [[
Data on various demographic and health-related factors were collected through NHANES household interviews. This encompassed details such as age, gender, race/ethnicity, educational level, family income, smoking status, alcohol consumption, and disease status. Body Mass Index (BMI) was computed by dividing weight (in kilograms) by the square of height (in meters). Participants were classified as normal weight (< 25 kg/m
According to NHANES analytic guidelines, statistical analyses were performed with appropriate sampling weights and accounting for complex multistage cluster surveys. Continuous variables were presented as means ± standard deviations, while categorical variables were expressed as percentages. Participants, categorized based on the TyG index quartiles, were compared utilizing a weighted Student's t-test for continuous variables or a weighted chi-square test for categorical variables. Multivariate logistic regression was utilized to investigate the association between the TyG index (independent variable) and the risk of CVD (dependent variable) through three distinct models for statistical inference. In model 1, no covariates were adjusted. Model 2 was adjusted for gender, age, and race. Model 3 involved adjustments for age, gender, race, education level, family income-poverty ratio (PIR), BMI, serum creatinine, serum uric acid, TC, LDL-C, HDL-C, ACR, eGFR, systolic blood pressure, diastolic blood pressure, hypertension, smoking status, and alcohol consumption. In sensitivity analyses, we categorized the TyG index into quartiles to assess the robustness of the results and examined the risk of CVD across these quartiles. Additionally, we employed restricted cubic spline (RCS) analysis with three piecewise points to explore potential nonlinear relationships between the TyG index and the CVD risk. For subgroup analysis concerning the association between the TyG index and the likelihood of CVD, we stratified the data by gender (male/female), BMI (normal weight/overweight/obesity), hypertension (yes/no), alcohol use (yes/no) and smoking status (never/former/now). These stratified factors were also considered as potential effect modifiers. In addition, we also used multivariate logistic regression to investigate the association between the TyG index and the likelihood of MetS in individuals under 65 years of age with prediabetes or diabetes. No were adjusted in Model 1. Model 2 was adjusted for gender, age, and race. Model 3 was adjusted for age, gender, race, education level, PIR, BMI, serum creatinine, serum uric acid, ACR, eGFR, smoking, and alcohol consumption status. For subgroup analysis concerning the association between the TyG index and MetS in individuals under 65 years of age with prediabetes or diabetes, we stratified the data by gender (male/female), BMI (normal weight/overweight/obesity), alcohol use (yes/no) and smoking status (never/former/now). A significance level of two-sided P < 0.05 was utilized to indicate statistical significance. All analyses were performed using R version 4.3.2 (http://www.R-project.org, The R Foundation).
In this study, 4340 participants were enrolled, with an average age of 48.48 ± 0.24 years. Among them, 54.96% were male, and 45.04% were female. The mean TyG index was 9.02 ± 0.02. The overall prevalence of CVD was 10.38% and this prevalence decreased as the TyG index increased across quartiles (Quartile 1: 7.35%; Quartile 2: 10.04%; Quartile 3: 10.71%; Quartile 4: 13.65%). Participants in higher TyG quartiles exhibited elevated rates of stroke (Quartile 1: 2.18%; Quartile 2: 2.24%; Quartile 3: 2.52%; Quartile 4: 4.69%), CHF (Quartile 1: 1.40%; Quartile 2: 2.96%; Quartile 3: 3.06%; Quartile 4: 4.04%), CHD (Quartile 1: 2.82%; Quartile 2: 4.22%; Quartile 3: 4.41%; Quartile 4: 5.46%), and ASCVD (Quartile 1: 6.87%; Quartile 2: 9.04%; Quartile 3:9.96%; Quartile 4: 12.08%).
Various factors, including age, gender, race, education level, BMI, serum uric acid, TC, LDL-C, HDL-C, ACR, eGFR, systolic blood pressure, diastolic blood pressure, hypertension, smoking status, and alcohol consumption, exhibited significant differences among the TyG index quartiles (all p < 0.05). Compared to the lowest TyG index group, participants in the higher TyG index group were significantly more likely to have diabetes and hypertension, elevated serum uric acid, TC, LDL-C, fasting glucose, triglyceride, HbA1c%, BMI, ACR, systolic blood pressure, diastolic blood pressure, and decreased eGFR and HDL-C, more likely to be male, Mexican American, poor education level, and former smokers. No statistically significant differences were observed between quartiles in serum creatinine, PIR, and the risk of heart attack and angina (all p > 0.05) (Table 1).
Table 1 Weighted baseline characteristics of the study population
TyG index All participants Quartile 1 (5.65–8.56) Quartile 2 (8.56–8.97) Quartile 3 (8.97–9.44) Quartile 4 (9.44–12.84) Age (year) 48.48 (0.24) 46.79 (0.43) 48.31 (0.54) 49.11 (0.47) 49.71 (0.47) Serum creatinine (mg/dL) 0.88 (0.01) 0.89 (0.02) 0.88 (0.02) 0.87 (0.01) 1.02 (0.02) 0.62 Serum uric acid (umol/L) 344.82 (1.89) 321.29 (3.25) 349.25 (3.17) 352.57 (3.71) 356.94 (3.94) Total cholesterol (mg/dL) 197.56 (1.09) 182.35 (1.62) 191.70 (1.67) 200.38 (1.66) 216.77 (2.04) HDL-C (mg/dL) 48.95 (0.35) 58.57 (0.61) 50.56 (0.51) 45.43 (0.43) 40.88 (0.45) LDL-C (mg/dL) 117.23 (0.87) 109.65 (1.40) 118.35 (1.46) 121.96 (1.55) 119.31 (1.96) Triglyceride (mg/dL) 164.91 (3.57) 70.64 (0.57) 113.82 (0.99) 164.93 (1.44) 318.45 (9.16) Fast glucose (mg/dL) 130.21 (0.83) 109.18 (0.73) 117.96 (1.06) 124.28 (1.11) 171.84 (2.83) HbA1c (%) 6,26 (0.03) 5.72 (0.03) 5.93 (0.05) 6.14 (0.04) 7.31 (0.08) BMI (Kg/m2) 32.39 (0.19) 29.96 (0.33) 32.62 (0.31) 33.31 (0.32) 33.74 (0.29) ACR (mg/g) 70.32 (7.68) 39.40 (10.57) 40.73 (9.40) 51.99 (11.26) 153.75 (24.96) eGFR (mL/min/1.73 m2) 95.88 (0.44) 97.34 (0.72) 96.23 (0.84) 95.75 (0.75) 94.25 (0.71) Systolic blood pressure (mmHg) 124.92 (0.33) 122.76 (0.67) 123.48 (0.70) 125.73 (0.73) 127.87 (0.64) Diastolic blood pressure (mmHg) 73.53 (0.26) 72.01 (0.47) 73.40 (0.54) 73.62 (0.48) 75.21 (0.48) TyG index 9.02 (0.02) 8.20 (0.01) 8.77 (0.00) 9.18 (0.01) 9.99 (0.02) Gender, % (SE) Female 45.04 (1.86) 50.46 (2.00) 48.13 (2.03) 41.94 (1.91) 39.05 (1.83) Male 54.96 (1.86) 49.54 (2.00) 51.87 (2.03) 58.06 (1.91) 60.95 (1.83) Races, % (SE) Mexican American 11.32 (1.12) 8.91 (1.10) 11.08 (1.20) 11.35 (1.08) 14.10 (1.51) Non-hispanic black 11.23 (1.08) 16.70 (1.34) 12.23 (1.10) 7.80 (0.80) 8.11 (0.94) Non-hispanic white 62.56 (1.82) 59.12 (2.22) 62.69 (2.17) 66.34 (1.79) 61.98 (2.58) Others 14.89 (1.34) 15.27 (1.29) 14.00 (1.22) 14.50 (1.29) 15.81 (1.67) Educational levels, % (SE) Less than 9th grade 7.52 (0.87) 5.08 (0.75) 7.50 (0.83) 7.55 (0.84) 10.14 (1.13) 9–11th grade 12.12 (1.15) 9.69 (0.87) 11.83 (1.24) 13.05 (1.44) 14.05 (1.42) High school graduate 25.35 (1.92) 26.35 (1.97) 24.21 (1.97) 26.13 (2.07) 24.73 (1.88) Some college or AA degree 31.49 (1.91) 31.32 (2.22) 33.66 (1.93) 28.78 (1.80) 32.11 (1.96) College graduate or above 23.52 (2.17) 27.56 (2.38) 22.80 (1.92) 24.50 (1.97) 18.97 (1.92) PIR, % (SE) 0.53 < 1 14.26 (1.44) 15.18 (1.35) 16.94 (1.72) 12.58 (1.27) 16.90 (1.57) 1–4 45.75 (2.21) 48.74 (2.30) 48.63 (2.41) 50.54 (2.49) 48.71 (2.45) > 4 33.01 (2.38) 36.08 (2.22) 34.43 (2.65) 36.88 (2.48) 34.39 (2.76) BMI, % (SE) Normal weight 14.09 (1.82) 25.58 (1.88) 13.46 (1.24) 10.13 (1.24) 7.46 (1.01) Overweight 28.61 (1.91) 33.11 (1.86) 30.62 (2.05) 25.67 (1.76) 26.48 (1.72) Obesity 56.01 (2.10) 41.31 (2.04) 55.93 (2.29) 64.20 (2.05) 66.06 (2.00) Smoke, % (SE) Never 50.62 (2.12) 56.94 (2.23) 50.46 (2.42) 49.44 (2.23) 45.36 (1.83) Former 26.96 (1.88) 21.64 (1.85) 27.38 (1.97) 29.36 (2.08) 29.75 (1.61) Now 22.38 (1.54) 21.42 (1.69) 22.16 (1.63) 21.19 (1.52) 24.89 (1.51) Alcohol use, % (SE) 77.31 (1.65) 81.65 (1.59) 76.99 (1.67) 76.12 (1.72) 74.31 (1.73) Hypertension, % (SE) 52.89 (1.93) 41.00 (2.16) 53.02 (1.86) 56.56 (2.38) 61.45 (2.38) CVD, % (SE) 10.38 (1.18) 7.35 (1.04) 10.04 (1.10) 10.71 (1.32) 13.65 (1.35) Stroke, % (SE) 2.88 (0.54) 2.18 (0.47) 2.24 (0.55) 2.52 (0.61) 4.69 (0.79) CHD, % (SE) 4.21 (0.78) 2.82 (0.65) 4.22 (0.71) 4.41 (0.89) 5.46 (0.96) CHF, % (SE) 2.85 (0.46) 1.40 (0.29) 2.96 (0.69) 3.06 (0.65) 4.04 (0.83) ASCVD, % (SE) 9.44 (1.21) 6.87 (1.04) 9.04 (1.11) 9.96 (1.17) 12.08 (1.29) Heart attack, % (SE) 4.50 (0.85) 3.44 (0.81) 9.32 (0.83) 4.34 (0.86) 4.75 (0.95) 0.39 Angina, % (SE) 3.36 (0.79) 1.87 (0.57) 4.69 (0.89) 3.20 (0.81) 3.74 (0.78) 0.06 DM, % (SE) 45.41 (2.05) 29.33 (1.80) 37.48 (2.12) 49.29 (2.07) 66.53 (2.31) PreDM, % (SE) 54.59 (2.13) 70.67 (1.80) 62.52 (2.12) 50.71 (2.07) 33.47 (2.31)
LDL-C, Low-density lipoprotein cholesterol; HDL-C, High-density lipoprotein cholesterol; ACR, Urinary albumin: creatinine ratio; eGFR, estimated-glomerular filtration rate; BMI, Body mass index; PIR, Family income-poverty ratio; DM, Diabetes; PreDM, Prediabetes. CHF, Congestive heart failure; CVD, Cardiovascular disease; CHD, Congestive heart disease; ASCVD, Atherosclerotic cardiovascular disease Bold value indicates the statistical significance
Table 2 illustrates the relationship between the TyG index and the risk of CVD. Our findings revealed that a higher TyG index was linked to an elevated risk of CVD. This association was significant in both our unadjusted model (OR = 1.46, 95%CI 1.26–1.68, p < 0.0001) and the minimally adjusted model (OR = 1.48, 95%CI 1.26–1.77, p < 0.0001). Following full adjustment, a positive association between the TyG index and the CVD risk remained consistent (OR = 1.65, 95%CI 1.20–2.25, p = 0.002), signifying that each unit of the TyG index was associated with a 65% increase in CVD risk. When categorizing the TyG index into quartiles, in the fully adjusted models, participants in the highest TyG index showed a significant 63% increased risk of CVD compared to those in the lowest quartiles (OR = 1.63, 95%CI 1.03–2.56, p < 0.001).
Table 2 The association between TyG index and the risk of CVD
CVD OR (95%CI) Model 1 Model 2 Model 3 TyG index (continuous) 1.46 (1.26, 1.68), 1.48 (1.26, 1.77), 1.65 (1.20, 2.25) TyG index (quartiles) Quartile 1 Reference Reference Reference Quartile 2 1.51 (1.05, 2.17), 1.45 (1.01, 2.11), 1.37 (1.08, 1.66), Quartile 3 1.64 (1.11, 2.16), 1.51 (1.07, 1.94), 1.49 (1.17, 1.80), Quartile 4 1.99 (1.39, 2.86) 1.97 (1.34, 2.90), 1.63 (1.03, 2.56),
Model 1: No covariates were adjusted Model 2: Age, gender, and race were adjusted Model 3: Age, gender, race, education level, PIR, BMI, serum creatinine, serum uric acid, total cholesterol, HDL-C, LDL-C, ACR, eGFR, systolic blood pressure, diastolic blood pressure, hypertension, smoking and alcohol consumption status were adjusted OR, odds ratio; 95%CI, 95% Confidence interval Bold value indicates the statistical significance
No significant association between the TyG index and the risk of stroke, heart attack, and angina was found in this study (Tables 3, 4, and 5).
Table 3 The association between TyG index and the risk of Stroke
Stroke OR (95%CI) Model 1 Model 2 Model 3 TyG index (continuous) 1.56 (1.23, 1.97), 1.62 (1.26, 2.07), 1.54 (0.89, 2.67) TyG index (quartiles) Quartile 1 Reference Reference Reference Quartile 2 0.97 (0.55, 1.72), 0.98 (0.54, 1.75), 0.64 (0.33, 1.26), Quartile 3 1.13 (0.58, 2.19), 1.09 (0.55, 2.18), 1.15 (0.73, 1.56), Quartile 4 2.15 (1.24, 3.75) 2.23 (1.28, 3.88), 1.63 (0.75, 3.57),
Model 1: No covariates were adjusted Model 2: Age, gender, and race were adjusted Model 3: Age, gender, race, education level, PIR, BMI, serum creatinine, serum uric acid, total cholesterol, HDL-C, LDL-C, ACR, eGFR, systolic blood pressure, diastolic blood pressure, hypertension, smoking and alcohol consumption status were adjusted OR, Odds ratio; 95%CI, 95% Confidence interval Bold value indicates the statistical significance
Table 4 The association between TyG index and the risk of heart attack
Heart attack OR (95%CI) Model 1 Model 2 Model 3 TyG index (continuous) 1.20 (0.92, 1.56), 1.17 (0.86, 1.60), 1.24 (0.89, 1.74) TyG index (quartiles) Quartile 1 Reference Reference Reference Quartile 2 1.38 (0.94, 2.87), 1.51 (0.86, 2.64), 1.28 (0.72, 2.29), Quartile 3 1.43 (0.67, 2.42), 1.16 (0.62, 2.20), 1.11 (0.59, 2.09), Quartile 4 1.72 (0.76, 2.59) 1.31 (0.70, 2.46), 1.30 (0.66, 2.56),
Model 1: No covariates were adjusted Model 2: Age, gender, and race were adjusted Model 3: Age, gender, race, education level, PIR, BMI, serum creatinine, serum uric acid, total cholesterol, HDL-C, LDL-C, ACR, eGFR, systolic blood pressure, diastolic blood pressure, hypertension, smoking and alcohol consumption status were adjusted OR, Odds ratio; 95%CI, 95% Confidence interval
Table 5 The association between TyG index and the risk of angina
Angina OR (95%CI) Model 1 Model 2 Model 3 TyG index (continuous) 1.30 (1.01, 1.67), 1.28 (0.95, 1.73), 1.03 (0.66, 1.61) TyG index (quartiles) Quartile 1 Reference Reference Reference Quartile 2 1.59 (1.31, 3.13), 1.42 (0.21, 2.85), 1.09 (0.51, 2.33), Quartile 3 1.74 (1.23, 3.75), 1.52 (0.69, 3.35), 1.23 (0.52, 22.86), Quartile 4 2.04 (1.35, 3.95) 1.89 (0.96, 3.74), 1.94 (0.91, 4.11),
Model 1: No covariates were adjusted Model 2: Age, gender, and race were adjusted Model 3: Age, gender, race, education level, PIR, BMI, serum creatinine, serum uric acid, total cholesterol, HDL-C, LDL-C, ACR, eGFR, systolic blood pressure, diastolic blood pressure, hypertension, smoking and alcohol consumption status were adjusted OR, Odds ratio; 95%CI, 95% Confidence interval Bold value indicates the statistical significance
For CHF, our study identified a positive association between the TyG index and an elevated likelihood of CHF with statistical significance (Table 6). In both our unadjusted model and minimally adjusted model, participants with higher TyG index levels exhibited an increased risk of CHF (Model 1: OR = 1.60, 95%CI 1.26–2.03, p < 0.001; Model 2: OR = 1.67, 95%CI 1.29–2.16, p < 0.001). After full adjustment, each unit increase in the TyG index was linked to a 47% increase in CHF risk (Model 3: OR = 1.47, 95%CI 1.03–2.09, p = 0.03). Even when considering the TyG index as quartiles, a statistically significant association persisted. Participants in the highest TyG index experienced a significant 107% higher risk compared to those in the lowest TyG index quartile (OR = 2.07, 95%CI 1.03–4.14, p = 0.04).
Table 6 The association between TyG index and the risk of CHF
CHF OR (95%CI) Model 1 Model 2 Model 3 TyG index (continuous) 1.60 (1.26, 2.03), 1.67 (1.29, 2.16), 1.47 (1.03, 2.09) TyG index (quartiles) Quartile 1 Reference Reference Reference Quartile 2 2.15 (1.17, 3.94), 2.10 (1.16, 3.81), 1.59 (1.11, 2.07), Quartile 3 2.22 (1.23, 4.00), 2.18 (1.20, 3.94), 1.63 (1.08, 2.18), Quartile 4 2.96 (1.65, 5.31) 3.03 (1.69, 5.45), 2.07 (1.03, 4.14),
Model 1: No covariates were adjusted Model 2: Age, gender, and race were adjusted Model 3: Age, gender, race, education level, PIR, BMI, serum creatinine, serum uric acid, total cholesterol, HDL-C, LDL-C, ACR, eGFR, systolic blood pressure, diastolic blood pressure, hypertension, smoking and alcohol consumption status were adjusted OR, Odds ratio; 95%CI, 95% Confidence interval
Tables 7 and 8 revealed a significant risk increase between the TyG index and the risk of CHD (Model 3: OR 1.51, 95%CI 1.14–2.00, p = 0.005) and ASCVD (Model 3: OR 1.37, 95%CI 1.06–1.76, p = 0.02). In the sensitivity analyses, in fully adjusted Model 3, the highest quartile of the TyG index demonstrated an increase in the risk of both CHD (OR 1.70, 95%CI 1.12–3.19, p = 0.02) and ASCVD (OR 1.48, 95%CI 1.19–2.09, p = 0.01).
Table 7 The association between TyG index and the risk of CHD
CHD OR (95%CI) Model 1 Model 2 Model 3 TyG index (continuous) 1.46 (1.19, 1.81), 1.47 (1.14, 1.889), 1.51 (1.14, 2.00) TyG index (quartiles) Quartile 1 Reference Reference Reference Quartile 2 1.52 (1.09, 1.95), 1.31 (1.05, 1.57), 1.11 (1.02, 2.00), Quartile 3 1.59 (1.11, 2.07), 1.38 (1.08, 1.68), 1.35 (1.09, 2.50), Quartile 4 1.99 (1.13, 3.51) 1.77 (1.11, 3.17), 1.70 (1.12, 3.19),
Model 1: No covariates were adjusted Model 2: Age, gender, and race were adjusted Model 3: Age, gender, race, education level, PIR, BMI, serum creatinine, serum uric acid, total cholesterol, HDL-C, LDL-C, ACR, eGFR, systolic blood pressure, diastolic blood pressure, hypertension, smoking and alcohol consumption status were adjusted OR, Odds ratio; 95%CI, 95% Confidence interval Bold value indicates the statistical significance
Table 8 The association between TyG index and the risk of ASCVD
ASCVD OR (95%CI) Model 1 Model 2 Model 3 TyG index (continuous) 1.41 (1.20, 1.65), 1.42 (1.18, 1.72), 1.37 (1.06, 1.76) TyG index (quartiles) Quartile 1 Reference Reference Reference Quartile 2 1.50 (1.05, 2.15), 1.42 (1.08, 2.06), 1.17 (1.06, 1.28), Quartile 3 1.56 (1.09, 2.04), 1.63 (1.14, 2.18), 1.43 (1.15, 1.74), Quartile 4 1.86 (1.25, 2.78) 1.81 (1.18, 2.76), 1.48 (1.19, 2.09),
Model 1: No covariates were adjusted Model 2: Age, gender, and race were adjusted Model 3: Age, gender, race, education level, PIR, BMI, serum creatinine, serum uric acid, total cholesterol, HDL-C, LDL-C, ACR, eGFR, systolic blood pressure, diastolic blood pressure, hypertension, smoking and alcohol consumption status were adjusted OR, Odds ratio; 95%CI, 95% Confidence interval Bold value indicates the statistical significance
We utilized restricted cubic spline (RCS) curves to assess potential nonlinearity in the relationship between the TyG index and the risk of CVD, CHD, CHF, and ASCVD, as illustrated in Figs. 2, 3, 4, and 5. Our results indicated that there was an approximately linear relationship between the TyG index and the risk of CVD (P overall = 0.0001, P nonlinear = 0.4961), CHD (P overall = 0.0076, P nonlinear = 0.816), CHF (P overall = 0.0309, P nonlinear = 0.9812), and ASCVD (P overall = 0.001, P nonlinear = 0.3509).
Graph: Fig. 2 The restricted cubic spline (RCS) analysis between the TyG index and the risk of CVD
Graph: Fig. 3 The restricted cubic spline (RCS) analysis between the TyG index and the risk of CHD
Graph: Fig. 4 The restricted cubic spline (RCS) analysis between the TyG index and the risk of CHF
Graph: Fig. 5 The restricted cubic spline (RCS) analysis between the TyG index and the risk of ASCVD
Besides, in our study, we examined the population of individuals with diabetes and pre-diabetes separately. The results revealed a U-shaped relationship between the TyG index and both the risk of CVD (P nonlinear = 0.02583) and CHF (P nonlinear = 0.0208) in individuals with diabetes (Supplemental Figs. 1 and 3). The relationship between the TyG index and the risk of CHD (P nonlinear = 0.6958) and ASCVD (P nonlinear = 0.4331) was linear in patients with diabetes (Supplemental Figs. 2 and 4).
For the patients with pre-diabetes, the TyG index and the risk of CVD (P nonlinear = 0.6193), CHD (P nonlinear = 0.6768), CHF (P nonlinear = 0.1515), and ASCVD (P nonlinear = 0.9134) exhibited an approximately linear relationship.
In our subgroup analyses and interaction tests, we investigated the relationship between the TyG index and the risk of CVD, CHD, CHF, and ASCVD across different population subgroups (Fig. 6).
Graph: Fig. 6 Subgroup analysis for the association between the TyG index and the risk of CVD, CHD, CHF, and ASCVD. A Subgroup analysis for the association between the TyG index and the risk of CVD. B Subgroup analysis for the association between the TyG index and the risk of CHD. C Subgroup analysis for the association between the TyG index and the risk of CHF. D Subgroup analysis for the association between the TyG index and the risk of ASCVD
The risk of CVD increased in participants who were male (OR 1.670, 95%CI 1.262–2.210, p < 0.001), alcohol users (OR 1.649, 95%CI 1.262–2.154, p < 0.001), former smokers (OR 1.705, 95%CI 1.063–2.736, p = 0.027), overweight (OR 1.852, 95%CI 1.159–2.959, p = 0.010) and obese (OR 1.367, 95%CI 1.016–1.840, p = 0.039). Individuals without hypertension (OR 2.550, 95%CI 1.600–4.063, p < 0.001) also had an elevated risk of CVD.
For the risk of CHD, a positive relationship was observed in participants who were male (OR 1.490, 95%CI 1.028–2.160, p = 0.035) obese (OR 1.643, 95%CI 1.135–2.380, p = 0.009), and with a history of alcohol consumption (OR 1.666, 95%CI 1.153–2.406, p = 0.007). Alcohol users (OR 1.673, 95%CI 1.049–2.669, p = 0.031) were also reported to experience an increased risk of CHF.
Regarding ASCVD, the TyG index was associated with an increased risk of ASCVD, with significant correlations observed in males (OR 1.629, 95%CI 1.204–2.204, p = 0.002), those who were alcohol users (OR 1.559, 95%CI 1.153–2.107, p = 0.004), former smokers (OR 1.684, 95%CI 1.044–2.717, p = 0.033), overweight (OR 1.783, 95%CI 1.102–2.884, p = 0.019), and individuals without hypertension (OR 2.696, 95%CI 1.665–4.365, p < 0.0001).
Interaction tests did not reveal any significant influence of gender, BMI, alcohol use, smoking status, or hypertension on the association between the TyG index and the risk of CVD, CHD, CHF, and ASCVD (all P for interaction > 0.05). In conclusion, there was no significant interaction between the baseline TyG index and stratified variables.
In our present study, the association between the TyG index and comorbid MetS in the U.S. population under 65 years of age with prediabetes or diabetes was further analyzed. For MetS, a positive correlation was observed between the TyG index and the likelihood of MetS with statistical significance (Supplemental Table 1). Both our crude (Model 1: OR = 6.84, 95%CI 5.63–8.30) and minimally adjusted models (Model 2: OR = 7.63, 95%CI 6.21–9.37) indicated that a higher TyG index was associated with an elevated likelihood of MetS. With full adjustment, a positive association between the TyG index and Mets remained stable (Model 3: OR = 7.22, 95%CI 5.75–9.06). Notably, even when stratifying the TyG index into quartiles, a significant association still persisted (Model 3: Quartile 2: OR = 1.45, 95%CI 1.14–1.83; Quartile 3: OR = 7.68, 95%CI 5.59–10.57; Quartile 4: OR = 26.54, 95%CI 18.57–37.92).
Stratified analysis was conducted across different gender, BMI, smoking, and drinking status groups to investigate potential heterogeneities. Supplemental Table 2 presents the associations between the TyG index and the likelihood of MetS within these groups. For MetS, a positive association was observed in both females (OR = 8.147, 95%CI 5.872–11.304) and males (OR = 7.708, 95%CI 5.560–10.685), both alcohol users (OR = 8.792, 95%CI 6.703–11.533) and non-alcohol users (OR = 4.521, 95%CI 3.111–6.571), among individuals with normal weight (OR = 25.781, 95%CI 11.241–59.128) and those classified as overweight (OR = 6.495, 95%CI 4.367–9.661) and obese (OR = 6.935, 95%CI 4.255–10.789), as well as those who were never smokers (OR = 7.440, 95%CI 5.556–9.963), former smokers (OR = 6.775, 95%CI 4.255–10.789) and current smokers (OR = 10.485, 95%CI 6.049–18.174). Additionally, the interaction test did not suggest significant differences among different stratifications, indicating that this positive association was not significantly influenced by gender, BMI, smoking, and drinking status (All P for interaction > 0.05) (Supplemental Table 2).
In this study, which included 4340 participants, we found that a higher TyG index was independently associated with an increased risk of CVD. In addition, we found a non-linear correlation between the TyG index and the risk of CVD and CHF in the diabetic population, as can be seen in the figure the threshold of the TyG index is 9.18, and the risk of CVD and CHF episodes increased significantly when the TyG index exceeded 9.18. There was no significant interaction between the baseline TyG index and stratified variables. In conclusion, our findings demonstrate that the TyG index can be used as a valid predictor of early cardiovascular risk in diabetic and prediabetic patients under 65 years of age in the United States. Besides, our study also revealed a positive association between the TyG index and comorbid MetS in the U.S. population under 65 years of age with prediabetes or diabetes. And results from the subgroup analysis suggested that this positive correlation was similar in different population settings.
Previous studies have evaluated the association between the TyG index and CVD risk in various populations. Liu et al. revealed that a heightened TyG index at baseline correlated with an increased risk of future CVD in postmenopausal women [[
The influence of age on the correlation between the TyG index and CVD risk remains unclear, with most investigations focusing on middle-aged and older cohorts. Li et al. highlighted the utility of the TyG index in predicting CVD risk among individuals aged 60 and above [[
There appears to be a threshold effect in the association between the TyG index and CVD risk, indicating that both excessively high and low TyG levels may negatively impact health prognosis. One study has identified a negative correlation with the risk of CVD mortality when the TyG index falls below the threshold of 8.84, and a positive correlation with CVD mortality when the TyG index surpasses the threshold of 8.84 [[
The underlying mechanism of the predictive role of the TyG index for future CVD risk is unknown but may be related to the following factors. The prolonged hyperglycemic state resulting from decreased sensitivity to insulin in insulin resistance (IR) can initiate heightened glycosylation. This process contributes to collagen deposition and the formation of chronic fibrosis in myocardial tissues, resulting in the deterioration of cardiac function [[
We also evaluated the association between the TyG index and comorbid MetS in the U.S. population under 65 years of age with prediabetes or diabetes. MetS is composed of a spectrum of metabolic disorders including central obesity, hypertension, abnormal glucose metabolism, dyslipidemia, etc., which has become a global public health problem due to its increasing prevalence [[
The strength of our research is based on the NHANES database. The expansive size of the sample ensures robust statistical power, while its representation of the entire U.S. population assures a heightened level of external validity. Additionally, all variables are meticulously gathered in a standardized and homogeneous manner. We mitigated confounding bias through the meticulous adjustment of covariates, thereby augmenting the robustness of our findings. However, it is imperative to acknowledge certain limitations inherent in our study. Firstly, it is imperative to note that our research is confined to a single-center observational study, precluding the establishment of causality. Additionally, the study cohort predominantly comprises individuals from the United States, thereby constraining its generalizability on a global scale. Secondly, despite our efforts to account for potential confounding covariates, the possibility of residual confounders cannot be completely dismissed. Moreover, the TyG index was derived from a solitary baseline blood sample, leaving the impact of variations in the TyG index throughout the follow-up period on the risk of cardiovascular events uncertain.
The findings of our study indicate that the TyG index could serve as a potentially valuable predictor of CVD risk in individuals under 65 years of age with diabetes or prediabetes in the United States. Furthermore, our results revealed a nonlinear relationship between the TyG index and the risk of both CVD and CHF in patients with diabetes. This has important clinical implications for the early identification of CVD risk in non-elderly individuals with diabetes and prediabetes. In addition, our study also revealed a positive association between the TyG index and comorbid MetS in the U.S. population under 65 years of age with prediabetes or diabetes. The study's insights suggest that future research should be directed towards exploring whether interventions targeting the TyG index can lead to improvements in clinical prognosis for younger patients with diabetes and prediabetes.
The authors thank the staff and the participants of the NHANES study for their valuable contributions.
CL analyzed the data and wrote the primary manuscript. DL reviewed and revised the manuscript. All the authors have approved the manuscript for publication.
Not applicable.
Publicly available datasets were analyzed in this study. This data can be found here: https://
The studies involving human participants were reviewed and approved by the NCHS Research Ethics Review Board (ERB). Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements.
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be recognized as a potential conflict of interest.
Graph: Supplementary Material 1.
Graph: Supplementary Material 2.
Graph: Supplementary Material 3.
Graph: Supplementary Material 4.
Graph: Supplementary Material 5.
Graph: Supplementary Material 6.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
By Chang Liu and Dan Liang
Reported by Author; Author