The study population for this cross-sectional study was identified from high-risk cardiovascular disease subjects in the pilot screening survey of the longitudinal study (China PEACE Million Persons Project) initiated by the National Center for Cardiovascular Disease in Zhejiang Province, Wenzhou City, from 2019 to 2020 (20). This project was jointly approved by the Central Ethics Committee of the National Center for Cardiovascular Diseases and the Ethics Committee of Wenzhou People's Hospital. Since its inception in 2014, the China PEACE Million Persons Project has been continuously carried out nationwide, increasing to 383 project sites and screening over a million people by 2023. The main screening targets of the study were permanent residents of the project sites, with the inclusion criteria being aged between 35 and 75 years old, and having lived in the selected area for at least six months in the past 12 months with formal household registration (a formal record for identifying residents of the area). All participants provided written informed consent. Data were missing for less than 5% of participants, and audit results indicated high overall data quality.

In the preliminary screening phase, experienced project team members, trained in cardiovascular health, identified participants’ cardiovascular risk factors through initial assessments, including inquiries about cardiovascular health history, physical examinations, and rapid blood glucose and lipid level tests. They assessed the likelihood of participants developing cardiovascular diseases to determine if they were part of the high-risk group. Participants were considered high-risk individuals if they met any of the following criteria: (1)

Past medical history (meeting any of the following conditions is sufficient): (a)

History of myocardial infarction;

The assessment considered the following factors: (a)

Age;

During general examinations, general characteristics of participants were recorded, including age, gender, body mass index (BMI), systolic blood pressure (SBP), and diastolic blood pressure (DBP). In blood tests, blood glucose levels (BG), triglycerides (TG), glycated hemoglobin (HbA 1c), low-density lipoprotein cholesterol (LDL-C), and high-density lipoprotein cholesterol (HDL-C) were included, using fasting blood samples. Participants’ long-term medication history (antihypertensive drugs, lipid-lowering drugs, hypoglycemic drugs) and lifestyle (including smoking and drinking habits) were obtained through standardized self-administered questionnaires. Hypertension is defined as self-reported hypertension, use of antihypertensive drugs, SBP ≥ 140 mmHg, DBP ≥ 90 mmHg (or both) (22, 23). Diabetes is defined as the use of hypoglycemic drugs, glycated hemoglobin levels ≥6.5%, and fasting blood glucose levels ≥7 mmol/L (24). According to the definition by the National Institute on Alcohol Abuse and Alcoholism, drinking habits are divided into never drinking, light drinking (1–2 cups/day), moderate drinking (3–4 cups/day), or heavy drinking (>5 cups/day) (25). Smoking status is divided into non-smoker, former smoker, and current smoker.

According to previous studies (26), the formula for calculating the TyG index in this study is as follows:TyG=ln(TG(mg/dL)×FBG(mg/dL)2)

Carotid ultrasound examinations were conducted by trained and experienced sonographers who were blinded to the baseline characteristics and laboratory test results of the subjects. Equipment settings included a high-end color Doppler ultrasound device with a linear array transducer at a frequency of 5–12 MHz. Observation site: the actual observation of the entire carotid artery plaque occurrence, including the internal carotid artery, external carotid artery, carotid bifurcation, and the main trunk of the common carotid artery. Examination process: first take the transverse section, then take the longitudinal section, starting from the proximal end of the common carotid artery, from near to far, gradually sweeping towards the distal end, observing the intima-media display of the common carotid artery, bifurcation, internal and external carotid arteries. First, check the left side, then the right side. Intima-media thickness (IMT) definition: refers to the vertical distance between the upper edge of the intima and the lower edge of the media of the carotid artery (27). Carotid plaque is defined as local thickening of the intima-media ≥1.5 millimeters or thickening >50% of the surrounding IMT value (28).

Participants were categorized into four groups based on the quartiles of the TyG index, and their characteristics were described. Categorical variables were represented by numbers (N) and percentages (%), and group comparisons were assessed using the chi-square test. Continuous variables were described using means and standard deviations if they were normally distributed; otherwise, medians and interquartile ranges were used. Group comparisons for normally distributed continuous variables were made using one-way analysis of variance, while the Mann–Whitney U test was used for non-normally distributed continuous variables.

Single-factor and multiple-factor logistic regression models were employed to analyze the relationship between the TyG index and carotid plaque. Model Ⅰ was adjusted for age, sex, BMI, smoking, drinking, and waistline. Model Ⅱ further adjusted for additional physiological and biochemical indicators, including blood pressure, liver function indicators (AST/ALT), blood lipid levels (TC, LDL-C, HDL-C), and renal function indicators (creatinine), antilipidemic medication use, as well as socioeconomic factors such as education level and occupation, to more comprehensively assess the association between the TyG index and carotid plaque and control for potential confounding variables. The selection of confounding factors was based on three principles: first, variables were chosen according to their clinical significance; second, relevant studies were reviewed for reference (12, 29, 30); and third, covariates were adjusted if their inclusion changed the odds ratios by less than 10% (31). These three criteria were used to determine the necessary adjustments for confounding factors.

A generalized additive model was utilized to evaluate the nonlinear relationship between the TyG index and carotid plaques. In cases where a nonlinear association was detected, a two-piece linear regression model was applied to calculate the threshold effect of the TyG index on carotid plaques (32). When a clear relationship between the TyG index and carotid plaques was evident on a smoothed curve, recursive methods were automatically employed by the software to determine inflection points after adjusting for potential confounders. A likelihood ratio test was used to examine the interaction effect of gender on the relationship between the TyG index and carotid plaques. The statistical analyses of this study were performed via R, version 4.2.0 (R Foundation) and EmpowerStats (http://www.empowerstats.com, X&Y Solutions, Inc., Boston, MA). The level of statistical significance was set at p < 0.05.

This cross-sectional study evaluated a total of 1,085 participants (Figure 1). The quartiles of the TyG index were used for classification (Table 1). The TyG index increased significantly with each ascending quartile, and concurrent significant changes were observed in waist circumference, BMI, systolic and diastolic blood pressure, ALT and AST levels, blood glucose, HDL-c, LDL-c, TC, and TG. The distribution of gender indicated a higher proportion of male participants with higher TyG index values. The prevalence of smoking and drinking habits, as well as hypertension and diabetes, also significantly changed with an increase in the TyG index. In addition, the use of antihypertensive and antidiabetic medications became more common with the increase in the TyG index. Associations between urinary protein levels, occupation type, and education level with the TyG index were not significant or only had weak statistical relevance.

Firstly, potential risk factors or protective factors were included in the univariate logistic regression analysis. In the results, the TyG index, gender, age, waist circumference, systolic blood pressure, diastolic blood pressure, creatinine, and blood glucose levels were significantly associated with carotid plaque. When included as a continuous variable, an increase of one unit in the TyG index increased the risk of carotid plaque formation by 30% (OR:1.30, 95% CI: 1.07–1.58, p = 0.009) (Table 2). Subsequently, to assess the degree of association between the TyG index and carotid plaque, a multivariate logistic regression analysis was conducted. Table 3 demonstrated the relationship between the TyG index and carotid plaque using a multivariate logistic regression model. In Model I, the association between the TyG index and carotid plaque was slightly weakened but remained significant (OR 1.30, 95% CI 1.03–1.64, P = 0.028). However, in Model II, which further adjusted for smoking, alcohol, education, occupation, SBP, DBP, AST/ALT, creatinine, TC, HDL-c, LDL-c, antilipemic medication, antihypertension medication, diabetes, urinary protein and history of myocardial infarction, the association between the TyG index and carotid plaque was no longer significant (OR = 1.46, 95% CI 0.98–2.18, P = 0.066).

We employed generalized additive models (GAM) and smooth curve fitting to further investigate the potential nonlinear relationship between the TyG index and carotid plaques in the overall population and across different gender groups (adjusted for age, sex, BMI, waistline, smoking, alcohol, education, occupation, SBP, DBP, AST/ALT, creatinine, TC, HDL-c, LDL-c, antilipemic medication, antihypertension medication, diabetes, urinary protein and history of myocardial infarction) (Figure 2). Figure 2A illustrates that as the TyG index increases, the probability of forming carotid plaques also increases and exhibits a nonlinear correlation. Interestingly, as shown in Figure 2B, unlike in females, the relationship between the TyG index and carotid plaques in males follows an S-shaped curve. The interaction effect of gender was statistically significant (P = 0.030). By using a recursive algorithm, we further identified two different inflection points (8.42, P = 0.008 and 9.23, P = 0.079) (Table 4). When the TyG index is below 8.42, each unit increase is significantly associated with an increased risk of carotid plaque (OR = 5.63, 95% CI 1.16–27.26, P = 0.032); however, when the TyG index is above 8.42, the risk increase is not significant (P = 0.055). Another inflection point occurs at 9.23, where the risk ratio for values below this threshold is 1.06, which is not significant (P = 0.805), whereas for values above 9.23, the risk significantly escalates to 5.68 (95% CI:1.58–20.47, P = 0.008).

The effect modification of gender on the relationship between TyG and carotid plaques across different subgroups is presented in Supplementary Figure S1. The results indicate that gender differences persist across various subgroups. In the hypertensive population, the interaction between gender and TyG is significant (P for interaction = 0.029).