Abstract
Background
Cardiovascular disease (CVD) is a leading cause of death and disability worldwide, particularly among middle-aged and elderly populations. And early risk assessment is essential for early intervention and effective prevention. This study aims to investigate and compare the associations of the triglyceride, total cholesterol and body weight index (TCBI), a novel nutritional index, and the widely researched insulin resistance index-triglyceride glucose-body mass index (TyG-BMI) with CVD risk.
Methods
Using data from the CHARLS (2011–2020), including 8,104 participants without CVD at baseline. The associations of TCBI and TyG-BMI with CVD risk were assessed using Cox proportional hazards models and restricted cubic spline (RCS) analysis. Additionally, receiver operating characteristic (ROC) analysis, subgroup analysis and sensitivity analyses were also conducted.
Results
After 9 years, 1,840 participants developed CVD. Each 1-unit increase in log-transformed TCBI (LgTCBI) was associated with a 59% higher risk of CVD (HR 1.59, 95%CI 1.31–1.92, p < 0.001), while a 10-unit increase in TyG-BMI corresponded to a 6% higher risk (HR 1.06, 95% CI 1.04–1.07, p < 0.001). RCS analysis revealed a nonlinear relationship for TCBI and a linear one for TyG-BMI. ROC analysis indicated limited independent predictive accuracy for both indices, with area under the curve (AUC) values ranging from 0.556 to 0.576. In short-term prediction (2–4 years), TCBI and TyG-BMI showed similar discriminatory ability, whereas TyG-BMI performed slightly better in long-term prediction (7–9 years). The subgroup analysis indicated that there was no interaction between the subgroups and the two indices (p for interaction > 0.05). Sensitivity analyses confirmed the robustness of the results.
Conclusions
TCBI and TyG-BMI were significantly associated with an increased CVD risk. Both indices demonstrated limited discriminative ability as standalone predictive tools, but TCBI performed comparably to TyG-BMI in short-term risk assessment, TyG-BMI held a slight advantage in long-term prediction. These findings suggest that TCBI and TyG-BMI may serve as simple, complementary screening tools to help identify high-risk individuals warranting further comprehensive clinical evaluation, rather than as standalone primary screening instruments and their clinical applicability requires further validation in diverse cohorts.
Supplementary Information
The online version contains supplementary material available at 10.1186/s41043-025-01234-1.
Keywords: Cardiovascular disease, Triglyceride, Total cholesterol and body weight index, Risk assessment, Triglyceride glucose-body mass index
Introduction
Cardiovascular disease (CVD) is the foremost cause of death globally, responsible for approximately 17.9 million annual deaths, representing 32% of worldwide mortality [1, 2]. Aging populations, urbanization, and metabolic disorders contribute to rising CVD incidence and mortality, especially in developing nations [3]. In China, the mortality rate of CVD has surpassed 40% of total deaths [4]. This severe situation not only puts immense pressure on the healthcare system but also creates a heavy socio-economic burden. Therefore, developing accurate and accessible early risk assessment tools is vital for clinical and public health efforts.
Traditional models like the Framingham Risk Score and the Atherosclerotic Cardiovascular Disease Risk Assessment Tool (ASSIGN) have been used in CVD risk assessment [1]. However, while established, these models may have limitations in generalizing to Asian populations and their reliance on a multitude of clinical parameters can limit their scalability and ease of use in resource-constrained settings [5]. This has spurred the search for simpler, more cost-effective predictive biomarkers. Consequently, combined metrics capturing metabolic dysregulation have emerged as promising tools, as they may better reflect the complex, multifactorial nature of CVD pathogenesis [6, 7].
The triglyceride glucose-body mass index (TyG-BMI), as a composite indicator reflecting insulin resistance and obesity, has been shown by multiple studies to be significantly associated with CVD risk [6, 8, 9]. A meta-analysis of a cohort study involving 871,728 participants showed that compared to the lowest TyG-BMI index category, the highest TyG-BMI index was associated with a higher incidence of CVD (HR 1.62, 95% CI 1.35–1.95) [10]. However, TyG-BMI primarily reflects the synergistic effects of glucose metabolism, triglyceride abnormalities and obesity, with insufficient consideration of other important components in the lipid profile, such as total cholesterol (TC). As a classic risk factor for atherosclerosis, the role of TC in CVD development has been supported by a large number of studies [11–13].
Recently, the triglyceride, total cholesterol and body mass index (TCBI) has been proposed as a novel nutritional index that integrates three parameters: triglycerides (TG), TC and body weight (BW), potentially providing a more comprehensive perspective for metabolic risk assessment. Compared to conventional lipid parameters such as low-density lipoprotein (LDL), high-density lipoprotein (HDL) and the TG/HDL ratio, TCBI offers the advantage of combining lipid profile with body composition (BW). It captures the synergistic effects of metabolic and nutritional status through a single composite score, thereby simplifying risk stratification [14]. Researches have shown that TCBI could serve as a prognostic indicator for CVD and was significantly related to various diseases such as heart failure (HF), cardiomyopathy, and coronary artery disease (CAD) [15–18]. However, it is noteworthy that existing researches on TCBI primarily focus on its value as a nutritional assessment indicator, often compared with traditional nutritional indicators like the geriatric nutritional risk index (GNRI) and prognostic nutritional index (PNI) [19, 20]. There is still a lack of large-scale prospective evidence supporting the application value of TCBI in CVD risk assessment, especially regarding its comparative predictive efficacy with existing indicators such as TyG-BMI.
This study, based on the nationally representative large cohort data from the China Health and Retirement Longitudinal Study (CHARLS), aims to investigate the associations between TCBI, TyG-BMI and CVD risk as well as their predictive values. The findings may offer insights into the understanding of metabolic risk assessment in CVD by evaluating these novel composite indices and may help lay the groundwork for targeted preventive strategies.
Materials and methods
Data source and study population
Data came from CHARLS (http://charls.pku.edu.cn/), a nationally representative cohort using multistage stratified sampling. The study employed a scientifically rigorous multi-stage stratified probability sampling method. During the baseline survey conducted in 2011–2012, a nationally representative sample of middle-aged and elderly individuals aged 45 and above was recruited from 450 communities in 150 counties across 28 provincial-level administrative regions of China [21]. The initial response rate of the survey was 80.5%, with follow-up assessments conducted every two years [21]. The study covered various dimensions, including socio-demographic information, health status assessment, physical examinations, and laboratory tests.
This analysis utilized CHARLS survey data from 2011 to 2020, including 8,104 participants aged 45 or older who had complete data on TC, TG, fasting blood glucose (FBG), BW, and height. Individuals with incomplete CVD data or a history of CVD at baseline were excluded. The flowchart of the study was shown in Fig. 1.
Fig. 1.
The flowchart of this study. CVD, cardiovascular disease; FBG, fasting blood glucose; TG, triglyceride; TC, total cholesterol
The CHARLS was approved by the Biomedical Ethics Review Committee of Peking University (IRB00001052-11015). All participants provided informed consent. The study followed the Declaration of Helsinki and STROBE guidelines [22, 23].
Data collection
The data and survey instruments (questionnaires) used in this study were obtained from the CHARLS national database. The data collection of CHARLS involved questionnaires, physical exams and blood tests. The questionnaires, covering demographics, socioeconomics, health status and healthcare utilization were developed based on the well-established Health and Retirement Study (HRS) model and have been widely used and validated in the Chinese context [21]. During data collection, interviewers used the Computer-Assisted Personal Interviewing (CAPI) system to gather all personal, health and lifestyle data [21]. Health professionals measured blood pressure, height, and weight. BMI was calculated as weight/height² (kg/m²). Fasting blood samples were analyzed at Youan Hospital, Capital Medical University, with < 2% coefficient variation for TC, TG, and FBG [24].
Collected data included demographics: gender, age, education (illiterate, primary school and below, middle school and above), residence (urban/rural), marital status (single/married). Lifestyle: smoking and drinking. Physical measures: systolic blood pressure (SBP), diastolic blood pressure (DBP) and BMI. Health conditions: hypertension, diabetes mellitus (DM), cancer, lung disease, psychiatric disorders, depression, arthritis, dyslipidemia, hepatic diseases, kidney diseases, digestive diseases, asthma and memory-related diseases. Laboratory tests: C-reactive protein (CRP), uric acid (UA), FBG, TG, TC, HDL and LDL. Hypertension was diagnosed when the SBP exceeded 140 mmHg and/or the DBP exceeded 90 mmHg, DM was defined when fasting plasma glucose ≥ 126 mg/dL and/or a HbA1c ≥ 6.5% and dyslipidemia was defined as TC ≥ 6.2 mg/dL, TG ≥ 200 mg/dL, LDL ≥ 160 mg/dL and/or HDL ≤ 40 mg/dL according to Chinese guidelines [25]. Additionally, self-reported hypertension, diabetes, and dyslipidemia were also employed for diagnosis. Depression was assessed using the 10-item Center for Epidemiologic Studies Depression Scale (CES-D-10) [26]. Other health conditions were all determined through self-reported.
Definitions of TCBI and TyG-BMI assessment
The calculation formula for TCBI is: TCBI = TG (mg/dL) × TC (mg/dL) × BW (kg)/1,000 [27]. The calculation formula for TyG-BMI is: ln (TG (mg/dL) × FBG (mg/dL)/2) × BMI [6].
Assessment of CVD
Incident CVD was defined as new self-reported heart disease or stroke during follow-up, confirmed via interview: “Has a doctor diagnosed you with stroke/heart attack/CHD/angina/heart failure/other heart problems?” Affirmative answers indicated CVD.
Missing data processing
The extent of missing data for each variable is presented in Supplementary Table 1. Although the proportion of missing values for all variables was below 10%, multiple imputation with 5 replications was performed to reduce potential bias and preserve data completeness.
Statistical analysis
Participants were grouped by CVD occurrence. Standardized mean difference (SMD) was used to assess the balance of baseline characteristics between groups. With reference to Cohen’s criteria, an |SMD|<0.1 was considered a negligible between-group difference, 0.1–0.2 a small difference, 0.2–0.5 a moderate difference and ≥ 0.5 a large difference. Categorical data are expressed as n (%), continuous data are shown as median (IQR) or mean ± SD. The χ2 test was used for categorical variables and independent sample t-tests or Mann-Whitney U tests were used for comparisons of continuous variables between groups.
Cox proportional hazards models were used to assess TCBI and TyG-BMI relations to CVD risk, with results expressed as hazard ratios (HRs) and corresponding 95% confidence intervals (CIs). LgTCBI and TyG-BMI were analyzed both as continuous variables and categorical variables based on quartiles: LgTCBI (Q1: 1.43–2.86; Q2: 2.86–3.05; Q3: 3.05–3.27; Q4: ≥3.27) and TyG-BMI (Q1: 102.42–174.87.42.87; Q2: 174.87–198.00; Q3: 198.00–226.64.00.64; Q4: ≥226.64). Model 1 was unadjusted, Model 2 was adjusted for age, gender, marital status, education, residence, Model 3 was further adjusted for smoking status, drinking status, CRP, UA, HDL, LDL and all health conditions. The proportional hazards assumption for the Cox models was assessed using Schoenfeld residual tests. Additionally, to assess potential multicollinearity among the exposure variables and key continuous covariates, we calculated variance inflation factors (VIFs) for all variables included in the fully adjusted Cox model (Model 3). A VIF value exceeding 10 was considered indicative of substantial multicollinearity. Additionally, we examined the correlation matrix between the core components of the indices (TG, TC, BW, BMI, FBG) using Pearson or Spearman correlation coefficients as appropriate.
After adjusting for the same covariates, RCS was conducted to analyze dose-response relationships. Subgroup analyses were performed across strata defined by age, gender, education, residence, hypertension, DM, smoking and BMI to evaluate potential effect modification, with all covariates from Model 3 (except the stratification variable itself) adjusted for to avoid over-adjustment. Interaction was tested using the likelihood ratio test. Furthermore, to compare the predictive performance of TCBI and TyG-BMI in assessing CVD risk, we conducted receiver operating characteristic (ROC) analysis. We plotted the ROC curves for two-year, four-year, seven-year and nine-year intervals to assess the predictive performance at different time points and applied the DeLong test to compare the area under the curve (AUC). Finally, a series of sensitivity analyses were conducted to evaluate the robustness of the findings.
The statistical analysis and plotting of this study were performed using SPSS 26.0, R4.4.1 and MedCalc software. Statistical significance was p < 0.05 (two-tailed).
Results
Baseline characteristics of participants
This study included a total of 8,104 participants, with 1,840 individuals diagnosed with CVD. Among them, 3,827 (47.22%) were male and 4,277 (57.78%) were female, with an average age of 59.12 ± 9.35 years. Table 1 presented the baseline characteristics of the participants based on the occurrence of CVD events after 9 years of follow-up. The SMD was less than 0.1 for a majority of variables (15/30, 50.00%), indicating a generally comparable distribution of baseline characteristics between the two groups. However, SMD values exceeding 0.2 were observed for hypertension (SMD = 0.293), BMI (SMD = 0.249) and depression (SMD = 0.201), suggesting the presence of moderate differences for these variables. Notably, TCBI (SMD = 0.207) and TyG-BMI (SMD = 0.266) also demonstrated moderate baseline differences, with magnitudes comparable to those of traditional risk factors. As all these variables are established CVD risk factors, their higher levels in the CVD group are consistent with the expected clinical profile. The results showed that CVD patients were older, had a higher proportion of females, and exhibited higher BMI, LDL, FBG, CRP, TG and TC levels, while HDL levels were lower. The proportion of current smokers and drinkers was also lower in CVD patients. Furthermore, CVD patients were significantly more likely to have depression, hypertension, DM, lung disease, arthritis, dyslipidemia, hepatic disease, digestive diseases and asthma.
Table 1.
The baseline characteristics of participants
| Variables | Overall | Non-CVD | CVD | SMD | P |
|---|---|---|---|---|---|
| Overall | 8104 (100.00) | 6264 (77.30) | 1840 (22.70) | ||
| Age (years) | 59.12 ± 9.35 | 58.74 ± 9.48 | 60.41 ± 8.77 | 0.184 | < 0.001 |
| Gender | 0.111 | < 0.001 | |||
| Male | 3827 (47.22) | 3037 (48.48) | 790 (42.93) | ||
| Female | 4277 (52.78) | 3227 (51.52) | 1050 (57.07) | ||
| Education | 0.064 | 0.055 | |||
| Illiterate | 2370 (29.24) | 1792 (28.61) | 578 (31.41) | ||
| Primary school and below | 3326 (41.04) | 2584 (41.25) | 742 (40.33) | ||
| Middle school and above | 2408 (29.71) | 1888 (30.14) | 520 (28.26) | ||
| Marital status | 0.022 | 0.438 | |||
| Single | 956 (11.80) | 729 (11.64) | 227 (12.34) | ||
| Married | 7148 (88.20) | 5535 (88.36) | 1613 (87.66) | ||
| Smoking status | 0.092 | 0.003 | |||
| Never | 4924 (60.76) | 3772 (60.22) | 1152 (62.61) | ||
| Former | 653 (8.06) | 484 (7.73) | 169 (9.18) | ||
| Current | 2527 (31.18) | 2008 (32.06) | 519 (28.21) | ||
| Residence | 0.017 | 0.555 | |||
| Rural | 7544 (93.09) | 5825 (92.99) | 1719 (93.42) | ||
| Urban | 560 (6.91) | 439 (7.01) | 121 (6.58) | ||
| Drinking status | 0.102 | < 0.001 | |||
| No | 5351 (66.03) | 4068 (64.94) | 1283 (69.73) | ||
| Yes | 2753 (33.97) | 2196 (35.06) | 557 (30.27) | ||
| Depression | 2942 (36.30) | 2135 (34.08) | 807 (43.86) | 0.201 | < 0.001 |
| Hypertension | 3106 (38.33) | 2197 (35.07) | 909 (49.40) | 0.293 | < 0.001 |
| DM | 574 (7.08) | 384 (6.13) | 190 (10.33) | 0.153 | < 0.001 |
| Cancer | 71 (0.88) | 57 (0.91) | 14 (0.76) | 0.016 | 0.645 |
| Lung diseases | 883 (10.90) | 607 (9.69) | 276 (15.00) | 0.162 | < 0.001 |
| Asthma | 251 (3.10) | 163 (2.60) | 88 (4.78) | 0.116 | < 0.001 |
| Dyslipidemia | 3354 (41.39) | 2492 (39.78) | 862 (46.85) | 0.143 | < 0.001 |
| Hepatic diseases | 289 (3.57) | 199 (3.18) | 90 (4.89) | 0.087 | 0.001 |
| Kidney diseases | 475 (5.86) | 352 (5.62) | 123 (6.68) | 0.044 | 0.098 |
| Digestive diseases | 1829 (22.57) | 1367 (21.82) | 462 (25.11) | 0.078 | 0.003 |
| Psychiatric disorders | 143 (1.76) | 113 (1.80) | 30 (1.63) | 0.013 | 0.692 |
| Memory related diseases | 67 (0.83) | 46 (0.73) | 21 (1.14) | 0.042 | 0.122 |
| Arthritis | 2772 (34.21) | 2077 (33.16) | 695 (37.77) | 0.097 | < 0.001 |
| BMI (kg/m2) | 23.32 ± 3.57 | 23.11 ± 3.50 | 24.01 ± 3.72 | 0.249 | < 0.001 |
| HDL (mg/dL) |
49.48 (40.98, 60.31) |
49.87 (40.98, 60.70) |
48.71 (40.21, 58.76) |
0.095 | < 0.001 |
| LDL (mg/dL) |
114.05 (93.17, 137.63) |
113.27 (92.40, 136.08) |
116.95 (95.88, 140.72) |
0.117 | < 0.001 |
| CRP (mg/L) |
1.01 (0.54, 2.12) |
0.96 (0.53, 2.04) |
1.19 (0.59, 2.32) |
0.007 | < 0.001 |
| UA (mg/dL) |
4.30 (3.57, 5.15) |
4.31 (3.59, 5.15) |
4.27 (3.54, 5.17) |
0.013 | 0.373 |
| FBG (mg/dL) |
102.42 (94.32, 113.04) |
102.06 (94.14, 112.50) |
103.32 (95.04, 115.02) |
0.112 | < 0.001 |
| TG (mg/dL) |
104.43 (74.34, 151.34) |
101.78 (72.57, 148.68) |
111.51 (80.54, 160.18) |
0.092 | < 0.001 |
| TC (mg/dL) |
190.59 (167.40, 215.72) |
189.43 (166.24, 214.56) |
194.07 (170.49, 219.98) |
0.120 | < 0.001 |
| LgTCBI | 3.08 ± 0.31 | 3.06 ± 0.31 | 3.13 ± 0.30 | 0.207 | < 0.001 |
| TyG-BMI | 202.87 ± 38.80 | 200.50 ± 37.92 | 210.97 ± 40.64 | 0.266 | < 0.001 |
Continuous variables are expressed as mean ± standard deviation or as median (Quartile1, Quartile 3), Categorical variables are presented as n (%)
SMD, standard mean difference; DM, diabetes mellitus; HDL, high-density lipoprotein cholesterol; LDL, low-density lipoprotein cholesterol; CRP, c-reactive protein; UA, uric acid; FBG, fasting blood glucose; TG, triglyceride; TC, total cholesterol; BMI, body mass index; TCBI, triglyceride-total cholesterol-body weight index; TyG-BMI, triglyceride glucose-body mass index
The LgTCBI and TyG-BMI indices followed a normal distribution in the study population and were significantly higher in CVD patients, with mean (SD) values of 3.13 ± 0.30 and 210.97 ± 40.64 respectively, compared to 3.06 ± 0.31 and 200.50 ± 37.92 in the non-CVD group. There were no statistically significant differences between the two groups in terms of education, residence, marital status, cancer, kidney disease, psychiatric disorders, memory-related diseases and UA (all p > 0.05).
The incidence rate of CVD
Table 2 showed that 1,840 participants had CVD. The overall CVD incidence rate was 274.72 per 10,000 person-years. The CVD incidence rates for participants in the LgTCBI quartiles were: Q1: 206.69/10,000 person-years, Q2: 250.60/10,000 person-years, Q3: 307.21/10,000 person-years, Q4: 338.78/10,000 person-years. The CVD incidence rates for participants in the TyG-BMI quartiles were: Q1: 200.41/10,000 person-years, Q2: 251.23/10,000 person-years, Q3: 282.45/10,000 person-years, Q4: 368.53/10,000 person-years. The overall CVD incidence rate was 22.70%. The incidence rates for each LgTCBI quartile were: Q1: 17.38%, Q2: 20.83%, Q3: 25.23%, Q4: 27.52% (Fig. 2A) and the incidence rates for each TyG-BMI quartile were: Q1: 16.83%, Q2: 20.98%, Q3: 23.30%, Q4: 29.71% (Fig. 2B). Participants with lower LgTCBI and TyG-BMI had markedly lower CVD incidence compared to those with higher LgTCBI and TyG-BMI, suggesting positive associations between higher levels of these indices and CVD risk.
Table 2.
Incidence rate of CVD (% or per 10000 person-year)
| Variables | Participants (n) | CVD events (n) | Incidence rate (95% CI)(%) |
Per 10,000 person-year |
|---|---|---|---|---|
| Total | 8104 | 1840 | 22.70 (21.79–23.61) | 274.72 |
| LgTCBI | ||||
| Q1 (1.43–2.86) | 2066 | 359 | 17.38 (15.75–19.01) | 206.69 |
| Q2 (2.86–3.05) | 2002 | 417 | 20.83 (19.05–22.61) | 250.60 |
| Q3 (3.05–3.27) | 2041 | 515 | 25.23 (23.35–27.11) | 307.21 |
| Q4 (≥ 3.27) | 1995 | 549 | 27.52 (25.56–29.48) | 338.78 |
| TyG-BMI | ||||
| Q1 (102.42–174.87.42.87) | 2026 | 341 | 16.83 (15.20–18.46.20.46) | 200.41 |
| Q2 (174.87–198.00) | 2026 | 425 | 20.98 (19.21–22.75) | 251.23 |
| Q3 (198.00–226.64.00.64) | 2026 | 472 | 23.30 (21.46–25.14) | 282.45 |
| Q4 (≥ 226.64) | 2026 | 602 | 29.71 (27.72–31.70) | 368.53 |
TCBI, triglyceride-total cholesterol-body weight index; TyG-BMI, triglyceride glucose-body mass index; CI, confidence; Q, quartile
Fig. 2.
The incidence of CVD across different quartiles of LgTCBI and TyG-BMI. (A) Incidence of CVD across quartiles (Q1-Q4) of the LgTCBI. (B) Incidence of CVD across quartiles (Q1-Q4) of the TyG-BMI. In both panels, Q1 represents the group with the lowest values of the respective index and Q4 represents the group with the highest values
Cox regression models for assessing the relationship between TCBI, TyG-BMI and CVD
Before reporting the Cox regression estimates, we assessed the proportional hazards assumption using Schoenfeld residual tests. For TCBI and TyG-BMI, analyzed as both continuous and categorical variables, no significant violations were detected (p > 0.05) (Supplementary Tables 2–3). However, the global test for the fully adjusted models indicated minor violations attributable to covariates (marital status, residence, and arthritis). Consequently, we performed sensitivity analyses treating these covariates as stratification factors, which yielded hazard ratios consistent with those from the primary models, confirming the robustness of our findings (Supplementary Table 4).
Moreover, to assess potential multicollinearity in the fully adjusted model, we calculated VIFs. All VIFs were below 3 (Supplementary Table 5), well under the conventional threshold of 10, indicating that multicollinearity did not substantially affect the stability of the estimates. As expected, moderate to strong correlations were observed among some biological components that constitute the composite indices. Notably, the highest correlations were found between variables that are integral to the same index: TG and LgTCBI (r = 0.92), BMI and TyG-BMI (r = 0.92) and BMI and weight (r = 0.80) (Supplementary Fig. 1). These correlations are inherent to the physiological relationships and the mathematical construction of the indices. In conclusion, these expected intercorrelations at the component level did not translate into problematic multicollinearity within the regression model, thereby not compromising the stability or interpretability of the hazard ratio estimates for TCBI and TyG-BMI.
Cox regression analysis showed that each 10-unit increase in the TyG-BMI index was associated with a 6% higher CVD risk (Model 1: HR 1.06, 95% CI 1.05–1.07, p < 0.001). This positive association remained significant after multivariable adjustment in Model 2 (HR 1.07, 95% CI 1.05–1.08, p < 0.001) and Model 3 (HR 1.06, 95% CI 1.04–1.07, p < 0.001). Similarly, each 1-unit increase in the LgTCBI index corresponded to a 73% elevation in CVD risk in the unadjusted model 1 (Model 1: HR 1.73, 95% CI 1.50–1.99, p < 0.001), with consistent associations observed in Model 2 (HR 1.83, 95% CI 1.59–2.12, p < 0.001) and Model 3(HR 1.59, 95% CI 1.31–1.92, p < 0.001). And the quartiles of TyG-BMI and LgTCBI index (Q2, Q3 and Q4) all showed gradual increase in CVD risk across all three models (Table 3).
Table 3.
Cox regression models for the association between the TyG-BMI, TCBI and CVD risk
| Categories | Model 1 | Model 2 | Model 3 | |||
|---|---|---|---|---|---|---|
| HR (95% CI) |
P | HR (95% CI) |
P | HR (95%CI) | P | |
|
LgTCBI (per 1 unit) |
1.73 (1.50–1.99) | < 0.001 | 1.83 (1.59–2.12) | < 0.001 | 1.59 (1.31–1.92) | < 0.001 |
| LgTCBI quartile | ||||||
| Q1 | Reference | Reference | Reference | |||
| Q2 | 1.22 (1.06–1.41) | 0.005 | 1.22 (1.06–1.41) | 0.006 | 1.20 (1.04–1.39) | 0.013 |
| Q3 | 1.51 (1.32–1.73) | < 0.001 | 1.52 (1.33–1.74) | < 0.001 | 1.42 (1.22–1.65) | < 0.001 |
| Q4 | 1.68 (1.47–1.91) | < 0.001 | 1.74 (1.52–1.99) | < 0.001 | 1.55 (1.32–1.84) | < 0.001 |
|
TyG-BMI (per 10 units) |
1.06 (1.05–1.07) | < 0.001 | 1.07 (1.05–1.08) | < 0.001 | 1.06 (1.04–1.07) | < 0.001 |
| TyG-BMI quartile | ||||||
| Q1 | Reference | Reference | Reference | |||
| Q2 | 1.26 (1.10–1.46) | 0.001 | 1.32 (1.15–1.53) | < 0.001 | 1.30 (1.13–1.51) | < 0.001 |
| Q3 | 1.43 (1.24–1.64) | < 0.001 | 1.51 (1.31–1.74) | < 0.001 | 1.45 (1.24–1.68) | < 0.001 |
| Q4 | 1.89 (1.66–2.16) | < 0.001 | 2.04 (1.78–2.34) | < 0.001 | 1.85 (1.57–2.18) | < 0.001 |
HR, Hazard Ratio; CI, Confidence Interval; TCBI, triglyceride-total cholesterol-body weight index; TyG-BMI, triglyceride glucose-body mass index; Q, quartile
Model 1: Not adjusted; Model 2: Adjusted for age, gender, education, marital status and residence; Model 3: Adjusted for Model2, smoking status, drinking status, hypertension, diabetes mellitus, dyslipidemia, depression, Cancer, Lung disease, asthma, hepatic disease, kidney disease, digestive disease, psychiatric disorders, memory-related disease, arthritis, UA, CRP, HDL and LDL
Cumulative risk of CVD: analysis using Kaplan–Meier survival curves
Figure 3A showed that the cumulative CVD risk increased with the LgTCBI quartiles, and this trend was statistically significant (Log-rank test: Chi-square = 69.3532, p < 0.001). Similarly, Fig. 3B showed that the cumulative CVD risk also increased with the TyG-BMI quartiles, which was also statistically significant (Log-rank test: Chi-square = 98.8722, p < 0.001).
Fig. 3.
Kaplan-Meier survival curves for cumulative CVD risk by LgTCBI (A) and TyG-BMI (B). Cumulative incidence (1 - survival probability) is shown on the y-axis, and years of follow-up are shown on the x-axis. Each panel displays four curves, representing the Q1 (lowest), Q2, Q3, and Q4 (highest) quartile groups of the respective index. The log-rank test was used to assess the overall difference in survival distributions across the four groups
RCS analysis of the relationships between TCBI, TyG-BMI and CVD risk
Figure 4 showed two fully adjusted RCS models. Figure 4A indicated a positive nonlinear relationship between the LgTCBI index and CVD risk (p for overall < 0.001; p for nonlinear = 0.009). Figure 4B showed a positive linear relationship between the TyG-BMI index and CVD risk (p for overall < 0.001; p for nonlinear = 0.295).
Fig. 4.
RCS models for the associations between the LgTCBI (A), TyG-BMI (B) and CVD risk. All models were adjusted for the full set of covariates in Model 3 (adjusted for age, gender, marital status, education, residence, smoking status, drinking status, CRP, UA, HDL, LDL and all health conditions). P values for overall association and nonlinearity are displayed within each panel
Threshold-effect analysis of the relationship between the TCBI and CVD risk
Threshold analysis revealed a significant threshold effect in the association between LgTCBI and CVD (p for likelihood ratio test = 0.016). The inflection point was identified at LgTCBI = 3.28. When LgTCBI was below this threshold, it showed a positive association with CVD [HR 2.07, 95% CI 1.48–2.88, p < 0.01]. In contrast, no significant association was observed when LgTCBI exceeded 3.28 [HR 0.87, 95% CI 0.53–1.43, p = 0.591] (Table 4). These findings suggested a nonlinear, saturating association between TCBI and CVD risk. To examine whether the association was driven by extreme values, a sensitivity analysis was performed: after Winsorizing the top and bottom 1% of TCBI values, the nonlinear p-value from the refitted RCS model remained significant (p = 0.018) and the shape of the curve was consistent with the original analysis (Supplementary Fig. 2).
Table 4.
Threshold-effect analysis of the relationship between TCBI and CVD risk
| Models | effect | P |
|---|---|---|
| Model 1 Fitting model by standard linear regression | 1.55 (1.28–1.88) | < 0.001 |
| Model 2 Fitting model by two-piecewise linear regression | ||
| Inflection point | 3.28 | |
| <3.28 | 2.07 (1.48–2.88) | < 0.001 |
| ≥3.28 | 0.87 (0.53–1.43) | 0.591 |
| P for likelihood test | 0.016 |
Comparison of predictive efficacy of TCBI and TyG-BMI for CVD risk at various time points
Figure 5A compared the predictive ability of the TCBI and TyG-BMI indices for CVD risk within two years, showing no significant difference: the AUC for the LgTCBI index was 0.559 and for the TyG-BMI index was 0.571 (p = 0.395, Delong’s test). Figure 5B examined the results over four years, finding similar results, with AUC values of 0.556 for the LgTCBI index and 0.562 for the TyG-BMI index (p = 0.465, Delong’s test). Figure 5C assessed the seven-years follow-up period, showing a significant difference, with AUC values of 0.558 for the LgTCBI index and 0.574 for the TyG-BMI index (p = 0.046, Delong’s test). Finally, Fig. 5D evaluated the nine-years follow-up period and also showed a significant difference, with the AUC values of 0.565 for the LgTCBI index and 0.576 for the TyG-BMI index (p = 0.045, Delong’s test). The results indicated that TCBI and TyG-BMI exhibited comparable predictive ability in short-term (2–4 years) prediction, whereas TyG-BMI performed slightly better in long-term (7–9 years) prediction. However, both indices demonstrated limited independent predictive ability, with AUC values < 0.6.
Fig. 5.
Receiver operating characteristic curves compared the predictive efficacy of the TCBI, TyG-BMI for CVD risk events at different time points. (A) 2-year follow-up. (B) 4-years follow-up. (C) 7-year follow-up. (D) 9-year follow-up. For each curve, the area under the curve (AUC) values are provided in the corresponding panel. The statistical comparison between the two AUCs at each time point was performed using DeLong’s test
Subgroup analysis of TCBI and TyG-BMI indices in predicting CVD risk
Figure 6 illustrated the subgroup analysis of TCBI in predicting CVD risk. The results showed that the indicator and CVD risk was not influenced by age, gender, residence, education, hypertension, diabetes, smoking status or BMI. In other words, the interactions between these variables and TCBI were not statistically significant (p for interaction > 0.05). Similarly, Fig. 7 presented the subgroup analysis for TyG-BMI, yielding similar results. It was found that TyG-BMI and CVD risk were also not significantly influenced by factors such as age, gender, residence, education, hypertension, diabetes, smoking status or BMI. (p > 0.05 for interaction).
Fig. 6.
Subgroup analysis of the association between TCBI and incident CVD risk. All subgroup-specific HRs were derived from Cox regression models that included the full set of covariates from Model 3 (adjusted for age, gender, marital status, education, residence, smoking status, drinking status, CRP, UA, HDL, LDL and all health conditions), except for the stratification variable itself (to avoid over-adjustment)
Fig. 7.
Subgroup analysis of the association between TyG-BMI and incident CVD risk. All subgroup-specific HRs were derived from Cox regression models that included the full set of covariates from Model 3 (adjusted for age, gender, marital status, education, residence, smoking status, drinking status, CRP, UA, HDL, LDL and all health conditions), except for the stratification variable itself (to avoid over-adjustment)
Sensitivity analysis
To assess the robustness of the findings, a series of sensitivity analyses were performed. Firstly, after excluding individuals with missing data, the associations of TCBI and TyG-BMI quartiles with CVD risk remained significant (all p < 0.05) (Table 5). Secondly, to address concerns regarding the single baseline measurement of metabolic indices and potential reverse causality, we excluded participants who developed CVD within the first two years of follow-up. The associations for both TCBI and TyG-BMI with CVD remained significant (all p < 0.05) (Supplementary Table 6), arguing against this bias and supporting the robustness of the results. Furthermore, to evaluate whether the observed associations might be driven by misclassification or confounding from pre-existing multimorbidity, participants with three or more comorbidities at baseline were excluded. The associations between TCBI, TyG-BMI and CVD risk remained statistically significant (Supplementary Table 7), supporting the robustness of the primary findings against this source of bias. Overall, these sensitivity analyses confirmed robust associations of TCBI and TyG-BMI with CVD risk.
Table 5.
Sensitivity analysis: associations between TCBI, TyG-BMI and CVD after exclusion of individuals with missing data
| Categories | Model 1 | Model 2 | Model 3 | |||
|---|---|---|---|---|---|---|
| HR (95% CI) |
P | HR (95% CI) |
P | HR (95%CI) | P | |
|
LgTCBI (per 1 unit) |
1.83 (1.58–2.12) | < 0.001 | 1.94 (1.66–2.25) | < 0.001 | 1.63 (1.31–2.02) | < 0.001 |
| LgTCBI quartile | ||||||
| Q1 | Reference | Reference | Reference | |||
| Q2 |
1.25 (1.08–1.44) |
0.003 | 1.25 (1.08–1.44) | 0.006 |
1.21 (1.05–1.40) |
0.010 |
| Q3 | 1.50 (1.30–1.73) | < 0.001 | 1.51 (1.31–1.74) | < 0.001 | 1.38 (1.17–1.62) | < 0.001 |
| Q4 | 1.73 (1.50–1.99) | < 0.001 |
1.79 (1.55–2.06) |
< 0.001 | 1.54 (1.27–1.85) | < 0.001 |
| TyG-BMI | 1.28 (1.20–1.36) | < 0.001 |
1.28 (1.20–1.37) |
< 0.001 |
1.17 (1.06–1.30) |
0.002 |
| TyG-BMI quartile | ||||||
| Q1 | Reference | Reference | Reference | |||
| Q2 |
1.38 (1.20–1.59) |
< 0.001 | 1.38 (1.19–1.59) | < 0.001 |
1.29 (1.11–1.49) |
< 0.001 |
| Q3 |
1.63 (1.42–1.87) |
< 0.001 | 1.61 (1.40–1.86) | < 0.001 | 1.42 (1.20–1.67) | < 0.001 |
| Q4 | 1.71 (1.48–1.98) | < 0.001 |
1.71 (1.47–1.99) |
< 0.001 |
1.38 (1.13–1.68) |
< 0.001 |
HR, Hazard Ratio; CI, Confidence Interval; TCBI, triglyceride-total cholesterol-body weight index; TyG-BMI, triglyceride glucose-body mass index; Q, quartile
Model 1: Not adjusted; Model 2: Adjusted for strata (age), gender, education, strata (marital status) and strata (residence); Model 3: Adjusted for Model2, smoking status, drinking status, hypertension, diabetes mellitus, dyslipidemia, depression, Cancer, Lung disease, asthma, hepatic disease, kidney disease, digestive disease, psychiatric disorders, memory-related disease, strata (arthritis), UA, CRP, HDL and LDL
Discussion
In this study, we directly compared the associations of a novel nutritional index-TCBI, and an established insulin resistance index-TyG-BMI, with CVD risk. Our findings indicated that both indices were significantly associated with an increased CVD risk among middle-aged and older adults, and these associations remained robust after multivariable adjustment. Both indices demonstrated limited predictive ability for CVD risk (AUC < 0.60), but TCBI showed performance comparable to TyG-BMI in short-term prediction (2–4 years), whereas TyG-BMI exhibited a slight but consistent advantage in long-term prediction (7–9 years). These results suggest that the primary value of these indices may lie not in precise individual-level prediction, but rather as potential auxiliary tools for identifying high-risk individuals who require further clinical assessment.
Firstly, this study found that both traditional CVD risk factors and the composite indicators TCBI and TyG-BMI exhibited moderate baseline elevations in CVD. Such consistency supports the plausibility of TCBI and TyG-BMI as biomarkers for CVD risk assessment. The observed SMDs for these indicators (0.207 and 0.266, respectively) were comparable to those of established risk factors such as hypertension (0.293) and BMI (0.249), suggesting that they may capture similar dimensions of CVD pathophysiology.
Furthermore, the robust linear association observed between TyG-BMI and CVD risk in our study confirms and extends prior evidence. As a reliable surrogate for IR, the significant positive correlation of TyG-BMI with CVD risk is consistent with previous reports [10]. This index integrates fasting glucose (a direct marker of glycemic homeostasis), lipids and BMI (a more stable measure of adiposity over time than BW), collectively reflecting a sustained state of metabolic dysregulation, particularly the core pathophysiological pathway of IR in CVD pathogenesis [28, 29]. IR can promote cause inflammatory cell infiltration in adipose tissue, which secretes large amounts of pro-inflammatory cytokines and these cytokines can inhibit insulin signaling, leading to systemic low-grade inflammation [30, 31]. Both IR and inflammation can trigger endothelial dysfunction, damaging arterial integrity, accelerating atherosclerosis progression and increasing CVD risk [32, 33]. Additionally, IR also can lead to increased oxidative stress, causing excessive production of reactive oxygen species, which further damages vascular endothelial cells and exacerbates cardiovascular pathology [34]. The above mechanism may be the main reason why TyG BMI performs better in long-term prediction. And multiple studies supported the value of TyG-BMI in CVD risk, further affirming its reliability as a metabolic marker [6, 10, 35].
In contrast, researches on TCBI in relation to CVD risk remain limited and inconsistent. Some studies have reported that lower TCBI is associated with poor prognosis in patients with acute ischemic stroke, acute decompensated heart failure, or coronary artery disease [16, 18, 36], while another prospective cohort linked higher TCBI to increased CVD mortality [1]. Most existing studies have focused on the prognostic value of TCBI, whereas our study examined its association with CVD risk in a community-based middle-aged and older Chinese population, highlighting its potential role in primary risk assessment.
TCBI was designed to overcome the clinical utility limitations of existing nutritional indices. Unlike tools such as the GNRI or CONUT score, which require multiple complex parameters, TCBI incorporates only three readily available clinical measures: triglycerides, total cholesterol, and body weight, thereby enhancing its practicality in daily CVD assessments [15]. And its multiplicative formula may reflect a synergistic amplification between dyslipidemia and overall metabolic load. Elevated TC and TG are established atherogenic risk factors and high BW indicates increased metabolic burden [11, 37, 38]. In obesity, dysfunctional adipose tissue can exacerbate systemic inflammation and lipid disorders, potentially amplifying the vascular injury caused by high TG and TC [31, 39, 40]. Notably, low TC and TG have been linked to poor outcomes in some nutritional assessment frameworks-analogous to the “obesity paradox” concerning BW and mortality [15]. But TCBI aims to capture the synergistic effect of high-level metabolic risk rather than the linear contribution of any single component.
We further identified a nonlinear, saturating relationship between TCBI and CVD risk, with a sharp increase at lower levels followed by a plateau. This suggests a possible threshold effect beyond which risk acceleration diminishes, potentially due to metabolic compensation or measurement limits of the index. Furthermore, by using BW instead of BMI, TCBI concurrently conveys information on both metabolic risk and nutritional status. In older adults, changes in BW, particularly unintentional loss, serve as a strong, accessible indicator of nutritional status and overall prognosis [41, 42]. BW can more directly and sensitively reflect recent energy balance or muscle loss (such as sarcopenia) than BMI, which is crucial in geriatric assessment [42]. Thus, TCBI may indicate not only atherogenic risk but also nutrition-related changes in body composition, offering distinct value in clinical settings where both metabolic and nutritional evaluations are relevant. Its stronger performance in short-term prediction may stem from sensitivity to acute “metabolic-nutritional” imbalance, making it particularly useful for identifying individuals at near-term high risk.
In summary, the differential performance of the two indices reflects their ability to capture distinct aspects and phases of metabolic risk. TyG-BMI is more aligned with sustained metabolic dysregulation centered on IR, whereas TCBI appears more sensitive to immediate risk driven by the synergy between lipid abnormalities and BW load. Given their limited standalone predictive accuracy, neither index is suitable as a first-line screening tool. However, as simple, low-cost auxiliary measures, TyG-BMI may help inform long-term risk, while TCBI could offer quick reference in short-term assessment scenarios that require integrated lipids and BW information, especially when nutritional status is a concern. Their primary utility likely lies in identifying individuals who require a more comprehensive risk assessment.
Strengths and limitations
This study presents several notable advantages. Firstly, to our knowledge, this is the first study to compare the novel TCBI index and the widely recognized TyG-BMI index with CVD risk, which may offer a fresh evaluation perspective on the utility of these indices. Secondly, the study is based on a broad and representative cohort from the general population in China, with long-term monitoring, which is crucial for accurately examining longitudinal associations. Moreover, we developed comprehensive evaluation models using various statistical methods, which allowed for a more robust and multifaceted analysis. Finally, by extending the application scope of the TCBI index, our study could provide a potential reference for clinicians in CVD risk assessment and contributes to the evidence base that might inform future public health strategies.
However, this study also has several limitations. Firstly, the sample consisted of Chinese middle‑aged and older adults, whether the findings generalize to younger or other ethnic populations requires further investigation. Secondly, The diagnosis of CVD was based on self-reported questionnaires without objective evidence such as ECG, imaging or medical record linkage, which may introduce non-differential misclassification and thus attenuate the hazard ratios. Although sensitivity analyses suggested the robustness of the main findings, future validation using clinically confirmed endpoints remains necessary. Thirdly, although we adjusted for multiple confounders, factors such as income, medication use, dietary habits, exercise and genetic predispositions were not included, which may affect the observed associations. Incorporating these factors into future research will help to gain a more comprehensive understanding of the relationship between TCBI, TyG-BMI and CVD. Additionally, TCBI and TyG-BMI were measured only at baseline. Since metabolic indicators can change over time, a single measurement may not fully reflect long-term exposure levels, potentially introducing non-differential misclassification and leading to an underestimation of their true association with CVD risk. Notably, sensitivity analyses excluding events occurring within the first two years of follow-up showed that the associations remained significant, suggesting that reverse causality was unlikely to be the main driver and supporting the robustness of the primary findings. Future studies with more frequent measurements would be valuable for elucidating the impact of metabolic changes over time on CVD risk. Finally, as an observational study, it can only demonstrate correlation but cannot prove causality between the indices and CVD risk. And stronger causal inference designs are needed in future research.
Conclusion
TCBI and TyG-BMI were significantly associated with an increased CVD risk. Both indices demonstrated limited discriminative ability as standalone predictive tools, but TCBI performed comparably to TyG-BMI in short-term risk assessment, TyG-BMI held a slight advantage in long-term prediction. These findings suggest that TCBI and TyG-BMI may serve as simple, complementary screening tools to help identify high-risk individuals warranting further comprehensive clinical evaluation, rather than as standalone primary screening instruments and their clinical applicability requires further validation in diverse cohorts.
Supplementary Information
Acknowledgements
The authors would like to extend our heartfelt gratitude to the researchers and staff of the China Health and Retirement Longitudinal Study (CHARLS), as well as the participants of the study.
Abbreviations
- CHARLS
China health and retirement longitudinal study
- CVD
Cardiovascular disease
- IR
Insulin resistance
- TyG-BMI
Triglyceride glucose-body mass index
- TCBI
Triglyceride, total cholesterol and body mass index
- DM
Diabetes mellitus
- TG
Triglyceride
- TC
Total cholesterol
- HDL
High-density lipoprotein
- LDL
Low-density lipoprotein
- UA
Uric acid
- RCS
Restricted cubic spline
- ROC
Receiver operating characteristic
- OR
Odds ratios
- CI
Confidence intervals
- CRP
C-reactive protein
Author contributions
T.W. and X.Y. contributed to the conception and design of the study. Y. W. and D.M. was responsible for statistics analysis, preparation of charts, results interpretation and manuscript writing. W.H and Y.Z. contributed to formal analysis and data curation. Z.L. and H.M. contributed to data reconciliation and results visualization. N.J. contributed to data collection and management. X.Y. performed article review and revision. All authors reviewed the manuscript.
Funding
This work was supported by the Shandong Provincial Natural Science Foundation (ZR2024MG047), NHC Key Laboratory of Cardiopulmonary Rehabilitation and Functional Recovery (University of Health and Rehabilitation Sciences), Shandong Health Science and Technology Innovation Team Construction Project-Cardiopulmonary Comorbidity, and Qingdao Key Laboratory of Respiratory Comorbidity Remodeling and Precision Prevention.
Data availability
The datasets analyzed for this study can be obtained from the CHARLS database (http://charls.pku.edu.cn/).
Declarations
Ethics approval and consent to participate
The CHARLS was approved by the Biomedical Ethics Review Committee of Peking University (IRB00001052-11015), and informed consent was obtained from all participants.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Yaoyao Wang and Degang Mo contributed equally to this work.
Contributor Information
Tao Wang, Email: wangtao@uor.edu.cn.
Xinjuan Yu, Email: yxj4501@126.com.
References
- 1.Rezaee M, Kamrani F, Imannezhad M, Shahri HH, Saihood WK, Rezvani A, et al. Beyond traditional metrics: evaluating the triglyceride-total cholesterol-body weight index (TCBI) in cardiovascular risk assessment. BMC Cardiovasc Disord. 2025;25(1):39. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Qiao YN, Zou YL, Guo SD. Low-density lipoprotein particles in atherosclerosis. Front Physiol. 2022;13:931931. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Roth GA, Mensah GA, Johnson CO, Addolorato G, Ammirati E, Baddour LM, et al. Global burden of cardiovascular diseases and risk Factors, 1990–2019: update from the GBD 2019 study. J Am Coll Cardiol. 2020;76(25):2982–3021. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Wang W, Liu Y, Ye P, Liu J, Yin P, Qi J, et al. Trends and associated factors in place of death among individuals with cardiovascular disease in China, 2008–2020: a population-based study. Lancet Reg Health Am. 2022;21:100383. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Sacramento-Pacheco J, Duarte-Climents G, Gomez-Salgado J, Romero-Martin M, Sanchez-Gomez MB. Cardiovascular risk assessment tools: a scoping review. Aust Crit Care. 2019;32(6):540–59. [DOI] [PubMed] [Google Scholar]
- 6.Li F, Wang Y, Shi B, Sun S, Wang S, Pang S, et al. Association between the cumulative average triglyceride glucose-body mass index and cardiovascular disease incidence among the middle-aged and older population: a prospective nationwide cohort study in China. Cardiovasc Diabetol. 2024;23(1):16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Fernandez-Macias JC, Ochoa-Martinez AC, Varela-Silva JA, Perez-Maldonado IN. Atherogenic index of plasma: novel predictive biomarker for cardiovascular illnesses. Arch Med Res. 2019;50(5):285–94. [DOI] [PubMed] [Google Scholar]
- 8.Wang R, Cheng X, Tao W. Association between triglyceride glucose body mass index and cardiovascular disease in adults: evidence from NHANES 2011–2020. Front Endocrinol (Lausanne). 2024;15:1362667. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Zhuang Y, Qiu L, Han D, Qiao Z, Wang F, Jiang Q, et al. The association between triglyceride-glucose index and related parameters and risk of cardiovascular disease in American adults under different glucose metabolic states. Diabetol Metab Syndr. 2024;16(1):102. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Rao X, Xin Z, Yu Q, Feng L, Shi Y, Tang T, et al. Triglyceride-glucose-body mass index and the incidence of cardiovascular diseases: a meta-analysis of cohort studies. Cardiovasc Diabetol. 2025;24(1):34. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Jung E, Kong SY, Ro YS, Ryu HH, Shin SD. Serum cholesterol levels and risk of cardiovascular death: a systematic review and a dose-response meta-analysis of prospective cohort studies. Int J Environ Res Public Health. 2022;19(14):8272. 10.3390/ijerph19148272. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Yang Y, Li JX, Chen JC, Cao J, Lu XF, Chen SF, et al. Effect of elevated total cholesterol level and hypertension on the risk of fatal cardiovascular disease: a cohort study of Chinese steelworkers. Chin Med J (Engl). 2011;124(22):3702–6. [PubMed] [Google Scholar]
- 13.Hedayatnia M, Asadi Z, Zare-Feyzabadi R, Yaghooti-Khorasani M, Ghazizadeh H, Ghaffarian-Zirak R, et al. Dyslipidemia and cardiovascular disease risk among the MASHAD study population. Lipids Health Dis. 2020;19(1):42. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Ding H, Li C, Zhao X. The relationship between intrinsic capacity and sarcopenia in middle-aged and older Chinese populations: the mediating influence of a novel nutritional index. Front Public Health. 2025;13:1605158. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Doi S, Iwata H, Wada H, Funamizu T, Shitara J, Endo H, et al. A novel and simply calculated nutritional index serves as a useful prognostic indicator in patients with coronary artery disease. Int J Cardiol. 2018;262:92–8. [DOI] [PubMed] [Google Scholar]
- 16.Ishiwata S, Yatsu S, Kasai T, Sato A, Matsumoto H, Shitara J, et al. Prognostic effect of a novel simply calculated nutritional index in acute decompensated heartfailure. Nutrients. 2020;12(11):3311. 10.3390/nu12113311. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Wang C, Qing Y, Chen W, Li G. A novel nutritional index as a predictor of mortality in dilated cardiomyopathy: a retrospective study. PeerJ. 2022;10:e12704. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Maruyama S, Ebisawa S, Miura T, Yui H, Kashiwagi D, Nagae A, et al. Impact of nutritional index on long-term outcomes of elderly patients with coronary artery disease: sub-analysis of the SHINANO 5 year registry. Heart Vessels. 2021;36(1):7–13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Fan X, He Q, Zhang K, Lan X, Li Y, Zhang H. Comparison of the value of four objective nutritional indices in assessing the long-term prognosis of elderly patients with heart failure with preserved ejection fraction. Rev Cardiovasc Med. 2024;25(6):201. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.He Q, Fan X, Li B, He Q, Cao Y, Zhang H. Prognostic value of four objective nutritional indices in predicting long-term prognosis in elderly patients with atrial fibrillation: a retrospective cohort study. Clin Interv Aging. 2024;19:2043–56. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Zhao Y, Hu Y, Smith JP, Strauss J, Yang G. Cohort profile: the China health and retirement longitudinal study (CHARLS). Int J Epidemiol. 2014;43(1):61–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Huo RR, Liao Q, Zhai L, You XM, Zuo YL. Interacting and joint effects of triglyceride-glucose index (TyG) and body mass index on stroke risk and the mediating role of TyG in middle-aged and older Chinese adults: a nationwide prospective cohort study. Cardiovasc Diabetol. 2024;23(1):30. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.von Elm E, Altman DG, Egger M, Pocock SJ, Gotzsche PC, Vandenbroucke JP, et al. The strengthening the reporting of observational studies in epidemiology (STROBE) statement: guidelines for reporting observational studies. Epidemiology. 2007;18(6):800–4. [DOI] [PubMed] [Google Scholar]
- 24.Mo D, Zhang P, Zhang M, Dai H, Guan J. Cholesterol, high-density lipoprotein, and glucose index versus triglyceride-glucose index in predicting cardiovascular disease risk: a cohort study. Cardiovasc Diabetol. 2025;24(1):116. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Han Y, Zhou Z, Zhang Y, Zhao G, Xu B. The association of surrogates of insulin resistance with hyperuricemia among middle-aged and older individuals: a population-based nationwide cohort study. Nutrients. 2023;15(14):3139. 10.3390/nu15143139. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Zhou P, Wang S, Yan Y, Lu Q, Pei J, Guo W, et al. Association between chronic diseases and depression in the middle-aged and older adult Chinese population-a seven-year follow-up study based on CHARLS. Front Public Health. 2023;11:1176669. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Liu G, Zhang J. Association of a novel nutritional index with cognitive impairment in middle-aged and elderly Chinese adults: a cross-sectional analysis from the China Health and Retirement Longitudinal Study. Front Nutr. 2025;12:1486917. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Lee J, Kim B, Kim W, Ahn C, Choi HY, Kim JG, et al. Lipid indices as simple and clinically useful surrogate markers for insulin resistance in the U.S. population. Sci Rep. 2021;11(1):2366. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Moshkovits Y, Rott D, Chetrit A, Dankner R. The association between insulin sensitivity indices, ECG findings and mortality: a 40-year cohort study. Cardiovasc Diabetol. 2021;20(1):97. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Puschel GP, Klauder J, Henkel J. Macrophages, low-grade inflammation, insulin resistance and hyperinsulinemia: a mutual ambiguous relationship in the development of metabolic disease. J Clin Med. 2022;11(15):4358. 10.3390/jcm11154358. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Szukiewicz D. Molecular mechanisms for the vicious cycle between insulin resistance and the inflammatory response in obesity. Int J Mol Sci. 2023;24(12):9818. 10.3390/ijms24129818. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Jin A, Wang S, Li J, Wang M, Lin J, Li H, et al. Mediation of systemic inflammation on insulin resistance and prognosis of nondiabetic patients with ischemic stroke. Stroke. 2023;54(3):759–69. [DOI] [PubMed] [Google Scholar]
- 33.Yang T, Li G, Wang C, Xu G, Li Q, Yang Y, et al. Insulin resistance and coronary inflammation in patients with coronary artery disease: a cross-sectional study. Cardiovasc Diabetol. 2024;23(1):79. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Kosmas CE, Bousvarou MD, Kostara CE, Papakonstantinou EJ, Salamou E, Guzman E. Insulin resistance and cardiovascular disease. J Int Med Res. 2023;51(3):3000605231164548. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Li W, Shen C, Kong W, Zhou X, Fan H, Zhang Y, et al. Association between the triglyceride glucose-body mass index and future cardiovascular disease risk in a population with Cardiovascular-Kidney-Metabolic syndrome stage 0–3: a nationwide prospective cohort study. Cardiovasc Diabetol. 2024;23(1):292. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Zhang G, Pan Y, Zhang R, Wang M, Meng X, Li Z, et al. A novel nutritional index and adverse outcomes in ischemic stroke: results from the third China National Stroke Registry. Nutr Metab Cardiovasc Dis. 2022;32(6):1477–84. [DOI] [PubMed] [Google Scholar]
- 37.Wazir M, Olanrewaju OA, Yahya M, Kumari J, Kumar N, Singh J, et al. Lipid disorders and cardiovascular risk: a comprehensive analysis of current perspectives. Cureus. 2023;15(12):e51395. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Hasan B, Nayfeh T, Alzuabi M, Wang Z, Kuchkuntla AR, Prokop LJ, et al. Weight loss and serum lipids in overweight and obese adults: a systematic review and meta-analysis. J Clin Endocrinol Metab. 2020;105(12):dgaa673. 10.1210/clinem/dgaa673. [DOI] [PubMed] [Google Scholar]
- 39.Ahmed B, Sultana R, Greene MW. Adipose tissue and insulin resistance in obese. Biomed Pharmacother. 2021;137:111315. [DOI] [PubMed] [Google Scholar]
- 40.Lumeng CN, Saltiel AR. Inflammatory links between obesity and metabolic disease. J Clin Invest. 2011;121(6):2111–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Hussain SM, Newman AB, Beilin LJ, Tonkin AM, Woods RL, Neumann JT, et al. Associations of change in body size with all-cause and cause-specific mortality among healthy older adults. JAMA Netw Open. 2023;6(4):e237482. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Wang J, Zhuang Q, Tan S, Xu J, Zhang Y, Yan M, et al. Loss of body weight and skeletal muscle negatively affect postoperative outcomes after major abdominal surgery in geriatric patients with cancer. Nutrition. 2023;106:111907. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The datasets analyzed for this study can be obtained from the CHARLS database (http://charls.pku.edu.cn/).