Abstract
Background and aims
The triglyceride-glucose (TyG) index has been recognized as a surrogate marker for insulin resistance (IR) and an independent risk factor for cardiovascular disease (CVD). However, the combined effect of the TyG index and visceral obesity on CVD incidence remains unclear. We aimed to investigate the interaction, joint association, and potential mediators between the TyG index and comprehensive anthropometric indices with CVD risk in middle-aged and older adults.
Methods
We analyzed 7046 participants aged ≥ 45 years without baseline CVD from the China Health and Retirement Longitudinal Study (CHARLS) over a 9-year follow-up period. Retrospective collection included sociodemographic details, health status, physical examination results, and blood biomarkers. Adjusted Cox proportional hazards models were used to examine the interaction between TyG levels and anthropometric indices and their joint associations with CVD incidence. Subgroup analyses were conducted to evaluate the associations across different populations, and mediation analysis was performed to identify potential mediating pathways. The predictive value was determined using the area under the curve (AUC) of receiver operating characteristic curves. In addition, we validated the findings in the Multi-Ethnic Study of Atherosclerosis (MESA) cohort.
Results
In the CHARLS study, 1768 (25.1%) participants developed CVD. All TyG–anthropometric indices exhibited significant positive associations with the incidence of CVD. TyG–waist-to-height ratio (WHtR) showed the strongest association, with each 1-SD increase correlating with a 25% increase in CVD risk. Elevated systolic and diastolic blood pressure (BP) levels partially mediated these associations. TyG–weight-adjusted waist index (WWI) indicated the highest predictive performance in CHARLS, while TyG–conicity index (ConI) was the most predictive in the MESA cohort. Across both cohorts, TyG–WHtR provided the most substantial incremental improvement to the baseline model. Models combining TyG–anthropometric indices showed higher prediction accuracy and goodness of fit than the basic model combining TyG or anthropometric indices alone. Decision curve analysis showed that TyG–WC and TyG–ConI yielded the superior net clinical benefits for CVD prediction. Subgroup analyses demonstrated consistent associations between TyG–anthropometric indices and CVD incidence across different clinical characteristics and sociodemographic groups.
Conclusions
The integration of TyG with anthropometric indices strengthened its association with CVD incidence. While TyG–WWI and TyG–ConI exhibited the highest predictive ability in the CHARLS and MESA cohorts, respectively, TyG–WHtR consistently yielded the greatest improvement to traditional risk models. Elevated BP levels partially mediated this association. Early intervention targeting visceral adiposity and impaired insulin sensitivity is crucial for mitigating CVD incidence in middle-aged and older adults.
Graphical Abstract
Supplementary Information
The online version contains supplementary material available at 10.1186/s40001-026-03907-x.
Keywords: Insulin resistance, Visceral adiposity, Triglyceride-glucose (TyG) index, Cardiovascular disease
Background
Cardiovascular disease (CVD) constitutes the foremost cause of mortality worldwide, responsible for approximately 17.9 million deaths each year, equating to 32% of total global mortality. More than 80% of these fatalities occur in individuals aged 50 and older [1]. Insulin resistance (IR), an important component of metabolic dysregulation, significantly increases the risk of CVD incidence [2, 3]. The triglyceride-glucose (TyG) index, serving as an IR surrogate, has demonstrated an independent association with atherosclerosis, coronary artery disease (CAD), and stroke, underscoring its value in CVD risk stratification [4–7].
The TyG index strongly predicts CVD risk in obesity, implying potential synergistic effects [8]. However, BMI is limited in its ability to distinguish visceral adiposity, a limitation underscored by the recent European Association for the Study of Obesity (EASO) framework defining obesity by abnormal adipose accumulation rather than weight alone [9]. While newer anthropometric indices better characterize central fat distribution [10–15], their interaction with TyG remains under-investigated. Furthermore, systematic comparisons to determine which TyG–anthropometric combination provides the most accurate risk stratification are currently absent.
We hypothesized that integrating TyG with novel visceral adiposity indices—which offer a more refined assessment of central fat distribution—would yield stronger predictive value for CVD incidence than combinations relying on BMI. To test this hypothesis and identify the optimal composite marker, we systematically evaluated and compared different TyG–anthropometric indices using data from the China Health and Retirement Longitudinal Study (CHARLS). We examined their joint associations, interactions, and mediating pathways with CVD risk. Furthermore, to ensure the robustness and generalizability of our findings across different populations, we performed an external validation using the Multi-Ethnic Study of Atherosclerosis (MESA) cohort, a large, diverse population with adjudicated CVD incidence.
Methods
Study population
This study conducted a secondary analysis of the CHARLS, a nationally representative cohort investigating the economic, social, and health status of Chinese adults aged 45 years and older. The baseline survey, carried out between June 2011 and March 2012, included 17,708 participants from 450 villages in 150 counties or districts across 28 provinces. This study used data from survey waves conducted from 2011 to 2020. Participants aged 45 years and older with complete data on fasting blood glucose (FBG), triglycerides (TGs), height, weight, and waist circumference (WC) were eligible for inclusion. Individuals with a prior CVD diagnosis, incomplete follow-up data, and missing anthropometric measurements were excluded from the analysis. Ultimately, a total of 7046 participants met the final inclusion criteria. This study also used data from 6472 participants in the MESA cohort between 2000 and 2015 as a validation population. This study was conducted in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines [16]. The participant selection process is detailed in Supplemental Fig. 1.
Data assessment
At the Chinese Center for Disease Control and Prevention, fasting venous blood samples were gathered and examined by medical personnel using a standard protocol. The samples were stored and frozen at − 20 °C before they were transported. After completing the requisite laboratory tests, the samples were stored in a deep freezer at a temperature of − 80 °C.
TG and FBG levels were measured using an enzyme colorimetric assay. A random blood sample was taken from each participant to measure cholesterol, C-reactive protein, and creatinine levels using a Beckman Coulter AU5800 analyzer. Eight indicators—BMI, WC, waist-to-height ratio (WHtR), body shape index (ABSI), relative fat mass (RFM), body roundness index (BRI), waist-to-weight index (WWI), and conicity index (ConI)—were used to assess general or visceral adiposity. The TyG index was determined using the formula: TyG index = Ln [TG (mg/dL) × FBG (mg/dL/2)]. Height was measured with a vertical altimeter, and weight was taken using a scale during the physical examination. Standing WC was measured at the navel using a soft ruler. The TyG–adiposity indices were calculated using the following equations:
TyG–WC = Ln [TG (mg/dL) × glucose (mg/dL)/2] × WC (m);
TyG–WHtR = Ln (TG × glucose/2) × (WC/height);
TyG–BMI = Ln (TG × glucose/2) × weight/height2);
TyG–ABSI = Ln (TG × glucose/2) × (WC/BMI2/3 × height/2);
TyG–RFM = Ln (TG × glucose/2) × [64—(20 × (height/WC) + 12 × 0 (0, male; 1, female))];
TyG–BRI = Ln (TG × glucose/2) × (364.2—365.5 (1—[WC/2π]2 / [0.5 × height]2)/2];
TyG–WWI = Ln (TG × glucose/2) × ((WC/weight/2);
TyG–ConI = Ln (TG × glucose/2) × [(WC/(0.109 × weight/height)/2].
Outcomes
The primary outcome of the study was the development of newly diagnosed CVD events. CVD events were identified based on participant responses to standardized follow-up questionnaires, which inquired about physician diagnoses of conditions such as myocardial infarction, angina pectoris, congestive heart failure (CHF), CAD, and stroke. Participants who confirmed a new CVD diagnosis during follow-up were classified as having experienced an incident event, with the self-reported date used to determine the time of onset. For these individuals, time to event was calculated from the baseline survey to the onset date. In contrast, for participants who did not report any CVD during the study period, follow-up time was defined as the interval between the baseline visit and their last available follow-up. Furthermore, the incident CVD events in the MESA cohort (2000–2015) were adjudicated by an independent committee using medical records and ICD-10 codes.
Covariates
During the baseline survey, trained staff used a standardized questionnaire to gather comprehensive data on the participants’ sociodemographic details and health conditions. The assessed covariates included age, sex, type of residence (urban or rural), self-reported educational level (categorized as no formal education, primary school, junior high school, high school, or higher), and marital status (married vs. other). Lifestyle factors such as smoking and alcohol consumption were also recorded. Additionally, medical history was obtained for physician-diagnosed hypertension, diabetes, and kidney disease, alongside measurements of high-sensitivity C-reactive protein (hs-CRP). Diabetes was defined based on one or more of the following: fasting blood glucose (FBG) ≥ 126 mg/dL, glycated hemoglobin (HbA1c) ≥ 6.5%, current use of glucose-lowering medications, or a documented history.
Statistical analyses
For continuous variables, non-normally distributed data are reported as median (IQR), and normally distributed data are reported as mean ± standard deviation (SD). Percentages are used to present categorical variables. To guarantee the accuracy and integrity of these assessments, only available non-imputed data were used for baseline characteristic comparisons. Continuous variables were analyzed using the Kruskal–Wallis or Wilcoxon rank sum tests, and categorical variables were assessed with the Chi-square or Fisher’s exact tests. To compare groups, TyG and TyG–adiposity indices were organized into quartiles, and multivariate Cox regression analysis was used to evaluate CVD risk. The proportional hazards assumption was verified using Schoenfeld residuals, and no violations were observed. Covariates were selected based on their biological plausibility and established associations with CVD in prior epidemiological evidence. The cut-off values for TyG and TyG–adiposity indices were determined through Kaplan–Meier analyses. Restricted cubic splines (RCS) were used in multivariable Cox models to investigate potential nonlinear relationships between basal TyG, anthropometric indices, and outcomes. The knots in the spline models were determined by minimizing the Akaike information criterion (AIC) values. To explore the mediating role of BP levels (systolic and diastolic) in the association between TyG–anthropometric indices and CVD, causal mediation analyses were performed based on a counterfactual framework. After adjusting for confounders, we estimated the proportion of mediators and total effect. Proportion mediated was calculated as: (indirect effect/total effect) × 100%. Additionally, receiver operating characteristic (ROC) curves and areas under the curve (AUC) were computed to quantify the incremental predictive values of TyG–anthropometric indices for CVD events. The overall model fit was evaluated using the root-mean-square error (RMSE), coefficient of determination (Nagelkerke’s R2), AIC, and Bayesian information criterion (BIC). Finally, decision curve analysis (DCA) was employed to assess the clinical benefits and represent the net benefits across a range of threshold probabilities.
A sensitivity analysis was conducted to enhance the robustness of the findings. Given the influence of age on the risk of CVD events, we re-evaluated the association between TyG–anthropometric incidence and CVD risk in patients aged > 65 years, as classified by the World Health Organization [19]. Subgroup analyses were performed based on sex (female or male), residence, hypertension, diabetes, kidney disease, and dyslipidemia to identify susceptible populations. To further reduce the risk of reverse causality, individuals diagnosed with CVD events within the first 2 years of follow-up were excluded from the sensitivity analysis. Additionally, glycated hemoglobin (HbA1c) levels were adjusted to mitigate the potential impact of non-fasting glucose levels. SPSS (version 26.0; SPSS Inc., Chicago, IL) and R Statistics (version 4.5.2; R Foundation for Statistical Computing, Vienna, Austria) were used for all statistical analyses. The R packages utilized in this study include survival, survminer, RMS, RMDA, mediation, and pROC.
Results
Population characteristics
Baseline characteristics of the 7046 participants are summarized in Table 1. Over a nine-year follow-up, 1768 (25.1%) participants developed CVD. Compared to those without CVD, participants who developed CVD were older, had a higher proportion of females, and exhibited a higher prevalence of metabolic comorbidities than those without CVD. Significantly higher baseline levels of TyG and all TyG–anthropometric indices were found in the CVD group (P < 0.001). There were no statistical differences between the groups in terms of lifestyle factors like alcohol consumption or smoking. Baseline characteristics of the 6472 participants in the MESA cohort are shown in Supplemental Table 1.
Table 1.
Basic characteristics of the study population
| Characteristic | Total population N = 70461 |
Incidence of CVD | P-value | |
|---|---|---|---|---|
| No, N = 52781 | Yes, N = 17681 | |||
| Age (years) | 58 (51, 65) | 57 (51, 64) | 59 (54, 66) | < 0.0012 |
| Gender | < 0.0013 | |||
| Male | 3317 (47.1%) | 2561 (48.5%) | 756 (42.8%) | |
| Female | 3,729 (52.9%) | 2717 (51.5%) | 1012 (57.2%) | |
| Marital status | 0.0323 | |||
| Married | 6210 (88.1%) | 4677 (88.6%) | 1533 (86.7%) | |
| Other | 836 (11.9%) | 601 (11.4%) | 235 (13.3%) | |
| Education degree | 0.2913 | |||
| College or above | 77 (1.1%) | 57 (1.1%) | 20 (1.1%) | |
| Middle or high school | 2,007 (28.5%) | 1,535 (29.1%) | 472 (26.7%) | |
| No formal education | 3451 (49.0%) | 2561 (48.5%) | 890 (50.3%) | |
| Primary school | 1511 (21.4%) | 1125 (21.3%) | 386 (21.8%) | |
| Resident | 0.7933 | |||
| Urban community | 2409 (34.2%) | 1800 (34.1%) | 609 (34.4%) | |
| Rural village | 4637 (65.8%) | 3478 (65.9%) | 1159 (65.6%) | |
| Drinking status | 0.0643 | |||
| Never | 4295 (61.0%) | 3190 (60.4%) | 1105 (62.5%) | |
| Former | 2751 (39.0%) | 2088 (39.6%) | 663 (37.5%) | |
| Current | 2184 (31.0%) | 1696 (32.1%) | 488 (27.6%) | |
| Smoking status | 0.1093 | |||
| Never | 4234 (60.1%) | 3143 (59.5%) | 1091 (61.7%) | |
| Former | 2812 (39.9%) | 2135 (40.5%) | 677 (38.3%) | |
| Current | 2416 (34.3%) | 1876 (35.5%) | 540 (30.5%) | |
| Blood pressure | ||||
| Systolic BP, mmHg | 130 ± 21 | 129 ± 21 | 134 ± 22 | < 0.0012 |
| Diastolic BP, mmHg | 76 ± 12 | 75 ± 12 | 77 ± 12 | < 0.0012 |
| History of comorbidities | ||||
| Hypertension | 1603 (22.8%) | 993 (18.8%) | 610 (34.5%) | < 0.0013 |
| Diabetes | 364 (5.2%) | 234 (4.4%) | 130 (7.4%) | < 0.0013 |
| Dyslipidemia | 550 (7.8%) | 327 (6.2%) | 179 (10.1%) | < 0.0013 |
| Kidney disease | 341 (4.8%) | 232 (4.4%) | 109 (6.2%) | 0.0033 |
| Laboratory values | ||||
| Fasting glucose, mg/dL | 102 (95, 112) | 102 (94, 111) | 103 (95, 114) | < 0.0012 |
| Creatinine, mg/dL | 0.76 (0.64, 0.88) | 0.76 (0.64, 0.88) | 0.75 (0.64, 0.87) | 0.2602 |
| Total cholesterol, mg/dL | 191 (168, 216) | 191 (167, 215) | 194 (170, 220) | < 0.0012 |
| Triglycerides, mg/dL | 103 (73, 148) | 100 (72, 145) | 109 (79, 156) | < 0.0012 |
| HDL cholesterol, mg/dL | 50 (41, 61) | 50 (41, 61) | 49 (41, 59) | < 0.0012 |
| LDL cholesterol, mg/dL | 115 (94, 138) | 114 (94, 137) | 117 (97, 141) | < 0.0012 |
| Hs-CRP, mg/dL | 1.00 (0.54, 2.10) | 0.95 (0.53, 1.99) | 1.14 (0.59, 2.29) | < 0.0012 |
| Glycated hemoglobin, % | 5.10 (4.90, 5.40) | 5.10 (4.90, 5.40) | 5.20 (4.90, 5.50) | < 0.0012 |
| TyG-related indexes | ||||
| TyG index | 8.57 (8.20, 9.00) | 8.54 (8.18, 8.97) | 8.65 (8.29, 9.07) | < 0.0012 |
| TyG–WC index | 7.20 (6.47, 8.10) | 7.12 (6.40, 7.98) | 7.46 (6.66, 8.38) | < 0.0012 |
| TyG–WHtR index | 3.12 (2.74, 3.59) | 3.08 (2.72, 3.53) | 3.24 (2.83, 3.76) | < 0.0012 |
| TyG–BMI index | 198 (175, 226) | 195 (173, 222) | 206 (180, 237) | < 0.0012 |
| TyG–WWI index | 95 (88, 103) | 95 (88, 103) | 98 (90, 105) | < 0.0012 |
| TyG–ABSI index | 0.71 (0.66, 0.76) | 0.71 (0.66, 0.76) | 0.72 (0.67, 0.77) | < 0.0012 |
| TyG–RFM index | 283 (211, 348) | 275 (207, 343) | 302 (228, 363) | < 0.0012 |
| TyG–BRI index | 34 (26, 44) | 33 (26, 43) | 37 (29, 48) | < 0.0012 |
| TyG–ConI index | 11.0 (10.2, 11.9) | 10.9 (10.2, 11.8) | 11.2 (10.4, 12.1) | < 0.0012 |
1Median (IQR); mean ± SD; n (%); 2Wilcoxon rank sum test; 3Pearson’s Chi-squared test
Interaction analyses
Figure 1 shows the distribution of TyG, TyG–BMI, and TyG–visceral adiposity indices. Cox regression analyses of interaction terms revealed a potential negative interaction between TyG and anthropometric indices. Statistically significant interactions were observed for TyG*WC (hazard ratio [HR]: 0.94, 95%CI: 0.89–0.99), TyG*RFM (HR: 0.95, 95% CI: 0.90–1.00), and TyG*BRI (HR: 0.95, 95%CI: 0.91–1.00), suggesting that the incremental risk associated with elevated TyG may be attenuated in individuals with already high levels of visceral adiposity, potentially reflecting a risk saturation effect. No statistical interactions were observed for WHtR, BMI, ABSI, WWI, or ConI (Table 2). Furthermore, consistent negative interaction trends were also observed across all TyG*anthropometric indices within the MESA cohort, with the interaction between TyG and WWI showing statistical significance (HR: 0.93, 95% CI: 0.87–1.00; Supplemental Table 2).
Fig. 1.
Distribution of T TyG and TyG–anthropometric indices. BMI, body mass index; WC, waist circumference; WHtR, waist-to-height ratio; ABSI, body shape index; RFM, relative fat mass; BRI, body roundness index; WWI, weight-waist index; ConI, conicity index
Table 2.
Cox regression analysis of interaction between TyG and adiposity indices
| Characteristic | N | Event N | HR | 95% CI | P for interaction |
|---|---|---|---|---|---|
| TyG * WC | 6873 | 1595 | 0.94 | 0.89, 0.99 | 0.018 |
| TyG * WHtR | 6873 | 1595 | 0.97 | 0.93, 1.01 | 0.101 |
| TyG * BMI | 6873 | 1595 | 0.99 | 0.98, 1.00 | 0.056 |
| TyG * ABSI | 6873 | 1595 | 0.98 | 0.94, 1.03 | 0.486 |
| TyG * RFM | 6873 | 1595 | 0.95 | 0.90, 1.00 | 0.038 |
| TyG * BRI | 6873 | 1595 | 0.95 | 0.91, 1.00 | 0.044 |
| TyG * WWI | 6873 | 1595 | 0.97 | 0.93, 1.02 | 0.238 |
| TyG * CI | 6873 | 1595 | 0.97 | 0.93, 1.02 | 0.226 |
HR, hazard ratio; CI, confidence interval
Association of TyG–anthropometric indices with CVD incidence
After adjusted for potential covariates, all TyG–anthropometric indices (except TyG–ABSI) were positively associated with incident CVD (Table 3). In the fully adjusted model (Model 3), TyG–WHtR demonstrated the strongest association in multivariable Cox regression analyses, with a 25% increase in CVD risk for each 1-SD increase (aHR: 1.25, 95% CI: 1.20–1.30). This was followed closely by TyG–WC and TyG–BMI. Continuous analyses confirmed that elevated levels of these indices significantly predicted CVD events across all models.
Table 3.
Association between TyG and TyG–adiposity indices and CVD incidence
| Characteristics Per SD |
Model 1 | Model 2 | Model 3 | ||||||
|---|---|---|---|---|---|---|---|---|---|
| HR1 | 95% CI | p-value | HR | 95% CI | p-value | HR | 95% CI | p-value | |
| TyG | 1.14 | 1.09, 1.19 | < 0.001 | 1.14 | 1.09, 1.19 | < 0.001 | 1.12 | 1.07, 1.17 | 0.007 |
| TyG-WC | 1.24 | 1.18, 1.30 | < 0.001 | 1.24 | 1.18, 1.30 | < 0.001 | 1.23 | 1.17, 1.29 | < 0.001 |
| TyG–WHtR | 1.25 | 1.20, 1.31 | < 0.001 | 1.26 | 1.21, 1.31 | < 0.001 | 1.25 | 1.20, 1.30 | < 0.001 |
| TyG–BMI | 1.23 | 1.18, 1.28 | < 0.001 | 1.23 | 1.18, 1.28 | < 0.001 | 1.22 | 1.17, 1.27 | < 0.001 |
| TyG–ABSI | 1.05 | 1.01, 1.09 | 0.007 | 1.05 | 1.01, 1.09 | 0.007 | 1.04 | 1.00, 1.09 | 0.033 |
| TyG–RFM | 1.15 | 1.07, 1.24 | < 0.001 | 1.15 | 1.07, 1.24 | < 0.001 | 1.13 | 1.05, 1.21 | 0.001 |
| TyG–BRI | 1.21 | 1.15, 1.26 | < 0.001 | 1.21 | 1.16, 1.27 | < 0.001 | 1.20 | 1.14, 1.25 | < 0.001 |
| TyG–WWI | 1.10 | 1.05, 1.15 | < 0.001 | 1.10 | 1.05, 1.15 | < 0.001 | 1.08 | 1.03, 1.14 | < 0.001 |
| TyG–ConI | 1.11 | 1.06, 1.16 | < 0.001 | 1.11 | 1.06, 1.16 | < 0.001 | 1.10 | 1.05, 1.15 | < 0.001 |
HR, hazard ratio; CI, confidence interval
Categorical analyses using quartiles further corroborated these findings. Kaplan–Meier curves indicated a significantly higher cumulative risk of CVD for participants in the highest quartile of all indices (Fig. 2). Consistently, Cox models revealed a clear risk gradient across quartiles. Comparing the highest versus lowest quartiles, TyG–RFM showed the highest risk estimate, followed by TyG–BMI and TyG–WHtR. Moderate but statistically significant associations were also observed for TyG–BRI, TyG–ConI, TyG–WWI, and TyG–ABSI (Fig. 3). The parallel risk trends persisted in the MESA cohort, where TyG–ABSI demonstrated the highest risk estimate, followed by TyG–RFM and TyG–WHtR (Supplemental Fig. 2).
Fig. 2.
Kaplan–Meier analyses of CVD incidence by TyG and TyG–anthropometric indices. BMI, body mass index; WC, waist circumference; WHtR, waist-to-height ratio; ABSI, body shape index; RFM, relative fat mass; BRI, body roundness index; WWI, weight-waist index; ConI, conicity index
Fig. 3.
Adjusted HRs and 95%CIs for CVD incidence across quintiles of indices. TyG, triglyceride-glucose index; WC, waist circumference; WHtR, waist-to-height ratio; BMI, body mass index; ABSI, body shape index; RFM, relative fat mass; BRI, body roundness index; WWI: weight-waist index; ConI, conicity index. HR, hazard ratio; CI, confidence interval
Dose–response relationships and threshold analyses
Restricted cubic spline (RCS) analyses visualized the dose–response relationships between the indices and CVD risk. After adjustment for covariates, J-shaped non-linear associations were observed for TyG–WC, TyG–WHtR, TyG–BMI, TyG–RFM, TyG–BRI, TyG–WWI, and TyG–ConI (Fig. 4). Threshold analysis identified specific inflection points for these indices, suggesting a potential saturation effect of visceral adiposity on CVD risk (Table 4).
Fig. 4.
Dose–response relationships of TyG and TyG–anthropometric indices with CVD incidence. Solid lines represent HRs and shaded areas represent 95% CIs. The reference point was set at the 25th percentile of each index. TyG, triglyceride-glucose index; WC, waist circumference; WHtR, waist-to-height ratio; BMI, body mass index; ABSI, body shape index; RFM, relative fat mass; BRI, body roundness index; WWI: weight-waist index; ConI, conicity index. HR, hazard ratio; CI, confidence interval
Table 4.
Threshold effect analysis of variables on CVD
| Characteristics | Adjusted HR (95% CI) * | Log likelihood ratio | p-value |
|---|---|---|---|
| TyG index | |||
| TyG index < 8.40 | 1.44 (1.13, 1.85) | 0.098 | 0.004 |
| TyG index ≥ 8.40 | 1.13 (1.03, 1.24) | 0.013 | |
| TyG–WC | |||
| TyG–WC < 5.71 | 0.88 (0.79, 0.98) | < 0.001 | 0.019 |
| TyG–WC ≥ 5.71 | 1.22 (1.17, 1.27) | < 0.001 | |
| TyG–WHtR | |||
| TyG–WHtR < 4.36 | 1.55 (1.42, 1.68) | < 0.001 | < 0.001 |
| TyG–WHtR ≥ 4.36 | 0.99 (0.78, 1.24) | 0.910 | |
| TyG–BMI | |||
| TyG–BMI < 274.49 | 1.01 (1.01, 1.01) | < 0.001 | < 0.001 |
| TyG–BMI ≥ 274.49 | 1.00 (1.00, 1.00) | 0.602 | |
| TyG–ABSI | |||
| TyG–ABSI < 0.60 | 0.67 (0.25, 1.76) | 0.093 | 0.415 |
| TyG–ABSI ≥ 0.60 | 1.77 (1.22, 2.58) | 0.003 | |
| TyG–RFM | |||
| TyG–RFM < 155.81 | 1.00 (1.00, 1.00) | < 0.001 | 0.005 |
| TyG–RFM ≥ 155.81 | 1.00 (1.00, 1.00) | < 0.001 | |
| TyG–BRI | |||
| TyG–BRI < 19.50 | 0.99 (0.97, 1.00) | < 0.001 | 0.038 |
| TyG–BRI ≥ 19.50 | 1.02 (1.01, 1.02) | < 0.001 | |
| TyG–WWI | |||
| TyG–WWI < 78.63 | 0.99 (0.99, 1.00) | 0.006 | 0.196 |
| TyG–WWI ≥ 78.63 | 1.01 (1.00, 1.01) | < 0.001 | |
| TyG–ConI | |||
| TyG–ConI < 9.14 | 0.96 (0.90, 1.02) | 0.002 | 0.174 |
| TyG–ConI ≥ 9.14 | 1.09 (1.05, 1.12) | < 0.001 |
*Adjusted for: age, gender, education, married, resident, smoking status, alcohol use, diabetes, kidney disease, and hs-CRP level
Threshold analysis revealed that TyG–anthropometric indices exhibited unique threshold points beyond which the risk trajectory changed (Table 4). These findings suggest that IR and visceral adiposity have a saturated effect on the risk of CVD. Because TyG and TyG–ABSI demonstrated a linear dose–response relationship with CVD risk, no threshold effect was observed (Fig. 5).
Fig. 5.
Roles of visceral adiposity indices in the association between TyG and CVD incidence. a-h. Mediation models illustrating the direct and indirect associations of TyG and TyG–anthropometric indices with CVD incidence. Mediation models illustrating direct and indirect effects. TyG, triglyceride-glucose index; WC, waist circumference; WHtR, waist-to-height ratio; BMI, body mass index; ABSI, body shape index; RFM, relative fat mass; BRI, body roundness index; WWI, weight-waist index; ConI, conicity index; SBP, systolic blood pressure; DBP, diastolic blood pressure
Mediation analysis
As Supplemental Fig. 3 shows, anthropometric indices, especially visceral-adiposity indices, were significant partial mediators in the relationship between TyG and CVD, with WHtR and WC mediating 43.2% and 36.7% of the association, respectively. BMI and BRI also accounted for a significant proportion of the risk (35.0% and 25.7%, respectively), whereas the mediating proportion of RFM was relatively minor (5.3%). SBP and DBP each mediated approximately 18% of the TyG–CVD association. Furthermore, the association between TyG–anthropometric indices and CVD incidence was also consistently mediated by SBP and DBP. The magnitude of the indirect effect via BP levels varied across TyG–anthropometric indices, ranging from the lowest for TyG–WHtR (12.7% for SBP; 13.5% for DBP) to the highest for TyG–ABSI (31.1% for SBP; 32.0% for DBP) (all P < 0.001).
Predictive performance of anthropometric indices
Figure 6 and Supplemental Fig. 4 show the incremental predictive values of the TyG–anthropometric indices when added to the basic prediction model. Paired comparisons revealed that adding TyG–visceral adiposity indices, especially TyG–WHtR, improved CVD risk prediction (AUC increase = 0.030 in CHARLS; AUC increase = 0.024 in MESA) and outperformed traditional risk factors combined with the TyG index (0.617 vs. 0.593 and 0.711 vs. 0.695, P < 0.001). Among TyG–adiposity indices, TyG–WWI (AUC = 0.578, 95%CI: 0.563–0.594) and TyG–ConI (AUC = 0.599, 95%CI: 0.577–0.622) showed the highest discriminatory ability; see Fig. 7.
Fig. 6.
Incremental predictive value of indices beyond the basic model for CVD incidence. BMI, body mass index; WC, waist circumference; WHtR, waist-to-height ratio; BMI, body mass index; ABSI, body shape index; RFM, relative fat mass; BRI, body roundness index; WWI, weight-waist index; ConI, conicity index. *Basic model incorporating age, sex, education, smoking status, marital status, residence, drinking status, diabetes, kidney disease, and hs-CRP level
Fig. 7.
ROC curves for CVD prediction in the CHARLS and MESA cohorts. a. Predictive value of TyG and TyG–anthropometric indices for CVD incidence in CHARLS cohort. b. Predictive value of TyG and TyG–anthropometric indices for CVD incidence in MESA cohort. BMI, body mass index; WC, waist circumference; WHtR, waist-to-height ratio; BMI, body mass index; ABSI, body shape index; RFM, relative fat mass; BRI, body roundness index; WWI, weight-waist index; ConI, conicity index; ROC, receiver operating characteristic curve. CHARLS: the China Health and Retirement Longitudinal Study. MESA: the Multi-Ethnic Study of Atherosclerosis.
Calibration curves assessed the agreement between predicted and observed probabilities within a range of 0.1 to 0.6. In CHARLS, TyG–WC, TyG–WHtR, and TyG–BRI demonstrated the highest degree of alignment between apparent, bias-corrected, and ideal curves; see Fig. 8. In MESA, TyG–WWI, TyG–ConI, and TyG–WC yielded optimal calibration, with three curves closely approximating the ideal 45-degree line; see Supplemental Fig. 5.
Fig.8.
Calibration curves for predicted versus observed CVD probability in the CHARLS cohort. CHARLS: the China Health and Retirement Longitudinal Study
Performance metrics including Nagelkerke’s R2, AIC, BIC, and RMSE assessed model fit, predictive accuracy, and error rates. Integrating TyG with visceral adiposity indices yielded superior accuracy and enhanced goodness of fit compared to models utilizing either index independently (all P < 0.05 for model comparison, see Supplemental Table 3), without escalating prediction error (Fig. 9; Supplemental Fig. 6). TyG–WHtR and TyG–WWI achieved the highest Nagelkerke R2 values (0.046 and 0.081, respectively) across the two cohorts, indicating improved explanatory power for CVD incidence. Reduced AIC and BIC values further confirmed that TyG–visceral adiposity indices provide a more optimal balance between model complexity and fit. Furthermore, the RMSE remained stable and comparable to the baseline TyG model, demonstrating the robustness and precision of these composite indices in CVD risk assessment.
Fig. 9.
Comparison of model performance metrics for CVD prediction. The overall model fit was evaluated using the root-mean-square error (RMSE), the coefficient of determination (Nagelkerke’s R2), Akaike information criterion (AIC), and Bayesian information criterion (BIC). BMI, body mass index; WC, waist circumference; WHtR, waist-to-height ratio; BMI, body mass index; ABSI, body shape index; RFM, relative fat mass; BRI, body roundness index; WWI, weight-waist index; ConI, conicity index. *Basic model incorporating age, sex, education, smoking status, marital status, residence, drinking status, diabetes, kidney disease, and hs-CRP level
DCA was performed to evaluate the clinical utility of the models. Within threshold probabilities of approximately 0.1 to 0.5, TyG–visceral adiposity indices, particularly TyG–WC and TyG–WWI, provided a higher net benefit compared to the “all” or “none” strategies. However, the net benefit plateaued at higher thresholds, suggesting the clinical utility of these indices is most pronounced for screening purposes within specific risk ranges (Fig. 10 and Supplemental Fig. 7).
Fig. 10.
Decision curve analysis (DCA) evaluating the clinical utility of TyG–visceral adiposity indices. The x-axis represents threshold probabilities, and the y-axis represents the net benefit. Combined models are compared against the “All” and “None” strategies. DCA, decision curve analysis
Sensitivity and subgroup analyses
Further sensitivity analyses were performed to ensure the robustness of results. Similar results were obtained in the sensitivity analysis upon the exclusion of outcomes from the initial two years of follow-up (Supplemental Table 4) and adjusting for HbA1c to reduce the potential effect of non-FBG (Supplemental Table 5). Elevated baseline TyG and TyG–anthropometric indices were associated with higher risk ratios for CVD development in people younger than 65 years of age and higher risk ratios in men and were more pronounced in patients with baseline diabetes and kidney disease (Supplemental Fig. 8–16).
The results of the subgroup analyses stratified by residence, history of hypertension, and use of baseline dyslipidemia medication were consistent with the results of the primary analysis (Supplemental Fig. 8–16). No significant interactions were found between the subgroups except for the age subgroup (PInteraction > 0.05), suggesting that TyG–visceral adiposity indices have a consistent effect on CVD risk.
Discussion
To the best of our knowledge, this study is the first to confirm the joint association of TyG and comprehensive anthropometric indices on CVD incidence in a national cohort. To ensure the robustness and generalizability of our findings in this study, we performed an external validation using data from the MESA cohort. The consistency of results across both the CHARLS and the MESA suggests that the observed associations between TyG–anthropometric indices and CVD risk are robust across different genetic backgrounds and environmental settings. Second, this study systematically evaluates the joint predictive value of the TyG index combined with a comprehensive array of novel visceral adiposity indices, providing a holistic assessment of cardiometabolic risk. Third, we employed rigorous statistical approaches, including mediation analysis based on a counterfactual framework, to elucidate the mechanistic pathways linking metabolic dysregulation to cardiovascular outcomes.
Our results show that when combined with anthropometric indices, especially visceral adiposity indices, the hazard ratio for CVD incidence was higher than that of the TyG index alone, and the predictive ability was statistically improved. Specifically, the combination of the TyG index with WC or WWI exhibited better predictive capabilities for CVD incidence compared with TyG–BMI. Furthermore, in both cohorts, TyG–WHtR improved the basic prediction model more effectively than TyG alone or TyG–BMI. These findings suggest that TyG–anthropometric indices may function as sensitive, accessible markers of metabolically associated cardiovascular risk, particularly in individuals with normal BMI but high visceral fat accumulation.
Our findings further expand previous epidemiological evidence suggesting that adiposity indices incorporating visceral fat distribution offer robust predictive value [13–15, 17, 18]. While visceral adiposity and IR are known major components of metabolic syndrome [19, 20], few studies have evaluated their combined utility in a longitudinal setting. Previous research has indicated that indices such as ABSI, WWI, and BRI outperform BMI in predicting metabolic outcomes [14, 15, 18]. For instance, WWI has been linked to arterial stiffness, and BRI provides a remarkable estimation of body fat percentage compared to BMI. By integrating the TyG index—a reliable, cost-effective surrogate for IR [21, 22]—with these advanced anthropometric measures, our study bridges the gap between metabolic and morphological risk assessment. The better performance of composite indices like TyG–WC and TyG–WWI in our study supports the notion that capturing both metabolic dysfunction and central adiposity provides a more holistic view of cardiovascular risk than either component alone.
We also observed a modest negative interaction between the TyG index and visceral adiposity indices, which indicates a sub-additive effect on CVD risk. This phenomenon may be related to the overlapping metabolic pathways or metabolic saturation of the two indices. Since both the TyG index and visceral adiposity share common pathophysiological pathways—including chronic inflammation, oxidative stress, and adipokine imbalance [23, 24]—their combined predictive information overlaps partially. Specifically, elevated TyG levels reflect established insulin resistance and lipotoxicity. In this state, the pathogenic pathways leading to endothelial dysfunction and atherosclerosis may already be activated near their physiological maximum. Consequently, the marginal contribution of excess visceral adiposity to CVD risk is attenuated in the presence of high TyG, compared to its stronger impact in individuals with a normal metabolic profile [25]. These observations may suggest a potential saturation effect, wherein the incremental risk contributed by adiposity is attenuated once a certain metabolic threshold—represented by elevated TyG levels—is surpassed, possibly due to convergent metabolic pathways.
Another novel contribution of this study is revealing the mediating role of BP levels in the impact of TyG–anthropometric indices on CVD development. This finding strengthens the mechanistic connection between IR, visceral adiposity, and cardiovascular pathology. Our mediation analysis, based on a counterfactual framework, indicated that SBP and DBP significantly mediated the association between all TyG–anthropometric indices and CVD risk. Biologically, IR and visceral adiposity can interact synergistically to promote hypertension through multiple mechanisms, including endothelial dysfunction, dysregulation of the renin–angiotensin–aldosterone system (RAAS), and sympathetic nervous system activation [26, 27]. These processes lead to vasoconstriction, sodium retention, and vascular remodeling. Chronic hypertension exacerbates shear stress, resulting in endothelial injury and accelerated atherosclerotic progression [28]. By identifying blood pressure as a key mediator, our study suggests that the adverse cardiovascular effects of high TyG and visceral adiposity are, to a significant extent, channeled through hemodynamic alterations.
Our sensitivity analysis further validated the robustness of these findings. The association between higher baseline TyG–anthropometric indices and CVD risk remained significant after excluding participants with < 2 years of follow-up (mitigating reverse causality) and adjusting for glycated hemoglobin. Subgroup analyses revealed stronger associations in men, individuals under 65 years of age, and those without a history of hypertension. The stronger predictive value in younger and non-hypertensive populations highlights the utility of these indices for early screening and primary prevention, identifying at-risk individuals before overt clinical signs appear.
In this study, several limitations should be acknowledged. Firstly, in the CHARLS derivation cohort, CVD outcomes were ascertained based on self-reported information, which may introduce recall bias. To mitigate this impact, we performed external validation using the MESA cohort, yielding results that were consistent with our study’s outcomes, thereby further corroborating the validity of our findings. Secondly, given the observational nature of the study, we cannot establish causality. Although we adjusted for a comprehensive set of covariates, the possibility of unmeasured or residual confounding remains. Specifically, we lacked precise data on physical activity levels and detailed medication dosages (such as antihypertensives or statins). Meanwhile, due to data constraints on menopausal status and hormone levels, the impact of sex-specific biological factors on predictive performance and optimal thresholds remains unevaluated. Thirdly, we did not apply corrections for multiple testing due to the high correlation between the anthropometric indices, which may increase the risk of type I errors. To mitigate this, we focused on the consistency of associations across different models rather than isolated statistical significance. Finally, both study populations were restricted to middle-aged and older adults (aged 45 years and older). Consequently, the findings may not be fully generalizable to younger populations, warranting future research across broader age spectrums. Future investigations should focus on examining the longitudinal trajectories of these indices and exploring whether interventions targeting the adiposity–insulin resistance axis (such as lifestyle modifications or GLP-1 receptor agonists) can effectively reduce CVD incidence in high-risk subgroups defined by TyG–anthropometric indices.
Conclusions
We determined the association between TyG–visceral adiposity indices and incident CVD among middle-aged and older populations of different ethnicities, emphasizing improved predictive value compared to TyG or TyG–BMI. In addition, elevated BP levels mediate the relationship between IR, visceral adiposity, and CVD events. Future research should validate these findings and investigate their clinical implications in early CVD risk stratification and personalized prevention.
Supplementary Information
Acknowledgements
This study uses nationally representative data from the China Health and Retirement Longitudinal Study (CHARLS) and the Multi-Ethnic Study of Atherosclerosis (MESA). We thank the investigators, staff, and participants of the CHARLS and MESA studies for their invaluable contributions to this research.
Abbreviations
- ABSI
A body shape index
- AUC
Area under the curve
- BRI
Body roundness index
- BMI
Body mass index
- BP
Blood pressure
- CDC
Chinese Center for Disease Control and Prevention
- CHARLS
China Health and Retirement Longitudinal Study
- CHF
Congestive heart failure
- CI
Confidence interval
- ConI
Conicity index
- DBP
Diastolic blood pressure
- FBG
Fasting blood glucose
- IR
Insulin resistance
- MESA
Multi-Ethnic Study of Atherosclerosis
- RFM
Relative fat mass
- TyG
Triglyceride-glucose index
- WWI
Weight-adjusted waist index
- WC
Waist circumference
- WHtR
Waist-to-height ratio
Author contributions
Y. Y. and J. W. Z. were responsible for study design and manuscript drafting. X. L. (Xu Liu), L. L. W X. L. (Xue Li) and Q. Y. L. contributed to the data collection and analysis procedures. Y. G. L., Z. P. H. and Z. L. critically revised the manuscript for important intellectual content. R.X., S. Q. N., and Y. L. W. were involved in the study design and manuscript revision.
Funding
This research was partially supported by the Noncommunicable Chronic Diseases-National Science and Technology Major Project (No. 2024ZD0521500, No. 2024ZD0521504).
Availability of data and materials
The data utilized in this study are publicly accessible through the China Health and Retirement Longitudinal Study (CHARLS) online repository (https://charls.pku.edu.cn/). De-identified patient data from MESA are accessible upon authorization by the National Heart, Lung, and Blood Institute’s Biologic Specimen and Data Repository Information Coordinating Center (BioLINCC) (https://biolincc.nhlbi.nih.gov/studies/mesa/).
Declarations
Ethics approval and consent to participate
The CHARLS study received approval from the Institutional Review Board of Peking University (IRB00001052-11015), and ethical clearance for this analysis was granted by Peking University’s Ethics Review Committee. The MESA study was approved by the Institutional Review Boards of all participating field centers and the National Heart, Lung, and Blood Institute (NHLBI). Written informed consent was obtained from all participants in both cohorts prior to enrollment.
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.
Yan Yin and Jiawei Zhang contributed equally to this work.
Contributor Information
Ran Xiong, Email: xiongran1981@aliyun.com.
Shangqiu Ning, Email: ningshangqiu@163.com.
Yunlong Wang, Email: yunlong76818@126.com.
References
- 1.World Health Organization. Cardiovasc Dis (CVDs). 2021. https://www.who.int/news-room/fact-sheets/detail/cardiovascular-diseases-(cvds).
- 2.Bostick B, Aroor AR, Habibi J, Durante W, Ma L, DeMarco VG, et al. Daily exercise prevents diastolic dysfunction and oxidative stress in a female mouse model of western diet induced obesity by maintaining cardiac heme oxygenase-1 levels. Metabolism. 2017;66:14–22. 10.1016/j.metabol.2016.09.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Jiang Z, Lu Z, Wang T, Wang Y, Chu J, Chen K, et al. Lipid-lowering efficacy of Kuding tea in patients with metabolic disorders: a systematic review and meta-analysis of randomized controlled trials. Front Nutr. 2022;9:802687. 10.3389/fnut.2022.802687. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Huang R, Wang Z, Chen J, Bao X, Xu N, Guo S, et al. Prognostic value of triglyceride glucose (TyG) index in patients with acute decompensated heart failure. Cardiovasc Diabetol. 2022;21:88. 10.1186/s12933-022-01507-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Ding X, Wang X, Wu J, Zhang M, Cui M. Triglyceride-glucose index and the incidence of atherosclerotic cardiovascular diseases: a meta-analysis of cohort studies. Cardiovasc Diabetol. 2021;20:76. 10.1186/s12933-021-01268-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Lee EY, Yang HK, Lee J, Kang B, Yang Y, Lee SH, et al. Triglyceride glucose index, a marker of insulin resistance, is associated with coronary artery stenosis in asymptomatic subjects with type 2 diabetes. Lipids Health Dis. 2016;15:155. 10.1186/s12944-016-0324-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Sánchez-Iñigo L, Navarro-González D, Fernández-Montero A, Pastrana-Delgado J, Martínez JA. Risk of incident ischemic stroke according to the metabolic health and obesity states in the Vascular-Metabolic CUN cohort. Int J Stroke. 2017;12:187–91. 10.1177/1747493016672083. [DOI] [PubMed] [Google Scholar]
- 8.Du L, Xu X, Wu Y, Yao H. Association between the triglyceride glucose index and cardiovascular mortality in obese population. Nutr Metab Cardiovasc Dis. 2024;34(1):107–11. 10.1016/j.numecd.2023.08.007. [DOI] [PubMed] [Google Scholar]
- 9.Busetto L, Dicker D, Frühbeck G, et al. A new framework for the diagnosis, staging and management of obesity in adults. Nat Med. 2024;30(9):2395–9. 10.1038/s41591-024-03095-3. [DOI] [PubMed] [Google Scholar]
- 10.Chandramouli C, Tay WT, Bamadhaj NS, Tromp J, Teng TK, Yap JJL, et al. Association of obesity with heart failure outcomes in 11 Asian regions: A cohort study. PLOS Med. 2019;16:e1002916. 10.1371/journal.pmed.1002916. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Ashwell M, Gunn P, Gibson S. Waist-to-height ratio is a better screening tool than waist circumference and BMI for adult cardiometabolic risk factors: systematic review and meta-analysis. Obes Rev. 2012;13:275–86. 10.1111/j.1467-789X.2011.00952.x. [DOI] [PubMed] [Google Scholar]
- 12.Woolcott OO, Bergman RN. Relative fat mass (RFM) as a new estimator of whole-body fat percentage – a cross-sectional study in American adult individuals. Sci Rep. 2018;8:10980. 10.1038/s41598-018-29362-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Ryu K, Suliman ME, Qureshi AR, Chen Z, Avesani CM, Brismar TB, et al. Central obesity as assessed by conicity index and a-body shape index associates with cardiovascular risk factors and mortality in kidney failure patients. Front Nutr. 2023;10:1035343. 10.3389/fnut.2023.1035343. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Maessen MF, Eijsvogels TM, Verheggen RJ, Hopman MT, Verbeek AL, de Vegt F. Entering a new era of body indices: the feasibility of a body shape index and body roundness index to identify cardiovascular health status. PLoS ONE. 2014;9:e107212. 10.1371/journal.pone.0107212. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Liu S, Yu J, Wang L, Zhang X, Wang F, Zhu Y. Weight-adjusted waist index as a practical predictor for diabetes, cardiovascular disease, and non-accidental mortality risk. Nutr Metab Cardiovasc Dis. 2024;34:2498–510. 10.1016/j.numecd.2024.06.012. [DOI] [PubMed] [Google Scholar]
- 16.Von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, Vandenbroucke JP. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. J Clin Epidemiol. 2008;61:344–9. 10.1016/j.jclinepi.2007.11.008. [DOI] [PubMed] [Google Scholar]
- 17.Zwartkruis VW, Suthahar N, Idema DL, Mahmoud B, van Deutekom C, Rutten FH, et al. Relative fat mass and prediction of incident atrial fibrillation, heart failure and coronary artery disease in the general population. Int J Obes (Lond). 2023;47:1256–62. 10.1038/s41366-023-01380-8. [DOI] [PubMed] [Google Scholar]
- 18.Kajikawa M, Maruhashi T, Kishimoto S, Yamaji T, Harada T, Saito Y, et al. A body shape index as a simple anthropometric marker of abdominal obesity and risk of cardiovascular events. J Clin Endocrinol Metab. 2024;109:3272–81. 10.1210/clinem/dgae282. [DOI] [PubMed] [Google Scholar]
- 19.Martini D, Chiavaroli L, González-Sarrías A, Bresciani L, Palma-Duran SA, Dall’Asta M, et al. Impact of foods and dietary supplements containing hydroxycinnamic acids on cardiometabolic biomarkers: a systematic review to explore inter-individual variability. Nutrients. 2019;11:1805. 10.3390/nu11081805. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Capra ME, Pederiva C, Viggiano C, De Santis R, Banderali G, Biasucci G. Nutritional approach to prevention and treatment of cardiovascular disease in childhood. Nutrients. 2021;13:2359. 10.3390/nu13072359. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Stewart RM, Clearkin LG. Insulin resistance and autoregulatory dysfunction in glaucoma and retinal vein occlusion. Am J Ophthalmol. 2008;145:394–6. 10.1016/j.ajo.2007.11.005. [DOI] [PubMed] [Google Scholar]
- 22.Tian X, Zuo Y, Chen S, Liu Q, Tao B, Wu S, et al. Triglyceride-glucose index is associated with the risk of myocardial infarction: an 11-year prospective study in the Kailuan cohort. Cardiovasc Diabetol. 2021;20:19. 10.1186/s12933-020-01210-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Lu Y, Hajifathalian K, Ezzati M, et al. Metabolic mediators of the effects of body-mass index, overweight, and obesity on coronary heart disease and stroke: a pooled analysis of 97 prospective cohorts with 1.8 million participants. Lancet. 2014;383(9921):970–83. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Sugiura T, Dohi Y, Yamashita S, Yamamoto K, Wakamatsu Y, Tanaka S, et al. Impact of lipid profile and high blood pressure on endothelial damage. J Clin Lipidol. 2011;5:460–6. 10.1016/j.jacl.2011.06.005. [DOI] [PubMed] [Google Scholar]
- 25.Dyer A, Brose U, Berti E, Rosenbaum B, Hirt MR. The travel speeds of large animals are limited by their heat-dissipation capacities. PLoS Biol. 2023;21(4):e3001820. 10.1371/journal.pbio.3001820. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.El Meouchy P, Wahoud M, Allam S, Chedid R, Karam W, Karam S. Hypertension related to obesity: pathogenesis, characteristics and factors for control. Int J Mol Sci. 2022;23:12305. 10.3390/ijms232012305. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Nemtsova V, Vischer AS, Burkard T. Hypertensive heart disease: a narrative review series-part 3: vasculature, biomarkers and the matrix of hypertensive heart disease. J Clin Med. 2024;13:505. 10.3390/jcm13020505. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Lobato NS, Filgueira FP, Akamine EH, Tostes RC, Carvalho MH, Fortes ZB. Mechanisms of endothelial dysfunction in obesity-associated hypertension. Braz J Med Biol Res. 2012;45:392–400. 10.1590/s0100-879x2012007500058. [DOI] [PMC free article] [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 data utilized in this study are publicly accessible through the China Health and Retirement Longitudinal Study (CHARLS) online repository (https://charls.pku.edu.cn/). De-identified patient data from MESA are accessible upon authorization by the National Heart, Lung, and Blood Institute’s Biologic Specimen and Data Repository Information Coordinating Center (BioLINCC) (https://biolincc.nhlbi.nih.gov/studies/mesa/).