Abstract
Background
Cardiovascular disease (CVD) remains the leading cause of mortality worldwide, posing a significant burden on public health systems. Identifying novel biomarkers to improve risk stratification is crucial for effective primary prevention. The cholesterol, high-density lipoprotein, and glucose (CHG) index and estimated pulse-wave velocity (ePWV) reflect metabolic dysregulation and vascular damage, respectively. However, their combined predictive value for CVD risk remains unexplored. This study aimed to investigate the associations of CHG index and ePWV with incident CVD and evaluate whether their combination improves risk prediction.
Methods
This prospective cohort study utilized data from the China Health and Retirement Longitudinal Study (CHARLS). Participants aged ≥ 45 years without CVD at baseline were included. CHG index and ePWV were calculated using baseline measurements of lipid profiles, fasting glucose, and blood pressure. CVD incidence was ascertained through follow-up questionnaires. Cox proportional hazards models, restricted cubic splines, and receiver operating characteristic (ROC) curves were employed to assess associations, nonlinear relationships, and predictive performance.
Results
A total of 5,431 participants were enrolled, with 1,428 incident CVD cases during follow-up. Both CHG index (HR = 1.40, 95% CI: 1.25–1.57) and ePWV (HR = 1.11, 95% CI: 1.08–1.13) were independently associated with CVD risk. Participants with both high CHG and high ePWV exhibited the highest risk (HR = 2.12, 95% CI: 1.83–2.44). Restricted cubic splines revealed a linear association for CHG index and a nonlinear association for ePWV with CVD risk. The combination of CHG index and ePWV demonstrated superior predictive performance (AUC = 0.611) compared to CHG index alone (AUC = 0.555) or ePWV alone (AUC = 0.603).
Conclusions
The CHG index and ePWV are independently and synergistically associated with CVD risk in Chinese adults aged ≥ 45 years. Their combination may serve as a novel and effective strategy for enhancing risk stratification in the primary prevention of CVD within this middle-aged and older population.
Clinical trial number
Not applicable.
Keywords: Cardiovascular disease; Cholesterol, high-density lipoprotein, and glucose index; Estimated pulse-wave velocity; Risk stratification; Primary prevention
Background
Cardiovascular disease (CVD) is a global public health issue. According to statistics, the prevalence of CVD will rise from 11.3% to 15.0% by 2050, and clinically confirmed CVD will affect 45 million adults [1]. According to data from the China Cardiovascular Health and Disease Report 2024, the incidence of CVD among Chinese residents aged 18 and above is approximately 620.33 per 100,000 population. In 2021, CVD remained the leading cause of mortality, surpassing cancer and other causes, posing a significant challenge to public health in China [2]. Primary prevention is of paramount importance in addressing CVD. Identifying high-risk populations and providing them with early lifestyle interventions or pharmacological treatments can significantly reduce the burden of CVD [3, 4]. Consequently, the identification of novel biomarkers to improve risk stratification in CVD is a critical priority.
Given this priority, the occurrence of CVD is associated with a variety of risk factors, including obesity, glucose metabolism disorders, lipid metabolism disorders, and advanced age [5–8]. Among these, metabolic dysregulation is a critical driver, leading to the investigation of composite biomarkers. The cholesterol, high-density lipoprotein, and glucose (CHG) index, which integrates parameters of glucose and lipid metabolism, has recently been implicated in the incidence and mortality of CVD [9–11]. Another research hotspot indicator, estimated pulse-wave velocity (ePWV), which integrates age and mean blood pressure (MBP), has recently been found to be closely related to CVD [12, 13], with its predictive power even surpassing that of the Framingham Risk Score [14].
Beyond individual biomarkers, contemporary frameworks for cardiovascular risk stratification have increasingly emphasized the integration of traditional and emerging metabolic risk factors. Insulin resistance (IR), a core pathophysiological driver of metabolic dysregulation, plays a pivotal role in linking lipid abnormalities, hyperglycemia, and vascular damage to adverse cardiovascular outcomes [15–17]. The dysmetabolic state induced by IR promotes atherogenic dyslipidemia, endothelial dysfunction, and chronic low-grade inflammation, thereby accelerating the initiation and progression of atherosclerosis [18, 19]. Recent comprehensive reviews have synthesized the evolving landscape of cardiovascular-metabolic risk integration, highlighting the need to incorporate novel metabolic indices alongside conventional risk factors to improve risk prediction [20, 21]. Furthermore, the mechanistic interplay between IR, lipid metabolism, and arterial stiffness has been increasingly recognized as a critical pathway underlying CVD development [22].
Despite the individual promise of the CHG index and ePWV, which reflect distinct pathological pathways, namely metabolic dysregulation and vascular damage respectively, their combined predictive value for CVD risk remains unexplored, particularly in middle-aged and elderly Chinese populations.
This study conducted a prospective cohort analysis using the China Health and Retirement Longitudinal Study (CHARLS) database to investigate the associations between the CHG index, ePWV, and the risk of incident CVD. Furthermore, it sought to validate whether the combination of these two indicators could enhance predictive capability. The findings hold significant clinical implications, as they may aid in identifying individuals at high risk for CVD, thereby enabling early intervention to reduce incidence rates and alleviate the socioeconomic burden.
Methods
Data source and study population
This cohort study utilized data from the CHARLS, a nationally representative longitudinal survey initiated in 2011 that focuses on Chinese residents aged 45 and above (http://charls.pku.edu.cn/). The survey employed a multistage stratified probability sampling strategy proportional to population size, covering 150 county-level units across 28 provinces. This survey was initiated in 2011, with follow-up assessments conducted in 2013, 2015, 2018, and 2020, resulting in a total of five waves [23].
Initially, a total of 17,708 participants from the 2011 baseline survey of the CHARLS were considered for this study. We applied a series of exclusion criteria sequentially. First, we excluded participants who had a self-reported physician diagnosis of heart disease (n = 2,130) or stroke (n = 372) at baseline. Heart disease was defined based on a positive response to the question: “Has a doctor ever told you that you have a heart attack, coronary heart disease, angina, congestive heart failure, or other heart problems?” Stroke was defined as a positive response to: “Has a doctor ever informed you that you had a stroke?” These definitions are consistent with standard epidemiological practices and have been widely used in previous CHARLS-based studies [24, 25]. Subsequently, individuals were excluded if they did not meet the fasting requirement (n = 6,116) or had missing data on key metabolic biomarkers, including blood glucose (n = 160) and lipid profiles [triglycerides, total cholesterol, high-density lipoprotein cholesterol (HDL-C), and low-density lipoprotein cholesterol (LDL-C) (n = 16)]. Furthermore, participants with incomplete blood pressure measurements (missing any of the three systolic/diastolic readings; n = 1,371) or missing age data (n = 27) were also removed. After these exclusions, 7,516 participants remained for the initial analysis. For the final longitudinal analysis, we further excluded participants who were lost to follow-up or did not participate in subsequent waves (n = 2085). Finally, a total of 5,431 participants were included in this study. Fig. 1 illustrates the study flowchart for this research.
Fig. 1.
Flowchart of participant selection from the CHARLS cohort
Ethics approval for the CHARLS was granted by the Peking University Institutional Review Board (Reference No. IRB00001052-11015), with all procedures performed in accordance with the Declaration of Helsinki. Every participant provided signed informed consent before taking part. The reporting of this work follows the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines.
Calculation formulas for exposure variables
Assays were carried out at the Youanmen Center for Clinical Laboratory, affiliated with Capital Medical University. Enzymatic colorimetric assays were employed to analyze biomarkers in archived frozen plasma or whole blood. The reliability of these measurements was underscored by the low coefficients of variation (all less than 2%) observed for triglycerides, cholesterol, HDL-C, and glucose [26]. Three separate blood pressure measurements were taken. The mean systolic blood pressure (SBP) and mean diastolic blood pressure (DBP) were calculated by averaging the three readings, respectively. The calculation formulas for this study are as follows:
Covariates
The covariates included in this study encompass sociodemographic characteristics, lifestyle factors, clinical indicators, and medical history information. Specifically, sociodemographic characteristics included age, gender, marital status, educational level, and residential area. Lifestyle factors were defined based on self-reported smoking and drinking status at baseline (2011). Smoking status was categorized as “current smoker” if the participant responded “yes” to the question: “Do you still have the habit of smoking?” (smoking was defined as having smoked more than 100 cigarettes in a lifetime). All other participants (never smokers or former smokers) were classified as non-smokers. Drinking status was categorized as “current drinker” if the participant reported drinking alcoholic beverages more than once a month in the past year. Participants who drank less than once a month or never drank were classified as non-drinkers. lifestyle factors included smoking and drinking status. Clinical indicators comprised body mass index (BMI), SBP, DBP, ePWV, and CHG index. Medical history included hypertension, diabetes, and dyslipidemia. Specifically, hypertension was defined as a positive response to the question: “Has a doctor ever told you that you have hypertension?” (CHARLS question DA007:1). Diabetes was defined as a positive response to the question: “Has a doctor ever told you that you have diabetes or high blood sugar?” (CHARLS question DA007:3). Dyslipidemia was defined as a positive response to the question: “Has a doctor ever told you that you have dyslipidemia (elevation of low density lipoprotein, triglycerides, and total cholesterol, or a low high density lipoprotein level)?” (CHARLS question DA007:2). All these variables were collected at baseline survey and were adjusted for as potential confounders in the analyses.
Definition of outcome events
The outcome of this study was the incidence of CVD, which encompassed heart disease or stroke. Diagnoses of heart disease and stroke were ascertained using standardized questionnaires adapted from previous studies. Specifically, participants were asked, “Has a doctor ever told you that you had a heart attack, coronary heart disease, angina, congestive heart failure, or other heart problems?” For stroke, the assessment was based on the question, “Has a doctor ever informed you that you had a stroke?” If a participant responded “Yes”, it was defined as the occurrence of CVD.
Statistical analysis
We used descriptive statistics to summarize the baseline characteristics of the study participants. The normality of the data was assessed using the Kolmogorov-Smirnov test. In this study, continuous variables were found to be non-normally distributed and were accordingly presented as medians (Q1, Q3), while categorical variables were expressed as counts (percentages). Between-group comparisons for categorical variables were conducted using the chi-square test, and differences in continuous variables were examined using the Kruskal-Wallis H test.
Cox proportional hazards models were employed to investigate the association between the CHG index, ePWV, and the risk of CVD. ePWV and the CHG index were included in the models as both continuous and categorical variables. CHG index and ePWV were each dichotomized into high and low groups based on their median values (CHG index: 5.26; ePWV: 12.58), and their pairwise combinations yielded four distinct categories for inclusion in the Cox regression model. We constructed two Cox proportional hazards models. Model 1 was unadjusted. Model 2 was adjusted for gender, BMI, marital status, smoking status and drinking status. The Kaplan-Meier survival curve was used to visually analyze the differences in the cumulative risk of CVD over time among the four groups formed by the combination of ePWV and the CHG index. In constructing the model, the time axis was defined as years 2, 4, 7, and 9 of the follow-up period. The vertical axis represented the cumulative risk of CVD. The log-rank test was employed to compare the differences among the groups. Restricted cubic spline (RCS) regression was employed to explore the potential non-linear relationships of ePWV and the CHG index with CVD risk. The covariates used for adjustment in the RCS analysis were identical to those in the Cox proportional hazards models. Receiver operating characteristic (ROC) curve analysis and DeLong’s test were used to assess whether the combination of ePWV and the CHG index could improve predictive ability. Net reclassification improvement (NRI) was calculated to assess whether the addition of the CHG index to ePWV improved risk classification for incident CVD. The continuous NRI was used to quantify the correctness of reclassification across all possible risk categories. R (version 4.5.0) and MedCalc (version 20.215) were used to perform statistical analyses. All tests were two-sided; P < 0.05 was considered statistically significant.
Results
Population characteristics
Based on Table 1, a total of 5,431 participants were included in this study. The median age was 57.00 (51.00, 63.00) years, and 55.81% of the participants were female. The median BMI was 23.13 (20.89, 25.71) kg/m², and the median SBP and DBP were 125.33 (113.67, 139.67) mmHg and 74.33 (66.67, 82.33) mmHg, respectively. The overall median ePWV was 12.58 (11.13, 14.45), and the median CHG index was 5.26 (5.02, 5.53). The majority of participants were married (90.79%), resided in rural areas (68.62%), and had an education level below primary school (48.41%). The proportions of current smokers and drinkers were 29.20% and 33.31%, respectively. Regarding comorbidities, the prevalence of hypertension, diabetes, and dyslipidemia was 21.21%, 4.60%, and 7.83%, respectively. During follow-up, 1,428 participants (26.29%) developed incident CVD. Significant differences were observed across all baseline characteristics among the four groups defined by ePWV and CHG index combinations (all P < 0.001).
Table 1.
Basic characteristics of the population included in this study
| Variables | Total | Low ePWV & Low CHG | High ePWV & Low CHG | Low ePWV & High CHG | High ePWV & High CHG | P |
|---|---|---|---|---|---|---|
| (n = 5431) | (n = 1803) | (n = 1167) | (n = 1262) | (n = 1199) | ||
| Age, year | 57.00 (51.00, 63.00) | 53.00 (48.00,58.00) | 64.00 (59.00,70.00) | 53.00 (48.00,57.00) | 63.00 (58.00,68.00) | < 0.001 |
| BMI, (kg/m2) | 23.13 (20.89, 25.71) | 22.25 (20.31,24.26) | 22.01 (20.08,24.45) | 24.41 (22.16,26.86) | 24.64 (22.19,27.15) | < 0.001 |
| SBP, mmHg | 125.33 (113.67, 139.67) | 114.33 (107.33,122.33) | 140.33 (128.83,152.67) | 118.33 (110.00,126.00) | 142.33 (132.67,155.00) | < 0.001 |
| DBP, mmHg | 74.33 (66.67, 82.33) | 69.33 (63.00,76.00) | 78.33 (70.67,87.00) | 72.33 (66.00,79.25) | 80.33 (73.00,88.67) | < 0.001 |
| ePWV | 12.58 (11.13, 14.45) | 11.21 (10.32,12.04) | 14.82 (13.74,16.31) | 11.45 (10.59,12.20) | 14.82 (13.73,16.43) | < 0.001 |
| CHG | 5.26 (5.02, 5.53) | 5.04 (4.88,5.17) | 5.06 (4.90,5.19) | 5.56 (5.41,5.76) | 5.58 (5.44,5.80) | < 0.001 |
| Gender | < 0.001 | |||||
| Male | 2400 (44.19) | 736 (40.82) | 575 (49.27) | 562 (44.53) | 527 (43.95) | |
| Female | 3031 (55.81) | 1067 (59.18) | 592 (50.73) | 700 (55.47) | 672 (56.05) | |
| Marital Status | < 0.001 | |||||
| Others | 500 (9.21) | 103 (5.71) | 181 (15.51) | 57 (4.52) | 159 (13.26) | |
| Married | 4931 (90.79) | 1700 (94.29) | 986 (84.49) | 1205 (95.48) | 1040 (86.74) | |
| Education level | < 0.001 | |||||
| Below primary school | 2629 (48.41) | 811 (44.98) | 670 (57.41) | 490 (38.83) | 658 (54.88) | |
| Primary school | 1186 (21.84) | 357 (19.80) | 286 (24.51) | 257 (20.36) | 286 (23.85) | |
| Middle school | 1091 (20.09) | 415 (23.02) | 155 (13.28) | 345 (27.34) | 176 (14.68) | |
| High school and above | 525 (9.67) | 220 (12.20) | 56 (4.80) | 170 (13.47) | 79 (6.59) | |
| Residential area | < 0.001 | |||||
| Urban | 1704 (31.38) | 511 (28.34) | 305 (26.14) | 462 (36.61) | 426 (35.53) | |
| Rural | 3727 (68.62) | 1292 (71.66) | 862 (73.86) | 800 (63.39) | 773 (64.47) | |
| Smoking Status | < 0.001 | |||||
| No | 3845 (70.80) | 1288 (71.44) | 773 (66.24) | 903 (71.55) | 881 (73.48) | |
| Yes | 1586 (29.20) | 515 (28.56) | 394 (33.76) | 359 (28.45) | 318 (26.52) | |
| Drinking Status | < 0.001 | |||||
| No | 3622 (66.69) | 1187 (65.83) | 748 (64.10) | 818 (64.82) | 869 (72.48) | |
| Yes | 1809 (33.31) | 616 (34.17) | 419 (35.90) | 444 (35.18) | 330 (27.52) | |
| Hypertension | < 0.001 | |||||
| No | 4279 (78.79) | 1681 (93.23) | 803 (68.81) | 1084 (85.90) | 711 (59.30) | |
| Yes | 1152 (21.21) | 122 (6.77) | 364 (31.19) | 178 (14.10) | 488 (40.70) | |
| Diabetes | < 0.001 | |||||
| No | 5181 (95.40) | 1783 (98.89) | 1142 (97.86) | 1177 (93.26) | 1079 (89.99) | |
| Yes | 250 (4.60) | 20 (1.11) | 25 (2.14) | 85 (6.74) | 120 (10.01) | |
| Dyslipidemia | < 0.001 | |||||
| No | 5006 (92.17) | 1718 (95.29) | 1099 (94.17) | 1140 (90.33) | 1049 (87.49) | |
| Yes | 425 (7.83) | 85 (4.71) | 68 (5.83) | 122 (9.67) | 150 (12.51) | |
| Incident CVD | < 0.001 | |||||
| No | 4003 (73.71) | 1470 (81.53) | 817 (70.01) | 954 (75.59) | 762 (63.55) | |
| Yes | 1428 (26.29) | 333 (18.47) | 350 (29.99) | 308 (24.41) | 437 (36.45) |
Continuous variables are presented as medians (Q1, Q3) and were compared using the Kruskal-Wallis H test. Categorical variables are presented as counts (percentages) and were compared using the Chi-square test
Abbreviation: ePWV Estimated Pulse Wave Velocity, CHG Cholesterol, High-density lipoprotein, and Glucose, BMI Body mass index, SBP Systolic blood pressure, DBP Diastolic blood pressure, CVD Cardiovascular Disease
Association of ePWV, CHG index, and their joint effect on CVD risk
Cox proportional hazards regression analysis revealed that both ePWV and the CHG index were significantly associated with the risk of CVD. In the fully adjusted model (Model 2), each unit increase in ePWV was associated with a 11% higher risk of CVD [Hazard ratio (HR) = 1.11, 95% Confidence Interval (CI): 1.08–1.13, P < 0.001]. Compared with the low ePWV group, participants in the high ePWV group had a 69% increased risk of CVD (HR = 1.69, 95% CI: 1.52–1.88, P < 0.001). For the CHG index, each unit increase was associated with a 40% higher risk of CVD (HR = 1.40, 95% CI: 1.25–1.57, P < 0.001), and individuals in the high CHG group exhibited a 34% increased risk relative to those in the low CHG group (HR = 1.34, 95% CI: 1.21–1.49, P < 0.001).
When examining the joint effects of ePWV and the CHG index, using the low ePWV and low CHG group as the reference, the fully adjusted model demonstrated that participants in the high ePWV and low CHG group had a 75% higher risk of CVD (HR = 1.75, 95% CI: 1.50–2.04, P < 0.001). Those in the low ePWV and high CHG group showed a 36% increased risk (HR = 1.36, 95% CI: 1.17–1.59, P < 0.001). Notably, individuals in the high ePWV and high CHG group experienced the highest risk, with a 112% increase in CVD incidence (HR = 2.12, 95% CI: 1.83–2.44, P < 0.001) (Table 2).
Table 2.
Association of ePWV, CHG index and combination with CVD risk
| Categories | Model 1 | Model 2 | ||||
|---|---|---|---|---|---|---|
| HR | 95% CI | P | HR | 95% CI | P | |
| ePWV | ||||||
| Per unit | 1.10 | 1.08–1.13 | < 0.001 | 1.11 | 1.08–1.13 | < 0.001 |
| Low ePWV | 1 (Reference) | 1 (Reference) | ||||
| High ePWV | 1.68 | 1.52–1.87 | < 0.001 | 1.69 | 1.52–1.88 | < 0.001 |
| CHG | ||||||
| Per unit | 1.40 | 1.25–1.58 | < 0.001 | 1.40 | 1.25–1.57 | < 0.001 |
| Low CHG | 1(Reference) | 1(Reference) | ||||
| High CHG | 1.35 | 1.22–1.50 | < 0.001 | 1.34 | 1.21–1.49 | < 0.001 |
| ePWV & CHG | ||||||
| Low ePWV& Low CHG | 1(Reference) | 1(Reference) | ||||
| High ePWV & Low CHG | 1.72 | 1.48-2.00 | < 0.001 | 1.75 | 1.50–2.04 | < 0.001 |
| Low ePWV& High CHG | 1.35 | 1.16–1.58 | < 0.001 | 1.36 | 1.17–1.59 | < 0.001 |
| High ePWV & High CHG | 2.12 | 1.84–2.45 | < 0.001 | 2.12 | 1.83–2.44 | < 0.001 |
Model 1: unadjusted
Model 2: adjusted for Gender, BMI, Marital Status, Smoking Status and Drinking Status
Abbreviation: ePWV Estimated Pulse Wave Velocity, CHG Cholesterol, High-density lipoprotein, and Glucose, BMI Body mass index, CVD Cardiovascular Disease, HR Hazard Ratio, CI Confidence Interval
Cumulative risk of CVD: analysis using Kaplan-Meier survival curves
The results of the Kaplan-Meier survival analysis further demonstrated that participants in the high CHG index and high ePWV group exhibited the highest CVD risk (log-rank test, P < 0.001) (Fig. 2).
Fig. 2.
Kaplan–Meier survival curves for cardiovascular disease (CVD) incidence according to combined categories of CHG index and ePWV. Participants were stratified into four groups based on median cutoffs of CHG index and ePWV: low CHG/low ePWV, low CHG/high ePWV, high CHG/low ePWV, and high CHG/high ePWV. The high CHG/high ePWV group exhibited the highest cumulative risk of CVD (log-rank test, P<0.001)
Nonlinearity test between CHG, ePWV, and CVD risk
RCS analysis (Fig. 3) showed that, after full adjustment, the CHG index was linearly associated with CVD risk (P for nonlinear = 0.371), while ePWV was nonlinearly associated with CVD risk (P for nonlinear < 0.001).
Fig. 3.
Restricted cubic spline (RCS) analysis of the association between CHG index, ePWV, and CVD risk. Panels A and B illustrate the dose-response relationship between the CHG index and CVD risk, while panels C and D depict the association between ePWV and CVD risk. Models were sequentially adjusted: (A, C) unadjusted; (B, D) adjusted for gender, body mass index (BMI), marital status, smoking status, and drinking status. The CHG index demonstrated a linear association with CVD risk across all models, whereas ePWV exhibited a nonlinear association
Predictive efficacy of ePWV and CHG index for CVD risk
Figure 4 presented the ROC curves and the corresponding areas under the curve (AUC) for three indicators—CHG index, ePWV, and the combined ePWV-CHG—in predicting a certain outcome. The results revealed that the combined ePWV-CHG index exhibited the highest predictive ability, with an AUC of 0.611, followed by ePWV alone (AUC = 0.603), while CHG alone showed the weakest predictive performance (AUC = 0.555). Pairwise comparisons using DeLong’s test indicated that all differences between the indicators were statistically significant. Specifically, the difference between CHG and ePWV (0.0474) was significant, and the difference between CHG and ePWV-CHG (0.0552) was also significant (P < 0.0001). Furthermore, although the absolute difference between ePWV and ePWV-CHG was small (0.0078), it remained statistically significant (P = 0.0189). The continuous NRI for adding the CHG index to ePWV was 0.1218 (95% CI: 0.0617–0.182, P < 0.001), indicating that approximately 12.18% of individuals were correctly reclassified after inclusion of the CHG index.
Fig. 4.
Receiver operating characteristic (ROC) curves for predicting CVD risk using CHG index, ePWV, and their combination
Discussion
This study, based on a prospective middle-aged and elderly Chinese cohort, provides the first systematic investigation into the combined predictive value of the CHG index and ePWV for the risk of new-onset CVD. The results confirm that both the CHG index and ePWV are independent risk factors for CVD. More importantly, a significant synergistic effect was observed in predicting CVD risk: individuals with concurrently high CHG index and high ePWV exhibited a 112% increased risk of incident CVD compared to the reference group. ROC analysis further demonstrated that the combination of these two indicators yielded significantly superior predictive efficacy compared to either indicator alone. These findings suggest that integrating the CHG index, which reflects metabolic dysregulation, with ePWV, which indicates vascular damage, may provide a more precise risk assessment strategy for the primary prevention of CVD.
These findings are consistent and complementary with a series of recently published studies. Regarding the CHG index, Mansoori et al., in a 10-year follow-up of 7,433 participants from the Mashhad Stroke and Heart Atherosclerotic Disorder (MASHAD) cohort study, first proposed that the CHG index could serve as a diagnostic marker for CVD, reporting an AUC of 0.660 for predicting CVD, with predictive capability superior to traditional metabolic indicators such as the triglyceride-glucose (TyG) index, atherogenic index of plasma (AIP), and lipid accumulation product (LAP) [27]. Subsequently, Mo et al., also conducting a cohort study based on the CHARLS database, found that the predictive value of the CHG index for CVD risk was comparable to that of the TyG index, with each unit increase in the CHG index associated with a 35% higher risk of CVD (HR = 1.35, 95%CI: 1.21–1.51) [9]. Research by Song et al. in myocardial infarction survivors further confirmed that the CHG index was significantly positively associated with all-cause mortality (HR = 1.39, 95%CI: 1.33–1.41) and cardiovascular mortality (HR = 1.42, 95%CI: 1.14–1.74), demonstrating a linear association [31]. Tian et al. found that the high cumulative CHG was associated with an increased risk of CVD and all-cause mortality, demonstrating superior predictive performance compared to both baseline CHG and other cumulative metabolic indices [11]. Wei et al. observed a U-shaped nonlinear relationship between the CHG index and mortality risk in patients with metabolic syndrome [10]. This suggests that the effect pattern of the CHG index may vary across different populations. For ePWV, Cheng et al. found that its predictive ability for all-cause and CVD mortality was even superior to that of the Framingham Risk Score [14]. A meta-analysis by Li et al. incorporated multiple studies and systematically confirmed that ePWV is significantly associated with the risk of CVD incidence and mortality [13]. In patients with heart failure with preserved ejection fraction, Xue et al. discovered that each 1 m/s increase in ePWV was associated with a 16% higher risk of all-cause mortality (HR = 1.16, 95% CI: 1.10–1.23) [32]. Utilizing the CHARLS population, Liu et al. found that ePWV mediates the relationship between physical activity and CVD risk [12]. The latest research by Chen et al., based on the staging of cardiovascular-kidney-metabolic syndrome, further revealed that ePWV trajectories are significantly correlated with CVD risk; the high ePWV trajectory group exhibited a 1.4-fold higher CVD risk compared to the low trajectory group [33]. However, previous studies have predominantly focused on the independent predictive value of single indicators. This study is the first to conduct a combined analysis of the CHG index, which reflects metabolic disturbances, and ePWV, an indicator of vascular damage. Our findings demonstrate that their combination significantly enhances the predictive ability for CVD risk, thereby filling a research gap in this field.
Regarding the clinical utility of our combined model, we acknowledge that the AUC increase from 0.603 (ePWV alone) to 0.611 (ePWV and CHG index) is statistically significant but modest in absolute terms. However, in primary prevention settings, even small gains in discrimination can translate into meaningful population-level benefits when applied to large-scale screening [34]. More importantly, our reclassification analysis showed a continuous NRI of 0.1218 (95% CI: 0.0617–0.182, P < 0.001), indicating that approximately 12.18% of individuals were correctly reclassified after adding the CHG index to ePWV. This suggests that the clinical value of our combined model lies more in refining individual risk classification than in substantially improving overall discrimination. To contextualize our model’s performance, we compared it with established risk assessment tools. The Pooled Cohort Equations (PCEs) incorporate traditional risk factors (age, sex, blood pressure, cholesterol, smoking, and diabetes) for 10-year atherosclerotic cardiovascular disease risk prediction [35, 36]. In contrast, our model achieved an AUC of 0.611 using only two simple indices (CHG index and ePWV) derived from routine clinical measurements. While our AUC is lower than that of the PCEs (0.679) [37], our model is not intended to replace comprehensive risk scores but rather to serve as a simplified screening tool in resource-limited settings or as an adjunct to existing tools. Consistent with a recent study by Wang et al. [21] demonstrating the value of integrating metabolic features into prediction models, our findings support that combining metabolic dysregulation (CHG index) with vascular damage (ePWV) enhances risk stratification. Future studies should explore whether adding the CHG index and ePWV to existing risk scores such as the PCEs provides incremental predictive value beyond traditional risk factors.
The mechanism by which the combination of the CHG index and ePWV enhances the predictive capacity for CVD risk can be elucidated from the perspective that they represent two core yet distinct pathophysiological dimensions in the pathogenesis of CVD. Specifically, the CHG index primarily reflects the “upstream root cause” of metabolic disturbance, whereas ePWV mainly embodies the “downstream consequence” of vascular injury. The CHG index integrates three metabolic parameters: total cholesterol, fasting blood glucose, and high-density lipoprotein cholesterol. Its core pathological basis is the disruption of glucose and lipid metabolism induced by IR [15]. In the state of IR, enhanced lipolysis in adipose tissue leads to elevated free fatty acids, increased hepatic synthesis of very-low-density lipoprotein, and decreased HDL-C levels, forming a typical atherogenic lipid profile [16]. Concurrently, hyperglycemia promotes endothelial dysfunction and inflammatory responses through pathways such as oxidative stress and the accumulation of advanced glycation end-products [17, 18], thereby laying the foundation for the initiation and progression of atherosclerosis. In contrast, ePWV, as a surrogate marker of arterial stiffness, primarily reflects vascular structural damage resulting from the long-term effects of age and blood pressure [19]. Increased arterial stiffness, on one hand, directly leads to elevated SBP and widened pulse pressure, thereby increasing cardiac afterload. On the other hand, it enhances the transmission of pulsatile flow to the microvasculature, potentially causing damage to target organs such as the kidneys and brain [22, 38]. Notably, metabolic abnormalities such as diabetes mellitus and chronic kidney disease can alter the relationship between blood pressure and PWV, affecting the accuracy of ePWV [39, 40]. This, in turn, indirectly corroborates the existence of an interaction between metabolic factors and arterial stiffness. Therefore, the CHG index and ePWV characterize the pathophysiological process of CVD from two distinct dimensions—namely, “metabolic origin” and “vascular consequence” respectively. The combined effect of the CHG index and ePWV observed in this study represents a clinical manifestation of this cascade reaction, thereby providing a robust biological basis for the combined application of these two indicators in CVD risk assessment.
The mechanistic basis for the synergistic effect of the CHG index and ePWV lies in their reflection of two interconnected pathophysiological dimensions: metabolic disturbance and vascular injury. Vascular calcification, a key driver of arterial stiffening, is an active process involving the phenotypic transformation of vascular smooth muscle cells into osteoblast-like cells, regulated by BMP-Smad, Wnt/β-catenin, MAPK, and NF-κB pathways, while protective factors such as SIRT1/SIRT6 and matrix Gla protein inhibit this process [41]. Metabolic abnormalities, including IR and hyperglycemia, activate these pro-calcific pathways through oxidative stress and inflammation, thereby promoting endothelial dysfunction and arterial stiffening [15–18]. Panvascular disease offers a broader framework, viewing atherosclerosis as a systemic process affecting multiple arterial beds [42]. This perspective aligns with our finding that metabolic dysregulation (CHG index) and arterial stiffness (ePWV) jointly contribute to CVD risk across the vascular tree, and supports the integration of metabolic and vascular biomarkers for personalized risk stratification [42].
This study possesses several strengths. First, to the best of our knowledge, this is the first prospective cohort study to investigate the combined predictive value of the CHG index and ePWV for CVD risk, thereby offering a novel integrated strategy for CVD risk assessment. Second, the data were derived from the nationally representative CHARLS cohort, characterized by a large sample size, long-term follow-up, and a controllable loss to follow-up rate, which enhances the generalizability of our findings. Third, the study not only evaluated the independent effects of each indicator but also revealed their synergistic effect through joint grouping analysis, with the incremental predictive value of the combined model validated using ROC curves and the DeLong test. Fourth, RCS analysis elucidated a linear relationship between the CHG index and CVD risk, as well as a non-linear relationship for ePWV, providing important reference points for interpreting these distinct indicators in clinical practice.
Nonetheless, several limitations should be acknowledged. First, CVD outcomes were based on self-reported physician diagnoses, which may introduce recall or reporting bias and cause outcome misclassification. Some participants may have misreported diagnoses, while others with undiagnosed prevalent CVD might have been incorrectly classified as event-free at baseline, and some incident cases could have been missed during follow-up. Non-differential misclassification with respect to exposure status (CHG index and ePWV) would bias hazard ratios toward the null, underestimating the true associations, whereas differential misclassification could either inflate or attenuate the effect estimates. Future studies should employ more conservative outcome definitions (e.g., self-report confirmed by medication or hospitalization records) or restrict analyses to hard endpoints such as myocardial infarction and stroke in sensitivity analyses to validate our findings. Second, the study population was restricted to middle-aged and elderly Chinese individuals aged 45 years and older; therefore, caution is warranted when extrapolating these conclusions to younger populations or other ethnic groups. Third, while various potential confounders were adjusted for in our models, the possibility of residual confounding from unmeasured factors, such as cardiovascular medications (e.g., antihypertensive agents, lipid-lowering drugs, and antiplatelet therapy) cannot be entirely excluded. These medications may substantially modulate the relationship between the CHG index, ePWV, and CVD risk. For instance, statins and antihypertensive drugs could alter lipid profiles and blood pressure levels, thereby potentially influencing the calculation of both the CHG index and ePWV, as well as their associated risk estimates. Future studies with available medication data should incorporate medication use as a covariate in sensitivity analyses to further validate the robustness of our findings. Fourth, the AUC for the CHG index combined with ePWV was only 0.611. Although this represents a significant improvement over single indicators, the overall predictive performance remains moderate. This suggests that the pathogenesis of CVD is complex, and future research should consider incorporating additional novel biomarkers or genetic information to further enhance the accuracy of predictive models.
Conclusion
In this prospective cohort study of Chinese adults aged ≥ 45 years, we demonstrated that both the CHG index and ePWV are independent risk factors for incident CVD, and their combination exhibits a significant synergistic effect in predicting CVD risk. Integrating these two indicators provides a novel strategy for enhancing risk stratification in primary prevention. However, given that both indices incorporate age as a key component, their predictive performance may vary across demographic groups. Moreover, our findings are derived exclusively from Chinese adults aged ≥ 45 years. Therefore, extrapolation to younger populations or other ethnic groups warrants caution. Future studies should validate this combined strategy in younger and diverse populations to establish its generalizability.
Acknowledgements
We express our gratitude to all participants in the CHARLS study and the project team.
Abbreviations
- CVD
Cardiovascular disease
- CHG
Cholesterol, high-density lipoprotein, and glucose
- ePWV
Estimated pulse-wave velocity
- MBP
Mean blood pressure
- CHARLS
China Health and Retirement Longitudinal Study
- HDL-C
High-density lipoprotein cholesterol
- LDL-C
Low-density lipoprotein cholesterol
- DBP
Diastolic blood pressure
- SBP
Systolic blood pressure
- TC
Total cholesterol
- FBG
Fasting blood glucose
- BMI
Body mass index
- RCS
Restricted cubic spline
- ROC
Receiver operating characteristic
- NRI
Net reclassification improvement
- HR
Hazard ratio
- CI
Confidence Interval
- AUC
Area under the curves
- IR
Insulin resistance
- TyG
Triglyceride-glucose
- AIP
Atherogenic index of plasma
- PCEs
Pooled cohort equations
Authors’ contributions
D.C. and Y.D. were responsible for the design and conceptualization of the study. X.L. and L.W. contributed to data collecting, statistical analysis, result interpretation and data visualization. D.C. and Y.D. and Y.R. read and revised the manuscript. All authors read and approved the final manuscript.
Funding
No funding.
Data availability
The datasets produced and analyzed in this research are accessible in the CHARLS database (http://charls.pku.edu.cn/).
Declarations
Ethics approval and consent to participate
The CHARLS study was conducted in line with the principles stated in the Declaration of Helsinki and received approval from the Institutional Review Board of Peking University (IRB00001052-11015). Prior to their involvement in the CHARLS study, all participants gave their written informed consent. The research adhered to the STROBE guidelines for reporting observational studies in epidemiology.
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.
Dengyong Chen and Yuting Deng contributed equally to this study.
References
- 1.Joynt Maddox KE, Elkind MSV, Aparicio HJ, et al. Forecasting the Burden of Cardiovascular Disease and Stroke in the United States Through 2050-Prevalence of Risk Factors and Disease: A Presidential Advisory From the American Heart Association. Circulation. 2024;150(4):e65–88. 10.1161/CIR.0000000000001256. [DOI] [PubMed] [Google Scholar]
- 2.Liu M, He X, Yang X et al. 2024 annual report on cardiovascular health and diseases in China: Data and trend. Chin Med J (Engl). 2025;138(23):3037-49. 10.1097/CM9.0000000000003902. [DOI] [PMC free article] [PubMed]
- 3.Pradhan A, Thandi P, Mahajan K, et al. 2025 ESC/EAS Dyslipidaemia Guidelines Focused Update: Intensifying Prevention, Risk Stratification, and Therapy. Am J Cardiol. 2026. 10.1016/j.amjcard.2026.02.059. [DOI] [PubMed] [Google Scholar]
- 4.Siegrist M, Bakker EA, Timmermans EJ, et al. The role of physical activity in an obesogenic environment for cardiovascular risk reduction across the lifespan. A Scientific Statement of the European Association of Preventive Cardiology of the ESC. Eur J Prev Cardiol. 2026. 10.1093/eurjpc/zwag033. [DOI] [PubMed] [Google Scholar]
- 5.Welsh A, Hammad M, Pina IL, et al. Obesity and cardiovascular health. Eur J Prev Cardiol. 2024;31(8):1026–35. 10.1093/eurjpc/zwae025. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Tan MY, Zhang YJ, Zhu SX, et al. The prognostic significance of stress hyperglycemia ratio in evaluating all-cause and cardiovascular mortality risk among individuals across stages 0–3 of cardiovascular-kidney-metabolic syndrome: evidence from two cohort studies. Cardiovasc Diabetol. 2025;24(1):137. 10.1186/s12933-025-02689-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Katzke VA, Sookthai D, Johnson T, et al. Blood lipids and lipoproteins in relation to incidence and mortality risks for CVD and cancer in the prospective EPIC-Heidelberg cohort. BMC Med. 2017;15(1):218. 10.1186/s12916-017-0976-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Wang C, Yuan Y, Zheng M, et al. Association of Age of Onset of Hypertension With Cardiovascular Diseases and Mortality. J Am Coll Cardiol. 2020;75(23):2921–30. 10.1016/j.jacc.2020.04.038. [DOI] [PubMed] [Google Scholar]
- 9.Mo D, Zhang P, Zhang M, et al. 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. 10.1186/s12933-025-02675-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Wei S, Yu S, Qian C, et al. Exploring the U-shaped nonlinear relationship of CHG index with metabolic syndrome and mortality risks in metabolic syndrome patients. Lipids Health Dis. 2025;24(1):294. 10.1186/s12944-025-02714-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Tian R, Tian X, Chen S, et al. Association of cumulative exposure to cholesterol, high-density lipoprotein, and glucose index with the risk of cardiovascular disease and all-cause mortality: A longitudinal cohort study. Diabetes Obes Metab. 2026;28(2):1453–62. 10.1111/dom.70338. [DOI] [PubMed] [Google Scholar]
- 12.Liu Z, Kuang J, Yan W, et al. Mediating role of arterial stiffness in the association between physical activity and cardiovascular disease risk: a prospective cohort study. Sci Rep. 2025;15(1):38740. 10.1038/s41598-025-22442-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Li J, Gao F, Cao F, et al. Association of estimated pulse wave velocity with cardiovascular disease outcomes and all-cause death-a systematic review and meta-analysis. Front Cardiovasc Med. 2025;12:1641697. 10.3389/fcvm.2025.1641697. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Cheng W, Kong F, Pan H, et al. Superior predictive value of estimated pulse wave velocity for all-cause and cardiovascular disease mortality risk in U.S. general adults. BMC Public Health. 2024;24(1):600. 10.1186/s12889-024-18071-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Tajik A, Ghayour-Mobarhan M, Darroudi S, et al. Association of four insulin resistance surrogate indices and hypertension in the MASHAD cohort study population: a cross-sectional study. Eur J Med Res. 2026. 10.1186/s40001-026-03843-w. [DOI] [PubMed] [Google Scholar]
- 16.Ginsberg HN, Huang LS. The insulin resistance syndrome: impact on lipoprotein metabolism and atherothrombosis. J Cardiovasc Risk. 2000;7(5):325–31. 10.1177/204748730000700505. [DOI] [PubMed] [Google Scholar]
- 17.Hussain A. Chronic hyperglycemia and cardiovascular dysfunction: an in-depth exploration of metabolic and cellular pathways in type 2 diabetes mellitus. Cardiovasc Diabetol Endocrinol Rep. 2025;11(1):39. 10.1186/s40842-025-00247-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Pan Y, Si J, Endothelial Metabolic Reprogramming Links Diabetes to Atherosclerosis. Diabetes Metab Syndr Obes. 2026;19:565805. 10.2147/DMSO.S565805. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Boutouyrie P, Estimating Is Not Measuring. The Lessons About Estimated Pulse Wave Velocity. J Am Heart Assoc. 2022;11(10):e025830. 10.1161/JAHA.122.025830. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Little PJ, Offermanns S, Cho H, et al. Insulin Resistance: An Update on Biochemical and Pathophysiological Mechanisms and Impact on Various Diseases. iNew Med. 2026. 10.1002/inm3.70028. n/a(n/a):e70028. [Google Scholar]
- 21.Wang Z, Chen S, Xu S, et al. Risk Factors for Cardiovascular and Metabolic Disease: Integrating Traditional and Novel Paradigms. iNew Med. 2025;1(1):e70000. 10.1002/inm3.70000. [Google Scholar]
- 22.Climie RE, Gallo A, Picone DS, et al. Measuring the Interaction Between the Macro- and Micro-Vasculature. Front Cardiovasc Med. 2019;6:169. 10.3389/fcvm.2019.00169. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Zhao Y, Hu Y, Smith JP, et al. Cohort profile: the China Health and Retirement Longitudinal Study (CHARLS). Int J Epidemiol. 2014;43(1):61–8. 10.1093/ije/dys203. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Zhang Y, Liu B, Zhu Y, et al. Associations of cumulative exposure and dynamic trajectories of cholesterol-HDL-glucose (CHG) index with cardiovascular disease in middle-aged and older Chinese adults: a longitudinal analysis. Cardiovasc Diabetol. 2026;25(1). 10.1186/s12933-026-03137-9. [DOI] [PMC free article] [PubMed]
- 25.Ma X, Ma X, Wang Y, et al. Associations of cumulative exposure and dynamic trajectories of the C-reactive protein-triglyceride-glucose index with incident cardiovascular disease in middle-aged and older Chinese adults: a nationwide cohort study. Cardiovasc Diabetol. 2025;24(1):303. 10.1186/s12933-025-02869-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Chen X, Crimmins E, Hu PP, et al. Venous Blood-Based Biomarkers in the China Health and Retirement Longitudinal Study: Rationale, Design, and Results From the 2015 Wave. Am J Epidemiol. 2019;188(11):1871–7. 10.1093/aje/kwz170. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Mansoori A, Nosrati M, Dorchin M, et al. A novel index for diagnosis of type 2 diabetes mellitus: Cholesterol, High density lipoprotein, and Glucose (CHG) index. J Diabetes Investig. 2025;16(2):309–14. 10.1111/jdi.14343. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Sainas G, Milia R, Palazzolo G, et al. Mean Blood Pressure Assessment during Post-Exercise: Result from Two Different Methods of Calculation. J Sports Sci Med. 2016;15(3):424–33. [PMC free article] [PubMed] [Google Scholar]
- 29.Bao W, Chen C, Chen C, et al. Association between estimated pulse wave velocity and risk of diabetes: A large sample size cohort study. Nutr Metab Cardiovasc Dis. 2023;33(9):1716–24. 10.1016/j.numecd.2023.05.032. [DOI] [PubMed] [Google Scholar]
- 30.Xiao Q, Xue A, Huang W, et al. Triglyceride-glucose index combined with estimated pulse wave velocity improves stroke risk stratification in the CHARLS cohort. Sci Rep. 2025;15(1):32639. 10.1038/s41598-025-18668-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Song Y, Chen X, Chang Z, et al. The cholesterol, high-density lipoprotein, and glucose (CHG) index as a novel metabolic marker for predicting adverse outcomes in myocardial infarction survivors: insights from two large prospective cohorts. Cardiovasc Diabetol. 2026;25(1). 10.1186/s12933-026-03104-4. [DOI] [PMC free article] [PubMed]
- 32.Xue R, Zhang J, Zhen Z, et al. Estimated pulse wave velocity predicts mortality in patients with heart failure with preserved ejection fraction. Hellenic J Cardiol. 2025;86:51–62. 10.1016/j.hjc.2024.05.013. [DOI] [PubMed] [Google Scholar]
- 33.Chen T, Yin H, Zhou Y, et al. Relationship between estimated pulse wave velocity trajectories and cardiovascular disease risk in patients with cardiovascular-kidney-metabolic syndrome stages 0–3. Nutr Metab Cardiovasc Dis. 2025;35(11):104192. 10.1016/j.numecd.2025.104192. [DOI] [PubMed] [Google Scholar]
- 34.Pencina MJ, D’Agostino RB, Sr. Steyerberg EW Extensions of net reclassification improvement calculations to measure usefulness of new biomarkers. Stat Med. 2011;30(1):11–21. 10.1002/sim.4085. [DOI] [PMC free article] [PubMed]
- 35.Andrus B, Lacaille D. 2013 ACC/AHA guideline on the assessment of cardiovascular risk. J Am Coll Cardiol. 2014; 63(25 Pt A):2886. 10.1016/j.jacc.2014.02.606. [DOI] [PubMed]
- 36.Yadlowsky S, Hayward RA, Sussman JB, et al. Clinical Implications of Revised Pooled Cohort Equations for Estimating Atherosclerotic Cardiovascular Disease Risk. Ann Intern Med. 2018;169(1):20–9. 10.7326/M17-3011. [DOI] [PubMed] [Google Scholar]
- 37.Campos-Staffico AM, Cordwin D, Murthy VL, et al. Comparative performance of the two pooled cohort equations for predicting atherosclerotic cardiovascular disease. Atherosclerosis. 2021. 10.1016/j.atherosclerosis.2021.08.034. 334:23 – 9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Herzog MJ, Muller P, Lechner K, et al. Arterial stiffness and vascular aging: mechanisms, prevention, and therapy. Signal Transduct Target Ther. 2025;10(1):282. 10.1038/s41392-025-02346-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Cote N, Fortier C, Jafari K, et al. Estimated versus measured aortic stiffness: implications of diabetes, chronic kidney disease, sex and height. J Hypertens. 2024;42(12):2115–21. 10.1097/HJH.0000000000003845. [DOI] [PubMed] [Google Scholar]
- 40.Li W, Xu R, Yan H, et al. Nonlinear relationship between estimated pulse wave velocity and chronic kidney disease: analyses of NHANES 1999–2020. Front Med (Lausanne). 2025;12:1560272. 10.3389/fmed.2025.1560272. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Zhu W, Yang S, Chen Z, et al. Vascular Calcification: Mechanisms, Models, and Therapies. iNew Med. 2026;n/a(n/a):e70035. 10.1002/inm3.70035. [Google Scholar]
- 42.Li Z, Dai Y, Ge J, Panvascular Disease. Opportunities and Challenges. iNew Med. 2025;1(1):e70001. 10.1002/inm3.70001. [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The datasets produced and analyzed in this research are accessible in the CHARLS database (http://charls.pku.edu.cn/).