Abstract
Background
Stroke is one of the advanced outcomes of cardiovascular kidney metabolic syndrome (CKM). The combined effects of inflammation, insulin resistance, and hypertension in stroke remain to be fully elucidated in population studies. This study investigates the association between the composite TyG inflammation index (c-reactive protein-triglyceride glucose index, CTI) and hypertension with long-term stroke.
Methods
This longitudinal cohort study included 9082 participants from the China Health and Retirement Longitudinal Study (CHARLS). The associations with stroke risk were assessed using Cox proportional hazards models and Kaplan–Meier analysis, while predictive performance was evaluated using ROC curves and time-dependent AUC. The contribution of each component was determined by Weighted Quantile Sum (WQS) regression. Subgroup analysis and sensitivity analysis were also conducted.
Results
Over a median 9-year follow-up, 811 (8.9%) incident stroke cases occurred. A significant dose–response relationship was observed, with the highest CTI-hypertension group exhibiting a fully adjusted hazard ratio of 3.19 (95% CI 2.62–3.88) for stroke compared to the lowest group. The combined CTI-hypertension model demonstrated superior predictive performance (AUC = 0.672) versus hypertension-alone (AUC = 0.661) or CTI-alone (AUC = 0.651) models. WQS analysis identified C-reactive protein (38.8%) and hypertension (31.7%) as the predominant risk factors.
Conclusion
Integrating CTI with hypertension significantly improves stroke risk stratification in CKM stage 0–3 populations, supporting its potential for early identification of high-risk individuals.
Graphical abstract
Supplementary Information
The online version contains supplementary material available at 10.1186/s12933-026-03075-6.
Keywords: Cardiovascular kidney metabolic (CKM) syndrome, C-reactive protein triglyceride glucose index (CTI), Hypertension, Long term stroke, CHARLS
Research Insights
What is currently known about this topic?
Stroke is a major advanced outcome in the progression of the Cardiovascular-Kidney-Metabolic (CKM) syndrome.
Individually, systemic inflammation, insulin resistance, and hypertension are established risk factors for stroke.
However, the combined prognostic impact and potential synergistic effect of these metabolic-inflammatory components with hypertension on long-term stroke risk in the general CKM population remain inadequately quantified in large-scale prospective studies.
What is the key research question?
Does the integration of a composite index reflecting inflammation and insulin resistance (CTI) with hypertension status provide superior risk stratification for incident stroke compared to either factor alone in a CKM stage 0-3 population, and what are the predominant contributing components?
What is new?
This large, prospective cohort study establishes a strong, dose–response relationship between the combined CTI-hypertension status and long-term stroke risk, with the highest-risk group having over a threefold increased hazard.
It demonstrates that a model combining CTI and hypertension offers significantly improved predictive performance for stroke compared to models using hypertension or the metabolic-inflammatory index alone.
The analysis identifies C-reactive protein (inflammation) and hypertension as the two foremost weighted contributors to the composite stroke risk, providing mechanistic insight into the driving factors behind the observed association.
How might this study influence clinical practice?
The findings support a practical, integrative approach to stroke risk assessment in the broad CKM population. Moving beyond assessing hypertension in isolation, the combined evaluation of readily available clinical markers of inflammation (CRP) and insulin resistance (TyG index) could enhance early identification of individuals at substantially elevated risk. This may facilitate more targeted and intensive preventive strategies in clinical practice.
Introduction
Cardiovascular-kidney-metabolic (CKM) syndrome was proposed by the American Heart Association (AHA) in 2023 [1, 2]. It emphasizes the interconnections between obesity, metabolic disorders, cardiovascular disease, and chronic kidney disease. CKM syndrome is categorized into five stages (0–4), with progression typically driven by multiple factors including hypertension, metabolic abnormalities, and insulin resistance. CKM syndrome poses a significant global burden, with 90% of U.S. adults meeting the criteria for stage 1 or higher [3]. The final stage (stage 4) is characterized by clinical cardiovascular events, including stroke.
Stroke was a major contributor to CKM syndrome burden in 2025 [4, 5]. Stroke is also the second leading cause of death and the third leading cause of disability globally among non-communicable diseases (NCDs) [5, 6]. A recent study in The Lancet highlights an urgent need to enhance primary prevention strategies for stroke and cardiovascular diseases [7, 8]. Therefore, identifying reliable and clinically useful risk indicators in individuals at CKM stages 0–3 who have not yet developed overt cardiovascular disease is crucial for early detection and prevention.
The c-reactive protein-triglyceride glucose index (CTI) is a novel index based on the combination of CRP (inflammatory marker) and the TyG index [insulin resistance (IR) marker] [9, 10]. It simultaneously reflects inflammatory and metabolic status, demonstrating a strong ability to predict cancer mortality and cardiovascular events in the general population [11, 12]. Previous studies have shown that elevated CTI levels are significantly associated with higher stroke prevalence across populations and its predictive performance is better than that of any single indicator [13, 14].
Hypertension is also one of the most important risk factors for stroke. Patients with hypertension often exhibit chronic inflammation and metabolic disorders [15, 16], creating a vicious cycle that synergistically increases stroke risk. Although the individual predictive value of CTI and hypertension for stroke has been established, no prospective study has yet evaluated their combined predictive power.
To bridge this gap, we used nationally representative data from the China Health and Retirement Longitudinal Study (CHARLS) to investigate the relationship between inflammatory-metabolic status and blood pressure regulation with stroke. We explore the joint effect of CTI and hypertension (HTN) on long-term stroke within the framework of CKM stages 0–3. This study aims to provide a scientific basis for the early identification of high-risk CKM stages 0–3 populations and to inform more precise and effective stroke prevention strategies.
Methods
Data source and study population
The China Health and Retirement Longitudinal Study (CHARLS) is a nationally representative, population-based prospective cohort study targeting middle-aged and older adults, designed to systematically track multidimensional health dynamics among community-dwelling Chinese residents. The baseline survey was conducted in 2011, enrolling 17,708 participants from 10,257 households across 150 counties and 450 villages (https://charls.pku.edu.cn/en/) [17]. The study adhered to the principles of the Declaration of Helsinki. The protocol for the CHARLS cohort was approved by the Peking University Institutional Review Board (IRB00001052-1015), and all participants provided written informed consent at the time of enrollment. This study defined participants who completed the baseline survey between 2011 and 2012 as the starting point of the cohort, with follow-ups conducted in 2013, 2015, 2018, and 2020. The study population consisted of individuals aged ≥ 45 years at baseline who were in CKM stages 0–3 and had complete follow-up records. A total of 9,082 participants were ultimately included in the analysis (Fig. 1).
Fig. 1.
Flow chart for participants’ selection. 17,708 participants from the 2011 CHARLS cohort were initially included. After excluding individuals with incomplete data, missing biomarkers, CKM Stage 4, or follow-up loss, 9,082 participants were included in the final analysis
Definition of CKM stages
According to the American Heart Association’s scientific statement, the CKM stages (assessed at the baseline wave) were defined as follows [1].
Stage 0: Absence of all CKM syndrome risk factors.
Stage 1: Overweight/obesity, abdominal obesity, or dysfunctional adipose tissue (manifested as prediabetes), in the absence of other metabolic risk factors or CKD.
Stage 2: Presence of metabolic risk factors (e.g., hypertriglyceridemia, hypertension, metabolic syndrome, diabetes) and/or moderate- to high-risk CKD.
Stage 3: Subclinical cardiovascular disease. In this study, the operational definition of CKM Stage 3 subclinical cardiovascular disease was as follows: a 10-year cardiovascular risk ≥ 20% as calculated by the PREVENT equation, or the presence of CKD G4/G5 [18].
Stage 4: Clinical cardiovascular disease, identified based on self-reported CVD history.
Data collection
Body measurements were obtained using the following methods[17]: height was measured with a Seca™ 213 stadiometer, weight was determined using an Omron™ HN-286 scale, blood pressure was assessed with an Omron™ HEM-7200 monitor, and waist circumference was measured using a soft measuring tape. Age, physician-diagnosed chronic diseases, lifestyle, and health-related behaviors (smoking, alcohol consumption, physical activity) were collected through questionnaires.
Laboratory Assessments and Anthropometric Measurements [19]: After an overnight fast, venous blood samples were collected from participants in the morning. The samples were promptly aliquoted, transported to the Chinese Center for Disease Control and Prevention, and stored at −70 °C for subsequent analysis. Metabolic parameters were measured using standardized laboratory procedures, including: Lipid profile: total cholesterol (TC), triglycerides (TG), high-density lipoprotein cholesterol (HDL-C), and low-density lipoprotein cholesterol (LDL-C); Glucose homeostasis markers: fasting plasma glucose (FPG) and glycated hemoglobin (HbA1c); Kidney function indicators: blood urea nitrogen (BUN), uric acid (UA), cystatin C (CysC), and serum creatinine (Scr); Inflammatory biomarker: high-sensitivity C-reactive protein (CRP). Except for CRP, which is reported in mg/L, all other blood parameters are expressed in mg/dL.
Assessment of baseline exposure variables
The C-reactive protein-triglyceride glucose index (CTI) was calculated using the following formula [20]: CTI = 0.412 × Ln (CRP [mg/L]) + Ln (TG [mg/dL] × FPG [mg/dL]/2).
Hypertension was defined as systolic blood pressure ≥ 140 mmHg and/or diastolic blood pressure ≥ 90 mmHg, and/or a self-reported history of hypertension diagnosis, and/or current use of anti-hypertensive medications [21].
Outcome definition
The primary outcome of this study was the occurrence of a stroke event. Stroke was assessed based on the response to a key question: “Have you ever been diagnosed with stroke by a physician?” Participants who answered “yes” were classified as having experienced a stroke. All participants were followed from 2011 onward, with interviews conducted in each survey round until either the occurrence of a stroke or the end of the follow-up period in 2020 [22].
Statistical analysis
Continuous variables following a normal distribution were described using mean ± standard deviation, while those not conforming to normality were summarized as median (interquartile range). Categorical variables were expressed as frequency (n) and percentage (%). Group comparisons of baseline characteristics were performed using the chi-square test (χ2) for categorical variables, analysis of variance (ANOVA) for normally distributed continuous variables, and the Kruskal–Wallis test for non-normally distributed variables.
Given the high-dimensional nature of the dataset, which encompassed extensive demographic, clinical, and laboratory indicators, a two-stage machine learning feature selection strategy was employed to identify the most relevant predictors associated with stroke events. First, the Boruta algorithm based on random forests was applied for comprehensive feature screening, whereby features were classified as “confirmed”, “tentative”, or “rejected” by comparing the importance of original features with shadow features [23]. Subsequently, variables confirmed by Boruta underwent least absolute shrinkage and selection operator (LASSO) regression analysis with fivefold cross-validation (Supplementary Fig. 1a and b). By applying L1 regularization, the coefficients of redundant variables were shrunk to zero, yielding a robust set of predictive features [24]. Integrating these two machine learning algorithms identified hypertension and the CTI as the most critical predictors of stroke (Supplementary Fig. 1c and d), which were therefore included as core exposure variables in the final model.
Participants were categorized into four classes based on the CTI threshold (break point = 8.65, determined via restricted cubic spline (RCS) analysis shown in Supplementary Fig. 2) and the presence of hypertension (HTN): Class 1: CTI < 8.65 without HTN (reference); Class 2: CTI ≥ 8.65 without HTN; Class 3: CTI < 8.65 with HTN; Class 4: CTI ≥ 8.65 with HTN. Stroke incidence was assessed using Kaplan–Meier curves and the log-rank test. To account for death during follow-up as a competing risk, a competing risk model was constructed, and group comparisons were performed using cumulative incidence function (CIF) curves and Gray’s test. The association between the CTI-HTN combined groups and stroke incidence was examined using Cox regression models. Three Cox regression models were developed: Model 1 was unadjusted; Model 2 was adjusted for age and sex; and Model 3 was further adjusted for marital status, physical activity, educational level, residence, diabetes, dyslipidemia, smoking, alcohol consumption, and BMI. Multicollinearity diagnostics confirmed all variance inflation factors (VIF) remained below 5, indicating that the interdependencies among covariates were negligible (Supplementary Table 1). The proportional hazards assumption for the Cox models was tested using Schoenfeld residuals. The global test for all models was non-significant (P > 0.05), indicating that the assumption was not violated (Supplementary Table 2 and Supplementary Fig. 3).
To further investigate whether the relationship between CTI and stroke risk was modified by hypertension status, Cox proportional hazards models incorporating RCS curves were used, and the statistical significance of the interaction term was assessed using the likelihood ratio test. Additionally, from a public health perspective, additive interactions were quantified by calculating the relative excess risk due to interaction (RERI), the attributable proportion due to interaction (AP), and the synergy index (SI), which evaluate the excess risk due to co-exposure, the proportion of risk attributable to interaction in individuals with both exposures, and the magnitude of synergistic effects, respectively.
The predictive performance of CTI, hypertension, and their combination for stroke occurrence was evaluated by plotting receiver operating characteristic (ROC) curves and calculating the area under the curve (AUC). Model calibration was also assessed using calibration plots, which visualize the agreement between predicted probabilities and observed event frequencies. The ideal line of perfect fit (45-degree line) was used as a reference. To obtain robust, overfitting-corrected estimates of model performance and evaluate the stability of the calibration curves, we performed an internal validation using the bootstrap method with 200 resamples from the original dataset. The calibration intercept (a value of 0 indicates perfect calibration-in-the-large) and calibration slope (a value of 1 indicates perfect calibration slope) were calculated and reported for each model after this bootstrap correction. Furthermore, to comprehensively quantify the improvement in risk prediction, we calculated the Continuous Net Reclassification Index (NRI) and the Integrated Discrimination Improvement (IDI). Statistical significance for the IDI and NRI was determined by examining their 95% confidence intervals, derived from 1000 bootstrap replicates; intervals that did not include zero were considered indicative of a statistically significant improvement at the α = 0.05 level.
Weighted quantile sum (WQS) regression was applied to assess the contribution weights of components within the CTI index (CRP, TG, FPG) and the presence of hypertension to stroke risk. Subgroup analysis was conducted to examine whether the association between CTI-HTN combined groups and stroke occurrence differed across various covariate strata, and multiple sensitivity analyzes were performed to verify the robustness of the primary findings. A two-sided P-value < 0.05 was considered statistically significant. All analyzes were performed using R statistical software (version 4.5.1).
Results
Baseline characteristics of the participants
A total of 9082 participants were analyzed, including 4743 males and 4339 females, with a mean age of 58.74 ± 9.25 years. Table 1 showed the demographic and clinical characteristics of the participants categorized by CTI index and hypertension status. Compared to Class 1, Classes 2, 3, and 4 showed significantly higher values in age, SBP, DBP, FPG, TC, TG, and LDL-C, along with lower eGFR and HDL-C levels, and a higher burden of CKM syndrome (all P < 0.001). During the median follow-up period of 9 years (approximately 110 months), 811 participants (8.9%) experienced stroke events, with 99.22 cases per 10,000 person-years. The population with stroke exhibited significantly higher CTI values (8.96 ± 0.88 vs. 8.71 ± 0.84, P < 0.001) and a higher prevalence of hypertension (40.3% vs. 20.7%, P < 0.001) compared to the non-stroke population (Supplementary Fig. 4). Supplementary Table 3 presents the demographic and baseline clinical characteristics of the study population stratified by CKM) stages 0–3 at baseline.
Table 1.
Subjects demographics and baseline characteristics in 2011
| Characteristic | Combined grouping of CTI and hypertension | P value | |||
|---|---|---|---|---|---|
| Class 1 N = 3805 |
Class 2 N = 3234 |
Class 3 N = 755 |
Class 4 N = 1288 |
||
| Age, year | 58 ± 9 | 59 ± 9 | 61 ± 9 | 61 ± 9 | < 0.001 |
| Gender (male, %) | 1891 (49.7%) | 1718 (53.1%) | 395 (52.3%) | 739 (57.4%) | < 0.001 |
| Education level (n, %) | 0.113 | ||||
| Below primary school | 1783 (46.9%) | 1519 (47.0%) | 394 (52.2%) | 624 (48.4%) | |
| Primary or Middle school | 825 (21.7%) | 663 (20.5%) | 156 (20.7%) | 288 (22.4%) | |
| High school | 785 (20.6%) | 693 (21.4%) | 136 (18.0%) | 241 (18.7%) | |
| College or above | 412 (10.8%) | 359 (11.1%) | 69 (9.1%) | 135 (10.5%) | |
| Regular exercise, n (%) | 2568 (67.5%) | 2294 (70.9%) | 529 (70.1%) | 935 (72.6%) | < 0.001 |
| Marital status (married, %) | 3413 (89.7%) | 2862 (88.5%) | 648 (85.8%) | 1129 (87.7%) | 0.009 |
| Living in rural (n, %) | 2656 (69.8%) | 1990 (61.5%) | 510 (67.5%) | 778 (60.4%) | < 0.001 |
| Alcohol consumption (n, %) | 1559 (41.0%) | 1244 (38.5%) | 325 (43.0%) | 486 (37.7%) | 0.017 |
| Current smoking (n, %) | 2,246 (59.0%) | 1990 (61.5%) | 430 (57.0%) | 802 (62.3%) | 0.003 |
| CKM stage | < 0.001 | ||||
| Stage 0 | 1010 (26.5%) | 347 (10.7%) | 97 (12.8%) | 50 (3.9%) | |
| Stage 1 | 997 (26.2%) | 269 (8.3%) | 92 (12.2%) | 44 (3.4%) | |
| Stage 2 | 1311 (34.5%) | 1947 (60.2%) | 340 (45.0%) | 635 (49.3%) | |
| Stage 3 | 487 (12.8%) | 671 (20.7%) | 226 (29.9%) | 559 (43.4%) | |
| SBP, mmHg | 121 ± 17 | 127 ± 18 | 141 ± 22 | 143 ± 21 | < 0.001 |
| DBP, mmHg | 72 ± 10 | 75 ± 11 | 80 ± 13 | 81 ± 12 | < 0.001 |
| Pulse, beats/min | 71 ± 10 | 73 ± 9 | 71 ± 10 | 73 ± 10 | < 0.001 |
| Height, metre | 1.58 ± 0.09 | 1.58 ± 0.09 | 1.57 ± 0.10 | 1.58 ± 0.09 | 0.095 |
| Weight, Kg | 56 ± 10 | 60 ± 11 | 58 ± 11 | 64 ± 11 | < 0.001 |
| Waist, cm | 81 ± 10 | 85 ± 12 | 84 ± 11 | 90 ± 12 | < 0.001 |
| BMI, kg/m2 | 22.9 ± 14.2 | 24.7 ± 42.9 | 24.2 ± 19.5 | 26.1 ± 13.8 | 0.002 |
| BUN, mg/dL | 15.3 (12.7, 18.5) | 14.9 (12.3, 17.9) | 15.6 (12.6, 18.9) | 15.0 (12.7, 18.1) | < 0.001 |
| Scr, mg/dL | 0.75 (0.64, 0.86) | 0.76 (0.64, 0.88) | 0.76 (0.66, 0.89) | 0.78 (0.67, 0.93) | < 0.001 |
| TC, mg/dL | 186 ± 34 | 199 ± 41 | 187 ± 34 | 204 ± 40 | < 0.001 |
| TG, mg/dL | 83 ± 29 | 177 ± 124 | 85 ± 29 | 196 ± 158 | < 0.001 |
| HDL-C, mg/dL | 57 ± 15 | 46 ± 14 | 57 ± 14 | 44 ± 13 | < 0.001 |
| LDL-C, mg/dL | 114 ± 31 | 116 ± 38 | 115 ± 31 | 120 ± 39 | < 0.001 |
| FPG, mg/dL | 98 (91, 105) | 106 (98, 121) | 99 (92, 107) | 109 (100, 126) | < 0.001 |
| HbA1c(%) | 5.10 (4.80, 5.30) | 5.20 (4.90, 5.50) | 5.00 (4.80, 5.30) | 5.20 (5.00, 5.60) | < 0.001 |
| CRP, mg/L | 0.60 (0.39, 1.01) | 1.72 (0.93, 3.61) | 0.67 (0.42, 1.11) | 1.88 (1.05, 3.77) | < 0.001 |
| eGFR, mL/min/1.73m2 | 103 (89, 120) | 100 (84, 116) | 97 (83, 114) | 94 (79, 107) | < 0.001 |
| Long-term stroke (n, %) | 204 (5.4%) | 280 (8.7%) | 107 (14.2%) | 220 (17.1%) | < 0.001 |
The Wilcoxon rank sum test or Pearson's Chi-squared test is used to test non-normally distributed variables. Non-normally distributed variables are expressed as median (interquartile range). All other values are expressed as mean or SD. SBP, systolic blood pressure; DBP, diastolic blood pressure; BMI: body mass index; BUN, blood urea nitrogen; TC, total cholesterol; TG, triglyceride; HDL-C, high density lipoprotein cholesterol; LDL-C, low density lipoprotein cholesterol; FPG, fasting plasma glucose; HbA1c, glycosylated hemoglobin, type A1c; CRP, c-reactive protein; eGFR, estimated glomerular filtration rate
Association between CTI-HTN combined groups and long-term stroke
To explore the associations between CTI, HTN and CTI-HTN combined groups with stroke incidence, three Cox proportional risk regression models were developed, as detailed in Table 2. After multiple adjustments, hypertension and CTI were respectively associated with a significantly increased risk of long-term stroke incidence, with the hazard ratio (HR) and 95% confidence interval (CI) were 2.17 (1.88–2.52) and 1.25 (1.17, 1.33) (all P < 0.001). A gradient increase in stroke risk was observed across the CTI-HTN combined groups. After fully adjusted (Model 3), the adjusted HR and 95% CI were 1.62 (1.35–1.94) for Class 2, 2.59 (2.04–3.27) for Class 3, and 3.19 (2.62–3.88) for Class 4 (all P < 0.001) compared with Class 1, respectively. A significant trend was observed across all models (P for trend < 0.001). Kaplan–Meier survival analysis further revealed that hypertension status, higher grades of CTI groups and CTI-HTN combined groups were significantly associated with an increased incidence of stroke (log-rank test, P < 0.001) (Fig. 2a–c).
Table 2.
Association between CTI and hypertension with long-term stroke
| Characteristic | Model 1 | Model 2 | Model 3 | |||
|---|---|---|---|---|---|---|
| HR (95% CI) | P value | HR(95% CI) | P value | HR (95% CI) | P value | |
| CTI | 1.29 (1.22, 1.38) | < 0.001 | 1.29 (1.21, 1.38) | < 0.001 | 1.25 (1.17, 1.33) | < 0.001 |
| Hypertension | 2.46 (2.14, 2.84) | < 0.001 | 2.33 (2.02, 2.69) | < 0.001 | 2.17 (1.88, 2.52) | < 0.001 |
| CTI-HTN combined groups | ||||||
| Class 1 | – | – | – | – | – | – |
| Class 2 | 1.65 (1.38, 1.98) | < 0.001 | 1.63 (1.36, 1.95) | < 0.001 | 1.62 (1.35, 1.94) | < 0.001 |
| Class 3 | 2.80 (2.21, 3.54) | < 0.001 | 2.62 (2.07, 3.31) | < 0.001 | 2.59 (2.04, 3.27) | < 0.001 |
| Class 4 | 3.43 (2.83, 4.14) | < 0.001 | 3.25 (2.68, 3.94) | < 0.001 | 3.19 (2.62, 3.88) | < 0.001 |
| P for trend | < 0.001 | < 0.001 | < 0.001 | |||
HR = hazard ratio, CI = confidence interval. Model 1: unadjusted; Model 2: adjusted for age and gender; Model 3: further adjusted for marital status, education level, residence, history of hypertension, diabetes, smoking status, alcohol consumption, and BMI
Fig. 2.
Association between the CTI and hypertension with long-term stroke in CKM stages 0–3. a Association between the CTI—HTN combined groups with long-term stroke in CKM stages 0–3. b Association between hypertension with long-term stroke in CKM stages 0–3. c Association between the CTI groups with long-term stroke in CKM stages 0–3. Adjustment for age, gender, marital status, education level, SBP current smoking, alcohol consumption and BMI
We further divided the population into CKM stages 0–2 and stage 3 and also found that the combined CTI-HTN group (Class 4) was closely associated with an increased incidence of long-term stroke. The adjusted HR was 3.17 (95% CI 2.47–4.07) in the CKM stages 0–2 and 2.01 (95% CI 1.38–2.93) in the CKM stage 3 (Supplementary Fig. 5).
The interaction between CTI and HTN on stroke
The RCS plot based on Cox regression showed that the multiplicative interaction between CTI and HTN on stroke did not reach statistical significance across various adjusted models (P for interaction = 0.0799 in the fully adjusted model (Fig. 3). Supplementary Table 4 presents measures of interaction assessed on the additive scale. The relative excess risk due to interaction (RERI) was −0.126 (95% CI −0.796, 0.545), the attributable proportion due to interaction (AP) was −0.042 (95% CI −0.269, 0.185), and the synergy index (SI) was 0.941 (95% CI 0.637, 1.244). This indicated that no statistically significant additive interaction was observed.
Fig. 3.
The interaction between hypertension and CTI on stroke. a Unadjusted model; b Adjustment for age and gender; c Multiple adjustments
Predictive value of CTI-HTN combined groups on stroke
ROC curve analysis was performed to evaluate the predictive performance of each model for stroke events (Fig. 4a). The model combining CTI as a continuous variable with hypertension achieved an AUC of 0.672, which was numerically higher than that of the hypertension-only model (AUC = 0.661), the CTI-only model (AUC = 0.651), and the base model (AUC = 0.638). DeLong’s test indicated that the combined model had a statistically significantly higher discriminatory ability than each single model and the base model (all P < 0.05, Supplementary Table 5). In addition, time-dependent AUC (tAUC) curves revealed that the full model (CTI-HTN combined group) consistently exhibited better predictive capability than the base model at all evaluated time points, with a mean AUC improvement of approximately 0.034 (base model tAUC: 0.670 ± 0.028 vs. full model tAUC: 0.704 ± 0.022) (Fig. 4b).
Fig. 4.
Predictive Value of CTI and Hypertension for Stroke Risk. a ROC curves of prediction models for stroke; b Time-dependent AUC (tAUC) of the base and full model over follow-up
We also conducted a comprehensive evaluation of the model's performance. The calibration curve was shown in Supplementary Fig. 6 and calibration intercept, calibration slope and brier score were listed in Supplementary Table 6, all models showed well-calibrated predictions. The full model (Base + CTI + Hypertension) demonstrated superior calibration, with relatively better intercept and slope values (closer to 0 and 1) compared to other models. Furthermore, CTI offered a measurable, though often smaller, incremental contribution even after accounting for hypertension, as supported by positive IDI estimates in most comparisons. Conversely, hypertension also added significant predictive value beyond a model containing only CTI, particularly at the longer follow-up interval (Supplementary Table 7). The Decision Curve Analysis shown in Supplementary Fig. 7, which indicated that the combined model provided a consistently higher net benefit than the other models across most reasonable risk thresholds during the 4-, 7-, and 9-year follow-up periods. In addition, the CTI demonstrated a statistically similar, though numerically slightly higher, discriminatory ability for stroke risk compared to other advanced indices like TyG-WHtR in the comparative analysis (Supplementary Fig. 8).
WQS regression analysis
WQS regression model was applied to assess the contribution weights of the components of the CTI index (CRP, TG, FPG) and hypertension status to stroke risk. Consistency of the results across different quantile levels (q) was verified using 1000 bootstrap iterations. The WQS regression analysis results with q = 10 indicated that CRP had the highest contribution weight (38.8%), followed by hypertension (31.7%), TG (26.5%), and FPG (3.0%) (Supplementary Fig. 9).
Sensitivity analysis and subgroup analysis
Considering that 642 participants (7.0%) experienced a competing event (death) during follow-up (Supplementary Fig. 10), we also employed the Fine–Gray competing risks model. The cumulative incidence function (CIF) also indicated a higher incidence of stroke in group with combined HTN and high CTI (Gray’s test, P < 0.001) (Supplementary Fig. 11). After multiple adjustments, the competing risk model showed a graded increase in the subdistribution hazard ratio (sHR) and 95%CI across the classes: Class 2 [1.60 (1.33–1.91)], Class 3 [2.49 (1.97–3.16)], and Class 4 [2.96 (2.43–3.61)]. A comparative analysis revealed that the estimates for the CTI-HTN combined groups and stroke association were closely aligned between the competing risk and Cox proportional hazards models. In the fully adjusted model, the maximum difference between the HR and the SHR was 0.23, and the maximum HR/SHR ratio was 1.077 (Supplementary Table 8).
To further validate the robustness of our primary findings, multiple sensitivity analyzes were conducted. The robust association between stroke risk and Class 4 (HTN with higher CTI) was consistently observed, when using uninterpoluted data (Supplementary Table 9), remaining significant after excluding stage 0 individuals (Supplementary Table 10), after adjusting for medication use (Supplementary Table 11), after excluding participants who died during follow-up (Supplementary Table 12), after additional adjustment for eGFR (Supplementary Table 13), after converting the units of the CTI (Supplementary Fig. 12 and Supplementary Table 14), after excluding those with CRP of 10 or greater (Supplementary Table 15) and after using a criterion of 130/80 mmHg for hypertension (Supplementary Table 16).
We conducted subgroup analyzes to assess the relationship between the CTI-HTN combined groups and stroke incidence across the following variables: age, gender, marital status, educational level, smoking status, alcohol consumption, BMI, diabetes, and dyslipidemia. The results revealed a significant and consistent relationship between CTI-HTN combined groups and long-term stroke risk across all prespecified subgroups (P for trend < 0.05 for all, Table 3). Interaction analyses indicated no significant effect by age, sex, or other clinical subgroups (all P for interaction > 0.05).
Table 3.
Subgroup analyses of association between CTI and hypertension with long-term stroke
| Subgroups | Status | HR (95% CI) | P value | P for interaction | ||
|---|---|---|---|---|---|---|
| Class 2 | Class 3 | Class 4 | ||||
| Age | < 65 | 1.86 (1.49, 2.33) | 2.95 (2.18, 3.98) | 3.33 (2.59, 4.29) | < 0.001 | 0.075 |
| ≥ 65 | 1.16 (0.85, 1.60) | 1.81 (1.24, 2.65) | 2.17 (1.56, 3.02) | < 0.001 | ||
| Gender | Female | 1.46 (1.13, 1.90) | 1.94 (1.36, 2.78) | 2.65 (2.01, 3.50) | < 0.001 | 0.360 |
| Male | 1.71 (1.33, 2.21) | 3.13 (2.28, 4.29) | 3.32 (2.49, 4.43) | < 0.001 | ||
| BMI | < 24 | 1.49 (1.18, 1.89) | 2.34 (1.73, 3.16) | 2.86 (2.14, 3.83) | < 0.001 | 0.493 |
| ≥ 24 | 1.35 (0.99, 1.84) | 2.23 (1.51, 3.31) | 2.26 (1.64, 3.10) | < 0.001 | ||
| Diabetes | No | 1.59 (1.32, 1.91) | 2.55 (2.00, 3.24) | 2.98 (2.43, 3.66) | < 0.001 | 0.858 |
| Yes | 1.87 (0.62, 5.63) | 1.63 (0.40, 6.68) | 3.18 (1.09, 9.27) | 0.012 | ||
| Dyslipidemia | No | 1.54 (1.27, 1.85) | 2.42 (1.88, 3.10) | 3.00 (2.43, 3.70) | < 0.001 | 0.411 |
| Yes | 2.68 (1.18, 6.11) | 3.75 (1.51, 9.29) | 3.71 (1.65, 8.32) | 0.002 | ||
| Current smoking | No | 1.46 (1.15, 1.87) | 2.26 (1.64, 3.11) | 2.61 (2.01, 3.39) | < 0.001 | 0.663 |
| Yes | 1.78 (1.35, 2.34) | 2.85 (2.00, 4.06) | 3.58 (2.63, 4.89) | < 0.001 | ||
| Alcohol consumption | No | 1.43 (1.13, 1.83) | 2.07 (1.48, 2.89) | 2.94 (2.26, 3.81) | < 0.001 | 0.212 |
| Yes | 1.86 (1.41, 2.45) | 3.09 (2.20, 4.33) | 3.03 (2.21, 4.15) | < 0.001 | ||
| Marital status | Unmarried | 1.28 (0.81, 2.03) | 1.39 (0.73, 2.63) | 1.85 (1.09, 3.13) | 0.023 | 0.088 |
| Married | 1.68 (1.38, 2.05) | 2.77 (2.15, 3.58) | 3.25 (2.62, 4.04) | < 0.001 | ||
| Residence | Urban | 1.99 (1.40, 2.83) | 3.54 (2.28, 5.51) | 3.82 (2.63, 5.56) | < 0.001 | 0.202 |
| Rural | 1.47 (1.19, 1.83) | 2.21 (1.66, 2.93) | 2.69 (2.12, 3.43) | < 0.001 | ||
The Class 1 as a reference. BMI: body mass index. HR = hazard ratio, CI = confidence interval
Discussion
In this longitudinal CHARLS cohort analysis, we systematically evaluated the relationship between inflammation, insulin resistance and hypertension with long-term incidence of stroke. First, a graded relationship exists between the combined exposure to CTI and hypertension and the incidence of long-term stroke. Second, the lack of a significant interaction between CTI and hypertension indicates their effects are independent. Third, a model combining both factors significantly outperformed all single-factor models in predictive accuracy. Finally, CRP and hypertension were quantified as the predominant risk factors. These results underscore the value of integrating inflammatory and metabolic markers with hypertension status to optimize early stroke risk stratification.
Mounting evidence underscores the advantage of composite indicators, such as obesity parameters, blood pressure, and blood glucose, over single parameters in assessing cardiometabolic risk [25, 26]. The CTI, integrating the TyG index with CRP, further enhances the comprehensive characterization ability for cardiovascular diseases [12]. While prior studies established CTI as an independent stroke predictor in specific cohorts [27, 28], our study extends this by systematically evaluating its synergistic effect with hypertension (HTN). Our findings demonstrated that the combination of CTI and HTN exhibited superior predictive performance for stroke compared to single-factor models. Furthermore, no interaction was observed between CTI and hypertension in our study, consistent with previous findings that showed no significant interactive effects between the TyG index and body shape indices such as BRI and CVAI in terms of stroke risk [29, 30]. This suggests that these risk factors may contribute to stroke incidence through independent pathophysiological pathways, highlighting the clinical importance of multidimensional combined assessment.
The observed additive interaction suggests that the combined effect of CTI and hypertension on stroke risk is approximately equal to the sum of their individual effects. This indicates that in individuals exposed to both high-risk CTI and hypertension, the excess stroke risk can be viewed as a simple summation of the risk increments independently attributable to each factor, without evidence of a significant synergistic amplification effect. It is important to interpret this finding with caution. Statistical power for detecting additive interactions is generally limited. The non-significant result in this study could be attributable to the constrained sample size, or it may accurately reflect the underlying biological relationship within the target population. Further validation in larger prospective cohorts is warranted. From a public health perspective, interventions aimed at reducing the population-level stroke burden may therefore approach “blood pressure control” and “CTI management” as two distinct, yet equally critical, foundational pillars.
In contrast to prior approaches that viewed SBP as a mediator [25, 31], we positioned hypertension as a co-exposure alongside CTI within the CKM syndrome model, where inflammation, metabolic dysregulation, and hypertension jointly promote vascular injury and disease progression to stroke [32]. The WQS analysis revealed that CRP and hypertension status together account for more than 70% of the predictive weight, highlighting the critical role of the “inflammatory-hypertensive axis” in stroke risk prediction. Consequently, in clinical practice, alongside stringent blood pressure management, increased attention should be given to the monitoring and intervention of inflammatory factors to more comprehensively and effectively reduce the risk of stroke.
The pathophysiological link between the CTI-hypertension combination and stroke involves synergistic mechanisms. Insulin resistance (a core component of CTI) induces endothelial dysfunction and a pro-thrombotic state via PI3K/NO pathway suppression [33, 34]. while simultaneously promoting the systemic inflammation that drives atherosclerosis [35, 36]. Hypertension amplifies this injury by causing direct arterial damage and accelerating atherosclerotic progression. This convergence of metabolic, inflammatory, and hemodynamic insults creates a high-risk environment for stroke [37, 38].
Current stroke prevention paradigms are evolving beyond a sole focus on high-risk individuals, particularly since over 80% of events occur in those classified as low-risk [39]. This underscores the imperative for strategies that integrate population-wide prevention with refined risk stratification [40, 41]. In alignment with AHA guidelines prioritizing intervention in CKM stages 0–3 [8, 42], our study demonstrates that the combination of CTI and hypertension effectively stratifies stroke risk not only in stage 3 but also in the broader, lower-risk stage 0–2 population. Crucially, the predictive association remained consistent across all examined subgroups without significant interaction, indicating that the model demonstrates robust and stable performance in patients with different clinical characteristics.
Strengths of this study
This study had several strengths. First, it was the first investigation to examine the association between inflammation, insulin resistance and hypertension with the incidence of stroke in a rigorously defined population with CKM stages 0–3. Second, the utilization of the nationally representative China Health and Retirement Longitudinal Study (CHARLS) cohort enhanced the generalizability and reliability of our findings. Third, we accounted for the competing risk of death by incorporating competing risk models, which more accurately delineated the independent association between CTI and stroke risk and ensured the robustness of the conclusions. Fourth, multiple sensitivity analyzes were conducted to validate the consistency and reliability of the results.
Limitations
Several limitations should be acknowledged. First, stroke diagnoses were based on self-report without differentiation between subtypes (e.g., ischemic or hemorrhagic), which may affect the accuracy of the outcomes, and we must admit the existence of potential reporting bias. Secondly, the timing of stroke onset in the CHARLS dataset is interval-censored, as events are only known to have occurred between biennial survey waves. While the standard Cox model provides a robust approximation, our findings could be further validated in future studies using statistical methods specifically designed for interval-censored data. Thirdly, it must be acknowledged that the CHARLS dataset lacks direct measurements such as carotid intima-media thickness, plaque, and ankle-brachial index. However, the above definition based on risk equivalents represents the optimal and clinically reasonable alternative under the current data constraints, in alignment with international guidelines. Fourth, the study included only middle-aged and older Chinese adults, and caution should be exercised when generalizing the findings to other populations. Fifth, CTI and hypertension were measured only at baseline; changes over time were not captured, which may influence risk estimation. Sixth, the CTI cutoff value in this study was derived from the overall sample using restricted cubic splines and may not be directly applicable to other specific populations or subgroups. Finally, although multiple confounding factors were adjusted for, residual confounding cannot be entirely ruled out due to the observational nature of the study.
Conclusion
This study demonstrated that the combined assessment of inflammation, insulin resistance, and hypertension improves the prediction of long-term stroke risk. The CTI–HTN combined model provided superior risk stratification and should be integrated into the early stages (0–3) of the cardiovascular-kidney-metabolic (CKM) framework to enable targeted primary prevention, mitigating the future burden of stroke on healthcare systems.
Supplementary Information
Acknowledgements
The data used in this study were generously provided by the Institute of Social Science Survey at Peking University and the National School of Development at Peking University.
Author contributions
Ying Cui: Conceptualization, Methodology and Writing- Reviewing and Editing. Ze-Jiaxin Niu: Data curation, Writing- Original draft preparation and Methodology. Zi-Ang Liu: Writing- Original draft preparation and Methodology. Tian Wei: Investigation and Data analysis. Meng Dou: Fund assistance. Pu-Xun Tian: Supervision, Reviewing and Fund assistance.
Funding
This work was financially supported by the National Natural Science Foundation of China (Grant No. 82270791), Natural Science Basic Research Plan in Shaanxi Province (Grant No. 2025SF-YBXM-268).
Data availability
The datasets used and analyzed during the current study are available from the corresponding author on reasonable request.
Declarations
Ethical approval and consent to participate
The data utilized in this study were obtained from the China Health and Retirement Longitudinal Study (CHARLS), a publicly available national cohort. The original CHARLS survey protocol was reviewed and approved by the Peking University Institutional Review Board (IRB00001052–11015). All participants provided written informed consent prior to their inclusion. This analysis was performed in strict compliance with the principles outlined in the Declaration of Helsinki. The study was conducted using anonymized data, and all procedures adhered to relevant ethical guidelines for the use of human subject data.
Consent for publication
Not applicable.
Generative AI use
DeepSeek AI was employed for language refinement during the preparation of this manuscript. The tool assisted in enhancing the clarity and fluency of the textual content. All AI-generated content underwent thorough review and editing by the authors to ensure accuracy and appropriateness. The authors assume full responsibility for the final content, interpretation of results, and scholarly claims presented in this publication.
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.
Contributor Information
Pu-Xun Tian, Email: tianpuxunxa@163.com.
Ying Cui, Email: cuiyingcy10@163.com.
References
- 1.Ndumele CE, Rangaswami J, Chow SL, Neeland IJ, Tuttle KR, Khan SS, et al. Cardiovascular-Kidney-Metabolic health: a presidential advisory from the American Heart Association. Circulation. 2023;148(20):1606–35. [DOI] [PubMed] [Google Scholar]
- 2.Ndumele CE, Neeland IJ, Tuttle KR, Chow SL, Mathew RO, Khan SS, et al. A synopsis of the evidence for the science and clinical management of Cardiovascular-Kidney-Metabolic (CKM) syndrome: a scientific statement from the American Heart Association. Circulation. 2023;148(20):1636–64. [DOI] [PubMed] [Google Scholar]
- 3.Aggarwal R, Ostrominski JW, Vaduganathan M. Prevalence of Cardiovascular-Kidney-Metabolic syndrome stages in US adults, 2011–2020. JAMA. 2024;331(21):1858–60. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Xie Z, Yu C, Cui Q, Zhao X, Zhuang J, Chen S, et al. Global burden of the key components of Cardiovascular-Kidney-Metabolic syndrome. J Am Soc Nephrol. 2025;36(8):1572–84. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Feigin VL, Brainin M, Norrving B, Martins SO, Pandian J, Lindsay P, et al. World Stroke Organization: Global Stroke Fact Sheet 2025. Int J Stroke. 2025;20(2):132–44. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Collaborators G2S. Global, regional, and national burden of stroke and its risk factors, 1990–2019: a systematic analysis for the Global Burden of Disease Study 2019. Lancet Neurol. 2021;20(10):795–820. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Brainin M, Organization WS, Sliwa K, Federation WH. WSO and WHF joint position statement on population-wide prevention strategies. Lancet. 2020;396(10250):533–4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Owolabi MO, Thrift AG, Mahal A, Ishida M, Martins S, Johnson WD, et al. Primary stroke prevention worldwide: translating evidence into action. Lancet Public Health. 2022;7(1):e74–85. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Kaptoge S, Angelantonio ED, Lowe G, Pepys MB, Thompson SG, Collins R, et al. C-reactive protein concentration and risk of coronary heart disease, stroke, and mortality: an individual participant meta-analysis. Lancet. 2010;375(9709):132–40. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Tabas I, Tall A, Accili D. The impact of macrophage insulin resistance on advanced atherosclerotic plaque progression. Circ Res. 2010;106(1):58–67. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Zhao DF. Value of C-reactive protein-triglyceride glucose index in predicting cancer mortality in the general population: results from National Health and Nutrition Examination Survey. Nutr Cancer. 2023;75(10):1934–44. [DOI] [PubMed] [Google Scholar]
- 12.Ou H, Wei M, Li X, Xia X. C-reactive protein-triglyceride glucose index in evaluating cardiovascular disease and all-cause mortality incidence among individuals across stages 0–3 of cardiovascular-kidney-metabolic syndrome: a nationwide prospective cohort study. Cardiovasc Diabetol. 2025;24(1):296. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Huo G, Tang Y, Liu Z, Cao J, Yao Z, Zhou D. Association between C-reactive protein-triglyceride glucose index and stroke risk in different glycemic status: insights from the China Health and Retirement Longitudinal Study (CHARLS). Cardiovasc Diabetol. 2025;24(1):142. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Jiang L, Zhu T, Song W, Zhai Y, Tang Y, Ruan F, et al. Assessment of six insulin resistance surrogate indexes for predicting stroke incidence in Chinese middle-aged and elderly populations with abnormal glucose metabolism: a nationwide prospective cohort study. Cardiovasc Diabetol. 2025;24(1):56. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Chapman MJ, Sposito AC. Hypertension and dyslipidaemia in obesity and insulin resistance: pathophysiology, impact on atherosclerotic disease and pharmacotherapy. Pharmacol Ther. 2008;117(3):354–73. [DOI] [PubMed] [Google Scholar]
- 16.Guzik TJ, Nosalski R, Maffia P, Drummond GR. Immune and inflammatory mechanisms in hypertension. Nat Rev Cardiol. 2024;21(6):396–416. [DOI] [PubMed] [Google Scholar]
- 17.Zhao Y, Hu Y, Smith JP, Strauss J, Yang G. Cohort profile: the China Health and Retirement Longitudinal Study (CHARLS). Int J Epidemiol. 2014;43(1):61–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Khan SS, Matsushita K, Sang Y, Ballew SH, Grams ME, Surapaneni A, et al. Development and validation of the American Heart Association’s PREVENT equations. Circulation. 2024;149(6):430–49. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Chen X, Crimmins E, Hu PP, Kim JK, Meng Q, Strauss J, 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. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Ruan GT, Xie HL, Zhang HY, Liu CA, Ge YZ, Zhang Q, et al. A novel inflammation and insulin resistance related indicator to predict the survival of patients with cancer. Front Endocrinol (Lausanne). 2022;13:905266. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Min Q, Wu Z, Yao J, Wang S, Duan L, Liu S, et al. Association between atherogenic index of plasma control level and incident cardiovascular disease in middle-aged and elderly Chinese individuals with abnormal glucose metabolism. Cardiovasc Diabetol. 2024;23(1):54. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Liu Y, Li W, Zhou H, Zeng H, Jiang J, Wang Q, et al. Association between atherogenic index of plasma and new-onset stroke in a population with cardiovascular-kidney-metabolic syndrome stages 0-3: insights from CHARLS. Cardiovasc Diabetol. 2025;24(1):168. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Yan F, Chen X, Quan X, Wang L, Wei X, Zhu J. Association between the stress hyperglycemia ratio and 28-day all-cause mortality in critically ill patients with sepsis: a retrospective cohort study and predictive model establishment based on machine learning. Cardiovasc Diabetol. 2024;23(1):163. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Wang Q, Qiao W, Zhang H, Liu B, Li J, Zang C, et al. Nomogram established on account of Lasso-Cox regression for predicting recurrence in patients with early-stage hepatocellular carcinoma. Front Immunol. 2022;13:1019638. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Yue Y, Li P, Sun Z, Murayama R, Li Z, Hashimoto K, et al. Association of novel triglyceride-glucose-related indices with incident stroke in early-stage cardiovascular-kidney-metabolic syndrome. Cardiovasc Diabetol. 2025;24(1):301. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Wang X, Wen P, Liao Y, Wu T, Zeng L, Huang Y, et al. Association of atherogenic index of plasma and its modified indices with stroke risk in individuals with cardiovascular-kidney-metabolic syndrome stages 0-3: a longitudinal analysis based on CHARLS. Cardiovasc Diabetol. 2025;24(1):254. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Tang S, Wang H, Li K, Chen Y, Zheng Q, Meng J, et al. C-reactive protein-triglyceride glucose index predicts stroke incidence in a hypertensive population: a national cohort study. Diabetol Metab Syndr. 2024;16(1):277. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Xu Y, Chen S, Zhu J, Wang Q, Li W, Pan G, et al. C-reactive protein-triglyceride glucose index and stroke risk in early cardiovascular-kidney-metabolic syndrome: a national cohort study. BMC Cardiovasc Disord. 2025;25(1):634. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Yang Y, Li S, Ren Q, Qiu Y, Pan M, Liu G, et al. The interaction between triglyceride-glucose index and visceral adiposity in cardiovascular disease risk: findings from a nationwide Chinese cohort. Cardiovasc Diabetol. 2024;23(1):427. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Wang B, Li L, Tang Y, Ran X. Joint association of triglyceride glucose index (TyG) and body roundness index (BRI) with stroke incidence: a national cohort study. Cardiovasc Diabetol. 2025;24(1):164. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Qiu W, Cai A, Li L, Feng Y. Systolic blood pressure status modifies the associations between the triglyceride-glucose index and incident cardiovascular disease: a national cohort study in China. Cardiovasc Diabetol. 2024;23(1):135. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Cipolla MJ, Liebeskind DS, Chan SL. The importance of comorbidities in ischemic stroke: impact of hypertension on the cerebral circulation. J Cereb Blood Flow Metab. 2018;38(12):2129–49. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Muniyappa R, Chen H, Montagnani M, Sherman A, Quon MJ. Endothelial dysfunction due to selective insulin resistance in vascular endothelium: insights from mechanistic modeling. Am J Physiol Endocrinol Metab. 2020;319(3):E629–46. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Hierons SJ, Marsh JS, Wu D, Blindauer CA, Stewart AJ. The interplay between non-esterified fatty acids and plasma zinc and its influence on thrombotic risk in obesity and type 2 diabetes. Int J Mol Sci. 2021. 10.3390/ijms221810140. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.DeFronzo RA. Insulin resistance, lipotoxicity, type 2 diabetes and atherosclerosis: the missing links. The Claude Bernard Lecture 2009. Diabetologia. 2010;53(7):1270–87. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Henein MY, Vancheri S, Longo G, Vancheri F. The role of inflammation in cardiovascular disease. Int J Mol Sci. 2022. 10.3390/ijms232112906. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Laurent S, Boutouyrie P, Lacolley P. Structural and genetic bases of arterial stiffness. Hypertension. 2005;45(6):1050–5. [DOI] [PubMed] [Google Scholar]
- 38.Yu JG, Zhou RR, Cai GJ. From hypertension to stroke: mechanisms and potential prevention strategies. CNS Neurosci Ther. 2011;17(5):577–84. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Feigin VL. Primary stroke prevention needs overhaul. Int J Stroke. 2017;12(1):5–6. [DOI] [PubMed] [Google Scholar]
- 40.Feigin VL, Martins SC, Brainin M, Norrving B, Kamenova S, Giniyat A, et al. Twenty years on from the introduction of the high risk strategy for stroke and cardiovascular disease prevention: a systematic scoping review. Eur J Neurol. 2024;31(3):e16157. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Feigin VL, Brainin M, Norrving B, Gorelick PB, Dichgans M, Wang W, et al. What is the best mix of population-wide and high-risk targeted strategies of primary stroke and cardiovascular disease prevention. J Am Heart Assoc. 2020;9(3):e014494. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Feigin VL, Owolabi MO, Group WSONCSC. Pragmatic solutions to reduce the global burden of stroke: a World Stroke Organization-Lancet Neurology Commission. Lancet Neurol. 2023;22(12):1160–206. [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 datasets used and analyzed during the current study are available from the corresponding author on reasonable request.