Association of C-reactive protein-triglyceride glucose index and its modified indices with stroke risk in individuals at cardiovascular-kidney-metabolic syndrome stages 0–3: a population-based cohort study

Abstract

Stroke pathogenesis in Cardiovascular-Kidney-Metabolic (CKM) is driven by a complex interplay of insulin resistance (IR), systemic inflammation, and adiposity. This study investigated whether integrating adiposity-related measures with the inflammation-IR axis could refine stroke risk prediction in individuals at CKM stage 0–3. A total of 3,348 participants (aged ≥ 45) at CKM stages 0–3 were included from the China Health and Retirement Longitudinal Study (CHARLS). Baseline levels, cumulative exposures, and longitudinal changes were assessed for the original C-reactive protein-triglyceride glucose index (CTI) and seven adiposity-modified CTI indices. Associations with stroke risk were evaluated using Cox proportional hazards models, and potential non-linear relationships were examined using restricted cubic splines. Predictive performance at 3 and 5 years was assessed using fixed-horizon receiver operating characteristic (ROC) analysis. Discrimination was quantified by the area under the curve (AUC) and compared using DeLong’s test, while incremental predictive value was further evaluated using the integrated discrimination improvement (IDI) index. Mediation analysis was used to evaluate the potential role of estimated pulse wave velocity (ePWV) in these associations. Subgroup and sensitivity analyses were conducted to test result robustness, and weighted quantile sum (WQS) regression was performed to determine the relative contribution of components within CTI-related indices. During follow-up, 263 new cases of stroke were observed (7.9%). In fully adjusted models, participants in the highest tertile of all CTI-related indices had significantly increased stroke risk, with hazard ratios (HRs) ranging from 1.92 (95%CI 1.35–2.73) for original CTI to 2.52 (95%CI 1.76–3.61) for CTI-WHtR. K-means clustering of longitudinal changes identified three trajectory groups; compared with the stable group, both markedly increasing and slowly increasing trajectories of CTI and its modified indices were significantly associated with higher stroke risk. In fixed-horizon ROC analyses, the 3-year AUCs for all indices exceeded 0.70. Among baseline indices, CTI-BRI, CTI-BMI, CTI-WC, and CTI-WHtR showed significantly higher AUCs than the original CTI, whereas only CTI-BMI demonstrated a small but significant improvement in IDI over CTI (IDI = 0.0021, 95% CI 0.0004–0.0038; P = 0.012). Among cumulative indices, cumCTI-BMI and cumCTI-WHtR showed significantly higher 3-year AUCs than cumCTI. CTIs-stroke associations were stronger in advanced CKM stages (2–3). Furthermore, component analysis identified CVAI (29.0% in 2012; 29.3% in 2015) as the primary driver of predictive power. ePWV partially mediated the associations between CTI-related indices and incident stroke, with mediation proportions of 12.44%–21.20%. CTI and adiposity-modified CTI indices were independently associated with incident stroke among adults at CKM stages 0–3, especially in advanced CKM stages. Incorporating adiposity-related information provided modest and inconsistent improvement in short-term prediction, with CVAI contributing most prominently and ePWV suggesting a potential arterial stiffness pathway. These findings support the relevance of combined inflammatory, IR, and adiposity burden in stroke risk assessment, but the clinical use of complex composite indices should be balanced against screening feasibility.