CVMI-age quantifies multidimensional cardiovascular aging and predicts cardiovascular events: a Chinese community-based cohort study
H Xie, F F Fan, J Jia, Y Gong, Y ZhangAbstract
Background
Cardiovascular (CV) aging is a complex biological process driven by hemodynamic stress, metabolic dysregulation, and chronic inflammation. Traditional vascular aging models relying solely on arterial stiffness often fail to capture the systemic immuno-metabolic burden, leaving a gap in quantifying the residual CV risk in the general population.
Purpose
This study aims to construct a composite Cardiovascular-Metabolic-Inflammatory Age (CVMI-Age), identify distinct aging phenotypes and determine their independent predictive value for major adverse CV events (MACE) over standard risk scores and simple stiffness-based metrics.
Methods
8,096 participants without CV disease were analyzed from a Chinese community-based cohort. CVMI-Age was constructed using GAM and LASSO regression. Predictors included: (1) Vascular: brachial-ankle pulse wave velocity (baPWV), carotid intima-media thickness, ankle-brachial index, and blood pressure; (2) Metabolic: triglyceride-glucose index, lipid accumulation product index, cardiometabolic index, atherogenic index of plasma, and weight-adjusted-waist index; and (3) Inflammatory: systemic immune-inflammation index, neutrophil-to-lymphocyte ratio, and neutrophil-to-HDL ratio. ΔCVMI-Age was defined as the residual obtained from a regression of CVMI-Age on chronological age. Participants were stratified into Decelerated (DCA, <10th percentile), Normal (NCA, 10th-90th), and Accelerated (ACA, >90th) CV aging phenotypes. The endpoint was MACE, comprising CV mortality, non-fatal stroke, and non-fatal myocardial infarction. Statistical analyses included restricted cubic splines (RCS), Kaplan-Meier curves, Cox proportional hazard models, and reclassification metrics (C-statistic, NRI and IDI).
Results
Mean age was 56.6 ± 8.9 years, and 5,226 were female (64.2%). Over a median follow-up of 9.31 years, 815 (10.1%) participants experienced MACE. RCS analysis revealed a J-shaped non-linear association between ΔCVMI-Age and MACE (P <0.001, P for non-linear <0.001). Kaplan-Meier curves showed significant divergence in cumulative MACE incidence among CVMI-Age phenotypes (Log-rank P<0.001). After adjusting for age, sex, and the China-PAR CV risk score, the ACA phenotype was independently associated with increased MACE risk (HR 1.70, 95% CI 1.42–2.03) compared with NCA, while DCA was protective (HR 0.71, 95% CI 0.54–0.93). Notably, the addition of ΔCVMI-Age significantly improved risk discrimination and reclassification beyond the combined model incorporating both China-PAR and baPWV-derived age (C-statistic increase 0.0033; NRI 5.56%; IDI 0.37%; all P<0.05).
Conclusion
CVMI-Age is a potential biological aging marker integrating vascular structure with systemic immuno-metabolic profiles. The ACA phenotype identifies high-risk individuals with hidden burdens missed by traditional scores, offering significant incremental value for risk stratification.