Synergistic effect of AIP and cystatin C: a study on cardiovascular risk prediction in populations with different glycemic statuses.

BACKGROUND

Cardiovascular disease (CVD) poses a major global health burden. The Atherogenic Index of Plasma (AIP) and Cystatin C are novel biomarkers reflecting lipid metabolism disorders and renal/micro-inflammatory status, respectively. Their combined indicator (AIP-Cys, the product of AIP × Cystatin C) may provide improved predictive value beyond either marker alone. This study aimed to investigate the predictive role of AIP-Cys for the risk of incident CVD across different glycemic statuses: normal glucose regulation (NGR), pre-diabetes (preDM), and diabetes mellitus (DM).

MATERIALS AND

METHODS

This study was based on the prospective China Health and Retirement Longitudinal Study (CHARLS) cohort. A total of 6,035 participants aged ≥ 45 years without a history of CVD at baseline (2011) were included. The primary exposure was baseline AIP-Cys (the product of AIP × Cystatin C), and the outcome was the first self-reported incident heart disease or stroke event during follow-up (until 2020). Glycemic status was defined based on fasting plasma glucose and glycated hemoglobin levels. Multivariable Cox proportional hazards models were used to estimate hazard ratios (HRs). Restricted cubic splines and piecewise linear models were applied to analyze nonlinear relationships. Mediation and sensitivity analyses were conducted to verify the robustness of the results.

RESULTS

During a median follow-up of 9.0 years, 651 incident CVD cases occurred. In the fully adjusted model, AIP-Cys was independently associated with CVD risk in the overall (HR per unit increase: 1.34; 95% CI: 1.06-1.69), NGR (HR: 1.60; 95% CI: 1.13-2.26), and preDM (HR: 1.32; 95% CI: 1.02-1.70) populations, but not in the DM population. RCS analysis revealed a significant nonlinear relationship in the NGR and preDM groups (P for nonlinearity < 0.05). Piecewise regression identified inflection points (0.47 for NGR; 0.37 for preDM), below which risk increased sharply. Quartile analysis showed the highest risk in Q3 rather than Q4, suggesting a saturation effect. Mediation analysis indicated that systolic blood pressure partially mediated this association (mediating proportion: 14.6% in overall population). The predictive performance of AIP-Cys was modest for short-term risk (1-year AUC: 0.64-0.67) but limited for long-term prediction.

CONCLUSIONS

The composite indicator AIP-Cys, derived from AIP and Cystatin C, is an independent risk factor for incident CVD in individuals with NGR and preDM. The association exhibits a nonlinear threshold pattern. However, its predictive value is limited in the DM population, and its long-term predictive performance is modest. Additionally, the self-reported nature of cardiovascular outcomes represents a limitation that should be considered when interpreting these findings. By integrating lipid metabolism and renal/inflammatory pathways, this indicator offers a novel biomarker-based approach for early cardiovascular risk identification in the pre-diabetes stage. Future research should further explore its potential for clinical translation and intervention.