Combined Metabolic-Frailty Score Triples Cardiovascular Risk in Older Chinese Adults

Cardiovascular disease remains the world's leading cause of death, yet conventional risk factors often fail to capture the full complexity of cardiometabolic vulnerability—particularly in aging populations where metabolic dysregulation and physical decline frequently coexist. This prospective study from the China Health and Retirement Longitudinal Study (CHARLS) introduces the AIP-FI, a novel composite index multiplying the Atherogenic Index of Plasma (AIP = log10[TG/HDL-C]) by the Frailty Index (FI), to assess whether integrating lipid metabolic imbalance with cumulative physiological deficits improves cardiovascular risk prediction.

The study enrolled 6,896 participants aged 45 and older without baseline CVD, stroke, or heart disease, drawn from a nationally representative Chinese cohort. The Frailty Index was constructed from 28 health variables spanning chronic diseases, physical function, depressive symptoms, disability, and cognitive performance. Participants were followed across five survey waves from 2011 to 2020. Cox proportional hazards models and restricted cubic spline (RCS) analyses were applied across three progressively adjusted models to examine the association between AIP-FI and incident CVD, stroke, and heart disease.

Over the 9-year follow-up, 1,648 CVD events (23.9%), 548 strokes (7.9%), and 1,280 heart disease events (18.6%) occurred. Each 1-unit increment in AIP-FI was associated with an approximately threefold increase in CVD risk (HR: 2.95, 95% CI: 2.15–4.05), a more than threefold increase in stroke risk (HR: 3.14, 95% CI: 1.88–5.26), and a nearly threefold increase in heart disease risk (HR: 2.72, 95% CI: 1.89–3.92), all in the fully adjusted model. RCS analyses confirmed a significant positive nonlinear dose-response relationship for all three outcomes, indicating that risk accelerates disproportionately at higher AIP-FI values.

Subgroup and interaction analyses were conducted across sex, age, residence, smoking, drinking, hypertension, diabetes, dyslipidemia, and BMI strata, supporting the robustness of the association. Three sensitivity analyses—excluding participants with missing values, removing non-fasting individuals, and removing deceased participants—consistently confirmed the primary findings. E-value calculations suggested that unmeasured confounding would need to be implausibly strong to explain away the observed associations.

These findings suggest that AIP-FI captures a synergistic interaction between atherogenic lipid profiles and biological aging that neither marker alone fully conveys. As both AIP and FI can be derived from routinely collected clinical data, AIP-FI represents a low-cost, scalable tool for early identification of high-risk individuals in community and clinical settings. The study is particularly relevant for aging populations in China and similar demographic contexts globally, where cardiovascular burden is rising rapidly.