Abstract
BACKGROUND: Cardiovascular-kidney-metabolic (CKM) syndrome is a major health burden. Stroke, the third leading cause of death globally, is strongly linked to insulin resistance (IR). The cholesterol-high-density lipoprotein-glucose (CHG) index offers superior diagnostic accuracy for diabetes, but its association with stroke in patients with early CKM syndrome (stages 0-3) is uncertain. METHODS: A total of 6,836 adults with CKM stages 0-3 from CHARLS (2011-2020) were analysed after excluding participants with missing baseline data for exposures or covariates, yielding an analytic sample with a somewhat higher proportion of men and CKM stage 3 than the parent CHARLS cohort. Multivariable Cox models were used to estimate the associations of the CHG index and its derivatives (CHG-BMI, CHG-WC, and CHG-WHtR) with incident stroke. Dose-response associations were examined using restricted cubic splines (RCS). When non-linear patterns were detected, two-piecewise Cox models were fitted to quantify threshold effects. Discrimination was summarised using ROC curves. This selection may limit the generalisability of the findings to the broader CHARLS population. RESULTS: Over a 9-year follow-up period, 575 strokes occurred (8.41%). In multivariable models, higher CHG-related indices were modestly but consistently associated with incident stroke. RCS analyses indicated an approximately linear association for CHG, a saturation pattern for CHG-BMI (risk rising below but flattening above its inflection point), and threshold-like non-linear patterns for CHG-WC and CHG-WHtR, with little change in risk at low values and increasing hazards beyond their inflection points. Subgroup analyses revealed generally consistent associations by age and sex and stronger effects in CKM stages 2-3, without significant interactions. ROC performance was modest and comparable across indices (AUCs ≈0.64-0.65). CONCLUSION: The CHG index and its derivatives were consistently associated with incident stroke amongst adults in the early CKM stages, and their clinical adoption may enhance early detection and prevention of stroke events in populations with vulnerable metabolism.