Associations of the CHG index and GLMI with incident heart disease in middle-aged and older Chinese adults: findings from CHARLS

Abstract

Cardiovascular disease (CVD) remains a major global health challenge. The cholesterol–high-density lipoprotein–glucose (CHG) index and glucose-lipid metabolic index (GLMI) have been associated with cardiovascular outcomes, but studies on their associations with incident heart disease in Chinese populations remain scarce. This study compared their associations with incident heart disease and assessed their standalone discriminative ability in middle-aged and older Chinese adults. This cohort study included 5,549 adults aged ≥ 45 years from the China Health and Retirement Longitudinal Study (CHARLS). Baseline data (2011) were used to calculate the CHG index and GLMI; incident heart disease was ascertained over 9 years. Cox proportional hazards models with sequential adjustment, restricted cubic splines (RCS), Kaplan–Meier (KM) curves, time-dependent receiver operating characteristic (ROC) analyses, and subgroup analyses were used to examine associations and describe standalone discrimination. During 9 years of follow-up, 1,404 (25.3%) participants developed heart disease. In unadjusted models, each 1-standard deviation (1-SD) increase in the CHG index was associated with an 11% higher hazard (HR = 1.11, 95% CI: 1.05–1.16), and each 1-SD increase in GLMI was associated with a 6% higher hazard (HR = 1.06, 95% CI: 1.03–1.10; both P < 0.001). The highest quartile had the strongest associations (CHG: HR = 1.31; GLMI: HR = 1.42). After adjustment for age, sex, BMI, smoking, and drinking (Model 2), both associations became non-significant; further adjustment (Model 3) yielded similar results. In unadjusted KM analyses, higher quartiles had higher cumulative incidence (log-rank P < 0.001). RCS analyses showed no significant overall or nonlinear associations after adjustment. Time-dependent ROC analyses yielded an area under the curve (AUC) of 0.517–0.561, indicating poor standalone discrimination. DeLong tests showed occasional significant differences, but absolute AUC differences were small. Subgroup analyses found no significant interactions (all P for interaction > 0.05). The CHG index and GLMI were associated with incident heart disease in unadjusted analyses, but these associations attenuated after sequential adjustment for conventional risk factors. Both indices showed poor standalone discrimination (AUCs 0.517–0.561), with numerically small differences between them. Current evidence is not yet sufficient to support these indices as standalone screening tools for heart disease in this Chinese middle-aged and older population. Future studies should focus on specific cardiac endpoints and incorporate comprehensive metabolomic profiling to clarify their role in risk stratification. Not applicable.

The authors' abstract, as published at the source. BMC Cardiovascular Disorders, 2026 · DOI ↗