From association to prediction: Advancing risk assessment for diabetes and diabetic kidney disease with CHG and CHG-related indices

Introduction The cholesterol, high density lipoprotein, and glucose (CHG) index has been linked to diabetes mellitus (DM) and diabetic kidney disease (DKD), while its combined use with glycolipid markers for disease prediction remains underexplored.Theory The CHG index captures glucose‑lipid dysregulation driving insulin resistance and inflammation. Given its modest discriminative ability, we constructed composite indices by multiplying CHG with eight metabolic markers (e.g., TyG, CVAI, and WC). This multiplicative approach improves discrimination by combining related biological parameters, a strategy supported by prior metabolic studies.Method This cross-sectional study included 9,125 participants derived from the NHANES. Logistic regression analysis assessed the associations between CHG, CHG-related indices and DM, DKD. Receiver operating characteristic (ROC) curves showed predictive performance, restricted cubic spline (RCS) explored nonlinear relationships and subgroup analyses tested the stability of these associations.Results Elevated CHG was independently associated with higher risks of DM (OR = 3.01, 95% CI: 2.44–3.72) and DKD (OR = 2.98, 95% CI: 2.06–4.31). Among the eight composite indices, CHG-TyG and CHG-CVAI achieved the highest AUC values in the total population for DM (AUC = 0.75, 95% CI: 0.74–0.76) and DKD (CHG-TyG: AUC = 0.72, 95% CI: 0.70–0.75; CHG-CVAI: AUC = 0.75, 95% CI: 0.73–0.77). CHG-WC also showed good discriminative ability. RCS analyses revealed predominantly nonlinear relationships and subgroup analyses confirmed the robustness of these associations.Discussion These findings suggested that CHG-related indices, especially CHG-TyG, CHG-CVAI, and CHG-WC were associated with DM and DKD. Incorporating these indicators into risk stratification frameworks may help identify individuals at higher risk.