Comparison of Novel Metabolic Indices in Estimation of Chronic Kidney Diseases in a Southern Chinese Population

BACKGROUND: To determine the optimal cut-off values and evaluate the associations of product of triacylglycerol and glucose (TyG), lipid accumulation product (LAPI), visceral adiposity index (VAI) with chronic kidney diseases (CKD) stratified by sex. METHODS: From January to April 2018, our team had conducted a large-scale cross-sectional survey that contained 2720 individuals on the southern coast of China. Logistic regression analysis and receiver operating characteristic (ROC) analyses were used to evaluate the optimal cut-off and value of TyG, LAPI, VAI for predicting CKD. RESULTS: A multivariate logistic regression analysis found that the TyG had the better value of prediction for the presence of CKD for the highest quartile vs the lowest quartile in both males (OR: 3.65; 95% CI, 2.04–6.52; p<0.001) and females (OR: 3.50; 95% CI, 2.20–5.56; p<0.001), followed by LAPI and VAI, when further adjusted for cofounder factors, LAPI and VAI both lost their independence, and only TyG remains its significant association with CKD in both males (OR: 2.81; 95% CI, 1.25–6.30; p<0.001) and females (OR: 3.22; 95% CI, 1.56–6.61; p<0.001). ROC curve showed that TyG had the highest AUC for predicting CKD in males (AUC: 0.618). TyG (AUC: 0.670) and LAPI (AUC: 0.670) both had the highest AUC in females. United predicted models which contain TyG were conducted for predicting CKD in males (AUC: 758) and females (AUC: 0.773) and results indicated that multivariate analysis of TyG and other traditional factors can impressively improve the accuracy of predictive probability for CKD. CONCLUSION: TyG is a priority to the other two novel indices and may become valuable makers and have strong predictive power for predicting CKD, especially in females.