The association between estimated glucose disposal rate and the prevalence and mortality of chronic kidney disease: a cross-sectional study with linked mortality follow-up.

BACKGROUND: Metabolic disorders represented by insulin resistance (IR) are at risk of chronic kidney disease (CKD). Estimated glucose disposal rate (eGDR) reflects IR. The relationship between eGDR and CKD was unclear. This study aimed at discussing the association between eGDR and the prevalence of CKD in general population and the mortality of CKD patients, and compare it with other IR indicators. METHODS: The data from the National Health and Nutrition Examination Survey (NHANES) were used to conduct a cross-sectional study with linked mortality follow-up. The association between eGDR and CKD prevalence was determined using logistic regression, restricted cubic spline (RCS) analysis and stratified analysis. Receiver-operating characteristic (ROC) curves, weighted quantile sum (WQS) model, random forest and extreme gradient boost (XGBoost) machine learning models were performed to explore the importance of IR indicators components and CKD risk factors. The association between eGDR and mortality was analyzed by sub-distribution hazard model in CKD patients. RESULTS: Among 29,621 participants finally included, the median eGDR was 8.64 mg/kg/min and the CKD prevalence was 12.47%. Logistic regression and stratified analysis showed eGDR was associated with CKD prevalence independently, especially in people aged 40-60 years, with overweight or impaired glucose tolerance. RCS curve indicated the association between decreased eGDR and increased CKD risk was a U-shaped curve. ROC analysis showed eGDR assessed the CKD prevalence better. WQS model implied blood glucose control level was the main influencing factor in IR components. In machine learning models, the weights of age, eGDR, uric acid and heart failure were high. During 71 months follow-up, the all-cause mortality was 23.33% and cardiovascular disease (CVD) mortality was 8.9%. Sub-distribution hazard model showed eGDR independently predicted all-cause mortality rather CVD mortality in CKD patients after adjusting for confounding factors. CONCLUSIONS: eGDR was a better indicator to assess CKD risk in general population and could predict all-cause mortality rather CVD mortality in CKD patients.