Association between exposure to a mixture of metals and chronic kidney disease: Comparison of statistical models

Abstract

Abstract Previous studies have identified several genetic and environmental risk factors for chronic kidney disease (CKD), but little is known about the relationship between blood metals and CKD risk. Herein, we examined associations between serum levels of metals and the risk of CKD among 100 medical examiners and 443 patients with CKD participating in the medical center of the First Hospital Affiliated to China Medical University. Therefore, we aimed to conduct some statistical approaches, machine learning, logistics regression, Bayesian Kernel Machine Regression (BKMR) and serial mediation model, to explore the prediction and effect of metals exposure on CKD. In this cross-sectional study, the concentrations serum of metals mixtures was measured using inductively coupled plasma mass spectrometry (ICP-MS). The result suggested that exposure to K, Na and Ca lead to CKD increasing and Se and Mo lead to CKD decreasing. A significant negative effect of metal mixtures on CKD when metal mixtures concentrations were all from 30th to 45th percentile compared to the median, whereas the opposite was true for the 55th to 60th percentiles. A change in blood K concentration from the 25th to the 75th percentile is associated with a significant increase in CKD disease of 5.15(1.77,8.53), 13.62(8.91,18.33) and 31.81(14.03,49.58) when other metals are fixed at the 25th, 50th and 75th percentiles, respectively. Finally, our findings suggested that metal cumulative exposures and particularly double-exposure of K and Se might impact CKD. Machine learning could verify the external validation of the metal factors.