Machine Learning Methods to Analyze the Relationship Between Dietary Trace Elements and Kidney Stones: From NHANES 2007–2020

The early detection of kidney stones usually improves treatment outcomes. This cross-sectional survey involved 33,267 adult participants. The aim was to investigate the relationship between dietary elements (P, Mg, Cu, K, Ca, Zn) and kidney stones. Studies have shown that an increase in trace minerals in the diet is linked to a decrease in the frequency of kidney stones. Specific saturation effects for dietary elements and kidney stones were determined using smoothed curve fitting and found to have an effect inflection point of P, Mg, Cu, K, Ca, Zn. Subgroup analyses showed that the majority of the dietary trace metal element intake remained consistent with the correlation with kidney stone risk. The key dietary elements identified using Boruta and LASSO regression were Cu, Mg, P, and Ca. Among them, the element Mg had a significant effect on the column-line model, and the ROC of the established nomogram model was 0.718 for the training set and 0.702 for the validation set, which demonstrates moderate discriminative ability, stability, and effectiveness, significantly enhancing the prediction accuracy of the association between trace metal element deficiencies and nephrolithiasis. This study suggests that adequate dietary intake or supplementation of these elements may aid in preventing kidney stones.