Scholarlink

Background <p>The Zhejiang University (ZJU) Index has emerged as a comprehensive metabolic indicator and demonstrated significant association with various diseases. The goal of this study was to investigate the potential relationship between ZJU index and kidney stones.</p> Methods <p>A cross-sectional study analyzed participants’ demographic, socioeconomic, and laboratory data from NHANES 2007–2018. Weighted multivariate logistic regression, restricted cubic spline (RCS) models, and stratified analysis were applied to validate the relationship between ZJU index and kidney stone. Machine learning based analysis was employed to further improve the predictive performance and identify key predictors.</p> Results <p>A total of 11,317 participants were enrolled in our study and 1,115 were classified as kidney stone former. Significant differences were observed between the kidney stone formers and non-kidney stone formers in variables such as gender, race, age, education, marital status, recreational activities, hypertension, diabetes mellitus and BMI. Weighted logistic regression analysis revealed a significant positive association between ZJU index and kidney stone risk (OR = 1.03, 95% CI: 1.01–1.04) after maximal adjustment for the covariates. Participants in the highest ZJU tertile faced a 74% higher odds of nephrolithiasis than those in the lowest tertile (OR = 1.74, 95% CI: 1.34–2.26). RCS analysis indicated ZJU index raise the risk of stone formation in a non-linear dose-response manner. In the stratified analysis, we observed that the positive association was maintained across most subgroups, except for individuals younger than 40 or from other race. A significant interaction between ZJU index and marital status was detected (<i>P</i><sub>interaction</sub>=0.042). Among the three machine learning models, XGBoost model exhibited the best predictive performance, with an area under the curve (AUC) of 0.638. SHAP analysis ranked ZJU index as the most influential predictor for nephrolithiasis.</p> Conclusion <p>Our study provided additional evidence supporting the role of ZJU index as an effective metabolic biomarker for kidney stone risk prediction. Further clinical and epidemiological study should be warranted to unveil a more precise cause-effect relationship between them.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

The association between ZJU index and kidney stone risk: a machine learning approach on NHANES 2007–2018

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年3月11日

查看全文
  • Yiwei Lin,
  • Jiatong Zhou,
  • Jianchen Lv,
  • Baihua Shen

摘要

Background

The Zhejiang University (ZJU) Index has emerged as a comprehensive metabolic indicator and demonstrated significant association with various diseases. The goal of this study was to investigate the potential relationship between ZJU index and kidney stones.

Methods

A cross-sectional study analyzed participants’ demographic, socioeconomic, and laboratory data from NHANES 2007–2018. Weighted multivariate logistic regression, restricted cubic spline (RCS) models, and stratified analysis were applied to validate the relationship between ZJU index and kidney stone. Machine learning based analysis was employed to further improve the predictive performance and identify key predictors.

Results

A total of 11,317 participants were enrolled in our study and 1,115 were classified as kidney stone former. Significant differences were observed between the kidney stone formers and non-kidney stone formers in variables such as gender, race, age, education, marital status, recreational activities, hypertension, diabetes mellitus and BMI. Weighted logistic regression analysis revealed a significant positive association between ZJU index and kidney stone risk (OR = 1.03, 95% CI: 1.01–1.04) after maximal adjustment for the covariates. Participants in the highest ZJU tertile faced a 74% higher odds of nephrolithiasis than those in the lowest tertile (OR = 1.74, 95% CI: 1.34–2.26). RCS analysis indicated ZJU index raise the risk of stone formation in a non-linear dose-response manner. In the stratified analysis, we observed that the positive association was maintained across most subgroups, except for individuals younger than 40 or from other race. A significant interaction between ZJU index and marital status was detected (Pinteraction=0.042). Among the three machine learning models, XGBoost model exhibited the best predictive performance, with an area under the curve (AUC) of 0.638. SHAP analysis ranked ZJU index as the most influential predictor for nephrolithiasis.

Conclusion

Our study provided additional evidence supporting the role of ZJU index as an effective metabolic biomarker for kidney stone risk prediction. Further clinical and epidemiological study should be warranted to unveil a more precise cause-effect relationship between them.