Chronic kidney disease (CKD) is a growing global public health concern. This study aimed to identify and compare demographic factors and blood biomarkers associated with CKD in hypertensive and non-hypertensive individuals. A cross-sectional analysis was conducted using data from the 2009 China Health and Nutrition Survey. CKD was defined as an estimated glomerular filtration rate 2. After propensity score matching, 1,823 hypertensive and 1,823 non-hypertensive participants were included. The prevalence of CKD was higher in hypertensive individuals (14.7%) than in non-hypertensive individuals (10.8%). Although age, female sex, hyperuricemia, elevated total protein, soluble transferrin receptor, blood urea nitrogen, and low transferrin were shared factors across groups, logistic regression and decision tree analyses revealed different group-specific variables: urban residence and low education level were more relevant in hypertensive individuals, while elevated glucose was more important in non-hypertensives. The extreme gradient boosting (XGBoost) with SHapley Additive exPlanations (SHAP) analyses further showed that triglyceride and residence had greater importance in hypertensive individuals, whereas glucose and magnesium were more influential in non-hypertensive individuals. The area under the receiver operating characteristic curve (AUC) values of the XGBoost model in hypertensive and non-hypertensive individuals using the top 10 variables were 0.930 (95% CI: 0.915-0.945) and 0.945 (95% CI: 0.933-0.958), respectively. In conclusion, although CKD shares several common factors across groups, the relative importance and distribution of predictors differ between hypertensive and non-hypertensive individuals. Stratified monitoring based on hypertension status may enhance early detection and targeted prevention of CKD.
Zhu et al. (Wed,) studied this question.