Abstract: FR-PO0329
Development and Validation of Cluster-Specific Risk Stratification Models for Osteoporosis in Patients with CKD Based on NHANES Data
Session Information
- CKD-MBD: Mechanistic Insights and Management
October 23, 2026 | Location: Exhibit Hall A, Convention Center
Abstract Time: 10:00 AM - 12:00 PM
Category: Bone and Mineral Metabolism
- 502 Bone and Mineral Metabolism: Clinical
Authors
- Liu, Zicheng, The Third Affiliated Hospital of Southern Medical University, Guangzhou, Guangdong, China
- Lu, Yi, The Third Affiliated Hospital of Southern Medical University, Guangzhou, Guangdong, China
- Chen, Junzhe, The Third Affiliated Hospital of Southern Medical University, Guangzhou, Guangdong, China
- Zhou, Lili, Southern Medical University Nanfang Hospital, Guangzhou, Guangdong, China
- Nie, Sheng, Southern Medical University Nanfang Hospital, Guangzhou, Guangdong, China
- Tang, Ying, The Third Affiliated Hospital of Southern Medical University, Guangzhou, Guangdong, China
Background
CKD-mineral and bone disorder (CKD-MBD) exhibits substantial heterogeneity. Conventional uniform risk stratification approaches fail to capture distinct pathophysiological phenotypes. We developed cluster-specific predictive models for osteoporosis risk in CKD patients using routine clinical parameters.
Methods
Data were obtained from the NHANES database. A total of 3,806 CKD patients were enrolled and randomly allocated to training (70%, n=2,667) and validation (30%, n=1,139) cohorts. K-means clustering was performed to identify distinct subgroups. Weighted univariate logistic regression, Lasso and Boruta were employed for feature selection. Final risk stratification models were constructed using weighted multivariable logistic regression. Model performance was assessed using receiver operating characteristic (ROC) curves, calibration curves, and decision curve analysis (DCA).
Results
Three distinct CKD phenotypic clusters were identified, of which two with adequate sample sizes were selected for model development. For Group A, six variables (body mass index, serum creatinine, high-density lipoprotein, alkaline phosphatase, red blood cell count, and fracture history) were retained as final risk factors, achieving an AUC of 0.880 in the training cohort and 0.839 in the validation cohort. For Group B, seven variables (age, body mass index, height, high-density lipoprotein, alkaline phosphatase, estimated glomerular filtration rate, and fracture history) were selected, yielding an AUC of 0.824 in training and 0.806 in validation. Both models demonstrated favorable calibration and clinical utility as evidenced by DCA.
Conclusion
Cluster-specific risk stratification models based on weighted logistic regression exhibited robust discriminative performance for osteoporosis risk stratification in CKD patients.
Acknowledgment
We acknowledge the data from the NHANES.