Abstract: SA-PO0403
Claims-Based Modeling Identifies High-Risk Cardiovascular-Kidney-Metabolic Profiles in CKD Without Laboratory Data
Session Information
- CKM: Clinical - Epidemiology and Outcomes
October 24, 2026 | Location: Exhibit Hall A, Convention Center
Abstract Time: 10:00 AM - 12:00 PM
Category: Cardiovascular-Kidney-Metabolic Health
- 602 Cardiovascular-Kidney-Metabolic Health: Clinical
Authors
- Schwab, Sebastian, Rheinische Friedrich-Wilhelms-Universitat Bonn, Bonn, NRW, Germany
- Stoy, Volquart, AOK Rheinland Hamburg, Düsseldorf, NRW, Germany
- Scheidt, Uta, Rheinische Friedrich-Wilhelms-Universitat Bonn, Bonn, NRW, Germany
- Raudzus, Janika, Universitat Heidelberg, Heidelberg, BW, Germany
- Mayrhofer, Thomas, Universitat Heidelberg, Heidelberg, BW, Germany
- Von Vietinghoff, Sibylle, Rheinische Friedrich-Wilhelms-Universitat Bonn, Bonn, NRW, Germany
Background
Chronic kidney disease (CKD) is a central component of the cardiovascular–kidney–metabolic (CKM) syndrome, yet integrated risk stratification in real-world settings is often limited by missing laboratory data. We assessed whether claims data reflect CKM risk and predict adverse renal and cardiovascular outcomes.
Methods
We studied 29,144 individuals with CKD stage III–IV insured by a German statutory health insurance in 2016. Claims-derived variables were used to predict renal replacement therapy (RRT), mortality, acute kidney injury (AKI), heart failure hospitalization, and ICU admission over six years. Discrimination was evaluated by AUC.
Results
CKD IV patients exhibited a more advanced cardiovascular comorbidity profile with higher cardiovascular comorbidity burden. During follow-up, mortality (52.3% vs. 66.4%), RRT (3.8% vs. 14.3%), AKI (7.0% vs. 13.3%), and heart failure hospitalizations were consistently higher in CKD IV than CKD III. Claims-based models showed high discrimination for mortality (AUC 0.9783) and RRT (0.8715), and moderate performance for heart failure hospitalization (0.7664), AKI (0.6817), and ICU admission (0.6744). Addition of CKD stage to the equation significantly improved all models and independently increased risks of renal and cardiovascular events. Engagement in structured care and longitudinal monitoring was associated with lower risks across CKM outcomes. Renin–angiotensin system inhibitors and SGLT2 inhibitors associated with reduced renal and cardiovascular risk, whereas diuretic therapy identified higher-risk phenotypes. Claims data further captured CKD-related metabolic complications and were associated with subsequent treatment with good discrimination.
Conclusion
Administrative claims data reflect the cardiovascular–kidney–metabolic risk profile in CKD and enable prediction of renal and cardiovascular outcomes in the absence of laboratory data. This approach supports scalable risk stratification in real-world populations.