Abstract: TH-PO0489
Inflammatory Cytokines Associated with Apparent Treatment-Resistant Hypertension in the Maintenance Hemodialysis Population: Insights from Machine Learning
Session Information
- Hemodialysis: Novel Markers and Case Reports
November 06, 2025 | Location: Exhibit Hall, Convention Center
Abstract Time: 10:00 AM - 12:00 PM
Category: Dialysis
- 801 Dialysis: Hemodialysis and Frequent Dialysis
Authors
- Luo, Yuan, Center for Kidney Diseases, the Second Affiliated Hospital of Nanjing Medical University, Nanjing, Jiangsu, China
- Ye, Hong, Center for Kidney Diseases, the Second Affiliated Hospital of Nanjing Medical University, Nanjing, Jiangsu, China
- Dai, Chunsun, Center for Kidney Diseases, the Second Affiliated Hospital of Nanjing Medical University, Nanjing, Jiangsu, China
Background
The prevalence of apparent treatment-resistant hypertension (aTRH) among individuals undergoing maintenance hemodialysis (MHD) ranges from 18% to 42%, significantly increasing cardiovascular risks and mortality. Chronic inflammation may contribute to resistant hypertension; however, the specific inflammatory cytokine profiles associated with aTRH in the dialysis population remain unclear.
Methods
The discovery cohort included 78 participants from the CARE-MHD study (February–May 2022), while the validation cohort consisted of 168 participants from the PURE-HD study (October 2023–April 2024). Twenty-seven cytokines were measured using a multiplex immunoassay (discovery) and validated by ELISA (validation). Participants were classified into aTRH, non-aTRH hypertension, and normotension groups. Logistic regression (LR) and machine learning (ML) models assessed cytokine associations and predictive capabilities.
Results
Among the analyzed cytokines, IL-4, IL-9, Eotaxin, and RANTES were significantly elevated in aTRH participants compared to both non-aTRH hypertensive and normotensive groups. An LR model incorporating all four cytokines exhibited the best variance explanation ability (adjusted Cox and Snell R2 = 0.411) among other LR models, confirming the association between these cytokines and aTRH. LR analyses demonstrated the strong predictive potential of RANTES and Eotaxin. A multivariate model incorporating all four cytokines showed superior predictive performance. ML models consistently identified RANTES and Eotaxin as critical predictive biomarkers for aTRH. Additionally, protein-protein interaction network analysis revealed that all four cytokines are involved in inflammatory pathways linked to vascular dysfunction and immune cell activation.
Conclusion
A distinct inflammatory signature (IL-4, IL-9, Eotaxin, RANTES) is associated with aTRH in the MHD population. Among these, RANTES and Eotaxin may serve as non-invasive biomarkers for early diagnosis and personalized therapeutic strategies in this high-risk population.
Funding
- Government Support – Non-U.S.