Abstract: SA-PO1195
Characterizing the Serum Metabolome Landscape in Nondiabetic Patients with CKD
Session Information
- CKD: Biomarkers and Emerging Tools for Diagnosis and Monitoring
November 08, 2025 | Location: Exhibit Hall, Convention Center
Abstract Time: 10:00 AM - 12:00 PM
Category: CKD (Non-Dialysis)
- 2302 CKD (Non-Dialysis): Clinical, Outcomes, and Trials
Authors
- Zhou, Yuqi, University of Connecticut, Storrs, Connecticut, United States
- Gui, Yuan, University of Connecticut School of Medicine, Farmington, Connecticut, United States
Background
By 2040, CKD is projected to become the fifth leading cause of death globally. Currently, CKD diagnosis relies primarily on kidney biopsy and conventional urine or serum biomarkers such as creatinine. This study aims to characterize the serum metabolomic landscape in individuals with non-diabetic CKD to uncover novel biomarkers and provide insights for more precise monitoring of disease progression.
Methods
A total of 989 participants were enrolled from a single medical center, including 463 healthy controls and 526 CKD patients (stages 1–4). Individuals with a history of diabetes or undergoing dialysis were excluded. Serum samples were collected and subjected to large-scale untargeted metabolomics. Bioinformatics tools and five machine learning algorithms were applied to analyze the data.
Results
A total of 84 differential serum metabolites were identified across CKD stages. However, only 6 and 12 metabolites were altered in stages 1 and 2, respectively, changes that preceded detectable increases in serum creatinine. From stage 3 onward, the number of differential metabolites markedly increased, correlating with elevated serum creatinine levels. Among all stages, glyceric acid was consistently and significantly decreased, suggesting disturbances in metabolic regulation, redox balance, and energy homeostasis. Pathway analysis revealed that pentose and glucuronate interconversions, which are involved in nucleotide biosynthesis, detoxification, and cellular repair, were the most perturbed in stages 1–3, reflecting early compensatory metabolic adaptations. In stage 4, a distinct shift toward glyoxylate and dicarboxylate metabolism was observed, signaling mitochondrial dysfunction and impaired organic acid handling characteristic of advanced metabolic stress. Importantly, glycerol-3-galactoside and D-tagatose levels were significantly changed in stages 3 and 4 and outperformed serum creatinine in predicting CKD progression across multiple machine learning models if combined. In addition, by stage 4, the serum metabolome profiles of diabetic and non-diabetic CKD patients converged, suggesting a common metabolic endpoint in advanced kidney disease.
Conclusion
Our study highlights the potential of metabolomics to uncover novel biomarkers and pathogenic pathways in CKD progression. The dataset provides a valuable, open-access resource to support improved clinical monitoring and management of CKD.