Evaluation of kidney biological age based on IgG N-glycosylation profiles and its association with chronic kidney disease in the context of predictive, preventive and personalized medicine
Abstract
Background: Chronic kidney disease (CKD) is associated with premature aging, which reflects in the difference between biological age and chronological age. Immunoglobulin G (IgG) N-glycosylation profiles have emerged as promising biomarkers for biological aging. From the perspective of predictive, preventive, and personalized medicine (PPPM/3PM), we assumed that the evaluation of kidney-specific biological age based on IgG N-glycosylation profiles provides a better tool for targeted prevention and personalized intervention of CKD by monitoring kidney aging. Methods: This study analyzed data from the Beijing Health Management Cohort. Plasma IgG N-glycosylation profiles were quantified into 24 glycan peaks (GPs), and feature selection was conducted using adaptive elastic net followed by logistic regression. IgG N-glycosylation kidney biological age (GlyKage) was calculated using linear regression, and the difference between GlyKage and chronological age (GlyKageDiff) was calculated. The associations of GlyKage and GlyKageDiff with CKD were evaluated using the adjusted multivariable logistic regression. Odds ratio (OR) and 95% confidence interval (CI) were calculated. Diagnostic models were developed using an 8:2 train-test split of the dataset, by incorporating different predictor variables and using support vector machine, XGBoost, and LightGBM. Results: From 3123 participants with blood samples collected during 2014–2015, we selected 2382 participants for evaluating GlyKage and developing the CKD diagnostic models. We selected four GPs associated with CKD, included GP3, GP11, GP13, GP24, to calculate GlyKage. In adjusted models, each one-unit increase of GlyKage was associated with higher CKD risk (OR = 1.136, 95% CI: 1.114–1.159), and individuals with high GlyKageDiff (the top 25% values) had a higher CKD risk (OR = 12.179, 95% CI: 6.530–25.364). In the test set of the diagnostic model, compared to chronological age, GlyKage and GlyKageDiff increased the area under curve (AUC) value by 10.8% and 15.2% respectively. The AUC value of the final model was 0.945 (95% CI: 0.916–0.971). Conclusions: GlyKage and GlyKageDiff are associated with a higher risk of CKD, and show a substantial value in the diagnosis of CKD. In the context of PPPM/3PM, evaluating GlyKage helps identify individuals at high risk for CKD, facilitating early intervention and management. Furthermore, evaluating an individual’s kidney aging status meets the need for personalized healthcare.
Keywords
Kidney-specific biological age, chronic kidney disease, feature selection, IgG n-glycosylation kidney biological age (GlyKage), machine learning, organ-specific aging, personalized healthcare, predictive preventive and personalized medicine (PPPM / 3PM)
Document Type
Journal Article
Date of Publication
1-1-2026
Publication Title
EPMA Journal
Publisher
Springer
School
School of Medical and Health Sciences
Funders
National Natural Science Foundation of China (82373683) / Program of Health Scientific Fund in the Capital (2024-1G-4261)
Copyright
subscription content
Comments
Liu, H., Lv, S., Li, H., Yang, R., Tian, Y., Wang, J., Li, X., Chen, S., Zhang, G., & Guo, X. (2026). Evaluation of kidney biological age based on IgG N-glycosylation profiles and its association with chronic kidney disease in the context of predictive, preventive and personalized medicine. EPMA Journal. Advance online publication. https://doi.org/10.1007/s13167-026-00440-z