BDNF-related biomarkers and cognitive impairment in chronic kidney disease: associations and machine learning-based discrimination analysis
- 1Taizhou Fourth People’s Hospital, Taizhou, Jiangsu, China
- 2College of Pharmacy, Taizhou University, Taizhou, China
- 3Eastern Theater Command Centers for Disease Control and Prevention, Nanjing, China
Abstract
Background: Cognitive impairment is a common complication among patients with chronic kidney disease (CKD). However, biomarkers associated with cognitive impairment in this population remain insufficiently characterized. Brain-derived neurotrophic factor (BDNF) may be involved in neurobiological processes related to cognitive function, but its association with cognitive impairment among patients with CKD remains unclear.
Methods: In this cross-sectional study, 240 patients with CKD underwent cognitive assessment and serum BDNF measurement during the same clinical evaluation. Multivariable logistic regression models were used to examine the associations of BDNF and the BDNF/eGFR ratio with prevalent cognitive impairment. Restricted cubic spline analyses were performed to assess potential dose–response relationships, and subgroup analyses were conducted to explore potential effect modification. Additionally, nine machine learning algorithms were applied to evaluate the discrimination ability of BDNF-related indicators for distinguishing participants with and without cognitive impairment. SHapley Additive exPlanations (SHAP) analysis was used to interpret the contribution of variables in the best-performing model.
Results: Higher levels of BDNF and BDNF/eGFR ratio were significantly associated with lower odds of prevalent cognitive impairment after multivariable adjustment (BDNF: OR = 0.72, 95% CI: 0.63–0.82; BDNF/eGFR ratio: OR = 0.62, 95% CI: 0.45–0.82). Quartile analyses showed consistent inverse associations, and restricted cubic spline analyses indicated linear relationships. Significant effect modification was observed by sex for both markers and by stroke history for the BDNF/eGFR ratio. Among the evaluated machine learning approaches, the random forest model demonstrated the highest discrimination performance, with AUCs of 0.803 for BDNF and 0.844 for BDNF/eGFR ratio. SHAP analysis indicated that BDNF-related markers contributed substantially to the classification of cognitive impairment status.
Conclusion: BDNF and BDNF/eGFR ratio were inversely associated with prevalent cognitive impairment in patients with CKD. These findings suggest that BDNF-related markers may be associated with cognitive impairment status and provide potential biological insights into CKD-related cognitive dysfunction. Further longitudinal studies are required to determine whether these biomarkers can contribute to future risk assessment or clinical decision-making.
Related
- Complement C4b and factor B in astrocyte-derived exosomes as novel biomarkers for amnestic mild cognitive impairment
- Associations of physical function and multimorbidity with cognitive impairment in community-dwelling Chinese octogenarians
- Longitudinal associations between neuropsychiatric symptoms, structural and perfusion MRI changes in older adults with Mild Cognitive Impairment
- Associations Between the Brain Care Score and Structural Brain Changes and Cognitive Impairment Using Data From the Hamburg City Health Study
- EEG-centred computational biomarkers and complementary physiological signals for mild cognitive impairment and early Alzheimer’s disease: A systematic review 2005–2025
BibTeX
@article{li2026bdnfrelated,
title = {{BDNF-related biomarkers and cognitive impairment in chronic kidney disease: associations and machine learning-based discrimination analysis}},
author = {Xiaoran Li and Hong Qian and Wenwen Xiao},
journal = {Frontiers in Medicine},
year = {2026},
month = oct,
doi = {10.3389/fmed.2026.1951634},
url = {https://doi.org/10.3389/fmed.2026.1951634},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - BDNF-related biomarkers and cognitive impairment in chronic kidney disease: associations and machine learning-based discrimination analysis
AU - Xiaoran Li
AU - Hong Qian
AU - Wenwen Xiao
PY - 2026
DA - 2026/10/21
JO - Frontiers in Medicine
DO - 10.3389/fmed.2026.1951634
UR - https://doi.org/10.3389/fmed.2026.1951634
AB - Background: Cognitive impairment is a common complication among patients with chronic kidney disease (CKD). However, biomarkers associated with cognitive impairment in this population remain insufficiently characterized. Brain-derived neurotrophic factor (BDNF) may be involved in neurobiological processes related to cognitive function, but its association with cognitive impairment among patients with CKD remains unclear. Methods: In this cross-sectional study, 240 patients with CKD underwent cognitive assessment and serum BDNF measurement during the same clinical evaluation. Multivariable logistic regression models were used to examine the associations of BDNF and the BDNF/eGFR ratio with prevalent cognitive impairment. Restricted cubic spline analyses were performed to assess potential dose–response relationships, and subgroup analyses were conducted to explore potential effect modification. Additionally, nine machine learning algorithms were applied to evaluate the discrimination ability of BDNF-related indicators for distinguishing participants with and without cognitive impairment. SHapley Additive exPlanations (SHAP) analysis was used to interpret the contribution of variables in the best-performing model. Results: Higher levels of BDNF and BDNF/eGFR ratio were significantly associated with lower odds of prevalent cognitive impairment after multivariable adjustment (BDNF: OR = 0.72, 95% CI: 0.63–0.82; BDNF/eGFR ratio: OR = 0.62, 95% CI: 0.45–0.82). Quartile analyses showed consistent inverse associations, and restricted cubic spline analyses indicated linear relationships. Significant effect modification was observed by sex for both markers and by stroke history for the BDNF/eGFR ratio. Among the evaluated machine learning approaches, the random forest model demonstrated the highest discrimination performance, with AUCs of 0.803 for BDNF and 0.844 for BDNF/eGFR ratio. SHAP analysis indicated that BDNF-related markers contributed substantially to the classification of cognitive impairment status. Conclusion: BDNF and BDNF/eGFR ratio were inversely associated with prevalent cognitive impairment in patients with CKD. These findings suggest that BDNF-related markers may be associated with cognitive impairment status and provide potential biological insights into CKD-related cognitive dysfunction. Further longitudinal studies are required to determine whether these biomarkers can contribute to future risk assessment or clinical decision-making.
SN - 2296-858X
ER -