Objective To explore the predictive value of macular volume, central macular thickness, and retinal nerve fiber layer (RNFL) thickness on the occurrence risk of cognitive dysfunction in elderly patients with leukoaraiosis (LA). Methods A total of 122 elderly inpatients from the Department of Neurology were selected as the research subjects. According to the diagnostic criteria for white matter hyperintensities and combined with MRI findings, patients who presented with symmetric punctate, patchy, or linear long T1 and long T2 abnormal hyperintensities around the anterior horns, posterior horns, and bodies of the lateral ventricles on T2WI and fluid⁃attenuated inversion recovery sequences, without obvious periventricular white matter atrophy or bilateral ventricular enlargement, were assigned to the LA group (n=82); patients without obvious abnormal hyperintensities in the periventricular and deep white matter regions on T2WI and fluid⁃attenuated inversion recovery sequences, and without other imaging features associated with white matter lesions, were assigned to the control group (n=40). Cognitive function (the Chinese version of Montreal Cognitive Assessment ［MOCA］ scale), cognitive domain (Chinese Auditory Verbal Learning Test ［CAVLT］, Symbol Digit Modalities Test ［SDMT］, Color Trails Test ［CTT］, the Chinese version of Boston Naming Test⁃15 ［BNT⁃15］), mental state (Mini⁃Mental State Examination ［MMSE］ scale, Hamilton Depression Scale ［HAMD］, Hamilton Anxiety Scale ［HAMA］), and the optical coherence tomography parameters (macular volume, CMT, and RNFL thickness) were compared between patients of the two groups. Pearson analysis was employed to analyze the correlation between age, cognitive function, cognitive domain, and optical coherence tomography parameters in LA patients. According to the presence of cognitive dysfunction or not, the elderly LA patients were further divided into cognitive dysfunction group (MOCA scale score<26, n=45) or normal cognitive function group (MOCA scale score≥26, n=37). Multivariate Logistic regression model was used to analyze the influencing factors for the occurrence of cognitive dysfunction in elderly patients with LA, and a risk prediction model for cognitive dysfunction was constructed. The receiver operating characteristic (ROC) curve was drawn to evaluate the predictive efficiency of the model. Results Compared with the control group, the CMT and RNFL thickness of the LA group were thinner, the macular volume was smaller, the scores of MOCA scale, CAVLT and MMSE scale were lower, the correct number of SDMT and BNT⁃15 was less, and the CTT completion time was longer (P<0.05). Pearson analysis results revealed that the MOCA scale score, CAVLT score, correct number of SDMT and correct number of BNT⁃15 positively correlated with CMT, macular volume and RNFL thickness in elderly LA patients, and age negatively correlated with CMT, macular volume and RNFL thickness. The completion time of CTT negatively correlated with macular volume and RNFL thickness (P<0.05). Compared with the normal cognitive function group, the cognitive dysfunction group was older, the macular volume was smaller, and the CMT and RNFL thickness were thinner (P<0.05). Multivariate Logistic regression analysis results indicated that age, macular volume, CMT, and RNFL thickness were independent influencing factors for the occurrence of cognitive dysfunction in elderly LA patients (P<0.05). The results of ROC curve analysis elucidated that the area under the curve of the risk prediction model for cognitive dysfunction in elderly LA patients constructed based on the above influencing factors was 0.843. Conclusion Elderly patients with LA have cognitive impairment in multiple domains, accompanied by decreased macular volume, and attenuated CMT and RNFL thickness. Age, macular volume, CMT and RNFL thickness are the influencing factors for the occurrence of cognitive dysfunction in elderly LA patients. The risk prediction model for cognitive dysfunction in elderly LA patients constructed based on the above influencing factors has a good prediction efficiency.

无