Retinal Neurosensory Function Assessed by Electroretinography Predicts Cognition in Older Adults with Diabetic Retinopathy or Advanced Heart Failure

Introduction The retina is the only directly visualizable part of the central nervous system (CNS), and researchers are increasingly examining retinal functions as a proxy measure of CNS structure and function in neuropsychiatric disorders. Ophthalmological techniques such as electroretinography (ERG) test electrical responses to light stimuli of each retinal neural cell type and rapidly measures neural function in the retina while optical coherence tomography (OCT) and OCT Angiography (OCTA) generate high resolution images of retinal neural layers and visualize the retinal microvasculature, respectively. Combined, these exams allow for a reliable and rapid assessment of retinal structure and function as potential CNS biomarkers for neuropsychiatric symptoms in stroke, multiple sclerosis, Alzheimer's disease and schizophrenia. Previous longitudinal studies have demonstrated an association between retinal diseases and increased risk of dementia. ERG, OCT, and OCTA exams are known to be abnormal in diabetic retinopathy, but few studies have examined retinal neurosensory functions as potential biomarkers for cognitive decline in patients with diabetic retinopathy. Additionally, patients with heart failure (HF) are known to have a high prevalence of cognitive dysfunction, postulated to be mediated by poor cerebral perfusion, but few studies have utilized ERG, OCT, and OCTA to examine retinal neurovascular function as a proxy measure of the CNS neurovascular function related to cognitive dysfunction in heart failure. Methods We analyzed the baseline data of an ongoing prospective study that compares the associations among retinal health, cerebral blood flow, and neuropsychiatric symptoms in three groups of subjects: 1. Non-proliferative diabetic retinopathy (NPDR Group), 2. Advanced HF (NYHA Class II-IV) on medical treatment (HF Group), and 3. Advanced HF on left-ventricular assist device (LVAD group). This preliminary analysis based on the first 30 subjects (61.5 ± 12.8 years; 23% female) includes only the NPDR group (n=17) and the HF group (n = 13). Exclusion criteria included a history of delirium, stroke, or dementia. We used Pearson correlations to determine the association between Montreal Cognitive Assessment (MoCA), Beck Depression Inventory – Fast Screen (BDI-FS) and Generalized Anxiety Disorder (GAD)-7 scores, and measurements from retinal exams using ERG, OCT, and OCTA. Hierarchical regression analysis was used to determine the degree to which ERG indices predicted MoCA scores over and above age and years of education. Results There were no significant differences between the HF and NPDR groups on sociodemographic characteristics (age, gender, race, marital status and employment) and neuropsychic measures (MoCA, BDI-FS, and GAD-7). Interestingly, retinal exam abnormalities were just as common among HF group as they were in NPDR groups (Table 2). In both groups, ERG values (flicker amplitude; photopic negative response amplitude (PhNR)) were significantly correlated with total MoCA scores (Table 3). However, flicker amplitude and PhNR did not account for significant variance in MoCA scores over and above age and years of education (Table 4). No significant correlation was observed between other retinal markers (OCT, OCTA) and BDI-FS, GAD-7, or MoCA. Conclusions Neurovascular structure and functions of retina measured by ERG, OCT, and OCTA are similarly impaired in HF patients as they are in NPDR patients. Retinal neurosensory function measured by ERG is significantly associated with severity of cognition impairment among both HF and NPDR patients. ERG is a reliable, quick exam that takes minutes that can be performed in ophthalmology offices, and it may offer a simple, objective biomarker to predict cognitive impairment. This research was funded by Chernowitz Medical Research Foundation