Associations between multimodal retinal measurements and cognitive functions in patients with cerebral small vessel disease

Purpose To investigate the associations between retinal parameters and cognitive impairment and identify retinal biomarkers for detecting cognitive dysfunction in patients with cerebral small vessel disease (CSVD). Methods This cross-sectional study enrolled 125 patients with CSVD-related white matter hyperintensities. Participants were independent in daily activities and free of significant ophthalmic diseases. Cognition was assessed using the Mini-Mental State Examination (MMSE). Multimodal retinal imaging was employed to evaluate retinal structural and vascular features, including peripapillary retinal nerve fibre layer (pRNFL) and ganglion cell complex thicknesses and vessel densities (VDs) of the radial peripapillary capillary (RPC) network, macular superficial retinal capillary plexus (SRCP), and deep RCP (DRCP). The central retina vascular diameter, mean angle, angle tortuosity, arc length tortuosity, and fractal dimension of the retinal arteries and veins were quantified using a fully automated vessel segmentation and parameter calculation method. Linear and logistic regression models were applied to assess the associations between retinal parameters with general cognitive domains (general cognition, memory, visuospatial space, attention and numeracy, language, and orientation). Receiver operating characteristic (ROC) curves and areas under the ROC curve (AUCs) were used to evaluate the predictive performance of the retinal biomarkers. Results A smaller central retinal arteriolar equivalent/central retinal venular equivalent ratio [odds ratio (OR) = 0.013, p = 0.035] and lower VDs in whole macular SRCP (OR = 0.864, p = 0.029), perifoveal SRCP (OR = 0.862, p = 0.027), perifoveal superior SRCP (OR = 0.855, p = 0.014), perifoveal nasal SRCP (OR = 0.833, p = 0.008), and inside-disc capillary RPC network (OR = 0.901, p = 0.028) were significantly associated with lower MMSE scores. Attention and orientation scores were significantly correlated with the VDs of the SRCP ( β = 0.734, p = 0.027 and β = −9.460, p = 0.037, respectively) and DRCP ( β = −0.553, p = 0.004 and p = 0.044, respectively), whereas visuospatial and language function scores were associated with the VD of the inside-disc RPC network ( β = 0.018, p = 0.026 and β = 0.068, p = 0.012, respectively). The VDs of the whole macular, perifoveal, perifoveal superior, and nasal SRCPs achieved AUCs of 0.71–0.74 for screening cognitive impairment. Conclusion Specific retina parameters are associated with cognitive decline in patients with CSVD. These findings suggest that multimodal retinal evaluation might provide an objective, imaging-based adjunct to conventional subjective cognitive tests for screening cognitive impairment in patients with CSVD. Clinical trial registration https://www.chictr.org.cn/ , identifier ChiCTR 2100043346.