Association between retinal vessel density and neuroimaging features and in cerebral small vessel disease

Aims: This study aimed to investigate the associations among retinal vessel density (RVD), neuroimaging features and cognitive impairment in patients with sporadic cerebral small vessel disease (CSVD). Methods: This was a prospective observational study. A total of forty-nine patients with CSVD were recruited. The CSVD imaging burden was calculated by using a scoring system with a total score of 4 that assigns one point each for severe white matter hyperintensities (WMH), lacune, microbleeds (MBs), and basal-ganglia perivascular space (BG-PVS). Patients with a burden score >= 2 were classified as having a moderate/severe burden, and those with a score <= 1 were classified as the having a none/mild burden. The RVD in the superficial retinal capillary plexus (SRCP) and deep retinal capillary plexus (DRCP) was evaluated by using optical coherence tomography angiography (OCTA). The associations among the RVD values, CSVD imaging features, and cognitive impairment were evaluated. Results: Patients with a moderate/severe CSVD burden showed lower RVD values in the para-fovea and peri-fovea areas of the left DRCP than patients with none/mild burden (para-fovea, 8 coefficient= -0.185 [-0.351--0.015], P = 0.003; peri-fovea, 8 coefficient= -0.113 [-0.208--0.018], P = 0.021). The RVD values in the para-fovea and peri-fovea areas of the left DRCP were significantly associated with the CSVD burden score after adjusting for age and vascular risk factors (P = 0.030 and P = 0.021, respectively) and with severe WMH (para-fovea, R = -0.398, P = 0.005; peri-fovea, R= -0.443, P = 0.001) and BG-PVS (para-fovea, R = -0.445, P = 0.001; peri-fovea, R= -0.396, P = 0.005). Neither para-fovea nor peri-fovea RVD values had a marked association with cognition. Conclusion: The RVD in the left DRCP may reflect the presence of cerebral small vessel lesions and might be a useful tool for predicting the neuroimaging-based burden of CSVD.