Characterization of Retinal Architecture in Spinocerebellar Ataxia Type 3 and Correlation with Disease Severity

Neurodegeneration affects the brain and peripheral nervous system in spinocerebellar ataxia type 3 (SCA3). As the retina is also involved, studying the retinal architecture in a cohort of patients could reveal clinically relevant biomarkers.The aim is to investigate retinal architecture in SCA3 to identify potential biomarkers.We evaluated 38 patients with SCA3 and 25 healthy age-matched controls, who underwent visual acuity assessment, intraocular pressure measurement, and fundoscopy and macular and peripapillary spectral domain optical coherence tomography (SD-OCT). We measured the peripapillary retinal nerve fiber layer (pRNFL) thickness in each quadrant of the temporal-superior-nasal-inferior-temporal chart and the macular layer thicknesses in each sector of the inner circle of the Early Treatment Diabetic Retinopathy Study (IC-ETDRS) grid. Linear regression analysis was employed to test the associations between retinal parameters and age, disease duration, CAG repeats, and SARA (Scale of the Assessment and Rating of Ataxia) and ICARS (International Cooperative Ataxia Rating Scale) scores in SCA3.In all sectors, except for the temporal quadrant, pRNFL was significantly thinner in SCA3 patients than in controls. Average total macular, ganglion cell layer (GCL), and inner plexiform layer (IPL) thicknesses were significantly decreased in SCA3 patients in comparison to controls. The average total macular thickness and the average thicknesses of RNFL, GCL, and IPL negatively correlated with ICARS scores, whereas average GCL and IPL thicknesses negatively correlated with SARA scores.The retinal ganglion cells, their dendrites, and axons are selectively affected in SCA3 patients. The RNFL, GCL, and IPL thicknesses in SD-OCT correlate with the clinical phenotype and represent potential biomarkers for future clinical trials and natural history studies. © 2021 International Parkinson and Movement Disorder Society.