Optical coherence tomography angiography parameters in Marfan syndrome: Genetic determinants and associations with cardiovascular manifestations

Abstract

In addition to aortic manifestations, Marfan syndrome can affect retinal vessels. Our aim was to evaluate retinal circulation, and its correlation with genotype and cardiovascular manifestation identifying predictors of aortic involvement (dilation and/or dissection). In the retrospective, cross-sectional study, 39 Marfan syndrome patients with optical coherence tomography angiography records were included. Retinal thickness, superficial and deep vessel density in total retina, fovea, parafovea, perifovea, area and perimeter of foveal avascular zone, fractal dimension were measured. Two groups were created by mutation type: haploinsufficient, dominant negative. Latter were divided into two subgroups according to whether mutation resulted in cysteine elimination. Subjects were assigned into cardiovascular risk based on previous aortic surgery. Retina of haploinsufficient patients was thinner in total, foveal, parafoveal, perifoveal areas compared to dominant negative subjects (p ≤ 0.047). Retinal thickness of haploinsufficient individuals was thinner in total, parafoveal areas compared to dominant negative without (LSD p = 0.038, Bonferroni p = 0.027, respectively) and with cysteine elimination variants (LSD p = 0.032, Bonferroni p = 0.002, respectively). In fovea and perifovea, retinal thickness was decreased in haploinsufficient patients in comparison to dominant negative with cysteine elimination group (Bonferroni p ≤ 0.029). Total, parafoveal, perifoveal superficial and total, parafoveal deep vessel density of subjects who underwent aortic surgery were lower compared to non-operated patients (p ≤ 0.043). To the best of our knowledge, our study is the first to describe a relationship between genotype and optical coherence tomography angiography parameters in Marfan syndrome. These findings along with correlations between genetics and cardiovascular manifestations reported previously, suggest that these parameters may be indirect predictors of increased cardiovascular risk. Here we demonstrated associations between these parameters and aortic involvement.