Background/aimsOptical coherence tomography angiography (OCTA) allows the study of vessel density (VD). We intended to perform a systematic review of studies focusing on longitudinal changes in peripapillary and macular VD measurements in glaucoma.MethodsA search was performed across MEDLINE, Scopus, ISI Web of Science and Google Scholar, using the following query from inception until 20 September 2019: ((“optical coherence tomography angiography”[tiab]) OR (optical coherence tomography angiography[MeSH]) OR (“OCTA”[tiab]) OR (“OCT-A”[tiab]) OR (“angio-OCT”[tiab]) OR (“OCT- angiography”[tiab]) OR (“OCT-angio”[tiab]) OR (“OCT-angiographie”[tiab])) AND (glaucom*[tiab] OR glaucoma[MeSH]). Prospective studies that quantitatively assessed the longitudinal changes in VD in glaucoma with at least 3 months of follow-up were included.ResultsTen out of 4516 studies were included. The rate of VD change in glaucoma varied from 0.036/year to 1.08/year and 1.3% to 3.2% per year, with significantly different rates between glaucoma and healthy controls. Five studies assessed VD change after glaucoma surgery, obtaining variable results, ranging from a temporary VD decrease to increase after 3 months. Meta-analysis was not possible due to a wide variation in methods, measurements and region of VD.ConclusionOCTA is a non-invasive technology, which shows promise in glaucoma. Measures should be taken to increase the quality and standardise the methodology of VD measures in OCTA longitudinal studies, for future meta-analyses.

The authors investigated the influence of spatial frequencies in foveal vision in glaucomatous patients in a recognition task of facial expressions. Nineteen patients, 16 age-matched and 14 young controls saw centrally presented photographs of faces. Participants categorized the facial expressions as happy, angry or neutral. Two versions were tested: filtered faces of either low (LSF) or high (HSF) spatial frequency content and hybrid faces constructed from a face with LSF content superimposed on a face with HSF content with differing facial expressions. Compared to age-matched controls a significant deficit was observed both on HSF and on LSF filtered faces for patients. Controls, but not patients, were biased towards the HSF component of the hybrid faces. Different spatial frequencies are normally used to recognize different facial expressions. Impaired processing of a spatial frequency bandwidth due to an ocular pathology can therefore lead to misinterpretation of emotions, and impact on social interactions.