Posterior Segment Optical Coherence Tomography Findings in Patients with COVID-19 COVID -19 Hastalarında Arka Segment Optik Koherens Tomografi Bulguları

Objective: COVID–19 which is still the leading health problem in the world, is not yet fully understood, therefore the study aims to investigate the posterior segment of the eye in patients with COVID–19 using optical coherence tomography. Material and Method: The present study was designed prospective and cross-sectional. Pa - tients with a positive test result of COVID–19 and healthy controls who were older than 18 years were included in the study. Excluded from the study were the patients with any systemic comorbidity and ocular disease. The demographic and clinical data, including ocular findings and the optical coherence tomography data, including hyper-reflective lesions, retinal nerve fiber layer, ganglion cell complex, and macular thicknesses were recorded as outcomes and analyzed. Results: Twenty patients with SARS-CoV-2 and 20 healthy subjects were included in this study. There was no difference between COVID-19 patients and healthy controls in terms of the retinal nerve fiber layer, ganglion cell complex, and macular thicknesses (p>0,05). Hyper-reflective lesions were detected in 18(45%) eyes of 9(45%) patients with COVID – 19 and 15(37,5%) eyes of 8(40%) healthy


INTRODUCTION
Coronavirus disease (COVID- 19), first reported in December 2019 in Wuhan, China, has been spreading worldwide since then.The pathogen causing this disease is a non-segmented positive-sense RNA virus called severe acute respiratory syndrome coronavirus-2 (SARS CoV-2), which has a similar binding receptor (angiotensin-converting enzyme -2) and similar pathologic features and epidemiological characteristics with SARS CoV (1,2).It is believed that Coronaviridae are not always confined to the respiratory tract, and neurological signs, such as headache, nausea, and vomiting suggest the CNS involvement of these viruses (3,4) Although it is very well known that SARS CoV mainly affects the respiratory system, a few studies have underlined the eye as a potential site for replication and transmission (5,6).
Recently, Marinho et al. (7) released a new report showing hyper-reflective (HR) lesions at the level of ganglion cells and inner plexiform layers of the papillomacular bundle using spectral domain-optical coherence tomography (SD-OCT) in 12 healthcare professionals with COVID -19.However this study had several major limitations.One of the most important limitations was that the study had no controls.There are also case reports which indicate that posterior segment manifestations could be seen in COVID -19 patients' eyes (8,9).In this study, we aimed primarily to investigate the posterior segment of the eyes in patients with COVID -19 using SD-OCT and secondly, show these HR lesions if there are any and search for associations between the presence of these lesions and the thicknesses of the retinal nerve fiber layer (RNFL), ganglion cell complex (GCC) and macula.

Demographic and clinical data
The study design was prospective and cross-sectional.Twenty healthy subjects and 26 COVID-19 patients with mild to moderate severity of pneumonia who were hospitalized between 1st to 30th April 2020 for diagnosis and treatment were included in this study.The demographic features, including age and gender, and clinical features, including symptoms, the severity of pneumonia according to the Turkish Ministry of Health, Scientific Committee, Guidelines for CO-VID-19 (10), elapsed time since the first symptom occurred, elapsed time since hospitalization, the time from the first symptom to hospitalization, and the treatment applied was recorded for each case.All participants underwent a complete ophthalmological examination, including best-corrected visual acuity (BCVA) with a Snellen chart, intraocular pressure measurement with a non-contact air-puff tonometer, slitlamp biomicroscopy examination, and dilated fundoscopy along with SD-OCT (RTVue-100 OCT, Optovue Inc, Fremont, CA) scanning.
The study included patients older than 18 years.Patients with a positive test result of COVID-19 according to the real-time reverse transcriptase-polymerase chain reaction (rRT-PCR, using nasal or oral swabs) test were included in the COVID -19 patients group.Excluded from the study were patients with severe COVID-19, those with systemic comorbidities (diabetes mellitus, hypertension, chronic obstructive lung diseases, such as asthma, connective tissue disorders, and autoimmune diseases), and those with ocular diseases (any ocular surgery history, spherical equivalent > ±3 D, corneal astigmatism > ±3 D, cataract, corneal diseases, glaucoma, retinal diseases, such as retinal vascular obstruction and age-related macular degeneration).
SD-OCT scanning An RTVue-100 OCT device (Optovue Inc, Fremont, CA) was used for scanning.The RNFLT, MRT, and GCCT of both eyes of the study participants were evaluated.The RNFLT map was measured along a circle of 3.45 mm in diameter centered on the optic disc.A 3D disc protocol was used to register the edge of the optic nerve head.The global (average) RNFL thickness, the RNFL thickness of the superior and inferior hemifields, and the temporal, superior, nasal, and inferior quadrants were recorded for the analyses.The GCC protocol explored parameters within a circle of 6 mm diameter with the center of the GCC scan being shifted approximately 1 mm temporal to the fovea to improve the sampling of temporal peripheral nerve fibers.The GCC thickness of the superior and inferior hemifields and the global (average) GCC thickness were recorded from these scans.The macular thicknesses of the same fields as GCC were also obtained from the same scans.The enhanced deep imaging mode of SD-OCT was used to evaluate the macula and papillomacular bundle for the HR lesions which are described as HR foci or plaque seen in the papillomacular bundle at the level of ganglion cell and inner plexiform layers.(7)

Statistical analyses
The Statistical Package for the Social Sciences v. 21 (SPSS, Inc., Chicago, IL, USA) was used for statistical analysis.Quantitative variables were defined as mean and standard deviation (SD) and qualitative variables as percentages.Power calculation was not required due to the exploratory nature of the study.The Shapiro-Wilk test was used to determine whether the sample came from a normally distributed population.The Pearson chi-square test was used to analyze the gender and symptom distribution among the groups and the presence of HR lesions.According to the normality test results, the Mann-Whitney U test or Student's t-test was used to compare the age between COVID -19 patients and controls.All the eye-related parameters were compared using generalized estimating equations (GEE).For each parameter [RNFL, GCC and macular thicknesses, and intraocular pressure (IOP)], a single age-and gender-corrected GEE model with the correlation matrix "independent" was created.In these models, eye-related parameters were chosen as the dependent variables, gender and grouping as factors, and age as the covariant.At the end of this process, the main effects of grouping on eye-related models were calculated.The GEE results were given with the correlation coefficient (B), 95% Wald confidence interval (CI), and P value.A P value less than 0.05 was considered statistically significant.
The groups showed no significant difference in terms of IOP (P = 0.426).The mean LogMAR-converted BCVA was 0.00 for both groups.No conjunctivitis or anterior segment involvement was seen in any patient.With the dilated fundus examinations, no sign of retinal involvement was presented in any patient and healthy subject.
Analyses of SD-OCT scans; RNFL The RNFL thickness values of the participants and the results of GEE analyses are given in Table 2. RNFL of all quadrants, independent of age and gender, were not significantly differed between patients and healthy subjects.(P > 0.05 for all quadrants)  Analyses of SD-OCT scans; Macula and GCC The macular and GCC thickness values of the participants and the results of GEE analyses are given in Table 3.There was no significant difference in terms of the global macular thickness of that of the superior and inferior hemifields (P = 0.861, 0.720, and 0.369, respectively).Although the GCC of the patients with COVID -19 was thicker than healthy subjects, the differences were not significant (P = 0.303, 0.218, and 0.265, respectively).

DISCUSSION
In the present study, we investigated COVID-19 patients with mild to moderate pneumonia and healthy subjects for HR lesions, which Marinho et al. (7) described.Twenty of 26 patients and 20 healthy subjects met the inclusion criteria.We detected HR lesions in 18 eyes of 9 patients and 15 eyes of 8 healthy subjects.There was no significant difference between the COVID -19 patients and healthy individuals.These results suggest that the HR lesions which were presented in the papillomacular bundle could only be retinal vessels' tangentially cut on OCT B-scans.In our study, we also found no significant differences between healthy and COVID -19 patients in terms of the thicknesses of RNFL, GCC, and macula.Although it is almost impossible to draw definite conclusions because of such a small sample size, we believe that the retinal involvement is not likely in COVID-19.
Recently, Marinho et al. presented a report that revealed HR lesions in the retina of 12 COVID-19 patients, especially in the papillomacular bundle and at the level of ganglion cells and inner plexiform layers.This report was the first to show the posterior segment/retinal involvement of SARS-CoV-2.Four of the patients in that report also had subtle cotton wool spots and microhemorrhages along the retinal arcade.All the patients had the symptoms of fever, asthenia, and dyspnea, and 11 also presented with anosmia.The main limitation of this study is that no comparison was made with the control group.Marinho et al. assumed that the presence of these HR lesions were proof of the retinal or a reflection of central nervous system (CNS) involvement, however, it is impossible to support these assumptions without a control group.In the current study including 20 cases, nine patients had similar HR lesions, however the same lesions were seen in 8 healthy individuals.Therefore, we agree with the various criticisms that have been made of the work of Marinho et al. (17)(18)(19)(20)(21) In addition to the report of Marinho et al., we also searched for associations of the OCT data including RNFL, GCC, and macular thicknesses with COVID-19.However, we found no significant differences in any of these parameters.Gündoğan et al. (22) also investigated the posterior segment involvement of the COVID-19 and concluded the same with our study.However, Pereira et al. (23) reported cotton wool spots and flame-shaped hemorrhages on the retina of the COVID-19 patients.They investigated the posterior segments of the hospitalized severe patients and this could be the reason for the inconsistency.
It is believed that SARS-CoV and MERS-CoV are not always confined to the respiratory tract, and neurological signs, such as headache, nausea, and vomiting suggest the CNS involvement of these viruses (3,4).Furthermore, the neurological invasion of SARS-CoV has been shown in both patients and experimental animals (24)(25)(26).The high similarity between SARS-CoV and SARS-CoV-2 leads to the idea that the same neurological concerns may be valid for SARS-CoV-2, thus playing a role in the respiratory failure of COVID-19 patients (27).Lastly, Moriguchi et al (28).reported the first case of meningitis and encephalitis associated with SARS-CoV-2.There are also some animal coronaviruses, such as mouse hepatitis virus (MHV) with both neurological and ocular manifestations.MHV is used for the creation of virally induced optic neuritis and retinal degeneration models (29-31).MHV has also been believed to be one of the etiological factors of multiple sclerosis (29).In the light of these investigations and reports, we can hypothesize that possible neurological findings may have reflections on the posterior segment of the eye, such as the swelling of the optic nerve head and RNFL.In our investigations, with such a small sample size, we found no signs of retinal or CNS involvement.Further experimental studies are needed to verify or reject this hypothesis.
The small sample size is the main limitation of this study, which did not allow us to analyze the effect of treatment strategies on the retina.Secondly, due to the study's cross-sectional design, it is almost impossible to draw definite conclusions.Another limitation was we could not use any additional imaging techniques due to pandemic reasons such as color fundus photography or fluorescein angiography which could give us more information about the posterior segment.And lastly, we also did not search for the associations with serum inflammation biomarkers which could give us more information about the disease severity.

CONCLUSION
In conclusion, the hyper-reflective lesions which were seen in the OCT images could only be retinal vessels tangential cuts seen in OCT B-Scans.On the other hand, COVID-19 is caused by a virus that is not yet fully understood.Including our study, there are not enough data to assume that the SARS-CoV-2 could affect the posterior segment of the eye, therefore, physicians should be aware of neurological and ocular manifestations of this disease.Future follow-up and experimental studies are needed to provide an understanding of the true nature of this disease.
This study was carried out under the principles of the Declaration of Helsinki after obtaining the permission of the Turkish Ministry of Health and the approval of the Clinical Research Ethics Committee of Umraniye Training and Research Hospital, Istanbul, Turkey (Approval number: B.10.1.THK.4.34.H. GP.0.01/205).Written informed consent was obtained from each patient.

Figure I .
Figure I.The samples of HR lesions which were detected using optical coherence tomography of the macula.White arrowheads; HR dots, red arrowheads; HR plaques.HR: hyperreflective.

Table II .
RNFL thickness values of the participants and the results of GEE analyses RNFL; retinal nerve fiber layer, GEE; generalized estimating equations, B; coefficient, CI; confidence interval, sup; superior, h; hemisphere, inf; inferior.Significant P values are shown in bold.

Table III .
Macular and GCC thickness values of the participants and the results of GEE analyses GCC; ganglion cell complex, GEE; generalized estimating equations, B; coefficient, CI; confidence interval, sup; superior, h; hemisphere, inf; inferior.