“Vision Loss” and COVID-19 Infection: A Systematic Review and Meta-Analysis

Background: Visual impairment in terms of reduced visual acuity and “visual loss” has been reported as an atypical symptom in patients with severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) infection. This systematic review and meta-analysis aims to assess the cumulative incidence of “visual loss” during coronavirus disease 2019 (COVID-19) and review the current evidence regarding “visual loss” caused by SARS-CoV-2 infection. Methods: We performed a systematic review and meta-analysis of studies following Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines. We systematically searched the PubMed, Embase, and Scopus databases for relevant studies published that clearly described “vision loss” and SARS-CoV-2 infection. All studies reporting concomitant “vision loss” and laboratory-confirmed SARS-CoV-2 infection were included. Meta-analyses were conducted using the measurement of risk and a 95% confidence interval for each study. Results: Our search identified 1143 manuscripts published in the English language. After study screening, twenty-nine articles were selected: two cross-sectional studies, twenty-four case reports, and three case series. A random-effect meta-analysis demonstrated that the pooled “visual loss” cumulative incidence in COVID-19 patients was 0.16 (95% CI 0.12–0.21). The quality rating of the cross-sectional studies averaged four out of the maximum score on the Newcastle–Ottawa scale. Conclusions: COVID-19 infection might cause “visual loss”. Even if the current evidence is limited, ophthalmological assessment should be promptly provided to all patients experiencing visual impairment symptoms during SARS-CoV-2 infection.

The World Health Organization (WHO) declared the coronavirus disease 2019 (COVID-19) outbreak a global pandemic on 11 March 2020, which led to a significant economic and healthcare burden [2]. Current available diagnostic tests to detect COVID-19 include a triad of complementary approaches. The polymerase chain reaction (PCR) is the most highly sensitive and specific molecular test to detect SARS-CoV-2 nucleic acids' presence, representing the gold standard technique because of its sensitivity and specificity. It uses primers matching a segment of the SARS-CoV-2 genetic material to detect COVID-19 [3].
After exposure, the average incubation period may range from four to five days [4]. A wide range of symptoms has been associated with the SARS-CoV-2 infection, whose severity may vary from asymptomatic to death. Although most patients either remain asymptomatic or experience common viral infection symptoms such as fever, cough, and fatigue, some patients may develop atypical symptoms such as neurological (headaches, loss of taste, smell) and ophthalmological symptoms (conjunctivitis, epiphora, and "vision loss") [5]. Based on multiple cross-sectional studies, the incidence of ocular manifestation in COVID-19 patients might be as high as 30% [6]. At the beginning of the pandemic, many physicians reported eye redness and irritation in patients, describing "conjunctival congestion" in Wuhan, China. In a recent systematic review and meta-analysis, Inomata et al. reported clinical and prodromal ocular symptoms in patients with COVID-19. The most common ocular findings among COVID-19 patients were conjunctivitis (86.4%), ocular pain (34.4%), dry eye (33.3%), and floaters (6.7%) [7]. "Visual loss" in COVID-19 patients was reported in a few articles. Nonetheless, it has been observed that its onset may be due to viral neurotropism and indirect immunologic and neurovascular effects .
Furthermore, despite extensive research on sensory manifestations of COVID-19 since the start of the pandemic, only a few articles and no meta-analysis papers have assessed "vision loss" as a symptom during the SARS-CoV-2 infection.
The present paper intends to systematically review the current evidence regarding "visual loss" caused by SARS-CoV-2 and to determine its cumulative incidence through a meta-analysis. In addition, we further aimed to identify the characteristics of the "visual loss", thus evaluating factors that could contribute to understanding the association between COVID-19 and "visual loss".

Search Strategy
Three databases (PubMed, Embase, and Scopus) were checked from inception until 9 June 2022, using free text and controlled vocabulary (MeSH or Emtree) to analyze the relationship between visual impairment and SARS-CoV-2 infection.
The search strategy combined the controlled vocabulary and the keywords according to the indications from each database. The Medical Subject Headings (MeSH) controlled vocabulary were used to search for articles in PubMed, and the Embase Subject Headings (EMTREE) was used in the EMBASE. The keywords were selected based on readings related to the study's subject. The controlled vocabularies and keywords were used with Boolean operators to extend and direct the search. For addition and restriction, the Boolean operators OR and AND were used. In addition, the investigation was conducted using recognized and extended vocabulary without database filters to achieve a significant sample with a decreased potential loss. Our core search comprised the following terms: "COVID-19" AND "blindness" or relevant synonyms, such as "vision loss"; and "SARS-CoV-2". In addition, we also hand-searched the bibliographies of included articles to identify further studies that were not found in the initial database search. Figure 1 illustrates a flow diagram of the literature search and screening results. This review is reported following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. The International Prospective Register of Systematic Reviews (PROSPERO number CRD42022339189, registered on 17 June 2022) has been used to register this systematic review. After the protocol registration, no changes were made. We included the study protocol of the synthesis in Supplementary Material S1. The detailed search strategy and PRISMA Checklist are reported in Supplementary Materials S2 and S3.

Study Selection Data Extraction and Data Synthesis
Articles reporting "vision loss" developed during laboratory-confirmed COVID-19 infection were included.
"Visual loss" was defined according to The International Classification of Diseases 11 (2018), (distance mild visual impairment: visual acuity between <0.5 but ≥0.3 using a decimal scale; distance moderate visual impairment: visual acuity between <0.3 but ≥0.1 using a decimal scale; severe visual impairment: visual acuity between <0.1 but ≥0.05 using a decimal scale; and blindness: visual acuity < 0.5 using a decimal scale or near vision impairment: near visual acuity worse than N6 at 40 cm with existing correction). This review is reported following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. The International Prospective Register of Systematic Reviews (PROSPERO number CRD42022339189, registered on 17 June 2022) has been used to register this systematic review. After the protocol registration, no changes were made. We included the study protocol of the synthesis in Supplementary Material S1. The detailed search strategy and PRISMA Checklist are reported in Supplementary Materials S2 and S3.

Study Selection Data Extraction and Data Synthesis
Articles reporting "vision loss" developed during laboratory-confirmed COVID-19 infection were included.
"Visual loss" was defined according to The International Classification of Diseases 11 (2018), (distance mild visual impairment: visual acuity between <0.5 but ≥0.3 using a decimal scale; distance moderate visual impairment: visual acuity between <0.3 but ≥0.1 using a decimal scale; severe visual impairment: visual acuity between <0.1 but ≥0.05 using a decimal scale; and blindness: visual acuity < 0.5 using a decimal scale or near vision impairment: near visual acuity worse than N6 at 40 cm with existing correction).
All studies reporting patients with neither laboratory-confirmed SARS-CoV-2 infection nor COVID-19-related "visual loss" were excluded. None of the studies reporting "visual loss" developed after COVID-19 infection or vaccination were included. Furthermore, articles were excluded if they were unavailable in the English language or assessed the visual impairment as a prodrome of SARS-CoV-2 infection in otherwise healthy patients. In addition, literature review studies, theses, and dissertations; book chapters; and conference abstracts were not included in our analysis. Reasons for exclusion were documented. In addition, all articles that reported data on "visual loss" cumulative incidence among COVID-19 patients were included in the quantitative analysis.
We contacted the corresponding authors of eligible studies whenever the article could not be retrieved, or we needed to obtain additional information that was not available in the article or online Supplementary Files. Thus, the information was extracted directly from the included studies or provided by the corresponding authors.
Two investigators (M.R. and C.S.) independently extracted baseline and outcome data. If consensus could not be reached, two co-authors (P.A and S.R.) were consulted for adjudication. We extracted the following data from each article: the first author, publication date, country, study design, sample size, study design, average age, gender, visual impairment description, "laterality", duration between the onset of COVID-19 symptoms and ocular symptoms, comorbidities, number of COVID-19 affected patients and diagnosis. We used Covidence systematic review software© (Veritas Health Innovation, Melbourne, Australia), available at www.covidence.org (accessed on 9 June 2022) [37], to record and evaluate the study data between 22 May 2022 and 09 June 2022.

Risk of Bias Assessment
Two authors (M.R. and C.S.) independently appraised the methodological quality of each cross-sectional and case-report study by using the Newcastle-Ottawa scale (NOS) and the Joanna Briggs Institute (JBI) Critical Appraisal Checklist for Case Reports, which consist of eight yes/no/unclear questions. The JBI critical appraisal checklist for the case series was used for the quality assessment of the case series [38].
Quality assessment data individually appraised by each of the reviewers were compared. If consensus could not be achieved, M.R. and C.S. discussed the discrepancies for adjudication. The data from each reviewer's quality assessment were compared. M.R. and C.S. discussed the inconsistencies for adjudication if consensus could not be reached.

Statistical Analysis
A random-effects meta-analysis of pooled prevalence and their 95% confidence intervals of COVID-19-affected patients who developed "visual loss" was obtained based on the exact binomial distributions (i.e., number of "events" versus number of "non-events" in a sample) with Freeman-Tukey double-arcsine transformation using the "metaprop" command in STATA (STATA Corp, College Station, TX, USA), version 17.0.
According to Barker et al., a high I 2 in the context of proportional meta-analysis does not necessarily mean that data are inconsistent, and the results of this test should be interpreted conservatively. Therefore, we did not perform further analysis. Tests to evaluate publication bias, such as Egger's test and funnel plots, were not performed due to the low number of studies analyzed. Furthermore, Egger's test and funnel plots were developed in the context of comparative data, and there is no evidence that proportional data adequately adjust for these tests [39]. Statistical significance was determined by a two-sided p-value of 0.05. for final eligibility. Furthermore, 57 articles were excluded because the "visual loss" was developed before or just after the SARS-CoV-2 infection, or COVID-19 was not detected by PCR test. In addition, potential studies were most often excluded due to not fulfilling the study design criteria or being a pre-print and not yet published.

Study Selection
Finally, 29 studies were included in the systematic review  and two in the meta-analysis [9,23]. The included studies provided data on the number of participants with "visual loss" developed during COVID-19, and the meta-analysis included two studies [9,23] with 288 COVID-19 patients.
"Visual loss" due to fungal infection (mucormycosis) in corticosteroid-treated p
"Visual loss" due to fungal infection (mucormycosis) in corticosteroid-treated patients was described in three case reports and one cross-sectional study [9,14,27,32].
Reich et al. and Clarke et al. reported "visual loss" following assisted mechanical ventilation and prone position [24,35].
Where provided, severe visual impairment (visual acuity between <0.1 but ≥0.05 using a decimal scale) and blindness (visual acuity < 0.5 using a decimal scale) incidences were reported.

Meta-Analysis
A proportional random meta-analysis was performed to estimate "visual loss" cumulative incidence among COVID-19 patients. The total population was equal to 280, and the sample size varied between 89 and 199. The pooled prevalence of "visual loss" among COVID-19 patients was equal to 0.16 with a confidence interval (CI) between 0.12 and 0.21 (I 2 = 84.10, z: 12.79, p ≤ 0.001) (Figure 3).

Meta-Analysis
A proportional random meta-analysis was performed to estimate "visual loss" cumulative incidence among COVID-19 patients. The total population was equal to 280, and the sample size varied between 89 and 199. The pooled prevalence of "visual loss" among COVID-19 patients was equal to 0.16 with a confidence interval (CI) between 0.12 and 0.21 (I 2 = 84.10, z: 12.79, p ≤ 0.001) (Figure 3). Figure 3. Proportional meta-analysis of cumulative "visual loss" incidence in COVID-19 patients [9,23].

Risk of Bias and Publication Bias
Supplementary Material S4 summarizes all studies' risk of bias evaluation. The quality rating of the cross-sectional studies [9,23] averaged four out of the maximum score on the Newcastle-Ottawa Scale. Overall, the two cross-sectional studies reached a total score of four.
The studies did not report the response rate or the characteristics of the responders and the non-responders. No statistical tests were used to assess the "visual loss" prevalence among the patients. According to the JBI Critical Appraisal Checklist for Case Reports and JBI Critical Appraisal Checklist for Case Series, the quality of the included studies was moderate to good. Most case reports scored 6 out of 8 quality criteria or higher.

Risk of Bias and Publication Bias
Supplementary Material S4 summarizes all studies' risk of bias evaluation. The quality rating of the cross-sectional studies [9,23] averaged four out of the maximum score on the Newcastle-Ottawa Scale. Overall, the two cross-sectional studies reached a total score of four.
The studies did not report the response rate or the characteristics of the responders and the non-responders. No statistical tests were used to assess the "visual loss" prevalence among the patients. According to the JBI Critical Appraisal Checklist for Case Reports and JBI Critical Appraisal Checklist for Case Series, the quality of the included studies was moderate to good. Most case reports scored 6 out of 8 quality criteria or higher. All case series scored 6 out of 10 quality criteria or higher. Notably, two case series scored 8 out of 10 quality criteria [14,26], whereas one case series scored only six, as it did not provide information regarding follow-up results [35].

Discussion
Our systematic review and meta-analysis aimed to identify and describe the characteristics of the "visual loss" developed during SARS-CoV-2 infection. Furthermore, we aimed to determine the cumulative incidence of "visual loss" during SARS-CoV-2 infection.
In our meta-analysis, we found that the cumulative incidence of "visual loss" is 16% (CIs: 0.12-0.21) in confirmed cases of COVID-19. This number should be interpreted with precaution because of the low level of evidence (one study provided data of "visual loss" in patients with concomitant COVID-19 and rhino-orbital-mucormycosis) and high heterogeneity between the papers [9].
Among the sensory impairment symptoms during COVID-19, "visual loss" is less frequently reported than olfactory and gustatory changes. Indeed, few reports reporting "visual loss" during SARS-CoV-2 infection have been published since the pandemic outbreak. Despite many papers reporting several ocular symptoms such as conjunctivitis, epiphora, pain, and redness, few articles report "vision loss" [40].
Many patients experiencing "vision loss" underwent an ischemic or inflammatory optic neuropathy, mucormycosis, uveitis, or cerebrovascular accidents. COVID-19 infection has been associated with prothrombotic effects due to virusinduced cytokine storm that can activate and upregulate the coagulation, triggering the formation of a thrombus that may lead to ophthalmic artery occlusion (CRAO), central retinal artery occlusion, central retinal vein occlusion, ischemic optic neuropathy, occipital cortical infarct, or acute macular neuroretinopathy [41]. Furthermore, SARS-CoV-2 may engage the endothelium, increasing the permeability of the blood-brain barrier and leading to encephalopathy, encephalitis, and thrombosis [42].
Murchinson et al. first reported CRAO as the initial manifestation of COVID-19 in a patient with an unremarkable neurologic exam [34]. Cyr et al. reported two cases of COVID-19-positive patients with severe bilateral "vision loss" due to an acute bilateral occipital territorial ischemic infarct in a 61-year-old patient with a seven-day history of COVID-19-like symptoms and a chronic infarction in the right temporal-parietal lobe and bilateral medial occipital lobes in a 34-year-old woman with a history of systemic lupus erythematosus [26]. Khan et al. first described a bilateral occipital stroke leading to bilateral "vision loss" in a 60-year-old man with no previous risk factors for the cerebrovascular incident [19].
Furthermore, the virus-induced cytokine storm leads to a systemic inflammatory state, causing endothelial dysfunction and determining vascular leakage and edema formation, and endothelial activation resulting in the release of the immunogenic and vasoactive substance [43]. Kaya et al. and Elhassan et al. described a bilateral reversible cortical blindness and Anton's cortical blindness in patients affected by posterior reversible leukoencephalopathy with modest blood pressure fluctuations and no hypertension history [11,15].
The leading cause of visual loss across the studies was optic neuropathy. SARS-CoV-2 can affect the nervous system through different routes. It can enter the nervous system hematogenously by infecting the choroid plexus or meninges or spreading through the olfactory nerves. Moreover, a mechanism of molecular mimicry in which viral antigens would induce an immune response against self-proteins may be responsible for tissue injury [44]. Zhou et al. illustrated a case of SARS-CoV-2 infection followed by myelin oligodendrocyte glycoprotein (MOG)-IgG-related optic neuritis and myelitis that strengthened the immune-based pathogenesis. Furthermore, the upregulated coagulation may raise small capillary ischemic events, leading to ischemic optic neuritis [29].
Clarke et al. first described a case of non-arteritic ischaemic optic neuropathy after mechanical ventilation in the prone position. Indeed, prone positioning might alter ocular hemodynamics, raising intraocular pressure (IOP) and thus reducing optic nerve perfusion. The patient underwent eight episodes of prone position during mechanical ventilation to treat COVID-19-related acute respiratory distress syndrome (ARDS), and just after awakening, he reported bilateral "vision loss" [24]. Reich et al. described three cases of "visual loss" following assisted mechanical ventilation. Fundus examination revealed ischemic lesions of the retina, atrophy of inner retinal layers, and optic atrophy [35].
Benito Pascual et al. described a case of panuveitis and optic neuritis preceded by conjunctivitis prior to the onset of pulmonary symptoms [21]. Subsequently, Liu et al. described acute viral retinitis and optic neuritis followed by cataract and glaucoma due to COVID-19 infection. This inflammation might be due either to direct infiltration of the virus via ACE2 or an intraocular autoimmune response [20].
COVID-19 may lead to various opportunistic infections. Indeed, the altered immune response and the use of corticosteroids may increase the risk of superadded infections after a prolonged period in intensive care units. During COVID-19, many cytokines such as IL-6, IL-10, and TNF-α are markedly higher, whereas T lymphocytes are much lower. In patients with predisposing factors, COVID-19 may raise superinfections [9]. Reactivation of Herpes simplex virus and mucormycosis infection during COVID-19 were described as potential conditions that determine "visual loss". COVID-19 direct injury to human islet cells, determining beta cell damage and the endogenous insulin secretion's reduction, as well as the cytokine storm, lead to insulin resistance. In addition, commonly used drugs such as glucocorticoids and remdesivir further alter the glucose homeostasis, predisposing the patient to opportunistic infections. Notably, fungal infections such as mucormycosis may be promoted by ketoacidosis-induced free-iron availability [9].
Mani et al. reported that 19% of patients experienced "visual loss" due to mucormycosis [9], whereas Crane et al. described a case of Klebsiella endophthalmitis in a patient with multiple comorbidities such as liver cirrhosis, diabetes, and emphysematous prostatitis [25]. Moreover, Gonzalez et al. described a reactivation of HSV causing acute retinal necrosis in a patient with a prior history of necrotizing herpetic retinitis in the fellow eye [16].
This study has several limitations. First, the meta-analysis involved only two studies with different sample sizes. Furthermore, one study only evaluated the "visual loss" in ROMC-COVID-19-affected patients. Thus, the pooled cumulative incidence should be interpreted cautiously, and our findings should be interpreted while keeping in mind this significant limitation. Second, we could not evaluate the follow-up, as many studies lack this critical data. Third, the majority of the studies were subjected to a qualitative analysis. Fourth, the cross-sectional studies scored four out of the maximum score on the Newcastle-Ottawa Scale. However, to the best of our knowledge, this is the first study that deeply analyzes the association between COVID-19 and "visual loss". In addition, many systematically included studies involving a high number of patients from different countries make our findings generalizable and represent one of our study's strengths. Nonetheless, future research that aims to prevent any COVID-19-related blindness disease should be further conducted. Studies with larger sample sizes are needed to further investigate the pooled cumulative incidence of "visual loss" during SARS-CoV-2 infection.

Conclusions
"Visual loss" during SARS-CoV-2 infection is a rare finding. Despite the low incidence, many cases have been reported in the literature. Indeed, COVID-19 might cause "visual loss" through several mechanisms. Therefore, COVID-19 should be considered in patients who have recently developed "visual loss", and clinicians should be aware of this uncommon event to avoid blindness in everyday clinical practice.