Posterior segment ophthalmic manifestations in Ebola survivors, Sierra Leone

The largest Ebola virus disease (EVD) outbreak occurred from 2013–2016 in West Africa, leading to over 28,600 cases and 11,300 deaths and resulted in the largest cohort of EVD survivors to date.¹ Another recent outbreak in Democratic Republic of the Congo (DRC) has resulted in 3,481 cases with over 1,170 survivors since August 2018.² In studies of large EVD survivor cohorts, uveitis is the most common ocular manifestation with a 13–34% reported prevalence.^3,4 Cataract, with or without uveitis, is the second most common finding and has been reported in 10% of EVD survivors.^3,4 The Ebola Virus Persistence in Ocular Tissues (EVICT) study was a cross-sectional study which supported the safety of cataract surgery in EVD survivors5, but may not be generalizable to vitreoretinal surgery. Posterior segment findings in EVD survivors have been described primarily in relationship to uveitis and include vitritis, chorioretinitis and chorioretinal scarring.^3,6 The prevalence of posterior segment findings in EVD survivors, particularly those that may require surgical intervention, is of particular interest given the risk of Ebola virus (EBOV) persistence in ocular tissues and fluids.⁷

This is a secondary analysis of baseline data from patients screened for the EVICT study, which evaluated EBOV RNA persistence in the eyes prior to cataract surgery in EVD survivors. Institutional Review Board approval was obtained from Emory University and the Sierra Leone Ministry of Health. Human research was conducted according to the Tenets of the Declaration of Helsinki. Informed consent was obtained with Sierra Leonean interpreters in the dialect of patients evaluated.

Patients screened for the EVICT study included EVD survivors referred from local eye clinics for vision impairment and/or cataract evaluation.⁵ All patients had documented serologic evidence of prior EBOV infection. Full ophthalmic exams were performed and included corrected visual acuity (VA), slit lamp and dilated funduscopic examination (DFE). Active uveitis was defined as the presence of inflammation in the anterior chamber, keratic precipitates, vitreous haze, and retinal or choroidal infiltrates. Per the original EVICT study, visually significant cataract was defined as Snellen VA of 20/200 or worse or affecting activities of daily living.⁵ A B-scan ultrasound was performed for dense media opacity precluding DFE.

Statistical analysis was performed with Stata (StataCorp). Demographic data and ocular findings were summarized as frequencies or medians with interquartile ranges. Snellen VAs were converted to logMAR values for all analyses (Table 1). For eyes with low vision (counting fingers or worse), VA was converted to logMAR. Per the original EVICT study, eyes with NLP vision (n=5) were excluded from analysis to avoid a bias towards rejection of the null hypothesis.⁵ A p-value<0.05 was considered significant for univariate and multivariate analyses.

Table 1.

Statistical analysis of ocular findings and association with baseline visual acuity

Univariate analysis
Ocular Finding, N (% of eyes) *	Median baseline logMAR Visual Acuity	Significance†
Visually significant cataract, 68 (27)	3.00	P<0.001
Posterior synechiae, 52 (21)	3.00	P<0.001
Any type of uveitis, 86 (34)	2.00	P<0.001
Anterior uveitis, 54 (22)	3.00	P<0.001
Intermediate uveitis. 4 (2)	2.00	P=0.032
Posterior uveitis, 7 (3)	0.30	P=0.890
Panuveitis, 7 (3)	1.30	P=0.241
Any posterior segment findings‡, 48 (19)	1.00	P<0.001
Chorioretinal scarring, 22 (10)	0.24	P=0.231
Optic neuropathy, 7 (3)	2.00	P=0.005
Vitreous opacities, 7 (3)	3.00	P=0.84
Retinal detachment, 6 (2)	3.00	P<0.001
Epiretinal membrane, 2 (1)	0.48	P=0.332
Vitreomacular traction, 2(1)	3.00	P=0.019
Drusen, 1(0)	0.00	P=0.400
Retinal pigment epithelium atrophy, 1(0)	0.00	P=0.140
Surgical vitreoretinal findings‡‡, 17 (7)	3.00	P<0.001
Multivariate Analysis
Ocular Finding		Significance
Visually significant cataract		P<0.001
Posterior synechiae		P=0.012
Anterior uveitis		P=0.574
Intermediate uveitis		P=0.171
Optic neuropathy		P<0.001
Retinal detachment		P=0.016
Vitreomacular traction		P=0.113

^*Two-hundred fifty eyes of 125 patients had data sufficient for analysis (i.e. 137 patients screened in the EVICT Study)

^†Univariate analysis conducted using Wilcoxon-Mann-Whitney test; multivariate analysis conducted using ANOVA. Analyses compared logMAR visual acuity of patients with and without the ocular finding of interest.

^‡Includes any vitreous, optic nerve or retinal findings

^‡‡Includes retinal detachment, epiretinal membrane, vitreous opacities and vitreomacular traction

Two hundred and fifty eyes of 125 EVD survivors from Sierra Leone were examined. Sixty-five percent were female and the median age was 28 years (IQR 17–45). Median baseline logMAR VA was 0.2 (Snellen 20/31, IQR 0–1.8) and 26.8% of eyes had baseline logMAR worse than 1.3 (20/400). The frequency and spectrum of ocular findings are summarized in Table 1.

Chorioretinal scarring, the most common vitreoretinal finding, was observed in 22 of 220 eyes (10.0%) that had sufficiently clear media for DFE. Retinal detachment (RD) was present in six of 250 eyes (2.4%) and included tractional (3), rhegmatogenous (1) and serous RD (1). The presence of RD was associated with significantly worse median VA at logMAR 3.0 compared to logMAR 0.0 in individuals without RD (p<0.001).

In total, 48 of 250 eyes (19.2%) demonstrated posterior segment pathology that was associated with significantly worse VA (median logMAR 1.0, VA 20/200) compared to those without posterior segment disease (median logMAR 0.0, VA 20/20; p<0.001). Nine individuals demonstrated bilateral posterior segment pathology (23.1%), six of whom had bilateral chorioretinal scarring. Seventeen eyes (6.8%) showed disease that could benefit from vitreoretinal surgery (e.g. RD, epiretinal membrane [ERM], and vitreomacular traction) and also showed worse VA (median logMAR 3.0, VA HM) than individuals without surgical disease (median logMAR 0.0, VA 20/20; p<0.001). Cataract (p<0.001), posterior synechiae (p<0.05) and RD (p<0.05) were independently associated with poorer VA with multivariate analysis.

Within this cohort of EVD survivors, posterior segment pathology was present in 19.2% of eyes examined. Chorioretinal scarring was the most common posterior segment finding and present in 10% of eyes. Chorioretinal scarring has ranged from 15–60% in other studies of EVD survivors. In a prior case-control study of EVD survivors, 10 different types of chorioretinal lesions were described, with one distinct subtype observed in 15% of EVD survivors while not observed in controls.⁶

Other vitreoretinal findings observed in this cohort included RD, persistent vitreous opacities, ERM, vitreomacular traction and one patient with a posteriorly dislocated crystalline lens. Unlike chorioretinal scarring which may result in permanent vision loss, surgical intervention may improve vision in these conditions, specifically RD, which was present in 2.4% of eyes. Importantly, the young age of our cohort has significant implications for quality-of-life-years if left untreated. In a resource-limited setting, challenges to vitreoretinal surgery include the scarcity of subspecialty-trained personnel, retinal equipment and instrumentation.. Vision health systems strengthening is thus needed to avert vision loss due to potentially treatable vitreoretinal disease.

Limitations of this study include the nature of a post hoc, secondary analysis of the cohort, and potential selection bias as patients were identified via national and local screening programs. Specifically, it is possible that the incidence of vision loss is overestimated as symptomatic patients are more likely to present for screening. In addition, given the resource-limited setting, fundus photography and other advanced imaging modalities were limited, potentially reducing the identification of more subtle retinal abnormalities.

Nonetheless, this study demonstrates a significant burden of posterior segment disease approaching 20% of eyes evaluated. Given the relatively high rate of uveitis in the West African population, it is unclear if these findings are associated with EVD pathogenesis or reflect sequelae from other endemic diseases. However, given that EVD survivors demonstrate vitreoretinal pathology that would benefit from surgical intervention and the known risk of EBOV persistence in the eye, further studies are needed to understand the potential surgical risks.⁷ Just as the EVICT study assessed the safety of cataract surgery on EVD survivors, understanding the potential for EBOV persistence within the posterior segment is warranted for both patient outcomes and provider safety measures.

Supplementary Material

2. Online Supplemental Figure 1. Spectrum of chorioretinal scarring and retinal disease observed in EVD survivors.

Fundus photographs taken with an iPhone and 28-diopter condensing lens shows a spectrum of chorioretinal scarring. (A) Retinal image shows dense retinal pigment epithelial hyperplasia with atrophic center along superotemporal arcade with associated foveal granularity. (B) Retinal image shows peripapillary chorioretinal scarring and multifocal, macular hyperpigmentary changes. There is also subtle granularity to the retinal pigment epithelium along the inferotemporal arcade. (C) Nummular-appearing chorioretinal scarring involving fovea with adjacent area of hypopigmentary change. (D) Chronic rhegmatogenous retinal detachment with proliferative vitreoretinopathy.

3. Online Supplemental Figure 2. B-scan ultrasound in EVD survivors with dense cataract.

(A) B-scan ultrasound of the right eye (L9) shows vitreoretinal traction peripherally and macula-involved tractional retinal detachment. (B) B-scan ultrasound of an EVD survivor shows focal elevation and tractional retinal detachment isolated to the macula within the posterior pole.