Abstract
Introduction
Cytomegalovirus (CMV) retinitis, historically a leading cause of blindness in HIV-positive patients before the widespread use of combination antiretroviral therapy, now primarily affects other immunosuppressed populations, including transplant recipients and patients on high-dose corticosteroids or immunomodulatory therapies.1–4 CMV is characterized by 3 classic phenotypes: a wedge-shaped sector of necrotizing retinitis with an advancing border that tracks along retinal vessels; a hemorrhagic fulminant pattern with confluent whitening, prominent intraretinal hemorrhages, and vessel obscuration; and a granular indolent pattern of small whitish lesions with minimal hemorrhage that often begins in the periphery and spreads slowly. 5 If left untreated, CMV retinitis can spread along the vascular arcades, ultimately resulting in macular involvement, retinal detachment (RD), and irreversible visual loss. 6 Polymerase chain reaction (PCR) testing of aqueous or vitreous fluid has become the gold standard for diagnosis and facilitates prompt initiation of antiviral therapy. 7 Treatment involves systemic antivirals such as ganciclovir, valganciclovir, or foscarnet, often combined with intravitreal (IVT) antiviral medications.8–10
Although direct viral-induced retinal necrosis and inflammation constitute the hallmark manifestations of CMV, vascular complications represent a critical but relatively less emphasized component of its clinical spectrum. CMV-associated vasculitis is known to predispose affected eyes to severe vascular occlusive events, including branch retinal artery occlusion (BRAO), central retinal artery occlusion (CRAO), branch retinal vein occlusion (BRVO), and central retinal vein occlusion (CRVO). 11 These vascular complications can potentially influence clinical outcomes by rapidly exacerbating visual impairment, often independent of the extent of the retinitis. CMV retinitis can cause retinal ischemia and subsequent neovascular complications secondary to ischemic processes.12,13 The risk and severity of these complications may be increased by vascular occlusive events, prompting the need for treatments such as antivascular endothelial growth factor (anti-VEGF) injections and laser photocoagulation. 14 The long-term management implications of these secondary complications, including increased treatment burden and potential for further vision deterioration, remain poorly characterized. Furthermore, it remains unclear whether patients who experience vascular occlusive events secondary to CMV retinitis have different therapeutic outcomes or treatment requirements compared with those without such events, and whether there are any predictive risk features remains to be determined. This literature gap underscores the need for targeted research.
The primary purpose of the current study was to comprehensively describe the demographic and clinical characteristics, outcomes, and complications of patients with CMV retinitis, including a comparative analysis of outcomes between HIV-positive patients and iatrogenically immunosuppressed patients. To better inform clinical management and mitigate vision loss in this vulnerable population, a second objective of this study was to systematically evaluate the incidence, risk factors, clinical features, and visual prognosis of vascular occlusive events in this patient population and elucidate their impact on long-term treatment strategies.
Methods
This retrospective cohort analysis included patients diagnosed with CMV retinitis at the Emory Eye Center in Atlanta, GA, from 2014 to 2024. All patients had CMV positivity confirmed by PCR test in aqueous humor samples taken by anterior chamber paracentesis. Collected variables included demographic information (age, sex, gender identity, and self-reported race), date of diagnosis, visual acuity (VA, at presentation, at 1 month, 3 months, 6 months, and at the final recorded follow-up or 12 months), immune status (eg, immunocompetent, HIV-positive, or iatrogenically immunosuppressed), and the use of any immunosuppressive or immunomodulatory medications. Common systemic comorbidities were documented, including cardiovascular disease, hypertension, diabetes mellitus (DM), and hyperlipidemia.
Ocular findings of interest included the baseline zone of CMV retinitis, defined by the anatomic location of retinal involvement (zone 1: within 1500 µm of the optic disc or 3000 µm of the fovea; zone 2: between zone 1 and the vortex veins; zone 3: anterior to the vortex veins) 15 and the extent of retinitis in clock hours, representing the circumferential area of retinal involvement. Treatment duration was defined as the interval in days between the first and last IVT antiviral injection. Systemic antiviral therapy was recorded separately but was not used to define treatment duration. The total number of IVT injections reflected individualized treatment decisions by the treating physician, with injections discontinued once clinical resolution of active retinitis was observed. Ischemic complications were assessed using fluorescein angiography, which was performed on all patients. All patients underwent fluorescein fundus angiography, which was reviewed by 2 retina specialists (A.H., K.S.) to diagnose vascular occlusions, including BRAO, CRAO, BRVO, and CRVO. Additional complications, such as cystoid macular edema (CME), neovascularization (NV), vitreous hemorrhage (VH), and retinal detachment (RD), were recorded based on multimodal imaging and documented clinical examinations. Treatments were categorized into oral antivirals (eg, valganciclovir, letermovir, maribavir, valacyclovir), total number of IVT injections (foscarnet, ganciclovir, or a combination), topical or regional steroid use, and surgical interventions (eg, pars plana vitrectomy). In cases with bilateral involvement, each eye was considered a separate unit of analysis, and no statistical adjustment was made for potential inter-eye correlation.
All statistical analyses were performed using SPSS software (version 29.0, IBM SPSS Statistics for Windows). In parameters with normal distributions, Student t test was used; otherwise, nonparametric t tests (Mann-Whitney U test) were used. The χ2 test and Fisher exact test were used for the analyses of features seen on qualitative spectral-domain optical coherence tomography. Univariate and multivariate regression analysis was performed to determine the factors associated with VA outcomes and treatment period. Institutional review board approval was obtained, and the study adhered to the principles of the Declaration of Helsinki.
Results
Demographics and Comorbidities
Twenty-two eyes of 17 patients met the inclusion criteria (Table 1). The mean age at diagnosis of CMV retinitis was 49 years (range, 33-77). Ten patients were women (58.8%), 6 were men (35.3%), and 1 patient identified as nonbinary (5.9%). Self-reported racial/ethnic backgrounds included Black/African American (n = 5 [29.4%]), White/Caucasian (n = 5 [29.4%]), Asian (n = 3 [17.6%]), Hispanic (n = 1 [5.9%]), and mixed/other heritage (n = 3 [17.6%]).
Patient Demographics.
Only 1 patient (5.9%) was immunocompetent. The remaining patients were immunocompromised, with iatrogenic immunosuppression accounting for the largest proportion (n = 10 [58.8%]) and HIV infection for the remainder (n = 6 [35.3%]). Hypertension (n = 11 [64.7%]) and DM (n = 9 [52.9%]) were the most common systemic comorbidities, followed by cardiovascular disease and hyperlipidemia (n = 7 [41.2% each]). Primary open-angle glaucoma was present in 2 patients (11.8%).
Clinical Presentation
At presentation, patients’ VA ranged from 20/20 to light perception (Figure 1). The mean VA was 0.845 logMAR (20/140); by 12 months or final follow-up, VA declined to 1.345 logMAR (20/440). Retinitis involved the posterior pole (zone 1) at presentation in a substantial proportion of eyes (n = 9 [40.9%]), while other eyes presented with involvement of zone 2 (n = 9 [40.9%]) or zone 3 (n = 4 [18.2%]) (Table 2). The mean extent of retinitis was 2.9 ± 3.7 clock hours. Two eyes demonstrated involvement of the optic nerve (9.1%). Eight eyes (36.4%) were found to have vascular occlusive events. The mean CD4+ T-cell counts at presentation were 96.5 ± 13.4 cells/µL in iatrogenically immunosuppressed patients and 65.5 ± 28.2 cells/µL in HIV-positive patients. The CD4+ T-cell count in the immunocompetent patient was 550 cells/µL.
Cytomegalovirus retinitis. Acute stage: (A) Fundus photography shows characteristic retinal whitening with retinal hemorrhages along with vascular sheathing involving both arteries and veins. (B) Fundus autofluorescence demonstrates a hyperautofluorescent signal corresponding to active retinitis. (C) Optical coherence tomography (OCT) shows diffuse retinal thickening with hyperreflective foci in the vitreous, indicative of inflammation. Healed stage: (D) Fundus photography shows retinal scars at the sites of previous retinitis. (E) Fundus autofluorescence demonstrates hypoautofluorescence consistent with resolved inflammation. (F) OCT shows diffuse retinal atrophy.
Ocular Findings and Complications.
Treatment Modalities
All 22 eyes received IVT antiviral injections. Foscarnet was used alone in 68.18% of cases and a combination of foscarnet and ganciclovir in 31.82%. The mean number of injections per eye was 7.0 (range, 1-34) over 1 to 7 months. Systemic medications included oral valganciclovir in 95.5% of cases, with letermovir (18.2%) or maribavir (13.6%) added for patients with refractory CMV or intolerance to standard therapy. Topical steroids were used in 40.9% of cases and oral steroids in 63.6%, mainly for coexisting inflammatory conditions.
Immune Status
A total of 22 eyes of patients were analyzed, including 8 from HIV-positive patients, 13 from non–HIV-immunosuppressed patients (receiving immunosuppressive or immunomodulatory therapy), and 1 immunocompetent patient. The HIV-positive group was significantly younger than the non–HIV-immunosuppressed group, with a mean age of 48.6 ± 9.9 years compared with 61.9 ± 14.2 years (P = .021) (Table 3). The HIV-positive group showed significantly higher zone 1 involvement (75%) compared with the non–HIV-immunosuppressed group (23%), while involvement of zones 2 or 3 was more prevalent in non–HIV-immunosuppressed patients (77%) compared with HIV-positive cases (25%) (P = .032). Additionally, the HIV-positive group received fewer total IVT antiviral injections than the non–HIV-immunosuppressed group (2.1 ± 1.7 vs 8.9 ± 9.3) (P = .014). Furthermore, the HIV-positive group had a significantly shorter duration of IVT treatment, with a mean of 8.3 ± 8.9 days between the first and last injection, compared with 86.7 ± 102.9 days in the non–HIV-immunosuppressed group (P = .028).
Characteristics Stratified by Immune Status and Presence of Vascular Events.
P < .05 was considered to be statistically significant.
Abbreviation: VA, visual acuity.
HIV-positive patients had worse VA at diagnosis (logMAR 1.3 ± 1.0) than non–HIV-immunosuppressed patients (0.4 ± 0.4) (P = .058), and this disparity persisted through follow-up. At 1 month, 3 months, and 6 months, VA remained significantly worse in the HIV-positive group (all P < .05). At final follow-up, the difference remained clinically meaningful, although not statistically significant (1.8 ± 1.5 vs 0.8 ± 0.8) (P = .088). The degree of VA changes over time (ie, VA improvement or decline) did not differ significantly between the groups at any time point (all P > .4).
Vascular Occlusive Events and Other Complications
Vascular occlusive events were documented in 8 eyes (36.4%) belonging to 8 patients (Figure 2). Three eyes presented with BRAO, 2 with CRAO, and 3 with BRVO. Over the 12 months following the initial diagnosis, other complications included optical coherence tomography (OCT)-confirmed CME in 7 eyes (31.8%) and ischemic complications such as NV or VH in 4 eyes (18.2%) (Table 2).
Cytomegalovirus retinitis complicated by a branch retinal artery occlusion. (A) Fundus photography shows a peripheral superotemporal area of retinitis with sclerotic vessels in the superior retina. (B) Fluorescein angiography demonstrates hyperfluorescent optic disc and extensive area of capillary nonperfusion in the superior retina, indicative of occlusion and ischemic changes.
When patients with and without vascular occlusive events were compared, age showed a trend toward significance, with patients experiencing vascular events being, on average, 11.6 years older (P = .059). Logistic regression analysis demonstrated a nonsignificant trend suggesting that older age in patients with CMV retinitis may be associated with a slightly increased risk of vascular events (odds ratio, 1.067, 95% CI, 0.992-1.147; P = .079), with each additional year of age associated with a 6.7% increase in the odds of experiencing a vascular event. Although this did not reach conventional statistical significance, the directionality of the association suggests that age may be a clinically relevant risk factor (Table 3).
Among the 8 patients who experienced vascular occlusive events, a diagnosis of diabetes was the only systemic risk factor significantly associated with the development of vascular events, found in 88% of affected patients compared with 29% of unaffected patients (P = .024). VA outcomes at diagnosis and follow-up intervals (1, 3, 6, and 12 months) did not significantly differ between the groups. In addition, changes in VA over time were not significantly associated with the occurrence of vascular events, nor were treatment-related variables, including the total number of IVT antiviral injections and the injection interval between first and last injections significantly different between patients with and without vascular events.
Intravitreal Treatment Duration
Stepwise multivariate linear regression analysis found several independent predictors of IVT treatment duration in CMV retinitis (Figure 3). A higher total number of IVT antiviral injections was significantly associated with a longer treatment period (β = 9.205, 95% CI, 8.006-10.403) (P < .001). Long-term (12 months) visual decline was associated with longer duration of IVT treatment (β = 34.792, 95% CI, 22.705-46.880) (P < .001). In contrast, greater improvement in VA within the first month was independently associated with shorter duration of IVT treatment (β = −30.059, 95% CI, −46.538 to −13.580) (P = .002).
Factors associated with intravitreal treatment duration in patients with cytomegalovirus retinitis, based on multivariate linear regression. Bars represent the standardized β coefficients (effect sizes) of clinical variables on the duration between the first and last intravitreal (IVT) antiviral injection. More IVT antiviral injections (β = 9.2) and greater long-term visual decline at 12 months (β = +34.8) were associated with longer duration of IVT treatment. In contrast, greater early visual improvement at 1 month (β = −30.1), presence of vascular events (β = −33.0), and HIV-positive status (β = −27.4) were significantly associated with shorter durations of treatment.
The presence of vascular occlusive events was significantly associated with a shorter treatment period (β = −32.964, 95% CI, −52.177 to −13.751) (P = .002). Similarly, HIV-positive immune status was independently associated with shorter duration of IVT treatment (β = −27.436, 95% CI, −48.076 to −6.797) (P = .013). Overall, we found a strong association between these clinical characteristics and treatment length, highlighting both functional response and systemic status as key determinants.
Visual Outcomes
VA remained relatively stable during the first 3 months, with no significant differences between baseline and 1 month (P = .408) or baseline and 3 months (P = .575). The decline in VA became evident starting from the third month onward, with significant worsening detected between 3 and 6 months (P = .013), and between 3 and 12 months (P = .028). The comparison of VA between 1 and 12 months also trended toward significance (P = .029), further indicating delayed but progressive vision loss (Figure 4).
Changes in mean logMAR visual acuity (VA) from baseline through 12 months in patients with cytomegalovirus retinitis.
Among the 22 eyes, 7 (32%) maintained stable vision, 3 (14%) showed improvement, and 12 (54%) experienced a decrease in VA of 2 or more lines. An analysis of patients with vision loss revealed multiple contributing factors. Three patients had zone 1 involvement that progressed to macular atrophy, while 2 had baseline low vision with extensive retinitis. Three patients developed complications associated with vascular occlusive events, including neovascular glaucoma and VH. Two cases had severe CME, 1 patient developed cataract, and another had hypotony due to extensive retinitis. One case of unexplained vision loss was also noted.
Multivariate linear regression analysis identified several factors associated with final VA outcomes (Figure 5). A greater degree of vision loss within the first month was significantly associated with worse final VA (β = 0.950, 95% CI, 0.420-1.480) (P = .002), indicating that early decline in vision was a strong predictor of poor long-term outcome. Longer duration of IVT treatment was also associated with worse final VA (β = 0.015, 95% CI, 0.001-0.029 (P = .038). In contrast, a higher number of IVT antiviral injections (β = −0.154) was slightly associated with better final VA (β = −0.154, 95% CI, −0.307 to −0.001 (P = .048). In the stepwise regression models, vision change at 1 month remained a strong and consistent predictor of final VA (model 1: β = 1.033, 95% CI, 0.460-1.606, P = .001; model 2: β = 1.001, 95% CI, 0.481-1.522, P < .001). HIV-positive immune status was also independently associated with worse final VA in the adjusted model (β = 0.864, 95% CI, 0.070-1.658) (P = .035).
Factors associated with final logMAR visual acuity (VA) in patients with cytomegalovirus retinitis. Standardized β coefficients represent the strength and direction of association between each variable and final VA, with higher logMAR values indicating worse vision. A greater degree of vision loss within the first month (β = 0.950) and HIV-positive immune status (β = 0.864) were significantly associated with worse final VA. In contrast, a higher number of intravitreal antiviral injections (β = −0.154) was slightly associated with better final VA.
Other variables, including the presence of vascular occlusive events, ischemic complications, and retinitis zone involvement, were not significantly associated with final VA outcomes. Patients with a decline in VA of 2 or more lines had significantly greater worsening in longitudinal logMAR VA, particularly at 1, 3, and 6 months, as well as in the cumulative change from baseline to month 12. In contrast to initial expectations, there were no significant differences in baseline VA, patient age, extent of retinitis, or total number of antiviral injections between patients with and without a decline in VA of 2 or more lines. Although not statistically significant, a trend toward increased ischemic complications was observed in eyes with a decline of 2 or more lines, suggesting a potential association (P = .099).
Conclusions
Our findings underscore the multifaceted impact of CMV retinitis on immunocompromised patients, particularly those with advanced HIV/AIDS or iatrogenic immunosuppression. Historically, CMV retinitis was mostly associated with AIDS in HIV-positive individuals. 16 However, because effective combination antiretroviral therapy has reduced its incidence among people living with HIV, the disease has become increasingly recognized in non–HIV-immunosuppressed populations, such as solid organ transplant recipients or patients receiving immunomodulatory therapy for autoimmune conditions. 17
Our study found that although HIV-positive patients demonstrated significantly greater extension of retinitis, both in terms of clock hour involvement and increased frequency of zone 1 (posterior pole) disease, paradoxically they received fewer IVT antiviral injections and had shorter treatment durations compared with HIV-negative patients. Despite this shorter and less intensive treatment course, their VA outcomes were consistently worse at all time points. This suggests that the disease in HIV-positive patients may present more acutely with severe posterior involvement, limiting visual prognosis from the outset. The more posterior location of retinitis (zone 1) likely accounts for the poorer VA outcomes because central vision is critically affected by this region. Furthermore, the shorter treatment duration and fewer total injections in the HIV-positive group may reflect either more rapid disease resolution due to effective systemic control or, alternatively, a ceiling effect on visual recovery due to early irreversible macular damage. These findings underscore the distinct clinical trajectory of CMV retinitis in HIV-positive patients, characterized by posterior-predominant, aggressive retinal involvement with limited visual recovery despite a shorter treatment duration.
In our cohort, nearly 60% of the individuals were iatrogenically immunosuppressed rather than HIV positive, reflecting a shift in the epidemiology of CMV retinitis that has been reported in other recent studies.18,19 This highlights the importance of vigilance for CMV retinitis in any patient with significant immunosuppression presenting with new visual symptoms, regardless of HIV status. In the current study, iatrogenically immunosuppressed patients required significantly longer durations of treatment, highlighting a potential unmet need in this growing patient population. Unlike HIV-positive individuals, whose therapy may be de-escalated after immune recovery on combination antiretroviral therapy, patients receiving chronic immunosuppressive therapy (eg, for organ transplantation or autoimmune conditions) may require sustained antiviral treatment. This observation underscores the need for longer-acting therapeutic strategies, such as extended-release implants, that are tailored to the specific needs of non–HIV-immunosuppressed individuals. We believe that larger, prospective studies are needed to further characterize the treatment course in this population and guide long-term management.
Vascular occlusive events were observed in more than one-third of eyes in our study and were significantly associated with older age and DM. Vascular events were also associated with short treatment durations and did not correlate with the extent or zone of retinitis, suggesting that these events are not strongly correlated. In our study, vascular events appeared most frequently at presentation. Vascular occlusion was not directly associated with worse final VA, but it was significantly linked to ischemic complications, including NV and VH, both of which were disproportionately present among patients with vision loss of 2 or more lines. Although the correlation did not reach statistical significance, it did show a trend toward poorer visual outcomes (P = .099).
Another key observation in our study was the high incidence of vascular complications, affecting 41% of the eyes. Previous research has shown that CMV retinitis can precipitate both arterial and venous occlusions, likely through a combination of direct viral endothelial damage and inflammatory mechanisms. 12 Indeed, several eyes in our cohort presented with BRAO, CRAO, or BRVO, while a new BRAO developed in 1 eye over the 12-month follow-up period. These occlusive events often correlate with abrupt declines in VA and can significantly worsen overall prognosis. They also raise the risk of secondary complications such as NV, VH, and RD.20,21 Our findings are consistent with previous work by Davis et al, 12 who reported a particularly high frequency of retinal arteriolar occlusions in older, non–HIV-infected patients with CMV retinitis. In that study, 71% of older immunocompetent patients experienced vascular occlusion compared with just 5% of younger HIV-infected individuals. These results support the hypothesis that age-related vascular vulnerability, rather than CMV retinitis severity, per se, may predispose certain patients to ischemic events. Together, these data emphasize the need for clinicians to remain vigilant for vascular complications in older patients or patients with diabetes with CMV retinitis, even when the retinitis appears limited in extent.
In addition to vascular events, we observed a considerable burden of CME, which occurred in approximately one-third of the eyes. CME in the setting of CMV retinitis may arise from viral-induced inflammation, vascular leakage, or as a result of retinal ischemia. 22 Its presence often portends worse visual outcomes and necessitates timely intervention. Similarly, ischemic complications, NV, and RD, although less common, further underscore the severity of this disease. 23
From baseline until 3 months, VA remained relatively stable; however, progressive vision loss became evident thereafter, with significant deterioration between months 3 and 6 and further decline through 12 months. Overall, 55% of eyes experienced a decline in VA of 2 or more lines, with a variety of contributing complications, including vascular events, CME, cataract, hypotonia, and macular atrophy. Multivariate regression analysis revealed that early functional decline (VA loss at 1 month), prolonged IVT treatment duration, and higher number of antiviral injections were all independently associated with worse final VA. HIV status also remained an independent predictor of poor visual outcome. These findings highlight the complex interplay between immune status, retinal involvement, and functional response in shaping long-term vision outcomes.
Therapeutically, our data highlight the role of IVT injections (primarily ganciclovir or foscarnet) in controlling active infection and preventing progression, with a mean of 7.0 IVT injections per eye over 12 months. IVT therapy offers the advantage of high local drug concentrations while minimizing systemic toxicity, which is especially relevant in patients who may already be burdened by multiple systemic medications. 24 Systemic therapy, predominantly with oral valganciclovir, remains an essential adjunct, particularly for patients with extraocular CMV involvement or bilateral disease. However, toxicity profiles and the emergence of resistant strains necessitate the exploration of newer agents, such as letermovir and maribavir, which were used in select cases in our cohort.
Optimal care for CMV retinitis is inherently multidisciplinary. Coordination among ophthalmologists, infectious disease specialists, transplant physicians, and pharmacists is essential to balance antiviral efficacy, systemic toxicity, and the tempo of immune reconstitution. Cost-effectiveness analyses and consensus guidelines will be important to standardize such surveillance strategies.
This study has several limitations, including its retrospective design, which may introduce selection bias and limit the ability to establish causality. The relatively small sample size reduces statistical power, particularly in subgroup analyses, and the heterogeneity of immune status and comorbidities may confound comparisons. Additionally, treatment approaches and follow-up protocols were not standardized across all cases, which may have impacted outcomes. Despite these limitations, this is the first original study to statistically analyze and directly compare vascular occlusive events and immune status subgroups in patients with CMV retinitis, offering valuable insight into an understudied and clinically significant dimension of the disease. By integrating both descriptive and inferential statistics, this study provides a nuanced understanding of the relationship between immune suppression, disease distribution, treatment course, and vascular complications. These findings address an unmet need in the CMV retinitis literature and may help inform risk stratification and management strategies in diverse immunosuppressed populations.
Our results underscore that CMV retinitis remains a sight-threatening condition with a high rate of vascular, inflammatory, and ischemic complications. Advances in imaging technologies, such as wide-field fluorescein angiography and optical coherence tomography angiography, may also enhance the early detection of vascular involvement and guide treatment intensity. Importantly, vascular occlusive events were not associated with the extent or zone of retinitis and appeared most commonly at presentation, particularly in older patients and patients with diabetes. While not directly predictive of poor VA outcomes, these events were significantly associated with ischemic complications that contributed to vision loss. Aggressive antiviral treatment remains the cornerstone of management, but our findings highlight the need for tailored approaches in non–HIV-immunosuppressed individuals, who often require longer durations of IVT treatment. Future research should focus on prospective, multicenter studies to refine therapeutic protocols and explore the use of emerging antivirals and sustained-release therapies.
In conclusion, the current study contributes novel insights into the complex interplay between immune status, vascular occlusive events, and treatment outcomes in CMV retinitis. It draws attention to the evolving epidemiology of the disease and highlights the need for ongoing research in this vulnerable and rapidly growing patient population.
Footnotes
Authors’ Note
Mr. Akova and Dr. Sarici contributed equally to this work.
Ethical Approval
Institutional research ethics board approval was received from Emory School of Medicine. This study adheres to the tenets of the Declaration of Helsinki.
Statement of Informed Consent
All patients provided written informed consent before the study.
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Research to Prevent Blindness Challenge Grant; National Institutes of Health (NIH) P30EY006360; Sitaraman Family; Eberle Family.
Data Availability Statement
De-identified data supporting the findings of this study are available from the corresponding author upon reasonable request.