Visual Acuity, Vitreous Hemorrhage, and Other Ocular Outcomes After Vitrectomy vs Aflibercept for Vitreous Hemorrhage Due to Diabetic Retinopathy: A Secondary Analysis of a Randomized Clinical Trial

Abstract

Importance: Although there were no differences in mean visual acuity (VA) over 24 weeks after vitrectomy with panretinal photocoagulation (PRP) vs aflibercept in a randomized clinical trial among eyes with vitreous hemorrhage due to proliferative diabetic retinopathy (PDR), post hoc analyses may influence treatment choices.

Objective: To compare exploratory outcomes between treatment groups that may affect treatment choices for patients with vitreous hemorrhage due to PDR.

Design, setting, and participants: This post hoc analysis of a randomized clinical trial conducted at 39 DRCR Retina Network sites included adults with vision loss due to PDR-related vitreous hemorrhage for whom vitrectomy was considered. Data were collected from November 2016 to January 2020.

Interventions: Random assignment to 4 monthly injections of aflibercept vs vitrectomy with PRP. Both groups could receive aflibercept or vitrectomy during follow-up based on protocol-specific criteria.

Main outcomes and measures: Visual acuity area under the curve (adjusted for baseline VA) and clearance of vitreous hemorrhage.

Results: A total of 205 eyes were included in the analysis (115 male [56%] and 90 [44%] female participants; mean [SD] age, 57 [11] years). Among 89 eyes with a baseline VA of 20/32 to 20/160 (47 receiving aflibercept, including 4 [9%] that had undergone vitrectomy; 42 undergoing vitrectomy, including 3 [7%] that had received aflibercept), the adjusted mean difference in VA letter score over 24 weeks between the aflibercept and vitrectomy groups was -4.3 (95% CI, -10.6 to 1.9) compared with -16.7 (95% CI, -24.4 to -9.1) among 59 eyes with baseline VA worse than 20/800 (P = .02 for interaction; 26 in the aflibercept group, including 6 [23%] that had undergone vitrectomy; 33 in the vitrectomy group, including 8 [24%] that had received aflibercept). In the full cohort, the median time to clearance of the initial vitreous hemorrhage was 36 (interquartile range [IQR], 24-52) weeks in the aflibercept group vs 4 (IQR, 4-4) weeks in the vitrectomy group (difference, 32 [95% CI, 20-32] weeks; P < .001).

Conclusions and relevance: Both initial aflibercept and vitrectomy with PRP are viable treatment approaches for PDR-related vitreous hemorrhage. Although this study did not find a significant difference between groups in the primary outcome of mean VA over 24 weeks of follow-up, eyes receiving initial vitrectomy with PRP had faster recovery of vision over 24 weeks when baseline VA was worse than 20/800 and faster vitreous hemorrhage clearance. Approximately one-third of the eyes in each group received the alternative treatment (aflibercept or vitrectomy with PRP). These factors may influence treatment decisions for patients initiating therapy for PDR-related vitreous hemorrhage.

Trial registration: ClinicalTrials.gov Identifier: NCT02858076.

Conflict of interest statement

Conflict of Interest Disclosures: Mr Glassman reported receiving grants from the National Eye Institute (NEI), Regeneron Pharmaceuticals, Inc, and Genentech, Inc. Dr Beaulieu reported receiving grants to his institution from the NEI, Regeneron Pharmaceuticals, Inc, and Genentech, Inc. Dr Maguire reported receiving grants to her institution from the NEI, Regeneron Pharmaceuticals, Inc, and Genentech, Inc, and professional fees paid to her from Genentech, Inc. Dr Antoszyk reported receiving grants from Roche Genentech and personal/consultant fees from Jaeb Center for Health Research, Opthea, Clearside Biomedical, Inc, and Roche Genentech. Dr Elman reported receiving professional fees from Allergan plc, MEDIforce DME, and Genentech, Inc, and clinical/epidemiological research support from Apellis Pharmaceuticals, Inc, Graybug Vision, Inc, National Institutes of Health (NIH), Neurotech FPO, Novartis AG, Optos, and Global Life Science Ltd. Dr Jampol reported receiving grants from the NEI and personal fees from Sanofi SA. Dr Sun reported receiving grants from Roche Genentech, Juvenile Diabetes Research Foundation, and KalVista Pharmaceuticals, Inc; personal fees from Current Diabetes Reports, JAMA Ophthalmology, and Merck & Co, Inc; and nonfinancial support from Optovue (equipment loaned for research), Roche Genentech (food/beverage), and KalVista Pharmaceuticals, Inc (travel support). No other disclosures were reported.

Figures

Figure 1.. Mean Visual Acuity (VA) Over 2 Years by Treatment Group and Baseline VA Subgroup

Error bars represent 95% CIs. Best-corrected VA was collected after protocol-defined refraction. Visual acuity was measured with the electronic Early Treatment Diabetic Retinopathy Study VA test on a scale from 100 (Snellen equivalent, 20/10) to 0 letters (Snellen equivalent, worse than 20/800). PRP indicates panretinal photocoagulation.

Figure 2.. Presence of Vitreous Hemorrhage Over 2 Years by Treatment Group

Presence of vitreous hemorrhage was assessed by the investigator during clinical examination. Error bars represent 95% CIs for the proportion of eyes with vitreous hemorrhage. PRP indicates panretinal photocoagulation.

Figure 3.. Time to Clearance of Initial Vitreous Hemorrhage by Treatment Group

Presence of vitreous hemorrhage was assessed by the investigator during clinical examination. PRP indicates panretinal photocoagulation.

Figure 4.. Time to Resolution of Retinal Neovascularization by Treatment Group

Retinal neovascularization was defined as neovascularization of the disc or elsewhere as assessed by the investigator during clinical examination. PRP indicates panretinal photocoagulation.

Figure 5.. Presence of Retinal Neovascularization Over 2 Years by Treatment Group

Presence of retinal neovascularization was assessed by the investigator during clinical examination. Error bars represent 95% CIs for the proportion of eyes with retinal neovascularization. Retinal neovascularization could not be assessed at 4, 12, 24, 36, 52, 68, 84, and 104 weeks in 42, 25, 12, 5, 6, 9, 7, and 2 eyes in the aflibercept group and 15, 7, 6, 2, 3, 4, 2, and 4 eyes in the vitrectomy with PRP group, respectively (these eyes are excluded from the denominator of the percentage).