Purpose
To evaluate intravitreal aflibercept injections (IAI; also called VEGF Trap-Eye) for patients with macular edema secondary to central retinal vein occlusion (CRVO).
Design
Randomized controlled trial.
Methods
This multicenter study randomized 189 patients (1 eye/patient) with macular edema secondary to CRVO to receive 6 monthly injections of either 2 mg intravitreal aflibercept (IAI 2Q4) (n = 115) or sham (n = 74). From week 24 to week 52, all patients received 2 mg intravitreal aflibercept as needed (IAI 2Q4 + PRN and sham + IAI PRN) according to retreatment criteria. The primary endpoint was the proportion of patients who gained ≥15 ETDRS letters from baseline at week 24. Additional endpoints included visual, anatomic, and quality-of-life NEI VFQ-25 outcomes at weeks 24 and 52.
Results
At week 24, 56.1% of IAI 2Q4 patients gained ≥15 letters from baseline compared with 12.3% of sham patients ( P < .001). At week 52, 55.3% of IAI 2Q4 + PRN patients gained ≥15 letters compared with 30.1% of sham + IAI PRN patients ( P < .001). At week 52, IAI 2Q4 + PRN patients gained a mean of 16.2 letters of vision vs 3.8 letters for sham + IAI PRN ( P < .001). The most common adverse events for both groups were conjunctival hemorrhage, eye pain, reduced visual acuity, and increased intraocular pressure.
Conclusions
Monthly injections of 2 mg intravitreal aflibercept for patients with macular edema secondary to CRVO resulted in a statistically significant improvement in visual acuity at week 24, which was largely maintained through week 52 with intravitreal aflibercept PRN dosing. Intravitreal aflibercept injection was generally well tolerated.
The management of eyes with vision loss from macular edema secondary to central retinal venous occlusion (CRVO) has been entirely transformed in recent years. Both anti–vascular endothelial growth factor (VEGF) and steroidal pharmacotherapies have been developed, tested, and put into clinical practice, targeting the vascular permeability and leakage that frequently develops following blockage of the central retinal vein. Intravitreal aflibercept injection, also known as VEGF Trap-Eye (Regeneron Pharmaceuticals Inc, Tarrytown, New York, USA) is a 115-kDa decoy receptor fusion protein, composed of the second domain of human VEGF receptor 1 and the third domain of VEGF receptor 2 fused to the Fc domain of human IgG1. The binding affinity of aflibercept for VEGF is substantially greater than that of either bevacizumab or ranibizumab, and mathematical modeling has predicted its potential for a longer duration of action in the eye.
The current phase 3 randomized, sham-controlled clinical trial (COPERNICUS) recently reported that aflibercept, given as a monthly intravitreal injection in eyes with macular edema attributable to CRVO, improved visual acuity (VA) and central retinal thickness, was associated with no progression to neovascularization, and had a low rate of ocular adverse events at 24 weeks. Beginning at week 24, patients in both groups were eligible to receive 2 mg of intravitreal aflibercept injection (IAI) as needed (pro re nata; PRN). Therefore the patients in the sham + IAI PRN group received a different dosing regimen from weeks 24 onward compared with the start of treatment in those patients who had received intravitreal aflibercept initially (every 4 weeks). This report follows these patients to 1 year.
Methods
The COPERNICUS study is an ongoing 2-year, phase 3, prospective, randomized, double-masked trial. This multicenter study was conducted across 70 sites in the United States, Canada, Colombia, India, and Israel. The study protocol of the COPERNICUS trial was approved by the institutional review board or ethics committee at each participating clinical center before the start of the study. This trial was registered with ClinicalTrials.gov (identifier #NCT00943072 ). All patients signed a written consent form before initiation of the study-specific procedures. This study was conducted in compliance with regulations of the Health Insurance Portability and Accountability Act and the Declaration of Helsinki.
Treatments
Patients were randomly allocated using a 3:2 ratio to receive intravitreal aflibercept injections, 2 mg (IAI 2Q4), or sham injections, every 4 weeks for 24 weeks, for a total of 6 monthly treatments ( Supplemental Figure 1 , available at AJO.com ). Between weeks 24 and 52, patients in both treatment groups were evaluated monthly and were injected with intravitreal aflibercept as needed if they met protocol-specified retreatment criteria. They received a sham injection if retreatment was not indicated. After the first year of masked dosing, patients are continuing in a 1-year extension phase with PRN dosing. Data for this 52-week report were collected between July 2009 and April 2011.
Randomization was stratified by geographic region (North America [Canada and the United States] vs the rest of the world [Colombia, India, and Israel]) and by using a baseline best-corrected visual acuity (BCVA) score (>20/200, ie, 35 to 73 Early Treatment Diabetic Retinopathy Study [ETDRS] letters; and ≤20/200 ie, 24 to 34 ETDRS letters). Only 1 eye per patient was included in the randomization. All patients were eligible to receive panretinal photocoagulation at any time during the study if they progressed to neovascularization of the anterior or posterior segment.
Participants
The study enrolled patients aged ≥18 years with center-involved macular edema secondary to CRVO diagnosed within 9 months of study initiation. All study eyes had mean central subfield retinal thickness ≥250 μm using optical coherence tomography (OCT) from Zeiss Stratus OCT (Version 4.0 or later; Carl Zeiss Meditec, Jena, Germany), and protocol refracted ETDRS BCVA of 20/40 to 20/320 (73 to 24 letters).
Key exclusion criteria for the study eye included: any previous treatment with antiangiogenic drugs; prior panretinal or macular laser photocoagulation; and any ocular disorders that could confound interpretation of study results. Exclusion criteria with respect to both eyes included: previous use of intraocular corticosteroids or use of periocular corticosteroids within the 3 months prior to day 1; iris neovascularization, vitreous hemorrhage, traction retinal detachment, or preretinal fibrosis involving the macula; history or presence of age-related macular degeneration (AMD; dry or wet form) that significantly affected central vision; diabetic macular edema or diabetic retinopathy, defined as eyes of diabetic subjects with more than 1 microaneurysm outside the area of the vein occlusion; and infectious blepharitis, keratitis, scleritis, or conjunctivitis.
Endpoints and Assessments
The primary efficacy endpoint was the proportion of eyes with a gain of ≥15 ETDRS letters in BCVA from baseline to week 24. Secondary efficacy endpoints (all assessed at week 24) were: change from baseline in BCVA scores; change from baseline in central retinal thickness (CRT); proportion of patients progressing to neovascularization of anterior segment, optic disc, or elsewhere in the retina; and change from baseline in the National Eye Institute 25-item Visual Function Questionnaire (NEI VFQ-25) in total and subscale scores (distance activities, near activities, and vision dependency). The tertiary efficacy endpoints were all of the parameters mentioned above measured at 52 weeks.
A masked physician was assigned to assess adverse events (AEs), supervise the masked assessment of efficacy, and decide on the need for retreatment during the PRN phase. Visual acuity assessors were masked as to treatment assignment.
Assessments were conducted at regularly scheduled clinic visits on day 1 and every 4 weeks from weeks 4 to 52. BCVA was evaluated by certified examiners using the ETDRS refraction protocol. Retinal characteristics from OCT scans were assessed at a masked independent central reading center (Duke Reading Center, Durham, North Carolina, USA). Central retinal thickness was defined as the thickness of the center subfield (the area of the retina using a 1-mm diameter around the center of the macula). Fundus photography and fluorescein angiography were used to evaluate the anatomy of the retinal vasculature. Angiographic images were reviewed by masked graders at an independent reading center (Digital Angiographic Reading Center, New York, New York, USA). Vision-related quality of life was assessed using the NEI VFQ-25, which was administered by masked site personnel prior to intravitreal injection.
Eyes were evaluated, starting at week 24, for retreatment and received an injection of intravitreal aflibercept if any of the following retreatment criteria were met: a >50 μm increase in CRT on OCT compared with lowest previous measurement; new or persistent cystic retinal changes or subretinal fluid on OCT, or persistent diffuse edema ≥250 μm in the central subfield on OCT; a loss of ≥5 letters from the best prior measurement in conjunction with any increase in CRT on OCT; or an increase of VA between the current and most recent visit of ≥5 letters. If none of these retreatment criteria were met, patients received a sham injection.
Safety was monitored with the recording of ocular and nonocular AEs and laboratory measures.
Statistical Analyses
The full analysis set (FAS), on which the primary efficacy analyses were conducted, included all randomized patients who received any study medication, had a baseline efficacy assessment, and had at least 1 postbaseline efficacy assessment. In the primary endpoint analysis, patients who discontinued prematurely (prior to week 24) and had fewer than 5 injections were evaluated as nonresponders. The last-observation-carried-forward method was used to impute missing values. The proportions of patients who gained 15 letters were compared with a 2-sided Cochran-Mantel-Haenszel test and randomization was stratified by region and baseline BCVA. Secondary endpoint analyses were conducted sequentially according to the order in which the variables were predefined to preserve an alpha of 0.05. Proportions were analyzed using the Cochran-Mantel-Haenszel test. Time to first injection was analyzed using Kaplan-Meier methodology. The Cox proportional hazards model was used to quantify the differences in the rate of time to first injection between treatment groups.
Continuous variables were analyzed with an analysis of covariance main effects model with treatment group, region, and baseline BCVA as fixed factors, and the respective baseline variable as a covariate.
Ocular serious adverse events (SAEs) included any AE that: caused a decrease in VA of >30 letters (compared with the most recent assessment) or a decrease in VA to the level of light perception or worse that lasted >1 hour; required medical or surgical intervention to prevent permanent loss of sight; or was associated with severe intraocular inflammation.
The sample size calculation was based on the assumptions that the dropout rate would be 9% for each arm and the difference in the proportion of eyes gaining at least 15 letters of vision would be 25% (ie, 15% in the sham group and 40% in the intravitreal aflibercept 2Q4 group. Therefore, a total sample size of 165 eyes was required to detect a difference in the primary analysis with 90% power at a significance level of 5% using a 2-sided Fisher exact test or a Cochran-Mantel-Haenszel test.
Results
Patient Disposition
A total of 189 patients were randomized to intravitreal aflibercept 2Q4 + PRN (n = 115) and sham + IAI PRN (n = 74). With the exception of 1 patient in the IAI 2Q4 + PRN group, all randomized patients received study drugs. The majority of patients (57/74, 77.0% sham + IAI PRN and 107/115, 93.0% IAI 2Q4 + PRN) completed the first 52 weeks of the study ( Supplemental Figure 2 , available at AJO.com ). The primary reasons for premature discontinuation from the study before week 52 were withdrawal of consent (5/115, 4.3%) for the IAI 2Q4 + PRN group and adverse event (4/74, 5.4%) and treatment failure (4/74, 5.4%) in the sham + IAI PRN group. Treatment failures and AEs were the main reasons for the larger proportion of discontinuations in the sham + IAI PRN group compared with the IAI 2Q4 + PRN patient group, which had no discontinuations attributable to AEs or treatment failures during the first 52 weeks of the study. Adverse events resulting in the study discontinuation in the sham + IAI PRN group included vitreous and retinal hemorrhages, reduced visual acuity, and iris neovascularization, which all occurred before week 24 and were consistent with the complications of CRVO.
Demographics and Baseline Disease Characteristics
Most patients were male (107/187; 57%), white (147/187; 78.6%), and originating from North America (159/187; 85%) ( Table 1 ). The majority of patients (127/187; 67.9%) had fewer than 10 disc areas of nonperfusion on reading center evaluation.
| IAI 2Q4 + PRN (n = 114) d | Sham + IAI PRN (n = 73) | Total (n = 187) e | |
|---|---|---|---|
| Age (y), mean (SD) | 65.5 (13.57) | 67.5 (14.29) | 66.3 (13.85) |
| Sex, n (%) | |||
| Female | 45 (39) | 35 (48) | 80 (43) |
| Male | 69 (61) | 38 (52) | 107 (57) |
| Race, n (%) | |||
| White | 88 (77.2) | 59 (80.8) | 147 (78.6) |
| Black | 5 (4.4) | 5 (6.8) | 10 (5.3) |
| Asian | 7 (6.1) | 2 (2.7) | 9 (4.8) |
| Other a | 14 (12.3) | 7 (9.6) | 21 (11.2) |
| Geographic region, n (%) | |||
| North America | 95 (83.3) | 64 (87.7) | 159 (85.0) |
| Rest of world | 19 (16.7) | 9 (12.3) | 28 (15.0) |
| Visual acuity (ETDRS) | |||
| Mean (SD) | 50.7 (13.90) | 48.9 (14.42) | 50.0 (14.09) |
| BCVA >20/200 (letters read ≥35) | 86 (75.4) | 55 (75.3) | 141 (75.4) |
| BCVA ≤20/200 (letters read ≤34) | 28 (24.6) | 18 (24.7) | 46 (24.6) |
| Retinal perfusion status, n (%) | |||
| Perfused b | 77 (67.5) | 50 (68.5) | 127 (67.9) |
| Nonperfused | 17 (14.9) | 12 (16.4) | 29 (15.5) |
| Indeterminate | 20 (17.5) | 11 (15.1) | 31 (16.6) |
| Retinal thickness (μm) | |||
| Mean (SD) | 661.7 (237.37) | 672.4 (245.33) | 665.8 (239.82) |
| IOP (mm Hg), mean (SD) | 15.1 (3.26) | 15.0 (2.81) | 15.1 (3.08) |
| Time since CRVO diagnosis (mo) | |||
| Mean (SD) | 2.73 (3.09) | 1.88 (2.19) | 2.40 (2.796) |
| ≤2 months | 64 (56.1) | 52 (71.2) | 116 (62.0) |
| >2 months | 49 (43.0) | 21 (28.8) | 70 (37.4) |
| NEI VFQ-25 total score, c mean (SD) | 77.39 (16.176) | 77.38 (16.602) | 77.39 (16.299) |
| NEI VFQ-25 near activities score, c mean (SD) | 69.96 (21.939) | 70.72 (20.222) | 70.25 (21.234) |
| NEI VFQ-25 distance activities score, mean (SD) | 75.99 (21.255) | 78.08 (21.258) | 76.80 (21.224) |
| Vision dependency score, mean (SD) | 83.26 (25.511) | 82.76 (27.405) | 83.07 (26.195) |
a Included not reported and multiple races.
b Less than 10 disc areas of nonperfusion.
c Baseline total and near activities subscale scores changed slightly from weeks 24 to 52 as a result of multiple-choice options in the questionnaires changing from 4 possible responses at week 24 to 5 possible responses at week 52.
d 113 for the time since CRVO diagnosis.
Efficacy
At week 52, the proportion of patients who gained at least 15 letters in BCVA was 55.3% in the IAI 2Q4 + PRN group vs 30.1% in the sham + IAI PRN group ( Figure 1 ), demonstrating that IAI 2Q4 + PRN continued to be superior at week 52. For comparison, at week 24, 56.1% of patients in the IAI 2Q4 group had gained ≥15 letters from baseline compared with 12.3% of patients in the sham group ( P < .001). The majority of patients in both groups gained ≥0 letters of vision (92.1% of patients in the IAI 2Q4 + PRN group compared with 68.5% of those in the sham + IAI PRN group at week 52) ( Table 2 ).
| Week 24 | Week 52 | |||
|---|---|---|---|---|
| IAI 2Q4 + PRN (n = 114) | Sham + IAI PRN (n = 73) | IAI 2Q4 + PRN (n = 114) | Sham + IAI PRN (n = 73) | |
| Letter gain, n (%) | ||||
| ≥15 letters a | 64 (56.1) | 9 (12.3) | 63 (55.3) | 22 (30.1) |
| ≥10 letters | 87 (76.3) | 16 (21.9) | 88 (77.2) | 34 (46.6) |
| ≥5 letters | 97 (85.1) | 29 (39.7) | 93 (81.6) | 43 (58.9) |
| ≥0 letters | 107 (93.9) | 38 (52.1) | 105 (92.1) | 50 (68.5) |
| Letter loss, n (%) | ||||
| >0 letter | 7 (6.1) | 35 (47.9) | 9 (7.9) | 23 (31.5) |
| ≥5 letters | 5 (4.4) | 29 (39.7) | 8 (7.0) | 17 (23.3) |
| ≥10 letters | 2 (1.8) | 22 (30.1) | 6 (5.3) | 13 (17.8) |
| ≥15 letters | 2 (1.8) | 20 (27.4) | 6 (5.3) | 11 (15.1) |
At week 52, patients in the IAI 2Q4 + PRN group showed a mean change from baseline BCVA of 16.2 ETDRS letters ( Figure 2 ). For comparison, at week 24, the mean change from baseline BCVA was 17.3 ETDRS letters in the IAI 2Q4 + PRN group and −4.0 ETDRS letters in the sham group ( P < .001). When the sham group was later eligible to receive intravitreal aflibercept (following an as-needed dosing regimen), mean change from baseline in BCVA improved in this group from −4.0 letters at week 24 to +3.8 letters at week 52, a gain of 7.8 letters. The waterfall analysis ( Supplemental Figure 3 , Top and Bottom panels) (Supplemental Material available at AJO.com ) of individual patient responses found that only 7.9% of patients who were originally randomized to the IAI 2Q4 treatment group experienced a loss of vision at week 52 (vs baseline) compared with 31.5% of patients in the sham + IAI PRN group.
At week 52, a gain of ≥15 letters was noted in 60.7% vs 22.2% of patients with a baseline BCVA ≤20/200 in the IAI 2Q4 + PRN group vs the sham + IAI PRN group and in 53.5% vs 32.7% of patients, respectively, with a baseline BCVA >20/200. For patients with a baseline BCVA ≤20/200, the mean change from baseline at week 52 in BCVA letter score was +19.9 vs +5.1 letters for IAI 2Q4 + PRN compared with sham + IAI PRN. Patients who had a baseline BCVA of >20/200 had improvements in BCVA of +14.9 compared with +3.5 letters.
In eyes with over 10 disc areas of posterior nonperfusion (that seen on 7 standard field fluorescein angiogram) at baseline, the proportion of eyes gaining ≥15 letters at week 24 was 51.4% vs 4.3% for IAI 2Q4 + PRN vs sham + IAI PRN treatment and 58.4% vs 16.0% in eyes with less than 10 disc areas of posterior nonperfusion at baseline, respectively. At 52 weeks, these proportions were 48.6% vs 30.4% for IAI 2Q4 + PRN vs sham + IAI PRN treatment in eyes with posterior nonperfusion, and 58.4% vs 30.0% in eyes without posterior nonperfusion eyes, respectively.
If the diagnosis was within 2 months of treatment, the proportions of eyes gaining ≥15 letters were 64.1% vs 34.6% for IAI 2Q4 + PRN vs sham + IAI PRN treatment at 52 weeks (difference of 29.4%), and 42.9% vs 19.0% (difference of 23.8%) if the time since diagnosis was greater than 2 months.
The rapid reduction in CRT observed in the IAI 2Q4 + PRN group through week 24 was largely maintained through week 52 (−457.2 μm and −413.0 μm, respectively) ( Figure 3 ). At week 52, mean CRT reductions were similar in both treatment groups (413.0 μm for IAI 2Q4 + PRN vs 381.8 μm for sham + IAI PRN).
During the first 52 weeks, no eyes in the IAI 2Q4 + PRN group developed any neovascularization compared with 5 eyes (6.8%, all in the anterior segment) for the sham + IAI PRN group ( P = .006 by Cochran-Mantel-Haenszel test). Panretinal photocoagulation was performed for 4 of the patients (5.5%) in the sham + IAI PRN group.
The between-group differences in the mean NEI VFQ-25 total and subscale scores were significant at week 24. A clinically relevant improvement in the mean NEI VFQ-25 total score was observed in both IAI 2Q4 + PRN (7.5 points) and sham + IAI PRN (5.1 points) groups at week 52 ( Figure 4 ). A significant difference between the treatment arms was not seen at week 52 because of the crossover to intravitreal aflibercept PRN in the sham group.
Treatment Experience
Beginning at week 24, all patients were eligible to receive an injection of aflibercept if any of the protocol-defined retreatment criteria were met. During the PRN phase of the study, 57 of the 60 patients in the sham + IAI PRN group and 102 of the 110 patients in the IAI 2Q4 + PRN group had at least 1 injection ( Figure 5 , Top panel). Eighty-three percent of the sham + IAI PRN patients compared with 30% of IAI 2Q4 + PRN patients had their first injection at week 24. The median time to first injection in the PRN phase was 29 days for the sham + IAI PRN group and 68 days for the IAI 2Q4 + PRN group. The mean numbers of injections were 3.9 (SE = 0.3) and 2.7 (SE = 0.2) in the sham + IAI PRN and IAI 2Q4 + PRN groups, respectively ( Figure 5 , Bottom panel).
