To investigate 5-year outcomes in eyes initially treated with aflibercept or bevacizumab for macular edema due to central retinal or hemiretinal vein occlusion.
Long-term follow-up (LTF) after a randomized clinical trial from 64 centers in the United States. Participants were followed up to 60 months and treated at investigator discretion after completing the 12-month treatment protocol. Main outcomes were visual acuity letter score (VALS) and central subfield thickness (CST) on optical coherence tomography.
Seventy-five percent (248/330) of eligible participants completed at least 1 visit between months 24 and 60, and 45% completed the month 60 visit. Among participants completing month 60, overall mean VALS improvement over baseline was 13.5 (95% CI: 9.6, 17.5), less than the mean improvement of 20.6 (95% CI: 18.7, 22.4) observed at month 12, with no significant differences between originally assigned study groups. Further, 66% (99/150) had at least 1 treatment between months 48 and 60 with a mean (SD) of 3.41 (3.69) treatments over this period. Mean CST was 671 μm at baseline and 261 μm (95% CI: 241.2, 280.9) at month 60.
Although VALS improved substantially when patients were treated per protocol through month 12, improvement lessened when treatment was at investigator discretion and fewer treatments were received although VALS remained markedly improved over baseline through year 5. Most patients continued to receive treatment in year 5. This suggests that continued monitoring and, if warranted, treatment with anti-VEGF therapy benefits patients with macular edema associated with central retinal or hemiretinal vein occlusion. Publication of this article is sponsored by the American Ophthalmological Society.
Retinal vein occlusion is the most common retinal vascular disease after diabetic retinopathy, affecting 1% to 2% of the population over the age of 40 years , and 16 million people worldwide. Macular edema is the most common cause of visual impairment in patients with retinal vein occlusion. Although focal grid laser treatment for macular edema was demonstrated to be associated with improved visual acuity outcomes in patients with branch retinal vein occlusion (BRVO) and reduced leakage on fluorescein angiography in patients with central retinal vein occlusion (CRVO), no significant visual acuity benefit of grid laser was found in patients with CRVO.
Intravitreal triamcinolone for macular edema showed no significant differences to focal grid laser treatment with respect to visual acuity outcomes in patients with BRVO and was superior to observation with respect to visual acuity outcomes in patients with CRVO. Intravitreal dexamethasone implant was associated with a reduced risk of vision loss and an increased incidence of visual acuity improvement in 2 multicenter, sham-controlled trials that included patients with BRVO and patients with CRVO. , The triamcinolone and dexamethasone trials demonstrated an increased risk of intraocular pressure elevation and cataract in the steroid treatment arms. More recently, several industry-sponsored phase III trials demonstrated significant visual acuity benefit of anti–vascular endothelial growth factor (anti-VEGF) therapy for treatment of macular edema associated with BRVO or CRVO.
The primary outcome results of the S tudy of CO mparative Treatments for RE tinal Vein Occlusion 2 (SCORE2) demonstrated that bevacizumab, a frequently used off-label anti-VEGF treatment, is noninferior to aflibercept, an anti-VEGF treatment approved by the United States Food and Drug Administration, with respect to visual acuity at 6 months in eyes with macular edema associated with CRVO or hemiretinal vein occlusion (HRVO). SCORE2 also showed that at 24 months postrandomization (12 months after cessation of the SCORE2 protocol-defined treatment schedule), participants randomized to aflibercept and those randomized to bevacizumab had similar visual acuity and central retinal thickness outcomes. The current investigation evaluates treatment patterns and longer-term outcomes among SCORE2 participants 5 years after initiation of treatment, which is 4 years after cessation of the SCORE2 protocol-defined treatment schedule.
The study adhered to the tenets of the Declaration of Helsinki and is registered on www.clinicaltrials.gov (identifier: NCT01969708). The study was approved by either a site-specific or a centralized institutional review board (IRB), and written informed consent was obtained from all participants. Between September 17, 2014, and November 18, 2015, a total of 305 patients with CRVO and 57 patients with HRVO (362 study participants in total) were randomly assigned to intravitreal injection of bevacizumab (1.25 mg) or aflibercept (2.0 mg) at randomization and every 4 weeks through month 5.
The primary outcome was change from baseline in best-corrected electronic Early Treatment Diabetic Retinopathy Study (e-ETDRS) visual acuity letter score (VALS) at month 6, with a noninferiority margin of 5. After primary outcome assessment at month 6, participants originally assigned to aflibercept who met the protocol-defined criteria for a good response were rerandomized to either continuing aflibercept every 4 weeks (n=79) vs changing to a treat and extend regimen (n=80); 15 participants with a protocol-defined poor or marginal response at 6 months were assigned to receive a dexamethasone implant. Participants originally assigned to bevacizumab who met the protocol-defined criteria for a good response were rerandomized to either continuing bevacizumab every 4 weeks (n=67) vs changing to a treat and extend regimen (n=67); participants (n=39) with a protocol-defined poor or marginal response at 6 months were assigned to receive aflibercept.
The last visit as part of the SCORE2 protocol-defined treatment schedule was at month 12; after month 12, patients were treated at investigator discretion using any commercially available drug (including nonstudy drug or no drug) and followed at annual visits through month 60 as part of the SCORE2 Long-term Follow-up (SCORE2 LTF). The current report primarily focuses on the 248 SCORE2 participants who completed at least 1 visit in the SCORE2 LTF, of which 73 SCORE2 participants initially randomized to aflibercept at month 0 and 77 participants initially randomized to bevacizumab at month 0 completed the month 60 visit (defined as month 60 completers). A graphical description of the SCORE2 participant flow through month 60 is provided in Figure 1 .
Study data included all interventions administered to the study eye for treatment of macular edema secondary to CRVO or HRVO (including treatments given at nonstudy offices, provided they were documented in the medical record). At months 0, 6, 12, 24, 36, 48, and 60, data were collected on best-corrected e-ETDRS VALS, central subfield thickness (CST) assessed by spectral domain optical coherence tomography (SD-OCT), and eye examinations. The SD-OCT images were sent to the Reading Center at the University of Wisconsin–Madison for grading. At each annual in-person visit with the participant, and via telephone call with the site clinical coordinator at months 18, 30, 42, and 54, there was a medical record review of new ocular conditions, procedures, and other new health conditions occurring since the preceding annual visit. Data for these analyses were frozen on May 6, 2021.
In this article, we define complete case analysis as the analysis of all nonmissing data. We also define macular edema resolution to be CST <300 µm with no subretinal or intraretinal fluid, and no cystoid spaces within the ETDRS grid, based on reading center evaluation. We used fully conditional specification multiple-imputation to test the sensitivity of our analyses to missing data under the missing at random (MAR) assumption. Roughly speaking, MAR assumes that a missing value for a participant can be imputed from nonmissing values of other participants with similar nonmissing covariates.
To account for the longitudinal aspect of the data, we implemented the “just another variable” approach. Our imputation model incorporated month 0 to month 60 VALS and CST, together with covariates of disease type (HRVO vs CRVO), anti-VEGF treatment prior to randomization, treatment assignment, gender, race, and age. After imputation but before analysis, deaths were modeled by deleting imputed postmortem records, and VALS or CST changes from baseline were calculated if required.
Our analysis model was a longitudinal mixed one that predicted the outcome of interest (eg, VALS) as a function of baseline VALS, and for the CST outcome analysis, baseline CST, and treatment assignment, categorical month, and assignment by month interaction were included in all models. Temporal autocorrelation between months t and s was modeled as <SPAN role=presentation tabIndex=0 id=MathJax-Element-1-Frame class=MathJax style="POSITION: relative" data-mathml='ρ|t−s|’>𝜌|𝑡−𝑠|ρ|t−s|
ρ | t − s |
. MAR estimates for the outcome variables (eg, VALS, CST) at various months were obtained from the analysis model by contrasts. The Supplementary Material explores sensitivity of conclusions under missing not at random “shift” hypotheses that missing VALS values should be shifted from their MAR imputed values by an amount that may differ by treatment arm but not time.
SCORE2 enrolled 362 study eyes at 66 sites and randomized 180 study eyes to receive aflibercept and 182 eyes to receive bevacizumab. Of these, 330 participants were followed through month 12 and became eligible to be enrolled into the SCORE2 LTF. During the SCORE2 LTF through month 60, 36 (11%) of the 330 participants are known to have died ( Figure 1 ). Further, 236 (72%) of the 330 participants targeted for the SCORE2 LTF at 64 sites completed the month 24 visit, 209 (63%) participants completed the month 36 visit, 170 (52%) participants completed the month 48 visit, and 150 (45%) completed the month 60 visit. During the SCORE2 LTF, 248 (75%) participants completed at least 1 SCORE2 LTF annual visit.
Among the 150 month 60 completers (mean [SD] age = 68.0 [11.6] years, 44% [66 of 150] female and 11% [16 of 150] black), 73 (49%) had initially been randomized at month 0 to receive aflibercept and 77 (51%) had originally been randomized to receive bevacizumab as part of SCORE2 ( Table 1 ). The mean baseline VALS was 51.1 (Snellen equivalent, 20/100; SD=14.8) and CST based on SD-OCT was 671 μm (SD=232). Month 60 completers were similar to month 60 noncompleters with respect to baseline factors of VALS, SD-OCT CST, receipt of anti-VEGF treatment prior to randomization, months from diagnosis of macular edema to randomization, disease type (CRVO vs HRVO diagnosis), demographics, and medical history characteristics, but received on average 0.8 more anti-VEGF treatments between month 0 and up to month 12 (mean [SD] = 10.8 [1.5] vs 10.1 [2.6]; P = .001) than month 60 noncompleters.
|Month 60 Completer a (n=150)||Month 60 Non completer |
|P -value b|
|Study eye treatment arm||.74|
|Aflibercept||180 (50)||73 (49)||107 (50)|
|Bevacizumab||182 (50)||77 (51)||105 (50)|
|Age, years, mean (SD)||68.9 (12.0)||68.0 (11.6)||69.5 (12.2)||.23|
|Study eye received prior anti-VEGF treatment at randomization||121 (33)||49 (33)||72 (34)||.80|
|Months between diagnosis of macular edema and randomization in study eye, mean (SD)||6.6 (13.9)||7.1 (15.0)||6.2 (13.1)||.51|
|Female sex||157 (43)||66 (44)||91 (43)||.84|
|Hispanic or Latino||38 (10)||15 (10)||23 (11)||.80|
|Black race||54 (15)||16 (11)||38 (18)||.16|
|Hemiretinal vein occlusion in study eye||57 (16)||23 (15)||34 (16)||.87|
|Diabetes mellitus, type 2||113 (31)||45 (30)||68 (32)||.63|
|Hypertension||278 (77)||110 (73)||168 (79)||.19|
|Coronary heart disease||56 (15)||24 (16)||32 (15)||.81|
|Visual acuity letter score in study eye; mean (SD)||50.3 (15.2)||51.1 (14.8)||49.8 (15.5)||.41|
|SD-OCT central subfield thickness in study eye, μm, mean (SD)||665 (223.2)||671 (232)||660 (217)||.64|
|Number of anti-VEGF treatments between baseline and up to month 12 in study eye; mean (SD)||10.4 (2.3)||10.9 (1.5)||10.1 (2.7)||.001 c|
Treatment for macular edema:
Treatment for macular edema in SCORE2 was dictated by protocol for the initial 12 months, with most participants receiving monthly treatment in the study eye up to the 6-month visit. The mean (SD) number of anti-VEGF injections between month 0 up to month 12 was 10.65 (1.52) for participants originally randomized to aflibercept and 11.21 (1.19) for those originally randomized to bevacizumab ( Table 2 ). When the SCORE2 participants exited the clinical trial and were followed in the LTF, treatment for macular edema was at the discretion of the physician; the mean number of treatments combining the 2 treatment groups was 4.04 (3.38) (range 0-13) between months 12 and 24, 3.92 (3.31) (range 0-13) between months 24 and 36, 3.75 (3.50) (range 0-12) between months 36 and 48, and 3.41 (3.69) (range 0-19) between months 48 and 60.
|Initially Randomized to Aflibercept||Initially Randomized to Bevacizumab||Total|
|Treatments in Study Eye|
|Month 0-6 (under monthly treatment protocol)||175||5.94||6||0.32||4||7||173||5.92||6||0.27||5||6||348||5.93||6||0.30||4||7|
|Month 6-12 (under TAE or monthly treatment protocol)||167||4.69||5||1.48||1||7||168||5.29||6||1.13||1||7||335||4.99||6||1.35||1||7|
|No. and PercentageWith Any Treatment in Study Eye|
|No. Treated/Total||% Treated||No. Treated/Total||% Treated||No. Treated/Total||% Treated|
|Among Month 60 Completers: Number of Treatments in Study Eye From Month 12 to 60|
|Treatments in study eye||73||14.45||13||11.60||0||53||77||17.22||18||12.56||0||44||150||15.87||15||12.14||0||53|
|Any steroid treatment||9||12||15||19||24||16|
|Among Month 60 Completers: Number of Treatments in Study Eye From Month 48 to 60|
|Any steroid treatment||4||5.5||8||10||12||8.0|