Abstract
Purpose:
To evaluate long-term visual and anatomical outcomes of anti–vascular endothelial growth factor (VEGF) therapy for macular edema (ME) secondary to retinal vein occlusion (RVO) in routine clinical practice.
Methods:
Patients with ME secondary to hemi-RVO (HRVO), central RVO (CRVO), or branch RVO (BRVO) after initiating anti-VEGF therapy were followed for at least 36 months. Main outcomes were change in best visual acuity (BVA) and mean absolute change in central subfield thickness (CST) at 12, 24, 36, and 48 months.
Results:
Patients with BRVO showed significant increases in BVA that were maintained after 12, 24, 36, and 48 months (+11.03, +12.06, +10.71, and +9.26 letters, respectively; P < .05). CST significantly decreased after 12, 24, 36, and 48 months (−83.51, −67.93, −97.52, −127.85 µm, respectively; P < .05). In patients with CRVO/HRVO, significant improvements in BVA were seen at 12 and 24 months (+9.39 and +8.54 letters, respectively; P = .023). At 36 and 48 months, the visual gain was not significant (+2.64 and +3.42 letters, respectively; P > .05). For CST changes, there were significant decreases at 12, 24, and 36 months (−146.23, −149.54, and −166.44 µm, respectively; P < .05). At 48 months (−97.66 µm, P = .130), changes in CST were not significant.
Conclusions:
In routine clinical practice, visual and anatomical benefits of anti-VEGF agents in patients with BRVO were sustained at 36 and 48 months. For patients with CRVO/HRVO, anatomical improvements were maintained for 36, but not 48 months, while visual improvements were no longer maintained by 36 months.
Introduction
Retinal vein occlusion (RVO) is one of the most common retinal vascular diseases, with an estimated 16 million people in the United States, Asia, Europe, and Australia being affected and can be classified as central RVO (CRVO) or branch RVO (BRVO), based on the type of vessel occluded. 1 In patients with CRVO, blockage of the main outflow vessel of the retina results in variable amounts of hemorrhage, edema, and retinal nonperfusion. In BRVO, occlusion of one of the primary, secondary, or tertiary branches of the central retinal vein results in similar findings as those for CRVO but only in the portion of retina drained by the branch vein. The ischemic retina produces vascular endothelial growth factor (VEGF) in an attempt to restore perfusion, but this results in increased vascular permeability and angiogenesis that can lead to macular edema (ME) and retinal hemorrhage. 2
Intravitreal injections of anti-VEGF are the first-line therapy for ME secondary to RVO. These drugs have proven to be beneficial in the initial treatment of both BRVO and CRVO. 3 –8 The BRAVO and CRUISE studies evaluated ranibizumab (Lucentis; Genentech, Inc, South San Francisco, California) in treatment of ME due to BRVO or CRVO. 3,5 Patients were randomized 1:1:1 to receive either ranibizumab (0.3 or 0.5 mg) administered monthly or a monthly sham injection, and a 6-month observation period in which the ranibizumab treatment groups could continue to receive their assigned treatment pro re nata (prn). Patients were monitored monthly during the observation phase, and ranibizumab was administered based on prespecified treatment criteria. These studies showed a mean visual acuity (VA) improvement of 18.3 letters in BRAVO and 14.9 letters in CRUISE, at a ranibizumab dose of 0.5 mg at the 6-month primary end point. They also showed a mean decrease from baseline at 6 months in central subfield thickness (CST) of 345 and 452 μm, respectively. However, a significant limitation of both of these studies is that patients were followed for only a 12-month period, and a mandated period of injections and monthly follow-up visits does not always follow routine clinical practice.
Few studies have looked at the long-term outcomes of patients with RVO treated with anti-VEGF injections. 9 –11 The most substantial of these was the RETAIN study, an open-label, single-arm, multicenter extension trial of BRAVO and CRUISE studies, in which patients were followed for an additional 24 months. 9 Patients were seen every month in the first year and at least every 3 months in second year. The study demonstrated maintenance of visual and anatomical improvements achieved at the end of the BRAVO and CRUISE trials through 48 months in patients treated with ranibizumab. 9 Other studies with patients treated in a routine clinical practice regimen showed that ranibizumab or bevacizumab produces an initial improvement in VA that gradually lessens over time. 10,11 This could potentially be due to the differences between the fixed dosing regimen and/or close monitoring that occurs in clinical trials and the treat and extend or PRN schedule that tends to be used in clinical practice. There is also a lack of data in the long-term anatomic and clinical outcomes of other anti-VEGF agents. The aim of this study was to evaluate the long-term efficacy of anti-VEGF therapy for RVO in a routine clinical practice setting.
Methods
Study Design
The study was performed at Cole Eye Institute, Cleveland, Ohio, and received approval from the Cleveland Clinic Investigational Review Board. Because of the retrospective nature of the study, written informed consent was not required. All study-related procedures were performed in accordance with the Health Insurance Portability and Accountability Act.
Participants
Patients seen at the Cole Eye Institute between January 2011 and May 2016 were included if they met the following criteria: (1) a new diagnosis based on International Classification of Diseases, Ninth Revision codes of CRVO, hemi-RVO (HRVO), or BRVO (362.35, 362.36, 362.37), (2) follow-up visit at 36 months (±1 month) after their first anti-VEGF injection, (3) spectral-domain optical coherence tomography (SD-OCT) scan at the time of diagnosis and at follow-up, (4) presence of ME at the time of first injection, and (5) age 18 years or older.
Only 1 eye per patient was accepted for study. When both eyes were eligible, 1 eye was chosen randomly. Eyes were excluded if any of the following were present: confounding retinal or ocular disease (eg, proliferative diabetic retinopathy, macular degeneration, macular hole, advanced cataract classified by the treating physician as the significant reason for vision decline), history of recent major ocular surgery (eg, pars plana vitrectomy), and prior history of anti-VEGF or steroids injections in order to select a treatment-naive population.
Of the initial cohort of 235 patients/eyes identified, 52 were included in the study. Patients were treated with anti-VEGF injections at the discretion of the treating physician and were followed and retreated based on physician determination following a complete ophthalmologic examination and SD-OCT images.
Study End Points and Data Collection
The main outcomes assessed in this study were the mean change in VA and the mean change from baseline in CST as measured by SD-OCT at 12, 24, 36, and 48 months. A time window of ±1 month was allowed for all visits. Secondary outcomes included the total number of anti-VEGF and steroid injections per year during the period of follow-up and the predictive factors of visual and anatomic improvement at 36 months. Based on the stratifications from previous major randomized controlled trials, patients were stratified into groups by their diagnosis of BRVO or CRVO/HRVO. The VAs were recorded and converted from Snellen to approximate Early Treatment Diabetic Retinopathy Study (ETDRS) letter scores for statistical analyses using the formula: ETDRS letter score = 85 + 50 × log (Snellen fraction). 12 Visits were recorded at yearly intervals from initial injection.
Statistical Methods
Continuous measures were summarized with means and standard deviations, while categorical factors were described using frequencies and percentages. To estimate changes in ETDRS and CST over time, overall and by RVO group, linear mixed effect models were fit. Mean changes with 95% confidence intervals are provided. To evaluate predictors of change at 36 months in ETDRS and CST, 2-sample t tests and Pearson correlations were fit. Analyses were performed using SAS software (version 9.4; Cary, North Carolina).
Results
A total of 52 eyes (52 patients) were identified at baseline from the inclusion and exclusion criteria, of which 25 were diagnosed with CRVO/HRVO and 27 were diagnosed with BRVO. Fifty-one patients were examined at 12 months, 50 had a visit at 24 months, and all patients were examined at 36 months. At 48 months, only 20 patients remained in the study either due to timing of initial presentation or loss to follow up. Table 1 summarizes the overall and RVO group-specific baseline characteristics. The mean (SD) age of the whole cohort was 67 years (11.27), and 56% (n = 23) of patients were male. Table 2 discerns the percentage of anti-VEGF drugs (ranibizumab, bevacizumab, and aflibercept) employed at the first encounter and yearly through the follow-up. No patient was solely treated with aflibercept during the study.
Baseline Measures for the CRVO/HRVO and BRVO Groups.
Abbreviations: ME, macular edema; FA, fluorescein angiogram; OCT, optical coherence tomography; PRP, panretinal photocoagulation; CME, cystoid macular edema; RVO, retinal vein occlusion.
Percentage of Anti-VEGF Drug Employed.
Abbreviation: VEGF, vascular endothelial growth factor.
Patients With BRVO
Twenty-seven eyes were diagnosed with BRVO. The average age for the BRVO cohort was 64.9 ± 11.8 years. The mean (SD) ETDRS letter score at baseline was 59.01 (0.22), and mean (SD) baseline CST was 404.41 μm (156.53). Figure 1 shows the mean VA at baseline and at recorded follow-up visits through 48 months for all identified patients with BRVO. At 12, 24, 36, and 48 months of follow-up visits, the mean change in best visual acuity (BVA) from baseline was +11.03 (95% confidence interval [CI]: 5.16-16.90; P < .001), +12.06 (95% CI: 6.12-18.00; P < .001), +10.71 (95% CI: 4.91-16.51; P < .001), and +9.26 (95% CI: 1.34-17.17; P = .022), respectively.
Mean best visual acuity (BVA) in ETDRS letters of patients with BRVO plotted at 12-month intervals from initial diagnosis and treatment though 48 months. ETDRS indicates Early Treatment Diabetic Retinopathy Study; BRVO, branch retinal vein occlusion.
There was an improvement from baseline in CST at months 12, 24, 36, and 48 of −83.51 μm (95% CI: −131.83 to −35.18; P < .001), −67.93 μm (95% CI: −116.26 to −19.59; P = .006), −97.52 (95% CI: −144.75 to −50.29; P < .001), and −127.85 μm (95% CI: −191.98 to −63.71; P < .001), respectively (Figure 2). The mean (SD) number of anti-VEGF injections was 6.67 (2.20), 3.44 (2.93), 3.43 (2.85), and 4.55 (3.88) in the first, second, third, and fourth years of study, respectively (Table 3). Only 1 patient received 1 focal laser treatment during follow-up.
Mean central subfield thickness (CST) of patients with BRVO plotted at 12-month intervals from initial diagnosis and treatment through 48 months. BRVO indicates branch retinal vein occlusion.
Mean Number of Injections Over 4 Years for BRVO and CRVO/HRVO Groups.
Abbreviations: VEGF, vascular endothelial growth factor; BRVO, branch retinal vein occlusion; CRVO, central RVO; HRVO, hemi-RVO.
Patients With CRVO/HRVO
Twenty-five eyes were diagnosed with either CRVO or HRVO. The average age for the CRVO/HRVO cohort was 69.8 ± 10.3 years. The mean ETDRS letters score at baseline was 54.77 (SD), and the mean CST at baseline was 496.72 μm.
The mean change from baseline in BVA at 12, 24, 36, and 48 months of follow-up visits was +9.39 (95% CI: 2.06-16.73; P = .013), +8.54 (95% CI: 1.20-15.87; P = .023), +2.64 (95% CI: −4.70-9.97; P = .48), and −3.54 (95% CI: −14.06-6.98; P = .51), respectively (Figure 3).
Mean best visual acuity (BVA) in ETDRS letters of patients with CRVO/HRVO plotted at 12-month intervals from initial diagnosis and treatment though 48 months. ETDRS indicates Early Treatment Diabetic Retinopathy Study; CRVO, central retinal vein occlusion; HRVO, hemi-RVO.
There was an improvement in CST from baseline at months 12, 24, 36, and 48 of −146.23 μm (95% CI: −239.26 to −53.19; P = .002), −149.54 (95% CI: −240.36 to −58.71; P = .002), −166.44 (95% CI: −256.27 to −76.61; P < .001), and −97.66 (95% CI: −223.01 to 27.69; P = .13), respectively (Figure 4). The mean (SD) number of anti-VEGF injections was 6.96 (3.18), 4.00 (3.06), 3.80 (3.08), and 4.25 (3.49) in the first, second, third, and fourth years of study, respectively (Table 2).
Mean central subfield thickness (CST) of patients with CRVO/HRVO plotted at 12-month intervals from initial diagnosis and treatment through 48 months. CRVO indicates central retinal vein occlusion; HRVO, hemi-RVO.
Predictors of Visual and Anatomic Outcomes at 36 Months
Baseline predictive factors that influenced visual and anatomic outcomes at 36 months such as sex, diabetes, glaucoma, fluorescein angiogram (when present), age, CST, and ETDRS letter score were identified. For the BRVO cohort, baseline ETDRS letter score was the only significant predictor of visual outcome at month 36 (ρ = −0.45; 95% CI: −0.71 to −0.08; P = .019). The association was negative, indicating that as the baseline VA decreased, the improvement in VA at month 36 increased. In other words, the lower the initial ETDRS letter score, the greater the improvement that was seen at month 36. Both the baseline ETDRS letter score (ρ = 0.46; 95% CI: 0.09-0.71; P = .016) and baseline CST (ρ = −0.64; 95% CI: −0.82 to −0.34; P < .001) were predictive of anatomic outcomes at month 36. Greater reductions in CST at month 48 were seen in patients with worse vision and higher baseline CSTs.
In patients diagnosed with CRVO/HRVO, baseline ETDRS letters score (ρ = −0.63; 95% CI: −0.82 to −0.31; P < .001) and baseline CST (ρ = −0.74; 95% CI: −0.88, −0.48; P < .001) were the only significant predictive factors for ETDRS and CST at 36 months, respectively. Similar to findings in the patients with BRVO, this means that a lower initial ETDRS letter score is predictive of a greater improvement of BVA at 36 months and a lower initial CST is predictive of a lower CST at 36 months.
Discussion
While multiple studies have shown short-term visual and anatomic benefits in treatment of RVO with intravitreal injections, relatively fewer have explored this association with long-term outcomes. In this study, we retrospectively examined the outcomes of using intravitreal injections in a routine clinical setting. The analyses from this study allow us to assess if dosing regimens in routine clinical practice achieve the same benefits as randomized controlled trials. The only controlled trial to analyze long-term outcomes was the RETAIN study, which used patients from BRAVO/CRUISE, but excluded patients with a poor initial VA (>20/320 for CRUISE and >20/400 for BRAVO). 3,5 In clinical practice, these patients are often not ignored and are still treated in hopes of achieving optimal visual outcomes.
The current study demonstrates that both patients with BRVO and patients with CRVO/HRVO have had continued benefits of anti-VEGF treatment at 2 years. In patients with BRVO, visual and anatomical improvements were sustained up to 48 months from event, while in patients with CRVO/HRVO, anatomical benefits were seen up to 36 months. These findings are similar to those found at 36 months in both the RETAIN study and those found by Rezar et al, but with slightly less improvement in BVA and CST. 9,10 Rezar et al showed that patients with BRVO treated with ranibizumab demonstrated a continued BVA improvement over the 60-month period (+19 letters at month 12 and +28 letters at month 60). 10 Similarly, in the RETAIN study, patients with BRVO and CRVO demonstrated sustained improvements in vision 4 years after the BRAVO/CRUISE study baseline assessments (BRVO: +20.1 letters at month 48 vs +18.6 letters at month 6, P = .6; CRVO: +14 letters at month 48 vs +13.1 at month 12, P = .3). The differences in findings between the RETAIN trial compared to the present study for the CRVO group could be due to the structured visit schedule, the monotherapy with ranibizumab, and the mandatory scatter photocoagulation when patients required an anti-VEGF injection in 2 consecutive visits, present in the latter. Additionally, the small sample size at year 4 in the present study could be responsible for the nondetection of small treatment benefits in the CRVO group.
When treating patients in the clinical setting, it is important to understand factors that are predictive of long-term visual outcomes. In looking at the predictive factors of outcomes at 3 years, we found that there is a strong relationship with initial BVA and CST in both BRVO and CRVO/HRVO diagnoses. For both groups of patients, those with lower ETDRS letter score at baseline showed greater visual improvement at 3 years, whereas a higher CST on initial presentation was predictive of greater anatomic improvement at year 3. This corroborates the outcomes observed in studies that examined short-term predictive factors. Wai et al stratified patients based on initial VA and found that the greatest improvement in letter gains tended to occur in patients with a VA of 20/320 or worse. 13 Januschowski et al treated patients with bevacizumab and looked at multiple baseline patient characteristics in a diagnosis of CRVO and found that a low BVA and high CRT were indicative of greater visual improvement at weeks 24 to 48. 14 These findings may be due to the increased capacity for recovery in patients with initially low BVA and high CRT.
Limitations of this study include the retrospective nature and relatively small sample size of the individual cohorts. Because the study was retrospective, patients did not always have comprehensive complementary exams data, regular follow-up periods and some were lost to follow-up. The relatively small sample sizes of the separate groups may have also increased the probability of type II errors. There was also a possibility of selection bias with the sample: patients with successful outcomes might not have needed treatment, and patients with adverse outcomes might have discontinued treatment. Additionally, patients with conditions that could confuse the diagnosis were not included in the study. Furthermore, since the study was performed in a single tertiary care institution, patients may not be representative of most eye clinics.
Nevertheless, this study analyzes long-term treatment of RVO among a diverse population of physicians at an academic center and the results contribute to the understanding of prognosis and treatment of patients with RVO.
Footnotes
Authors’ Note
This study was presented at the Association for Research in Vision and Ophthalmology (ARVO) Annual Meeting, May 7 to 11, 2016, in Baltimore, Maryland. The authors received a waiver for publication in journals other than IOVS.
Declaration of Conflicting Interests
The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: SKS received personal fees from Regeneron, Bausch and Lomb, and grants and personal fees from Allergan, outside the submitted work; JPE received grants from Regeneron, grants and personal fees from Genentech, Alcon, Thrombogenics, and personal fees from Zeiss, Leica, Santen, Alimera, outside the submitted work; AR received personal fees from Allergan, outside the submitted work; PKK received personal fees from Allergan, Alcon, Bayer, Regeneron, Novartis, Ohr, outside the submitted work; AS is a part-time employee of the state of Ohio, and he received personal fees from Elsevier and American Academy of Ophthalmology; AB received personal fees from Allergan, MCME Global, Johnson & Johnson, outside the submitted work; RPS is a consultant for Regeneron, Genentech, Alcon, Shire, and Optos, and he received research funding from Zeiss, Novartis, Apellis, Roche, and Regeneron.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.