Changes in Retinal Microvasculature and Visual Acuity After Antivascular Endothelial Growth Factor Therapy in Retinal Vein Occlusion

Abstract

Purpose: To investigate the changes in the retinal microvasculature during the course of anti-VEGF therapy in eyes with macular edema due to retinal vein occlusion (RVO) and their association with visual outcomes.

Methods: The vessel density (VD) and foveal avascular zone (FAZ) area in the superficial capillary plexus (SCP) and deep capillary plexus (DCP) were quantitatively measured by optical coherence tomography angiography (OCTA) in 48 consecutive eyes with RVO before and 1, 3, 6, 9, and 12 months after anti-VEGF therapy. Anti-VEGF therapy was performed either with ranibizumab or aflibercept following a pro re nata (PRN) regimen. The correlation between post-treatment best-corrected visual acuity (BCVA) and changes in the retinal microvasculature evaluated by OCTA were assessed.

Results: The BCVA improved significantly at 12 months (P < 0.001). Better BCVA at 12 months was significantly associated with a better VD in the SCP and DCP both at baseline (R² = 0.524, P < 0.001 and R² = 0.457, P < 0.001, respectively) and at 12 months (R² = 0.521, P < 0.001 and R² = 0.662, P < 0.001, respectively). Overall, both VD and FAZ did not change significantly during the 12 months. However, the progression of nonperfusion was observed in the SCP in 6 (13%) eyes and in the DCP in 10 (21%) eyes. The number of macular edema recurrence was significantly associated with a decrease in the VD (P = 0.006 [SCP] and P < 0.001 [DCP]) and less visual gain (P = 0.02) after treatment.

Conclusions: Anti-VEGF therapy maintains retinal perfusion in most patients with RVO. Preserving retinal perfusion is crucial for better visual outcomes.

Retinal vein occlusion (RVO) is a relatively common retinal vascular disorder in middle-aged and elderly people with an estimated prevalence of over 16 million worldwide.¹ Predisposing factors include diabetes mellitus, hypertension, smoking, and open-angle glaucoma.^2–5 Patients with RVO develop retinal hemorrhage, vascular tortuosity, retinal ischemia, and macular edema due to obstruction of the retinal veins that drain blood from the retina.^6,7 Macular edema induced by increased production of intraocular VEGF is a predominant cause of vision loss in patients with RVO.⁸ The anti-VEGF therapy with bevacizumab, ranibizumab, and aflibercept has been reported to be effective in reducing macular edema.^9–15 However, the visual outcomes after anti-VEGF therapy for macular edema varies significantly among patients ranging from complete visual recovery to less improvement in the visual acuity. Recently, the vascular perfusion status has been shown to be a possible reason for the differences in the visual recovery after resolution of the macular edema. The visual improvement after treatment of macular edema significantly correlates with better retinal perfusion and less retinal ischemia in patients with RVO.^16–21 Therefore, the effects of repeated anti-VEGF therapy to the retinal perfusion should be investigated for a better long-term management of RVO. 

Optical coherence tomography angiography (OCTA) enables noninvasive visualization of different retinal layers, such as the superficial and deep retinal capillaries by segmentation of each layer.^22–25 In addition, the OCTA has the potential to quantitatively analyze the retinal vessel density (VD) and the foveal avascular zone (FAZ) area.^26,27 Thus, OCTA has provided a better understanding of the relationship between the layer-specific retinal microvascular changes and the visual outcomes in patients with RVO. To date, several studies have reported the macular VD and FAZ before and after anti-VEGF therapy.^28–30 However, the results are conflicting because of the short follow-up periods and a limited number of patients studied. In addition, the possible associations between changes in the retinal VD and visual outcomes have not been reported. 

The purpose of this study was to quantitatively evaluate the changes in the retinal VD and FAZ using OCTA before and after anti-VEGF therapy in patients with RVO, and assess their relationship with the 12-month visual outcomes in a consecutive series of patients. 

Methods

The institutional review board of Osaka University Graduate School of Medicine approved this retrospective, consecutive, interventional case series (10039). The procedures used in this study conformed to the tenets of the Declaration of Helsinki. All patients were informed of the nature and possible consequences of the procedures, and signed informed consent was obtained from all patients. Patients who were treated with anti-VEGF therapy (ranibizumab or aflibercept) for the macular edema associated with RVO between October 2014 and March 2016 and followed monthly for at least 12 months were enrolled. All patients were diagnosed with branched retinal vein occlusion (BRVO) or central retinal vein occlusion (CRVO) with retinal hemorrhage and macular edema extending to the fovea confirmed by spectral-domain (SD) OCT. Patients were excluded if they had a pre-existing macular disease, media opacity, or diabetes mellitus that could interfere with the OCTA examinations and imaging interpretations. All patients underwent comprehensive ophthalmic evaluation, including measurement of the best-corrected visual acuity (BCVA), binocular indirect ophthalmoscopy, contact lens slit-lamp biomicroscopy, fundus photography, SD-OCT, and OCTA at monthly intervals before and after treatment. 

Retinal Microvasculature Imaging by OCTA

AngioVue system (Optovue RTVue XR Avanti; Optovue, Inc., Fremont, CA, USA) was used for all OCTA images before and monthly after anti-VEGF therapy. Each 3 × 3-mm image acquired had the fovea at the center. The device software automatically delineated the superficial capillary plexus (SCP) from the deep capillary plexus (DCP) in addition to the outer retina and the choroid. The AngioVue software determined the SCP to be the layer extending from 3 mm below the internal limiting membrane to 15 μm below the inner plexiform layer (IPL), whereas the DCP extended from 15 to 70 μm below the IPL. Based on the acquired images, VDs in the SCP and DCP were automatically calculated by AngioAnalytics software as the percentage area occupied by blood vessels within a 3 × 3-mm area, with the blood vessels being defined as pixels above the threshold level. The FAZ area in the SCP and DCP were automatically calculated using the nonflow function in the software. OCTA images with poor quality (a signal strength index ≤ 50), motion artifacts, or incorrect autosegmentation were excluded from the data analysis. 

Anti-VEGF Treatment

All patients were examined at the Osaka University hospital and were given intravitreal anti-VEGF injections (either 0.5 mg ranibizumab or aflibercept) for 12 months. The patients received 1 to 3 initial monthly intravitreal anti-VEGF injections followed by a pro re nata (PRN) regimen with monthly monitoring. Retreatment with ranibizumab or aflibercept was considered if a patient showed recurrence of macular edema on SD-OCT imaging that exceeded 300 μm of central retinal thickness (CRT). 

Spectral-Domain Optical Coherence Tomography

SD-OCT images were acquired using a Cirrus SD-OCT (Carl Zeiss Meditec, Inc., Dublin, CA, USA). CRT was measured as the average retinal thickness in the central area within a diameter of 1 mm. 

Data Collection and Statistical Analyses

The collected data included a complete medical and ophthalmic history, BCVA, CRT, and retinal microvasculature (VD and FAZ) evaluated by OCTA before treatment and at 1, 3, 6, 9, and 12 months after treatment. The main outcome measures were post-treatment BCVA at 12 months and changes in the retinal microvasculature evaluated by OCTA. 

For statistical analysis, the BCVA was measured using the Landolt C acuity chart and analyzed on a logMAR scale. 

One-way ANOVA was used when quantitative parameters, such as BCVA, VD, and FAZ area, were compared among baseline, 1, 3, 6, 9, and 12 months. If the parameter was not normally distributed, the nonparametric Kruskal-Wallis one-way ANOVA on Ranks with pairwise multiple comparison by Dunn's test was applied. Patients were divided into 3 subgroups, depending on the changes in the VD before and 12 months after anti-VEGF therapy, which included “increased VD” defined as increase in the VD by 10% or more at 12 months from baseline, “decreased VD” defined as decrease in the VD by 10% or more at 12 months from baseline, and “unchanged VD” defined as changes in the VD within 10% at 12 months from baseline. Univariate and multivariate regression analyses were also performed to investigate the associations between logMAR BCVA at 12 months and OCTA parameters, such as vascular area and FAZ area. All analyses were conducted using SigmaStat software version 3.1 (SPSS, Inc., Chicago, IL, USA) and JMP Pro Software (SAS, Inc., Cary, NC, USA). P < 0.05 indicated statistical significance. 

Results

Sixty-one eyes of 61 consecutive patients who were treated with anti-VEGF therapy for macular edema associated with RVO and followed for at least 12 months were enrolled initially. Thirteen eyes were excluded because of diabetic retinopathy (1 eye), epiretinal membrane (1 eye), significant autosegmentation error (4 eyes), motion artifact (6 eyes), and low image quality potentially due to progressed cataract (1 eye). Therefore, 48 eyes of 48 patients met the study criteria for subsequent data analysis. The baseline characteristics of the 48 patients are summarized in Table 1. The mean patient age was 69.9 ± 9.4 (range, 50–86) years. Thirty-eight eyes had BRVO while 10 had CRVO. The CRT before treatment was 538 ± 192 μm (range, 324–1019 μm). The mean number of anti-VEGF injections during the 12-month period was 3.7 ± 1.4 (range, 1–7). 

Table 1

 

Patient Characteristics

 

Visual Outcomes

The changes in logMAR BCVA are shown in Table 2. The mean logMAR BCVA values were 0.36 ± 0.34 (pretreatment), 0.23 ± 0.30 (1 month), 0.18 ± 0.28 (3 months), 0.15 ± 0.30 (6 months), 0.14 ± 0.29 (9 months), and 0.13 ± 0.29 (12 months), respectively, indicating that the BCVA continuously improved during the treatment period (P < 0.001). Similarly, the CRT significantly reduced from 538 ± 192 (before treatment) to 250 ± 36 (1 month), 248 ± 38 (3 months), 249 ± 50 (6 months), 252 ± 51 (9 months), and 248 ± 33 μm (12 months) after treatment (P < 0.001), respectively. 

Table 2

 

Visual and Optical Coherence Tomography Angiography Outcomes

 

Retinal Microvascular Changes After Anti-VEGF therapy

At baseline before treatment, the area of macular edema corresponded to the capillary nonperfusion areas in both the SCP and DCP, indicating coexistence of intraretinal fluid and nonperfusion areas (Figs. 1–4). After anti-VEGF therapy, the nonperfusion areas remained in most cases at the areas of resolved cystoid spaces. 

Figure 1

 

Changes in the retinal microvasculature in the SCP and DCP in a 61-year-old male with BRVO. The patient had slight increase in the vessel density after anti-VEGF therapy. Fundus photograph (A), horizontal SD-OCT (B), and en face image (C) segmented at the SCP at presentation before treatment. The area of macular edema was delineated in red line. (D–I) OCTA images in the SCP and corresponding structural OCT image (J–O) at baseline, 1, 3, 6, 9, and 12 months after initial anti-VEGF therapy. The contours of the macular edema were overlaid in (D, E). VD in the SCP was 45.03% (baseline), 45.59% (1 month), 47.16% (3 months), 53.50% (6 months), 52.26% (9 months), and 54.20% (12 months). FAZ in the SCP was 0.224 (baseline), 0.358 (1 month), 0.314 (3 months), 0.323 (6 months), 0.334 (9 months), and 0.289 mm² (12 months). (P) En face image segmented at the DCP at presentation before treatment. (Q–V) OCTA images in the DCP and corresponding structural OCT image (W–AB) at baseline, 1, 3, 6, 9, and 12 months after initial anti-VEGF therapy. The contours of the macular edema were overlaid in (Q, R). The area of macular edema corresponds to nonperfusion area (Q). At 1 month, the most area of resolved edema remain nonperfused but some retinal capillaries reappeared (yellow arrows). VD in the DCP was 45.05% (baseline), 49.03% (1 month), 49.55% (3 months), 55.40% (6 months), 54.58% (9 months), and 53.55% (12 months). The FAZ in the DCP was 0.422 (baseline), 0.418 (1 month), 0.369 (3 months), 0.355 (6 months), 0.437 (9 months), and 0.382 mm² (12 months). The patient had two times of recurrence of macular edema and three anti-VEGF injections. Decimal visual acuity improved from 0.9 (baseline) to 0.9 (1 month), 1.0 (3 months), 1.2 (6 months), 1.5 (9 months), and 1.5 (12 months) after treatment.

Figure 2

 

Changes in the retinal microvasculature in the SCP and DCP in a 54-year-old male with BRVO. The patient had capillary loss superior to the fovea but without enlargement of the nonperfusion during anti-VEGF therapy. Fundus photograph (A), horizontal SD-OCT (B), and en face image (C) segmented at the SCP at presentation before treatment. The area of macular edema was delineated in red line (D–H, J). OCTA images in the SCP and corresponding structural OCT image (K–P) at baseline, 1, 3, 6, 9, and 12 months after initial anti-VEGF therapy. The contours of the macular edema were overlaid in (D). The area of the macular edema recurrence was delineated in the orange line in (I). The dashed red line corresponds to pre-existing macular edema before treatment in (C), indicating that the area of macular edema recurrence was within the area of pre-existing macular edema. VD in the SCP was 43.53% (baseline), 43.10% (1 month), 42.32% (3 months), 43.88% (6 months), 44.84% (9 months), and 43.69% (12 months). The FAZ in the SCP was 0.163 (baseline), 0.348 (1 month), 0.342 (3 months), 0.348 (6 months), 0.350 (9 months), and 0.349 mm² (12 months). VD was maintained and nonperfusion area superior to the fovea did not show enlargement. (Q) En face image segmented at the DCP at presentation before treatment with the red line delineating the area of macular edema (R–V, X). OCTA images in the DCP and corresponding structural OCT image (Y–AD) at baseline, 1, 3, 6, 9, and 12 months after initial anti-VEGF therapy. The contours of the macular edema were overlaid in (R). The area of macular edema corresponds to nonperfusion area. (S) At 1 month, most area of resolved edema remains nonperfused. VD in the DCP was 43.79% (baseline), 44.21% (1 month), 44.47% (3 months), 48.23% (6 months), 46.50% (9 months), and 46.15% (12 months). The FAZ in the DCP was 0.649 (baseline), 0.858 (1 month), 0.393 (3 months), 0.355 (6 months), 0.343 (9 months), and 0.647 (12 months). The patient had two times of recurrence of macular edema and a total of four injections during 12 months. Decimal visual acuity improved from 0.9 (baseline) to 1.2 (1 month), 1.2 (3 months), 1.2 (6 months), 1.2 (9 months), and 1.2 (12 months) after treatment.

Figure 3

 

Changes in the retinal microvasculature in the SCP and DCP in a 69-year-old female with BRVO. The patient had enlargement of the nonperfusion during anti-VEGF therapy. Fundus photograph (A), horizontal SD-OCT (B), and en face image (C) segmented at the SCP at presentation before treatment. The area of macular edema was delineated in the red line (D–F, H–J). OCTA images in the SCP and corresponding structural OCT image (K–P) at baseline, 1, 3, 6, 9, and 12 months after initial anti-VEGF therapy. The contours of the macular edema were overlaid in (D). The area of the macular edema recurrence was delineated in the orange line in (G). The dashed red line corresponds to pre-existing macular edema before treatment, indicating that the macular edema recurred beyond the area of pre-existing macular edema. VD in the SCP was 50.43% (baseline), 47.48% (1 month), 45.86% (3 months), 47.43% (6 months), 47.47% (9 months), and 47.00% (12 months). FAZ in the SCP was 0.502 (baseline), 0.227 (1 month), 0.245 (3 months), 0.238 (6 months), 0.226 (9 months), and 0.218 (12 months). Although the FAZ did not show enlargement, the VD slightly decreased because of the nonperfusion area associated with recurrent macular edema (yellow arrows). (Q) En face image segmented at the DCP at presentation before treatment (R–T, V–X). OCTA images in the DCP and corresponding structural OCT image (Y–AD) at baseline, 1, 3, 6, 9, and 12 months after initial anti-VEGF therapy. VD in the DCP was 52.66% (baseline), 53.36% (1 month), 50.96% (3 months), 51.98% (6 months), 52.04% (9 months), and 50.60% (12 months). The FAZ in the DCP was 0.460 (baseline), 0.254 (1 month), 0.252 (3 months), 0.456 (6 months), 0.447 (9 months), and 0.232 (12 months). VD slightly decreased at 12 months because of the enlargement of nonperfusion area associated with recurrent macular edema (yellow arrow). The patient had four times of recurrence of macular edema and a total of five injections during 12 months. Decimal visual acuity was 1.0 (baseline) to 1.0 (1 month), 0.9 (3 months), 1.2 (6 months), 0.9 (9 months), and 0.9 (12 months) after treatment.

Figure 4

 

Changes in the retinal microvasculature in the SCP and DCP in a 58-year-old male with CRVO. The patient had progressive nonperfusion during anti-VEGF therapy. Fundus photograph (A), horizontal SD-OCT (B), and en face image (C) segmented at the SCP at presentation before treatment. The area of macular edema was delineated in the red line (D–G, I, J). OCTA images in the SCP and corresponding structural OCT image (K–P) at baseline, 1, 3, 6, 9, and 12 months after initial anti-VEGF therapy. The contours of the macular edema were overlaid in (D). The area of the macular edema recurrence at 9 months was delineated in the orange line in (H). The dashed red line corresponds to pre-existing macular edema before treatment, indicating that the macular edema recurred beyond the area of pre-existing macular edema. VD in the SCP was 41.64% (baseline), 40.21% (1 month), 37.45% (3 months), 36.43% (6 months), 38.38% (9 months), and 36.52% (12 months). FAZ in the SCP was 0.497 (baseline), 0.658 (1 month), 0.933 (3 months), 0.924 (6 months), 1.004 (9 months), and 1.024 mm² (12 months). VD decreased and the FAZ enlarged with recurrence of the macular edema (yellow arrows). (Q) En face image segmented at the DCP at presentation before treatment (R–U, W, X). OCTA images in the DCP and corresponding structural OCT image (Y–AD) at baseline, 1, 3, 6, 9, and 12 months after initial anti-VEGF therapy. The contours of the macular edema were overlaid in (R). The large area of the macular edema recurrence at 9 months was delineated in the orange line in (V). Similar to the findings in the SCP, macular edema recurred beyond the area of pre-existing macular edema indicated in the dashed red line. VD in the DCP was 47.98% (baseline), 47.31% (1 month), 41.45% (3 months), 39.44% (6 months), 39.07% (9 months), and 40.69% (12 months). FAZ in the DCP was 0.870 (baseline), 1.419 (1 month), 1.579 (3 months), 1.626 (6 months), 1.947 (9 months), and 1.788 mm² (12 months). VD decreased and the FAZ enlarged with recurrence of the macular edema (yellow arrows). The patient had four times of recurrence of macular edema and a total of seven injections during 12 months. Decimal visual acuity was 0.4 (baseline) to 0.3 (1 month), 0.3 (3 months), 0.8 (6 months), 0.5 (9 months), and 0.3 (12 months) after treatment.

Table 2 shows the VD and FAZ within a 3 × 3-mm area before and after anti-VEGF therapy. The overall mean VD and FAZ did not change significantly in both the SCP and DCP during the 12 months. Pretreatment and 12-month values for VD in SCP were 44.9 ± 4.5% (range, 29.1%–55.9%) and 45.1 ± 5.5% (range, 28.2%–54.2%), respectively (P = 0.940). Pretreatment and 12-month values for VD in DCP were 46.9 ± 5.7% (range, 28.5%–58.3%) and 47.2 ± 5.9 % (range, 29.9%–57.0%), respectively (P = 0.997). Pretreatment and 12-month values for FAZ area in SCP were 0.48 ± 0.42 mm² (range, 0.16–2.48 mm²) and 0.56 ± 0.47 mm² (range, 0.15–2.59 mm²), respectively (P = 0.776). Pretreatment and 12-month values for FAZ area in DCP were 0.74 ± 0.49 mm² (range, 0.23–2.81 mm²) and 0.78 ± 0.7 mm² (range, 0.23–4.45 mm²), respectively (P = 0.997). 

Association Between the 12-Month VA and Retinal Microvasculature

Based on a univariate regression analysis, better BCVA at 12 months was significantly associated with younger age (P = 0.016), better pretreatment logMAR BCVA (P < 0.001), BRVO but not CRVO (P < 0.001), better pretreatment VD in the SCP and DCP (P < 0.001), better post-treatment VD at 12 months in the SCP and DCP (P < 0.001), and a smaller FAZ area post treatment at 12 months in the SCP and DCP (P = 0.020 and P = 0.021, respectively; Table 3). Among these factors, a multivariate regression analysis showed that younger age (P = 0.031), better pretreatment logMAR BCVA (P < 0.001), BRVO but not CRVO (P = 0.016), and better VD post treatment at 12 months in the DCP (P = 0.014) were significantly associated with BCVA at 12 months. 

Table 3

 

Univariate and Multivariate Regression Analyses of the Association Between Posttreatment logMAR BCVA at 12 Months and Variables

 

Factors Associated With Visual Gains at 12 Months

The mean visual gain was +0.23 ± 0.21 at 12 months when compared with the pretreatment readings. A lower rate of recurrence of macular edema (coefficient −0.063, R² = 0.115, P = 0.020) and a worse pretreatment BCVA (coefficient 0.310, R² = 0.257, P < 0.001) were significantly associated with visual gain at 12 months. The number of anti-VEGF injections, age, sex, disease type, VD, and FAZ area was not significantly associated with visual gain at 12 months. 

Factors Associated With Changes in Vessel Density and FAZ

Compared with the baseline, the VD in the SCP increased by more than 10% in seven eyes (15%; Fig. 1), remained unchanged in 35 eyes (73; Fig. 2), and decreased by more than 10% in six eyes (13%) at 12 months (Table 4). The VD in the DCP increased by more than 10% in 10 eyes (21%; Fig. 1), remained unchanged in 28 eyes (58%; Fig. 2), and decreased by more than 10% in 10 eyes (21%; Fig. 4) at 12 months (Table 5). When we divided the patients into three groups based on the change in the VD, eyes with a decrease in perfusion in the SCP and DCP had a greater number of recurrences of macular edema (P = 0.029 and P = 0.004, respectively) and tended to have less visual gain at 12 months, although the differences were not statistically significant. 

Table 4

 

Characteristics of the Eyes With Increased, Unchanged, or Decreased Vessel Density in the SCP at 12 Months From Baseline

 

Table 5

 

Characteristics of the Eyes With Increased, Unchanged, or Decreased Vessel Density in the DCP at 12 Months From Baseline

 

The mean changes in the VD at 12 months after anti-VEGF therapy were +0.14 ± 4.15% (range, −13.59 to +11.96%) in the SCP and +0.33 ± 4.49% (range, −12.06 to +9.46%) in the DCP. The mean changes in the FAZ area at 12 months after anti-VEGF therapy were +0.08 ± 0.26 mm² (range, −0.28 to 1.62 mm²) in the SCP and +0.04 ± 0.63 mm² (range, −0.74 to 3.95 mm²). Univariate linear regression analysis also showed that VD changes in the SCP and DCP were significantly associated with the number of recurrences of macular edema during the 12 months (coefficient −1.494, R² = 0.158, P = 0.006 in the SCP and coefficient −1.895, R² = 0.217, P < 0.001 in the DCP, respectively). In other words, eyes with a greater number of recurrences of macular edema showed a significant decrease in the VD over the 12 months follow-up period. The recurrence of macular edema especially beyond the pre-existing nonperfusion area seemed to be associated with both an enlargement of the nonperfusion area and a decrease in VD (Figs. 3–5). 

Figure 5

 

Schematic representation of the changes in the retinal microvasculature before and after anti-VEGF therapy. (A) Macular edema and associated capillary nonperfusion in retinal vein occlusion. (B) After anti-VEGF therapy, macular edema disappeared. The nonperfusion areas remained at the areas of resolved macular edema. (C) Recurrence of macular edema within the area of pre-existing macular edema and nonperfusion area. (D) After anti-VEGF therapy, macular edema resolved without enlargement of nonperfusion area. (E) Recurrence of macular edema beyond the pre-existing macular edema and nonperfusion area. (F) After anti-VEGF therapy, macular edema resolved but with further enlargement of the nonperfusion area. Visual gain tended to be poor even after resolution of the macular edema.

The enlargement of the FAZ in the SCP and DCP was not associated with the number of recurrences of macular edema (P = 0.413 in the SCP and P = 0.105 in the DCP, respectively), but was significantly associated with the disease type, particularly CRVO. That is, eyes with CRVO more frequently showed enlargement of FAZ compared with those with BRVO. 

Discussion

There have been conflicting theories regarding the influence of anti-VEGF therapy on retinal perfusion. While some reports demonstrate adverse vasoconstrictive effects on retinal perfusion that include a reduction in retinal vessel diameters and flow velocities,^31–34 there are others that demonstrate positive effects of anti-VEGF therapy on retinal perfusion, such as prevention of retinal nonperfusion.^35,36 In the current study, quantitative OCTA imaging identified that both VD and FAZ within the 3 × 3-mm area did not differ significantly during the course of anti-VEGF therapy for 12 months in patients with RVO. In addition, number of the anti-VEGF injections was not significantly associated with changes in the VD and FAZ area. Our results indicate that repeated anti-VEGF therapy with a mean of 3.7 injections during 12 months may not deteriorate or improve blood flow at least to the macular area in most eyes with RVO. Therefore, perfused retinal capillaries can be mostly maintained for at least 12 months with anti-VEGF therapy as indicated by the previous reports.^29,30 Alternatively, the once-occluded capillaries may not achieve complete reperfusion even after successful treatment of the macular edema. 

A better VD in both the SCP and DCP prior to the anti-VEGF treatment was found to associate significantly with a better pretreatment BCVA. Visual acuity may be preserved in cases with mild venous occlusion where the retinal capillary damage is not prominent. In contrast, a lower VD before treatment was found to associate with a worse BCVA, indicating that retinal capillary ischemia may cause acute damage to the retinal tissue, resulting in decreased VA. Therefore, the VD at the initial presentation may reflect the severity of venous occlusion and visual impairment in patients with RVO. The VD in the SCP and DCP before treatment also correlated with BCVA at 12 months after treatment. Thus, the initial VD not only correlates with the initial BCVA but also may be useful for predicting the visual outcomes following anti-VEGF therapy in patients with RVO. 

In our series, the anti-VEGF therapy significantly improved BCVA and reduced CRT during the 12-month follow-up as expected. The visual outcomes 12 months after anti-VEGF therapy were associated with a better VD in the SCP and DCP, and a smaller FAZ in the SCP and DCP. Among these OCTA parameters, VD in the DCP was the most important parameter that significantly correlated with BCVA. It has been reported that nonperfusion, microaneurysms, and enlargement of the nonperfusion area occur more frequently in the DCP compared with the SCP.^37,38 The DCP seems to be particularly important for visual acuity after anti-VEGF therapy. Our results are consistent with previous studies that addressed the favorable effects of retinal perfusion in the DCP on visual acuity in patients with RVO.¹⁸ Therefore, strategies that preserve retinal perfusion especially in the DCP should be explored to improve visual outcomes after RVO. 

Although the overall mean VD and FAZ did not differ significantly among 48 patients during the course of anti-VEGF therapy for 12 months, substantial increase or decrease in the VD was observed in some patients. Among the 48 eyes, 7 (15%) showed an increased VD in the SCP, and 10 (21%) showed it in the DCP from the time of presentation to 12 months of therapy. The eyes with increased VD tended to have a better visual gain at 12 months compared with those with a decrease in the VD. Although we did not find any case of complete reperfusion after anti-VEGF therapy within the area of coexisting edema and nonperfusion, the retinal perfusion may be improved with resolution of the edema in some cases. Suzuki et al.²⁸ have reported decreased nonperfusion after anti-VEGF therapy in patients with RVO. de Carlo et al.³⁹ also have reported the reappearance of retinal microvasculature after anti-VEGF treatment in the areas previously shown to have flow voids due to intraretinal fluid in diabetic macular edema (DME). They speculated that either the vessels that were poorly perfused prior to treatment, were re-perfused after an anti-VEGF treatment or the retinal vessels displaced or masked by the cystoid spaces reappeared after fluid resolution. In contrast, Spaide et al.⁴⁰ showed that anti-VEGF therapy results in resolution of edema without changes in the nonperfusion area. Mané et al.⁴¹ have also indicated that anti-VEGF–associated reperfusion is unlikely to occur but the capillaries already present above the cystoid spaces became visible in eyes with resolved DME. Based on the current study, we speculate that in eyes with macular edema with well-established areas of capillary dropout before treatment, the retinal capillaries may not return even after resolution of the edema. In such cases, the VD may not increase after anti-VEGF therapy. However, in eyes with acute macular edema with possible residual flow below the threshold, the retinal capillaries may have reappeared after resolution of the edema or hemorrhage as shown in Figure 1. We did not find any evidence of a significant capillary regeneration where there were no retinal capillaries before treatment. Further studies are needed to investigate whether early anti-VEGF injections during the reversible phase of retinal nonperfusion may prevent the progression of ischemia in RVO. 

The progression of the ischemia identified as enlargement of the capillary nonperfusion area is a significant cause of irreversible vision loss.³⁵ In the current study, the progression of nonperfusion was observed in the SCP in 6 (13%) eyes and in the DCP in 10 (21%) eyes during the 12-month treatment period. The number of recurrences of macular edema during the follow-up was identified as a significant cause of progressive nonperfusion. In particular, when the macular edema recurred beyond the pre-existing nonperfusion area, the nonperfusion area enlarged at the area of the recurrent macular edema as shown in Figures 3 and 4. In addition, accumulation of the intraretinal fluid associated with recurrence of macular edema may compress the remaining retinal capillaries, resulting in further progression of the capillary dropout. In such cases, visual gain tended to be poor even after resolution of the macular edema. In contrast, the nonperfusion area seemed to unlikely extend if the recurrence of the edema was limited within the area of pre-existing macular edema and nonperfusion as shown in Figure 2. Therefore, strategies to prevent re-occlusion and re-accumulation of the fluid beyond the baseline are important to prevent progression of capillary ischemia and hence preserve vision in eyes with RVO (Fig. 5). 

High intraocular levels of VEGF have been shown to be responsible not only for the development of macular edema but also the progression of ischemia in RVO.^35,42 The high levels of VEGF increase the expression of intercellular adhesion molecule-1 (ICAM-1) in the retinal endothelial cells and promote ICAM-1–mediated binding of leukocytes to the vasculature and their entrapment in the retinal microcirculation.^43–46 Progressive leukocyte entrapment in retinal capillaries contributes to subsequent downstream nonperfusion. Theoretically, anti-VEGF therapy reduces VEGF-induced leukocyte adhesion and leukocyte-mediated resistance of the retinal blood flow. Thus, long-term administration of intravitreal anti-VEGF injections has been shown to improve or at least maintain retinal perfusion in patients with RVO and diabetic retinopathy.^36,47 Because anti-VEGF injections were based on a PRN regimen in the current study, some eyes with a decrease in retinal perfusion during the 12-month period may have had less than sufficient suppression of VEGF to prevent leukostasis and subsequent nonperfusion. However, the progression of capillary ischemia may primarily depend on the degree of venous occlusion and may not be completely prevented by repeated anti-VEGF treatment. Further studies are needed to determine whether aggressive anti-VEGF treatments, such as continuous monthly injections can prevent the worsening of ischemia and hence preserve long-term vision in patients with RVO. 

The limitations of the current study include its retrospective design, different anti-VEGF agents used for treating macular edema, the small number of patients with CRVO, and the limited area (3 × 3 mm) analyzed for VD and FAZ. In addition, blockage of the signal from superficial retinal hemorrhage, motion artifacts from poor fixation, projection artifacts, and inability to completely eliminate segmentation errors may have influenced some of the automatic measurements of VD. Furthermore, we could only evaluate the SCP and the DCP, but not intermediate capillary plexus. Therefore, further studies using different devices and algorithms are necessary to validate our results and to advance our understanding of microvascular changes associated with RVO. Nevertheless, this study revealed that a PRN regimen of anti-VEGF therapy maintains vascular perfusion for at least 12 months in most patients with RVO. We also found that a higher number of recurrences of macular edema significantly decrease the retinal perfusion and tend to prevent visual gain after anti-VEGF therapy. These findings may be valuable for maintaining capillary perfusion and improving the long-term management of RVO. 

In conclusion, our study identified that the mean VD and FAZ area was maintained for at least 12 months of anti-VEGF therapy in most patients with RVO. A higher VD and a smaller FAZ were associated with better BCVA at 12 months. The number of recurrences of macular edema was also found to be associated with a decrease in VD and poor visual gain after treatment. Further studies are expected to investigate strategies to maintain or increase the retinal perfusion for better visual prognosis in patients with RVO. 

Acknowledgments

Disclosure: A. Winegarner, None; T. Wakabayashi, None; Y. Fukushima, None; T. Sato, None; C. Hara-Ueno, None; C. Busch, None; I. Nishiyama, None; N. Shiraki, None; K. Sayanagi, None; Ke. Nishida, None; H. Sakaguchi, None; Ko. Nishida, None