En face OCTA vessel density maps at the level of the superficial vascular complex (A), deep vascular complex (B), and choriocapillaris (C) are shown.
En face OCTA vessel density maps at the level of the superficial vascular complex (A), deep vascular complex (B), and choriocapillaris (C) are shown. During 1 year, there was no change in central macular vessel density in the eyes of patients with proliferative diabetic retinopathy and no diabetic macular edema treated with monthly or quarterly aflibercept therapy.
Customize your JAMA Network experience by selecting one or more topics from the list below.
Alagorie AR, Nittala MG, Velaga S, et al. Association of Intravitreal Aflibercept With Optical Coherence Tomography Angiography Vessel Density in Patients With Proliferative Diabetic Retinopathy: A Secondary Analysis of a Randomized Clinical Trial. JAMA Ophthalmol. 2020;138(8):851–857. doi:10.1001/jamaophthalmol.2020.2130
What is the association of macular vessel density change with intravitreal aflibercept injection in patients with proliferative diabetic retinopathy without diabetic macular edema?
In this post hoc analysis of a randomized clinical trial of 40 patients, the macular vessel density of the superficial capillary plexus, deep capillary plexus, and choriocapillaris was not associated with a change after monthly or quarterly intravitreal injections of aflibercept during 12 months of treatment.
The findings suggest that anti–vascular endothelial growth factor therapy may be useful for patients with retinal nonperfusion without exacerbating the nonperfusion.
Although previous studies have evaluated the association between anti–vascular endothelial growth factor therapy and macular vessel density, they were confounded by the presence of macular edema, which may be associated with artifacts and segmentation errors in optical coherence tomography angiography (OCTA).
To evaluate the association of intravitreal aflibercept with changes in macular vascular density using OCTA in patients with proliferative diabetic retinopathy without diabetic macular edema.
Design, Setting, and Participants
This post hoc analysis of a randomized clinical trial used data on 40 eyes of 40 patients with proliferative diabetic retinopathy without diabetic macular edema who were enrolled in the Intravitreal Aflibercept for Retinal Nonperfusion in Proliferative Diabetic Retinopathy (RECOVERY) clinical trial from August 1, 2016, to June 31, 2017. Three patients were lost to follow-up at month 12, and 5 patients were excluded from analysis because of poor OCTA image quality, leaving 16 patients in each cohort in the final analysis. Data analysis was performed from March 1, 2018, to January 15, 2019.
In the RECOVERY trial, patients were randomized into cohorts receiving 2 mg of aflibercept injections monthly (n = 20) or quarterly (n = 20) and treated for 12 months.
Main Outcomes and Measures
The percentage of vascular density (in total scan and foveal and parafoveal regions) was compared before and after 12 months of therapy.
The sample for this OCTA analysis included 32 eyes from 32 patients (mean [SD] age, 48.37 [12.30] years; 17 [53.1%] male). The mean (SD) total scan vascular density for the superficial vascular complex was 42.28% (4.03%; 95% CI, 40.63%-43.93%) at baseline and 39.64% (4.01%; 95% CI, 37.91%-41.37%) at month 12 (P = .69). For the deep vascular complex, the mean (SD) vascular density was 48.42% (4.99%; 95% CI, 46.36%-50.47%) at baseline and 45.69% (4.63%; 95% CI, 43.69%-47.70%) at month 12 (P = .40). For the choriocapillaris, the mean (SD) vascular density was 64.42% (3.36%; 95% CI, 63.04%-65.81%) at baseline and 62.55% (4.79%; 95% CI, 60.48%-64.62%) at month 12 (P = .16). There was no difference in vascular density parameters between monthly and quarterly injection arms at month 12.
Conclusions and Relevance
In this study, macular vascular density did not change after 12 months of intravitreal aflibercept therapy. Because nonperfusion is expected to progress in diabetic retinopathy, this finding may represent a beneficial association between anti–vascular endothelial growth factor therapy and macular vascular density.
ClinicalTrials.gov Identifier: NCT02863354
Proliferative diabetic retinopathy (PDR) and associated complications are a common cause of profound visual loss in patients with diabetes.1,2 Since 1981, panretinal photocoagulation has been a criterion standard of treatment for PDR.3 However, panretinal photocoagulation can be associated with adverse effects, including visual field constriction, decreased night vision, and worsening of coexisting diabetic macular edema (DME).4,5 More recently, the randomized Diabetic Retinopathy Clinical Research Network (DRCR.net) Protocol S clinical trial found that ranibizumab intravitreal therapy was noninferior to panretinal photocoagulation for the management of PDR. In that study, the ranibizumab group had better visual acuity and better visual field outcomes through at least 2 years compared with the panretinal photocoagulation group, with less need for vitrectomy.6
Studies7-11 evaluating the association of anti–vascular endothelial growth factor (VEGF) therapy with macular capillary perfusion using fluorescein angiography (FA) have reported inconsistent results, with some studies12,13 identifying a worsening of retinal nonperfusion and others8,14 suggesting an improvement in perfusion. These studies,8,12-14 however, were primarily based on FA, and the accuracy of FA for quantifying perfusion may be affected by image quality and potential obscuration by adjacent hemorrhage or leakage. In addition, FA does not allow the inner choroidal circulation to be evaluated in detail (at least at the capillary level). In comparison, optical coherence tomography angiography (OCTA) readily allows the retinal and choriocapillary circulations to be evaluated in detail in a depth-resolved fashion.15
Prior studies16,17 have considered vessel density changes using OCTA in patients with diabetes after anti-VEGF therapy. However, these studies16,17 have been limited by concurrent DME, which may cause segmentation errors and in particular inconsistent segmentation selection because the extent of edema changes with anti-VEGF therapy. In addition, signal attenuation from macular exudation, a common finding in eyes with DME, may also confound vessel density measurements. One study16 included mixed groups of macular edema secondary to diabetes and retinal venous occlusive disease after a single anti-VEGF injection, whereas another study17 included only 3 patients with short-term follow-up. In the current study, change in macular vessel density and flow area were evaluated before and after anti-VEGF intravitreal injection in patients with PDR without DME during 12 months using 2 different treatment regimens of monthly or quarterly aflibercept.
This post hoc analysis of a randomized clinical trial used data from the prospective, longitudinal, open-label Intravitreal Aflibercept for Retinal Nonperfusion in Proliferative Diabetic Retinopathy (RECOVERY) trial.18 The primary aim of the RECOVERY trial was assessment of the effect of intravitreal aflibercept injections given monthly or every 12 weeks on retinal capillary nonperfusion associated with PDR. In that trial, 40 participants were recruited from the clinical centers of the Retina Consultants of Houston (Houston, Katy, and Woodlands, Texas) from August 1, 2016, to June 31, 2017. The study was performed in accordance with the Health Insurance Portability and Accountability Act and adhered to the principles of the Declaration of Helsinki.19 Written informed consent was obtained from all of the patients before enrollment in the study, and all data were deidentified. No incentive or compensation was offered to participants. The study was approved by the Houston Methodist Institutional Review Board/Ethics Committee. For the present analysis, data collection and analysis were approved by the institutional review board of the University of California at Los Angeles.
Patients were included in the RECOVERY study if they had treatment-naive PDR with Early Treatment Diabetic Retinopathy Study (ETDRS) best-corrected visual acuity of at least 19 (20/400 Snellen equivalent) measured by trial lens refraction at a 4-m distance and with retinal nonperfusion of more than 20 disc areas on ultra-widefield FA. All patients underwent a comprehensive ophthalmic examination, including best-corrected visual acuity, slitlamp biomicroscopy, intraocular pressure measurement, dilated ophthalmoscopy, and spectral domain OCTA using the Heidelberg Spectralis HRA+OCT device (Heidelberg Engineering) with a volume or cube acquisition protocol (20° × 20°, 49 lines, 768 A-scans per line) and 9 × image average. Only 1 eye from each patient was included in this study.
Patients were excluded from the RECOVERY study if they had DME, vitreoretinal traction, vitreous hemorrhage, uveitis, uncontrolled glaucoma, macular laser, macular fibrovascular proliferation, media opacities that could affect image quality, and other retinal comorbidities, such as high myopia, age-related macular degeneration and retinal vascular disorders, history of anti-VEGF treatment, vitrectomy, or panretinal photocoagulation.
In the RECOVERY trial, patients were randomized into 1 of 2 treatment arms. Cohort 1 (n = 20) was treated with 0.05-mL intravitreal aflibercept injections (2 mg) every month (mean [SD], 28  days) for 12 months. Cohort 2 (n = 20) was treated with intravitreal aflibercept injections every 3 months for 12 months (baseline and weeks 12, 24, and 36).
For the present study, 3 patients were unavailable for follow-up at month 12, and 5 patients were excluded from analysis because of poor OCTA image quality (eg, low signal strength or motion artifact), leaving 16 patients in each cohort in the final analysis.
Patient eyes were dilated by instillation of 1% tropicamide and 2.5% phenylephrine before acquisition of ultra-widefield pseudocolor fundus and FA images using the Optos 200Tx (Optos plc) centered on the fovea and steered peripherally.20 The spectral-domain OCTA images were captured using the RTVue-XR 100 Avanti OCTA device (Optovue Inc). The instrument has an A-scan rate of 70 000 scans per second and uses a light source spectrum centered at an 840-nm wavelength and a bandwidth of 45 nm. The axial tissue resolution is 5 μm with a 22-μm beam width. Each OCTA 3 × 3-mm volume contains 304 × 304 A-scans. For OCTA, 2 consecutive B-scans are acquired at a given position before proceeding to the next location. Split-spectrum, amplitude-decorrelation angiography was used to extract the OCTA information. The volumes were registered and the decorrelation was calculated by comparison of the successive B-scan images and viewed as the maximal projection image of blood flow.21,22 The manufacturer-recommended default segmentation settings were used for layer and en face image extraction. Only images that were free of motion artifact were used for subsequent quantitative analysis. To generate the en face OCTA sections of the superficial vascular complex (SVC), the inner segmentation boundary was set at 3 μm below the internal limiting membrane and the outer boundary was set at 15 μm beneath the inner plexiform layer. For isolation of the deep vascular complex (DVC), the inner boundary was set at 15 μm beneath the inner plexiform layer and the outer boundary was set at 70 μm beneath the inner plexiform layer.23 The choriocapillaris (CC) was segmented as a 10-μm-thick section starting 31 μm posterior to the retinal pigment epithelium reference line.21 The segmentation boundaries were inspected by 2 certified Doheny Image Reading Center OCTA graders (A.R.A., M.G.N., and S.V.), and manual corrections were performed if needed to ensure accurate and consistent segmentation.
The following parameters were measured for SVC, DVC, and CC using the AngioVue Analytics algorithm: macular vessel density (for the total scan and foveal and parafoveal regions) and flow area of the total scan (Figure 1 and Figure 2). The foveal region was defined as a 1-mm ring centered on the fovea, and the parafoveal region was considered to be the zone between the 1-mm and 3-mm concentric rings centered on the fovea. Measurements were compared before and after 12 months of therapy.
With a sample of 32 eyes in the final analysis cohort, the study was 75% powered with a type I error of 0.05 and type II error of 0.25. A paired t test was performed to test the difference in the macular vessel density and flow area before and after 12 months of therapy, and an unpaired t test was performed to test the difference between the 2 cohorts at the month 12 visit. No adjustment was made for multiple analyses. A 2-sided P < .05 was considered to be statistically significant. Data analysis was performed from March 1, 2018, to January 15, 2019. Statistical analyses were performed using SPSS, version 20 (IBM Inc).
The sample for this OCTA analysis included 32 eyes from 32 patients (mean [SD] age, 48.37 [12.30] years; 17 [53.1%] male). Cohort 1 included 16 eyes, and cohort 2 included 16 eyes. At baseline, the median glycated hemoglobin level was 8.75% (95% CI, 8.62%-10.21%) (to convert to proportion of total hemoglobin, multiply by 0.01), and the median duration of diabetes mellitus was 16.01 years (95% CI, 13.19-18.96 years). The mean (SD) baseline best-corrected visual acuity was a 79.15 (7.55) ETDRS letter score (approximately 20/25 Snellen equivalent).
The mean (SD) central retinal thickness in all eyes decreased from 262.68 (29.42) μm (95% CI, 205-306 μm) to 231.31 (25.75) μm (95% CI, 159-270 μm) after treatment (P < .001). In further analysis of the 2 cohorts, the mean (SD) central retinal thickness in the monthly cohort decreased from 263.58 (33.66) μm (95% CI, 205-302 μm) to 225.84 (29.92) μm (95% CI, 159-256 μm) after treatment (P = .003). In the quarterly cohort, the mean (SD) decrease in mean central retinal thickness from 261.55 (24.53) μm (95% CI, 229-306 μm) to 237.28 (19.98) μm (95% CI, 201-270 μm) was not statistically significant (P = .06).
At the month 12 visit, the mean (SD) ETDRS letter score in the entire population improved from 79.15 (7.63) (20/25 Snellen equivalent) to 82.36 (8.22) (20/25 Snellen equivalent). In the analysis of each group, the ETDRS letter score in the monthly cohort improved from 78.60 (7.62) (20/32 Snellen equivalent) to 83.46 (8.89) (20/25 Snellen equivalent), whereas the ETDRS letter score in the quarterly cohort improved from 79.83 (7.93) (20/25 Snellen equivalent) to 81.17 (7.63) (20/25 Snellen equivalent).
The total scan, foveal, and parafoveal mean (SD) vessel density percentage for the SVC, DVC, and CC did not differ between baseline and month 12 for the entire population (SVC baseline vs month 12: 42.28% [4.03%] vs 39.64% [4.01%] for total scan, 24.28% [6.45%] vs 20.11% [5.83%] for fovea, and 43.58% [4.06%] vs 41.24% [3.72%] for parafovea; DVC baseline vs month 12: 48.42% [4.99%] vs 45.69% [4.63%] for total scan, 25.07% [6.13%] vs 24.08% [5.62%] for fovea, and 50.80% [4.90%] vs 47.59% [5.06%] for parafovea; and CC baseline vs month 12: 64.42% [3.36%] vs 62.55% [4.79%] for total scan, 62.92% [4.99%] vs 58.71% [7.41%] for fovea, and 64.40% [3.53%] vs 62.66% [4.78%] for parafovea) or the monthly injection (SVC baseline vs month 12: 42.57% [3.76%] vs 38.80% [4.26%] for total scan, 25.16% [6.67%] vs 20.40% [6.38%] for fovea, and 43.89% [3.94%] vs 40.60% [4.07%] for parafovea; DVC baseline vs month 12: 48.85% [3.79%] vs 44.33% [5.02%] for total scan, 25.98% [5.74%] vs 25.73% [5.94%] for fovea, and 51.23% [3.89%] vs 46.20% [5.70%] for parafovea; and CC baseline vs month 12: 64.93% [2.71%] vs 62.94% [4.25%] for total scan, 62.89% [5.81%] vs 59.23% [7.12%] for fovea, and 65.05% [2.55%] vs 63.19% [4.04%] for parafovea) and quarterly injection (SVC baseline vs month 12: 41.92% [4.44%] vs 40.57% [3.68%] for total scan, 23.18% [6.29%] vs 19.80% [5.47%] for fovea, and 43.18% [4.38%] vs 41.93% [3.35%] for parafovea; DVC baseline vs month 12: 47.86% [6.36%] vs 47.18% [3.86%] for total scan, 23.91% [6.67%] vs 22.29% [4.88%] for fovea, and 50.26% [6.10%] vs 49.11% [3.96%] for parafovea; and CC baseline vs 12 months: 63.78% [4.10%] vs 62.12% [5.50%] for total scan, 62.95% [4.01%] vs 58.14% [8.01%] for fovea, and 63.57% [4.48%] vs 62.09% [5.63%] for parafovea) cohorts (Table 1, Table 2, and Table 3). The mean (SD) flow area of the SVC, DVC, and CC did not differ between baseline and month 12 for the entire population (SVC: 3.44 [0.59] vs 3.30 [0.75] mm2; DVC: 3.35 [1.15] vs 3.27 [1.22] mm2; and CC: 5.40 [0.25] vs 5.39 [0.25] mm2) as well as for the monthly injection (SVC: 3.43 [0.65] vs 3.08 [0.91] mm2; DVC: 3.32 [1.24] vs 2.88 [1.49] mm2; and CC: 5.40 [0.28] vs 5.37 [0.26] mm2) and quarterly injection (SVC: 3.45 [0.53] vs 3.59 [0.36] mm2; DVC: 3.39 [1.10] vs 3.79 [0.40] mm2; and CC: 5.40 [0.22] vs 5.41 [0.25] mm2) cohorts (Tables 1-3). Change analysis (computing the difference between baseline and month 12 for each subject) of the vessel density percentage and flow area between the monthly and quarterly injection cohorts at month 12 showed no difference for any of the parameters.
In this study, the macular vessel density and flow area of the SVC, DVC, and CC did not change after monthly or quarterly intravitreal injections of aflibercept during 12 months. The observations from our study suggest that anti-VEGF therapy can be used in patients with retinal nonperfusion without exacerbating the nonperfusion. Our results are in line with the post hoc analysis of the Ranibizumab Injection in Subjects With Clinically Significant Macular Edema With Center Involvement Secondary to Diabetes Mellitus (RISE/RIDE) multicenter randomized clinical trial performed by Reddy et al,24 who reported no significant difference in the macular nonperfusion area between the 0.3-mg ranibizumab group and the sham group at months 12 and 24. Ghasemi Falavarjani et al16 reported unchanged retinal capillary density of SVC and DVC after a single intravitreal anti-VEGF injection. We observed the same findings after 12 months regardless of whether a patient was treated monthly or quarterly with aflibercept injections.
The mechanism by which VEGF can lead to retinal ischemia is not certain; occlusion of monkey eye retinal capillaries has been reported after VEGF injection as a result of increased expression of intercellular adhesion molecule 1 by endothelial cells.25 Upregulation of intercellular adhesion molecule 1 can increase the adhesion of activated monocytes and granulocytes to the endothelial cells, which can result in capillary occlusion.25-27 The subsequent retinal ischemia can lead to the release of additional VEGF, creating a positive feedback loop resulting in progressive vascular nonperfusion and retinal ischemia. On the basis of this occurrence, one might theorize that anti-VEGF therapy could reduce nonperfusion or at least the progression of nonperfusion by interrupting this cycle.24 The current results appear to support the association of anti-VEGF therapy with slowed development of additional areas of nonperfusion because nonperfusion (at least in the central macula) did not progress during the 12-month period of this study.
Despite the fact that the patients in this study did not have DME at baseline, the foveal thickness of the entire cohort significantly decreased at month 12 compared with baseline. The mild progressive thinning of the retina after anti-VEGF therapy despite the absence of baseline edema was a finding of interest. A separate analysis28 sought to assess this further by segmenting individual retinal layers and found that multiple layers (and not only the inner retina) demonstrated thinning. The mechanism of this thinning is not known; thus, the findings from our OCTA analysis may be important because they suggest that progressive nonperfusion and ischemia are not the underlying mechanism responsible for retinal layer thinning. It is possible that subtle, diffuse (noncystoid) edema or thickening was not apparent on clinical OCT evaluation. Despite the lack of worsening nonperfusion on OCTA, it is possible there was still some worsening of the flow or velocity that was below the detection sensitivity of OCTA. This possibility may require further investigations using techniques such as adaptive optics scanning laser ophthalmoscopy–based FA.29
This study has several strengths, including its prospective design, standardized acquisition protocols, and the evaluation of vascular density by OCTA rather than FA. These characteristics provided greater detail, lack of obscuration by leakage, and the ability to study the circulation in 3 dimensions.
This study has limitations. Because of the small sample size, the study was underpowered to detect smaller differences in nonperfusion over time or between cohorts. However, the current study was prospective and larger than previous studies16,17 that reported outcomes to 1 year of follow-up. Another limitation is related to the change in retinal thickness over time because this factor can affect the section selection. We had expected that a lack of the thickness change confounder would be a potential advantage of our analysis of patients without DME compared with previous reports17,30 that included patients with DME. The change in thickness, albeit smaller than in prior DME studies,31,32 could have had an effect because the section segmentations were based on a fixed offset. However, by considering the SVC and DVC, the entire retinal circulation was assessed; thus, it is unlikely that inconsistent segmentation would have significantly affected the findings. In addition, our study was limited to the central 3 mm2 of the macula. We could not determine whether there were significant changes in vessel density over time in more peripheral regions. We were unable to determine whether capillary perfusion remained stable longer than the 1-year duration of the current study. Larger longitudinal studies are needed in the future to confirm our findings.
During 1 year, we did not observe a change in central macular vessel density in eyes of patients with PDR without apparent DME treated with monthly or quarterly aflibercept therapy. The lack of change may reflect a beneficial association between anti-VEGF therapy and macular vessel density. Future studies with larger widefield OCTA are needed to confirm whether these observations extend to more peripheral regions of the retina.
Accepted for Publication: April 28, 2020.
Corresponding Author: SriniVas R. Sadda, MD, Doheny Image Reading Center, Doheny Eye Institute, 1355 San Pablo St, Ste 211, Los Angeles, CA 90033 (firstname.lastname@example.org).
Published Online: June 25, 2020. doi:10.1001/jamaophthalmol.2020.2130
Author Contributions: Dr Alagorie and Mr Nittala contributed equally to the study and are considered co–first authors. Drs Wykoff and Sadda had full access to all the data in the study and take the responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: Alagorie, Nittala, Sadda.
Acquisition, analysis, or interpretation of data: Alagorie, Nittala, Velaga, Zhou, Rusakevich, Wykoff.
Drafting of the manuscript: Alagorie, Zhou.
Critical revision of the manuscript for important intellectual content: Nittala, Velaga, Rusakevich, Wykoff, Sadda.
Statistical analysis: Alagorie, Nittala, Zhou.
Administrative, technical, or material support: Velaga, Zhou, Rusakevich.
Supervision: Nittala, Sadda.
Conflict of Interest Disclosures: Dr Zhou reported receiving grants from Regeneron during the conduct of the study. Dr Wykoff reported receiving grants and personal fees from Adverum, Allergan, Chengdu Kanghong, Clearside Biomedical, Genentech/Roche, Kodiak, Novartis, Regeneron, and Regenxbio; receiving grants from Opthea, Samsung, and Xbrane Biopharma; and receiving personal fees from Bayer and Alimera Sciences outside the submitted work. Dr Sadda reported receiving personal fees from Amgen, Allergan, Novartis, Roche/Genentech, Merck, 4DMT, Regeneron, Optos, Centervue, and Hedelberg and receiving nonfinancial support from Carl Zeiss Meditec, Nidek, and Topcon outside the submitted work. No other disclosures were reported.
Funding/Support: This study was supported by Regeneron (Dr Wykoff).
Role of the Funder/Sponsor: The funding source had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.