Article Text
PDF
Abstract
Aims To compare anti-vascular endothelial growth factor (VEGF) treatment outcomes for macular oedema (ME) secondary to retinal vein occlusion (RVO) based on vitreoretinal interface (VRI) status.
Methods This retrospective case series includes treatment-naive eyes diagnosed with RVO and treated with anti-VEGF injections. Eyes were stratified based on international VRI classification schema at baseline into three groups—vitreomacular traction (group A), no posterior vitreous detachment (PVD) (group B) and PVD without vitreomacular attachment (group C). Fifty-two eyes were identified based on inclusion/exclusion criteria. The primary endpoint was change in central subfield thickness (CST) on optical coherence tomography at 6 months.
Results There were no statistically significant differences in baseline characteristics of patients with RVO when stratified by VRI subgroups. After 6 months of treatment, there was no statistically significant difference in the change in CST from baseline between VRI cohorts (p=0.11). There was a trend demonstrating the greatest improvement in CST in eyes in group A compared with eyes in groups B and C (−224.13 μm, −160.88 μm and −50.92 μm, respectively, p=0.11 between cohorts). Mean change in logarithm of the minimum angle of resolution visual acuity from baseline to month 6 in group A compared with groups B and C was −0.25, −0.14 and −0.13, respectively (p=0.64 between cohorts).
Conclusions We did not identify an association between VRI status and treatment outcomes with anti-VEGF agents for ME secondary to RVO.
- Macula
- Retina
- Vitreous
Statistics from Altmetric.com
Introduction
Retinal vein occlusion (RVO) is a common retinal vascular disease that affects 1–2% of patients over the age of 40 and 16 million patients worldwide.1 Though the disease is characterised by blockage of either the central retinal vein or its branches, most vision loss in patients with RVO results from macular oedema (ME).1 Although the exact mechanism for the development of ME in eyes with RVO is unknown, it is widely accepted that inflammatory factors such as vascular endothelial growth factor (VEGF), cytokines and chemokines are the root cause of the ME that subsequently develops.2–4 Studies conducted by Noma et al found that vitreous levels of VEGF and interleukin (IL)-6 specifically were significantly higher in patients with ME secondary to branch retinal vein occlusion (BRVO) and central retinal vein occlusion (CRVO) than in controls, suggesting a plausible role for vitreous inflammatory mediators in the pathogenesis of ME in RVO.5–7 Contributing factors may include retinal ischaemia (which induces the production of cytokines in the occluded region affected by anoxia), damage to endothelial cells and subsequent impairment of the blood–retinal barrier and increased rigidity of a crossing artery causing compression of the underlying vein.8 Likewise, another proposed explanation is that RVO results in vitreoretinal adhesion and traction on the retina leading to vascular leakage and subsequently ME.9
Previous studies conducted prior to the advent of optical coherence tomography (OCT) have evaluated the role of the vitreoretinal interface (VRI) in treatment outcomes for RVO.7 ,10 ,11 While the vitreous is initially attached to the retina in its entirety, ageing results in the weakening of vitreoretinal adhesion and subsequently a progressive separation; this detachment typically begins in the macula but can progress through various stages of attachment before a total posterior vitreous detachment (PVD) occurs. A 1984 study conducted by Hikichi et al7 showed that in non-ischemic eyes with RVO the prevalence of posterior vitreous adhesion defined by clinical exam was significantly higher in patients with persistent ME than in those without. Similarly, a study conducted by Kado et al10 on central RVO in 1990 observed a lower incidence of ME in eyes with PVD than in eyes without vitreous separation from the macula again using clinical examination techniques to separate patients. A subsequent study done by Avunduk et al11 in 1997 indicated that a total PVD reduced the incidence and persistence of both retinal neovascularization and ME in eyes with RVO. These studies provide precedence in investigating the use of modern clinical diagnostics and treatments that could induce PVD to inhibit persistent ME secondary to RVO.12 ,13
The aim of this retrospective analysis is to approximate whether there is a significant difference in treatment outcomes for ME secondary to RVO when considering patients' VRI status at baseline in the modern era of OCT and anti-VEGF treatment.
Materials and methods
Institutional Review Board approval was obtained from the Cleveland Clinic for this retrospective study. Because of the retrospective nature of the study, written informed consent was not required. Patients were seen at the Cole Eye Institute between January 2011 and June 2014. All study-related procedures were performed in accordance with best practices and adhered to the Health Insurance Portability and Accountability Act. Patients were included in the record review if they met the following criteria: (1) a new diagnosis based on the International Classification of Diseases, ninth revision codes of central/hemiretinal, or branch vein occlusion (362.35 or 362.36), (2) the completion of an spectral domain OCT (SD-OCT) at the time of the initial exam and (3) 18 years of age or older. Exclusion criteria included patients who had been referred to or seen at Cole Eye Institute with an existing diagnosis of RVO, patients who had been treated in any capacity for RVO, any prior intravitreal injection treatment, uncontrolled glaucoma, presence of proliferative retinopathy, presence of epiretinal membrane based on review of medical records and OCT, neovascular age-related macular degeneration, history of retinal detachment, prior vitrectomy or prior injection of a vitreolysis agent.
The international classification system of VRI disorders was used to grade the initial OCT findings (Zeiss Cirrus SDOCT, V.6.0, Carl Zeiss Meditech, Dublin, California, USA).14 Two independent graders reviewed the entire cube scan of the OCT; any discrepancy was resolved by a third reviewer. Patient eyes were sorted into the following groups: vitreomacular traction (VMT) (group A), no PVD (group B) and PVD without vitreomacular attachment (VMA) (group C). VMA, as defined by the classification scheme, was identified based on perifoveal vitreous cortex detachment from the retinal surface, macular attachment of the vitreous cortex within a 3 mm radius of the fovea and no detectable change in foveal contour or underlying retinal tissues.14 VMT (group A) was also identified based on the classification definition of perifoveal vitreous cortex detachment from the retinal surface, macular attachment of the vitreous cortex within a 3 mm radius of the fovea and association of attachment with distortion of the foveal surface, intraretinal structural changes and/or elevation of the fovea above the retinal pigment epithelium, without a full-thickness interruption of retinal layers (eg, macular hole).14 A complete PVD was not explicitly defined in the classification schema but was described as a complete separation of the vitreous from the macula and optic nerve. Given this definition, diagnosis of complete PVD would be impossible based on OCT macular cube scans alone; as such, patient eyes were grouped into a ‘PVD without VMA’ group (group C) if they had a clinical diagnosis of PVD and absence of VMA or VMT on OCT. Eyes were assumed to have a complete vitreoretinal adhesion, labelled in this study as ‘no PVD’ (group B) if they had no evidence of vitreous detachment on exam or OCT defined as increased reflectivity of the hyaloid visible on the inner retinal surface.14
A total of 114 eyes were identified and divided into the following initial groups: no PVD (n=61), PVD without VMA (n=25), VMA (n=7) and VMT (n=21), determined by analysis of OCT recorded at baseline. Of the 114 initial patients that qualified for record review, 52 patients met the entry criteria for treatment analysis that included the following: concurrent diagnosis of ME defined by a central subfield thickness (CST) >300 μm, baseline visual acuity (VA) between 20/20 and 20/400 and treatment with anti-VEGF injection at baseline. No patient with VMA met entry criteria for the analysis. The treatment protocol was determined by retina specialists at a single institute based on comprehensive ophthalmic examination and OCT findings (Zeiss Cirrus SDOCT, V.6.0). Baseline demographics and clinical variables including Snellen VA and OCT parameters were recorded. SD-OCT parameters including CST, cube volume (CV), cube average thickness (CAT) and presence of cystoid ME were also documented from visits corresponding closest to 3 and 6 months.
Snellen VA measurements were converted to logarithm of the minimum angle of resolution (logMAR) values for statistical analysis. Pearson's χ2 and analysis of variance (ANOVA) tests were used to compare baseline characteristics between groups. Treatment analysis was conducted to assess clinical and anatomical differences in treatment outcomes for ME secondary to RVO based on vitreoretinal status at baseline. Statistical analyses included the conduction of t-tests and ANOVA to compare VA and OCT parameters between cohorts; a 0.05 significance level was assumed.
Results
Baseline characteristics and comparisons
Fifty-two eyes met the entry criteria for treatment analysis and were divided into the following groups: VMT (n=15) (group A), no PVD (n=24) (group B) and PVD without VMA (n=13) (group C). Table 1 summarises the baseline characteristics of patients who met the inclusion criteria for treatment analysis, as defined in the ‘Materials and methods’ section. The mean age at baseline was 70.3±12.8 years and 57.7% (n=30) were male. The mean baseline logMAR VA was 0.64 (SD=0.37; Snellen equivalent of 20/87) and the mean CST at baseline was 483.0±179.9 μm. The average follow-up at 6 months was 174.14±20.29 days for group A, 176.52±13.32 days for group B and 181.50±18.07 days for group C. Because of the need for the initial presence of ME, no patients with VMA met the inclusion for treatment analysis as patients with VMA and foveal cystoid ME were categorised as VMT by the international classification of VRI. There were also no patients in the VMT group at baseline that switched groups to VMA following treatment, suggesting alterations in the foveal architecture in the VMT group were largely due to VRI status and not due to ME from RVO.
Baseline characteristics of patients qualified for treatment analysis
When comparing baseline demographics across groups stratified by VRI status, patients in group A were significantly more likely to be phakic than those in group B (p=0.027). There were no significant differences between cohorts when considering age (p=0.16), sex (p=0.89) and laterality (p=0.24). Similarly, no significant differences in RVO subtype (hemiretinal vein occlusion, BRVO, CRVO) between groups were present (p=0.73). Baseline measures including logMAR best-corrected visual acuity (BCVA) and OCT parameters were also compared among groups; no significant differences were observed (table 1).
Clinical and anatomic treatment outcomes
Improvement in BCVA from baseline to month 6 was statistically similar for patients irrespective of vitreoretinal status (figure 1; p=0.64 between groups). Only patients in group A made significant improvements in vision at all time points (p=0.009 at 3 months; p=0.013 at 6 months). Final vision at 6 months was statistically similar between cohorts (Snellen equivalent: 20/41, 20/56, 20/71, in groups A, B and C, respectively; p=0.22 between cohorts).
Change in logarithm of the minimum angle of resolution (logMAR) best-corrected visual acuity from baseline. Only vitreomacular traction (VMT) patients made significant improvements in logMAR BCVA at 3 and 6 months. There was no statistically significant difference in change from baseline between cohorts at 6 months (p=0.64). PVD, posterior vitreous detachment; VMA, vitreomacular attachment.
Reduction of CST from baseline to 6 months was comparable between groups despite differences in vitreoretinal status (figure 2; p=0.11). Group A showed the trend with the greatest reduction in CST (baseline to month 6: −224.13 μm, p=0.005). Final CST at 6 months was statistically similar between cohorts (CST: 306.21 μm, 327.39 μm, and 328.25 μm in the group A, B and C cohorts, respectively; p=0.80 between cohorts).
Change in central subfield thickness (CST) from baseline. There was no statistically significant difference in change from baseline between cohorts at 6 months (p=0.11). PVD, posterior vitreous detachment; VMA, vitreomacular attachment; VMT, vitreomacular traction.
Reductions in CV and CAT were similarly commensurate between cohorts irrespective of vitreoretinal status (p=0.78 and 0.77, respectively). There was no statistically significant difference between CV and CAT values at 6 months despite differences in vitreoretinal status (table 2; p=0.84 and 0.86, respectively).
CST, CV and CAT at baseline and 6 months
The mean number of anti-VEGF injections over 6 months across the VRI groups was similar (group A=3.13±1.68, group B=4.33±1.58, group C=3.62±2.06, respectively; p=0.11).
Discussion
Previous studies by both Hikichi et al7 and Avunduk et al11 showed that patients with RVO with a vitreous separation had reduced incidence of neovascularization and reduction of ME compared with patients without a PVD. A significant drawback to these studies was classification of the VRI using either a Goldmann lens or el Bayadi-Kajiura lens alone and diagnosis of ME based on biomicroscopy and fluorescein angiography (FA) alone. An international classification scheme for the VRI was recently published to set guideline terms for identifying VRI using OCT criteria.14 Using these guidelines, we aimed to re-create these studies in the era of OCT and found quite different results.
Both Hikichi et al7 and Avunduk et al11 proposed that the vitreous was responsible for the persistent ME in patients with traction on Müller cells and the delivery of pharmacological factors within the vitreous to the retina. Although the exact mechanism for the development of ME in eyes with RVO remains unclear, additional studies substantiate these theories. Noma et al5–7 found that vitreous levels of VEGF and IL-6 were significantly higher in patients with ME secondary to BRVO and CRVO than in controls while studies by Schepens et al15 and Sebag16 suggest that centripetal traction caused by vitreous contraction after an RVO causes cellular swelling leading to ME. Given these findings, it was expected that retinal traction from VMT in eyes with ME secondary to RVO would preclude improvement with anti-VEGF injections. However, results from this pilot study indicate that, in patients with PVD, the absence of VMT provides no statistically significant benefit in terms of treatment outcomes with anti-VEGF therapy.
There are several differences in methodology that may account for the discrepancy between results of the present study with previous studies. While both Hikichi and Avunduk measured ME treatment outcomes by either the presence or absence of leakage on FA at final follow-up, the present study used BCVA and SD-OCT measurements to measure treatment outcomes. Importantly, what is being measured in these two tests, that is, macular leakage in FA and macular thickness in OCT, is not always concordant.17 A recent comparison done by Jittpoonkuson et al18 showed that SD-OCT demonstrates greater sensitivity than FA in detecting cystoid macular oedema (CME), particularly in patients with RVO. Subsequently, results of this study could vary from past results due to differences in both measurement of treatment outcomes and even classification of ME at baseline.
Additionally, while the studies performed by Hikichi and Avunduk et al primarily used observation and laser photocoagulation as treatment methods, the analysis conducted for this study was strictly limited to patients being treated with anti-VEGF injections. Panretinal laser argon photocoagulation therapy has been shown to induce a total PVD over time.19 There is conflicting evidence on whether anti-VEGF injections induces changes in VRI status; a prospective observational study by Geck et al20 including 61 eyes with a variety of retinal diseases showed development of a PVD in 15 out of 61 eyes (24.5%) while a cohort study by Veloso et al21 investigating the development of a PVD in 396 eyes with age-related macular degeneration treated with anti-VEGF showed development of a PVD rarely (5.6% of eyes after an average of 8.3 injections).
While this study tracked treatment outcomes at 3 and 6 months, studies by Avunduk et al11 followed patients for 1–2 years and studies by Hikichi et al7 followed patients anywhere between 1 and 7 years with a median of 4 years. Therefore, while PVD may improve treatment outcomes in patients with RVO given certain treatments or time scales, the present study indicates that using modern treatment regimens PVD may only be an epiphenomenon that has no cause and effect relationship to BCVA improvement in patients with ME secondary to RVO.
In this study, eyes with VMT counterintuitively demonstrated the greatest gains in BCVA and greatest reductions in CST at month 6. However, these findings are consistent with recent literature examining the role of VMA on visual and anatomic outcomes for both diabetic ME and BRVO treated with anti-VEGF. A retrospective study done by Sadiq et al22 based on data from the READ-3 trial found that diabetic ME patients with VMA had significantly greater gains in BCVA and improvement in OCT compared with VMA-negative patients. Another study by Terao et al23 looked at the effect of VMA on treatment outcomes for BRVO with anti-VEGF therapy and found that the VMA-positive group showed more improvement in BCVA than did the VMA-negative group (p=0.0150), and a greater decrease in CST after adjusting for age (p=0.0019). These studies are supportive of the conclusions within this study albeit with different entry criteria and with other retinovascular diseases. Finally, while it is possible that release of VMT during the course of anti-VEGF injections (due to the mechanics of an intravitreal injection) could have resulted in the VMA/VMT groups experiencing greater improvements in CST, patients’ VRI classification did not change when VRI status was examined at 6 months.
Weaknesses of this study include uncertainty in classification of vitreous status and both known and unknown confounding factors. Regarding classification of the vitreous status: while the designation of eyes with VMT was made strictly based on definitions from the international VRI classification scheme, the classification of eyes as ‘PVD without VMA’ and ‘no PVD’ may have been affected by reliance on OCT findings and recording of PVD in the chart versus ultrasound findings. Additionally, known confounding factors including phakic status as well as unknown confounding variables such as time from RVO onset to treatment and differences in occlusion site may have affected treatment outcomes between the VRI groups.
The strengths of this study include a review of patients from a single institution, the inclusion of only newly diagnosed, treatment-naive patients, the strict exclusion of concurrent ocular disease and the analysis of anatomic factors such as CME and vitreous attachment on OCT. The limitations of this study include a small sample size and its retrospective nature. Our available sample size was small, especially after dividing the cohort by diagnosis. Adequate statistical power exists only to find significance for large differences between groups on our endpoints, limiting the meaning of negative findings. However, the similarity of the mean levels across groups on all of endpoints at 6 months lends credence to the negative finding that PVD does not affect 6-month outcomes for ME secondary to RVO. The small sample size precluded the ability to analyse patients based on subtypes of RVO. It also limited the study to 6 months of follow-up data. It also may have led to a selection bias regarding the type of RVO studied; a study by Rogers et al24 looking at the prevalence of RVO in the population showed that BRVO occurs approximately 5.5 times more often than CRVO in the general population. Our study has a distribution of around 1:1 patients with BRVO to CRVO. However, several studies have shown similar outcomes when BRVO and CRVO patients are treated with anti-VEGF for ME.25 Other limitations include variation in physician treatment style, which may have introduced bias related to the nature and frequency of treatment chosen. Exclusion of patients with a baseline BCVA of counting fingers, hand motion and light perception from treatment analysis based on the unreliability of those measurements may have created a potential bias of including patients with better vision at baseline.
In conclusion, this pilot study does not support earlier findings that PVD leads to enhanced treatment outcomes for ME secondary to RVO. The power of the study may be a limitation; larger, prospective studies will be needed to validate these findings and to elucidate the primary predictive factors for improvement in BCVA in patients with RVO and subsequent ME.
References
Footnotes
Contributors All authors met the ICJME criteria: (1) substantial contributions to conception and design, acquisition of data or analysis and interpretation of data; (2) drafting the article or revising it critically for important intellectual content and (3) final approval of the version to be published.
Funding Support provided by Research to Prevent Blindness (Cole Eye Institutional Grant).
Competing interests RPS: Regeneron (grants and personal fees), Alcon (grants and personal fees), Genentech (grants and personal fees), Shire (personal fees), Zeiss (grants), Biogen (personal fees), during the conduct of the study; JPE: Thrombogenics (grants and personal fees), Alcon (personal fees), Zeiss (personal fees), Leica (personal fees), Genentech (grants), Regeneron (grants), Santen (personal fees), Alimera (personal fees), outside the submitted work; APS: Cleveland Clinic (full-time employee), State of Ohio (part-time employee), Elsevier (Royalties), American Academy of Ophthalmology (Honorarium), Easton Capital (possible future payments, nothing to date); SKS: Santen (personal fees), Bausch and Lomb (grants and personal fees), Synergetics (personal fees), Zeiss (grants and personal fees), Sanofi (grants and personal fees), Optos (personal fees), Regeneron (grants and personal fees), Allergan (grants and personal fees), outside the submitted work; PKK: Alcon (personal fees), Bayer (personal fees), Regeneron (personal fees), Novartis (personal fees), Kanghong (personal fees), Thrombgenics (personal fees), outside the submitted work.
Ethics approval Cleveland Clinic Institutional Review Board.
Provenance and peer review Not commissioned; externally peer reviewed.