Effects of the epiretinal membrane on the outcomes of intravitreal dexamethasone implantation for macular edema secondary to branch retinal vein occlusion

ABSTRACT Purpose: To investigate the effects of epiretinal membrane formation on the clinical outcomes of intravitreal dexamethasone implantation for macular edema secondary to branch retinal vein occlusion. Methods: This retrospective interventional case series includes the treatment of naive patients with macular edema secondary to non-ischemic branch retinal vein occlusion who underwent intravitreal dexamethasone implantation. The patients were divided into two groups as follows: Group 1 (n=25), comprised of patients with macular edema secondary to branch retinal vein occlusion without epiretinal membrane, and Group 2 (n=16), comprised of patients with macular edema secondary to branch retinal vein occlusion with an epiretinal membrane. Corrected visual acuity, central macular thickness, and central macular volume values were measured before and after treatment. The clinical outcomes of the groups were compared. Results: Mean age and male-to-female ratio were similar between the two groups (p>0.05, for both). The baseline and final corrected visual acuity values, central macular thickness, and central macular volumes of the groups were similar (p>0.05, for all). All the parameters were significantly improved after intravitreal dexamethasone implantation treatment (p<0.001, for all). The changes in central macular thickness and volume were also similar (p>0.05, for both). The mean number of intravitreal dexamethasone implantations was 2.1 ± 1.0 (range, 1-4) in Group 1 and 3.0 ± 1.2 (range, 1-5) in Group 2 (p=0.043). Conclusion: Epiretinal membrane formation had no effects on the baseline and final clinical parameters, including corrected visual acuity and central macular thickness and volume. The only parameter affected by the presence of epiretinal membrane formation is the number of intravitreal dexamethasone implantations, a greater number of which is needed for macular edema secondary to branch retinal vein occlusion with an epiretinal membrane.

ABSTRACT | Purpose: To investigate the effects of epi retinal membrane formation on the clinical outcomes of intravitreal dexamethasone implantation for macular edema secondary to branch retinal vein occlusion. Methods: This retrospective interventional case series includes the treatment of naive patients with macular edema secondary to nonischemic branch retinal vein occlusion who underwent intravitreal dexamethasone implantation. The patients were divided into two groups as follows: Group 1 (n=25), comprised of patients with macular edema secondary to branch retinal vein occlusion without epiretinal membrane, and Group 2 (n=16), comprised of patients with macular edema secondary to branch retinal vein occlusion with an epiretinal membrane. Corrected visual acuity, central macular thickness, and central macular volume values were measured before and after treatment. The clinical outcomes of the groups were compared. Results: Mean age and maletofemale ratio were similar between the two groups (p>0.05, for both). The baseline and final corrected visual acuity values, central macular thickness, and central macular volumes of the groups were similar (p>0.05, for all). All the parameters were significantly improved after intravitreal dexamethasone implantation treatment (p<0.001, for all). The changes in cen tral macular thickness and volume were also similar (p>0.05, for both). The mean number of intravitreal dexamethasone implantations was 2.1 ± 1.0 (range, 14) in Group 1 and 3.0 ± 1.2 (range, 15) in Group 2 (p=0.043). Conclusion: Epiretinal membrane formation had no effects on the baseline and final clinical parameters, including corrected visual acuity and central macular thickness and volume. The only parameter affected by the presence of epiretinal membrane formation is the number of intravitreal dexamethasone implantations, a greater number of which is needed for macular edema secondary to branch retinal vein occlusion with an epiretinal membrane.

INTRODUCTION
Retinal vein occlusion (RVO) is one of the most com mon reasons for visual loss associated with retinal vascu lar disease (1) . It is prevalent in 12% of people aged >40 years, and the prevalence of branch RVO (BRVO) is four times greater than that of central RVO (1) . Macular edema (ME) is a common complication of BRVO and has the po tential to permanently disrupt the macular architecture if left untreated (2) . In the past, treatment options were highly limited for ME secondary to BRVO (3) . Subsequent randomized controlled studies demonstrated improve ment in the clinical outcomes of ME secondary to BRVO treated with intravitreal administrations of antivascular endothelial growth factors (antiVEGFs) and steroids (48) . Intravitreal dexamethasone implantation (IDI) was found to be more effective than sham injections (9) . The Geneva study (10) reported significant improvement in visual acuity in ME secondary to RVO.
An epiretinal membrane (ERM) is a disease of the vi treomacular interface involving both the macular and perimacular regions and can cause visual impairment or metamorphopsia. Anomalous posterior vitreous detach ment resulting in vitreoschisis and vitreoretinal traction has been widely understood to be the most important pathophysiological mechanism (11) . Secondary ERM can be associated with inflammatory and retinal vascular di seases or retinal detachments (12) . The progression of ERM is generally slow and is not always clinically important; however, in association with other retinal conditions, mechanical vitreoretinal traction may change the course of underlying diseases and affect their treatment respon se (12) . The aim of this study was to investigate the effects of ERM formation on the anatomical and functional outcomes of IDI for ME secondary to BRVO by evaluating realworld data.

METHODS
This retrospective interventional case series was conducted in a single tertiary referral hospital between June 2015 and June 2019. Approval was obtained from the local research ethics committee (Ankara Numune Education and Research Hospital). After a detailed explanation of the protocol, written informed consent was obtained before IDI was performed. All the proce dures were performed in accordance with the ethical standards of the Declaration of Helsinki for human subjects.Medical records documenting the treatment of naive Caucasian patients with ME secondary to non ischemic BRVO who underwent IDI as firstline therapy were investigated. The patient inclusion criteria were as follows: 1) age >18 years; 2) clinically (presence of intraretinal microvascular abnormalities or anastomo tic vessels, localized retinal edema, venous dilation or sheathing within the retinal quadrant corresponding to the obstructed vein, and superficial or deep retinal he morrhage) (1) and angiographically (delayed armretinal transit time, late staining of the vein, nonperfusion or hyperpermeability of the retinal capillary bed in the is chemic area, petaloid pattern hyperfluorescence in the cystoid ME without macular ischemia associated with the increased foveal avascular zone) (13) documented nonischemic (nonperfusion area of the retinal capillary ≤5 discdiameters on fluorescein angiography) (14) BRVO history ≤6 months; 3) central macular thickness (CMT) >300 µm; 4) cataract surgery; and 5) at least 3 months of followup after IDI. The exclusion criteria were as follows: 1) history or clinical findings of other retinal diseases (e.g., diabetic retinopathy, agerelated macular dystrophy, degenerative myopia, retinitis pigmentosa, or uveitis); 2) history of previous retinal treatment (e.g., vi trectomy, intravitreal injection or implantation, or laser photocoagulation); 3) history of increased intraocular pressure or antiglaucomatous use and other risk factors of glaucoma (e.g., glaucoma history in a family member or thin central corneal thickness); 4) media opacity (e.g., corneal opacity, hyphema, or vitreous hemorrhage); 5) loss of vision due to other causes (e.g., neuroophthalmo logical diseases, retinal artery occlusion, or amblyopia); and 6) other reasons for secondary ERM (e.g., ocular trauma or primary vitreoretinal diseases).
Medical history and other ocular findings, including corrected visual acuity (CVA), were obtained from the patients' medical records. CVA was determined using a Snellen chart, and the data were converted to logMAR. Colored fundus photographs, fundus autofluorescence, and fundus fluorescein angiograms were evaluated using a scanning laser ophthalmoscope (Heidelberg Retina An giography 2, Heidelberg Engineering, Heidelberg, Ger many). Macular configuration, vitreomacular interface, and quantitative analysis of CMT and central macular volume (CMV) were measured using spectraldo main OCT (Spectralis, Heidelberg, Germany), and quality scores ≥20 were considered acceptable. CMT was measured as the thickness of the central fovea, and CMV was measu red in both the fovea and 6mm perifoveal circular area. The patients were divided into two groups according to the presence or absence of ERM before treatment initia tion. ERM was diagnosed as a hyperreflective membrane formation on the innermost layer of the retina on OCT. Group 1 was comprised of patients with ME secondary to BRVO without ERM, and Group 2 was comprised of pa tients with ME secondary to BRVO with ERM.
IDI (Ozurdex, Allergan, Inc., Irvine, CA, USA) was performed under sterile conditions as a firstline treat ment for the all the patients. Then, the patients were instructed to use topical 0.5% moxifloxacin for a week. Retreatment was performed at least 3 months after the previous implantation if the ME persisted and the CVA did not improve as compared with the initial visit. Peripheric scatter retinal laser photocoagulation was applied in one or more sessions if evidence of peripheral retinal ischemia or neovascularization was found. Similarly, focal laser photocoagulation was performed if a focal ischemic area was observed close to the RVO region on angiography. IDI treatment was terminated in the following conditions: 1) complete ME regression and CVA stability in consecu tive followups; 2) incidence of adverse events, and 3) switching to another treatment option.
The Statistical Package for the Social Sciences (SPSS) 22.0 software (IBM Corp., New York, USA) was used for the statistical analysis. Descriptive data were pre sented as mean ± standard deviation (range). The KolmogorovSmirnov test was used to check the normal distribution of the variables. The MannWhitney U test was used to compare the groups, as the numerical data did not conform to a normal distribution. The statistical significance was set at p<0.05. The Sample Size Calcu lator software (ClinCalc LLC, Indianapolis, IN, USA) was used for the power analyses of the parameters, which showed significant differences.

RESULTS
Group 1 included 25 eyes of 25 patients, and Group 2 included 16 eyes of 16 patients. The mean age of the patients was 54.6 ± 6.4 years (range, 4468 years) in Group 1 and 59.3 ± 9.1 years (range, 4671 years) in Group 2. The maletofemale ratio was 16:9 in Group 1 and 9:7 in Group 2. The demographic characteristics of the two groups were similar (p>0.05, for both).
Diabetes mellitus was the most common systemic comorbidity in both groups. Ten participants in Group 1 had diabetes mellitus, with a mean disease duration of 4.2 ± 3.6 years (range, 212 years), whereas five par ticipants in Group 2 had diabetes mellitus, with a mean disease duarion of 5.5 ± 3.4 years (range, 211 years). Systemic hypertension and coronary artery diseases were the other most common systemic comorbidities after diabetes mellitus. The baseline clinical characteris tics of the patients in the groups were similar (p>0.05, for all), as shown table 1. Peripheric scatter retinal laser photocoagulation was applied in one eye, and focal laser photocoagulation was also applied in one eye in Group 1. Focal laser photo coagulation was applied in one eye in Group 2. None of the eyes developed ERM after laser photocoagulation in Group 1 during the followup period.
The mean followup time was 8.7 ± 3.0 months (range, 322 months) in Group 1 and 11.4 ± 3.9 months (range, 324 months) in Group 2 (p>0.05). The mean number of IDIs and duration between IDIs were 2.1 ± 1.0 months (range, 14 months) and 4.1 ± 1.4 months (range, 36 months) in Group 1 and 3.0 ± 1.2 months (range, 15 months) and 3.9 ± 1.3 months (range, 36 months) in Group 2, respectively. The mean number of IDIs was statistically higher in Group 2 (p=0.043), while the mean duration between IDIs was similar between the two groups (p>0.05). The clinical outcomes are summarized in table 2.
According to the power analysis results, in the com parison of two independent samples for the mean num ber of IDIs, a power of 80% could be attained if each group consisted of a minimum number of 16 patients (enrollment ratio 1:1, α=0.05, and β=0.2). The sample sizes of the groups in this study were in accordance with this condition.

DISCUSSION
Currently, the most commonly used treatments for ME secondary to BRVO are intravitreal administration of antiVEGF agents and dexamethasone. Many studies have compared the clinical outcomes of antiVEGF and dexamethasone treatments. Comparable results have been reported in terms of anatomical and functional impro vements, especially at mid and longterm followups (15,16) . Conversely, some adverse events such as cataract forma tion and intraocular pressure increase are more likely to occur after intravitreal dexamethasone treatment (15,16) . Therefore, many physicians conclude that dexametha sone treatment may be a more suitable alternative for pseudophakic patients, especially when considering the need for less frequent intravitreal administration (15,16) . In this study, IDI was preferred as a firstline treatment solely for pseudophakic patients without a history of intraocular pressure increases. Moreover, no persistent intraocular pressure increases requiring antiglaucoma tous medications were observed during the early or longterm followup period in this study.
Pars plana vitrectomy, membrane peeling, and the intraocular gas tamponade injection protocol are gene rally considered standard treatment options for patients with symptomatic ERM, with generally quite satisfying anatomical outcomes (17) . Sometimes, residual intrareti nal edema may persist, and adjuvant pharmacological therapy with intravitreal injection of steroid derivatives in addition to vitreoretinal surgery may accelerate the resolution of the associated intraretinal edema and hasten the recovery of visual function. This adjuvant thera py may be administered during or after surgery (1820) . However, ERMassociated retinal comorbidities may better respond to less invasive treatments, and intravi treal pharmacotherapy may be used for some cases in place of vitreoretinal surgery. For instance, intravitreal antiVEGF agent injection is accepted as a firstline the rapy for diabetic ME patients with ERM formation (21) . In travitreal drug administration may increase the vitreous volume due to vitreous liquefaction. Thus, spontaneous ERM separation may occurr, and this process is similar to the action mechanism of the mechanical relief of traction in pars plana vitrectomy (22) .
Baseline clinical characteristics are thought to be worse in ME patients with ERM. Mechanical vitreoretinal traction may increase CMT and CMV, and the optical barrier effect of a thicker membrane may play an addi tional role in decreasing CVA. Yiu et al. (23) reported that patients with RVOrelated ME had worse baseline CVA when ERM formation was present. In the study by Wong et al. (24) ME caused by another etiology and similar base line CVA and CMT values between diabetic ME with and without ERM were reported. In this study, the baseline clinical characteristics, including CVA, CMT, CMV, and duration before the first IDI, were similar between the two groups with and without ERM. Therefore, ERM was not an important additional risk factor for the worsening of baseline clinical characteristics or the need of earlier treatment in this study.
Only a limited number of studies have investigated the effects of intravitreal treatment on different patient groups with and without ERM. In diabetic ME patients with and without ERM, in the early term or midterm after intravitreal antiVEGF injection, the CVA and CMT results were not fully consistent. Maryam et al. (25) reported that 1 month after intravitreal bevacizumab injection, CMT improved only in diabetic ME patients without ERM, and CVA improved only in diabetic ME pa tients with ERM. Ercalik et al. (26) reported improvements in CVA and CMT in diabetic ME patients without ERM, but only CMT improved in diabetic ME patients with ERM at 3month followup. Wong et al. (24) reported worse CVA results 12 months after intravitreal ranibizumab therapy in diabetic ME patients with ERM. In neovascu lar agerelated macular degeneration, which is another common indication for intravitreal treatment, limited responses to the intravitreal aflibercept injection may be observed in patients with and without ERM especially at short followup time points. Cho et al. (27) reported worse 12month CMT results in patients with neovascular agerelated macular degeneration with ERM after they received intravitreal aflibercept injection. Followup time should be considered before evaluating the clinical response to intravitreal treatment in patients with and without ERM because early and longterm outcomes may differ. In this study, the clinical parameters after IDI and their changes were similar between the patients with ME secondary to BRVO with and without ERM at a relatively long followup period. One important reason for the similar outcomes between the groups may be their similarities in baseline characteristics because ba seline clinical parameters have been identified before as possible predictors of longterm outcomes (28) .
Similar to longterm outcomes, the need for a greater amount of intravitreal dexamethasone therapy may be predicted by considering some clinical data. Baseline visual acuity and early treatment response have been reported as two predictors in patients with ME secon dary to BRVO (28) . In this study, the baseline clinical para meters and their changes were quite similar. However, the patients with ME secondary to BRVO with ERM formation needed more IDIs than the patients without ERM formation. Theoretically, the reason why some ME patients with ERM require more intravitreal treatments may be reduced perfusion of the drug into the retina due to the mechanical barrier effect of the ERM structu re (26) . This hypothesis is compatible with the finding of this study because the only difference between the two groups was the presence of ERM formation.
Randomized controlled studies are considered a "gold standard" for clarifying the efficacy and safety of treatment modalities for any diseases. The outcomes of these studies are quite accurate because they include carefully selected, highly homogeneous study groups with strict treatment and followup schedules. However, the outcomes of realworld studies may not completely correlate with the outcomes of randomized controlled studies. In this regard, realword data have better ex ternal validity despite having a lower certainty level (29) . The most important aspect of this study is that it directly reflects the longterm, realworld results of the effects of ERM on clinical outcomes after IDI for ME secondary to BRVO, which no other study has previously done as far as we know. Moreover, the evaluation of CMV is another essential aspect of this study. CMV is more associated with diffuse ME than with focal edema and can provide a more accurate information about the treatment outcomes in patients with ME (3032) . The CMV results in this study could not be compared with those reported in the litera ture because, to the best of our knowledge, no previous studies have reported CMV results of patients with ME secondary to BRVO with ERM who underwent dexame thasone treatment. However, we observed that both the baseline and final CMVs and their changes were compa tible with the values of the other clinical parameters. We presumed that CMV will be an important parameter for evaluating ME in future OCTbased studies.
This study has some limitations, including its small sample size (despite having a high statistical power) and retrospective design. Owing to being a realworld study and the longer efficacy period of dexamethasone than that of antiVEGF agents, the posttreatment control time points during the followup period were not standardized for all the patients, and only baseline and final clinical data could be included in the statistical analyses. In addi tion, the ERM pattern and ellipsoid zone or the external limiting membrane integrity were not evaluated.
In conclusion, ERM formation had no effects on the baseline and final values of the clinical parameters, including CVA, CMT, and CMV, and on the changes in these parameters and the duration before the first IDI for ME secondary to BRVO. The only parameter affected by the presence of ERM formation was the number of IDIs, of which a more significant number is needed for ME secondary to BRVO with ERM. sortium. The prevalence of retinal vein occlusion: pooled data from population studies from the United States, Europe, Asia, and Australia. Ophthalmology. 2010;117(2):3139.