FormalPara Key Summary Points

Why carry out this study?

To investigate changes in the vitreoretinal interface after anti-vascular endothelial growth factor (anti-VEGF) treatment in highly myopic eyes.

What was learned from the study?

Changes in the vitreoretinal interface, including progression, onset, and improvement of macular retinoschisis (MRS), were observed after anti-VEGF treatment in highly myopic eyes.

Outer schisis and lamellar holes were risk factors for progression and onset of MRS.

Intravitreal injection of ranibizumab and retinal hemorrhage were protective factors for progression to vision-threatening MRS that required surgical intervention.

Introduction

Myopic choroidal neovascularization (mCNV) is a vision-threatening complication in eyes with pathologic myopia [1]. The prevalence of mCNV ranges from 5% to 11% among individuals with high myopia [2,3,4]. In addition to mCNV, intraretinal splitting of the inner and outer retinal layers, recognized as myopic macular retinoschisis (MRS), has been noted in 9% to 34% of highly myopic eyes [5, 6]. The terminology used in discussing this condition has not been consistent, and Panozzo and Mercanti proposed the term “myopic traction maculopathy” (MTM) to describe the changes in the vitreoretinal interface caused by traction forces from the epiretinal membrane or by residual focal vitreoretinal adhesion combined with posterior staphyloma and progressive scleral stretching in myopic eyes [7]. Optical computed tomography (OCT) is essential for the diagnosis of MTM, which manifests as three sometimes overlapping subtypes, i.e., MRS, foveal detachment (FD), and macular holes [8, 9]. The impairments in visual acuity associated with each subtype have been described in previous publications [8, 10]. The natural course of MRS varies, with some eyes with MRS experiencing progressive visual deterioration, while others may have stable visual acuity for years [11, 12]. Surgery is indicated when there is decreased visual acuity, FD, full-thickness macular hole (FTMH), or a schisis shaped like a champagne flute. [12, 13]

The incidence of myopia has increased rapidly in recent decades in many countries, particularly in East Asia [14,15,16] and, consequently, the management of associated comorbidities has become increasingly important. Intravitreal (IV) injection of anti-vascular endothelial growth factor (anti-VEGF) has been shown to effectively improve vision in patients with mCNV [17, 18]. The progression and onset of MRS in mCNV eyes after anti-VEGF treatment have been reported [19,20,21,22], but the risk and protective factors have not yet been studied in a large series. In the present study, we collected cases of eyes with mCNV treated with anti-VEGF, investigated the clinical features, and identified the risk and protective factors for vitreoretinal interface changes after anti-VEGF treatment in highly myopic eyes.

Methods

Medical records were retrospectively reviewed on consecutive patients with high myopia and mCNV who had applied for National Health Insurance coverage of anti-VEGF treatment at the National Taiwan University Hospital between July 2016 and September 2021. The inclusion criteria were: (1) presence of high myopia, defined as a refractive spherical equivalent (SE) of at least — 6.00 diopters (D) or an axial length (AL) of > 26.0 mm; (2) diagnosis of active subfoveal, juxtafoveal, or extrafoveal CNV secondary to pathologic myopia confirmed by complete ocular examination; (3) a minimum follow-up duration of 3 months. Patients who had undergone macular surgery, such as peeling of epiretinal membrane (ERM) and/or the internal limiting membrane (ILM), or photodynamic therapy, were excluded. Patients with other diseases causing retinoschisis or cystoid macular edema, such as retinal vein occlusion, age-related macular degeneration, and uveitis, were also excluded.

All patients underwent a complete ophthalmic examination, which included noncycloplegic refraction assessment using an autorefractometer (KR-800TM; Topcon, Tokyo, Japan), measurement of the best corrected visual acuity (BCVA; in Snellen visual acuity ratios and in logarithms of the minimum angle of resolution [logMAR] when used for statistical analyses), slit-lamp biomicroscopy, indirect ophthalmoscopy, AL measurement using the Lenstar™ (LS 900; Haag-Streit, Mason, OH, USA), color fundus photography, spectral-domain-OCT (SD-OCT; RTVue XR or 100; Optovue, Fremont, CA, USA), and fluorescein angiography (HRA 2; Heidelberg Engineering, Heidelberg, Germany). The horizontal and vertical scans of OCT were used for serial follow-up and comparison. Relevant OCT parameters, such as vitreomacular traction (VMT), vitreomacular adhesion (VMA), the presence of ERM, foveoschisis, extrafoveoschisis, inner retinal schisis, outer retinal schisis, lamellar macular hole (LMH), FD, central foveal thickness (CFT), FTMH, dome-shaped macula, and the location of CNV (subfoveal, juxtafoveal or extrafoveal), were documented for analysis. CFTs were generated automatically by the OCT device. In this study, splitting of the inner and outer retinal layers resulting in intraretinal, hyporeflective cystoid spaces on the OCT is identified as MRS, and splitting within the outer plexiform layer and the outer nuclear layer is defined as outer retinal schisis. Schisis cavities at the level of the inner plexiform, ganglion cell layer, and retinal nerve fiber layer are defined as inner retinal schisis [20, 23, 24]. Progression of MRS is defined as the development of an inner LMH, FD, or FTMH, or an increase of > 100 μm in the height of retinoschisis and an enlargement of the retinoschisis towards the area not having a macular retinoschisis [12]. Onset of MRS is defined as the absence of MRS at baseline but with the detection of signs of MRS during follow-ups. Improvement of MRS is defined as a decrease in the height or extent of the macular retinoschisis without the development of an inner lamellar MH, FD, or FTMH. Myopic maculopathy is classified into five categories: 0, indicating no myopic retinal degenerative lesion; 1, indicating tessellated fundus; 2, indicating diffuse chorioretinal atrophy; 3, indicating patchy chorioretinal atrophy; and 4, indicating macular atrophy [25]. The images were reviewed and analyzed independently by two of the authors (MCT and YTH). In cases of doubt or disagreement, the senior author (TCH) was consulted, and a panel discussion was held to reach a consensus.

This study was approved by the Research Ethics Committee of National Taiwan University Hospital and was conducted in accordance with the Declaration of Helsinki of 1964 and its later amendments.

Statistical Analysis

Means and standard deviations (SD) of quantitative variables and frequencies and percentages of categorical variables were calculated. Continuous variables were compared by using the t-test or Mann–Whitney U-test, and categorical variables were analyzed by using the Chi-square test or Fisher’s exact test. Logistic regression was used for investigating factors relevant to progression and onset of MRS. p < 0.05 was considered to be statistically significant. The statistical analysis of the data was carried out by using Statistical Package for Social Sciences (SPSS) version 22 software (SPSS IBM, Armonk, NY, USA). Descriptive statistics was calculated in terms of the mean, SD, median, and range.

Results

A total of 295 eyes with mCNV from 254 patients were included in this study. The mean (± SD) age of the patients was 57.4 ± 13.3 years (range 19–92; median 58 years) at the diagnosis of mCNV. The mean AL was 29.4 ± 6 mm. Prevalence of MRS in mCNV eyes was 25.4%. Of the 295 mCNV eyes, 95 eyes (32.2%) developed progression or onset of MRS at 11.6 ± 14.8 months (range 1–60 months, median 5 months) after the initiation of IV anti-VEGF treatment. The cumulative incidences of progression and onset of MRS were 15.2% at 3 months, 19.3% at 6 months, 23.4% at 1 year, 28.5% at 2 years, 29.2% at 3 years, and 32.2% at 5 years. Eyes with progression or onset of MRS, compared to those without progression or onset of MRS, had a higher ratio of presence of MRS (63.2% vs. 9.5%, p < 0.0001), foveoschisis (36.8% vs. 4.5%, p < 0.0001), extrafoveoschisis (61.1% vs. 10.0%, p < 0.0001), inner retinal schisis (49.5% vs. 9.0%, p < 0.0001), outer retinal schisis (58.9% vs. 6.5%, p < 0.0001), and LMH (16.8% vs. 1.5%, p < 0.0001) at baseline, more MRS in the fellow eye (39.1% vs. 16.1%, p < 0.0001), and greater initial CFT (322.4 ± 89.3 vs. 301.7 ± 77.6 µm, p = 0.047). For multivariate logistic regression analysis adjusted for foveoschisis, extrafoveoschisis, inner and outer retinal schisis, and LMH at baseline, only outer retinal schisis (β = 8.586, p = 0.003) and LMH (β = 5.015, p = 0.043) were identified as variables associated with progression or new onset of MRS (Table 1).

Table 1 Risk factors for progression and onset of myopic macular retinoschisis
Full size table

Of the 295 eyes, 13 showed progression, with 15.4% of these progressing to FTMH (1 case is shown in Fig. 1), 15.4% to FD (1 case is shown in Fig. 2), 30.8% to macular hole retinal detachment (MHRD), 7.7% to LMH, and 30.8% to other conditions at risk of vision threatening, such as foveoschisis shaped like a champagne flute [12, 13, 26] (1 case is shown in Fig. 3). Consequently, these eyes received surgical intervention at 12.7 ± 9.1 months (range 2–34 months, median 10 months) after initial injection. (Table 2). One eye (Fig. 3) received IV perfluoropropane (C3F8) [27, 28] and the other 12 eyes received pars plana vitrectomy combined with the inverted ILM flap technique [29, 30] (5 eyes; 1 case is shown in Fig. 1), the center non-peeling ILM surgery technique [31,32,33,34] (3 eyes), or simple ILM peeling (4 eyes; 1 case is shown in Fig. 2). Of the 13 eyes with progression, 11 (84.6%) had MRS at baseline. Risk factors for vision-threatening MRS requiring surgical intervention in eyes with progression and onset of MRS were evaluated by using logistic regression models (Table 3). AL, AL of fellow eye, initial CFT, myopic maculopathy category at baseline, and presence of foveoschisis, inner retinal schisis, ERM, and LMH at baseline were identified to be risk factors, while intravitreal ranibizumab (IVR) and presence of retinal hemorrhage at baseline were protective factors. Eyes requiring surgical intervention compared to those with no need of surgical intervention had poorer BCVA at 1 month after injection (logMAR 0.757 ± 0.363 vs. logMAR 0.510 ± 0.410, p = 0.026), at 6 months (logMAR 0.799 ± 0.402 vs. logMAR 0.521 ± 0.451, p = 0.011), at 1 year (logMAR 0.882 ± 0.528 vs. logMAR 0.530 ± 0.413, p = 0.012), and at the last follow-up visit (logMAR 1.038 ± 0.767 vs. logMAR 0.667 ± 0.624, p = 0.030), and had greater CFT at 1 month (370.7 ± 147.5 vs. 274.9 ± 57.9 µm, p = 0.009), at 3 months (353.4 ± 129.9 vs. 268.8 ± 56.6 µm, p = 0.003), and at 6 months (415.8 ± 196.5 vs. 274.5 ± 52.1 µm, p = 0.001). In addition, the former had higher ratios of categories 3 and 4 myopic maculopathy at the last follow-up visit. We also found that there was no association of age, gender, initial BCVA, location of CNV, the number of total injections, the number of intravitreal bevacizumab (IVB), IVR, and intravitreal aflibercept (IVA) injections, as well as presence of lacquer cracks, punctate inner choroidopathy (PIC), staphyloma, MRS, outer retinal schisis, VMT, VMA, and dome-shaped macula at baseline (p > 0.05 for all) (Table 3). In eyes with MRS progression requiring surgery, the progression was found in the following ratios: 53.8% in the outer retinal layers, 23.1% presence of LMH, 7.7% in both the outer and inner retinal layers, 7.7% in inner retinal layers, and 7.7% presence of FTMH.

Fig. 1
figure 1

Progression of myopic macular retinoschisis (MRS) in the right eye of a 56-year-old woman with myopic choroidal neovascularization (mCNV). a, b Before anti-vascular endothelial growth factor (anti-VEGF) treatment, hyperfluoresence, and leakage of mCNV on late phase of the fluorescein angiography (a) and category 2 myopic maculopathy on color photo (b) are shown. c 7 years after operation, macular atrophy (category 4 myopic maculopathy) was noted on color photo. d, e Before treatment, myopic macular retinoschisis (MRS) with lamellar macular hole (LMH) and juxtafoveal mCNV (arrow) are shown on optical computed tomography (OCT). f 2 years after intravitreal bevacizumab (IVB) treatment, the size of mCNV became smaller (arrow), and best-corrected visual acuity (BCVA) was 20/40. g 32 months after IVB treatment, full-thickness macular hole (FTMH) and new mCNV (arrow) were noted on OCT, and BCVA was 20/200. h 2 years after pars plana vitrectomy and the inverted internal limiting membrane (ILM) flap technique, the FTMH sealed. i 7 years after surgical intervention, the FTMH remained sealed but there were progressive chorioretinal atrophy. BCVA was 20/400

Full size image
Fig. 2
figure 2

Progression of MRS in the right eye of a 53-year-old woman with mCNV. a Before anti-VEGF treatment, subfoveal mCNV (arrow) and outer retinal schisis were shown on OCT. BCVA was 20/100. b 2 months after 2 injections of IVB, mCNV became smaller (arrow), BCVA improved to 20/40, and MRS remained stationary. c 6 months after injections, MRS progression to foveal detachment (FD) (asterisk) was observed, and BCVA decreased to 20/67. d 6 months after pars plana vitrectomy and ILM peeling, FD and MRS improved. e 1 year after operation, resolution of FD and improvement of MRS were noted on OCT. f 2 years after operation, attached retina without MRS was observed and BCVA was 20/50

Full size image
Fig. 3
figure 3

Progression of MRS in the right eye of a 44-year-old woman with mCNV. a Before treatment, subfoveal mCNV (arrow), inner and outer retinal schisis, and posterior hyaloid membrane (arrowheads) were observed on OCT. BCVA was 20/50. b 3 months after 2 injections of IVR, mCNV became smaller (arrow) and inner and outer retinal schisis slightly increased. c 6 months after IVR, progression of outer retinal schisis to foveoschisis shaped like a champagne flute was observed on OCT. BCVA decreased to 20/100. d, e 2 weeks (d) and 1 year (e) after intravitreal injection of C3F8 0.3 ml, MRS improved. BCVA returned to 20/29 at 1 year

Full size image
Table 2 Characteristics of each eye requiring surgical intervention in this study
Full size table
Table 3 Risk factors for vision-threatening myopic macular retinoschisis requiring surgical intervention
Full size table

Table 4 shows the characteristics of eyes with or without MRS at baseline. Compared to eyes without MRS at baseline, eyes with MRS at baseline were associated with older age, fewer total injections of anti-VEGF, fewer injections of ranibizumab, longer AL, longer AL of the fellow eye, higher proportion of staphyloma, ERM (62.0% vs. 26.9%, p < 0.0001), LMH, MRS in the fellow eye at baseline, and greater CFT at baseline, at 1 month, at 3 months, at 6 months, at 1 year, and at the last follow-up visit.

Table 4 Characteristics of eyes with or without myopic macular retinoschisis at baseline
Full size table

During follow-up, the onset of MRS developed in 35 of the 216 eyes (16.2%) of high myopic eyes without MRS at baseline at 17.9 ± 17.2 months (range 1–60 months; median 12 months) after initiation of anti-VEGF treatment. New MRS developed more likely in the inner retinal layers (31.4%, 11/35), followed by in the outer retinal layers (25.7%, 9/35), in both the outer and inner retinal layers (20.0%, 7/35), presence of LMH (17.1%. 6/35), and presence of FTMH (5.7%, 2/35). Onset of MRS was not associated with age, gender, laterality, the total number of injections of anti-VEGF, the numbers of injections of bevacizumab, ranibizumab or aflibercept, the numbers of injections at 3 months, 6 months, 1 year, and yearly until 7 years, AL, initial BCVA, initial CFT, dome-shaped macula, BCVA at 3 months, 6 months, 1 year and yearly until 5 years, and BCVA at the last follow-up visit. However, eyes with onset of MRS had significantly longer duration of follow-up (62.3 ± 33.0 vs. 45.8 ± 34.1 months, p = 0.009) and better BCVA at 1 month (logMAR 0.467 ± 0.312 vs. logMAR 0.643 ± 0.536, p = 0.018). Of the 35 eyes (5.7%) with onset of MRS, two had MRS progression to a FTMH or MHRD and received surgical intervention at 12 and 14 months, respectively.

Of the 79 eyes with MRS at baseline, 60 eyes (75.9%) had MRS progression at 7.2 ± 10.1 months (range 1–48 months, median 3 months) after IV anti-VEGF. MRS progression was associated with male sex (odds ratio [OR] 9.000, p = 0.039), and outer retinal schisis at baseline (OR 5.250, p = 0.010). MRS progression was found in the following ratios: 48.3% in the outer retinal layers, 31.7% in both the outer and inner retinal layers, 13.3% presence of LMH, and 6.7% in the inner retinal layers. Of the 79 eyes, 11 required surgical intervention at 11.1 ± 8.6 months; the surgical indications were FTMH (9.1%), MHRD (27.3%), FD (18.2%), LMH (9.1%), and others (36.4%). Five of the 79 eyes (6.3%) had improvement of MRS at 9.6 ± 7.1 months (range 2–19 months, median 12 months) after IV anti-VEGF (Fig. 4). The height or extent of MRS on OCT decreased in four of the five eyes, while complete resolution was achieved in the other one eye. MRS improvement was first detected in the outer retinal layers (60.0%) and in both the outer and inner retinal layers (40.0%). ILM disruption was observed in 3 of the 5 eyes.

Fig. 4
figure 4

Improvement of MRS in the right eye of a 47-year-old woman. a, b Color photos before treatment (a) and 2 years after anti-VEGF treatment (b) showed macular atrophy (category 4 myopic maculopathy). c Before treatment, outer retinal schisis and suspected disruption of ILM (arrowhead) were shown on OCT. BCVA was 20/67. d 1 year after intravitreal bevacizumab and aflibercept, MRS remained stable. e 19 months after IV anti-VEGF treatment, improvement of MRS and suspected disruption of ILM (arrowhead) were observed on OCT. BCVA became 20/25. f 2 years after treatment, near total resolution of MRS and suspected disruption of ILM (arrowhead) were noted on OCT. BCVA was 20/22

Full size image

In this study, BCVA improved significantly from logMAR 0.785 ± 0.492 to logMAR 0.608 ± 0.505 at 1 month (p < 0.0001), to logMAR 0.568 ± 0.498 at 3 months (p < 0.0001), to logMAR 0.587 ± 0.502 at 6 months (p < 0.0001), to logMAR 0.617 ± 502 at 1 year (p < 0.0001), to logMAR 0.641 ± 0.549 at 2 years (p < 0.0001), and to logMAR 0.693 ± 0.622 at 3 years (p = 0.016) after anti-VEGF treatment.

Discussion

In this retrospective study, vitreoretinal interface changes, including progression and onset of MRS after IV anti-VEGF treatment for eyes with mCNV, were evaluated. Factors contributing to progression of MRS have been reported to be diverse and to include AL, macular chorioretinal atrophy, VMT, and intrinsic noncompliance of the ILM [6, 12, 23, 35, 36]. Whether IV anti-VEGF treatment for mCNV increases the risk of MRS incidence or progression remains controversial. Shimada et al. reported that 5.4% of 74 eyes that had IVB for the treatment of mCNV developed macular retinal detachment following treatment, and four eyes had foveoschisis around the CNV [12]. Huang et al. found progression of MRS in seven of the 11 patients with preexisting ERM and MRS, and no progression of MRS in the six contralateral eyes which had MRS and ERM but were not treated with IV anti-VEGF [21]. Lai et al. reported that two of the 37 eyes had progression of MRS and another one eye had FTMH after IVB or IVR [22]. Ceklic et al. observed new MRS in eight of 474 eyes (1.7%) in the outer retinal layers, and new MRS developed in seven eyes treated with IVR but only in one untreated fellow eye. None of the eight eyes had progression of MRS to a FTMH or FD during the follow-up period of 12 months [20]. Zhou et al. reported ten eyes had MRS progression and two eyes had new-onset MRS in 122 mCNV eyes treated with Conbercept [19].

In our study, we observed 94 of 295 eyes (31.9%) had progression or onset of MRS. Of 216 eyes, 35 (16.2%) without MRS at baseline developed MRS, and 60 of 79 eyes (75.9%) with MRS at baseline had progression of MRS. The rates of progression and incidence of MRS were higher than those previously reported. The reasons for our results might be that we did not include fellow eyes, which usually did not have pathological changes, and that with a better understanding of MRS, we could differentiate MRS from macular edema and recognize it earlier with OCT. In addition, both the average AL and the average duration of follow-up were relatively longer and the average age was older in our study compared to other studies. Comparison of the higher rate of progression of MRS in our study to the 11.6–69.0% reported in studies of natural course of MRS [12, 13] implies that IV anti-VEGF might play a role in changing the vitreoretinal interface. The injection procedure may cause vitreous disturbance and induce posterior vitreous detachment (PVD) [37]. The rapid influx of the drug may lead to vitreous perturbation [19]. The anti-VEGF drugs themselves may enter the space between the vitreous and the retina and mechanically influence the adhesion or traction of the vitreous on the retina. The drug fluid mixed with the vitreous body potentially change the density of vitreous body. Furthermore, shrinkage of the fibrovascular tissue induced by anti-VEGF agents may result in tractional forces and enhance the separation of different retinal layers, which may result in the incidence or progression of MRS. The risk factors for progression and new onset of MRS were outer retinal schisis and LMH at baseline in the multivariate logistic regression model, while the risk factors for progression of MRS were outer schisis and male sex using univariate logistic regression analysis. Outer retinal schisis represents disruption of attachment of outer processes of glial Müller cells from photoreceptors, leading to an unstable structure and an increase in the risk of MRS progression. The result is consistent with that of outer schisis, classified in stage 2 in the staging system for myopic traction maculopathy recently proposed by Parolini et al. [38], and might further progress into macular detachment (stage 3). LMH was found to be another risk factor for progression and onset of MRS in our study, and this echoes the report that LMH with MRS is more susceptible to anatomical progression [39]. In our study, male sex was identified as another risk factor for MRS progression in eyes with mCNV, which is a female-predominant disease [40]. Although no study has yet addressed the gender predilection for MRS progression, several studies have shown that male sex was associated with VMA [41,42,43,44], and Palacio et al. found that males had significantly larger area of VMA. The reasons for this may be related to a lower concentration of sodium hyaluronate in women, leading to earlier PVD [45]. It should be noted that MRS progression was first detected in the outer retinal layers while onset of MRS was more likely to occur in the inner retinal layers, and this again confirms the staging system that stage 1 indicates inner or inner-outer schisis and stage 2 indicates predominantly outer schisis [38].

The mechanisms of MRS improvement are also diverse. Polito et al. reported that a complete PVD occurred before the resolution of macular RD and suggested that the PVD contributed to resolution of MRS [46]. Shimada et al. suggested that rupture of ILM reduces traction and leads to resolution of MRS and that disruption of the ILM may occur when ILM is detached from the underlying retina and is stretched extensively [12]. Lai et al. reported spontaneous resolution of MRS in 8 eyes without evidence of PVD and ILM disruption [47]. In our study, 5 eyes (6.3%) with MRS at baseline had improvements of MRS, and ILM disruption was detected in 3 of the 5 eyes. This is, to the best of our knowledge, the first report of MRS improvement after IV anti-VEGF treatment for mCNV. Likewise, the higher rate of MRS improvement in our study compared to 3.9% reported in the study of natural course of MRS [12] indicates that IV anti-VEGF did play a role in changing the vitreoretinal interface. We suspect that the profibrotic effect of anti-VEGF may either increase tractional force causing the extremely-stretched ILM to break or consolidate the connections among different retinal layers. In either ways, the traction on the retina is released or decreased, leading to improvement of MRS [12].

The baseline prevalence of MRS was 26.8% in this study, which is also higher than the 6–18% reported in studies including fellow eyes without mCNV [20, 48,49,50], but is similar to the prevalence (23.8%) in a study of mCNV [19]. Our data also confirm that MRS was more prevalent in older patients with high myopia [5, 20] and that eyes with MRS at baseline had more staphyloma and greater AL [6, 7]. Patients with MRS at baseline were reported to have received more injections than the other patients without MRS (MRS [n = 15 eyes]: 5.8 ± 2.1 vs. non-MRS [n = 207 eyes]: 4.0 ± 2.9; p = 0.0001) [20]. However, in our study, eyes with MRS at baseline received fewer total injections (p = 0.037), and fewer IVR (p = 0.038) than eyes without MRS at baseline. We attributed the treatment protocol to be the major contributor to this discrepancy. In the RADIANCE trial, two injections of ranibizumab were given in the first month and treatments were continued if there was a loss in visual acuity, intraretinal or subretinal fluid, or active mCNV. In early stages, MRS might be mistaken as intraretinal cysts or fluids, and injections were continually given. In addition, loss in visual acuity might be caused by progression of MRS. However, injections were repeatedly given until stable visual acuity was reached. In addition, the number of cases in our study was higher and we might have much experience of detecting MRS.

In this study, BCVA improved significantly at 1 month, 3 months, 6 months, and yearly until 3 years after treatment. This finding in turn suggests a beneficial effect of anti-VEGF treatment in eyes with mCNV despite the possibility of MRS progression or incidence.

Interestingly, IVR and presence of retinal hemorrhage at baseline were found to be protective factors of surgical intervention for vision-threatening MRS. Although no previous studies on the difference in the risk of MRS among different anti-VEGF agents are available, we suspected that the different fibrotic effects of anti-VEGF agents are associated with different risks of progression to vision-threatening MRS. Gillies et al. reported no significant differences in the rate of development or growth of macular atrophy over 24 months between ranibizumab and aflibercept in patients with neovascular age-related macular degeneration (nAMD) treated using an treat-and-extend regimen [51]. However, some studies have shown that the CNV in ranibizumab-treated nAMD patients demonstrated less fibrosis on OCT [52] and that ranibizumab prevented epidural fibrosis in a postlaminectomy rat model [53]. The result that IVR was a protective factor of vision-threatening MRS requiring surgical intervention in our study supports the notion that ranibizumab might have less fibrotic effect. Further studies with long-term treatment using a single anti-VEGF agent are needed to accurately identify the effect of different agents on the development of vision-threatening MRS. On the other hand, blood components have long been used to improve the hole closure rate for macular hole [54,55,56,57]. Blood contains many growth factors that promote wound healing [58] and cell migration [59]. Therefore, retinal hemorrhage may play a role in cell migration and healing in MRS and may help consolidate the retinal structure. Eyes with onset of MRS had better BCVA at 1 month after IV anti-VEGF treatment, and we suspect that these eyes responded more positively to anti-VEGF agents. Thus, the profibrotic effects of anti-VEGF may be more obvious in these eyes. The stronger the profibrotic effects induced by anti-VEGF agents, the greater the effect of vitreoretinal interface changes developing.

Our study is limited by the retrospective nature of the study and the various treatment regimens of anti-VEGF. The lack of control group is another limitation. The optimal control group (eyes with mCNV without anti-VEGF treatment) was unavailable. Instead, we compared the data in our study to data from other studies on natural course of MRS, which is not an ideal solution. Nevertheless, this is the largest real-world study to evaluate vitreoretinal interface changes after anti-VEGF treatment for mCNV. A number of eyes (16.2%) without MRS at baseline developed new MRS, and 75.9% of eyes with MRS at baseline had progression of MRS. Furthermore, improvement of MRS was first reported after anti-VEGF treatment. Risk factors of progression and onset of MRS were outer retinal schisis and LMH while risk factors for MRS progression were outer schisis and male sex. IVR and presence of retinal hemorrhage were protective factors for vision-threatening MRS requiring surgical intervention. An understanding of the vitreoretinal interface changes after anti-VEGF treatment for eyes with mCNV may help in guiding follow-up schedule, surgical planning, and management.