Postoperative Outcomes and Prognostic Factors of High Myopia-Epiretinal Membrane Associated with Retinoschisis after Vitrectomy Open Access

High myopia (HM) is defined as myopia with ametropia <−6.00 diopters with or without axial length (AXL) exceeding 26 mm. With the increase in AXL, the risk of macular diseases gradually escalates, such as epiretinal membrane (ERM), macular holes, retinoschisis, choroidal neovascularization, and choroidal retinal atrophy, potentially resulting in blindness [1, 2]. The ERM is a common vitreoretinal interface disease, which is characterized by fibrous tissue formed by the proliferation of glial cells, fibrocytes, hyalocytes, macrophages, inflammatory cells, and vascular components on the surface of the internal limiting membrane (ILM) through the gap of ILM [3, 4].

Recent studies have reported an association between HM and ERM. The long AXL is related to an early occurrence of posterior vitreous detachment, which is a risk factor for ERM. Meanwhile, the longer the AXL, the greater the vitreoretinal traction. So, it is easier to form ERM in HM eyes [5, 6]. Therefore, the increase in myopia and the elongation of the AXL are positively correlated with the incidence of ERM [4, 5, 7]. The anterior tractional forces from ERM and the posterior forces from elongated AXL are significant factors contributing to the incidence of retinoschisis [8‒10].

Although the surgical treatment for idiopathic ERM is common, the postoperative outcomes in HM eyes with ERM were rarely reported. For idiopathic ERM patients with obvious symptoms, pars plana vitrectomy (PPV) with ERM peeling is the first choice which has great effect in visual and anatomical recovery [6, 11, 12]. Previous studies have shown that ERM removal produces similar anatomical and visual results in both high myopic and non-high myopic eyes. However, in these studies, all high myopia-epiretinal membrane (HM-ERM) eyes exhibited only mild or moderate myopic retinal degeneration, with no significant alterations in macular foveal structure observed on OCT, such as foveoschisis or foveal retinal detachment [11‒13]. Once HM patients develop retinal detachment and foveoschisis, it can adversely affect their visual function, leading to metamorphopsia and visual impairment.

Thus, the purpose of the present study was to analyze the effects of retinoschisis on the anatomical and visual outcomes in HM-ERM patients undergoing PPV and ERM peeling. Identifying prognostic factors for these patients may assist in more effectively determining the timing of surgical intervention.

This retrospective, consecutive case series study was conducted in Beijing Tongren Hospital and was in accordance with the Declaration of Helsinki and its later amendments. Patients in this study were recruited from May 2012 to November 2021, and all signed an informed consent form.The inclusion criterion was that the patients had been clinically diagnosed with HM with unilateral or bilateral ERM and underwent surgical removal with a minimum follow-up of 1 month. The exclusion criteria were as follows: (1) eyes with secondary ERM, such as diabetic retinopathy, venous occlusion, retinal detachment, uveitis, or trauma; (2) eyes with macular hole or anterior abnormality with refractive stroma turbidity affecting SD-OCT imaging; (3) eyes with history of vitreoretinal surgery or intraocular injection; (4) eyes with any other conditions that can cause visual impairment, such as amblyopia, severe glaucoma, optic neuritis, and so on.

Every case in this study recorded the baseline characteristics, including age, gender, symptom duration and surgical procedures. In addition, the clinical ophthalmic characteristics were also recorded, incorporating best corrected visual acuity (BCVA), AXL, intraocular pressure (IOP), lens status, slit-lamp examination, fundus photographs, an d spectral-domain OCT (SD-OCT).

Preoperative and postoperative foveal microstructures were evaluated by SD-OCT (Cirrus high-definition OCT; Carl Zeiss Meditec, Dublin, CA, USA), using Macular Cube 512*128 protocol and the HD 5-line raster. Intraretinal splitting of the inner and outer retinal layers resulting in intraretinal, hyporeflective cystoid spaces on OCT was identified as macular retinoschisis (MRS). These HM-ERMs were divided into four groups according to the classification by Ceklic [14, 15] based on the layer involvement of MRS (shown in Fig. 1): (A) without retinoschisis; (B) with outer retinoschisis; (C) with inner retinoschisis; (D) with both inner and outer retinoschisis. Inner retinoschisis can occur in inner nuclear layer, inner plexiform layer, ganglion cell layer, and retinal fiber layer. Outer retinoschisis can occur in outer plexiform layer and outer nuclear layer. Besides, OCT images record central fovea thinkness (CFT) which was calculated automatically as the mean value within the innermost central 1 mm circle, according to the Early Treatment Diabetic Retinopathy Study (ETDRS) sectors. The presence of lamller macular hole (LMH), foveal retinal detachment, epiretinal proliferation (EP) were also recorded. EP is a result of neuroglial cell proliferation following retinal stimulation [16]. The integrity of the ellipsoid zone (EZ) and interdigitation zone (IZ) at the fovea was additionally evaluated based on SD-OCT images. A completely invisible or discontinuous EZ or IZ was defined as a loss of integrity.

All patients underwent a standard 23-G, 3-port vitrectomy with ERM and ILM peeling by one experienced surgeon (W.L.). All cases included were immature cataracts that did not significantly affect vision. However, phacoemulsification cataract surgery was performed in phakic eyes first because the degree of cataract was already impacting their vision or affecting surgical procedures, in order to achieve better visual outcomes postoperatively. After core vitrectomy, the posterior vitreous detachment was created if it was not present. The surgeon peeled all visible ERM and then peeled ILM up to the vascular arcades with 0.25% indocyanine green staining the ILM. Laser photocoagulation was applied any retinal degeneration and breaks were observed. At the final step, fluid-air exchange was performed, then these eyes were filled with sterile air or 0.5 mL of 14% perfluoropropane (C3F8) gas. The vitreous cavity was filled with C3F8 in eyes with wide-field retinoschisis, foveal retinal detachment, or lamellar macular hole. The patients filled with sterile air maintained a face-down position for 5 days, while patients filled with C3F8 for 7 days. Except for necessary breaks such as meals and eye drops, patients were asked to keep a face-down posture during the day and sleep at night.

Continuous variables were indicated as mean ± standard deviation. These variables were analyzed using Fisher’s exact for categorical data and Kruskal-Wallis H Test for nonnormal distribution among the 4 groups to detect the differences, and post hoc pairwise comparison was utilized with the Bonferroni correction. ANOVA was used for normal distribution. The variation between the initial and final BCVA was evaluated using the Wilcoxon signed-rank test. And the paired t-test was used to assess the statistical distinction between the primal and the final CFT. Pearson correlation analysis for normal distribution and Spearman correlation analysis for unnormal data were used to estimate the factors that affected the BCVA at the last visit. A Kaplan-Meier survival probability curve was plotted to estimate the resolution of retinoschisis after PPV. A p < 0.05 was considered statistically significant.

Seventy-three highly myopia patients (76 eyes) diagnosed with ERM (27 male patients, and 46 female patients) were enrolled in the study. According to the preoperative OCT findings, we found that the outer retinoschisis predominantly occurred at the central fovea, while the inner retinoschisis was located outside the central fovea. At the time of surgery, 12 were already pseudophakic and 1 was aphakia. In the other 63 phakic eyes, 48 eyes (63.2%) underwent vitrectomy combined with phacoemulsification cataract surgery. During surgery, C3F8 was used to fill 6 eyes, while sterile air was used in 70 eyes.

The detailed comparison of the baseline characteristics among the four groups were summarized in Table 1. Both the eyes in group A and in group C had a significantly less presence of LMH (p_AB < 0.05, p_AD < 0.05, p_CB < 0.05, p_CD < 0.05) and a shorter AXL (p_AB = 0.046, p_AD = 0.005, p_BC = 0.015, p_CD = 0.001) than that in group B and group D. Moreover, the presence of macular fovea detachment in group C was significantly less than group D (p_CD < 0.05). The CFT of eyes in group A was significantly thinner than in group C and group D (p_AC = 0.009, 95% confidence interval [CI], 22.87, 156.61; p_AD = 0.035, CI, 5.81, 150.06).

We confirmed via OCT that the ERM was completely removed during the postoperative period. Out of 59 eyes with retinoschisis, 39 cases (66.1%) showed complete disappearance of schisis after surgery, including 11 eyes (52.4%) in group B, 20 eyes (90.9%) in group C, and 8 eyes (50%) in group D. At the final follow-up, one eye in group A still had the schisis of the retinal nerve fiber layer, and 4 eyes in group B had recurrences of inner retinoschisis, 3 of which showed the disappearance of outer retinoschisis. In group D, 5 eyes only had outer retinoschisis at the last follow-up time, while the other 3 eyes continued to present with both inner and outer retinoschisis. At the final follow-up visit, there was no significant difference in preoperative and postoperative BCVA between schisis-recovered eyes and non-recovered eyes in every group (p_B-pre = 0.659, p_B-post = 0.797, p_C-Pre = 1.00, p_C-post = 0.526, p_D-pre = 0.548, p_D-post = 0.278).

Furthermore, Kaplan-Meier survival curve analysis (Fig. 2) shows that the resolution of retinoschisis among groups B, C, and D has statistically significant differences (χ² = 17.59, p < 0.001). Additionally, pairwise comparisons of the survival time distributions among the three groups were conducted, with the significance level adjusted using the Bonferroni correction, resulting in an adjusted significance level (α) of 0.0167. It was found that there were statistically significant differences between group B and group C and also between group C and group D (χ² = 13.32, p_BC < 0.001; χ² = 11.261, p_CD = 0.001; χ² = 0.003, p_BD = 0.956).

Among all patients, the mean CFT showed a significant reduction from a baseline of 476.00 ± 15.79 μm to 318.74 ± 9.80 μm at the final follow-up examination (p < 0.001). Postoperative measurements of CFT for each group are detailed in Table 2.

Of the 43 eyes with a disrupted EZ line at baseline, 11 eyes (25.6%) recovered their continuity at the last follow-up examination, including 4 eyes (36.4%) in group A, 3 eyes (27.3%) in group B, 3 eyes (27.3%) in group C, and 1 eyes (9.0%) in group D. No statistically significant difference was observed at the final follow-up among the four groups in terms of the recovery of the EZ (p = 0.896).

The mean BCVA improved to 0.3 ± 0.3 LogMAR (Snellen: 20/40) at the final follow-up, which is significantly better than that at baseline (p < 0.001, CI, 0.22, 0.39). The changes in BCVA after surgery for each group are shown in Figure 3 and Table 3. At the final follow-up, BCVA was better than 20/40 in 59 eyes (A = 16, B = 16, C = 15, D = 12), and BCVA was 20/25 or better in 29 eyes (A = 9, B = 8, C = 6, D = 6).

After conducting a logistic regression analysis, we found no association between retinoschisis resolution and various baseline factors, including course of disease(p = 0.088), layer of retinoschisis (p = 0.451), AXL (p = 0.471), preoperative CFT (p = 0.069), preoperative presence of LMH (p = 0.249), preoperative presence of central foveal detachment (p = 0.067), preoperative presence of EP (p = 0.601), and preoperative visual acuity (p = 0.361).

Based on binary logistic regression analysis, a worse baseline BCVA (p = 0.02, CI, 0.004, 0.619) and the presence of LMH (p = 0.028, CI, 0.067, 0.856) were significantly associated with improved BCVA following surgery. Correlation analysis indicated that the presence of EP (p = 0.014), baseline BCVA (p = 0.001), and the postoperative continuity of EZ (p = 0.001) and IZ (p = 0.008) were related factors for the final BCVA. Among them, only preoperative BCVA showed a positive correlation with the BCVA at the last follow-up. Furthermore, neither the resolution of the retinoschisis (p = 0.394) nor the type of retinoschisis (p = 0.073) was associated with the final BCVA outcome.

This study assessed the postoperative anatomical and visual outcomes of 76 eyes diagnosed with HM-ERM. Eyes were categorized into four groups based on the affected layer of retinoschisis. Most of eyes achieved good visual and anatomical results after surgery. Among those with HM-ERM, patients with outer retinoschisis demonstrated a significantly longer AXL. There is no significant difference in the resolution of retinoschisis among the 3 groups.

This study focused on eyes with an AXL exceeding 26 mm, where the prevalence of ERM was significantly higher due to posterior vitreous detachment and vitreoretinal traction [7, 15, 17]. ERM is recognized as a risk factor for poor visual outcomes in eyes with HM [15, 18]. The study observed improved vision after surgery in eyes with HM-ERM and retinoschisis, but no significant difference in BCVA between those with and without retinoschisis. The relationship between retinoschisis and vision is inconclusive [8, 10, 19]. However, progression of retinoschisis can lead to macular holes or RD, with a high risk of nearly 50% within a few years, causing severe vision loss [8, 18, 19].

The development of retinoschisis is propelled by several opposing forces: anterior tractional forces from the ERM pulling on the macular region, vitreomacular traction, and retinal arteriolar traction; posterior forces including posterior staphyloma and elongated AXL. The combined force leads to the separation of the retinal neuroepithelial layer in the central foveal area, termed myopic MRS [8]. Shinohara et al. [9] suggested that in outer retinoschisis, the main factor was the presence of posterior tractional forces, such as posterior staphyloma, whereas in inner retinoschisis, the schisis typically occurred at the level of the inner plexiform layer and were associated with anterior tractional forces, such as ERM, vitreous traction, and vascular microfolds.

Ceklic et al. [15] observed 8 eyes with newly occurrence of MRS in their follow-up eyes. They found that in 2 eyes with inner foveoschisis, new outer foveoschisis appeared over time at the sites previously affected by the inner foveoschisis. In our study, we observed that eyes in group C had shorter AXLs, corresponding to reduced vitreoretinal traction, resulting in a lower incidence of lamellar holes and foveal retinal detachment. Additionally, in group C, no postoperative occurrences of outer retinal schisis were observed; only two eyes exhibited incomplete resolution of schisis, although this was with a reduced extent and severity compared to the preoperative state. In group D, schisis remained unresolved in seven eyes, five of which continued to present with outer retinoschisis postoperatively. The impact of axial elongation is relatively minor in eyes with shorter AXLs, primarily due to damage caused by ERM contraction. Once the traction between the vitreoretinal interface, including the vitreous, ERM, and ILM, is removed, such eyes are more likely to experience inner retinoschisis, making postoperative schisis recovery more achievable. In contrast, in eyes with longer AXLs, in addition to the abnormal vitreoretinal interface, factors, such as longer AXLs, posterior staphyloma can also affect the retinal structure.

In this study, postoperative visual acuity improved overall, but declined in 5 eyes – 3 in group A and 2 in group B. In group A, two eyes experienced final visual acuity declines of 0.05 and 0.1, respectively, and were lost to follow-up 1 month and 4 months after surgery. The third one did not undergo combined cataract surgery. Based on OCT and medical records, we believe that the progression of cataracts following PPV was the cause of the decrease in visual acuity to preoperative levels during the 10-month follow-up. To eliminate cataracts as a confounding factor, we only included this eye’s follow-up data up to 4 months postoperatively in our analysis. Fujimoto et al. [1] observed that postoperative improvements in BCVA were closely correlated with the recovery of photoreceptor cells, which gradually regained function as CFT reduced. Lehmann et al. [20] proposed that while CFT dropped notably 1 month postoperatively, BCVA only significantly improved by 3 months after PPV, indicating that the recovery of visual function might be delayed relative to the achievement of anatomical improvements. In group B, final follow-up showed a decrease in BCVA for two eyes. One eye had a preoperative LMH and EP, leading to outer retinoschisis and EZ disruption at the final check. The other eye continued to have EP and LMH, with added retinal atrophy and EZ disruption noted at the last follow-up [1]. Our study also found that eyes with intact EZ band which was the inner and outer segment of the photoreceptors had better preoperative and postoperative BCVA at each follow-up point.

In the current study, postoperative BCVA influencing factors included integrity of the EZ at final follow-up, integrity of the IZ at final follow-up, preoperative BCVA, and presence of EP. Patients with better preoperative BCVA tend to have better BCVA at their last follow-up, suggesting that early surgical intervention is a preferable option for this group of patients with visual impairment. The first three factors are consistent with previous findings in studies on idiopathic ERM and MRS [6, 12]. Thus, similar to many other macular diseases underwent PPV, ERM should be removed in the early stages of visual impairment in patients to achieve better final visual acuity. However, the impact of EP remains controversial. Omoto et al. [21] study on patients with LMHs and epiretinal membrane foveoschisis demonstrated no significant correlation between both baseline and postoperative BCVA and the presence of EP. Hetzel et al. [22] also discovered that in patients with epiretinal membrane foveoschisis, EP was not a factor that affected BCVA. However, Choi et al. [23] suggested that postoperative VA recovery was better in patients without EP compared to those with EP in lamellar macular hole patients. Our current results are consistent with the latter, indicating that the presence of EP is unrelated to preoperative visual acuity, but final BCVA is significantly correlated with the presence of EP, with eyes having EP preoperatively showing poorer visual acuity at final follow-up.

This study has several limitations. First, the study was limited by its retrospective design, which poses certain limitations in establishing clear causal relationships. And there are inherent limitations, including potential selection bias and information bias. Second, its retrospective design resulted in an inconsistent samples size among the four groups, but within a comparable range. The patients enrolled in this study had different follow-up times. To verify the comparability of the groups, we categorized the eyes into two groups: one with a follow-up period of 4 months or less, and the other with a follow-up period exceeding 4 months. The analysis revealed that the only parameter in which a statistically significant difference was observed between the two groups was gender (χ² = 7.856, p = 0.005 < 0.05). No other preoperative or postoperative parameters exhibited statistically significant differences. Additionally, not all enrolled cases underwent cataract surgery. In the comparison between patients who underwent cataract surgery and those who did not, we found only a statistical difference in the follow-up time (10.0 ± 1.3 months vs. 16.7 ± 1.8 months, p = 0.007 < 0.05) with no statistical differences in other postoperative anatomical parameters and visual acuity. Therefore, patients who did not undergo cataract surgery may not significantly influence the results of this study. Finally, in future research, we need to employ a prospective design to more accurately determine the causal relationships related to our research questions and mitigate the impact of biases and confounding factors.

In summary, we focused on patients with HM-ERM accompanied by retinoschisis, for which PPV could provide satisfactory anatomical and functional outcomes. PPV with ERM peeling is beneficial for the disappearance of retinoschisis, especially for inner retinoschisis. The presence of retinoschisis did not affect the final outcome for HM-ERM eyes. Preoperative BCVA and disruption of the EZ and IZ were the main factors influencing final BCVA. Therefore, surgery should still be considered when visual impairment is mild in patients with HM-ERM to control disease progression and achieve better BCVA. Our research suggested that patients with EP have poorer final BCVA. Whether the presence of EP affects postoperative VA in patients with HM-ERM requires larger sample sizes and further prospective studies to validate our results.

All procedures conducted in this study, which involved human subjects, were in strict accordance with the ethical standards of the Helsinki Declaration and its subsequent amendments. This study protocol was reviewed and approved by the Ethical Review Committee of Beijing Tongren Hospital, Capital Medical University, Approval No. TREC2024-KYS076. Prior to their inclusion in the study, written informed consent was obtained from all participants. They also provided consent for the publication of their data and photographs. This research was conducted as a retrospective, observational study.

The authors have no conflicts of interest to declare.

This study was supported by National Natural Science Foundation of China, Grant/Award No. 8240040319. Sponsor: Xiaohan Yang.

Conceptualization, methodology, and original draft preparation: Xiao Feng and Xiaohan Yang; data collection and analysis: Xiao Feng and Xiaohan Yang; review and editing: Xiao Feng, Xiaohan Yang, Biying Qi, and Wu Liu; support of statistical techniques: Xijin Wu, Xinbo Wang, and Qinlang Jia; supervision and surgical support: Wu Liu; and approval of the manuscript: all authors.

Feng Xiao and Xiaohan Yang contributed equally to this paper and should be considered as joint 1st authors.

All data generated or analyzed during this study are included in this article. Further inquiries can be directed to the corresponding author.