Inverted Internal Limiting Membrane Flap versus Internal Limiting Membrane Peeling for <400 μm Macular Hole: A Meta-Analysis and Systematic Review

Macular hole (MH) is a condition characterized by a tissue defect that extends from the inner limiting membrane of the retina to the photoreceptor layer within the macular area. This condition often leads to significant visual decline, visual distortion, central scotoma, and other visual impairments. In 1988, Gass [1] proposed a theory suggesting that the development of a MH occurs due to a tangential traction force on the surface of the retina. According to this theory, to facilitate the closure of MHs, it is necessary to relieve the traction in the macular area. By alleviating this traction force, the retinal tissue is given the opportunity to heal and close the hole, leading to the retina recovery for patients. Since the study by Eckardt et al. [2] in 1997, where they first introduced internal limiting membrane (ILM) peeling as a treatment for full-thickness MHs, there has been a remarkable improvement in the closure rate of MHs. As a result, the combination of PPV with ILM peeling, along with gas-liquid exchange during the surgical procedure, has become the standard surgical approach for the management of MHs [3‒6].

The International Vitreomacular Traction Study Group has established a categorization system for MHs based on the size of the minimal diameter of the MH. According to this classification, MHs are categorized as follows: small (≤250 μm), medium (251–400 μm), and large (>400 μm) [7]. Michalewska et al. [8] introduced the ILM flap technique as an alternative approach to treat large full-thickness macular holes. This technique represents a refinement of the standard ILM peeling procedure. According to their findings, the closure rate achieved with the ILM flap technique was reported to be 76%, whereas traditional ILM peeling yielded a closure rate of only 52%. In a study by Kannan et al. [9], it was reported that patients with large MH treated using the ILM flap technique also showed favourable results in terms of visual acuity improvement. The use of the ILM flap technique has demonstrated beneficial outcomes for large, traumatic, myopic, and chronic full-thickness macular holes. Consequently, an increasing number of surgeons are opting for this operation. However, there has been an ongoing debate regarding the effectiveness of the ILM flap technique for small and medium-sized MHs. Some studies have suggested that traditional ILM peeling yields a high closure rate for small and medium-sized MHs, and implementing the inverted ILM flap technique can make the procedure more challenging. However, some more recent studies have indicated that the inverted ILM flap technique can achieve similar visual outcomes compared to traditional ILM peeling. Moreover, it has shown advantages such as faster visual acuity improvement and anatomical recovery during the early postoperative period [10‒12]. These findings suggest that the inverted ILM flap technique may offer potential benefits for small and medium-sized MHs.

In this meta-analysis, we conducted a comprehensive and quantitative comparison between PPV combined with the ILM flap technique and ILM peeling for the treatment of small and medium-sized MHs. We aimed to explore the optimal surgical approach for the treatment of less than 400 μm.

This meta-analysis is written according to Preferred Reporting Items for Systematic Reviews statement for reporting systematic reviews and meta-analyses (online suppl. material; for all online suppl. material, see https://doi.org/10.1159/000534873).

The search strategy employed appropriate keywords and medical subject headings to target relevant experimental studies published prior to January 2023. To ensure a comprehensive search, we utilized multiple reliable databases, including PubMed, Web of Science, Embase, and Cochrane Library. The search terms used were “macular hole” OR “MH” OR “macular break,” AND “inverted internal limiting membrane flap technique” OR “inverted ILM flap technique”, AND “internal limiting membrane peeling” OR “ILM peeling” OR “internal limiting membrane removal” OR “removing the ILM” OR “ILM peel”. We also searched manually to avoid any potential omission of relevant studies.

Based on the PICOS (Participants, Intervention, Comparison, Outcome, Study design) protocol, we have defined specific criteria as follows: (I) participants: individuals diagnosed with MH and having a minimum diameter of the hole less than 400 μm; (II) intervention measures: the experimental group received PPV combined with the ILM flap technique. And only single-layer flap covers the hole. The control group received PPV combined with ILM peeling; (III) outcome measures: the outcomes assessed included the rate of MH closure, preoperative and postoperative best corrected visual acuity, and postoperative assessment results of the external limiting membrane (ELM) and ellipsoid zone (EZ); (IV) study design: prospective randomized controlled trials (RCTs) and non-randomized concurrent control trials (n-RCTs) were included.

The exclusion criteria were as follows: (I) combination with severe cataract, glaucoma, myopia, retinal detachment, and other eye diseases; (II) did not provide the data required for this meta-analysis and did not yield such information upon request, as well as the original text of the literature obtained by the method; (III) follow-up time of fewer than 3 months; (IV) poor quality of literature, missing data, duplicate reports; (V) systematic reviews, meta-analysis, or case reports.

The data extraction process was carried out independently by two investigators. Any discrepancies that arose during the extraction were resolved through discussions between the two investigators or through consultation with a third researcher. The content extracted included essential information such as the study title, first author, publication year, country of origin, trial type, participant age, baseline best corrected visual acuity, follow-up time, and mean diameter of the MH.

The quality assessment of the RCTs included in the meta-analysis was performed using the “risk of bias” tool as recommended by the Cochrane Handbook 5.1.0. The RCTs were assessed across seven key domains to determine the risk of bias: “random sequence generation,” “allocation concealment,” “blinding of participants and personnel,” “blinding of outcome assessment,” “incomplete outcome data,” “selective reporting” and “other bias.”

The methodological quality of the n-RCTs included in the meta-analysis was assessed using the Newcastle-Ottawa Scale (NOS). The NOS evaluates studies based on three main areas: selection, comparability, and exposure. The studies that receive six or more stars are considered to have relatively high methodological quality [13].

The statistical analysis was conducted using Review Manager 5.3 software. The Cochrane test and I² statistics were utilized to assess the heterogeneity among the included studies. The I² statistic quantifies the degree of heterogeneity. The values ≥50% indicate strong heterogeneity. The values between 25% and 50% suggest moderate heterogeneity. The values between 0% and 25% indicate mild heterogeneity. And an I² value of 0% suggests no observed heterogeneity among the studies [14]. When the I² statistic exceeded 50% and indicated a significant level of heterogeneity among the included studies, a random-effects model was employed for data synthesis. Conversely, when there was no significant heterogeneity observed (I² ≤ 50%), a fixed-effects model was utilized for data synthesis.

The dichotomous variable data, specifically the effective rate of MH closure and the postoperative assessment results of ELM and EZ, were analysed using odds ratios (OR) along with their corresponding 95% confidence intervals (CIs). Concurrently, the preoperative and postoperative BCVA data were assessed using standardized mean differences accompanied by their 95% CI.

Statistical significance was determined at a significance level of p < 0.05. Sensitivity analysis was employed to assess the robustness and credibility of the meta-analysis findings. Forest plots were utilized as a graphical tool to visually present the results of the meta-analysis. In addition, funnel plots were applied to assess the potential publication bias [15].

Figure 1 presents an overview of the included literature. A total of 204 studies were initially retrieved for this meta-analysis. Through the screening of titles and abstracts, 175 records were excluded, leaving 29 articles for more detailed evaluation. Subsequently, a thorough assessment led to the inclusion of 5 studies [16‒20], encompassing 269 eyes, with 146 eyes allocated to the ILM peeling group and 123 eyes to the inverted ILM flap group. Of these studies, three [18‒20] adopted an n-RCT design, while the remaining two [16, 17] were designed as RCTs. A similar surgical technique was employed across all five studies, involving either 23-G or 25-G PPV with the utilization of indocyanine green for ILM staining. Following fluid-air exchange, a 20% sulphur hexafluoride tamponade was applied. To ensure optimal outcomes, patients were instructed to maintain a face-down positioning for a minimum of 48 h postoperatively. Table 1 provides a summary of the key characteristics of the included studies, offering insights into the various parameters examined in the meta-analysis.

Four studies [16‒18, 20] were included in the meta-analysis to compare the MH closure rates between the inverted ILM flap and ILM peeling groups. Only closed cases were selected for analysis in this study [19]. So, we did not include it. The inverted ILM flap group demonstrated a closure rate of 97.0% (96/99 eyes), while the ILM peeling group showed a closure rate of 99.0% (109/110 eyes). The analysis, using the fixed-effects model due to an I² value of 0%, revealed no statistically significant difference in the MH closure rate between the two groups (OR = 0.29, 95% CI: 0.04–1.96, p = 0.33, Fig. 2).

Among the four trials included in the meta-analysis, one trial [17] recorded distance-corrected visual acuity rather than preoperative and postoperative BCVA. Therefore, data from the remaining four trials [16, 18‒20] were subjected to analysis. Pooling the data from these four trials, no significant heterogeneity was observed for preoperative BCVA (I² = 0%, p = 0.94) and postoperative BCVA (I² = 13%, p = 0.33). Consequently, the fixed-effects model was utilized to combine the results and estimate the overall effect size for these outcome measures. The results revealed that there were no statistically significant differences in preoperative BCVA (weighted mean difference (WMD = 0.06, 95% CI: −0.20–0.31, p = 0.94, Fig. 3a) and postoperative BCVA (WMD = −0.09, 95% CI: −0.34–0.16, p = 0.33, Fig. 3b) between the inverted ILM flap and ILM peeling groups.

In four of the five included studies [16, 18‒20], the integrity of the ELM was assessed using spectral-domain optical coherence tomography. We compared the ELM integrity between the ILM flap and ILM peeling groups at the time points of 3 months and 12 months after surgery. These analyses aimed to evaluate any differences in the ELM preservation between the two treatment approaches for MHs. The heterogeneity observed in the meta-analysis was below the threshold of 50%, with an I² value of 46% at 3 months and 33% at 12 months. Therefore, the fixed-effects model was deemed appropriate for the analysis. At the 3-month follow-up, the integrity of the ELM was found to be intact in 94 out of 109 eyes (86.2%) in the inverted ILM flap group and in 127 out of 144 eyes (88.2%) in the ILM peeling group. At the 12-month follow-up, ELM integrity was restored in 102 out of 109 eyes (93.6%) in the inverted ILM flap group and in 134 out of 144 eyes (93.0%) in the ILM peeling group. The meta-analysis results indicated no statistically significant difference in the integrity of the ELM between patients who underwent the inverted ILM flap technique and those who underwent ILM peeling, both at the 3-month follow-up (OR = 0.88, 95% CI: 0.39–1.97, p = 0.16, Figure 4a) and the 12-month follow-up (OR = 0.96, 95% CI: 0.30–3.07, p = 0.21, Fig. 4b).

Among the four studies [16, 18‒20] that investigated the structure of the EZ, data were collected at 3 months and 12 months postoperatively. The heterogeneity analysis revealed an I² value of 27% after 3 months and 43% after 12 months, both below the predefined threshold of 50%. Thus, the fixed-effects model was utilized for data synthesis. After 3 months, the EZ remained intact in 32 out of 109 eyes (29.4%) in the inverted ILM flap group, while 53 out of 144 eyes (36.8%) in the ILM peeling group demonstrated intact EZ. At the 12-month follow-up, the EZ integrity was restored in 76 out of 109 eyes (69.7%) in the inverted ILM flap group and in 96 out of 144 eyes (66.7%) in the ILM peeling group. The meta-analysis results indicated no significant differences in postoperative EZ recovery between the inverted ILM flap and ILM peeling groups, both at 3 months (OR = 0.85, 95% CI: 0.47–1.53, p = 0.25, Figure 5a) and 12 months (OR = 1.39, 95% CI: 0.79–2.46, p = 0.15, Fig. 5b).

The two RCTs [16, 17] underwent a comprehensive evaluation using the risk of bias (RoB) tool recommended by the Cochrane collaboration. Due to the nature of the surgical intervention, blinding of the operators and patients was not feasible in these studies. However, potential sources of bias such as performance bias and selective bias were deemed to have minimal impact on the results. Notably, one study (Leisser et al. [17]) missed some follow-up data, indicating a low-quality trial. The remaining studies were considered to be of high quality, with only one item rated as having an “unclear risk of bias” (Fig. 6, 7). The quality assessment of the three non-RCTss [18‒20] was conducted using the Newcastle-Ottawa Scale (NOS). These studies received high-quality ratings, as they achieved a score of 6 stars or higher in the assessment, indicating robust methodological standards (Table 2).

The presence of heterogeneity among the included studies was acknowledged in the analysis. To ensure the robustness and stability of the results, a sensitivity analysis was conducted by excluding studies with the highest and lowest weights. Notably, it was observed that the rate of MH closure and the visual acuity before and after surgery did not exhibit significant heterogeneity, hence precluding the need for a sensitivity analysis for these outcomes. In the case of the postoperative 3-month and 12-month assessment of the ELM and EZ, each study was individually excluded in the sensitivity analysis. However, it was found that the exclusion of any specific study did not result in changes in the overall heterogeneity. In terms of the sensitivity analysis, it can be concluded that all the studies included in our meta-analysis are deemed reliable.

Publication bias was assessed using funnel plots. In our analysis, no statistically significant evidence of publication bias was detected, indicating that the included studies were representative and comprehensive in addressing the research question.

With the development of knowledge and treatment options for MH, there has been a notable improvement in the outcomes of MH surgery. Among these medical interventions, the inverted ILM flap technique has emerged as a new effective approach, particularly in the management of large, complex, and myopic MHs [9, 21‒24]. Several meta-analyses and systematic reviews have highlighted the advantages associated with the inverted ILM flap technique in these specific MH cases [22‒24]. Inverted flap is indeed unanimously considered an excellent technique for potentially most MH in any stage. However, some controversy exist in the literature regarding the comparative efficacy of the inverted ILM flap technique versus traditional ILM peeling in the treatment of small and medium-sized MHs. This prompted us to conduct the present meta-analysis, which represents the first systematic attempt to compare the effectiveness of hole closure and visual acuity improvement between the two treatment approaches specifically in the context of small and medium-sized MHs. We believe our research fills a void in the meta-analysis regarding the specific outcome of small and medium-sized MH treated with two different surgical techniques.

This systematic review and meta-analysis aimed to provide a comprehensive evaluation of the current evidence by comparing the anatomical and visual outcomes of PPV with the inverted ILM flap technique and ILM peeling for small and medium-sized MHs. Upon pooling the results of the MH closure rate, we found that the rate of MH closure did not differ significantly between the inverted ILM flap and ILM peeling groups. This suggests that both techniques are similarly effective in achieving MH closure. Additionally, the improvement in visual acuity was evident in the studies analysed [16, 18‒20]; however, there was no conclusive evidence of a significant difference between the two treatment approaches. Furthermore, the integrity of the ELM and EZ was assessed in a subset of studies [16, 18‒20]. Our analysis revealed no significant variation in ELM or EZ integrity at both 3 months and 12 months postoperatively between the inverted ILM flap and conventional peeling groups. This indicates that both techniques are comparable in preserving the structural integrity of the retina. Overall, our findings suggest that both the inverted ILM flap technique and ILM peeling are equally effective in treating small and medium-sized MHs.

The introduction of the inverted ILM flap technique by Michalewska et al. [25] in 2010 has significantly advanced the treatment of large MHs. This technique involves the use of a peeled-off ILM as a flap to cover the hole, harnessing the potential of Müller cell fragments and gliosis induction to enhance closure. The inverted ILM flap also provides a structural scaffold for tissue proliferation, as cells require a basement membrane for proliferation to occur [25]. These findings are further supported by Spiteri et al. [26], who emphasize the role of Müller cells in providing structural support to the retina. The exact mechanisms underlying the efficacy of the inverted ILM flap technique are varied. Several studies have suggested some potential contributing factors. Experiments have demonstrated the presence of type IV collagen, protein fibre connections, and laminin in the ILM, which may promote the proliferation and migration of Müller cells. This, in turn, induces the movement of the retina toward the centre and facilitates hole closure [27]. Additionally, the separation of the outer retina from the vitreous by the inverted ILM flap reduces the effects of hydration, allowing for the reshaping of the central fovea and improving the visual outcomes [27]. Despite the advantages of the inverted ILM flap technique, there are ongoing debates surrounding its application due to the concerns of potential complications and outcomes. For instance, Shiode et al. [27] have suggested that the flap may displace the hole during surgery, leading to incomplete closure. And meanwhile, Iwasaki et al. [28] have observed excessive proliferation above the hole with the ILM flap technique. These variations in outcomes may be attributed to differences in surgical approaches, such as intentional filling of the hole with ILM, which can cause anatomical damage and scar tissue formation, potentially influencing the final results. In our meta-analysis, we aimed to address these concerns and minimize confounding influences by excluding studies with additional factors and focusing on those utilizing a single flap. The results of our meta-analysis suggest that the inverted ILM flap technique is equally effective as ILM peeling in the treatment of small and medium-sized MHs.

The evaluation of ELM and EZ recovery after surgery at 3 months and 12 months in both groups provides valuable insights into the short-term and relatively long-term effects of the inverted ILM flap technique. The EZ has been identified as a significant marker band for assessing the viability and activity of photoreceptor cells [29]. The integrity of the ELM and EZ is crucial for the restoration of photoreceptor microstructures and has been recognized as an important prognostic factor for visual outcomes following MH surgery [30‒33]. Although our meta-analysis did not reveal a significant difference in ELM and EZ integrity between the inverted ILM flap group and the ILM peeling group, a continuous ELM could be observed postoperatively of most patients in a short time. A fully regenerated ELM was identified to be essential for EZ regeneration [34, 35]. Our study also validated this finding that has emphasized the importance of ELM and EZ restoration for visual recovery after MH surgery. Chou et al. [18] reported an earlier restoration of the ELM in the inverted flap group compared to conventional ILM peeling, with higher ELM restoration rates at 1 and 3 months. However, the rates became comparable between the two groups at 12 months. Similar conclusions have been drawn in other studies [36, 37]. Ramtohul et al. [37] demonstrated that the reconstruction of the foveal ELM at 3 months postoperatively can predict the subsequent restoration of the EZ and potentially lead to improved visual outcomes. Horiguchi et al. [38] compared operated eyes with inverted ILM flap and normal fellow eyes by multifocal electroretinograms. They involved creating a large (2–3 disc diameter) ILM flap on the superior side of the MH, which was then inverted and cover the hole. The superior retina did not have any ILM, while the inferior retina retained a two-layer ILM. This study investigated the functional outcomes of the upper retina without the ILM flap and the lower retina covered by the ILM flap. The parameters such as the peak time and amplitude of N1, P1, N2 (components of the multifocal electroretinogram) and sensitivity were evaluated. These findings indicate that the use of an inverted ILM flap technique, as described in the study, did between the upper and lower retinas. But Iuliano et al. [39] analysed the foveal sensitivity (4°) and a global macular retinal sensitivity (10°) in small MH. They reported that inverted ILM flap probably delays the improvement of the foveal sensitivity (4°) and the fixation stability, but it did not interfere with a good final visual outcome and a global macular retinal sensitivity (10°).

Our study acknowledges several limitations that should be considered. Firstly, the inclusion of only two RCTs and predominantly n-RCTs may introduce potential bias and limit the overall strength of the evidence. Further studies with a larger number of high-quality RCTs are warranted to enhance the validity and reliability of the results. Secondly, we cannot compare the speed of recovery because of insufficient follow-up time, which the retain have almost recovered after 1 month. Further studies with more frequent follow-up time, such as 1, 2, or 3 weeks after surgery, may be necessary to better evaluate anatomical recovery in the early period. Thirdly, we only analysed the efficacy of the traditional ILM peel and did not analyse the improved operation, such as cauliflower with a 360 peeling, concentric circles peeling, retain the centre of peeling, and so on. Further studies could analyse the efficacy of modified surgical methods on postoperative recovery.

Our meta-analysis results indicate that the inverted ILM flap technique is comparable to conventional ILM peeling in terms of its impact on BCVA and retinal anatomical outcomes. Based on the current evidence, the findings may suggest that the technique of encouraging the using of inverted ILM flaps in small and medium-sized MH is not necessary. So far, there are few studies comparing postoperative macular sensitivity between the two ways. Further long-term research, possibly supplemented with postoperative macular sensitivity and electrophysiologic data, may support surgeons in taking more sensible decisions.

We sincerely acknowledge the technical assistance provided by the Renmin Hospital of Wuhan University, which greatly contributed to the successful completion of this study. Their support and expertise were invaluable in conducting this research.

All analyses were exclusively based on previously published articles; thus, ethical approval was not required.

The authors declare no competing interests.

This work was supported by the Grant No. 81600740 from the National Natural Science Foundation of China for L.L. and the Grant No. 2042019kf0071 from the Independent Scientific Research of Wuhan University for L.L.

Ping-Ping Li conceived and designed the study. Jian-Hua Wu and Lu Li searched the literature and contributed to data acquisition and analysis. Ping-Ping Li was responsible for writing of original manuscript. Lu Li contributed to revising and reviewing. All the authors reviewed and approved the final manuscript.

The data underlying this article are available in the article. Further enquiries can be directed to the corresponding author.