Background
Macular hole (MH) is an anatomical defect in the fovea of retina that cause severe visual impairment. It was regarded as an untreatable disease in poor prognosis until the first describtion of vitrectomy to treat MH by Kelly and Wendel [
1] in 1991. The success rate of MH surgery increased to 98% [
2‐
4] of cases with the introduction of internal limiting membrane (ILM) peeling by Eckardt et al. [
5], which was thought to be one of the most effective surgical procedures [
6]. However, the anatomical success rate of MH that are larger than 400 μm is less likely to close and has been as low as 40% [
7,
8], regardless of whether the ILM has been removed or not during vitrectomy. Michalewska et al. [
9] first presented a novel technique of inverted ILM flap for the treatment of large MH, contributing to a relatively high MH closure rates. Recently, a number of clinical studies have suggested that inverted ILM flap technique achieved better anatomical and visual outcomes than ILM peeling [
10‐
15]. However because of lacking appropriate controls, uncontrollable elements, or insufficient samples in most of these studies, the reliable evidences to support such a view were limited. As far as we know, double-arm study to compare anatomical and visual outcomes between these two methods in MH larger than 400 μm have not been systematically reviewed and published. Thus, we conducted a comprehensive meta-analysis to evaluate the efficacy of vitrectomy with inverted ILM flap technique and ILM peeling.
Methods
Search strategy
We cautiously searched for studies that used inverted ILM flap technique or ILM peeling to treat MH larger than 400 μm. The Pubmed, Embase, Cochrane Library databases were systematically searched for all articles including relevant prospective and retrospective clinical trials published before December 2018. The terms used for systematic search were:(macular hole OR macula hole OR MH OR macular break OR macular fissure OR retinal perforations OR retinal break OR retinal hole OR retinal tear) AND (inner limiting membrane OR internal limiting membrane OR ILM OR limiting membrane) AND (peeling OR peel OR removal OR IP OR SIP) AND (inverted OR inversion OR invert OR flap OR flap technique OR IF OR IFT). Furthermore, We enlarged retrieval coverage and manually searched reference lists of original studies, gray literatures and records, without language or publication year restrictions.
Inclusion and exclusion criteria
All abstracts, studies and citations were reviewed and assessed. The inclusion criteria for eligibility were as follows: (1) double-arm studies; (2) studies included cases among patients with MH larger than 400 μm who had been treated with the inverted ILM flap technique or ILM peeling; (3) anatomical hole closure rate and visual acuity (VA) were observed after the treatments; (4) the relevant statistics were provided, such as age, gender, duration of disease, hole size, and follow-up time; (5) prospective randomized control trial or retrospective case series. Exclusion criteria were as follows: (1) non - controlled study; (2) patients with macular retinoschisis, age-related macular degeneration, retinal detachment, or proliferative diabetic retinopathy; (3) treatments with modified inverted ILM flap techniques; (4) short -term follow-up that less than three months; (5) reviews or case reports.
Data were independently extracted and reviewed from each included study by two reviewers (YS and XQL). Any discrepancy between data extractions were resolved by the discussion or consulted by the expert. The following data were extracted: first author, year of publication, type of trials, country, surgical procedure, number of eyes involved, patient demographics, age, duration of disease, minimum diameter of MH, hole closure rate, preoperative and postoperative VA, and follow-up time.
Quality of assessment
This review included eight studies: 4 randomized control trials (RCTs) and 4 retrospective case series. The included RCTs were evaluated for quality in accordance with the “risk of bias” tool recommended by the Cochrane Handbook 5.1.0 [
16].Seven items were assessed: “random sequence generation,” “allocation concealment,” “blinding of participants and personnel,” “blinding of outcome assessment,” “incomplete outcome data,” “selective reporting,” and “other bias.” According to whether the included studies fully meet the above criteria, we assessed the quality of trials. The methodological quality of each study was assessed based on the Newcastle-Ottawa Scale (NOS) [
17] (range, 0 to 9 stars) for quality of case control studies in meta-analysis. Studies were rated in three areas, including selection, comparability and exposure. Scores ≥5 indicated that the quality of research were relatively high. All items were independently assessed by two investigators (YS and XQL), with consensus reached after discussion or expert consultation.
Statistical analysis
Data analysis was collated and analyzed by Review Manager 5.3 software (RevMan 5.3, The Cochrane Collaboration, Oxford, UK). For the rate of MH closure, odds ratios (ORs) and 95% confidence intervals (Cl) were calculated by using Mantel–Haenszel (M-H) method. To compare the evaluation of VA, the mean difference (MD) of preoperative and postoperative measurements between the two surgical treatments were compared using weighted MD and 95% CI. The sofeware estimated statistical heterogeneity among studies using I2 statistic. When I2 > 50%, it suggested there was significant heterogeneity. A random-effects model was used for data synthesis in the presence of significant heterogeneity, while a fixed-effects model was used when there was no significant heterogeneity. The results of meta - analysis was visually examined by forest plot, and the potential publication bias was showed by funnel plot. P < 0.05 was considered statistically significant.
Discussion
We conducted this systematic review and double arm meta-analysis of literature to summarize current evidence and compare the anatomical and visual outcomes of vitrectomy with inverted ILM flap technique and ILM peeling. After pooling the results of the rate of MH closure in these included studies, we performed sensitivity analysis to find the source of heterogeneity. This study [
14] was considered to be remove, because its limitations included small sample size, non-standardized measurements, the presence of multiple surgeons, the utilization of multiple OCT machines, and so on. After eliminating the interference of heterogeneity, the result indicated that the rate of MH closure was significantly higher in vitrectomy with inverted ILM flap than that in vitrectomy with ILM peeling.
In histopathologic findings, when the margins of the ILM were left attached to the edges of the hole, it served as a gap-free natural scaffold for gliosis, and provided a basement membrane for proliferateglial glial cells to maintain anatomic structure of the foveola [
9,
10]. Shiode et al. showed that the neurotrophic factors and basic fibroblast growth factors (bFGF) on the surface of the ILM flap promoted the proliferation and migration of the Müller cells, contributing to fill the MH and enhance closure [
11]. Activated Müller cells also produced neurotrophic factors and growth factors that may promote the survival of retinal neurons [
18]. Moreover, the ILM flaps create a closed compartment enabling the RPE to pump out fluid and keep hole dry [
10]. These findings explained to the better anatomical results in the inverted ILM flap technique group.
As for the outcomes of visual functions, the preoperative and postoperative VA were observation indexes in this review. There were not significant in the preoperative VA between the two groups, which reduced pre-intervention effects to postoperative VA. Considering the heterogeneity in postoperative VA between these trials, the included studies divided into two subgroups based upon follow-up duration to observe the short and long-term visual efficacy. Compared with ILM peeling, postoperative VA was better in the group of inverted ILM flap technique at the follow-up of 3 months. It seemed that flap technique had a signifcant visual gain in the short term. It was believed that the ILM flaps was also a bridge leading photoreceptor to migrate into the retina defect [
19]. Michalewska et al. [
9] found that glial cells proliferated and produced an environment to transfer light from the retinal surface to the photoreceptor cell layer. In the cases after the surgery of inverted ILM flap technique, retinal tissue regenerated from the external limiting membrane (ELM), then regrowth of the ellipsoid layer was observed over the next few months [
4,
20].
However, contrary to expectation, our results indicated no difference in functional outcomes between the two groups of different surgical treatments at relatively long-term follow-up over 6 months. Hayashi et al. [
21] suggested that the photoreceptor layer of fovea might be destroyed and had irreversible damage. Ota et al. [
22] showed the fovea losed its original stratified structure in large MH. The IS/OS and ELM would returned gradually after surgery, but not completely or in all cases. The improvement of the foveal structure did not include the restoration of the normal layered structure of retina. There were no significant differences in structural changes over time after inverted ILM flap technique or ILM peeling. Moreover, the differences in the baseline characteristics such as MH size related to the results [
14,
22]. These findings might explain the limitations in visual recovery after vitrectomy with inverted ILM flap technique at a longer follow-up.
For all we know, double-arm study to compare the efficacy in large MH after treatments between inverted ILM flap technique and ILM peeling was reviewed and analyzed firstly in this meta-analysis. However, several limitations were inevitable and should be taken into account when citing the results of the meta-analysis. First, the number of included studies and available data were limited. Second, this systematic review included not only RCTs but also retrospective studies, which might have potential sources of selection bias. Third, significant heterogeneity among the studies was detected in the comparison of primary and secondary end points. We alleviated but not eliminated completely the heterogeneity through sensitivity analysis and subgroup analysis. The pooled data were from the relative long-term follow-up durations of the studies, introducing a potential heterogeneity. Follow-up durations in the included trials were not long enough for better observations of VA recovery in long term. Furthermore, other influence factors, such as dye for ILM stained or the specific forms of inverted ILM flap technique among the studies might serve as the points of heterogeneity. Due to the deficiency of these available data, we did not perform subgroup analysis.
Conclusions
Our meta-analysis indicated that vitrectomy with inverted ILM flap technique had a better anatomical outcome than ILM peeling in large MH. Flap technique had a signifcant visual gain in the short term, but the limitations in visual recovery at a longer follow-up was found. Moreover, there is an urgent need for long follow-up duration and large prospective randomized study to further confirm the efficacy of inverted ILM flap technique and ILM peeling.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.