Table 1
Study Characteristics Of The Included Studies
Author | Year | Study location | Study design | Follow-up months (mean) | Groups | Eyes, n | Mean age, years | Gender (male/female) | Minimum Mean MH diameter (µm) | Hole closure rate, % |
Kannan et al. [17] | 2018 | India | Prospective randomized control trial | 6 | ILM flap | 30 | 59.37 ± 6.71 | 11/19 | 803.33 ± 120.65 | 90 |
ILM peeling | 30 | 61.17 ± 7.42 | 17/13 | 759.97 ± 85.01 | 76.7 |
Iovino et al. [24] | 2018 | Italy | Randomized Clinical Trial | 6 | ILM flap | 20 | 71 ± 8.77 (54–83) | 9/11 | 666.95 ± 63.7 | 100 |
ILM peeling | 20 | 69 ± 10.4 (49–88) | 7/13 | 664.6 ± 71.1 | 100 |
Manasa et al. [[15] | 2018 | India | Prospective, randomized study | 3 | ILM flap | 43 | 63.41 | 20/23 | 673 | 95.3 |
ILM peeling | 48 | 60.95 | 22/26 | 657.5 | 87.5 |
Velez-Montoya et al. [23] | 2018 | Mexico | Prospective randomized controlled trial | 3 | ILM flap | 12 | 64.2 ± 6.7 | NA | 608.89 ± 213 | 91.67 |
ILM peeling | 12 | 61.8 ± 9.6 | NA | 522.22 ± 82.73 | 91.67 |
Michalewska et al. [14] | 2010 | Poland | Prospective, randomized clinical trial | 12 | ILM flap | 50 | 66 | 13/33 | 759 | 98 |
ILM peeling | 51 | 65 | 8/32 | 698 | 88 |
Ventre et al. [26] | 2022 | Italy | Randomized Clinical Trial | 12 | ILM flap | 25 | 62 ± 5 | 14/11 | 269 ± 52 | 100 |
ILM peeling | 25 | 64 ± 5 | 13/12 | 254 ± 70 | 100 |
Agrawal et al. [16] | 2022 | India | Randomized Clinical Trial | 12 | ILM flap | 75 | 65.39 ± 4.87 | NA | 765.6 ± 77.01 | 100 |
ILM peeling | 75 | 64.17 ± 5.95 | NA | 749.7 ± 167.6 | 93.33 |
Yamada et al. [30] | 2022 | Japan | Retrospective, nonrandomized, comparative study | 12 | ILM flap | 21 | 66.2 ± 10.6 | NA | 278.6 ± 80.7 | 90.5 |
ILM peeling | 21 | 66.6 ± 7.0 | NA | 276.0 ± 84.5 | 100 |
Yilmaz et al. [25] | 2021 | Turkey | Retrospective, observational, single center study | 12 | ILM flap | 32 | 68.22 ± 5.38 | 12/20 | 494.48 ± 226.5 | 100 |
ILM peeling | 25 | 67.04 ± 7.95 | 7/18 | 460.13 ± 160.89 | 100 |
Yan et al. [34] | 2021 | China | Retrospective, non-ran-domized comparative study | 6 | ILM flap | 29 | 64.28 ± 6.24 | 13/16 | 533.41 ± 245.03 | 100 |
ILM peeling | 19 | 67.89 ± 6.19 | 3/16 | 502.75 ± 128.59 | 94.7 |
Friedrich et al. [35] | 2021 | German | Retrospective observational study | 6 | ILM flap | 30 | 65 | NA | 408.4 ± 157.5 | 100 |
ILM peeling | 15 | 72 | NA | 287.4 ± 104.9 | 100 |
Baumann et al. [21] | 2020 | United Kingdom | Retrospective, nonrandomized, case series | 12 | ILM flap | 68 | NA | NA | 560 ± 104 | 98.53 |
ILM peeling | 49 | NA | NA | 504 ± 106 | 87.76 |
Bottoni et al. [29] | 2020 | Italy | Retrospective, nonrandomized comparative study | 12 | ILM flap | 24 | 69.12 ± 9.75 | 6/18 | 652.5 ± 202.85 | 95.83 |
ILM peeling | 17 | 67.58 ± 8.04 | 6/11 | 632.11 ± 133.90 | 100 |
Ramtohul et al. [13] | 2020 | France | Retrospective, nonrandomized comparative study | 6 | ILM flap | 23 | 68.03 ± 9.50 | 13/10 | 657.33 ± 172.36 | 95.65 |
ILM peeling | 23 | 65.69 ± 10.25 | 16/7 | 574.09 ± 164.68 | 69.56 |
Álvarez et al. [36] | 2020 | Spain | Retrospective case series study | 6 | ILM flap | 12 | 59.5 (56–67) | 6/6 | 307 (170–453) | 91.66 |
ILM peeling | 16 | 60 (57–67) | 5/11 | 264 (116–325) | 81.25 |
Lumi et al. [37] | 2020 | Slovenia | Retrospective study | NA | ILM flap | 21 | 69.8 ± 7.1 | 6/15 | 299.2 ± 151.6 | 95.2 |
ILM peeling | 17 | 68.8 ± 5.4 | 6/9 | 458.1 ± 130.8 | 100 |
Rizzo et al. [38] | 2018 | Italy | Retrospective, consecutive, nonrandomized comparative study | 9.3 ± 2.03 | ILM flap | 320 | NA | NA | NA | 91.3 |
ILM peeling | 300 | NA | NA | NA | 78.6 |
Iturburu et al. [39] | 2019 | Spain | Retrospective, nonrandomized, comparative study | 6 | ILM flap | 27 | NA | NA | NA | 92.6 |
ILM peeling | 39 | NA | NA | NA | 94.8 |
Wu et al. [22] | 2018 | Taiwan | Retrospective study | 17.29 ± 20.20 | ILM flap | 6 | 62.33 ± 4.18 | 3/3 | 297.83 ± 56.55 | 100 |
ILM peeling | 8 | 57.88 ± 12.58 | 2/6 | 264.25 ± 87.63 | 75 |
Hu et al. [40] | 2019 | China | Retrospective study | NA | ILM flap | 10 | 58.8 ± 13.8 | 3/16 | NA | 100 |
ILM peeling | 11 | 59.9 ± 8.5 | 4/15 | NA | 81.8 |
Mete et al. [41] | 2017 | Italy | Retrospective study | 6 | ILM flap | 34 | 60.8 | NA | NA | 94.1 |
ILM peeling | 36 | 57.8 | NA | NA | 61.1 |
Pak K Y et al. [20] | 2017 | Korea | Retrospective case series | 3 | ILM flap | 41 | 65.7 ± 7.5 | 9/32 | 590.8 ± 113.1 | 100 |
ILM peeling | 51 | 66.1 ± 6.6 | 17/34 | 558.9 ± 106.8 | 88.2 |
Yamashita T et al. [19] | 2018 | Japan | Retrospective case series | 6 | ILM flap | 60 | NA | NA | 624 ± 108 | 100 |
ILM peeling | 105 | NA | NA | 544 ± 119 | 92.4 |
Iwasaki et al. [28] | 2018 | Japan | Retrospective case series | 10.0 ± 4.1 | ILM flap | 14 | 65.4 ± 9.0 | 5/9 | 655.2 ± 112.1 | 21.4 |
ILM peeling | 10 | 69.9 ± 6.6 | 2/8 | 551.1 ± 99.5 | 70 |
Narayanan et al. [18] | 2018 | St. Louis | Retrospective case series | 6 | ILM flap | 18 | 60.22 ± 12.09 | 5/13 | 577.4 ± 159.4 | 88.9 |
ILM peeling | 18 | 67.50 ± 7.78 | 8/10 | 493.8 ± 170.5 | 77.8 |
Leisser et al. [27] | 2022 | Austria | Prospective randomized trial | 3 | ILM flap | 7 | 71 ± 7 | 3/4 | 275 ± 90 | 100 |
ILM peeling | 9 | 67 ± 5 | 2/7 | 244 ± 101 | 100 |
Note. ILM internal limiting membrane, MH macular hole, NA not available, n number, µm micrometer, % percentage. |
Table 2
New Castle Ottawa Scale For Assessment Of Publication Bias Of Non-RCT Studies
Study | Selection | Comparability | Exposure | Total |
Pak 2017 [20] | 2 | 2 | 2 | 6 |
Mete 2017 [41] | 2 | 2 | 2 | 6 |
Yamashita 2018[19] | 2 | 2 | 1 | 5 |
Wu 2018 [22] | 2 | 2 | 2 | 6 |
Iwasaki 2018 [28] | 2 | 1 | 1 | 4 |
Rizzo 2018 [38] | 1 | 1 | 2 | 4 |
Narayanan 2018 [18] | 2 | 2 | 2 | 6 |
Iturburu 2019 [39] | 3 | 2 | 2 | 7 |
Hu 2019 [40] | 2 | 2 | 2 | 6 |
Alvarez 2020 [36] | 3 | 2 | 2 | 7 |
Bauman 2020 [21] | 2 | 2 | 2 | 6 |
Bottoni 2020 [29] | 2 | 2 | 2 | 6 |
Ramtohul 2020[13] | 2 | 2 | 2 | 6 |
Yilmaz 2021 [25] | 2 | 2 | 3 | 7 |
Yan 2021 [34] | 2 | 2 | 2 | 6 |
Friedrich 2021[35] | 2 | 2 | 3 | 7 |
Yamada 2022 [30] | 3 | 2 | 2 | 6 |
Lumi 2020 [37] | 3 | 2 | 2 | 7 |
Study Characteristics
The detailed flowchart of the study selection process is shown in Fig. 1. A total of 153 articles were found after the preliminary literature search, after removing the duplicates the articles were shortlisted first by the topics, then by reading the abstracts and finally by the way of full text review; this left us with a total 26 studies [13–30, 34–41] which were included in this meta-analysis.
The main characteristics of the included trials are presented in Table 1. Out of the included studies 18 were of retrospective in nature while 8 were randomized controlled trials. A total of 2102 eyes were included in this meta-analysis 1050 in the ILM peeling group and 1052 in the ILM flap group. The mean age ranged from 58.8 to 71 years in the ILM flap group and from 57.8 to 72 years in the ILM peeling group, the average minimum diameter of the MH ranged from as low as 269 µm to as high as 803.33µm in the flap group and 244 to 759.97µm in the peeling group. The mean follow-up months duration ranged from 3 to 17.29 months.
Quality Assessment
We assessed the quality of the 8 RCTs using the Cochrane risk of bias tool and overall, all the studies were found to have low risk of bias and are comprehensively shown in Fig. 2. Six out of eight studies were deemed high quality as they had low risk of bias in all assessing criteria. However, the remaining two studies [14, 24] had unclear risk in performance, detection or attrition bias.
As of the 18 retrospective studies all of which were qualitatively assessed using NOS and all studies got 5 or more stars out of 9 except these two studies [28, 38] got 4 out of 9 stars as they had problems with reporting of selection of cases and comparability shown comprehensively in Table 2.
Macular Hole Closure Rate
All the studies reported this outcome, the rate of MH closure was 94.5% (995/1052 eyes) in the ILM flap group while 85.4% (897/1050 eyes) in the ILM peeling group. The overall pooled data showed significantly higher MH closure rate with inverted ILM flap technique in comparison with the ILM peeling group (OR = 2.74, CI = 1.62 to 4.66; P = 0.0002; Fig. 3).
Overall heterogeneity came out to be 28% (Fig. 3), for which sensitivity analysis was performed which revealed this study by Iwasaki et al. [28] to be the source for it as it had a lot of limitations. After removing this trial, the overall MH closure rate came out to be significantly higher with ILM flap technique (OR = 3.44, CI = 2.41 to 4.90; P < 0.00001; Fig. 4) and the heterogeneity lowered down to 0%, Fig. 4.
Furthermore, the forest plot showed this retrospective study conducted by Rizzo et al.[38] carried significant weightage on the overall result therefore to confirm the robustness of the results leave one out analysis was performed which validated the conclusion in favor of ILM flap technique for both the subgroup (OR = 2.81, CI = 0.98 to 8.01; P = 0.005; Fig. 5) and the overall result (OR = 2.59, CI = 1.39 to 4.85; P = 0.003; 5).
Subgroup analysis was performed based on the different macular hole sizes taken in the included RCTs and Retrospective studies:
In Retrospective subgroup with mixed macular hole size (OR = 3.37, 95% CI = 2.19 to 5.18; P < 0.00001; Fig. 3) the result favored ILM flap technique over ILM peeling as the p value reached a statistical significance.
While in the RCT subgroup with macular hole sizes < 400µm there was no difference identified in MH closure rate between the two techniques, Fig. 3.
In the RCT subgroup with macular hole sizes greater than 400µm (OR = 3.36, 95% CI = 1.39 to 8.13; P = 0.007; Fig. 3) and in the retrospective subgroup with macular hole sizes > 400 µ m and < 400µm the MH closure rate was not significantly different between the ILM flap technique group and ILM peeling group (OR = 2.65, 95% CI = 0.61 to 11.56; P = 0.20; Fig. 3) and (OR = 0.92, 95% CI = 0.04 to 23.79; P = 0.96; Fig. 3)respectively.
Preoperative Visual Acuity
All the included studies reported the preoperative visual acuity outcome after treatment with ILM peeling or inverted ILM flap technique except the study by Michalewska et al. [14] didn't mention SD of preoperative VA. The forest plot for preoperative visual acuity showed no significant difference between ILM flap and ILM peeling technique (MD = 0.03, 95% CI= -0.01 to 0.07; P = 0.15; Fig. 6). Heterogeneity came out to be high (I2 = 55%).
Postoperative Visual Acuity
All the included studies reported the postoperative visual acuity outcome after treatment with ILM peeling or inverted ILM flap technique except the study by Michalewska et al. [14] which didn't mention SD of postoperative VA.
Subgroup analysis was performed for the comparison of ILM flap group and ILM peeling group based on study type and macular hole size.
Chiefly only the RCT group with macular hole sizes greater than 400µm showed statistically significant result in favor of ILM flap technique (MD = -0.13, 95% CI = -0.17 to -0.08; P < 0.00001; Fig. 7) with mild heterogeneity (I2 = 13%). For this concern we successfully performed a sensitivity analysis and after removing this study by Iovino et al. [24] shown in; Fig. 9, the heterogeneity was resolved giving a statistically significant result in favor of ILM flap (MD = -0.13, 95% CI = -0.16 to -0.09; P < 0.00001; Fig. 9)
While in all the other subgroups that are RCTs with macular hole sizes < 400µm (MD = 0.00, 95% CI = -0.06 to 0.07; P = 0.90; Fig. 7) and similarly for retrospective studies with macular hole sizes greater than 400µm (MD =- 0.01, 95% CI = -0.10 to 0.08; P = 0.81; Fig. 7) and lesser than 400µm (MD = -0.16, 95% CI = -0.18 to 0.49; P = 0.36; Fig. 7) as well as for mixed macular hole sizes (MD = 0.11, 95% CI = -0.01 to 0.24; P = 0.08; Fig. 7) the results were found to be statistically insignificant. The meta-analysis of overall pooled data showed postoperative visual acuity with inverted ILM flap technique group was statistically not significant in comparison with the ILM peeling group (MD = 0.01, 95% CI = -0.04 to 0.06; P = 0.74). The overall heterogeneity was also high.
Furthermore, to account for the impact of follow-up duration on the results, we conducted a subgroup analysis that categorized all the RCTs with > 400µm sizes [15–17, 23, 24][14] based on their follow-up duration except the study by Michalewska et al. [14]. Studies were divided into three categories: follow-up duration at 3,6 and 12 months. The results showed that postoperative visual acuity was significantly better in ILM flap technique than ILM peeling in all the subgroups at 3 months(MD=-0.11, 95% CI = -0.19 to -0.04; P = 0.003; Fig. 8), 6 months (MD=-0.11, 95% CI = -0.18 to -0.04; P = 0.002; Fig. 8) and 12 months (MD=-0.12, 95% CI = -0.16 to -0.08;P < 0.00001; Fig. 8) respectively. The overall pooled effect was also statistically significant (MD=-0.11, 95% CI = -0.14 to -0.08; P < 0.00001; Fig. 8). Additionally, there was mild to moderate heterogeneity in all the pooled results.
Publication Bias
Three funnel plots for the outcomes of MH closure rate (Fig. 10), preoperative VA (Fig. 11) and postoperative VA (Fig. 12) revealed that almost all of the studies shown by scattered points in the funnel plot were distributed in the middle and top of the baseline and were located in the range of the inverted funnel which implies that there is no significant publication bias and the results from this study are reliable.
Furthermore, to confirm our assessment of publication bias we performed an eggers test as the number of articles exceeded 10. Egger’s test for a regression intercept gave a p-value of 0.417 for the MH closure rate, indicating no evidence of publication bias.