Browse Subject Areas

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Vitrectomy with or without internal limiting membrane peeling for idiopathic epiretinal membrane: A meta-analysis

Vitrectomy with or without internal limiting membrane peeling for idiopathic epiretinal membrane: A meta-analysis

  • Wei-Cheng Chang, 
  • Chin Lin, 
  • Cho-Hao Lee, 
  • Tzu-Ling Sung, 
  • Tao-Hsin Tung, 
  • Jorn-Hon Liu



Studies on vitrectomy with and without internal limiting membrane (ILM) peeling for idiopathic epiretinal membrane (ERM) have yielded uncertain results regarding clinical outcomes and recurrence rates.


To compare the clinical outcomes of vitrectomy with and without ILM peeling for idiopathic ERM.


Databases, including PubMed, Embase, Cochrane, Web of Science, Google Scholar, CNKI databases,, and, published until July 2016, were searched to identify studies comparing the clinical outcomes following vitrectomy with ERM and ILM peeling and with only ERM peeling, for treating idiopathic ERM. Studies with sufficient data were selected. Pooled results were expressed as mean differences (MDs) and risk ratios (RRs) with corresponding 95% confidence intervals (CI) for vitrectomy with and without ILM peeling with regard to postoperative best corrected visual acuity (BCVA), central retinal thickness (CRT), and ERM recurrence rate.


Eleven retrospective studies and one randomized controlled trial involving 756 eyes were identified. This demonstrated that the postoperative BCVA within 12 months was significantly better in the non-ILM peeling group (MD = 0.04, 95% CI: 0.00 to 0.08; P = 0.0460), but that the patients in the ILM peeling group had significantly better postoperative BCVA after 18 months (MD = −0.13, 95% CI: −0.23 to −0.04; P = 0.0049) than did those in the non-ILM peeling group. The non-ILM peeling group exhibited a higher reduction in postoperative CRT (MD = 51.55, 95% CI:−84.23 to −18.88; P = 0.0020) and a higher recurrence rate of ERM (RR = 0.34, 95% CI:0.16 to 0.72; P = 0.0048) than did the ILM peeling group. However, the improvement rates of BCVA (RR = 1.03, 95% CI:0.72 to 1.47; P = 0.8802) and postoperative CRTs (MD = 18.15, 95% CI:−2.29 to 38.60; P = 0.0818) were similar between the two groups.


Vitrectomy with ILM peeling results in better visual improvement in long-term follow-ups and lower ERM recurrence rates, and vitrectomy with only ERM peeling is more efficacious in reduction of CRT than is vitrectomy with ILM peeling.

1. Introduction

An epiretinal membrane (ERM), also known as a macular pucker, is a condition affecting the avascular fibrocellular membrane over the central macular area between the vitreous and internal limiting membrane (ILM). Its pathogenic mechanism has an unknown etiology and can be idiopathic or secondary to other ocular diseases, trauma, or previous intraocular operation. The incidence of idiopathic ERM reportedly ranges from 2% in patients younger than 60 years to 12%–20% in those older than 70 years [1]. It may reduce visual acuity (VA) and cause micropsia, macropsia, monocular diplopia, metamorphopsia, or even progressive vision loss [2]. Some hypotheses of the pathogenesis of ERM have involved postulating the proliferation of fibroblasts, glial cells, and astrocytes after ILM disruption, following posterior vitreous detachment [3, 4]. Sebag et al. involve speculating that a residual posterior vitreous cortex (vitreouschisis), attached to the macula during the liquefying process of the vitreous body, may play a role in ERM development [5]. Kishi and Shimizu reported that premacular vitreous cortex, which forms the posterior wall of the premacular liquefied pocket, plays a key role in the development of idiopathic preretinal macular fibrosis in eyes with or without posterior vitreous detachment [6].

Pars plana vitrectomy with membrane peeling has been effectively used for the surgical treatment of ERM since 1978 [7]. A high visual improvement rate, up to 90%, and a recurrence rate of 1%–16% have been reported after successful surgery [3, 811]. Currently, the surgical methods for membrane peeling have evolved because of the use of dyes. Triamcinolone stained the cortical vitreous and ERM although not the ILM [12], whereas indocyanine green (ICG), trypan blue, and brilliant blue G (BBG) were used to stain the ILM [13]. Some ILM peeling reports have revealed results such as improved VA outcomes, lower recurrence rates, and reduced retinal striae [14, 15]. The ILM peeling procedure is increasingly being used by retinal surgeons from 25% in 2008 to 44% in 2010 [16]. Although an increasing number of vitrectomy with ILM peeling has been reported, ILM peeling is believed to cause functional and mechanical damage to the Muller cells because the ILM is the basal lamina connected to the end feet of the Muller cells [1719]. Whether to consider peeling of the ILM a surgical method for treating idiopathic ERM disease continues to be debated among vitreoretinal surgeons.

Because of the uncertainty of the effectiveness of ILM peeling in idiopathic ERM surgery, many studies have investigated the single peeling (only ERM) and double peeling (ERM and ILM peeling) surgical techniques for idiopathic ERM in recent 20 years since 1993 [15]. Most studies were retrospective comparative studies; however, only one new randomized control trial with a sufficiently large sample size and appropriate design was published in 2016 [20]. Until now, only one meta-analysis has enrolled eight retrospective studies to compare the effectiveness of pars plana vitrectomy with and without ILM peeling for idiopathic ERM removal [21]. The results reported by Liu et al. revealed uncertainty in postoperative VA outcomes and nonsignificant differences in ERM recurrence rates in the meta-analysis because of the limited sample size. In our study, we attempted to include retrospective studies published in recent years and recruited the results of the randomized control trial for detailed discussion and comparison. Eleven retrospective studies and one randomized control trial were meta-analyzed to compare existing evidence regarding the efficacy and clinical outcomes of vitrectomy with and without ILM peeling for idiopathic ERM.

2. Materials and methods

2.1 Literature search methods

The PRISMA checklist is described in S1 Table [22]. We selected relevant prospective or retrospective studies published before July 2016, by searching PubMed, Embase, Cochrane, Web of Science, Google Scholar, CNKI databases,, and The studies that compared the outcomes following vitrectomy, with and without ILM peeling, for treatment of idiopathic ERM disease were included. We did not apply language restrictions. We used relevant text words and medical subject headings that included all possible spellings of ERM and ILM peeling (detailed search strategy and records are shown in S2 Table). We considered all potentially eligible studies for review, irrespective of the primary outcome or language. Reference lists of all retrieved articles were searched manually to broaden the search. All scanned abstracts, studies, and citations were reviewed.

2.2 Criteria for inclusion, exclusion, and outcomes of interest

We included studies in this meta-analysis if they were randomized controlled trials or retrospective studies involving adults with idiopathic ERM, compared vitrectomy with and without ILM peeling, entailed pre- and postoperation monitoring of at least 3 months, and reported one change in the best corrected visual acuity (BCVA), central retinal thickness (CRT), or the rate of ERM recurrence at the end of the follow-up. The exclusion criteria were as follows: case series studies; studies with less than 3-month durations of follow up, studies including cases with a history of retinal detachment, proliferative diabetic retinopathy, retinal vascular occlusion disease that could lead to secondary macular pucker, and significant differences in the preoperative BCVA between the two groups. The outcomes assessed were extracted as follows: postoperative BCVA, rate of vision improvement, rate of ERM recurrence, and postoperative CRT at the end of follow-up, and change in CRT between baseline and the end of follow up.

2.3 Data extraction and quality assessment

Two investigators (WCC and CHL) independently reviewed study titles and abstracts, and studies that satisfied the inclusion criteria were retrieved for full-text assessment. Studies selected for detailed analysis and data extraction were analyzed by two investigators (WCC and CHL); disagreements were resolved by a third investigator (JHL). For each selected article, we extracted the following data: first author’s name, year of publication, ethnicity of the study population, participant numbers of both groups, and population characteristics including mean age, sex, follow-up duration, change in BCVA (mean [SD]), change in CRT (mean [SD]), and ERM recurrence rate at the end of the follow up. Because most of the selected studies were nonrandomized surgical studies, two independent reviewers (WCC and CHL) assessed the quality of eleven included retrospective trials according to the Newcastle–Ottawa Scale [23], and all the studies received scores >6 points. Only one enrolled randomized controlled trial was assessed according to the Cochrane Collaboration Reviewers’ Handbook for Systematic Reviews of Interventions [24] and revealed a low risk of bias.

2.4 Statistical analysis

We assessed the effect of vitrectomy with ILM peeling or not on five outcomes: data of functional efficacy, as assessed by postoperative BCVA and the rate of visual improvement; data of anatomical efficacy, as assessed by postoperative CRT and change of CRT; and the rate of ERM recurrence. We analyzed BCVA and CRT as continuous variables and reported mean difference (MD) with 95% CIs. For the analysis of the proportion of participants achieving VA improvement and those having recurrence of ERM at the end of follow up, we calculated risk ratios (RR) with 95% CIs. The I2 statistic, which was estimated using the DerSimonian–Laird method, was used to assess heterogeneity. An I2 value >50% indicated a moderate to high heterogeneity [25]. We presented pooled results from the fixed-effects and random-effects models to compare the differences based on the suggestions of a previous study [26]. If the treatment effect difference estimation on the basis of the fixed-effect and random-effect models were similar, we used the random-effects model to conclude because it can address interstudy heterogeneity, and is more conservative than the fixed-effect model. Egger’s regression and a funnel plot were used to test the symmetry of the pooled results [27].

For a high heterogeneity result, a meta-regression using an average summary value was used to determine the source of heterogeneity. Possible moderators (study type, quality score, age, and follow-up duration) were tested to explore heterogeneity. This study considered a P value of <0.05 to be significant for all analyses. Statistical analyses were conducted using the “metafor” [28] and “meta” [29] packages of R software, version 3.2.3.

3. Results

3.1 Study selection

We identified 797 studies from PubMed, Embase, Cochrane, Web of Science, Google Scholar, CNKI databases,, and, of which 12 (with data on 756 eyes) were included in our analysis. Fig 1 depicts the overall study identification process. For our analysis, 755 studies were unsuitable because they included duplicate studies, in vitro studies or animal studies, case reports, and review articles irrelevant to our topic. Forty records remained and 26 records were unrelated to the topic. Of the remaining 14 papers, 1 paper did not contain sufficiently detailed data for analysis and one had significant differences in potential confounders (preoperative BCVA). After screening all the titles, 11 retrospective studies and a randomized controlled trial (RCT) were included in our meta-analysis [20, 3040]. A total of 12 retrospective studies were retrieved for more detailed evaluation; detailed data are presented in Tables 1 and 2.

3.2 Study characteristics

The 12 studies were published between 2005 and 2016 (one was published in 2016). In total, 756 eyes, comprising 410 and 346 eyes without and with ILM peeling, respectively, that underwent treatment of idiopathic macular pucker were included in this meta-analysis. Of the included studies, 11 studies were retrospective, and 1 was an RCT; and seven, two, and three studies were conducted in Asia, Europe, and USA, respectively. The sample size of each study varied from 34 to 139 eyes. The baseline characteristics of each included study, such as duration of follow up, preoperative BCVA, and preoperative CRT are listed in Table 1. Table 2 also presents postoperative BCVA, postoperative CRT, change in CRT reduction, vision improvement rate (which is defined by an improvement in VA of ≥2 Snellen lines in three studies), and ERM recurrence rates (which are defined as any evidence of a recurrent macular ERM on spectral domain optical coherence tomography). Five studies mentioning postoperative complications, and retinal detachment, vitreous hemorrhage, punctate retinal hemorrhage, cataract, and retinal tear are listed in Table 3. Of these 12 studies, seven used ICG, one used triamcinolone and ICG, and two used the BBG assistant for ILM staining. Two studies did not mention the type of stain in the context. Phakia and pseudophakia were also noted in some studies. The 11 retrospective studies contained the confounding factor of lens status without strict inclusion criteria for prior lens variability; however, the RCT did have a strict inclusion criteria for prior lens variability.

Table 3. Surgery-related features of studies included in the meta-analysis.

3.3 Efficacy analysis

Fig 2 depicts the six main outcomes from our meta-analysis. In a pooled analysis of nine trials involving 482 eyes, five trials indicated that the non-ILM peeling group exhibited better BCVA in the <12-month follow-up period after surgery than did the ILM peeling group (mean difference [MD] = 0.04, 95% CI: 0.0007 to 0.008; P = 0.0460), and no significant heterogeneity was observed between the two groups (heterogeneity P = 0.3147, I2 = 15.7%) (Fig 2). Although the RCT only revealed that the differences were not significant, the combined retrospective studies with RCT rendered the differences significant.

Fig 2. Main results from the meta-analysis of vitrectomy with epiretinal membrane (ERM) and ILM peeling or with only ERM peeling in idiopathic ERM.

(A) Best-corrected visual acuity (BCVA) <12 months follow-up duration after surgery; vitrectomy with only ERM peeling yielded significantly better results; (B) BCVA >18 months follow-up duration after surgery; vitrectomy with ERM+ILM peeling yielded significantly better results; (C) rate of improvement in visual acuity, defined as ≥2 Snellen lines at the end of follow-up; the difference between the two groups was nonsignificant (D) ERM recurrence rate; was significantly lower in the ERM + ILM peeling group than in the ERM peeling only group (E) central retinal thickness (CRT) at the end of follow-up; the difference between the two groups was nonsignificant; (F) CRT reduction at the end of follow-up; CRT reduction was significantly higher in the ERM peeling group than in the ERM+ ILM peeling group.

In the other four trials, the ILM peeling group revealed better BCVA in a follow-up time >18 months greater than that of the non-ILM peeling group (MD = −0.13, 95% CI: −0.23 to −0.04; P = 0.0049), with no significant heterogeneity between studies (heterogeneity P = 0.6060, I2 = 0%; Fig 2).

In the evaluation of the rate of VA improvement after surgery, the pooled data of five studies including 186 eyes showed no significant difference between the two groups (RR = 1.03, 95% CI:0.72 to 1.47; P = 0.8802). The heterogeneity between studies was statistically significantly moderate (heterogeneity P = 0.0296, I2 = 62.8%; Fig 2).

The postoperative CRT was assessed in seven studies, involving a total of 404 eyes. Pooling the data of these studies showed no significant differences between the ILM peeling and non-ILM peeling groups (MD = 18.15, 95% CI: −2.29 to 38.60; P = 0.0818). The heterogeneity between studies in this analysis was statistically significant (heterogeneity P = 0.0202, I2 = 60%; Fig 2).

We assessed the change in CRT after operation in three studies, involving a total of 159 eyes. Pooling the data from these studies showed that the non-ILM peeling group had a higher mean reduction in CRT than did the ILM peeling group (MD = −51.55, 95% CI: −84.23 to −18.88; P = 0.0020), with no statistically significant heterogeneity between studies (heterogeneity P = 0.8699, I2 = 0%; Fig 2).

The rate of ERM recurrence after surgery was reported in 10 studies involving a total of 674 eyes. By using a random-effects model, the rate of ERM recurrence after initial surgery was revealed to be higher in the non-ILM peeling group than that in the ILM peeling group (RR = 0.34, 95% CI: 0.16 to 0.72; P = 0.0048; Fig 2). The heterogeneity between the studies was not statistically significant (heterogeneity P = 0.5737, I2 = 0%).

Funnel plot and Egger’s regression were used to determine the potential bias of pooled results. In all six funnel plots, we found no evidence of asymmetry from a visual observation and Egger’s regression yielded the same result (P value of Egger’s regression > 0.05; S1 Fig). Egger’s regression test indicated no evidence of publication bias among the included studies or pooled results in this meta-analysis. A summary of all the results is presented in Table 4.

Because high heterogeneities were found in the rate of VA improvement and postoperative CRT, we further analyzed the sources of heterogeneities using meta-regression. Table 5 lists the moderator effects from study type, quality score, age, and follow-up duration. However, all of the potential moderators showed no significant effects on heterogeneity in the meta-regression analysis.

Table 5. Meta-regression analysis of heterogeneity and moderator effects of study characteristics on high heterogeneity outcomes.

4. Discussion

Our results revealed that—compared with vitrectomy with ILM peeling—vitrectomy without ILM peeling achieved better BCVA in a follow-up duration shorter than 12 months. By contrast, better BCVA was observed in the ILM peeling group in studies in which the follow-up duration was longer than 18 months. Furthermore, non-ILM peeling group exhibited higher reduction in postoperation CRT and a higher rate of ERM recurrence than did the ILM peeling group. These data thus support that vitrectomy with ILM peeling is a preferable surgical treatment that can improve long-term postoperative BCVA of idiopathic macular pucker and lower the recurrence rate of ERM.

In the management of idiopathic macular pucker, pars planar posterior vitrectomy with ERM peeling and ILM peeling have been reported to successfully remove macular pucker, improve VA, and relieve metamorphopsia [41]. However, the safety and efficacy of ILM peeling remain controversial. Some studies have mentioned that ILM peeling may not result in improved postoperative BCVA compared with only peeling the ERM, and have reported complications, including the formation of macular holes [42]. To date, only one meta-analysis has discussed the postoperative BCVA, vision improvement, and ERM recurrence rates with limited data and unspecified conclusions [21]. This meta-analysis showed no significant difference in ERM recurrence rates between the ILM peeling and non-ILM peeling groups. Furthermore, some anatomical damage after ILM peeling including dissociation of the nerve fiber layer and inner retinal dimpling was reported by recent studies [4345]. The functional significance of the anatomical damage is still controversial [37, 43, 46, 47]. In this context, we performed a more complete meta-analysis to compare the postoperative BCVA, vision improvement rates, ERM recurrence rates, postoperative CRT, and CRT reduction between the two groups by including more retrospective studies and RCTs, and providing more powerful evidence in the visual and recurrence outcomes discussion.

Our meta-analysis has highlighted significant findings that support the benefits of ILM peeling for the treatment of idiopathic ERM with a lower recurrence rates in the ILM peeling group than in the non-ILM peeling. De Bustros et al. promoted a hypothesis regarding ILM as a growth bridge for proliferation of fibroblasts, glial cells, and astrocytes from the retina to form the ERM [3]. Some studies have reported the benefits of ILM peeling in the ERM surgery in reducing the risk of ERM recurrence. Our trial is the first meta-analysis to present evidence to indicate that the ERM recurrence rate is related to ILM peeling. Compared with the single randomized controlled trial [20], no recurrence in both the groups was observed at the end of 12-month follow-up. Ripandelli et al. mentioned that the result may be related to the limited follow-up time, and that a longer observation period is necessary. We enrolled 10 studies investigating ERM recurrence with a follow-up time range from 8.9 to 47.9 months, which is longer than the follow up time of RCT. A Trial Sequential Analysis (TSA) was performed for calculating the required sample size for this factor. The number of necessary samples is 1263; we collected 674 samples in this meta-analysis. Although this number was insufficient, Fig 3 illustrates that the cumulative z-curve crosses the traditional boundary near the trial sequential monitoring boundary and tends to reach the required information size. We can fairly suppose that if we had more samples for this factor, the results would be consistent with our findings and exhibit strong validity. Our results have a crucial role in the discussion of ERM recurrence with or without ILM peeling. However, regarding the study design and methodology of our enrolled studies, the definition of the ERM recurrence differed among the studies. More highly qualified RCTs are necessary to provide robust evidence to prove our hypothesis.

Fig 3. Trial Sequential Analysis (TSA) of the ERM recurrence rate outcome in this meta-analysis.

TSA is a methodology that includes a sample size calculation for a meta-analysis with the threshold of statistical significance. The detailed settings of this TSA were shown as follows: Significance level = 0.05; Power = 0.80; incidence of control = 2.3; relative risk reduction = 30%; I2 = 0%. Finally, the number of required samples is 1263 but this meta-analysis collected only 674 samples.

Another significant finding from our meta-analysis is a more efficient effect on the reduction of CRT in the non-ILM peeling group than in the ILM peeling group. This effect was also noted in the RCT conducted by Ripandelli et al, in which faster reduction of foveal thickness and higher central foveal cube volume were observed in the non-ILM peeling group than in the ILM peeling group [20]. Moreover, this higher reduction in CRT was not achieved along with a significantly thinner postoperative CRT, better BCVA, or pronounced higher improvement rate in the VA between both the groups. These findings were consistent with those of Ripandelli et al. These findings were also reported by Chang et al. They mentioned a higher proportional decrease in central macular thickness in the single-peeling group than in the double-peeling group [16]. Chang et al. considered that it may be related to an increased disruption or swelling of the inner retinal layers postoperatively in the double-peeling group. Wollensak et al. also demonstrated that ILM may be the structure that contributes to the biomechanical strength of the retina [48], thereby explaining the ILM maintaining the retinal tension. Despite the residual ILM function, we thought that the longer operation time and more complex peeling procedure in the double-peeling group leading to more turbulent retinal tissue may contribute to the ILM peeling group exhibiting an decreased reduction in CRT change in comparison with the non-ILM peeling group (with the effect of potential retinal edematous changes due to increased damage). Longer operation time may cause more potentially phototoxicity but we did not find enough evidence to prove the association between light damage and CRT change currently [4951]. Moreover, the postoperative CRTs at the end of follow-up were not significantly different between the two groups. The high heterogeneity between studies in this analysis reflected the differences in study design. The different CRT definition and measurement by different OCT machine with different protocol exist in each studies. TSA indicated the lack of numbers of required samples which ensure that the results are credible and significant. Although the RCT showed significant CRT thinning in the non-ILM peeling group, we still believed that increased and larger RCTs are necessary to verify whether the postoperative CRT outcome or CRT reduction were affected by ILM peeling.

In our meta-analysis, the postoperative BCVA significantly better within the <12 months follow-up time in the non-ILM peeling group, but showed converse results in the longer follow-up >18 months in the ILM peeling group. This result was consistent with that of Liu et al [21]. We maintained a conservative attitude toward these results because of the marked influence from confounding factors such as post-vitrectomy cataract formation, variations in the VA measurement method in each trial, and the conversion of different units to logMAR. Among our enrolled studies, only the RCT adopt Early Treatment Diabetic Retinopathy Study (ETDRS) chart to evaluate VA in comparison of other retrospective studies using Snellen chart, which may lead to some inconsistency for the BCVA outcomes. The cases included in the Ripandelli et al. had the visual acuity better than 20/200 which is showed less pronounced difference between ETDRS and Snellen chart [52]. Anatomical and functional changes in the photoreceptor cells postoperation, particularly in the ILM peeling method, were estimated by microperimetry in many trials. These factors can potentially negatively affect the improvement of BCVA [20, 34, 39, 40, 5360]. In the RCT, Ripandelli et al. demonstrated no difference in BCVA between the groups until 12 months after ERM surgery and provided more substantial evidence of BCVA change at 12 months follow-up [20]. Thus, we cannot draw conclusions regarding the postoperative BCVA between these two different surgery methods in the current limited evidence-based context. More precise and controlled confounding factors and longer RCTs are necessary to confirm the visual outcomes post vitrectomy with and without ILM peeling surgery for the treatment of idiopathic ERM disease. Furthermore, our meta-analysis revealed that VA improvement rates were not significantly different between the two groups. The improvement of BCVA in patients undergoing vitrectomy with and without ILM peeling were similar. We performed TSA and determined that even if more case numbers are enrolled in the trials, the result remains nonsignificantly different between the with and without ILM groups. The lack of differences in VA improvement between the groups is likely related to the broad definition of ≥2 Snellen lines in each trial. This may have resulted from variations in the VA improving range in each case. This outcome resulted in no comparison of the surgical efficacy between the two groups. We can only consider vitrectomy with and without ILM peeling as two beneficial surgical methods for treating patients with idiopathic ERM that result in postoperative VA outcome improvement.

Several limitations of the present meta-analysis should be acknowledged. First, most of the studies available for this meta-analysis were retrospective studies; hence, evident selection bias and observer bias with regard to the adoption of the operative approach are possible. Second, as is known, we relied on the tabulated data for the meta-analysis, rather than on individual patient data. However, meta-analyses have increased power compared with individual studies, and provide more accurate confidence intervals. Third, all the enrolled trials did not prove the use of single-peeling or double-peeling procedures with pathological tissue proof. Some errors in ERM peeling or ILM peeling depend on individual experiences. Gaudric et al. have histologically compared functional and anatomical results between single-and double-peeling groups [15]. It may resolve the variations in peeling techniques among different surgeons. However, the problem of intersurgeon variability, including surgical duration and peeling technique, and different vitrectomy gauge usage were difficult to solve despite being commonly encountered. The difference of the microincision vitrectomy surgery (MIVS) had been reported faster visual recovery and less postoperative inflammation in smaller gauge MIVS observed in days to weeks follow-up duration [6163]. However, our visual outcomes were long term result rather than short term recovery. Our BCVA outcomes and VA improvement rate may not be interfered with the different MIVS procedures bias. Fourth, the confounding factor, postoperative cataract formation, was not eliminated in most trials in our meta-analysis. Different ratios of preoperative patients with phakia and aphakia were observed among different trials. Patients with phakia may exhibit lens opacity, thereby influencing VA following ERM surgery. Additional RCTs with strict inclusion criteria should be conducted to support the evidence in discussion of post BCVA in ERM surgery. Furthermore, the method of using ICG stain assistance for ILM peeling varied in each trial. To date, the toxicity of different stain durations and osmolality toward retinal tissue remains controversial and uncertain [6466]. Some reports revealed ICG might lead to optic nerve atrophy in the long-term and persistent visual field defects [6769]. We supposed that some effects of differences in stain usage on the outcomes of BCVA of each trial are possible. However, it still lacked of strong evidence to support the association between ICG toxicity and the expression of BCVA or CRT outcomes. Additionally, the different measurement of the visual acuity, the different definition of the ERM recurrence and the different OCT machine with different protocol measuring CRT may all cause the bias. Finally, the adequate duration of follow-up time to reveal the effect of operation was unknown. Our meta-analysis included follow-up times from at least 3 months up to 41.9 months. Although in the meta-regression, our analysis of the trials in the meta-analysis resulted in nonsignificant influences on the six main outcomes on the follow-up duration, we believe that higher numbers of follow-up trials of a longer duration are required to reveal the actual outcomes; this applies particularly to the recurrence rate and CRT because iatrogenic effects, such as residual ERM and postoperative retinal edema, influence the outcome of surgery.

In conclusion, our meta-analysis provides important evidence for exploring vitrectomy with or without ILM peeling in idiopathic ERM treatment. Vitrectomy with ILM peeling might provide higher VA improvement after long-term follow up and might reveal poorer results in CRT reduction than does vitrectomy without ILM peeling. ERM recurrence rate is significantly lower after vitrectomy with ILM peeling than after vitrectomy without ILM peeling according to currently limited data. Further studies are necessary to establish the optimal visual and anatomical outcome of this surgery. Large, prospective randomized clinical trials are necessary to draw valid results and confirm our current conclusion.

Supporting information

S2 Table. Search strategies and detailed records.


S1 Fig. Funnel plots of the six main outcomes.


Author Contributions

  1. Conceptualization: W-CC.
  2. Data curation: W-CC C-HL.
  3. Formal analysis: CL.
  4. Investigation: W-CC C-HL.
  5. Methodology: CL.
  6. Project administration: W-CC J-HL.
  7. Resources: CL T-HT.
  8. Software: CL C-HL T-LS T-HT.
  9. Supervision: J-HL.
  10. Validation: CL C-HL.
  11. Visualization: CL C-HL T-LS.
  12. Writing – original draft: W-CC.
  13. Writing – review & editing: W-CC J-HL.


  1. 1. Mitchell P, Smith W, Chey T, Wang JJ, Chang A. Prevalence and associations of epiretinal membranes. The Blue Mountains Eye Study, Australia. Ophthalmology. 1997;104(6):1033–40. pmid:9186446.
  2. 2. Ting FS, Kwok AK. Treatment of epiretinal membrane: an update. Hong Kong medical journal = Xianggang yi xue za zhi / Hong Kong Academy of Medicine. 2005;11(6):496–502. Epub 2005/12/13. pmid:16340027.
  3. 3. de Bustros S, Thompson JT, Michels RG, Rice TA, Glaser BM. Vitrectomy for idiopathic epiretinal membranes causing macular pucker. The British journal of ophthalmology. 1988;72(9):692–5. pmid:3179258;
  4. 4. Smiddy WE, Maguire AM, Green WR, Michels RG, de la Cruz Z, Enger C, et al. Idiopathic epiretinal membranes. Ultrastructural characteristics and clinicopathologic correlation. Ophthalmology. 1989;96(6):811–20; discussion 21. pmid:2740079.
  5. 5. Sebag J, Gupta P, Rosen RR, Garcia P, Sadun AA. Macular holes and macular pucker: the role of vitreoschisis as imaged by optical coherence tomography/scanning laser ophthalmoscopy. Trans Am Ophthalmol Soc. 2007;105:121–9; discusion 9–31. pmid:18427601;
  6. 6. Kishi S, Shimizu K. Oval defect in detached posterior hyaloid membrane in idiopathic preretinal macular fibrosis. Am J Ophthalmol. 1994;118(4):451–6. pmid:7943122.
  7. 7. Machemer R. [The surgical removal of epiretinal macular membranes (macular puckers) (author's transl)]. Klinische Monatsblatter fur Augenheilkunde. 1978;173(1):36–42. pmid:692034.
  8. 8. Margherio RR, Cox MS Jr., Trese MT, Murphy PL, Johnson J, Minor LA. Removal of epimacular membranes. Ophthalmology. 1985;92(8):1075–83. pmid:4047601.
  9. 9. Poliner LS, Olk RJ, Grand MG, Escoffery RF, Okun E, Boniuk I. Surgical management of premacular fibroplasia. Arch Ophthalmol. 1988;106(6):761–4. pmid:3370001.
  10. 10. Grewing R, Mester U. Results of surgery for epiretinal membranes and their recurrences. The British journal of ophthalmology. 1996;80(4):323–6. pmid:8703883;
  11. 11. Shimada H, Nakashizuka H, Hattori T, Mori R, Mizutani Y, Yuzawa M. Double staining with brilliant blue G and double peeling for epiretinal membranes. Ophthalmology. 2009;116(7):1370–6. Epub 2009/05/12. pmid:19427701.
  12. 12. Tognetto D, Zenoni S, Sanguinetti G, Haritoglou C, Ravalico G. Staining of the internal limiting membrane with intravitreal triamcinolone acetonide. Retina (Philadelphia, Pa). 2005;25(4):462–7.
  13. 13. Farah ME, Maia M, Rodrigues EB. Dyes in ocular surgery: principles for use in chromovitrectomy. American journal of ophthalmology. 2009;148(3):332–40. pmid:19477708.
  14. 14. Park DW, Dugel PU, Garda J, Sipperley JO, Thach A, Sneed SR, et al. Macular pucker removal with and without internal limiting membrane peeling: Pilot study. Ophthalmology. 2003;110(1):62–4. pmid:12511347
  15. 15. Gaudric A, Fardeau C, Goberville M, Cohen D, Paques M, Mikol J. Internal limiting membrane peeling, macular unfolding and visual outcome in idiopathic epimacular membrane surgery. Journal francais d'ophtalmologie. 1993;16(11):571–6.
  16. 16. Chang S, Gregory-Roberts EM, Park S, Laud K, Smith SD, Hoang QV. Double peeling during vitrectomy for macular pucker: the Charles L. Schepens Lecture. JAMA ophthalmology. 2013;131(4):525–30. Epub 2013/04/13. pmid:23579603.
  17. 17. Uemoto R, Yamamoto S, Aoki T, Tsukahara I, Yamamoto T, Takeuchi S. Macular configuration determined by optical coherence tomography after idiopathic macular hole surgery with or without internal limiting membrane peeling. The British journal of ophthalmology. 2002;86(11):1240–2. pmid:12386079;
  18. 18. Terasaki H, Miyake Y, Nomura R, Piao CH, Hori K, Niwa T, et al. Focal macular ERGs in eyes after removal of macular ILM during macular hole surgery. Investigative ophthalmology & visual science. 2001;42(1):229–34. pmid:11133873.
  19. 19. Uemura A, Kanda S, Sakamoto Y, Kita H. Visual field defects after uneventful vitrectomy for epiretinal membrane with indocyanine green-assisted internal limiting membrane peeling. American journal of ophthalmology. 2003;136(2):252–7. Epub 2003/07/31. pmid:12888046.
  20. 20. Ripandelli G, Scarinci F, Piaggi P, Guidi G, Pileri M, Cupo G, et al. Macular pucker: to peel or not to peel the internal limiting membrane? A microperimetric response. Retina. 2015;35(3):498–507. pmid:25158943.
  21. 21. Liu H, Zuo S, Ding C, Dai X, Zhu X. Comparison of the Effectiveness of Pars Plana Vitrectomy with and without Internal Limiting Membrane Peeling for Idiopathic Retinal Membrane Removal: A Meta-Analysis. J Ophthalmol. 2015;2015:974568. pmid:26693348;
  22. 22. Moher D, Liberati A, Tetzlaff J, Altman DG, Group P. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. PLoS Med. 2009;6(7):e1000097. pmid:19621072;
  23. 23. Wells G SB, O' connell D, Peterson J, Welch V, Losos M, et al. The Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in meta-analyses 2000 [cited 2016 15 November].
  24. 24. Higgins JP, Altman DG, Gotzsche PC, Juni P, Moher D, Oxman AD, et al. The Cochrane Collaboration's tool for assessing risk of bias in randomised trials. BMJ. 2011;343:d5928. pmid:22008217;
  25. 25. Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ. 2003;327(7414):557–60. pmid:12958120;
  26. 26. Sterne JA, Sutton AJ, Ioannidis JP, Terrin N, Jones DR, Lau J, et al. Recommendations for examining and interpreting funnel plot asymmetry in meta-analyses of randomised controlled trials. BMJ. 2011;343:d4002. pmid:21784880.
  27. 27. Egger M, Davey Smith G, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ. 1997;315(7109):629–34. pmid:9310563;
  28. 28. Viechtbauer W. Conducting meta-analyses in R with the metafor package. Journal of Statistical Software. 2010;36(3):1–48.
  29. 29. Schwarzer G. Pakage "meta": General Package for Meta-Analysis. 2016.
  30. 30. Kwok A, Lai TY, Yuen KS. Epiretinal membrane surgery with or without internal limiting membrane peeling. Clinical & experimental ophthalmology. 2005;33(4):379–85. Epub 2005/07/22. pmid:16033350.
  31. 31. Mason JO Iii, Roberts B, Emond TL, Feist RM, McGwin G Jr. Visual outcomes and recurrence rates of macular pucker surgery with and without internal limiting membrane peeling. Clinical and Surgical Ophthalmology. 2006;24(3):93+6–8.
  32. 32. Lee JW, Kim IT. Outcomes of idiopathic macular epiretinal membrane removal with and without internal limiting membrane peeling: A comparative study. Japanese journal of ophthalmology. 2010;54(2):129–34. pmid:20401561
  33. 33. Pournaras CJ, Emarah A, Petropoulos IK. Idiopathic macular epiretinal membrane surgery and ILM peeling: Anatomical and functional outcomes. Seminars in ophthalmology. 2011;26(2):42–6. pmid:21469962
  34. 34. Oh HN, Lee JE, Kim HW, Yun IH. Clinical outcomes of double staining and additional ILM peeling during ERM surgery. Korean journal of ophthalmology: KJO. 2013;27(4):256–60. Epub 2013/08/03. pmid:23908571;
  35. 35. Ahn SJ, Ahn J, Woo SJ, Park KH. Photoreceptor change and visual outcome after idiopathic epiretinal membrane removal with or without additional internal limiting membrane peeling. Retina (Philadelphia, Pa). 2014;34(1):172–81. Epub 2013/06/08. pmid:23743638.
  36. 36. Reilly G, Melamud A, Lipscomb P, Toussaint B. Surgical outcomes in patients with macular pucker and good preoperative visual acuity after vitrectomy with membrane peeling. Retina (Philadelphia, Pa). 2015;35(9):1817–21.
  38. 38. Chuang LH W N, Chen YP, Hwang YS, Chen KJ, Yeung L, Wu WC, Chen TL, Lai CC. Comparison of visual outcomes after epiretinal membrane surgery. Taiwan Journal of Ophthalmology. 2012;2:56–9.
  39. 39. Kim TW S S, Chung H, Yu HG. Internal limiting membrane peeling in surgical treatment of macular epiretinal membrane. J Korean Ophthalmol Soc. 2005;46(6):989–94.
  40. 40. Liu YL MT . Epiretinal membrane surgery with or without internal limiting membrane peeling. Chinese Journal of Practice Ophthalmology. 2005;23(No.12):1278–80.
  41. 41. Khokhar AR, Faiz-ur-Rab K, Akhtar HU. Management of macular pucker. Journal of the College of Physicians and Surgeons Pakistan. 2003;13(12):701–3. pmid:15569556
  42. 42. Amouyal F, Shah SU, Pan CK, Schwartz SD, Hubschman JP. Morphologic features and evolution of inner retinal dimples on optical coherence tomography after internal limiting membrane peeling. Retina (Philadelphia, Pa). 2014;34(10):2096–102. pmid:24859477.
  43. 43. Ito Y, Terasaki H, Takahashi A, Yamakoshi T, Kondo M, Nakamura M. Dissociated optic nerve fiber layer appearance after internal limiting membrane peeling for idiopathic macular holes. Ophthalmology. 2005;112(8):1415–20. pmid:16061095.
  44. 44. Tadayoni R, Paques M, Massin P, Mouki-Benani S, Mikol J, Gaudric A. Dissociated optic nerve fiber layer appearance of the fundus after idiopathic epiretinal membrane removal. Ophthalmology. 2001;108(12):2279–83. pmid:11733271.
  45. 45. Spaide RF. "Dissociated optic nerve fiber layer appearance" after internal limiting membrane removal is inner retinal dimpling. Retina. 2012;32(9):1719–26. pmid:23007669.
  46. 46. Mitamura Y, Ohtsuka K. Relationship of dissociated optic nerve fiber layer appearance to internal limiting membrane peeling. Ophthalmology. 2005;112(10):1766–70. pmid:16095706.
  47. 47. Mitamura Y, Suzuki T, Kinoshita T, Miyano N, Tashimo A, Ohtsuka K. Optical coherence tomographic findings of dissociated optic nerve fiber layer appearance. Am J Ophthalmol. 2004;137(6):1155–6. pmid:15183817.
  48. 48. Wollensak G, Spoerl E. Biomechanical characteristics of retina. Retina (Philadelphia, Pa). 2004;24(6):967–70. pmid:15579999.
  49. 49. Michels M, Lewis H, Abrams GW, Han DP, Mieler WF, Neitz J. Macular phototoxicity caused by fiberoptic endoillumination during pars plana vitrectomy. Am J Ophthalmol. 1992;114(3):287–96. pmid:1524116.
  50. 50. Kuhn F, Morris R, Massey M. Photic retinal injury from endoillumination during vitrectomy. Am J Ophthalmol. 1991;111(1):42–6. pmid:1985488.
  51. 51. van den Biesen PR, Berenschot T, Verdaasdonk RM, van Weelden H, van Norren D. Endoillumination during vitrectomy and phototoxicity thresholds. Br J Ophthalmol. 2000;84(12):1372–5. pmid:11090475;
  52. 52. Falkenstein IA, Cochran DE, Azen SP, Dustin L, Tammewar AM, Kozak I, et al. Comparison of visual acuity in macular degeneration patients measured with snellen and early treatment diabetic retinopathy study charts. Ophthalmology. 2008;115(2):319–23. pmid:17706288;
  53. 53. Goto M, Nishimura A, Shirao Y. Scanning Laser Ophthalmoscopic Microperimetry on Idiopathic Epiretinal Membrane and Vitreomacular Traction Syndrome. Japanese journal of ophthalmology. 2001;45(1):115. pmid:11341904.
  54. 54. Pilli S, Lim P, Zawadzki RJ, Choi SS, Werner JS, Park SS. Fourier-domain optical coherence tomography of eyes with idiopathic epiretinal membrane: correlation between macular morphology and visual function. Eye (London, England). 2011;25(6):775–83. pmid:21436847;
  55. 55. Zhou J, Liu W, Li Q, Li X, Hao J, Wang XL. [A comparative study of macular thickness and function before and after idiopathic macular epiretinal membrane surgery]. [Zhonghua yan ke za zhi] Chinese journal of ophthalmology. 2012;48(2):119–23. pmid:22490946.
  56. 56. Mayer WJ, Vogel M, Neubauer A, Kernt M, Kampik A, Wolf A, et al. Pars Plana Vitrectomy and internal limiting membrane peeling in Epimacular membranes: Correlation of function and morphology across the macula. Ophthalmologica Journal international d'ophtalmologie International journal of ophthalmology Zeitschrift fur Augenheilkunde. 2013;230(1):9–17. pmid:23711891
  57. 57. Ozeki-Jingami Y, Ooto S, Hangai M, Tsujikawa A, Yoshimura N. Features and surgical outcome of retinal nerve fiber layer alterations associated with epiretinal membrane. Retin Cases Brief Rep. 2013;7(3):198–203. pmid:25391105.
  58. 58. Grimbert P, Lebreton O, Weber M. [Optical coherence tomography and microperimetry after internal limiting membrane peeling for epiretinal membrane]. Journal francais d'ophtalmologie. 2014;37(6):434–41. Epub 2014/05/06. pmid:24792397.
  59. 59. Dal Vecchio M, Lavia C, Nassisi M, Grignolo FM, Fea AM. Microperimetric Assessment after Epiretinal Membrane Surgery: 4-Year Follow-Up. J Ophthalmol. 2016;2016:7030791. pmid:27088008;
  60. 60. Deltour JB, Grimbert P, Masse H, Lebreton O, Weber M. DETRIMENTAL EFFECTS OF ACTIVE INTERNAL LIMITING MEMBRANE PEELING DURING EPIRETINAL MEMBRANE SURGERY: Microperimetric Analysis. Retina (Philadelphia, Pa). 2016.
  61. 61. Akcay BI, Uyar OM, Akkan F, Eltutar K. Outcomes of 23-gauge pars plana vitrectomy in vitreoretinal diseases. Clin Ophthalmol. 2011;5:1771–6. pmid:22267910;
  62. 62. Kim M, Park YS, Lee DH, Koh HJ, Lee SC, Kim SS. Comparison of Surgical Outcome of 23-Gauge and 25-Gauge Microincision Vitrectomy Surgery for Management of Idiopathic Epiretinal Membrane in Pseudophakic Eyes. Retina. 2015;35(10):2115–20. pmid:25978731.
  63. 63. Rizzo S, Genovesi-Ebert F, Murri S, Belting C, Vento A, Cresti F, et al. 25-gauge, sutureless vitrectomy and standard 20-gauge pars plana vitrectomy in idiopathic epiretinal membrane surgery: a comparative pilot study. Graefes Arch Clin Exp Ophthalmol. 2006;244(4):472–9. pmid:16421743.
  64. 64. Da Mata AP, Burk SE, Riemann CD, Rosa RH Jr., Snyder ME, Petersen MR, et al. Indocyanine green-assisted peeling of the retinal internal limiting membrane during vitrectomy surgery for macular hole repair. Ophthalmology. 2001;108(7):1187–92. pmid:11425673.
  65. 65. Kwok AK, Lai TY, Yew DT, Li WW. Internal limiting membrane staining with various concentrations of indocyanine green dye under air in macular surgeries. American journal of ophthalmology [Internet]. 2003; 136(2):[223–30 pp.]. Available from:
  66. 66. Lai TY, Kwok AK, Au AW, Lam DS. Assessment of macular function by multifocal electroretinography following epiretinal membrane surgery with indocyanine green-assisted internal limiting membrane peeling. Graefe's archive for clinical and experimental ophthalmology = Albrecht von Graefes Archiv fur klinische und experimentelle Ophthalmologie. 2007;245(1):148–54. Epub 2006/05/18. pmid:16705448.
  67. 67. Haritoglou C, Gandorfer A, Gass CA, Schaumberger M, Ulbig MW, Kampik A. The effect of indocyanine-green on functional outcome of macular pucker surgery. Am J Ophthalmol. 2003;135(3):328–37. pmid:12614750.
  68. 68. Haritoglou C, Gandorfer A, Gass CA, Schaumberger M, Ulbig MW, Kampik A. Indocyanine green-assisted peeling of the internal limiting membrane in macular hole surgery affects visual outcome: a clinicopathologic correlation. Am J Ophthalmol. 2002;134(6):836–41. pmid:12470751.
  69. 69. von Jagow B, Hoing A, Gandorfer A, Rudolph G, Kohnen T, Kampik A, et al. Functional outcome of indocyanine green-assisted macular surgery: 7-year follow-up. Retina. 2009;29(9):1249–56. pmid:19934819.