Secondary Logo

Journal Logo



Do We Gain by Reoperating?—A Systematic Review and Meta-analysis

Reid, Gerard A. MB, BCh, BAO, MSc*; McDonagh, Niamh BSc, PGCE; Wright, David M. BSc, MSc, PhD; Yek, John T. O. BBMed, MChD; Essex, Rohan W. MBBS, MBiostat, FRANZCO§; Lois, Noemi MD, PhD, FRCS (Ed), FRCOphth*,†

Author Information
doi: 10.1097/IAE.0000000000002564
  • Open

An idiopathic full-thickness macular hole (iFTMH) represents a defect of all neurosensory retinal layers involving the fovea and can result in metamorphopsia and reduced central vision.1 The Beaver Dam Study found an FTMH prevalence of 0.3% in the population; the risk increased with age with a prevalence of 0.7% among persons aged 75 years and older and zero prevalence in the 43- to 54-year age group.2 Optical coherence tomography–guided examination at 20-year follow-up of the Beaver Dam population found an FTMH prevalence of 0.4%.3 There is greater incidence among women with a population-based study finding an incidence of 10.9 per 100,000 population per year in women compared with 4.3 per 100,000 population per year in men.4

Idiopathic full-thickness macular holes result from changes at the vitreomacular interface, which, in turn, lead to perifoveal cortical vitreous traction.5 Clinical observations based on slit-lamp examination allowed Gass6 to classify iFTMHs in four stages, based on the size of the macular hole and the status of the posterior vitreous, whether attached or detached. Using optical coherence tomography, the International Vitreomacular Traction Study Group proposed a classification for macular holes in terms of size (i.e., small, medium, or large), presence or absence of vitreomacular traction, and cause (primary [idiopathic] or secondary [subsequent to other anomalies, such as myopia or after trauma]).7

If untreated, most iFTMHs will progress in size and grade with increasing central visual loss.8 Closure of FTMHs surgically was first described by Kelly and Wendel9 who in 1991 aimed to reattach the retina around the macular hole by performing pars plana vitrectomy, peeling of epiretinal membranes at the macula, gas tamponade, and face-down positioning for 1 week after surgery. In subsequent years, internal limiting membrane (ILM) peeling gained popularity as an adjuvant maneuver to macular hole surgery (reviewed by Abdelkader and Lois10). Internal limiting membrane peeling is now used routinely for the treatment of most iFTMH; a systematic review and individual participant data (IPD) meta-analysis demonstrated superiority of ILM peeling versus no peel for the treatment of iFTMHs of any size both in terms of macular hole closure and visual acuity improvement.11,12 Thus, ILM peeling increased the odds of primary and final macular hole closure (odds ratio 9.27, 95% confidence interval [CI] 4.98–17.24 and odds ratio 3.99, CI 1.63–9.75, respectively); this applied to any stage of macular hole, from 2 to 4.11,12 The chance for an improvement in best-corrected visual acuity (BCVA) at 3 months after surgery was also increased by peeling the ILM (mean difference −0.09, 95% CI −0.17 to −0.02). Differences in BCVA were not observed at 6 or 12 months; this, however, could be explained by the fact that most patients in the no ILM peel group would have received by then ILM peeling given that, following ethical considerations and standard clinical practice, ILM peeling was allowed in the nonpeel group in all randomized controlled trials included if the hole had not closed following the primary procedure. Only one study included in this systematic review and IPD meta-analysis evaluated near vision13; this study did not find differences in near vision between ILM peeling and no peel groups.

Published rates of macular hole closure following a primary surgical procedure range from 70% to 100%.10,13–16 A study using data from the national U.K. electronic database17 found that over an 8-year period, further vitreoretinal surgery had been undertaken in 4.2% of primary macular hole repairs. Furthermore, it has been recognized that iFTMHs which initially closed after the primary surgery may later reopen.18 For patients in whom primary macular hole closure failed and for those in whom macular hole reopening occurs, repeated surgery could be considered.

The aim of this study was to systematically review available evidence on functional and anatomical outcomes after repeated surgery 1) after primary surgical failure and 2) after initial closure and subsequent macular hole reopening. This information would be useful to clinicians for the counseling of patients before repeated surgery and for patients to make informed decisions with regards to proceeding with further treatment.


Study Eligibility Criteria, Definitions, and Outcome Measures

All study designs were considered eligible for studies presenting outcomes of reoperation, whichever the procedure, for iFTMH that failed to close (FTC) following the primary procedure or that initially closed but later reopened (RO), with the exception of case reports and small case series including less than five cases. Only articles in English were eligible for inclusion. Studies presenting data on both FTC and RO iFTMHs were included only if separate data for these two groups were provided. Studies evaluating outcomes of non-iFTMHs (e.g., traumatic, myopic, etc) were excluded. Studies presenting data with both non-iFTMHs and iFTMHs were included only if data on the iFTMH group were provided separately. Abstracts of unpublished studies were not included.

For the purpose of this review, macular hole closure was defined as full closure of the macular hole with complete apposition of the hole margins. Flattening of the cuff of subretinal fluid around the hole alone was not considered hole closure if there was still a neurosensory retinal defect present. Studies in which this definition of macular hole closure was not used were excluded (e.g., the following studies; Ezra et al, Christmas et al, and Lappas et al).19–21

The following outcomes were determined: 1) anatomical success rate after secondary surgery, defined as complete apposition of the macular hole margins with no remaining neurosensory retinal defect; 2) postoperative BCVA and BCVA improvement comparing presecond surgery and final BCVA values; 3) any reported intraoperative and postoperative complications; and 4) any patient-reported outcomes (PROs). Outcomes were measured at 6, 12, 18, and ≥24 months (as available); a window of ±3 months was allowed for each time point. Best-corrected visual acuity values were transformed to the LogMAR scale where necessary for analysis.

Search Strategy

Systematic searches were conducted in PubMed (1946–2018) and Cochrane Library databases (1999–2018) with no date restrictions (latest search updated January 1, 2018). Two literature searches were conducted to identify studies of 1) iFTMHs that FTC following the primary procedure (FTC group) and 2) iFTMHs that initially closed but reopened thereafter (RO group).

Search terms combined the following key words

  1. To identify iFTMHs that FTC after primary surgery: “Macular hole” AND “Persistent” Or “Refractory” “Macular hole surgery” and “Primary failure” “Failure of macular hole surgery” “Idiopathic full-thickness macular hole” AND “persistent,” “Gas injection” and “Macular hole surgery failure,” “Primary macular hole” AND “Vitrectomy” AND “Surgical failure” “Primary macular hole” AND “Reoperation” AND “Vitrectomy” “Macular hole” AND “Resurgery” “Primary macular hole” AND “Vitrectomy” AND “Reoperation” “Unclosed macular hole” AND “Reoperation” “Persistent” AND “macular hole” AND “Vitrectomy”;
  2. To identify iFTMHs that initially closed but reopened after primary surgery: “Reopened macular hole” “Reopened macular hole” AND “Vitrectomy” AND “Reoperation” “Primary macular hole” AND “Vitrectomy” AND “Failure” “Primary macular hole” AND “Reopen” AND “Surgery” “Reopened macular hole” AND “Reoperation” AND “Successful Repair” “Macular hole” AND “Initial closure” AND “Vitrectomy” “Primary macular hole” AND “Surgery” AND “Reoperation” “Macular hole surgery” AND “Initial success” AND “Repair” “Primary macular hole” AND “Surgery” AND “Recurrence.”

Reference lists of identified studies were also reviewed; additional relevant studies identified were also included.

Selection of Studies and Data Extraction

One reviewer (N.M.) conducted searches and examined studies for eligibility, and subsequently, a second reviewer (G.A.R.) repeated the searches and consideration of eligibility; a librarian experienced with undertaking systematic searches provided guidance, training, and support to these reviewers. A third investigator (N.L.) was consulted regarding disagreements which were solved through discussion. Thus, two reviewers (G.A.R. and N.M.) independently reviewed the titles identified, and articles were classified as ineligible or potentially eligible; duplicate articles were removed. Reviewers obtained and screened abstracts of potentially eligible studies in a similar manner, classifying them as ineligible or potentially eligible. Three reviewers (G.A.R., N.M., and N.L.) obtained and reviewed full-text articles of potentially eligible studies. Full-text articles presenting studies that did not meet the inclusion criteria were excluded.

One reviewer (G.A.R. or N.M.) extracted the following data from included studies, and a second reviewer (N.L.) corroborated these data: author/year of publication, study design, number of patients receiving intervention, patient demographics (age and sex), characteristics of the iFTMHs, details of both primary and repeated surgical procedure, proportion of eyes with anatomical closure after repeated surgery and mean BCVA preoperatively and postoperatively. Information on number and type of intraoperative and postoperative complications, as well as PRO scores, where available, was also recorded at the follow-up intervals set in the outcomes section below.

We contacted the authors of two of the largest series identified (Valldeperas and Wong and Yek et al)22,23 to enquire about the availability of IPD and their willingness to share it for the purpose of this review.

We categorized studies investigating outcomes after reintervention for iFTMHs as either those addressing iFTMH that FTC or those addressing iFTMH that RO after initial closure.

No detailed assessment of the quality of identified studies (e.g., using the Cochrane Risk of Bias Tool) was undertaken for the purpose of this review.

Data Synthesis and Analysis

Two types of data were available for synthesis: aggregate data consisting of summary statistics reported in the reviewed studies and IPD either extracted from reviewed studies or subsequently obtained from study authors. IPD was used in preference to aggregate data wherever possible, so that the influence of several covariates that varied within studies (follow-up duration and ILM peeling at first and second surgeries) on outcomes could be assessed.

Two outcomes were considered for meta-analysis: anatomical closure of iFTMHs and visual acuity. There was considerable variation in BCVA preoperatively and postoperatively among studies, and Standard Deviations (SD) for changes in BCVA were often unreported. Therefore, a binary indicator of improvement in BCVA by two or more Snellen lines or equivalent (i.e., 10 letters, 0.2 LogMAR) was used as the visual acuity outcome.

For the FTC group, an initial analysis was conducted to determine whether either outcome was associated with follow-up duration or whether ILM peeling was conducted at either or both surgeries. Meta-regressions were conducted using IPD alone because it provides greater statistical power to detect patient-level associations. This was particularly relevant for follow-up duration as mean values demonstrated considerable variation within studies (Table 1). These models revealed no statistically significant associations between visual acuity or anatomical hole closure and any of the following covariates: follow-up duration, ILM peeling at first surgery, ILM peeling at second surgery (data not shown, available on request). Therefore, simple intercept only models were selected for the main analysis. For the meta-analysis, measurements at the point at which they were reported (e.g., 6, 12 months, etc.) were used. Follow-up time was not included in any of the final meta-analysis models (given the lack of evidence of associations in the aggregate or individual-level data between follow-up time and either of the outcomes of interest, as explained above).

Table 1-a
Table 1-a:
Characteristics, Surgical Techniques and Outcome Measures for Studies Included in the FTC Group
Table 1-b
Table 1-b:
Characteristics, Surgical Techniques and Outcome Measures for Studies Included in the FTC Group
Table 1-c
Table 1-c:
Characteristics, Surgical Techniques and Outcome Measures for Studies Included in the FTC Group

Meta-analysis was conducted for each outcome combining both aggregate data and IPD. Summary statistics were calculated for each study reporting IPD, and all studies were included in a single model. Heterogeneity among studies was calculated using the I2 statistic. Where there was evidence of high heterogeneity, a random-effects model was fitted; where heterogeneity was low, a fixed-effects model was used. Meta-analysis was performed using the metafor package in R version 3.4.24


Literature Search

A total of 1,023 study titles were identified across both literature searches (Figure 1) (715 titles in the FTC group and 310 titles in the RO group; 2 titles were identified in both of literature searches and referred to 2 studies that presented data for both FTC and RO groups separately), and 977 titles (677 titles in FTC and 300 titles in RO) were excluded as duplicates or not relevant to the search criteria. Forty-six abstracts were retrieved as potentially eligible; eight abstracts were excluded as not meeting inclusion criteria. Two further articles from the group where iFTMH failed to close were identified through the reference lists when full articles were reviewed and one additional article was forwarded by the author just before its publication after presentation of some of the current work at a meeting. Thus, a total of 41 full-text articles (35 in FTC and 8 in RO, with 2 articles presenting data for both FTC and RO groups) were reviewed. Thirteen full-text articles (9 FTC and 4 RO) were excluded for the following reasons: case report of <5 cases; definition of FTMH closure did not meet criteria; results for FTC and RO groups not distinct; diagnosis other than iFTMHs; and not second FTMH procedure. Twenty-eight articles were eligible and included, dating from 1993 to 2017. Twenty-six studies were included in the FTC group, and four studies were included in the RO group (Figure 1). As stated above, two studies (Rao et al and Valldeperas and Wong) provided separate data for both the FTC group and RO groups and were included in both groups.22,25

Fig. 1
Fig. 1:
Flow diagram of studies identified on iFTMHs that FTC after the primary surgical procedure (FTC, left) and those that initially closed which later reopened (RO, right). *Two studies identified through reference lists of included studies, one study identified after presentation of the systematic review at the Fernie Vitreoretinal Meeting, Fernie, Canada, January 26, 2017, and following discussion with the author.

In the FTC group, there was one randomized controlled trial26 and one pilot study,27 both of which were prospective and 24 case series, 3 of which were prospective and 21 retrospective (Table 1). The RO group consisted of four studies all of which were retrospective case series (Table 2).

Table 2
Table 2:
Characteristics, Surgical Techniques, and Outcome Measures for Studies Included in the (RO) group

The populations of included studies had an overall predominance of women versus men (313 vs. 149), with mean ages ranging from 60 to 75 years (weighted mean 68 years).

Outcomes: Failed to Close Group

Macular hole closure

Data on anatomical outcomes after repeat surgery to repair iFTMHs which FTC after the primary surgery were collected from 26 identified studies and included 520 eyes. Meta-analysis was conducted combining aggregate data (13 studies, 331 eyes)22,25,26,28–37 and IPD (13 studies, 189 eyes).23,27,38–48 There was moderate heterogeneity among studies (I2 = 54.9), and so, a random-effects meta-analysis was used to estimate the proportion of patients with anatomical closure following a second surgical procedure. The pooled estimated proportion of macular holes that closed after repeat surgery in FTC eyes was 0.78 (95% CI 0.71–0.84, Figure 2).

Fig. 2
Fig. 2:
Proportion of iFTMHs that closed after repeated surgery. Estimates and 95% CIs for 26 studies along with overall estimate from meta-analysis and I2 measure of among study heterogeneity. Point size is proportional to weighting of individual studies in the meta-analysis.

The proportion of eyes with anatomical closure after repeated surgery ranged from 44% to 100% (weighted mean 75%, 389 of 520 eyes, 26 studies) (Table 1). Anatomical closure at a mean follow-up of 6 months ranged from 60% to 100% (weighted mean 84%, 62/74 eyes; 7 studies),31,36,38,43,45,47,48 at 12 months ranged from 52% to 100% (weighted mean 76%, 166/219 eyes; 12 studies),22,26–30,37,39,41,42,44,46 at 18 months, mean follow-up ranged from of 68% to 100% (weighted mean 76%, 28/37 eyes; 2 studies),32,40 and at ≥24 months, mean follow-up (range 24–58 months) mean iFTMH closure ranged from 62% to 81% (weighted mean 72%, 113/156 eyes; 3 studies).23,25,33

Best-corrected visual acuity

Data on BCVA were provided in 23 studies (475 eyes) of the FTC group; however, there was considerable variation among studies in BCVA both before and after surgery, indicating heterogeneity in terms of measured outcomes among studies and patient characteristics. Most studies reported the mean and SD of BCVA measurements before and after the second surgery along with the mean change in BCVA. However, the SD of the change in BCVA was frequently not reported; so, it was not possible to use change in BCVA as the outcome of a meta-analysis (meta-analysis requires measures of both central tendency and dispersion for each study). Therefore, it was decided to use the categorical variable “improvement in BCVA by two Snellen lines” (or 10 letters, 0.2 LogMAR) as the outcome for meta-analysis as this measure was reported or could be derived for a greater number of studies. Data on improvement in BCVA by 2 Snellen lines were available from 17 studies, and meta-analysis was conducted combining aggregate data (4 studies, 110 eyes) and IPD (13 studies, 189 eyes). Due to high heterogeneity among studies (I2 = 64.3%), a random-effects model was fitted. The pooled estimate of the proportion of patients with a ≥2 Snellen line improvement in BCVA was 0.58 (CI 0.45–0.71) (approx. 6/24 Snellen equivalent) (Figure 3). Of the 189 eyes for which IPD was available, visual acuity of 6/12 Snellen equivalent or better was achieved in 15% (28 eyes, 13 studies) (Table 1).

Fig. 3
Fig. 3:
Proportion of iFTMHs exhibiting a ≥2 Snellen line gain in BCVA after repeated surgery. Estimates and 95% CIs for 17 studies along with overall estimate from meta-analysis and I2 measure of among study heterogeneity. Point size is proportional to weighting of individual studies in the meta-analysis.

Mean BCVA from each study before the secondary surgical procedure ranged from 1.52 to 0.84 logMAR (weighted mean 1.07 [approx. 6/72 Snellen equivalent]; 460 eyes); while postoperative mean BCVA ranged from 1.18 to 0.28 logMAR (weighted mean 0.76 [approx. 6/36 Snellen equivalent]; 460 eyes). Best-corrected visual acuity before the second surgery was not provided by 2 studies30,32 who instead provided mean BCVA before the first surgery and after the second procedure.

For the studies where preoperative BCVA was reported (n = 16), the mean improvement from presecond surgery to final postoperative BCVA ranged from 0.06 to 0.56 (weighted mean 0.31, approx. 6/12 Snellen equivalent) LogMAR (460 eyes). Best-corrected visual acuity gain by ≥2 Snellen lines was presented in 17 of 26 studies (246 eyes) with a range of 10% to 100% (weighted mean 41%) achieving this improvement (Table 1). Rizzo et al and Yek et al used a three Snellen line improvement and found 83% (19 of 23 eyes) and 22% (11 of 53 eyes) achieving at least three Snellen lines of BCVA improvement respectively.23,28

Best-corrected visual acuity at 6 months postop was reported in 6 studies; preop and 6-month mean BCVA ranged from 1.52 to 0.84 and 1.10 to 0.38 LogMAR, respectively (weighted mean 0.97–0.65, approx. 6/60–6/24 Snellen equivalent; 67 eyes).31,38,43,45,47,48 Best-corrected visual acuity at 12 months postop was reported in 12 studies; preop and 12-month mean BCVA ranged from 1.40 to 0.94 and 1.18 to 0.41 LogMAR, respectively (weighted mean 1.11–0.78, approx. 6/75–6/36 Snellen equivalent; 213 eyes).22,26–30,37,39,41,42,44,46 Mean BCVA at 18-month follow-up was recorded by Hillenkamp et al40 preop and postop at 1.10 and 1.04 LogMAR (6/75–6/60). At ≥24-month follow-up, preop and postop mean BCVA was reported in 2 studies and ranged from 1.26 to 1.00 and 0.78 to 0.60 LogMAR, respectively (weighted mean 1.11–0.68, approx. 6/75–6/30 Snellen equivalent; 127 eyes).23,33

Surgical techniques

All surgical techniques used in identified studies have been summarized in Table 1. Of all eyes included in the review, ILM peeling was undertaken at the primary surgery in 419 of 520 eyes (81%) (23 of 26 studies, 2005–2017).

During the second surgery, 101 eyes (7 studies) underwent further ILM peeling (note that in Valldeperas and Wong,22 Hejsek et al,30 Moisseiev et al,37 and Yek et al,23 the ILM peeling was performed for the first time in 96, 17, 14, and 6% of eyes, respectively, during second surgery). Fluid–gas exchange alone was performed as a secondary procedure in 108 eyes.25,34,35,42–44 Autologous ILM graft transplants were used in 30 eyes (3 studies),39,45,46 and one study38 used lens capsule transplantation (9 eyes). Additional techniques described were as follows: autologous blood products, used in 183 eyes; transforming growth factor-beta-2, used in 10 eyes; laser photocoagulation, used in 11 eyes; and an induced macular detachment in 15 eyes (Table 1).

The endotamponade agents used included air in 1 eye (0.2%), C3F8 in 250 eyes (48%), C2F6 in 77 eyes (15%), SF6 in 79 eyes (15%), and silicone oil (SO) in 56 eyes (11%); in 57 eyes, the type of tamponade used was not specified.


Data on complications were recorded in 17 studies (Table 1).

Progression of cataract was reported in 3 studies (35 phakic eyes),27,41,44 with a weighted mean of 69% (at a median follow-up of 12 months). Cataract extraction was reported in 6 studies,22,23,28,37,44 with a weighted mean of 59%. At a median follow-up of 13 months, retinal detachment was reported in 2 studies and occurred in 4 of 380 eyes (1.5%).33,40 Transient intraocular pressure (IOP)-related complications were reported in 14 of 400 eyes (3.5%) (Table 1).

Two studies25,29 did not differentiate second surgery complications between FTC and RO groups. Rao et al25 reported cataract extraction in 34 phakic eyes (94%), retinal detachments in 2 eyes (6%), and IOP complications in three eyes (8%). D'Souza et al29 report IOP complications in 17 eyes (68%).

Patient-reported outcomes

No studies provided information on PROs after repeated surgery for FTC iFTMH.

Outcomes: Reopened Group

Proportion of eyes with anatomical closure of macular hole

Meta-analysis was performed on the aggregate data sets of the 4 studies, 49 eyes22,25,49,50 in the RO group. There was low to moderate heterogeneity among the studies (I2 = 42.8), and a fixed-effects meta-analysis model was used to estimate a pooled effect size of iFTMH anatomical closure following a second surgical procedure of 0.80 (CI 0.66–0.89) (Figure 4).

Fig. 4
Fig. 4:
Proportion of iFTMHs that closed and then RO after initial surgery that subsequently closed after repeated surgery. Estimates and 95% CIs for four studies along with overall estimate from meta-analysis and I2 measure of among study heterogeneity. Point size is proportional to weighting of individual studies in the meta-analysis.

The proportion of eye with iFTMH reopening following the primary vitreoretinal procedure ranged from 1.3% to 9.2%, occurring in 51 of 1,512 eyes (weighted mean 3.4%). Of these iFHTMs, outcomes following a second surgical procedure were presented for 49 eyes. Anatomical closure after the second surgery ranged from 57% to 95% (weighted mean 80%, 39 eyes) (Table 2).

Best-corrected visual acuity

Data on BCVA were available in 2 studies (34 eyes).22,49 Meta-analysis of ≥2-line Snellen improvement following a secondary procedure was not possible in the RO group as only Valldeperas and Wong22 provided data on this outcome.

Mean BCVA before a second procedure ranged from 0.88 to 0.85 logMAR (weighted mean 0.86 [approx. 6/45 Snellen equivalent]), and postoperative ranged from 0.78 to 0.50 logMAR (weighted mean 0.61 [approx. 6/24 Snellen equivalent]) in Abbey et al and Valldeperas and Wong, respectively.22,49 The mean improvement from preoperative to postoperative levels ranged from 0.1 to 0.35 LogMAR (weighted mean 0.25, approx. 6/10 Snellen equivalent). Only Valldeperas and Wong22 included data on gain of ≥2 Snellen line BCVA which occurred in 76% (16/20 eyes) at 12 months. Mean BCVA preop and at 6 months and 12 months postop as recorded by Valldeperas and Wong was 0.85, 0.48, and 0.50 LogMAR, respectively (approx. 6/45, 6/18, and 6/19 Snellen equivalent, respectively).22,50 No study recorded BCVA at 18 months; however, Abbey et al presented preop BCVA and at ≥24-month follow-up with values of 0.88 preop and 0.78 LogMAR, respectively (approx. 6/45 and 6/36, Snellen equivalent, respectively).49

Surgical techniques

Paques et al (8 eyes) did not undertake any ILM peeling at either initial or secondary vitrectomy. An initial pars plana vitrectomy plus ILM peeling was performed by Rao et al and Abbey et al in all eyes (20 eyes), whereas Valldeperas and Wong performed only a limited number of ILM peels (2 eyes, 12%) (Table 2). During the secondary procedure, Valldeperas and Wong performed ILM peeling on 12 eyes (57%), whereas Abbey et al “stained and attempted” further ILM peel (13 eyes). Rao et al performed fluid–gas exchange only without further vitrectomy (7 eyes). Autologous platelets were used as adjuvant in Valldeperas and Wong (Table 2).


In the RO group, Valldeperas and Wong22 reported cataract surgery after second surgery in 16 of 19 phakic (84%) eyes at 12-month follow-up. Valldeperas and Wong and Abbey et al reported retinal breaks/tears in 2 eyes (10%) and 1 eye (8%), respectively.22,49

Patient-reported outcomes

No studies provided information on PROs.


Meta-analysis of probability of iFTMH anatomical closure following a second surgical procedure generated a pooled estimate of 78% (CI 71–84%) in the FTC group and 80% (CI 66–89%) in the RO group. Meta-analysis of the probability of gain in BCVA by ≥2 Snellen lines was only possible in the FTC group; this generated a pooled estimate of 58% (45–71%), with 15% of eyes achieving visual acuity of 6/12 or better at the last recorded follow-up. Based on these data, reoperation for iFTMH seems to be justified; available evidence, however, should be discussed with patients before surgery to ensure they have realistic expectations to what can be achieved following a repeated procedure. No data were provided on PROs in any of the identified studies. Thus, it is difficult to determine whether or not the benefits of further surgery were relevant to patients. It would seem important in future studies to take this into consideration and include health-related and visual-related quality-of-life testing as outcome measures.

The proportion of patients with an improvement of ≥2 Snellen lines was significantly higher in the single study reporting this outcome in the RO group (i.e., 76%, Valldeperas and Wong22) than our pooled estimate across the FTC group (58%). Valldeperas and Wong and Rao et al provided a comparison between FTC and RO in their series.22,25 Valldeperas and Wong stated that differences between the two groups (RO and FTC) were not statistically significant before the second intervention; anatomical success (95 vs. 76%, P = 0.012) and 6-month BCVA (0.48 vs. 0.92 LogMAR [6.18 vs. 6/50]) were significantly better in the RO group when compared with the FTC group, although statistically significant differences in BCVA at 12 months were no longer present (0.50 vs. 0.78 LogMAR [6.19 vs. 6/36]).22 Rao et al25 did not attempt a statistical comparison between these groups but reported anatomical success rates that were similar between the two groups (57 vs. 62%); unfortunately, visual acuity data were not presented separately.

At primary macular hole surgery, an increased period before the surgical repair51–53 and a larger preop macular hole size54,55 have been associated with worse outcomes both in terms of anatomical closure and postop BCVA. It has been suggested that an increased time between first and second surgeries may contribute to poorer visual and anatomical outcomes.29,35,37 In addition, increased size of the iFTMH may also negatively impact surgical outcomes33,37. These findings, however, were not consistent across all reviewed studies,40 and as noted by Yek et al,23 it is difficult to assess these variables in studies with small sample sizes.

There was variation among studies in the surgical maneuvers undertaken at the repeated procedure (listed in Tables 1 and 2), including the postop endotamponade agent used, the use of autologous blood products, grafting with autologous ILM or lens capsule, and macular detachment. The large variation in surgical techniques and small study sizes made it impossible to determine which ones may have contributed most to the success of the second surgery. There were multiple postoperative tamponade agents used among the studies we reviewed; however, Cillino et al26 (FTC group) compared C2F6 gas with SO in a randomized clinical trial. The SO group achieved greater anatomical closure and BCVA at 12 months. It was, however, not stated whether the ILM had been peeled at the time of the primary surgery or whether ILM peel was performed or extended during repeated surgery. Silicone oil also requires an additional surgical procedure for its removal with the potential of complications such as retinal detachment and unexplained visual loss can occur after removal of SO.40

Recently described maneuvers included the use of an autologous ILM flaps or free grafts, which are suggested to provide a scaffold for glial cell proliferation and macular hole closure.56 Inverted ILM flap can be successful in closing large iFTMHs57; however, at a secondary procedure, a free ILM graft may be required, as an ILM peel will have been previously undertaken in most cases; free ILM grafts are more challenging given the possibility of them to dislodge at the time of fluid–air exchange.58 Anatomical closure after ILM grafting was reported in 3 studies and ranged from 91% to 100% (28 eyes); functional improvement, however, varied.39,45,46 Alternatively, grafting with lens capsule has also been described with possible greater ease of manipulation intrasurgically. During repeated iFTHM surgery, this technique closed 67% of cases (6 of 9 eyes).38

This study has several limitations. Thus, most studies identified were retrospective, and many were relatively small case series. This fact limits the quality of the primary evidence identified to the lower end of the evidence hierarchy and allows for the potential entry of biases through, for example, case selection, lack of masking, attrition bias (incomplete reporting of outcome data), and reporting bias (selective reporting), among others. A further limitation of this review was the heterogeneity of studies included with multiple surgical techniques used, making it impossible to determine what may be the best surgical approach to treat FTC or RO iFTMH. On this regard, this systematic review highlights the need for high-quality studies especially those investigating the relative effectiveness of the different surgical techniques used in the repair of iFTMH that FTC or RO. Available data, however, suggest repeated macular hole surgery is beneficial in terms of anatomical closure and functional improvement. The strengths of this review include the systematic identification of studies, standardized data extraction, and the meta-analysis on anatomical and functional outcomes undertaken in the FTC group and on anatomical outcomes in the RO group, providing useful information for the counseling of patients before repeated surgery.


Our review and meta-analysis indicate that just more than half of the patients undergoing a second surgery for FTC iFTMH are likely to experience visual acuity improvement of ≥2 Snellen lines after a second surgery, with a small risk of complications. Only a small proportion of these (15%), however, may achieve very good vision (≥6/12). We found weak evidence that visual success may be higher in the RO group. Repeated surgery is likely to lead to anatomical closure of the iFTMH in a high proportion of patients with FTC and RO iFTMHs (estimated at 78 and 80%, respectively). This information is useful for the counseling of patients before surgery.


The authors thank Mr. Richard Fallis for his support to the undertaking of systematic searches for this review.


1. Ho AC, Guyer DR, Fine SL. Macular hole. Surv Ophthalmol 1998;42:393–416.
2. Klein R, Klein BE, Wang Q, Moss SE. The epidemiology of epiretinal membranes. Trans Am Ophthalmol Soc 1994;92:403–430.
3. Meuer SM, Myers CE, Klein BEK, et al. The epidemiology of vitreoretinal interface abnormalities as detected by spectral-domain optical coherence tomography: the beaver dam eye study. Ophthalmology 2015;122:787–795.
4. McCannel CA, Ensminger JL, Diehl NN, Hodge DN. Population-based incidence of macular holes. Ophthalmology 2009;116:1366–1369.
5. Gass JD. Idiopathic senile macular hole. Its early stages and pathogenesis. Arch Ophthalmol 1988;106:629–639.
6. Gass J. Reappraisal of biomicroscopic classification of stages of development of a macular hole. Am J Ophthalmol 1995;119:752–759.
7. Duker JS, Kaiser PK, Binder S, et al. The International vitreomacular traction study group classification of vitreomacular adhesion, traction, and macular hole. Ophthalmology 2013;120:2611–2619.
8. Casuso LA, Scott IU, Flynn HW, et al. Long-term follow-up of unoperated macular holes. Ophthalmology 2001;108:1150–1155.
9. Kelly NE, Wendel RT. Vitreous surgery for idiopathic macular holes. Results of a pilot study. Arch Ophthalmol 1991;109:654–659.
10. Abdelkader E, Lois N. Internal limiting membrane peeling in vitreo-retinal surgery. Surv Ophthalmol 2008;53:368–396.
11. Spiteri Cornish K, Lois N, Scott N, et al. Vitrectomy with internal limiting membrane (ILM) peeling versus vitrectomy with no peeling for idiopathic full-thickness macular hole (FTMH). Cochrane Database Syst Rev 2013;5:CD009306.
12. Spiteri Cornish K, Lois N, Scott NW, et al. Vitrectomy with internal limiting membrane peeling versus no peeling for idiopathic full-thickness macular hole. Ophthalmology 2014;121:649–655.
13. Lois N, Burr J, Norrie J, et al. Internal limiting membrane peeling versus no peeling for idiopathic full-thickness macular hole: a pragmatic randomized controlled trial. Invest Ophthalmol Vis Sci 2011;52:1586–1587.
14. Kwok AKH, Lai TYY, Wong VWY. Idiopathic macular hole surgery in Chinese patients: a randomised study to compare indocyanine green-assisted internal limiting membrane peeling with no internal limiting membrane peeling. Hong Kong Med J 2005;11:259–266.
15. Christensen UC, Kroyer K, Sander B, et al. Value of internal limiting membrane peeling in surgery for idiopathic macular hole stage 2 and 3: a randomised clinical trial. Br J Ophthalmol 2009;93:1005–1015.
16. Essex RW, Kingston ZS, Moreno-Betancur M, et al. The effect of postoperative face-down positioning and of long-versus short-acting gas in macular hole surgery: results of a registry-based study. Ophthalmology 2016;123:1129–1136.
17. Jackson TL, Donachie PHJ, Sparrow JM, Johnston RL. United Kingdom National Ophthalmology Database study of vitreoretinal surgery: report 2, macular hole. Ophthalmology 2013;120:629–634.
18. Duker JS, Wendel R, Patel AC, Puliafito CA. Late re-opening of macular holes after initially successful treatment with vitreous surgery. Ophthalmology 1994;101:1373–1378.
19. Ezra E, Aylward WG, Gregor ZJ. Membranectomy and autologous serum for the retreatment of full-thickness macular holes. Arch Ophthalmol 1997;115:1276–1280.
20. Christmas NJ, Smiddy WE, Flynn HW. Reopening of macular holes after initially successful repair. Ophthalmology 1998;105:1835–1838.
21. Lappas A, Foerster AM, Kirchhof B. Use of heavy silicone oil (Densiron-68) in the treatment of persistent macular holes. Acta Ophthalmol 2009;87:866–870.
22. Valldeperas X, Wong D. Is it worth reoperating on macular holes? Ophthalmology 2008;115:158–163.
23. Yek J, Hunyor AP, Campbell WG, et al. Outcomes of eyes with failed primary surgery for idiopathic macular hole. Ophthalmol Retina 2018;2:757–764.
24. Viechtbauer W. Conducting meta-analyses in R with the metafor package. J Stat Softw 2010;36:1–48.
25. Rao X, Wang NK, Chen YP, et al. Outcomes of outpatient fluid-gas exchange for open macular hole after vitrectomy. Am J Ophthalmol 2013;156:326–333.
26. Cillino S, Cillino G, Ferraro LL, Casuccio A. Treatment of persistently open macular holes with heavy silicone oil (Densiron 68) versus C2F6: a prospective randomized study. Retina 2016;36:688–694.
27. Ie D, Glaser BM, Thompson JT, et al. Retreatment of full-thickness macular holes persisting after prior vitrectomy: a pilot study. Ophthalmology 1993;100:1787–1793.
28. Rizzo S, Genovesi-Ebert F, Vento A, et al. Heavy silicone oil (Densiron-68) for the treatment of persistent macular holes: Densiron-68 endotamponade for persistent macular holes. Graefes Arch Clin Exp Ophthalmol 2009;247:1471–1476.
29. D'Souza MJ, Chaudhary V, Devenyi R, et al. Re-operation of idiopathic full-thickness macular holes after initial surgery with internal limiting membrane peel. Br J Ophthalmol 2011;95:1564–1567.
30. Hejsek L, Dusova J, Stepanov A, Rozsival P. Re-operation of idiopathic macular hole after failed initial surgery. Biomed Pap Med Fac Univ Palacky Olomouc Czech Repub 2014;158:596–599.
31. Dimopoulos S, William A, Voykov B, et al. Anatomical and visual outcomes of autologous thrombocyte serum concentrate in the treatment of persistent full-thickness idiopathic macular hole after ILM peeling with brilliant blue G and membrane blue dual. Acta Ophthalmol 2017;95:429–430.
32. Hagiwara A, Baba T, Tatsumi T, et al. Functional and morphologic outcomes after reoperation for persistent idiopathic macular hole. Eur J Ophthalmol 2017;27:231–234.
33. Purtskhvanidze K, Frühsorger B, Bartsch S, et al. Persistent full-thickness idiopathic macular hole: anatomical and functional outcome of revitrectomy with autologous platelet concentrate or autologous whole blood. Ophthalmologica 2018;239:19–26.
34. Modi A, Giridhar A, Gopalakrishnan M. Sulfurhexafluoride (SF6) Versus Perfluoropropane (C3F8) gas as tamponade in macular hole surgery. Retina 2017;37:283–290.
35. Patel R, Gopalakrishnan M, Giridhar A. Timing and outcome of surgery for persistent macular hole. Retina 2017;39:314–318.
36. Gurelik G, Sul S, Kiliç G, Özsaygili C. A modified foveal advancement technique in the treatment of persistent large macular holes. Ophthalmic Surg Lasers Imaging Retina 2017;48:793–798.
37. Moisseiev E, Fabian ID, Moisseiev J, Barak A. Outcomes of repeated pars plana vitrectomy for persistent macular holes. Retina 2013;33:1137–1143.
38. Chen SN, Yang CM. Lens capsular flap transplantation in the management of refractory macular hole form multiple etiologies. Retina 2016;36:163–170.
39. Dai Y, Dong F, Zhang X, Yang Z. Internal limiting membrane transplantation for unclosed and large macular holes. Graefes Arch Clin Exp Ophthalmol 2016;254:2095–2099.
40. Hillenkamp J, Kraus J, Framme C, et al. Retreatment of full-thickness macular hole: predictive value of optical coherence tomography. Br J Ophthalmol 2007;91:1445–1449.
41. Ikuno Y, Kamei M, Saito Y, et al. Photocoagulation and fluid-gas exchange to treat persistent macular holes after prior vitrectomy: a pilot study. Ophthalmol 1998;105:1411–1418.
42. Imai M, Gotoh T, Iijima H. Additional intravitreal gas injection in the early postoperative period for an unclosed macular hole treated with internal limiting membrane peeling. Retina 2005;25:158–161.
43. Iwase T, Sugiyama K. Additional gas injection after failed macular hole surgery with internal limiting membrane peeling. Clin Exp Ophthalmol 2007;35:214–219.
44. Johnson RN, McDonald HR, Schatz H, Ai E. Outpatient postoperative fluid-gas exchange after early failed vitrectomy surgery for macular hole. Ophthalmology 1997;104:2009–2013.
45. Ozdek S, Baskaran P, Karabas L, Neves PP. A modified perfluoro-n-octane-assisted autologous internal limiting membrane transplant for failed macular hole reintervention: a case series. Ophthalmic Surg Lasers Imaging Retina 2017;48:416–420.
46. Pires J, Nadal J, Gomes NL. Internal limiting membrane translocation for refractory macular holes. Br J Ophthalmol 2017;101:377–382.
47. Saeed MU, Heimann H, Wong D, Gibran SK. Heavy silicone oil tamponade after failed macular hole surgery with perfluoropropane (C3F8): a report of five cases. Graefes Arch Clin Exp Ophthalmol 2009;247:707–709.
48. Szigiato AA, Gilani F, Walsh MK, et al. Induction of macular detachment for the treatment of persistent or recurrent idiopathic macular holes. Retina 2016;36:1694–1698.
49. Abbey AM, Van Laere L, Shah AR, Hassan TS. Recurrent macular holes in the era of small-gauge vitrectomy: a review of incidence, risk factors, and outcomes. Retina 2017;37:921–924.
50. Paques M, Massin P, Blain P, et al. Long-term incidence of reopening of macular holes. Ophthalmology 2000;107:760–766.
51. Park DW, Sipperley JO, Sneed SR, et al. Macular hole surgery with internal-limiting membrane peeling and intravitreous air. Ophthalmology 1999;106:1392–1398.
52. Jaycock PD, Bunce C, Xing W, et al. Outcomes of macular hole surgery: implications for surgical management and clinical governance. Eye (Lond) 2004;19:879–884.
53. Tognetto D, Grandin R, Sanguinetti G, et al. Internal limiting membrane removal during macular hole surgery. Ophthalmology 2006;113:1401–1410.
54. Ullrich S, Haritoglou C, Gass C, et al. Macular hole size as a prognostic factor in macular hole surgery. Br J Ophthalmol 2002;86:390–393.
55. Cheng L, Azen SP, El-Bradey MH, et al. Effects of preoperative and postoperative epiretinal membranes on macular hole closure and visual restoration. Ophthalmology 2002;109:1514–1520.
56. Shiode Y, Morizane Y, Matoba R, et al. The role of inverted internal limiting membrane flap in macular hole closure. Invest Ophthalmol Vis Sci 2017;58:4847–4855.
57. Michalewska Z, Michalewski J, Adelman RA, Nawrocki J. Inverted internal limiting membrane flap technique for large macular holes. Ophthalmology 2010;117:2018–2025.
58. Morizane Y, Shiraga F, Kimura S, et al. Autologous transplantation of the internal limiting membrane for refractory macular holes. Am J Ophthalmol 2014;157:861–869.

failed primary procedure; failure to close; failed surgery; idiopathic full-thickness macular hole; internal limiting membrane; ILM flap; ILM peeling; redo surgery; reoperation; reopening; vitreoretinal surgery

© 2020 by Ophthalmic Communications Society, Inc.