Inverted Internal Limiting Membrane Flap vs Complete Internal Limiting Membrane Peeling for Large Macular Holes in Nonposturing Surgery

Abstract

Purpose:

This study compared anatomical and functional outcomes of the inverted internal limiting membrane flap (ILMF) technique with complete ILM peeling (ILMP) in nonposturing surgery for large, full-thickness macular holes (MHs).

Methods:

This was a retrospective, comparative, single-surgeon study. Eyes with idiopathic large full-thickness MHs (minimum diameter > 400 μm) were included. A total of 46 patients including 22 ILMF cases and 24 ILMP cases were analyzed. No positioning instructions were advised postoperatively.

Results:

Primary hole closure was achieved in all patients (100%) in both groups. Mean logMAR visual acuity (VA) improved significantly in both groups compared with the preoperative values (ILMF: 0.60 [SD, 0.26] postoperatively, vs 0.93 [SD, 0.3] preoperatively, P = .032; ILMP: 0.43 [SD, 0.22] postoperatively vs 0.83 [SD, 0.16] preoperatively, P < .01). However, ILMP showed a statistically significant improvement in VA compared with ILMF (P = .02).

Conclusions:

All MHs in both groups closed after surgery (100%). There was a statistically significant improvement of VA in the ILMP group compared with the ILMF group (P = .02). Not posturing after surgery did not compromise surgical success in both groups.

Introduction

Macular hole (MH) represents a vision-threatening event requiring prompt treatment. Since first described by Kelly and Wendel in 1991 as a treatment of full-thickness MHs (FTMHs), pars plana vitrectomy (PPV) has evolved with an increasing success rate. ¹ Improvements in surgical techniques and instrumentation have enhanced the anatomical and functional outcomes of MH surgery.

Currently, PPV combined with internal limiting membrane (ILM) removal and gas tamponade is considered the standard procedure in the treatment of idiopathic MHs. This provides major visual improvements in most cases, with an MH closure rate of approximately 90%. ^2,3 However, in more challenging cases, such as large MHs, chronic MHs, and secondary MHs resulting from ocular trauma or severe myopia, the surgical closure rate is lower. ^{4 -6} MHs with a diameter of 400 µm or greater have unsatisfactory surgical outcomes, even with ILM peeling (ILMP). ⁷

In 2010, Michalewska et al introduced the inverted ILM flap (ILMF) technique, improving both anatomical and visual outcomes in large and complicated MHs. ⁸ This technique is believed to induce proliferation of glial cells on the surface of the retina that then fill the hole and facilitate its closure. A variety of ILMF techniques have been proposed, such as temporal inverted ILM, the cabbage leaf technique, folded inverted flaps, pedicle ILM transposition flap, and the free-flap technique. ^{9 -11} However, there is still controversy about which technique gives the best outcome.

Postoperative prone posturing is believed to intensify the buoyancy force exerted to the posterior pole, which theoretically contributes to the success of MH surgery. However, recent publications indicate that long-lasting posturing is not necessary for MH closure after surgery. ^{12 -14}

The purpose of this study was to compare the anatomical and functional outcomes of the inverted ILMF technique and the complete ILM removal in the treatment of large idiopathic MHs (diameter > 400 μm) in patients with no face-down posturing.

Methods

Patient Inclusion and Exclusion Criteria

This was a comparative cohort study conducted at Calderdale & Huddersfield National Health Service Trust from October 2016 to October 2019.

All patients who underwent PPV with either the ILMF or complete ILMP for idiopathic large FTMHs (minimum diameter ≥ 400 µm) were included. The surgeon used the ILMF technique for all consecutive large holes (≥ 400 µm) and compared them with a cohort of consecutive patients who had ILMP technique performed previously for large holes. Patients were excluded if they had a retinal detachment associated with an MH, traumatic MH, or if they exhibited high myopia (a spherical equivalent ≥ –6.0 D).

Surgical Techniques

All operations were performed by the same surgeon (R.R.). All patients underwent phacovitrectomy unless they were pseudophakic on presentation. After insertion of pars plana ports, phacoemulsification was performed through a 2.2-mm limbal incision. PPV (27 gauge) was performed using a Constellation Vision system (27 Total plus Combined procedure pack; Alcon) and Resight 700 viewing system (Carl Zeiss Meditec). There was complete removal of posterior vitreous, followed by ILMF or ILMP and fluid/air exchange. At this point a 1-piece acrylic intraocular lens (Akreos MICS, MI60, Bausch + Lomb) was implanted into the capsular bag followed by an injection of perfluoroethane (C₂F₆) gas endotamponade. The ILM was stained with membrane blue dual (DORC International).

In the ILMF group, the ILM was grasped with an ILM forceps and peeled off in a circular fashion approximately 2 disc diameters around the MH. During the circumferential peeling, the ILM was not removed completely from the retina but was left attached to the edges of the MH. A rolled segment of the peeled ILM remained hanging in the vitreous cavity. Next, a peripheral piece of the hanging ILM was trimmed with a vitreous cutter, and the central part of the ILM was left in place. The residual flap was not tucked inside the MH. In the ILMP group, ILM was peeled approximately 2 disc diameters around the MH and completely detached from the margins of the hole.

No postoperative positioning instructions were given other than to avoid lying supine. Anatomical MH closure was assessed by spectral-domain optical coherence tomography (OCT; Spectralis, Heidelberg Engineering). Anatomical success was defined as flattening of the MH with resolution of the subretinal cuff of fluid and neurosensory retina completely covering the fovea (flat-closed configuration), as shown on the spectral-domain OCT images.

Ocular Parameters

All patients underwent comprehensive ophthalmologic examinations, including best-corrected visual acuity (BCVA), intraocular pressure measurements, slitlamp examination, indirect ophthalmoscopy, and macular OCT; these examinations were performed at baseline, at 1 month, and at last visit postoperatively, which was usually at 6 months. BCVA was measured using the Snellen chart and converted to the logMAR scale for statistical analysis.

Preoperative OCT scans were analyzed by one researcher (R.R.) to avoid interobserver variation. The OCT examinations were performed using the concurrent 24-line radial scanning protocol (24 B-scans with 7.5° interscan spacing) through the center of the macula. Radial scans were used to measure the minimum linear diameter (MLD) and base diameter (BD). All measurements on OCT scan were performed using the embedded manual caliper function of the OCT machine. A large MH was defined as an MLD of 400 µm or greater. Postoperatively, hole closure and foveal configuration were studied. The foveal configuration was characterized as U shaped or V shaped. The U shape was defined as a contour similar to that of a healthy fovea, whereas the V shape was defined as a steep contour with a thin fovea centralis (Figures 1 and 2). ¹⁵

Figure 1.

U-shaped configuration post macular hole closure in the internal limiting membrane peel group.

Figure 2.

V-shaped configuration with disorganization of retinal layers in the internal limiting membrane flap group.

Statistical Analysis

Statistical analysis was performed using the statistical add-ins in Microsoft Excel 365. The t test for dependent variables was used to compare preoperative and postoperative BCVA in the same group. The t test for independent variables was used to compare different parameters between groups, including BCVA, BD and MLD of the MH, age, sex, and lens status. A P value of less than .05 was considered statistically significant.

Results

A total of 46 eyes were included in this study; 22 eyes composed the ILMF group and 24 eyes composed the ILMP group. The demographic characteristics of the 2 groups are shown in Table 1. There were no statistically significant differences in age (71 vs 70 years, respectively) BCVA (0.93 vs 0.83 logMAR, respectively), and MLD (486 µm vs 484 μm, respectively) and BD of the MHs (962 µm vs 849 μm, respectively).

Table 1.

Baseline Characteristics of the 2 Groups.^a

	ILMF	ILMP	P Value
No. of eyes	22	24
Age, y	71 (±6)	70 (±7)	.72
Sex, male:female	7:15	8:16	.92
Lens status, phakic:pseudophakic	16:6	23:1	.04
BCVA, logMAR	0.93 ± 0.3	0.83 ± 0.16	.18
Minimum linear diameter of MH, μm	486 ± 164	484 ± 90	.96
Base diameter of MH, μm	962 ± 281	849 ± 113	.10

Abbreviations: BCVA, best-corrected visual acuity; ILMF, internal limiting membrane flap; ILMP, internal limiting membrane peel; MH, macular hole.

^aValues are presented as mean ±SD unless otherwise indicated.

Postoperative OCT scans confirmed flat-closed closure in all eyes in both groups. OCT scan at the last follow-up in the ILMF group showed the foveal configuration as V shaped in 19 eyes (86%) and U shaped in 3 eyes (14%). In the ILMP group, 8 (33%) and 16 (67%) eyes demonstrated V-shaped and U-shaped foveal configuration, respectively.

Complete recovery of the external limiting membrane (ELM) was present in 83% (20 of 24) in the ILMP group vs 31% (7 of 22) in the ILMF group. Complete ellipsoid zone (EZ) recovery was shown in 30% (7 of 24) of the ILMP group compared with 14% (3 of 22) of the ILMF group. Cross-sectional OCT images demonstrated a difference in the mean residual EZ defect size between the 2 groups (ILMF, 219 μm; ILMP, 181 μm). However, in the ILMF group, the ELM and EZ were both found to be disorganized in 7 eyes (31%), as shown in Table 2. No eyes represented such changes in the ILMP group.

Table 2.

Summary of Postoperative Macular Microstructural Changes in Both Groups.

	ILMF (n = 22)	ILMP (n = 24)
V shape	19 (86%)	8 (33%)
U shape	3 (13%)	16 (66%)
Complete recovery of ELM	7 (31%)	17 (83%)
Complete recovery of EZ	3 (14%)	7 (30%)
ELM and EZ disorganized	7 (31%)	0
Foveal thickness, μm, mean ± SD	208 ± 58	198 ± 43
Residual EZ defects, µm, mean ± SD	219 ± 125	181 ± 164

Abbreviations: ELM, external limiting membrane; EZ, ellipsoid zone; ILMF, internal limiting membrane flap; ILMP, internal limiting membrane peel.

The mean BCVA was 0.60 logMAR in the ILMF group and 0.43 logMAR in the ILMP group at a mean follow-up of 6 months; BCVA significantly improved in both groups compared with preoperative vision. However, there was a statistically significant improvement of VA in the ILMP compared with the ILMF (mean BCVA: ILMP, 0.43; ILMF, 0.6; P = .02) as shown in Table 3. Furthermore, 46% of the ILMP group achieved 6/12 Snellen vision postoperatively compared with only 18% in the ILMF group (P = .11).

Table 3.

Summary of Anatomical and Visual Outcomes in Both Groups.

	ILMF	ILMP	P Value
No. of eyes	22	24
MH, flat closed	22 (100%)	24 (100%)
Postoperative FUP, mo, mean ± SD	14 ± 7	11 ± 7
BCVA, logMAR, mean ± SD	0.60 ± 0.26	0.43 ± 0.22	.02
6/12 or better vision postoperatively	18%	46%

Abbreviations: BCVA, best-corrected visual acuity; FUP, follow-up; ILMF, internal limiting membrane flap; ILMP, internal limiting membrane peel; MH, macular hole.

Conclusions

Vitrectomy for MH as first described ¹ comprised PPV, removal of adherent cortical vitreous, peeling of the epiretinal membranes, and intraocular tamponade with gas followed by strict face-down posturing. However, its effectiveness may be reduced in the case of large, chronic, or myopic MHs, for which a lower closure rate has been reported. ^{4 -6}

Michalewska et al proposed a novel surgical management, namely the ILMF technique, and demonstrated its efficacy in large idiopathic MHs (diameter > 400 μm) and myopic MHs, improving both anatomical and functional outcomes. ⁸ In the original study, 50 patients randomly assigned to ILMF had a 98% success rate (flat-open MH was considered a success), whereas the standard vitrectomy with ILMP had an 88% success rate. When flat-open FTMHs were excluded, the success rate of standard surgery dropped to 69%, which was significantly lower than our results. However, a major reason for this low success rate could have been that only air tamponade was used in the standard technique, whereas we used C₂F₆, which provides a much longer tamponade.

The exact mechanism of the improved surgical results using the inverted ILMF technique is not precisely understood. In ILMP, it is thought that macrophage-like cells, which are stimulated when the retina is mechanically damaged by ILMP, may infiltrate the retinal tissue and activate the Müller cells, thereby inducing glial cell proliferation that facilitates hole closure. In the ILMF technique, preservation of the ILM, which contains Müller cell fragments around the hole, can help induce gliosis and provide a scaffold for glial cell proliferation on the surface that grows into the hole and retina. This ILMF can thus serve as a barrier, preventing fluid from entering the hole and encouraging photoreceptors to move into the correct position.

Since the initial description, there has been a lot of emphasis on ILMFs for large MHs, with various studies having a success rate between 85% and 100%. ^{16 -19} The largest comparative study on ILMP vs ILMF was published in 2018 by Rizzo et al. ²⁰ The authors compared 300 patients undergoing standard MH surgery (ILMP) with 320 patients undergoing ILMF who had FTMHs of all sizes. The hole closure rate was 78.75% in the standard surgery group, whereas it was 91.93% in the ILMF group. There was no difference between the 2 techniques in the MHs less than 400 μm, whereas for FTMHs larger than 400 μm, the closure rates were 78.6% and 95.6% in the standard ILMP and ILMF groups, respectively, the difference being statistically significant. The authors did not indicate how many patients with FTMHs larger than 400 μm had standard surgery. However, they stated that ILMFs were reserved for large MHs between 2011 and 2013 and that from February 2013, ILMF was also carried out for FTMHs smaller than 400 μm to enhance the closure rate. This statement points to the inherent bias of the surgeons toward flaps and indicates that not many patients with large MHs had standard surgery.

Michalewska et al ⁸ described flap massage over the hole without inserting it inside the hole. Rizzo et al, ²⁰ however, inserted the flap inside the hole. In their study, ²⁰ the ILMF end was inserted into the hole after being trimmed under a balanced salt solution. We found no previous study in which the flap was not manipulated after shortening like described in the present case series. Final results in both groups in Rizzo et al showed comparable VAs postoperatively. Interestingly, no microstructure studies were performed.

Park et al compared insertion vs inverted flap techniques. In their study they showed comparable hole closure results, but the inverted technique was more successful for functional results. ²¹ In our study no attempt was made to insert the ILMF inside the MH.

In the present study, both ILMF and ILMP techniques were successful in causing complete closure of idiopathic large MHs. Furthermore, both groups showed significant postoperative VA improvement compared with preoperative measurements. Nevertheless, the ILMP technique showed greater efficacy in anatomical and functional outcome compared with the ILMF. There was a statistically significant improvement of VA compared with the ILMF (mean BCVA: ILMP, 0.43; ILMF, 0.6; P = .02). Furthermore, 48% of patients in the ILMP group reached 0.3 logMAR or better vision (≥ 6/12 Snellen chart) compared with only 18% in the ILMF group.

On OCT scans of closed MHs, the U-shaped foveal configuration was more frequently observed in ILMP compared with ILMF (66% vs 13%). In contrast, a recent study comparing ILMP vs ILMF techniques for large MHs found a U-shaped closure rate of 50% in the ILMP group compared with 58% in the ILMF group. ²²

Furthermore, as shown in Table 2, complete recovery of the ELM and EZ was more common in the ILMP group than ILMF. Seven eyes (31%) in the ILMF group showed disorganization of the ELM and EZ compared with none in the ILMP group. This abnormal architecture of the external retina in the ILMF is believed to be correlated with worse vision results.

To our knowledge, this is the first study that compared ILMF with ILMP in large MHs without postoperative posturing. One review of the subject had concluded that no posturing may be acceptable for small holes, but posturing does appear to have a role for large holes. ²³ In our series, a good gas fill resulting from combined phacovitrectomy may have adequately isolated the macula despite patients not posturing.

We accept that our study lacks randomization, but nevertheless we have presented well-matched groups in a single-surgeon series to reduce bias. We have used both MLD and BD compared with previous studies that used only 1 parameter (MLD) for preoperative measurements of MHs. Duration of an MH is subjective and unreliable as a prognostic factor because often binocular vision compensates for the central blurring/distortion and patients may only realize any changes at the optician when the good eye is covered. Therefore, we did not include it as the preoperative baseline characteristic of MHs in our study. All the MHs were operated on between 3 and 4 months of presentation. None of the MHs had features of chronicity such as atrophy or pigmentation. However, 50% of patients in the ILMF group had posterior vitreous detachment (stage 4 holes) compared with only 8% in the ILMP group, which may indicate that the MHs were more advanced in the ILMF group.

The power of our study could have been insufficient, considering the number of included patients; however, it was enough to draw a definitive conclusion with statistically significant results. This might suggest the superiority of the ILMP technique in the treatment of large MHs in nonposturing surgery.

All MHs in both groups closed after surgery (100%). There was a statistically significant improvement of vision in both groups (post-VA vs pre-VA, P < .01).There was a statistically significant improvement of VA in the ILMP group compared with the ILMF group (P = .02). Not posturing after surgery with both techniques did not compromise surgical success.

Larger comparative studies need to be performed to conclusively demonstrate any significant benefit of the inverted ILMF technique, especially in MHs with an MLD greater than 650 µm, which previously have been shown to have a lower rate of success with standard surgery. ²⁴