Internal Limiting Membrane Peeling in Idiopathic Epiretinal Membrane

Abstract

The primary management for epiretinal membrane (ERM) is membrane peel after pars plana vitrectomy. However, the rates of postoperative recurrence of epiretinal membrane reported range from 10 to 21%. Internal limiting membrane (ILM) peeling combined with ERM removal has been introduced in an attempt to diminish this recurrence. Some studies showed that this method largely prevented the recurrence compared with those without ILM peeling. Conversely, other studies demonstrated that combined ERM and ILM peeling did not provide a lower recurrence rate. Since the ILM is formed by the basal lamina of Muller cells, removal of this structure must be pondered due to possible mechanical and functional damage to those important cells. In this chapter, current data on this topic are covered.

1. Introduction

Epiretinal membrane is a prevalent disease [1, 2]. Studies that incorporated ocular coherence tomography (OCT) for detection found a higher prevalence of this pathology, ranging from 3.4 [3] to 34.1% [4].

Most epiretinal membrane (ERM) is idiopathic and increasing age is the most important risk factor, with most patients presenting over 50 years and a peak prevalence in the 7th decade [5, 6]. There is great variability in the reported prevalence of ERM among different racial groups, although studies using similar methodologies reported a higher prevalence in Asians [7, 8].

Vitrectomy with membrane peeling remains the mainstay of treatment for symptomatic ERMs. First, a three ports pars plana vitrectomy is performed and then the ERM is peeled. Dyes are often used to better visualize the ERM and the internal limiting membrane (ILM).

2. Epiretinal membrane

Idiopathic ERMs occur when there are no associated ocular pathologies. Secondary ERMs are considered those associated with an ocular pathology and account for about 30% of ERMs [8].

Regardless of its etiology, an epiretinal membrane is formed by an innermost single or multilayer of cells and an outermost noncellular layer, which is in contact with the ILM. The cellular layer constituents include retinal glial cells, hyalocytes, retinal pigment epithelial cells, and fibroblasts, and these cells originate myofibroblasts through transdifferentiation [9, 10, 11]. The main component of the outermost layer is different types of extracellular collagen divided into native vitreous collagen, reminiscent on the retinal surface after posterior vitreous detachment (PVD), and newly formed collagen, synthesized and secreted by the cellular layer [12, 13].

Since a PVD is present in the majority of cases [14, 15] it has been suggested its participation during idiopathic ERM formation. After PVD, it is theorized that reminiscent of hyalocytes on the retinal surface starts a process of metaplasia and ends up forming the ERM [16].

Another etiology proposed, although less accepted, is the migration of retinal glial cells to the retinal surface through defects on ILM after a PVD [17].

Classification based on clinical findings proposed by Gass [18] is still widely used. Along with developments in OCT (optical coherence tomography), several classifications based on this technology have been proposed based on the identification of associated retinal anatomic changes [19, 20, 21, 22, 23].

Based on new OCT findings, including ectopic inner foveal layers, a recent grading scheme had been proposed [24]. Stage 1 was defined as the presence of an ERM, seen as a hyperreflective line above the inner retina, with negligible retinal anatomic changes. Stage 2 was defined as the presence of ERM causing loss of foveal depression and stretching of the outer nuclear layer. Stage 3 was defined as the presence of an ERM with continuous ectopic inner foveal layers added to stage 2 findings. Stage 4 was defined as the presence of stage 3 findings added to the disorganization of retinal layers.

3. Internal limiting membrane

The ILM is composed of an innermost structure formed by a meshwork of collagen fibers, glycosaminoglycans, laminin, and fibronectin called cuticular layer, and an outmost structure, facing the retina surface, formed by the footplates of Müller cells [25, 26]. The internal limiting membrane has a smooth vitreal side and an irregular retinal side where folds are in apposition to Muller cells footplates [27].

Muller cells are the main glial cell of the retina. They give structural stability to the fovea and have an important role in metabolic functions, such as regulating the balance of relevant ions, removing metabolic waste, and providing trophic substances to neurons [28]. The inner processes of Muller cells, formed by its footplates (Figure 1), participate in the outmost structure of ILM [25]. The outer processes of Muller cells surround the somata of photoreceptor cells and together constitute the external limiting membrane (ELM) (Figure 1).

Figure 1.

Muller cells of foveal walls, Muller cells cone, external limiting membrane (ELM), and internal limiting membrane (ILM). Schematic drawing by Luciana S.Q. Makarczyk.

Muller cells are differentiated into two groups according to their fovea location: Müller cell cone and outer processes of the Müller cells of the foveal walls [29, 30]. The Muller cell cone acts as a plug binding together the receptor cells in the foveola, increasing resistance against mechanical stress, since the radiating nerve fibers would be highly susceptible to disruption in this region [31]. It has been suggested that the Muller cells of the foveal wall give stability to the outer layers of the fovea and parafovea [32, 33].

Muller cells of the foveal walls display a z-shaped pattern as a result of the centrifugal displacement of the inner retina and the centripetal displacement of the outer retina and photoreceptors. Muller cells’ vertical processes run from the ILM to the inner nuclear layer (INL), then diagonally within Henle’s fiber layer (HFL) and, once more, vertically to the outer limiting membrane (OLM). The morphology of this type of Müller cell may absorb mechanical tension according to a physical model study [34]. This model considered those cells as the main determinants of force transmission and suggested their importance in the structural stability of the parafovea by increasing retinal compliance to mechanical stress.

The thickness of the ILM within 400 μm from the foveal center ranges from 0.050 to 0.2 μm. From 400 to 600 μm from the center of the fovea, its thickness varies from 0.08 to 1.0 μm. At a distance of 600 to 900 μm from the foveal center, ILM showed a thickness from 1.4 to 2.4 μm. Between 900 and 1050 μm from the foveal center, ILM reaches 4.0 μm of thickness. At the disc, its thickness ranges from 0.07 to 0.1 μm and it is thicker in the posterior pole than the equator [27, 35].

Analyses of ILM stiffness matched thickness findings according to geographical distribution, meaning that higher stiffness values were found in the posterior pole compared with the mid-peripheral quadrants [27].

Biomechanical analyses of the ILM showed that it provides a protective function to the retina. Removal of the ILM showed to reduce the mean strength of the central retina by 53.6% [36].

Several diseases affecting the vitreomacular interface could be attributed to foveal mechanical instability: vitreous adhesion, a higher thickness and stiffness of the ILM at the macular area added to a known absence of cellular connections between the cells of the Müller cell cone and the outer processes of the Müller cells of the foveal walls [30, 37].

3.1 Ectopic inner foveal layers

Ectopic inner foveal layers have been defined by Govetto et al. [37, 38] as displacement of retina architecture with gliosis and proliferation of Muller cells, characterized on OCT as a continuous, homogenous, hypo or hyper-reflective band, extending from the inner nuclear layer and inner plexiform layer across the foveal region (Figure 2). This ectopic layer does not possess contractile characteristics.

Figure 2.

Epiretinal membrane causing tractional lamellar hole. a. Before surgery. Ectopic inner foveal layer (asterisks). Stretched Muller cell cone processes (arrow). b. Two months after ERM and ILM peeling and absorption of C3F8 gas - Luciana S.Q. Makarczyk, MD, PhD.

4. Epiretinal membrane and internal limiting membrane peeling

The aim of surgery is to remove tractions caused by ERM and ectopic inner foveal layers. Three-port pars plana vitrectomy followed by an ERM peeling, using 23-, 25-, or 26-gauge instruments is the standard of care. The use of smaller gauges allows for a faster surgery recovery [39].

Since the separation of the ILM has been shown to be compatible with a good visual function [40, 41, 42], intentional removal of the ILM has been performed in order to increase macula elasticity, which could be beneficial for treating several macular disorders, such as macular holes, chronic diabetic macular edema, vitreomacular traction, and myopic foveschisis [28, 29]. Removal of the macular ILM has been demonstrated to greatly improve the anatomical success rate of the surgical treatment of macular holes [43, 44].

The ILM is suspected to provide a scaffold for cellular proliferation and its active peeling has become a common practice during ERM removal. Simultaneous ILM peel is a frequent occurrence during ERM surgery when there is a complete ERM macula adhesion on OCT [45].

Postoperative recurrence of epiretinal membrane range from 10 to 21%. Specimens of recurrent ERM evaluated by electron microscopic demonstrate mainly myofibroblasts, but also fibrocytes, retinal pigment cells, fragments of ILM, and new collagen [45, 46]. Also, some studies had demonstrated remnants of ERM or cell remnants on ILM vitreous surface specimens [47, 48, 49].

Thus, ILM peeling combined with ERM removal has been performed in an attempt to remove cells from the retina surface that would serve as potential sites for ERM reproliferation [50].

Chromovitrectomy is frequently used as an adjuvant to facilitate ILM identification. It involves the use of dyes intended to improve ERM and ILM visibility. Trypan blue (TB), brilliant blue G (BBG), and indocyanine green (ICG) are the most frequently used dyes.

Trypan blue stains mainly the ERM while BBG stains mainly the ILM. ICG is an excellent dye for ILM stain, although it has significant evidence of retina and optic nerve toxicity [51, 52].

Despite its toxicity evidence, ICG is still used since it has been described to facilitate ILM peeling by adding an ILM stiffening effect through a collagen IV cross-linking [53] and also having a higher ILM staining property than BBG [54].

A thicker and stiffer ILM facilitates surgical removal and according to previously described, a region within a foveal distance of roughly 1000 μm would be ideal for starting the peeling. Also, in this region, a better stain during chromovitrectomy is expected [27].

The use of BBG has been proven helpful in better visualizing ILM and decreasing ERM fragments, which may be the source of recurrent ERMs [55].

5. Updated data on maintaining ILM versus ILM peeling in ERM treatment

Even more than 10 years after additional ILM peeling has been introduced, there is still a discussion of the best surgical approach for ERM treatment in order to prevent its recurrence.

Some studies showed that ILM peeling prevented ERM recurrence [48, 56, 57]. According to those studies, groups that had ILM peeling had no recurrence of ERM, while in groups where double peeling was not performed, the recurrence reached 17,6%.

On contrary, some other studies showed that patients who had ERM peeling without removing the ILM had an even better anatomical and visual outcome [58, 59, 60].

Ahn et al. [60] compared the OCT postoperative status and visual acuity after ERM removal with or without additional ILM peeling. They found that patients who had epiretinal membrane surgery with additional ILM peeling showed worse visual acuity and most severe cone outer segment tips and inner and outer segment junction line defects. Those photoreceptors defects were gradually restored at 12 months postoperatively. A possible reason for this finding is that by removing the ILM, the footplates of Muller cells are also being removed, leading to mechanical and functional damage to those important cells.

There is a supposition that ILM peeling spares the ganglion cells and is limited to the Müller cell footplates. Some studies have suggested that injury to Müller cells was due to direct and indirect traumas, caused by surgical instruments and traction during membrane removal, respectively [61, 62, 63].

Deltour et al. [64], in a retrospective study, demonstrated that active peeling induces more numerous and deeper microscotomas than spontaneous peeling. Later, a prospective, randomized, controlled, single-blind, multicentered trial with two parallel arms investigated differences between active ILM peeling and passive ILM peeling, and showed similar findings [65].

A recent meta-analysis of randomized and controlled trials showed no statistically significant difference in final visual acuity and recurrence rate of ERM when ILM peeling was performed. As for the central macular thickness (CMT), it was thicker within 6 months postoperatively in those patients who had additional ILM peeling. A possible reason for a thicker CMT in a combined ERM and ILM peeling is that it could induce retinal edema. Also, this meta-analysis data showed that ILM peeling seemed to result in more complications, including intraoperative retinal breaks and nonarteritic anterior ischemic optic neuropathy [66].

Distinct spectral domain optical coherence tomography (SD-OCT) findings on the inner retina have been associated with ILM peeling. Swelling of the arcuate fiber layer (SANFL) is an earlier pathology that is characterized on B scan OCT as hyperreflective images and on infrared as hyperfluorescent arcuate areas. SANFL is followed by dissociated optic nerve fiber layer defect (DONFL) [67]. DONFL (Figure 3) presents on SD-OCT as irregular depressions in the nerve fiber layer due to thinning of the ganglion cell layer [67, 68] and more visible with shorter wavelength illumination [69, 70]. It is thought to be due to trauma to Muller cells after the ILM is peeled, disconnecting their attached foot plates [62, 70, 71], and exposing the retinal nerve fiber layers. Trauma to Muller footplates would possibly result in changes in the bundling of the nerve fiber layer and also a volume reduction in the ganglion cell layer where the Muller cell bodies are located [72].

Figure 3.

DONFL after double peeling for stage 4 ERM. a. Enface OCT showing numerous dark areas corresponding to retinal cavitations. b. B scan OCT showing increased foveal thickness due to hyporeflective spaces—images shared by Michelle Gantois, MD.

It also had been reported that (DONFL) would also cause a reduction in central retinal sensitivity and paracentral scotomas. Terasaki et al. also reported a delayed recovery of the b-wave amplitude of the focal macula electroretinogram after ILM peeling during macular hole surgery [73].

It is still unclear the cause of DONFL development, but it has been also associated with surgical trauma during ILM peeling [74].

6. Conclusion

The surgical procedures for removing ERM are well-established and safe. However, there is still not enough data to affirm if an additional ILM peeling is always an important surgical step or would be better to be performed only in selected ERM cases.

There are proven morphological and functional retina modifications with an additional ILM peeling, although is not known whether those changes could cause progressive retinal damage in the medium or long term.

In my own experience, recurrence rates of ERM are not common, even when ILM is not removed. And since ILM peeling can lead to complications, in my point of view, it is not essential, unless when associated with macular holes. In addition, ILM removal could also be needed when epiretinal proliferation is present under the ERM. Evaluation of ILM status after removal of ERM is also important to decide if it is worth to removed.

Intraoperative optical coherence tomography is an emerging technology that has been evolving and improving in reproducibility and motion artifacts. Soon it will help guide decision-making in the surgical approach for treating ERM, including better visualization of debris left on the retina surface that needs to be removed, as well as when ILM should not be left in place.