Clinical Pearls for Performing ILM Peeling in Vitreoretinal Surgery

SARA J. HAUG, MD, PhD • H. RICHARD MCDONALD, MD

Macular hole surgery was pioneered by Kelly and Wendel, who published the first surgical case series in 1991. The authors presented 52 patients with stage 3 or stage 4 macular holes who underwent pars plana vitrectomy, removal of adherent cortical vitreous, stripping of epiretinal membranes, and gas-fluid exchange.¹ Thirty of these 52 patients, or 58%, were surgical successes, with postoperative macular hole closure.

More recently, peeling of the internal limiting membrane has been suggested as an additional surgical maneuver, improving hole closure rates to 90% or greater.^2-7 This article will discuss the techniques for ILM removal, as well as indications for the procedure.

HISTORICAL REVIEW OF ILM PEELING

The ILM is produced by Müller glial cells during development of the retina, and it represents the structural boundary between the retina and the posterior vitreous.⁸ The membrane has a smooth vitreal surface, with an irregular retinal surface in close apposition with Müller cells.⁹ However, under pathologic conditions, the ILM provides an extracellular matrix for cellular migration and contraction.^8,10

Removal of ILM was not investigated surgically until vitrectomy and ERM removal became routine procedures in the 1980s. It was noted in histological studies of epimacular proliferations that ILM fragments were often identified, which led to the concept of intentional ILM removal.^9,10

Morris and colleagues presented the first case of ILM peeling in 1990 in patients with Terson’s syndrome and sub-ILM macular hemorrhage. The blood and detached ILM were removed without clinically visible surface reproliferation over a 32-month follow-up.^9,11

INDICATIONS FOR ILM PEELING

The rationale for ILM peeling came out of the pathogenesis underlying macular hole formation. Donald Gass, MD, reported that the most plausible explanation for an idiopathic macular hole to develop is a biologic change in the prefoveal vitreous cortex, causing it to shrink, which in turn causes anterior traction detachment of the retina, first in the foveolar and then in the foveal area. Soon thereafter, tangential traction and foveolar hole formation occur.¹²

The tangential traction that is relieved by ILM peeling is thought to be the reason for the increased rate of success of macular hole closure. Now it is accepted that tangential traction plays a role in a number of pathological conditions and the indications for ILM peeling have expanded.

Idiopathic Full-thickness Macular Hole

The classic indication for ILM peeling is for idiopathic full-thickness macular holes (FTMHs). A meta-analysis of the literature, reviewing 31 studies involving 1,654 eyes with FTMHs, compared eyes undergoing vitrectomy surgery with ILM peeling vs vitrectomy surgery without ILM peeling. The authors concluded that ILM removal resulted in statistically significantly better (P<.0001) anatomical and functional outcomes over no ILM peeling.²

A more recent meta-analysis was performed, looking at randomized, controlled trials analyzing surgical success and visual results following vitrectomy with ILM peeling (158 patients) vs vitrectomy with no ILM peeling (149 patients) for stage 2 or greater idiopathic FTMHs. Internal limiting membrane peeling was found to result in significantly higher primary macular hole closure (P<.00001) for all stages of macular holes.⁷

There was no difference in intraoperative or postoperative complications between the groups. Given the decreased need for additional surgery in the ILM peeling group, the procedure was concluded to be more cost-effective. Final visual acuity results between groups were not different.⁷

Myopic FTMH repair also plays a role for ILM peeling. Although the pathogenic mechanism that leads to the FTMH in myopia is not clear, it is presumed that tangential vitreoretinal traction participates in its development. A number of studies have indicated increased success in hole closure with ILM peeling.^13,14

Epiretinal Membrane

Epiretinal membranes arise from a variety of causes, including retinal tears, retinal vascular occlusions, and trauma. Idiopathic ERMs are often associated with a posterior vitreous detachment. Surgical specimens of removed ERMs often have ILM fragments removed as well, indicating that the ILM allows a scaffold on which ERMs proliferate and contract.^15,16

In a pilot study looking at intentional ILM removal in addition to ERM peeling, Park et al found that none of the eyes that underwent ILM peeling had recurrence of the ERM, whereas 21% of eyes that did not receive intentional ILM removal had recurrence.¹⁷

Follow-up studies have shown similar results,^18,19 although no randomized, prospective studies have yet been published. Final VA has not been found to be different between patients undergoing ILM peeling and those not undergoing peeling.^19,20

Diabetic Macular Edema

The pathogenesis of diffuse diabetic macular edema likely involves multiple mechanisms, including leakage from dilated capillary beds, deep leakage from the retinal pigment epithelium, and tractional forces at the vitreoretinal interface.^9,21

Previous studies have noted improvement in diffuse refractory macular edema following vitrectomy surgery with posterior hyaloid removal, indicating a significant contribution from vitreoretinal traction in the development of DME.^21-23

Histopathological analysis of the ILM in patients with diabetes found that the mean thickness was significantly greater than that in nondiabetic patients (4.8 µm vs 1.8 µm, respectively).

In addition, extracellular components, such as heparan sulfate, were found in association with the ILM in increased amounts in diabetic patients, leading the authors to hypothesize that the thickening of the ILM could contribute to a disturbance in water movement between the retina and the vitreous, resulting in fluid accumulation within the retina.²⁴ Therefore, the additional removal of the ILM at the time of PPV surgery may allow for more complete removal of all tractional forces and inhibit the reproliferation of fibrous astrocytes.^9,21

Gandorfer and colleagues published in 2000 the first series of vitrectomy surgery plus ILM removal and found that the procedure expedited resolution of diffuse DME and improved VA without subsequent ERM formation.²¹ However, this study did not have a control group undergoing vitrectomy alone without ILM peeling.

Another study involved a randomized, controlled trial and found that vitrectomy with ILM peeling was significantly superior to observation, but again, it did not compare the results to vitrectomy with no ILM peeling.²⁵

Case-control and randomized studies comparing the results of vitrectomy with induction of a PVD to vitrectomy with ILM peeling have found no significant differences in outcomes of VA or improvement of macular edema.^9,26-28

Myopic Foveoschisis

Myopic foveoschisis is a common complication in pathologic myopia. Evidence suggests that vitrectomy with or without gas tamponade is an effective treatment for myopic foveoschisis, but the role of peeling the ILM is not established.

A number of studies have indicated that peeling the ILM results in better visual and anatomic results.^29-31 A study by Sayanagi in 2006 also indicated that the rate of reoperation for persistent myopic foveoschisis was much higher in eyes that did not undergo ILM peeling, compared to eyes that underwent both vitrectomy and ILM peeling.³² However, the rate of macular hole formation after vitrectomy with ILM peeling to repair myopic foveoschisis was found to be high (19%), compared to the rate following vitrectomy performed for rhegmatogenous retinal detachments (0.5–1.1%). The authors speculated that this difference might have been due to potential trauma to underlying Müller cells due to the ILM peel, resulting in weakening of the macular glial structure.³³

Other indications for ILM peeling include edema associated with branch retinal vein occlusion or central retinal vein occlusion. It has also been applied in cases of Terson’s syndrome and optic disc pit maculopathy.⁹ The efficacy of ILM peeling in these conditions remains to be established.

SURGICAL TECHNIQUE

Over the last couple of decades, a variety of surgical maneuvers have been introduced that have improved the closure rate and reduced the recurrence rate of macular holes following surgical repair, as discussed above.

It was debated initially whether manipulation of the retinal surface had any beneficial effects at all. It came to be accepted that removal of ERM and/or ILM was a key component of successful vitrectomy for macular hole. So how best can the surgeon accomplish the maneuver?

Stains and Marking of the ILM

Indocyanine green is an effective dye for staining the ILM. Once a portion of ILM has been engaged and removed, it is clear where ILM remnants have been left. But evidence has accumulated that there are toxicity issues with the stain.^34-38

We use it infrequently in our practice, primarily for removing ILM in some eyes with myopic macular schisis. We feel there is a value in removing the ILM in such eyes, and we have seen recurrent schisis in some eyes in which it has been left. Brilliant Blue G is not approved for use in the United States, but it may eventually become a mainstay of ILM peeling technique.

We have used intravitreal triamcinolone acetonide (Kenalog, Bristol-Myers Squibb, New York, NY) for a number of years, in association with 25-gauge vitrectomy. The triamcinolone crystals are placed to mark the surface of the ILM, much as the surface of water is lightly sprayed during high dive competitions.

The amount of triamcinolone layered on the macular surface is important. Too much clumped in one area obscures the retinal surface. Not enough means that you are essentially performing your dissection without any staining or marking of the retinal surface.

We have found that diluting the triamcinolone 50:50 with balanced salt solution renders a more homogeneous dispersal pattern of crystals (Figure 1). Triamcinolone is placed at the beginning of each vitrectomy to assist in the disinsertion of the vitreous from the optic nerve head and to ensure that there is no residual vitreous cortex adherent to the macula. Once the vitrectomy is complete, a small amount of 50:50 triamcinolone/BSS is again injected over the macula.

Figure 1. Following vitrectomy, a 50-50 mixture of triamcinolone and basic salt solution (BSS) is distributed over the macula.

An important consideration in covering the macula with an ideal pattern of triamcinolone crystals is the use of valved cannula systems. The triamcinolone will serve to mark the ILM, but as the dissection begins, and there is ILM splitting, the margins of peeled areas can only be seen if the triamcinolone remains present over the unpeeled region.

In the days before valved cannula systems, untimely removal of an instrument, especially if there was momentary high infusion pressure in use, would result in the triamcinolone crystals being blown away from the macular surface. It was occasionally difficult to identify the peeled area from the unpeeled area, resulting in unnecessary retinal trauma as attempts were made to restart the dissection.

Of course, many surgeons use no stains or marking agents at all. But marking makes the process easier. Removing the ILM from a highly myopic, pale fundus can be an additional challenge, and we have found marking with triamcinolone particularly helpful in these eyes.

Our Technique

The technique for removing ERMs and ILM has changed through the years. Although Steve Charles, MD, discussed the concept of pinching and elevating the ERM with forceps in the early 1980s,³⁹ most surgeons used a barbed MVR blade to engage and initiate the ERM dissection.

We have transitioned, with the introduction of small-gauge surgery, to using the “pinch” technique for both ERM and ILM peeling. We peel the ILM in all cases of macular hole surgery, and we attempt to remove the ILM along with, or after, ERM removal.

The technique we employ is to pinch the inner surface of the retina with a 25-gauge ILM forceps (Figure 2, top). A preferred area for us is either superior or inferior to the fovea, inside the arcades, at a distance of approximately one disc diameter. Once the inner retina is grasped, a slow, small, back-and-forth rocking movement is used with subtle elevations of the ILM. The back-and-forth movement gradually increases in distance until there is a splitting of the ILM at this focal point (Figure 2, inset).

Figure 2. The partially opened jaws of the ILM forceps are placed on the surface of the retina pushing enough to indent the surface. (inset) The forceps are closed to pinch down on the innermost layer of the tissue and gently elevated slightly above the surface of the retina. A back and forth motion is then employed with gradual elongation of the side-to-side movements and increased elevation. This technique results in a break in the ILM. Once the ILM is disrupted by this maneuver, the forceps’ hold on the ILM is released and the incised ILM is engaged on the edge of the incised tissue.

During this flap initiation procedure, the deeper structures of the retina are viewed to see how much movement is occurring. Too much retinal vascular movement is a warning signal to release and regrasp. Of course, hemorrhage also signals that the tissue must be released.

Once there is an initial split in the ILM, we release and grasp next to the start point, sequentially regrasping and elevating ILM to each side of the initial ILM split. We continue to “go wide” until a broad front of elevated ILM has been fashioned (Figure 3).

Figure 3. This maneuver is repeated, creating an elongated front of incised ILM. By “going wide” at either end of the ILM incision, a broad flap of tissue is created.

Only after this leading edge of ILM has been made is there any attempt to peel the ILM as a sheet away from the retinal surface (Figure 4). We prefer to elevate the broad sheet of ILM before peeling over the fovea. This allows for good visualization of the ILM coming off the fovea and reduces the risk of having the ILM split at the fovea, necessitating reinitiation of the ILM dissection at the foveal edge.

Figure 4. One end of this linear front of incised ILM is grasped and a maculorrhexis is created by pulling the ILM in a circumferential manner. The area where the ILM has been removed is devoid of triamcinolone crystals.

Obviously, this cannot always be controlled and ILM splitting can occur at the fovea. In these cases, we prefer, if possible, to initiate ILM dissection away from the fovea. If the original dissection started inferiorly, then we would attempt to elevate the superior ILM and bring it down to the fovea. However, many times the split angles are such that the ILM must be regrasped at the fovea.

A helpful maneuver during the “maculorrhexis” involves splitting the ILM in a circular manner from the rest of the macula (Figure 4). This is easily accomplished by grasping the sheet of ILM already elevated at its widest margin and pulling in a circumferential motion.

Once the ILM has been removed in this manner, there is always the question of how much more should be peeled. We peel out to roughly one disc diameter from the vascular arcades, although it is unknown what is necessary and sufficient to achieve macular hole closure.

Once the maculorrhexis is complete, the area where the ILM has been removed will have a whitish sheen and will be free of overlying triamcinolone. If there are steroid crystals in the macular hole, even quite a few, there is no need to attempt evacuation with aspiration or forceps removal. Neither the closure rate nor subsequent visual result appears to be affected.

A VIEW OF THE FUTURE

An intriguing concept currently being studied is the possible role of intraoperative optical coherence tomography in ILM dissection.⁴⁰ It could be a great help to know with certainty whether there are areas of ILM still adherent to the retina, particularly in those eyes with ERMs in which the surgeon is attempting to remove the ERM and ILM at the same time. This technique has been called the “tablecloth and placemat maneuver.” It would be of benefit to know whether the tablecloth (the ILM) really came away with the placemat (the ERM). Intraoperative OCT would be valuable in such cases.

We expect that better stains will be available in the future. Stains serve primarily to let us know “where we’ve been” during ILM dissection. Brilliant Blue G is available as a compounded agent, and it is currently being evaluated by the FDA. While it is known to be safe in the eye in animal studies and is approved for use in Europe, it has not been approved in the United States for human use in any form (unlike indocyanine green). This is the reason hospital pharmacies are not keen to allow it in hospital operating rooms. RP

REFERENCES