Advancing Macular Hole Surgery

Advancing Macular Hole Surgery


Macular holes were untreatable until Kelly and Wendel1 developed the concept of core vitrectomy followed by fluid-gas exchange using an iso-expansive mixture of air and SF6 or C3F8 gas. Their goal was to reattach the cuff of subretinal fluid around a full-thickness macular hole and thereby eliminate the relative scotoma that surrounds the absolute scotoma. Serendipitously, they soon discovered that the macular hole often closed with remarkable improvement in vision.

Leading surgeons initially discounted this discovery, but fortunately it has become the standard of care. The prevailing thought was that substantial neural tissue was avulsed by a posterior vitreous detachment; an operculum was often seen, seemingly validating this notion. Examination of surgically removed operculums using the electron microscope revealed very little neural tissue; most of the operculum proved to be glial tissue, ex plaining why substantial visual improvement was possible.2,3 Subsequently, OCT demonstrated restoration of near normal or normal foveal anatomy in successfully operated cases. Clearly, “closure” of macular hole is quite different than the term “closure” when used in the context of retinal detachment repair.


Much emphasis is placed on the substantial work of Don Gass,4 which involved clinical observation of the evolution of macular holes, theoretical considerations concerning their pathogenesis, and a classification system. Although all surgeons agree that the posterior vitreous cortex is somehow involved and the elasticity of the ILM plays a role, the pathogenesis of macular holes remains unknown.

Presurgical classification, even using spectral-domain OCT, is incapable of reliably determining if residual vitreous cortex is adherent to the retinal surface, rendering the classification system virtually useless.5 Macular holes are three times as common in females as they are in males, but there is no explanation for this interesting observation.

OCT is essential to determine if the hole is partial thickness or full thickness; not infrequently, clinical examination is inadequate to detect a very small full-thickness hole at the base of a large-diameter partial thickness hole. Occasionally, a very small inner-layer full-thickness hole will overlay a large full-thickness outer-layer hole, producing an appearance suggesting a macular cyst. SD-OCT is essential in the evaluation of macular disease; time-domain OCT is no longer adequate. Mackenzie and colleagues6 have shown that 30% to 50% of partial thickness (lamellar, stage 1) macular holes spontaneously close if observed over the long term. A clinical trial done before the availability of OCT or ILM peeling did not demonstrate the benefit of operating on a partial-thickness hole to prevent progression to a full-thickness hole. Neither ultrasound nor OCT can predict whether a partial-thickness hole will progress to become a full-thickness hole, and the status of the other eye is not helpful because holes are bilateral less than 10% of the time.

Size of the macular hole is the only preoperative factor that has been shown to drive surgical closure rates; duration is not a determining factor if the size is controlled for.7 Although duration has an influence on visual outcomes, assessment of the subhole RPE using spectral-domain OCT and confocal autofluorescence imaging will probably prove to be more effective in predicting visual outcomes in longer-duration holes. Some studies have shown reasonable visual outcomes after operating on macular holes of relatively long duration.8

Macular holes originating from macular cysts due to chronic macular edema typically have relatively poor visual outcomes because of macular ischemia related to underlying diabetic retinopathy or retinal vein occlusion. Similarly, macular holes arising from chronic inflammatory macular edema have a relatively poor visual prognoses.

Traumatic macular holes can spontaneously close in the first four to six weeks, suggesting that a period of observation should precede determination of operability. An afferent pupillary defect (APD) should be considered a contraindication to surgical repair because it indicates associated optic nerve damage. As mentioned above regarding longer-duration macular holes, assessment of the subhole RPE using SD-OCT and confocal autofluorescence will probably prove to be more effective in predicting visual outcomes in cases of traumatic macular holes. Presence of a choroidal rupture in the papillomacular bundle is a relative contraindication to repair of a traumatic macular hole; often, these patients will have an APD.

My management of partial-thickness holes has changed over the years; I began operating on lamellar holes after Arthur Willis introduced the concept of macular hole prevention surgery. I stopped macular hole prevention surgery after the results of the Macular Hole Study and the Charteris paper were published. My current practice is to determine the need for surgery on partial-thickness holes based on symptoms. The goal is to improve visual function — not prevent a full-thickness hole — because progression to a full-thickness hole cannot be predicted. I have determined in recent years that surface tension management using air — or better yet SF6 — combined with ILM peeling is required to restore normal or near-normal foveal anatomy and improve or eliminate symptoms when operating on symptomatic partial-thickness holes.


The conventional explanation of the role of surgical steps in surgical hole closure is incomplete and, to an extent, probably incorrect. It is often stated that the role of PVD creation is to eliminate vitreous traction on the macula, but in fact the posterior vitreous is usually attached to the optic nerve head, nasal and midperipheral retina, but not attached to the macula or retina within the temporal arcades. Careful PVD creation reduces the likelihood of inferior retinal breaks caused by interaction of the bubble with residual vitreous — similar to what may occur with pneumatic retinopexy. The role of core vitrectomy is to enable exchange of vitreous for a large gas bubble and to ensure that a thin layer of vitreous does not prevent contact of the gas bubble with the hole.9,10

Virtually all surgeons agree that ILM peeling improves closure rates, but why this is true has a complex answer. Clearly, ILM peeling ensures removal of tangential traction due to residual vitreous on the retinal surface, which, although rare, can occur even with an apparent PVD as evidenced by a Weiss ring. In addition, ILM peeling guarantees successful removal of epiretinal membranes, which are occasionally present — typically, in eyes with striae and elliptical holes without cuffs. I strongly believe that ILM peeling has an additional and very important role: ILM peeling increases retinal elasticity by over 50%, which enables lateral surface-tension forces to close the hole as soon as the bubble comes into contact with the hole. I believe that it is also likely that ILM peeling initiates mechanical signaling to the astrocytes to heal the hole margins days after it is closed by lateral surface tension.

Most surgeons use the meaningless term “tamponade” when describing the mechanism of action of air, gas and silicone oil bubbles. To some, tamponade means to “press,” but the word tamponade is from the French and means to “seal.” It is more accurate and descriptive to use the term “surface tension management” or “interfacial tension management” to describe the function of the bubble. It is obvious that a bubble eliminates transhole flow; this is the reason it is used in retinal detachment repair. An additional and crucial function of the bubble, as Vincent Reppucci has pointed out,11 is the lateral surface-tension effect. Surface-tension forces act along an interface of two immiscible substances, effectively pulling the surface inward. Surface tension makes a droplet nearly spherical as it falls away from a faucet and makes a soap bubble spherical. Reppucci describes the bubble as bridging the hole, but the key concept is that the bubble pulls the hole margins together as soon as the patient is positioned with bubble in contact with the hole.

Yet another function of the bubble, as I have pointed out, is prevention of transretinal flow (uveal-scleral outflow). Tornambe introduced the hydration hypothesis, making note of the marked edema surrounding the hole as seen on OCT;8 edema-mediated retinal thickening coupled with the elasticity of the ILM causes eversion of the hole margins. Elimination of the edema facilitates restoration of near-normal foveal anatomy, resolution of the edge eversion, and approximation of the hole margins. The bubble dries out the retinal surface, which probably signals the astrocytes to heal the hole days after approximation of hole margins via the lateral surface-tension effect.


Many retinal surgeons use a lateral motion of the extrusion cannula, soft-tip cannula or vitreous cutter to create a PVD. This method creates shear force at the vitreous base, potentially leading to iatrogenic retinal breaks. A better method is to position the vitreous cutter at the nasal, superior, and inferior margins disk margins, with the port oriented away from the center of the disk, and to pull anteriorly (toward the cornea), using the vacuum-only mode (Figure 1). The anterior-pull disk-margin method safely, reliably and quickly produces a PVD using 25-g cutters. It is a huge misconception that higher flow rates help produce PVDs or that 25-g cutters cannot produce a PVD; it is all about technique.

Figure 1. Creation of a PVD using vacuum-only mode, with the port oriented away from the center of the disk.


Unfortunately, many surgeons equate ILM peeling with ICG staining or with alternative dyes or particulate marking with triamcinolone acetonide. I have never used ICG because of my concerns about toxicity, as well as the potential for mixing errors, contamination and increased photo-toxicity risk, and because it is simply unnecessary. Many papers using ICG for ILM peeling report the typical 90% closure rate but unacceptable visual outcomes, suggesting toxicity, and laboratory studies demonstrate neural, glial and RPE cell toxicity. One factor driving dangerous and unnecessary ICG use is using noncontact optical systems, such as the BIOM and EIBOS, which decrease both axial resolution and lateral resolution, making it more difficult to visualize the ILM. There is significant anecdotal evidence for triamcinolone reducing closure rates and particles being trapped within the hole or in the subretinal space.

I use Alcon 25-g DSP ILM end-grasping forceps (Figure 2) but never use pics, MVR blades, or membrane scrapers. These often result in unnecessary retinal surface damage as attempts are made to find or construct an “edge.” Forceps are still required to peel the ILM when pics, scrapers, or MVR blades are used to construct and edge, making this a two-step approach. An additional factor driving ICG use is using inadequate forceps incapable of true end-grasping. Forceps used for ILM peeling should be end-grasping, 25- or 23-g, and not asymmetric or built with a gripping surface extending more that 120 µm along the blades. Reusable forceps rapidly lose the ability to grip at the leading edge of the gripping surface, as the surface is eroded and the blades warped outward by use, cleaning, and sterilization. Reuse of disposable forceps rapidly degrades the ability to grip at the leading edge as well.

Figure 2. Internal limiting membrane peeling with 25-g end-grasping forceps.


I used C3F8 for many years, thinking incorrectly that glial cells pulled the hole together and longer-term surface-tension management would improve outcomes. I converted to SF6 three years ago (Figure 3), after learning of Reppucci's surface-tension concept and observing that holes not closed on the first postop day rarely if ever closed at all. That glial cells do not pull the hole margins together follows from observations that normal or near-normal foveal anat omy without a glial scar is the typical postop OCT appearance.

Figure 3. Surface-tension management using SF6.

Positioning is controversial, in part because of semantics and also from inappropriate pandering to the patients, suggesting that positioning is not necessary with “their” technique. If patients remained supine continuously after surgery, the hole would never close because surface tension from bubble contact is required to pull the hole margins together. It is unknown how long it takes for enough adherence to occur that exposure to fluid would not cause the hole to reopen, but it is probably about three days. Positioning is necessary in phakic eyes to prevent immediate gas cataract.12

Some surgeons advocate silicone oil for patients unable to position after surgery. This is a seriously flawed concept because silicone oil has roughly one-third the surface tension of a gas-fluid (actually gas-retina) interface. A silicone oil bubble to a significant extent conforms to the hole and RPE in the hole's base, reducing the likelihood of closure.13

Vitrectomy produces accelerated nuclear sclerosis progression rates in approximately 90% of cases, although it does not produce cataract in cases with preoperative clear lenses — rarely the case with macular hole patients. As Nancy Holekamp14 has shown, nuclear sclerosis progression is most likely due to a permanent increase in oxygen tension of 7 to 12 mm Hg. Combining cataract surgery with vitrectomy for macular holes, or even removing lenses with minimal or no cataract, is advocated by some surgeons.

I do not recommend phacovit for most macular hole surgery because it produces less precise refractive outcomes and increases the likelihood of posterior synechia and other complications. Refractive outcomes are less precise because axial-length determination in the presence of a macular hole is problematic and because most vitreoretinal surgeons do not perform high-volume cataract surgery or advanced combined refractive technology and techniques. Phaco before vitrectomy can produce slight corneal edema or astigmatism, which can interfere with visualization of the ILM. If phaco is performed after vitrectomy, phaco is more difficult and the sclerotomies must be re-entered for fluid-gas exchange. Gas bubbles increase the likelihood of iris capsule adhesions (synechia) during the postoperative course.


Macular hole surgery using ILM peeling without toxic ICG staining produces hole closure in about 90% of cases. SF6 appears to have an equal success rate to the much longer acting C3F8. I much prefer end-grasping Alcon DSP ILM forceps to pics, MVR blades and membrane scrapers. Using 25-g microincisional vitrectomy with the highest possible cutting rates is preferred to 20-g sutured surgery. PVD creation should be done by use of the vitreous cutter positioned with the port facing outward at the nasal, superior, and inferior margins of the optic disk and pulling anteriorly, not laterally without cutting. RP

Steve Charles, MD, is clinical professor of ophthalmology at the University of Tennessee, College of Medicine, Memphis. Dr. Charles reports significant financial interest in Alcon. He can be reached via e-mail at


  1. Kelly NE, Wendel RT. Vitreous surgery for idiopathic macular holes. Results of a pilot study. Arch Ophthalmol. 1991;109:654-659.
  2. Madreperla SA, McCuen BW 2nd, Hickinbotham D, et al. Clinicopathologic correlation of surgically removed macular hole opercula. Am J Ophthalmol. 1995;120:197-207.
  3. Ezra E, Munro PM, Charteris DG, Aylward WG, Luthert PJ, Gregor ZJ. Macular hole opercula. Ultrastructural features and clinicopathological correlation. Arch Ophthalmol. 1997;115:1381-1387.
  4. Gass JD. Idiopathic senile macular hole. Its early stages and pathogenesis. Arch Ophthalmol. 1988;106:629-639.
  5. Gass JD. Reappraisal of biomicroscopic classification of stages of development of a macular hole. Am J Ophthalmol. 1995;119:752-759.
  6. Mackenzie SE, Gandorfer A, Rohleder M, et al. Ultrastructure and retinal imaging of internal limiting membrane: a clinicopathologic correlation of trypan blue stain in macular hole surgery. Retina. 2009 Dec 4. [Epub ahead of print]
  7. Sjaarda RN, Frank DA, Glaser BM, Thompson JT, Murphy R P. Assessment of vision in idiopathic macular holes with macular microperimetry using the scanning laser ophthalmoscope. Ophthalmology. 1993;100:1513-1518.
  8. Tornambe PE, Poliner LS, Grote K. Macular hole surgery without face down positioning. A pilot study. Retina. 1997;17:179-185.
  9. Glaser BM, Michels RG, Kupperman BD, Sjaarda RN, Pena RA. Transforming growth factor-beta 2 for the treatment of full thickness macular holes. A prospective randomized study. Ophthalmology. 1992;99:1162-1173.
  10. Liggett PE, Skolik DS, Horio B, Saito Y, Alfaro V, Mieler W. Human autologous serum for the treatment of full thickness macular holes: A preliminary study. Ophthalmology. 1995;102:1071-1076.
  11. Chang S, Reppucci V, Zimmerman NJ, Heinemann MH, Coleman DJ. Perfluorocarbon liquids in the management of traumatic retinal detachments. Ophthalmology. 1989;96:785-791; discussion 791-792.
  12. Thompson JT, Smiddy WE, Williams GA, et al. Comparison of recombinant transforming growth factor beta-2 and placebo as an adjuvant agent for macular hole surgery. Ophthalmology. 1998;105:700-706.
  13. Thompson JT,Sjaarda RN,Lansing MB. The results of vitreous surgery for chronic macular hole. Retina. 1997;17:493-501.
  14. Harocopos GJ, Shui YB, McKinnon M, Holekamp NM, Gordon MO, Beebe DC. Importance of vitreous liquefaction in age-related cataract. Invest Ophthalmol Vis Sci. 2004;45:77-85.