Diseases affecting primarily the vitreous and vitreoretinal interface are associated with a variety of macular lesions causing loss of central vision. These lesions may be detected ophthalmoscopically and biomicroscopically and should be differentiated from the other causes of macular dysfunction. In no other area of macular disease is the use of a fundus contact lens more important to detect and define the anatomic changes. Recently introduced techniques that improve our ability in this regard are optical coherence tomography (OCT) and kinetic ultrasonography.
Anatomic Considerations
The vitreous is a semisolid gel containing a hyaluronic acid network interspersed in a framework of randomly spaced collagen fibrils. The framework is most apparent histologically in the region of the pars plana, where it is strongly anchored to the ciliary epithelium in an area referred to as the vitreous base. Posterior to the pars plana the concentration of collagen and hyaluronic acid is greatest in the ill-defined outer part of the vitreous gel, referred to as the vitreous cortex, that lies along the inner retinal surface. The collagen fibrils, which are condensed to form an outer layer of the vitreous cortex, are adherent to the internal limiting membrane (ILM: basement membrane, or basal lamina of the Müller cells) of the retina ( Figures 7.01 and 7.02 ). The basal lamina thickness increases from the vitreous base posteriorly, to where it reaches maximal thickness at the crest of the foveal clivus. From there it rapidly becomes thinner, reaching a thickness of 200 Å or less in the foveal center. At the margin of the optic disc, the basal lamina abruptly thins to approximately 450 Å, where it covers the disc surface. Here the basal lamina is associated with multiple gaps associated with glial epipapillary membranes that are probably of developmental origin. Attachment plaques or hemidesmosomes are evident electron microscopically along the vitreoretinal junction in the peripheral and equatorial zones but are absent posteriorly except in the foveal area. These findings indicate greater adherence of the basal lamina to the Müller cells in these zones with attachment plaques but do not necessarily reflect greater adherence of the vitreous to the basal lamina in these areas. There is other evidence, however, to indicate greater adherence of cortical vitreous to the basal lamina in these areas, including the central macular region. Progressive liquefaction of the posterior vitreous occurs with aging, giving rise to a large optically empty cavity of liquified vitreous in the premacular area, referred to as the premacular bursa, or prefoveolar pocket ( Figure 7.03 ). The thin layer of the posterior cortical vitreous gel lying on the inner surface of the macula is not visible biomicroscopically, and the anterior interface of the bursa may be visible and misinterpreted biomicroscopically as the posterior hyaloid of the separated vitreous.
The degree of vitreoretinal adherence varies with age as well as location in the eye. OCT illustrates the change in the contour of the vitreous attachment to the posterior pole with age. Children and young adults show no separation of the posterior hyaloid from the retinal surface ( Figure 7.04D ). From the fifth decade onwards the posterior hyaloid shows a gentle curve away from the retina, still being attached to the fovea and the optic disc ( Figure 7.04E , arrows). There are no visible effects of traction on any structure at this stage. As the posterior hyaloid tries to separate from the retina, various configurations occur in different eyes that include incomplete separation with residual vitreofoveal or vitreopapillary traction, and anomalous separation with vitreoschisis resulting in epiretinal membrane and full-thickness or lamellar macular hole ( Figure 7.05A–E, G, I, and J ). A normal vitreous separation shows the posterior hyaloid membrane separated from the retina with a normal foveal contour ( Figure 7.04F ). Generally the adherence decreases with age. The attachment of the vitreous to the retina is greatest at those sites where the ILM of the retina is the thinnest ( Figure 7.01 ). These sites include the vitreous base, the major retinal vessels, the optic nerve head, the 1500-μm-diameter rim surrounding the fovea, and the 500-μm-diameter foveola. The latter two sites of attachment are probably important in the development of idiopathic age-related macular hole. Forces generated by movement of the vitreous and the premacular bursa as the eye moves may also play a role in the pathogenesis of posterior vitreous detachment (PVD), epiretinal membranes, and macular holes ( Figure 7.03 ).
Kishi and coworkers found anatomic evidence that the prefoveolar vitreous cortex (PVC) may be focally condensed and tightly adherent to the inner surface of the foveolar retina. They examined 59 eyes with spontaneous PVD with scanning electron microscopy. In 44% of the eyes they found three patterns of vitreous remnants on the surface of the foveolar area. The most common pattern, type 1, found in one-half of these eyes, was a 500-μm-diameter disc of condensed cortical vitreous adherent to the foveolar retina ( Figure 7.06A ). In 30% of cases (type 2) a 500-μm-diameter ring of remnants was found adherent to the margin of the foveolar retina ( Figure 7.06B and C ). In some eyes the authors also noted a 1500-μm-diameter ring of vitreous remnants at the foveal margin ( Figure 7.06A ). Twenty percent of the eyes (type 3) showed a pseudocyst formation consisting of a focal 200–300-μm-diameter disc of contracted vitreous cortex bridging the foveolar area. These findings suggest that the structures of the PVC and the vitreoretinal interface in the foveolar area are probably different from that elsewhere in the macular area.
The cells that are part of the normal vitreous are widely scattered within the vitreous cortex along the surface of the retina and ciliary body. Their concentration is maximal within the vitreous base and near the posterior pole. These cells, termed “hyalocytes,” show phagocytic properties, have a high metabolic activity, and may be responsible for both the formation and the maintenance of several vitreous components. They are probably mesenchymal cells with macrophage-like properties. When properly stimulated, they are capable of cell migration, proliferation, collagen formation, and membrane contraction. The membrane contraction (collagen) may be mediated via transforming growth factor (TGF)-β 2 since using anti-TGF-β 2 neutralizing antibodies experimentally can block the contraction. This capability of fibrous metaplasia may supplement the process of collapse, condensation, and shrinkage of the normal collagen framework in the production of pathologic vitreous membranes. Much of the posterior vitreous becomes liquefied by the seventh decade (synchysis senilis). This process of syneresis may be accompanied by spontaneous separation of the vitreous cortex from the retina, a process referred to as PVD. Following vitreous separation, there is condensation and realignment of the collagen molecules comprising the outer surface of the cortical vitreous to form a distinct membrane, the so-called posterior hyaloid membrane, which may be visible biomicroscopically and histologically ( Figure 7.02 ). PVD is present in over 25% of persons by the seventh decade and approximately 65% by the eighth decade. It is more common in women. PVD most frequently begins in the macular region following spontaneous dehiscence of the posterior hyaloid near the center of the macula. It may, however, begin more peripherally. In most patients separation of the posterior hyaloid face from the retina occurs rapidly and smoothly and may or may not be accompanied by symptoms of photopsia and floaters. Slit-lamp examination reveals anterior displacement of the posterior hyaloid membrane. A gray-white ring of vitreous condensation (Weiss ring) that marks the site of previous vitreous attachment to the margins of the optic disc is usually visible and is the single most important biomicroscopic sign of posterior vitreous separation from the optic disc and macular area ( Figure 7.04 ). In cases where the posterior hyaloid face tears near the site of attachment to the crest of the fovea, a similar condensation ring may lie in front of the macula ( Figure 7.06 ). These condensation rings are often distorted and twisted. PVD usually occurs without producing any visible alterations in the retina. As the vitreous separates, traction on the inner surface of the optic disc, along the major vascular arcades, or near the vitreous base may occasionally produce a focal intraretinal, preretinal, or diffuse vitreous hemorrhage ( Figure 7.07A and B ).
PVD is the primary cause of peripheral retinal tears and rhegmatogenous retinal detachment. Pathologic alterations in the vitreous gel unrelated to aging may be responsible for vitreous shrinkage and premature PVD. Patients with high myopia are more likely to develop PVD early. Anomalous PVD may result in either vitreoschisis leading to macular hole, macular pucker or diabetic traction detachment (see next section) or partial (incomplete) PVD with residual traction in the periphery causing peripheral retinal tears, in the macula causing vitreomacular traction, or the optic disc resulting in vitreopapillary traction.
Vitreoschisis
Posterior vitreous cortex is composed of lamellae running tangential to the ILM. These lamellae are the site of potential cleavage when the vitreous detaches. Anomalous PVD results in vitreoschisis whereby a split occurs within the posterior hyaloid leaving a membrane adherent to, or in close proximity to, the retinal ILM. Contraction of this membrane may result in anteroposterior or tangential traction at various points of adherence to the inner retinal surface ( Figure 7.08A–C ). The pathogenesis of macular hole and macular pucker may be explained at least in part via this phenomenon. Hyalocytes are located approximately 50 μm from the retinal surface within the cortical vitreous. It has been postulated that the split occurs at different levels in patients with macular hole and pucker. A split posterior to the hyalocytes is likely to play a role in macular hole formation where a thin acellular membrane is left adherent to the ILM. Taut contraction of this membrane over the fovea likely pops up the foveal depression which may cause a dehiscence in the continuity of the foveal tissue, initiating the process of a macular hole. A split in the cortical vitreous anterior to the hyalocytes leaves behind a thicker cellular membrane on the retinal surface. Hyalocytes stimulate migration of monocytes from the circulation and glial cells from the retina. Cytokines, platelet-derived growth factor, and other chemokines stimulate proliferation of these cells, resulting in hypercellular epiretinal membranes. Hyalocytes are also known to cause collagen contraction, resulting in a pucker. Recurrence of epiretinal membranes may also be explained by vitreoschisis wherein the anterior wall may be removed at surgery and the cells in the residual posterior wall proliferate and contract. OCT is able to detect these membranes and their interplay with the retina, except when the membranes are extremely thin.
Two distinct clinicopathological features of membranes removed from eyes with vitreomacular traction syndrome without obvious PVD suggest different forms of epiretinal fibrocellular proliferation: (1) multilayered cellular membranes separated from the ILM by a layer of intervening native vitreous collagen in eyes with visible epiretinal membrane; and (2) single cells or a cellular monolayer directly on the ILM with no visible epiretinal membrane. The predominant cell type is myofibroblast that contributes to the contractility, in both types. The higher number in the multilayered membranes may explain the cystoid macular edema and progressive vitreomacular traction characteristic of this disorder. Overall, it appears that the location of the split in the posterior vitreous cortex and variable cellular proliferation determine the nature and severity of vitreofoveal traction and epiretinal membrane formation.
Vitreous Traction Maculopathies
Changes in the vitreous gel may cause traction on the retinal surface and macular distortion through several different mechanisms, including: (1) incomplete PVD in which the vitreous remains attached focally to the macular surface, resulting in macular cysts or macular detachment; (2) anomalous posterior vitreous separation resulting in vitreoschisis with continued broad foveal traction causing macular cysts; (3) vitreoschisis with proliferation of the posterior layer into epiretinal membrane causing macular distortion; (4) vitreous gel condensation and shrinkage caused by inflammatory, vascular, and metabolic diseases in the absence of hyaloid separation; (5) complete PVD with subsequent epiretinal membrane formation; and (6) a peculiar form of traction maculopathy that is related to focal tangential contraction of the PVC, anterior displacement of the foveal retina, resulting in idiopathic macular hole.
Traction Maculopathy Caused by Incomplete Posterior Vitreous Detachment
Approximately 30% of patients (average age 60 years) who develop a symptomatic PVD will have evidence of vitreous hemorrhage or a peripheral retinal tear or both. This affects twice as many women as men. In all, 10–15% of these patients will develop a PVD in the second eye, usually within 2 years. Vitreous hemorrhage that is usually caused by a demonstrable peripheral full- or partial-thickness retinal tear is the most frequent cause of transient loss of vision in patients after an acute PVD ( Figures 7.05D and 7.07A ). In most cases the macula is unaffected by the PVD. Small hemorrhages around the optic disc, along the major vascular arcades, and less frequently in the macula may be the only sign of microtrauma to the retina caused by the PVD ( Figure 7.07A and B ). When the PVD is impeded by abnormal vitreoretinal adhesions in the macular area, traction and distortion of the macula may cause blurring of vision, metamorphopsia, and occasionally a scotoma ( Figures 7.05A, G, I, and J , and 7.07C–F and J–L ). Ophthalmoscopic and biomicroscopic examination reveals a partial PVD and tenting of the retina at the site of the vitreoretinal adhesion. This site may be localized in the parafoveal region rather than directly in the foveal area. If the onset of symptoms is recent, the vitreoretinal adhesion may separate in a matter of days or weeks and visual function may be restored to normal ( Figure 7.07E, F, and L ). These patients, however, may subsequently develop evidence of epiretinal membrane ( Figure 7.07G ). In a few cases an epiretinal membrane may develop before PVD (a visible Weiss ring) occurs ( Figure 7.08A–C ).
In some patients vitreoretinal adhesion in the macular area is sufficiently dense that prolonged traction causes distortion, cystic edema, degeneration, and detachment of the macula. This may be caused by a linear area of attachment of the posterior hyaloid to the retinal surface ( Figure 7.07C ), a single condensed strand of vitreous attached to the paracentral retina, a cone-shaped mass of condensed vitreous with attachment to the entire foveal inner surface ( Figures 7.05C and 7.07J and L ), and paracentral traction at the major vascular arcades ( Figure 7.07A and H ). Vitreoschisis with splitting of the posterior hyaloid into two layers and persistent broad attachment of the posterior layer to the retina may be responsible for this ( Figure 7.08I ). When the partly detached vitreous remains attached to the center of the macula, the retina is tented anteriorly, causing a localized tractional serous retinal detachment that is surrounded by radiating retinal folds ( Figures 7.05C , and 7.07J and L , 7.08C, E–H ). Cystic changes are often evident centrally ( Figure 7.08B ). Prolonged vitreous traction may be associated with angiographic evidence of retinal capillary permeability alterations, and development of an epiretinal membrane in the area of vitreoretinal adhesion. Spontaneous separation of the adhesion may eventually occur ( Figure 7.07J–L ). Surgical separation of the vitreoretinal attachment may be required to reattach the macula ( Figure 7.09J–L ). Lysis of vitreoretinal adhesions may be accomplished in some cases with Q-switched neodymium laser.
Vitreous traction at the site of a major retinal vessel may cause not only a tractional retinal detachment that extends into the macula but also avulsion of the blood vessel ( Figures 7.05D and 7.07H ), vitreous hemorrhage, proliferative retinopathy ( Figure 7.09 ), and infrequently a full-thickness retinal hole.
Vitreous traction on the optic nerve head and juxtapapillary retina may cause a fundus picture that simulates papilledema, optic disc capillary angioma, astrocytoma, or combined retinal pigment epithelium (RPE) and retinal hamartoma ( Figures 7.05E and 7.07I , 7.09D–G , and see Figure 7.26 ).
In some cases with unusual adherence of the center of the fovea to the vitreous, a PVD that begins in the extramacular area may cause either a partial (lamellar) or full-thickness macular hole as it extends through the macular area ( Figures 7.05E and 7.09H, J–L ). This, however, is an infrequent mechanism for causing a macular hole (see discussion of idiopathic macular hole in a subsequent section).
Idiopathic Traction Maculopathy Unassociated with Posterior Vitreous Detachment
There is some evidence to suggest that subtle changes may occasionally occur in the vitreous body, causing it to contract and to exert anterior traction on the retinal surface posteriorly, without any biomicroscopic evidence of posterior separation, or discrete vitreous bands attached to the inner retinal surface ( Figure 7.10E–I ). This traction may be associated with cystoid macular edema and angiographic evidence of retinal capillary leakage in the macular area or serous detachment of the sensory retina. (See discussion of diabetic traction maculopathy, p. 544, and congenital pit of the optic nerve head, p. 1260.) Vascular leakage or inflammation with or without secondary vascular permeability alteration can result in vitreous contraction without posterior separation in conditions such as pars planitis ( Figure 7.10A–D ), retinal capillary hemangioma (see Chapter 13 ), Coats’ disease, and others.
Traction Maculopathy Caused by Spontaneous Contraction of the Prefoveolar Vitreous Cortex Unassociated with a Posterior Vitreous Detachment
Idiopathic Age-Related Macular Hole
Idiopathic age-related macular hole, referred to henceforth in this section as macular hole, affects predominantly older patients, more often women at a ratio of 2 or 3:1. They often discover blurred vision and metamorphopsia when they cover the fellow normal eye. Most patients report that both eyes were normal during their last eye examination 1 or 2 years previously. From the pathogenetic and therapeutic standpoint, it is important to differentiate idiopathic age-related macular hole from the less common causes of macular hole, such as trauma or macular traction resulting from incomplete posterior vitreous separation, transvitreal bands of vitreous condensation, or neighboring epiretinal membranes. An understanding of the structure of the vitreous and aging changes in its structure discussed previously in this chapter, and the ultrastructure of the foveolar retina ( Figure 7.08L ) are important in considering the pathogenesis of age-related macular hole, which typically begins in eyes with an optically empty liquefied premacular vitreous and no evidence of posterior vitreous separation.
It is likely no other condition in ophthalmology has elicited as much debate and controversy about its pathogenesis as macular hole, since the original description by Gass. It is important to bear in mind that Gass described the mechanism much before OCT was available. With the advent of OCT we began to view the interplay between the fovea and the vitreous, a concept that Don Gass had visualized by his keen observation with a fundus contact lens and interpretation. The earliest observations with the first-generation OCT were made by Alain Gaudric, where he introduced the terms “foveal cyst” and “anteroposterior traction forces” at the vitreofoveal junction associated with localized perifoveal vitreous detachment. Subsequent reports have continued to use this as the basis of macular hole formation. Gass, who constantly reappraised the understanding of various diseases, in 1999 wrote about the forgotten “Müller cell cone” ( Figure 7.11B ) and postulated dissolution of the Müller cells in eyes destined to develop a macular hole. An inverted cone of Müller cells occupying the inner half of the fovea centralis was first shown by Yamada et al. histopathologically in a 45-year-old woman ( Figure 7.08L ). Gass postulated breakdown of the compact arrangement of the Müller cells and perhaps their movement into the prefoveal vitreous cortex initiating centripetal contraction that pops the foveal depression up ( Figure 7.10J–L ). It is likely that the dissolution of the Müller cell cone weakens the fovea which splits by the contracting forces of the overlying vitreous. The concentric contraction of the prefoveal vitreous cortex can explain the localized perifoveal vitreous detachment seen on OCT ( Figure 7.11C ) by foreshortening the vitreous cortex. Subsequent reports have alluded to this concept. The role of the Müller cell cone changes in eyes destined to form a macular hole is further strengthened by the observation of various other anatomic appearances such as foveal cysts, foveal detachment, and diffuse foveal/macular thickening ( Figures 7.07 and 7.08 ), in eyes with vitreomacular traction alone, none of which leads to a macular hole. Figure 7.11F also illustrates that tangential, rather than anteroposterior, forces play a role in macular hole formation; the operculum is lying close to the foveola rather than being drawn more anteriorly which should have occurred if anteroposterior traction forces were predominant.
Table 7.1 and the diagrams in Figures 7.10 (J–L) and 7.11 summarize the characteristic biomicroscopic features and the presumed anatomic changes accompanying each of the stages of development of a macular hole.
Stage | Biomicroscopic findings | Anatomic interpretation |
---|---|---|
1-A (impending hole) | Central yellow spot, loss of foveolar depression, no vitreofoveolar separation | Early serous detachment of foveolar retina |
1-B (impending or occult hole) | Yellow ring with bridging interface, loss of foveolar depression, no vitreofoveolar separation | Small ring – serous foveolar detachment with lateral displacement of xanthophyll. Large ring – central occult foveolar hole with centrifugal displacement of foveolar retina and xanthophyll, with bridging contracted prefoveolar vitreous cortex. Cannot detect transition from impending to occult hole |
2 | Eccentric oval, crescent, or horseshoe retinal defect inside edge of yellow ring | Hole (tear) in periphery of contracted prefoveolar vitreous cortex bridging round retinal hole, no loss of foveolar retina. Central round retinal defect with rim of elevated retina |
With prefoveolar opacity | Hole with pseudo-operculum, * rim of retinal detachment, no posterior vitreous detachment from optic disc and macula | |
Without prefoveolar opacity | Hole without pseudo-operculum or posterior vitreous detachment | |
3 | Central round ≥400-μm-diameter retinal defect, no Weiss’s ring, rim of elevated retina | |
With prefoveolar opacity | Hole with pseudo-operculum, no posterior vitreous detachment from optic disc and periphery of macula | |
Without prefoveolar opacity | Hole without pseudo-operculum, no posterior vitreous detachment from optic disc and macula | |
4 | Central round retinal defect, rim of elevated retina, Weiss’s ring | |
With small vitreous opacity near temporal edge of ring | Hole with pseudo-operculum, posterior vitreous detachment from optic disc and macula with mobile Weiss ring and pseudo-operculum † | |
Without small opacity | Hole and posterior vitreous detachment from optic disc and macula without pseudo-operculum |
* Pseudo-operculum contains no retinal receptors.
Stage 1-A: Impending Macular Hole
Although the earliest precipitating event responsible for the progression of changes leading to a macular hole has not been identified, the author believes that proliferation of Müller cells located in the center of the normal foveola ( Figures 7.8 and 7.10J–L ) and their extension through the ILM at the umbo into the outer part of the layer of formed PVC is most likely responsible for causing contraction, condensation, and partial loss of transparency of the outer part of vitreous cortex in the foveolar and perifoveolar region. Retinal astrocytes and vitreocytes would seem to be less likely candidates as cells responsible for inducing contracture of the PVC. Tangential contraction of the outer part of the prefoveolar cortical vitreous causes anterior displacement and serous detachment of the foveolar retina ( Figures 7.10K and 7.11B ). Biomicroscopically a yellow spot appears centrally ( Figures 7.12A and 7.13A ). This spot is caused by greater visibility of the retinal xanthophyll, which is highly concentrated in the receptor cells and nerve fiber layer in the foveolar region. It is more apparent as the retina separates from the RPE. The patient, particularly if he or she has a macular hole in the fellow eye, may experience the abrupt onset of metamorphopsia unassociated with photopsia or floaters. The visual acuity may be almost normal. Distortion of the Amsler grid is usually present. Biomicroscopically, there is no evidence of a PVD but there is loss of the normal foveolar depression and the foveal reflex. Fluorescein angiography often shows a focal area of faint fluorescence centrally ( Figure 7.13B ).
Stage 1-B: Impending Macular Hole
As the foveal retina elevates to the level of the surrounding thick perifoveal retina ( Figures 7.10L and 7.11C ), the retinal receptor layer is put on stretch and thinning of the foveolar retina around the umbo causes a change in the biomicroscopic appearance from a yellow spot to a small donut-shaped yellow ring lesion ( Figures 7.12D and 7.14A ). Although a yellow spot occurs with foveal detachment from other causes, e.g., idiopathic central serous chorioretinopathy, the change from a spot to a ring is peculiar to patients developing a macular hole.
Stage 1-B: Occult Macular Hole
Whereas the small central area of translucence in the center of the yellow spot may result from attenuation of the foveolar retina, it is probable that the clearly defined yellow ring that develops soon afterward is caused by a break in the continuity of the receptor cell layer at the umbo, structurally the thinnest and weakest site in the retina. This is followed by centrifugal movement of the foveolar retinal receptor cells, their radiating nerve fibers, the Müller cells, and the xanthophyll beneath the ILM of the retina and the contracted PVC ( Figure 7.11A(d and e) ). Initially the ILM of the foveolar retina and the thin layer of horizontally oriented Müller cell processes separating it from the retinal receptor cells may not be involved in the central retinal break. Regardless, as long as the contracted prefoveolar vitreous cortex bridges the hole, it may be visible biomicroscopically as a semitranslucent interface. Thus the change from a stage 1-B impending hole to a stage 1-B occult hole cannot be detected biomicroscopically. Reactive proliferation of Müller cells and retinal astrocytes occurring within the area of the receptor cell dehiscence probably contributes to the opacification of the tissue bridging the defect and, in some cases, may cause ruffled edges of the retinal dehiscence surrounded by fine radiating retinal folds ( Figure 7.12G and H ).
Fluorescein angiography in stage 1-B lesions of all sizes usually shows hyperfluorescence of variable intensity centrally. Although a high intensity of fluorescence is more suggestive that a full-thickness hole is present, angiography is not reliable in this regard.
Stage 2 Hole
Spontaneous vitreofoveal separation may occur soon after the central retinal dehiscence, and the contracted PVC becomes visible as a semitranslucent prehole opacity or operculum-like structure lying anterior to a small foveolar hole ( Figure 7.11A(f) ). Initially the diameter of this opacity is often larger than that of the foveolar hole. In a few patients with early stage 1-B lesions, separation of the PVC may be accompanied by avulsion of part of the foveolar retina, resulting in a true operculum formation. Biomicroscopy, however, cannot determine the presence or absence of retinal tissue in the prehole opacity. In some patients the contracted PVC, either while it remains attached to and bridges the macular hole or after it separates from the perifoveolar retina, may be transparent and undetectable biomicroscopically. In such cases very small stage 2 holes without a prehole opacity may be evident. In most patients, however, the contracted vitreous cortex is semitransparent and remains attached to the inner retinal surface surrounding the retinal hole as the foveolar retina continues to retract centrifugally ( Figure 7.11A(e) ). Biomicroscopically there is progressive enlargement of the yellow ring, which may become serrated along its inner margin that corresponds to the edge of the occult round retinal hole ( Figure 7.12H–K ). Eventually the first biomicroscopic evidence of a dehiscence may occur in the semitransparent vitreous cortex at the inner edge of the yellow ring ( Figures 7.11A(g) and 7.12G ). In the area of the dehiscence, the serration of the yellow ring disappears, presumably because of relief of traction on the edge of the retinal hole, and the yellow pigmentation fades, possibly as a result of diffusion of xanthophyll out of the retina in this area. Over a period of days or weeks, further enlargement of the macular hole and additional contraction of the PVC cause a can opener-type 360° tear in the contracted PVC, separating it from the less condensed outer vitreous cortex at the edge of the retinal hole ( Figures 7.11A(h) and 7.12G–L ). The contracted prefoveal vitreous cortex is visible biomicroscopically as an operculum-like opacity (pseudo-operculum) suspended anterior to the hole on the posterior surface of the layer of transparent vitreous gel that bridges the hole and lies along the inner retinal surface in the macula. This prefoveolar opacity oscillates slightly with eye movements. It is usually not possible to detect biomicroscopically an interface caused by the layer of transparent vitreous cortical gel surrounding the pseudo-operculum.
Stage 3 Hole
Centrifugal retraction of the foveolar retinal receptors continues until the hole becomes fully developed and its diameter in all but a few cases reaches 400–600 μm ( Figures 7.11A(h) , 7.12F and L, and 7.13C, G, I, and K ). All stages of progressive enlargement of the hole are considered as stage 2 holes. Since the ultimate diameter of the hole is variable, for purposes of classification the author suggests that all holes less than 400 μm in diameter be considered stage 2 holes.
Stage 3 holes are associated with a mean visual acuity of 20/200 with a range from 20/40 to 5/200. The sharply outlined 400–600-μm-diameter hole is typically surrounded by a 1000–1500-μm-diameter gray rim of retinal detachment. The patient describes metamorphopsia on the Amsler grid. A well-defined central scotoma, however, is usually difficult to demonstrate on the grid. Microperimetry using the scanning laser ophthalmoscope demonstrates an absolute scotoma corresponding with a macular hole, and a relative scotoma corresponding with the rim of retinal detachment surrounding the hole. With further refinement of the perimetric technique, it is probable that visual defects that extend beyond the area corresponding with the rim of detachment will be detected. Ninety-five percent of patients report a gap in a narrow slit beam of light (positive slit-beam sign, Watzke sign) when the slit of light is directed through a fundus contact lens into the center of the macular hole. A 50-μm krypton or argon laser aiming spot placed within a stage 3 or 4 hole will not be seen in almost 100% of patients (positive laser beam sign). Thinning and depigmentation of the RPE develop within the area of the hole. A pigmented demarcation ring may occur ( Figure 7.15F ). Within most holes there are several yellow nodular opacities at the level of the RPE ( Figures 7.12E, F, and I , 7.13K , and 7.15C and D ). These opacities change in number and distribution from one examination to the other. The RPE and choroid surrounding the hole typically appear normal, although some patients may have drusen. Several small intraretinal cysts may be evident near the margin of the hole. In 10–20% of patients, fine crinkling of the inner retinal surface caused by an epiretinal membrane develops around the hole. The membrane occasionally may distort the contour of the hole ( Figure 7.13K ). An operculum-like structure (contracted PVC) is suspended on the posterior surface of the hyaloid membrane in front of the hole in 75–85% of cases ( Figures 7.12B, F, and L, and 7.13C and I ). In these cases there is no evidence of a PVD except in the foveal area.
Stage 4 Hole
After separation of the vitreous from the entire macular surface and optic disc, irrespective of its diameter, the hole is designated stage 4 ( Figures 7.11A(i) and 7.12C ). The operculum-like opacity can often be found attached to the mobile posterior hyaloid membrane near the temporal side of the Weiss ring.
Fluorescein angiography in patients with stages 2, 3, and 4 holes typically demonstrates prominent early hyperfluorescence caused primarily by the absence of xanthophyll in the area of the hole but is also caused by RPE thinning, depigmentation of the RPE, and slight loss of transparency of the retina immediately surrounding the hole. In patients with a long-standing hole, the central zone of hyperfluorescence may be surrounded by a rim of faint fluorescence corresponding with the rim of retinal detachment and the underlying hypopigmented RPE ( Figure 7.13L ). In a few patients with a heavily pigmented choroid, the fluorescence may be minimal or absent. The yellow deposits within the depth of the hole and the operculum overlying the macular hole ( Figure 7.13J and L ) usually appear nonfluorescent or hypofluorescent.
If the anatomic interpretations summarized in Figure 7.11A are correct, the implications include the following: (1) most macular holes develop as the result of a central retinal dehiscence at the umbo, followed by centrifugal displacement of the relatively normal complement of retinal receptors; (2) this dehiscence occurs soon after the change from a yellow spot (stage 1-A impending hole) to a yellow ring lesion (stage 1-B impending hole), but in most cases it is not detectable with a thin slit beam as a defect in the center of the ring because of the presence of the semitranslucent condensed cortical vitreous bridging the hole (stage 1-B occult hole); (3) most of the prehole opacities overlying stage 2 and stage 3 holes are condensed prefoveal vitreous cortex (pseudo-opercula), not opercula; and (4) following successful vitreous surgery, which includes tamponade of the hole with an intravitreal gas bubble, if done within 1 year of commencement of hole formation, the anatomy of the central retina and its visual function may be restored to nearly normal levels in some patients as a result of retinal reattachment and centripetal repositioning of the retinal receptors. If these concepts of the anatomic changes occurring in macular hole development are correct, histopathologic examination of the prehole opacities should determine that most of them contain no retinal receptor cells but are composed of vitreous collagen, reactive Müller cell and astrocytic proliferation, and in some cases ILM of the retina. Although retinal opercula have been described histopathologically in two eyes, one with posttraumatic and the other with an idiopathic macular hole, it is uncertain whether the “opercula” contained retinal receptor cells. Opercula have not been observed in most idiopathic macular holes studied histopathologically.
Spontaneous Abortion of Macular Hole Formation
In approximately 50% of cases, patients with stage 1-A and early stage 1-B lesions may experience rapid improvement in visual symptoms because of spontaneous separation of the vitreous from the fovea without developing a full-thickness macular hole ( Figure 7.14 ). In such cases, the patient usually notices improvement in the symptoms and biomicroscopy may show several different pictures, all of which are accompanied by return of the foveal depression and a good visual prognosis.
Vitreofoveal Separation and Pseudo-operculum Formation
The foveal area returns to a normal appearance except for the presence of a semitranslucent, operculum-like structure or pseudo-operculum (condensed, contracted prefoveal vitreous cortex) immediately in front of the fovea ( Figure 7.14A–C ). When viewed obliquely with a thin slit beam, the pseudo-operculum in a few patients will cast a yellow shadow on the pigment epithelium. Some patients with a pseudo-operculum will notice a small scotoma when reading and a few will describe it as having a yellow color.
Vitreofoveal Separation without Pseudo-operculum Formation
After spontaneous vitreofoveal separation the fundus returns to a normal appearance and no pseudo-operculum is evident.
Vitreofoveal Separation and Lamellar Hole Formation
Separation of the vitreous from the fovea in these patients is associated with a break in the continuity of the ILM and the biomicroscopic appearance of one or more sharply defined reddish defects in the inner retinal surface in the foveolar area ( Figures 7.14D–F and I and J , and 7.15H ). An operculum is usually evident overlying the defect. The defect may be minute and simulate that seen after sun gazing or in patients with no recognizable cause ( Figure 7.14J ). Larger lamellar holes often have a scalloped border ( Figure 7.14D ). Unlike full-thickness holes there is no rim of retinal detachment. The visual acuity is usually 20/30 or better. Fluorescein angiography shows minimal or no fluorescence in the area of the lamellar hole ( Figure 7.14E ). The demonstration of a focus of bright fluorescence within the area of the lamellar hole suggests the possible presence of a full-thickness hole without a rim of detachment ( Figure 7.13D–F ). This type of full-thickness hole appears identical biomicroscopically to a lamellar hole, may be associated with visual acuity of 20/30 or better, and is likely to develop a rim of retinal detachment and be associated with visual loss at a later date. The visual prognosis for patients with a lamellar hole is excellent.
Incomplete Separation of the Contracted PrefoveolarVitreous Cortex
A portion or all of the contracted PVC may remain as a small stellate opacity on the surface of the center of the foveolar retina and be associated with fine stellate retinal folds simulating X-linked foveomacular schisis ( Figure 7.14K and L ).
It is important to use a fundus contact lens to look for signs of vitreofoveal separation not only in the symptomatic eye but also in the fellow eye of any patient with evidence of a macular hole in one eye. Spectral domain OCT is very useful in detecting the relation of the posterior hyaloid to the fovea. Fellow eyes with evidence of vitreofoveal separation probably have less than a 5% chance of developing a macular hole. Some patients who have reportedly developed a hole after demonstration of a PVD probably had small occult holes that developed at the time of PVD. Others probably had residual vitreous cortex on the inner retinal surface centrally after separation of the vitreous from the optic disc and paracentral macular area.
Natural Course
The time course from the development of symptoms and stage 1 impending macular hole to a fully developed stage 3 or 4 hole varies but in most patients is within 6 months. In some patients the course may be complete within a matter of weeks, and in others hole formation may not have progressed beyond stage 2 till several years later. The visual acuity usually stabilizes after the first 6–12 months at a mean level of 20/200. A few patients with stage 3 and 4 holes may maintain excellent visual function of 20/40 to 20/50 for years.
Spontaneous reattachment of the retina surrounding the hole may occur ( Figures 7.15G and 7.16 ) and the biomicroscopic appearance may be identical with that of a lamellar macular hole. In some cases the hole may disappear and recovery of vision may be excellent. Closure of a macular hole occurs occasionally as the result of development of an epiretinal membrane ( Figure 7.16 ).
Approximately 25% of patients with a macular hole have evidence of posterior vitreous separation from the optic disc and macula in the fellow eye. The macula of the asymptomatic eye is usually normal but it may show evidence of previous spontaneous separation of the vitreous that is limited to the foveal area. In addition, other minor changes may occur at the vitreoretinal interface, including epiretinal membrane formation, small irregular folds of the inner retinal surface, and absence of the foveal reflex. Fluorescein angiography in the “asymptomatic eye” is typically normal. The value of focal electroretinography in detecting predilection for hole development in the fellow eye is uncertain. The only finding of definite prognostic significance in the fellow eye is the presence or absence of a PVD. The reported risk for development of a hole in the normal fellow eye has varied from 1% to 22%. The probable risk is between 10% and 15%. The presence of a PVD or vitreofoveal separation probably reduces the risk of developing a hole to 1% or less. Possible explanations for the occasional patient with a PVD who develops a hole include: a tear in the posterior hyaloid at the time of PVD, leaving vitreous cortex attached to the central macular area, and a subclinical full-thickness microhole caused by traction during the PVD. In most patients the second eye becomes involved within 2 years. Those with bilateral involvement usually retain moderately useful central vision, and most can read successfully with high-power spectacles.
Pathology and Pathogenesis
Immunocytochemical labeling and electron microscopic examination of vitreous removed at the time of surgery for impending macular holes has demonstrated cortical vitreous containing RPE and glial cells. Histopathologic examination of a macular hole has failed to demonstrate any evidence of either retinal or choroidal vascular disease as a cause for the development of a macular hole ( Figure 7.15 ). The edges of the hole are typically rounded, some cystic spaces in the outer plexiform and inner nuclear layers are often present, and there is frequently cellular proliferation from the edges of the hole on to the neighboring inner retinal surface ( Figure 7.15E ). A cellular prehole opacity may occasionally be observed ( Figure 7.15A and B ). It is not known whether or not retinal receptor cells are included.
Nodular proliferations of the RPE overlying an eosinophilic material probably account for the yellowish deposits noted biomicroscopically in the depth of the macular hole ( Figure 7.15C and D ). These appear to be identical in structure to the reactive, proliferative type of drusen noted histopathologically in eyes with long-standing retinal detachment. Proliferation of the RPE is probably caused by loss of the RPE’s contact with the outer segments of the retinal photoreceptors as well as its exposure to the vitreous. There are two reports in which histopathologic examination of the second eye in patients with unilateral macular hole found cystic spaces in the outer plexiform layer in the paracentral macular area. The fact that fluorescein angiography in asymptomatic second eyes, as well as in the affected eyes, shows no permeability alterations suggests that these cysts probably are not caused by abnormal retinal vascular permeability. In spite of these findings, which suggest that a slow process of cystic degeneration of the center of the fovea may predate the development of a macular hole, Gass’s observations suggest that macular hole development is not preceded by a gradual change in either the appearance of the macula or visual function. On the contrary, its formation begins abruptly, although its full development usually occurs over a period of 2–3 months. There is no evidence to incriminate either the underlying RPE or the choroid in the pathogenesis of macular hole formation. As discussed previously the primary tissues involved in macular hole formation involve the vitreoretinal interface region in the foveolar area. Electron microscopic study of a series of prefoveolar operculum-like structures collected at the time of macular hole surgery are needed to confirm the anatomic interpretation of the biomicroscopic stages of hole development suggested here by the author.
The predilection for hole development to occur in women has suggested that ingestion of estrogenic compounds may be of importance in the pathogenesis of macular hole.
Treatment
In 1988 surgical separation of the PVC was suggested as a possible treatment to prevent hole formation in patients with stage 1 impending macula holes. Uncontrolled pilot studies of vitreous surgery for treatment of impending macular holes suggested that the surgery might be of benefit. The criteria used for an impending hole by these authors, however, were not confined to those of a stage 1 impending hole as defined by this author. Furthermore, there are many lesions that may simulate a stage 1 hole and misdiagnosis of an impending hole is frequent. A randomized, multicenter clinical trial to evaluate the effectiveness of surgical peeling of the vitreous for treatment of an impending macular hole in one eye in patients with a stage 3 or 4 hole in the fellow eye was organized in 1988. The results of this study of 62 patients showed that approximately 40% of eyes in both groups developed a full-thickness hole. The study was discontinued before definite conclusions could be reached because of a dramatic drop in patient recruitment coinciding with enthusiastic reports concerning treatment of full-thickness holes. With the following available information: (1) that 40–60% of stage 1 holes spontaneously abort; (2) that there is a high incidence of misdiagnosis of stage 1 holes; and (3) the apparently favorable results of surgery for full-thickness holes, it is probably prudent to observe patients with a stage 1 impending hole, particularly when the fellow eye is normal.
Kelly and Wendel in 1991, Glaser and colleagues in 1992, Poliner and Tornambe in 1992, and others, in uncontrolled pilot studies, reported successful closure of macular holes and visual improvement using pars plana vitrectomy, intraocular gas, and 1–2 weeks of face-down positioning ( Figure 7.18 ). Glaser’s group used a tissue growth factor, TGF-β, to stimulate proliferation of glial cells to seal the hole. Some surgeons have used the patient’s serum in lieu of TGF-β. The visual results obtained by all of these investigators were similar in those eyes with successful reattachment of the retina. Approximately 70% obtained improvement of two lines of visual acuity or better, and 20–40% regained 20/40 or better. These same authors, who originally obtained only approximately a 50% reattachment rate, have reported at recent meetings successful reattachment in 90–95% of cases and return of acuity to 20/40 or better in approximately 50% of cases ( Figure 7.19A–F ). Reoperation of eyes with failed macular hole surgery may result in visual improvement.
Since first conceptualized in 1999, removal of the ILM at the time of the vitrectomy has vastly improved the rate of hole closure and decreased the need for repeat surgery. Various methods to stain or delineate the transparent ILM have been employed using indocyanine green (ICG), trypan blue, and triamcinolone. Toxicity with use of ICG appears to be variable and related to the longer duration of the surgery and the higher intensity of the illuminating light. Surgical complications have included retinal tears with and without retinal detachment; retinal vascular occlusion; light toxicity pigment epitheliopathy; cataract; acute permanent loss of temporal visual field, probably caused by damage to the nasal part of the optic disc during fluid–gas exchange; and reopening of the macular hole. A controlled clinical trial randomizing patients with full-thickness holes to surgery or to observation is under way, and over 150 patients have been randomized.
Following successful reattachment of the retina after surgery for a macular hole, the macula may return to a nearly normal appearance, the hyperfluorescence corresponding with the hole preoperatively often disappears, and a central scotoma may no longer be demonstrable. A few patients may regain 20/20 visual acuity. These surgical results were difficult to explain on the basis of the initial anatomic interpretation of the biomicroscopic stages of hole development. The prehole opacity seen in 75–80% of patients was originally thought to represent an operculum containing the foveolar retina; this operculum was thought to be derived from a circumferential tear occurring in the periphery of a stage 1-B lesion, and significant improvement of vision even with reattachment of the retina around the hole was thought unlikely. Evidence now strongly suggests that nearly all macular holes begin as an occult central foveolar dehiscence at the umbo and that hole development is the result of centrifugal sliding and retraction of the receptors away from the center of the hole analogous to the opening of a lens diaphragm. It is understandable that surgical reattachment of the retina around the hole, when accompanied by reactive glial cell proliferation and contraction, may result in “closure of the lens diaphragm” and return of the foveal retina to its near-normal anatomic position and function in some patients ( Figures 7.18 and 7.19 ). The disappearance of the focal hyperfluorescence corresponding with the hole, and of the absolute central scotomas that occur in some patients after successful macular hole surgery, is further evidence in support of the concept of centripetal movement of the paracentral retinal receptors and the xanthophyll.
The criteria for recommending surgery for a macular hole have evolved over the past 20 years. At the start most patients undergoing surgery had symptoms for a year or less, visual acuity of 20/70 or worse, and a large stage 2 or stage 3 or 4 macular hole. The results of a randomized study and other reports broadened to include early stage 2 holes and holes of longer duration. Although the likelihood for progression to a stage 2 hole is probably directly related to the diameter of the stage 1-B yellow ring and the level of visual acuity loss, there is no reliable information or method at this time to determine which stage 1-B lesions will progress to hole formation.
The patient considering surgery for a macular hole in one eye and having normal function in the fellow eye should be aware of the following: (1) chances of developing a hole in the fellow eye are 10–15% and probably less than 5% in the presence of vitreofoveal separation; and (2) treatment usually involves two operations, including cataract surgery.
Laser treatment to the edge of macular holes has had minimal success in improving visual function. The treatment has no rationale as far as preventing further retinal detachment in patients with senile macular holes is concerned since the detachment involves only the central macular area and is unlikely to cause extensive detachment in these patients with relatively emmetropic eyes. There is less chance for visual improvement following surgical treatment of macular holes associated with other diseases, such as diabetic retinopathy, Behçet’s disease, or for holes caused by trauma.
Macular or paramacular holes do not cause rhegmatogenous retinal detachment unless they are associated with a posterior staphyloma and high myopia, or with vitreous bands causing traction on the surface of the posterior retina. In such cases the detachment infrequently extends beyond the equator. Treatment of such holes requires either one or a combination of permanent or temporary scleral buckling techniques; for example, the Klotti clip; vitrectomy; intravitreal injection of air or gas; and cryotherapy, diathermy, or photocoagulation.
Macular holes may develop in several clinical settings, some of which are unusual, e.g., Best’s disease, adult vitelliform foveomacular dystrophy, high myopia with posterior staphyloma, posterior microphthalmos, congenital arteriovenous aneurysm, hypertensive retinopathy, and following commencement of topical pilocarpine therapy, pneumatic retinopexy, and Nd-YAG posterior capsulotomy. Macular holes occurring in association with rhegmatogenous retinal detachment caused by peripheral retinal tears, trauma, myopia, contraction of an epiretinal membrane, and solar retinopathy are discussed elsewhere in this text.
Differential Diagnosis
Most patients referred to the author with a diagnosis of a stage 1-A impending hole have had a foveolar yellow lesion caused by one of the following: solitary drusen, small RPE detachment, idiopathic central serous chorioretinopathy, foveolar detachment with epiretinal membrane, bilateral idiopathic juxtafoveolar retinal telangiectasis, pattern dystrophy, cystoid macular edema, and solar maculopathy ( Figure 7.17 ).
Lesions that may simulate a full-thickness macula hole include an inner lamellar macular hole ( Figure 7.14D ), a hole in an epiretinal membrane ( Figures 7.17J–L and see Figure 7.21A ), geographic atrophy of the RPE ( Figure 7.17G–I ), choroidal neovascularization, a small focal area of central serous chorioretinopathy, cystoid macular edema with a large central cyst ( Figure 7.17A–C ), focal retinal atrophy associated with bilateral juxtafoveal retinal telangiectasis ( Figure 7.17D and E ) congenital optic pit, and a solitary macular cyst, a lesion that rarely occurs. Features of a full-thickness macular hole that differentiate it from most simulating lesions are the presence of a halo of retinal detachment surrounding the hole, yellow deposits within the depth of the hole, and a zone of hyperfluorescence corresponding to the size of the hole during the early stages of angiography. Use of the slit-beam test (Watzke sign), 50-μm-size aiming beam laser perimetry, OCT, and fluorescein angiography are helpful adjuncts to contact lens examination in arriving at the correct diagnosis. Echography is capable of detecting PVD and the presence of pseudo-opercula but appears to be no better than contact lens examination in this regard.