The etiologies of macular hole and macular pucker are poorly understood. Disparate in symptoms as well as in ophthalmoscopic features and imaging, these two conditions have long been considered distinctly different entities. Past theories of the pathogenesis of macular holes included ocular trauma,¹ cystic degeneration,² and ocular angiospasm.³ In recent years, emergent theories proposed that vitreous plays an important role in the formation of macular holes.^4,5 Similarly, theories on the development of macular pucker have evolved over time. One theory is that microbreaks in the retina caused by posterior vitreous detachment (PVD) allow for the migration of fibroblasts, glial cells, and astrocytes from the retina to the surface of the internal limiting lamina, where they proliferate.⁶

The most recent hypothesis is the unifying concept of anomalous PVD,⁵ which proposes that the first event in the pathophysiology of both macular hole and macular pucker is the same. Collapse of the liquefied vitreous body without sufficient dehiscence at the vitreoretinal interface can induce a split within the posterior vitreous cortex (vitreoschisis), leaving the outermost layer of the posterior vitreous cortex attached to the macula. To date, attempts to elucidate pathogenesis have been hampered by a lack of adequate imaging technology.

Combined optical coherence tomography/scanning laser ophthalmoscopy (OCT/SLO) was used to study eyes with macular holes and macular pucker in search of vitreoschisis. Quantitative analysis of the topographic features of these 2 conditions was undertaken using the unique features of this imaging technology.

In coronal plane imaging, the grayscale SLO confocal fundus image (Figure 1A) and the threshold color OCT image in the coronal plane (Figure 1C) can be superimposed. There is point-to-point correspondence between the 2 images (coronal OCT and SLO; Figure 1B), since they are obtained simultaneously using parallel detector systems. The superimposed images were especially useful for identifying centers of retinal contraction in each subject. A pucker contraction center was defined as an area where radially oriented retinal striations converged. The center of an area of retinal contraction was determined using 3-dimensional rendering (manufacturer-provided) of the intersection between a longitudinal OCT scan and the SLO image (Figure 2). The point on the longitudinal OCT scan that corresponded to the macular pucker contraction center, as visualized by convergent striations, was defined as the center of retinal contraction. This point was used to measure retinal thickness from the internal limiting lamina of the retina to the anterior aspect of the outer layer of the double laminar structure representing the choriocapillaris/retinal pigment epithelium interface,⁷ using the calibrated calipers of the OCT/SLO software (Figure 3).

FIGURE 1 Analysis of topographic features in macular pucker. A, Grayscale SLO image of the fundus in a subject’s left eye. B, Overlay of the coronal OCT image (color photo C) superimposed on the SLO fundus image (grayscale photo A) with point-to-point (more ...)

FIGURE 2 Intersecting planes analysis of macular hole. Three-dimensional reconstruction of a transverse OCT image at 45 degrees intersecting with the SLO fundus image of a patient with a macular hole. The point that was identified as the center of contraction (more ...)

FIGURE 3 Quantitative analysis of retinal contraction in macular hole. Longitudinal OCT imaging was performed at an axis of 45 degrees (red line with arrowhead on the small grayscale SLO images of fundus; to the right). A, White arrow points to the center of eccentric (more ...)

The SLO fundus images with superimposed coronal plane OCT scans were analyzed quantitatively with Adobe PhotoShop software (Figure 4). The diameter of the contraction center was determined by visualizing the region of the pucker where the striations converged in confluence. The ruler function of the software was used to measure the diameter. This value was then divided in half to yield the radius, which was used to calculate the surface area of pucker. The ruler was also used to measure the distance between the retinal contraction centers and the center of the macula.

FIGURE 4 Quantitative analysis of coronal plane OCT images in macular holes. Measurement of the distance from the center of contraction to the center of the macular hole and the diameter of retinal contraction area was performed using Adobe Photoshop software. (more ...)

Investigations into the pathogenesis of macular holes and macular pucker have been hampered by the lack of experimental models and adequate imaging technologies. Using combined OCT/SLO imaging, heretofore unidentified features of the vitreoretinal interface in macular hole and macular pucker could be analyzed both qualitatively and quantitatively. Vitreoschisis was found in half of all eyes. This is likely related to the underlying multilamellar structure of the posterior vitreous cortex (Figure 8), which can split into separate layers during anomalous PVD. Previously detected in proliferative diabetic retinopathy,^8–10 vitreoschisis has not been described in macular hole and macular pucker and was probably found in this study owing to the increased resolution of the combined OCT/SLO technology. The “lambda” (Y-shaped) configuration seen in Figures 6 and 7 demonstrates that 2 layers arise from a split in the lamellar posterior vitreous cortex. Although the remaining cases in this study did not feature the “lambda” sign, there were many eyes that had a visible posterior vitreous cortex anterior to the retina with a distinct membrane attached to the anterior surface of the retina. Although this was not considered vitreoschisis in this study, future studies with higher-resolution imaging might be able to ascertain whether the incidence of vitreoschisis is even greater than identified in this investigation.

FIGURE 8 Lamellar structure of posterior vitreous cortex. Immufluorescent micrograph of human vitreoretinal interface. Using anti-ABA lectin binding, the internal limiting lamina stains intensely as a horizontal line in the middle of this photograph, and the lamellae (more ...)

These studies found that 40% of eyes with macular hole had evidence of an eccentric premacular membrane with retinal contraction. However, there were more cases with multiple foci of retinal contraction in eyes with macular pucker (45.5%) than in eyes with macular hole (28%). What’s more, retinal contraction centers in eyes with macular hole were 3 times smaller (23 vs 63 mm²; P = .006) than the retinal contraction centers in eyes with macular pucker. Thus, if anomalous PVD were indeed an important event in macular hole and macular pucker, there must be some other factors to account for the dissimilar clinical and topographic features of the two conditions.

One difference between macular hole and macular pucker could be the level at which vitreoschisis splitting occurs during anomalous PVD.⁴ As shown in Figure 8, there are many lamellae within the posterior vitreous cortex. Also present within the posterior vitreous cortex are hyalocytes (Figure 9). These mononuclear phagocytes are embedded within the posterior vitreous cortex,^11–13 approximately 50 μm from the internal limiting lamina of the retina.^14,15 If the level of the vitreoschisis split in the posterior vitreous cortex occurs posterior to the level of hyalocytes, these cells will detach away from the retinal surface along with the anterior portion of the vitreous cortex. This is likely what happens in macular hole, leaving a relatively hypocellular layer of vitreous attached to the macula. The relative paucity of cells left on the surface of the retina could account for the findings reported herein: that a minority (40%) of eyes with macular hole have retinal contraction; that in those eyes with macular hole with retinal contraction, the number of contraction centers is never more than 2 (compared to 3 and 4 centers in macular pucker); and the surface area of these retinal contraction centers is relatively small. At surgery, these membranes are relatively thin and more difficult to engage than are the membranes found in macular pucker.

FIGURE 9 Top left, Human hyalocytes in situ. Dark-field slit microscopy of the posterior vitreous in a 59-year-old man. Arrow indicates hyalocytes embedded within the posterior vitreous cortex. Top right, Explanted human posterior vitreous cortex with hyalocytes. (more ...)

In eyes with macular pucker, the vitreoschisis split likely occurs more anteriorly, leaving hyalocytes attached to the macula. As mononuclear phagocytes of the reticuloendothelial cell system, sentinel hyalocytes can stimulate the migration and proliferation of monocytes from the systemic circulation and glial cells from the retina, resulting in more pucker centers and larger areas of retinal contraction, as was found in the present study. Indeed, studies¹⁶ have shown that hyalocytes may play an important role in proliferative disorders of the retina via the release of connective tissue growth factor. Other studies¹⁷ have furthermore shown that hyalocytes can induce collagen gel contraction in response to platelet-derived growth factor and other cytokines. Thus, hyalocytes are likely important in stimulating cell proliferation as well as inducing tangential vitreoretinal contraction, the two main components of macular pucker.

In conclusion, the findings of this study support the hypothesis that vitreoschisis occurs in macular holes and macular pucker, probably as a result of anomalous PVD. Future studies should be undertaken to correlate histopathologic analysis of tissue removed at surgery with clinical and OCT/SLO topography, so as to determine whether the cellular compositions are consistent with the differences noted on clinical evaluation and imaging. This should test the hypothesis that vitreoschisis in macular holes splits posterior to the level of hyalocytes, whereas splitting of the posterior vitreous cortex in macular pucker occurs anterior to the level of hylaocytes. Indeed, while vitreomacular pathophysiology may be more complex than previously thought, and while we wish to understand it as basically as possible, we should attempt to follow Albert Einstein’s sage advice: “Make everything as simple as possible, but not simpler.”

So-called anomalous posterior vitreous detachment (PVD)¹ is a concept that has been postulated by Dr Sebag for several years and includes (1) vitreous liquefaction with an accumulation of fluids in the premacular portion of the vitreous cavity; (2) varying degrees of persistent adhesion between the posterior cortical vitreous and the internal limiting membrane (ILM); (3) splitting, or “vitreoschisis,” of the posterior cortical vitreous at the time of PVD; and (4) residual islands of cortical vitreous attached to the ILM following separation of the majority of the gel from the surface of the retina. He has described this entity as important in the pathogenesis of a variety of important disorders, including macular holes (MH), idiopathic epiretinal membrane (ERM) formation, vitreomacular traction syndrome, proliferative diabetic retinopathy, diabetic macular edema, retinal tears, and additional problems.

In this report the authors compared ERMs in eyes with MH to those with idiopathic ERM alone, termed macular pucker (MP) cases. ERMs were present in 40% of eyes with MH. Compared to idiopathic MP cases, ERMs in MH eyes were smaller and their centers were located farther from the fovea. Similar retinal thicknesses were discovered beneath the centers of the ERMs in both groups of cases. Only 2 (11%) of the 18 MH cases had a complete PVD, whereas this was observed in 18 (75%) of the24 MP eyes. Visual acuities were substantially better in the MP cases.

Apparent vitreoschisis was identified in 24 (53%) of 45 eyes with MH and 19 (43%) of 44 MP cases. The authors propose (1) that vitreoschisis at varying levels within the posterior vitreous cortex may be responsible for the differences in ERMs associated with these entities, and (2) that the ERMs are derived from hyalocytes or that these cells stimulate ERM formation by other types of cells.

The burning question posed by a reading of this manuscript is the viability of the authors’ proposition of anomalous PVD as the best explanation for these macular disorders. It is now generally accepted that some degree of PVD plays an important role in the pathogenesis of both MH and ERM formation and that islands of residual vitreous cortex frequently remain on the inner surface of the retina following both spontaneous and surgically created PVD. Still, the debate regarding the types of cells that produce ERMs and the means by which they reach the retinal surface has continued for decades, and this study does not provide information that assists in solving the puzzle.

Ultrastructural and immunocytochemical analyses are more likely than imaging studies to improve our understanding regarding ERM formation. Published data that are at odds with the authors’ opinions include (1) a study by Messmer and associates² that demonstrated similar electron microscopic features of ERMs associated with both MP and MH cases; (2) a report by Chen and coworkers³ demonstrating a lack of correlation between residual islands of retained vitreous cortex at surgery and postoperative ERM formation; and (3) a publication by Snead and colleagues⁴ that implicated glial cells as responsible for ERMs associated with both MP and MH. Still, I suppose it may be possible that metaplastic hyalocytes theoretically could assume characteristics of glial cells.

The authors’ conclusions appear to me to invoke a more complicated explanation or theory than is necessary, and this might appear to be a violation of “Ockham’s razor” that “entities should not be multiplied unnecessarily,” ie, the simplest viable explanation may be most likely to be the best. A simpler explanation has long been offered, and this is that PVD typically begins posteriorly near the foveloa and then extends more peripherally. The speed with which this occurs has recently been brought into question in Mark Johnson’s excellent 2005 AOS thesis, and transient or persistent foveolar and macular adhesions to the cortical vitreous may be responsible for a variety of maculopathies. Regardless, following even small amounts of PVD, cells can gain access to the retinal surface through a MH or through breaks in the ILM caused by PVD. I believe this rather old theory continues to provide a rational explanation of ERM formation, although it could be incorrect. My primary question for the authors, what are the weaknesses in the premise that this older explanation may be the best?

The authors are to be congratulated and thanked for introducing us to this exciting new imaging technique and for stimulating us with interesting material. My negative comments are intended as “constructive criticism,” and one should bear “Hanlon’s razor” in mind: “Never attribute to malice that which can be adequately explained by stupidity.”

Financial Disclosures: None.

Concerning the questions of Dr. Jampol, he wanted to know if it was possible to distinguish between epiretinal tissue, posterior vitreous cortex, and other membranes with this technology. I believe that the resolution of this current technology is just at the border of making that differentiation, Very shortly, in fact in two weeks, I will be receiving the spectral domain OCT. It is my expectation, based on observations I have made with the time domain version, that the higher resolution spectral domain will provide a more precise and reliable method to distinguish between these various tissues. John Thompson questions whether the difference in the age of the macular hole may have relevance in terms of the formation of these membranes. He asks if they are a secondary event or part of the primary pathophysiology. It is my opinion that these membranes are an essential part of the pathophysiology. There can be further thickening with older membranes and the spectral domain instrument will likely provide more exact measurements. We will be able to measure these and, with a sufficiently large data base, also be able to perform statistical analysis and to correlate the thickness of these membranes with the age of the macular hole. I believe that these membranes exist at the very beginning of the pathophysiologic process in this condition. Tim Stout questions whether integrins are present on the surface of hyalocytes. I believe resident macrophages are usually in a quiescent stage and function as sentinels that undergo up-regulation in response to certain noxious situations. I believe that future studies should be done to identify the up-regulation of integrins. This process is extremely important, not only in explaining the movement of these cells, but in understanding the stimulation, migration, and proliferation of the other cells that were previously mentioned.

Lastly, I do speak to patients in terms of “posterior vitreous separation” because I have had the same problem. Your point is well taken and our literature should reflect our clinical experience.

Funding/Support: None.

Financial Disclosures: Dr Rosen and Dr Garcia are consultants to OTI, Inc.

Author Contributions: Design and conduct of the study (J.S., P. Gupta.); Collection, management, analysis, and interpretation of data (J.S., P. Gupta, P. Garcia, R.R.R., A.A.S.); Preparation, review, and approval of the manuscript (J.S., P. Gupta, P. Garcia, R.R.R., A.A.S.)