Surgical Treatment of Diabetic Macular Edema

Pathophysiological and clinical evidence with a proposed classification scheme

Peer Reviewed

Surgical Treatment of Diabetic Macular Edema

Pathophysiological and clinical evidence with a proposed classification scheme.

Apurva K. Patel, MD, is a second-year retina fellow at the New York Eye and Ear Infirmary. Ronald C. Gentile, MD, FACS, is professor of ophthalmology at the New York Eye and Ear Infirmary, New York Medical College (Valhalla, NY) and attending surgeon at Winthrop University Hospital (Mineola, NY). Neither author reports any financial interests in any products mentioned in this article. Dr. Patel can be reached via e-mail at

Apurva K. Patel, MD • Ronald C. Gentile, MD, FACS

Diabetic retinopathy is the leading cause of blindness among working-age adults in the industrialized world, with clinically significant diabetic macular edema affecting 770,000 people in the United States.1-3

Without treatment, 50% of eyes with DME will lose 2 or more lines of vision within two years of diagnosis.4 Even though the incidence of DME can be reduced with systemic control of serum glucose, hypertension, and hypercholesterolemia, the need for ophthalmic treatment of DME is common.

Even with the availability of multiple therapeutic agents, including intravitreal anti-VEGF agents, focal laser, and, in some selective cases, corticosteroids, refractory DME continues to challenge retina specialists around the world. However, subsets of patients may benefit from a surgical approach to treatment of DME.

In this article, we review the pathophysiological and clinical evidence that support the surgical management of DME, and we discuss how to identify appropriate candidates. We also outline a proposed classification scheme for vitreoretinal interface abnormalities in patients with DME, along with illustrative cases.

This classification scheme can help in interpreting the literature and in assessing eligible patients for surgery. This is useful given the challenges in reviewing this topic as a result of heterogeneous study populations and the mix of pre-OCT and post-OCT reports in the literature.


The pathogenesis of DME is multifactorial and includes a cascade of complex biochemical processes. One theory is that the primary insult of hyperglycemia leads to the development and acceleration of advanced glycation end products (AGEs). AGEs form on proteins, lipids, and DNA, modifying their structures and functions. In addition to promoting leukostasis and causing damage to both pericytes and endothelial cells, AGEs are believed to promote mechanical changes in the vitreous and at the vitreous-retinal interface. Vitrectomy may help remove the AGEs from the vitreoretinal interface or remove abnormally adherent vitreous that could promote DME.5,6

“Even with the availability of multiple therapeutic agents, cases of refractory DME continue to challenge retina specialists.”

Posterior vitreous detachment has been associated with spontaneous resolution of DME.7 Eyes with DME have been shown to have a lower incidence of PVD compared to eyes without DME.8 Theoretically, vitrectomy, with or without membrane removal, may help improve DME via multiple mechanisms. These mechanisms include the release of abnormal vitreomacular adhesions, elimination of free and bound VEGF loads, and improved diffusion of oxygen to the retina from the vitreous cavity.9

Although vitrectomy may clear the vitreous of its VEGF load, one caveat is that intravitreal medications may also be cleared more quickly once the vitreous is removed. For example, intravitreal triamcinolone acetonide is cleared 1.5 times faster in vitrectomized eyes than nonvitrectomized eyes.10

In our practice, these medications have become the most frequently used treatment for DME. Decreasing their halflife early on in the course of a patient’s treatment could necessitate more frequent injections to maintain VEGF blockade or require medication via sustained-release devices.


When evaluating an eye as a potential candidate for surgical intervention, one should classify DME as not only as focal or diffuse, but also by the status of the vitreous and the vitreomacular interface. Evaluation of the vitreomacular interface is best accomplished by funduscopic examination and OCT imaging.

Surgical treatment for DME is usually reserved for mechanical causes of the edema. Surgery includes vitrectomy with detachment of the posterior hyaloid, if present, and peeling of any associated epiretinal membranes, with or without internal limiting membrane removal.

Table 1. Surgical Classification of DME

DME with abnormal vitreomacular adhesions and/or proliferation

a. Taut hyaloid
b. Vitreomacular and/or vitreofoveal traction
c. Epiretinal membrane (avascular and/or vascular)

Postvitrectomy taut ILM syndrome
Recalcitrant DME without abnormal vitreomacular adhesions

We have divided cases of DME into three major categories with five subcategories or entities (Table 1). The first category we have termed DME with abnormal vitreomacular adhesions or proliferation, or both. This category is separated into three different entities, based on differences in clinical appearance and OCT findings.

The second category is an entity referred to as postvitrectomy taut ILM syndrome, which causes new-onset DME following diabetic vitrectomy related to tangential traction along the ILM.

The final category is recalcitrant DME without any apparent vitreomacular interface abnormalities. All five entities have different biomicroscopic and OCT findings (Table 2).

Table 2. DME Surgical Categories/Entities: Biomicroscopic and OCT findings
DME Classification Biomicroscopy OCT
DME with abnormal vitreomacular adhesions and/or proliferation
             Taut posterior hyaloid Posterior hyaloid mostly attached with glistening reflex Striae in posterior hyaloid (not in the retina) can be present Low-lying and uniform partial posterior vitreous detachment. Underlying diffuse retinal edema and thickening
             Vitreomacular and/or vitreofoveal traction Posterior hyaloid partially detached with irregular contour visible at the sites of vitreoretinal adhesions Variable and uneven partial posterior vitreous detachment Retinal edema greatest below vitreoretinal adhesions Vitreofoveal traction associated with greater focal foveal distortion
             Epiretinal membrane (vascular or avascular) Posterior hyaloid detached or attached Retinal striae/vessels distortion present Vascular: whitish, opaque with varying degrees of vascularity Hyper-reflective membrane overlying the macular surface with variable degrees of membrane and/or posterior vitreous detachment
Postvitrectomy taut ILM syndrome Hyaloid previously removed Hyper-reflective (less dense than epiretinal membrane) to normal reflective ILM without vitreoretinal adhesions. Diffuse retinal edema
Recalcitrant DME without abnormal vitreomacular adhesion Posterior hyaloid detached or attached Cystic or diffuse macular edema without vitreoretinal adhesions


Figure 1. Fundus photo (top) of a patient with chronic DME unresponsive to intravitreal bevacizumab and triamcinolone. Visual acuity was 20/200. The posterior hyaloid was thickened and taut with striae. The corresponding OCT (inset) revealed a low-lying partial hyaloid detachment with multiple vitreomacular adhesions over the fovea with underlying retinal edema. Postoperative OCT (bottom) two months after vitrectomy with hyaloid removal demonstrated marked improvement in the DME and restoration of the foveal contour with some IS/OS disruption centrally. Vision improved to 20/80.


Abnormal vitreomacular adhesions or proliferation, or both, start with various manifestations of PVD. The earliest stage of PVD is a separation of the perifoveal hyaloid from the ILM, with persistent adherence to the fovea itself. In time, the hyaloid separates from the entire macula, then toward the peripheral retina, and finally from the optic nerve, creating a Weiss ring.11,12

If the foveal or macular separation of the posterior hyaloid does not occur, persistent traction on the area of adhesion can, in a diabetic eye, result in macular edema. This process can also induce glial cell proliferation with subsequent contraction.13

We have divided DME with abnormal vitreomacular adhesions or proliferation, or both, into three distinct entities based on both the clinical and OCT findings. These entities include: 1) taut hyaloid; 2) vitreomacular or vitreofoveal traction, or both; and 3) epiretinal membrane (avascular and vascular).

Even though we separate them into three entities, they are, in fact, a spectrum of disease. Many eyes have features of all three. This overlap has made the classification of these disease processes confusing in the literature. In addition, the original description of these entities predates the development of OCT. Because of this, there has not yet been adequate correlation between the ophthalmoscopic and OCT findings of these subtypes.

Taut Hyaloid

Lewis et al. were the first authors to use the term “taut hyaloid” to describe a subset of eyes with DME related to the posterior hyaloid. These eyes were observed to have a thickened and taut premacular posterior hyaloid.14 OCT was not available at the time of the original publication, and the funduscopic features included a glistening sheen and striae to the posterior hyaloid (Figure 1). The striae were limited to the hyaloid and did not involve the retina or associated retinal vessels, differentiating it from eyes with ERMs.

Lewis et al. reported that vitrectomy with removal of the posterior hyaloid was associated with complete resolution of DME in 80% of eyes and improvement in macular edema and vision in nine out of 10 patients.


Figure 2. Fundus photo (top) of a patient with DME and clinically visible partial posterior hyaloid detachment with broad based macular traction. Visual acuity was 20/150 with metamorphopsia. OCT (inset) revealed a partial hyaloid detachment with asymmetric hyaloid separation and multiple vitreomacular adhesions around the fovea with underlying retinal edema. Postoperative OCT (bottom) one week after vitrectomy with hyaloid removal. The ILM was also peeled. Vision improved to 20/80, and the metamorphopsia improved.

A larger study by Pendergast et al., with 55 eyes, had similar anatomic outcomes, with DME completely resolving in 81% of eyes after a mean follow-up of 4.5 months. Mean visual acuity in this cohort improved from 20/160 to 20/80. Better preoperative acuity was associated with better postoperative acuity, and the most common reason for loss of vision during follow-up was progressive cataract in 63% (24/38) of the phakic eyes.15

Harbour et al. studied a similar group of eyes with results that were not as favorable. They had visual acuity improvement in four out of 10 patients following vitrectomy for DME with a taut hyaloid. In this series, shorter duration of edema was associated with better visual outcomes, and macular ischemia, atrophy, and postoperative ERMs were associated with poorer outcomes.16

All of these three studies were published before the standard use of OCT. Using an early generation OCT (Humphrey Zeiss, Inc., San Leandro, CA), Kaiser et al. found that eyes with a clinical diagnosis of a taut hyaloid had shallow elevation of a partially detached hyaloid with vitreomacular adhesions.


Figure 3. Fundus photos and OCTs of a monocular patient with a history of PDR and panretinal photocoagulation. Visual acuity at presentation was 20/80 due to vitreofoveal traction and DME. Fundus photo (above) and OCT (inset) revealed a partial hyaloid detachment with a focal vitreofoveal adhesion and underlying retinal edema. The patient elected to be observed. Three years later, visual acuity decreased to 20/400, and OCT revealed persistence of the focal vitreofoveal adhesion with thickening of the detached hyaloid and an increase in retinal edema (top right). Epiretinal proliferation was also noted. During pars plana vitrectomy, the membranes were found to be extremely adherent, and membrane peeling was limited to the macular area. Postoperative fundus photo and OCT (right) three months following surgery revealed resolution of the DME and associated vitreofoveal traction with restoration of the foveal contour and some S/OS disruption centrally. Vision improved to 20/100.

Even though they described the majority of eyes as having a tractional macular detachment, the majority of eyes in their study had partial-thickness retinal elevations. The mean retinal thickness in their cases was approximately 550 μm.17,18

An additional study by Massin found that six eyes of seven patients with refractory DME and a taut hyaloid had vitreomacular traction on OCT. Furthermore, OCT detected traction in at least one patient in whom clinical examination showed only a taut hyaloid.19

Additional studies of OCT findings in patients with taut hyaloids could help shed light on the vitreoretinal interface in these patients. This clarification would be helpful in determining how to apply the older literature on taut hyaloids in patients in currentday practice.

Vitreomacular and/or Vitreofoveal Traction

The terms vitreomacular traction and vitreofoveal traction have been used to describe a clinical entity in which the posterior hyaloid detaches from the posterior retina but remains attached focally to the macula and or fovea. This can be a cause of DME in eyes with diabetic retinopathy.

The etiology is similar to taut hyaloid, and many retina specialists lump these three entities into the same group. The terms “vitreomacular traction” and “vitreofoveal traction” are occasionally mistaken by early trainees for “vitreomacular traction syndrome.” Actually, vitreomacular traction syndrome is a different entity. It was initially described with the help of B-scan ultrasonography in eyes without diabetic retinopathy, which had the hyaloid detached peripherally but still adherent to the macula.

Vitreomacular traction has similar, but less dramatic, vitreoretinal interface changes, and it is commonly diagnosed using OCT. Clinically, vitreomacular traction and vitreofoveal traction produce greater distortion to the retinal surface and have a noticeable elevation to the posterior hyaloid, as opposed to a taut hyaloid, in which the elevation of the hyaloid is clinically difficult to discern from the surface of the retina.

Broad areas of multiple adhesions cause vitreomacular traction (VMT) (Figure 2), while a small area of adhesion at the fovea causes vitreofoveal traction (VFT) (Figure 3). Because the tractional forces are distributed over a smaller area, vitreofoveal adhesion often induces more distortion in the retinal contour.20

Even though OCT findings for all three entities reveal abnormal vitreoretinal adhesions, VMT and VFT tend to have greater separation of the posterior hyaloid from the surface of the retina when compared to eyes with the classic taut hyaloids. This greater hyaloid separation in VMT and VFT also results in a greater angle of incidence between the hyaloid and retinal surface.

In 2010, the Diabetic Retinopathy Clinical Research Network published the largest series of vitrectomy for DME with VMT. studied 87 eyes with DME felt to be related to VMT. These eyes underwent pars plana vitrectomy with ERM or ILM peeling performed at the surgeon’s discretion.

At six months, they found a reduction in OCT central subfield thickness (CST) of at least 50% in 68% of patients. The visual acuity results were variable, with 38% improving by ≥10 letters and 22% worsening by ≥10 letters. 21 Differentiating between a taut hyaloid and vitreomacula or vitreofoveal traction, or both, may not be clinically significant because each may represent different presentations of a similar phenomenon, only differentiated by the degree of separation of the hyaloid and angle of vitreoretinal traction.


Figure 4. Fundus photo (top) of a patient with partially opaque vascularized preretinal membrane with DME and proliferative retinopathy. The membrane was fibrotic and originated from regressed preretinal neovascularization. Visual acuity was 20/400. OCT (inset) revealed a thickened and partially detached hyper-reflective membrane. There was sub-retinal fluid and lamellar hole within the center of the fovea. Postoperative OCT (bottom) six months after vitrectomy with ERM peeling. The DME and subretinal fluid resolved with thinning of the retina due to atrophy. Vision improved to 20/150.

Each of these entities has varying degrees of vitreoretinal traction that contributes to the formation of DME, and many eyes have features of all three entities. Published studies have not rigorously distinguished among them.

In the future, we hope that standardized OCT interpretations of the vitreoretinal interface will become available and help determine which eyes benefit most from surgical treatment.

Epiretinal Membrane

In patients with or without diabetes, epiretinal membranes can cause macular edema. Diabetes has been associated with an increased risk of ERMs, independent of diabetic retinopathy.22,23

Posterior vitreous detachment is also important in the pathogenesis of ERMs.24 In the early stages of PVD, and especially in areas of vitreoschisis, hyalocytes retained on the retinal surface can proliferate and grow onto the surface of the attached hyaloid. This can strengthen the vitreoretinal adhesion and exacerbate vitreoretinal traction. 25,26

One can observe these ERMs on ophthalmoscopy. They can produce retinal striae with or without retinal folds. OCT shows a hyper-reflective membrane that may distort the normal contour of the inner retina.

The majority of these membranes are avascular. They can become thickened and vascularized if neovascularization grows within the membrane. The membranes form in response to the growth and regression of preretinal neovascularization into the posterior hyaloid. On ophthalmoscopy, they appear white. On OCT, they are much thicker than avascular membranes (Figure 4).

Few studies have specifically looked at the role of ERM peeling in eyes with DME. studied a group of 241 eyes with DME (87 with VMT, 154 without VMT) that underwent pars plana vitrectomy. When the subgroup of 87 patients with vitreomacular traction was excluded, 94 (61%) of the additional 154 had ERM peeling, which was associated with improved visual acuity.

However this group did not have a corresponding reduction in macular thickness on OCT. The authors suggested that the improvement in visual acuity might have resulted from improvement in visual distortion rather than DME. Not unique to this study, this disconnect appears to be due to the lack of correlation between visual acuity and macular thickness.


Postvitrectomy taut ILM is an uncommon cause of new-onset diffuse DME after diabetic vitrectomy with prior hyaloid removal.27 One can differentiate this entity from a taut posterior hyaloid and vitreomacular or vitreofoveal traction, or both, by the absence of an attached or partially attached hyaloid.

It occurs more than six weeks after vitrectomy with hyaloid removal and is unresponsive to laser, corticosteroids and anti-VEGF. Two forms have been reported: focal and diffuse. The former has ILM striae, and the latter a glistening sheen over the posterior pole.

Optical coherence tomography shows outer retinal edema with no evidence of definite traction (Figure 5). Fluorescein angiography reveals diffuse and deep fluorescein leakage.

The pathophysiology of postvitrectomy taut ILM is postulated to be from tangential traction along the ILM by contractile cellular elements. Because the ILM is a sheet of basement membrane without elasticity, it is capable of transmitting tangential traction onto the retina with distant effects.

Pathological specimens of two eyes with postvitrectomy taut ILM revealed a monolayer of cells of retinal pigment epithelium or glial cell origin, or both, on the ILM, which expressed smooth muscle actin.27

Treatment includes ILM peeling over the fovea to release tangential traction. Macular ILM peeling could be indicated in cases of persistent diffuse DME in which the ILM is taut. This mechanism may be responsible for some cases of persistent DME following vitrectomy.


Although vitrectomy may be useful for eyes with vitreomacular interface abnormalities, the majority of eyes with refractory DME have a normal interface. Thomas et al. used OCT to image 140 patients with DME unresponsive to laser. Partial vitreomacular separation was seen in 10%, and taut hyaloid was seen in 4%.27

The authors concluded the majority of eyes with DME unresponsive to laser did not have a taut hyaloid or significant abnormal vitreomacular adhesion.28

Many case series have been published on the surgical treatment of DME without vitreomacular interface abnormalities. In addition to removal of the posterior hyaloid, removal of the ILM has been advocated by some authors.

Most studies of vitrectomy, with or without ILM peeling, for DME without abnormal vitreomacular adhesion were undertaken before the era of anti-VEGF agents and were mainly reserved for eyes unresponsive to laser treatment. The rationale for surgical treatment was that it released tractionalmediated effects of the hyaloid and ILM and allowed for clearance of a permeability barrier. It was also felt to inhibit reproliferation and to improve transvitreal oxygenation.

One of the first studies that seemed to hold promise for this treatment was published by Gandorfer et al. in 2000.29 They found that removing the ILM improved visual acuity by at least 2 lines in 11 out of 12 patients with DME unresponsive to laser. Even though others have had similar results, most studies have yielded variable visual results not as good as the anatomical results.

A randomized trial reported by Thomas and colleagues found no benefit when comparing vitrectomy to additional laser in 40 patients with DME without VMT in terms of macular thickness or visual acuity.30 Other small randomized trials have also not been able to show a visual benefit of vitrectomy with ILM peeling over laser, despite improvement in macular thickening.31

Shah et al. performed a prospective study on vitrectomy for DME unresponsive to previous focal grid laser and found that the only factor predicting visual gain was the presence of macular traction on OCT.32 The lack of correlation between anatomic resolution of edema and improved vision remains a major outstanding question in understanding surgical treatments for DME.


Vitrectomy for the treatment of DME has been shown to benefit some eyes with mechanical causes of the edema. Unfortunately, we still have no fail-safe way of determining which eyes will benefit from vitrectomy and which will not.


Figure 5. Red-free photograph (top) of a patient with new-onset DME three years after pars plana vitrectomy and posterior hyaloid removal. There is a glistening sheen to the posterior pole and a lamellar macular hole. The DME was unresponsive to focal laser, intravitreal bevacizumab, and intravitreal triamcinolone. Visual acuity was 20/100. OCT (middle) revealed marked retinal edema with lamellar hole and a hyper-reflective ILM. Postoperative OCT (bottom) two weeks after ILM peeling revealed resolution of the DME and restoration of a normal foveal contour. Vision improved to 20/50.


In the absence of a mechanical cause for DME, vitrectomy results have been variable, with visual outcomes much less impressive than OCT results. The diagnosis of tractional causes of DME can most often be made clinically and confirmed by OCT findings. Clinical features consistent with the presence of a taut hyaloid, vitreomacular/vitreofoveal traction, and ERM support the role of vitrectomy with removal of the tractional elements.

It is not clear what role the addition of ILM peeling has in the treatment of these entities. Theoretically, ILM removal can ensure complete removal of the cortical vitreous and other tractional elements. This is especially true in the presence of vitreoschisis.

In contrast, ILM peeling has been shown to be the treatment of choice in a small subgroup of eyes with postvitrectomy taut ILM syndrome.27 Lack of response to laser, steroids, and anti-VEGF agents is a requirement for this diagnosis. Even though OCT does not clearly image a mechanical cause of this entity, its etiology is believed to be caused by tangential traction across the ILM by contractile cellular elements.

Because most of the studies on vitrectomy for DME were undertaken before the widespread use of anti-VEGF agents, it is not known whether anti-VEGF treatment can have any effect on the mechanical causes of DME. The authors suspect that both tractional (mechanical) and nontractional (biochemical) factors play roles in some eyes with DME.

In eyes in which retinal edema and thickening seem out of proportion to clinical and OCT findings of traction, a trial of multiple intravitreal pharmacological treatments should be considered prior to surgery. If patients do not respond to medical therapy, and there seems to be preoperative evidence suggesting a tractional component to DME, pars plana vitrectomy can be considered.

However, in addition to the surgical risks of vitrectomy, the increased clearance of anti-VEGF and other pharmacological agents in a vitrectomized eye should be considered. OCT findings of traction should always be correlated with clinical findings, because abnormal vitreomacular adhesions are not uncommon and may, in some cases, not be the primary cause of DME.

Grading and quantifying clinically significant vs insignificant abnormal vitreomacular adhesions on OCT have yet to be performed. We believe our classification scheme is a step toward a better understanding of the vitreoretinal interface in DME.

We hope that with more imaging studies and OCT surgical correlations, we will be better able to predict which eyes with DME are best treated with surgery and, in the future, even pharmacologic vitreolysis. RP


1. Sivaprasad S, Gupta B, Crosby-Nwaobi R, et al. Prevalence of diabetic retinopathy in various ethnic groups: A worldwide perspective. Surv Ophthalmol. 2012;57:347-370.

2. Ding J, Wong TY. Current epidemiology of diabetic retinopathy and diabetic macular edema. Curr Diab Rep. 2012;12:346-354.

3. Centers for Disease Control and Prevention. National Diabetes Fact Sheet: National Estimates and General Information on Diabetes and Prediabetes in the United States. Atlanta, GA: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention; 2011.

4. Ferris FL 3rd, Patz A. Macular edema. A complication of diabetic retinopathy. Surv Ophthalmol. 1984;28(Suppl):452-461.

5. Bhagat N, Grigorian RA, Tutela A, et al. Diabetic macular edema: Pathogenesis and treatment. Surv Ophthamol. 2009;54:1-32.

6. Joussen AM, Murata T, Tsujikawa A, et al. Leukocyte-mediated endothelial cell injury and death in the diabetic retina. Am J Pathol. 2001;158:147-152.

7. Hikichi T, Fujio N, Akiba J, et al. Association between the short-term natural history of diabetic macular edema and the vitreomacular relationship in type II diabetes mellitus. Ophthalmology. 1997;104:473-478.

8. Nasrallah FP, Jalkh AE, Van Coppenolle F, et al. The role of the vitreous in diabetic macular edema. Ophthalmology. 1988;95:1335-1339.

9. Stefansson E. The therapeutic effects of retinal laser treatment and vitrectomy. A theory based on oxygen and vascular physiology. Acta Ophthalmol Scand. 2001;79:435-440.

10. Chin HS, Park TS, Moon YS, et al. Difference in clearance of intravitreal triamcinolone acetonide between vitrectomized and nonvitrectomized eyes. Retina. 2005;25:556-560.

11. Johnson MW. Posterior vitreous detachment: Evolution and complications of its early stages. Am J Ophthamol. 2010;149:371, 382.e1.

12. Johnson MW. Posterior vitreous detachment: Evolution and role in macular disease. Retina. 2012;32(Suppl 2):S174-S178.

13. Gandorfer A, Rohleder M, Grosselfinger S, et al. Epiretinal pathology of diffuse diabetic macular edema associated with vitreomacular traction. Am J Ophthamol. 2005;139:638-652.

14. Lewis H, Abrams GW, Blumenkranz MS, et al. Vitrectomy for diabetic macular traction and edema associated with posterior hyaloidal traction. Ophthalmology. 1992;99:753-759.

15. Pendergast SD, Hassan TS, Williams GA, et al. Vitrectomy for diffuse diabetic macular edema associated with a taut premacular posterior hyaloid. Am J Ophthamol. 2000;130:178-186.

16. Harbour JW, Smiddy WE, Flynn HW Jr, et al. Vitrectomy for diabetic macular edema associated with a thickeΪned and taut posterior hyaloid membrane. Am J Ophthamol. 1996;121:405-413.

17. Kaiser PK, Riemann CD, Sears JE, et al. Macular traction detachment and diabetic macular edema associated with posterior hyaloidal traction. Am J Ophthamol. 2001;131:44-49.

18. Lewis H. The role of vitrectomy in the treatment of diabetic macular edema. Am J Ophthamol. 2001;131:123-125.

19. Massin P, Duguid G, Erginay A, et al. Optical coherence tomography for evaluating diabetic macular edema before and after vitrectomy. Am J Ophthamol. 2003;135:169-177.

20. Spaide RF, Wong D, Fisher Y, et al. Correlation of vitreous attachment and foveal deformation in early macular hole states. Am J Ophthamol. 2002;133:226-229.

21. Diabetic Retinopathy Clinical Research Network Writing Committee, Haller JA, Qin H, et al. Vitrectomy outcomes in eyes with diabetic macular edema and vitreomacular traction. Ophthalmology. 2010;117:1087, 1093.e3.

22. Mitchell P, Smith W, Chey T, et al. Prevalence and associations of epiretinal membranes. the blue mountains eye study, Australia. Ophthalmology. 1997;104:1033-1040.

23. Klein R, Klein BE, Wang Q, et al. The epidemiology of epiretinal membranes. Trans Am Ophthalmol Soc. 1994;92:403-425; discussion 425-430.

24. Kampik A. Pathology of epiretinal membrane, idiopathic macular hole, and vitreomacular traction syndrome. Retina. 2012;32(Suppl 2):S194-S199.

25. Itakura H, Kishi S. Aging changes of vitreomacular interface. Retina. 2011;31:1400-1404.

26. Spaide RF. Vitreomacular traction syndrome. Retina. 2012;32(Suppl 2):S187-S193.

27. Gentile RC, Milman T, Eliott D, et al. Taut internal limiting membrane causing diffuse diabetic macular edema after vitrectomy: Clinicopathological correlation. Ophthalmologica. 2011;226:64-70.

28. Thomas D, Bunce C, Moorman C, et al. Frequency and associations of a taut thickened posterior hyaloid, partial vitreomacular separation, and subretinal fluid in patients with diabetic macular edema. Retina. 2005;25:883-888.

29. Gandorfer A, Messmer EM, Ulbig MW, et al. Resolution of diabetic macular edema after surgical removal of the posterior hyaloid and the inner limiting membrane. Retina. 2000;20:126-133.

30. Thomas D, Bunce C, Moorman C, et al. A randomised controlled feasibility trial of vitrectomy versus laser for diabetic macular oedema. Br J Ophthamol. 2005;89:81-86.

31. Kumar A, Sinha S, Azad R, et al. Comparative evaluation of vitrectomy and dye-enhanced ILM peel with grid laser in diffuse diabetic macular edema. Graefes Arch Clin Exp Ophthamol. 2007;245:360-368.

32. Shah SP, Patel M, Thomas D, et al. Factors predicting outcome of vitrectomy for diabetic macular oedema: Results of a prospective study. Br J Ophthamol. 2006;90:33-36.