Vitrectomy is reported to be effective in resolving macular edema, but the data sets have been limited, and the published studies to date have, for the most part, not been well-controlled, prospective, randomized trials.1-4 An important question is: Why should vitrectomy work at all? One theory is that mechanical stress or stretch causes leakage. Another is that there is a physiologic improvement in oxygenation in the vitreous, and a third is that intravitreal VEGF levels are reduced.5,6

EARLY STUDIES
In the 1992 paper that first described the potential benefits of vitrectomy for macular edema, we reported on 10 patients who were failures of prior photocoagulation.1 We postulated that an abnormally thickened hyaloid was exerting traction on the vitreoretinal interface and causing leakage. Some of these eyes had epiretinal membranes, and approximately 60% were judged to be successes.

Several years later, I reviewed my results on a second cohort of 15 patients.7 Before surgery, none of these patients had visual acuity better than 20/80, but after surgery, about two-thirds of them had better than 20/80 vision. In fact, 20% had better than 20/50, suggesting that vitrectomy was helping these patients.

Interestingly, these findings did presage some later understandings. With monochromatic blue-light photography, one could see 2 x 2 disc diameter shiny, circular zones in the macula that corresponded to leakage in the later fluorescein angiogram and which seemed distinct from a macular pucker (Figure 1). At that time, we described it without postulating what it was. Now, we recognize it is the contracted internal limiting membrane (ILM). Subsequently, we performed en face histologic examination after vitrectomy on peeled membranes after placing them on a glass slide rather than embedding them in paraffin. The impact was to examine them as we would with an ophthalmoscope as opposed to conventional histology, which is analogous to optical coherence tomography (OCT). We could see that these very modestly hypocellular membranes were composed primarily of ILM of the retina (Figure 2).7

Another interesting finding was that many membranes contained very small vessels. These were, in effect, vascularized epiretinal membranes that were causing localized traction. Now, we suspect that the ILM in DME is probably important. Some vitreoretinal surgeons globally consider vitrectomy a primary form of therapy that may precede focal or grid photocoagulation in selected settings. This view, however, is not a standard of care in the United States and is considered controversial.

PROSPECTIVE COHORT STUDY
Data should drive best clinical practice, and the Diabetic Retinopathy Clinical Research Network (DRCR.net) has pioneered a host of studies on treatment of diabetic retinopathy. One is the Evaluation of Vitrectomy for Diabetic Macular Edema Study.5 This study provides an opportunity to collect data prospectively, using standardized protocols to assess benefits and risks. The results can be used to stimulate the design and execution of a prospective, randomized clinical trial with Level 1 evidence, and it also can be used to help design other studies.

A series of patients, aged ≥18 years old, were enrolled in one such study. Figure 3 summarizes eligibility criteria. Surgery was performed according to the investigator's usual routine. About half of the patients had small-gauge vitrectomy, and the other half had standard 20-gauge surgery. About two-thirds of the patients had epiretinal membranes removed. Clearly, the surgeons felt some degree of traction was present, although not in every patient. There are those who argue that performing a vitrectomy alone, with or without ILM peeling, still helps DME.

This prospective, nonrandomized but controlled study showed no significant change in mean visual acuity from baseline to 3 months or to 6 months. Although some patients improved, and more patients improved at 6 months than at 3 months, some became worse. So the aggregate effect was that there was no improvement in the mean visual acuity overall. It is unknown how these eyes would have fared if they had not received vitrectomy, so the proportion with improvement may be better than no vitrectomy, and the proportion with loss may be better than no vitrectomy. At the least, this information provides some guidance regarding expectations for patients considering this surgery for these circumstances.

There was a benefit in the OCT-measured macular thickness. There was a considerable decrease in thickness from about 500 μm to just under 300 μm. All factors being equal, we would have assumed that if the mean macular thickness decreased, the mean visual acuity would improve, yet it did not. Cataract could have been a factor, or perhaps these maculas did not have reasonable potential for visual improvement because of prior damage from diabetes.

When an epiretinal membrane is present, we know that time-domain OCT in the autocalculation mode shows the retina to be thicker than it might really be. So the OCT central subfield thicknesses may not reflect as much reduction in retina thickness; much of the reduced thickness may have been from removing the epiretinal membranes.

WEIGHING BENEFITS/RISKS
Whenever we judge the efficacy of a therapy, we must take into account the positive benefits and subtract the complications. The net of those two determines whether or not treatment is beneficial. Many treatments are beneficial but have unacceptably high risk, while others are not especially beneficial but they have minimal risk. The treating physician and the patient must decide how to achieve the best balance for the patient.

These data show the surgery was fairly successful in the sense that it was not likely to lead to serious adverse events, but we also know that most patients older than 40 years will develop a cataract after vitrectomy, and a small percentage will experience a detachment or other complications.6

What these data show is that we do not know everything we need to know to make an informed judgment about whether vitrectomy is or is not safe and effective for DME. They do suggest, however, that vitrectomy reduces macular thickness. So a conclusive recommendation remains an open issue at this time, and there should be additional studies looking at this. My hope is that within several years we will have more information to guide us. ■

1. Lewis H, Abrams GW, Blumenkranz MS, Campo RV. Vitrectomy for diabetic macular traction and edema associated with posterior hyaloidal traction. Ophthalmology. 1992;99:753-759.
2. Harbour JW, Smiddy WE, Flynn HW Jr, Rubsamen PE. Vitrectomy for diabetic macular edema associated with a thickened and taut posterior hyaloid membrane. Am J Ophthalmol. 1996;121:405-413.
3. Tachi N, Ogino N. Vitrectomy for diffuse macular edema in cases of diabetic retinopathy. Am J Ophthalmol. 1996;122:258-260.
4. Pendergast SD. Vitrectomy for diabetic macular edema associated with a taut premacular posterior hyaloid. Curr Opin Ophthalmol. 1998;9:71-75.
5. Stefánsson E. Ocular oxygenation and the treatment of diabetic retinopathy. Surv Ophthalmol. 2006; 51:364-380.
6. Funatsu H, Yamashita H, Ikeda T, Mimura T, Eguchi S, Hori S. Vitreous levels of interleukin-6 and vascular endothelial growth factor are related to diabetic macular edema. Ophthalmology. 2003;110:1690-1696.
7. Blumenkranz MS. Paper presented at: Retinal Subspecialty Day; October 1991; Jules Stein Eye Institute, Los Angeles, CA.