Challenges in Diabetic Vitrectomy

Learn the latest about four issues that affect your outcomes.

Challenges in Diabetic Vitrectomy
Learn the latest about four issues that affect your outcomes.
By Dean Eliott, MD

The goals of diabetic vitrectomy are to relieve anteroposterior and tangential traction, while avoiding or minimizing iatrogenic retinal breaks and maintaining hemostasis. To achieve this, we use a variety of techniques, most notably combining segmentation and delamination using a tissue manipulator and the MPC scissors.

Despite using the latest surgical advances, we continue to face challenges in diabetic vitrectomy, including: The timing of intervention for premacular subhyaloid hemorrhage; the indications for retinectomy; the risk of rubeosis after lensectomy; and the role of pharmacological intervention to inhibit postoperative fibrous proliferation.

Premacular subhyaloid hemorrhage with tightly confined blood and opacified posterior hyaloid.


Premacular subhyaloid hemorrhage is blood tightly confined between the macula and the posterior vitreous face. The posterior hyaloid is often opaque, indicating that cellular proliferation has occurred. In these cases, the hyaloid is attached at the disc and from the arcades to the periphery, and there is a low-lying vitreomacular separation.

My colleagues and I investigated whether the timing of vitrectomy affects outcomes, evaluating 400 primary vitrectomies to find 40 cases where the blood was tightly confined (10% of diabetic vitrectomies in our practice). We found that when vitrectomy was performed within 1 month of the onset of hemorrhage, preoperative vision was count fingers (20/400 to hand motion) and postoperative vision was relatively good at 20/50. But in cases with comparable preoperative vision where we waited longer than 1 month, postoperative vision was 20/100.

The group undergoing early vitrectomy had a more favorable outcome than the group undergoing later vitrectomy. Poor prognostic factors included macular edema, foveal traction detachment, fibrosis and macular ischemia. We believe the group that underwent earlier surgery did better because removal of hemorrhage enabled timely treatment of macular edema. In addition, foveal traction was relieved with surgery, and fibrosis was removed.

Regarding macular ischemia, there is no effective treatment, and it was associated with a poor outcome. However, given the other three factors -- macular edema, foveal traction and fibrosis -- we obtained a more favorable outcome when we intervened earlier.

Arrows show small posterior focal retinectomy. 


The most common type of retinectomy is a small posterior focal retinectomy, which is usually performed to remove fibrovascular plaques with adjacent preexisting or iatrogenic retinal breaks. We also occasionally perform small posterior focal retinectomies to gain access to subretinal membranes. In contrast, a large peripheral retinectomy is required when there is unrelieved peripheral traction despite membrane peeling and scleral buckle placement.

In our series, we performed retinectomy in approximately 5% of primary diabetic vitrectomies, and we found no relationship to outcome. When membrane reproliferation caused redetachment requiring reoperation, retinectomy was performed in 30%. In those cases, the membranes were tightly adherent and the retina was atrophic, making it very difficult to remove the fibrous membranes without creating retinal breaks. Patients who required retinectomies during reoperations did poorly compared to those who did not require retinectomy.


Has the advent of endolaser therapy altered the risk of rubeosis after lensectomy?

My colleagues and I perform lensectomy to improve intraoperative visualization or to gain access to peripheral fibrovascular plaques. We found that eyes that underwent lensectomy had less postoperative rubeosis than eyes that remained phakic. Eyes that underwent lensectomy typically had more advanced disease; however, they were able to receive more thorough treatment and actually had a lower incidence of postoperative rubeosis.

Arrows show posterior edge of large peripheral retinectomy.

These results can be explained since lensectomy enables more complete removal of anterior vitreous and membranes. After lensectomy, we were able to perform careful scleral depression, dissect the vitreous base, and use endolaser and/or the laser-indirect ophthalmoscope to perform thorough photocoagulation from the arcades to the ora serrata.


Currently, my colleagues and I are conducting a randomized study using 5-fluorouracil (5-FU) and low molecular weight heparin (LMWH) in diabetic patients who are at high risk of postoperative fibrous membrane reproliferation.

The 5-FU inhibits cellular proliferation and collagen production, while the LMWH inhibits growth factors and extracellular matrix contraction. The endpoints of our study are visual acuity and reoperation rate.

With the information we gain from this study, as well as other new advances, we hope to achieve better outcomes for our patients who undergo diabetic vitrectomy.

Dr. Eliott is an associate professor of ophthalmology at the Kresge Eye Institute at Wayne State University in Detroit.