Surgical management of tractional retinal detachment (TRD) in young diabetic patients can be a challenge for even the best vitreoretinal surgeons, with moderate success rates and a high degree of surgical difficulty. The eyes of these patients often have thin, ischemic retinas, and complications can include intraoperative bleeding, iatrogenic retinal breaks, postoperative bleeding, and retinal redetachment.

Historically, young diabetic patients seen in retinal practices have type 1 diabetes— what used to be called juvenile-onset diabetes. Increasingly, however, with the epidemic of obesity in the developed world,1 more and more of the young patients we see with diabetes have type 2. If we define young patients as those less than 40 years of age, retina practice has moved in recent years much more toward type 2 diabetes in young patients.

In general, the greater the success of local screening and community support programs for diabetics, the fewer patients with type 1 diabetes are seen in retina practices. If a young type 1 patient presents to us, however, it is likely he or she has other diabetes-related complications such as renal or peripheral neuropathy. These are patients whose diabetes is not under good control.

Often the neovascularization in these young people is active, and if a TRD is present it is likely to be part of a progressive deterioration in their retina. Therefore, there is a need to intervene quickly. In addition, during surgery the eyes of young type 1 diabetic patients tend to have more hyperdynamic circulation than type 2 diabetic patients. They often have “hot” eyes, going fairly rapidly out of control, and it is our task to get on top of the situation. For all of these reasons, surgery in young diabetic patients is challenging (Figure 1).

Recently, intravitreal injection of the vascular endothelial growth factor (VEGF) inhibitor bevacizumab (Avastin, Genentech) has been employed prior to surgery for TRD in diabetic eyes with the hope of quieting the eyes and reducing intraoperative bleeding and other complications. Several authors have reported that pretreatement with bevacizumab reduces neovascularization, is well tolerated, and facilitates vitrectomy in diabetic eyes.2-4

In diabetic eyes with TRD we now routinely administer intravitreal bevacizumab the week before surgery. This makes the operation much easier; the membranes are much crisper and less vascular during surgery than they would have been without the anti-VEGF agent, and if a vessel is cut there is less hemorrhage and fewer attendant problems. There also seems to be a reduction in bleeding in the postoperative period. Whether that reduction is related to the less traumatic course of the surgery or to the continued effects of the anti-VEGF agent is not clear.

One of the difficulties in management of these cases is that, when panretinal photocoagulation (PRP) is applied at the time of surgery, it takes roughly 3 to 4 weeks for its effect to kick in. The preoperative anti- VEGF agent quiets the eye down more quickly and buys some time for the PRP to start to work (Figure 2).

Surgery for TRD begins with removal of the core of the vitreous. The taut peripheral gel attached to the TRD should be retained, as the traction it produces elevates the membranes off the retina, facilitating dissection. One can create a hole in the peripheral cortical gel to allow access of instruments, or start in the macular area and work outward.5

A key maneuver is the separation of the posterior hyaloid membrane (PHM) from the surface of the retina (internal limiting membrane, ILM). The cortical vitreous is often split into an outer layer lying on the retina (which includes the PHM) and an inner layer on the vitreous (so-called vitreoschisis). The outer vitreous cortex and PHM must be found and elevated, creating a cleavage plane between the PHM and the ILM. One should search for the outer layer on the slope of the TRD. In contrast to other eyes that we operate on, such as those with macular holes, in severe TRD disease the PHM often thins out and disappears near the equator. If this layer is not found, the surgeon will be cutting through the cortical gel and will leave membrane on the retina, allowing postoperative reproliferation.

Once this separation is achieved, the membranes can be elevated and dissected. Points of adhesion that cannot be lifted must be cut. Cutting pegs of adhesion with the scissors results in less bleeding than pulling the membranes off. Visualization should be kept to the maximum, and any elevated membrane should be trimmed with the vitreous cutter. Any anteroposterior traction should be maintained for as long as possible until the membranes have been removed. Once the dissection is complete, vitrectomy can be completed and PRP performed.

Some surgeons take aggressive steps at the time of surgery to try to prevent rebleeding, such as peripheral cryotherapy at the sclerotomy sites on the anterior retina or laser application right out to the ora serrata. There is a belief among some that neovascularization at the vitreous base is responsible for rebleeds, and these measures are attempts to address these areas prophylactically. I am not convinced this is the case, and I do not perform peripheral cryotherapy.

For a tamponade in these cases I prefer a long-acting gas such as perfluoropropane (C3F8) for its support of retinal breaks and because of potential oxygenating effects on the retina. Silicone oil use should be avoided in primary surgery for TRD, as it seems to adversely affect outcomes, stimulating membrane formation and formation of spontaneous breaks. If holes are made in the retina—and they frequently are in this type of surgery— subretinal oil can result postoperatively. Functional success rates with silicone oil are low, and the capacity for silicone oil removal is restricted. Silicone oil is impervious to oxygen, thereby exacerbating ischemia, and oil also increases VEGF concentration in the thin aqueous layer near the retina.

If there is recurrent detachment and a secondary surgery is needed 3 to 4 weeks after the primary surgery, silicone oil may be used at that time when the PRP has had a chance to reduce oxygen consumption by the retina and reduce VEGF production.

In the past 2 to 3 years we have moved increasingly to the use of 23-gauge instrumentation for this type of complex diabetic surgery. The instrumentation for sutureless surgery is much better now than it was when originally introduced. The scissors, although less angled with small-gauge systems, are very good. The 23-gauge high speed cutters are in many ways better than the 20-gauge cutters: nicely designed technology that is well suited for use in TRD. Modern cutters have an orifice near the end of the shaft allowing direct dissection or shaving of membranes with the cutter.

I tend not to use chandelier illumination on every case. I will use the chandelier if there is a combined TRD and rhegmatogenous RD, with the need to peel membranes from mobile retina. I want two hands for that, one holding forceps and the other holding 45° or 30° curved scissors. For the most part I perform onehanded surgery (the other holding the light pipe), but I believe that as more young surgeons are trained in bimanual surgery it will become more common.

The above comments pertain to young patients with type 1 diabetes in particular. In young patients with type 2 diabetes, TRDs tend not to be so aggressive or so active. Indeed, in some older patients with type 2 diabetes, surgery may not be needed because the TRD may be past its active stage (in a way analogous to diabetic nephropathy that “burns out”). The differences in these two types of patients should be kept in mind when considering surgery.

In type 1 patients, postoperative retinal detachment after surgery for TRD can be a serious complication. Our standard follow-up schedule for vitrectomy is to see patients at postoperative day 1, then at 2 and 6 weeks. We are beginning to think that this is not the right schedule for diabetic patients with complex TRD. If you miss a redetachment or another problem coming on, with that schedule you miss the short window you have to address the problem before the eye becomes much more difficult to operate on, or indeed inoperable. For difficult cases like this, I now want to examine the patient as often as weekly until I am sure the retina is safe, so that if a complication appears I can do something about it quickly.

This surgery has many potential hazards but is rewarding if done well. Although these patients can expect a 60% chance of improvement in vision, it is important to manage patients' expectations because you may be trying to save their vision rather than restore it.

Tom H. Williamson, MD, MBChB, FRCS, FRCOphth, is a Senior Consultant Ophthalmologist at St. Thomas' Hospital, London, Queen Mary's Hospital, Sidcup, Kent, and an Honorary Clinical Lecturer at King's College, London. Dr. Williamson states that he has no financial relationships to disclose. He can be reached at +1 44 207 188 4320; or via e-mail at