Complications of Diabetic Vitreoretinal Surgery
Things to look out for and tips for prevention
STEVE CHARLES, MD
Vitreoretinal surgery has made remarkable progress in the almost four decades since the first patient was operated on in 1970. Outcomes have markedly improved; many diseases — such as diabetic traction retinal detachments, vitreomacular traction syndrome, and dense vitreous hemorrhage — that were formerly untreatable are managed with high success rates. Vitreoretinal surgery utilized in combination with anti-vascular endothelial growth factor (VEGF) therapy and laser photocoagulation has significantly improved outcomes in the management of diabetic retinopathy. Although tremendous progress has been made in the vitreoretinal technique and technology arena, complications are ever present — some technique-driven, others a consequence of biology.
CONJUNCTIVA
Conjunctival damage from using various instruments to stabilize the eye during 23-gauge trocar insertion (Figure 1) is preventable by using a constant relative orientation insertion technique, changing the insertion path as the eye compresses orbital fat instead of stabilizing the eye. Poor closure technique with overlapping margins and irritating Vicryl suture material remain problems when using largely outdated sutured surgery. The author utilizes 25-g, sutureless transconjunctival surgery for all cases.
Figure 1. Conjunctival damage can arise from using various instruments to stabilize the eye during 23-gauge trocar insertion. Inadvertent suprachoroidal infusion can result from using initial 5° trajectory; 15° single plane approach is preferred.
CORNEA
Patients with diabetes have decreased corneal sensitivity and are more prone to corneal epithelial defects during and after surgery. The author never removes the epithelium in diabetic patients. Removal of the corneal epithelium is required in less than 1% of the author's cases. Corneal epithelial defects are largely technique-driven; Gelfoam, excessive use of topical anesthetics, inappropriate fundus/contact lens coupling fluids, excessive exposure to povidone-iodine, and inattention to corneal irrigation result in iatrogenic damage to the corneal epithelium.
Corneal endothelial failure is also largely preventable; cell loss is caused by poor quality infusion fluid (lactated Ringers, BSS instead of BSS Plus, etc.), prolonged operating times with excessive fluid throughout (more frequent with 20-g surgery), and inappropriate use of various additives (dextrose, bicarbonate, epinephrine, antibiotics). It is unnecessary and potentially dangerous to add dextrose to the infusion fluid during vitrectomy on diabetic patients. This is because patients are essentially euglycemic during surgery with the availability of real-time serum glucose monitoring. As vitreoretinal surgery is rarely elective, patients with low cell counts must still be operated on; careful attention to technique and avoiding anterior chamber intervention will minimize further cell loss.
LENS AND IRIS
Cataract (Figure 2) is not an inevitable consequence of vitrectomy. Patients with clear lenses preoperatively will usually retain clear lenses if the optimal infusion fluid without additive is used (BSS Plus), fluid throughput is minimized, and intensive patient education is used to prevent gas bubble contact with the lens. Progression of preexisting nuclear sclerosis is thought to be due to depletion of ascorbate associated with hyaluronan gel and vitreous collagen fiber matrix. Ascorbate absorbs oxygen; the partial pressure of oxygen permanently increases by 7 mm to 12 mm Hg after vitrectomy (Holekamp). Development of posterior capsular cataract is related to the factors described above, long-term use of intravitreal steroids (not intraoperative), diabetes management, and disease-based factors.
Figure 2. Cataract is a frequent, but not an inevitable, consequence of vitrectomy.
Iris damage from iris retractors, sphincterotomies, iris sutures, and phaco probe contact result in cosmetic problems and patient complaints but rarely have functional implications.
GLAUCOMA
Glaucoma remains the most common and significant anterior-segment complication of posterior vitrectomy. Steroid glaucoma was once thought to affect only 6% of the population; this notion was based on studies utilizing 1% topical prednisolone. In-office use of intravitreal triamcinolone acetonide produces a roughly 30% incidence of glaucoma, and the fluocinolone acetonide intravitreal implant Retisert (Bausch+Lomb, Rochester, NY) results in a ~90% incidence, with over 30% of patients requiring a glaucoma filtration procedure.
Many surgeons, including the author, used subconjunctival triamcinolone in all patients undergoing vitrectomy who were not thought to be steroid responders. The author abandoned this practice in recent years and now uses subconjunctival dexamethasone because of the shorter duration of action.
Prior to the use of intravitreal and subconjunctival triamcinolone, it was thought that steroid glaucoma was reversible when the drug was gone, but this has not proved to be the case. Neovascular glaucoma is VEGF-driven and occurs after vitrectomy, mostly in patients with diabetic retinopathy or retinal vein occlusions. Vitreous, lens, IOL and trabecular meshwork are all diffusion barriers for VEGF, as well as substrates for neovascularization.
Posterior vitrectomy in the presence of ischemic retina results in anterior vitreous cortex fibrovascular proliferation, unless the eye is aphakic, and in trabecular meshwork and iris neovascularization because of anterior diffusion of VEGF. Panretinal photocoagulation results in permanent decrease in VEGF production and is very effective in combination with anti-VEGF agents (eg, bevacizumab).
Retinal detachment in the context of diabetic retinopathy increases VEGF production; retinal reattachment is effective in reducing VEGF release from the retina. The author has coined the term "particulate glaucoma" for increased IOP secondary to hemolyzed blood, erythroclasts (ghost cells), lens particles and associated inflammatory cells, micelles due to silicone oil emulsification, inflammatory cells, pigment due to pigment dispersion syndrome, and liquid perfluorocarbon droplets. In general, the treatment of particulate glaucoma is to remove the particulate materials and treat the inflammation.
Intraocular lens dislocation (Figure 3) and decentration can result from gas bubble-pupillary block, iris retractors and iris manipulation. For unknown reasons, low intra ocular pressure causes severe miosis; careful attention to IOP control reduces the need for iris retractors. The Alcon Constellation Vision System with pressurized infusion and IOP compensation can reduce the incidence of intraoperative low IOP and secondary miosis. Synechia between the IOL, capsule, and iris can result from gas-pupillary block combined with inflammation.
Figure 3. Intraocular lens dislocation can result from gas bubble–pupillary block, iris retractors and iris manipulation.
REFRACTIVE
Refractive changes from encircling buckles remain a significant issue in vitreoretinal surgery. Encircling buckles are, in the author's opinion, significantly overutilized in conjunction with vitrectomy. Encircling buckles result in approximately 3 D of axial myopia, which is unacceptable in the broader context of emmetropic cataract surgery, LASIK, refractive lens exchange, toric IOLs, limbal relaxing incisions, etc. In addition, encircling buckles cause strabismus (50% of patients have increased tropias and phorias, Michels et al). Encircling buckles also cause ptosis because of damage to the levator aponeurosis and Mueller's muscle, as well as longer operating time and substantially more pain.
Silicone oil results in refractive change because the index of refraction of silicone oil differs from aqueous and vitreous. Preoperative determination of whether the oil will ultimately be removed or retained behind an IOL can help determine required IOL power. It is a misconception that oil impairs vision; a normal retina will produce perfect vision though oil if appropriate refraction is in place. It is also a mistake to think silicone oil is toxic and must be removed; emulsification glaucoma and corneal endothelial damage occur in about 3% of cases and virtually never if high-quality silicone oil (Alcon 1,000 cSt) is contained behind an IOL. The emulsification and complication rates of 1,000-cSt oil are identical to 5,000-cSt oil; emulsification is related to interfacial surface tension, which is not identical to viscosity.
OTHER COMPLICATIONS
Rebleeding after diabetic vitrectomy is very common in spite of careful hemostasis using endophotocoagulation (preferred to diathermy) and leaving the eye normotensive at the end of the case. These patients should be treated with anti-VEGF agents and followed with ultrasound to make certain a retinal detachment is not hidden by the hemorrhage.
Peripheral retinal breaks occur after approximately 0.5% to 1.0 % of vitrectomies performed for any diagnosis; diabetic hemorrhage and vitreomacular traction syndrome are no exception. Iatrogenic retinal breaks occur when operating for diabetic traction retinal detachments, especially if peeling instead of cutter or scissors delamination techniques are used.
The author was the first to note that glial tissue can recur after removal in diabetic tractional retinal detachment cases. The incidence of recurrence is higher when segmentation is used instead of delamination.
Fibrovascular ingrowth, in the author's opinion, never occurs; in reality, the choroid, retinal pigment epithelium and sclera undergo wound healing. Vessels in the wound exposed to high VEGF levels are a common cause of bleeding, emphasizing the need to inject an anti-VEGF agent at the end of most diabetic vitrectomies.
SUMMARY
Many complications of diabetic vitreoretinal surgery are preventable by performing fast, efficient surgery and paying strict attention to technical details. Others, such as progression of pre-existing nuclear sclerosis, are physiologic and cannot be prevented but must be taken into account. RP
Steve Charles, MD, is clinical professor of ophthalmology at the University of Tennessee College of Medicine in Memphis. Dr. Charles reports significant financial interest in Alcon. He can be reached via e-mail at scharles@att.net. |