Complications of Vitreoretinal Surgery

STEVE CHARLES, MD

Vitreoretinal surgery techniques and technology have dramatically evolved since the first pars plana vitrectomy was performed by Robert Machemer in 1970. As the safety and efficacy of vitreoretinal surgery has improved, these techniques have been utilized, not only for patients with poor preoperative visual acuity, but also for cases of patients with good preoperative vision, and not just blind eyes. In addition, surgeons as well as patients have come to have much higher expectations. Complications, both straightforward and complex, can occur with all types of surgery, but with the appropriate preparation and management techniques, vitreoretinal surgical complications can be minimized and resolved. The intent of this article is to provide some insight into the pathogenesis, prevention, and treatment of complications of vitreoretinal surgery.

GLAUCOMA

The incidence of erythroclastic glaucoma secondary to intraocular hemorrhage can be reduced by meticulous removal of all preretinal blood with a straight extrusion cannula, as well as removal of blood entrapped in the peripheral edge of the resected posterior vitreous cortex (vitreous skirt).¹

Inflammatory glaucoma can occur from increased protein and cellular content in the anterior chamber, as well as presumed trabeculitis. Pressure increase can be seen with a subtotal gas fill on the first postoperative day with no blood in the anterior chamber, a deep chamber, and minimal cells and flare. Although many surgeons will blame the gas bubble, a subtotal gas bubble cannot cause a pressure increase without pupillary block. The pressure increase must instead be from unrecognized trabeculitis.

An expanding gas bubble can cause increased intraocular pressure (IOP), which can lead to occlusion of the central retinal artery if the IOP is above the critical closing pressure. Elevated IOP can occur from a mixing error, such as confusing cubic centimeters in the syringe for percentage, pupillary block, or an unwise decision to use an expanding gas concentration in a total-fill surgical situation.²

Emulsification glaucoma is a relatively uncommon, delayed complication of silicone oil use. Viscoelastics, bleeding, and inflammation increase the risk of emulsification; topical steroids should be continued in nonresponders as long as silicone oil remains in the eye. In contrast to what some surgeons believe, the incidence of emulsification is the same with 1000 cS and 5000 cS oil.³ Emulsification glaucoma can be treated by silicone-silicone exchange if long-term silicone tamponade is needed for chronic inferior retinal detachment.

Steroid glaucoma is a significant complication because both dose and efficacy are proportional to dose. In the past there was too much emphasis on genetic susceptibility; in fact, all patients are susceptible to steroid glaucoma as proven by the Retisert fluocinolone acetonide implant (Bausch & Lomb, Rochester, NY) data.⁴ Earlier thinking indicated that steroid glaucoma was reversible, but patients with the fluocinolone device or those who are injected with triamcinolone often develop glaucoma that persists once the steroid is no longer present in the eye.

Hyperoxygenation of the vitreous cavity and, secondarily, the aqueous humor have been shown to occur after vitrectomy. Stanley Chang, MD, has postulated that hyperoxygenation is responsible for causing damage to the trabecular meshwork.⁵

Neovascular glaucoma is primarily produced by vascular endothelial growth factor (VEGF), most often in the context of proliferative diabetic retinopathy or central retinal vein occlusion.⁶

CATARACT

Progression of pre-existing nuclear sclerosis occurs after nearly all vitrectomies. It is probably related to significantly higher oxygen levels in the vitreous cavity, as reported by Nancy Holekamp, MD.⁷ No evidence exists that nuclear sclerosis is caused by vitrectomy if the lens is clear before surgery.

Posterior subcapsular cataract can occur from gas bubble contact if the anterior vitreous cortex has been removed, and it can also be caused by low-quality infusion fluid or additives to the infusion fluid, prolonged surgery with high fluid throughput, or lens bump, which is most common with wide-angle visualization.

TECHNIQUE-DRIVEN COMPLICATIONS

Silicone oil removal and replacement is unnecessary and increases trauma and complications. Redetachment from proliferative vitreoretinopathy (PVR), as well as epimacular membranes, are best treated by interface vitrectomy with the oil in place, as presented by the author in 2006 at the American Society of Retinal Surgeons (ASRS) meeting in Cannes, France, and the Club Jules Gonin meeting in Cape Town, South Africa.⁸

Combined phacoemulsification and intraocular lens procedures in uveitis cases increase cystoid macular edema rates and are, in my opinion, overutilized. Most eyes affected with uveitis severe enough to require vitrectomy are better served by performing simultaneous lensectomy and removing the entire lens with forceps without iris manipulation.

Combined phaco procedures cause miosis and hypotony, as well as more inflammation and corneal edema than is seen with staged procedures. Although combined phaco and vitrectomy can be done efficiently by superb surgeons, I believe overutilization has resulted in significant complications.

Iris retractors can cause iritis, fibrin syndrome, postop dilation, or pupillary distortion. It is usually better to use wide-angle visualization to see through a small pupil than to use retractors.

INFLAMMATION AND TOXICITY

Fibrin syndrome is similar to toxic anterior segment syndrome as well as sterile endophthalmitis described with intravitreal triamcinolone and hyaluronidase. Excessive retinopexy, especially cryopexy, panretinal photocoagulation (PRP) on formerly detached, edematous retina, and iris manipulation are common causes of fibrin syndrome, which is usually best treated with tissue plasminogen activator and intravitreal or periocular steroids.

Toxicity can be caused by infusion of fluid additives such as dextrose and bicarbonate, which are unnecessary when using BSS Plus (Alcon, Fort Worth, TX) and operating on euglycemic patients, as we do currently. Preservative-free epinephrine should be pulsed in the infusion port if absolutely necessary, not added to infusion fluid. Wide-angle visualization obviates the need for epinephrine in most cases. Antibiotics in the infusion fluid are capable of causing retinal toxicity and severe visual loss; this problem is exacerbated if mixing errors occur. Triamcinolone itself, and not just the preservative, have been shown by Baruch Kuppermann⁹ to be toxic in tissue culture assays. Triamcinolone is excellent for particulate marking of residual vitreous but should be used with caution.

Figure 1. Endophthalmitis.

Indocyanine green (ICG) produces both chemical toxicity and phototoxicity. It is my opinion that ICG is unnecessary for internal limiting membrane (ILM) peeling. The Alcon 25G DSP ILM forceps facilitate ILM peeling without ICG, pics, bent MVR blades, or membrane scrapers.

BLEEDING

Laser hemostasis is effective for vascular attachment points severed during delamination and avoids the larger area of retinal necrosis associated with bipolar diathermy resulting in late atrophic retinal breaks. Closing the eye with normal IOP applies internal pressure to the scleral tunnel produced by angulated wound construction with 23-g or 25-g and reduces postoperative–hypotony-related hemorrhage. Systemic hypertension, especially during the 24 hours after surgery, increases bleeding, especially if peeling, segmentation, or delamination of vascularized epiretinal membranes was performed.

CORNEAL COMPLICATIONS

Corneal edema secondary to endothelial damage from low-quality irrigating fluid and/or infusion fluid additives, complications of combined phaco, or high fluid throughput is less common than in the early years of vitrectomy and largely preventable with good technique. Corneal epithelial damage can be produced by poor-quality contact lens irrigation fluid, excessive scrapping of the epithelium, and damage from the preparation or drying. The author performs about 650 vitrectomies per year and has 1 or 2 cases at most of postoperative corneal problems.

WOUND LEAKS AND HYPOTONY

Wound leaks are more frequent in 23-g than 25-g surgery. Angulated wound construction is obligatory for 23-g surgery, but I advise it for all 25-g surgery as well. The surgeon should not hesitate to suture one or more wounds if there is any doubt about wound integrity; 8-0 Biosorb (Alcon) creates less inflammation than polyglactin (Vicryl, Novartis) and is ideal for full-thickness conjunctival-scleral wound closure. Choroidals are very rare with angulated wound construction but can be treated with 1 to 2 days of pressure patching.

Figure 2. Phototoxicity.

ENDOPHTHALMITIS

There is a perception that 23-g or 25-g surgery results in a higher incidence of postoperative endophthalmitis (Figure 1). I believe that vitreous wicks and the unwise omission of subconjunctival antibiotics are the reasons for this problem. No evidence exists that topical antibiotics produce minimum inhibitory concentration levels in the vitreous, so surgeons should use both subconjunctival ceftazidime and gentamycin or tobramycin; the spectrum of ceftazidime alone is insufficient for the most virulent organisms.

SYSTEMIC ANTICOAGULATION ISSUES

Stopping anticoagulants before surgery is unnecessary and increases risk of stroke and myocardial infarction (Steve Charles, MD, Philip Rosenfeld, MD, and Stephen Gayer, MD, unpublished data, June 2007). There is no scientific evidence that anticoagulants increase the risk of ocular bleeding.

IATROGENIC RETINAL BREAKS

Peeling of highly adherent epiretinal membranes in diabetic traction retinal detachment surgery or retinopathy of prematurity often produces retinal breaks. These cases should be managed by conformal cutter delamination, foldback cutter, delamination, or scissors delamination.¹⁰ Port-based flow limiting reduces incidence of iatrogenic retinal breaks; always use the highest possible cutting rates, and increase proportional vacuum until vitreous/tissue removal rate is sufficient; high flow rates increase vitreoretinal traction even during core vitrectomy.¹⁰ The cutter should never be pulled away from the vitreous while suction is applied. The continuous engage-and-advance method I advocate reduces both core vitrectomy and iatrogenic retinal breaks.

PHOTOTOXICITY

Xenon and mercury-vapor high-intensity light sources, while essential for small-gauge surgery, have increased the likelihood of phototoxicity (Figure 2). The surgeon should always start with low intensity and increase until the illumination is sufficient. This is especially crucial after lamp replacement because xenon light output decreases during lamp life. Starting with lower intensity is also crucial when the previous case was 25-g and current case is 20-g because of the greater light loss associated with small-gauge tools.

SUMMARY

The list of potential complications is long and at times daunting, but fortunately most are manageable. Many complications are preventable by strict attention to detail. Brighter light sources, use of off-label drugs, and the advent of a variety of new technologies will benefit patient outcomes, but the surgeon must remain vigilant to avoid complications. Meticulous analysis of patients with poor outcomes is essential along with real-time monitoring of all outcomes. Convergence on best practice should occur throughout the vitreoretinal community if surgeons remain open minded and continue their dialogue. RP

REFERENCES

1. Brucker AJ, Michels RG, Green WR. Pars plana vitrectomy in the management of blood-induced glaucoma with vitreous hemorrhage. Am J Ophthlamol. 1978;10:1427.

2. Campbell DG, Simmons RI, Tolentino FL, McMeel JW. Glaucoma occurring after closed vitrectomy. Am J Ophthalmol. 1977;83:63.

3. Scott IU, Flynn HW Jr, Murray TG, Smiddy WE, Davis JL, Feuer WJ. Outcomes of complex retinal detachment repair using 1000- vs 5000-centistoke silicone oil. Arch Ophthalmol. 2005;123:473-478.

4. Jaffe GJ, Martin D, Callanan D, Pearson PA, Levy B, Constock T; Fluocinolone Acetonide Uveitis Study Group. Fluocinolone acetonide implant (Retisert) for noninfectious posterior uveitis; thirty-four-week results of a multicenter randomized clinical study. Ophthalmology. 2006;113:10-1027. Epub 2006. May 9.

5. Chang S. Later open-angle glaucoma after vitrectomy: The Jackson Memorial Lecture. Presented at the 2005 Annual Meeting of the American Academy of Ophthalmology; October 16, 2005; Chicago, IL.

6. Aiello LP, Avery RL, Arrigg PG, et al. Vascular endothelial growth factor in ocular fluid of patients with diabetic retinopathy and other retinal disorders. N Engl J Med. 1994;331:1480-1487.

7. Holekamp NM, Shui YB, Beebe DC. Vitrectomy surgery increases oxygen exposure to the lens: a possible mechanism for nuclear cataract formation. Am J Ophthalmol. 2005;139:302-310.

8. Charles S. Interface vitrectomy. Presented at the 24th annual American Society of Retina Specialists/6th Annual European Vireo Retinal Society meeting; September 9-13, 2006; Cannes, France.

9. Narayanan R, Mungcal JK, Kenney MC, Seigel GM, Kuppermann BD. Toxicity of triamcinolone acetonide on retinal neurosensory and pigment epithelial cells. Invest Ophthalmol Vis Sci. 2006;47:722-728.

10. Charles S. Complications. In: Charles S, Calzada J, Wood B, eds. Vitreous Microsurgery. 4th Edition. Philadelphia, PA: Lippincott, Williams & Wilkins; 2006:215-219.