Chapter 225 – Neovascular Glaucoma
MAHER M. FANOUS
• A unique form of glaucoma resulting from ocular or extraocular disease that produces ischemia of the eye.
• Neovascularization of the iris or anterior chamber.
• Elevated intraocular pressure.
• Decreased vision.
• Bullous keratopathy.
• Ectropion uveae.
• Anterior chamber inflammation.
• Synechial angle closure.
• Cupping of the optic nerve.
Neovascular glaucoma (NVG) is a relatively common and potentially devastating disorder that occurs when new vessels proliferate on the iris and anterior chamber angle. The common insults that promote NVG are severe retinal hypoxia and retinal capillary nonperfusion. Invasion of the anterior chamber by a fibrovascular membrane initially obstructs aqueous outflow in an open-angle form and later contracts to produce secondary synechial angle-closure glaucoma. The key to successful treatment is early detection of the neovascularization process. Panretinal photocoagulation remains the most effective initial therapy, which reduces the stimulus for ocular neovascularization and the development of secondary angle-closure glaucoma. Once the anterior chamber is closed by synechiae, however, filtering should be considered.
EPIDEMIOLOGY AND PATHOGENESIS
In 1906, Coats described new vessel formation in irides of eyes that had central retinal vein occlusion (CRVO). The term neovascular glaucoma was proposed by Weiss et al. and is based on the presence of new vessels on the iris. Knowledge of disorders that predispose to the development of neovascularization of the iris helps the ophthalmologist identify patients at risk for developing NVG. Diabetic retinopathy and CRVO are by far the most common conditions that predispose to the development of ocular neovascularization. Other common causes are uveitis, central retinal artery occlusion, long-standing retinal detachment, and intraocular tumors; there are also two extraocular causes—carotid artery obstructive disease and carotid cavernous fistula. Brown et al. concluded that 36% of all cases of NVG arise from CRVO, 32% from proliferative diabetic retinopathy, and 13% from carotid artery occlusive disease. About one third of patients who have ischemic CRVO develop NVG. The incidence of neovascularization of the iris in patients who have diabetic retinopathy correlates with the extent of capillary dropout and retinal hypoxia. Diabetic patients who undergo vitrectomy, lensectomy, intracapsular cataract extraction, and extracapsular cataract extraction with posterior capsulotomy are at a higher risk of developing NVG.
The most widely accepted theory explaining the development of neovascularization is that the hypoxic retina produces a diffusible angiogenic factor that stimulates new vessel proliferation. Since 1948, many investigators have searched for this postulated angiogenic factor, and recent advances in molecular biology have led to the identification of several possible factors. Vascular endothelial growth factor (VEGF) represents the leading candidate. VEGF receptors are present on retinal capillary endothelial cells, and VEGF triggers growth in these cells. Endothelial cells have not been proved to produce VEGF thus far. In a nonhuman primate model with experimentally induced CRVO and neovascularization of the iris, VEGF and VEGF-mRNA levels were markedly elevated in ischemic retina. A strong temporal and spatial correlation occurs between intraocular VEGF protein levels and the extent of neovascularization. Elevated intraocular VEGF levels have been found in humans who have proliferative diabetic retinopathy and neovascularization of the iris.
Growing evidence supports a central role for VEGF in the process of iris neovascularization. Tolentino et al. showed that injection of recombinant human VEGF is sufficient to produce neovascularization in a nonhuman primate model. Neutralizing anti-VEGF antibodies prevent the development of iris neovascularization induced by retinal ischemia in a monkey model. Systemic administration of interferon-a in cynomolgus monkeys causes inhibition and regression of neovascularization.
The earliest sign of vascular proliferation appears at the pupillary margin. Neovascularization of the iris may be difficult to detect in its earliest stages. Slit-lamp biomicroscopy reveals fine, tortuous, randomly oriented tufts of vessels on the surface of the iris, near the pupillary margin. These tufts may be obscured in dark irides and more obvious in lighter irides. Neovascularization characteristically progresses from the pupillary margin toward the angle ( Fig. 225-1 ) of undilated pupils, but angle neovascularization in the absence of pupillary involvement may occur. Repeated gonioscopy is indicated in eyes at high risk for the development of NVG. As vascular proliferation develops, biomicroscopy of the anterior chamber shows cells and flare. Gonioscopy reveals new vessels that grow from the circumferential artery of the ciliary body onto the surface of the iris and onto the surface of the wall of the angle.
The vessels cross the angle recess and grow forward over the ciliary body band and scleral spur onto the trabecular meshwork, which imparts a characteristic red flush ( Fig. 225-2 ). Early in the course of anterior segment neovascularization, the intraocular pressure (IOP) often is normal. The new blood vessels
Figure 225-1 Neovascularization of the iris. Note the florid neovascular proliferation at the pupillary margin, which grows in random orientation on the iris surface.
Figure 225-2 Neovascularization of the angle. Gonioscopic view of new vessels that cover the trabecular meshwork and impart a characteristic red flush.
arborize to form a fibrovascular membrane (invisible on gonioscopy) that gives rise to a secondary open-angle glaucoma. The final stage is characterized by contraction of the fibrovascular membrane, which pulls the peripheral iris over the trabecular meshwork and results in variable degrees of synechial angle closure. Ectropion uveae and hyphema occur frequently. Ectropion uveae results from radial traction along the surface of the iris, which pulls the posterior pigmented layer of the iris around the pupillary margin onto the anterior iris surface. It is at this stage that patients may present with the dramatic onset of pain secondary to elevated IOP. Patients experience severely reduced visual acuity (to counting fingers), accompanied by corneal edema and anterior chamber inflammation.
Evaluation of the medical history is crucial to the identification of patients at risk for the development of NVG. Diabetes mellitus, hypertension, arteriosclerosis, and a previous history of vision loss indicative of CRVO or retinal detachment are important and must be identified, as these disorders predispose to the development of NVG. Recent surgery may increase the risk in predisposed individuals. It is imperative that a posterior segment examination be performed in all patients to identify concomitant retinal disease. The diagnosis of NVG is made based on the clinical examination. Careful slit-lamp and gonioscopic examinations are usually sufficient to make the diagnosis. An undilated pupil is helpful. The goal is to establish the diagnosis well before angle structures become involved and elevated IOP or synechial angle closure occurs.
Differential Diagnosis of Neovascular Glaucoma
Postsurgical dilatation of iris vessels
Dilatation of iris vessels secondary to uveitis
Fuchs’ heterochromic iridocyclitis
Essential iris atrophy
Acute angle-closure glaucoma
Congenital iris tufts
Lightly pigmented irides
Retinopathy of prematurity
Involvement of the anterior chamber angle sometimes occurs before the appearance of neovascularization of the iris. These vessels typically run on the iris surface, follow a nonradial course, and may cross the scleral spur. Thus, gonioscopy must be performed on every patient at risk for the development of NVG. In most instances, however, small tufts of neovascularization are noted first at the pupillary margin. This tendency for initial involvement of the pupillary margin appears to result from aqueous flow dynamics, whereby angiogenic factors produced in the posterior segment have the most contact with the pupillary margin.
Occasionally, early neovascularization may be missed when the new vessels are fine and thin, the iris is darkly pigmented, or pressure from the gonioscopy lens reduces the caliber of the new vessels and renders them clinically inapparent. Frequent follow-up of patients at high risk for NVG enables the earliest detection of new vessels in difficult cases. Fluorescein angiography of the iris may be used to demonstrate the presence of new vessels before they become apparent at the slit lamp.
The b wave–a wave amplitude ratio of the bright-flash, dark-adapted electroretinogram may help predict which eyes will develop NVG following CRVO. This ratio was found to be less than 1.0 (average 0.84) in eyes that developed NVG after ischemic CRVO. In contrast, the b wave–a wave amplitude ratio was always greater than 1.0 in eyes that did not develop NVG following CRVO.
Several entities present with prominent iris vessels and elevated IOP ( Box 225-1 ). In these entities, the cause of elevated IOP is a mechanism other than growth of new vessels and an associated contractile membrane over the chamber angle. A detailed history and slit-lamp examination can, in the majority of cases, distinguish these entities from NVG.
Postsurgical engorgement of iris vessels is one of the more common entities that may be confused with NVG. After intraocular surgery, vessel dilatation secondary to intraocular inflammation may occur, which might be confused with NVG in diabetic patients. Postsurgical dilatation of iris vessels usually resolves as the secondary uveitis is treated with topical corticosteroids, whereas neovascularization of the iris does not respond to such therapy. Similarly, vascular congestion of normal iris vessels secondary to uveitis from various causes may simulate NVG because of concomitant IOP elevation.
Fuchs’ heterochromic iridocyclitis may present with elevated IOP and iris neovascularization. The glaucoma in this entity is secondary to uveitis intrinsic to the disease process and not a result of the presence of new vessels. Neovascularization has been found in the chamber angle as well as on the iris. The new vessels are thin and fragile and may result in a hyphema after paracentesis (Amsler’s sign). It is thought that neovascularization occurs as a result of iris ischemia.
Essential iris atrophy may present with elevated IOP associated with new vessel formation (see Chapters 223 and 230 ). 
Acute angle-closure glaucoma may be confused with advanced, late-stage NVG. Both may present with markedly elevated IOP
and an edematous cornea. In acute angle-closure glaucoma, engorged iris vessels may take on the appearance of rubeosis iridis, which suggests advanced NVG. Gonioscopy of the fellow eye may help differentiate the two entities, as the fellow eye tends to be narrow or potentially occludable in angle-closure glaucoma.
Congenital iris tufts and prominent iris vessels in lightly pigmented irides may sometimes mimic neovascularization. Prominent iris vessels typically are radial, lie within the iris stroma, and, when in the chamber angle, do not cross the scleral spur. In contrast, new vessels typically do not follow a radial course, lie on the surface of the iris, and may cross the scleral spur.
Retinopathy of prematurity may present with engorged iris vessels as a result of plus disease and concomitant angle-closure glaucoma from a retrolental membrane. Plus disease refers to arteriolar tortuosity and venous engorgement of the posterior pole and iris secondary to vascular shunting. Retinopathy of prematurity is distinguished from NVG on the basis of characteristic fundus findings and the clinical setting.
The histology of angle structures in eyes with NVG does not differ significantly with regard to the different causes of the disorder. The neovascular process begins at the pupillary margin and occasionally at the iris periphery, where new vessels arise from preexisting capillaries. These vessels are thin walled, have fenestrated endothelial cells, and leak fluorescein on angiography. This is in contrast to normal blood vessels in the iris, which are thick walled and impermeable to fluorescein because of the presence of endothelial tight junctions. New vessels initially may proliferate on the iris surface with little structural support. However, as the disease progresses, an iridic membrane that consists of myofibroblasts forms over the new vessels and iris surface. The iridic membrane may contract, because of the presence of myofibroblasts, which results in synechial angle closure, ectropion uveae, and flattening of the normal topography of the anterior iris surface ( Fig. 225-3 ). This myofibroblastic membrane
Figure 225-3 Neovascularization of the iris. A, Note the peripheral anterior synechia, secondary angle closure, and tissue anterior to the anterior border layer of the iris (the last of which constitutes iris neovascularization). B, High-power view of iris neovascularization and ectropion uveae in A. (From Yanoff M, Fine BS. Ocular pathology, ed 5. St. Louis: Mosby, 2002.)
is present whenever new vessels are seen but is invisible on gonioscopy. The presence of this clinically inapparent membrane is why many eyes that have an apparent open angle and only mild or no angle neovascularization or synechia formation have elevated IOP. In some cases, the iridic membrane may stimulate, perhaps by contractile force, growth of corneal endothelium and Descemet’s membrane over a pseudoangle formed by synechial closure of the chamber angle.
The key to successful management of NVG is early diagnosis. Recognition of neovascularization of the iris is crucial so that preventive treatment can be initiated before the anterior chamber angle is closed by peripheral anterior synechiae. Once the florid, intractable final stage is established, the eye is blind, with very high IOP and painful bullous keratopathy. If the NVG is secondary to carotid artery or other systemic disease, it is important to evaluate and treat the primary systemic condition.
Panretinal photocoagulation (PRP) is the first line of therapy in almost all cases of NVG. Prompt application of PRP has been shown to effect regression of anterior and posterior segment neovascularization and to reduce the risk of developing neovascularization of the iris in eyes that have retinal vascular disease. In the open-angle glaucoma stage and early angle-closure glaucoma stage, PRP may reverse IOP elevation. For eyes that have advanced synechial angle closure of the anterior chamber with some potentially useful vision, PRP may eliminate the stimulus for neovascularization, which prepares the eye for filtering surgery and the prevention of further visual loss.
Panretinal cryotherapy is an alternative to PRP in eyes that have cloudy media and in eyes for which complete PRP fails to halt the progression of neovascularization. Goniophotocoagulation may be used as an adjunct to PRP to reduce neovascularization in the angle before it is closed by synechiae, but generally the effects are only temporary.
The treatment of NVG is directed by the visual potential. Any usable vision, even 20/400 (6/120) or less in a monocular patient, is worth preserving.
The role of glaucoma filtering surgery in NVG is to prevent pressure-induced injury to the optic nerve and, theoretically, to improve vascular perfusion. Before glaucoma surgery, every attempt is made to reduce or eliminate the stimulus for angiogenesis using PRP. Enough time must elapse between PRP and glaucoma surgery for the eye to quiet down; this reduces the risk of intraoperative and postoperative bleeding and severe intraocular inflammation. The importance of complete preoperative PRP to the success of glaucoma filtering surgery in patients who have NVG cannot be overstressed.
Of special importance is the use of intraoperative cautery to achieve hemostasis and avoid bleeding. Direct cauterization of the peripheral iris before iridectomy may reduce the risk of bleeding. Variable success rates have been reported after conventional filtering surgery in patients who have NVG. Allen et al. reported IOP control in 67% of patients who had NVG and underwent trabeculectomy or posterior lip sclerectomy after PRP. Tsai et al.  reported a high risk of long-term failure with 5-fluorouracil filtering surgery; 12 of 34 NVG patients (35%) lost light perception vision, and 8 patients (24%) developed phthisis bulbi over a 5-year follow-up period. Younger age (=50 years) and type I diabetes mellitus are significant risk factors for early surgical failure. Skuta et al. described the use of mitomycin C with trabeculectomy in patients who had NVG.
Glaucoma drainage implants are also used for the primary surgical treatment of NVG. Seton procedures place the effective sclerostomy inside the anterior chamber away from the angle, which maintains a patent fistula between the anterior chamber and an equatorial bleb. Sidoti et al. cited life-table success rates of 79% and 56% at 12 and 18 months, respectively, following Baerveldt glaucoma implantation surgery for NVG. Success was
defined as a final IOP of 6–21?mmHg (0.8–2.8?kPa) without additional glaucoma surgery or devastating complication. Loss of light perception occurred in 31% of patients. Cox model regression survival analysis demonstrated young patient age and poorer preoperative visual acuity as significant predictors of surgical failure. Another study that evaluated the use of the single-plate Molteno implant for NVG reported success rates of 62% at 1 year, 52.9% at 2 years, 43.1% at 3 years, 30.8% at 4 years, and 10.3% at 5 years; loss of light perception was seen in 29 of 60 eyes (48%), and phthisis bulbi developed in 11 eyes (18%).
Noninvasive techniques are employed to achieve patient comfort if the eye has no visual potential. Topically administered corticosteroids and cycloplegics may relieve ocular discomfort. Topical ß-adrenergic blockers, a-adrenergic agonists, and carbonic anhydrase inhibitors may be used to reduce IOP. Miotic therapy is avoided, as it may aggravate intraocular inflammation and pain. Once the anterior chamber angle is closed, medical therapy alone may not provide long-term IOP control, and surgical intervention becomes necessary. Cyclocryotherapy has been advocated in the treatment of elevated IOP in NVG. Although the IOP can be controlled in a high percentage of patients who undergo the procedure, the long-term visual outcome is dismal; loss of light perception occurs in 58.5% of patients. In addition, the high incidence of major complications, which include anterior segment necrosis and phthisis bulbi (34%), means that its use in eyes that have visual potential is limited. Other modalities of therapy to control IOP include diode and neodymium:yttrium-aluminum-garnet trans-scleral cyclophotocoagulation. Schuman et al.  reported a 39% success rate using the latter in patients who had advanced NVG.
In painful eyes that have poor visual potential, atropine and corticosteroid drops or cycloablation, retrobulbar alcohol injection, and enucleation may help achieve comfort.
COURSE AND OUTCOME
The natural course of NVG is uniformly one of complete loss of vision and the development of intractable, severe pain. The high degree of ocular morbidity and mortality in patients who have NVG emphasizes the severity of the underlying systemic conditions associated with diabetic retinopathy and CRVO, the main causes of this disorder. Usually, NVG occurs in patients burdened with serious systemic disease. Krupin et al. and Mermoud et al. cited mortality rates of 22% and 15% in patients who have NVG. Diabetes mellitus was the underlying cause of NVG in the majority of patients reported by the Diabetes Control and Complication Research Group, which highlights the importance of effective blood-sugar control in patients who have diabetic retinopathy. The risk of progression of mild diabetic retinopathy and the development of proliferative diabetic retinopathy is reduced to half in patients using effective blood-sugar control.
NVG does not invariably follow the development of neovascularization of the iris. When such neovascularization is detected, it behooves the clinician to follow patients carefully with repeated slit-lamp examinations and undilated gonioscopy. Neovascularization of the iris has been reported to develop in 50% of patients who have proliferative diabetic retinopathy and in 60% of those who have the ischemic type of CRVO. It is imperative that PRP be applied promptly to ischemic retina to eliminate the stimulus for further neovascularization.
Visual loss in NVG is common and may be attributed to a combination of causes, including severe ocular ischemia with progression of the underlying retinal disease, glaucomatous optic nerve damage, cataract formation, corneal decompensation, and phthisis bulbi. The most common cause of surgical failure in patients who have NVG is related to progression of the underlying retinal disease, not to uncontrolled IOP.  
The ultimate solution for patients who have NVG lies in the development of new modalities of treatment designed to prevent the initiation of neovascularization. Murata et al. recently showed that thiazolidinediones, a novel class of drugs that can be used to improve insulin resistance in type II diabetes, inhibit angiogenic responses to VEGF in vitro. Current research to develop pharmacological therapies targeted at the inhibition of angiogenic factors offers hope for the preservation of vision in patients at risk.
1. Coats G. Further cases of thrombosis of the central vein. R London Ophthalmol Hosp Rep. 1906;16:516–64.
2. Weiss DI, Shaffner RN, Nehrenberg TR. Neovascular glaucoma complicating carotid-cavernous fistula. Arch Ophthalmol. 1963;69:304–7.
3. Brown GC, Magargal LE, Schachat A, Shah H. Neovascular glaucoma. Etiologic considerations. Ophthalmology. 1984;91:315–20.
4. Adamis AP, Shima DT, Tolentino MJ, et al. Inhibition of vascular endothelial growth factor prevents retinal ischemia-associated iris neovascularization in a nonhuman primate. Arch Ophthalmol. 1996;114:66–71.
5. Adamis AP, Miller JW, Bernal M-T, et al. Increased vascular endothelial growth factor levels in the vitreous of eyes with proliferative diabetic retinopathy. Am J Ophthalmol. 1994;118:445–50.
6. Tolentino MJ, Miller JW, Gragoudas ES, et al. Vascular endothelial growth factor is sufficient to produce iris neovascularization and neovascular glaucoma in a nonhuman primate. Arch Ophthalmol. 1996;114:964–70.
7. Miller JW, Stinson WG, Folkman J. Regression of experimental iris neovascularization with systemic alpha-interferon. Ophthalmology. 1993;100:9–14.
8. Sabates R, Hirose T, McNeel JW. Electroretinography in the prognosis and classification of central retinal vein occlusion. Arch Ophthalmol. 1983;101:232–5.
9. Perry HD, Yanoff M, Scheie NG. Rubeosis in Fuchs’ heterochromic iridocyclitis. Arch Ophthalmol. 1975;93:337–9.
10. Jampol LM, Rosser MJ, Sears ML. Unusual aspects of progressive essential iris atrophy. Am J Ophthalmol. 1974;77:353–7.
11. John T, Sassani JW, Eagle RC. The myofibroblastic component of rubeosis iridis. Ophthalmology. 1983;90:721–8.
12. Cashwell LF, Marks WP. Panretinal photocoagulation in the management of neovascular glaucoma. South Med J. 1988;81:1364–8.
13. Vernon SA, Cheng H. Panretinal cryotherapy in neovascular disease. Br J Ophthalmol. 1988;72:401–5.
14. Simmons RJ, Deppermann SR, Dueker DK. The role of gonio-photocoagulation in neovascularization of the anterior chamber angle. Ophthalmology. 1980; 87:79–82.
15. Allen RC, Bellows AR, Hutchinson BT, Murphy SD. Filtration surgery in the treatment of neovascular glaucoma. Ophthalmology. 1982;89:1181–7.
16. Tsai JC, Feuer WJ, Parrish RK II, Grajewski AL. 5-Fluorouracil filtering surgery and neovascular glaucoma. Long-term follow-up of the original study. Ophthalmology. 1995;102:887–93.
17. Skuta GL, Beeson CC, Higginbotham EJ, et al. Intraoperative mitomycin versus postoperative 5-fluorouracil in high-risk glaucoma filtering surgery. Ophthalmology. 1992;99:438–44.
18. Sidoti PA, Dunphy TR, Baerveldt G, et al. Experience with the Baerveldt glaucoma implant in treating neovascular glaucoma. Ophthalmology. 1995;102:1107–18.
19. Mermoud A, Salmon JF, Alexander P, et al. Molteno tube implantation for neovascular glaucoma. Long-term results and factors influencing the outcome. Ophthalmology. 1993;100:897–902.
20. Krupin T, Mitchell KB, Becker B. Cyclocryotherapy in neovascular glaucoma. Am J Ophthalmol. 1978;86:24–6.
21. Schuman JS, Bellows AR, Shingleton BJ, et al. Contact transscleral Nd:YAG laser cyclophotocoagulation: midterm results. Ophthalmology. 1992;99:1089–95.
22. Krupin T, Kaufman P, Mandell AI, et al. Long-term results of valve implants in filtering surgery for eyes with neovascular glaucoma. Am J Ophthalmol. 1983;95: 775–82.
23. Diabetes Control and Complications Research Group. The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus. N Engl J Med. 1993;329:977–86.
24. Murata T, Hata Y, Ishibashi T, et al. Response of experimental retinal neovascularization to thiazolidinediones. Arch Ophthalmol. 2001;119:709–17.