11.5 Anterior Segment Neovascularization in Central Retinal Vein Occlusion

11.5Anterior Segment Neovascularization in Central Retinal Vein Occlusion

When ASNV follows ischemic CRVO, the extent of capillary nonperfusion is correlated with the level of aqueous VEGF.19 ASNV develops when aqueous VEGF concentrations reach the range 849–1,569 pg/ml.9 The threshold for ASNV disappearance after PRP is 378 pg/ml.9 Therefore, a lower concentration of VEGF is necessary for maintenance than for induction of ASNV.9

Anterior segment neovascularization develops only in the ischemic form of CRVO.19 It has been reported to occur after 21–70%, median 33%, (Table 11.4) of cases of ischemic CRVO, a wide range reflecting the poor reproducibility in determining ischemia (see Chap. 8) and the various methods for detecting ASNV.11,17,29,30,32,46,56,83 When ASNV follows CRVO, it occurs at the pupillary

margin in 44%, in the angle in 12%, and at both the pupillary margin and in the angle in 44%.11 Two studies showed that angle neovascularization can develop before pupillary margin neovascularization.11,72 Thus, undilated slit lamp examination and gonioscopy for the first 6 months are indicated in eyes with CRVO and signs of ischemia with lengthened follow-up after 6 months passes and no evidence of ASN.11,72

The time course for development of ASNV has been documented in many studies (Fig. 11.13). In the Central Vein Occlusion Study (CVOS), the median time to development of ASNV was 61 days with a range of 6 days to 8 months. In a large case series from the Wills Eye Hospital, the time of onset of NVI after onset of CRVO was 2 weeks to 2 years with a mean time to NVI of 5 months.46 In a consecutive series of 12 cases developing NVI after acute CRVO, the median time to development was 2 months, range 3 weeks to 7 months.63 In another study, 90% of ASNV developed within

Iris NV

Angle NV

NV Glaucoma

Time in days

9 months of acute CRVO.11 Approximately 75% of cases of the various forms of ocular neovascularization following ischemic CRVO will develop by 3 months and 85% by 1 year, but the remainder may continue to develop over an additional 2 years. Thus, careful monitoring for ASNV is most critical in the first year after CRVO, but subsequent follow-up of predisposed eyes at less frequent intervals is also necessary.23,31

At all follow-up times, the order of frequency of the various manifestations of ASNV after ischemic CRVO is NVG < NVA < NVI (see Tables 9.2 and 11.4). Thus, not every patient with ASNV goes on to develop NVG. The percentage of eyes with ASNV after CRVO not developing NVG has been reported to be 33% of all eyes with NVI and 22% to 25% of all eyes with either NVI or NVA.18,25

ASNV occurred in 35% of eyes categorized at baseline as ischemic or indeterminate and 10% of eyes categorized as nonischemic using FA criteria alone.46 Looked at from a different perspective, 49% of eyes with CRVO that progressed to NVI had a baseline FA classified as nonischemic, 31% as ischemic, and 20% as indeterminate.46 Thus, a nonischemic categorization of an eye with CRVO is not a basis for complacence on checking for subsequent ASNV.71 Likewise, the risk of NVG in CRVOs classified as nonischemic at diagnosis was 1%. One infers that eyes classified as nonischemic CRVOs at baseline that later develop NVG had to have converted to ischemic CRVOs at some unobserved point.47 Thirty-three percent of eyes beginning as nonischemic CRVOs that later convert to ischemic CRVOs eventually develop NVG.29

Predicting development of NVI after CRVO has been the goal of many studies.33 The relative afferent pupillary defect (RAPD) gives useful prognostic information throughout the disease course and is uninfluenced by the degree of intraretinal hemorrhage.25 In one series, no eye developed ASNV without a RAPD of 1.2 log units or greater.62 The extent of capillary nonperfusion present on FA has also been useful, when the angiogram is interpretable. The greater the baseline area of capillary nonperfusion, the higher the risk of ASNV.12 The risk of ASNV is

16% when 11–29 disc areas (DAs) of capillary nonperfusion are present and 52% when 75 or more DAs of capillary nonperfusion are present.70 For eyes with CRVO of duration less than 1 month, the baseline VA contains all the predictive information relative to subsequent risk of ASNV.71 For a doubling of the baseline visual angle, the risk of subsequent ASNV increases by a factor of 1.7.71

Patients developing NVG generally

•Have a shorter duration of symptoms

•Are older

•Are more likely to have hypertension, diabetes, atherosclerotic cardiovascular disease, and POAG

•Are more likely to have moderate-to-severe venous tortuosity

•Are more likely to have a ratio of electroretinogram b-wave amplitude in the involved eye to fellow eye less than 0.7 than patients not developing NVG, although there is some inconsistency across studies in these factors.11,18,46,63,69,71,76 Patients developing ASNV have not been found to differ in systemic associations compared to patients not developing ASNV.36,57

Iris fluorescein angiography has been examined as a predictor of later NVI.43 Normal iris vessels become leaky after ischemic CRVO and before development of NVI.40 The best predictor

from baseline iris FA was relative area of iris late fluorescein staining, but the specificity of the test was too poor, and few have adopted iris fluorescein angiography for this purpose.43

Central retinal venous blood flow velocity has been used to predict the later development of NVI.82 In patients examined within 3 months of onset of CRVO, a cutoff point exists that detects later NVI with 76% sensitivity and 86% specificity.82 Few clinicians have access to such technology, making its use for such a purpose unlikely in actual practice.

Associations of ASNV after CRVO with certain laboratory tests have also been found (Table 11.5). The result for factor VII is paradoxical to intuition, as factor VII is a procoagulant protein. The results for antithrombin III and TPA are intuitive as ATIII is an anticoagulant and TPA is profibrinolytic.

Presence of optic disc collateral vessels is predictive that ASNV will not develop.20 Eyes developing ASNV were one twenty-fifth as likely to have optic disc collateral vessels as eyes not developing ASNV.20 Presumably, optic disc collaterals confer some protection against ischemia. A lesser protective factor against development of ASNV in one study was macular lipid.11 The biological plausibility of this purported association is more suspect.

The prevalence of NVI after CRVO does not depend on the status of the vitreous. In a case series of 52 eyes with ischemic CRVO, 38 had a complete PVD and 14 did not. Of the 38 with complete PVD, 22% went on to develop NVI compared to 21% without complete PVD.32 Treatment of macular edema in CRVO with intravitreal triamcinolone did not reduce the incidence of ASNV in the SCORE CRVO study.12

Untreated eyes developing NVG after CRVO go on to blindness in 76% and 28% go on to phthisis bulbi.23 Because a high proportion of eyes developing ASNV go on to develop NVG, most clinicians treat all cases of ASNV after CRVO. A minority disagree with this approach on grounds that this commits some patients not destined to NVG to iatrogenic field loss from laser panretinal photocoagulation (PRP).

The standard treatment for anterior segment neovascularization is PRP.9,18,45,46,72 Because ASNV can progress rapidly from the open-angle stage to the closed-angle stage, PRP should be applied promptly after diagnosis of ASNV.23,46 When the view is inadequate, cryotherapy can be used, but is not the first choice because it is associated with greater inflammation and more pain during treatment. For short-term regression of anterior segment neovascularization, intravitreal or intracameral injection of bevacizumab is effective and can cause rapid regression of new vessels, reduction in intraocular pressure, and clearing of corneal edema, which can allow the more permanent measure of PRP to be applied successfully.8

There is a good correlation between the effectiveness of laser PRP in causing regression in NVI and in causing a reduction of the aqueous concentration of VEGF.9 In one series, eyes with NVI following CRVO had a median aqueous VEGF concentration of 1,201 pg/ml. Following laser PRP with successful regression of NVI, the median aqueous VEGF concentration was 204 pg/ml. Those with persistence of NVI despite laser PRP had median aqueous VEGF of 1,295 pg/ml.9