Branch Vein Occlusion

Published on 08/03/2015 by admin

Filed under Opthalmology

Last modified 22/04/2025

Print this page

This article have been viewed 3537 times

Chapter 53 Branch Vein Occlusion

Alex Yuan, Peter K. Kaiser

Introduction

Branch retinal vein occlusion (BRVO) is a common cause of retinal vascular disease.¹ The Beaver Dam Study estimated the 15-year cumulative incidence of retinal vein occlusions (RVO) at 2.3% in the population, with a majority of these (78%) being BRVO.² BRVO affects males and females equally and occurs most frequently between the ages of 60 and 70. The pathologic interruption of venous flow in these eyes almost always occurs at a retinal arteriovenous intersection, where a retinal artery crosses over a retinal vein.³^–⁷ Systemic vascular diseases such as hypertension and arteriosclerosis are risk factors for BRVO, probably because they lead to thickening of the retinal artery.^3,⁷ Other risk factors for BRVO include diabetes, smoking, hyperlipidemia, glaucoma, and ocular inflammatory disease.⁸ Antiphospholipid antibodies, elevated plasma homocysteine levels, and low serum folate levels have also been associated with increased risk of vein occlusion.⁹^–¹¹ A decreased risk is present in those with higher serum levels of high-density lipoprotein and light to moderate alcohol consumption.⁸ Other studies have suggested an increased risk of BRVO in eyes with shorter axial lengths.¹²^–¹⁵ This section discusses the pathophysiology, clinical features, evaluation, and treatment of patients with BRVO.

Pathogenesis

Because BRVO mostly occurs at arteriovenous crossings,^4,^8,¹⁶ underlying arterial disease may play a causative role. In 99% of 106 eyes with BRVO, the artery was located anterior to the vein at the obstructed site.^7,⁸ Histopathologically, the retinal artery and vein share a common adventitial sheath, and in some cases, a common medium.¹⁷ The lumen of the vein may be compressed up to 33% at the crossing site.^7,¹⁸ The vitreous may also play a role in compression of susceptible arteriovenous crossing sites, as evidenced by studies demonstrating that eyes with decreased axial length and higher likelihood of vitreomacular attachment at the arteriovenous crossing are at increased risk of BRVO.^13,¹⁹

Some have postulated that turbulent blood flow at the crossing site causes focal swelling of the endothelium and deeper vein wall tissue, leading to venous obstruction.^17,^19,²⁰ Other reports have demonstrated actual venous thrombus formation at the point of occlusion.^3,¹⁸ The resulting venous obstruction leads to elevation of venous pressure that may overload the collateral drainage capacity²¹ and lead to macular edema and ischemia by mechanisms that are still under investigation. Unrelieved venous pressure can also result in rupture of the vein wall with intraretinal hemorrhage.¹⁷ Vision loss from RVOs is typically due to macular ischemia, macular edema, or complications from neovascular disease.

Clinical features

Symptoms

Patients with BRVO present with sudden painless loss of vision or a visual field defect. Subclinical presentations may occur if a tributary distal to the macula or a nasal retinal vein is involved. Rarely, patients with BRVO will present with floaters from a vitreous hemorrhage if the initial vein occlusion was unrecognized and retinal neovascularization has occurred.

Signs

Patients typically present with a wedge-shaped distribution of intraretinal hemorrhage that is less marked if the occlusion is perfused (or nonischemic), and more extensive if the occlusion is nonperfused (or ischemic) and associated with retinal capillary nonperfusion. The Branch Vein Occlusion Study Group defined ischemic BRVO as those with greater than a total of five disc diameters of nonperfusion on fluorescein angiography (FA).¹ The location of the venous blockage determines the distribution of the intraretinal hemorrhage; if the venous obstruction is at the optic nerve head, two quadrants of the fundus may be involved, whereas if the occlusion is peripheral to the disc, one quadrant or less may be involved with the intraretinal hemorrhage. If the venous blockage is peripheral to tributary veins draining the macula, there may be no macular involvement and no decrease in visual acuity. The most common location for BRVOs is in the superotemporal quadrant.^5,^6,²² This favored location may be attributed to a larger number of arteriovenous crossings in the superotemporal quadrant.

Figure 53.1 demonstrates the typical acute appearance of a BRVO involving the superotemporal quadrant of the right eye. A narrowed branch retinal vein passing under a retinal artery can sometimes be identified proximal to the hemorrhage. Rarely, a patient may present initially with very little intraretinal hemorrhage, which then becomes more extensive in the succeeding weeks to months. In these instances, it is presumed that an incomplete block at the arteriovenous crossing has progressed to more complete occlusion.

Fig. 53.1 Acute branch retinal vein occlusion. Intraretinal hemorrhages in a wedge-shaped pattern delineating the area drained by the occluded vein. The occluded vessel is often seen passing underneath a retinal artery (arrowhead). Cotton-wool spots (asterisk) are sometimes seen. Note the dilated and tortuous occluded vein (arrow) compared to the normal retinal vein in the inferior arcade.

Over time the intraretinal hemorrhage may completely resorb. Without the characteristic segmental distribution of intraretinal hemorrhage, the ophthalmoscopic diagnosis may be more difficult, but the segmental distribution of retinal vascular abnormalities that occurred during the acute phase will persist and be apparent on FA. In many cases, macular edema may also be detected by optical coherence tomography (OCT). Consequently, in the chronic phase of the disease, after intraretinal hemorrhage absorption, the diagnosis may depend on detecting a segmental distribution of retinal vascular abnormalities that may include capillary nonperfusion, dilation of capillaries, microaneurysms, telangectatic vessels, and collateral vessel formation (Fig. 53.2).

Fig. 53.2 Chronic branch retinal vein occlusion. (A) Color fundus photograph showing microaneurysms, exudates, and a sclerosed retinal vein (arrowhead) draining into a sheathed vessel (arrow). (B) Corresponding mid- to late-phase fluorescein angiogram shows abundant collaterals (arrowhead) and highlights the microvascular abnormalities.

Nonocular signs such as systemic hypertension have been associated with BRVOs.^2,^8,²³ Thus, systemic blood pressure may be elevated. In bilateral cases or cases involving young patients, systemic manifestations of infectious disease, inflammatory or autoimmune conditions, neoplasm, or hypercoagulable states may be present.

Complications

There are three common vision-limiting complications of BRVO: (1) macular edema; (2) macular ischemia; and (3) sequelae of neovascularization. It is important to appreciate the variability of these complications before considering the benefits of treatment.²⁴^–²⁶

During the acute phase, extensive intraretinal hemorrhages may obscure macular ischemia and macular leakage on FA. Under these circumstances it is impossible to evaluate the perfusion status by FA because the hemorrhage itself blocks the view of the vasculature. In addition, the hemorrhage in the foveal center may reduce visual acuity independently of any macular edema or ischemia. Since this reduction in visual acuity may recover completely if there is no other cause for the visual loss, such as macular edema or macular capillary nonperfusion, observation in these cases can be considered. When there is extensive foveal hemorrhage, OCT is an important ancillary test to look for macular edema. Although it may be difficult to provide a prognosis in the acute phase, it is helpful to recognize that about one-third to one-half of patients with BRVO have a return of vision to 20/40 or better without any therapy.

Retinal and iris neovascularization, vitreous hemorrhage, traction retinal detachment, and neovascular glaucoma are complications that manifest late in the course of the disease due to ischemia. With the exception of macular ischemia, these complications can largely be treated or prevented. Thus, it is important that patients with BRVO be closely followed. From the Branch Vein Occlusion Study, 31–41% of patients with ischemic BRVO (defined as >5 disc diameters of nonperfusion on FA) developed neovascularization or vitreous hemorrhage, compared with 11% of patients with nonischemic BRVO.¹

Clinical evaluation

Clinical examination

A complete ophthalmic examination should be performed, paying particular attention to the history of glaucoma and signs of intraocular inflammation, since these are risk factors for BRVO. Careful examination of the iris and angle should be performed in appropriate cases to monitor for early signs of rubeosis or neovascular glaucoma. Initially, when the risk of macular edema and neovascularization is higher, patients should be followed every month. Once stable, and if visually significant macular edema and other complications are not present, follow-up can be extended.

Fluorescein angiography

To help verify the diagnosis and evaluate for complications, FA should be obtained to delineate the retinal vascular characteristics that may have prognostic significance: macular leakage and edema, macular ischemia, and large segments of capillary nonperfusion that may portend eventual neovascularization. FA is the only technique that will accurately define the capillary abnormalities in BRVO; it is therefore particularly important that high-quality angiography be obtained (Fig. 53.3).

Fig. 53.3 Fluorescein angiogram of branch retinal vein occlusion. (A) Blocked fluorescence from intraretinal hemorrhage is common in acute branch retinal vein occlusion. Note the telangiectatic vessels forming collaterals across the horizontal raphe. The hemorrhages obscure underlying areas of capillary nonperfusion and edema. (B) Six months later, the hemorrhages have cleared, revealing small patches of nonperfusion and macular edema.

The characteristic finding on FA is delayed filling of the occluded retinal vein. Varying amounts of capillary nonperfusion, blockage from intraretinal hemorrhages, microaneurysms, telangiectatic collateral vessels, and dye extravasation from macular edema or retinal neovascularization are other features encountered. In chronic cases, when the hemorrhages have resolved, microvascular changes on FA may provide the only clues of a previous BRVO.

When FA demonstrates macular leakage and edema with cystoid involvement of the fovea, but no capillary nonperfusion, it is presumed that the macular edema is the cause of vision loss. Under these circumstances, about one-third of patients will spontaneously regain some vision. However, patients who have had decreased vision for over 1 year as a result of macular edema are much less likely to regain vision spontaneously. When macular edema is present ophthalmoscopically within the first 6 months after a BRVO and there is little or no leakage on FA, macular ischemia may be the cause of the macular edema. In such circumstances, the edema almost always spontaneously resorbs in the first year after the occlusion, often with return of vision.²⁷

Unfortunately, acute BRVOs with dense intraretinal hemorrhages may make FA interpretation challenging due to blockage of fluorescence by the hemorrhages. Thus, it is advisable to obtain FA only after the intraretinal hemorrhages have cleared significantly from the macula. Other diagnostic tests, such as OCT, can be obtained in the acute phase to aid in the diagnosis of macular edema.

Wide-field angiography

Wide-field angiography is not a commonly used imaging modality for patients with BRVO, but a recent study supports its utility. Ultrawide-field FA using the Optos C200MA revealed a correlation between nonperfusion in the peripheral retina with macular edema and retinal neovascularization.²⁸ Future studies to determine if laser photocoagulation of the nonperfused peripheral retina decreases macular edema and neovascularization may alter the therapeutic paradigm. In patients with recalcitrant macular edema or retinal neovascularization, wide-field angiography may reveal peripheral areas of nonperfusion helping to guide targeted laser photocoagulation.

Optical coherence tomography

OCT has arguably become the most important imaging modality in the treatment of patients with BRVO and macular edema. OCT offers a noninvasive and rapid method of quantitatively measuring macular edema. OCT is frequently used to monitor the response to treatment of macular edema and has been used in place of FA in some treatment trials for BRVO. Unlike FA, intraretinal hemorrhages have a minimal effect on the interpretation of OCT, making this imaging modality helpful, even in the acute setting with foveal hemorrhage. The characteristic findings of BRVO on OCT are cystoid edema, intraretinal hyperreflectivity from hemorrhages, shadowing from edema and hemorrhages, and occasionally subretinal fluid ^29,³⁰ (Fig. 53.4A). In chronic cases, photoreceptor inner-segment–outer-segment junction abnormalities from long-standing macular ischemia and macular edema may also be seen (Fig. 53.4B).

Fig. 53.4 Spectral-domain optical coherence tomography of an eye with a branch retinal vein occlusion (BRVO). (A) Raster scan of the BRVO in Figure 53.1 reveals cystoid macular edema, intraretinal fluid, and shadowing (between the arrowheads). Intraretinal heme (arrow) appears hyperreflective and produces a shadow on optical coherence tomography. (B) Raster scan of a chronic BRVO with inner-segment–outer-segment abnormalities and a large cyst.

Diagnostic workup

Treatment options

Medical treatment

In cases where a hypercoagulopathy has been identified, anticoagulation may be considered in consultation with an internist. In most cases, however, anticoagulant therapy has not been shown to be beneficial in either the prevention or the management of BRVO. Since the systemic administration of anticoagulants can be associated with systemic complications, and could, in theory, increase the severity of intraretinal hemorrhage occurring in the acute phase, such therapy is not recommended. Treatment of BRVO has focused on the management of vision-limiting complications.

Laser treatment

Branch Vein Occlusion Study for macular edema

The collaborative Branch Vein Occlusion Study (BVOS),³⁸ a multicenter randomized clinical trial supported by the National Eye Institute, reported that argon laser photocoagulation may reduce visual loss from macular edema for those eyes that meet study eligibility criteria and are treated according to that protocol. Important eligibility criteria included fluorescein-proven perfused macular edema involving the foveal center, absorption of intraretinal hemorrhage from the foveal center, recent BRVO (usually 3–18 months’ duration), no diabetic retinopathy, and vision reduced to 20/40 or worse after best refraction.

In the BVOS,³⁸ argon laser photocoagulation was applied in a grid pattern throughout the leaking area demonstrated by FA (Fig. 53.5). Coagulation extended no closer to the fovea than the edge of the capillary-free zone and no further into the periphery than the major vascular arcade. Recommended treatment parameters included a duration of 0.1 second, a 100-µm diameter spot size, and a power setting sufficient to produce a “medium” white burn. FA was repeated 2–4 months after the treatment, and additional photocoagulation was applied to residual areas of leakage if reduced visual acuity persisted. Improvement in visual acuity was assessed in several ways.³⁸ When improvement was defined as reading two or more Snellen lines (beyond baseline) at two consecutive visits, treated eyes showed visual improvement more often than untreated eyes. After 3 years of follow-up, 63% of treated eyes gained two or more lines of vision, compared to 36% of untreated eyes. The average gain in visual acuity for treated eyes was one more Snellen line than in untreated eyes.

Fig. 53.5 Grid macular laser for macular edema. (A) Fluorescein angiogram, late phase, demonstrating macular edema with foveal involvement. (B) Immediate posttreatment fundus photograph showing grid pattern of laser photocoagulation.

Before laser photocoagulation is performed, it is important to obtain high-quality FAs of the macula; the FA must demonstrate that the macular edema involves the center of the fovea and that there is not a large amount of capillary nonperfusion adjacent to the capillary-free zone that could explain the visual loss. In addition, it is important to follow patients for a length of time sufficient to ascertain that macular edema is not resolving spontaneously. During this period of follow-up, it should be demonstrated that there is clearing of intraretinal hemorrhage and that there is no hemorrhage in the center of the fovea that could account for a spontaneously reversible cause of visual loss. In the application of the grid photocoagulation, laser absorption occurs at the level of the pigment epithelium; photocoagulation is not applied to close the leaking and dilated capillary vasculature directly and immediately. Although it is not understood how the laser treatment may act in lessening edema, it is interesting to note that preliminary experimental studies in the normal primate have shown a decrease in capillary diameter when this form of therapy is used and when laser absorption occurs at the level of the pigment epithelium.³⁹ One explanation for the effect of grid pattern photocoagulation is that it results in a thinning of the retina (in particular the outer retina), reducing oxygen consumption and increasing choroidal delivery of oxygen to the inner retina, producing a consequent autoregulatory constriction of the retinal vasculature in the leaking area and thereby decreasing the edema.

In the application of grid pattern laser photocoagulation, it is crucial to obtain good definition of landmarks so that the center of the fovea can be identified and avoided. Since landmarks frequently may be obscured in the macula after BRVO, such cases can be managed more effectively and safely by treating well peripheral to the capillary-free zone in the first sitting. When the patient returns in 2 months for follow-up evaluation, a repeat FA may identify more clearly the amount of further treatment that needs to be applied closer to the edge of the capillary-free zone, because the pigmentation of the previous treatment is then visible. Consequently, treatment in this next sitting may be advanced closer to the edge of the capillary-free zone, if that is deemed necessary because of persistent foveal edema and vision loss. The placement of grid laser treatment in this repetitively staged fashion may be safer and appears to be just as effective as a single treatment. It has never been established that macular edema must be treated quickly or that long-standing edema produces irreversible macular damage in the first 2–3 years.

For the grid treatment used in the BVOS, the argon blue-green wavelength was employed.³⁸ This is the only wavelength that has been proven effective and it is unknown whether argon green and krypton red photocoagulation are equally effective. In other diseases, when laser treatment is applied inside the capillary-free zone, it is recognized that krypton red and argon green laser photocoagulation are absorbed less than blue-green by the xanthophyll pigment of the inner retina that is present in increasing concentrations close to the foveal center. However, because the grid treatment never comes closer to the fovea than the capillary-free zone, the BVOS did not encounter any problems with the argon blue-green laser in this region; consequently, this laser continues to be recommended.

The summary recommendations for management of acute branch vein occlusion from the BVOS emphasize waiting at least 3–6 months before considering laser therapy. If the vision is reduced to 20/40 or worse, wait 3–6 months for sufficient clearing of retinal hemorrhage to permit high-quality FA and then evaluate for macular edema and macular ischemia. If perfused macular edema accounts for the visual loss, and vision continues to be 20/40 or worse without spontaneous improvement, consider grid macular photocoagulation. However, this conclusion needs to be balanced against the improvements in vision seen with recent anti-vascular endothelial growth factor (VEGF) agents. If macular ischemia accounts for the visual loss, no laser treatment is recommended to improve vision.

Branch Vein Occlusion Study for neovascularization

A separate group of patients in the BVOS were randomized to receive scatter panretinal photocoagulation to prevent neovascular complications. The BVOS demonstrated that prophylactic scatter laser photocoagulation can lessen subsequent neovascularization and, if neovascularization already exists, that peripheral scatter laser photocoagulation can lessen subsequent vitreous hemorrhage.¹ Only eyes with the type of BRVO that shows large areas (>5 disc diameters) of retinal capillary nonperfusion are at risk for developing neovascularization. About 40% of these eyes develop neovascularization, and of this 40%, about 60% will experience periodic vitreous hemorrhage. Retinal or disc neovascularization, or both, may develop at any time within the first 3 years after an occlusion but are most likely to appear within the first 6–12 months after the occlusion. If peripheral scatter laser photocoagulation is applied in eyes with large areas of nonperfusion, the incidence of neovascularization can be reduced from about 40% to 20%. However, if one were to treat prophylactically, many eyes (60%) that would never develop neovascularization would receive peripheral scatter laser photocoagulation. For this reason, it is recommended that laser photocoagulation be applied only after neovascularization is observed.

Iris neovascularization is a rare complication of BRVO; it appears, however, that diabetes (with or without retinopathy) may increase this risk. Retinal neovascularization is particularly difficult to recognize in BRVO because the collaterals that develop frequently may mimic neovascularization. Arising presumably from pre-existing capillaries, these collaterals occur as vein-to-vein channels around the blockage site, across the temporal raphe, and in other locations to bypass the blocked retinal segment. These collaterals frequently become quite tortuous, mimicking the appearance of neovascularization if they are evaluated by ophthalmoscopy alone. When it is unclear whether an abnormal vascular pattern represents collateral formation or true neovascularization, the FA (Fig. 53.6) can be helpful because leakage from neovascularization is more prominent than from collateral vessels.

Fig. 53.6 Retinal neovascularization with leakage (arrow) can be differentiated from collaterals that are not leaking (arrowhead).

The BVOS data strongly suggest that photocoagulation after the development of neovascularization is as effective in preventing vitreous hemorrhage as is photocoagulation before the development of neovascularization.¹ When neovascularization is unequivocally confirmed by FA, peripheral scatter laser photocoagulation can reduce the likelihood of vitreous hemorrhage from about 60% to 30%. As demonstrated in Figure 53.7, the scatter laser photocoagulation can be applied with argon blue-green laser to achieve “medium” white burns (200–500 µm in diameter) spaced one burn width apart and covering the entire area of capillary nonperfusion, as defined by FA, but extending no closer than two disc diameters from the center of the fovea and extending peripherally at least to the equator. Retrobulbar anesthesia is used as needed for discomfort associated with the scatter photocoagulation.

Fig. 53.7 Immediate posttreatment fundus photograph showing pattern of peripheral scatter photocoagulation.

Of patients who develop neovascularization, approximately 60% experience episodes of vitreous hemorrhage if the condition is left untreated. The short- and long-term visual consequences of vitreous hemorrhage in BRVO have not been carefully studied. In some cases, the hemorrhage may be mild or may clear spontaneously without causing permanent visual impairment. However, in some patients, vitreous hemorrhage from neovascularization can lead to prolonged visual disability in the affected eye. When the hemorrhage is dense, B-scan ultrasonography may help rule out an associated traction retinal detachment. Most eyes can be observed. If the vitreous hemorrhage does not spontaneously clear in a few months, a pars plana vitrectomy with sector endolaser photocoagulation should be considered.

Familiarity with the laser treatment technique is required to individualize the treatment. Important variables, such as residual intraretinal hemorrhage, thickness of the retina from edema, location of collaterals, and presence of retinal traction, influence the exact mode of therapy within the above general treatment guidelines for the management of macular edema and neovascularization. There are numerous complications of laser photocoagulation; however, it is generally recognized that with proper attention to detail, complications are infrequent. Side-effects of treatment, including generation of scotoma, merit careful consideration and discussion with the patient before initiation of treatment. It is particularly important to recognize that laser photocoagulation should never be placed over extensive intraretinal hemorrhage in the acute phase of branch vein occlusion because the laser energy will be absorbed by the intraretinal hemorrhage rather than at the level of the pigment epithelium, likely damaging the nerve fiber layer and possibly enhancing the development of preretinal fibrosis.

Steroid treatment

Macular edema results from increased vascular permeability mediated at least in part by an increase in VEGF.^40,⁴¹ Corticosteroids have been shown to inhibit the expression of VEGF and therefore reduce macular edema in animal models.^42,⁴³ The anti-inflammatory effects of corticosteroids may further potentiate its anti-VEGF effects and help attenuate macular edema. Intraocular corticosteroids, however, have significant side-effects, including cataract formation and glaucoma. Several trials have evaluated the use of corticosteroids in the treatment of macular edema in BRVO.

SCORE (triamcinolone) study

In the Standard care vs Corticosteroid for Retinal vein occlusion (SCORE) BRVO study, the effectiveness and safety of intravitreal triamcinolone acetate (IVTA) for the treatment of macular edema from BRVO were evaluated.⁴⁴ In this multicenter, randomized controlled study, 411 patients were randomized to receive macular grid laser, 1 mg IVTA, or 4 mg IVTA. Retreatment was allowed every 4 months for each group unless the treatment was successful, futile, or contraindicated. There was no significant difference in vision or the reduction of macular edema measured by OCT at the end of 12 months between each group. Respectively, 29%, 26%, and 27% of eyes in the laser, 1 mg IVTA, and 4 mg IVTA groups gained a visual acuity score of ≥15 ETDRS letters. Subgroup analysis of pseudophakic eyes also failed to demonstrate a significant difference in vision. Median (interquartile range) center point thicknesses measured by OCT were 228 (163–364) µm, 354 (183–486) µm, and 274 (180–481) µm in the laser, 1 mg, and 4 mg IVTA groups, respectively. Three-year results from 128 patients suggested that the laser group maintained a significantly greater average increase in vision (12.9 letters) compared with the two IVTA groups (4.4 letters, 1 mg and 8.0 letters, 4 mg). Significant side-effects from IVTA included cataract formation and elevation of intraocular pressure requiring treatment. Both side-effects were dose-dependent.

As a result of this study, IVTA is not recommended as first-line therapy for macular edema in BRVO. However it can be considered in patients where macular grid laser or other therapies are ineffective, as the treatment was found to be relatively safe, especially in pseudophakic eyes.

GENEVA (dexamethasone implant) study

The Global Evaluation of implantable dexamethasone in retinal Vein occlusion with macular edema (GENEVA) study evaluated a sustained-release, biodegradable, dexamethasone intravitreal implant (Ozurdex, Allergan, Irvine, CA) for the treatment of macular edema in central retinal vein occlusion (CRVO) and BRVO patients.⁴⁵ Ozurdex is a biodegradable copolymer of poly (d,l-lactide-co-glycolide) acid (PLGA) containing micronized dexamethasone. It is injected intravitreally through a pars plana route using a 23-gauge custom injector, and it gradually releases the total dose of dexamethasone over several months via Krebs cycle breakdown of the PLGA into lactic and glycolic acid, and finally into water and carbon dioxide. In this multicenter, randomized controlled study, an increase in best-corrected visual acuity (BCVA) of ≥15 ETDRS letters was achieved in 30% of the Ozurdex 0.7 mg group (n = 291), 26% of the 0.35 mg group (n = 260), and 13% of the sham group (n = 279) 60 days after injection (peak response) in patients with BRVO (P < 0.001 for each group versus sham). A statistically significant difference between both Ozurdex groups and sham was seen up to 90 days after injection. At 90 days after injection, there was a significant improvement (P < 0.001) in central retinal thickness measured by OCT in both Ozurdex groups, compared with the sham group. The mean ± sd decrease in central retinal thickness at 90 days was 208 ± 201 µm, 177 ± 197 µm, and 85 ± 173 µm in the 0.7 mg, 0.35 mg, and sham groups, respectively. The OCT results are from pooled data including both BRVO and CRVO patients. The only complications that were significantly greater in the Ozurdex groups compared with sham were elevated intraocular pressure and anterior-chamber cell. Most eyes with elevated intraocular pressures were successfully managed with topical therapy, but five eyes required a procedure to lower the pressure adequately. In the 6 months of this study, there was no difference in the rate of cataract formation, and no endophthalmitis cases were reported. A long-term study of repeated treatments is currently underway and will help determine the safety and optimal interval for retreatment.

A major difference between the GENEVA study and other recent BRVO studies is the absence of a macular grid laser group, or rescue laser treatment for the sham group. The GENEVA study showed that the dexamethasone implant is an alternative treatment to macular grid laser in the appropriate patient population (i.e., no glaucoma, pseudophakic) and is approved by the Food and Drug Administration (FDA) for this indication.

Anti-VEGF treatment

Macular edema results from increased vascular permeability as a response to retinal nonperfusion. In patients with BRVO, retinal ischemia leads to the secretion of VEGF, which leads to increased vascular permeability, vasodilation, migration of endothelial cells, and neovascularization.^40,^41,⁴⁶ Increased vascular permeability and perhaps vasodilation lead to retinal edema. Thus, inhibition of VEGF is an attractive treatment for macular edema from BRVO. There are several anti-VEGF agents currently being investigated for use in treatment of RVOs. We will discuss the use of ranibizumab (Lucentis), bevacizumab (Avastin), pegaptanib (Macugen), and aflibercept (Eylea). Bevacizumab is a full-length, humanized monoclonal antibody that binds all VEGF-A isoforms and is FDA-approved for colorectal cancer, but is used off-label in the eye. Ranibizumab is an affinity-purified, humanized monoclonal antibody fragment (Fab) that binds all VEGF-A isoforms. Pegaptanib is an aptamer targeted against only the VEGF 165 isoform, and no other isoforms. Aflibercept is a fusion protein composed of key binding domains from VEGF receptors 1 and 2 fused to the Fc portion of human immunoglobulin G. Aflibercept binds with high affinity to all VEGF-A isoforms and placental growth factor. At this time, only ranibizumab is FDA-approved for the treatment of RVO.

BRAVO (ranibizumab) study

The Branch Retinal Vein Occlusion (BRAVO) study was a prospective, multicenter, randomized controlled study to evaluate the efficacy and safety of ranibizumab in the treatment of macular edema from BRVO.⁴⁷ Patients were randomized into three groups: (1) sham injection (n = 132); (2) 0.3 mg ranibizumab (n = 134); and (3) 0.5 mg ranibizumab (n = 131). In the first 6 months, injections were given monthly. A 28-day screening period excluded patients with spontaneous and rapid improvement in vision of >10 ETDRS letters. At month 3, a patient was eligible for rescue laser if a gain of <5 ETDRS letters, or improvement of <50 µm in central subfield thickness was observed compared with the visit 3 months prior. If rescue laser was not applied at month 3, the same criteria were used to determine eligibility for rescue laser at each subsequent monthly visit. At 6 months, both ranibizumab groups gained +16.6 and +18.3 ETDRS letters (0.3 mg and 0.5 mg groups, respectively) compared with a gain of +7.3 letters in the control group (P < 0.0001 for each group versus sham). The percentage of patients who improved greater than 15 ETDRS letters was 55.2% and 61.1% (0.3 mg and 0.5 mg groups, respectively) compared with 28.8% in the control (P < 0.0001 for each group versus sham). Concurrent with the improvement in visual acuity, the mean decrease in OCT central retinal thickness was –337.3 µm and –345.2 µm (0.3 mg and 0.5 mg groups, respectively) compared with –157.7 µm in the control (P < 0.0001 for each group versus sham). During the first 6 months, 54.5% of the control group required rescue laser therapy compared with 18.7% in the 0.3 mg and 19.8% in the 0.5 mg ranibizumab groups.

After the first 6 months, all three groups were allowed to receive “as needed” (PRN) intravitreal ranibizumab at monthly intervals if they had vision ≤ 20/40 or mean central foveal thickness ≥250 µm. Despite receiving only PRN treatments, patients in both ranibizumab groups maintained their vision gain at 12 months (unpublished results). Although the control group showed a benefit from the PRN treatment regimen, the final vision gained at 12 months was not equivalent in all three groups (unpublished results). Not surprisingly, the mean change in central foveal thickness was maintained in both ranibizumab groups at 12 months while there was a significant improvement observed in the control group after PRN treatment was initiated (unpublished results). At the end of 12 months, the incidence of adverse events in all groups was similar. One patient in the 0.5 mg ranibizumab group suffered from endophthalmitis, which is a known complication of intravitreal injections.

The BRAVO study showed that ranibizumab is superior to traditional laser treatment for macular edema from BRVO with little risk of adverse events. The current recommendation is therefore to treat patients diagnosed with macular edema from BRVO with monthly 0.5 mg ranibizumab. If treatment fails after 3 months (<5 ETDRS letter gain, or improvement of <50 µm in central subfield thickness), then traditional grid macular laser should be performed. The BRAVO study showed that PRN treatment did not adversely affect the visual outcome after five scheduled monthly injections. However, the timing of when to switch to PRN treatment was not evaluated in the BRAVO study and thus the decision to switch to PRN dosing should be based on factors such as improvement in visual acuity, residual macular edema on OCT imaging, success of prior injections, and expectations of the patient.

Other anti-VEGF inhibitors

Bevacizumab

There are numerous case series and small prospective studies evaluating the efficacy of intravitreal bevacizumab in treating macular edema from BRVO.⁴⁸^–⁵⁵ All of these studies show that bevacizumab is effective at improving visual acuity and decreasing macular edema, as measured by OCT. One of the larger retrospective studies compared two different doses of bevacizumab, 1.25 and 2.5 mg.⁵⁰ Forty-five patients with an average 35.2 weeks of follow-up completed the study. There were no functional or anatomical differences between the two dosages, and both required similar numbers of injections. At 6 months the 1.25 mg group improved by an average +5.1 lines compared with +4.8 lines in the 2.5 mg group. The central macular thickness decreased by –184 µm in the 1.25 mg group and –145 µm in the 2.5 mg group. A second large retrospective review with 1 year of follow-up showed similar results with a mean improvement of +13 letters at 6 months and +15 letters at 1 year with a decrease in central retinal thickness of –161 µm at 6 months and –205 µm at 1 year.⁵³

Although there are no studies directly comparing the efficacy of bevacizumab with ranibizumab or macular grid laser, the results reported in the BRAVO study are roughly equivalent to those reported in multiple studies with bevacizumab. It is reasonable to extrapolate that bevacizumab may be effective in treating macular edema from BRVO and may be a viable alternative to macular grid laser, corticosteroids, and other anti-VEGF agents.

Pegaptanib

A phase II study of pegaptanib for the treatment of BRVO showed significant improvements in vision and central macular thickness.⁵⁶ However, due to limited enrollment with the increasing use of bevacizumab and ranibizumab, larger trials with pegaptanib have not been conducted. For the most part, with the availability of bevacizumab and ranibizumab, pegaptanib is not used as a primary treatment for macular edema in BRVO patients.

Aflibercept

Currently, there are no trials evaluating intravitreal aflibercept specifically for macular edema from BRVO. However, two multicenter, randomized controlled phase III trials (COPERNICUS and GALILEO) are evaluating its use in patients with macular edema secondary to CRVO. In the COPERNICUS study, patients with macular edema from CRVO were randomized to receive 2 mg aflibercept (n = 114) or sham (n = 74). Only patients with vision between 20/40 and 20/320 with central retinal thickness of ≥250 µm were included. Patients were retreated every 4 weeks with aflibercept. Preliminary results of the COPERNICUS study ⁵⁷ showed that, after 24 weeks, 56.1% of the treatment group gained ≥15 ETDRS letters compared with 12.3% of the sham group (P < 0.0001). There was an average of +17.3 ETDRS letters gained in the treatment group compared with –4.0 letters lost in the sham (P < 0.001). There was a reduction of –457.8 µm in central retinal thickness in the treatment group compared with –144.8 µm in the sham group (P < 0.001). A parallel phase III clinical trial (GALILEO) is currently underway with preliminary results that are similar to the COPERNICUS study.

Given the results of ranibizumab in the BRAVO study, and the results of ranibizumab in the phase III study for CRVO (CRUISE), it is likely that the results of aflibercept in COPERNICUS and GALILEO can be extrapolated to BRVO. Currently, aflibercept is not FDA-approved for either indication.

Experimental treatments

FAVOR (iluvien) study

A sustained-release, non-erodable, intravitreal implant of fluocinolone acetonide (Iluvien) is currently under investigation for use in patients with macular edema due to BRVO or CRVO of >3 months’ duration in a phase II trial. The implant is introduced through a 25-gauge needle using a special injector during an in-office procedure, and releases 0.2 µg/day of the steroid. The implant is designed to sustain release for 36 months.

Surgical management

Vitrectomy with or without sheathotomy

The majority of the venous lesions in BRVO occur downstream from the arteriovenous crossing site. In a retrospective review of color photographs and FAs of patients with BRVO, Kumar and associates ¹⁹ identified venous narrowing at the crossing site, and in the majority of cases, evidence of downstream hemodynamic changes on angiogram, including venous-phase leakage, abnormal flow, and presumed thrombi. The authors also suggested that removal of the compressive factor by sectioning the adventitial sheath (sheathotomy) may be an effective treatment for BRVO.

In the first report of sheathotomy for BRVO, Osterloh and Charles ⁵⁸ reported significant visual improvement in the one case presented (20/200 to 20/25+ over 8 months). In the second report, Opremcak and Bruce⁵⁹ reported equal or improved visual acuity in 12 of 15 patients (80%). Ten of those patients (67%) had improved postoperative visual acuities, with an average gain of four lines of vision. Three patients had a decline in visual acuity, with an average of two lines of vision lost. All patients had marked resolution of the intraretinal hemorrhage and edema. Visual symptoms from BRVO ranged from 1 to 12 months, with an average of 3.3 months. In 2 patients, intraoperative retinal vascular bleeding was controlled with intraocular diathermy. No patient in the series developed worsening edema, ischemia, retinal neovascularization, or secondary vitreous hemorrhage. Mester and Dillinger reported 43 cases of BRVO treated with sheathotomy with similar results. In 16 of the cases, removal of the internal limiting membrane in the area of the arteriovenous crossing was also performed.⁶⁰ In contrast, Cahill and colleagues reported 27 cases of BRVO treated with vitrectomy and sheathotomy without a statistically significant improvement in postoperative median visual acuity.⁶¹

Other authors have experienced difficulty in separating the artery from the vein at the crossing site. Han and colleagues reported 20 cases of vitrectomy and attempted sheathotomy. While the visual outcome results were similar to those reported by Opremcak and Bruce, in 19 of the 20 cases, the authors were unable to separate the artery from the vein.⁶² No randomized, controlled study evaluating the benefit of sheathotomy has been published.⁶³

There is evidence that vitreomacular attachment itself may contribute to the development of macular edema in BRVO.⁶⁴ Saika and coworkers reported reduction in macular edema and restoration of normal foveal contour in 10 of 19 eyes after vitrectomy, posterior hyaloid separation, and intraocular gas tamponade.⁶⁵ Possible explanations for the clinical improvements in these studies include removal of vitreous traction, increased oxygenation of the macula, and tamponade of the macula by intraocular gas.

Due to the risk of intraoperative complications and the availability of less invasive alternatives, vitrectomy with or without sheathotomy has limited clinical use as a first-line therapy.