Epidemiology

The true incidence of CRAOs is unknown. The estimated incidence of CRAO is reported to be roughly 1 in 10 000 cases at tertiary referral centers.^2,3 The incidence may be even lower for the general population, at approximately 8.5 cases per 100 000.³ Similar to other vascular disorders, this condition is largely seen in older adults but cases in children and young adults have also been reported.^4,5 The average age at presentation is in the early sixties, with greater than 90% of patients presenting at over 40 years of age. Men are affected more frequently than women. No predilection for one eye over the other has been reported; however, 1–2% of cases may manifest bilateral involvement.⁶

Clinical features

Typically, patients with acute CRAO present with monocular, painless, severe loss of vision which occurs acutely, possibly over the span of a few seconds. In some cases, premonitory episodes of amaurosis fugax may be reported. Amaurosis fugax represents transient acute retinal ischemia and typically suggests an embolic source of occlusion; however, it has also been reported in association with giant cell arteritis.⁷ Presence of amaurosis fugax also has a higher correlation with stroke compared with retinal emboli alone.⁸ The risk of CRAO after amaurosis fugax is estimated to be only 1% per year.⁹ Although CRAO rarely presents simultaneously in both eyes, it may occur sequentially.¹⁰

Visual acuity at the time of initial presentation ranges from counting fingers to light perception in 74–90% of eyes.^6,11 Central visual acuity may be near normal in patients who have a transient CRAO or a cilioretinal artery providing sufficient vascular supply to the fovea. Connolly et al. reported a trend toward better visual acuities in patients with monocular CRAO secondary to giant cell arteritis as compared with those with CRAO due to other, largely embolic causes.¹² The presence of an embolus is usually associated with poorer vision. The absence of light perception is rare; therefore, in such cases, concomitant choroidal circulation deficit (e.g., due to ophthalmic artery occlusion) or optic nerve involvement should be considered.⁶ Visual acuity tends only to improve within the first week of onset with minimal chance for appreciable improvement subsequently.¹¹ Visual recovery after treatment has been shown to correlate with presenting visual acuity and the duration of visual impairment.¹³ Although visual acuity may spontaneously improve in up to 22% of patients with nonarteritic CRAO,¹¹ less than 10% of patients report a meaningful recovery of vision.¹⁴

Typically an afferent pupillary defect develops within seconds following obstruction of the central retinal artery regardless of macular sparing.⁴ Intraocular pressure is often normal at presentation but may become elevated in the setting of rubeosis iridis.

In a majority of cases of acute CRAO, the anterior-segment exam is normal initially. If rubeosis iridis is present acutely, ocular ischemia secondary to the presence of a concomitant carotid artery obstruction should be considered. The incidence of rubeosis in CRAO is 16.6–18.8%.^15–17 As compared to central retinal vein occlusions (CRVO), iris rubeosis tends to occur earlier after CRAO – at a mean of 4–5 weeks after onset compared to 5 months after onset in CRVO. Not surprisingly, rubeosis is more common in more severe and complete obstructions with extensive nonperfusion. The risk of developing iris neovascularization is greater for obstruction lasting greater than 1 week versus those lasting only a few days after onset.^15–17 In the case of carotid obstruction, rubeosis iridis can cause elevated intraocular pressure to the point of exceeding the perfusion pressure in the central retinal artery and lead to an obstruction via this mechanism. Laser panretinal photocoagulation causes successful regression in approximately 65% of cases.¹⁸

In his 1891 report, Nettleship described in detail the ophthalmoscopic appearance of CRAO. “The classic dense, white haze of the central region in the retina with a well-marked clear patch at the yellow-spot was very well shown (Fig. 51.1); there were no hemorrhages; the arteries and veins were of about normal size, but no pulsation could be produced in any of them by pressure with the finger upon the globe.”¹⁹ Hayreh and Zimmerman investigated fundus changes in CRAOs in a large retrospective review in 2007 of 248 eyes of 240 patients. In the acute phase, his group noted a cherry-red spot (90%), posterior pole retinal opacity or whitening (58%), box-carring of retinal arteries and veins (19% and 20% respectively), retinal arterial attenuation (32%), optic disc edema (22%), and optic nerve pallor (39%). The retinal findings were predominantly located in the posterior pole with a normal-appearing periphery.²⁰

Fig. 51.1 Acute central retinal artery obstruction. Superficial retinal whitening or opacification is noted in the posterior pole with evidence of a central cherry-red spot. The opacification is most prominent in the peripheral fovea.

(Courtesy of Dr Steven Yeh, Emory Eye Center, Atlanta.)

Typically, retinal whitening in the posterior pole and a cherry-red spot are the earliest characteristic changes in CRAO. Both of these findings are clinical signs for which ophthalmologists have a high degree of clinical agreement.²¹ The retinal whitening corresponds to ischemic damage to the inner half of the retina and is due to opacification of the retinal nerve fiber and ganglion cell layer as a result of cessation of axoplasmic transport caused by the acute ischemic insult. The opacification is visible ophthalmoscopically where the ganglion cell layer is more than one cell thick, i.e., the macula, except in the foveal region, where a cherry-red spot is seen. The outer nuclear and plexiform layers and photoreceptors remain intact, as demonstrated in histologic studies.^22,23 The size of the cherry-red spot is variable and is dependent on the width of the foveola. The cherry-red spot is actually normal-appearing retina and is observed in high contrast against the surrounding opacified retina because the thin retina in this location is nourished by the underlying choroidal circulation and as a result does not become hypoxic or opacified, permitting continued visualization of the underlying retinal pigment epithelium (RPE) and choroid.²⁴ Similarly, the retinal periphery in CRAO cases appears normal because the retina is also thinner with a single layer of ganglion cells, such that the nutrition of the inner retinal layers can be maintained by the choroidal circulation alone. In experimental primate models of transiently induced CRAOs, retinal opacification was observed as early as 7 minutes after complete occlusion. The probability of a cherry-red spot still being present decreases with increasing duration from the onset of CRAO: 88% after 1 week, 59% after 2 weeks, 47% after 3 weeks, and 19% after 4 weeks after onset. Typically, the retinal opacification resolves over a period of 4–6 weeks, although at least some retinal whitening is noted in 17% of patients with complete, nontransient CRAO after 1 month.²⁰ Pathologically, this evolution corresponds to a resolution of initial acute ischemia-induced intracellular edema with subsequent loss of neuronal cells and the development of an acellular scar of the inner retinal cell layers.²⁵

A patent cilioretinal artery supplying some or all of the papillomacular bundle is seen in approximately one-third of cases. In cases of cilioretinal sparing, retinal whitening will be clearly demarcated around the area of preserved macula perfused by the cilioretinal circulation (Fig. 51.2A). In these cases, the visual acuity will be dictated by the location and extent of the area of the papillomacular bundle perfused by the patent cilioretinal artery.^11,26–28 Sparing of the fovea may be associated with excellent visual acuity, albeit with a significant visual field deficit corresponding to the topography of the occlusion.

Fig. 51.2 Acute central retinal artery occlusion with cilioretinal artery sparing. (A) Retinal opacification is clearly demarcated around the preserved macula perfused by the patent cilioretinal artery. The perfused area includes the fovea; therefore the patient’s presenting visual acuity was good. (B) At 19 seconds, the cilioretinal artery is perfused with retrograde flow into the retinal veins. (C) At 47 seconds, incomplete appearance of fluorescein dye is noted in the retinal arteries. (D) At 9 minutes the arterial circulation is still not completely filled, demonstrating a severely delayed arteriovenous transit time.

(Courtesy of Dr John Payne, Emory Eye Center, Atlanta.)

The appearance of the retinal vasculature can be quite variable soon after the onset of CRAO, and the presence of a normal-appearing vasculature should not exclude the diagnosis. In Nettleship’s 19th-century description of CRAO, he noted the appearance of the retinal vasculature to have a stagnated arterial blood column without attenuation.¹⁹ In the natural history study by Hayreh and Zimmerman,²⁰ only 15% of acute CRAO had normal-appearing retinal arteries. Box-carring or segmentation of the blood column of both the arteries and veins can occur secondary to separation of blood serum from erythrocytes in a stacked or rouleaux formation.

Retinal emboli are visible in 20–40% of eyes with CRAO.^4,21 Retinal emboli are the most common cause of nonarteritic CRAO and BRAO.²⁹ The most common variant is a yellow, refractile cholesterol embolus (Hollenhorst plaque) (Fig. 51.3). According to a study by Arruga and Sanders, retinal emboli consist of cholesterol in 74% of cases, calcified material in 15.5%, and platelet and fibrin in 15.5%.³⁰ These cholesterol emboli typically originate from the carotid arteries in the setting of atherosclerotic disease, but can also arise from the aortic arch, ophthalmic artery, or proximal central retinal artery. Cholesterol emboli are often small, do not completely obstruct retinal arterial blood flow, and are frequently found at bifurcation sites (Fig. 51.3). Emboli can often be asymptomatic and migration with disappearance of retinal emboli is common.²⁰ Calcific emboli are less common than cholesterol emboli but are typically larger and cause more severe or complete obstruction. They most commonly originate from the cardiac valves.^24,31

Fig. 51.3 Cholesterol embolus at an arterial bifurcation along the superotemporal arcade.

The optic nerve is acutely edematous in nearly all cases of arteritic CRAO as a result of the associated anterior ischemic optic neuropathy that is typically observed in these patients. In the acute phase of nonarteritic CRAO, the disc may be normal, hyperemic, edematous, and, rarely, pale. Acute-phase optic nerve pallor is due to ischemic opacification of the surface nerve fiber layer since this layer of the optic nerve is supplied by retinal circulation.³² Neovascularization of the disc in acute CRAO is rare but has been reported, typically in association with a chronically hypoxic retina (e.g., concomitant diabetic retinopathy, ischemic CRVO, or ocular ischemia).^19,20

The most frequent findings in the chronic stage of eyes with CRAO are optic atrophy (91%), retinal arterial attenuation (58%), cilioretinal collaterals (18%), macular RPE changes (11%), and cotton-wool spots (3%) (Fig. 51.4). In experimental studies in rhesus monkeys, the extent of optic nerve and nerve fiber layer damage has been correlated with the duration of CRAO.³³ Chronic-phase optic nerve pallor is due to optic atrophy and nerve fiber loss (Fig. 51.5). In arteritic CRAO, the associated anterior ischemic optic neuropathy also contributes to the development of pallor.²⁰ In chronic CRAO, neovascularization of the optic disc rarely occurs, presumably because nonviable tissue is less likely to elaborate angiogenic factors compared with chronically ischemic but viable retinal tissue seen in cases of diabetic retinopathy or retinal vein occlusion.³⁴ Incidence of neovascularization of the disc in one retrospective study was only 1.8%.³⁵ Concomitant rubeosis iridis may also occur in the setting of chronic CRAO. Retinal arterial attenuation is more common in the chronic phase of disease than in the acute phase.²⁰ Months after an acute CRAO, cilioretinal collaterals may develop as a result of a compensatory enlargement of capillary anastomoses between retinal capillaries on the surface of the disc and ciliary capillaries in deeper parts of the optic nerve head. The probability of developing cilioretinal disc collaterals was 4% at 1 month and 18% at 3 months from onset of CRAO in one study.²⁰ Macular RPE changes are seen with CRAO but are much less common than in ophthalmic artery obstruction, where the choroidal circulation is also involved.

Fig. 51.4 Chronic central retinal artery occlusion. Optic nerve pallor from atrophy and severe arterial attenuation is seen in this patient with a remote central artery occlusion.

Fig. 51.5 Histopathologic findings in central retinal artery obstruction. In the chronic stage of the disease, the inner retinal layers show significant atrophy. Hematoxylin and eosin ×25.

(Courtesy of Dr Hans Grossniklaus, Emory Eye Center, Atlanta.)

Ancillary studies

Fluorescein angiography in CRAO almost always initially shows some variable residual retinal circulation with delayed and sluggish filling of the retinal vasculature. Complete absence of retinal filling is rare. Appearance of dye in the central retinal artery is typically delayed by 5–20 seconds. However, the delay in retinal arterial branches is even more substantial. The fluorescein dye lines the arterial walls in a pattern similar to the laminar flow filling of normal retinal veins. In cases with visible intra-arterial emboli, the arteriovenous transit time can be even further delayed (Fig. 51.2B–D).³⁶ The severity of obstruction and retinal ischemia correlates with less favorable initial visual acuities. Staining of the optic disc can be variable; however, staining of the retinal vessels is rare. Areas of delayed choroidal perfusion, characterized by a delay in filling of more than 10 seconds compared to adjacent normal choroid, may be seen in about 11% of eyes with acute CRAO.⁶ Leakage of fluorescein dye at the level of the RPE is generally not seen with CRAO unless the choroidal circulation is involved.³⁷ Typically, the retinal circulation is re-established after an acute CRAO but the inner retinal tissue has generally already infarcted by then. Therefore, although the fluorescein angiogram may return to relatively normal appearance, the vision loss, optic nerve atrophy, and arterial narrowing persist.²⁴ For patients with a normalized fluorescein angiogram who do not go on to develop optic atrophy, the diagnosis of a true CRAO should be called into question, though it is possible that some individuals may have reperfusion before the irreversible damage has occurred.

In the acute stage, optical coherence tomography (OCT) shows an irregular macular contour with increased reflectivity of the inner retina. This corresponds to intracellular edema and explains the lack of intraretinal, hyporeflective fluid spaces in cases of CRAO or BRAO. The reflectivity of the outer retinal layers and RPE is blocked by the highly reflective inner retinal layer. No retinal thickening secondary to the accumulation of serous fluid escaping from retinal capillaries into the extracellular space is seen (Fig. 51.6). OCT images of chronic CRAO show thinning and atrophy of the inner retina. OCT can be helpful in cases of chronic CRAO where the fundus may appear featureless but the OCT shows inner retinal atrophy with preservation of the outer retina.^38–40

Fig. 51.6 Spectral-domain optical coherence tomography (SD-OCT) of acute central retinal artery occlusion (CRAO). An irregular foveal contour is seen with a highly reflective, thickened inner retina. The outer retina is relatively hyporeflective because of blocking from the thickened inner retina. The inner-segment–outer-segment junction and the external limiting membrane are intact, as would be expected with acute CRAO.

Central scotoma is the most common defect observed on macular visual field testing followed by paracentral scotoma. Patients with cilioretinal sparing show a preserved central island of vision corresponding to the area perfused by the patent cilioretinal artery. Peripheral constriction is the most common visual field deficit noted in these patients.¹¹ A preserved temporal island may be seen in some patients, presumably secondary to choroid-derived perfusion of the nasal retina.²⁶ Visual field defects improve in approximately 28% of patients, remain stable in 57%, and worsen in 7%.¹¹

In a CRAO, electroretinography typically demonstrates more severe attenuation of the b-wave than the a-wave since the inner retinal layers are more affected – this produces a characteristic negative waveform with the scotopic white stimulus (Fig. 51.7). Diminution of the a-wave and b-wave may suggest outer retinal damage secondary to choroidal vascular hypoperfusion in the setting of an ophthalmic artery occlusion in addition to a CRAO.⁴¹

Fig. 51.7 Electroretinogram. The responses from a normal eye are seen on the left. The tracings on the right are from an eye with a central retinal artery occlusion (CRAO). The bottom right tracing shows the maximum response to the scotopic white stimulus. The a-wave is of normal amplitude; however, the b-wave does not reach the baseline, yielding a negative waveform, as is typically seen with a CRAO.

Autofluorescence imaging in the area supplied by the occluded retinal artery acutely shows decreased autofluorescence due to blockage of the normal autofluorescence of the RPE by the thickened inner retina. This blockage resolves over time and may evolve into a window defect with increased autofluorescence in the chronic phase as areas of significant inner retinal thinning develop.³⁹

Systemic associations

The distribution of systemic associations for CRAO varies depending on age. Overall, embolism from carotid artery atherosclerosis is the most common etiology for retinal arterial occlusion; however, carotid disease is relatively rare in patients under the age of 40 in whom cardiac embolism is the most common etiology.^4,5,42 The systemic and ocular abnormalities that have been associated with retinal arterial occlusions are summarized in Box 51.1.

The Beaver Dam Eye Study, a large population-based study in Wisconsin, found a 10-year cumulative incidence of retinal emboli of 1.5%. The incidence of retinal emboli varied with age; persons who were 65 years of age or older at baseline were 2.4 times as likely to develop a retinal embolus compared with persons 43–54 years of age at baseline. Retinal emboli were more likely to occur in men than in women. Incidence of bilateral emboli was rare; however, multiple emboli in the same eye may be seen in up to one-third of cases (Fig. 51.8). Persons with retinal emboli in Beaver Dam were 2.4 times as likely to have a diagnosis of stroke on their death certificate over an 11-year period compared with those without retinal emboli.⁴³ The large population-based studies, the Atherosclerosis Risk in Communities (North Carolina) and the Cardiovascular Health Study (Australia), both looked at a biracial population and showed that retinal arteriolar emboli were associated with hypertension, higher systolic blood pressure, carotid artery plaque, increased plasma fibrinogen levels, coronary heart disease, increased plasma lipoprotein levels, and current cigarette smoking. In multivariable models, significant independent predictors were carotid artery plaque, hypertension status, and current cigarette smoking.⁴⁴ In the heart, sources of emboli are aortic or mitral valve lesions, patent foramen ovale, left atrial tumor, and myxoma.^45–47

Fig. 51.8 Multiple cholesterol emboli are seen in the same eye (arrows) within the inferotemporal arcade.

The presence of a Hollenhorst plaque or retinal artery occlusion is associated with a low prevalence of carotid atherosclerosis requiring carotid endarterectomy. Furthermore, in contrast to amaurosis fugax, these ocular findings are not associated with a high risk for hemispheric neurological events.⁴⁸ However, patients with retinal emboli do have an associated higher mortality rate.^43,45,49,50 Pooled data from the Beaver Dam Eye Study and the Blue Mountain Eye study of two older populations suggests that retinal emboli predict a modest increase in all-cause and stroke-related mortality independent of cardiovascular risk factors.⁵¹ The prevalence of diabetes, hypertension, ischemic heart disease, cerebrovascular accidents, and smoking is significantly higher in patients with retinal arterial occlusions.^{6,42–44,48,52–54}

Evaluation

Patients with CRAO typically present to a physician several days after the acute onset; therefore, the etiologic workup is generally recommended on an outpatient basis along with a primary care physician. The only true emergency in such a circumstance would be to rule out giant cell arteritis in patients older than 50 years with a positive review of systems. Evaluation for giant cell arteritis includes compete blood count, including platelets, erythrocyte sedimentation rate, and C-reactive protein. If suspicion is high, the patient should be started on steroid therapy and scheduled for a temporal artery biopsy. Rarely, a patient with acute CRAO may present within the first few hours of visual deficit. These patients should be admitted for observation, treatment, and immediate workup as their risk is higher for cerebral infarction.¹⁴

The evaluation of embolic source often includes carotid Doppler imaging and echocardiography since the most common sources of retinal emboli are from the carotid artery or the heart. As most retinal emboli are relatively small, from the ophthalmic point of view, when evaluating the results of carotid Doppler ultrasonography, the presence or absence of plaque is more important than whether a hemodynamically significant stenosis is present; the latter is more important in determining the need for carotid endarterectomy. Carotid Doppler also has its limitations, including the lack of imaging of the thoracic and intracranial portion of the carotid artery and poor resolution for detection of microemboli.⁵⁵

Although a cardiogenic cause is less common, identifying this is important as chronic anticoagulation may be indicated to prevent more serious adverse events.

A cardiac evaluation is especially important in young patients and those with calcific emboli. In patients with acute retinal arterial obstruction at low cardioembolic risk, transthoracic echocardiography (TTE) results in anticoagulation or cardiac surgery in only 1.5% of patients.⁵⁶ Transesophageal echocardiography (TEE) has a higher yield than the transthoracic approach in the cardiovascular evaluation of patients with retinal artery occlusion. In patients without a cardiac history and a normal TTE examination (including normal size and function of cardiac chambers and normal valves with no calcification, along with the absence of atrial fibrillation), the yield of subsequent TEE for identifying intracardiac pathology is low. However, in patients in whom clinical and TTE findings are suggestive of possible intracardiac or aortic thrombus, TEE should be considered as an adjunct to the systemic workup.⁵⁷ A computed tomography (CT) angiogram or magnetic resonance imaging angiogram should be considered in special cases such as suspected carotid or aortic dissection. Since the cardiac morbidity and mortality are significant in patients with retinal artery occlusion, a baseline electrocardiogram is recommended.^10,14

A hypercoagulability evaluation should be considered for patients less than 50 years of age with a suggestive history (e.g., prior thrombosis, miscarriage, or family history) or unknown embolic source. Workup includes blood tests for factor V Leiden mutation; protein C, protein S, and antithrombin III deficiencies; homocysteine levels; sickle-cell disease; and antiphospholipid antibodies. Other tests for monoclonal gammopathy, cancer, infection, and disseminated intravascular coagulation may be ordered depending on the clinical circumstance.¹⁴

Treatment

CRAO continues to be a challenging disease entity to treat. Typically treatment is either conservative or invasive. Most reports of treatment outcome are anecdotal as a result of low incidence.⁵⁸ Spontaneous resolution can occur in up to 22% of patients¹¹ and has been reported to occur up to 3 days after initial onset.⁵⁹ However, less than 10% of patients report meaningful visual recovery.^60,61 Rarely do patients have complete spontaneous recovery.⁶²

Based on experimental models of CRAO in elderly and atherosclerotic rhesus monkeys, the retina suffers no damage up to 97 minutes after an acute CRAO but after 4 hours the retina suffers massive irreversible damage. Therefore, no treatment instituted after about 4 hours from onset can logically restore any vision in the setting of complete obstruction. Additionally, this model showed the longer the ischemia, the longer the time to recovery.¹¹ However, unlike the animal model, humans rarely have complete obstruction. As a result, treatment for CRAO has been recommended within 24 hours of symptom onset. Given the relative rarity of CRAO and variability in time to presentation, therapeutic trials have been limited in sample size, thus reducing the power to detect small treatment benefits. Thus, for a new therapy to have a major impact on the management of this disease, it will likely need to double or triple the success rate of current conventional therapy yet still maintain a low risk for morbidity and mortality.

Current conventional therapy consists of dislodging emboli, reducing intraocular pressure and increasing retinal blood flow, vasodilating the ocular blood supply, improving retinal circulation, decreasing retinal edema, maintaining retinal oxygenation until spontaneous reperfusion, and acting on the thrombus.¹⁴ None of these treatments have proven effective and their use is largely based on anecdotal reports and small case series. In a small study of 11 patients, Rumelt et al. found that a systematic regimen involving multiple, sequential treatment steps had better visual outcomes than arbitrary treatment with conservative measures. The protocol included ocular massage, sublingual isosorbide dinitrate, intravenous acetazolamide, intravenous mannitol or oral glycerol, anterior-chamber paracentesis, intravenous methylprednisolone, streptokinase, and retrobulbar tolazine. Treatment with one or two conservative modalities was usually insufficient.^3,63,64 In a Cochrane Controlled Trials Register comparing any treatment for CRAO with another treatment, Fraser and Siriwardena found no randomized controlled trials meeting their inclusion criteria. They concluded that insufficient data existed to decide if any beneficial treatments existed for CRAO.⁶⁵

Ocular massage is performed using either a Goldmann contact lens or digital massage to apply ocular pressure with an in-and-out movement to dislodge a possibly obstructing embolus. Repeated massage with 10–15 seconds of pressure followed by a sudden release is recommended.⁶⁶ This maneuver can produce retinal arterial vasodilation, thereby improving retinal blood flow.^67,68 A mixture of 95% oxygen and 5% carbon dioxide (carbogen) can be provided to induce vasodilation and improve oxygenation, but efficacy has not been proven.⁶⁹ Hyperbaric oxygen provides oxygen at levels of atmospheric pressure. The purpose of hyperbaric oxygen is to preserve the retina in an oxygenated state until recanalization and reperfusion occur, typically at 72 hours. The hyperbaric oxygen increases the arterial oxygen pressure and thereby increases nitric oxide synthesis, leading to vasodilation. Case reports of successful treatment have been published,^70–72 but currently treatment for ocular diseases is an off-label use of hyperbaric oxygen.⁷³ Anterior-chamber paracentesis causes a sudden decrease in intraocular pressure, possibly causing the arterial perfusion pressure behind the obstruction to force an obstructing embolus downstream. This treatment has shown some success in retrospective analyses;^74,75 however, a study by Atebara et al. compared anterior-chamber paracentesis and carbogen therapy with no intervention and found no significant difference in visual outcome.⁶⁹ Vasodilating medications that have been utilized to increase retinal blood flow in retinal arterial occlusion include pentoxifylline, nitroglycerin, and isosorbide dinitrate.^3,76,77 Isovolemic hemodilution is used to increase oxygen supply to the retinal tissue. It includes the replacement of 500 mL of blood with the same volume of hydroxy-ethyl starch. In vascular disease, the limiting factor in tissue oxygenation is blood flow and not oxygen-carrying capacity. In this situation lowering blood viscosity by decreasing the hematocrit and plasma viscosity will improve tissue oxygen levels.^65,78

Various surgical techniques have also been explored for treatment of CRAO. Neodymium : yttrium aluminum garnet (Nd-YAG) laser arteriotomy in patients with CRAO has been reported to result in extrusion of an embolus, reopening of the central retinal artery, and return of vision. A fundus contact lens is used with the laser in single-burst mode. The laser is focused slightly deep to the vessel wall at the site of the embolus to avoid photodisruption and opacification of the overlying nerve fiber layer. In patients with small emboli, the laser is focused on the center of the plaque. In patients with elongated emboli, the laser is focused slightly on the distal or downstream end to reduce the chance of hemorrhage. Pulses are delivered directly to the emboli, beginning with the lowest power setting and then with increasing energy until either (1) achieving photofragmentation of the embolus within the arteriole without creating an opening in the vessel wall and without vitreous hemorrhage or (2) creating visible removal of the embolus from within arteriole into the vitreous cavity, typically associated with a limited vitreous hemorrhage. Digital pressure can be applied to the globe to help stop bleeding, if it occurs.⁷⁹

Corticosteroids should only be used when arteritic CRAO from giant cell arteritis is suspected.¹⁴ Anticoagulants should be reserved for secondary prevention of cerebral and ocular infarction in those rare patients who have an underlying systemic disease such as atrial fibrillation, acute internal carotid artery dissection, or a hypercoagulable condition.¹⁴

With the common use of thrombolytics for acute cerebrovascular accidents, their potential application in the setting of acute CRAO has been considered. Intravenously or intra-arterially administered thrombolytics currently in use include streptokinase, urokinase, and tissue plasminogen activator (t-PA). Intravenous administration is relatively easy and workup prior to treatment is minimal, including blood tests and a brain CT; however intravenous thrombolytics do increase the risk of systemic hemorrhage. Because of the increased systemic risks, most thrombolysis strategies currently in use are intra-arterial.¹⁴ A meta-analysis of all published literature pertaining to intra-arterial thrombolysis for treatment of acute CRAO done in 2000 found that this treatment did not improve visual acuity beyond the disease’s natural course. No prospective randomized studies were available to be included at that time.⁸⁰ In 2010 the European Assessment Group for Lysis in the Eye (EAGLE) study group published the results of the first prospective, randomized clinical trial evaluating the effect of intra-arterial t-PA compared with conservative treatment. Of note, no true placebo arm was included in the study. At 1 month, the mean best-corrected visual acuity improved significantly in both groups but no significant difference was noted between groups. Clinically significant visual improvement (0.3 logMAR) was noted in 60.0% of patients in the conventional therapy group and 57.1% of patients in the thrombolysis group. The trial was stopped early because of the apparent similarity in efficacy between groups and the higher rate of adverse events, namely cerebral hemorrhage, in the intra-arterial t-PA group.⁷⁸ The EAGLE study highlights the importance of careful randomized controlled trials, as the rate of visual improvement in the conventional therapy group was higher than one might expect from prior retrospective studies.