Acute central retinal artery occlusion (CRAO) is a visually debilitating condition resulting in ischaemia of the inner retina with subsequent necrosis.1 Accounting for approximately 1/10 000 visits to ophthalmologists each year, it is a recognised ophthalmic emergency with a poor visual prognosis.2^–4

Managing CRAO remains a therapeutic dilemma. A number of treatment modalities for CRAO have been proposed and practised, all with limited success.5 Conventional conservative treatment options include anterior chamber paracentesis, acetazolamide, aspirin, ocular massage and carbogen inhalation (95% oxygen, 5% carbon dioxide).2^–8 However, a recent Cochrane meta-analysis by Mueller et al9 concluded that these regimens do not improve visual outcome or alter the natural course of CRAO. More recently, a case report of translumenal Nd:YAG embolysis has shown positive treatment effect.10 Because of the dismal prognosis and low rate of adverse affects from treatment, most physicians entertain some combination of therapies.

In recent years, a number of encouraging reports have emerged in the literature regarding the use of catheter-assisted super-selective intra-arterial thrombolysis (IAT) for the treatment of acute CRAO.11^–18 This recent attention is motivated in part by the documented success of thrombolysis in treating acute myocardial infarction and ischaemic stroke.15 Based on similar principles, it has been assumed that thrombolytic interventions could theoretically restore retinal blood flow in CRAO by dissolving occlusive thromboemboli, thereby restoring perfusion to the inner retina and abating ischaemic damage.14

In the early 1990s, Schmidt and Schumacher pioneered IAT as a treatment modality for CRAO.16^–18 Using a femoral artery approach, fibrinolysis was performed using a microcatheter directed superselectively to the proximal portion of the ophthalmic artery. A range of outcomes were reported, from no effect to complete restoration of visual acuity. These studies sparked interest in the use of thrombolysis for the treatment of CRAO as a method for improving the visual prognosis of such patients. Throughout the 1990s, several more reports emerged in the literature purporting favourable results of IAT for CRAO. These studies were summarised in the meta-analysis and review by Beatty and Au Eong which found somewhat favourable results from non-randomised case series and retrospective cohort studies.14 19

In the interim, several more studies investigating effectiveness of IAT for CRAO have emerged, but there have been no attempts at rigorously reviewing the recent data on this topic. The present study systematically reviews the evidence supporting the use of IAT for the treatment of acute CRAO.

METHODS

Ovid MEDLINE literature searches were conducted for the following medical subject headings: central retinal artery occlusion; thrombolysis; intra-arterial thrombolysis; fibrinolysis; and randomised controlled trials. A Cochrane Database search was also performed for randomised control trials, systematic reviews and meta-analyses using the same medical subject headings. Only studies with non-arteritic CRAO were included. In our final analysis, case reports or studies with fewer than five patients and animal studies were excluded.

RESULTS

One meta-analysis14 published in 2000 was found with no systematic reviews published since then. All published studies retrieved were non-randomised, and most were retrospective in nature (see table 1). Of all studies retrieved, eight fit our inclusion criteria and were analysed systematically.

In one of the original investigations of significant patient size, Schumacher et al reported an apparent benefit of IAT in a heterogeneous cohort of 23 patients with angiographically confirmed non-arteritic CRAO.18 The time delay between the onset of symptoms and the initiation of treatment varied from just under 4 h to 2.5 days. Following treatment, 26% of patients (6/23) demonstrated marked improvement or total recovery of visual acuity (20/50 or better). The mean time between the onset of symptoms and treatment for this group was 6 h. Another 48% of patients (11/23) demonstrated partial recovery (LP up to 20/60), evident within a few hours to 2 days. The mean time to treatment for this group was 15 h. Six patients showed nominal to no change in vision; the mean time to treatment in this latter group was 20 h.

In 1998, Weber et al published results of a non-randomised retrospective cohort analysis comparing outcomes in non-arteritic CRAO treated with IAT (17 patients) with those treated with a variety of conservative methods (15 patients).20 The mean time delay between visual loss and treatment in the thrombolysis group was 4.2 h (range 1–6 h); all control patients were also treated within 6 h of onset of symptoms. Visual outcomes for those treated with IAT were superior to the controls, with 29% of patients (5/17) demonstrating marked visual improvement, with a final visual acuity of 20/30 or better, compared with 0% in the control group (p = 0.01). Also, 67% (10/15) in the control group showed no improvement in vision compared with 35% (6/17) in the IAT group (p = 0.01). Unfortunately, fluorescein angiography was not performed, and so patients were not classified into CRAO subtypes.

Weill et al published an investigation of IAT after acute CRAO in a non-comparative case series involving seven patients.21 In this study, IAT was administered between 9 and 20 h after the onset of symptoms (mean of 12.5 h). Visual acuity before treatment was limited to light perception in four patients, hand motions in two patients, and 20/50 in one patient. Post-treatment visual acuity was 20/20 in 43% (3/7) patients, 20/40 or better in 29% (2/7), with no recovery seen in 29% (2/7) of patients.

A retrospective non-comparative case series performed by Richard et al in 1999 investigated IAT with rTPA in 53 patients (46 with CRAO and seven with branch retinal arterial occlusion).22 Fluorescein angiography was also performed. The average time to treatment was 14 h (SD 10) h (range 3–50 h). Following IAT, visual acuity improved in 66% (35/53) patients. Compared with baseline, 47.2% (28/53) showed an improvement of greater than two lines (p<0.001). Visual acuity did not improve in 33% of patients (18/53).

In 2002, Schmidt and colleagues reported a non-randomised retrospective cohort study of 178 patients comparing conservative treatment methods (116 patients) to IAT with urokinase (62 patients) after angiographically confirmed non-arteritic CRAO.11 The mean time to treatment was similar between both groups (24.2 (40.4) h for the control group and 10.8 (9.5) h for the IAT group, p = 0.5). Patients treated with IAT demonstrated a significantly greater chance of visual acuity improvement when compared with those treated conservatively (58% vs 29%, p = 0.0022). This improvement was more marked in patients treated within 6 h of the onset of visual loss and only occurred in those with subtotal or incomplete CRAO’s.

In a 2003 non-comparative case series, Butz et al reported outcomes from 22 patients after IAT using both urokinase and rtPA in acute non-arteritic CRAO.23 The mean latency period was 7.6 (1.8) h. Following treatment with IAT, one patient achieved a visual acuity of 20/20, 36% (8/22) showed some improvement (hand movements to 20/32) while 59% of patients (13/22) demonstrated no improvement in visual acuity. This study lacked angiographic confirmation.

A large, retrospective case-control study by Arnold et al in 2005 compared IAT using urokinase (37 patients) to conventional treatment (19 patients) in non-angiographically confirmed non-arteritic CRAO.12 The median time to treatment was 240 min for both groups (221 (53) min for the control group and 240 (62) min for the IAT group). In the IAT group, 22% (8/37) of patients regained a visual acuity of >0.6 logMAR (20/80 or better) versus 0% (0/19) in the conventionally treated group (p = 0.04). Visual improvement was more likely in younger patients treated with IAT (p = 0.012).

Finally, a retrospective study carried out in Calgary24 showed that 100% of patients (6/6) treated for CRAO by IAT showed an improvement in visual acuity (three demonstrated an improvement of two Snellen lines or more, and three showed an improvement of one line).

Of the eight studies included in the present analysis, the overall complication rate was 4.5% (10/220), including: local haemorrhage (1), transient ischaemic attack (5), hypertensive crises (1), intracerebral haemorrhage (1) and stroke (2).

DISCUSSION

Of the eight studies reviewed, seven suggested some benefit to IAT compared with conventional conservative therapies (see table 1). Of those purporting a benefit, 93% of patients (147/158) demonstrated some level of visual improvement, with 13% achieving 20/20 or better, 25% achieving 20/40 or better and 41% achieving 20/200 or better. These results echo findings of the review performed by Beatty and Au Eong in 2000.14 One hundred patients from five retrospective cohort and case series studies were reviewed, and they found that after IAT, 14% achieved 20/20 or better, 27% achieved 20/40 or better, and 37% achieved 20/200 or better.

Although it may appear that IAT import a beneficial effect on visual outcomes after acute non-arteritic CRAO, the results must be interpreted with caution. To date, all published studies are non-randomised, retrospective case series and cohort studies. Many published studies are of questionable quality, with important deficiencies. Indeed, in the present study, a meta-analysis was not performed, as the total number of patients was insufficient for meaningful results, randomised trials were not available, and the studies reviewed were not of similar design.

Perhaps the most important error in previously published work has been the confounding of IAT treatment success with the natural history of visual improvement after CRAO. Indeed, some eyes with CRAO undergo spontaneous visual recovery, with most improvement occurring within 1 week postevent.4 The amount of improvement is largely dependent on the type of CRAO, with eyes suffering from transient non-arteritic CRAO showing the most improvement. Hayreh and Zimmerman showed that in eyes with counting fingers vision or worse, visual acuity improves spontaneously in 82% of eyes with transient CRAO, compared with 67% for eyes with cilioretinal artery sparing and 22% for eyes with total CRAO. Actual visual outcomes include 6.5% achieving 20/20 or better, 16% achieving 20/40 or better and 29% achieving 20/200 or better. Improvements in visual fields can occur in about 1/4 to 1/3 of patients.4 Therefore, non-comparative trials demonstrating positive treatment effects from IAT given up to 24 h after the onset of symptoms (eg, Richard et al22) may be describing the natural outcome with visual improvements, from learnt behaviours such as eccentric viewing or natural improvements from incomplete occlusions. Nevertheless, the combined data from the studies analysed in the present review suggest that outcomes after tPA may be superior to natural history outcomes.

The actual therapeutic time window in which IAT is beneficial is unknown. Estimates from studies in primates suggest that after 240 min of ischaemia, the retina suffers massive, irreparable damage.1 Animal studies have also shown that if retinal blood flow is restored between 2 and 4 h of occlusion, the improvement in retinal function was not evident until up to 24 h after restoration of circulation. In humans, fluorescein angiography studies reveal that most cases of CRAO exhibit partial residual blood flow, theoretically prolonging retinal tolerance time. In the present study, we selected studies which instituted IAT within an 8 h time window of symptoms, as clinical observation suggests that IAT is may be beneficial within that time frame.

Previous work has also been limited by the failure to group patients according to CRAO subtypes (eg, complete, incomplete, cilio-retinal sparing) through fluorescein angiography. This is particularly important, as patients presenting with CRAO represent a clinically heterogeneous cohort. CRAO may be arteritic or non-arteritic; it may be complete, incomplete or transient; and the actual occlusion can be of different constituents. For example, only 15% of retinal emboli are composed of platelet fibrin, with the remainder being cholesterol (75%) or caclific (10%) emboli, although the latter two can potentially develop secondary platelet thrombi. Up to 25% of patients have a macular cilio-retinal artery, which influences final visual acuity. The effect of these variables on outcome has not been investigated systematically in good-quality studies, and further confounds the interpretation of treatment effect in existing studies.

Another important consideration in the therapeutic use of IAT for CRAO is that of local and systemic complications. Potential adverse events include local site haemorrhage, intraretinal and intracerebral bleeding, transient ischaemic attacks and stroke. These complications are felt to occur more commonly in elderly patients (>80 years of age) and those with significant cardiovascular risk factors.25

Finally, instituting an IAT programme in itself has many practical considerations. In order to offer timely administration, 24 h access to fluorescein angiography and interventional radiology suites is required. Assessments by the emergency department, Ophthalmology and Interventional Neuroradiology must be streamlined. When more data regarding efficacy of IAT for CRAO become available, a cost–benefit analysis and cost–utility analysis are warranted to determine the feasibility of instituting such a programme.

In summary, the literature to date remains replete with non-randomised retrospective cohort studies. Without evidence from well-designed randomised controlled trials, it is difficult to develop guidelines explaining the role of IAT in the management of CRAO. The first large, multi-centred randomised controlled clinical trial comparing conservative medical treatment to IAT in patients with CRAO is under way, having been commenced in 2002.26 Termed the EAGLE study (European Assessment Group for Lysis in the Eye), it is hoped that this trial, despite being criticised for some design flaws,27 will provide more conclusive data regarding the efficacy of IAT for CRAO.26

CONCLUSION

CRAO is a visually debilitating condition with poor treatment options and generally poor visual outcomes. Although no current treatment has been definitively proven to improve outcomes in acute non-arteritic CRAO, our review of the available data suggests that IAT may produce superior visual outcomes compared with conventional treatments when instituted in certain situations. At the moment, there is insufficient evidence to support the routine use of IAT to treat CRAO. Data from randomised clinical trials are required before evidence-based guidelines can be developed.